words	labels
Molecular	O
Dissection	O
of	O
Xyloglucan	O
Recognition	O
in	O
a	O
Prominent	O
Human	O
Gut	O
Symbiont	O
	O
Polysaccharide	O
utilization	O
loci	O
(	O
PUL	O
)	O
within	O
the	O
genomes	O
of	O
resident	O
human	O
gut	O
Bacteroidetes	O
are	O
central	O
to	O
the	O
metabolism	O
of	O
the	O
otherwise	O
indigestible	O
complex	O
carbohydrates	O
known	O
as	O
“	O
dietary	O
fiber	O
.”	O
However	O
,	O
functional	O
characterization	O
of	O
PUL	O
lags	O
significantly	O
behind	O
sequencing	O
efforts	O
,	O
which	O
limits	O
physiological	O
understanding	O
of	O
the	O
human	O
-	O
bacterial	O
symbiosis	O
.	O
	O
In	O
particular	O
,	O
the	O
molecular	O
basis	O
of	O
complex	O
polysaccharide	O
recognition	O
,	O
an	O
essential	O
prerequisite	O
to	O
hydrolysis	O
by	O
cell	O
surface	O
glycosidases	O
and	O
subsequent	O
metabolism	O
,	O
is	O
generally	O
poorly	O
understood	O
.	O
	O
Here	O
,	O
we	O
present	O
the	O
biochemical	O
,	O
structural	O
,	O
and	O
reverse	O
genetic	O
characterization	O
of	O
two	O
unique	O
cell	O
surface	O
glycan	O
-	O
binding	O
proteins	O
(	O
SGBPs	O
)	O
encoded	O
by	O
a	O
xyloglucan	O
utilization	O
locus	O
(	O
XyGUL	O
)	O
from	O
Bacteroides	O
ovatus	O
,	O
which	O
are	O
integral	O
to	O
growth	O
on	O
this	O
key	O
dietary	O
vegetable	O
polysaccharide	O
.	O
	O
Biochemical	O
analysis	O
reveals	O
that	O
these	O
outer	O
membrane	O
-	O
anchored	O
proteins	O
are	O
in	O
fact	O
exquisitely	O
specific	O
for	O
the	O
highly	O
branched	O
xyloglucan	O
(	O
XyG	O
)	O
polysaccharide	O
.	O
	O
The	O
crystal	O
structure	O
of	O
SGBP	O
-	O
A	O
,	O
a	O
SusD	O
homolog	O
,	O
with	O
a	O
bound	O
XyG	O
tetradecasaccharide	O
reveals	O
an	O
extended	O
carbohydrate	O
-	O
binding	O
platform	O
that	O
primarily	O
relies	O
on	O
recognition	O
of	O
the	O
β	O
-	O
glucan	O
backbone	O
.	O
	O
The	O
unique	O
,	O
tetra	O
-	O
modular	O
structure	O
of	O
SGBP	O
-	O
B	O
is	O
comprised	O
of	O
tandem	O
Ig	O
-	O
like	O
folds	O
,	O
with	O
XyG	O
binding	O
mediated	O
at	O
the	O
distal	O
C	O
-	O
terminal	O
domain	O
.	O
	O
Despite	O
displaying	O
similar	O
affinities	O
for	O
XyG	O
,	O
reverse	O
-	O
genetic	O
analysis	O
reveals	O
that	O
SGBP	O
-	O
B	O
is	O
only	O
required	O
for	O
the	O
efficient	O
capture	O
of	O
smaller	O
oligosaccharides	O
,	O
whereas	O
the	O
presence	O
of	O
SGBP	O
-	O
A	O
is	O
more	O
critical	O
than	O
its	O
carbohydrate	O
-	O
binding	O
ability	O
for	O
growth	O
on	O
XyG	O
.	O
Together	O
,	O
these	O
data	O
demonstrate	O
that	O
SGBP	O
-	O
A	O
and	O
SGBP	O
-	O
B	O
play	O
complementary	O
,	O
specialized	O
roles	O
in	O
carbohydrate	O
capture	O
by	O
B	O
.	O
ovatus	O
and	O
elaborate	O
a	O
model	O
of	O
how	O
vegetable	O
xyloglucans	O
are	O
accessed	O
by	O
the	O
Bacteroidetes	O
.	O
	O
Our	O
combined	O
analysis	O
illuminates	O
new	O
fundamental	O
aspects	O
of	O
complex	O
polysaccharide	O
recognition	O
,	O
cleavage	O
,	O
and	O
import	O
at	O
the	O
Bacteroidetes	O
cell	O
surface	O
that	O
may	O
facilitate	O
the	O
development	O
of	O
prebiotics	O
to	O
target	O
this	O
phylum	O
of	O
gut	O
bacteria	O
.	O
	O
This	O
microbial	O
community	O
is	O
largely	O
bacterial	O
,	O
with	O
the	O
Bacteroidetes	O
,	O
Firmicutes	O
,	O
and	O
Actinobacteria	O
comprising	O
the	O
dominant	O
phyla	O
.	O
	O
However	O
,	O
there	O
is	O
a	O
paucity	O
of	O
data	O
regarding	O
how	O
the	O
vast	O
array	O
of	O
complex	O
carbohydrate	O
structures	O
are	O
selectively	O
recognized	O
and	O
imported	O
by	O
members	O
of	O
the	O
microbiota	O
,	O
a	O
critical	O
process	O
that	O
enables	O
these	O
organisms	O
to	O
thrive	O
in	O
the	O
competitive	O
gut	O
environment	O
.	O
	O
The	O
human	O
gut	O
bacteria	O
Bacteroidetes	O
share	O
a	O
profound	O
capacity	O
for	O
dietary	O
glycan	O
degradation	O
,	O
with	O
many	O
species	O
containing	O
>	O
250	O
predicted	O
carbohydrate	O
-	O
active	O
enzymes	O
(	O
CAZymes	O
),	O
compared	O
to	O
50	O
to	O
100	O
within	O
many	O
Firmicutes	O
and	O
only	O
17	O
in	O
the	O
human	O
genome	O
devoted	O
toward	O
carbohydrate	O
utilization	O
.	O
	O
A	O
remarkable	O
feature	O
of	O
the	O
Bacteroidetes	O
is	O
the	O
packaging	O
of	O
genes	O
for	O
carbohydrate	O
catabolism	O
into	O
discrete	O
polysaccharide	O
utilization	O
loci	O
(	O
PUL	O
),	O
which	O
are	O
transcriptionally	O
regulated	O
by	O
specific	O
substrate	O
signatures	O
.	O
	O
The	O
Sus	O
includes	O
a	O
lipid	O
-	O
anchored	O
,	O
outer	O
membrane	O
endo	O
-	O
amylase	O
,	O
SusG	O
;	O
a	O
TonB	O
-	O
dependent	O
transporter	O
(	O
TBDT	O
),	O
SusC	O
,	O
which	O
imports	O
oligosaccharides	O
with	O
the	O
help	O
of	O
an	O
associated	O
starch	O
-	O
binding	O
protein	O
,	O
SusD	O
;	O
two	O
additional	O
carbohydrate	O
-	O
binding	O
lipoproteins	O
,	O
SusE	O
and	O
SusF	O
;	O
and	O
two	O
periplasmic	O
exo	O
-	O
glucosidases	O
,	O
SusA	O
and	O
SusB	O
,	O
which	O
generate	O
glucose	O
for	O
transport	O
into	O
the	O
cytoplasm	O
.	O
	O
The	O
importance	O
of	O
PUL	O
as	O
a	O
successful	O
evolutionary	O
strategy	O
is	O
underscored	O
by	O
the	O
observation	O
that	O
Bacteroidetes	O
such	O
as	O
B	O
.	O
thetaiotaomicron	O
and	O
Bacteroides	O
ovatus	O
devote	O
~	O
18	O
%	O
of	O
their	O
genomes	O
to	O
these	O
systems	O
.	O
	O
Moving	O
beyond	O
seminal	O
genomic	O
and	O
transcriptomic	O
analyses	O
,	O
the	O
current	O
state	O
-	O
of	O
-	O
the	O
-	O
art	O
PUL	O
characterization	O
involves	O
combined	O
reverse	O
-	O
genetic	O
,	O
biochemical	O
,	O
and	O
structural	O
studies	O
to	O
illuminate	O
the	O
molecular	O
details	O
of	O
PUL	O
function	O
.	O
	O
Cleavage	O
sites	O
for	O
BoXyGUL	O
glycosidases	O
(	O
GHs	O
)	O
are	O
indicated	O
for	O
solanaceous	O
xyloglucan	O
.	O
(	O
B	O
)	O
BtSus	O
and	O
BoXyGUL	O
.	O
(	O
C	O
)	O
Localization	O
of	O
BoXyGUL	O
-	O
encoded	O
proteins	O
in	O
cellular	O
membranes	O
and	O
concerted	O
modes	O
of	O
action	O
in	O
the	O
degradation	O
of	O
xyloglucans	O
to	O
monosaccharides	O
.	O
	O
XyG	O
variants	O
(	O
Fig	O
.	O
1A	O
)	O
constitute	O
up	O
to	O
25	O
%	O
of	O
the	O
dry	O
weight	O
of	O
common	O
vegetables	O
.	O
	O
Analogous	O
to	O
the	O
Sus	O
locus	O
,	O
the	O
xyloglucan	O
utilization	O
locus	O
(	O
XyGUL	O
)	O
encodes	O
a	O
cohort	O
of	O
carbohydrate	O
-	O
binding	O
,	O
-	O
hydrolyzing	O
,	O
and	O
-	O
importing	O
proteins	O
(	O
Fig	O
.	O
1B	O
and	O
C	O
).	O
	O
The	O
number	O
of	O
glycoside	O
hydrolases	O
(	O
GHs	O
)	O
encoded	O
by	O
the	O
XyGUL	O
is	O
,	O
however	O
,	O
more	O
expansive	O
than	O
that	O
by	O
the	O
Sus	O
locus	O
(	O
Fig	O
.	O
1B	O
),	O
which	O
reflects	O
the	O
greater	O
complexity	O
of	O
glycosidic	O
linkages	O
found	O
in	O
XyG	O
vis	O
-	O
à	O
-	O
vis	O
starch	O
.	O
	O
In	O
the	O
archetypal	O
starch	O
utilization	O
system	O
of	O
B	O
.	O
thetaiotaomicron	O
,	O
starch	O
binding	O
to	O
the	O
cell	O
surface	O
is	O
mediated	O
at	O
eight	O
distinct	O
starch	O
-	O
binding	O
sites	O
distributed	O
among	O
four	O
surface	O
glycan	O
-	O
binding	O
proteins	O
(	O
SGBPs	O
):	O
two	O
within	O
the	O
amylase	O
SusG	O
,	O
one	O
within	O
SusD	O
,	O
two	O
within	O
SusE	O
,	O
and	O
three	O
within	O
SusF	O
.	O
The	O
functional	O
redundancy	O
of	O
many	O
of	O
these	O
sites	O
is	O
high	O
:	O
whereas	O
SusD	O
is	O
essential	O
for	O
growth	O
on	O
starch	O
,	O
combined	O
mutations	O
of	O
the	O
SusE	O
,	O
SusF	O
,	O
and	O
SusG	O
binding	O
sites	O
are	O
required	O
to	O
impair	O
growth	O
on	O
the	O
polysaccharide	O
.	O
	O
Bacteroidetes	O
PUL	O
ubiquitously	O
encode	O
homologs	O
of	O
SusC	O
and	O
SusD	O
,	O
as	O
well	O
as	O
proteins	O
whose	O
genes	O
are	O
immediately	O
downstream	O
of	O
susD	O
,	O
akin	O
to	O
susE	O
/	O
F	O
,	O
and	O
these	O
are	O
typically	O
annotated	O
as	O
“	O
putative	O
lipoproteins	O
”.	O
	O
The	O
genes	O
coding	O
for	O
these	O
proteins	O
,	O
sometimes	O
referred	O
to	O
as	O
“	O
susE	O
/	O
F	O
positioned	O
,”	O
display	O
products	O
with	O
a	O
wide	O
variation	O
in	O
amino	O
acid	O
sequence	O
and	O
which	O
have	O
little	O
or	O
no	O
homology	O
to	O
other	O
PUL	O
-	O
encoded	O
proteins	O
or	O
known	O
carbohydrate	O
-	O
binding	O
proteins	O
.	O
	O
We	O
describe	O
here	O
the	O
detailed	O
functional	O
and	O
structural	O
characterization	O
of	O
the	O
noncatalytic	O
SGBPs	O
encoded	O
by	O
Bacova_02651	O
and	O
Bacova_02650	O
of	O
the	O
XyGUL	O
,	O
here	O
referred	O
to	O
as	O
SGBP	O
-	O
A	O
and	O
SGBP	O
-	O
B	O
,	O
to	O
elucidate	O
their	O
molecular	O
roles	O
in	O
carbohydrate	O
acquisition	O
in	O
vivo	O
.	O
	O
Combined	O
biochemical	O
,	O
structural	O
,	O
and	O
reverse	O
-	O
genetic	O
approaches	O
clearly	O
illuminate	O
the	O
distinct	O
,	O
yet	O
complementary	O
,	O
functions	O
that	O
these	O
two	O
proteins	O
play	O
in	O
XyG	O
recognition	O
as	O
it	O
impacts	O
the	O
physiology	O
of	O
B	O
.	O
ovatus	O
.	O
	O
SGBP	O
-	O
A	O
and	O
SGBP	O
-	O
B	O
are	O
cell	O
-	O
surface	O
-	O
localized	O
,	O
xyloglucan	O
-	O
specific	O
binding	O
proteins	O
.	O
	O
SGBP	O
-	O
A	O
,	O
encoded	O
by	O
the	O
XyGUL	O
locus	O
tag	O
Bacova_02651	O
(	O
Fig	O
.	O
1B	O
),	O
shares	O
26	O
%	O
amino	O
acid	O
sequence	O
identity	O
(	O
40	O
%	O
similarity	O
)	O
with	O
its	O
homolog	O
,	O
B	O
.	O
thetaiotaomicron	O
SusD	O
,	O
and	O
similar	O
homology	O
with	O
the	O
SusD	O
-	O
like	O
proteins	O
encoded	O
within	O
syntenic	O
XyGUL	O
identified	O
in	O
our	O
earlier	O
work	O
.	O
	O
In	O
contrast	O
,	O
SGBP	O
-	O
B	O
,	O
encoded	O
by	O
locus	O
tag	O
Bacova_02650	O
,	O
displays	O
little	O
sequence	O
similarity	O
to	O
the	O
products	O
of	O
similarly	O
positioned	O
genes	O
in	O
syntenic	O
XyGUL	O
nor	O
to	O
any	O
other	O
gene	O
product	O
among	O
the	O
diversity	O
of	O
Bacteroidetes	O
PUL	O
.	O
	O
Whereas	O
sequence	O
similarity	O
among	O
SusC	O
/	O
SusD	O
homolog	O
pairs	O
often	O
serves	O
as	O
a	O
hallmark	O
for	O
PUL	O
identification	O
,	O
the	O
sequence	O
similarities	O
of	O
downstream	O
genes	O
encoding	O
SGBPs	O
are	O
generally	O
too	O
low	O
to	O
allow	O
reliable	O
bioinformatic	O
classification	O
of	O
their	O
products	O
into	O
protein	O
families	O
,	O
let	O
alone	O
prediction	O
of	O
function	O
.	O
	O
Hence	O
,	O
there	O
is	O
a	O
critical	O
need	O
for	O
the	O
elucidation	O
of	O
detailed	O
structure	O
-	O
function	O
relationships	O
among	O
PUL	O
SGBPs	O
,	O
in	O
light	O
of	O
the	O
manifold	O
glycan	O
structures	O
in	O
nature	O
.	O
	O
Immunofluorescence	O
of	O
formaldehyde	O
-	O
fixed	O
,	O
nonpermeabilized	O
cells	O
grown	O
in	O
minimal	O
medium	O
with	O
XyG	O
as	O
the	O
sole	O
carbon	O
source	O
to	O
induce	O
XyGUL	O
expression	O
,	O
reveals	O
that	O
both	O
SGBP	O
-	O
A	O
and	O
SGBP	O
-	O
B	O
are	O
presented	O
on	O
the	O
cell	O
surface	O
by	O
N	O
-	O
terminal	O
lipidation	O
,	O
as	O
predicted	O
by	O
signal	O
peptide	O
analysis	O
with	O
SignalP	O
(	O
Fig	O
.	O
2	O
).	O
	O
SGBP	O
-	O
A	O
and	O
SGBP	O
-	O
B	O
visualized	O
by	O
immunofluorescence	O
.	O
	O
Formalin	O
-	O
fixed	O
,	O
nonpermeabilized	O
B	O
.	O
ovatus	O
cells	O
were	O
grown	O
in	O
minimal	O
medium	O
plus	O
XyG	O
,	O
probed	O
with	O
custom	O
rabbit	O
antibodies	O
to	O
SGBP	O
-	O
A	O
or	O
SGBP	O
-	O
B	O
,	O
and	O
then	O
stained	O
with	O
Alexa	O
Fluor	O
488	O
goat	O
anti	O
-	O
rabbit	O
IgG	O
.	O
(	O
A	O
)	O
Overlay	O
of	O
bright	O
-	O
field	O
and	O
FITC	O
images	O
of	O
B	O
.	O
ovatus	O
cells	O
labeled	O
with	O
anti	O
-	O
SGBP	O
-	O
A	O
.	O
(	O
B	O
)	O
Overlay	O
of	O
bright	O
-	O
field	O
and	O
FITC	O
images	O
of	O
B	O
.	O
ovatus	O
cells	O
labeled	O
with	O
anti	O
-	O
SGBP	O
-	O
B	O
.	O
(	O
C	O
)	O
Bright	O
-	O
field	O
image	O
of	O
ΔSGBP	B-mutant
-	I-mutant
B	I-mutant
cells	O
labeled	O
with	O
anti	O
-	O
SGBP	O
-	O
B	O
antibodies	O
.	O
	O
(	O
D	O
)	O
FITC	O
images	O
of	O
ΔSGBP	B-mutant
-	I-mutant
B	I-mutant
cells	O
labeled	O
with	O
anti	O
-	O
SGBP	O
-	O
B	O
antibodies	O
.	O
	O
Cells	O
lacking	O
SGBP	O
-	O
A	O
(	O
ΔSGBP	B-mutant
-	I-mutant
A	I-mutant
)	O
do	O
not	O
grow	O
on	O
XyG	O
and	O
therefore	O
could	O
not	O
be	O
tested	O
in	O
parallel	O
.	O
	O
Additional	O
affinity	O
PAGE	O
analysis	O
(	O
Fig	O
.	O
3	O
)	O
demonstrates	O
that	O
SGBP	O
-	O
A	O
also	O
has	O
moderate	O
affinity	O
for	O
the	O
artificial	O
soluble	O
cellulose	O
derivative	O
hydroxyethyl	O
cellulose	O
[	O
HEC	O
;	O
a	O
β	O
(	O
1	O
→	O
4	O
)-	O
glucan	O
]	O
and	O
limited	O
affinity	O
for	O
mixed	O
-	O
linkage	O
β	O
(	O
1	O
→	O
3	O
)/	O
β	O
(	O
1	O
→	O
4	O
)-	O
glucan	O
(	O
MLG	O
)	O
and	O
glucomannan	O
(	O
GM	O
;	O
mixed	O
glucosyl	O
and	O
mannosyl	O
backbone	O
),	O
which	O
together	O
indicate	O
general	O
binding	O
to	O
polysaccharide	O
backbone	O
residues	O
and	O
major	O
contributions	O
from	O
side	O
-	O
chain	O
recognition	O
.	O
	O
In	O
contrast	O
,	O
SGBP	O
-	O
B	O
bound	O
to	O
HEC	O
more	O
weakly	O
than	O
SGBP	O
-	O
A	O
and	O
did	O
not	O
bind	O
to	O
MLG	O
or	O
GM	O
.	O
	O
Neither	O
SGBP	O
recognized	O
galactomannan	O
(	O
GGM	O
),	O
starch	O
,	O
carboxymethylcellulose	O
,	O
or	O
mucin	O
(	O
see	O
Fig	O
.	O
S1	O
in	O
the	O
supplemental	O
material	O
).	O
	O
Notably	O
,	O
the	O
absence	O
of	O
carbohydrate	O
-	O
binding	O
modules	O
in	O
the	O
GHs	O
encoded	O
by	O
the	O
XyGUL	O
implies	O
that	O
noncatalytic	O
recognition	O
of	O
xyloglucan	O
is	O
mediated	O
entirely	O
by	O
SGBP	O
-	O
A	O
and	O
-	O
B	O
.	O
	O
SGBP	O
-	O
A	O
and	O
SGBP	O
-	O
B	O
preferentially	O
bind	O
xyloglucan	O
.	O
	O
Affinity	O
electrophoresis	O
(	O
10	O
%	O
acrylamide	O
)	O
of	O
SGBP	O
-	O
A	O
and	O
SGBP	O
-	O
B	O
with	O
BSA	O
as	O
a	O
control	O
protein	O
.	O
	O
All	O
samples	O
were	O
loaded	O
on	O
the	O
same	O
gel	O
next	O
to	O
the	O
BSA	O
controls	O
;	O
thin	O
black	O
lines	O
indicate	O
where	O
intervening	O
lanes	O
were	O
removed	O
from	O
the	O
final	O
image	O
for	O
both	O
space	O
and	O
clarity	O
.	O
	O
The	O
percentage	O
of	O
polysaccharide	O
incorporated	O
into	O
each	O
native	O
gel	O
is	O
displayed	O
.	O
	O
The	O
vanguard	O
endo	O
-	O
xyloglucanase	O
of	O
the	O
XyGUL	O
,	O
BoGH5	O
,	O
preferentially	O
cleaves	O
the	O
polysaccharide	O
at	O
unbranched	O
glucosyl	O
residues	O
to	O
generate	O
xylogluco	O
-	O
oligosaccharides	O
(	O
XyGOs	O
)	O
comprising	O
a	O
Glc4	O
backbone	O
with	O
variable	O
side	O
-	O
chain	O
galactosylation	O
(	O
XyGO1	O
)	O
(	O
Fig	O
.	O
1A	O
;	O
n	O
=	O
1	O
)	O
as	O
the	O
limit	O
of	O
digestion	O
products	O
in	O
vitro	O
;	O
controlled	O
digestion	O
and	O
fractionation	O
by	O
size	O
exclusion	O
chromatography	O
allow	O
the	O
production	O
of	O
higher	O
-	O
order	O
oligosaccharides	O
(	O
e	O
.	O
g	O
.,	O
XyGO2	O
)	O
(	O
Fig	O
.	O
1A	O
;	O
n	O
=	O
2	O
).	O
	O
ITC	O
demonstrates	O
that	O
SGBP	O
-	O
A	O
binds	O
to	O
XyG	O
polysaccharide	O
and	O
XyGO2	O
(	O
based	O
on	O
a	O
Glc8	O
backbone	O
)	O
with	O
essentially	O
equal	O
affinities	O
,	O
while	O
no	O
binding	O
of	O
XyGO1	O
(	O
Glc4	O
backbone	O
)	O
was	O
detectable	O
(	O
Table	O
1	O
;	O
see	O
Fig	O
.	O
S2	O
and	O
S3	O
in	O
the	O
supplemental	O
material	O
).	O
	O
Together	O
,	O
these	O
data	O
clearly	O
suggest	O
that	O
polysaccharide	O
binding	O
of	O
both	O
SGBPs	O
is	O
fulfilled	O
by	O
a	O
dimer	O
of	O
the	O
minimal	O
repeat	O
,	O
corresponding	O
to	O
XyGO2	O
(	O
cf	O
.	O
	O
The	O
observation	O
by	O
affinity	O
PAGE	O
that	O
these	O
proteins	O
specifically	O
recognize	O
XyG	O
is	O
further	O
substantiated	O
by	O
their	O
lack	O
of	O
binding	O
for	O
the	O
undecorated	O
oligosaccharide	O
cellotetraose	O
(	O
Table	O
1	O
;	O
see	O
Fig	O
.	O
S3	O
).	O
	O
Furthermore	O
,	O
SGBP	O
-	O
A	O
binds	O
cellohexaose	O
with	O
~	O
770	O
-	O
fold	O
weaker	O
affinity	O
than	O
XyG	O
,	O
while	O
SGBP	O
-	O
B	O
displays	O
no	O
detectable	O
binding	O
to	O
this	O
linear	O
hexasaccharide	O
.	O
	O
To	O
provide	O
molecular	O
-	O
level	O
insight	O
into	O
how	O
the	O
XyGUL	O
SGBPs	O
equip	O
B	O
.	O
ovatus	O
to	O
specifically	O
harvest	O
XyG	O
from	O
the	O
gut	O
environment	O
,	O
we	O
performed	O
X	O
-	O
ray	O
crystallography	O
analysis	O
of	O
both	O
SGBP	O
-	O
A	O
and	O
SGPB	O
-	O
B	O
in	O
oligosaccharide	O
-	O
complex	O
forms	O
.	O
	O
Summary	O
of	O
thermodynamic	O
parameters	O
for	O
wild	O
-	O
type	O
SGBP	O
-	O
A	O
and	O
SGBP	O
-	O
B	O
obtained	O
by	O
isothermal	O
titration	O
calorimetry	O
at	O
25	O
°	O
Ca	O
	O
Carbohydrate	O
Ka	O
(	O
M	O
−	O
1	O
)	O
ΔG	O
(	O
kcal	O
⋅	O
mol	O
−	O
1	O
)	O
ΔH	O
(	O
kcal	O
⋅	O
mol	O
−	O
1	O
)	O
TΔS	O
(	O
kcal	O
⋅	O
mol	O
−	O
1	O
)	O
SGBP	O
-	O
A	O
SGBP	O
-	O
B	O
SGBP	O
-	O
A	O
SGBP	O
-	O
B	O
SGBP	O
-	O
A	O
SGBP	O
-	O
B	O
SGBP	O
-	O
A	O
SGBP	O
-	O
B	O
XyGb	O
(	O
4	O
.	O
4	O
±	O
0	O
.	O
1	O
)	O
×	O
105	O
(	O
5	O
.	O
7	O
±	O
0	O
.	O
2	O
)	O
×	O
104	O
−	O
7	O
.	O
7	O
−	O
6	O
.	O
5	O
−	O
14	O
±	O
3	O
−	O
14	O
±	O
2	O
−	O
6	O
.	O
5	O
−	O
7	O
.	O
6	O
XyGO2c	O
3	O
.	O
0	O
×	O
105	O
2	O
.	O
0	O
×	O
104	O
−	O
7	O
.	O
5	O
−	O
5	O
.	O
9	O
−	O
17	O
.	O
2	O
−	O
17	O
.	O
6	O
−	O
9	O
.	O
7	O
−	O
11	O
.	O
7	O
XyGO1	O
NBd	O
(	O
2	O
.	O
4	O
±	O
0	O
.	O
1	O
)	O
×	O
103	O
NB	O
−	O
4	O
.	O
6	O
NB	O
−	O
4	O
.	O
4	O
±	O
0	O
.	O
2	O
NB	O
0	O
.	O
2	O
Cellohexaose	O
568	O
.	O
0	O
±	O
291	O
.	O
0	O
NB	O
−	O
3	O
.	O
8	O
NB	O
−	O
16	O
±	O
8	O
NB	O
−	O
12	O
.	O
7	O
NB	O
Cellotetraose	O
NB	O
NB	O
NB	O
NB	O
NB	O
NB	O
NB	O
NB	O
	O
SGBP	O
-	O
A	O
is	O
a	O
SusD	O
homolog	O
with	O
an	O
extensive	O
glycan	O
-	O
binding	O
platform	O
.	O
	O
Specifically	O
,	O
SGBP	O
-	O
A	O
overlays	O
B	O
.	O
thetaiotaomicron	O
SusD	O
(	O
BtSusD	O
)	O
with	O
a	O
root	O
mean	O
square	O
deviation	O
(	O
RMSD	O
)	O
value	O
of	O
2	O
.	O
2	O
Å	O
for	O
363	O
Cα	O
pairs	O
,	O
which	O
is	O
notable	O
given	O
the	O
26	O
%	O
amino	O
acid	O
identity	O
(	O
40	O
%	O
similarity	O
)	O
between	O
these	O
homologs	O
(	O
Fig	O
.	O
4C	O
).	O
	O
The	O
SGBP	O
-	O
A	O
:	O
XyGO2	O
complex	O
superimposes	O
closely	O
with	O
the	O
apo	O
structure	O
(	O
RMSD	O
of	O
0	O
.	O
6	O
Å	O
)	O
and	O
demonstrates	O
that	O
no	O
major	O
conformational	O
change	O
occurs	O
upon	O
substrate	O
binding	O
;	O
small	O
deviations	O
in	O
the	O
orientation	O
of	O
several	O
surface	O
loops	O
are	O
likely	O
the	O
result	O
of	O
differential	O
crystal	O
packing	O
.	O
	O
It	O
is	O
particularly	O
notable	O
that	O
although	O
the	O
location	O
of	O
the	O
ligand	O
-	O
binding	O
site	O
is	O
conserved	O
between	O
SGBP	O
-	O
A	O
and	O
SusD	O
,	O
that	O
of	O
SGBP	O
-	O
A	O
displays	O
an	O
~	O
29	O
-	O
Å	O
-	O
long	O
aromatic	O
platform	O
to	O
accommodate	O
the	O
extended	O
,	O
linear	O
XyG	O
chain	O
(	O
see	O
reference	O
for	O
a	O
review	O
of	O
XyG	O
secondary	O
structure	O
),	O
versus	O
the	O
shorter	O
,	O
~	O
18	O
-	O
Å	O
-	O
long	O
,	O
site	O
within	O
SusD	O
that	O
complements	O
the	O
helical	O
conformation	O
of	O
amylose	O
(	O
Fig	O
.	O
4C	O
and	O
D	O
).	O
	O
Molecular	O
structure	O
of	O
SGBP	O
-	O
A	O
(	O
Bacova_02651	O
).	O
(	O
A	O
)	O
Overlay	O
of	O
SGBP	O
-	O
A	O
from	O
the	O
apo	O
(	O
rainbow	O
)	O
and	O
XyGO2	O
(	O
gray	O
)	O
structures	O
.	O
	O
An	O
omit	O
map	O
(	O
2σ	O
)	O
for	O
XyGO2	O
(	O
orange	O
and	O
red	O
sticks	O
)	O
is	O
displayed	O
.	O
	O
(	O
B	O
)	O
Close	O
-	O
up	O
view	O
of	O
the	O
omit	O
map	O
as	O
in	O
panel	O
A	O
,	O
rotated	O
90	O
°	O
clockwise	O
.	O
	O
(	O
C	O
)	O
Overlay	O
of	O
the	O
Cα	O
backbones	O
of	O
SGBP	O
-	O
A	O
(	O
black	O
)	O
with	O
XyGO2	O
(	O
orange	O
and	O
red	O
spheres	O
)	O
and	O
BtSusD	O
(	O
blue	O
)	O
with	O
maltoheptaose	O
(	O
pink	O
and	O
red	O
spheres	O
),	O
highlighting	O
the	O
conservation	O
of	O
the	O
glycan	O
-	O
binding	O
site	O
location	O
.	O
	O
(	O
D	O
)	O
Close	O
-	O
up	O
of	O
the	O
SGBP	O
-	O
A	O
(	O
black	O
and	O
orange	O
)	O
and	O
SusD	O
(	O
blue	O
and	O
pink	O
)	O
glycan	O
-	O
binding	O
sites	O
.	O
	O
The	O
backbone	O
glucose	O
residues	O
are	O
numbered	O
from	O
the	O
nonreducing	O
end	O
;	O
xylose	O
residues	O
are	O
labeled	O
X1	O
and	O
X2	O
.	O
	O
Indeed	O
,	O
the	O
electron	O
density	O
for	O
the	O
ligand	O
suggests	O
some	O
disorder	O
,	O
which	O
may	O
arise	O
from	O
multiple	O
oligosaccharide	O
orientations	O
along	O
the	O
binding	O
site	O
.	O
	O
Three	O
aromatic	O
residues	O
—	O
W82	O
,	O
W283	O
,	O
W306	O
—	O
comprise	O
the	O
flat	O
platform	O
that	O
stacks	O
along	O
the	O
naturally	O
twisted	O
β	O
-	O
glucan	O
backbone	O
(	O
Fig	O
.	O
4E	O
).	O
	O
Contrasting	O
with	O
the	O
clear	O
importance	O
of	O
these	O
hydrophobic	O
interactions	O
,	O
there	O
are	O
remarkably	O
few	O
hydrogen	O
-	O
bonding	O
interactions	O
with	O
the	O
ligand	O
,	O
which	O
are	O
provided	O
by	O
R65	O
,	O
N83	O
,	O
and	O
S308	O
,	O
which	O
are	O
proximal	O
to	O
Glc5	O
and	O
Glc3	O
.	O
	O
Most	O
surprising	O
in	O
light	O
of	O
the	O
saccharide	O
-	O
binding	O
data	O
,	O
however	O
,	O
was	O
a	O
lack	O
of	O
extensive	O
recognition	O
of	O
the	O
XyG	O
side	O
chains	O
;	O
only	O
Y84	O
appeared	O
to	O
provide	O
a	O
hydrophobic	O
interface	O
for	O
a	O
xylosyl	O
residue	O
(	O
Xyl1	O
).	O
	O
Summary	O
of	O
thermodynamic	O
parameters	O
for	O
site	O
-	O
directed	O
mutants	O
of	O
SGBP	O
-	O
A	O
and	O
SGBP	O
-	O
B	O
obtained	O
by	O
ITC	O
with	O
XyG	O
at	O
25	O
°	O
Ca	O
	O
Weak	O
binding	O
represents	O
a	O
Ka	O
of	O
<	O
500	O
M	O
−	O
1	O
.	O
	O
Ka	O
fold	O
change	O
=	O
Ka	O
of	O
wild	O
-	O
type	O
protein	O
/	O
Ka	O
of	O
mutant	O
protein	O
for	O
xyloglucan	O
binding	O
.	O
	O
SGBP	O
-	O
B	O
has	O
a	O
multimodular	O
structure	O
with	O
a	O
single	O
,	O
C	O
-	O
terminal	O
glycan	O
-	O
binding	O
domain	O
.	O
	O
The	O
crystal	O
structure	O
of	O
full	O
-	O
length	O
SGBP	O
-	O
B	O
in	O
complex	O
with	O
XyGO2	O
(	O
2	O
.	O
37	O
Å	O
,	O
Rwork	O
=	O
19	O
.	O
9	O
%,	O
Rfree	O
=	O
23	O
.	O
9	O
%,	O
residues	O
34	O
to	O
489	O
)	O
(	O
Table	O
2	O
)	O
revealed	O
an	O
extended	O
structure	O
composed	O
of	O
three	O
tandem	O
immunoglobulin	O
(	O
Ig	O
)-	O
like	O
domains	O
(	O
domains	O
A	O
,	O
B	O
,	O
and	O
C	O
)	O
followed	O
at	O
the	O
C	O
terminus	O
by	O
a	O
novel	O
xyloglucan	O
-	O
binding	O
domain	O
(	O
domain	O
D	O
)	O
(	O
Fig	O
.	O
5A	O
).	O
	O
These	O
domains	O
also	O
display	O
similarity	O
to	O
the	O
C	O
-	O
terminal	O
β	O
-	O
sandwich	O
domains	O
of	O
many	O
GH13	O
enzymes	O
,	O
including	O
the	O
cyclodextrin	O
glucanotransferase	O
of	O
Geobacillus	O
stearothermophilus	O
(	O
Fig	O
.	O
5B	O
).	O
	O
Such	O
domains	O
are	O
not	O
typically	O
involved	O
in	O
carbohydrate	O
binding	O
.	O
	O
Indeed	O
,	O
visual	O
inspection	O
of	O
the	O
SGBP	O
-	O
B	O
structure	O
,	O
as	O
well	O
as	O
individual	O
production	O
of	O
the	O
A	O
and	O
B	O
domains	O
and	O
affinity	O
PAGE	O
analysis	O
(	O
see	O
Fig	O
.	O
S5	O
in	O
the	O
supplemental	O
material	O
),	O
indicates	O
that	O
these	O
domains	O
do	O
not	O
contribute	O
to	O
XyG	O
capture	O
.	O
	O
On	O
the	O
other	O
hand	O
,	O
production	O
of	O
the	O
fused	B-mutant
domains	I-mutant
C	I-mutant
and	I-mutant
D	I-mutant
in	O
tandem	O
(	O
SGBP	O
-	O
B	O
residues	O
230	O
to	O
489	O
)	O
retains	O
complete	O
binding	O
of	O
xyloglucan	O
in	O
vitro	O
,	O
with	O
the	O
observed	O
slight	O
increase	O
in	O
affinity	O
likely	O
arising	O
from	O
a	O
reduced	O
potential	O
for	O
steric	O
hindrance	O
of	O
the	O
smaller	O
protein	O
construct	O
during	O
polysaccharide	O
interactions	O
(	O
Table	O
3	O
).	O
	O
While	O
neither	O
the	O
full	O
-	O
length	O
protein	O
nor	O
domain	O
D	O
displays	O
structural	O
homology	O
to	O
known	O
XyG	O
-	O
binding	O
proteins	O
,	O
the	O
topology	O
of	O
SGBP	O
-	O
B	O
resembles	O
the	O
xylan	O
-	O
binding	O
protein	O
Bacova_04391	O
(	O
PDB	O
3ORJ	O
)	O
encoded	O
within	O
a	O
xylan	O
-	O
targeting	O
PUL	O
of	O
B	O
.	O
ovatus	O
(	O
Fig	O
.	O
5C	O
).	O
	O
The	O
structure	O
-	O
based	O
alignment	O
of	O
these	O
proteins	O
reveals	O
17	O
%	O
sequence	O
identity	O
,	O
with	O
a	O
core	O
RMSD	O
of	O
3	O
.	O
6	O
Å	O
for	O
253	O
aligned	O
residues	O
.	O
	O
Multimodular	O
structure	O
of	O
SGBP	O
-	O
B	O
(	O
Bacova_02650	O
).	O
(	O
A	O
)	O
Full	O
-	O
length	O
structure	O
of	O
SGBP	O
-	O
B	O
,	O
color	O
coded	O
by	O
domain	O
as	O
indicated	O
.	O
	O
An	O
omit	O
map	O
(	O
2σ	O
)	O
for	O
XyGO2	O
is	O
displayed	O
to	O
highlight	O
the	O
location	O
of	O
the	O
glycan	O
-	O
binding	O
site	O
.	O
	O
(	O
B	O
)	O
Overlay	O
of	O
SGBP	O
-	O
B	O
domains	O
A	O
,	O
B	O
,	O
and	O
C	O
(	O
colored	O
as	O
in	O
panel	O
A	O
),	O
with	O
a	O
C	O
-	O
terminal	O
Ig	O
-	O
like	O
domain	O
of	O
the	O
G	O
.	O
stearothermophilus	O
cyclodextrin	O
glucanotransferase	O
(	O
PDB	O
1CYG	O
[	O
residues	O
375	O
to	O
493	O
])	O
in	O
green	O
.	O
(	O
C	O
)	O
Cα	O
overlay	O
of	O
SGBP	O
-	O
B	O
(	O
gray	O
)	O
and	O
Bacova_04391	O
(	O
PDB	O
3ORJ	O
)	O
(	O
pink	O
).	O
	O
(	O
D	O
)	O
Close	O
-	O
up	O
omit	O
map	O
for	O
the	O
XyGO2	O
ligand	O
,	O
contoured	O
at	O
2σ	O
.	O
(	O
E	O
)	O
Stereo	O
view	O
of	O
the	O
xyloglucan	O
-	O
binding	O
site	O
of	O
SGBP	O
-	O
B	O
,	O
displaying	O
all	O
residues	O
within	O
4	O
Å	O
of	O
the	O
ligand	O
.	O
	O
The	O
backbone	O
glucose	O
residues	O
are	O
numbered	O
from	O
the	O
nonreducing	O
end	O
,	O
xylose	O
residues	O
are	O
shown	O
as	O
X1	O
,	O
X2	O
,	O
and	O
X3	O
,	O
potential	O
hydrogen	O
-	O
bonding	O
interactions	O
are	O
shown	O
as	O
dashed	O
lines	O
,	O
and	O
the	O
distance	O
is	O
shown	O
in	O
angstroms	O
.	O
	O
Domains	O
A	O
,	O
B	O
,	O
and	O
C	O
do	O
not	O
pack	O
against	O
each	O
other	O
.	O
	O
Moreover	O
,	O
the	O
five	O
-	O
residue	O
linkers	O
between	O
these	O
first	O
three	O
domains	O
all	O
feature	O
a	O
proline	O
as	O
the	O
middle	O
residue	O
,	O
suggesting	O
significant	O
conformational	O
rigidity	O
(	O
Fig	O
.	O
5A	O
).	O
	O
Any	O
mobility	O
of	O
SGBP	O
-	O
B	O
on	O
the	O
surface	O
of	O
the	O
cell	O
(	O
beyond	O
lateral	O
diffusion	O
within	O
the	O
membrane	O
)	O
is	O
likely	O
imparted	O
by	O
the	O
eight	O
-	O
residue	O
linker	O
that	O
spans	O
the	O
predicted	O
lipidated	O
Cys	O
(	O
C28	O
)	O
and	O
the	O
first	O
β	O
-	O
strand	O
of	O
domain	O
A	O
.	O
Other	O
outer	O
membrane	O
proteins	O
from	O
various	O
Sus	O
-	O
like	O
systems	O
possess	O
a	O
similar	O
10	O
-	O
to	O
20	O
-	O
amino	O
-	O
acid	O
flexible	O
linker	O
between	O
the	O
lipidated	O
Cys	O
that	O
tethers	O
the	O
protein	O
to	O
the	O
outside	O
the	O
cell	O
and	O
the	O
first	O
secondary	O
structure	O
element	O
.	O
	O
Analogously	O
,	O
the	O
outer	O
membrane	O
-	O
anchored	O
endo	O
-	O
xyloglucanase	O
BoGH5	O
of	O
the	O
XyGUL	O
contains	O
a	O
100	O
-	O
amino	O
-	O
acid	O
,	O
all	O
-	O
β	O
-	O
strand	O
,	O
N	O
-	O
terminal	O
module	O
and	O
flexible	O
linker	O
that	O
imparts	O
conformational	O
flexibility	O
and	O
distances	O
the	O
catalytic	O
module	O
from	O
the	O
cell	O
surface	O
.	O
	O
XyG	O
binds	O
to	O
domain	O
D	O
of	O
SGBP	O
-	O
B	O
at	O
the	O
concave	O
interface	O
of	O
the	O
top	O
β	O
-	O
sheet	O
,	O
with	O
binding	O
mediated	O
by	O
loops	O
connecting	O
the	O
β	O
-	O
strands	O
.	O
	O
Six	O
glucosyl	O
residues	O
,	O
comprising	O
the	O
main	O
chain	O
,	O
and	O
three	O
branching	O
xylosyl	O
residues	O
of	O
XyGO2	O
can	O
be	O
modeled	O
in	O
the	O
density	O
(	O
Fig	O
.	O
5D	O
;	O
cf	O
.	O
	O
The	O
aromatic	O
platform	O
created	O
by	O
W330	O
,	O
W364	O
,	O
and	O
Y363	O
spans	O
four	O
glucosyl	O
residues	O
,	O
compared	O
to	O
the	O
longer	O
platform	O
of	O
SGBP	O
-	O
A	O
,	O
which	O
supports	O
six	O
glucosyl	O
residues	O
(	O
Fig	O
.	O
5E	O
).	O
	O
The	O
Y363A	B-mutant
site	O
-	O
directed	O
mutant	O
of	O
SGBP	O
-	O
B	O
displays	O
a	O
20	O
-	O
fold	O
decrease	O
in	O
the	O
Ka	O
for	O
XyG	O
,	O
while	O
the	O
W364A	B-mutant
mutant	O
lacks	O
XyG	O
binding	O
(	O
Table	O
3	O
;	O
see	O
Fig	O
.	O
S6	O
in	O
the	O
supplemental	O
material	O
).	O
	O
There	O
are	O
no	O
additional	O
contacts	O
between	O
the	O
protein	O
and	O
the	O
β	O
-	O
glucan	O
backbone	O
and	O
surprisingly	O
few	O
interactions	O
with	O
the	O
side	O
-	O
chain	O
xylosyl	O
residues	O
,	O
despite	O
that	O
fact	O
that	O
ITC	O
data	O
demonstrate	O
that	O
SGBP	O
-	O
B	O
does	O
not	O
measurably	O
bind	O
the	O
cellohexaose	O
(	O
Table	O
1	O
).	O
	O
F414	O
stacks	O
with	O
the	O
xylosyl	O
residue	O
of	O
Glc3	O
,	O
while	O
Q407	O
is	O
positioned	O
for	O
hydrogen	O
bonding	O
with	O
the	O
O4	O
of	O
xylosyl	O
residue	O
Xyl1	O
.	O
	O
Surprisingly	O
,	O
an	O
F414A	B-mutant
mutant	O
of	O
SGBP	O
-	O
B	O
displays	O
only	O
a	O
mild	O
3	O
-	O
fold	O
decrease	O
in	O
the	O
Ka	O
value	O
for	O
XyG	O
,	O
again	O
suggesting	O
that	O
glycan	O
recognition	O
is	O
primarily	O
mediated	O
via	O
contact	O
with	O
the	O
β	O
-	O
glucan	O
backbone	O
(	O
Table	O
3	O
;	O
see	O
Fig	O
.	O
S6	O
).	O
	O
Additional	O
residues	O
surrounding	O
the	O
binding	O
site	O
,	O
including	O
Y369	O
and	O
E412	O
,	O
may	O
contribute	O
to	O
the	O
recognition	O
of	O
more	O
highly	O
decorated	O
XyG	O
,	O
but	O
precisely	O
how	O
this	O
is	O
mediated	O
is	O
presently	O
unclear	O
.	O
	O
The	O
CD	O
domains	O
of	O
the	O
truncated	O
and	O
full	O
-	O
length	O
proteins	O
superimpose	O
with	O
a	O
0	O
.	O
4	O
-	O
Å	O
RMSD	O
of	O
the	O
Cα	O
backbone	O
,	O
with	O
no	O
differences	O
in	O
the	O
position	O
of	O
any	O
of	O
the	O
glycan	O
-	O
binding	O
residues	O
(	O
see	O
Fig	O
.	O
S7A	O
in	O
the	O
supplemental	O
material	O
).	O
	O
While	O
density	O
is	O
observed	O
for	O
XyGO2	O
,	O
the	O
ligand	O
could	O
not	O
be	O
unambiguously	O
modeled	O
into	O
this	O
density	O
to	O
achieve	O
a	O
reasonable	O
fit	O
between	O
the	O
X	O
-	O
ray	O
data	O
and	O
the	O
known	O
stereochemistry	O
of	O
the	O
sugar	O
(	O
see	O
Fig	O
.	O
S7B	O
and	O
C	O
).	O
	O
SGBP	O
-	O
A	O
and	O
SGBP	O
-	O
B	O
have	O
distinct	O
,	O
coordinated	O
functions	O
in	O
vivo	O
.	O
	O
To	O
disentangle	O
the	O
functions	O
of	O
SGBP	O
-	O
A	O
and	O
SGBP	O
-	O
B	O
in	O
XyG	O
recognition	O
and	O
uptake	O
,	O
we	O
created	O
individual	O
in	O
-	O
frame	O
deletion	O
and	O
complementation	O
mutant	O
strains	O
of	O
B	O
.	O
ovatus	O
.	O
	O
Growth	O
on	O
glucose	O
displayed	O
the	O
shortest	O
lag	O
time	O
for	O
each	O
strain	O
,	O
and	O
so	O
lag	O
times	O
were	O
normalized	O
for	O
each	O
carbohydrate	O
by	O
subtracting	O
the	O
lag	O
time	O
of	O
that	O
strain	O
in	O
glucose	O
(	O
Fig	O
.	O
6	O
;	O
see	O
Fig	O
.	O
S8	O
in	O
the	O
supplemental	O
material	O
).	O
	O
A	O
strain	O
in	O
which	O
the	O
entire	O
XyGUL	O
is	O
deleted	O
displays	O
a	O
lag	O
of	O
24	O
.	O
5	O
h	O
during	O
growth	O
on	O
glucose	O
compared	O
to	O
the	O
isogenic	O
parental	O
wild	O
-	O
type	O
(	O
WT	O
)	O
Δtdk	B-mutant
strain	O
,	O
for	O
which	O
exponential	O
growth	O
lags	O
for	O
19	O
.	O
8	O
h	O
(	O
see	O
Fig	O
.	O
S8D	O
).	O
	O
It	O
is	O
unknown	O
whether	O
this	O
is	O
because	O
cultures	O
were	O
not	O
normalized	O
by	O
the	O
starting	O
optical	O
density	O
(	O
OD	O
)	O
or	O
viable	O
cells	O
or	O
reflects	O
a	O
minor	O
defect	O
for	O
glucose	O
utilization	O
.	O
	O
The	O
former	O
seems	O
more	O
likely	O
as	O
the	O
growth	O
rates	O
are	O
nearly	O
identical	O
for	O
these	O
strains	O
on	O
glucose	O
and	O
xylose	O
.	O
	O
The	O
ΔXyGUL	B-mutant
and	O
WT	O
Δtdk	B-mutant
strains	O
display	O
normalized	O
lag	O
times	O
on	O
xylose	O
within	O
experimental	O
error	O
,	O
and	O
curiously	O
some	O
of	O
the	O
mutant	O
and	O
complemented	O
strains	O
display	O
a	O
nominally	O
shorter	O
lag	O
time	O
on	O
xylose	O
than	O
the	O
WT	O
Δtdk	B-mutant
strain	O
.	O
	O
The	O
reason	O
for	O
this	O
observation	O
on	O
XyGO2	O
is	O
unclear	O
,	O
as	O
the	O
ΔSGBP	B-mutant
-	I-mutant
B	I-mutant
mutant	O
does	O
not	O
have	O
a	O
significantly	O
different	O
growth	O
rate	O
from	O
the	O
WT	O
on	O
XyGO2	O
.	O
	O
Growth	O
of	O
select	O
XyGUL	O
mutants	O
on	O
xyloglucan	O
and	O
oligosaccharides	O
.	O
	O
B	O
.	O
ovatus	O
mutants	O
were	O
created	O
in	O
a	O
thymidine	B-mutant
kinase	I-mutant
deletion	I-mutant
(	O
Δtdk	B-mutant
)	O
mutant	O
as	O
described	O
previously	O
.	O
	O
SGBP	B-mutant
-	I-mutant
A	I-mutant
*	I-mutant
denotes	O
the	O
Bacova_02651	O
(	O
W82A	B-mutant
W283A	B-mutant
W306A	B-mutant
)	O
allele	O
,	O
and	O
the	O
GH9	O
gene	O
is	O
Bacova_02649	O
.	O
	O
Solid	O
bars	O
indicate	O
conditions	O
that	O
are	O
not	O
statistically	O
significant	O
from	O
the	O
WT	O
Δtdk	B-mutant
cultures	O
grown	O
on	O
the	O
indicated	O
carbohydrate	O
,	O
while	O
open	O
bars	O
indicate	O
a	O
P	O
value	O
of	O
<	O
0	O
.	O
005	O
compared	O
to	O
the	O
WT	O
Δtdk	B-mutant
strain	O
.	O
	O
Conditions	O
denoted	O
by	O
the	O
same	O
letter	O
(	O
b	O
,	O
c	O
,	O
or	O
d	O
)	O
are	O
not	O
statistically	O
significant	O
from	O
each	O
other	O
but	O
are	O
significantly	O
different	O
from	O
the	O
condition	O
labeled	O
“	O
a	O
.”	O
Complementation	O
of	O
ΔSGBP	B-mutant
-	I-mutant
A	I-mutant
and	O
ΔSBGP	B-mutant
-	I-mutant
B	I-mutant
was	O
performed	O
by	O
allelic	O
exchange	O
of	O
the	O
wild	O
-	O
type	O
genes	O
back	O
into	O
the	O
genome	O
for	O
expression	O
via	O
the	O
native	O
promoter	O
:	O
these	O
growth	O
curves	O
,	O
quantified	O
rates	O
and	O
lag	O
times	O
are	O
displayed	O
in	O
Fig	O
.	O
S8	O
in	O
the	O
supplemental	O
material	O
.	O
	O
Fig	O
.	O
1B	O
)	O
was	O
completely	O
incapable	O
of	O
growth	O
on	O
XyG	O
,	O
XyGO1	O
,	O
and	O
XyGO2	O
,	O
indicating	O
that	O
SGBP	O
-	O
A	O
is	O
essential	O
for	O
XyG	O
utilization	O
(	O
Fig	O
.	O
6	O
).	O
	O
This	O
result	O
mirrors	O
our	O
previous	O
data	O
for	O
the	O
canonical	O
Sus	O
of	O
B	O
.	O
thetaiotaomicron	O
,	O
which	O
revealed	O
that	O
a	O
homologous	O
ΔsusD	B-mutant
mutant	O
is	O
unable	O
to	O
grow	O
on	O
starch	O
or	O
malto	O
-	O
oligosaccharides	O
,	O
despite	O
normal	O
cell	O
surface	O
expression	O
of	O
all	O
other	O
PUL	O
-	O
encoded	O
proteins	O
.	O
	O
More	O
recently	O
,	O
we	O
demonstrated	O
that	O
this	O
phenotype	O
is	O
due	O
to	O
the	O
loss	O
of	O
the	O
physical	O
presence	O
of	O
SusD	O
;	O
complementation	O
of	O
ΔsusD	B-mutant
with	O
SusD	B-mutant
*,	I-mutant
a	O
triple	O
site	O
-	O
directed	O
mutant	O
(	O
W96A	B-mutant
W320A	B-mutant
Y296A	B-mutant
)	O
that	O
ablates	O
glycan	O
binding	O
,	O
restores	O
B	O
.	O
thetaiotaomicron	O
growth	O
on	O
malto	O
-	O
oligosaccharides	O
and	O
starch	O
when	O
sus	O
transcription	O
is	O
induced	O
by	O
maltose	O
addition	O
.	O
	O
Similarly	O
,	O
the	O
function	O
of	O
SGBP	O
-	O
A	O
extends	O
beyond	O
glycan	O
binding	O
.	O
	O
Complementation	O
of	O
ΔSGBP	B-mutant
-	I-mutant
A	I-mutant
with	O
the	O
SGBP	B-mutant
-	I-mutant
A	I-mutant
*	I-mutant
(	O
W82A	B-mutant
W283A	B-mutant
W306A	B-mutant
)	O
variant	O
,	O
which	O
does	O
not	O
bind	O
XyG	O
,	O
supports	O
growth	O
on	O
XyG	O
and	O
XyGOs	O
(	O
Fig	O
.	O
6	O
;	O
ΔSGBP	B-mutant
-	I-mutant
A	I-mutant
::	O
SGBP	B-mutant
-	I-mutant
A	I-mutant
*),	I-mutant
with	O
growth	O
rates	O
that	O
are	O
~	O
70	O
%	O
that	O
of	O
the	O
WT	O
.	O
	O
In	O
previous	O
studies	O
,	O
we	O
observed	O
that	O
carbohydrate	O
binding	O
by	O
SusD	O
enhanced	O
the	O
sensitivity	O
of	O
the	O
cells	O
to	O
limiting	O
concentrations	O
of	O
malto	O
-	O
oligosaccharides	O
by	O
several	O
orders	O
of	O
magnitude	O
,	O
such	O
that	O
the	O
addition	O
of	O
0	O
.	O
5	O
g	O
/	O
liter	O
maltose	O
was	O
required	O
to	O
restore	O
growth	O
of	O
the	O
ΔsusD	B-mutant
::	O
SusD	B-mutant
*	I-mutant
strain	O
on	O
starch	O
,	O
which	O
nonetheless	O
occurred	O
following	O
an	O
extended	O
lag	O
phase	O
.	O
	O
In	O
contrast	O
,	O
the	O
ΔSGBP	B-mutant
-	I-mutant
A	I-mutant
::	O
SGBP	B-mutant
-	I-mutant
A	I-mutant
*	I-mutant
strain	O
does	O
not	O
display	O
an	O
extended	O
lag	O
time	O
on	O
any	O
of	O
the	O
xyloglucan	O
substrates	O
compared	O
to	O
the	O
WT	O
(	O
Fig	O
.	O
6	O
).	O
	O
However	O
,	O
the	O
modest	O
rate	O
defect	O
displayed	O
by	O
the	O
SGBP	O
-	O
A	O
::	O
SGBP	B-mutant
-	I-mutant
A	I-mutant
*	I-mutant
strain	O
suggests	O
that	O
recognition	O
of	O
XyG	O
and	O
product	O
import	O
is	O
somewhat	O
less	O
efficient	O
in	O
these	O
cells	O
.	O
	O
10	O
-	O
fold	O
more	O
weakly	O
than	O
XyGO2	O
and	O
XyG	O
(	O
Fig	O
.	O
6	O
;	O
Table	O
1	O
).	O
	O
As	O
such	O
,	O
the	O
data	O
suggest	O
that	O
SGBP	O
-	O
A	O
can	O
compensate	O
for	O
the	O
loss	O
of	O
function	O
of	O
SGBP	O
-	O
B	O
on	O
longer	O
oligo	O
-	O
and	O
polysaccharides	O
,	O
while	O
SGBP	O
-	O
B	O
may	O
adapt	O
the	O
cell	O
to	O
recognize	O
smaller	O
oligosaccharides	O
efficiently	O
.	O
	O
Indeed	O
,	O
a	O
double	O
mutant	O
,	O
consisting	O
of	O
a	O
crippled	O
SGBP	O
-	O
A	O
and	O
a	O
deletion	O
of	O
SGBP	O
-	O
B	O
(	O
ΔSGBP	B-mutant
-	I-mutant
A	I-mutant
::	O
SGBP	B-mutant
-	I-mutant
A	I-mutant
*/	I-mutant
ΔSGBP	B-mutant
-	I-mutant
B	I-mutant
),	O
exhibits	O
an	O
extended	O
lag	O
time	O
on	O
both	O
XyG	O
and	O
XyGO2	O
,	O
as	O
well	O
as	O
XyGO1	O
.	O
	O
This	O
additional	O
role	O
of	O
SGBP	O
-	O
B	O
is	O
especially	O
notable	O
in	O
the	O
context	O
of	O
studies	O
on	O
BtSusE	O
and	O
BtSusF	O
(	O
positioned	O
similarly	O
in	O
the	O
archetypal	O
Sus	O
locus	O
)	O
(	O
Fig	O
.	O
1B	O
),	O
for	O
which	O
growth	O
defects	O
on	O
starch	O
or	O
malto	O
-	O
oligosaccharides	O
have	O
never	O
been	O
observed	O
.	O
	O
However	O
,	O
combined	O
deletion	O
of	O
the	O
genes	O
encoding	O
GH9	O
(	O
encoded	O
by	O
Bacova_02649	O
)	O
and	O
SGBP	O
-	O
B	O
does	O
not	O
exacerbate	O
the	O
growth	O
defect	O
on	O
XyGO1	O
(	O
Fig	O
.	O
6	O
;	O
ΔSGBP	B-mutant
-	I-mutant
B	I-mutant
/	O
ΔGH9	B-mutant
).	O
	O
The	O
necessity	O
of	O
SGBP	O
-	O
B	O
is	O
elevated	O
in	O
the	O
SGBP	B-mutant
-	I-mutant
A	I-mutant
*	I-mutant
strain	O
,	O
as	O
the	O
ΔSGBP	B-mutant
-	I-mutant
A	I-mutant
::	O
SGBP	B-mutant
-	I-mutant
A	I-mutant
*/	I-mutant
ΔSGBP	B-mutant
-	I-mutant
B	I-mutant
mutant	O
displays	O
an	O
extended	O
lag	O
during	O
growth	O
on	O
XyG	O
and	O
xylogluco	O
-	O
oligosaccharides	O
,	O
while	O
growth	O
rate	O
differences	O
are	O
more	O
subtle	O
.	O
	O
The	O
precise	O
reason	O
for	O
this	O
lag	O
is	O
unclear	O
,	O
but	O
recapitulating	O
our	O
findings	O
on	O
the	O
role	O
of	O
SusD	O
in	O
malto	O
-	O
oligosaccharide	O
sensing	O
in	O
B	O
.	O
thetaiotaomicron	O
,	O
this	O
extended	O
lag	O
may	O
be	O
due	O
to	O
inefficient	O
import	O
and	O
thus	O
sensing	O
of	O
xyloglucan	O
in	O
the	O
environment	O
in	O
the	O
absence	O
of	O
glycan	O
binding	O
by	O
essential	O
SGBPs	O
.	O
	O
Our	O
previous	O
work	O
demonstrates	O
that	O
B	O
.	O
ovatus	O
cells	O
grown	O
in	O
minimal	O
medium	O
plus	O
glucose	O
express	O
low	O
levels	O
of	O
the	O
XyGUL	O
transcript	O
.	O
	O
Thus	O
,	O
in	O
our	O
experiments	O
,	O
we	O
presume	O
that	O
each	O
strain	O
,	O
initially	O
grown	O
in	O
glucose	O
,	O
expresses	O
low	O
levels	O
of	O
the	O
XyGUL	O
transcript	O
and	O
thus	O
low	O
levels	O
of	O
the	O
XyGUL	O
-	O
encoded	O
surface	O
proteins	O
,	O
including	O
the	O
vanguard	O
GH5	O
.	O
	O
Presumably	O
without	O
glycan	O
binding	O
by	O
the	O
SGBPs	O
,	O
the	O
GH5	O
protein	O
cannot	O
efficiently	O
process	O
xyloglucan	O
,	O
and	O
/	O
or	O
the	O
lack	O
of	O
SGBP	O
function	O
prevents	O
efficient	O
capture	O
and	O
import	O
of	O
the	O
processed	O
oligosaccharides	O
.	O
	O
In	O
the	O
BtSus	O
,	O
SusD	O
and	O
the	O
TBDT	O
SusC	O
interact	O
,	O
and	O
we	O
speculate	O
that	O
this	O
interaction	O
is	O
necessary	O
for	O
glycan	O
uptake	O
,	O
as	O
suggested	O
by	O
the	O
fact	O
that	O
a	O
ΔsusD	B-mutant
mutant	O
cannot	O
grow	O
on	O
starch	O
,	O
but	O
a	O
ΔsusD	B-mutant
::	O
SusD	B-mutant
*	I-mutant
strain	O
regains	O
this	O
ability	O
if	O
a	O
transcriptional	O
activator	O
of	O
the	O
sus	O
operon	O
is	O
supplied	O
.	O
	O
However	O
,	O
unlike	O
the	O
Sus	O
,	O
in	O
which	O
elimination	O
of	O
SusE	O
and	O
SusF	O
does	O
not	O
affect	O
growth	O
on	O
starch	O
,	O
SGBP	O
-	O
B	O
appears	O
to	O
have	O
a	O
dedicated	O
role	O
in	O
growth	O
on	O
small	O
xylogluco	O
-	O
oligosaccharides	O
.	O
	O
The	O
ability	O
of	O
gut	O
-	O
adapted	O
microorganisms	O
to	O
thrive	O
in	O
the	O
gastrointestinal	O
tract	O
is	O
critically	O
dependent	O
upon	O
their	O
ability	O
to	O
efficiently	O
recognize	O
,	O
cleave	O
,	O
and	O
import	O
glycans	O
.	O
	O
The	O
human	O
gut	O
,	O
in	O
particular	O
,	O
is	O
a	O
densely	O
packed	O
ecosystem	O
with	O
hundreds	O
of	O
species	O
,	O
in	O
which	O
there	O
is	O
potential	O
for	O
both	O
competition	O
and	O
synergy	O
in	O
the	O
utilization	O
of	O
different	O
substrates	O
.	O
	O
Recent	O
work	O
has	O
elucidated	O
that	O
Bacteroidetes	O
cross	O
-	O
feed	O
during	O
growth	O
on	O
many	O
glycans	O
;	O
the	O
glycoside	O
hydrolases	O
expressed	O
by	O
one	O
species	O
liberate	O
oligosaccharides	O
for	O
consumption	O
by	O
other	O
members	O
of	O
the	O
community	O
.	O
	O
Here	O
,	O
we	O
demonstrate	O
that	O
the	O
surface	O
glycan	O
binding	O
proteins	O
encoded	O
within	O
the	O
BoXyGUL	O
play	O
unique	O
and	O
essential	O
roles	O
in	O
the	O
acquisition	O
of	O
the	O
ubiquitous	O
and	O
abundant	O
vegetable	O
polysaccharide	O
xyloglucan	O
.	O
	O
Yet	O
,	O
a	O
number	O
of	O
questions	O
remain	O
regarding	O
the	O
molecular	O
interplay	O
of	O
SGBPs	O
with	O
their	O
cotranscribed	O
cohort	O
of	O
glycoside	O
hydrolases	O
and	O
TonB	O
-	O
dependent	O
transporters	O
.	O
	O
A	O
direct	O
interaction	O
between	O
the	O
BtSusC	O
TBDT	O
and	O
the	O
SusD	O
SGBP	O
has	O
been	O
previously	O
demonstrated	O
,	O
as	O
has	O
an	O
interaction	O
between	O
the	O
homologous	O
components	O
encoded	O
by	O
an	O
N	O
-	O
glycan	O
-	O
scavenging	O
PUL	O
of	O
Capnocytophaga	O
canimorsus	O
.	O
	O
It	O
is	O
yet	O
presently	O
unclear	O
whether	O
this	O
interaction	O
is	O
static	O
or	O
dynamic	O
and	O
to	O
what	O
extent	O
the	O
association	O
of	O
cognate	O
TBDT	O
/	O
SGBPs	O
is	O
dependent	O
upon	O
the	O
structure	O
of	O
the	O
carbohydrate	O
to	O
be	O
imported	O
.	O
	O
On	O
the	O
other	O
hand	O
,	O
there	O
is	O
clear	O
evidence	O
for	O
independent	O
TBDTs	O
in	O
Bacteroidetes	O
that	O
do	O
not	O
require	O
SGBP	O
association	O
for	O
activity	O
.	O
	O
For	O
example	O
,	O
it	O
was	O
recently	O
demonstrated	O
that	O
expression	O
of	O
nanO	O
,	O
which	O
encodes	O
a	O
SusC	O
-	O
like	O
TBDT	O
as	O
part	O
of	O
a	O
sialic	O
-	O
acid	O
-	O
targeting	O
PUL	O
from	O
B	O
.	O
fragilis	O
,	O
restored	O
growth	O
on	O
this	O
monosaccharide	O
in	O
a	O
mutant	O
strain	O
of	O
E	O
.	O
coli	O
.	O
	O
In	O
this	O
instance	O
,	O
coexpression	O
of	O
the	O
susD	O
-	O
like	O
gene	O
nanU	O
was	O
not	O
required	O
,	O
nor	O
did	O
the	O
expression	O
of	O
the	O
nanU	O
gene	O
enhance	O
growth	O
kinetics	O
.	O
	O
Thus	O
,	O
the	O
strict	O
dependence	O
on	O
a	O
SusD	O
-	O
like	O
SGBP	O
for	O
glycan	O
uptake	O
in	O
the	O
Bacteroidetes	O
may	O
be	O
variable	O
and	O
substrate	O
dependent	O
.	O
	O
Such	O
is	O
the	O
case	O
for	O
XyGUL	O
from	O
related	O
Bacteroides	O
species	O
,	O
which	O
may	O
encode	O
either	O
one	O
or	O
two	O
of	O
these	O
predicted	O
SGBPs	O
,	O
and	O
these	O
proteins	O
vary	O
considerably	O
in	O
length	O
.	O
	O
The	O
extremely	O
low	O
similarity	O
of	O
these	O
SGBPs	O
is	O
striking	O
in	O
light	O
of	O
the	O
moderate	O
sequence	O
conservation	O
observed	O
among	O
homologous	O
GHs	O
in	O
syntenic	O
PUL	O
.	O
	O
This	O
,	O
together	O
with	O
the	O
observation	O
that	O
these	O
SGBPs	O
,	O
as	O
exemplified	O
by	O
BtSusE	O
and	O
BtSusF	O
and	O
the	O
XyGUL	O
SGBP	O
-	O
B	O
of	O
the	O
present	O
study	O
,	O
are	O
expendable	O
for	O
polysaccharide	O
growth	O
,	O
implies	O
a	O
high	O
degree	O
of	O
evolutionary	O
flexibility	O
to	O
enhance	O
glycan	O
capture	O
at	O
the	O
cell	O
surface	O
.	O
	O
However	O
,	O
the	O
natural	O
diversity	O
of	O
these	O
proteins	O
represents	O
a	O
rich	O
source	O
for	O
the	O
discovery	O
of	O
unique	O
carbohydrate	O
-	O
binding	O
motifs	O
to	O
both	O
inform	O
gut	O
microbiology	O
and	O
generate	O
new	O
,	O
specific	O
carbohydrate	O
analytical	O
reagents	O
.	O
	O
In	O
conclusion	O
,	O
the	O
present	O
study	O
further	O
illuminates	O
the	O
essential	O
role	O
that	O
surface	O
-	O
glycan	O
binding	O
proteins	O
play	O
in	O
facilitating	O
the	O
catabolism	O
of	O
complex	O
dietary	O
carbohydrates	O
by	O
Bacteroidetes	O
.	O
	O
The	O
ability	O
of	O
our	O
resident	O
gut	O
bacteria	O
to	O
recognize	O
polysaccharides	O
is	O
the	O
first	O
committed	O
step	O
of	O
glycan	O
consumption	O
by	O
these	O
organisms	O
,	O
a	O
critical	O
process	O
that	O
influences	O
the	O
community	O
structure	O
and	O
thus	O
the	O
metabolic	O
output	O
(	O
i	O
.	O
e	O
.,	O
short	O
-	O
chain	O
fatty	O
acid	O
and	O
metabolite	O
profile	O
)	O
of	O
these	O
organisms	O
.	O
	O
Inhibiting	O
complex	O
IL	O
-	O
17A	O
and	O
IL	O
-	O
17RA	O
interactions	O
with	O
a	O
linear	O
peptide	O
	O
IL	O
-	O
17A	O
is	O
a	O
pro	O
-	O
inflammatory	O
cytokine	O
that	O
has	O
been	O
implicated	O
in	O
autoimmune	O
and	O
inflammatory	O
diseases	O
.	O
	O
HAP	O
binds	O
specifically	O
to	O
IL	O
-	O
17A	O
and	O
inhibits	O
the	O
interaction	O
of	O
the	O
cytokine	O
with	O
its	O
receptor	O
,	O
IL	O
-	O
17RA	O
.	O
	O
Crystal	O
structure	O
studies	O
revealed	O
that	O
two	O
HAP	O
molecules	O
bind	O
to	O
one	O
IL	O
-	O
17A	O
dimer	O
symmetrically	O
.	O
	O
The	O
N	O
-	O
terminal	O
portions	O
of	O
HAP	O
form	O
a	O
β	O
-	O
strand	O
that	O
inserts	O
between	O
two	O
IL	O
-	O
17A	O
monomers	O
while	O
the	O
C	O
-	O
terminal	O
section	O
forms	O
an	O
α	O
helix	O
that	O
directly	O
blocks	O
IL	O
-	O
17RA	O
from	O
binding	O
to	O
the	O
same	O
region	O
of	O
IL	O
-	O
17A	O
.	O
	O
IL	O
-	O
17A	O
signals	O
through	O
a	O
specific	O
cell	O
surface	O
receptor	O
complex	O
which	O
consists	O
of	O
IL	O
-	O
17RA	O
and	O
IL	O
-	O
17RC	O
.	O
	O
IL	O
-	O
17A	O
’	O
s	O
downstream	O
signaling	O
leads	O
to	O
increased	O
production	O
of	O
inflammatory	O
cytokines	O
such	O
as	O
IL	O
-	O
6	O
,	O
IL	O
-	O
8	O
,	O
CCL	O
-	O
20	O
and	O
CXCL1	O
by	O
various	O
mechanisms	O
including	O
stimulation	O
of	O
transcription	O
and	O
stabilization	O
of	O
mRNA	O
.	O
	O
Although	O
various	O
cell	O
types	O
have	O
been	O
reported	O
to	O
express	O
IL	O
-	O
17RA	O
,	O
the	O
highest	O
responses	O
to	O
IL	O
-	O
17A	O
come	O
from	O
epithelial	O
cells	O
,	O
endothelial	O
cells	O
,	O
keratinocytes	O
and	O
fibroblasts	O
.	O
	O
IL	O
-	O
17A	O
and	O
its	O
signaling	O
is	O
important	O
in	O
host	O
defense	O
against	O
certain	O
fungal	O
and	O
bacterial	O
infections	O
as	O
demonstrated	O
by	O
patients	O
with	O
autoantibodies	O
against	O
IL	O
-	O
17A	O
and	O
IL	O
-	O
17F	O
,	O
or	O
with	O
inborn	O
errors	O
of	O
IL	O
-	O
17	O
immunity	O
.	O
	O
In	O
addition	O
to	O
its	O
physiological	O
role	O
,	O
IL	O
-	O
17A	O
is	O
a	O
key	O
pathogenic	O
factor	O
in	O
inflammatory	O
and	O
autoimmune	O
diseases	O
.	O
	O
In	O
phase	O
II	O
and	O
III	O
clinical	O
trials	O
,	O
neutralizing	O
monoclonal	O
antibodies	O
against	O
IL	O
-	O
17A	O
(	O
secukinumab	O
and	O
ixekizumab	O
)	O
or	O
its	O
receptor	O
IL	O
-	O
17RA	O
(	O
brodalumab	O
)	O
are	O
highly	O
efficacious	O
in	O
treating	O
moderate	O
to	O
severe	O
plaque	O
psoriasis	O
and	O
psoriatic	O
arthritis	O
.	O
	O
Secukinumab	O
has	O
been	O
approved	O
recently	O
as	O
a	O
new	O
psoriasis	O
drug	O
by	O
the	O
US	O
Food	O
and	O
Drug	O
Administration	O
(	O
Cosentyx	O
™).	O
	O
In	O
addition	O
to	O
psoriasis	O
and	O
psoriatic	O
arthritis	O
,	O
IL	O
-	O
17A	O
blockade	O
has	O
also	O
shown	O
preclinical	O
and	O
clinical	O
efficacies	O
in	O
ankylosing	O
spondylitis	O
and	O
rheumatoid	O
arthritis	O
.	O
	O
Among	O
IL	O
-	O
17	O
cytokines	O
,	O
IL	O
-	O
17A	O
and	O
IL	O
-	O
17F	O
share	O
the	O
highest	O
homology	O
.	O
	O
Structures	O
are	O
known	O
for	O
apo	O
IL	O
-	O
17F	O
and	O
its	O
complex	O
with	O
IL	O
-	O
17RA	O
,	O
for	O
apo	O
IL	O
-	O
17A	O
,	O
its	O
complex	O
with	O
an	O
antibody	O
Fab	O
,	O
and	O
its	O
complex	O
with	O
IL	O
-	O
17RA	O
.	O
	O
Developing	O
small	O
molecules	O
targeting	O
protein	O
-	O
protein	O
interactions	O
is	O
difficult	O
with	O
particular	O
challenges	O
associated	O
with	O
the	O
large	O
,	O
shallow	O
IL	O
-	O
17A	O
/	O
IL	O
-	O
17RA	O
interfaces	O
.	O
	O
Our	O
efforts	O
resulted	O
in	O
discovery	O
of	O
a	O
high	O
affinity	O
IL	O
-	O
17A	O
peptide	O
antagonist	O
(	O
HAP	O
),	O
which	O
we	O
attempted	O
to	O
increase	O
the	O
functional	O
production	O
and	O
pharmacokinetics	O
after	O
fusing	O
HAP	O
to	O
antibodies	O
for	O
evaluation	O
as	O
a	O
bispecific	O
therapeutic	O
in	O
animal	O
studies	O
.	O
	O
Unfortunately	O
,	O
this	O
past	O
work	O
revealed	O
stability	O
issues	O
of	O
the	O
uncapped	O
HAP	O
in	O
cell	O
culture	O
Here	O
,	O
we	O
provide	O
the	O
details	O
of	O
the	O
discovery	O
and	O
optimization	O
that	O
led	O
to	O
HAP	O
and	O
report	O
the	O
complex	O
structure	O
of	O
IL	O
-	O
17A	O
with	O
HAP	O
,	O
which	O
provides	O
structure	O
based	O
rationalization	O
of	O
peptide	O
optimization	O
and	O
structure	O
activity	O
relationship	O
(	O
SAR	O
).	O
	O
Single	O
clones	O
were	O
isolated	O
and	O
sub	O
-	O
cultured	O
in	O
growth	O
medium	O
,	O
and	O
culture	O
supernatants	O
were	O
used	O
in	O
an	O
enzyme	O
-	O
linked	O
immunosorbent	O
assay	O
(	O
ELISA	O
)	O
to	O
identify	O
specific	O
IL	O
-	O
17A	O
-	O
binding	O
clones	O
.	O
	O
Approximately	O
10	O
%	O
of	O
the	O
clones	O
that	O
specifically	O
bound	O
to	O
IL	O
-	O
17A	O
also	O
prevented	O
the	O
cytokine	O
from	O
binding	O
to	O
IL	O
-	O
17RA	O
.	O
	O
A	O
15	O
-	O
mer	O
linear	O
peptide	O
1	O
was	O
shown	O
to	O
block	O
IL	O
-	O
17A	O
/	O
IL	O
-	O
17RA	O
binding	O
with	O
an	O
IC50	O
of	O
80	O
nM	O
in	O
the	O
competition	O
ELISA	O
assay	O
(	O
Table	O
1	O
).	O
	O
This	O
peptide	O
was	O
then	O
tested	O
in	O
a	O
cell	O
-	O
based	O
functional	O
assay	O
wherein	O
production	O
of	O
GRO	O
-	O
α	O
in	O
BJ	O
human	O
fibroblast	O
cells	O
was	O
measured	O
as	O
a	O
function	O
of	O
IL	O
-	O
17A	O
stimulation	O
using	O
1	O
ng	O
/	O
ml	O
IL	O
-	O
17A	O
.	O
	O
Peptide	O
1	O
was	O
found	O
to	O
be	O
active	O
in	O
this	O
functional	O
assay	O
with	O
an	O
IC50	O
of	O
370	O
nM	O
.	O
	O
Optimization	O
of	O
IL	O
-	O
17A	O
peptide	O
inhibitors	O
	O
A	O
SAR	O
campaign	O
was	O
undertaken	O
to	O
improve	O
the	O
potency	O
of	O
peptide	O
1	O
.	O
	O
When	O
alanine	O
was	O
already	O
present	O
(	O
positions	O
7	O
and	O
15	O
),	O
substitution	O
was	O
made	O
with	O
lysine	O
(	O
Table	O
1	O
,	O
peptides	O
3	O
–	O
17	O
).	O
	O
Positions	O
1	O
,	O
2	O
,	O
4	O
,	O
5	O
,	O
7	O
,	O
14	O
and	O
15	O
were	O
shown	O
to	O
be	O
amenable	O
to	O
substitution	O
without	O
significant	O
loss	O
(	O
less	O
than	O
3	O
-	O
fold	O
)	O
of	O
binding	O
affinity	O
as	O
measured	O
by	O
the	O
IL	O
-	O
17A	O
/	O
IL	O
-	O
17RA	O
competition	O
ELISA	O
.	O
	O
In	O
order	O
to	O
rapidly	O
evaluate	O
the	O
effects	O
of	O
substitution	O
of	O
natural	O
amino	O
acids	O
at	O
tolerant	O
positions	O
identified	O
by	O
the	O
alanine	O
scan	O
,	O
the	O
lead	O
sequence	O
was	O
subjected	O
to	O
site	O
-	O
specific	O
saturation	O
mutagenesis	O
using	O
MBP	O
.	O
	O
Each	O
of	O
the	O
seven	O
positions	O
identified	O
by	O
the	O
alanine	O
scan	O
was	O
individually	O
modified	O
while	O
keeping	O
the	O
rest	O
of	O
the	O
sequence	O
constant	O
.	O
	O
Peptides	O
with	O
beneficial	O
point	O
mutations	O
at	O
positions	O
2	O
,	O
5	O
,	O
and	O
14	O
were	O
synthesized	O
and	O
evaluated	O
in	O
the	O
competition	O
ELISA	O
(	O
Table	O
1	O
).	O
	O
Two	O
of	O
the	O
changes	O
,	O
V2H	B-mutant
(	O
18	O
)	O
or	O
V2T	B-mutant
(	O
21	O
)	O
displayed	O
improved	O
binding	O
in	O
the	O
competition	O
ELISA	O
.	O
	O
Introduction	O
of	O
a	O
methionine	O
(	O
27	O
)	O
or	O
a	O
carboxamide	O
(	O
28	O
and	O
29	O
)	O
at	O
position	O
14	O
was	O
shown	O
to	O
improve	O
the	O
binding	O
affinity	O
of	O
the	O
lead	O
peptide	O
.	O
	O
In	O
general	O
,	O
there	O
was	O
good	O
agreement	O
between	O
the	O
respective	O
binding	O
affinities	O
of	O
the	O
synthesized	O
peptides	O
and	O
their	O
MBP	O
fusion	O
counterparts	O
,	O
except	O
for	O
substitution	O
of	O
valine	O
at	O
position	O
2	O
to	O
a	O
tryptophan	O
(	O
22	O
),	O
which	O
resulted	O
in	O
a	O
fivefold	O
loss	O
of	O
affinity	O
,	O
for	O
the	O
free	O
peptide	O
when	O
compared	O
with	O
the	O
MBP	O
fusion	O
.	O
	O
Combining	O
the	O
key	O
amino	O
-	O
acid	O
residues	O
identified	O
by	O
SAR	O
into	O
a	O
single	O
peptide	O
sequence	O
resulted	O
in	O
peptide	O
30	O
,	O
named	O
high	O
affinity	O
peptide	O
(	O
HAP	O
),	O
that	O
was	O
found	O
to	O
inhibit	O
IL	O
-	O
17A	O
signaling	O
in	O
a	O
BJ	O
human	O
fibroblast	O
cell	O
assay	O
with	O
an	O
IC50	O
of	O
17	O
nM	O
,	O
a	O
more	O
than	O
20	O
-	O
fold	O
improvement	O
over	O
the	O
phage	O
peptide	O
1	O
(	O
Table	O
2	O
and	O
Supplementary	O
Figure	O
S2	O
).	O
	O
We	O
also	O
examined	O
the	O
effect	O
of	O
removing	O
the	O
acetyl	O
group	O
at	O
the	O
N	O
-	O
terminus	O
of	O
HAP	O
(	O
which	O
is	O
present	O
in	O
all	O
the	O
peptides	O
made	O
,	O
see	O
Supplementary	O
Material	O
).	O
	O
The	O
un	O
-	O
capped	O
peptide	O
(	O
31	O
)	O
had	O
an	O
IC50	O
of	O
420	O
nM	O
in	O
the	O
cell	O
-	O
based	O
assay	O
.	O
	O
The	O
loss	O
of	O
cellular	O
activity	O
of	O
31	O
was	O
most	O
likely	O
due	O
to	O
the	O
degradation	O
of	O
the	O
N	O
-	O
terminus	O
of	O
31	O
,	O
since	O
peptide	O
31	O
was	O
shown	O
to	O
be	O
able	O
to	O
bind	O
to	O
IL	O
-	O
17A	O
with	O
similar	O
affinity	O
as	O
HAP	O
itself	O
.	O
	O
Furthermore	O
,	O
our	O
previous	O
work	O
had	O
reported	O
that	O
in	O
antibody	O
fusions	O
the	O
uncapped	O
peptide	O
was	O
degraded	O
under	O
cell	O
assay	O
conditions	O
with	O
removal	O
of	O
the	O
first	O
1	O
-	O
3	O
residues	O
to	O
inactive	O
products	O
with	O
the	O
same	O
N	O
-	O
terminal	O
sequences	O
as	O
peptides	O
32	O
–	O
34	O
.	O
	O
C	O
-	O
terminal	O
truncations	O
showed	O
a	O
more	O
gradual	O
reduction	O
in	O
activity	O
(	O
35	O
–	O
37	O
;	O
Table	O
2	O
).	O
	O
After	O
deletion	O
of	O
three	O
amino	O
acids	O
from	O
the	O
C	O
-	O
terminal	O
end	O
(	O
37	O
),	O
the	O
peptide	O
is	O
no	O
longer	O
active	O
.	O
	O
We	O
reasoned	O
that	O
since	O
the	O
IL	O
-	O
17A	O
protein	O
is	O
almost	O
exclusively	O
present	O
in	O
a	O
dimeric	O
form	O
,	O
dimerizing	O
the	O
IL	O
-	O
17A	O
binding	O
peptides	O
could	O
result	O
in	O
an	O
improvement	O
in	O
binding	O
affinity	O
and	O
inhibitory	O
activity	O
.	O
	O
Homodimers	O
of	O
HAP	O
were	O
made	O
through	O
attachment	O
of	O
polyethylene	O
glycol	O
(	O
PEG	O
)	O
spacers	O
of	O
different	O
lengths	O
at	O
amino	O
acids	O
4	O
,	O
7	O
and	O
14	O
,	O
as	O
these	O
positions	O
were	O
identified	O
in	O
the	O
alanine	O
scan	O
analysis	O
as	O
not	O
contributing	O
significantly	O
to	O
the	O
activity	O
,	O
and	O
at	O
each	O
N	O
-	O
terminus	O
(	O
Supplementary	O
Table	O
S2	O
).	O
	O
Due	O
to	O
the	O
high	O
reactivity	O
of	O
the	O
pentafluoroester	O
(	O
PFP	O
)	O
group	O
used	O
as	O
the	O
activating	O
group	O
in	O
the	O
PEG	O
,	O
the	O
histidine	O
at	O
position	O
2	O
and	O
the	O
lysine	O
at	O
position	O
15	O
were	O
replaced	O
with	O
threonine	O
and	O
dimethyllysine	O
respectively	O
to	O
prevent	O
formation	O
of	O
side	O
products	O
,	O
which	O
resulted	O
in	O
peptide	O
38	O
that	O
was	O
comparable	O
in	O
activity	O
with	O
HAP	O
.	O
	O
This	O
exercise	O
revealed	O
that	O
several	O
dimeric	O
peptides	O
with	O
the	O
longer	O
PEG21	O
spacer	O
were	O
significantly	O
more	O
potent	O
than	O
the	O
monomer	O
peptide	O
in	O
the	O
cell	O
-	O
based	O
assay	O
(	O
Supplementary	O
Table	O
S2	O
).	O
	O
The	O
species	O
cross	O
-	O
reactivity	O
of	O
the	O
dimeric	O
peptide	O
45	O
and	O
HAP	O
were	O
assessed	O
in	O
a	O
murine	O
functional	O
cell	O
assay	O
using	O
15	O
ng	O
/	O
ml	O
murine	O
IL	O
-	O
17A	O
.	O
	O
Peptide	O
45	O
blocked	O
the	O
receptor	O
binding	O
of	O
murine	O
IL	O
-	O
17A	O
although	O
with	O
potency	O
two	O
orders	O
of	O
magnitude	O
weaker	O
than	O
that	O
observed	O
against	O
human	O
IL	O
-	O
17A	O
(	O
IC50	O
=	O
41	O
nM	O
vs	O
IC50	O
=	O
0	O
.	O
1	O
nM	O
,	O
respectively	O
).	O
	O
The	O
monomer	O
HAP	O
was	O
much	O
weaker	O
(	O
IC50	O
>	O
1	O
μM	O
)	O
in	O
inhibiting	O
murine	O
IL	O
-	O
17A	O
signaling	O
(	O
Supplementary	O
Figure	O
S4	O
).	O
	O
Although	O
the	O
dimeric	O
peptide	O
45	O
is	O
much	O
more	O
potent	O
than	O
HAP	O
in	O
the	O
cell	O
-	O
based	O
assay	O
,	O
in	O
subsequent	O
studies	O
we	O
decided	O
to	O
focus	O
our	O
efforts	O
solely	O
on	O
characterizations	O
of	O
the	O
monomeric	O
peptide	O
HAP	O
in	O
hopes	O
to	O
identify	O
smaller	O
peptide	O
inhibitors	O
containing	O
the	O
best	O
minimal	O
functional	O
group	O
.	O
	O
HAP	O
binds	O
to	O
immobilized	O
human	O
IL	O
-	O
17A	O
homodimer	O
tightly	O
(	O
Table	O
3	O
).	O
	O
It	O
has	O
slightly	O
weaker	O
affinity	O
for	O
human	O
IL	O
-	O
17A	O
/	O
F	O
heterodimer	O
and	O
>	O
10	O
fold	O
weaker	O
affinity	O
for	O
mouse	O
IL	O
-	O
17A	O
(	O
Table	O
3	O
).	O
	O
HAP	O
does	O
not	O
show	O
significant	O
binding	O
to	O
immobilized	O
human	O
IL	O
-	O
17F	O
homodimer	O
or	O
IL	O
-	O
17RA	O
at	O
concentrations	O
up	O
to	O
100	O
nM	O
.	O
	O
Additionally	O
,	O
we	O
investigated	O
the	O
antagonism	O
of	O
the	O
human	O
IL	O
-	O
17A	O
/	O
IL	O
-	O
17RA	O
interaction	O
by	O
HAP	O
using	O
orthogonal	O
methods	O
including	O
SPR	O
and	O
Förster	O
resonance	O
energy	O
transfer	O
(	O
FRET	O
)	O
competition	O
assays	O
(	O
Fig	O
.	O
1B	O
,	O
C	O
).	O
	O
In	O
both	O
assays	O
,	O
incubation	O
of	O
IL	O
-	O
17A	O
with	O
HAP	O
effectively	O
blocks	O
the	O
binding	O
of	O
IL	O
-	O
17A	O
to	O
immobilized	O
IL	O
-	O
17RA	O
with	O
similar	O
sub	O
-	O
nM	O
IC50	O
(	O
Table	O
3	O
).	O
	O
While	O
either	O
IL	O
-	O
17A	O
or	O
TNF	O
-	O
α	O
alone	O
can	O
stimulate	O
the	O
release	O
of	O
multiple	O
inflammatory	O
cytokines	O
,	O
when	O
acting	O
together	O
they	O
can	O
synergistically	O
enhance	O
each	O
other	O
’	O
s	O
effects	O
(	O
Supplementary	O
Figure	O
S5	O
).	O
	O
These	O
integrative	O
responses	O
to	O
IL	O
-	O
17A	O
and	O
TNF	O
-	O
α	O
in	O
human	O
keratinocytes	O
have	O
been	O
reported	O
to	O
account	O
for	O
key	O
inflammatory	O
pathogenic	O
circuits	O
in	O
psoriasis	O
.	O
	O
Thus	O
,	O
we	O
chose	O
to	O
study	O
HAP	O
’	O
s	O
efficacy	O
in	O
blocking	O
the	O
production	O
of	O
IL	O
-	O
8	O
,	O
IL	O
-	O
6	O
and	O
CCL	O
-	O
20	O
by	O
primary	O
human	O
keratinocytes	O
stimulated	O
by	O
IL	O
-	O
17A	O
in	O
the	O
presence	O
of	O
TNF	O
-	O
α	O
,	O
an	O
assay	O
which	O
may	O
be	O
more	O
disease	O
-	O
relevant	O
.	O
	O
HAP	O
inhibits	O
the	O
production	O
of	O
all	O
three	O
cytokines	O
in	O
a	O
dose	O
-	O
dependent	O
fashion	O
(	O
Fig	O
.	O
1D	O
).	O
	O
Significantly	O
,	O
the	O
baseline	O
levels	O
of	O
IL	O
-	O
8	O
,	O
IL	O
-	O
6	O
and	O
CCL	O
-	O
20	O
stimulated	O
by	O
TNF	O
-	O
α	O
alone	O
are	O
not	O
inhibited	O
by	O
HAP	O
,	O
further	O
indicating	O
the	O
selectivity	O
of	O
HAP	O
(	O
Fig	O
.	O
1D	O
).	O
	O
Such	O
pharmacological	O
selectivity	O
may	O
be	O
important	O
to	O
suppress	O
inflammatory	O
pathogenic	O
circuits	O
in	O
psoriasis	O
,	O
while	O
sparing	O
the	O
anti	O
-	O
infectious	O
immune	O
responses	O
produced	O
by	O
TNF	O
-	O
α	O
.	O
	O
As	O
a	O
reference	O
,	O
a	O
commercial	O
anti	O
-	O
IL	O
-	O
17A	O
antibody	O
(	O
R	O
&	O
D	O
Systems	O
)	O
inhibits	O
the	O
production	O
of	O
IL	O
-	O
8	O
with	O
an	O
IC50	O
of	O
13	O
(±	O
6	O
)	O
nM	O
(	O
N	O
=	O
3	O
).	O
	O
Indeed	O
,	O
the	O
IC50	O
was	O
14	O
(±	O
9	O
)	O
nM	O
(	O
N	O
=	O
12	O
)	O
for	O
HAP	O
inhibition	O
of	O
IL	O
-	O
8	O
production	O
when	O
only	O
5	O
ng	O
/	O
ml	O
IL	O
-	O
17A	O
was	O
used	O
in	O
this	O
assay	O
.	O
	O
Similar	O
to	O
keratinocytes	O
assay	O
results	O
,	O
while	O
HAP	O
inhibits	O
IL	O
-	O
17A	O
stimulated	O
IL	O
-	O
6	O
production	O
by	O
BJ	O
human	O
fibroblast	O
potently	O
(	O
IC50	O
of	O
17	O
nM	O
),	O
it	O
does	O
not	O
inhibit	O
TNF	O
-	O
α	O
stimulated	O
IL	O
-	O
6	O
production	O
at	O
concentrations	O
up	O
to	O
10	O
μM	O
(	O
Supplementary	O
Figure	O
S2	O
).	O
	O
Extensive	O
crystallization	O
trials	O
,	O
either	O
by	O
co	O
-	O
crystallization	O
or	O
by	O
soaking	O
HAP	O
into	O
preformed	O
apo	O
IL	O
-	O
17A	O
crystals	O
,	O
failed	O
to	O
lead	O
to	O
an	O
IL	O
-	O
17A	O
/	O
HAP	O
complex	O
crystals	O
.	O
	O
We	O
theorized	O
that	O
HAP	O
binding	O
induced	O
large	O
conformational	O
changes	O
in	O
IL	O
-	O
17A	O
that	O
led	O
to	O
the	O
difficulty	O
of	O
getting	O
an	O
IL	O
-	O
17A	O
/	O
HAP	O
binary	O
complex	O
crystal	O
.	O
	O
We	O
hypothesized	O
that	O
HAP	O
may	O
target	O
the	O
N	O
-	O
terminal	O
of	O
IL	O
-	O
17A	O
which	O
is	O
known	O
to	O
be	O
more	O
flexible	O
than	O
its	O
C	O
-	O
terminal	O
and	O
conformational	O
changes	O
needed	O
for	O
HAP	O
binding	O
may	O
be	O
more	O
likely	O
there	O
.	O
	O
We	O
designed	O
an	O
antibody	O
Fab	O
known	O
to	O
target	O
the	O
C	O
-	O
terminal	O
half	O
of	O
IL	O
-	O
17A	O
based	O
on	O
a	O
published	O
IL	O
-	O
17A	O
/	O
Fab	O
complex	O
crystal	O
structure	O
,	O
and	O
produced	O
it	O
in	O
HEK293	O
cells	O
.	O
	O
In	O
an	O
SPR	O
assay	O
HAP	O
and	O
this	O
Fab	O
were	O
able	O
to	O
co	O
-	O
bind	O
IL	O
-	O
17A	O
without	O
large	O
changes	O
in	O
their	O
binding	O
affinities	O
and	O
kinetics	O
,	O
confirming	O
our	O
hypothesis	O
(	O
Supplementary	O
Figure	O
S6	O
).	O
	O
These	O
were	O
,	O
respectively	O
,	O
a	O
presumably	O
more	O
homogeneous	O
form	O
of	O
IL	O
-	O
17A	O
that	O
lacked	O
the	O
disordered	O
N	O
-	O
terminal	O
peptide	O
and	O
a	O
full	O
-	O
length	O
form	O
of	O
the	O
cytokine	O
with	O
a	O
full	O
complement	O
of	O
disulfide	O
bonds	O
.	O
	O
Both	O
complexes	O
crystallized	O
in	O
the	O
space	O
group	O
of	O
P321	O
,	O
with	O
half	O
the	O
complex	O
(	O
1	O
Fab	O
/	O
1	O
IL	O
-	O
17A	O
monomer	O
/	O
1	O
HAP	O
)	O
in	O
the	O
asymmetric	O
unit	O
.	O
	O
The	O
intact	O
complex	O
can	O
be	O
generated	O
by	O
applying	O
crystallographic	O
2	O
-	O
fold	O
symmetry	O
.	O
	O
Electron	O
densities	O
for	O
HAP	O
residues	O
Ile1	O
-	O
Asn14	O
were	O
readily	O
interpretable	O
with	O
the	O
exception	O
of	O
Lys15	O
,	O
which	O
is	O
disordered	O
.	O
	O
When	O
considering	O
the	O
protein	O
,	O
the	O
complex	O
structure	O
containing	O
the	O
full	O
length	O
IL	O
-	O
17A	O
is	O
identical	O
to	O
that	O
of	O
the	O
truncated	O
IL	O
-	O
17A	O
,	O
with	O
the	O
exception	O
of	O
Cys106	O
(	O
Ser106	O
in	O
the	O
truncated	O
IL	O
-	O
17A	O
),	O
which	O
is	O
disordered	O
.	O
	O
Cys106	O
is	O
covalently	O
linked	O
to	O
Cys10	O
that	O
resides	O
in	O
the	O
disordered	O
N	O
-	O
terminal	O
peptide	O
in	O
the	O
full	O
length	O
IL	O
-	O
17A	O
.	O
	O
Overall	O
structure	O
of	O
Fab	O
/	O
IL	O
-	O
17A	O
/	O
HAP	O
complex	O
	O
In	O
a	O
similar	O
manner	O
to	O
the	O
published	O
structure	O
of	O
Fab	O
/	O
IL	O
-	O
17A	O
complex	O
,	O
two	O
Fab	O
molecules	O
bind	O
symmetrically	O
to	O
the	O
C	O
-	O
terminal	O
of	O
the	O
cytokine	O
dimer	O
,	O
interacting	O
with	O
epitopes	O
from	O
both	O
monomers	O
(	O
Fig	O
.	O
2A	O
).	O
	O
Based	O
on	O
disclosed	O
epitopes	O
of	O
Secukinumab	O
and	O
Ixekizumab	O
,	O
HAP	O
binds	O
to	O
IL	O
-	O
17A	O
at	O
an	O
area	O
that	O
is	O
also	O
different	O
from	O
those	O
of	O
those	O
two	O
antibodies	O
.	O
	O
The	O
N	O
-	O
terminal	O
5	O
residues	O
of	O
HAP	O
,	O
1IHVTI	O
,	O
form	O
an	O
amphipathic	O
β	O
-	O
strand	O
that	O
inserts	O
between	O
β	O
-	O
strand	O
4	O
of	O
one	O
IL	O
-	O
17A	O
monomer	O
and	O
β	O
-	O
strand	O
0	O
(	O
the	O
first	O
ordered	O
peptide	O
of	O
IL	O
-	O
17A	O
)	O
of	O
the	O
second	O
monomer	O
.	O
	O
This	O
β	O
-	O
strand	O
is	O
parallel	O
to	O
both	O
strands	O
0	O
and	O
4	O
(	O
Fig	O
.	O
3B	O
).	O
	O
Strands	O
0	O
of	O
two	O
IL	O
-	O
17A	O
monomer	O
are	O
antiparallel	O
,	O
as	O
appeared	O
in	O
other	O
IL	O
-	O
17A	O
structures	O
.	O
	O
As	O
a	O
comparison	O
,	O
an	O
IL	O
-	O
17A	O
/	O
IL	O
-	O
17RA	O
complex	O
structure	O
(	O
PDB	O
code	O
4HSA	O
)	O
is	O
also	O
shown	O
with	O
IL	O
-	O
17A	O
in	O
the	O
same	O
orientation	O
(	O
Fig	O
.	O
2C	O
).	O
	O
IL	O
-	O
17RA	O
binds	O
IL	O
-	O
17A	O
at	O
three	O
regions	O
on	O
the	O
IL	O
-	O
17A	O
homodimer	O
.	O
	O
HAP	O
binds	O
IL	O
-	O
17A	O
at	O
region	O
I	O
.	O
Region	O
I	O
is	O
formed	O
by	O
residues	O
at	O
the	O
ends	O
of	O
β	O
strands	O
0	O
and	O
4	O
,	O
and	O
from	O
loops	O
1	O
–	O
2	O
and	O
3	O
–	O
4	O
of	O
IL	O
-	O
17A	O
(	O
Fig	O
.	O
2	O
).	O
	O
The	O
most	O
significant	O
interactions	O
between	O
the	O
α	O
helix	O
of	O
HAP	O
and	O
IL	O
-	O
17A	O
involve	O
Trp12	O
of	O
HAP	O
,	O
which	O
binds	O
in	O
a	O
hydrophobic	O
pocket	O
in	O
IL	O
-	O
17A	O
formed	O
by	O
the	O
side	O
chains	O
of	O
Phe110	O
,	O
Tyr62	O
,	O
Pro59	O
and	O
the	O
hydrophobic	O
portion	O
of	O
the	O
Arg101	O
side	O
chain	O
(	O
Fig	O
.	O
3A	O
).	O
	O
The	O
Trp12	O
side	O
chain	O
of	O
HAP	O
donates	O
a	O
hydrogen	O
bond	O
to	O
the	O
main	O
chain	O
oxygen	O
of	O
Pro69	O
of	O
IL	O
-	O
17A	O
.	O
	O
The	O
positively	O
charged	O
Arg101	O
side	O
chain	O
of	O
the	O
IL	O
-	O
17A	O
engages	O
in	O
a	O
charge	O
-	O
helix	O
dipole	O
interaction	O
with	O
the	O
main	O
chain	O
oxygen	O
of	O
Trp12	O
.	O
	O
Additionally	O
,	O
Leu9	O
and	O
Ile13	O
of	O
the	O
HAP	O
have	O
hydrophobic	O
interactions	O
with	O
IL	O
-	O
17A	O
,	O
and	O
the	O
Asp8	O
side	O
chain	O
has	O
hydrogen	O
bond	O
and	O
ion	O
pair	O
interactions	O
with	O
Tyr62	O
and	O
Lys114	O
of	O
IL	O
-	O
17A	O
,	O
respectively	O
.	O
	O
In	O
region	O
I	O
,	O
an	O
IL	O
-	O
17RA	O
peptide	O
interacts	O
with	O
IL	O
-	O
17A	O
in	O
a	O
very	O
similar	O
fashion	O
to	O
the	O
α	O
-	O
helix	O
of	O
HAP	O
.	O
	O
The	O
IL	O
-	O
17RA	O
peptide	O
has	O
sequences	O
of	O
27LDDSWI	O
,	O
and	O
part	O
of	O
the	O
peptide	O
is	O
also	O
α	O
-	O
helical	O
(	O
Fig	O
.	O
3B	O
).	O
	O
Leu7	O
,	O
Trp31	O
and	O
Ile32	O
of	O
IL	O
-	O
17RA	O
interact	O
very	O
similarly	O
with	O
the	O
same	O
residues	O
of	O
IL	O
-	O
17A	O
as	O
Leu9	O
,	O
Trp12	O
and	O
Ile13	O
of	O
HAP	O
(	O
Fig	O
.	O
3B	O
).	O
	O
In	O
this	O
sense	O
,	O
the	O
α	O
-	O
helix	O
of	O
HAP	O
with	O
a	O
sequence	O
of	O
9LWDWI	O
is	O
a	O
good	O
mimetic	O
of	O
the	O
27LDDSWI	O
peptide	O
of	O
IL	O
-	O
17RA	O
.	O
	O
The	O
β	O
-	O
strand	O
of	O
HAP	O
has	O
no	O
equivalent	O
in	O
IL	O
-	O
17RA	O
.	O
	O
The	O
amphipathic	O
β	O
-	O
strand	O
of	O
HAP	O
orients	O
the	O
hydrophilic	O
side	O
chains	O
of	O
His2	O
and	O
Thr4	O
outwards	O
,	O
and	O
the	O
hydrophobic	O
side	O
chains	O
of	O
Ile1	O
,	O
Val3	O
and	O
Ile5	O
inward	O
(	O
Fig	O
.	O
3A	O
).	O
	O
β	O
-	O
strand	O
0	O
in	O
IL	O
-	O
17A	O
is	O
also	O
amphipathic	O
with	O
the	O
sequence	O
of	O
21TVMVNLNI	O
.	O
	O
In	O
all	O
IL	O
-	O
17A	O
structures	O
obtained	O
to	O
date	O
,	O
β	O
-	O
strand	O
0	O
orients	O
the	O
hydrophilic	O
side	O
chains	O
of	O
Thr21	O
,	O
Asn25	O
and	O
Asn27	O
outward	O
,	O
and	O
the	O
hydrophobic	O
side	O
chains	O
of	O
Val22	O
,	O
Val24	O
,	O
Leu26	O
and	O
Ile28	O
inward	O
.	O
	O
The	O
binding	O
pocket	O
occupied	O
by	O
either	O
Trp12	O
of	O
HAP	O
or	O
Trp31	O
of	O
IL	O
-	O
17RA	O
is	O
not	O
formed	O
in	O
the	O
apo	O
IL	O
-	O
17A	O
structure	O
(	O
Fig	O
.	O
3C	O
).	O
	O
Particularly	O
for	O
HAP	O
,	O
β	O
-	O
strands	O
0	O
have	O
to	O
shift	O
out	O
of	O
the	O
hydrophobic	O
cleft	O
formed	O
by	O
the	O
main	O
body	O
of	O
the	O
IL	O
-	O
17A	O
by	O
as	O
much	O
as	O
10	O
Å	O
between	O
Cα	O
atoms	O
(	O
Fig	O
.	O
3C	O
).	O
	O
Disruptions	O
of	O
the	O
apo	O
IL	O
-	O
17A	O
structure	O
by	O
HAP	O
binding	O
are	O
apparently	O
compensated	O
for	O
by	O
formation	O
of	O
the	O
new	O
interactions	O
that	O
involve	O
almost	O
the	O
entire	O
HAP	O
molecule	O
(	O
Fig	O
.	O
3B	O
).	O
	O
The	O
IL	O
-	O
17A	O
/	O
HAP	O
complex	O
structure	O
obtained	O
is	O
very	O
consistent	O
with	O
the	O
observed	O
SAR	O
of	O
our	O
identified	O
peptide	O
inhibitors	O
,	O
explaining	O
well	O
how	O
the	O
evolution	O
of	O
the	O
initial	O
phage	O
peptide	O
1	O
to	O
HAP	O
and	O
45	O
improved	O
its	O
potency	O
(	O
Supplementary	O
Figure	O
S7	O
).	O
	O
The	O
important	O
interactions	O
involving	O
Trp12	O
of	O
HAP	O
explain	O
the	O
>	O
90	O
times	O
drop	O
in	O
potency	O
of	O
the	O
W12A	B-mutant
variant	O
(	O
6	O
vs	O
1	O
,	O
Table	O
1	O
).	O
	O
The	O
amphipathic	O
nature	O
of	O
the	O
HAP	O
β	O
-	O
strand	O
explains	O
the	O
preference	O
of	O
the	O
hydrophilic	O
residues	O
at	O
the	O
2	O
and	O
4	O
positions	O
of	O
peptides	O
(	O
14	O
,	O
18	O
,	O
19	O
,	O
21	O
and	O
23	O
vs	O
1	O
and	O
22	O
,	O
Table	O
1	O
).	O
	O
All	O
N	O
-	O
terminal	O
residues	O
of	O
HAP	O
are	O
part	O
of	O
the	O
β	O
-	O
sheet	O
with	O
β	O
-	O
stands	O
0	O
and	O
4	O
of	O
IL	O
-	O
17A	O
,	O
which	O
explains	O
why	O
removal	O
of	O
the	O
first	O
1	O
–	O
3	O
residues	O
completely	O
abolishes	O
the	O
ability	O
of	O
HAP	O
to	O
block	O
IL	O
-	O
17A	O
cell	O
signaling	O
(	O
31	O
,	O
32	O
and	O
33	O
,	O
Table	O
2	O
).	O
	O
Each	O
peptide	O
monomer	O
in	O
45	O
may	O
not	O
necessarily	O
be	O
more	O
potent	O
than	O
HAP	O
,	O
but	O
two	O
monomer	O
peptides	O
within	O
the	O
same	O
molecule	O
that	O
can	O
simultaneously	O
bind	O
to	O
IL	O
-	O
17A	O
can	O
greatly	O
improve	O
its	O
potency	O
due	O
to	O
avidity	O
effects	O
.	O
	O
HAP	O
targets	O
region	O
I	O
of	O
IL	O
-	O
17A	O
,	O
an	O
area	O
that	O
has	O
the	O
least	O
sequence	O
conservation	O
in	O
IL	O
-	O
17	O
cytokines	O
.	O
	O
This	O
lack	O
of	O
sequence	O
conservation	O
in	O
the	O
HAP	O
binding	O
site	O
explains	O
the	O
observed	O
specificity	O
of	O
HAP	O
binding	O
to	O
human	O
IL	O
-	O
17A	O
.	O
	O
This	O
Phe	O
-	O
Phe	O
motif	O
is	O
missing	O
in	O
IL	O
-	O
17A	O
.	O
	O
Sequence	O
alignments	O
between	O
human	O
and	O
mouse	O
IL	O
-	O
17A	O
indicated	O
that	O
among	O
IL	O
-	O
17A	O
residues	O
that	O
interacting	O
with	O
HAP	O
,	O
majority	O
differences	O
occur	O
in	O
strand	O
0	O
of	O
IL	O
-	O
17A	O
which	O
interacts	O
with	O
the	O
N	O
-	O
terminal	O
β	O
-	O
strand	O
of	O
HAP	O
.	O
	O
In	O
human	O
IL	O
-	O
17A	O
the	O
sequences	O
are	O
21TVMVNLNI	O
,	O
and	O
in	O
mouse	O
they	O
are	O
21NVKVNLKV	O
.	O
	O
Using	O
a	O
combination	O
of	O
phage	O
display	O
and	O
SAR	O
we	O
have	O
discovered	O
novel	O
peptides	O
that	O
are	O
IL	O
-	O
17A	O
antagonists	O
.	O
	O
One	O
of	O
those	O
peptides	O
,	O
HAP	O
,	O
also	O
shows	O
activity	O
in	O
inhibiting	O
the	O
production	O
of	O
multiple	O
inflammatory	O
cytokines	O
by	O
primary	O
human	O
keratinocytes	O
stimulated	O
by	O
IL	O
-	O
17A	O
and	O
TNF	O
-	O
α	O
,	O
a	O
disease	O
relevant	O
-	O
model	O
.	O
	O
With	O
two	O
HAP	O
molecules	O
covering	O
both	O
faces	O
of	O
the	O
IL	O
-	O
17A	O
dimer	O
,	O
HAP	O
can	O
block	O
IL	O
-	O
17RA	O
approaching	O
from	O
either	O
face	O
.	O
	O
To	O
form	O
the	O
1	O
:	O
2	O
complex	O
observed	O
in	O
crystal	O
structure	O
,	O
it	O
is	O
important	O
that	O
there	O
is	O
no	O
strong	O
negative	O
cooperativity	O
in	O
the	O
binding	O
of	O
two	O
HAP	O
molecules	O
.	O
	O
In	O
fact	O
,	O
in	O
native	O
electrospray	O
ionization	O
mass	O
spectrometry	O
analysis	O
only	O
1	O
:	O
2	O
IL	O
-	O
17A	O
/	O
HAP	O
complex	O
was	O
observed	O
even	O
when	O
IL	O
-	O
17A	O
was	O
in	O
excess	O
(	O
Supplementary	O
Figure	O
S8	O
),	O
indicating	O
a	O
positive	O
binding	O
cooperativity	O
that	O
favors	O
inhibition	O
of	O
IL	O
-	O
17RA	O
binding	O
by	O
HAP	O
.	O
	O
Due	O
to	O
the	O
discontinuous	O
nature	O
of	O
the	O
IL	O
-	O
17A	O
/	O
IL	O
-	O
17RA	O
binding	O
interface	O
,	O
it	O
is	O
classified	O
as	O
having	O
tertiary	O
structural	O
epitopes	O
on	O
both	O
binding	O
partners	O
,	O
and	O
is	O
therefore	O
hard	O
to	O
target	O
using	O
small	O
molecules	O
.	O
	O
Our	O
studies	O
of	O
HAP	O
demonstrated	O
an	O
uncommon	O
mode	O
of	O
action	O
for	O
a	O
peptide	O
in	O
inhibiting	O
such	O
a	O
difficult	O
protein	O
-	O
protein	O
interaction	O
target	O
,	O
and	O
suggest	O
further	O
possible	O
improvements	O
in	O
its	O
binding	O
potency	O
.	O
	O
Homo	O
-	O
dimerization	O
of	O
HAP	O
(	O
45	O
)	O
achieved	O
sub	O
-	O
nanomolar	O
potency	O
against	O
human	O
IL	O
-	O
17A	O
in	O
cell	O
assay	O
.	O
	O
In	O
the	O
crystal	O
structure	O
,	O
the	O
distance	O
between	O
the	O
carbonyl	O
of	O
Asn14	O
of	O
one	O
HAP	O
molecule	O
and	O
the	O
N	O
-	O
terminus	O
of	O
the	O
second	O
is	O
only	O
15	O
.	O
7	O
Å	O
,	O
suggesting	O
the	O
potential	O
for	O
more	O
potent	O
dimeric	O
peptides	O
to	O
be	O
designed	O
by	O
using	O
linkers	O
of	O
different	O
lengths	O
at	O
different	O
positions	O
.	O
	O
Another	O
direction	O
of	O
improving	O
HAP	O
is	O
by	O
reducing	O
its	O
size	O
.	O
	O
In	O
summary	O
,	O
these	O
peptide	O
-	O
based	O
anti	O
-	O
IL	O
-	O
17A	O
modalities	O
could	O
be	O
further	O
developed	O
as	O
alternative	O
therapeutic	O
options	O
to	O
the	O
reported	O
monoclonal	O
antibodies	O
.	O
	O
We	O
are	O
also	O
very	O
interested	O
in	O
finding	O
non	O
-	O
peptidic	O
small	O
molecule	O
IL	O
-	O
17A	O
antagonists	O
,	O
and	O
HAP	O
can	O
be	O
used	O
as	O
an	O
excellent	O
tool	O
peptide	O
.	O
	O
The	O
strategy	O
utilized	O
in	O
generating	O
the	O
complex	O
structures	O
of	O
HAP	O
may	O
also	O
be	O
useful	O
for	O
enabling	O
structure	O
based	O
design	O
of	O
some	O
known	O
small	O
molecule	O
IL	O
-	O
17A	O
antagonists	O
.	O
	O
Binding	O
of	O
HAP	O
to	O
IL	O
-	O
17A	O
and	O
inhibition	O
of	O
IL	O
-	O
17A	O
/	O
IL	O
-	O
17RA	O
are	O
measured	O
by	O
SPR	O
,	O
FRET	O
and	O
cell	O
-	O
based	O
assays	O
.	O
	O
(	O
A	O
)	O
Typical	O
SPR	O
sensorgrams	O
(	O
black	O
)	O
of	O
HAP	O
at	O
indicated	O
concentrations	O
binding	O
to	O
biotinylated	O
human	O
IL	O
-	O
17A	O
immobilized	O
on	O
a	O
streptavidin	O
chip	O
surface	O
,	O
fitted	O
with	O
single	O
site	O
binding	O
model	O
curves	O
(	O
red	O
).	O
	O
Kinetic	O
parameters	O
(	O
ka	O
,	O
kd	O
)	O
were	O
obtained	O
by	O
a	O
global	O
fit	O
using	O
three	O
concentrations	O
in	O
triplicate	O
.	O
	O
Data	O
are	O
mean	O
and	O
error	O
bars	O
of	O
+/−	O
standard	O
deviation	O
of	O
three	O
measurements	O
.	O
(	O
C	O
)	O
Inhibition	O
of	O
IL	O
-	O
17A	O
and	O
IL	O
-	O
17RA	O
binding	O
by	O
HAP	O
measured	O
by	O
FRET	O
assay	O
.	O
	O
Data	O
are	O
mean	O
and	O
error	O
bars	O
of	O
+/−	O
standard	O
deviation	O
from	O
299	O
experiments	O
,	O
each	O
performed	O
in	O
duplicate	O
.	O
(	O
D	O
)	O
Example	O
of	O
HAP	O
selective	O
inhibition	O
of	O
the	O
production	O
of	O
IL	O
-	O
8	O
(	O
triangles	O
),	O
IL	O
-	O
6	O
(	O
squares	O
)	O
and	O
CCL	O
-	O
20	O
(	O
circles	O
)	O
by	O
primary	O
human	O
keratinocyte	O
cells	O
synergistically	O
stimulated	O
by	O
100	O
ng	O
/	O
ml	O
IL	O
-	O
17A	O
and	O
10	O
ng	O
/	O
ml	O
TNF	O
-	O
α	O
.	O
	O
HAP	O
does	O
not	O
inhibit	O
the	O
baseline	O
production	O
of	O
IL	O
-	O
6	O
,	O
IL	O
-	O
8	O
and	O
CCL	O
-	O
20	O
stimulated	O
by	O
10	O
ng	O
/	O
ml	O
TNF	O
-	O
α	O
alone	O
(	O
gray	O
lines	O
and	O
symbols	O
).	O
	O
Two	O
HAP	O
molecules	O
are	O
colored	O
blue	O
and	O
red	O
,	O
and	O
IL	O
-	O
17A	O
monomers	O
are	O
colored	O
ice	O
blue	O
and	O
pink	O
,	O
respectively	O
.	O
	O
(	O
A	O
)	O
Overview	O
of	O
the	O
distinct	O
binding	O
sites	O
of	O
Fab	O
and	O
HAP	O
to	O
IL	O
-	O
17A	O
.	O
	O
(	O
B	O
)	O
Close	O
-	O
in	O
view	O
of	O
the	O
IL	O
-	O
17A	O
/	O
HAP	O
structure	O
.	O
	O
IL	O
-	O
17A	O
β	O
-	O
strands	O
are	O
labelled	O
.	O
	O
Each	O
of	O
the	O
two	O
bound	O
HAP	O
interacts	O
with	O
both	O
monomers	O
of	O
the	O
IL	O
-	O
17A	O
dimer	O
.	O
	O
Three	O
distinct	O
areas	O
IL	O
-	O
17A	O
/	O
IL	O
-	O
17RA	O
interface	O
are	O
labeled	O
.	O
	O
Mechanism	O
of	O
the	O
inhibition	O
of	O
the	O
IL	O
-	O
17A	O
/	O
IL	O
-	O
17RA	O
interaction	O
by	O
HAP	O
.	O
	O
IL	O
-	O
17A	O
dimer	O
is	O
in	O
surface	O
presentation	O
(	O
β	O
-	O
strands	O
0	O
shown	O
as	O
ribbons	O
for	O
clarity	O
).	O
	O
HAP	O
residues	O
as	O
well	O
as	O
key	O
IL	O
-	O
17A	O
residues	O
are	O
labeled	O
.	O
	O
For	O
clarity	O
,	O
a	O
few	O
HAP	O
residues	O
are	O
also	O
shown	O
in	O
stick	O
model	O
with	O
carbon	O
atoms	O
colored	O
green	O
,	O
oxygen	O
in	O
red	O
and	O
nitrogen	O
in	O
blue	O
.	O
	O
(	O
B	O
)	O
I	O
-	O
17RA	O
(	O
ribbon	O
in	O
gold	O
)	O
peptide	O
Leu27	O
-	O
Ile32	O
binds	O
to	O
the	O
same	O
area	O
as	O
the	O
HAP	O
α	O
-	O
helix	O
.	O
	O
Trp31	O
of	O
IL	O
-	O
17RA	O
binds	O
to	O
the	O
same	O
pocket	O
in	O
IL	O
-	O
17A	O
as	O
Trp12	O
of	O
HAP	O
.	O
(	O
C	O
)	O
As	O
illustrated	O
by	O
overlay	O
a	O
single	O
HAP	O
molecule	O
and	O
β	O
-	O
strands	O
0	O
(	O
grey	O
)	O
of	O
the	O
IL	O
-	O
17A	O
/	O
HAP	O
complex	O
in	O
the	O
apo	O
IL	O
-	O
17A	O
structure	O
,	O
conformational	O
changes	O
in	O
region	O
I	O
of	O
IL	O
-	O
17A	O
are	O
needed	O
for	O
binding	O
of	O
both	O
the	O
β	O
-	O
stand	O
and	O
α	O
-	O
helix	O
of	O
the	O
HAP	O
.	O
	O
Molecular	O
Basis	O
of	O
Ligand	O
-	O
Dependent	O
Regulation	O
of	O
NadR	O
,	O
the	O
Transcriptional	O
Repressor	O
of	O
Meningococcal	O
Virulence	O
Factor	O
NadA	O
	O
Neisseria	O
adhesin	O
A	O
(	O
NadA	O
)	O
is	O
present	O
on	O
the	O
meningococcal	O
surface	O
and	O
contributes	O
to	O
adhesion	O
to	O
and	O
invasion	O
of	O
human	O
cells	O
.	O
	O
NadA	O
is	O
also	O
one	O
of	O
three	O
recombinant	O
antigens	O
in	O
the	O
recently	O
-	O
approved	O
Bexsero	O
vaccine	O
,	O
which	O
protects	O
against	O
serogroup	O
B	O
meningococcus	O
.	O
	O
In	O
the	O
presence	O
of	O
4	O
-	O
hydroxyphenylacetate	O
(	O
4	O
-	O
HPA	O
),	O
a	O
catabolite	O
present	O
in	O
human	O
saliva	O
both	O
under	O
physiological	O
conditions	O
and	O
during	O
bacterial	O
infection	O
,	O
the	O
binding	O
of	O
NadR	O
to	O
the	O
nadA	O
promoter	O
is	O
attenuated	O
and	O
nadA	O
expression	O
is	O
induced	O
.	O
	O
To	O
gain	O
insights	O
into	O
the	O
regulation	O
of	O
NadR	O
mediated	O
by	O
4	O
-	O
HPA	O
,	O
we	O
combined	O
structural	O
,	O
biochemical	O
,	O
and	O
mutagenesis	O
studies	O
.	O
	O
In	O
particular	O
,	O
two	O
new	O
crystal	O
structures	O
of	O
ligand	O
-	O
free	O
and	O
ligand	O
-	O
bound	O
NadR	O
revealed	O
(	O
i	O
)	O
the	O
molecular	O
basis	O
of	O
‘	O
conformational	O
selection	O
’	O
by	O
which	O
a	O
single	O
molecule	O
of	O
4	O
-	O
HPA	O
binds	O
and	O
stabilizes	O
dimeric	O
NadR	O
in	O
a	O
conformation	O
unsuitable	O
for	O
DNA	O
-	O
binding	O
,	O
(	O
ii	O
)	O
molecular	O
explanations	O
for	O
the	O
binding	O
specificities	O
of	O
different	O
hydroxyphenylacetate	O
ligands	O
,	O
including	O
3Cl	O
,	O
4	O
-	O
HPA	O
which	O
is	O
produced	O
during	O
inflammation	O
,	O
(	O
iii	O
)	O
the	O
presence	O
of	O
a	O
leucine	O
residue	O
essential	O
for	O
dimerization	O
and	O
conserved	O
in	O
many	O
MarR	O
family	O
proteins	O
,	O
and	O
(	O
iv	O
)	O
four	O
residues	O
(	O
His7	O
,	O
Ser9	O
,	O
Asn11	O
and	O
Phe25	O
),	O
which	O
are	O
involved	O
in	O
binding	O
4	O
-	O
HPA	O
,	O
and	O
were	O
confirmed	O
in	O
vitro	O
to	O
have	O
key	O
roles	O
in	O
the	O
regulatory	O
mechanism	O
in	O
bacteria	O
.	O
	O
Overall	O
,	O
this	O
study	O
deepens	O
our	O
molecular	O
understanding	O
of	O
the	O
sophisticated	O
regulatory	O
mechanisms	O
of	O
the	O
expression	O
of	O
nadA	O
and	O
other	O
genes	O
governed	O
by	O
NadR	O
,	O
dependent	O
on	O
interactions	O
with	O
niche	O
-	O
specific	O
signal	O
molecules	O
that	O
may	O
play	O
important	O
roles	O
during	O
meningococcal	O
pathogenesis	O
.	O
	O
The	O
Bexsero	O
vaccine	O
protects	O
against	O
MenB	O
and	O
has	O
recently	O
been	O
approved	O
in	O
>	O
35	O
countries	O
worldwide	O
.	O
	O
Neisseria	O
adhesin	O
A	O
(	O
NadA	O
)	O
present	O
on	O
the	O
meningococcal	O
surface	O
can	O
mediate	O
binding	O
to	O
human	O
cells	O
and	O
is	O
one	O
of	O
the	O
three	O
MenB	O
vaccine	O
protein	O
antigens	O
.	O
	O
A	O
deep	O
understanding	O
of	O
nadA	O
expression	O
is	O
therefore	O
important	O
,	O
otherwise	O
the	O
contribution	O
of	O
NadA	O
to	O
vaccine	O
-	O
induced	O
protection	O
against	O
meningococcal	O
meningitis	O
may	O
be	O
underestimated	O
.	O
	O
These	O
findings	O
shed	O
light	O
on	O
the	O
regulation	O
of	O
NadR	O
,	O
a	O
key	O
MarR	O
-	O
family	O
virulence	O
factor	O
of	O
this	O
important	O
human	O
pathogen	O
.	O
	O
The	O
‘	O
Reverse	O
Vaccinology	O
’	O
approach	O
was	O
pioneered	O
to	O
identify	O
antigens	O
for	O
a	O
protein	O
-	O
based	O
vaccine	O
against	O
serogroup	O
B	O
Neisseria	O
meningitidis	O
(	O
MenB	O
),	O
a	O
human	O
pathogen	O
causing	O
potentially	O
-	O
fatal	O
sepsis	O
and	O
invasive	O
meningococcal	O
disease	O
.	O
	O
Indeed	O
,	O
Reverse	O
Vaccinology	O
identified	O
Neisseria	O
adhesin	O
A	O
(	O
NadA	O
),	O
a	O
surface	O
-	O
exposed	O
protein	O
involved	O
in	O
epithelial	O
cell	O
invasion	O
and	O
found	O
in	O
~	O
30	O
%	O
of	O
clinical	O
isolates	O
.	O
	O
Recently	O
,	O
we	O
reported	O
the	O
crystal	O
structure	O
of	O
NadA	O
,	O
providing	O
insights	O
into	O
its	O
biological	O
and	O
immunological	O
functions	O
.	O
	O
Recombinant	O
NadA	O
elicits	O
a	O
strong	O
bactericidal	O
immune	O
response	O
and	O
is	O
therefore	O
included	O
in	O
the	O
Bexsero	O
vaccine	O
that	O
protects	O
against	O
MenB	O
and	O
which	O
was	O
recently	O
approved	O
in	O
over	O
35	O
countries	O
worldwide	O
.	O
	O
Previous	O
studies	O
revealed	O
that	O
nadA	O
expression	O
levels	O
are	O
mainly	O
regulated	O
by	O
the	O
Neisseria	O
adhesin	O
A	O
Regulator	O
(	O
NadR	O
).	O
	O
Studies	O
of	O
NadR	O
also	O
have	O
broader	O
implications	O
,	O
since	O
a	O
genome	O
-	O
wide	O
analysis	O
of	O
MenB	O
wild	O
-	O
type	O
and	O
nadR	O
knock	O
-	O
out	O
strains	O
revealed	O
that	O
NadR	O
influences	O
the	O
regulation	O
of	O
>	O
30	O
genes	O
,	O
including	O
maf	O
genes	O
,	O
from	O
the	O
multiple	O
adhesin	O
family	O
.	O
	O
These	O
genes	O
encode	O
a	O
wide	O
variety	O
of	O
proteins	O
connected	O
to	O
many	O
biological	O
processes	O
contributing	O
to	O
bacterial	O
survival	O
,	O
adaptation	O
in	O
the	O
host	O
niche	O
,	O
colonization	O
and	O
invasion	O
.	O
	O
NadR	O
belongs	O
to	O
the	O
MarR	O
(	O
Multiple	O
Antibiotic	O
Resistance	O
Regulator	O
)	O
family	O
,	O
a	O
group	O
of	O
ligand	O
-	O
responsive	O
transcriptional	O
regulators	O
ubiquitous	O
in	O
bacteria	O
and	O
archaea	O
.	O
	O
MarR	O
family	O
proteins	O
can	O
promote	O
bacterial	O
survival	O
in	O
the	O
presence	O
of	O
antibiotics	O
,	O
toxic	O
chemicals	O
,	O
organic	O
solvents	O
or	O
reactive	O
oxygen	O
species	O
and	O
can	O
regulate	O
virulence	O
factor	O
expression	O
.	O
	O
MarR	O
homologues	O
can	O
act	O
either	O
as	O
transcriptional	O
repressors	O
or	O
as	O
activators	O
.	O
	O
Although	O
>	O
50	O
MarR	O
family	O
structures	O
are	O
known	O
,	O
a	O
molecular	O
understanding	O
of	O
their	O
ligand	O
-	O
dependent	O
regulatory	O
mechanisms	O
is	O
still	O
limited	O
,	O
often	O
hampered	O
by	O
lack	O
of	O
identification	O
of	O
their	O
ligands	O
and	O
/	O
or	O
DNA	O
targets	O
.	O
	O
A	O
potentially	O
interesting	O
exception	O
comes	O
from	O
the	O
ligand	O
-	O
free	O
and	O
salicylate	O
-	O
bound	O
forms	O
of	O
the	O
Methanobacterium	O
thermoautotrophicum	O
protein	O
MTH313	O
which	O
revealed	O
that	O
two	O
salicylate	O
molecules	O
bind	O
to	O
one	O
MTH313	O
dimer	O
and	O
induce	O
large	O
conformational	O
changes	O
,	O
apparently	O
sufficient	O
to	O
prevent	O
DNA	O
binding	O
.	O
	O
However	O
,	O
it	O
is	O
unknown	O
whether	O
salicylate	O
is	O
a	O
relevant	O
in	O
vivo	O
ligand	O
of	O
either	O
of	O
these	O
two	O
proteins	O
,	O
which	O
share	O
~	O
20	O
%	O
sequence	O
identity	O
with	O
NadR	O
,	O
rendering	O
unclear	O
the	O
interpretation	O
of	O
these	O
findings	O
in	O
relation	O
to	O
the	O
regulatory	O
mechanisms	O
of	O
NadR	O
or	O
other	O
MarR	O
family	O
proteins	O
.	O
	O
NadR	O
binds	O
nadA	O
on	O
three	O
different	O
operators	O
(	O
OpI	O
,	O
OpII	O
and	O
OpIII	O
).	O
	O
4	O
-	O
HPA	O
is	O
a	O
small	O
molecule	O
derived	O
from	O
mammalian	O
aromatic	O
amino	O
acid	O
catabolism	O
and	O
is	O
released	O
in	O
human	O
saliva	O
,	O
where	O
it	O
has	O
been	O
detected	O
at	O
micromolar	O
concentration	O
.	O
	O
In	O
the	O
presence	O
of	O
4	O
-	O
HPA	O
,	O
NadR	O
is	O
unable	O
to	O
bind	O
the	O
nadA	O
promoter	O
and	O
nadA	O
gene	O
expression	O
is	O
induced	O
.	O
	O
In	O
vivo	O
,	O
the	O
presence	O
of	O
4	O
-	O
HPA	O
in	O
the	O
host	O
niche	O
of	O
N	O
.	O
meningitidis	O
serves	O
as	O
an	O
inducer	O
of	O
NadA	O
production	O
,	O
thereby	O
promoting	O
bacterial	O
adhesion	O
to	O
host	O
cells	O
.	O
	O
Further	O
,	O
we	O
recently	O
reported	O
that	O
3Cl	O
,	O
4	O
-	O
HPA	O
,	O
produced	O
during	O
inflammation	O
,	O
is	O
another	O
inducer	O
of	O
nadA	O
expression	O
.	O
	O
Extending	O
our	O
previous	O
studies	O
based	O
on	O
hydrogen	O
-	O
deuterium	O
exchange	O
mass	O
spectrometry	O
(	O
HDX	O
-	O
MS	O
),	O
here	O
we	O
sought	O
to	O
reveal	O
the	O
molecular	O
mechanisms	O
and	O
effects	O
of	O
NadR	O
/	O
HPA	O
interactions	O
via	O
X	O
-	O
ray	O
crystallography	O
,	O
NMR	O
spectroscopy	O
and	O
complementary	O
biochemical	O
and	O
in	O
vivo	O
mutagenesis	O
studies	O
.	O
	O
Standard	O
chromatographic	O
techniques	O
were	O
used	O
to	O
obtain	O
a	O
highly	O
purified	O
sample	O
of	O
NadR	O
(	O
see	O
Materials	O
and	O
Methods	O
).	O
	O
These	O
data	O
showed	O
that	O
NadR	O
was	O
dimeric	O
in	O
solution	O
,	O
since	O
the	O
theoretical	O
molecular	O
mass	O
of	O
the	O
NadR	O
dimer	O
is	O
33	O
.	O
73	O
kDa	O
;	O
and	O
,	O
there	O
was	O
no	O
change	O
in	O
oligomeric	O
state	O
on	O
addition	O
of	O
4	O
-	O
HPA	O
.	O
	O
The	O
thermal	O
stability	O
of	O
NadR	O
was	O
examined	O
using	O
differential	O
scanning	O
calorimetry	O
(	O
DSC	O
).	O
	O
The	O
Tm	O
of	O
NadR	O
was	O
67	O
.	O
4	O
±	O
0	O
.	O
1	O
°	O
C	O
in	O
the	O
absence	O
of	O
ligand	O
,	O
and	O
was	O
unaffected	O
by	O
salicylate	O
.	O
	O
However	O
,	O
an	O
increased	O
thermal	O
stability	O
was	O
induced	O
by	O
4	O
-	O
HPA	O
and	O
,	O
to	O
a	O
lesser	O
extent	O
,	O
by	O
3	O
-	O
HPA	O
.	O
	O
Interestingly	O
,	O
NadR	O
displayed	O
the	O
greatest	O
Tm	O
increase	O
upon	O
addition	O
of	O
3Cl	O
,	O
4	O
-	O
HPA	O
(	O
Table	O
1	O
and	O
Fig	O
1B	O
).	O
	O
Stability	O
of	O
NadR	O
is	O
increased	O
by	O
small	O
molecule	O
ligands	O
.	O
	O
Melting	O
-	O
point	O
(	O
Tm	O
)	O
and	O
its	O
ligand	O
-	O
induced	O
increase	O
(	O
ΔTm	O
)	O
derived	O
from	O
DSC	O
thermostability	O
experiments	O
.	O
	O
Dissociation	O
constants	O
(	O
KD	O
)	O
of	O
the	O
NadR	O
/	O
ligand	O
interactions	O
from	O
SPR	O
steady	O
-	O
state	O
binding	O
experiments	O
.	O
	O
Ligand	O
Tm	O
(°	O
C	O
)	O
ΔTm	O
(°	O
C	O
)	O
KD	O
(	O
mM	O
)	O
No	O
ligand	O
67	O
.	O
4	O
±	O
0	O
.	O
1	O
n	O
.	O
a	O
.	O
n	O
.	O
a	O
.	O
	O
NadR	O
displays	O
distinct	O
binding	O
affinities	O
for	O
hydroxyphenylacetate	O
ligands	O
	O
The	O
SPR	O
sensorgrams	O
revealed	O
very	O
fast	O
association	O
and	O
dissociation	O
events	O
,	O
typical	O
of	O
small	O
molecule	O
ligands	O
,	O
thus	O
prohibiting	O
a	O
detailed	O
study	O
of	O
binding	O
kinetics	O
.	O
	O
3	O
-	O
HPA	O
showed	O
a	O
weaker	O
interaction	O
,	O
with	O
a	O
KD	O
of	O
2	O
.	O
7	O
mM	O
,	O
while	O
salicylate	O
showed	O
only	O
a	O
very	O
weak	O
response	O
that	O
did	O
not	O
reach	O
saturation	O
,	O
indicating	O
a	O
non	O
-	O
specific	O
interaction	O
with	O
NadR	O
.	O
A	O
ranking	O
of	O
these	O
KD	O
values	O
showed	O
that	O
3Cl	O
,	O
4	O
-	O
HPA	O
was	O
the	O
tightest	O
binder	O
,	O
and	O
thus	O
matched	O
the	O
ranking	O
of	O
ligand	O
-	O
induced	O
Tm	O
increases	O
observed	O
in	O
the	O
DSC	O
experiments	O
.	O
	O
Although	O
these	O
KD	O
values	O
indicate	O
rather	O
weak	O
interactions	O
,	O
they	O
are	O
similar	O
to	O
the	O
values	O
reported	O
previously	O
for	O
the	O
MarR	O
/	O
salicylate	O
interaction	O
(	O
KD	O
~	O
1	O
mM	O
)	O
and	O
the	O
MTH313	O
/	O
salicylate	O
interaction	O
(	O
KD	O
2	O
–	O
3	O
mM	O
),	O
and	O
approximately	O
20	O
-	O
fold	O
tighter	O
than	O
the	O
ST1710	O
/	O
salicylate	O
interaction	O
(	O
KD	O
~	O
20	O
mM	O
).	O
	O
Crystal	O
structures	O
of	O
holo	O
-	O
NadR	O
and	O
apo	O
-	O
NadR	O
	O
First	O
,	O
we	O
crystallized	O
NadR	O
(	O
a	O
selenomethionine	O
-	O
labelled	O
derivative	O
)	O
in	O
the	O
presence	O
of	O
a	O
200	O
-	O
fold	O
molar	O
excess	O
of	O
4	O
-	O
HPA	O
.	O
	O
The	O
structure	O
of	O
the	O
NadR	O
/	O
4	O
-	O
HPA	O
complex	O
was	O
determined	O
at	O
2	O
.	O
3	O
Å	O
resolution	O
using	O
a	O
combination	O
of	O
the	O
single	O
-	O
wavelength	O
anomalous	O
dispersion	O
(	O
SAD	O
)	O
and	O
molecular	O
replacement	O
(	O
MR	O
)	O
methods	O
,	O
and	O
was	O
refined	O
to	O
R	O
work	O
/	O
R	O
free	O
values	O
of	O
20	O
.	O
9	O
/	O
26	O
.	O
0	O
%	O
(	O
Table	O
2	O
).	O
	O
Despite	O
numerous	O
attempts	O
,	O
we	O
were	O
unable	O
to	O
obtain	O
high	O
-	O
quality	O
crystals	O
of	O
NadR	O
complexed	O
with	O
3Cl	O
,	O
4	O
-	O
HPA	O
,	O
3	O
,	O
4	O
-	O
HPA	O
,	O
3	O
-	O
HPA	O
or	O
DNA	O
targets	O
.	O
	O
However	O
,	O
it	O
was	O
eventually	O
possible	O
to	O
crystallize	O
apo	O
-	O
NadR	O
,	O
and	O
the	O
structure	O
was	O
determined	O
at	O
2	O
.	O
7	O
Å	O
resolution	O
by	O
MR	O
methods	O
using	O
the	O
NadR	O
/	O
4	O
-	O
HPA	O
complex	O
as	O
the	O
search	O
model	O
.	O
	O
The	O
apo	O
-	O
NadR	O
structure	O
was	O
refined	O
to	O
R	O
work	O
/	O
R	O
free	O
values	O
of	O
19	O
.	O
1	O
/	O
26	O
.	O
8	O
%	O
(	O
Table	O
2	O
).	O
	O
The	O
asymmetric	O
unit	O
of	O
the	O
NadR	O
/	O
4	O
-	O
HPA	O
crystals	O
(	O
holo	O
-	O
NadR	O
)	O
contained	O
one	O
NadR	O
homodimer	O
,	O
while	O
the	O
apo	O
-	O
NadR	O
crystals	O
contained	O
two	O
homodimers	O
.	O
	O
Moreover	O
,	O
our	O
SE	O
-	O
HPLC	O
/	O
MALLS	O
analyses	O
(	O
see	O
above	O
)	O
revealed	O
that	O
in	O
solution	O
NadR	O
is	O
dimeric	O
,	O
and	O
previous	O
studies	O
using	O
native	O
mass	O
spectrometry	O
(	O
MS	O
)	O
revealed	O
dimers	O
,	O
not	O
tetramers	O
.	O
	O
The	O
NadR	O
homodimer	O
bound	O
to	O
4	O
-	O
HPA	O
has	O
a	O
dimerization	O
interface	O
mostly	O
involving	O
the	O
top	O
of	O
its	O
‘	O
triangular	O
’	O
form	O
,	O
while	O
the	O
two	O
DNA	O
-	O
binding	O
domains	O
are	O
located	O
at	O
the	O
base	O
(	O
Fig	O
2A	O
).	O
	O
High	O
-	O
quality	O
electron	O
density	O
maps	O
allowed	O
clear	O
identification	O
of	O
the	O
bound	O
ligand	O
,	O
4	O
-	O
HPA	O
(	O
Fig	O
2B	O
).	O
	O
The	O
overall	O
structure	O
of	O
NadR	O
shows	O
dimensions	O
of	O
~	O
50	O
×	O
65	O
×	O
50	O
Å	O
and	O
a	O
large	O
homodimer	O
interface	O
that	O
buries	O
a	O
total	O
surface	O
area	O
of	O
~	O
4800	O
Å2	O
.	O
	O
Each	O
NadR	O
monomer	O
consists	O
of	O
six	O
α	O
-	O
helices	O
and	O
two	O
short	O
β	O
-	O
strands	O
,	O
with	O
helices	O
α1	O
,	O
α5	O
,	O
and	O
α6	O
forming	O
the	O
dimer	O
interface	O
.	O
	O
Together	O
,	O
these	O
structural	O
elements	O
constitute	O
the	O
winged	O
helix	O
-	O
turn	O
-	O
helix	O
(	O
wHTH	O
)	O
DNA	O
-	O
binding	O
domain	O
and	O
,	O
together	O
with	O
the	O
dimeric	O
organization	O
,	O
are	O
the	O
hallmarks	O
of	O
MarR	O
family	O
structures	O
.	O
	O
The	O
crystal	O
structure	O
of	O
NadR	O
in	O
complex	O
with	O
4	O
-	O
HPA	O
.	O
	O
(	O
A	O
)	O
The	O
holo	O
-	O
NadR	O
homodimer	O
is	O
depicted	O
in	O
green	O
and	O
blue	O
for	O
chains	O
A	O
and	O
B	O
respectively	O
,	O
while	O
yellow	O
sticks	O
depict	O
the	O
4	O
-	O
HPA	O
ligand	O
(	O
labelled	O
).	O
	O
For	O
simplicity	O
,	O
secondary	O
structure	O
elements	O
are	O
labelled	O
for	O
chain	O
B	O
only	O
.	O
	O
Red	O
dashes	O
show	O
hypothetical	O
positions	O
of	O
chain	O
B	O
residues	O
88	O
–	O
90	O
that	O
were	O
not	O
modeled	O
due	O
to	O
lack	O
of	O
electron	O
density	O
.	O
	O
(	O
B	O
)	O
A	O
zoom	O
into	O
the	O
pocket	O
occupied	O
by	O
4	O
-	O
HPA	O
shows	O
that	O
the	O
ligand	O
contacts	O
both	O
chains	O
A	O
and	O
B	O
;	O
blue	O
mesh	O
shows	O
electron	O
density	O
around	O
4	O
-	O
HPA	O
calculated	O
from	O
a	O
composite	O
omit	O
map	O
(	O
omitting	O
4	O
-	O
HPA	O
),	O
using	O
phenix	O
.	O
	O
A	O
single	O
conserved	O
leucine	O
residue	O
(	O
L130	O
)	O
is	O
crucial	O
for	O
dimerization	O
	O
The	O
NadR	O
dimer	O
interface	O
is	O
formed	O
by	O
at	O
least	O
32	O
residues	O
,	O
which	O
establish	O
numerous	O
inter	O
-	O
chain	O
salt	O
bridges	O
or	O
hydrogen	O
bonds	O
,	O
and	O
many	O
hydrophobic	O
packing	O
interactions	O
(	O
Fig	O
3A	O
and	O
3B	O
).	O
	O
To	O
determine	O
which	O
residues	O
were	O
most	O
important	O
for	O
dimerization	O
,	O
we	O
studied	O
the	O
interface	O
in	O
silico	O
and	O
identified	O
several	O
residues	O
as	O
potential	O
mediators	O
of	O
key	O
stabilizing	O
interactions	O
.	O
	O
Each	O
mutant	O
NadR	O
protein	O
was	O
purified	O
,	O
and	O
then	O
its	O
oligomeric	O
state	O
was	O
examined	O
by	O
analytical	O
SE	O
-	O
HPLC	O
.	O
	O
Almost	O
all	O
the	O
mutants	O
showed	O
the	O
same	O
elution	O
profile	O
as	O
the	O
wild	O
-	O
type	O
(	O
WT	O
)	O
NadR	O
protein	O
.	O
	O
Only	O
the	O
L130K	B-mutant
mutation	O
induced	O
a	O
notable	O
change	O
in	O
the	O
oligomeric	O
state	O
of	O
NadR	O
(	O
Fig	O
3C	O
).	O
	O
Further	O
,	O
in	O
SE	O
-	O
MALLS	O
analyses	O
,	O
the	O
L130K	B-mutant
mutant	O
displayed	O
two	O
distinct	O
species	O
in	O
solution	O
,	O
approximately	O
80	O
%	O
being	O
monomeric	O
(	O
a	O
19	O
kDa	O
species	O
),	O
and	O
only	O
20	O
%	O
retaining	O
the	O
typical	O
native	O
dimeric	O
state	O
(	O
a	O
35	O
kDa	O
species	O
)	O
(	O
Fig	O
3D	O
),	O
demonstrating	O
that	O
Leu130	O
is	O
crucial	O
for	O
stable	O
dimerization	O
.	O
	O
In	O
contrast	O
,	O
most	O
of	O
the	O
other	O
residues	O
identified	O
in	O
the	O
NadR	O
dimer	O
interface	O
were	O
poorly	O
conserved	O
in	O
the	O
MarR	O
family	O
.	O
	O
Analysis	O
of	O
the	O
NadR	O
dimer	O
interface	O
.	O
	O
(	O
A	O
)	O
Both	O
orientations	O
show	O
chain	O
A	O
,	O
green	O
backbone	O
ribbon	O
,	O
colored	O
red	O
to	O
highlight	O
all	O
locations	O
involved	O
in	O
dimerization	O
;	O
namely	O
,	O
inter	O
-	O
chain	O
salt	O
bridges	O
or	O
hydrogen	O
bonds	O
involving	O
Q4	O
,	O
S5	O
,	O
K6	O
,	O
H7	O
,	O
S9	O
,	O
I10	O
,	O
N11	O
,	O
I15	O
,	O
Q16	O
,	O
R18	O
,	O
D36	O
,	O
R43	O
,	O
A46	O
,	O
Q59	O
,	O
C61	O
,	O
Y104	O
,	O
D112	O
,	O
R114	O
,	O
Y115	O
,	O
D116	O
,	O
E119	O
,	O
K126	O
,	O
E136	O
,	O
E141	O
,	O
N145	O
,	O
and	O
the	O
hydrophobic	O
packing	O
interactions	O
involving	O
I10	O
,	O
I12	O
,	O
L14	O
,	O
I15	O
,	O
R18	O
,	O
Y115	O
,	O
I118	O
,	O
L130	O
,	O
L133	O
,	O
L134	O
and	O
L137	O
.	O
	O
Chain	O
B	O
,	O
grey	O
surface	O
,	O
is	O
marked	O
blue	O
to	O
highlight	O
residues	O
probed	O
by	O
site	O
-	O
directed	O
mutagenesis	O
(	O
E136	O
only	O
makes	O
a	O
salt	O
bridge	O
with	O
K126	O
,	O
therefore	O
it	O
was	O
sufficient	O
to	O
make	O
the	O
K126A	B-mutant
mutation	O
to	O
assess	O
the	O
importance	O
of	O
this	O
ionic	O
interaction	O
;	O
the	O
H7	O
position	O
is	O
labelled	O
for	O
monomer	O
A	O
,	O
since	O
electron	O
density	O
was	O
lacking	O
for	O
monomer	O
B	O
).	O
(	O
B	O
)	O
A	O
zoom	O
into	O
the	O
environment	O
of	O
helix	O
α6	O
to	O
show	O
how	O
residue	O
L130	O
chain	O
B	O
(	O
blue	O
side	O
chain	O
)	O
is	O
a	O
focus	O
of	O
hydrophobic	O
packing	O
interactions	O
with	O
L130	O
,	O
L133	O
,	O
L134	O
and	O
L137	O
of	O
chain	O
A	O
(	O
red	O
side	O
chains	O
).	O
	O
(	O
C	O
)	O
SE	O
-	O
HPLC	O
analyses	O
of	O
all	O
mutant	O
forms	O
of	O
NadR	O
are	O
compared	O
with	O
the	O
wild	O
-	O
type	O
(	O
WT	O
)	O
protein	O
.	O
	O
The	O
WT	O
and	O
most	O
of	O
the	O
mutants	O
show	O
a	O
single	O
elution	O
peak	O
with	O
an	O
absorbance	O
maximum	O
at	O
17	O
.	O
5	O
min	O
.	O
	O
The	O
holo	O
-	O
NadR	O
structure	O
presents	O
only	O
one	O
occupied	O
ligand	O
-	O
binding	O
pocket	O
	O
The	O
tunnel	O
was	O
lined	O
with	O
rather	O
hydrophobic	O
amino	O
acids	O
,	O
and	O
did	O
not	O
contain	O
water	O
molecules	O
.	O
	O
Unexpectedly	O
,	O
only	O
one	O
monomer	O
of	O
the	O
holo	O
-	O
NadR	O
homodimer	O
contained	O
4	O
-	O
HPA	O
in	O
the	O
binding	O
pocket	O
,	O
whereas	O
the	O
corresponding	O
pocket	O
of	O
the	O
other	O
monomer	O
was	O
unoccupied	O
by	O
ligand	O
,	O
despite	O
the	O
large	O
excess	O
of	O
4	O
-	O
HPA	O
used	O
in	O
the	O
crystallization	O
conditions	O
.	O
	O
Inspection	O
of	O
the	O
protein	O
-	O
ligand	O
interaction	O
network	O
revealed	O
no	O
bonds	O
from	O
NadR	O
backbone	O
groups	O
to	O
the	O
ligand	O
,	O
but	O
several	O
key	O
side	O
chain	O
mediated	O
hydrogen	O
(	O
H	O
)-	O
bonds	O
and	O
ionic	O
interactions	O
,	O
most	O
notably	O
between	O
the	O
carboxylate	O
group	O
of	O
4	O
-	O
HPA	O
and	O
Ser9	O
of	O
chain	O
A	O
(	O
SerA9	O
),	O
and	O
chain	O
B	O
residues	O
TrpB39	O
,	O
ArgB43	O
and	O
TyrB115	O
(	O
Fig	O
4A	O
).	O
	O
Atomic	O
details	O
of	O
NadR	O
/	O
HPA	O
interactions	O
.	O
	O
A	O
)	O
A	O
stereo	O
-	O
view	O
zoom	O
into	O
the	O
binding	O
pocket	O
showing	O
side	O
chain	O
sticks	O
for	O
all	O
interactions	O
between	O
NadR	O
and	O
4	O
-	O
HPA	O
.	O
	O
4	O
-	O
HPA	O
is	O
shown	O
in	O
yellow	O
sticks	O
,	O
with	O
oxygen	O
atoms	O
in	O
red	O
.	O
	O
A	O
water	O
molecule	O
is	O
shown	O
by	O
the	O
red	O
sphere	O
.	O
	O
H	O
-	O
bonds	O
up	O
to	O
3	O
.	O
6Å	O
are	O
shown	O
as	O
dashed	O
lines	O
.	O
	O
The	O
entire	O
set	O
of	O
residues	O
making	O
H	O
-	O
bonds	O
or	O
non	O
-	O
bonded	O
contacts	O
with	O
4	O
-	O
HPA	O
is	O
as	O
follows	O
:	O
SerA9	O
,	O
AsnA11	O
,	O
LeuB21	O
,	O
MetB22	O
,	O
PheB25	O
,	O
LeuB29	O
,	O
AspB36	O
,	O
TrpB39	O
,	O
ArgB43	O
,	O
ValB111	O
and	O
TyrB115	O
(	O
automated	O
analysis	O
performed	O
using	O
PDBsum	O
and	O
verified	O
manually	O
).	O
	O
Side	O
chains	O
mediating	O
hydrophobic	O
interactions	O
are	O
shown	O
in	O
orange	O
.	O
(	O
B	O
)	O
A	O
model	O
was	O
prepared	O
to	O
visualize	O
putative	O
interactions	O
of	O
3Cl	O
,	O
4	O
-	O
HPA	O
(	O
pink	O
)	O
with	O
NadR	O
,	O
revealing	O
the	O
potential	O
for	O
additional	O
contacts	O
(	O
dashed	O
lines	O
)	O
of	O
the	O
chloro	O
moiety	O
(	O
green	O
stick	O
)	O
with	O
LeuB29	O
and	O
AspB36	O
.	O
	O
Notably	O
,	O
the	O
phenyl	O
ring	O
of	O
PheB25	O
was	O
positioned	O
parallel	O
to	O
the	O
phenyl	O
ring	O
of	O
4	O
-	O
HPA	O
,	O
potentially	O
forming	O
π	O
-	O
π	O
parallel	O
-	O
displaced	O
stacking	O
interactions	O
.	O
	O
Consequently	O
,	O
residues	O
in	O
the	O
4	O
-	O
HPA	O
binding	O
pocket	O
are	O
mostly	O
contributed	O
by	O
NadR	O
chain	O
B	O
,	O
and	O
effectively	O
created	O
a	O
polar	O
‘	O
floor	O
’	O
and	O
a	O
hydrophobic	O
‘	O
ceiling	O
’,	O
which	O
house	O
the	O
ligand	O
.	O
	O
Collectively	O
,	O
this	O
mixed	O
network	O
of	O
polar	O
and	O
hydrophobic	O
interactions	O
endows	O
NadR	O
with	O
a	O
strong	O
recognition	O
pattern	O
for	O
HPAs	O
,	O
with	O
additional	O
medium	O
-	O
range	O
interactions	O
potentially	O
established	O
with	O
the	O
hydroxyl	O
group	O
at	O
the	O
4	O
-	O
position	O
.	O
	O
Structure	O
-	O
activity	O
relationships	O
:	O
molecular	O
basis	O
of	O
enhanced	O
stabilization	O
by	O
3Cl	O
,	O
4	O
-	O
HPA	O
	O
We	O
modelled	O
the	O
binding	O
of	O
other	O
HPAs	O
by	O
in	O
silico	O
superposition	O
onto	O
4	O
-	O
HPA	O
in	O
the	O
holo	O
-	O
NadR	O
structure	O
,	O
and	O
thereby	O
obtained	O
molecular	O
explanations	O
for	O
the	O
binding	O
specificities	O
of	O
diverse	O
ligands	O
.	O
	O
For	O
example	O
,	O
similar	O
to	O
4	O
-	O
HPA	O
,	O
the	O
binding	O
of	O
3Cl	O
,	O
4	O
-	O
HPA	O
could	O
involve	O
multiple	O
bonds	O
towards	O
the	O
carboxylate	O
group	O
of	O
the	O
ligand	O
and	O
some	O
to	O
the	O
4	O
-	O
hydroxyl	O
group	O
.	O
	O
Additionally	O
,	O
the	O
side	O
chains	O
of	O
LeuB29	O
and	O
AspB36	O
would	O
be	O
only	O
2	O
.	O
6	O
–	O
3	O
.	O
5	O
Å	O
from	O
the	O
chlorine	O
atom	O
,	O
thus	O
providing	O
van	O
der	O
Waals	O
’	O
interactions	O
or	O
H	O
-	O
bonds	O
to	O
generate	O
the	O
additional	O
binding	O
affinity	O
observed	O
for	O
3Cl	O
,	O
4	O
-	O
HPA	O
(	O
Fig	O
4B	O
).	O
	O
Finally	O
,	O
salicylate	O
is	O
presumably	O
unable	O
to	O
specifically	O
bind	O
NadR	O
due	O
to	O
the	O
2	O
-	O
hydroxyl	O
substitution	O
and	O
the	O
shorter	O
aliphatic	O
chain	O
connecting	O
its	O
carboxylate	O
group	O
(	O
Fig	O
1A	O
):	O
the	O
compound	O
simply	O
seems	O
too	O
small	O
to	O
simultaneously	O
establish	O
the	O
network	O
of	O
beneficial	O
bonds	O
observed	O
in	O
the	O
NadR	O
/	O
HPA	O
interactions	O
.	O
	O
Analysis	O
of	O
the	O
pockets	O
reveals	O
the	O
molecular	O
basis	O
for	O
asymmetric	O
binding	O
and	O
stoichiometry	O
	O
However	O
,	O
studies	O
based	O
on	O
tryptophan	O
fluorescence	O
were	O
confounded	O
by	O
the	O
fluorescence	O
of	O
the	O
HPA	O
ligands	O
,	O
and	O
isothermal	O
titration	O
calorimetry	O
(	O
ITC	O
)	O
was	O
unfeasible	O
due	O
to	O
the	O
need	O
for	O
very	O
high	O
concentrations	O
of	O
NadR	O
in	O
the	O
ITC	O
chamber	O
(	O
due	O
to	O
the	O
relatively	O
low	O
affinity	O
),	O
which	O
exceeded	O
the	O
solubility	O
limits	O
of	O
the	O
protein	O
.	O
	O
However	O
,	O
it	O
was	O
possible	O
to	O
calculate	O
the	O
binding	O
stoichiometry	O
of	O
the	O
NadR	O
-	O
HPA	O
interactions	O
using	O
an	O
SPR	O
-	O
based	O
approach	O
.	O
	O
This	O
approach	O
relies	O
on	O
the	O
assumption	O
that	O
the	O
captured	O
protein	O
(‘	O
the	O
ligand	O
’,	O
according	O
to	O
SPR	O
conventions	O
)	O
is	O
100	O
%	O
active	O
and	O
freely	O
-	O
accessible	O
to	O
potential	O
interactors	O
(‘	O
the	O
analytes	O
’).	O
	O
Overall	O
,	O
the	O
superposition	O
revealed	O
a	O
high	O
degree	O
of	O
structural	O
similarity	O
(	O
Cα	O
root	O
mean	O
square	O
deviation	O
(	O
rmsd	O
)	O
of	O
1	O
.	O
5Å	O
),	O
though	O
on	O
closer	O
inspection	O
a	O
rotational	O
difference	O
of	O
~	O
9	O
degrees	O
along	O
the	O
long	O
axis	O
of	O
helix	O
α6	O
was	O
observed	O
,	O
suggesting	O
that	O
4	O
-	O
HPA	O
induced	O
a	O
slight	O
conformational	O
change	O
(	O
Fig	O
5A	O
).	O
	O
Most	O
notably	O
,	O
atomic	O
clashes	O
between	O
the	O
ligand	O
and	O
the	O
side	O
chains	O
of	O
MetA22	O
,	O
PheA25	O
and	O
ArgA43	O
would	O
occur	O
if	O
4	O
-	O
HPA	O
were	O
present	O
in	O
the	O
monomer	O
A	O
pocket	O
(	O
Fig	O
5B	O
).	O
	O
Subsequently	O
,	O
analyses	O
of	O
the	O
pockets	O
in	O
apo	O
-	O
NadR	O
revealed	O
that	O
in	O
the	O
absence	O
of	O
ligand	O
the	O
long	O
Arg43	O
side	O
chain	O
was	O
always	O
in	O
the	O
open	O
‘	O
outward	O
’	O
position	O
compatible	O
with	O
binding	O
to	O
the	O
4	O
-	O
HPA	O
carboxylate	O
group	O
.	O
	O
Structural	O
differences	O
of	O
NadR	O
in	O
ligand	O
-	O
bound	O
or	O
free	O
forms	O
.	O
	O
(	O
A	O
)	O
Aligned	O
monomers	O
of	O
holo	O
-	O
NadR	O
(	O
chain	O
A	O
:	O
green	O
;	O
chain	O
B	O
:	O
blue	O
),	O
reveal	O
major	O
overall	O
differences	O
by	O
the	O
shift	O
of	O
helix	O
α6	O
.	O
(	O
B	O
)	O
Comparison	O
of	O
the	O
two	O
binding	O
pockets	O
in	O
holo	O
-	O
NadR	O
shows	O
that	O
in	O
the	O
ligand	O
-	O
free	O
monomer	O
A	O
(	O
green	O
)	O
residues	O
Met22	O
,	O
Phe25	O
and	O
Arg43	O
adopt	O
‘	O
inward	O
’	O
positions	O
(	O
highlighted	O
by	O
arrows	O
)	O
compared	O
to	O
the	O
ligand	O
-	O
occupied	O
pocket	O
(	O
blue	O
residues	O
);	O
these	O
‘	O
inward	O
’	O
conformations	O
appear	O
unfavorable	O
for	O
binding	O
of	O
4	O
-	O
HPA	O
due	O
to	O
clashes	O
with	O
the	O
4	O
-	O
hydroxyl	O
group	O
,	O
the	O
phenyl	O
ring	O
and	O
the	O
carboxylate	O
group	O
,	O
respectively	O
.	O
	O
In	O
these	O
crystals	O
,	O
the	O
ArgA43	O
side	O
chain	O
showed	O
two	O
alternate	O
conformations	O
,	O
modelled	O
with	O
50	O
%	O
occupancy	O
in	O
each	O
state	O
,	O
as	O
indicated	O
by	O
the	O
two	O
‘	O
mirrored	O
’	O
arrows	O
.	O
	O
Finally	O
,	O
we	O
applied	O
15N	O
heteronuclear	O
solution	O
NMR	O
spectroscopy	O
to	O
examine	O
the	O
interaction	O
of	O
4	O
-	O
HPA	O
with	O
apo	O
NadR	O
.	O
We	O
collected	O
NMR	O
spectra	O
on	O
NadR	O
in	O
the	O
presence	O
and	O
absence	O
of	O
4	O
-	O
HPA	O
(	O
see	O
Materials	O
and	O
Methods	O
).	O
	O
The	O
1H	O
-	O
15N	O
TROSY	O
-	O
HSQC	O
spectrum	O
of	O
apo	O
-	O
NadR	O
,	O
acquired	O
at	O
25	O
°	O
C	O
,	O
displayed	O
approximately	O
140	O
distinct	O
peaks	O
(	O
Fig	O
6A	O
),	O
most	O
of	O
which	O
correspond	O
to	O
backbone	O
amide	O
N	O
-	O
H	O
groups	O
.	O
	O
Upon	O
the	O
addition	O
of	O
4	O
-	O
HPA	O
,	O
over	O
45	O
peaks	O
showed	O
chemical	O
shift	O
perturbations	O
,	O
i	O
.	O
e	O
.	O
changed	O
position	O
in	O
the	O
spectrum	O
or	O
disappeared	O
,	O
while	O
the	O
remaining	O
peaks	O
remained	O
unchanged	O
.	O
	O
This	O
observation	O
showed	O
that	O
4	O
-	O
HPA	O
was	O
able	O
to	O
bind	O
NadR	O
and	O
induce	O
notable	O
changes	O
in	O
specific	O
regions	O
of	O
the	O
protein	O
.	O
	O
NMR	O
spectra	O
of	O
NadR	O
in	O
the	O
presence	O
and	O
absence	O
of	O
4	O
-	O
HPA	O
.	O
	O
(	O
A	O
)	O
Superposition	O
of	O
two	O
1H	O
-	O
15N	O
TROSY	O
-	O
HSQC	O
spectra	O
recorded	O
at	O
25	O
°	O
C	O
on	O
apo	O
-	O
NadR	O
(	O
cyan	O
)	O
and	O
on	O
NadR	O
in	O
the	O
presence	O
of	O
4	O
-	O
HPA	O
(	O
red	O
).	O
	O
The	O
spectra	O
acquired	O
at	O
10	O
°	O
C	O
are	O
excluded	O
from	O
panel	O
A	O
for	O
simplicity	O
.	O
	O
However	O
,	O
in	O
the	O
presence	O
of	O
4	O
-	O
HPA	O
,	O
the	O
1H	O
-	O
15N	O
TROSY	O
-	O
HSQC	O
spectrum	O
of	O
NadR	O
displayed	O
approximately	O
140	O
peaks	O
,	O
as	O
for	O
apo	O
-	O
NadR	O
,	O
i	O
.	O
e	O
.	O
two	O
distinct	O
stable	O
conformations	O
(	O
that	O
might	O
have	O
potentially	O
revealed	O
the	O
molecular	O
asymmetry	O
observed	O
crystallographically	O
)	O
were	O
not	O
notable	O
.	O
	O
These	O
doubled	O
peaks	O
may	O
therefore	O
reveal	O
that	O
the	O
cooler	O
temperature	O
partially	O
trapped	O
the	O
existence	O
in	O
solution	O
of	O
two	O
distinct	O
states	O
,	O
in	O
presence	O
or	O
absence	O
of	O
4	O
-	O
HPA	O
,	O
with	O
minor	O
conformational	O
differences	O
occurring	O
at	O
least	O
in	O
proximity	O
to	O
the	O
binding	O
pocket	O
.	O
	O
Apo	O
-	O
NadR	O
structures	O
reveal	O
intrinsic	O
conformational	O
flexibility	O
	O
The	O
apo	O
-	O
NadR	O
crystal	O
structure	O
contained	O
two	O
homodimers	O
in	O
the	O
asymmetric	O
unit	O
(	O
chains	O
A	O
+	O
B	O
and	O
chains	O
C	O
+	O
D	O
).	O
	O
Upon	O
overall	O
structural	O
superposition	O
,	O
these	O
dimers	O
revealed	O
a	O
few	O
minor	O
differences	O
in	O
the	O
α6	O
helix	O
(	O
a	O
major	O
component	O
of	O
the	O
dimer	O
interface	O
)	O
and	O
the	O
helices	O
α4	O
-	O
α5	O
(	O
the	O
DNA	O
binding	O
region	O
),	O
and	O
an	O
rmsd	O
of	O
1	O
.	O
55Å	O
(	O
Fig	O
7A	O
).	O
	O
The	O
slightly	O
larger	O
rmsd	O
between	O
the	O
two	O
apo	O
-	O
homodimers	O
,	O
rather	O
than	O
between	O
apo	O
-	O
and	O
holo	O
-	O
homodimers	O
,	O
further	O
indicate	O
that	O
apo	O
-	O
NadR	O
possesses	O
a	O
notable	O
degree	O
of	O
intrinsic	O
conformational	O
flexibility	O
.	O
	O
Overall	O
apo	O
-	O
and	O
holo	O
-	O
NadR	O
structures	O
are	O
similar	O
.	O
	O
(	O
A	O
)	O
Pairwise	O
alignment	O
of	O
the	O
two	O
distinct	O
apo	O
-	O
NadR	O
homodimers	O
(	O
AB	O
and	O
CD	O
)	O
present	O
in	O
the	O
apo	O
-	O
NadR	O
crystals	O
.	O
(	O
B	O
)	O
Alignment	O
of	O
the	O
holo	O
-	O
NadR	O
homodimer	O
(	O
green	O
and	O
blue	O
chains	O
)	O
onto	O
the	O
apo	O
-	O
NadR	O
homodimers	O
.	O
	O
Here	O
,	O
larger	O
differences	O
are	O
observed	O
in	O
the	O
α6	O
helices	O
(	O
top	O
).	O
	O
4	O
-	O
HPA	O
stabilizes	O
concerted	O
conformational	O
changes	O
in	O
NadR	O
that	O
prevent	O
DNA	O
-	O
binding	O
	O
To	O
further	O
investigate	O
the	O
conformational	O
rearrangements	O
of	O
NadR	O
,	O
we	O
performed	O
local	O
structural	O
alignments	O
using	O
only	O
a	O
subset	O
of	O
residues	O
in	O
the	O
DNA	O
-	O
binding	O
helix	O
(	O
α4	O
).	O
	O
By	O
selecting	O
and	O
aligning	O
residues	O
Arg64	O
-	O
Ala77	O
of	O
one	O
α4	O
helix	O
per	O
dimer	O
,	O
superposition	O
of	O
the	O
holo	O
-	O
homodimer	O
onto	O
the	O
two	O
apo	O
-	O
homodimers	O
revealed	O
differences	O
in	O
the	O
monomer	O
conformations	O
of	O
each	O
structure	O
.	O
	O
While	O
one	O
monomer	O
from	O
each	O
structure	O
was	O
closely	O
superimposable	O
(	O
Fig	O
8A	O
,	O
left	O
side	O
),	O
the	O
second	O
monomer	O
displayed	O
quite	O
large	O
differences	O
(	O
Fig	O
8A	O
,	O
right	O
side	O
).	O
	O
Most	O
notably	O
,	O
the	O
position	O
of	O
the	O
DNA	O
-	O
binding	O
helix	O
α4	O
shifted	O
by	O
as	O
much	O
as	O
6	O
Å	O
(	O
Fig	O
8B	O
).	O
	O
Accordingly	O
,	O
helix	O
α4	O
was	O
also	O
found	O
to	O
be	O
one	O
of	O
the	O
most	O
dynamic	O
regions	O
in	O
previous	O
HDX	O
-	O
MS	O
analyses	O
of	O
apo	O
-	O
NadR	O
in	O
solution	O
.	O
	O
The	O
α4	O
helices	O
aligned	O
closely	O
,	O
Cα	O
rmsd	O
0	O
.	O
2Å	O
for	O
14	O
residues	O
.	O
	O
(	O
B	O
)	O
The	O
relative	O
positions	O
of	O
the	O
α4	O
helices	O
of	O
the	O
4	O
-	O
HPA	O
-	O
bound	O
holo	O
homodimer	O
chain	O
B	O
(	O
blue	O
),	O
and	O
of	O
apo	O
homodimers	O
AB	O
and	O
CD	O
(	O
showing	O
chains	O
B	O
and	O
D	O
)	O
in	O
pale	O
blue	O
.	O
	O
Dashes	O
indicate	O
the	O
Ala77	O
Cα	O
atoms	O
,	O
in	O
the	O
most	O
highly	O
shifted	O
region	O
of	O
the	O
‘	O
non	O
-	O
fixed	O
’	O
α4	O
helix	O
.	O
	O
(	O
C	O
)	O
The	O
double	O
-	O
stranded	O
DNA	O
molecule	O
(	O
grey	O
cartoon	O
)	O
from	O
the	O
OhrR	O
-	O
ohrA	O
complex	O
is	O
shown	O
after	O
superposition	O
with	O
NadR	O
,	O
to	O
highlight	O
the	O
expected	O
positions	O
of	O
the	O
NadR	O
α4	O
helices	O
in	O
the	O
DNA	O
major	O
grooves	O
.	O
	O
For	O
clarity	O
,	O
only	O
the	O
α4	O
helices	O
are	O
shown	O
in	O
panels	O
(	O
B	O
)	O
and	O
(	O
C	O
).	O
(	O
D	O
)	O
Upon	O
comparison	O
with	O
the	O
experimentally	O
-	O
determined	O
OhrR	O
:	O
ohrA	O
structure	O
(	O
grey	O
),	O
the	O
α4	O
helix	O
of	O
holo	O
-	O
NadR	O
(	O
blue	O
)	O
is	O
shifted	O
~	O
8Å	O
out	O
of	O
the	O
major	O
groove	O
.	O
	O
In	O
summary	O
,	O
compared	O
to	O
ligand	O
-	O
stabilized	O
holo	O
-	O
NadR	O
,	O
apo	O
-	O
NadR	O
displayed	O
an	O
intrinsic	O
flexibility	O
focused	O
in	O
the	O
DNA	O
-	O
binding	O
region	O
.	O
	O
This	O
was	O
also	O
evident	O
in	O
the	O
greater	O
disorder	O
(	O
i	O
.	O
e	O
.	O
less	O
well	O
-	O
defined	O
electron	O
density	O
)	O
in	O
the	O
β1	O
-	O
β2	O
loops	O
of	O
the	O
apo	O
dimers	O
(	O
density	O
for	O
16	O
residues	O
per	O
dimer	O
was	O
missing	O
)	O
compared	O
to	O
the	O
holo	O
dimer	O
(	O
density	O
for	O
only	O
3	O
residues	O
was	O
missing	O
).	O
	O
In	O
holo	O
-	O
NadR	O
,	O
the	O
distance	O
separating	O
the	O
two	O
DNA	O
-	O
binding	O
α4	O
helices	O
was	O
32	O
Å	O
,	O
while	O
in	O
apo	O
-	O
NadR	O
it	O
was	O
29	O
Å	O
for	O
homodimer	O
AB	O
,	O
and	O
34	O
Å	O
for	O
homodimer	O
CD	O
(	O
Fig	O
8C	O
).	O
	O
Pairwise	O
superpositions	O
showed	O
that	O
the	O
NadR	O
apo	O
-	O
homodimer	O
AB	O
was	O
the	O
most	O
similar	O
to	O
OhrR	O
(	O
rmsd	O
2	O
.	O
6	O
Å	O
),	O
while	O
the	O
holo	O
-	O
homodimer	O
was	O
the	O
most	O
divergent	O
(	O
rmsd	O
3	O
.	O
3	O
Å	O
)	O
(	O
Fig	O
8C	O
).	O
	O
Assuming	O
the	O
same	O
DNA	O
-	O
binding	O
mechanism	O
is	O
used	O
by	O
OhrR	O
and	O
NadR	O
,	O
the	O
apo	O
-	O
homodimer	O
AB	O
seems	O
ideally	O
pre	O
-	O
configured	O
for	O
DNA	O
binding	O
,	O
while	O
4	O
-	O
HPA	O
appeared	O
to	O
stabilize	O
holo	O
-	O
NadR	O
in	O
a	O
conformation	O
poorly	O
suited	O
for	O
DNA	O
binding	O
.	O
	O
Specifically	O
,	O
in	O
addition	O
to	O
the	O
different	O
inter	O
-	O
helical	O
translational	O
distances	O
,	O
the	O
α4	O
helices	O
in	O
the	O
holo	O
-	O
NadR	O
homodimer	O
were	O
also	O
reoriented	O
,	O
resulting	O
in	O
movement	O
of	O
α4	O
out	O
of	O
the	O
major	O
groove	O
,	O
by	O
up	O
to	O
8Å	O
,	O
and	O
presumably	O
preventing	O
efficient	O
DNA	O
binding	O
in	O
the	O
presence	O
of	O
4	O
-	O
HPA	O
(	O
Fig	O
8D	O
).	O
	O
NadR	O
residues	O
His7	O
,	O
Ser9	O
,	O
Asn11	O
and	O
Phe25	O
are	O
essential	O
for	O
regulation	O
of	O
NadA	O
expression	O
in	O
vivo	O
	O
While	O
previous	O
studies	O
had	O
correctly	O
suggested	O
the	O
involvement	O
of	O
several	O
NadR	O
residues	O
in	O
ligand	O
binding	O
,	O
the	O
crystal	O
structures	O
presented	O
here	O
revealed	O
additional	O
residues	O
with	O
previously	O
unknown	O
roles	O
in	O
dimerization	O
and	O
/	O
or	O
binding	O
to	O
4	O
-	O
HPA	O
.	O
	O
To	O
explore	O
the	O
functional	O
involvement	O
of	O
these	O
residues	O
,	O
we	O
characterized	O
the	O
behavior	O
of	O
four	O
new	O
NadR	O
mutants	O
(	O
H7A	B-mutant
,	O
S9A	B-mutant
,	O
N11A	B-mutant
and	O
F25A	B-mutant
)	O
in	O
an	O
in	O
vivo	O
assay	O
using	O
the	O
previously	O
described	O
MC58	B-mutant
-	I-mutant
Δ1843	I-mutant
nadR	O
-	O
null	O
mutant	O
strain	O
,	O
which	O
was	O
complemented	O
either	O
by	O
wild	O
-	O
type	O
nadR	O
or	O
by	O
the	O
nadR	O
mutants	O
.	O
	O
NadA	O
protein	O
abundance	O
levels	O
were	O
assessed	O
by	O
Western	O
blotting	O
to	O
evaluate	O
the	O
ability	O
of	O
the	O
NadR	O
mutants	O
to	O
repress	O
the	O
nadA	O
promoter	O
,	O
in	O
the	O
presence	O
or	O
absence	O
of	O
4	O
-	O
HPA	O
.	O
	O
The	O
nadR	O
H7A	B-mutant
,	O
S9A	B-mutant
and	O
F25A	B-mutant
complemented	O
strains	O
showed	O
hyper	O
-	O
repression	O
of	O
nadA	O
expression	O
in	O
vivo	O
,	O
i	O
.	O
e	O
.	O
these	O
mutants	O
repressed	O
nadA	O
more	O
efficiently	O
than	O
the	O
NadR	O
WT	O
protein	O
,	O
either	O
in	O
the	O
presence	O
or	O
absence	O
of	O
4	O
-	O
HPA	O
,	O
while	O
complementation	O
with	O
wild	O
-	O
type	O
nadR	O
resulted	O
in	O
high	O
production	O
of	O
NadA	O
only	O
in	O
the	O
presence	O
of	O
4	O
-	O
HPA	O
(	O
Fig	O
9	O
).	O
	O
Interestingly	O
,	O
and	O
on	O
the	O
contrary	O
,	O
the	O
nadR	O
N11A	B-mutant
complemented	O
strain	O
showed	O
hypo	O
-	O
repression	O
(	O
i	O
.	O
e	O
.	O
exhibited	O
high	O
expression	O
of	O
nadA	O
both	O
in	O
absence	O
and	O
presence	O
of	O
4	O
-	O
HPA	O
).	O
	O
Structure	O
-	O
based	O
point	O
mutations	O
shed	O
light	O
on	O
ligand	O
-	O
induced	O
regulation	O
of	O
NadR	O
.	O
	O
Complementation	O
of	O
ΔNadR	B-mutant
with	O
WT	O
NadR	O
enables	O
induction	O
of	O
nadA	O
expression	O
by	O
4	O
-	O
HPA	O
.	O
	O
A	O
detailed	O
understanding	O
of	O
the	O
in	O
vitro	O
repression	O
of	O
nadA	O
expression	O
by	O
the	O
transcriptional	O
regulator	O
NadR	O
is	O
important	O
,	O
both	O
because	O
it	O
is	O
a	O
relevant	O
disease	O
-	O
related	O
model	O
of	O
how	O
small	O
-	O
molecule	O
ligands	O
can	O
regulate	O
MarR	O
family	O
proteins	O
and	O
thereby	O
impact	O
bacterial	O
virulence	O
,	O
and	O
because	O
nadA	O
expression	O
levels	O
are	O
linked	O
to	O
the	O
prediction	O
of	O
vaccine	O
coverage	O
.	O
	O
The	O
repressive	O
activity	O
of	O
NadR	O
can	O
be	O
relieved	O
by	O
hydroxyphenylacetate	O
(	O
HPA	O
)	O
ligands	O
,	O
and	O
HDX	O
-	O
MS	O
studies	O
previously	O
indicated	O
that	O
4	O
-	O
HPA	O
stabilizes	O
dimeric	O
NadR	O
in	O
a	O
configuration	O
incompatible	O
with	O
DNA	O
binding	O
.	O
	O
Despite	O
these	O
and	O
other	O
studies	O
,	O
the	O
molecular	O
mechanisms	O
by	O
which	O
ligands	O
regulate	O
MarR	O
family	O
proteins	O
are	O
relatively	O
poorly	O
understood	O
and	O
likely	O
differ	O
depending	O
on	O
the	O
specific	O
ligand	O
.	O
	O
Firstly	O
,	O
we	O
confirmed	O
that	O
NadR	O
is	O
dimeric	O
in	O
solution	O
and	O
demonstrated	O
that	O
it	O
retains	O
its	O
dimeric	O
state	O
in	O
the	O
presence	O
of	O
4	O
-	O
HPA	O
,	O
indicating	O
that	O
induction	O
of	O
a	O
monomeric	O
status	O
is	O
not	O
the	O
manner	O
by	O
which	O
4	O
-	O
HPA	O
regulates	O
NadR	O
.	O
These	O
observations	O
were	O
in	O
agreement	O
with	O
(	O
i	O
)	O
a	O
previous	O
study	O
of	O
NadR	O
performed	O
using	O
SEC	O
and	O
mass	O
spectrometry	O
,	O
and	O
(	O
ii	O
)	O
crystallographic	O
studies	O
showing	O
that	O
several	O
MarR	O
homologues	O
are	O
dimeric	O
.	O
	O
We	O
also	O
used	O
structure	O
-	O
guided	O
site	O
-	O
directed	O
mutagenesis	O
to	O
identify	O
an	O
important	O
conserved	O
residue	O
,	O
Leu130	O
,	O
which	O
stabilizes	O
the	O
NadR	O
dimer	O
interface	O
,	O
knowledge	O
of	O
which	O
may	O
also	O
inform	O
future	O
studies	O
to	O
explore	O
the	O
regulatory	O
mechanisms	O
of	O
other	O
MarR	O
family	O
proteins	O
.	O
	O
Secondly	O
,	O
we	O
assessed	O
the	O
thermal	O
stability	O
and	O
unfolding	O
of	O
NadR	O
in	O
the	O
presence	O
or	O
absence	O
of	O
ligands	O
.	O
	O
All	O
DSC	O
profiles	O
showed	O
a	O
single	O
peak	O
,	O
suggesting	O
that	O
a	O
single	O
unfolding	O
event	O
simultaneously	O
disrupted	O
the	O
dimer	O
and	O
the	O
monomer	O
.	O
	O
HPA	O
ligands	O
specifically	O
increased	O
the	O
stability	O
of	O
NadR	O
.	O
The	O
largest	O
effects	O
were	O
induced	O
by	O
the	O
naturally	O
-	O
occurring	O
compounds	O
4	O
-	O
HPA	O
and	O
3Cl	O
,	O
4	O
-	O
HPA	O
,	O
which	O
,	O
in	O
SPR	O
assays	O
,	O
were	O
found	O
to	O
bind	O
NadR	O
with	O
KD	O
values	O
of	O
1	O
.	O
5	O
mM	O
and	O
1	O
.	O
1	O
mM	O
,	O
respectively	O
.	O
	O
Although	O
these	O
NadR	O
/	O
HPA	O
interactions	O
appeared	O
rather	O
weak	O
,	O
their	O
distinct	O
affinities	O
and	O
specificities	O
matched	O
their	O
in	O
vitro	O
effects	O
and	O
their	O
biological	O
relevance	O
appears	O
similar	O
to	O
previous	O
proposals	O
that	O
certain	O
small	O
molecules	O
,	O
including	O
some	O
antibiotics	O
,	O
in	O
the	O
millimolar	O
concentration	O
range	O
may	O
be	O
broad	O
inhibitors	O
of	O
MarR	O
family	O
proteins	O
.	O
	O
Indeed	O
,	O
4	O
-	O
HPA	O
is	O
found	O
in	O
human	O
saliva	O
and	O
3Cl	O
,	O
4	O
-	O
HPA	O
is	O
produced	O
during	O
inflammatory	O
processes	O
,	O
suggesting	O
that	O
these	O
natural	O
ligands	O
are	O
encountered	O
by	O
N	O
.	O
meningitidis	O
in	O
the	O
mucosa	O
of	O
the	O
oropharynx	O
during	O
infections	O
.	O
	O
It	O
is	O
also	O
possible	O
that	O
NadR	O
responds	O
to	O
currently	O
unidentified	O
HPA	O
analogues	O
.	O
	O
Indeed	O
,	O
in	O
the	O
NadR	O
/	O
4	O
-	O
HPA	O
complex	O
there	O
was	O
a	O
water	O
molecule	O
close	O
to	O
the	O
carboxylate	O
group	O
and	O
also	O
a	O
small	O
unfilled	O
tunnel	O
~	O
5Å	O
long	O
,	O
both	O
factors	O
suggesting	O
that	O
alternative	O
larger	O
ligands	O
could	O
occupy	O
the	O
pocket	O
.	O
	O
The	O
ability	O
to	O
respond	O
to	O
various	O
ligands	O
might	O
enable	O
NadR	O
in	O
vivo	O
to	O
orchestrate	O
multiple	O
response	O
mechanisms	O
and	O
modulate	O
expression	O
of	O
genes	O
other	O
than	O
nadA	O
.	O
Ultimately	O
,	O
confirmation	O
of	O
the	O
relevance	O
of	O
each	O
ligand	O
will	O
require	O
a	O
deeper	O
understanding	O
of	O
the	O
available	O
concentration	O
in	O
vivo	O
in	O
the	O
host	O
niche	O
during	O
bacterial	O
colonization	O
and	O
inflammation	O
.	O
	O
Here	O
,	O
we	O
determined	O
the	O
first	O
crystal	O
structures	O
of	O
apo	O
-	O
NadR	O
and	O
holo	O
-	O
NadR	O
.	O
These	O
experimentally	O
-	O
determined	O
structures	O
enabled	O
a	O
new	O
detailed	O
characterization	O
of	O
the	O
ligand	O
-	O
binding	O
pocket	O
.	O
	O
In	O
holo	O
-	O
NadR	O
,	O
4	O
-	O
HPA	O
interacted	O
directly	O
with	O
at	O
least	O
11	O
polar	O
and	O
hydrophobic	O
residues	O
.	O
	O
Several	O
,	O
but	O
not	O
all	O
,	O
of	O
these	O
interactions	O
were	O
predicted	O
previously	O
by	O
homology	O
modelling	O
combined	O
with	O
ligand	O
docking	O
in	O
silico	O
.	O
	O
Subsequently	O
,	O
we	O
established	O
the	O
functional	O
importance	O
of	O
His7	O
,	O
Ser9	O
,	O
Asn11	O
and	O
Phe25	O
in	O
the	O
in	O
vitro	O
response	O
of	O
meningococcus	O
to	O
4	O
-	O
HPA	O
,	O
via	O
site	O
-	O
directed	O
mutagenesis	O
.	O
	O
More	O
unexpectedly	O
,	O
the	O
crystal	O
structure	O
revealed	O
that	O
only	O
one	O
molecule	O
of	O
4	O
-	O
HPA	O
was	O
bound	O
per	O
NadR	O
dimer	O
.	O
	O
We	O
also	O
used	O
heteronuclear	O
NMR	O
spectroscopy	O
to	O
detect	O
substantial	O
conformational	O
changes	O
of	O
NadR	O
occurring	O
in	O
solution	O
upon	O
addition	O
of	O
4	O
-	O
HPA	O
.	O
	O
Moreover	O
,	O
NMR	O
spectra	O
at	O
10	O
°	O
C	O
suggested	O
the	O
existence	O
of	O
two	O
distinct	O
conformations	O
of	O
NadR	O
in	O
the	O
vicinity	O
of	O
the	O
ligand	O
-	O
binding	O
pocket	O
.	O
	O
More	O
powerfully	O
,	O
our	O
unique	O
crystallographic	O
observation	O
of	O
this	O
‘	O
occupied	O
vs	O
unoccupied	O
site	O
’	O
asymmetry	O
in	O
the	O
NadR	O
/	O
4	O
-	O
HPA	O
interaction	O
is	O
,	O
to	O
our	O
knowledge	O
,	O
the	O
first	O
example	O
reported	O
for	O
a	O
MarR	O
family	O
protein	O
.	O
	O
Such	O
a	O
mechanism	O
indicates	O
negative	O
cooperativity	O
,	O
which	O
may	O
enhance	O
the	O
ligand	O
-	O
responsiveness	O
of	O
NadR	O
.	O
	O
Comparisons	O
of	O
the	O
NadR	O
/	O
4	O
-	O
HPA	O
complex	O
with	O
available	O
MarR	O
family	O
/	O
salicylate	O
complexes	O
revealed	O
that	O
4	O
-	O
HPA	O
has	O
a	O
previously	O
unobserved	O
binding	O
mode	O
.	O
	O
Briefly	O
,	O
in	O
the	O
M	O
.	O
thermoautotrophicum	O
MTH313	O
dimer	O
,	O
one	O
molecule	O
of	O
salicylate	O
binds	O
in	O
the	O
pocket	O
of	O
each	O
monomer	O
,	O
though	O
with	O
two	O
rather	O
different	O
positions	O
and	O
orientations	O
,	O
only	O
one	O
of	O
which	O
(	O
site	O
-	O
1	O
)	O
is	O
thought	O
to	O
be	O
biologically	O
relevant	O
(	O
Fig	O
10A	O
).	O
	O
In	O
NadR	O
,	O
the	O
single	O
molecule	O
of	O
4	O
-	O
HPA	O
binds	O
in	O
a	O
position	O
distinctly	O
different	O
from	O
the	O
salicylate	O
binding	O
site	O
:	O
translated	O
by	O
>	O
10	O
Å	O
and	O
with	O
a	O
180	O
°	O
inverted	O
orientation	O
(	O
Fig	O
10C	O
).	O
	O
(	O
A	O
)	O
A	O
structural	O
alignment	O
of	O
MTH313	O
chains	O
A	O
and	O
B	O
shows	O
that	O
salicylate	O
is	O
bound	O
in	O
distinct	O
locations	O
in	O
each	O
monomer	O
;	O
site	O
-	O
1	O
(	O
thought	O
to	O
be	O
the	O
biologically	O
relevant	O
site	O
)	O
and	O
site	O
-	O
2	O
differ	O
by	O
~	O
7Å	O
(	O
indicated	O
by	O
black	O
dotted	O
line	O
)	O
and	O
also	O
by	O
ligand	O
orientation	O
.	O
	O
(	O
C	O
)	O
Addition	O
of	O
holo	O
-	O
NadR	O
(	O
chain	O
B	O
,	O
blue	O
)	O
to	O
the	O
alignment	O
reveals	O
that	O
bound	O
4	O
-	O
HPA	O
differs	O
in	O
position	O
by	O
>	O
10	O
Å	O
compared	O
to	O
salicylate	O
,	O
and	O
adopts	O
a	O
novel	O
orientation	O
.	O
	O
Interestingly	O
,	O
a	O
crystal	O
structure	O
was	O
previously	O
reported	O
for	O
a	O
functionally	O
-	O
uncharacterized	O
meningococcal	O
homologue	O
of	O
NadR	O
,	O
termed	O
NMB1585	O
,	O
which	O
shares	O
16	O
%	O
sequence	O
identity	O
with	O
NadR	O
.	O
The	O
two	O
structures	O
can	O
be	O
closely	O
aligned	O
(	O
rmsd	O
2	O
.	O
3	O
Å	O
),	O
but	O
NMB1585	O
appears	O
unsuited	O
for	O
binding	O
HPAs	O
,	O
since	O
its	O
corresponding	O
‘	O
pocket	O
’	O
region	O
is	O
occupied	O
by	O
several	O
bulky	O
hydrophobic	O
side	O
chains	O
.	O
	O
It	O
can	O
be	O
speculated	O
that	O
MarR	O
family	O
members	O
have	O
evolved	O
separately	O
to	O
engage	O
distinct	O
signaling	O
molecules	O
,	O
thus	O
enabling	O
bacteria	O
to	O
use	O
the	O
overall	O
conserved	O
MarR	O
scaffold	O
to	O
adapt	O
and	O
respond	O
to	O
diverse	O
changing	O
environmental	O
stimuli	O
experienced	O
in	O
their	O
natural	O
niches	O
.	O
	O
The	O
apo	O
-	O
NadR	O
crystal	O
structures	O
revealed	O
two	O
dimers	O
with	O
slightly	O
different	O
conformations	O
,	O
most	O
divergent	O
in	O
the	O
DNA	O
-	O
binding	O
domain	O
.	O
	O
It	O
is	O
not	O
unusual	O
for	O
a	O
crystal	O
structure	O
to	O
reveal	O
multiple	O
copies	O
of	O
the	O
same	O
protein	O
in	O
very	O
slightly	O
different	O
conformations	O
,	O
which	O
are	O
likely	O
representative	O
of	O
the	O
lowest	O
-	O
energy	O
conformations	O
sampled	O
by	O
the	O
dynamic	O
ensemble	O
of	O
molecular	O
states	O
occurring	O
in	O
solution	O
,	O
and	O
which	O
likely	O
have	O
only	O
small	O
energetic	O
differences	O
,	O
as	O
described	O
previously	O
for	O
MexR	O
(	O
a	O
MarR	O
protein	O
)	O
or	O
more	O
recently	O
for	O
the	O
solute	O
-	O
binding	O
protein	O
FhuD2	O
.	O
	O
Further	O
,	O
the	O
holo	O
-	O
NadR	O
structure	O
was	O
overall	O
more	O
different	O
from	O
the	O
two	O
apo	O
-	O
NadR	O
structures	O
(	O
rmsd	O
values	O
~	O
1	O
.	O
3Å	O
),	O
suggesting	O
that	O
the	O
ligand	O
selected	O
and	O
stabilized	O
yet	O
another	O
conformation	O
of	O
NadR	O
.	O
These	O
observations	O
suggest	O
that	O
4	O
-	O
HPA	O
,	O
and	O
potentially	O
other	O
similar	O
ligands	O
,	O
can	O
shift	O
the	O
molecular	O
equilibrium	O
,	O
changing	O
the	O
energy	O
barriers	O
that	O
separate	O
active	O
and	O
inactive	O
states	O
,	O
and	O
stabilizing	O
the	O
specific	O
conformation	O
of	O
NadR	O
poorly	O
suited	O
to	O
bind	O
DNA	O
.	O
	O
Comparisons	O
of	O
the	O
apo	O
-	O
and	O
holo	O
-	O
NadR	O
structures	O
revealed	O
that	O
the	O
largest	O
differences	O
occurred	O
in	O
the	O
DNA	O
-	O
binding	O
helix	O
α4	O
.	O
	O
The	O
shift	O
of	O
helix	O
α4	O
in	O
holo	O
-	O
NadR	O
was	O
also	O
accompanied	O
by	O
rearrangements	O
at	O
the	O
dimer	O
interface	O
,	O
involving	O
helices	O
α1	O
,	O
α5	O
,	O
and	O
α6	O
,	O
and	O
this	O
holo	O
-	O
form	O
appeared	O
poorly	O
suited	O
for	O
DNA	O
-	O
binding	O
when	O
compared	O
with	O
the	O
known	O
OhrR	O
:	O
ohrA	O
complex	O
.	O
	O
One	O
of	O
the	O
two	O
conformations	O
of	O
apo	O
-	O
NadR	O
appeared	O
ideally	O
suited	O
for	O
DNA	O
-	O
binding	O
.	O
	O
Overall	O
,	O
these	O
analyses	O
suggest	O
that	O
the	O
apo	O
-	O
NadR	O
dimer	O
has	O
a	O
pre	O
-	O
existing	O
equilibrium	O
that	O
samples	O
a	O
variety	O
of	O
conformations	O
,	O
some	O
of	O
which	O
are	O
compatible	O
with	O
DNA	O
binding	O
.	O
	O
Subsequently	O
,	O
upon	O
ligand	O
binding	O
,	O
holo	O
-	O
NadR	O
adopts	O
a	O
structure	O
less	O
suited	O
for	O
DNA	O
-	O
binding	O
and	O
this	O
conformation	O
is	O
selected	O
and	O
stabilized	O
by	O
a	O
network	O
of	O
protein	O
-	O
ligand	O
interactions	O
and	O
concomitant	O
rearrangements	O
at	O
the	O
NadR	O
holo	O
dimer	O
interface	O
.	O
	O
In	O
an	O
alternative	O
and	O
less	O
extensive	O
manner	O
,	O
the	O
binding	O
of	O
two	O
salicylate	O
molecules	O
to	O
the	O
M	O
.	O
thermoautotrophicum	O
protein	O
MTH313	O
appeared	O
to	O
induce	O
large	O
changes	O
in	O
the	O
wHTH	O
domain	O
,	O
which	O
was	O
associated	O
with	O
reduced	O
DNA	O
-	O
binding	O
activity	O
.	O
	O
Here	O
we	O
have	O
presented	O
two	O
new	O
crystal	O
structures	O
of	O
the	O
transcription	O
factor	O
,	O
NadR	O
,	O
which	O
regulates	O
expression	O
of	O
the	O
meningococcal	O
surface	O
protein	O
,	O
virulence	O
factor	O
and	O
vaccine	O
antigen	O
NadA	O
.	O
Detailed	O
structural	O
analyses	O
provided	O
a	O
molecular	O
explanation	O
for	O
the	O
ligand	O
-	O
responsive	O
regulation	O
by	O
NadR	O
on	O
the	O
majority	O
of	O
the	O
promoters	O
of	O
meningococcal	O
genes	O
regulated	O
by	O
NadR	O
,	O
including	O
nadA	O
.	O
Intriguingly	O
,	O
NadR	O
exhibits	O
a	O
reversed	O
regulatory	O
mechanism	O
on	O
a	O
second	O
class	O
of	O
promoters	O
,	O
including	O
mafA	O
of	O
the	O
multiple	O
adhesin	O
family	O
–	O
i	O
.	O
e	O
.	O
NadR	O
represses	O
these	O
genes	O
in	O
the	O
presence	O
but	O
not	O
absence	O
of	O
4	O
-	O
HPA	O
.	O
	O
Ultimately	O
,	O
knowledge	O
of	O
the	O
ligand	O
-	O
dependent	O
activity	O
of	O
NadR	O
will	O
continue	O
to	O
deepen	O
our	O
understanding	O
of	O
nadA	O
expression	O
levels	O
,	O
which	O
influence	O
meningococcal	O
pathogenesis	O
.	O
	O
The	O
structure	O
of	O
NMB1585	O
,	O
a	O
MarR	O
-	O
family	O
regulator	O
from	O
Neisseria	O
meningitidis	O
	O
The	O
nuclear	O
hormone	O
receptor	O
RORγ	O
regulates	O
transcriptional	O
genes	O
involved	O
in	O
the	O
production	O
of	O
the	O
pro	O
-	O
inflammatory	O
interleukin	O
IL	O
-	O
17	O
which	O
has	O
been	O
linked	O
to	O
autoimmune	O
diseases	O
such	O
as	O
rheumatoid	O
arthritis	O
,	O
multiple	O
sclerosis	O
and	O
inflammatory	O
bowel	O
disease	O
.	O
	O
This	O
transcriptional	O
activity	O
of	O
RORγ	O
is	O
modulated	O
through	O
a	O
protein	O
-	O
protein	O
interaction	O
involving	O
the	O
activation	O
function	O
2	O
(	O
AF2	O
)	O
helix	O
on	O
the	O
ligand	O
binding	O
domain	O
of	O
RORγ	O
and	O
a	O
conserved	O
LXXLL	O
helix	O
motif	O
on	O
coactivator	O
proteins	O
.	O
	O
We	O
identified	O
a	O
novel	O
series	O
of	O
synthetic	O
benzoxazinone	O
ligands	O
having	O
an	O
agonist	O
(	O
BIO592	O
)	O
and	O
inverse	O
agonist	O
(	O
BIO399	O
)	O
mode	O
of	O
action	O
in	O
a	O
FRET	O
based	O
assay	O
.	O
	O
We	O
show	O
that	O
the	O
AF2	O
helix	O
of	O
RORγ	O
is	O
proteolytically	O
sensitive	O
when	O
inverse	O
agonist	O
BIO399	O
binds	O
.	O
	O
Using	O
x	O
-	O
ray	O
crystallography	O
we	O
show	O
how	O
small	O
modifications	O
on	O
the	O
benzoxazinone	O
agonist	O
BIO592	O
trigger	O
inverse	O
agonism	O
of	O
RORγ	O
.	O
	O
The	O
proteolytic	O
sensitivity	O
of	O
the	O
AF2	O
helix	O
of	O
RORγ	O
demonstrates	O
that	O
it	O
destabilizes	O
upon	O
BIO399	O
inverse	O
agonist	O
binding	O
perturbing	O
the	O
coactivator	O
protein	O
binding	O
site	O
.	O
	O
Even	O
though	O
a	O
high	O
degree	O
of	O
sequence	O
similarity	O
exists	O
between	O
the	O
RORs	O
,	O
their	O
functional	O
roles	O
in	O
regulation	O
for	O
physiological	O
processes	O
involved	O
in	O
development	O
and	O
immunity	O
are	O
distinct	O
.	O
	O
During	O
development	O
,	O
RORγ	O
regulates	O
the	O
transcriptional	O
genes	O
involved	O
in	O
the	O
functioning	O
of	O
multiple	O
pro	O
-	O
inflammatory	O
lymphocyte	O
lineages	O
including	O
T	O
helper	O
cells	O
(	O
TH17cells	O
)	O
which	O
are	O
necessary	O
for	O
IL	O
-	O
17	O
production	O
.	O
	O
IL	O
-	O
17	O
is	O
a	O
pro	O
-	O
inflammatory	O
interleukin	O
linked	O
to	O
autoimmune	O
diseases	O
such	O
as	O
rheumatoid	O
arthritis	O
,	O
multiple	O
sclerosis	O
and	O
inflammatory	O
bowel	O
disease	O
;	O
making	O
its	O
transcriptional	O
regulation	O
through	O
RORγ	O
an	O
attractive	O
therapeutic	O
target	O
.	O
	O
RORγ	O
consists	O
of	O
an	O
N	O
-	O
terminal	O
DNA	O
binding	O
domain	O
(	O
DBD	O
)	O
connected	O
to	O
a	O
C	O
-	O
terminal	O
ligand	O
binding	O
domain	O
(	O
LBD	O
)	O
via	O
a	O
flexible	O
hinge	O
region	O
.	O
	O
The	O
DBD	O
is	O
composed	O
of	O
two	O
zinc	O
fingers	O
that	O
allow	O
it	O
to	O
interact	O
with	O
specifically	O
encoded	O
regions	O
on	O
the	O
DNA	O
called	O
the	O
nuclear	O
receptor	O
response	O
elements	O
.	O
	O
The	O
LBD	O
consists	O
of	O
a	O
coactivator	O
protein	O
binding	O
pocket	O
and	O
a	O
hydrophobic	O
ligand	O
binding	O
site	O
(	O
LBS	O
)	O
which	O
are	O
responsible	O
for	O
regulating	O
transcription	O
.	O
	O
The	O
coactivator	O
binding	O
pocket	O
of	O
RORγ	O
recognizes	O
a	O
conserved	O
helix	O
motif	O
LXXLL	O
(	O
where	O
X	O
can	O
be	O
any	O
amino	O
acid	O
)	O
on	O
transcriptional	O
coactivator	O
complexes	O
and	O
recruits	O
it	O
to	O
activate	O
transcription	O
.	O
	O
In	O
RORγ	O
,	O
the	O
conformation	O
of	O
the	O
AF2	O
helix	O
required	O
to	O
form	O
the	O
coactivator	O
binding	O
pocket	O
is	O
mediated	O
by	O
a	O
salt	O
bridge	O
between	O
His479	O
and	O
Tyr502	O
in	O
addition	O
to	O
π	O
-	O
π	O
interactions	O
between	O
Tyr502	O
and	O
Phe506	O
.	O
	O
The	O
conformation	O
of	O
the	O
AF2	O
helix	O
can	O
be	O
modulated	O
through	O
targeted	O
ligands	O
which	O
bind	O
the	O
LBS	O
and	O
increase	O
the	O
binding	O
of	O
the	O
coactivator	O
protein	O
(	O
agonists	O
)	O
or	O
disrupt	O
binding	O
(	O
inverse	O
agonists	O
)	O
thereby	O
enhancing	O
or	O
inhibiting	O
transcription	O
.	O
	O
Since	O
RORγ	O
has	O
been	O
demonstrated	O
to	O
play	O
an	O
important	O
role	O
in	O
pro	O
-	O
inflammatory	O
gene	O
expression	O
patterns	O
implicated	O
in	O
several	O
major	O
autoimmune	O
diseases	O
,	O
our	O
aim	O
was	O
to	O
develop	O
RORγ	O
inverse	O
agonists	O
that	O
would	O
help	O
down	O
regulate	O
pro	O
-	O
inflammatory	O
gene	O
transcription	O
.	O
	O
Finally	O
,	O
comparing	O
binding	O
modes	O
of	O
our	O
benzoxazinone	O
RORγ	O
crystal	O
structures	O
to	O
other	O
ROR	O
structures	O
,	O
we	O
hypothesize	O
a	O
new	O
mode	O
of	O
action	O
for	O
achieving	O
inverse	O
agonism	O
and	O
selectivity	O
.	O
	O
Interestingly	O
,	O
the	O
structural	O
difference	O
between	O
the	O
agonist	O
BIO592	O
and	O
inverse	O
agonist	O
BIO399	O
was	O
minor	O
;	O
with	O
the	O
2	O
,	O
3	O
-	O
dihydrobenzo	O
[	O
1	O
,	O
4	O
]	O
oxazepin	O
-	O
4	O
-	O
one	O
ring	O
system	O
of	O
BIO399	O
being	O
3	O
atoms	O
larger	O
than	O
the	O
benzo	O
[	O
1	O
,	O
4	O
]	O
oxazine	O
-	O
3	O
-	O
one	O
ring	O
system	O
of	O
BIO592	O
.	O
	O
In	O
order	O
to	O
understand	O
how	O
small	O
changes	O
in	O
the	O
core	O
ring	O
system	O
leads	O
to	O
inverse	O
agonism	O
,	O
we	O
wanted	O
to	O
structurally	O
determine	O
the	O
binding	O
mode	O
of	O
both	O
BIO592	O
and	O
BIO399	O
in	O
the	O
LBS	O
of	O
RORγ	O
using	O
x	O
-	O
ray	O
crystallography	O
.	O
	O
Structure	O
of	O
the	O
RORγ518	O
-	O
BIO592	O
-	O
EBI96	O
ternary	O
complex	O
is	O
in	O
a	O
transcriptionally	O
active	O
conformation	O
	O
RORγ518	O
bound	O
to	O
agonist	O
BIO592	O
was	O
crystallized	O
with	O
a	O
truncated	O
form	O
of	O
the	O
coactivator	O
peptide	O
EBI96	O
to	O
a	O
resolution	O
of	O
2	O
.	O
6	O
Å	O
(	O
Fig	O
.	O
2a	O
).	O
	O
The	O
hydrogen	O
bond	O
between	O
His479	O
and	O
Tyr502	O
has	O
been	O
reported	O
to	O
be	O
critical	O
for	O
RORγ	O
agonist	O
activity	O
.	O
	O
Disrupting	O
this	O
interaction	O
through	O
mutagenesis	O
reduced	O
transcriptional	O
activity	O
of	O
RORγ	O
.	O
	O
This	O
reduced	O
transcriptional	O
activity	O
has	O
been	O
attributed	O
to	O
the	O
inability	O
of	O
the	O
AF2	O
helix	O
to	O
complete	O
the	O
formation	O
of	O
the	O
coactivator	O
binding	O
pocket	O
necessary	O
for	O
coactivator	O
proteins	O
to	O
bind	O
.	O
	O
This	O
interaction	O
is	O
further	O
stabilized	O
through	O
a	O
conserved	O
charged	O
clamp	O
wherein	O
the	O
backbone	O
amide	O
of	O
Tyr7	O
and	O
carbonyl	O
of	O
Leu11	O
of	O
EBI96	O
form	O
hydrogen	O
bonds	O
with	O
Glu504	O
(	O
helix12	O
)	O
and	O
Lys336	O
(	O
helix3	O
)	O
of	O
RORγ	O
.	O
	O
Formation	O
of	O
this	O
charged	O
clamp	O
is	O
essential	O
for	O
RORγ	O
’	O
s	O
function	O
for	O
playing	O
a	O
role	O
in	O
transcriptional	O
activation	O
and	O
this	O
has	O
been	O
corroborated	O
through	O
mutagenic	O
studies	O
in	O
this	O
region	O
.	O
	O
BIO592	O
binds	O
in	O
a	O
collapsed	O
conformation	O
stabilizing	O
the	O
agonist	O
conformation	O
of	O
RORγ	O
	O
a	O
Collapsed	O
binding	O
mode	O
of	O
agonist	O
BIO592	O
in	O
the	O
hydrophobic	O
LBS	O
of	O
RORγ	O
.	O
	O
The	O
sulfonyl	O
group	O
faces	O
the	O
entrance	O
of	O
the	O
pocket	O
,	O
while	O
the	O
CF3	O
makes	O
a	O
hydrophobic	O
contact	O
with	O
Ala327	O
.	O
	O
RORγ	O
AF2	O
helix	O
is	O
sensitive	O
to	O
proteolysis	O
in	O
the	O
presence	O
of	O
Inverse	O
Agonist	O
BIO399	O
	O
Next	O
,	O
we	O
attempted	O
co	O
-	O
crystallization	O
with	O
the	O
inverse	O
agonist	O
BIO399	O
.	O
	O
However	O
,	O
extensive	O
crystallization	O
efforts	O
with	O
BIO399	O
and	O
RORγ518	O
or	O
other	O
AF2	O
intact	O
constructs	O
did	O
not	O
produce	O
crystals	O
.	O
	O
We	O
hypothesized	O
that	O
the	O
RORγ518	O
coactivator	O
peptide	O
interaction	O
in	O
the	O
FRET	O
assay	O
was	O
disrupted	O
upon	O
BIO399	O
binding	O
and	O
that	O
a	O
conformational	O
rearrangement	O
of	O
the	O
AF2	O
helix	O
could	O
have	O
occurred	O
,	O
hindering	O
crystallization	O
.	O
	O
The	O
unfolding	O
of	O
the	O
AF2	O
helix	O
has	O
been	O
observed	O
for	O
other	O
nuclear	O
hormone	O
receptors	O
when	O
bound	O
to	O
an	O
inverse	O
agonist	O
or	O
antagonist	O
.	O
	O
We	O
used	O
partial	O
proteolysis	O
in	O
combination	O
with	O
mass	O
spectrometry	O
to	O
determine	O
if	O
BIO399	O
was	O
causing	O
the	O
AF2	O
helix	O
to	O
unfold	O
.	O
	O
Results	O
of	O
the	O
Actinase	O
E	O
proteolysis	O
experiments	O
on	O
RORγ518	O
,	O
the	O
ternary	O
complex	O
of	O
RORγ518	O
with	O
agonist	O
BIO592	O
and	O
coactivator	O
EBI96	O
,	O
or	O
in	O
the	O
presence	O
of	O
inverse	O
agonist	O
BIO399	O
supported	O
our	O
hypothesis	O
.	O
	O
Analysis	O
of	O
the	O
fragmentation	O
pattern	O
showed	O
minimal	O
proteolytic	O
removal	O
of	O
the	O
AF2	O
helix	O
by	O
Actinase	O
E	O
on	O
RORγ518	O
alone	O
(	O
ending	O
at	O
504	O
to	O
506	O
)	O
and	O
the	O
ternary	O
complex	O
remained	O
primarily	O
intact	O
(	O
ending	O
at	O
515	O
/	O
518	O
)	O
(	O
Additional	O
file	O
4	O
).	O
	O
We	O
attributed	O
the	O
inability	O
to	O
form	O
crystals	O
to	O
the	O
unfolding	O
of	O
the	O
AF2	O
helix	O
induced	O
by	O
BIO399	O
.	O
	O
AF2	O
truncated	O
RORγ	O
BIO399	O
complex	O
is	O
more	O
amenable	O
to	O
crystallization	O
	O
a	O
The	O
binary	O
structure	O
of	O
AF2	O
-	O
truncated	O
RORγ	O
and	O
BIO399	O
.	O
	O
b	O
The	O
superposition	O
of	O
inverse	O
agonist	O
BIO399	O
(	O
Cyan	O
)	O
and	O
agonist	O
BIO592	O
(	O
Green	O
).	O
	O
c	O
Movement	O
of	O
Met358	O
and	O
His479	O
in	O
the	O
BIO399	O
(	O
Cyan	O
)	O
and	O
BIO592	O
(	O
Green	O
)	O
structures	O
	O
The	O
Actinase	O
E	O
treated	O
RORγ518	O
BIO399	O
ternary	O
complex	O
(	O
aeRORγ493	O
/	O
4	O
)	O
co	O
-	O
crystallized	O
readily	O
in	O
several	O
PEG	O
based	O
conditions	O
.	O
	O
The	O
structure	O
of	O
aeRORγ493	O
/	O
4	O
BIO399	O
complex	O
was	O
solved	O
to	O
2	O
.	O
3	O
Å	O
and	O
adopted	O
a	O
similar	O
core	O
fold	O
to	O
the	O
BIO592	O
agonist	O
crystal	O
structure	O
(	O
Fig	O
.	O
5a	O
,	O
Additional	O
file	O
3	O
).	O
	O
Inverse	O
agonist	O
BIO399	O
uses	O
Met358	O
as	O
a	O
trigger	O
for	O
inverse	O
agonism	O
	O
The	O
majority	O
of	O
the	O
side	O
chains	O
within	O
4	O
Å	O
of	O
BIO399	O
and	O
BIO592	O
adopt	O
similar	O
rotomer	O
conformations	O
with	O
the	O
exceptions	O
of	O
Met358	O
and	O
His479	O
(	O
Fig	O
.	O
5c	O
).	O
	O
The	O
difference	O
density	O
map	O
showed	O
clear	O
positive	O
density	O
for	O
Met358	O
in	O
an	O
alternate	O
rotomer	O
conformation	O
compared	O
to	O
the	O
one	O
observed	O
in	O
the	O
molecular	O
replacement	O
model	O
or	O
the	O
other	O
agonist	O
containing	O
models	O
(	O
Additional	O
file	O
6	O
).	O
	O
We	O
tried	O
to	O
refine	O
Met358	O
in	O
the	O
same	O
conformation	O
as	O
the	O
molecular	O
replacement	O
model	O
or	O
the	O
other	O
agonist	O
containing	O
models	O
,	O
but	O
the	O
results	O
clearly	O
indicated	O
that	O
this	O
was	O
not	O
possible	O
,	O
thus	O
confirming	O
the	O
new	O
rotamer	O
conformation	O
for	O
the	O
Met358	O
sidechain	O
in	O
the	O
inverse	O
agonist	O
bound	O
structure	O
.	O
	O
The	O
change	O
in	O
rotomer	O
conformation	O
of	O
Met358	O
between	O
the	O
agonist	O
and	O
inverse	O
agonist	O
structures	O
is	O
attributed	O
to	O
the	O
gem	O
-	O
dimethyl	O
group	O
on	O
the	O
larger	O
7	O
membered	O
benzoxazinone	O
ring	O
system	O
of	O
BIO399	O
.	O
	O
The	O
comparison	O
of	O
the	O
two	O
structures	O
shows	O
that	O
the	O
agonist	O
conformation	O
observed	O
in	O
the	O
BIO592	O
structure	O
would	O
be	O
perturbed	O
by	O
BIO399	O
pushing	O
Met358	O
into	O
Phe506	O
of	O
the	O
AF2	O
helix	O
indicating	O
that	O
Met358	O
is	O
a	O
trigger	O
for	O
inducing	O
inverse	O
agonism	O
in	O
RORγ	O
(	O
Fig	O
.	O
5c	O
).	O
	O
b	O
Overlay	O
of	O
M358	O
in	O
RORγ	O
structure	O
BIO596	O
(	O
Green	O
),	O
BIO399	O
(	O
Cyan	O
),	O
Digoxin	O
(	O
Yellow	O
),	O
Compound	O
2	O
(	O
Grey	O
),	O
Compound	O
48	O
(	O
Salmon	O
)	O
and	O
Compound	O
4j	O
(	O
Orange	O
)	O
	O
The	O
co	O
-	O
crystal	O
structure	O
of	O
RORγ	O
with	O
T0901317	O
(	O
PDB	O
code	O
:	O
4NB6	O
),	O
an	O
inverse	O
agonist	O
of	O
RORγ	O
(	O
IC50	O
of	O
54nM	O
in	O
an	O
SRC1	O
displacement	O
FRET	O
assay	O
and	O
an	O
IC50	O
of	O
59nM	O
in	O
our	O
FRET	O
assay	O
(	O
Additional	O
file	O
7	O
))	O
shows	O
that	O
it	O
adopts	O
a	O
collapsed	O
conformation	O
similar	O
to	O
the	O
structure	O
of	O
BIO399	O
described	O
here	O
.	O
	O
The	O
two	O
compounds	O
superimpose	O
with	O
an	O
RMSD	O
of	O
0	O
.	O
81	O
Å	O
(	O
Fig	O
.	O
6a	O
).	O
	O
The	O
CF3	O
group	O
on	O
the	O
hexafluoropropanol	O
group	O
of	O
T0901317	O
was	O
reported	O
to	O
fit	O
the	O
electron	O
density	O
in	O
two	O
conformations	O
one	O
of	O
which	O
pushes	O
Met358	O
into	O
the	O
vicinity	O
of	O
Phe506	O
in	O
the	O
RORγ	O
BIO592	O
agonist	O
structure	O
.	O
	O
Co	O
-	O
crystal	O
structures	O
of	O
RORγ	O
have	O
been	O
generated	O
with	O
several	O
potent	O
inverse	O
agonists	O
adopting	O
a	O
linear	O
conformation	O
distinct	O
from	O
the	O
collapsed	O
conformations	O
seen	O
for	O
BIO399	O
and	O
T090131718	O
.	O
	O
BIO399	O
neither	O
orients	O
the	O
sidechain	O
of	O
Trp317	O
toward	O
Tyr502	O
nor	O
forms	O
a	O
hydrogen	O
bond	O
with	O
His479	O
suggesting	O
its	O
mode	O
of	O
action	O
is	O
distinct	O
from	O
linear	O
inverse	O
agonists	O
(	O
Additional	O
file	O
8	O
).	O
	O
In	O
the	O
linear	O
inverse	O
agonist	O
crystal	O
structures	O
the	O
side	O
chain	O
of	O
Met358	O
resides	O
in	O
a	O
similar	O
position	O
as	O
the	O
rotomer	O
observed	O
in	O
RORγ	O
agonist	O
structures	O
with	O
BIO592	O
described	O
here	O
or	O
as	O
observed	O
in	O
the	O
hydroxycholesterol	O
derivatives	O
and	O
therefore	O
would	O
not	O
trigger	O
inverse	O
agonism	O
with	O
these	O
ligands	O
(	O
Fig	O
.	O
6b	O
).	O
	O
BIO399	O
shows	O
selectivity	O
for	O
RORγ	O
over	O
RORα	O
and	O
RORβ	O
in	O
a	O
GAL4	O
Cellular	O
Reporter	O
Assay	O
	O
a	O
Overlay	O
of	O
RORα	O
(	O
yellow	O
),	O
β	O
(	O
pink	O
)	O
and	O
γ	O
(	O
cyan	O
)	O
showing	O
side	O
chain	O
differences	O
at	O
Met358	O
inverse	O
agonism	O
trigger	O
position	O
and	O
(	O
b	O
)	O
around	O
the	O
benzoxazinone	O
ring	O
system	O
of	O
BIO399	O
	O
In	O
order	O
to	O
assess	O
the	O
in	O
vivo	O
selectivity	O
profile	O
of	O
BIO399	O
a	O
cellular	O
reporter	O
assay	O
was	O
implemented	O
where	O
the	O
ligand	O
binding	O
domains	O
of	O
ROR	O
α	O
,	O
β	O
and	O
γ	O
were	O
fused	O
to	O
the	O
DNA	O
binding	O
domain	O
of	O
the	O
transcriptional	O
factor	O
GAL4	O
.	O
	O
The	O
ROR	O
-	O
GAL4	O
fusion	O
proteins	O
were	O
expressed	O
in	O
cells	O
with	O
the	O
luciferase	O
reporter	O
gene	O
under	O
the	O
control	O
of	O
a	O
GAL4	O
promoter	O
.	O
	O
BIO399	O
inhibited	O
the	O
luciferase	O
activity	O
when	O
added	O
to	O
the	O
cells	O
expressing	O
the	O
RORγ	O
-	O
GAL4	O
fusion	O
with	O
an	O
in	O
vivo	O
IC50	O
of	O
42	O
.	O
5nM	O
while	O
showing	O
>	O
235	O
and	O
28	O
fold	O
selectivity	O
over	O
cells	O
expressing	O
GAL4	O
fused	O
to	O
the	O
LBD	O
of	O
ROR	O
α	O
or	O
β	O
,	O
respectively	O
(	O
Table	O
1	O
).	O
	O
The	O
LBS	O
of	O
RORs	O
share	O
a	O
high	O
degree	O
of	O
similarity	O
.	O
	O
This	O
selectivity	O
profile	O
for	O
BIO399	O
is	O
attributed	O
to	O
the	O
shorter	O
leucine	O
side	O
chain	O
in	O
RORα	O
and	O
β	O
which	O
would	O
not	O
reach	O
the	O
phenylalanine	O
on	O
the	O
AF2	O
helix	O
further	O
underscoring	O
the	O
role	O
of	O
Met358	O
as	O
a	O
trigger	O
for	O
RORγ	O
specific	O
inverse	O
agonism	O
(	O
Fig	O
.	O
7a	O
).	O
	O
We	O
hypothesize	O
that	O
the	O
two	O
phenylalanine	O
residues	O
in	O
the	O
LBS	O
of	O
RORα	O
occlude	O
the	O
dihydrobenzoxazepinone	O
ring	O
system	O
of	O
BIO399	O
from	O
binding	O
it	O
and	O
responsible	O
for	O
the	O
increase	O
in	O
selectivity	O
for	O
RORα	O
over	O
β	O
.	O
	O
We	O
have	O
identified	O
a	O
novel	O
series	O
of	O
synthetic	O
benzoxazinone	O
ligands	O
which	O
modulate	O
the	O
transcriptional	O
activity	O
of	O
RORγ	O
in	O
a	O
FRET	O
based	O
assay	O
.	O
	O
Using	O
partial	O
proteolysis	O
we	O
show	O
a	O
conformational	O
change	O
which	O
destabilizes	O
the	O
AF2	O
helix	O
of	O
RORγ	O
when	O
the	O
inverse	O
agonist	O
BIO399	O
binds	O
.	O
	O
The	O
two	O
RORγ	O
co	O
-	O
crystal	O
structures	O
reported	O
here	O
show	O
how	O
a	O
small	O
change	O
to	O
the	O
core	O
ring	O
system	O
can	O
modulate	O
the	O
mode	O
of	O
action	O
from	O
agonist	O
(	O
BIO592	O
)	O
to	O
inverse	O
agonism	O
(	O
BIO399	O
).	O
	O
Finally	O
,	O
we	O
are	O
reporting	O
a	O
newly	O
identified	O
trigger	O
for	O
achieving	O
RORγ	O
specific	O
inverse	O
agonism	O
in	O
an	O
in	O
vivo	O
setting	O
through	O
Met358	O
which	O
perturbs	O
the	O
agonist	O
conformation	O
of	O
the	O
AF2	O
helix	O
and	O
prevents	O
coactivator	O
protein	O
binding	O
.	O
	O
Bacterial	O
Microcompartments	O
(	O
BMCs	O
)	O
are	O
proteinaceous	O
organelles	O
that	O
encapsulate	O
critical	O
segments	O
of	O
autotrophic	O
and	O
heterotrophic	O
metabolic	O
pathways	O
;	O
they	O
are	O
functionally	O
diverse	O
and	O
are	O
found	O
across	O
23	O
different	O
phyla	O
.	O
	O
The	O
core	O
enzyme	O
phosphotransacylase	O
(	O
PTAC	O
)	O
recycles	O
Coenzyme	O
A	O
and	O
generates	O
an	O
acyl	O
phosphate	O
that	O
can	O
serve	O
as	O
an	O
energy	O
source	O
.	O
	O
The	O
PTAC	O
predominantly	O
associated	O
with	O
metabolosomes	O
(	O
PduL	O
)	O
has	O
no	O
sequence	O
homology	O
to	O
the	O
PTAC	O
ubiquitous	O
among	O
fermentative	O
bacteria	O
(	O
Pta	O
).	O
	O
Here	O
,	O
we	O
report	O
two	O
high	O
-	O
resolution	O
PduL	O
crystal	O
structures	O
with	O
bound	O
substrates	O
.	O
	O
The	O
PduL	O
fold	O
is	O
unrelated	O
to	O
that	O
of	O
Pta	O
;	O
it	O
contains	O
a	O
dimetal	O
active	O
site	O
involved	O
in	O
a	O
catalytic	O
mechanism	O
distinct	O
from	O
that	O
of	O
the	O
housekeeping	O
PTAC	O
.	O
	O
The	O
PduL	O
structure	O
,	O
in	O
the	O
context	O
of	O
the	O
catalytic	O
core	O
,	O
completes	O
our	O
understanding	O
of	O
the	O
structural	O
basis	O
of	O
cofactor	O
recycling	O
in	O
the	O
metabolosome	O
lumen	O
.	O
	O
This	O
study	O
describes	O
the	O
structure	O
of	O
a	O
novel	O
phosphotransacylase	O
enzyme	O
that	O
facilitates	O
the	O
recycling	O
of	O
the	O
essential	O
cofactor	O
acetyl	O
-	O
CoA	O
within	O
a	O
bacterial	O
organelle	O
and	O
discusses	O
the	O
properties	O
of	O
the	O
enzyme	O
'	O
s	O
active	O
site	O
and	O
how	O
it	O
is	O
packaged	O
into	O
the	O
organelle	O
.	O
	O
The	O
phosphotransacylase	O
(	O
Pta	O
)	O
enzyme	O
catalyzes	O
the	O
conversion	O
between	O
acyl	O
-	O
CoA	O
and	O
acyl	O
-	O
phosphate	O
.	O
	O
This	O
reaction	O
directly	O
links	O
an	O
acyl	O
-	O
CoA	O
with	O
ATP	O
generation	O
via	O
substrate	O
-	O
level	O
phosphorylation	O
,	O
producing	O
short	O
-	O
chain	O
fatty	O
acids	O
(	O
e	O
.	O
g	O
.,	O
acetate	O
),	O
and	O
also	O
provides	O
a	O
path	O
for	O
short	O
-	O
chain	O
fatty	O
acids	O
to	O
enter	O
central	O
metabolism	O
.	O
	O
Due	O
to	O
this	O
key	O
function	O
,	O
Pta	O
is	O
conserved	O
across	O
the	O
bacterial	O
kingdom	O
.	O
	O
Recently	O
,	O
a	O
new	O
type	O
of	O
phosphotransacylase	O
was	O
described	O
that	O
shares	O
no	O
evolutionary	O
relation	O
to	O
Pta	O
.	O
	O
Not	O
only	O
does	O
PduL	O
facilitate	O
substrate	O
level	O
phosphorylation	O
,	O
but	O
it	O
also	O
is	O
critical	O
for	O
cofactor	O
recycling	O
within	O
,	O
and	O
product	O
efflux	O
from	O
,	O
the	O
organelle	O
.	O
	O
We	O
solved	O
the	O
structure	O
of	O
this	O
convergent	O
phosphotransacylase	O
and	O
show	O
that	O
it	O
is	O
completely	O
structurally	O
different	O
from	O
Pta	O
,	O
including	O
its	O
active	O
site	O
architecture	O
.	O
	O
Bacterial	O
Microcompartments	O
(	O
BMCs	O
)	O
are	O
organelles	O
that	O
encapsulate	O
enzymes	O
for	O
sequential	O
biochemical	O
reactions	O
within	O
a	O
protein	O
shell	O
.	O
	O
The	O
shell	O
is	O
typically	O
composed	O
of	O
three	O
types	O
of	O
protein	O
subunits	O
,	O
which	O
form	O
either	O
hexagonal	O
(	O
BMC	O
-	O
H	O
and	O
BMC	O
-	O
T	O
)	O
or	O
pentagonal	O
(	O
BMC	O
-	O
P	O
)	O
tiles	O
that	O
assemble	O
into	O
a	O
polyhedral	O
shell	O
.	O
	O
The	O
vitamin	O
B12	O
-	O
dependent	O
propanediol	O
-	O
utilizing	O
(	O
PDU	O
)	O
BMC	O
was	O
one	O
of	O
the	O
first	O
functionally	O
characterized	O
catabolic	O
BMCs	O
;	O
subsequently	O
,	O
other	O
types	O
have	O
been	O
implicated	O
in	O
the	O
degradation	O
of	O
ethanolamine	O
,	O
choline	O
,	O
fucose	O
,	O
rhamnose	O
,	O
and	O
ethanol	O
,	O
all	O
of	O
which	O
produce	O
different	O
aldehyde	O
intermediates	O
(	O
Table	O
1	O
).	O
	O
More	O
recently	O
,	O
bioinformatic	O
studies	O
have	O
demonstrated	O
the	O
widespread	O
distribution	O
of	O
BMCs	O
among	O
diverse	O
bacterial	O
phyla	O
and	O
grouped	O
them	O
into	O
23	O
different	O
functional	O
types	O
.	O
	O
The	O
reactions	O
carried	O
out	O
in	O
the	O
majority	O
of	O
catabolic	O
BMCs	O
(	O
also	O
known	O
as	O
metabolosomes	O
)	O
fit	O
a	O
generalized	O
biochemical	O
paradigm	O
for	O
the	O
oxidation	O
of	O
aldehydes	O
(	O
Fig	O
1	O
).	O
	O
This	O
involves	O
a	O
BMC	O
-	O
encapsulated	O
signature	O
enzyme	O
that	O
generates	O
a	O
toxic	O
and	O
/	O
or	O
volatile	O
aldehyde	O
that	O
the	O
BMC	O
shell	O
sequesters	O
from	O
the	O
cytosol	O
.	O
	O
These	O
two	O
cofactors	O
are	O
relatively	O
large	O
,	O
and	O
their	O
diffusion	O
across	O
the	O
protein	O
shell	O
is	O
thought	O
to	O
be	O
restricted	O
,	O
necessitating	O
their	O
regeneration	O
within	O
the	O
BMC	O
lumen	O
.	O
	O
The	O
final	O
product	O
of	O
the	O
BMC	O
,	O
an	O
acyl	O
-	O
phosphate	O
,	O
can	O
then	O
be	O
used	O
to	O
generate	O
ATP	O
via	O
acyl	O
kinase	O
,	O
or	O
revert	O
back	O
to	O
acyl	O
-	O
CoA	O
by	O
Pta	O
for	O
biosynthesis	O
.	O
	O
Collectively	O
,	O
the	O
aldehyde	O
and	O
alcohol	O
dehydrogenases	O
,	O
as	O
well	O
as	O
the	O
PTAC	O
,	O
constitute	O
the	O
common	O
metabolosome	O
core	O
.	O
	O
General	O
biochemical	O
model	O
of	O
aldehyde	O
-	O
degrading	O
BMCs	O
(	O
metabolosomes	O
)	O
illustrating	O
the	O
common	O
metabolosome	O
core	O
enzymes	O
and	O
reactions	O
.	O
	O
Characterized	O
and	O
predicted	O
catabolic	O
BMC	O
(	O
metabolosome	O
)	O
types	O
that	O
represent	O
the	O
aldehyde	O
-	O
degrading	O
paradigm	O
(	O
for	O
definition	O
of	O
types	O
see	O
Kerfeld	O
and	O
Erbilgin	O
).	O
	O
Name	O
PTAC	O
Type	O
Sequestered	O
Aldehyde	O
PDU	O
*	O
PduL	O
propionaldehyde	O
EUT1	O
PTA_PTB	O
acetaldehyde	O
EUT2	O
PduL	O
acetaldehyde	O
ETU	O
None	O
acetaldehyde	O
GRM1	O
/	O
CUT	O
PduL	O
acetaldehyde	O
GRM2	O
PduL	O
acetaldehyde	O
GRM3	O
*,	O
4	O
PduL	O
propionaldehyde	O
GRM5	O
/	O
GRP	O
PduL	O
propionaldehyde	O
PVM	O
*	O
PduL	O
lactaldehyde	O
RMM1	O
,	O
2	O
None	O
unknown	O
SPU	O
PduL	O
unknown	O
	O
*	O
PduL	O
from	O
these	O
functional	O
types	O
of	O
metabolosomes	O
were	O
purified	O
in	O
this	O
study	O
.	O
	O
The	O
concerted	O
functioning	O
of	O
a	O
PTAC	O
and	O
an	O
acetate	O
kinase	O
(	O
Ack	O
)	O
is	O
crucial	O
for	O
ATP	O
generation	O
in	O
the	O
fermentation	O
of	O
pyruvate	O
to	O
acetate	O
(	O
see	O
Reactions	O
1	O
and	O
2	O
).	O
	O
Both	O
enzymes	O
are	O
,	O
however	O
,	O
not	O
restricted	O
to	O
fermentative	O
organisms	O
.	O
	O
Reaction	O
1	O
:	O
acetyl	O
-	O
S	O
-	O
CoA	O
+	O
Pi	O
←→	O
acetyl	O
phosphate	O
+	O
CoA	O
-	O
SH	O
(	O
PTAC	O
)	O
	O
Reaction	O
2	O
:	O
acetyl	O
phosphate	O
+	O
ADP	O
←→	O
acetate	O
+	O
ATP	O
(	O
Ack	O
)	O
	O
Pta	O
has	O
been	O
extensively	O
characterized	O
due	O
to	O
its	O
key	O
role	O
in	O
fermentation	O
.	O
	O
More	O
recently	O
,	O
a	O
second	O
type	O
of	O
PTAC	O
without	O
any	O
sequence	O
homology	O
to	O
Pta	O
was	O
identified	O
.	O
	O
This	O
protein	O
,	O
PduL	O
(	O
Pfam	O
domain	O
PF06130	O
),	O
was	O
shown	O
to	O
catalyze	O
the	O
conversion	O
of	O
propionyl	O
-	O
CoA	O
to	O
propionyl	O
-	O
phosphate	O
and	O
is	O
associated	O
with	O
a	O
BMC	O
involved	O
in	O
propanediol	O
utilization	O
,	O
the	O
PDU	O
BMC	O
.	O
	O
Both	O
pduL	O
and	O
pta	O
genes	O
can	O
be	O
found	O
in	O
genetic	O
loci	O
of	O
functionally	O
distinct	O
BMCs	O
,	O
although	O
the	O
PduL	O
type	O
is	O
much	O
more	O
prevalent	O
,	O
being	O
found	O
in	O
all	O
but	O
one	O
type	O
of	O
metabolosome	O
locus	O
:	O
EUT1	O
(	O
Table	O
1	O
).	O
	O
Furthermore	O
,	O
in	O
the	O
Integrated	O
Microbial	O
Genomes	O
Database	O
,	O
91	O
%	O
of	O
genomes	O
that	O
encode	O
PF06130	O
also	O
encode	O
genes	O
for	O
shell	O
proteins	O
.	O
	O
As	O
a	O
member	O
of	O
the	O
core	O
biochemical	O
machinery	O
of	O
functionally	O
diverse	O
aldehyde	O
-	O
oxidizing	O
metabolosomes	O
,	O
PduL	O
must	O
have	O
a	O
certain	O
level	O
of	O
substrate	O
plasticity	O
(	O
see	O
Table	O
1	O
)	O
that	O
is	O
not	O
required	O
of	O
Pta	O
,	O
which	O
has	O
generally	O
been	O
observed	O
to	O
prefer	O
acetyl	O
-	O
CoA	O
.	O
PduL	O
from	O
the	O
PDU	O
BMC	O
of	O
Salmonella	O
enterica	O
favors	O
propionyl	O
-	O
CoA	O
over	O
acetyl	O
-	O
CoA	O
,	O
and	O
it	O
is	O
likely	O
that	O
PduL	O
orthologs	O
in	O
functionally	O
diverse	O
BMCs	O
would	O
have	O
substrate	O
preferences	O
for	O
other	O
CoA	O
derivatives	O
.	O
	O
EPs	O
have	O
also	O
been	O
observed	O
to	O
cause	O
proteins	O
to	O
aggregate	O
,	O
and	O
this	O
has	O
recently	O
been	O
suggested	O
to	O
be	O
functionally	O
relevant	O
as	O
an	O
initial	O
step	O
in	O
metabolosome	O
assembly	O
,	O
in	O
which	O
a	O
multifunctional	O
protein	O
core	O
is	O
formed	O
,	O
around	O
which	O
the	O
shell	O
assembles	O
.	O
	O
Of	O
the	O
three	O
common	O
metabolosome	O
core	O
enzymes	O
,	O
crystal	O
structures	O
are	O
available	O
for	O
both	O
the	O
alcohol	O
and	O
aldehyde	O
dehydrogenases	O
.	O
	O
In	O
contrast	O
,	O
the	O
structure	O
of	O
PduL	O
,	O
the	O
PTAC	O
found	O
in	O
the	O
vast	O
majority	O
of	O
catabolic	O
BMCs	O
,	O
has	O
not	O
been	O
determined	O
.	O
	O
This	O
is	O
a	O
major	O
gap	O
in	O
our	O
understanding	O
of	O
metabolosome	O
-	O
encapsulated	O
biochemistry	O
and	O
cofactor	O
recycling	O
.	O
	O
Moreover	O
,	O
it	O
will	O
be	O
useful	O
for	O
guiding	O
efforts	O
to	O
engineer	O
novel	O
BMC	O
cores	O
for	O
biotechnological	O
applications	O
.	O
	O
No	O
available	O
protein	O
structures	O
contain	O
the	O
PF06130	O
domain	O
,	O
and	O
homology	O
searches	O
using	O
the	O
primary	O
structure	O
of	O
PduL	O
do	O
not	O
return	O
any	O
significant	O
results	O
that	O
would	O
allow	O
prediction	O
of	O
the	O
structure	O
.	O
	O
Moreover	O
,	O
the	O
evident	O
novelty	O
of	O
PduL	O
makes	O
its	O
structure	O
interesting	O
in	O
the	O
context	O
of	O
convergent	O
evolution	O
of	O
PTAC	O
function	O
;	O
to	O
-	O
date	O
,	O
only	O
the	O
Pta	O
active	O
site	O
and	O
catalytic	O
mechanism	O
is	O
known	O
.	O
	O
We	O
propose	O
a	O
catalytic	O
mechanism	O
analogous	O
but	O
yet	O
distinct	O
from	O
the	O
ubiquitous	O
Pta	O
enzyme	O
,	O
highlighting	O
the	O
functional	O
convergence	O
of	O
two	O
enzymes	O
with	O
completely	O
different	O
structures	O
and	O
metal	O
requirements	O
.	O
	O
We	O
also	O
investigate	O
the	O
quaternary	O
structures	O
of	O
three	O
different	O
PduL	O
homologs	O
and	O
situate	O
our	O
findings	O
in	O
the	O
context	O
of	O
organelle	O
biogenesis	O
in	O
functionally	O
diverse	O
BMCs	O
.	O
	O
We	O
cloned	O
,	O
expressed	O
,	O
and	O
purified	O
three	O
different	O
PduL	O
homologs	O
from	O
functionally	O
distinct	O
BMCs	O
(	O
Table	O
1	O
):	O
from	O
the	O
well	O
-	O
studied	O
pdu	O
locus	O
in	O
S	O
.	O
enterica	O
Typhimurium	O
LT2	O
(	O
sPduL	O
),	O
from	O
the	O
recently	O
characterized	O
pvm	O
locus	O
in	O
Planctomyces	O
limnophilus	O
(	O
pPduL	O
),	O
and	O
from	O
the	O
grm3	O
locus	O
in	O
Rhodopseudomonas	O
palustris	O
BisB18	O
(	O
rPduL	O
).	O
	O
While	O
purifying	O
full	O
-	O
length	O
sPduL	O
,	O
we	O
observed	O
a	O
tendency	O
to	O
aggregation	O
as	O
described	O
previously	O
,	O
with	O
a	O
large	O
fraction	O
of	O
the	O
expressed	O
protein	O
found	O
in	O
the	O
insoluble	O
fraction	O
in	O
a	O
white	O
,	O
cake	O
-	O
like	O
pellet	O
.	O
	O
Similar	O
differences	O
in	O
solubility	O
were	O
observed	O
for	O
pPduL	O
and	O
rPduL	O
when	O
comparing	O
EP	O
-	O
truncated	O
forms	O
to	O
the	O
full	O
-	O
length	O
protein	O
,	O
but	O
none	O
were	O
quite	O
as	O
dramatic	O
as	O
for	O
sPduL	O
.	O
We	O
confirmed	O
that	O
all	O
homologs	O
were	O
active	O
(	O
S1a	O
and	O
S1b	O
Fig	O
).	O
	O
Among	O
these	O
,	O
we	O
were	O
only	O
able	O
to	O
obtain	O
diffraction	O
-	O
quality	O
crystals	O
of	O
rPduL	O
after	O
removing	O
the	O
N	O
-	O
terminal	O
putative	O
EP	O
(	O
33	O
amino	O
acids	O
,	O
also	O
see	O
Fig	O
2a	O
)	O
(	O
rPduLΔEP	B-mutant
).	O
	O
Truncated	O
rPduLΔEP	B-mutant
had	O
comparable	O
enzymatic	O
activity	O
to	O
the	O
full	O
-	O
length	O
enzyme	O
(	O
S1a	O
Fig	O
).	O
	O
Structural	O
overview	O
of	O
R	O
.	O
palustris	O
PduL	O
from	O
the	O
grm3	O
locus	O
.	O
	O
(	O
a	O
)	O
Primary	O
and	O
secondary	O
structure	O
of	O
rPduL	O
(	O
tubes	O
represent	O
α	O
-	O
helices	O
,	O
arrows	O
β	O
-	O
sheets	O
and	O
dashed	O
line	O
residues	O
disordered	O
in	O
the	O
structure	O
.	O
	O
The	O
first	O
33	O
amino	O
acids	O
are	O
present	O
only	O
in	O
the	O
wildtype	O
construct	O
and	O
contains	O
the	O
predicted	O
EP	O
alpha	O
helix	O
,	O
α0	O
);	O
the	O
truncated	O
rPduLΔEP	B-mutant
that	O
was	O
crystallized	O
begins	O
with	O
M	O
-	O
G	O
-	O
V	O
.	O
Coloring	O
is	O
according	O
to	O
structural	O
domains	O
(	O
domain	O
1	O
D36	O
-	O
N46	O
/	O
Q155	O
-	O
C224	O
,	O
blue	O
;	O
loop	O
insertion	O
G61	O
-	O
E81	O
,	O
grey	O
;	O
domain	O
2	O
R47	O
-	O
F60	O
/	O
E82	O
-	O
A154	O
,	O
red	O
).	O
	O
Metal	O
coordination	O
residues	O
are	O
highlighted	O
in	O
light	O
blue	O
and	O
CoA	O
contacting	O
residues	O
in	O
magenta	O
,	O
residues	O
contacting	O
the	O
CoA	O
of	O
the	O
other	O
chain	O
are	O
also	O
outlined	O
.	O
	O
(	O
b	O
)	O
Cartoon	O
representation	O
of	O
the	O
structure	O
colored	O
by	O
domains	O
and	O
including	O
secondary	O
structure	O
numbering	O
.	O
	O
Coenzyme	O
A	O
is	O
shown	O
in	O
magenta	O
sticks	O
and	O
Zinc	O
(	O
grey	O
)	O
as	O
spheres	O
.	O
	O
We	O
collected	O
a	O
native	O
dataset	O
from	O
rPduLΔEP	B-mutant
crystals	O
diffracting	O
to	O
a	O
resolution	O
of	O
1	O
.	O
54	O
Å	O
(	O
Table	O
2	O
).	O
	O
Using	O
a	O
mercury	O
-	O
derivative	O
crystal	O
form	O
diffracting	O
to	O
1	O
.	O
99	O
Å	O
(	O
Table	O
2	O
),	O
we	O
obtained	O
high	O
quality	O
electron	O
density	O
for	O
model	O
building	O
and	O
used	O
the	O
initial	O
model	O
to	O
refine	O
against	O
the	O
native	O
data	O
to	O
Rwork	O
/	O
Rfree	O
values	O
of	O
18	O
.	O
9	O
/	O
22	O
.	O
1	O
%.	O
	O
There	O
are	O
two	O
PduL	O
molecules	O
in	O
the	O
asymmetric	O
unit	O
of	O
the	O
P212121	O
unit	O
cell	O
.	O
	O
We	O
were	O
able	O
to	O
fit	O
all	O
of	O
the	O
primary	O
structure	O
of	O
PduLΔEP	B-mutant
into	O
the	O
electron	O
density	O
with	O
the	O
exception	O
of	O
three	O
amino	O
acids	O
at	O
the	O
N	O
-	O
terminus	O
and	O
two	O
amino	O
acids	O
at	O
the	O
C	O
-	O
terminus	O
(	O
Fig	O
2a	O
);	O
the	O
model	O
is	O
of	O
excellent	O
quality	O
(	O
Table	O
2	O
).	O
	O
Structurally	O
,	O
PduL	O
consists	O
of	O
two	O
domains	O
(	O
Fig	O
2	O
,	O
blue	O
/	O
red	O
),	O
each	O
a	O
beta	O
-	O
barrel	O
that	O
is	O
capped	O
on	O
both	O
ends	O
by	O
short	O
α	O
-	O
helices	O
.	O
	O
β	O
-	O
Barrel	O
2	O
consists	O
mainly	O
of	O
the	O
central	O
segment	O
of	O
primary	O
structure	O
(	O
β2	O
,	O
β5	O
–	O
β9	O
;	O
residues	O
47	O
–	O
60	O
and	O
82	O
–	O
154	O
)	O
(	O
Fig	O
2	O
,	O
red	O
),	O
but	O
is	O
interrupted	O
by	O
a	O
short	O
two	O
-	O
strand	O
beta	O
sheet	O
(	O
β3	O
-	O
β4	O
,	O
residues	O
61	O
–	O
81	O
).	O
	O
This	O
β	O
-	O
sheet	O
is	O
involved	O
in	O
contacts	O
between	O
the	O
two	O
domains	O
and	O
forms	O
a	O
lid	O
over	O
the	O
active	O
site	O
.	O
	O
Residues	O
in	O
this	O
region	O
(	O
Gln42	O
,	O
Pro43	O
,	O
Gly44	O
),	O
covering	O
the	O
active	O
site	O
,	O
are	O
strongly	O
conserved	O
(	O
Fig	O
3	O
).	O
	O
This	O
structural	O
arrangement	O
is	O
completely	O
different	O
from	O
the	O
functionally	O
related	O
Pta	O
,	O
which	O
is	O
composed	O
of	O
two	O
domains	O
,	O
each	O
consisting	O
of	O
a	O
central	O
flat	O
beta	O
sheet	O
with	O
alpha	O
-	O
helices	O
on	O
the	O
top	O
and	O
bottom	O
.	O
	O
Residues	O
100	O
%	O
conserved	O
across	O
all	O
PduL	O
homologs	O
in	O
our	O
dataset	O
are	O
noted	O
with	O
an	O
asterisk	O
,	O
and	O
residues	O
conserved	O
in	O
over	O
90	O
%	O
of	O
sequences	O
are	O
noted	O
with	O
a	O
colon	O
.	O
	O
There	O
are	O
two	O
PduL	O
molecules	O
in	O
the	O
asymmetric	O
unit	O
forming	O
a	O
butterfly	O
-	O
shaped	O
dimer	O
(	O
Fig	O
4c	O
).	O
	O
Consistent	O
with	O
this	O
,	O
results	O
from	O
size	O
exclusion	O
chromatography	O
of	O
rPduLΔEP	B-mutant
suggest	O
that	O
it	O
is	O
a	O
dimer	O
in	O
solution	O
(	O
Fig	O
5e	O
).	O
	O
The	O
interface	O
between	O
the	O
two	O
chains	O
buries	O
882	O
Å2	O
per	O
monomer	O
and	O
is	O
mainly	O
formed	O
by	O
α	O
-	O
helices	O
2	O
and	O
4	O
and	O
parts	O
of	O
β	O
-	O
sheets	O
12	O
and	O
14	O
,	O
as	O
well	O
as	O
a	O
π	O
–	O
π	O
stacking	O
of	O
the	O
adenine	O
moiety	O
of	O
CoA	O
with	O
Phe116	O
of	O
the	O
adjacent	O
chain	O
(	O
Fig	O
4c	O
).	O
	O
The	O
folds	O
of	O
the	O
two	O
chains	O
in	O
the	O
asymmetric	O
unit	O
are	O
very	O
similar	O
,	O
superimposing	O
with	O
a	O
rmsd	O
of	O
0	O
.	O
16	O
Å	O
over	O
2	O
,	O
306	O
aligned	O
atom	O
pairs	O
.	O
	O
The	O
peripheral	O
helices	O
and	O
the	O
short	O
antiparallel	O
β3	O
–	O
4	O
sheet	O
mediate	O
most	O
of	O
the	O
crystal	O
contacts	O
.	O
	O
Details	O
of	O
active	O
site	O
,	O
dimeric	O
assembly	O
,	O
and	O
sequence	O
conservation	O
of	O
PduL	O
.	O
	O
(	O
a	O
,	O
b	O
)	O
Proposed	O
active	O
site	O
of	O
PduL	O
with	O
relevant	O
residues	O
shown	O
as	O
sticks	O
in	O
atom	O
coloring	O
(	O
nitrogen	O
blue	O
,	O
oxygen	O
red	O
,	O
sulfur	O
yellow	O
),	O
zinc	O
as	O
grey	O
colored	O
spheres	O
and	O
coordinating	O
ordered	O
water	O
molecules	O
in	O
red	O
.	O
	O
(	O
c	O
)	O
View	O
of	O
the	O
dimer	O
in	O
the	O
asymmetric	O
unit	O
from	O
the	O
side	O
,	O
domains	O
1	O
and	O
2	O
colored	O
as	O
in	O
Fig	O
2	O
and	O
the	O
two	O
chains	O
differentiated	O
by	O
blue	O
/	O
red	O
versus	O
slate	O
/	O
firebrick	O
.	O
	O
(	O
d	O
)	O
Surface	O
representation	O
of	O
the	O
structure	O
with	O
indicated	O
conservation	O
(	O
red	O
:	O
high	O
,	O
white	O
:	O
intermediate	O
,	O
yellow	O
:	O
low	O
).	O
	O
All	O
chromatograms	O
are	O
cropped	O
to	O
show	O
only	O
the	O
linear	O
range	O
of	O
separation	O
based	O
on	O
standard	O
runs	O
,	O
shown	O
in	O
black	O
squares	O
with	O
a	O
dashed	O
linear	O
trend	O
line	O
.	O
	O
Active	O
Site	O
Properties	O
	O
CoA	O
and	O
the	O
metal	O
ions	O
bind	O
between	O
the	O
two	O
domains	O
,	O
presumably	O
in	O
the	O
active	O
site	O
(	O
Figs	O
2b	O
and	O
4a	O
).	O
	O
To	O
identify	O
the	O
bound	O
metals	O
,	O
we	O
performed	O
an	O
X	O
-	O
ray	O
fluorescence	O
scan	O
on	O
the	O
crystals	O
at	O
various	O
wavelengths	O
(	O
corresponding	O
to	O
the	O
K	O
-	O
edges	O
of	O
Mn	O
,	O
Fe	O
,	O
Co	O
,	O
Ni	O
,	O
Cu	O
,	O
and	O
Zn	O
).	O
	O
There	O
was	O
a	O
large	O
signal	O
at	O
the	O
zinc	O
edge	O
,	O
and	O
we	O
tested	O
for	O
the	O
presence	O
of	O
zinc	O
by	O
collecting	O
full	O
data	O
sets	O
before	O
and	O
after	O
the	O
Zn	O
K	O
-	O
edge	O
(	O
1	O
.	O
2861	O
and	O
1	O
.	O
2822	O
Å	O
,	O
respectively	O
).	O
	O
The	O
large	O
differences	O
between	O
the	O
anomalous	O
signals	O
confirm	O
the	O
presence	O
of	O
zinc	O
at	O
both	O
metal	O
sites	O
(	O
S3	O
Fig	O
).	O
	O
The	O
first	O
zinc	O
ion	O
(	O
Zn1	O
)	O
is	O
in	O
a	O
tetrahedral	O
coordination	O
state	O
with	O
His48	O
,	O
His50	O
,	O
Glu109	O
,	O
and	O
the	O
CoA	O
sulfur	O
(	O
Fig	O
4a	O
).	O
	O
The	O
nitrogen	O
atom	O
coordinating	O
the	O
zinc	O
is	O
the	O
Nε	O
in	O
each	O
histidine	O
residue	O
,	O
as	O
is	O
typical	O
for	O
this	O
interaction	O
.	O
	O
The	O
phosphate	O
-	O
bound	O
structure	O
aligns	O
well	O
with	O
the	O
CoA	O
-	O
bound	O
structure	O
(	O
0	O
.	O
43	O
Å	O
rmsd	O
over	O
2	O
,	O
361	O
atoms	O
for	O
the	O
monomer	O
,	O
0	O
.	O
83	O
Å	O
over	O
5	O
,	O
259	O
aligned	O
atoms	O
for	O
the	O
dimer	O
).	O
	O
Conserved	O
Arg103	O
seems	O
to	O
be	O
involved	O
in	O
maintaining	O
the	O
phosphate	O
in	O
that	O
position	O
.	O
	O
An	O
additional	O
phosphate	O
molecule	O
is	O
bound	O
at	O
a	O
crystal	O
contact	O
interface	O
,	O
perhaps	O
accounting	O
for	O
the	O
14	O
Å	O
shorter	O
c	O
-	O
axis	O
in	O
the	O
phosphate	O
-	O
bound	O
crystal	O
form	O
(	O
Table	O
2	O
).	O
	O
Interestingly	O
,	O
some	O
of	O
the	O
residues	O
important	O
for	O
dimerization	O
of	O
rPduL	O
,	O
particularly	O
Phe116	O
,	O
are	O
poorly	O
conserved	O
across	O
PduL	O
homologs	O
associated	O
with	O
functionally	O
diverse	O
BMCs	O
(	O
Figs	O
4c	O
and	O
3	O
),	O
suggesting	O
that	O
they	O
may	O
have	O
alternative	O
oligomeric	O
states	O
.	O
	O
We	O
tested	O
this	O
hypothesis	O
by	O
performing	O
size	O
exclusion	O
chromatography	O
on	O
both	O
full	O
-	O
length	O
and	O
truncated	O
variants	O
(	O
lacking	O
the	O
EP	O
,	O
ΔEP	B-mutant
)	O
of	O
sPduL	O
,	O
rPduL	O
,	O
and	O
pPduL	O
.	O
These	O
three	O
homologs	O
are	O
found	O
in	O
functionally	O
distinct	O
BMCs	O
(	O
Table	O
1	O
).	O
	O
It	O
has	O
been	O
proposed	O
that	O
the	O
catabolic	O
BMCs	O
may	O
assemble	O
in	O
a	O
core	O
-	O
first	O
manner	O
,	O
with	O
the	O
luminal	O
enzymes	O
(	O
signature	O
enzyme	O
,	O
aldehyde	O
,	O
and	O
alcohol	O
dehydrogenases	O
and	O
the	O
BMC	O
PTAC	O
)	O
forming	O
an	O
initial	O
bolus	O
,	O
or	O
prometabolosome	O
,	O
around	O
which	O
a	O
shell	O
assembles	O
.	O
	O
We	O
found	O
that	O
not	O
only	O
did	O
the	O
different	O
orthologs	O
appear	O
to	O
assemble	O
into	O
different	O
oligomeric	O
states	O
,	O
but	O
that	O
quaternary	O
structure	O
was	O
dependent	O
on	O
whether	O
or	O
not	O
the	O
EP	O
was	O
present	O
.	O
	O
Full	O
-	O
length	O
sPduL	O
was	O
unstable	O
in	O
solution	O
—	O
precipitating	O
over	O
time	O
—	O
and	O
eluted	O
throughout	O
the	O
entire	O
volume	O
of	O
a	O
size	O
exclusion	O
column	O
,	O
indicating	O
it	O
was	O
nonspecifically	O
aggregating	O
.	O
	O
However	O
,	O
when	O
the	O
putative	O
EP	O
(	O
residues	O
1	O
–	O
27	O
)	O
was	O
removed	O
(	O
sPduL	B-mutant
ΔEP	I-mutant
),	O
the	O
truncated	O
protein	O
was	O
stable	O
and	O
eluted	O
as	O
a	O
single	O
peak	O
(	O
Fig	O
5a	O
)	O
consistent	O
with	O
the	O
size	O
of	O
a	O
monomer	O
(	O
Fig	O
5d	O
,	O
blue	O
curve	O
).	O
	O
In	O
contrast	O
,	O
both	O
full	O
-	O
length	O
rPduL	O
and	O
pPduL	O
appeared	O
to	O
exist	O
in	O
two	O
distinct	O
oligomeric	O
states	O
(	O
Fig	O
5b	O
and	O
5c	O
respectively	O
,	O
orange	O
curves	O
),	O
one	O
form	O
of	O
the	O
approximate	O
size	O
of	O
a	O
dimer	O
and	O
the	O
second	O
,	O
a	O
higher	O
molecular	O
weight	O
oligomer	O
(~	O
150	O
kDa	O
).	O
	O
Upon	O
deletion	O
of	O
the	O
putative	O
EP	O
(	O
residues	O
1	O
–	O
47	O
for	O
rPduL	O
,	O
and	O
1	O
–	O
20	O
for	O
pPduL	O
),	O
there	O
was	O
a	O
distinct	O
change	O
in	O
the	O
elution	O
profiles	O
(	O
Fig	O
5b	O
and	O
5c	O
respectively	O
,	O
blue	O
curves	O
).	O
	O
In	O
contrast	O
,	O
rPduLΔEP	B-mutant
eluted	O
as	O
one	O
smaller	O
oligomer	O
,	O
possibly	O
a	O
dimer	O
.	O
	O
We	O
also	O
analyzed	O
purified	O
rPduL	O
and	O
rPduLΔEP	B-mutant
by	O
size	O
exclusion	O
chromatography	O
coupled	O
with	O
multiangle	O
light	O
scattering	O
(	O
SEC	O
-	O
MALS	O
)	O
for	O
a	O
complementary	O
approach	O
to	O
assessing	O
oligomeric	O
state	O
.	O
	O
SEC	O
-	O
MALS	O
analysis	O
of	O
rPdulΔEP	B-mutant
is	O
consistent	O
with	O
a	O
dimer	O
(	O
as	O
observed	O
in	O
the	O
crystal	O
structure	O
)	O
with	O
a	O
weighted	O
average	O
(	O
Mw	O
)	O
and	O
number	O
average	O
(	O
Mn	O
)	O
of	O
the	O
molar	O
mass	O
of	O
58	O
.	O
4	O
kDa	O
+/−	O
11	O
.	O
2	O
%	O
and	O
58	O
.	O
8	O
kDa	O
+/−	O
10	O
.	O
9	O
%,	O
respectively	O
(	O
S4a	O
Fig	O
).	O
	O
This	O
corresponds	O
to	O
an	O
oligomeric	O
state	O
of	O
six	O
subunits	O
(	O
calculated	O
molecular	O
weight	O
of	O
144	O
kDa	O
).	O
	O
Collectively	O
,	O
these	O
data	O
strongly	O
suggest	O
that	O
the	O
N	O
-	O
terminal	O
EP	O
of	O
PduL	O
plays	O
a	O
role	O
in	O
defining	O
the	O
quaternary	O
structure	O
of	O
the	O
protein	O
.	O
	O
The	O
BMC	O
shell	O
not	O
only	O
sequesters	O
specific	O
enzymes	O
but	O
also	O
their	O
cofactors	O
,	O
thereby	O
establishing	O
a	O
private	O
cofactor	O
pool	O
dedicated	O
to	O
the	O
encapsulated	O
reactions	O
.	O
	O
In	O
catabolic	O
BMCs	O
,	O
CoA	O
and	O
NAD	O
+	O
must	O
be	O
continually	O
recycled	O
within	O
the	O
organelle	O
(	O
Fig	O
1	O
).	O
	O
Curiously	O
,	O
while	O
the	O
housekeeping	O
Pta	O
could	O
provide	O
this	O
function	O
,	O
and	O
indeed	O
does	O
so	O
in	O
the	O
case	O
of	O
one	O
type	O
of	O
ethanolamine	O
-	O
utilizing	O
(	O
EUT	O
)	O
BMC	O
,	O
the	O
evolutionarily	O
unrelated	O
PduL	O
fulfills	O
this	O
function	O
for	O
the	O
majority	O
of	O
metabolosomes	O
using	O
a	O
novel	O
structure	O
and	O
active	O
site	O
for	O
convergent	O
evolution	O
of	O
function	O
.	O
	O
The	O
Tertiary	O
Structure	O
of	O
PduL	O
Is	O
Formed	O
by	O
Discontinuous	O
Segments	O
of	O
Primary	O
Structure	O
	O
The	O
structure	O
of	O
PduL	O
consists	O
of	O
two	O
β	O
-	O
barrel	O
domains	O
capped	O
by	O
short	O
alpha	O
helical	O
segments	O
(	O
Fig	O
2b	O
).	O
	O
The	O
two	O
domains	O
are	O
structurally	O
very	O
similar	O
(	O
superimposing	O
with	O
a	O
rmsd	O
of	O
1	O
.	O
34	O
Å	O
(	O
over	O
123	O
out	O
of	O
320	O
/	O
348	O
aligned	O
backbone	O
atoms	O
,	O
S5a	O
Fig	O
).	O
	O
However	O
,	O
the	O
amino	O
acid	O
sequences	O
of	O
the	O
two	O
domains	O
are	O
only	O
16	O
%	O
identical	O
(	O
mainly	O
the	O
RHxH	O
motif	O
,	O
β2	O
and	O
β10	O
),	O
and	O
34	O
%	O
similar	O
.	O
	O
Our	O
structure	O
reveals	O
that	O
the	O
two	O
assigned	O
PF06130	O
domains	O
(	O
Fig	O
3	O
)	O
do	O
not	O
form	O
structurally	O
discrete	O
units	O
;	O
this	O
reduces	O
the	O
apparent	O
sequence	O
conservation	O
at	O
the	O
level	O
of	O
primary	O
structure	O
.	O
	O
One	O
strand	O
of	O
the	O
domain	O
1	O
beta	O
barrel	O
(	O
shown	O
in	O
blue	O
in	O
Fig	O
2	O
)	O
is	O
contributed	O
by	O
the	O
N	O
-	O
terminus	O
,	O
while	O
the	O
rest	O
of	O
the	O
domain	O
is	O
formed	O
by	O
the	O
residues	O
from	O
the	O
C	O
-	O
terminal	O
half	O
of	O
the	O
protein	O
.	O
	O
When	O
aligned	O
by	O
structure	O
,	O
the	O
β1	O
strand	O
of	O
the	O
first	O
domain	O
(	O
Fig	O
2a	O
and	O
2b	O
,	O
blue	O
)	O
corresponds	O
to	O
the	O
final	O
strand	O
of	O
the	O
second	O
domain	O
(	O
β9	O
),	O
effectively	O
making	O
the	O
domains	O
continuous	O
if	O
the	O
first	O
strand	O
was	O
transplanted	O
to	O
the	O
C	O
-	O
terminus	O
.	O
	O
The	O
closest	O
structural	O
homolog	O
of	O
the	O
PduL	O
barrel	O
domain	O
is	O
a	O
subdomain	O
of	O
a	O
multienzyme	O
complex	O
,	O
the	O
alpha	O
subunit	O
of	O
ethylbenzene	O
dehydrogenase	O
(	O
S5b	O
Fig	O
,	O
rmsd	O
of	O
2	O
.	O
26	O
Å	O
over	O
226	O
aligned	O
atoms	O
consisting	O
of	O
one	O
beta	O
barrel	O
and	O
one	O
capping	O
helix	O
).	O
	O
In	O
contrast	O
to	O
PduL	O
,	O
there	O
is	O
only	O
one	O
barrel	O
present	O
in	O
ethylbenzene	O
dehydrogenase	O
,	O
and	O
there	O
is	O
no	O
comparable	O
active	O
site	O
arrangement	O
.	O
	O
The	O
PduL	O
signature	O
primary	O
structure	O
,	O
two	O
PF06130	O
domains	O
,	O
occurs	O
in	O
some	O
multidomain	O
proteins	O
,	O
most	O
of	O
them	O
annotated	O
as	O
Acks	O
,	O
suggesting	O
that	O
PduL	O
may	O
also	O
replace	O
Pta	O
in	O
variants	O
of	O
the	O
phosphotransacetylase	O
-	O
Ack	O
pathway	O
.	O
	O
These	O
PduL	O
homologs	O
lack	O
EPs	O
,	O
and	O
their	O
fusion	O
to	O
Ack	O
may	O
have	O
evolved	O
as	O
a	O
way	O
to	O
facilitate	O
substrate	O
channeling	O
between	O
the	O
two	O
enzymes	O
.	O
	O
For	O
BMC	O
-	O
encapsulated	O
proteins	O
to	O
properly	O
function	O
together	O
,	O
they	O
must	O
be	O
targeted	O
to	O
the	O
lumen	O
and	O
assemble	O
into	O
an	O
organization	O
that	O
facilitates	O
substrate	O
/	O
product	O
channeling	O
among	O
the	O
different	O
catalytic	O
sites	O
of	O
the	O
signature	O
and	O
core	O
enzymes	O
.	O
	O
The	O
N	O
-	O
terminal	O
extension	O
on	O
PduL	O
homologs	O
may	O
serve	O
both	O
of	O
these	O
functions	O
.	O
	O
The	O
extension	O
shares	O
many	O
features	O
with	O
previously	O
characterized	O
EPs	O
:	O
it	O
is	O
present	O
only	O
in	O
homologs	O
associated	O
with	O
BMC	O
loci	O
,	O
and	O
it	O
is	O
predicted	O
to	O
form	O
an	O
amphipathic	O
α	O
-	O
helix	O
.	O
	O
Moreover	O
,	O
its	O
removal	O
affects	O
the	O
oligomeric	O
state	O
of	O
the	O
protein	O
.	O
	O
EP	O
-	O
mediated	O
oligomerization	O
has	O
been	O
observed	O
for	O
the	O
signature	O
and	O
core	O
BMC	O
enzymes	O
;	O
for	O
example	O
,	O
full	O
-	O
length	O
propanediol	O
dehydratase	O
and	O
ethanolamine	O
ammonia	O
-	O
lyase	O
(	O
signature	O
enzymes	O
for	O
PDU	O
and	O
EUT	O
BMCs	O
)	O
subunits	O
are	O
also	O
insoluble	O
,	O
but	O
become	O
soluble	O
upon	O
removal	O
of	O
the	O
predicted	O
EP	O
.	O
	O
sPduL	O
has	O
also	O
previously	O
been	O
reported	O
to	O
localize	O
to	O
inclusion	O
bodies	O
when	O
overexpressed	O
;	O
we	O
show	O
here	O
that	O
this	O
is	O
dependent	O
on	O
the	O
presence	O
of	O
the	O
EP	O
.	O
	O
This	O
propensity	O
of	O
the	O
EP	O
to	O
cause	O
proteins	O
to	O
form	O
complexes	O
(	O
Fig	O
5	O
)	O
might	O
not	O
be	O
a	O
coincidence	O
,	O
but	O
could	O
be	O
a	O
necessary	O
step	O
in	O
the	O
assembly	O
of	O
BMCs	O
.	O
	O
Structured	O
aggregation	O
of	O
the	O
core	O
enzymes	O
has	O
been	O
proposed	O
to	O
be	O
the	O
initial	O
step	O
in	O
metabolosome	O
assembly	O
and	O
is	O
known	O
to	O
be	O
the	O
first	O
step	O
of	O
β	O
-	O
carboxysome	O
biogenesis	O
,	O
where	O
the	O
core	O
enzyme	O
Ribulose	O
Bisphosphate	O
Carboxylase	O
/	O
Oxygenase	O
(	O
RuBisCO	O
)	O
is	O
aggregated	O
by	O
the	O
CcmM	O
protein	O
.	O
	O
Likewise	O
,	O
CsoS2	O
,	O
a	O
protein	O
in	O
the	O
α	O
-	O
carboxysome	O
core	O
,	O
also	O
aggregates	O
when	O
purified	O
and	O
is	O
proposed	O
to	O
facilitate	O
the	O
nucleation	O
and	O
encapsulation	O
of	O
RuBisCO	O
molecules	O
in	O
the	O
lumen	O
of	O
the	O
organelle	O
.	O
	O
This	O
role	O
for	O
EPs	O
in	O
BMC	O
assembly	O
is	O
in	O
addition	O
to	O
their	O
interaction	O
with	O
shell	O
proteins	O
.	O
	O
Our	O
PduL	O
crystals	O
contained	O
CoA	O
that	O
was	O
captured	O
from	O
the	O
Escherichia	O
coli	O
cytosol	O
,	O
indicating	O
that	O
the	O
“	O
ground	O
state	O
”	O
of	O
PduL	O
is	O
in	O
the	O
CoA	O
-	O
bound	O
form	O
;	O
this	O
could	O
provide	O
an	O
elegantly	O
simple	O
means	O
of	O
guaranteeing	O
a	O
1	O
:	O
1	O
ratio	O
of	O
CoA	O
:	O
PduL	O
within	O
the	O
metabolosome	O
lumen	O
.	O
	O
Active	O
Site	O
Identification	O
and	O
Structural	O
Insights	O
into	O
Catalysis	O
	O
The	O
active	O
site	O
of	O
PduL	O
is	O
formed	O
at	O
the	O
interface	O
of	O
the	O
two	O
structural	O
domains	O
(	O
Fig	O
2b	O
).	O
	O
As	O
expected	O
,	O
the	O
amino	O
acid	O
sequence	O
conservation	O
is	O
highest	O
in	O
the	O
region	O
around	O
the	O
proposed	O
active	O
site	O
(	O
Fig	O
4d	O
);	O
highly	O
conserved	O
residues	O
are	O
also	O
involved	O
in	O
CoA	O
binding	O
(	O
Figs	O
2a	O
and	O
3	O
,	O
residues	O
Ser45	O
,	O
Lys70	O
,	O
Arg97	O
,	O
Leu99	O
,	O
His204	O
,	O
Asn211	O
).	O
	O
All	O
of	O
the	O
metal	O
-	O
coordinating	O
residues	O
(	O
Fig	O
2a	O
)	O
are	O
absolutely	O
conserved	O
,	O
implicating	O
them	O
in	O
catalysis	O
or	O
the	O
correct	O
spatial	O
orientation	O
of	O
the	O
substrates	O
.	O
	O
Arg103	O
,	O
which	O
contacts	O
the	O
phosphate	O
(	O
Fig	O
4b	O
),	O
is	O
present	O
in	O
all	O
PduL	O
homologs	O
.	O
	O
The	O
close	O
resemblance	O
between	O
the	O
structures	O
binding	O
CoA	O
and	O
phosphate	O
likely	O
indicates	O
that	O
no	O
large	O
changes	O
in	O
protein	O
conformation	O
are	O
involved	O
in	O
catalysis	O
,	O
and	O
that	O
our	O
crystal	O
structures	O
are	O
representative	O
of	O
the	O
active	O
form	O
.	O
	O
There	O
is	O
a	O
pocket	O
nearby	O
the	O
active	O
site	O
between	O
the	O
well	O
-	O
conserved	O
residues	O
Ser45	O
and	O
Ala154	O
,	O
which	O
could	O
accommodate	O
the	O
propionyl	O
group	O
(	O
S6	O
Fig	O
).	O
	O
A	O
homology	O
model	O
of	O
sPduL	O
indicates	O
that	O
the	O
residues	O
making	O
up	O
this	O
pocket	O
and	O
the	O
surrounding	O
active	O
site	O
region	O
are	O
identical	O
to	O
that	O
of	O
rPduL	O
,	O
which	O
is	O
not	O
surprising	O
,	O
because	O
these	O
two	O
homologs	O
presumably	O
have	O
the	O
same	O
propionyl	O
-	O
CoA	O
substrate	O
.	O
	O
The	O
homology	O
model	O
of	O
pPduL	O
also	O
has	O
identical	O
residues	O
making	O
up	O
the	O
pocket	O
,	O
but	O
with	O
a	O
key	O
difference	O
in	O
the	O
vicinity	O
of	O
the	O
active	O
site	O
:	O
Gln77	O
of	O
rPduL	O
is	O
replaced	O
by	O
a	O
tyrosine	O
(	O
Tyr77	O
)	O
in	O
pPduL	O
.	O
The	O
physiological	O
substrate	O
of	O
pPduL	O
(	O
Table	O
1	O
)	O
is	O
thought	O
to	O
be	O
lactyl	O
-	O
CoA	O
,	O
which	O
contains	O
an	O
additional	O
hydroxyl	O
group	O
relative	O
to	O
propionyl	O
-	O
CoA	O
.	O
The	O
presence	O
of	O
an	O
aromatic	O
residue	O
at	O
this	O
position	O
may	O
underlie	O
the	O
substrate	O
preference	O
of	O
the	O
PduL	O
enzyme	O
from	O
the	O
pvm	O
locus	O
.	O
	O
The	O
catalytic	O
mechanism	O
of	O
Pta	O
involves	O
the	O
abstraction	O
of	O
a	O
thiol	O
hydrogen	O
by	O
an	O
aspartate	O
residue	O
,	O
resulting	O
in	O
the	O
nucleophilic	O
attack	O
of	O
thiolate	O
upon	O
the	O
carbonyl	O
carbon	O
of	O
acetyl	O
-	O
phosphate	O
,	O
oriented	O
by	O
an	O
arginine	O
and	O
stabilized	O
by	O
a	O
serine	O
—	O
there	O
are	O
no	O
metals	O
involved	O
.	O
	O
In	O
contrast	O
,	O
in	O
the	O
rPduL	O
structure	O
,	O
there	O
are	O
no	O
conserved	O
aspartate	O
residues	O
in	O
or	O
around	O
the	O
active	O
site	O
,	O
and	O
the	O
only	O
well	O
-	O
conserved	O
glutamate	O
residue	O
in	O
the	O
active	O
site	O
is	O
involved	O
in	O
coordinating	O
one	O
of	O
the	O
metal	O
ions	O
.	O
	O
These	O
observations	O
strongly	O
suggest	O
that	O
an	O
acidic	O
residue	O
is	O
not	O
directly	O
involved	O
in	O
catalysis	O
by	O
PduL	O
.	O
Instead	O
,	O
the	O
dimetal	O
active	O
site	O
of	O
PduL	O
may	O
create	O
a	O
nucleophile	O
from	O
one	O
of	O
the	O
hydroxyl	O
groups	O
on	O
free	O
phosphate	O
to	O
attack	O
the	O
carbonyl	O
carbon	O
of	O
the	O
thioester	O
bond	O
of	O
an	O
acyl	O
-	O
CoA	O
.	O
In	O
the	O
reverse	O
direction	O
,	O
the	O
metal	O
ion	O
(	O
s	O
)	O
could	O
stabilize	O
the	O
thiolate	O
anion	O
that	O
would	O
attack	O
the	O
carbonyl	O
carbon	O
of	O
an	O
acyl	O
-	O
phosphate	O
;	O
a	O
similar	O
mechanism	O
has	O
been	O
described	O
for	O
phosphatases	O
where	O
hydroxyl	O
groups	O
or	O
hydroxide	O
ions	O
can	O
act	O
as	O
a	O
base	O
when	O
coordinated	O
by	O
a	O
dimetal	O
active	O
site	O
.	O
	O
Our	O
structures	O
provide	O
the	O
foundation	O
for	O
studies	O
to	O
elucidate	O
the	O
details	O
of	O
the	O
catalytic	O
mechanism	O
of	O
PduL	O
.	O
Conserved	O
residues	O
in	O
the	O
active	O
site	O
that	O
may	O
contribute	O
to	O
substrate	O
binding	O
and	O
/	O
or	O
transition	O
state	O
stabilization	O
include	O
Ser127	O
,	O
Arg103	O
,	O
Arg194	O
,	O
Gln107	O
,	O
Gln74	O
,	O
and	O
Gln	O
/	O
Glu77	O
.	O
	O
In	O
the	O
phosphate	O
-	O
bound	O
crystal	O
structure	O
,	O
Ser127	O
and	O
Arg103	O
appear	O
to	O
position	O
the	O
phosphate	O
(	O
Fig	O
4b	O
).	O
	O
Alternatively	O
,	O
Arg103	O
might	O
act	O
as	O
a	O
base	O
to	O
render	O
the	O
phosphate	O
more	O
nucleophilic	O
.	O
	O
The	O
functional	O
groups	O
of	O
Gln74	O
,	O
Gln	O
/	O
Glu77	O
,	O
and	O
Arg194	O
are	O
directed	O
away	O
from	O
the	O
active	O
site	O
in	O
both	O
CoA	O
and	O
phosphate	O
-	O
bound	O
crystal	O
structures	O
and	O
do	O
not	O
appear	O
to	O
be	O
involved	O
in	O
hydrogen	O
bonding	O
with	O
these	O
substrates	O
,	O
although	O
they	O
could	O
be	O
important	O
for	O
positioning	O
an	O
acyl	O
-	O
phosphate	O
.	O
	O
This	O
hypothesis	O
is	O
strengthened	O
by	O
the	O
fact	O
that	O
the	O
CoA	O
-	O
bound	O
crystals	O
were	O
obtained	O
without	O
added	O
CoA	O
,	O
indicating	O
that	O
the	O
protein	O
bound	O
CoA	O
from	O
the	O
E	O
.	O
coli	O
expression	O
strain	O
and	O
retained	O
it	O
throughout	O
purification	O
and	O
crystallization	O
.	O
	O
Functional	O
,	O
but	O
Not	O
Structural	O
,	O
Convergence	O
of	O
PduL	O
and	O
Pta	O
	O
PduL	O
and	O
Pta	O
are	O
mechanistically	O
and	O
structurally	O
distinct	O
enzymes	O
that	O
catalyze	O
the	O
same	O
reaction	O
,	O
a	O
prime	O
example	O
of	O
evolutionary	O
convergence	O
upon	O
a	O
function	O
.	O
	O
There	O
are	O
several	O
examples	O
of	O
such	O
functional	O
convergence	O
of	O
enzymes	O
,	O
although	O
typically	O
the	O
enzymes	O
have	O
independently	O
evolved	O
similar	O
,	O
or	O
even	O
identical	O
active	O
sites	O
;	O
for	O
example	O
,	O
the	O
carbonic	O
anhydrase	O
family	O
.	O
	O
However	O
,	O
apparently	O
less	O
frequent	O
is	O
functional	O
convergence	O
that	O
is	O
supported	O
by	O
distinctly	O
different	O
active	O
sites	O
and	O
accordingly	O
catalytic	O
mechanism	O
,	O
as	O
revealed	O
by	O
comparison	O
of	O
the	O
structures	O
of	O
Pta	O
and	O
PduL	O
.	O
One	O
well	O
-	O
studied	O
example	O
of	O
this	O
is	O
the	O
β	O
-	O
lactamase	O
family	O
of	O
enzymes	O
,	O
in	O
which	O
the	O
active	O
site	O
of	O
Class	O
A	O
and	O
Class	O
C	O
enzymes	O
involve	O
serine	O
-	O
based	O
catalysis	O
,	O
but	O
Class	O
B	O
enzymes	O
are	O
metalloproteins	O
.	O
	O
This	O
is	O
not	O
surprising	O
,	O
as	O
β	O
-	O
lactamases	O
are	O
not	O
so	O
widespread	O
among	O
bacteria	O
and	O
therefore	O
would	O
be	O
expected	O
to	O
have	O
evolved	O
independently	O
several	O
times	O
as	O
a	O
defense	O
mechanism	O
against	O
β	O
-	O
lactam	O
antibiotics	O
.	O
	O
However	O
,	O
nearly	O
all	O
bacteria	O
encode	O
Pta	O
,	O
and	O
it	O
is	O
not	O
immediately	O
clear	O
why	O
the	O
Pta	O
/	O
PduL	O
functional	O
convergence	O
should	O
have	O
evolved	O
:	O
it	O
would	O
seem	O
to	O
be	O
evolutionarily	O
more	O
resourceful	O
for	O
the	O
Pta	O
-	O
encoding	O
gene	O
to	O
be	O
duplicated	O
and	O
repurposed	O
for	O
BMCs	O
,	O
as	O
is	O
apparently	O
the	O
case	O
in	O
one	O
type	O
of	O
BMC	O
—	O
EUT1	O
(	O
Table	O
1	O
).	O
	O
Further	O
biochemical	O
comparison	O
between	O
the	O
two	O
PTACs	O
will	O
likely	O
yield	O
exciting	O
results	O
that	O
could	O
answer	O
this	O
evolutionary	O
question	O
.	O
	O
BMCs	O
are	O
now	O
known	O
to	O
be	O
widespread	O
among	O
the	O
bacteria	O
and	O
are	O
involved	O
in	O
critical	O
segments	O
of	O
both	O
autotrophic	O
and	O
heterotrophic	O
biochemical	O
pathways	O
that	O
confer	O
to	O
the	O
host	O
organism	O
a	O
competitive	O
(	O
metabolic	O
)	O
advantage	O
in	O
select	O
niches	O
.	O
	O
As	O
one	O
of	O
the	O
three	O
common	O
metabolosome	O
core	O
enzymes	O
,	O
the	O
structure	O
of	O
PduL	O
provides	O
a	O
key	O
missing	O
piece	O
to	O
our	O
structural	O
picture	O
of	O
the	O
shared	O
core	O
biochemistry	O
(	O
Fig	O
1	O
)	O
of	O
functionally	O
diverse	O
catabolic	O
BMCs	O
.	O
	O
We	O
have	O
observed	O
the	O
oligomeric	O
state	O
differences	O
of	O
PduL	O
to	O
correlate	O
with	O
the	O
presence	O
of	O
an	O
EP	O
,	O
providing	O
new	O
insight	O
into	O
the	O
function	O
of	O
this	O
sequence	O
extension	O
in	O
BMC	O
assembly	O
.	O
	O
Moreover	O
,	O
our	O
results	O
suggest	O
a	O
means	O
for	O
Coenzyme	O
A	O
incorporation	O
during	O
metabolosome	O
biogenesis	O
.	O
	O
The	O
fact	O
that	O
PduL	O
is	O
confined	O
almost	O
exclusively	O
to	O
metabolosomes	O
can	O
be	O
used	O
to	O
develop	O
an	O
inhibitor	O
that	O
blocks	O
only	O
PduL	O
and	O
not	O
Pta	O
as	O
a	O
way	O
to	O
selectively	O
disrupt	O
BMC	O
-	O
based	O
metabolism	O
,	O
while	O
not	O
affecting	O
most	O
commensal	O
organisms	O
that	O
require	O
PTAC	O
activity	O
.	O
	O
Biochemistry	O
and	O
Crystal	O
Structure	O
of	O
Ectoine	O
Synthase	O
:	O
A	O
Metal	O
-	O
Containing	O
Member	O
of	O
the	O
Cupin	O
Superfamily	O
	O
Ectoine	O
is	O
a	O
compatible	O
solute	O
and	O
chemical	O
chaperone	O
widely	O
used	O
by	O
members	O
of	O
the	O
Bacteria	O
and	O
a	O
few	O
Archaea	O
to	O
fend	O
-	O
off	O
the	O
detrimental	O
effects	O
of	O
high	O
external	O
osmolarity	O
on	O
cellular	O
physiology	O
and	O
growth	O
.	O
	O
Ectoine	O
synthase	O
(	O
EctC	O
)	O
catalyzes	O
the	O
last	O
step	O
in	O
ectoine	O
production	O
and	O
mediates	O
the	O
ring	O
closure	O
of	O
the	O
substrate	O
N	O
-	O
gamma	O
-	O
acetyl	O
-	O
L	O
-	O
2	O
,	O
4	O
-	O
diaminobutyric	O
acid	O
through	O
a	O
water	O
elimination	O
reaction	O
.	O
	O
However	O
,	O
the	O
crystal	O
structure	O
of	O
ectoine	O
synthase	O
is	O
not	O
known	O
and	O
a	O
clear	O
understanding	O
of	O
how	O
its	O
fold	O
contributes	O
to	O
enzyme	O
activity	O
is	O
thus	O
lacking	O
.	O
	O
Using	O
the	O
ectoine	O
synthase	O
from	O
the	O
cold	O
-	O
adapted	O
marine	O
bacterium	O
Sphingopyxis	O
alaskensis	O
(	O
Sa	O
),	O
we	O
report	O
here	O
both	O
a	O
detailed	O
biochemical	O
characterization	O
of	O
the	O
EctC	O
enzyme	O
and	O
the	O
high	O
-	O
resolution	O
crystal	O
structure	O
of	O
its	O
apo	O
-	O
form	O
.	O
	O
Structural	O
analysis	O
classified	O
the	O
(	O
Sa	O
)	O
EctC	O
protein	O
as	O
a	O
member	O
of	O
the	O
cupin	O
superfamily	O
.	O
	O
The	O
interface	O
of	O
the	O
dimer	O
assembly	O
is	O
shaped	O
through	O
backbone	O
-	O
contacts	O
and	O
weak	O
hydrophobic	O
interactions	O
mediated	O
by	O
two	O
beta	O
-	O
sheets	O
within	O
each	O
monomer	O
.	O
	O
We	O
found	O
that	O
EctC	O
not	O
only	O
effectively	O
converts	O
its	O
natural	O
substrate	O
N	O
-	O
gamma	O
-	O
acetyl	O
-	O
L	O
-	O
2	O
,	O
4	O
-	O
diaminobutyric	O
acid	O
into	O
ectoine	O
through	O
a	O
cyclocondensation	O
reaction	O
,	O
but	O
that	O
it	O
can	O
also	O
use	O
the	O
isomer	O
N	O
-	O
alpha	O
-	O
acetyl	O
-	O
L	O
-	O
2	O
,	O
4	O
-	O
diaminobutyric	O
acid	O
as	O
its	O
substrate	O
,	O
albeit	O
with	O
substantially	O
reduced	O
catalytic	O
efficiency	O
.	O
	O
An	O
assessment	O
of	O
enzyme	O
activity	O
and	O
iron	O
content	O
of	O
these	O
mutants	O
give	O
important	O
clues	O
for	O
understanding	O
the	O
architecture	O
of	O
the	O
active	O
site	O
positioned	O
within	O
the	O
core	O
of	O
the	O
EctC	O
cupin	O
barrel	O
.	O
	O
Ectoine	O
[(	O
S	O
)-	O
2	O
-	O
methyl	O
-	O
1	O
,	O
4	O
,	O
5	O
,	O
6	O
-	O
tetrahydropyrimidine	O
-	O
4	O
-	O
carboxylic	O
acid	O
]	O
and	O
its	O
derivative	O
5	O
-	O
hydroxyectoine	O
[(	O
4S	O
,	O
5S	O
)-	O
5	O
-	O
hydroxy	O
-	O
2	O
-	O
methyl	O
-	O
1	O
,	O
4	O
,	O
5	O
,	O
6	O
-	O
tetrahydropyrimidine	O
-	O
4	O
-	O
carboxylic	O
acid	O
]	O
are	O
such	O
compatible	O
solutes	O
.	O
	O
Both	O
marine	O
and	O
terrestrial	O
microorganisms	O
produce	O
them	O
widely	O
in	O
response	O
to	O
osmotic	O
or	O
temperature	O
stress	O
.	O
	O
Synthesis	O
of	O
ectoine	O
occurs	O
from	O
the	O
intermediate	O
metabolite	O
L	O
-	O
aspartate	O
-	O
ß	O
-	O
semialdehyde	O
and	O
comprises	O
the	O
sequential	O
activities	O
of	O
three	O
enzymes	O
:	O
L	O
-	O
2	O
,	O
4	O
-	O
diaminobutyrate	O
transaminase	O
(	O
EctB	O
;	O
EC	O
2	O
.	O
6	O
.	O
1	O
.	O
76	O
),	O
2	O
,	O
4	O
-	O
diaminobutyrate	O
acetyltransferase	O
(	O
EctA	O
;	O
EC	O
2	O
.	O
3	O
.	O
1	O
.	O
178	O
),	O
and	O
ectoine	O
synthase	O
(	O
EctC	O
;	O
EC	O
4	O
.	O
2	O
.	O
1	O
.	O
108	O
)	O
(	O
Fig	O
1	O
).	O
	O
The	O
ectoine	O
derivative	O
5	O
-	O
hydroxyectoine	O
,	O
a	O
highly	O
effective	O
stress	O
protectant	O
in	O
its	O
own	O
right	O
,	O
is	O
synthesized	O
by	O
a	O
substantial	O
subgroup	O
of	O
the	O
ectoine	O
producers	O
.	O
	O
The	O
remarkable	O
function	O
preserving	O
effects	O
of	O
ectoines	O
for	O
macromolecules	O
and	O
cells	O
,	O
frequently	O
also	O
addressed	O
as	O
chemical	O
chaperones	O
,	O
led	O
to	O
a	O
substantial	O
interest	O
in	O
exploiting	O
these	O
compounds	O
for	O
biotechnological	O
purposes	O
and	O
medical	O
applications	O
.	O
	O
Biosynthetic	O
routes	O
for	O
ectoine	O
and	O
5	O
-	O
hydroxyectoine	O
.	O
	O
Here	O
we	O
focus	O
on	O
ectoine	O
synthase	O
(	O
EctC	O
),	O
the	O
key	O
enzyme	O
of	O
the	O
ectoine	O
biosynthetic	O
route	O
(	O
Fig	O
1	O
).	O
	O
Biochemical	O
characterizations	O
of	O
ectoine	O
synthases	O
from	O
the	O
extremophiles	O
Halomonas	O
elongata	O
,	O
Methylomicrobium	O
alcaliphilum	O
,	O
and	O
Acidiphilium	O
cryptum	O
,	O
and	O
from	O
the	O
nitrifying	O
archaeon	O
Nitrosopumilus	O
maritimus	O
have	O
been	O
carried	O
out	O
.	O
	O
Each	O
of	O
these	O
enzymes	O
catalyzes	O
as	O
their	O
main	O
activity	O
the	O
cyclization	O
of	O
N	O
-	O
γ	O
-	O
acetyl	O
-	O
L	O
-	O
2	O
,	O
4	O
-	O
diaminobutyric	O
acid	O
(	O
N	O
-	O
γ	O
-	O
ADABA	O
),	O
the	O
reaction	O
product	O
of	O
the	O
2	O
,	O
4	O
-	O
diaminobutyrate	O
acetyltransferase	O
(	O
EctA	O
),	O
to	O
ectoine	O
with	O
the	O
concomitant	O
release	O
of	O
a	O
water	O
molecule	O
(	O
Fig	O
1	O
).	O
	O
In	O
side	O
reactions	O
,	O
EctC	O
can	O
promote	O
the	O
formation	O
of	O
the	O
synthetic	O
compatible	O
solute	O
5	O
-	O
amino	O
-	O
3	O
,	O
4	O
-	O
dihydro	O
-	O
2H	O
-	O
pyrrole	O
-	O
2	O
-	O
carboxylate	O
(	O
ADPC	O
)	O
through	O
the	O
cyclic	O
condensation	O
of	O
two	O
glutamine	O
molecules	O
and	O
it	O
also	O
possesses	O
a	O
minor	O
hydrolytic	O
activity	O
for	O
ectoine	O
and	O
synthetic	O
ectoine	O
derivatives	O
with	O
either	O
reduced	O
or	O
expanded	O
ring	O
sizes	O
.	O
	O
Although	O
progress	O
has	O
been	O
made	O
with	O
respect	O
to	O
the	O
biochemical	O
characterization	O
of	O
ectoine	O
synthase	O
,	O
a	O
clear	O
understanding	O
of	O
how	O
its	O
structure	O
contributes	O
to	O
its	O
enzyme	O
activity	O
and	O
reaction	O
mechanism	O
is	O
still	O
lacking	O
.	O
With	O
this	O
in	O
mind	O
,	O
we	O
have	O
biochemically	O
characterized	O
the	O
ectoine	O
synthase	O
from	O
the	O
cold	O
-	O
adapted	O
marine	O
bacterium	O
Sphingopyxis	O
alaskensis	O
(	O
Sa	O
).	O
	O
We	O
demonstrate	O
here	O
for	O
the	O
first	O
time	O
that	O
the	O
ectoine	O
synthase	O
is	O
a	O
metal	O
-	O
dependent	O
enzyme	O
,	O
with	O
iron	O
as	O
the	O
most	O
likely	O
physiologically	O
relevant	O
co	O
-	O
factor	O
.	O
	O
Overproduction	O
,	O
purification	O
and	O
oligomeric	O
state	O
of	O
the	O
ectoine	O
synthase	O
in	O
solution	O
	O
We	O
focused	O
our	O
biochemical	O
and	O
structural	O
studies	O
on	O
the	O
ectoine	O
synthase	O
from	O
S	O
.	O
alaskensis	O
[(	O
Sa	O
)	O
EctC	O
],	O
a	O
cold	O
-	O
adapted	O
marine	O
ultra	O
-	O
microbacterium	O
,	O
from	O
which	O
we	O
recently	O
also	O
determined	O
the	O
crystal	O
structure	O
of	O
the	O
ectoine	O
hydroxylase	O
(	O
EctD	O
)	O
in	O
complex	O
with	O
either	O
its	O
substrate	O
or	O
its	O
reaction	O
product	O
.	O
	O
We	O
expressed	O
a	O
codon	O
-	O
optimized	O
version	O
of	O
the	O
S	O
.	O
alaskensis	O
ectC	O
gene	O
in	O
E	O
.	O
coli	O
to	O
produce	O
a	O
recombinant	O
protein	O
with	O
a	O
carboxy	O
-	O
terminally	O
attached	O
Strep	O
-	O
tag	O
II	O
affinity	O
peptide	O
to	O
allow	O
purification	O
of	O
the	O
(	O
Sa	O
)	O
EctC	O
-	O
Strep	O
-	O
Tag	O
-	O
II	O
protein	O
by	O
affinity	O
chromatography	O
.	O
	O
Conventional	O
size	O
-	O
exclusion	O
chromatography	O
(	O
SEC	O
)	O
has	O
already	O
shown	O
that	O
(	O
Sa	O
)	O
EctC	O
preparations	O
produced	O
in	O
this	O
fashion	O
are	O
homogeneous	O
and	O
that	O
the	O
protein	O
forms	O
dimers	O
in	O
solution	O
.	O
	O
High	O
performance	O
liquid	O
chromatography	O
coupled	O
with	O
multi	O
-	O
angle	O
light	O
-	O
scattering	O
detection	O
(	O
HPLC	O
-	O
MALS	O
)	O
experiments	O
carried	O
out	O
here	O
confirmed	O
that	O
the	O
purified	O
(	O
Sa	O
)	O
EctC	O
protein	O
was	O
mono	O
-	O
disperse	O
and	O
possessed	O
a	O
molecular	O
mass	O
of	O
33	O
.	O
0	O
±	O
2	O
.	O
3	O
kDa	O
(	O
S2b	O
Fig	O
).	O
	O
This	O
value	O
corresponds	O
very	O
well	O
with	O
the	O
theoretically	O
calculated	O
molecular	O
mass	O
of	O
an	O
(	O
Sa	O
)	O
EctC	O
dimer	O
(	O
molecular	O
mass	O
of	O
the	O
monomer	O
,	O
including	O
the	O
Strep	O
-	O
tag	O
II	O
affinity	O
peptide	O
:	O
16	O
.	O
3	O
kDa	O
).	O
	O
Such	O
a	O
quaternary	O
assembly	O
as	O
dimer	O
has	O
also	O
been	O
reported	O
for	O
the	O
EctC	O
proteins	O
from	O
H	O
.	O
elongata	O
and	O
N	O
.	O
maritimus	O
.	O
	O
The	O
EctA	O
-	O
produced	O
substrate	O
of	O
the	O
ectoine	O
synthase	O
,	O
N	O
-	O
γ	O
-	O
acetyl	O
-	O
L	O
-	O
2	O
,	O
4	O
-	O
diaminobutyric	O
acid	O
(	O
N	O
-	O
γ	O
-	O
ADABA	O
)	O
(	O
Fig	O
1	O
),	O
is	O
commercially	O
not	O
available	O
.	O
	O
We	O
used	O
alkaline	O
hydrolysis	O
of	O
ectoine	O
and	O
subsequent	O
chromatography	O
on	O
silica	O
gel	O
columns	O
to	O
obtain	O
N	O
-	O
γ	O
-	O
ADABA	O
in	O
chemically	O
highly	O
purified	O
form	O
(	O
S1a	O
Fig	O
).	O
	O
This	O
procedure	O
also	O
yielded	O
the	O
isomer	O
of	O
N	O
-	O
γ	O
-	O
ADABA	O
,	O
N	O
-	O
α	O
-	O
acetyl	O
-	O
L	O
-	O
2	O
,	O
4	O
-	O
diaminobutyric	O
acid	O
(	O
N	O
-	O
α	O
-	O
ADABA	O
)	O
(	O
S1b	O
Fig	O
).	O
	O
Using	O
N	O
-	O
γ	O
-	O
ADABA	O
as	O
the	O
substrate	O
,	O
we	O
initially	O
evaluated	O
a	O
set	O
of	O
biochemical	O
parameters	O
of	O
the	O
recombinant	O
(	O
Sa	O
)	O
EctC	O
protein	O
.	O
	O
S	O
.	O
alaskensis	O
,	O
from	O
which	O
the	O
studied	O
ectoine	O
synthase	O
was	O
originally	O
derived	O
,	O
is	O
a	O
microorganism	O
that	O
is	O
well	O
-	O
adapted	O
to	O
a	O
life	O
in	O
permanently	O
cold	O
ocean	O
waters	O
.	O
	O
Consistent	O
with	O
the	O
physicochemical	O
attributes	O
of	O
this	O
habitat	O
,	O
the	O
(	O
Sa	O
)	O
EctC	O
protein	O
was	O
already	O
enzymatically	O
active	O
at	O
5	O
°	O
C	O
,	O
had	O
a	O
temperature	O
optimum	O
of	O
15	O
°	O
C	O
and	O
was	O
able	O
to	O
function	O
over	O
a	O
broad	O
range	O
of	O
temperatures	O
(	O
S3a	O
Fig	O
).	O
	O
It	O
possessed	O
an	O
alkaline	O
pH	O
optimum	O
of	O
8	O
.	O
5	O
(	O
S3b	O
Fig	O
),	O
a	O
value	O
similar	O
to	O
the	O
ectoine	O
synthases	O
from	O
the	O
halo	O
-	O
tolerant	O
H	O
.	O
elongata	O
(	O
pH	O
optimum	O
of	O
8	O
.	O
5	O
to	O
9	O
.	O
0	O
),	O
the	O
alkaliphile	O
M	O
.	O
alcaliphilum	O
(	O
pH	O
optimum	O
of	O
9	O
.	O
0	O
),	O
and	O
the	O
acidophile	O
Acidiphilium	O
cryptum	O
(	O
pH	O
optimum	O
of	O
8	O
.	O
5	O
to	O
9	O
.	O
0	O
),	O
whereas	O
the	O
EctC	O
protein	O
from	O
N	O
.	O
maritimus	O
has	O
a	O
neutral	O
pH	O
optimum	O
(	O
pH	O
7	O
.	O
0	O
).	O
	O
The	O
salinity	O
of	O
the	O
assay	O
buffer	O
had	O
a	O
significant	O
influence	O
on	O
the	O
maximal	O
enzyme	O
activity	O
of	O
the	O
(	O
Sa	O
)	O
EctC	O
protein	O
.	O
	O
An	O
increase	O
in	O
either	O
the	O
NaCl	O
or	O
the	O
KCl	O
concentration	O
led	O
to	O
an	O
approximately	O
5	O
-	O
fold	O
enhancement	O
of	O
the	O
ectoine	O
synthase	O
activity	O
.	O
	O
The	O
maximum	O
enzyme	O
activity	O
of	O
(	O
Sa	O
)	O
EctC	O
occurred	O
around	O
250	O
mM	O
NaCl	O
or	O
KCl	O
,	O
respectively	O
.	O
	O
Considerations	O
based	O
on	O
bioinformatics	O
suggests	O
that	O
EctC	O
belongs	O
to	O
the	O
cupin	O
superfamily	O
.	O
	O
Most	O
of	O
these	O
proteins	O
contain	O
catalytically	O
important	O
transition	O
state	O
metals	O
such	O
as	O
iron	O
,	O
copper	O
,	O
zinc	O
,	O
manganese	O
,	O
cobalt	O
,	O
or	O
nickel	O
.	O
	O
Cupins	O
contain	O
two	O
conserved	O
motifs	O
:	O
G	O
(	O
X	O
)	O
5HXH	O
(	O
X	O
)	O
3	O
,	O
4E	O
(	O
X	O
)	O
6G	O
and	O
G	O
(	O
X	O
)	O
5PXG	O
(	O
X	O
)	O
2H	O
(	O
X	O
)	O
3N	O
(	O
the	O
letters	O
in	O
bold	O
represent	O
those	O
residues	O
that	O
often	O
coordinate	O
the	O
metal	O
).	O
	O
Inspection	O
of	O
a	O
previous	O
alignment	O
of	O
the	O
amino	O
acid	O
sequences	O
of	O
440	O
EctC	O
-	O
type	O
proteins	O
revealed	O
that	O
the	O
canonical	O
metal	O
-	O
binding	O
motif	O
(	O
s	O
)	O
of	O
cupin	O
-	O
type	O
proteins	O
is	O
not	O
conserved	O
among	O
members	O
of	O
the	O
extended	O
ectoine	O
synthase	O
protein	O
family	O
.	O
	O
An	O
abbreviated	O
alignment	O
of	O
the	O
amino	O
acid	O
sequence	O
of	O
EctC	O
-	O
type	O
proteins	O
is	O
shown	O
in	O
Fig	O
2	O
.	O
	O
Abbreviated	O
alignment	O
of	O
EctC	O
-	O
type	O
proteins	O
.	O
	O
Strictly	O
conserved	O
amino	O
acid	O
residues	O
are	O
shown	O
in	O
yellow	O
.	O
	O
Dots	O
shown	O
above	O
the	O
(	O
Sa	O
)	O
EctC	O
protein	O
sequence	O
indicate	O
residues	O
likely	O
to	O
be	O
involved	O
in	O
iron	O
-	O
binding	O
(	O
red	O
),	O
ligand	O
-	O
binding	O
(	O
green	O
)	O
and	O
stabilization	O
of	O
the	O
loop	O
-	O
architecture	O
(	O
blue	O
).	O
	O
The	O
conserved	O
residue	O
Tyr	O
-	O
52	O
with	O
so	O
-	O
far	O
undefined	O
functions	O
is	O
indicated	O
by	O
a	O
green	O
dot	O
circled	O
in	O
red	O
.	O
	O
Secondary	O
structural	O
elements	O
(	O
α	O
-	O
helices	O
and	O
β	O
-	O
sheets	O
)	O
found	O
in	O
the	O
(	O
Sa	O
)	O
EctC	O
crystal	O
structure	O
are	O
projected	O
onto	O
the	O
amino	O
acid	O
sequences	O
of	O
EctC	O
-	O
type	O
proteins	O
.	O
	O
For	O
this	O
analysis	O
we	O
used	O
recombinant	O
(	O
Sa	O
)	O
EctC	O
preparations	O
from	O
three	O
independent	O
protein	O
overproduction	O
and	O
purification	O
experiments	O
.	O
	O
The	O
ICP	O
-	O
MS	O
analyses	O
yielded	O
an	O
iron	O
content	O
of	O
0	O
.	O
66	O
±	O
0	O
.	O
06	O
mol	O
iron	O
per	O
mol	O
of	O
protein	O
and	O
the	O
used	O
(	O
Sa	O
)	O
EctC	O
protein	O
preparations	O
also	O
contained	O
a	O
minor	O
amount	O
of	O
zinc	O
(	O
0	O
.	O
08	O
mol	O
zinc	O
per	O
mol	O
of	O
protein	O
).	O
	O
All	O
other	O
assayed	O
metals	O
(	O
copper	O
and	O
nickel	O
)	O
were	O
only	O
present	O
in	O
trace	O
amounts	O
(	O
0	O
.	O
01	O
mol	O
metal	O
per	O
mol	O
of	O
protein	O
,	O
respectively	O
).	O
	O
The	O
presence	O
of	O
iron	O
in	O
these	O
(	O
Sa	O
)	O
EctC	O
protein	O
preparations	O
was	O
further	O
confirmed	O
by	O
a	O
colorimetric	O
method	O
that	O
is	O
based	O
on	O
an	O
iron	O
-	O
complexing	O
reagent	O
;	O
this	O
procedure	O
yielded	O
an	O
iron	O
-	O
content	O
of	O
0	O
.	O
84	O
±	O
0	O
.	O
05	O
mol	O
per	O
mol	O
of	O
(	O
Sa	O
)	O
EctC	O
protein	O
.	O
	O
Hence	O
,	O
both	O
ICP	O
-	O
MS	O
and	O
the	O
colorimetric	O
method	O
clearly	O
established	O
that	O
the	O
recombinantly	O
produced	O
ectoine	O
synthase	O
from	O
S	O
.	O
alaskensis	O
is	O
an	O
iron	O
-	O
containing	O
protein	O
.	O
	O
The	O
reason	O
for	O
this	O
difference	O
is	O
not	O
known	O
,	O
but	O
indicates	O
that	O
the	O
well	O
established	O
colorimetric	O
assay	O
probably	O
overestimates	O
the	O
iron	O
content	O
of	O
(	O
Sa	O
)	O
EctC	O
protein	O
preparations	O
to	O
a	O
certain	O
degree	O
.	O
	O
The	O
iron	O
detected	O
in	O
the	O
(	O
Sa	O
)	O
EctC	O
protein	O
preparations	O
could	O
serve	O
a	O
structural	O
role	O
,	O
or	O
most	O
likely	O
,	O
could	O
be	O
critical	O
for	O
enzyme	O
catalysis	O
as	O
is	O
the	O
case	O
for	O
many	O
members	O
of	O
the	O
cupin	O
superfamily	O
.	O
	O
The	O
addition	O
of	O
very	O
low	O
concentrations	O
of	O
EDTA	O
(	O
0	O
.	O
05	O
mM	O
)	O
to	O
the	O
EctC	O
enzyme	O
already	O
led	O
to	O
a	O
noticeable	O
inhibition	O
of	O
the	O
ectoine	O
synthase	O
activity	O
and	O
the	O
presence	O
of	O
1	O
mM	O
EDTA	O
completely	O
inhibited	O
the	O
enzyme	O
(	O
Fig	O
3a	O
).	O
	O
(	O
a	O
)	O
Impact	O
of	O
the	O
iron	O
-	O
chelator	O
EDTA	O
on	O
the	O
enzyme	O
activity	O
of	O
the	O
purified	O
(	O
Sa	O
)	O
EctC	O
protein	O
.	O
	O
Metal	O
depletion	O
and	O
reconstitution	O
experiments	O
with	O
(	O
b	O
)	O
stoichiometric	O
and	O
(	O
c	O
)	O
excess	O
amounts	O
of	O
metals	O
.	O
	O
The	O
(	O
Sa	O
)	O
EctC	O
protein	O
was	O
present	O
at	O
a	O
concentration	O
of	O
10	O
μM	O
.	O
The	O
level	O
of	O
enzyme	O
activity	O
given	O
in	O
(	O
b	O
)	O
is	O
benchmarked	O
relative	O
to	O
that	O
of	O
ectoine	O
synthase	O
enzyme	O
assays	O
in	O
which	O
1	O
mM	O
FeCl2	O
was	O
added	O
.	O
	O
The	O
addition	O
of	O
FeCl2	O
to	O
the	O
enzyme	O
assay	O
restored	O
enzyme	O
activity	O
to	O
about	O
38	O
%,	O
whereas	O
the	O
addition	O
of	O
ZnCl2	O
or	O
CoCl2	O
rescued	O
(	O
Sa	O
)	O
EctC	O
enzyme	O
activity	O
only	O
to	O
5	O
%	O
and	O
3	O
%,	O
respectively	O
.	O
	O
All	O
other	O
tested	O
metals	O
,	O
including	O
Fe3	O
+,	O
were	O
unable	O
to	O
restore	O
activity	O
(	O
Fig	O
3b	O
).	O
	O
When	O
the	O
concentration	O
of	O
the	O
various	O
metals	O
in	O
the	O
enzyme	O
assay	O
was	O
increased	O
100	O
-	O
fold	O
,	O
Fe2	O
+	O
exhibited	O
again	O
the	O
strongest	O
stimulating	O
effect	O
on	O
enzyme	O
activity	O
,	O
and	O
rescued	O
enzyme	O
activity	O
to	O
a	O
degree	O
similar	O
to	O
that	O
exhibited	O
by	O
(	O
Sa	O
)	O
EctC	O
protein	O
preparations	O
that	O
had	O
not	O
been	O
inactivated	O
through	O
EDTA	O
treatment	O
(	O
Fig	O
3c	O
).	O
	O
However	O
,	O
a	O
large	O
molar	O
excess	O
of	O
other	O
transition	O
-	O
state	O
metals	O
(	O
zinc	O
,	O
cobalt	O
,	O
nickel	O
,	O
copper	O
,	O
and	O
manganese	O
)	O
typically	O
found	O
in	O
members	O
of	O
the	O
cupin	O
superfamily	O
allowed	O
the	O
partial	O
rescue	O
of	O
ectoine	O
synthase	O
activity	O
as	O
well	O
(	O
Fig	O
3c	O
).	O
	O
This	O
is	O
in	O
line	O
with	O
literature	O
data	O
showing	O
that	O
cupin	O
-	O
type	O
enzymes	O
are	O
often	O
promiscuous	O
with	O
respect	O
to	O
the	O
use	O
of	O
the	O
catalytically	O
important	O
metal	O
.	O
	O
Kinetic	O
parameters	O
of	O
EctC	O
for	O
N	O
-	O
γ	O
-	O
ADABA	O
and	O
N	O
-	O
α	O
-	O
ADABA	O
	O
Based	O
on	O
the	O
data	O
presented	O
in	O
S3	O
Fig	O
,	O
we	O
formulated	O
an	O
optimized	O
activity	O
assay	O
for	O
the	O
ectoine	O
synthase	O
of	O
S	O
.	O
alaskensis	O
and	O
used	O
it	O
to	O
determined	O
the	O
kinetic	O
parameters	O
for	O
the	O
(	O
Sa	O
)	O
EctC	O
enzyme	O
for	O
both	O
its	O
natural	O
substrate	O
N	O
-	O
γ	O
-	O
ADABA	O
and	O
the	O
isomer	O
N	O
-	O
α	O
-	O
ADABA	O
.	O
	O
Given	O
the	O
chemical	O
relatedness	O
of	O
N	O
-	O
α	O
-	O
ADABA	O
to	O
the	O
natural	O
substrate	O
(	O
N	O
-	O
γ	O
-	O
ADABA	O
)	O
of	O
the	O
ectoine	O
synthase	O
(	O
S1a	O
and	O
S1b	O
Fig	O
),	O
we	O
wondered	O
whether	O
(	O
Sa	O
)	O
EctC	O
could	O
also	O
use	O
N	O
-	O
α	O
-	O
ADABA	O
to	O
produce	O
ectoine	O
.	O
	O
However	O
,	O
both	O
the	O
affinity	O
(	O
Km	O
)	O
of	O
the	O
(	O
Sa	O
)	O
EctC	O
protein	O
and	O
its	O
catalytic	O
efficiency	O
(	O
kcat	O
/	O
Km	O
)	O
were	O
strongly	O
reduced	O
in	O
comparison	O
with	O
N	O
-	O
γ	O
-	O
ADABA	O
.	O
	O
Both	O
N	O
-	O
γ	O
-	O
ADABA	O
and	O
N	O
-	O
α	O
-	O
ADABA	O
are	O
concomitantly	O
formed	O
during	O
the	O
enzymatic	O
hydrolysis	O
of	O
the	O
ectoine	O
ring	O
during	O
catabolism	O
.	O
	O
Our	O
finding	O
that	O
N	O
-	O
α	O
-	O
ADABA	O
is	O
a	O
substrate	O
for	O
ectoine	O
synthase	O
has	O
bearings	O
for	O
an	O
understanding	O
of	O
the	O
physiology	O
of	O
those	O
microorganisms	O
that	O
can	O
both	O
synthesize	O
and	O
catabolize	O
ectoine	O
.	O
	O
However	O
,	O
these	O
types	O
of	O
microorganisms	O
should	O
still	O
be	O
able	O
to	O
largely	O
avoid	O
a	O
futile	O
cycle	O
since	O
the	O
affinity	O
of	O
ectoine	O
synthase	O
for	O
N	O
-	O
γ	O
-	O
ADABA	O
and	O
N	O
-	O
α	O
-	O
ADABA	O
,	O
and	O
its	O
catalytic	O
efficiency	O
for	O
the	O
two	O
compounds	O
,	O
differs	O
substantially	O
(	O
S4a	O
and	O
S4b	O
Fig	O
).	O
	O
Crystallization	O
of	O
the	O
(	O
Sa	O
)	O
EctC	O
protein	O
	O
Since	O
no	O
crystal	O
structure	O
of	O
ectoine	O
synthase	O
has	O
been	O
reported	O
,	O
we	O
set	O
out	O
to	O
crystallize	O
the	O
(	O
Sa	O
)	O
EctC	O
protein	O
.	O
	O
Attempts	O
to	O
obtain	O
crystals	O
of	O
(	O
Sa	O
)	O
EctC	O
in	O
complex	O
either	O
with	O
its	O
substrate	O
N	O
-	O
γ	O
-	O
ADABA	O
or	O
its	O
reaction	O
product	O
ectoine	O
were	O
not	O
successful	O
.	O
	O
Attempts	O
to	O
solve	O
the	O
crystal	O
structure	O
of	O
the	O
(	O
Sa	O
)	O
EctC	O
protein	O
by	O
molecular	O
replacement	O
has	O
previously	O
failed	O
.	O
	O
However	O
,	O
we	O
were	O
able	O
to	O
obtain	O
crystals	O
of	O
form	O
B	O
that	O
were	O
derivatized	O
with	O
mercury	O
and	O
these	O
diffracted	O
up	O
to	O
2	O
.	O
8	O
Å	O
(	O
S1	O
Table	O
).	O
	O
This	O
dataset	O
was	O
used	O
to	O
derive	O
an	O
initial	O
structural	O
model	O
of	O
the	O
(	O
Sa	O
)	O
EctC	O
protein	O
,	O
which	O
in	O
turn	O
was	O
employed	O
as	O
a	O
template	O
for	O
molecular	O
replacement	O
to	O
phase	O
the	O
native	O
dataset	O
(	O
2	O
.	O
0	O
Å	O
)	O
of	O
crystal	O
form	O
B	O
.	O
After	O
several	O
rounds	O
of	O
manual	O
model	O
building	O
and	O
refinement	O
,	O
four	O
monomers	O
of	O
(	O
Sa	O
)	O
EctC	O
were	O
identified	O
and	O
the	O
crystal	O
structure	O
was	O
refined	O
to	O
a	O
final	O
Rcryst	O
of	O
21	O
.	O
1	O
%	O
and	O
an	O
Rfree	O
of	O
24	O
.	O
8	O
%	O
(	O
S1	O
Table	O
).	O
	O
The	O
two	O
EctC	O
structures	O
that	O
we	O
determined	O
revealed	O
that	O
the	O
ectoine	O
synthase	O
belongs	O
to	O
the	O
cupin	O
superfamily	O
with	O
respect	O
to	O
its	O
overall	O
fold	O
(	O
Fig	O
4a	O
–	O
4c	O
).	O
	O
However	O
,	O
they	O
represent	O
two	O
different	O
states	O
of	O
the	O
137	O
amino	O
acids	O
comprising	O
(	O
Sa	O
)	O
EctC	O
protein	O
(	O
Fig	O
2	O
).	O
	O
First	O
,	O
the	O
1	O
.	O
2	O
Å	O
structure	O
reveals	O
the	O
spatial	O
configuration	O
of	O
the	O
(	O
Sa	O
)	O
EctC	O
protein	O
ranging	O
from	O
amino	O
acid	O
Met	O
-	O
1	O
to	O
Glu	O
-	O
115	O
;	O
hence	O
,	O
it	O
lacks	O
22	O
amino	O
acids	O
at	O
the	O
carboxy	O
-	O
terminus	O
of	O
the	O
authentic	O
(	O
Sa	O
)	O
EctC	O
protein	O
.	O
	O
In	O
this	O
structure	O
no	O
metal	O
co	O
-	O
factor	O
was	O
identified	O
.	O
	O
The	O
second	O
crystal	O
structure	O
of	O
the	O
(	O
Sa	O
)	O
EctC	O
protein	O
was	O
solved	O
at	O
a	O
resolution	O
of	O
2	O
.	O
0	O
Å	O
and	O
contained	O
four	O
molecules	O
of	O
the	O
protein	O
in	O
the	O
asymmetric	O
unit	O
of	O
which	O
protomer	O
A	O
comprised	O
amino	O
acid	O
Met	O
-	O
1	O
to	O
Gly	O
-	O
121	O
and	O
adopts	O
a	O
closed	O
conformation	O
.	O
	O
Hence	O
,	O
it	O
still	O
lacks	O
16	O
amino	O
acid	O
residues	O
of	O
the	O
carboxy	O
-	O
terminus	O
of	O
the	O
authentic	O
137	O
amino	O
acids	O
comprising	O
(	O
Sa	O
)	O
EctC	O
protein	O
(	O
Fig	O
2	O
).	O
	O
We	O
therefore	O
cannot	O
exclude	O
that	O
this	O
crystal	O
structure	O
does	O
not	O
represent	O
the	O
fully	O
closed	O
state	O
of	O
the	O
ectoine	O
synthase	O
;	O
consequently	O
,	O
we	O
tentatively	O
termed	O
it	O
the	O
“	O
semi	O
-	O
closed	O
”	O
(	O
Sa	O
)	O
EctC	O
structure	O
.	O
	O
Overall	O
structure	O
of	O
the	O
“	O
open	O
”	O
and	O
“	O
semi	O
-	O
closed	O
”	O
crystal	O
structures	O
of	O
(	O
Sa	O
)	O
EctC	O
.	O
	O
(	O
a	O
)	O
The	O
overall	O
structure	O
of	O
the	O
“	O
semi	O
-	O
closed	O
”	O
(	O
Sa	O
)	O
EctC	O
resolved	O
at	O
2	O
.	O
0	O
Å	O
is	O
depicted	O
in	O
green	O
in	O
a	O
cartoon	O
(	O
upper	O
panel	O
)	O
and	O
surface	O
(	O
lower	O
panel	O
)	O
representation	O
.	O
	O
(	O
b	O
)	O
The	O
overall	O
structure	O
of	O
the	O
“	O
open	O
”	O
(	O
Sa	O
)	O
EctC	O
was	O
resolved	O
at	O
1	O
.	O
2	O
Å	O
and	O
is	O
depicted	O
in	O
yellow	O
in	O
a	O
cartoon	O
(	O
upper	O
panel	O
)	O
and	O
surface	O
(	O
lower	O
panel	O
)	O
representation	O
.	O
	O
The	O
overall	O
structure	O
of	O
(	O
Sa	O
)	O
EctC	O
is	O
basically	O
the	O
same	O
in	O
both	O
crystals	O
except	O
for	O
the	O
carboxy	O
-	O
terminus	O
,	O
which	O
covers	O
the	O
entry	O
of	O
one	O
side	O
of	O
the	O
cupin	O
barrel	O
from	O
the	O
surroundings	O
in	O
monomer	O
A	O
in	O
the	O
“	O
semi	O
-	O
closed	O
”	O
structure	O
.	O
	O
This	O
is	O
reflected	O
by	O
the	O
calculated	O
root	O
mean	O
square	O
deviation	O
(	O
RMSD	O
)	O
of	O
the	O
Cα	O
atoms	O
that	O
was	O
about	O
0	O
.	O
56	O
Å	O
(	O
over	O
117	O
residues	O
)	O
when	O
the	O
four	O
“	O
open	O
”	O
monomers	O
were	O
compared	O
with	O
each	O
other	O
.	O
	O
However	O
,	O
the	O
“	O
semi	O
-	O
closed	O
”	O
monomer	O
has	O
a	O
slightly	O
higher	O
RMSD	O
of	O
1	O
.	O
4	O
Å	O
(	O
over	O
117	O
residues	O
)	O
when	O
compared	O
with	O
the	O
“	O
open	O
”	O
2	O
.	O
0	O
Å	O
structure	O
.	O
	O
Therefore	O
,	O
we	O
describe	O
in	O
the	O
following	O
the	O
overall	O
structure	O
for	O
the	O
“	O
semi	O
-	O
closed	O
”	O
form	O
of	O
the	O
(	O
Sa	O
)	O
EctC	O
protein	O
and	O
subsequently	O
highlight	O
the	O
structural	O
differences	O
between	O
the	O
“	O
open	O
”	O
and	O
“	O
semi	O
-	O
closed	O
”	O
forms	O
in	O
more	O
detail	O
.	O
	O
The	O
β	O
-	O
strands	O
form	O
two	O
anti	O
-	O
parallel	O
β	O
-	O
sheets	O
:	O
β2	O
β3	O
,	O
β4	O
,	O
β11	O
,	O
β6	O
,	O
and	O
β9	O
,	O
and	O
a	O
smaller	O
three	O
-	O
stranded	O
β	O
-	O
sheet	O
(	O
β7	O
,	O
β8	O
,	O
and	O
β10	O
),	O
respectively	O
.	O
	O
These	O
two	O
β	O
-	O
sheets	O
pack	O
against	O
each	O
other	O
,	O
forming	O
a	O
cup	O
-	O
shaped	O
β	O
-	O
sandwich	O
with	O
a	O
topology	O
characteristic	O
for	O
the	O
cupin	O
-	O
fold	O
.	O
	O
In	O
the	O
“	O
semi	O
-	O
closed	O
”	O
structure	O
,	O
a	O
longer	O
carboxy	O
-	O
terminal	O
tail	O
is	O
visible	O
in	O
the	O
electron	O
density	O
,	O
folding	O
into	O
a	O
small	O
helix	O
(	O
α	O
-	O
II	O
)	O
that	O
closes	O
the	O
active	O
site	O
of	O
the	O
(	O
Sa	O
)	O
EctC	O
protein	O
(	O
Fig	O
4a	O
).	O
	O
Structural	O
comparison	O
analyses	O
using	O
the	O
DALI	O
server	O
revealed	O
that	O
(	O
Sa	O
)	O
EctC	O
adopts	O
a	O
fold	O
similar	O
to	O
other	O
members	O
of	O
the	O
cupin	O
superfamily	O
.	O
	O
The	O
highest	O
structural	O
similarities	O
are	O
observed	O
for	O
the	O
Cupin	O
2	O
conserved	O
barrel	O
domain	O
protein	O
(	O
YP_751781	O
.	O
1	O
)	O
from	O
Shewanella	O
frigidimarina	O
(	O
PDB	O
accession	O
code	O
:	O
2PFW	O
)	O
with	O
a	O
Z	O
-	O
score	O
of	O
13	O
.	O
1	O
and	O
an	O
RMSD	O
of	O
2	O
.	O
2	O
Å	O
over	O
104	O
Cα	O
-	O
atoms	O
(	O
structural	O
data	O
for	O
this	O
protein	O
have	O
been	O
deposited	O
in	O
the	O
PDB	O
but	O
no	O
publication	O
connected	O
to	O
this	O
structure	O
is	O
currently	O
available	O
),	O
a	O
manganese	O
-	O
containing	O
cupin	O
(	O
TM1459	O
)	O
from	O
Thermotoga	O
maritima	O
(	O
PDB	O
accession	O
code	O
:	O
1VJ2	O
)	O
with	O
a	O
Z	O
-	O
score	O
of	O
12	O
.	O
8	O
and	O
an	O
RMSD	O
of	O
2	O
.	O
0	O
Å	O
over	O
103	O
Cα	O
-	O
atoms	O
,	O
the	O
cyclase	O
RemF	O
from	O
Streptomyces	O
resistomycificus	O
(	O
PDB	O
accession	O
code	O
:	O
3HT1	O
with	O
a	O
Z	O
-	O
score	O
of	O
11	O
.	O
9	O
and	O
an	O
RMSD	O
of	O
1	O
.	O
9	O
Å	O
over	O
102	O
Cα	O
-	O
atoms	O
),	O
and	O
an	O
auxin	O
-	O
binding	O
protein	O
1	O
from	O
Zea	O
mays	O
(	O
PDB	O
accession	O
code	O
:	O
1LR5	O
)	O
with	O
an	O
Z	O
-	O
score	O
of	O
11	O
.	O
8	O
and	O
an	O
RMSD	O
of	O
2	O
.	O
8	O
Å	O
over	O
104	O
Cα	O
-	O
atoms	O
).	O
	O
Next	O
to	O
RemF	O
and	O
the	O
aldos	O
-	O
2	O
-	O
ulose	O
dehydratase	O
/	O
isomerase	O
,	O
the	O
ectoine	O
synthase	O
is	O
only	O
the	O
third	O
characterized	O
dehydratase	O
within	O
the	O
cupin	O
superfamily	O
.	O
	O
In	O
the	O
“	O
semi	O
-	O
closed	O
”	O
crystal	O
structure	O
,	O
(	O
Sa	O
)	O
EctC	O
has	O
crystallized	O
as	O
a	O
dimer	O
of	O
dimers	O
within	O
the	O
asymmetric	O
unit	O
.	O
	O
This	O
dimer	O
(	O
Fig	O
5a	O
and	O
5b	O
)	O
is	O
composed	O
of	O
two	O
monomers	O
arranged	O
in	O
a	O
head	O
-	O
to	O
-	O
tail	O
orientation	O
and	O
is	O
stabilized	O
via	O
strong	O
interactions	O
mediated	O
by	O
two	O
antiparallel	O
β	O
-	O
strands	O
,	O
β	O
-	O
strand	O
β1	O
(	O
sequence	O
1MIVRN5	O
)	O
from	O
monomer	O
A	O
and	O
β	O
-	O
strand	O
β8	O
from	O
monomer	O
B	O
(	O
sequence	O
82GVMYAL87	O
)	O
(	O
Fig	O
5c	O
).	O
	O
The	O
strong	O
interactions	O
between	O
these	O
β	O
-	O
strands	O
rely	O
primarily	O
on	O
backbone	O
contacts	O
.	O
	O
In	O
addition	O
to	O
these	O
interactions	O
,	O
some	O
weaker	O
hydrophobic	O
interactions	O
are	O
also	O
observed	O
between	O
the	O
two	O
monomers	O
in	O
some	O
loops	O
connecting	O
the	O
β	O
-	O
strands	O
.	O
	O
Both	O
values	O
fall	O
within	O
the	O
range	O
for	O
known	O
functional	O
dimers	O
.	O
	O
Crystal	O
structure	O
of	O
(	O
Sa	O
)	O
EctC	O
.	O
	O
(	O
a	O
)	O
Top	O
-	O
view	O
of	O
the	O
dimer	O
of	O
the	O
(	O
Sa	O
)	O
EctC	O
protein	O
.	O
	O
The	O
position	O
of	O
the	O
water	O
molecule	O
,	O
described	O
in	O
detail	O
in	O
the	O
text	O
,	O
is	O
shown	O
in	O
one	O
of	O
the	O
monomers	O
as	O
an	O
orange	O
sphere	O
.	O
(	O
b	O
)	O
Side	O
-	O
view	O
of	O
a	O
(	O
Sa	O
)	O
EctC	O
dimer	O
allowing	O
an	O
assessment	O
of	O
the	O
dimer	O
interface	O
formed	O
by	O
two	O
β	O
-	O
strands	O
of	O
each	O
monomer	O
.	O
	O
(	O
c	O
)	O
Close	O
-	O
up	O
representation	O
of	O
the	O
dimer	O
interface	O
mediated	O
by	O
beta	O
-	O
strand	O
β1	O
and	O
β6	O
.	O
	O
Indeed	O
,	O
a	O
similar	O
dimer	O
configuration	O
to	O
the	O
one	O
described	O
for	O
the	O
“	O
semi	O
-	O
closed	O
”	O
(	O
Sa	O
)	O
EctC	O
structure	O
is	O
observed	O
with	O
the	O
same	O
monomer	O
-	O
monomer	O
interactions	O
mediated	O
by	O
the	O
two	O
β	O
-	O
sheets	O
.	O
	O
The	O
crystallographic	O
two	O
-	O
fold	O
axis	O
present	O
within	O
the	O
crystal	O
symmetry	O
is	O
located	O
exactly	O
in	O
between	O
the	O
two	O
monomers	O
,	O
resulting	O
in	O
a	O
monomer	O
within	O
the	O
asymmetric	O
unit	O
.	O
	O
Hence	O
,	O
the	O
same	O
dimer	O
observed	O
in	O
the	O
“	O
semi	O
-	O
closed	O
”	O
structure	O
of	O
(	O
Sa	O
)	O
EctC	O
can	O
also	O
be	O
observed	O
in	O
the	O
“	O
open	O
”	O
structure	O
.	O
	O
Interestingly	O
,	O
the	O
proteins	O
identified	O
by	O
the	O
above	O
-	O
described	O
DALI	O
search	O
not	O
only	O
have	O
folds	O
similar	O
to	O
EctC	O
,	O
but	O
are	O
also	O
functional	O
dimers	O
that	O
adopt	O
similar	O
monomer	O
-	O
monomer	O
interactions	O
within	O
the	O
dimer	O
assembly	O
as	O
deduced	O
from	O
the	O
inspection	O
of	O
the	O
corresponding	O
PDB	O
files	O
(	O
2PFW	O
,	O
3HT1	O
,	O
1VJ2	O
,	O
1LR5	O
).	O
	O
Structural	O
rearrangements	O
of	O
the	O
flexible	O
(	O
Sa	O
)	O
EctC	O
carboxy	O
-	O
terminus	O
	O
The	O
cupin	O
core	O
represents	O
the	O
structural	O
framework	O
of	O
ectoine	O
synthase	O
(	O
Figs	O
4	O
and	O
5	O
).	O
	O
The	O
major	O
difference	O
in	O
the	O
two	O
crystal	O
structures	O
of	O
the	O
(	O
Sa	O
)	O
EctC	O
protein	O
reported	O
here	O
is	O
the	O
orientation	O
of	O
the	O
carboxy	O
-	O
terminus	O
.	O
	O
Some	O
amino	O
acids	O
located	O
in	O
the	O
carboxy	O
-	O
terminal	O
region	O
of	O
the	O
137	O
amino	O
acids	O
comprising	O
(	O
Sa	O
)	O
EctC	O
protein	O
are	O
highly	O
conserved	O
(	O
Fig	O
2	O
)	O
within	O
the	O
extended	O
EctC	O
protein	O
family	O
.	O
	O
At	O
the	O
end	O
of	O
β	O
-	O
strand	O
β11	O
,	O
two	O
consecutive	O
conserved	O
proline	O
residues	O
(	O
Pro	O
-	O
109	O
and	O
Pro	O
-	O
110	O
)	O
are	O
present	O
that	O
are	O
responsible	O
for	O
a	O
turn	O
in	O
the	O
main	O
chain	O
of	O
the	O
(	O
Sa	O
)	O
EctC	O
protein	O
.	O
	O
In	O
the	O
“	O
semi	O
-	O
closed	O
”	O
(	O
Sa	O
)	O
EctC	O
structure	O
,	O
the	O
visible	O
electron	O
density	O
of	O
the	O
carboxy	O
-	O
terminus	O
is	O
extended	O
by	O
7	O
amino	O
acid	O
residues	O
and	O
ends	O
at	O
position	O
Gly	O
-	O
121	O
.	O
	O
Furthermore	O
,	O
this	O
helix	O
is	O
stabilized	O
via	O
interactions	O
with	O
the	O
loop	O
region	O
between	O
β	O
-	O
strands	O
β4	O
and	O
β6	O
,	O
thereby	O
inducing	O
a	O
structural	O
rearrangement	O
.	O
	O
This	O
induces	O
the	O
formation	O
of	O
β	O
-	O
strand	O
β5	O
,	O
which	O
is	O
not	O
present	O
when	O
the	O
small	O
C	O
-	O
terminal	O
helix	O
is	O
absent	O
as	O
observed	O
in	O
the	O
“	O
open	O
”	O
(	O
Sa	O
)	O
EctC	O
structure	O
.	O
	O
The	O
position	O
of	O
this	O
His	O
residue	O
is	O
slightly	O
shifted	O
in	O
both	O
(	O
Sa	O
)	O
EctC	O
structures	O
,	O
likely	O
the	O
result	O
of	O
the	O
formation	O
of	O
β	O
-	O
strand	O
β5	O
.	O
	O
The	O
consecutive	O
Pro	O
-	O
109	O
and	O
Pro	O
-	O
110	O
residues	O
found	O
at	O
the	O
end	O
of	O
β	O
-	O
strand	O
β11are	O
highly	O
conserved	O
in	O
EctC	O
-	O
type	O
proteins	O
(	O
Fig	O
2	O
).	O
	O
They	O
are	O
responsible	O
for	O
redirecting	O
the	O
main	O
chain	O
of	O
the	O
remaining	O
carboxy	O
-	O
terminus	O
(	O
27	O
amino	O
acid	O
residues	O
)	O
of	O
(	O
Sa	O
)	O
EctC	O
to	O
close	O
the	O
cupin	O
fold	O
.	O
	O
In	O
the	O
“	O
semi	O
-	O
closed	O
”	O
structure	O
this	O
results	O
in	O
a	O
complete	O
closure	O
of	O
the	O
entry	O
of	O
the	O
cupin	O
barrel	O
(	O
Fig	O
4a	O
to	O
4c	O
).	O
	O
A	O
search	O
for	O
partners	O
interacting	O
with	O
Pro	O
-	O
109	O
revealed	O
that	O
it	O
interacts	O
via	O
its	O
backbone	O
oxygen	O
with	O
the	O
side	O
chain	O
of	O
His	O
-	O
55	O
as	O
visible	O
in	O
both	O
the	O
“	O
open	O
”	O
and	O
“	O
semi	O
-	O
closed	O
”	O
(	O
Sa	O
)	O
EctC	O
structures	O
.	O
	O
The	O
Pro	O
-	O
109	O
/	O
His	O
-	O
55	O
interaction	O
ensures	O
the	O
stable	O
orientation	O
of	O
both	O
proline	O
residues	O
at	O
the	O
end	O
of	O
β	O
-	O
strand	O
β11	O
.	O
	O
In	O
addition	O
to	O
the	O
interactions	O
between	O
Pro	O
-	O
109	O
and	O
His	O
-	O
55	O
,	O
the	O
carboxy	O
-	O
terminal	O
region	O
of	O
(	O
Sa	O
)	O
EctC	O
is	O
held	O
in	O
position	O
via	O
an	O
interaction	O
of	O
Glu	O
-	O
115	O
with	O
His	O
-	O
55	O
,	O
which	O
stabilizes	O
the	O
conformation	O
of	O
the	O
small	O
helix	O
in	O
the	O
carboxy	O
-	O
terminus	O
further	O
.	O
	O
Architecture	O
of	O
the	O
presumed	O
metal	O
-	O
binding	O
site	O
of	O
the	O
(	O
Sa	O
)	O
EctC	O
protein	O
and	O
its	O
flexible	O
carboxy	O
-	O
terminus	O
.	O
	O
(	O
a	O
)	O
The	O
described	O
water	O
molecule	O
(	O
depicted	O
as	O
orange	O
sphere	O
)	O
is	O
bound	O
via	O
interactions	O
with	O
the	O
side	O
chains	O
of	O
Glu	O
-	O
57	O
,	O
Tyr	O
-	O
85	O
,	O
and	O
His	O
-	O
93	O
.	O
	O
The	O
position	O
occupied	O
by	O
this	O
water	O
molecule	O
represents	O
probably	O
the	O
position	O
of	O
the	O
Fe2	O
+	O
cofactor	O
in	O
the	O
active	O
side	O
of	O
the	O
ectoine	O
synthase	O
.	O
	O
His	O
-	O
55	O
interacts	O
with	O
the	O
double	O
proline	O
motif	O
(	O
Pro	O
-	O
109	O
and	O
Pro	O
-	O
110	O
).	O
	O
It	O
is	O
further	O
stabilized	O
via	O
an	O
interaction	O
with	O
the	O
side	O
chain	O
of	O
Glu	O
-	O
115	O
which	O
is	O
localized	O
in	O
the	O
flexible	O
carboxy	O
-	O
terminus	O
(	O
colored	O
in	O
orange	O
)	O
of	O
(	O
Sa	O
)	O
EctC	O
that	O
is	O
visible	O
in	O
the	O
“	O
semi	O
-	O
closed	O
”	O
(	O
Sa	O
)	O
EctC	O
structure	O
.	O
	O
Since	O
(	O
Sa	O
)	O
EctC	O
is	O
a	O
metal	O
containing	O
protein	O
(	O
Fig	O
3	O
),	O
we	O
tried	O
to	O
fit	O
either	O
Fe2	O
+,	O
or	O
Zn2	O
+	O
ions	O
into	O
this	O
density	O
and	O
also	O
refined	O
occupancy	O
.	O
	O
Only	O
the	O
refinement	O
of	O
Fe2	O
+	O
resulted	O
in	O
a	O
visibly	O
improved	O
electron	O
density	O
,	O
however	O
with	O
a	O
low	O
degree	O
of	O
occupancy	O
.	O
	O
This	O
possible	O
iron	O
molecule	O
is	O
bound	O
via	O
interactions	O
with	O
Glu	O
-	O
57	O
,	O
Tyr	O
-	O
85	O
and	O
His	O
-	O
93	O
(	O
Fig	O
6a	O
and	O
6b	O
).	O
	O
The	O
distance	O
between	O
the	O
side	O
chains	O
of	O
these	O
residues	O
and	O
the	O
(	O
putative	O
)	O
iron	O
co	O
-	O
factor	O
is	O
3	O
.	O
1	O
Å	O
for	O
Glu	O
-	O
57	O
,	O
2	O
.	O
9	O
Å	O
for	O
Tyr	O
-	O
85	O
,	O
and	O
2	O
.	O
9	O
Å	O
for	O
His	O
-	O
93	O
,	O
respectively	O
.	O
	O
The	O
position	O
of	O
this	O
water	O
molecule	O
is	O
described	O
in	O
more	O
detail	O
below	O
and	O
is	O
highlighted	O
in	O
Figs	O
5a	O
and	O
5b	O
and	O
6a	O
and	O
6b	O
as	O
a	O
sphere	O
.	O
	O
Interestingly	O
,	O
all	O
three	O
amino	O
acids	O
coordinating	O
this	O
water	O
molecule	O
are	O
strictly	O
conserved	O
within	O
an	O
alignment	O
of	O
440	O
members	O
of	O
the	O
EctC	O
protein	O
family	O
(	O
for	O
an	O
abbreviated	O
alignment	O
of	O
EctC	O
-	O
type	O
proteins	O
see	O
Fig	O
2	O
).	O
	O
In	O
the	O
“	O
open	O
”	O
structure	O
of	O
the	O
(	O
Sa	O
)	O
EctC	O
protein	O
,	O
electron	O
density	O
is	O
visible	O
where	O
the	O
presumptive	O
iron	O
is	O
positioned	O
in	O
the	O
“	O
semi	O
-	O
closed	O
”	O
structure	O
.	O
	O
However	O
,	O
this	O
electron	O
density	O
fits	O
perfectly	O
to	O
a	O
water	O
molecule	O
and	O
not	O
to	O
an	O
iron	O
,	O
and	O
the	O
water	O
molecule	O
was	O
clearly	O
visible	O
after	O
the	O
refinement	O
at	O
this	O
high	O
resolution	O
(	O
1	O
.	O
2	O
Å	O
)	O
of	O
the	O
“	O
open	O
”	O
(	O
Sa	O
)	O
EctC	O
structure	O
.	O
	O
In	O
a	O
superimposition	O
of	O
both	O
(	O
Sa	O
)	O
EctC	O
crystal	O
structures	O
,	O
the	O
spatial	O
arrangements	O
of	O
the	O
side	O
chains	O
of	O
the	O
three	O
amino	O
acids	O
(	O
Glu	O
-	O
57	O
,	O
Tyr	O
-	O
85	O
,	O
and	O
His	O
-	O
93	O
)	O
likely	O
to	O
contact	O
the	O
iron	O
in	O
the	O
“	O
semi	O
-	O
closed	O
”	O
structure	O
match	O
nicely	O
with	O
those	O
of	O
the	O
corresponding	O
residues	O
of	O
the	O
“	O
iron	O
-	O
free	O
”	O
“	O
open	O
”	O
structure	O
(	O
Fig	O
6b	O
).	O
	O
In	O
the	O
“	O
semi	O
-	O
closed	O
”	O
structure	O
,	O
the	O
hydroxyl	O
-	O
group	O
of	O
the	O
side	O
-	O
chain	O
of	O
Tyr	O
-	O
52	O
points	O
towards	O
the	O
iron	O
(	O
Fig	O
6a	O
and	O
6b	O
),	O
but	O
the	O
corresponding	O
distance	O
(	O
3	O
.	O
9	O
Å	O
)	O
makes	O
it	O
highly	O
unlikely	O
that	O
Tyr	O
-	O
52	O
is	O
directly	O
involved	O
in	O
metal	O
binding	O
.	O
	O
Nevertheless	O
,	O
its	O
substitution	O
by	O
an	O
Ala	O
residue	O
causes	O
a	O
strong	O
decrease	O
in	O
iron	O
-	O
content	O
and	O
enzyme	O
activity	O
of	O
the	O
mutant	O
protein	O
(	O
Table	O
1	O
).	O
	O
Since	O
Tyr	O
-	O
52	O
is	O
strictly	O
conserved	O
in	O
an	O
alignment	O
of	O
440	O
EctC	O
-	O
type	O
proteins	O
(	O
Fig	O
2	O
),	O
we	O
speculate	O
that	O
it	O
might	O
be	O
involved	O
in	O
contacting	O
the	O
substrate	O
of	O
the	O
ectoine	O
synthase	O
and	O
that	O
the	O
absence	O
of	O
N	O
-	O
γ	O
-	O
ADABA	O
in	O
our	O
(	O
Sa	O
)	O
EctC	O
crystal	O
structures	O
might	O
endow	O
the	O
side	O
chain	O
of	O
Tyr	O
-	O
52	O
with	O
extra	O
spatial	O
flexibility	O
.	O
	O
To	O
further	O
analyze	O
the	O
putative	O
iron	O
binding	O
site	O
(	O
Fig	O
6a	O
),	O
we	O
performed	O
structure	O
-	O
guided	O
site	O
-	O
directed	O
mutagenesis	O
and	O
assessed	O
the	O
resulting	O
(	O
Sa	O
)	O
EctC	O
variants	O
for	O
their	O
iron	O
content	O
and	O
studied	O
their	O
enzyme	O
activity	O
.	O
	O
When	O
those	O
three	O
residues	O
(	O
Glu	O
-	O
57	O
,	O
Tyr	O
-	O
85	O
,	O
His	O
-	O
93	O
)	O
that	O
likely	O
form	O
the	O
mono	O
-	O
nuclear	O
iron	O
center	O
in	O
the	O
(	O
Sa	O
)	O
EctC	O
crystal	O
structure	O
were	O
individually	O
replaced	O
by	O
an	O
Ala	O
residue	O
,	O
both	O
the	O
catalytic	O
activity	O
and	O
the	O
iron	O
content	O
of	O
the	O
mutant	O
proteins	O
was	O
strongly	O
reduced	O
(	O
Table	O
1	O
).	O
	O
For	O
some	O
of	O
the	O
presumptive	O
iron	O
-	O
coordinating	O
residues	O
,	O
additional	O
site	O
-	O
directed	O
mutagenesis	O
experiments	O
were	O
carried	O
out	O
.	O
	O
To	O
verify	O
the	O
importance	O
of	O
the	O
negative	O
charge	O
in	O
the	O
position	O
of	O
Glu	O
-	O
57	O
,	O
we	O
created	O
an	O
Asp	O
variant	O
.	O
	O
This	O
mutant	O
protein	O
rescued	O
the	O
enzyme	O
activity	O
and	O
iron	O
content	O
of	O
the	O
Ala	O
substitution	O
substantially	O
(	O
Table	O
1	O
).	O
	O
Collectively	O
,	O
these	O
data	O
suggest	O
that	O
the	O
hydroxyl	O
group	O
of	O
the	O
Tyr	O
-	O
85	O
side	O
chain	O
is	O
needed	O
for	O
the	O
binding	O
of	O
the	O
iron	O
(	O
Fig	O
6a	O
).	O
	O
We	O
also	O
replaced	O
the	O
presumptive	O
iron	O
-	O
binding	O
residue	O
His	O
-	O
93	O
by	O
an	O
Asn	O
residue	O
,	O
yielding	O
a	O
(	O
Sa	O
)	O
EctC	O
protein	O
variant	O
that	O
possessed	O
an	O
enzyme	O
activity	O
of	O
23	O
%	O
and	O
iron	O
content	O
of	O
only	O
14	O
%	O
relative	O
to	O
that	O
of	O
the	O
wild	O
-	O
type	O
protein	O
(	O
Table	O
1	O
).	O
	O
Collectively	O
,	O
the	O
data	O
addressing	O
the	O
functionality	O
of	O
the	O
putative	O
iron	O
-	O
coordinating	O
residues	O
(	O
Glu	O
-	O
57	O
,	O
Tyr	O
-	O
85	O
,	O
His	O
-	O
93	O
)	O
buttress	O
our	O
notion	O
that	O
the	O
Fe2	O
+	O
present	O
in	O
the	O
(	O
Sa	O
)	O
EctC	O
protein	O
is	O
of	O
catalytic	O
importance	O
.	O
	O
A	O
chemically	O
undefined	O
ligand	O
in	O
the	O
(	O
Sa	O
)	O
EctC	O
structure	O
provides	O
clues	O
for	O
the	O
binding	O
of	O
the	O
N	O
-	O
γ	O
-	O
ADABA	O
substrate	O
	O
Despite	O
considerable	O
efforts	O
,	O
either	O
by	O
trying	O
co	O
-	O
crystallization	O
or	O
soaking	O
experiments	O
,	O
we	O
were	O
not	O
able	O
to	O
obtain	O
a	O
(	O
Sa	O
)	O
EctC	O
crystal	O
structures	O
that	O
contained	O
either	O
the	O
substrate	O
N	O
-	O
γ	O
-	O
ADABA	O
,	O
or	O
ectoine	O
,	O
the	O
reaction	O
product	O
of	O
ectoine	O
synthase	O
(	O
Fig	O
1	O
).	O
	O
However	O
,	O
in	O
the	O
“	O
semi	O
-	O
closed	O
”	O
(	O
Sa	O
)	O
EctC	O
structure	O
where	O
the	O
carboxy	O
-	O
terminal	O
loop	O
is	O
largely	O
resolved	O
,	O
a	O
long	O
stretched	O
electron	O
density	O
feature	O
was	O
detected	O
in	O
the	O
predicted	O
active	O
site	O
of	O
the	O
enzyme	O
;	O
it	O
remained	O
visible	O
after	O
crystallographic	O
refinement	O
.	O
	O
We	O
tried	O
to	O
fit	O
all	O
compounds	O
used	O
in	O
the	O
buffers	O
during	O
purification	O
and	O
crystallization	O
into	O
the	O
observed	O
electron	O
density	O
,	O
but	O
none	O
matched	O
.	O
	O
This	O
observation	O
indicates	O
that	O
the	O
chemically	O
undefined	O
ligand	O
was	O
either	O
trapped	O
by	O
the	O
(	O
Sa	O
)	O
EctC	O
protein	O
during	O
its	O
heterologous	O
production	O
in	O
E	O
.	O
coli	O
or	O
during	O
crystallization	O
.	O
	O
Estimating	O
from	O
the	O
dimensions	O
of	O
the	O
electron	O
density	O
feature	O
,	O
we	O
modeled	O
the	O
chemically	O
undefined	O
compound	O
trapped	O
by	O
the	O
(	O
Sa	O
)	O
EctC	O
protein	O
as	O
a	O
hexane	O
-	O
1	O
,	O
6	O
-	O
diol	O
molecule	O
(	O
PDB	O
identifier	O
:	O
HEZ	O
)	O
to	O
best	O
fit	O
the	O
observed	O
electron	O
density	O
.	O
	O
However	O
,	O
to	O
the	O
best	O
of	O
our	O
knowledge	O
,	O
hexane	O
-	O
1	O
,	O
6	O
-	O
diol	O
is	O
not	O
part	O
of	O
the	O
E	O
.	O
coli	O
metabolome	O
.	O
	O
We	O
note	O
that	O
both	O
N	O
-	O
γ	O
-	O
ADABA	O
and	O
hexane	O
-	O
1	O
,	O
6	O
-	O
diol	O
are	O
both	O
C6	O
-	O
compounds	O
and	O
display	O
similar	O
length	O
(	O
Fig	O
7a	O
).	O
	O
A	O
chemically	O
undefined	O
ligand	O
is	O
captured	O
in	O
the	O
active	O
site	O
of	O
the	O
“	O
semi	O
-	O
closed	O
”	O
(	O
Sa	O
)	O
EctC	O
crystal	O
structure	O
.	O
	O
(	O
a	O
)	O
The	O
observed	O
electron	O
density	O
in	O
the	O
active	O
site	O
of	O
the	O
“	O
semi	O
-	O
closed	O
”	O
structure	O
of	O
(	O
Sa	O
)	O
EctC	O
is	O
modeled	O
as	O
a	O
hexane	O
-	O
1	O
,	O
6	O
-	O
diol	O
molecule	O
and	O
compared	O
with	O
the	O
electron	O
density	O
of	O
the	O
N	O
-	O
γ	O
-	O
ADABA	O
substrate	O
of	O
the	O
ectoine	O
synthase	O
to	O
emphasize	O
the	O
similarity	O
in	O
size	O
of	O
these	O
compounds	O
.	O
	O
(	O
b	O
)	O
The	O
presumable	O
binding	O
site	O
of	O
the	O
iron	O
co	O
-	O
factor	O
and	O
of	O
the	O
modeled	O
hexane	O
-	O
1	O
,	O
6	O
-	O
diol	O
molecule	O
is	O
depicted	O
.	O
	O
The	O
amino	O
acid	O
side	O
chains	O
involved	O
in	O
iron	O
-	O
ligand	O
binding	O
are	O
colored	O
in	O
blue	O
and	O
those	O
involved	O
in	O
the	O
binding	O
of	O
the	O
chemically	O
undefined	O
ligand	O
are	O
colored	O
in	O
green	O
using	O
a	O
ball	O
and	O
stick	O
representation	O
.	O
	O
The	O
flexible	O
carboxy	O
-	O
terminal	O
loop	O
of	O
(	O
Sa	O
)	O
EctC	O
is	O
highlighted	O
in	O
orange	O
.	O
	O
We	O
refined	O
the	O
(	O
Sa	O
)	O
EctC	O
structure	O
with	O
the	O
trapped	O
compound	O
,	O
and	O
by	O
doing	O
so	O
,	O
the	O
refinement	O
parameters	O
(	O
especially	O
R	O
-	O
and	O
Rfree	O
-	O
factor	O
)	O
dropped	O
by	O
1	O
.	O
5	O
%.	O
	O
Remarkably	O
,	O
all	O
of	O
these	O
residues	O
are	O
highly	O
conserved	O
throughout	O
the	O
extended	O
EctC	O
protein	O
family	O
(	O
Fig	O
2	O
).	O
	O
Structure	O
-	O
guided	O
site	O
-	O
directed	O
mutagenesis	O
of	O
the	O
catalytic	O
core	O
of	O
the	O
ectoine	O
synthase	O
	O
In	O
a	O
previous	O
alignment	O
of	O
the	O
amino	O
acid	O
sequences	O
of	O
440	O
EctC	O
-	O
type	O
proteins	O
,	O
13	O
amino	O
acids	O
were	O
identified	O
as	O
strictly	O
conserved	O
residues	O
.	O
	O
Amino	O
acid	O
residues	O
Gly	O
-	O
64	O
,	O
Pro	O
-	O
109	O
,	O
and	O
Gly	O
-	O
113	O
likely	O
fulfill	O
structural	O
roles	O
since	O
they	O
are	O
positioned	O
either	O
at	O
the	O
end	O
or	O
at	O
the	O
beginning	O
of	O
β	O
-	O
strands	O
and	O
α	O
-	O
helices	O
.	O
	O
We	O
considered	O
the	O
remaining	O
ten	O
residues	O
as	O
important	O
either	O
for	O
ligand	O
binding	O
,	O
for	O
catalysis	O
,	O
or	O
for	O
the	O
structurally	O
correct	O
orientation	O
of	O
the	O
flexible	O
carboxy	O
-	O
terminus	O
of	O
the	O
(	O
Sa	O
)	O
EctC	O
protein	O
.	O
	O
In	O
view	O
of	O
the	O
(	O
Sa	O
)	O
EctC	O
structure	O
with	O
the	O
serendipitously	O
trapped	O
compound	O
(	O
Fig	O
7b	O
),	O
we	O
probed	O
the	O
functional	O
importance	O
of	O
the	O
seven	O
residues	O
that	O
contact	O
this	O
ligand	O
by	O
structure	O
-	O
guided	O
site	O
-	O
directed	O
mutagenesis	O
(	O
Table	O
1	O
).	O
	O
We	O
benchmarked	O
the	O
activity	O
of	O
the	O
(	O
Sa	O
)	O
EctC	O
variants	O
in	O
a	O
single	O
time	O
-	O
point	O
enzyme	O
assay	O
under	O
conditions	O
where	O
10	O
μM	O
of	O
the	O
wild	O
-	O
type	O
(	O
Sa	O
)	O
EctC	O
protein	O
converted	O
almost	O
completely	O
the	O
supplied	O
10	O
mM	O
N	O
-	O
γ	O
-	O
ADABA	O
substrate	O
to	O
9	O
.	O
33	O
mM	O
ectoine	O
within	O
a	O
time	O
frame	O
of	O
20	O
min	O
.	O
	O
In	O
addition	O
,	O
we	O
determined	O
the	O
iron	O
content	O
of	O
each	O
of	O
the	O
mutant	O
(	O
Sa	O
)	O
EctC	O
protein	O
by	O
a	O
colorimetric	O
assay	O
(	O
Table	O
1	O
).	O
	O
The	O
side	O
chains	O
of	O
the	O
evolutionarily	O
conserved	O
Trp	O
-	O
21	O
,	O
Ser	O
-	O
23	O
,	O
Thr	O
-	O
40	O
,	O
Cys	O
-	O
105	O
,	O
and	O
Phe	O
-	O
107	O
residues	O
(	O
Fig	O
2	O
)	O
make	O
contacts	O
with	O
the	O
chemically	O
undefined	O
ligand	O
that	O
we	O
observed	O
in	O
the	O
“	O
semi	O
-	O
closed	O
”	O
(	O
Sa	O
)	O
EctC	O
structure	O
(	O
Fig	O
7b	O
).	O
	O
Thr	O
-	O
40	O
is	O
positioned	O
on	O
β	O
-	O
strand	O
β5	O
and	O
its	O
side	O
chain	O
protrudes	O
into	O
the	O
lumen	O
of	O
the	O
cupin	O
barrel	O
formed	O
by	O
the	O
(	O
Sa	O
)	O
EctC	O
protein	O
(	O
Fig	O
7b	O
).	O
	O
We	O
also	O
replaced	O
Phe	O
-	O
107	O
with	O
either	O
an	O
Tyr	O
or	O
an	O
Trp	O
residue	O
:	O
the	O
Phe	B-mutant
-	I-mutant
107	I-mutant
/	I-mutant
Tyr	I-mutant
substitution	O
possessed	O
near	O
wild	O
-	O
type	O
enzyme	O
activity	O
(	O
about	O
95	O
%)	O
and	O
the	O
full	O
iron	O
content	O
,	O
but	O
the	O
Phe	B-mutant
-	I-mutant
107	I-mutant
/	I-mutant
Trp	I-mutant
substitution	O
possessed	O
only	O
12	O
%	O
enzyme	O
activity	O
and	O
72	O
%	O
iron	O
content	O
compared	O
to	O
the	O
wild	O
-	O
type	O
protein	O
.	O
	O
The	O
properties	O
of	O
these	O
mutant	O
proteins	O
indicate	O
that	O
the	O
aromatic	O
side	O
chain	O
at	O
position	O
107	O
of	O
(	O
Sa	O
)	O
EctC	O
is	O
of	O
importance	O
but	O
that	O
a	O
substitution	O
with	O
a	O
bulky	O
aromatic	O
side	O
chain	O
is	O
strongly	O
detrimental	O
to	O
enzyme	O
activity	O
and	O
concomitantly	O
moderately	O
impairs	O
iron	O
binding	O
.	O
	O
Replacement	O
of	O
the	O
only	O
Cys	O
residue	O
in	O
(	O
Sa	O
)	O
EctC	O
(	O
Cys	O
-	O
105	O
;	O
Fig	O
2	O
)	O
by	O
a	O
Ser	O
residue	O
,	O
a	O
configuration	O
that	O
is	O
naturally	O
found	O
in	O
two	O
EctC	O
proteins	O
among	O
440	O
inspected	O
amino	O
acid	O
sequences	O
,	O
yielded	O
a	O
(	O
Sa	O
)	O
EctC	O
variant	O
with	O
84	O
%	O
wild	O
-	O
type	O
activity	O
and	O
an	O
iron	O
content	O
similar	O
to	O
that	O
of	O
the	O
wild	O
-	O
type	O
protein	O
.	O
	O
Since	O
the	O
side	O
-	O
chains	O
of	O
Cys	O
residues	O
are	O
chemically	O
reactive	O
and	O
often	O
participate	O
in	O
enzyme	O
catalysis	O
,	O
Cys	O
-	O
105	O
(	O
or	O
Ser	O
-	O
105	O
)	O
might	O
serve	O
such	O
a	O
role	O
for	O
ectoine	O
synthase	O
.	O
	O
Based	O
on	O
the	O
(	O
Sa	O
)	O
EctC	O
crystal	O
structures	O
that	O
we	O
present	O
here	O
,	O
we	O
can	O
currently	O
not	O
firmly	O
understand	O
why	O
the	O
replacement	O
of	O
Tyr	O
-	O
52	O
by	O
Ala	O
impairs	O
enzyme	O
function	O
and	O
iron	O
content	O
so	O
drastically	O
(	O
Table	O
1	O
).	O
	O
This	O
is	O
different	O
for	O
the	O
His	B-mutant
-	I-mutant
55	I-mutant
/	I-mutant
Ala	I-mutant
substitution	O
.	O
	O
The	O
individual	O
substitution	O
of	O
either	O
Glu	O
-	O
115	O
or	O
His	O
-	O
55	O
by	O
an	O
Ala	O
residue	O
is	O
predicted	O
to	O
disrupt	O
this	O
interactive	O
network	O
and	O
therefore	O
should	O
affect	O
enzyme	O
activity	O
.	O
	O
Indeed	O
,	O
the	O
Glu	B-mutant
-	I-mutant
115	I-mutant
/	I-mutant
Ala	I-mutant
and	O
the	O
His	B-mutant
-	I-mutant
55	I-mutant
/	I-mutant
Ala	I-mutant
substitutions	O
possessed	O
only	O
21	O
%	O
and	O
16	O
%	O
activity	O
of	O
the	O
wild	O
-	O
type	O
protein	O
,	O
respectively	O
(	O
Table	O
1	O
).	O
	O
We	O
also	O
replaced	O
Glu	O
-	O
115	O
with	O
a	O
negatively	O
charged	O
residue	O
(	O
Asp	O
);	O
this	O
(	O
Sa	O
)	O
EctC	O
variant	O
possessed	O
wild	O
-	O
type	O
levels	O
of	O
iron	O
and	O
still	O
exhibited	O
77	O
%	O
of	O
wild	O
-	O
type	O
enzyme	O
activity	O
.	O
	O
Collectively	O
,	O
these	O
data	O
suggest	O
that	O
the	O
correct	O
positioning	O
of	O
the	O
carboxy	O
-	O
terminus	O
of	O
the	O
(	O
Sa	O
)	O
EctC	O
protein	O
is	O
of	O
structural	O
and	O
functional	O
importance	O
for	O
the	O
activity	O
of	O
the	O
ectoine	O
synthase	O
.	O
	O
Residues	O
Leu	O
-	O
87	O
and	O
Asp	O
-	O
91	O
are	O
highly	O
conserved	O
in	O
the	O
ectoine	O
synthase	O
protein	O
family	O
.	O
	O
The	O
replacement	O
of	O
Leu	O
-	O
87	O
by	O
Ala	O
led	O
to	O
a	O
substantial	O
drop	O
in	O
enzyme	O
activity	O
(	O
Table	O
1	O
).	O
	O
Conversely	O
,	O
the	O
replacement	O
of	O
Asp	O
-	O
91	O
by	O
Ala	O
and	O
Glu	O
,	O
resulted	O
in	O
(	O
Sa	O
)	O
EctC	O
protein	O
variants	O
with	O
80	O
%	O
and	O
98	O
%	O
enzyme	O
activity	O
,	O
respectively	O
(	O
Table	O
1	O
).	O
	O
We	O
currently	O
cannot	O
comment	O
on	O
possible	O
functional	O
role	O
Asp	O
-	O
91	O
.	O
	O
However	O
,	O
Leu	O
-	O
87	O
is	O
positioned	O
at	O
the	O
end	O
of	O
one	O
of	O
the	O
β	O
-	O
sheets	O
that	O
form	O
the	O
dimer	O
interface	O
(	O
Fig	O
5c	O
)	O
and	O
it	O
might	O
therefore	O
possess	O
a	O
structural	O
role	O
.	O
	O
It	O
is	O
also	O
located	O
near	O
Tyr	O
-	O
85	O
,	O
one	O
of	O
the	O
residues	O
that	O
probably	O
coordinate	O
the	O
iron	O
molecule	O
with	O
in	O
the	O
(	O
Sa	O
)	O
EctC	O
active	O
site	O
(	O
Fig	O
6a	O
)	O
and	O
therefore	O
might	O
exert	O
indirect	O
effects	O
.	O
	O
We	O
note	O
that	O
His	O
-	O
117	O
is	O
located	O
close	O
to	O
the	O
chemically	O
undefined	O
ligand	O
in	O
the	O
(	O
Sa	O
)	O
EctC	O
structure	O
(	O
Fig	O
7b	O
)	O
and	O
might	O
thus	O
play	O
a	O
role	O
in	O
contacting	O
the	O
natural	O
substrate	O
of	O
the	O
ectoine	O
synthase	O
.	O
	O
Both	O
(	O
Sa	O
)	O
EctC	O
protein	O
variants	O
exhibited	O
wild	O
-	O
type	O
level	O
enzyme	O
activities	O
and	O
possessed	O
a	O
iron	O
content	O
matching	O
that	O
of	O
the	O
wild	O
-	O
type	O
(	O
Table	O
1	O
).	O
	O
This	O
illustrates	O
that	O
not	O
every	O
amino	O
acid	O
substitution	O
in	O
the	O
(	O
Sa	O
)	O
EctC	O
protein	O
leads	O
to	O
an	O
indiscriminate	O
impairment	O
of	O
enzyme	O
function	O
and	O
iron	O
content	O
.	O
	O
The	O
crystallographic	O
data	O
presented	O
here	O
firmly	O
identify	O
ectoine	O
synthase	O
(	O
EctC	O
),	O
an	O
enzyme	O
critical	O
for	O
the	O
production	O
of	O
the	O
microbial	O
cytoprotectant	O
and	O
chemical	O
chaperone	O
ectoine	O
,	O
as	O
a	O
new	O
member	O
of	O
the	O
cupin	O
superfamily	O
.	O
	O
The	O
overall	O
fold	O
and	O
bowl	O
shape	O
of	O
the	O
(	O
Sa	O
)	O
EctC	O
protein	O
(	O
Figs	O
4	O
and	O
5	O
)	O
with	O
its	O
11	O
β	O
-	O
strands	O
(	O
β1	O
-	O
β11	O
)	O
and	O
two	O
α	O
-	O
helices	O
(	O
α	O
-	O
I	O
and	O
α	O
-	O
II	O
)	O
closely	O
adheres	O
to	O
the	O
design	O
principles	O
typically	O
found	O
in	O
crystal	O
structures	O
of	O
cupins	O
.	O
	O
In	O
addition	O
to	O
the	O
ectoine	O
synthase	O
,	O
the	O
polyketide	O
cyclase	O
RemF	O
is	O
the	O
only	O
other	O
currently	O
known	O
cupin	O
-	O
related	O
enzyme	O
that	O
catalyze	O
a	O
cyclocondensation	O
reaction	O
although	O
the	O
substrates	O
of	O
EctC	O
and	O
RemF	O
are	O
rather	O
different	O
.	O
	O
The	O
pro	O
-	O
and	O
eukaryotic	O
members	O
of	O
the	O
cupin	O
superfamily	O
perform	O
a	O
variety	O
of	O
both	O
enzymatic	O
and	O
non	O
-	O
enzymatic	O
functions	O
that	O
are	O
built	O
upon	O
a	O
common	O
structural	O
scaffold	O
.	O
	O
Most	O
cupins	O
contain	O
transition	O
state	O
metals	O
that	O
can	O
promote	O
different	O
types	O
of	O
chemical	O
reactions	O
.	O
	O
We	O
report	O
here	O
for	O
the	O
first	O
time	O
that	O
the	O
ectoine	O
synthase	O
is	O
a	O
metal	O
-	O
dependent	O
enzyme	O
.	O
	O
ICP	O
-	O
MS	O
,	O
metal	O
-	O
depletion	O
and	O
reconstitution	O
experiments	O
(	O
Fig	O
3	O
)	O
consistently	O
identify	O
iron	O
as	O
the	O
biologically	O
most	O
relevant	O
metal	O
for	O
the	O
EctC	O
-	O
catalyzed	O
cyclocondensation	O
reaction	O
.	O
	O
However	O
,	O
as	O
observed	O
with	O
other	O
cupins	O
,	O
EctC	O
is	O
a	O
somewhat	O
promiscuous	O
enzyme	O
as	O
far	O
as	O
the	O
catalytically	O
important	O
metal	O
is	O
concerned	O
when	O
they	O
are	O
provided	O
in	O
large	O
molar	O
excess	O
(	O
Fig	O
3c	O
).	O
	O
Although	O
some	O
uncertainty	O
remains	O
with	O
respect	O
to	O
the	O
precise	O
identity	O
of	O
amino	O
acid	O
residues	O
that	O
participate	O
in	O
metal	O
binding	O
by	O
(	O
Sa	O
)	O
EctC	O
,	O
our	O
structure	O
-	O
guided	O
site	O
-	O
directed	O
mutagenesis	O
experiments	O
targeting	O
the	O
presumptive	O
iron	O
-	O
binding	O
residues	O
(	O
Fig	O
6a	O
and	O
6b	O
)	O
demonstrate	O
that	O
none	O
of	O
them	O
can	O
be	O
spared	O
(	O
Table	O
1	O
).	O
	O
The	O
three	O
residues	O
(	O
Glu	O
-	O
57	O
,	O
Tyr	O
-	O
85	O
,	O
His	O
-	O
93	O
)	O
that	O
we	O
deem	O
to	O
form	O
it	O
(	O
Figs	O
6	O
and	O
7b	O
)	O
are	O
strictly	O
conserved	O
in	O
a	O
large	O
collection	O
of	O
EctC	O
-	O
type	O
proteins	O
originating	O
from	O
16	O
bacterial	O
and	O
three	O
archaeal	O
phyla	O
(	O
Fig	O
2	O
).	O
	O
We	O
also	O
show	O
here	O
for	O
the	O
first	O
time	O
that	O
,	O
in	O
addition	O
to	O
its	O
natural	O
substrate	O
N	O
-	O
γ	O
-	O
ADABA	O
,	O
EctC	O
also	O
converts	O
the	O
isomer	O
N	O
-	O
α	O
-	O
ADABA	O
into	O
ectoine	O
,	O
albeit	O
with	O
a	O
73	O
-	O
fold	O
reduced	O
catalytic	O
efficiency	O
(	O
S3a	O
and	O
S3b	O
Fig	O
).	O
	O
Our	O
finding	O
that	O
N	O
-	O
α	O
-	O
ADABA	O
serves	O
as	O
a	O
substrate	O
for	O
ectoine	O
synthase	O
has	O
physiologically	O
relevant	O
ramifications	O
for	O
those	O
microorganisms	O
that	O
can	O
both	O
synthesize	O
and	O
catabolize	O
ectoine	O
,	O
since	O
they	O
need	O
to	O
prevent	O
a	O
futile	O
cycle	O
of	O
synthesis	O
and	O
degradation	O
when	O
N	O
-	O
α	O
-	O
ADABA	O
is	O
produced	O
as	O
an	O
intermediate	O
in	O
the	O
catabolic	O
route	O
.	O
	O
Although	O
we	O
cannot	O
identify	O
the	O
true	O
chemical	O
nature	O
of	O
the	O
C6	O
compound	O
that	O
was	O
trapped	O
in	O
the	O
(	O
Sa	O
)	O
EctC	O
structure	O
nor	O
its	O
precise	O
origin	O
,	O
we	O
treated	O
this	O
compound	O
as	O
a	O
proxy	O
for	O
the	O
natural	O
substrate	O
of	O
ectoine	O
synthase	O
,	O
which	O
is	O
a	O
C6	O
compound	O
as	O
well	O
(	O
Fig	O
7a	O
).	O
	O
Indeed	O
,	O
site	O
-	O
directed	O
mutagenesis	O
of	O
those	O
five	O
residues	O
that	O
contact	O
the	O
unknown	O
C6	O
compound	O
(	O
Fig	O
7b	O
)	O
yielded	O
(	O
Sa	O
)	O
EctC	O
variants	O
with	O
strongly	O
impaired	O
enzyme	O
function	O
but	O
near	O
wild	O
-	O
type	O
levels	O
of	O
iron	O
(	O
Table	O
1	O
).	O
	O
We	O
therefore	O
surmise	O
that	O
our	O
crystallographic	O
data	O
and	O
the	O
site	O
-	O
directed	O
mutagenesis	O
study	O
reported	O
here	O
provide	O
a	O
structural	O
and	O
functional	O
view	O
into	O
the	O
architecture	O
of	O
the	O
EctC	O
active	O
site	O
(	O
Fig	O
7b	O
).	O
	O
The	O
ectoine	O
synthase	O
from	O
the	O
cold	O
-	O
adapted	O
marine	O
bacterium	O
S	O
.	O
alaskensis	O
can	O
be	O
considered	O
as	O
a	O
psychrophilic	O
enzyme	O
(	O
S3a	O
Fig	O
),	O
types	O
of	O
proteins	O
with	O
a	O
considerable	O
structural	O
flexibility	O
.	O
	O
It	O
is	O
hoped	O
that	O
these	O
can	O
be	O
further	O
employed	O
to	O
obtain	O
EctC	O
crystal	O
structures	O
with	O
either	O
the	O
substrate	O
or	O
the	O
reaction	O
product	O
.	O
	O
Together	O
with	O
our	O
finding	O
that	O
ectoine	O
synthase	O
is	O
metal	O
dependent	O
,	O
these	O
crystal	O
structures	O
should	O
allow	O
a	O
more	O
detailed	O
understanding	O
of	O
the	O
chemistry	O
underlying	O
the	O
EctC	O
-	O
catalyzed	O
cyclocondensation	O
reaction	O
.	O
	O
Regnase	O
-	O
1	O
is	O
an	O
RNase	O
that	O
directly	O
cleaves	O
mRNAs	O
of	O
inflammatory	O
genes	O
such	O
as	O
IL	O
-	O
6	O
and	O
IL	O
-	O
12p40	O
,	O
and	O
negatively	O
regulates	O
cellular	O
inflammatory	O
responses	O
.	O
	O
Here	O
,	O
we	O
report	O
the	O
structures	O
of	O
four	O
domains	O
of	O
Regnase	O
-	O
1	O
from	O
Mus	O
musculus	O
—	O
the	O
N	O
-	O
terminal	O
domain	O
(	O
NTD	O
),	O
PilT	O
N	O
-	O
terminus	O
like	O
(	O
PIN	O
)	O
domain	O
,	O
zinc	O
finger	O
(	O
ZF	O
)	O
domain	O
and	O
C	O
-	O
terminal	O
domain	O
(	O
CTD	O
).	O
	O
The	O
PIN	O
domain	O
harbors	O
the	O
RNase	O
catalytic	O
center	O
;	O
however	O
,	O
it	O
is	O
insufficient	O
for	O
enzymatic	O
activity	O
.	O
	O
We	O
found	O
that	O
the	O
NTD	O
associates	O
with	O
the	O
PIN	O
domain	O
and	O
significantly	O
enhances	O
its	O
RNase	O
activity	O
.	O
	O
The	O
PIN	O
domain	O
forms	O
a	O
head	O
-	O
to	O
-	O
tail	O
oligomer	O
and	O
the	O
dimer	O
interface	O
overlaps	O
with	O
the	O
NTD	O
binding	O
site	O
.	O
	O
These	O
results	O
suggest	O
that	O
Regnase	O
-	O
1	O
RNase	O
activity	O
is	O
tightly	O
controlled	O
by	O
both	O
intramolecular	O
(	O
NTD	O
-	O
PIN	O
)	O
and	O
intermolecular	O
(	O
PIN	O
-	O
PIN	O
)	O
interactions	O
.	O
	O
The	O
initial	O
sensing	O
of	O
infection	O
is	O
mediated	O
by	O
a	O
set	O
of	O
pattern	O
-	O
recognition	O
receptors	O
(	O
PRRs	O
)	O
such	O
Toll	O
-	O
like	O
receptors	O
(	O
TLRs	O
)	O
and	O
the	O
intracellular	O
signaling	O
cascades	O
triggered	O
by	O
TLRs	O
evoke	O
transcriptional	O
expression	O
of	O
inflammatory	O
mediators	O
that	O
coordinate	O
the	O
elimination	O
of	O
pathogens	O
and	O
infected	O
cells	O
.	O
	O
Regnase	O
-	O
1	O
(	O
also	O
known	O
as	O
Zc3h12a	O
and	O
MCPIP1	O
)	O
is	O
an	O
RNase	O
whose	O
expression	O
level	O
is	O
stimulated	O
by	O
lipopolysaccharides	O
and	O
prevents	O
autoimmune	O
diseases	O
by	O
directly	O
controlling	O
the	O
stability	O
of	O
mRNAs	O
of	O
inflammatory	O
genes	O
such	O
as	O
interleukin	O
(	O
IL	O
)-	O
6	O
,	O
IL	O
-	O
1β	O
,	O
IL	O
-	O
2	O
,	O
and	O
IL	O
-	O
12p40	O
.	O
	O
Regnase	O
-	O
1	O
accelerates	O
target	O
mRNA	O
degradation	O
via	O
their	O
3	O
′-	O
terminal	O
untranslated	O
region	O
(	O
3	O
′	O
UTR	O
),	O
and	O
also	O
degrades	O
its	O
own	O
mRNA	O
.	O
	O
Recently	O
,	O
the	O
crystal	O
structure	O
of	O
the	O
Regnase	O
-	O
1	O
PIN	O
domain	O
derived	O
from	O
Homo	O
sapiens	O
was	O
reported	O
.	O
	O
The	O
structure	O
combined	O
with	O
functional	O
analyses	O
revealed	O
that	O
four	O
catalytically	O
important	O
Asp	O
residues	O
form	O
the	O
catalytic	O
center	O
and	O
stabilize	O
Mg2	O
+	O
binding	O
that	O
is	O
crucial	O
for	O
RNase	O
activity	O
.	O
	O
Several	O
CCCH	O
-	O
type	O
ZF	O
motifs	O
in	O
RNA	O
-	O
binding	O
proteins	O
have	O
been	O
reported	O
to	O
directly	O
bind	O
RNA	O
.	O
	O
In	O
addition	O
,	O
Regnase	O
-	O
1	O
has	O
been	O
predicted	O
to	O
possess	O
other	O
domains	O
in	O
the	O
N	O
-	O
and	O
C	O
-	O
terminal	O
regions	O
.	O
	O
However	O
,	O
the	O
structure	O
and	O
function	O
of	O
the	O
ZF	O
domain	O
,	O
N	O
-	O
terminal	O
domain	O
(	O
NTD	O
)	O
and	O
C	O
-	O
terminal	O
domain	O
(	O
CTD	O
)	O
of	O
Regnase	O
-	O
1	O
have	O
not	O
been	O
solved	O
.	O
	O
Here	O
,	O
we	O
performed	O
structural	O
and	O
functional	O
analyses	O
of	O
individual	O
domains	O
of	O
Regnase	O
-	O
1	O
derived	O
from	O
Mus	O
musculus	O
in	O
order	O
to	O
understand	O
the	O
catalytic	O
activity	O
in	O
vitro	O
.	O
	O
Our	O
data	O
revealed	O
that	O
the	O
catalytic	O
activity	O
of	O
Regnase	O
-	O
1	O
is	O
regulated	O
through	O
both	O
intra	O
and	O
intermolecular	O
domain	O
interactions	O
in	O
vitro	O
.	O
	O
The	O
NTD	O
plays	O
a	O
crucial	O
role	O
in	O
efficient	O
cleavage	O
of	O
target	O
mRNA	O
,	O
through	O
intramolecular	O
NTD	O
-	O
PIN	O
interactions	O
.	O
	O
Our	O
findings	O
suggest	O
that	O
Regnase	O
-	O
1	O
cleaves	O
its	O
target	O
mRNA	O
by	O
an	O
NTD	O
-	O
activated	O
functional	O
PIN	O
dimer	O
,	O
while	O
the	O
ZF	O
increases	O
RNA	O
affinity	O
in	O
the	O
vicinity	O
of	O
the	O
PIN	O
dimer	O
.	O
	O
We	O
analyzed	O
Rengase	O
-	O
1	O
derived	O
from	O
Mus	O
musculus	O
and	O
solved	O
the	O
structures	O
of	O
the	O
four	O
domains	O
;	O
NTD	O
,	O
PIN	O
,	O
ZF	O
,	O
and	O
CTD	O
individually	O
by	O
X	O
-	O
ray	O
crystallography	O
or	O
NMR	O
(	O
Fig	O
.	O
1a	O
–	O
e	O
).	O
	O
X	O
-	O
ray	O
crystallography	O
was	O
attempted	O
for	O
the	O
fragment	O
containing	O
both	O
the	O
PIN	O
and	O
ZF	O
domains	O
,	O
however	O
,	O
electron	O
density	O
was	O
observed	O
only	O
for	O
the	O
PIN	O
domain	O
(	O
Fig	O
.	O
1c	O
),	O
consistent	O
with	O
a	O
previous	O
report	O
on	O
Regnase	O
-	O
1	O
derived	O
from	O
Homo	O
sapiens	O
.	O
	O
This	O
suggests	O
that	O
the	O
PIN	O
and	O
ZF	O
domains	O
exist	O
independently	O
without	O
interacting	O
with	O
each	O
other	O
.	O
	O
The	O
NTD	O
and	O
CTD	O
are	O
both	O
composed	O
of	O
three	O
α	O
helices	O
,	O
and	O
structurally	O
resemble	O
ubiquitin	O
conjugating	O
enzyme	O
E2	O
K	O
(	O
PDB	O
ID	O
:	O
3K9O	O
)	O
and	O
ubiquitin	O
associated	O
protein	O
1	O
(	O
PDB	O
ID	O
:	O
4AE4	O
),	O
respectively	O
,	O
according	O
to	O
the	O
Dali	O
server	O
.	O
	O
Contribution	O
of	O
each	O
domain	O
of	O
Regnase	O
-	O
1	O
to	O
the	O
mRNA	O
binding	O
activity	O
	O
First	O
,	O
we	O
evaluated	O
a	O
role	O
of	O
the	O
NTD	O
and	O
ZF	O
domains	O
for	O
mRNA	O
binding	O
by	O
an	O
in	O
vitro	O
gel	O
shift	O
assay	O
(	O
Fig	O
.	O
1f	O
).	O
	O
Fluorescently	O
5	O
′-	O
labeled	O
RNA	O
corresponding	O
to	O
nucleotides	O
82	O
–	O
106	O
of	O
the	O
IL	O
-	O
6	O
mRNA	O
3	O
′	O
UTR	O
and	O
the	O
catalytically	O
inactive	O
mutant	O
(	O
D226N	B-mutant
and	O
D244N	B-mutant
)	O
of	O
Regnase	O
-	O
1	O
—	O
hereafter	O
referred	O
to	O
as	O
the	O
DDNN	B-mutant
mutant	O
—	O
were	O
utilized	O
.	O
	O
Upon	O
addition	O
of	O
a	O
larger	O
amount	O
of	O
Regnase	O
-	O
1	O
,	O
the	O
fluorescence	O
of	O
free	O
RNA	O
decreased	O
,	O
indicating	O
that	O
Regnase	O
-	O
1	O
bound	O
to	O
the	O
RNA	O
.	O
	O
While	O
the	O
RNA	O
binding	O
ability	O
was	O
not	O
significantly	O
changed	O
in	O
the	O
presence	O
of	O
NTD	O
,	O
it	O
increased	O
in	O
the	O
presence	O
of	O
the	O
ZF	O
domain	O
(	O
Fig	O
.	O
1f	O
,	O
g	O
and	O
Supplementary	O
Fig	O
.	O
1	O
).	O
	O
Direct	O
binding	O
of	O
the	O
ZF	O
domain	O
and	O
RNA	O
were	O
confirmed	O
by	O
NMR	O
spectral	O
changes	O
.	O
	O
The	O
fitting	O
of	O
the	O
titration	O
curve	O
of	O
Y314	O
resulted	O
in	O
an	O
apparent	O
dissociation	O
constant	O
(	O
Kd	O
)	O
of	O
10	O
±	O
1	O
.	O
1	O
μM	O
(	O
Supplementary	O
Fig	O
.	O
2	O
).	O
	O
In	O
order	O
to	O
characterize	O
the	O
role	O
of	O
each	O
domain	O
in	O
the	O
RNase	O
activity	O
of	O
Regnase	O
-	O
1	O
,	O
we	O
performed	O
an	O
in	O
vitro	O
cleavage	O
assay	O
using	O
fluorescently	O
5	O
′-	O
labeled	O
RNA	O
corresponding	O
to	O
nucleotides	O
82	O
–	O
106	O
of	O
the	O
IL	O
-	O
6	O
mRNA	O
3	O
′	O
UTR	O
(	O
Fig	O
.	O
1g	O
).	O
	O
Regnase	O
-	O
1	O
constructs	O
consisting	O
of	O
NTD	B-mutant
-	I-mutant
PIN	I-mutant
-	I-mutant
ZF	I-mutant
completely	O
cleaved	O
the	O
target	O
mRNA	O
and	O
generated	O
the	O
cleaved	O
products	O
.	O
	O
The	O
apparent	O
half	O
-	O
life	O
(	O
T1	O
/	O
2	O
)	O
of	O
the	O
RNase	O
activity	O
was	O
about	O
20	O
minutes	O
.	O
	O
Taken	O
together	O
with	O
the	O
results	O
in	O
the	O
previous	O
section	O
,	O
we	O
conclude	O
that	O
the	O
NTD	O
is	O
crucial	O
for	O
the	O
RNase	O
activity	O
of	O
Regnase	O
-	O
1	O
in	O
vitro	O
,	O
although	O
it	O
does	O
not	O
contribute	O
to	O
the	O
direct	O
mRNA	O
binding	O
.	O
	O
During	O
purification	O
by	O
gel	O
filtration	O
,	O
the	O
PIN	O
domain	O
exhibited	O
extremely	O
asymmetric	O
elution	O
peaks	O
in	O
a	O
concentration	O
dependent	O
manner	O
(	O
Fig	O
.	O
2a	O
).	O
	O
We	O
found	O
that	O
the	O
PIN	O
domain	O
formed	O
a	O
head	O
-	O
to	O
-	O
tail	O
oligomer	O
that	O
was	O
commonly	O
observed	O
in	O
all	O
three	O
crystal	O
forms	O
in	O
spite	O
of	O
the	O
different	O
crystallization	O
conditions	O
(	O
Supplementary	O
Fig	O
.	O
3	O
).	O
	O
On	O
the	O
other	O
hand	O
,	O
single	O
mutations	O
of	O
side	O
chains	O
involved	O
in	O
the	O
PIN	O
–	O
PIN	O
oligomeric	O
interaction	O
resulted	O
in	O
monomer	O
formation	O
,	O
judging	O
from	O
gel	O
filtration	O
(	O
Fig	O
.	O
2a	O
,	O
b	O
).	O
	O
Wild	O
type	O
and	O
monomeric	O
PIN	O
mutants	O
(	O
P212A	B-mutant
and	O
D278R	B-mutant
)	O
were	O
also	O
analyzed	O
by	O
NMR	O
.	O
	O
The	O
spectra	O
indicate	O
that	O
the	O
dimer	O
interface	O
of	O
the	O
wild	O
type	O
PIN	O
domain	O
were	O
significantly	O
broadened	O
compared	O
to	O
the	O
monomeric	O
mutants	O
(	O
Supplementary	O
Fig	O
.	O
4	O
).	O
	O
These	O
results	O
indicate	O
that	O
the	O
PIN	O
domain	O
forms	O
a	O
head	O
-	O
to	O
-	O
tail	O
oligomer	O
in	O
solution	O
similar	O
to	O
the	O
crystal	O
structure	O
.	O
	O
Interestingly	O
,	O
the	O
monomeric	O
PIN	O
mutants	O
P212A	B-mutant
,	O
R214A	B-mutant
,	O
and	O
D278R	B-mutant
had	O
no	O
significant	O
RNase	O
activity	O
for	O
IL	O
-	O
6	O
mRNA	O
in	O
vitro	O
(	O
Fig	O
.	O
2c	O
).	O
	O
The	O
side	O
chains	O
of	O
these	O
residues	O
point	O
away	O
from	O
the	O
catalytic	O
center	O
on	O
the	O
same	O
molecule	O
(	O
Fig	O
.	O
2b	O
).	O
	O
While	O
the	O
NTD	O
does	O
not	O
contribute	O
to	O
RNA	O
binding	O
(	O
Fig	O
.	O
1f	O
,	O
g	O
,	O
and	O
Supplementary	O
Fig	O
.	O
1	O
),	O
it	O
increases	O
the	O
RNase	O
activity	O
of	O
Regnase	O
-	O
1	O
(	O
Fig	O
.	O
1h	O
).	O
	O
In	O
order	O
to	O
gain	O
insight	O
into	O
the	O
molecular	O
mechanism	O
of	O
the	O
NTD	O
-	O
mediated	O
enhancement	O
of	O
Regnase	O
-	O
1	O
RNase	O
activity	O
,	O
we	O
further	O
investigated	O
the	O
domain	O
-	O
domain	O
interaction	O
between	O
the	O
NTD	O
and	O
the	O
PIN	O
domain	O
using	O
NMR	O
.	O
	O
We	O
used	O
the	O
catalytically	O
inactive	O
monomeric	O
PIN	O
mutant	O
possessing	O
both	O
the	O
DDNN	B-mutant
and	O
D278R	B-mutant
mutations	O
to	O
avoid	O
dimer	O
formation	O
of	O
the	O
PIN	O
domain	O
.	O
	O
The	O
NMR	O
signals	O
from	O
the	O
PIN	O
domain	O
(	O
residues	O
V177	O
,	O
F210	O
-	O
T211	O
,	O
R214	O
,	O
F228	O
-	O
L232	O
,	O
and	O
F234	O
-	O
S236	O
)	O
exhibited	O
significant	O
chemical	O
shift	O
changes	O
upon	O
addition	O
of	O
the	O
NTD	O
(	O
Fig	O
.	O
3a	O
).	O
	O
These	O
results	O
clearly	O
indicate	O
a	O
direct	O
interaction	O
between	O
the	O
PIN	O
domain	O
and	O
the	O
NTD	O
.	O
	O
Based	O
on	O
the	O
titration	O
curve	O
for	O
the	O
chemical	O
shift	O
changes	O
of	O
L58	O
,	O
the	O
apparent	O
Kd	O
between	O
the	O
isolated	O
NTD	O
and	O
PIN	O
was	O
estimated	O
to	O
be	O
110	O
±	O
5	O
.	O
8	O
μM	O
.	O
Considering	O
the	O
fact	O
that	O
the	O
NTD	O
and	O
PIN	O
domains	O
are	O
attached	O
by	O
a	O
linker	O
,	O
the	O
actual	O
binding	O
affinity	O
is	O
expected	O
much	O
higher	O
in	O
the	O
native	O
protein	O
.	O
	O
Mapping	O
the	O
residues	O
with	O
chemical	O
shift	O
changes	O
reveals	O
the	O
putative	O
PIN	O
/	O
NTD	O
interface	O
,	O
which	O
includes	O
a	O
helix	O
that	O
harbors	O
catalytic	O
residues	O
D225	O
and	O
D226	O
on	O
the	O
PIN	O
domain	O
(	O
Fig	O
.	O
3a	O
).	O
	O
Interestingly	O
,	O
the	O
putative	O
binding	O
site	O
for	O
the	O
NTD	O
overlaps	O
with	O
the	O
PIN	O
-	O
PIN	O
dimer	O
interface	O
,	O
implying	O
that	O
NTD	O
binding	O
can	O
“	O
terminate	O
”	O
PIN	O
-	O
PIN	O
oligomerization	O
(	O
Fig	O
.	O
2b	O
).	O
	O
An	O
in	O
silico	O
docking	O
of	O
the	O
NTD	O
and	O
PIN	O
domains	O
using	O
chemical	O
shift	O
restraints	O
provided	O
a	O
model	O
consistent	O
with	O
the	O
NMR	O
experiments	O
(	O
Fig	O
.	O
3c	O
).	O
	O
To	O
gain	O
insight	O
into	O
the	O
residues	O
critical	O
for	O
Regnase	O
-	O
1	O
RNase	O
activity	O
,	O
each	O
basic	O
or	O
aromatic	O
residue	O
located	O
around	O
the	O
catalytic	O
site	O
of	O
the	O
PIN	O
oligomer	O
was	O
mutated	O
to	O
alanine	O
,	O
and	O
the	O
oligomerization	O
and	O
RNase	O
activity	O
were	O
investigated	O
(	O
Fig	O
.	O
4	O
).	O
	O
From	O
the	O
gel	O
filtration	O
assays	O
,	O
all	O
mutants	O
except	O
R214A	B-mutant
formed	O
dimers	O
,	O
suggesting	O
that	O
any	O
lack	O
of	O
RNase	O
activity	O
in	O
the	O
mutants	O
,	O
except	O
R214A	B-mutant
,	O
was	O
directly	O
due	O
to	O
mutational	O
effects	O
of	O
the	O
specific	O
residues	O
and	O
not	O
to	O
abrogation	O
of	O
dimer	O
formation	O
.	O
	O
The	O
W182A	B-mutant
,	O
R183A	B-mutant
,	O
and	O
R214A	B-mutant
mutants	O
markedly	O
lost	O
cleavage	O
activity	O
for	O
IL	O
-	O
6	O
mRNA	O
as	O
well	O
as	O
for	O
Regnase	O
-	O
1	O
mRNA	O
.	O
	O
The	O
K184A	B-mutant
,	O
R215A	B-mutant
,	O
and	O
R220A	B-mutant
mutants	O
moderately	O
but	O
significantly	O
decreased	O
the	O
cleavage	O
activity	O
for	O
both	O
target	O
mRNAs	O
.	O
	O
The	O
importance	O
of	O
K219	O
and	O
R247	O
was	O
slightly	O
different	O
for	O
IL	O
-	O
6	O
and	O
Regnase	O
-	O
1	O
mRNA	O
;	O
both	O
K219	O
and	O
R247	O
were	O
more	O
important	O
in	O
the	O
cleavage	O
of	O
IL	O
-	O
6	O
mRNA	O
than	O
for	O
Regnase	O
-	O
1	O
mRNA	O
.	O
	O
In	O
contrast	O
,	O
R214	O
is	O
important	O
for	O
oligomerization	O
of	O
the	O
PIN	O
domain	O
and	O
the	O
“	O
secondary	O
”	O
chain	O
’	O
s	O
residue	O
R214	O
is	O
also	O
positioned	O
near	O
the	O
“	O
primary	O
”	O
chain	O
’	O
s	O
active	O
site	O
within	O
the	O
dimer	O
interface	O
.	O
	O
If	O
this	O
model	O
is	O
correct	O
,	O
then	O
we	O
reasoned	O
that	O
a	O
catalytically	O
inactive	O
PIN	O
and	O
a	O
PIN	O
lacking	O
the	O
putative	O
RNA	O
-	O
binding	O
residues	O
ought	O
to	O
be	O
inactive	O
in	O
isolation	O
but	O
become	O
active	O
when	O
mixed	O
together	O
.	O
	O
In	O
order	O
to	O
test	O
this	O
hypothesis	O
,	O
we	O
performed	O
in	O
vitro	O
cleavage	O
assays	O
using	O
combinations	O
of	O
Regnase	O
-	O
1	O
mutants	O
that	O
had	O
no	O
or	O
decreased	O
RNase	O
activities	O
by	O
themselves	O
(	O
Fig	O
.	O
5	O
).	O
	O
These	O
were	O
paired	O
with	O
a	O
DDNN	B-mutant
mutant	O
that	O
had	O
no	O
RNase	O
activity	O
by	O
itself	O
.	O
	O
When	O
any	O
members	O
of	O
the	O
two	O
groups	O
are	O
mixed	O
,	O
two	O
kinds	O
of	O
heterodimers	O
can	O
be	O
formed	O
:	O
one	O
is	O
composed	O
of	O
a	O
DDNN	B-mutant
primary	O
PIN	O
and	O
a	O
basic	O
residue	O
mutant	O
secondary	O
PIN	O
and	O
is	O
expected	O
to	O
exhibit	O
no	O
RNase	O
activity	O
;	O
the	O
other	O
is	O
composed	O
of	O
a	O
basic	O
residue	O
mutant	O
primary	O
PIN	O
and	O
a	O
DDNN	B-mutant
secondary	O
PIN	O
and	O
is	O
predicted	O
to	O
rescue	O
RNase	O
activity	O
(	O
Fig	O
.	O
5a	O
).	O
	O
When	O
we	O
compared	O
the	O
fluorescence	O
intensity	O
of	O
uncleaved	O
IL	O
-	O
6	O
mRNA	O
,	O
basic	O
residue	O
mutants	O
W182A	B-mutant
,	O
K184A	B-mutant
,	O
R214A	B-mutant
,	O
and	O
R220A	B-mutant
were	O
rescued	O
upon	O
addition	O
of	O
the	O
DDNN	B-mutant
mutant	O
(	O
Fig	O
.	O
5b	O
).	O
	O
Consistently	O
,	O
when	O
we	O
compared	O
the	O
fluorescence	O
intensity	O
of	O
the	O
uncleaved	O
Regnase	O
-	O
1	O
mRNA	O
,	O
basic	O
residue	O
mutants	O
K184A	B-mutant
and	O
R214A	B-mutant
were	O
rescued	O
upon	O
addition	O
of	O
the	O
DDNN	B-mutant
mutant	O
(	O
Fig	O
.	O
5c	O
).	O
	O
Rescue	O
of	O
K184A	B-mutant
and	O
R214A	B-mutant
by	O
the	O
DDNN	B-mutant
mutant	O
was	O
also	O
confirmed	O
by	O
a	O
significant	O
increase	O
in	O
the	O
cleaved	O
products	O
.	O
	O
This	O
is	O
particularly	O
significant	O
because	O
the	O
side	O
chains	O
of	O
K184	O
and	O
R214	O
in	O
the	O
primary	O
PIN	O
are	O
oriented	O
away	O
from	O
their	O
own	O
catalytic	O
center	O
,	O
while	O
those	O
in	O
the	O
secondary	O
PIN	O
face	O
toward	O
the	O
catalytic	O
center	O
of	O
the	O
primary	O
PIN	O
.	O
	O
R214	O
is	O
an	O
important	O
residue	O
for	O
dimer	O
formation	O
as	O
shown	O
in	O
Fig	O
.	O
2	O
,	O
therefore	O
,	O
R214A	B-mutant
in	O
the	O
secondary	O
PIN	O
cannot	O
dimerize	O
.	O
	O
Taken	O
together	O
,	O
the	O
rescue	O
experiments	O
above	O
support	O
the	O
proposed	O
model	O
in	O
which	O
the	O
head	O
-	O
to	O
-	O
tail	O
dimer	O
is	O
functional	O
in	O
vitro	O
.	O
	O
Although	O
the	O
function	O
of	O
the	O
CTD	O
remains	O
elusive	O
,	O
we	O
revealed	O
the	O
functions	O
of	O
the	O
NTD	O
,	O
PIN	O
,	O
and	O
ZF	O
domains	O
.	O
	O
A	O
Regnase	O
-	O
1	O
construct	O
consisting	O
of	O
PIN	O
and	O
ZF	O
domains	O
derived	O
from	O
Mus	O
musculus	O
was	O
crystallized	O
;	O
however	O
,	O
the	O
electron	O
density	O
of	O
the	O
ZF	O
domain	O
was	O
low	O
,	O
indicating	O
that	O
the	O
ZF	O
domain	O
is	O
highly	O
mobile	O
in	O
the	O
absence	O
of	O
target	O
mRNA	O
or	O
possibly	O
other	O
protein	O
-	O
protein	O
interactions	O
.	O
	O
Our	O
NMR	O
experiments	O
confirmed	O
direct	O
binding	O
of	O
the	O
ZF	O
domain	O
to	O
IL	O
-	O
6	O
mRNA	O
with	O
a	O
Kd	O
of	O
10	O
±	O
1	O
.	O
1	O
μM	O
.	O
Furthermore	O
,	O
an	O
in	O
vitro	O
gel	O
shift	O
assay	O
indicated	O
that	O
Regnase	O
-	O
1	O
containing	O
the	O
ZF	O
domain	O
enhanced	O
target	O
mRNA	O
-	O
binding	O
,	O
but	O
the	O
protein	O
-	O
RNA	O
complex	O
remained	O
in	O
the	O
bottom	O
of	O
the	O
well	O
without	O
entering	O
into	O
the	O
polyacrylamide	O
gel	O
.	O
	O
These	O
results	O
indicate	O
that	O
Regnase	O
-	O
1	O
directly	O
binds	O
to	O
RNA	O
and	O
precipitates	O
under	O
such	O
experimental	O
conditions	O
.	O
	O
Due	O
to	O
this	O
limitation	O
,	O
it	O
is	O
difficult	O
to	O
perform	O
further	O
structural	O
analyses	O
of	O
mRNA	O
-	O
Regnase	O
-	O
1	O
complexes	O
by	O
X	O
-	O
ray	O
crystallography	O
or	O
NMR	O
.	O
	O
Rao	O
and	O
co	O
-	O
workers	O
previously	O
argued	O
that	O
PIN	O
dimerization	O
is	O
likely	O
to	O
be	O
a	O
crystallographic	O
artifact	O
with	O
no	O
physiological	O
significance	O
,	O
since	O
monomers	O
were	O
dominant	O
in	O
their	O
analytical	O
ultra	O
-	O
centrifugation	O
experiments	O
.	O
	O
This	O
inconsistency	O
might	O
be	O
due	O
to	O
difference	O
in	O
the	O
analytical	O
methods	O
and	O
/	O
or	O
protein	O
concentrations	O
used	O
in	O
each	O
experiment	O
,	O
since	O
the	O
oligomer	O
formation	O
of	O
PIN	O
was	O
dependent	O
on	O
the	O
protein	O
concentration	O
in	O
our	O
study	O
.	O
	O
Single	O
mutations	O
to	O
residues	O
involved	O
in	O
the	O
putative	O
oligomeric	O
interaction	O
of	O
PIN	O
monomerized	O
as	O
expected	O
and	O
these	O
mutants	O
lost	O
their	O
RNase	O
activity	O
as	O
well	O
.	O
	O
Based	O
on	O
these	O
observations	O
,	O
we	O
concluded	O
that	O
PIN	O
-	O
PIN	O
dimer	O
formation	O
is	O
critical	O
for	O
Regnase	O
-	O
1	O
RNase	O
activity	O
in	O
vitro	O
.	O
	O
Within	O
the	O
crystal	O
structure	O
of	O
the	O
PIN	O
dimer	O
,	O
the	O
Regnase	O
-	O
1	O
specific	O
basic	O
regions	O
in	O
both	O
the	O
“	O
primary	O
”	O
and	O
“	O
secondary	O
”	O
PINs	O
are	O
located	O
around	O
the	O
catalytic	O
site	O
of	O
the	O
primary	O
PIN	O
(	O
Supplementary	O
Fig	O
.	O
6	O
).	O
	O
Moreover	O
,	O
our	O
structure	O
-	O
based	O
mutational	O
analyses	O
showed	O
these	O
two	O
Regnase	O
-	O
1	O
specific	O
basic	O
regions	O
were	O
essential	O
for	O
target	O
mRNA	O
cleavage	O
in	O
vitro	O
.	O
	O
The	O
cleavage	O
assay	O
also	O
showed	O
that	O
the	O
NTD	O
is	O
crucial	O
for	O
efficient	O
mRNA	O
cleavage	O
.	O
	O
While	O
further	O
analyses	O
are	O
necessary	O
to	O
prove	O
this	O
point	O
,	O
our	O
preliminary	O
docking	O
and	O
molecular	O
dynamics	O
simulations	O
indicate	O
that	O
NTD	O
-	O
binding	O
rearranges	O
the	O
catalytic	O
residues	O
of	O
the	O
PIN	O
domain	O
toward	O
an	O
active	O
conformation	O
suitable	O
for	O
binding	O
Mg2	O
+.	O
	O
In	O
this	O
context	O
,	O
it	O
is	O
interesting	O
that	O
,	O
in	O
response	O
to	O
TCR	O
stimulation	O
,	O
Malt1	O
cleaves	O
Regnase	O
-	O
1	O
at	O
R111	O
to	O
control	O
immune	O
responses	O
in	O
vivo	O
.	O
	O
This	O
result	O
is	O
consistent	O
with	O
a	O
model	O
in	O
which	O
the	O
NTD	O
acts	O
as	O
an	O
enhancer	O
,	O
and	O
cleavage	O
of	O
the	O
linker	O
lowers	O
enzymatic	O
activity	O
dramatically	O
.	O
	O
We	O
incorporated	O
information	O
from	O
the	O
cleavage	O
site	O
of	O
IL	O
-	O
6	O
mRNA	O
in	O
vitro	O
is	O
indicated	O
by	O
denaturing	O
polyacrylamide	O
gel	O
electrophoresis	O
(	O
Supplementary	O
Fig	O
.	O
7a	O
,	O
b	O
).	O
	O
The	O
docking	O
result	O
revealed	O
multiple	O
RNA	O
binding	O
modes	O
that	O
satisfied	O
the	O
experimental	O
results	O
in	O
vitro	O
(	O
Supplementary	O
Fig	O
.	O
7c	O
,	O
d	O
),	O
however	O
,	O
it	O
should	O
be	O
noted	O
that	O
,	O
in	O
vivo	O
,	O
there	O
would	O
likely	O
be	O
many	O
other	O
RNA	O
-	O
binding	O
proteins	O
that	O
would	O
protect	O
loop	O
regions	O
from	O
cleavage	O
by	O
Regnase	O
-	O
1	O
.	O
	O
In	O
the	O
absence	O
of	O
target	O
mRNA	O
,	O
the	O
PIN	O
domain	O
forms	O
head	O
-	O
to	O
-	O
tail	O
oligomers	O
at	O
high	O
concentration	O
.	O
	O
A	O
fully	O
active	O
catalytic	O
center	O
can	O
be	O
formed	O
only	O
when	O
the	O
NTD	O
associates	O
with	O
the	O
oligomer	O
surface	O
of	O
the	O
PIN	O
domain	O
,	O
which	O
terminates	O
the	O
head	O
-	O
to	O
-	O
tail	O
oligomer	O
formation	O
in	O
one	O
direction	O
(	O
primary	O
PIN	O
),	O
and	O
forms	O
a	O
functional	O
dimer	O
together	O
with	O
the	O
neighboring	O
PIN	O
(	O
secondary	O
PIN	O
).	O
	O
(	O
a	O
)	O
Domain	O
architecture	O
of	O
Regnase	O
-	O
1	O
.	O
(	O
b	O
)	O
Solution	O
structure	O
of	O
the	O
NTD	O
.	O
(	O
c	O
)	O
Crystal	O
structure	O
of	O
the	O
PIN	O
domain	O
.	O
	O
Catalytic	O
Asp	O
residues	O
were	O
shown	O
in	O
sticks	O
.	O
	O
(	O
d	O
)	O
Solution	O
structure	O
of	O
the	O
ZF	O
domain	O
.	O
	O
Three	O
Cys	O
residues	O
and	O
one	O
His	O
residue	O
responsible	O
for	O
Zn2	O
+-	O
binding	O
were	O
shown	O
in	O
sticks	O
.	O
	O
(	O
e	O
)	O
Solution	O
structure	O
of	O
the	O
CTD	O
.	O
	O
All	O
the	O
structures	O
were	O
colored	O
in	O
rainbow	O
from	O
N	O
-	O
terminus	O
(	O
blue	O
)	O
to	O
C	O
-	O
terminus	O
(	O
red	O
).	O
	O
(	O
f	O
)	O
In	O
vitro	O
gel	O
shift	O
binding	O
assay	O
between	O
Regnase	O
-	O
1	O
and	O
IL	O
-	O
6	O
mRNA	O
.	O
	O
(	O
g	O
)	O
Binding	O
of	O
Regnase	O
-	O
1	O
and	O
IL	O
-	O
6	O
mRNA	O
was	O
plotted	O
.	O
	O
(	O
h	O
)	O
In	O
vitro	O
cleavage	O
assay	O
of	O
Regnase	O
-	O
1	O
to	O
IL	O
-	O
6	O
mRNA	O
.	O
	O
Fluorescence	O
intensity	O
of	O
the	O
uncleaved	O
IL	O
-	O
6	O
mRNA	O
was	O
indicated	O
as	O
the	O
percentage	O
against	O
that	O
in	O
the	O
absence	O
of	O
Regnase	O
-	O
1	O
.	O
	O
Head	O
-	O
to	O
-	O
tail	O
oligomer	O
formation	O
of	O
the	O
PIN	O
domain	O
is	O
crucial	O
for	O
the	O
RNase	O
activity	O
of	O
Regnase	O
-	O
1	O
.	O
	O
(	O
a	O
)	O
Gel	O
filtration	O
analyses	O
of	O
the	O
PIN	O
domain	O
.	O
	O
Catalytic	O
residues	O
and	O
mutated	O
residues	O
were	O
shown	O
in	O
sticks	O
.	O
	O
Residues	O
important	O
for	O
the	O
oligomeric	O
interaction	O
were	O
colored	O
red	O
,	O
while	O
R215	O
that	O
was	O
dispensable	O
for	O
the	O
oligomeric	O
interaction	O
was	O
colored	O
blue	O
.	O
(	O
c	O
)	O
RNase	O
activity	O
of	O
monomeric	O
mutants	O
for	O
IL	O
-	O
6	O
mRNA	O
was	O
analyzed	O
.	O
	O
Pro	O
and	O
the	O
residues	O
without	O
analysis	O
were	O
colored	O
black	O
and	O
gray	O
,	O
respectively	O
.	O
	O
(	O
b	O
)	O
NMR	O
analyses	O
of	O
the	O
PIN	O
-	O
binding	O
to	O
the	O
NTD	O
.	O
	O
The	O
residues	O
with	O
significant	O
chemical	O
shift	O
changes	O
were	O
labeled	O
in	O
the	O
overlaid	O
spectra	O
(	O
left	O
)	O
and	O
colored	O
red	O
,	O
yellow	O
,	O
or	O
green	O
on	O
the	O
surface	O
and	O
ribbon	O
structure	O
of	O
the	O
NTD	O
.	O
	O
S62	O
was	O
colored	O
gray	O
and	O
excluded	O
from	O
the	O
analysis	O
,	O
due	O
to	O
low	O
signal	O
intensity	O
.	O
	O
(	O
c	O
)	O
Docking	O
model	O
of	O
the	O
NTD	O
and	O
the	O
PIN	O
domain	O
.	O
	O
Critical	O
residues	O
in	O
the	O
PIN	O
domain	O
for	O
the	O
RNase	O
activity	O
of	O
Regnase	O
-	O
1	O
.	O
	O
(	O
a	O
)	O
In	O
vitro	O
cleavage	O
assay	O
of	O
basic	O
residue	O
mutants	O
for	O
IL	O
-	O
6	O
mRNA	O
.	O
	O
The	O
fluorescence	O
intensity	O
of	O
the	O
uncleaved	O
mRNA	O
was	O
quantified	O
and	O
the	O
results	O
were	O
mapped	O
on	O
the	O
PIN	O
dimer	O
structure	O
.	O
	O
Heterodimer	O
formation	O
by	O
combination	O
of	O
the	O
Regnase	O
-	O
1	O
basic	O
residue	O
mutants	O
and	O
the	O
DDNN	B-mutant
mutant	O
restored	O
the	O
RNase	O
activity	O
.	O
	O
(	O
a	O
)	O
Cartoon	O
representation	O
of	O
the	O
concept	O
of	O
the	O
experiment	O
.	O
(	O
b	O
)	O
In	O
vitro	O
cleavage	O
assay	O
of	O
Regnase	O
-	O
1	O
for	O
IL	O
-	O
6	O
mRNA	O
.	O
	O
The	O
fluorescence	O
intensity	O
of	O
the	O
uncleaved	O
mRNA	O
was	O
quantified	O
and	O
the	O
results	O
were	O
mapped	O
on	O
the	O
PIN	O
dimer	O
.	O
	O
The	O
mutations	O
whose	O
RNase	O
activities	O
were	O
restored	O
in	O
the	O
presence	O
of	O
DDNN	B-mutant
mutant	O
were	O
colored	O
in	O
red	O
or	O
yellow	O
on	O
the	O
primary	O
PIN	O
.	O
	O
Schematic	O
representation	O
of	O
regulation	O
of	O
the	O
Regnase	O
-	O
1	O
catalytic	O
activity	O
through	O
the	O
domain	O
-	O
domain	O
interactions	O
.	O
	O
Clan	O
CD	O
cysteine	O
peptidases	O
,	O
a	O
structurally	O
related	O
group	O
of	O
peptidases	O
that	O
include	O
mammalian	O
caspases	O
,	O
exhibit	O
a	O
wide	O
range	O
of	O
important	O
functions	O
,	O
along	O
with	O
a	O
variety	O
of	O
specificities	O
and	O
activation	O
mechanisms	O
.	O
	O
However	O
,	O
for	O
the	O
clostripain	O
family	O
(	O
denoted	O
C11	O
),	O
little	O
is	O
currently	O
known	O
.	O
	O
Here	O
,	O
we	O
describe	O
the	O
first	O
crystal	O
structure	O
of	O
a	O
C11	O
protein	O
from	O
the	O
human	O
gut	O
bacterium	O
,	O
Parabacteroides	O
merdae	O
(	O
PmC11	O
),	O
determined	O
to	O
1	O
.	O
7	O
-	O
Å	O
resolution	O
.	O
	O
PmC11	O
is	O
a	O
monomeric	O
cysteine	O
peptidase	O
that	O
comprises	O
an	O
extended	O
caspase	O
-	O
like	O
α	O
/	O
β	O
/	O
α	O
sandwich	O
and	O
an	O
unusual	O
C	O
-	O
terminal	O
domain	O
.	O
	O
It	O
shares	O
core	O
structural	O
elements	O
with	O
clan	O
CD	O
cysteine	O
peptidases	O
but	O
otherwise	O
structurally	O
differs	O
from	O
the	O
other	O
families	O
in	O
the	O
clan	O
.	O
	O
These	O
studies	O
also	O
revealed	O
a	O
well	O
ordered	O
break	O
in	O
the	O
polypeptide	O
chain	O
at	O
Lys147	O
,	O
resulting	O
in	O
a	O
large	O
conformational	O
rearrangement	O
close	O
to	O
the	O
active	O
site	O
.	O
	O
Biochemical	O
and	O
kinetic	O
analysis	O
revealed	O
Lys147	O
to	O
be	O
an	O
intramolecular	O
processing	O
site	O
at	O
which	O
cleavage	O
is	O
required	O
for	O
full	O
activation	O
of	O
the	O
enzyme	O
,	O
suggesting	O
an	O
autoinhibitory	O
mechanism	O
for	O
self	O
-	O
preservation	O
.	O
	O
PmC11	O
has	O
an	O
acidic	O
binding	O
pocket	O
and	O
a	O
preference	O
for	O
basic	O
substrates	O
,	O
and	O
accepts	O
substrates	O
with	O
Arg	O
and	O
Lys	O
in	O
P1	O
and	O
does	O
not	O
require	O
Ca2	O
+	O
for	O
activity	O
.	O
	O
Collectively	O
,	O
these	O
data	O
provide	O
insights	O
into	O
the	O
mechanism	O
and	O
activity	O
of	O
PmC11	O
and	O
a	O
detailed	O
framework	O
for	O
studies	O
on	O
C11	O
peptidases	O
from	O
other	O
phylogenetic	O
kingdoms	O
.	O
	O
Cysteine	O
peptidases	O
play	O
crucial	O
roles	O
in	O
the	O
virulence	O
of	O
bacterial	O
and	O
other	O
eukaryotic	O
pathogens	O
.	O
	O
Clan	O
CD	O
families	O
are	O
typically	O
described	O
using	O
the	O
name	O
of	O
their	O
archetypal	O
,	O
or	O
founding	O
,	O
member	O
and	O
also	O
given	O
an	O
identification	O
number	O
preceded	O
by	O
a	O
“	O
C	O
,”	O
to	O
denote	O
cysteine	O
peptidase	O
.	O
	O
Although	O
seven	O
families	O
(	O
C14	O
is	O
additionally	O
split	O
into	O
three	O
subfamilies	O
)	O
have	O
been	O
described	O
for	O
this	O
clan	O
,	O
crystal	O
structures	O
have	O
only	O
been	O
determined	O
from	O
four	O
:	O
legumain	O
(	O
C13	O
),	O
caspase	O
(	O
C14a	O
),	O
paracaspase	O
(	O
C14b	O
(	O
P	O
),	O
metacaspase	O
(	O
C14b	O
(	O
M	O
),	O
gingipain	O
(	O
C25	O
),	O
and	O
the	O
cysteine	O
peptidase	O
domain	O
(	O
CPD	O
)	O
of	O
various	O
toxins	O
(	O
C80	O
).	O
	O
No	O
structural	O
information	O
is	O
available	O
for	O
clostripain	O
(	O
C11	O
),	O
separase	O
(	O
C50	O
),	O
or	O
PrtH	O
-	O
peptidase	O
(	O
C85	O
).	O
	O
Clan	O
CD	O
enzymes	O
have	O
a	O
highly	O
conserved	O
His	O
/	O
Cys	O
catalytic	O
dyad	O
and	O
exhibit	O
strict	O
specificity	O
for	O
the	O
P1	O
residue	O
of	O
their	O
substrates	O
.	O
	O
The	O
archetypal	O
and	O
arguably	O
most	O
notable	O
family	O
in	O
the	O
clan	O
is	O
that	O
of	O
the	O
mammalian	O
caspases	O
(	O
C14a	O
),	O
although	O
clan	O
CD	O
members	O
are	O
distributed	O
throughout	O
the	O
entire	O
phylogenetic	O
kingdom	O
and	O
are	O
often	O
required	O
in	O
fundamental	O
biological	O
processes	O
.	O
	O
Interestingly	O
,	O
little	O
is	O
known	O
about	O
the	O
structure	O
or	O
function	O
of	O
the	O
C11	O
proteins	O
,	O
despite	O
their	O
widespread	O
distribution	O
and	O
its	O
archetypal	O
member	O
,	O
clostripain	O
from	O
Clostridium	O
histolyticum	O
,	O
first	O
reported	O
in	O
the	O
literature	O
in	O
1938	O
.	O
	O
Clostripain	O
has	O
been	O
described	O
as	O
an	O
arginine	O
-	O
specific	O
peptidase	O
with	O
a	O
requirement	O
for	O
Ca2	O
+	O
and	O
loss	O
of	O
an	O
internal	O
nonapeptide	O
for	O
full	O
activation	O
;	O
lack	O
of	O
structural	O
information	O
on	O
the	O
family	O
appears	O
to	O
have	O
prohibited	O
further	O
investigation	O
.	O
	O
As	O
part	O
of	O
an	O
ongoing	O
project	O
to	O
characterize	O
commensal	O
bacteria	O
in	O
the	O
microbiome	O
that	O
inhabit	O
the	O
human	O
gut	O
,	O
the	O
structure	O
of	O
C11	O
peptidase	O
,	O
PmC11	O
,	O
from	O
Parabacteroides	O
merdae	O
was	O
determined	O
using	O
the	O
Joint	O
Center	O
for	O
Structural	O
Genomics	O
(	O
JCSG	O
)	O
4	O
HTP	O
structural	O
biology	O
pipeline	O
.	O
	O
A	O
single	O
cleavage	O
was	O
observed	O
in	O
the	O
polypeptide	O
chain	O
at	O
Lys147	O
(	O
Fig	O
.	O
1	O
,	O
A	O
and	O
B	O
),	O
where	O
both	O
ends	O
of	O
the	O
cleavage	O
site	O
are	O
fully	O
visible	O
and	O
well	O
ordered	O
in	O
the	O
electron	O
density	O
.	O
	O
Helix	O
α3	O
sits	O
at	O
the	O
end	O
of	O
the	O
loop	O
following	O
β5	O
(	O
L5	O
),	O
just	O
preceding	O
the	O
Lys147	O
cleavage	O
site	O
,	O
with	O
both	O
L5	O
and	O
α3	O
pointing	O
away	O
from	O
the	O
central	O
β	O
-	O
sheet	O
and	O
toward	O
the	O
CTD	O
,	O
which	O
starts	O
with	O
α8	O
.	O
	O
Crystal	O
structure	O
of	O
a	O
C11	O
peptidase	O
from	O
P	O
.	O
merdae	O
.	O
	O
A	O
,	O
primary	O
sequence	O
alignment	O
of	O
PmC11	O
(	O
Uniprot	O
ID	O
A7A9N3	O
)	O
and	O
clostripain	O
(	O
Uniprot	O
ID	O
P09870	O
)	O
from	O
C	O
.	O
histolyticum	O
with	O
identical	O
residues	O
highlighted	O
in	O
gray	O
shading	O
.	O
	O
The	O
secondary	O
structure	O
of	O
PmC11	O
from	O
the	O
crystal	O
structure	O
is	O
mapped	O
onto	O
its	O
sequence	O
with	O
the	O
position	O
of	O
the	O
PmC11	O
catalytic	O
dyad	O
,	O
autocatalytic	O
cleavage	O
site	O
(	O
Lys147	O
),	O
and	O
S1	O
binding	O
pocket	O
Asp	O
(	O
Asp177	O
)	O
highlighted	O
by	O
a	O
red	O
star	O
,	O
a	O
red	O
downturned	O
triangle	O
,	O
and	O
a	O
red	O
upturned	O
triangle	O
,	O
respectively	O
.	O
	O
Connecting	O
loops	O
are	O
colored	O
gray	O
,	O
the	O
main	O
β	O
-	O
sheet	O
is	O
in	O
orange	O
,	O
with	O
other	O
strands	O
in	O
olive	O
,	O
α	O
-	O
helices	O
are	O
in	O
blue	O
,	O
and	O
the	O
nonapeptide	O
linker	O
of	O
clostripain	O
that	O
is	O
excised	O
upon	O
autocleavage	O
is	O
underlined	O
in	O
red	O
.	O
	O
Sequences	O
around	O
the	O
catalytic	O
site	O
of	O
clostripain	O
and	O
PmC11	O
align	O
well	O
.	O
	O
B	O
,	O
topology	O
diagram	O
of	O
PmC11	O
colored	O
as	O
in	O
A	O
except	O
that	O
additional	O
(	O
non	O
-	O
core	O
)	O
β	O
-	O
strands	O
are	O
in	O
yellow	O
.	O
	O
Helices	O
found	O
on	O
either	O
side	O
of	O
the	O
central	O
β	O
-	O
sheet	O
are	O
shown	O
above	O
and	O
below	O
the	O
sheet	O
,	O
respectively	O
.	O
	O
The	O
position	O
of	O
the	O
catalytic	O
dyad	O
(	O
H	O
,	O
C	O
)	O
and	O
the	O
processing	O
site	O
(	O
Lys147	O
)	O
are	O
highlighted	O
.	O
	O
Helices	O
(	O
1	O
–	O
14	O
)	O
and	O
β	O
-	O
strands	O
(	O
1	O
–	O
9	O
and	O
A	O
-	O
F	O
)	O
are	O
numbered	O
from	O
the	O
N	O
terminus	O
.	O
	O
C	O
,	O
tertiary	O
structure	O
of	O
PmC11	O
.	O
	O
The	O
N	O
and	O
C	O
termini	O
(	O
N	O
and	O
C	O
)	O
of	O
PmC11	O
along	O
with	O
the	O
central	O
β	O
-	O
sheet	O
(	O
1	O
–	O
9	O
),	O
helix	O
α5	O
,	O
and	O
helices	O
α8	O
,	O
α11	O
,	O
and	O
α13	O
from	O
the	O
C	O
-	O
terminal	O
domain	O
,	O
are	O
all	O
labeled	O
.	O
	O
Loops	O
are	O
colored	O
gray	O
,	O
the	O
main	O
β	O
-	O
sheet	O
is	O
in	O
orange	O
,	O
with	O
other	O
β	O
-	O
strands	O
in	O
yellow	O
,	O
and	O
α	O
-	O
helices	O
are	O
in	O
blue	O
.	O
	O
Of	O
the	O
interacting	O
secondary	O
structure	O
elements	O
,	O
α5	O
is	O
perhaps	O
the	O
most	O
interesting	O
.	O
	O
This	O
helix	O
makes	O
a	O
total	O
of	O
eight	O
hydrogen	O
bonds	O
with	O
the	O
CTD	O
,	O
including	O
one	O
salt	O
bridge	O
(	O
Arg191	O
-	O
Asp255	O
)	O
and	O
is	O
surrounded	O
by	O
the	O
CTD	O
on	O
one	O
side	O
and	O
the	O
main	O
core	O
of	O
the	O
enzyme	O
on	O
the	O
other	O
,	O
acting	O
like	O
a	O
linchpin	O
holding	O
both	O
components	O
together	O
(	O
Fig	O
.	O
1C	O
).	O
	O
PmC11	O
is	O
,	O
as	O
expected	O
,	O
most	O
structurally	O
similar	O
to	O
other	O
members	O
of	O
clan	O
CD	O
with	O
the	O
top	O
hits	O
in	O
a	O
search	O
of	O
known	O
structures	O
being	O
caspase	O
-	O
7	O
,	O
gingipain	O
-	O
K	O
,	O
and	O
legumain	O
(	O
PBD	O
codes	O
4hq0	O
,	O
4tkx	O
,	O
and	O
4aw9	O
,	O
respectively	O
)	O
(	O
Table	O
2	O
).	O
	O
The	O
C	O
-	O
terminal	O
domain	O
is	O
unique	O
to	O
PmC11	O
within	O
clan	O
CD	O
and	O
structure	O
comparisons	O
for	O
this	O
domain	O
alone	O
does	O
not	O
produce	O
any	O
hits	O
in	O
the	O
PDB	O
(	O
DaliLite	O
,	O
PDBeFold	O
),	O
suggesting	O
a	O
completely	O
novel	O
fold	O
.	O
	O
As	O
the	O
archetypal	O
and	O
arguably	O
most	O
well	O
studied	O
member	O
of	O
clan	O
CD	O
,	O
the	O
caspases	O
were	O
used	O
as	O
the	O
basis	O
to	O
investigate	O
the	O
structure	O
/	O
function	O
relationships	O
in	O
PmC11	O
,	O
with	O
caspase	O
-	O
7	O
as	O
the	O
representative	O
member	O
.	O
	O
Six	O
of	O
the	O
central	O
β	O
-	O
strands	O
in	O
PmC11	O
(	O
β1	O
–	O
β2	O
and	O
β5	O
–	O
β8	O
)	O
share	O
the	O
same	O
topology	O
as	O
the	O
six	O
-	O
stranded	O
β	O
-	O
sheet	O
found	O
in	O
caspases	O
,	O
with	O
strands	O
β3	O
,	O
β4	O
,	O
and	O
β9	O
located	O
on	O
the	O
outside	O
of	O
this	O
core	O
structure	O
(	O
Fig	O
.	O
1B	O
,	O
box	O
).	O
	O
Summary	O
of	O
PDBeFOLD	O
superposition	O
of	O
structures	O
found	O
to	O
be	O
most	O
similar	O
to	O
PmC11	O
in	O
the	O
PBD	O
based	O
on	O
DaliLite	O
	O
Biochemical	O
and	O
structural	O
characterization	O
of	O
PmC11	O
.	O
	O
A	O
,	O
ribbon	O
representation	O
of	O
the	O
overall	O
structure	O
of	O
PmC11	O
illustrating	O
the	O
catalytic	O
site	O
,	O
cleavage	O
site	O
displacement	O
,	O
and	O
potential	O
S1	O
binding	O
site	O
.	O
	O
The	O
overall	O
structure	O
of	O
PmC11	O
is	O
shown	O
in	O
gray	O
,	O
looking	O
down	O
into	O
the	O
catalytic	O
site	O
with	O
the	O
catalytic	O
dyad	O
in	O
red	O
.	O
	O
The	O
two	O
ends	O
of	O
the	O
autolytic	O
cleavage	O
site	O
(	O
Lys147	O
and	O
Ala148	O
,	O
green	O
)	O
are	O
displaced	O
by	O
19	O
.	O
5	O
Å	O
(	O
thin	O
black	O
line	O
)	O
from	O
one	O
another	O
and	O
residues	O
in	O
the	O
potential	O
substrate	O
binding	O
pocket	O
are	O
highlighted	O
in	O
blue	O
.	O
	O
PmC11	O
migrates	O
as	O
a	O
monomer	O
with	O
a	O
molecular	O
mass	O
around	O
41	O
kDa	O
calculated	O
from	O
protein	O
standards	O
of	O
known	O
molecular	O
weights	O
.	O
	O
Elution	O
fractions	O
across	O
the	O
major	O
peak	O
(	O
1	O
–	O
6	O
)	O
were	O
analyzed	O
by	O
SDS	O
-	O
PAGE	O
on	O
a	O
4	O
–	O
12	O
%	O
gel	O
in	O
MES	O
buffer	O
.	O
	O
C	O
,	O
the	O
active	O
form	O
of	O
PmC11	O
and	O
two	O
mutants	O
,	O
PmC11C179A	B-mutant
(	O
C	O
)	O
and	O
PmC11K147A	B-mutant
(	O
K	O
),	O
were	O
examined	O
by	O
SDS	O
-	O
PAGE	O
(	O
lane	O
1	O
)	O
and	O
Western	O
blot	O
analysis	O
using	O
an	O
anti	O
-	O
His	O
antibody	O
(	O
lane	O
2	O
)	O
show	O
that	O
PmC11	O
autoprocesses	O
,	O
whereas	O
mutants	O
,	O
PmC11C179A	B-mutant
and	O
PmC11K147A	B-mutant
,	O
do	O
not	O
show	O
autoprocessing	O
in	O
vitro	O
.	O
	O
Km	O
and	O
Vmax	O
of	O
PmC11	O
and	O
K147A	B-mutant
mutant	O
were	O
determined	O
by	O
monitoring	O
change	O
in	O
the	O
fluorescence	O
corresponding	O
to	O
AMC	O
release	O
from	O
Bz	O
-	O
R	O
-	O
AMC	O
.	O
	O
E	O
,	O
intermolecular	O
processing	O
of	O
PmC11C179A	B-mutant
by	O
PmC11	O
.	O
	O
PmC11C179A	O
(	O
20	O
μg	O
)	O
was	O
incubated	O
overnight	O
at	O
37	O
°	O
C	O
with	O
increasing	O
amounts	O
of	O
processed	O
PmC11	O
and	O
analyzed	O
on	O
a	O
10	O
%	O
SDS	O
-	O
PAGE	O
gel	O
.	O
	O
A	O
single	O
lane	O
of	O
20	O
μg	O
of	O
active	O
PmC11	O
(	O
labeled	O
20	O
)	O
is	O
shown	O
for	O
comparison	O
.	O
	O
The	O
position	O
of	O
the	O
catalytic	O
dyad	O
,	O
one	O
potential	O
key	O
substrate	O
binding	O
residue	O
Asp177	O
,	O
and	O
the	O
ends	O
of	O
the	O
cleavage	O
site	O
Lys147	O
and	O
Ala148	O
are	O
indicated	O
.	O
	O
Five	O
of	O
the	O
α	O
-	O
helices	O
surrounding	O
the	O
β	O
-	O
sheet	O
of	O
PmC11	O
(	O
α1	O
,	O
α2	O
,	O
α4	O
,	O
α6	O
,	O
and	O
α7	O
)	O
are	O
found	O
in	O
similar	O
positions	O
to	O
the	O
five	O
structurally	O
conserved	O
helices	O
in	O
caspases	O
and	O
other	O
members	O
of	O
clan	O
CD	O
,	O
apart	O
from	O
family	O
C80	O
.	O
	O
Other	O
than	O
its	O
more	O
extended	O
β	O
-	O
sheet	O
,	O
PmC11	O
differs	O
most	O
significantly	O
from	O
other	O
clan	O
CD	O
members	O
at	O
its	O
C	O
terminus	O
,	O
where	O
the	O
CTD	O
contains	O
a	O
further	O
seven	O
α	O
-	O
helices	O
and	O
four	O
β	O
-	O
strands	O
after	O
β8	O
.	O
	O
Autoprocessing	O
of	O
PmC11	O
	O
Purification	O
of	O
recombinant	O
PmC11	O
(	O
molecular	O
mass	O
=	O
42	O
.	O
6	O
kDa	O
)	O
revealed	O
partial	O
processing	O
into	O
two	O
cleavage	O
products	O
of	O
26	O
.	O
4	O
and	O
16	O
.	O
2	O
kDa	O
,	O
related	O
to	O
the	O
observed	O
cleavage	O
at	O
Lys147	O
in	O
the	O
crystal	O
structure	O
(	O
Fig	O
.	O
2A	O
).	O
	O
The	O
single	O
cleavage	O
site	O
of	O
PmC11	O
at	O
Lys147	O
is	O
found	O
immediately	O
after	O
α3	O
,	O
in	O
loop	O
L5	O
within	O
the	O
central	O
β	O
-	O
sheet	O
(	O
Figs	O
.	O
1	O
,	O
A	O
and	O
B	O
,	O
and	O
2A	O
).	O
	O
The	O
two	O
ends	O
of	O
the	O
cleavage	O
site	O
are	O
remarkably	O
well	O
ordered	O
in	O
the	O
crystal	O
structure	O
and	O
displaced	O
from	O
one	O
another	O
by	O
19	O
.	O
5	O
Å	O
(	O
Fig	O
.	O
2A	O
).	O
	O
Moreover	O
,	O
the	O
C	O
-	O
terminal	O
side	O
of	O
the	O
cleavage	O
site	O
resides	O
near	O
the	O
catalytic	O
dyad	O
with	O
Ala148	O
being	O
4	O
.	O
5	O
and	O
5	O
.	O
7	O
Å	O
from	O
His133	O
and	O
Cys179	O
,	O
respectively	O
.	O
	O
Consequently	O
,	O
it	O
appears	O
feasible	O
that	O
the	O
helix	O
attached	O
to	O
Lys147	O
(	O
α3	O
)	O
could	O
be	O
responsible	O
for	O
steric	O
autoinhibition	O
of	O
PmC11	O
when	O
Lys147	O
is	O
covalently	O
bonded	O
to	O
Ala148	O
.	O
	O
Thus	O
,	O
the	O
cleavage	O
would	O
be	O
required	O
for	O
full	O
activation	O
of	O
PmC11	O
.	O
	O
To	O
investigate	O
this	O
possibility	O
,	O
two	O
mutant	O
forms	O
of	O
the	O
enzyme	O
were	O
created	O
:	O
PmC11C179A	B-mutant
(	O
a	O
catalytically	O
inactive	O
mutant	O
)	O
and	O
PmC11K147A	B-mutant
(	O
a	O
cleavage	O
-	O
site	O
mutant	O
).	O
	O
Initial	O
SDS	O
-	O
PAGE	O
and	O
Western	O
blot	O
analysis	O
of	O
both	O
mutants	O
revealed	O
no	O
discernible	O
processing	O
occurred	O
as	O
compared	O
with	O
active	O
PmC11	O
(	O
Fig	O
.	O
2C	O
).	O
	O
The	O
PmC11K147A	B-mutant
mutant	O
enzyme	O
had	O
a	O
markedly	O
different	O
reaction	O
rate	O
(	O
Vmax	O
)	O
compared	O
with	O
WT	O
,	O
where	O
the	O
reaction	O
velocity	O
of	O
PmC11	O
was	O
10	O
times	O
greater	O
than	O
that	O
of	O
PmC11K147A	B-mutant
(	O
Fig	O
.	O
2D	O
).	O
	O
Taken	O
together	O
,	O
these	O
data	O
reveal	O
that	O
PmC11	O
requires	O
processing	O
at	O
Lys147	O
for	O
optimum	O
activity	O
.	O
	O
To	O
investigate	O
whether	O
processing	O
is	O
a	O
result	O
of	O
intra	O
-	O
or	O
intermolecular	O
cleavage	O
,	O
the	O
PmC11C179A	B-mutant
mutant	O
was	O
incubated	O
with	O
increasing	O
concentrations	O
of	O
processed	O
and	O
activated	O
PmC11	O
.	O
	O
These	O
studies	O
revealed	O
that	O
there	O
was	O
no	O
apparent	O
cleavage	O
of	O
PmC11C179A	B-mutant
by	O
the	O
active	O
enzyme	O
at	O
low	O
concentrations	O
of	O
PmC11	O
and	O
that	O
only	O
limited	O
cleavage	O
was	O
observed	O
when	O
the	O
ratio	O
of	O
active	O
enzyme	O
(	O
PmC11	O
:	O
PmC11C179A	B-mutant
)	O
was	O
increased	O
to	O
∼	O
1	O
:	O
10	O
and	O
1	O
:	O
4	O
,	O
with	O
complete	O
cleavage	O
observed	O
at	O
a	O
ratio	O
of	O
1	O
:	O
1	O
(	O
Fig	O
.	O
2E	O
).	O
	O
This	O
suggests	O
that	O
cleavage	O
of	O
PmC11C179A	B-mutant
was	O
most	O
likely	O
an	O
effect	O
of	O
the	O
increasing	O
concentration	O
of	O
PmC11	O
and	O
intermolecular	O
cleavage	O
.	O
	O
Collectively	O
,	O
these	O
data	O
suggest	O
that	O
the	O
pro	O
-	O
form	O
of	O
PmC11	O
is	O
autoinhibited	O
by	O
a	O
section	O
of	O
L5	O
blocking	O
access	O
to	O
the	O
active	O
site	O
,	O
prior	O
to	O
intramolecular	O
cleavage	O
at	O
Lys147	O
.	O
	O
This	O
cleavage	O
subsequently	O
allows	O
movement	O
of	O
the	O
region	O
containing	O
Lys147	O
and	O
the	O
active	O
site	O
to	O
open	O
up	O
for	O
substrate	O
access	O
.	O
	O
Substrate	O
Specificity	O
of	O
PmC11	O
	O
The	O
autocatalytic	O
cleavage	O
of	O
PmC11	O
at	O
Lys147	O
(	O
sequence	O
KLK	O
∧	O
A	O
)	O
demonstrates	O
that	O
the	O
enzyme	O
accepts	O
substrates	O
with	O
Lys	O
in	O
the	O
P1	O
position	O
.	O
	O
The	O
catalytic	O
dyad	O
of	O
PmC11	O
sits	O
near	O
the	O
bottom	O
of	O
an	O
open	O
pocket	O
on	O
the	O
surface	O
of	O
the	O
enzyme	O
at	O
a	O
conserved	O
location	O
in	O
the	O
clan	O
CD	O
family	O
.	O
	O
The	O
PmC11	O
structure	O
reveals	O
that	O
the	O
catalytic	O
dyad	O
forms	O
part	O
of	O
a	O
large	O
acidic	O
pocket	O
(	O
Fig	O
.	O
2G	O
),	O
consistent	O
with	O
a	O
binding	O
site	O
for	O
a	O
basic	O
substrate	O
.	O
	O
This	O
pocket	O
is	O
lined	O
with	O
the	O
potential	O
functional	O
side	O
chains	O
of	O
Asn50	O
,	O
Asp177	O
,	O
and	O
Thr204	O
with	O
Gly134	O
,	O
Asp207	O
,	O
and	O
Met205	O
also	O
contributing	O
to	O
the	O
pocket	O
(	O
Fig	O
.	O
2A	O
).	O
	O
Interestingly	O
,	O
these	O
residues	O
are	O
in	O
regions	O
that	O
are	O
structurally	O
similar	O
to	O
those	O
involved	O
in	O
the	O
S1	O
binding	O
pockets	O
of	O
other	O
clan	O
CD	O
members	O
(	O
shown	O
in	O
Ref	O
.).	O
	O
Z	O
-	O
VRPR	O
-	O
FMK	O
was	O
also	O
shown	O
to	O
bind	O
to	O
the	O
enzyme	O
:	O
a	O
size	O
-	O
shift	O
was	O
observed	O
,	O
by	O
SDS	O
-	O
PAGE	O
analysis	O
,	O
in	O
the	O
larger	O
processed	O
product	O
of	O
PmC11	O
suggesting	O
that	O
the	O
inhibitor	O
bound	O
to	O
the	O
active	O
site	O
(	O
Fig	O
.	O
3B	O
).	O
	O
A	O
structure	O
overlay	O
of	O
PmC11	O
with	O
the	O
MALT1	O
-	O
paracacaspase	O
(	O
MALT1	O
-	O
P	O
),	O
in	O
complex	O
with	O
Z	O
-	O
VRPR	O
-	O
FMK	O
,	O
revealed	O
that	O
the	O
PmC11	O
dyad	O
sits	O
in	O
a	O
very	O
similar	O
position	O
to	O
that	O
of	O
active	O
MALT1	O
-	O
P	O
and	O
that	O
Asn50	O
,	O
Asp177	O
,	O
and	O
Asp207	O
superimpose	O
well	O
with	O
the	O
principal	O
MALT1	O
-	O
P	O
inhibitor	O
binding	O
residues	O
(	O
Asp365	O
,	O
Asp462	O
,	O
and	O
Glu500	O
,	O
respectively	O
(	O
VRPR	O
-	O
FMK	O
from	O
MALT1	O
-	O
P	O
with	O
the	O
corresponding	O
PmC11	O
residues	O
from	O
the	O
structural	O
overlay	O
is	O
shown	O
in	O
Fig	O
.	O
1D	O
),	O
as	O
described	O
in	O
Ref	O
.).	O
	O
However	O
,	O
this	O
loop	O
has	O
been	O
shown	O
to	O
be	O
important	O
for	O
substrate	O
binding	O
in	O
clan	O
CD	O
and	O
this	O
residue	O
could	O
easily	O
rotate	O
and	O
be	O
involved	O
in	O
substrate	O
binding	O
in	O
PmC11	O
.	O
	O
Thus	O
,	O
Asn50	O
,	O
Asp177	O
,	O
and	O
Asp207	O
are	O
most	O
likely	O
responsible	O
for	O
the	O
substrate	O
specificity	O
of	O
PmC11	O
.	O
	O
PmC11	O
binds	O
and	O
is	O
inhibited	O
by	O
Z	O
-	O
VRPR	O
-	O
FMK	O
and	O
does	O
not	O
require	O
Ca2	O
+	O
for	O
activity	O
.	O
	O
Cleavage	O
of	O
Bz	O
-	O
R	O
-	O
AMC	O
by	O
PmC11	O
was	O
measured	O
in	O
a	O
fluorometric	O
activity	O
assay	O
with	O
(+,	O
purple	O
)	O
and	O
without	O
(−,	O
red	O
)	O
Z	O
-	O
VRPR	O
-	O
FMK	O
.	O
	O
PmC11	O
was	O
incubated	O
with	O
(+)	O
or	O
without	O
(−)	O
Z	O
-	O
VRPR	O
-	O
FMK	O
and	O
the	O
samples	O
analyzed	O
on	O
a	O
10	O
%	O
SDS	O
-	O
PAGE	O
gel	O
.	O
	O
A	O
size	O
shift	O
can	O
be	O
observed	O
in	O
the	O
larger	O
processed	O
product	O
of	O
PmC11	O
(	O
26	O
.	O
1	O
kDa	O
).	O
	O
C	O
,	O
PmC11	O
with	O
the	O
Z	O
-	O
VRPR	O
-	O
FMK	O
from	O
the	O
MALT1	O
-	O
paracacaspase	O
(	O
MALT1	O
-	O
P	O
)	O
superimposed	O
.	O
	O
A	O
three	O
-	O
dimensional	O
structural	O
overlay	O
of	O
Z	O
-	O
VRPR	O
-	O
FMK	O
from	O
the	O
MALT1	O
-	O
P	O
complex	O
onto	O
PmC11	O
.	O
	O
Residues	O
surrounding	O
the	O
inhibitor	O
are	O
labeled	O
and	O
represent	O
potentially	O
important	O
binding	O
site	O
residues	O
,	O
labeled	O
in	O
black	O
and	O
shown	O
in	O
an	O
atomic	O
representation	O
.	O
	O
Furthermore	O
,	O
Cu2	O
+,	O
Fe2	O
+,	O
and	O
Zn2	O
+	O
appear	O
to	O
inhibit	O
PmC11	O
.	O
	O
Comparison	O
with	O
Clostripain	O
	O
Clostripain	O
from	O
C	O
.	O
histolyticum	O
is	O
the	O
founding	O
member	O
of	O
the	O
C11	O
family	O
of	O
peptidases	O
and	O
contains	O
an	O
additional	O
149	O
residues	O
compared	O
with	O
PmC11	O
.	O
	O
A	O
multiple	O
sequence	O
alignment	O
revealed	O
that	O
most	O
of	O
the	O
secondary	O
structural	O
elements	O
are	O
conserved	O
between	O
the	O
two	O
enzymes	O
,	O
although	O
they	O
are	O
only	O
∼	O
23	O
%	O
identical	O
(	O
Fig	O
.	O
1A	O
).	O
	O
Nevertheless	O
,	O
PmC11	O
may	O
be	O
a	O
good	O
model	O
for	O
the	O
core	O
structure	O
of	O
clostripain	O
.	O
	O
In	O
addition	O
,	O
the	O
predicted	O
primary	O
S1	O
-	O
binding	O
residue	O
in	O
PmC11	O
Asp177	O
also	O
overlays	O
with	O
the	O
residue	O
predicted	O
to	O
be	O
the	O
P1	O
specificity	O
determining	O
residue	O
in	O
clostripain	O
(	O
Asp229	O
,	O
Fig	O
.	O
1A	O
).	O
	O
Surprisingly	O
,	O
Ca2	O
+	O
did	O
not	O
enhance	O
PmC11	O
activity	O
and	O
,	O
furthermore	O
,	O
other	O
divalent	O
cations	O
,	O
Mg2	O
+,	O
Mn2	O
+,	O
Co2	O
+,	O
Fe2	O
+,	O
Zn2	O
+,	O
and	O
Cu2	O
+,	O
were	O
not	O
necessary	O
for	O
PmC11	O
activity	O
(	O
Fig	O
.	O
3D	O
).	O
	O
The	O
crystal	O
structure	O
of	O
PmC11	O
now	O
provides	O
three	O
-	O
dimensional	O
information	O
for	O
a	O
member	O
of	O
the	O
clostripain	O
C11	O
family	O
of	O
cysteine	O
peptidases	O
.	O
	O
The	O
structural	O
similarity	O
of	O
PmC11	O
with	O
its	O
nearest	O
structural	O
neighbors	O
in	O
the	O
PDB	O
is	O
decidedly	O
low	O
,	O
overlaying	O
better	O
with	O
six	O
-	O
stranded	O
caspase	O
-	O
7	O
than	O
any	O
of	O
the	O
other	O
larger	O
members	O
of	O
the	O
clan	O
(	O
Table	O
2	O
).	O
	O
The	O
substrate	O
specificity	O
of	O
PmC11	O
is	O
Arg	O
/	O
Lys	O
and	O
the	O
crystal	O
structure	O
revealed	O
an	O
acidic	O
pocket	O
for	O
specific	O
binding	O
of	O
such	O
basic	O
substrates	O
.	O
	O
PmC11	O
differs	O
from	O
clostripain	O
in	O
that	O
is	O
does	O
not	O
appear	O
to	O
require	O
divalent	O
cations	O
for	O
activation	O
.	O
	O
To	O
date	O
,	O
the	O
effector	O
caspases	O
are	O
the	O
only	O
group	O
of	O
enzymes	O
that	O
require	O
cleavage	O
of	O
a	O
loop	O
within	O
the	O
central	O
β	O
-	O
sheet	O
.	O
	O
All	O
other	O
clan	O
CD	O
members	O
requiring	O
cleavage	O
for	O
full	O
activation	O
do	O
so	O
at	O
sites	O
external	O
to	O
their	O
central	O
sheets	O
.	O
	O
In	O
addition	O
,	O
several	O
members	O
of	O
clan	O
CD	O
exhibit	O
self	O
-	O
inhibition	O
,	O
whereby	O
regions	O
of	O
the	O
enzyme	O
block	O
access	O
to	O
the	O
active	O
site	O
.	O
	O
The	O
structure	O
of	O
PmC11	O
gives	O
the	O
first	O
insight	O
into	O
this	O
class	O
of	O
relatively	O
unexplored	O
family	O
of	O
proteins	O
and	O
should	O
allow	O
important	O
catalytic	O
and	O
substrate	O
binding	O
residues	O
to	O
be	O
identified	O
in	O
a	O
variety	O
of	O
orthologues	O
.	O
	O
Ribosome	O
biogenesis	O
factor	O
Tsr3	O
is	O
the	O
aminocarboxypropyl	O
transferase	O
responsible	O
for	O
18S	O
rRNA	O
hypermodification	O
in	O
yeast	O
and	O
humans	O
	O
Here	O
we	O
identify	O
the	O
cytoplasmic	O
ribosome	O
biogenesis	O
protein	O
Tsr3	O
as	O
the	O
responsible	O
enzyme	O
in	O
yeast	O
and	O
human	O
cells	O
.	O
	O
In	O
functionally	O
impaired	O
Tsr3	O
-	O
mutants	O
,	O
a	O
reduced	O
level	O
of	O
acp	O
modification	O
directly	O
correlates	O
with	O
increased	O
20S	O
pre	O
-	O
rRNA	O
accumulation	O
.	O
	O
The	O
crystal	O
structure	O
of	O
archaeal	O
Tsr3	O
homologs	O
revealed	O
the	O
same	O
fold	O
as	O
in	O
SPOUT	O
-	O
class	O
RNA	O
-	O
methyltransferases	O
but	O
a	O
distinct	O
SAM	O
binding	O
mode	O
.	O
	O
This	O
unique	O
SAM	O
binding	O
mode	O
explains	O
why	O
Tsr3	O
transfers	O
the	O
acp	O
and	O
not	O
the	O
methyl	O
group	O
of	O
SAM	O
to	O
its	O
substrate	O
.	O
	O
Structurally	O
,	O
Tsr3	O
therefore	O
represents	O
a	O
novel	O
class	O
of	O
acp	O
transferase	O
enzymes	O
.	O
	O
Eukaryotic	O
ribosome	O
biogenesis	O
is	O
highly	O
complex	O
and	O
requires	O
a	O
large	O
number	O
of	O
non	O
-	O
ribosomal	O
proteins	O
and	O
small	O
non	O
-	O
coding	O
RNAs	O
in	O
addition	O
to	O
ribosomal	O
RNAs	O
(	O
rRNAs	O
)	O
and	O
proteins	O
.	O
	O
During	O
eukaryotic	O
ribosome	O
biogenesis	O
several	O
dozens	O
of	O
rRNA	O
nucleotides	O
become	O
chemically	O
modified	O
.	O
	O
The	O
most	O
abundant	O
rRNA	O
modifications	O
are	O
methylations	O
at	O
the	O
2	O
′-	O
OH	O
ribose	O
moieties	O
and	O
isomerizations	O
of	O
uridine	O
residues	O
to	O
pseudouridine	O
,	O
catalyzed	O
by	O
small	O
nucleolar	O
ribonucleoprotein	O
particles	O
(	O
snoRNPs	O
).	O
	O
Ribosomal	O
RNA	O
modifications	O
have	O
been	O
suggested	O
to	O
optimize	O
ribosome	O
function	O
,	O
although	O
in	O
most	O
cases	O
this	O
remains	O
to	O
be	O
clearly	O
established	O
.	O
	O
Most	O
modified	O
rRNA	O
nucleotides	O
cluster	O
in	O
the	O
vicinity	O
of	O
the	O
decoding	O
or	O
the	O
peptidyl	O
transferase	O
center	O
,	O
suggesting	O
an	O
influence	O
on	O
ribosome	O
functionality	O
and	O
stability	O
.	O
	O
The	O
chemically	O
most	O
complex	O
modification	O
is	O
located	O
in	O
the	O
loop	O
capping	O
helix	O
31	O
of	O
18S	O
rRNA	O
(	O
Supplementary	O
Figure	O
S1B	O
).	O
	O
There	O
a	O
uridine	O
(	O
U1191	O
in	O
yeast	O
)	O
is	O
modified	O
to	O
1	O
-	O
methyl	O
-	O
3	O
-(	O
3	O
-	O
amino	O
-	O
3	O
-	O
carboxypropyl	O
)-	O
pseudouridine	O
(	O
m1acp3Ψ	O
,	O
Figure	O
1A	O
).	O
	O
This	O
base	O
modification	O
was	O
first	O
described	O
in	O
1968	O
for	O
hamster	O
cells	O
and	O
is	O
conserved	O
in	O
eukaryotes	O
.	O
	O
This	O
hypermodified	O
nucleotide	O
,	O
which	O
is	O
located	O
at	O
the	O
P	O
-	O
site	O
tRNA	O
,	O
is	O
synthesized	O
in	O
three	O
steps	O
beginning	O
with	O
the	O
snR35	O
H	O
/	O
ACA	O
snoRNP	O
guided	O
conversion	O
of	O
uridine	O
into	O
pseudouridine	O
.	O
	O
Methylation	O
can	O
only	O
occur	O
once	O
pseudouridylation	O
has	O
taken	O
place	O
,	O
as	O
the	O
latter	O
reaction	O
generates	O
the	O
substrate	O
for	O
the	O
former	O
.	O
	O
The	O
final	O
acp	O
modification	O
leading	O
to	O
N1	O
-	O
methyl	O
-	O
N3	O
-	O
aminocarboxypropyl	O
-	O
pseudouridine	O
occurs	O
late	O
during	O
40S	O
biogenesis	O
in	O
the	O
cytoplasm	O
,	O
while	O
the	O
two	O
former	O
reactions	O
are	O
taking	O
place	O
in	O
the	O
nucleolus	O
and	O
nucleus	O
,	O
and	O
is	O
independent	O
from	O
pseudouridylation	O
or	O
methylation	O
.	O
	O
Both	O
the	O
methyl	O
and	O
the	O
acp	O
group	O
are	O
derived	O
from	O
S	O
-	O
adenosylmethionine	O
(	O
SAM	O
),	O
but	O
the	O
enzyme	O
responsible	O
for	O
acp	O
modification	O
remained	O
elusive	O
for	O
more	O
than	O
40	O
years	O
.	O
	O
Tsr3	O
is	O
necessary	O
for	O
acp	O
modification	O
of	O
18S	O
rRNA	O
in	O
yeast	O
and	O
human	O
.	O
(	O
A	O
)	O
Hypermodified	O
nucleotide	O
m1acp3Ψ	O
is	O
synthesized	O
in	O
three	O
steps	O
:	O
pseudouridylation	O
catalyzed	O
by	O
snoRNP35	O
,	O
N1	O
-	O
methylation	O
catalyzed	O
by	O
methyltransferase	O
Nep1	O
and	O
N3	O
-	O
acp	O
modification	O
catalyzed	O
by	O
Tsr3	O
.	O
	O
(	O
B	O
)	O
RP	O
-	O
HPLC	O
elution	O
profile	O
of	O
yeast	O
18S	O
rRNA	O
nucleosides	O
.	O
	O
Wild	O
type	O
(	O
WT	O
)	O
and	O
plasmid	O
encoded	O
18S	O
rRNA	O
(	O
U1191U	B-mutant
)	O
show	O
the	O
14C	O
-	O
acp	O
signal	O
,	O
whereas	O
the	O
14C	O
-	O
acp	O
signal	O
is	O
missing	O
in	O
the	O
U1191A	B-mutant
mutant	O
plasmid	O
encoded	O
18S	O
rRNA	O
(	O
U1191A	B-mutant
)	O
and	O
Δtsr3	B-mutant
mutants	O
(	O
Δtsr3	B-mutant
).	O
	O
All	O
samples	O
were	O
loaded	O
on	O
the	O
gel	O
with	O
two	O
different	O
amounts	O
of	O
5	O
and	O
10	O
μl	O
.	O
(	O
D	O
)	O
Primer	O
extension	O
analysis	O
of	O
acp	O
modification	O
in	O
yeast	O
18S	O
rRNA	O
(	O
right	O
gel	O
)	O
including	O
a	O
sequencing	O
ladder	O
(	O
left	O
gel	O
).	O
	O
The	O
primer	O
extension	O
stop	O
at	O
nucleotide	O
1191	O
is	O
missing	O
exclusively	O
in	O
Δtsr3	B-mutant
mutants	O
and	O
Δtsr3	B-mutant
Δsnr35	I-mutant
recombinants	O
.	O
	O
(	O
E	O
)	O
Primer	O
extension	O
analysis	O
of	O
human	O
18S	O
rRNA	O
after	O
siRNA	O
knockdown	O
of	O
HsNEP1	O
/	O
EMG1	O
(	O
541	O
,	O
542	O
and	O
543	O
)	O
and	O
HsTSR3	O
(	O
544	O
and	O
545	O
)	O
(	O
right	O
gel	O
),	O
including	O
a	O
sequencing	O
ladder	O
(	O
left	O
gel	O
).	O
	O
Only	O
a	O
few	O
acp	O
transferring	O
enzymes	O
have	O
been	O
characterized	O
until	O
now	O
.	O
	O
During	O
the	O
biosynthesis	O
of	O
wybutosine	O
,	O
a	O
tricyclic	O
nucleoside	O
present	O
in	O
eukaryotic	O
and	O
archaeal	O
phenylalanine	O
tRNA	O
,	O
Tyw2	O
(	O
Trm12	O
in	O
yeast	O
)	O
transfers	O
an	O
acp	O
group	O
from	O
SAM	O
to	O
an	O
acidic	O
carbon	O
atom	O
.	O
	O
Archaeal	O
Tyw2	O
has	O
a	O
structure	O
very	O
similar	O
to	O
Rossmann	O
-	O
fold	O
(	O
class	O
I	O
)	O
RNA	O
-	O
methyltransferases	O
,	O
but	O
its	O
distinctive	O
SAM	O
-	O
binding	O
mode	O
enables	O
the	O
transfer	O
of	O
the	O
acp	O
group	O
instead	O
of	O
the	O
methyl	O
group	O
of	O
the	O
cofactor	O
.	O
	O
Another	O
acp	O
modification	O
has	O
been	O
described	O
in	O
the	O
diphtamide	O
biosynthesis	O
pathway	O
,	O
where	O
an	O
acp	O
group	O
is	O
transferred	O
from	O
SAM	O
to	O
the	O
carbon	O
atom	O
of	O
a	O
histidine	O
residue	O
of	O
eukaryotic	O
translation	O
elongation	O
factor	O
2	O
by	O
use	O
of	O
a	O
radical	O
mechanism	O
.	O
	O
In	O
a	O
recent	O
bioinformatic	O
study	O
,	O
the	O
uncharacterized	O
yeast	O
gene	O
YOR006c	O
was	O
predicted	O
to	O
be	O
involved	O
in	O
ribosome	O
biogenesis	O
.	O
	O
It	O
is	O
highly	O
conserved	O
among	O
eukaryotes	O
and	O
archaea	O
(	O
Supplementary	O
Figure	O
S1A	O
)	O
and	O
its	O
deletion	O
leads	O
to	O
an	O
accumulation	O
of	O
the	O
20S	O
pre	O
-	O
rRNA	O
precursor	O
of	O
18S	O
rRNA	O
,	O
suggesting	O
an	O
influence	O
on	O
D	O
-	O
site	O
cleavage	O
during	O
the	O
maturation	O
of	O
the	O
small	O
ribosomal	O
subunit	O
.	O
	O
On	O
this	O
basis	O
,	O
YOR006C	O
was	O
renamed	O
‘	O
Twenty	O
S	O
rRNA	O
accumulation	O
3	O
′	O
(	O
TSR3	O
).	O
	O
However	O
,	O
its	O
function	O
remained	O
unclear	O
although	O
recently	O
a	O
putative	O
nuclease	O
function	O
during	O
18S	O
rRNA	O
maturation	O
was	O
predicted	O
.	O
	O
Here	O
,	O
we	O
identify	O
Tsr3	O
as	O
the	O
long	O
-	O
sought	O
acp	O
transferase	O
that	O
catalyzes	O
the	O
last	O
step	O
in	O
the	O
biosynthesis	O
of	O
the	O
hypermodified	O
nucleotide	O
m1acp3Ψ	O
in	O
yeast	O
and	O
human	O
cells	O
.	O
	O
Furthermore	O
using	O
catalytically	O
defective	O
mutants	O
of	O
yeast	O
Tsr3	O
we	O
demonstrated	O
that	O
the	O
acp	O
modification	O
is	O
required	O
for	O
18S	O
rRNA	O
maturation	O
.	O
	O
Surprisingly	O
,	O
the	O
crystal	O
structures	O
of	O
archaeal	O
homologs	O
revealed	O
that	O
Tsr3	O
is	O
structurally	O
similar	O
to	O
the	O
SPOUT	O
-	O
class	O
RNA	O
methyltransferases	O
.	O
	O
Interestingly	O
,	O
the	O
two	O
structurally	O
very	O
different	O
enzymes	O
use	O
similar	O
strategies	O
in	O
binding	O
the	O
SAM	O
-	O
cofactor	O
in	O
order	O
to	O
ensure	O
that	O
in	O
contrast	O
to	O
methyltransferases	O
the	O
acp	O
and	O
not	O
the	O
methyl	O
group	O
of	O
SAM	O
is	O
transferred	O
to	O
the	O
substrate	O
.	O
	O
Tsr3	O
is	O
the	O
enzyme	O
responsible	O
for	O
18S	O
rRNA	O
acp	O
modification	O
in	O
yeast	O
and	O
humans	O
	O
The	O
S	O
.	O
cerevisiae	O
18S	O
rRNA	O
acp	O
transferase	O
was	O
identified	O
in	O
a	O
systematic	O
genetic	O
screen	O
where	O
numerous	O
deletion	O
mutants	O
from	O
the	O
EUROSCARF	O
strain	O
collection	O
(	O
www	O
.	O
euroscarf	O
.	O
de	O
)	O
were	O
analyzed	O
by	O
HPLC	O
for	O
alterations	O
in	O
18S	O
rRNA	O
base	O
modifications	O
.	O
	O
For	O
the	O
Δtsr3	B-mutant
deletion	O
strain	O
the	O
HPLC	O
elution	O
profile	O
of	O
18S	O
rRNA	O
nucleosides	O
(	O
Figure	O
1B	O
)	O
was	O
very	O
similar	O
to	O
that	O
of	O
the	O
pseudouridine	O
-	O
N1	O
methyltransferase	O
mutant	O
Δnep1	B-mutant
,	O
where	O
a	O
shoulder	O
at	O
∼	O
7	O
.	O
4	O
min	O
elution	O
time	O
was	O
missing	O
in	O
the	O
elution	O
profile	O
.	O
	O
In	O
order	O
to	O
directly	O
analyze	O
the	O
presence	O
of	O
the	O
acp	O
modification	O
of	O
nucleotide	O
1191	O
we	O
used	O
an	O
in	O
vivo14C	O
incorporation	O
assay	O
with	O
1	O
-	O
14C	O
-	O
methionine	O
.	O
	O
No	O
radioactive	O
labeling	O
was	O
detected	O
in	O
the	O
18S	B-mutant
U1191A	I-mutant
mutant	O
which	O
served	O
as	O
a	O
control	O
for	O
the	O
specificity	O
of	O
the	O
14C	O
-	O
aminocarboxypropyl	O
incorporation	O
.	O
	O
As	O
previously	O
shown	O
,	O
only	O
the	O
acp	O
but	O
none	O
of	O
the	O
other	O
modifications	O
at	O
U1191	O
of	O
yeast	O
18S	O
rRNA	O
blocks	O
reverse	O
transcriptase	O
activity	O
.	O
	O
Therefore	O
the	O
presence	O
of	O
the	O
acp	O
modification	O
can	O
be	O
directly	O
assessed	O
by	O
primer	O
extension	O
.	O
	O
In	O
contrast	O
,	O
in	O
a	O
Δtsr3	B-mutant
mutant	O
no	O
primer	O
extension	O
stop	O
signal	O
was	O
present	O
at	O
this	O
position	O
.	O
	O
In	O
a	O
Δtsr3	B-mutant
Δsnr35	I-mutant
double	O
deletion	O
strain	O
the	O
18S	O
rRNA	O
contains	O
an	O
unmodified	O
U	O
and	O
the	O
primer	O
extension	O
stop	O
signal	O
was	O
missing	O
(	O
Figure	O
1D	O
).	O
	O
The	O
Tsr3	O
protein	O
is	O
highly	O
conserved	O
in	O
yeast	O
and	O
humans	O
(	O
50	O
%	O
identity	O
).	O
	O
After	O
siRNA	O
-	O
mediated	O
depletion	O
of	O
Tsr3	O
in	O
human	O
colon	O
carcinoma	O
HCT116	O
(+/+)	O
cells	O
the	O
acp	O
primer	O
extension	O
arrest	O
was	O
reduced	O
in	O
comparison	O
to	O
cells	O
transfected	O
with	O
a	O
non	O
-	O
targeting	O
scramble	O
siRNA	O
control	O
(	O
Figure	O
1E	O
,	O
compare	O
lanes	O
544	O
and	O
scramble	O
).	O
	O
This	O
suggests	O
that	O
low	O
residual	O
levels	O
of	O
HsTsr3	O
are	O
sufficient	O
to	O
modify	O
the	O
RNA	O
.	O
	O
Thus	O
,	O
HsTsr3	O
is	O
also	O
responsible	O
for	O
the	O
acp	O
modification	O
of	O
18S	O
rRNA	O
nucleotide	O
Ψ1248	O
in	O
helix	O
31	O
.	O
	O
Similar	O
to	O
yeast	O
,	O
siRNA	O
-	O
mediated	O
depletion	O
of	O
the	O
Ψ1248	O
N1	O
-	O
methyltransferase	O
Nep1	O
/	O
Emg1	O
had	O
no	O
influence	O
on	O
the	O
primer	O
extension	O
arrest	O
(	O
Figure	O
1E	O
).	O
	O
Although	O
the	O
acp	O
modification	O
of	O
18S	O
rRNA	O
is	O
highly	O
conserved	O
in	O
eukaryotes	O
,	O
yeast	O
Δtsr3	B-mutant
mutants	O
showed	O
only	O
a	O
minor	O
growth	O
defect	O
.	O
	O
However	O
,	O
the	O
Δtsr3	B-mutant
deletion	O
was	O
synthetic	O
sick	O
with	O
a	O
Δsnr35	B-mutant
deletion	O
preventing	O
pseudouridylation	O
and	O
Nep1	O
-	O
catalyzed	O
methylation	O
of	O
nucleotide	O
1191	O
(	O
Figure	O
2A	O
).	O
	O
Interestingly	O
,	O
no	O
increased	O
growth	O
defect	O
could	O
be	O
observed	O
for	O
Δtsr3	B-mutant
Δnep1	I-mutant
recombinants	O
containing	O
the	O
nep1	O
suppressor	O
mutation	O
Δnop6	B-mutant
as	O
well	O
as	O
for	O
Δtsr3	B-mutant
Δsnr35	I-mutant
Δnep1	I-mutant
recombinants	O
with	O
unmodified	O
U1191	O
(	O
Supplementary	O
Figure	O
S2D	O
and	O
E	O
).	O
	O
The	O
Δtsr3	B-mutant
deletion	O
is	O
synthetic	O
sick	O
with	O
a	O
Δsnr35	B-mutant
deletion	O
preventing	O
U1191	O
pseudouridylation	O
.	O
	O
(	O
B	O
)	O
In	O
agar	O
diffusion	O
assays	O
the	O
yeast	O
Δtsr3	B-mutant
deletion	O
mutant	O
shows	O
a	O
hypersensitivity	O
against	O
paromomycin	O
and	O
hygromycin	O
B	O
which	O
is	O
further	O
increased	O
by	O
recombination	O
with	O
Δsnr35	B-mutant
.	O
(	O
C	O
)	O
Northern	O
blot	O
analysis	O
with	O
an	O
ITS1	O
hybridization	O
probe	O
after	O
siRNA	O
depletion	O
of	O
HsTSR3	O
(	O
siRNAs	O
544	O
and	O
545	O
)	O
and	O
a	O
scrambled	O
siRNA	O
as	O
control	O
.	O
	O
The	O
accumulation	O
of	O
18SE	O
and	O
47S	O
and	O
/	O
or	O
45S	O
pre	O
-	O
RNAs	O
is	O
enforced	O
upon	O
HsTSR3	O
depletion	O
.	O
	O
Right	O
gel	O
:	O
Ethidium	O
bromide	O
staining	O
showing	O
18S	O
and	O
28S	O
rRNAs	O
.	O
	O
(	O
D	O
)	O
Cytoplasmic	O
localization	O
of	O
yeast	O
Tsr3	O
shown	O
by	O
fluorescence	O
microscopy	O
of	O
GFP	B-mutant
-	I-mutant
fused	I-mutant
Tsr3	I-mutant
.	O
	O
From	O
left	O
to	O
right	O
:	O
differential	O
interference	O
contrast	O
(	O
DIC	O
),	O
green	O
fluorescence	O
of	O
GFP	B-mutant
-	I-mutant
Tsr3	I-mutant
,	O
red	O
fluorescence	O
of	O
Nop56	B-mutant
-	I-mutant
mRFP	I-mutant
as	O
nucleolar	O
marker	O
,	O
and	O
merge	O
of	O
GFP	B-mutant
-	I-mutant
Tsr3	I-mutant
/	O
Nop56	B-mutant
-	I-mutant
mRFP	I-mutant
with	O
DIC	O
.	O
(	O
E	O
)	O
Elution	O
profile	O
(	O
A254	O
)	O
after	O
sucrose	O
gradient	O
separation	O
of	O
yeast	O
ribosomal	O
subunits	O
and	O
polysomes	O
(	O
upper	O
part	O
)	O
and	O
western	O
blot	O
analysis	O
of	O
3xHA	O
tagged	O
Tsr3	O
(	O
Tsr3	B-mutant
-	I-mutant
3xHA	I-mutant
)	O
after	O
SDS	O
-	O
PAGE	O
separation	O
of	O
polysome	O
profile	O
fractions	O
taken	O
every	O
20	O
s	O
(	O
lower	O
part	O
).	O
	O
The	O
influence	O
of	O
the	O
acp	O
modification	O
of	O
nucleotide	O
1191	O
on	O
ribosome	O
function	O
was	O
analyzed	O
by	O
treating	O
Δtsr3	B-mutant
mutants	O
with	O
protein	O
synthesis	O
inhibitors	O
.	O
	O
Similar	O
to	O
a	O
temperature	O
-	O
sensitive	O
nep1	O
mutant	O
,	O
the	O
Δtsr3	B-mutant
deletion	O
caused	O
hypersensitivity	O
to	O
paromomycin	O
and	O
,	O
to	O
a	O
lesser	O
extent	O
,	O
to	O
hygromycin	O
B	O
(	O
Figure	O
2B	O
),	O
but	O
not	O
to	O
G418	O
or	O
cycloheximide	O
(	O
data	O
not	O
shown	O
).	O
	O
In	O
a	O
yeast	O
Δtsr3	B-mutant
strain	O
as	O
well	O
as	O
in	O
the	O
Δtsr3	B-mutant
Δsnr35	I-mutant
recombinant	O
20S	O
pre	O
-	O
rRNA	O
accumulated	O
significantly	O
and	O
the	O
level	O
of	O
mature	O
18S	O
rRNA	O
was	O
reduced	O
(	O
Supplementary	O
Figures	O
S2C	O
and	O
S3D	O
),	O
as	O
reported	O
previously	O
.	O
	O
A	O
minor	O
effect	O
on	O
20S	O
rRNA	O
accumulation	O
was	O
also	O
observed	O
for	O
Δsnr35	B-mutant
,	O
but	O
-	O
probably	O
due	O
to	O
different	O
strain	O
backgrounds	O
–	O
to	O
a	O
weaker	O
extent	O
than	O
described	O
earlier	O
.	O
	O
In	O
human	O
cells	O
,	O
the	O
depletion	O
of	O
HsTsr3	O
in	O
HCT116	O
(+/+)	O
cells	O
caused	O
an	O
accumulation	O
of	O
the	O
human	O
20S	O
pre	O
-	O
rRNA	O
equivalent	O
18S	O
-	O
E	O
suggesting	O
an	O
evolutionary	O
conserved	O
role	O
of	O
Tsr3	O
in	O
the	O
late	O
steps	O
of	O
18S	O
rRNA	O
processing	O
(	O
Figure	O
2C	O
and	O
Supplementary	O
Figure	O
S2B	O
).	O
	O
In	O
polysome	O
profiles	O
,	O
a	O
reduced	O
level	O
of	O
80S	O
ribosomes	O
and	O
a	O
strong	O
signal	O
for	O
free	O
60S	O
subunits	O
was	O
observed	O
in	O
line	O
with	O
the	O
40S	O
subunit	O
deficiency	O
(	O
Supplementary	O
Figure	O
S2G	O
).	O
	O
Cellular	O
localization	O
of	O
Tsr3	O
in	O
S	O
.	O
cerevisiae	O
	O
Fluorescence	O
microscopy	O
of	O
GFP	O
-	O
tagged	O
Tsr3	O
localized	O
the	O
fusion	O
protein	O
in	O
the	O
cytoplasm	O
of	O
yeast	O
cells	O
and	O
no	O
co	O
-	O
localization	O
with	O
the	O
nucleolar	O
marker	O
protein	O
Nop56	O
could	O
be	O
observed	O
(	O
Figure	O
2D	O
).	O
	O
This	O
agrees	O
with	O
previous	O
biochemical	O
data	O
suggesting	O
that	O
the	O
acp	O
modification	O
of	O
18S	O
rRNA	O
occurs	O
late	O
during	O
40S	O
subunit	O
biogenesis	O
in	O
the	O
cytoplasm	O
,	O
and	O
makes	O
an	O
additional	O
nuclear	O
localization	O
as	O
reported	O
in	O
a	O
previous	O
large	O
-	O
scale	O
analysis	O
unlikely	O
.	O
	O
Searches	O
for	O
sequence	O
homologs	O
of	O
S	O
.	O
cerevisiae	O
Tsr3	O
(	O
ScTsr3	O
)	O
by	O
us	O
and	O
others	O
revealed	O
that	O
the	O
genomes	O
of	O
many	O
archaea	O
contain	O
genes	O
encoding	O
Tsr3	O
-	O
like	O
proteins	O
.	O
	O
To	O
locate	O
the	O
domains	O
most	O
important	O
for	O
Tsr3	O
activity	O
,	O
ScTsr3	O
fragments	O
of	O
different	O
lengths	O
containing	O
the	O
highly	O
conserved	O
central	O
part	O
were	O
expressed	O
in	O
a	O
Δtsr3	B-mutant
mutant	O
(	O
Figure	O
3A	O
)	O
and	O
analyzed	O
by	O
primer	O
extension	O
(	O
Figure	O
3B	O
)	O
and	O
Northern	O
blotting	O
(	O
Figure	O
3C	O
).	O
	O
TSR3	O
fragments	O
of	O
different	O
length	O
were	O
expressed	O
under	O
the	O
native	O
promotor	O
from	O
multicopy	O
plasmids	O
in	O
a	O
Δtsr3	B-mutant
deletion	O
strain	O
.	O
	O
(	O
B	O
)	O
Primer	O
extension	O
analysis	O
of	O
18S	O
rRNA	O
acp	O
modification	O
in	O
yeast	O
cells	O
expressing	O
the	O
indicated	O
TSR3	O
fragments	O
.	O
	O
N	O
-	O
terminal	O
deletions	O
of	O
36	O
or	O
45	O
amino	O
acids	O
and	O
C	O
-	O
terminal	O
deletions	O
of	O
43	O
or	O
76	O
residues	O
show	O
a	O
primer	O
extension	O
stop	O
comparable	O
to	O
the	O
wild	O
type	O
.	O
	O
Tsr3	O
fragments	O
37	O
–	O
223	O
or	O
46	O
–	O
223	O
cause	O
a	O
nearly	O
complete	O
loss	O
of	O
the	O
arrest	O
signal	O
.	O
	O
The	O
box	O
highlights	O
the	O
shortest	O
Tsr3	O
fragment	O
(	O
aa	O
46	O
–	O
270	O
)	O
with	O
wild	O
type	O
activity	O
(	O
strong	O
primer	O
extension	O
block	O
).	O
(	O
C	O
)	O
Northern	O
blot	O
analysis	O
of	O
20S	O
pre	O
-	O
rRNA	O
accumulation	O
.	O
	O
A	O
weak	O
20S	O
rRNA	O
signal	O
,	O
indicating	O
normal	O
processing	O
,	O
is	O
observed	O
for	O
Tsr3	O
fragment	O
46	O
–	O
270	O
(	O
highlighted	O
in	O
a	O
box	O
)	O
showing	O
its	O
functionality	O
.	O
	O
Thus	O
,	O
the	O
archaeal	O
homologs	O
correspond	O
to	O
the	O
functional	O
core	O
of	O
Tsr3	O
.	O
	O
In	O
order	O
to	O
define	O
the	O
structural	O
basis	O
for	O
Tsr3	O
function	O
,	O
homologs	O
from	O
thermophilic	O
archaea	O
were	O
screened	O
for	O
crystallization	O
.	O
	O
Well	O
diffracting	O
crystals	O
were	O
obtained	O
for	O
Tsr3	O
homologs	O
from	O
the	O
two	O
crenarchaeal	O
species	O
Vulcanisaeta	O
distributa	O
(	O
VdTsr3	O
)	O
and	O
Sulfolobus	O
solfataricus	O
(	O
SsTsr3	O
)	O
which	O
share	O
36	O
%	O
(	O
VdTsr3	O
)	O
and	O
38	O
%	O
(	O
SsTsr3	O
)	O
identity	O
with	O
the	O
ScTsr3	O
core	O
region	O
(	O
ScTsr3	O
aa	O
46	O
–	O
223	O
).	O
	O
While	O
for	O
S	O
.	O
solfataricus	O
the	O
existence	O
of	O
a	O
modified	O
nucleotide	O
of	O
unknown	O
chemical	O
composition	O
in	O
the	O
loop	O
capping	O
helix	O
31	O
of	O
its	O
16S	O
rRNA	O
has	O
been	O
demonstrated	O
,	O
no	O
information	O
regarding	O
rRNA	O
modifications	O
is	O
yet	O
available	O
for	O
V	O
.	O
distributa	O
.	O
	O
Crystals	O
of	O
VdTsr3	O
diffracted	O
to	O
a	O
resolution	O
of	O
1	O
.	O
6	O
Å	O
whereas	O
crystals	O
of	O
SsTsr3	O
diffracted	O
to	O
2	O
.	O
25	O
Å	O
.	O
Serendipitously	O
,	O
VdTsr3	O
was	O
purified	O
and	O
crystallized	O
in	O
complex	O
with	O
endogenous	O
(	O
E	O
.	O
coli	O
)	O
SAM	O
(	O
Supplementary	O
Figure	O
S4	O
)	O
while	O
SsTsr3	O
crystals	O
contained	O
the	O
protein	O
in	O
the	O
apo	O
state	O
.	O
	O
The	O
structure	O
of	O
VdTsr3	O
was	O
solved	O
ab	O
initio	O
,	O
by	O
single	O
-	O
wavelength	O
anomalous	O
diffraction	O
phasing	O
(	O
Se	O
-	O
SAD	O
)	O
with	O
Se	O
containing	O
derivatives	O
(	O
selenomethionine	O
and	O
seleno	O
-	O
substituted	O
SAM	O
).	O
	O
The	O
structure	O
of	O
SsTsr3	O
was	O
solved	O
by	O
molecular	O
replacement	O
using	O
VdTsr3	O
as	O
a	O
search	O
model	O
(	O
see	O
Supplementary	O
Table	O
S1	O
for	O
data	O
collection	O
and	O
refinement	O
statistics	O
).	O
	O
The	O
structure	O
of	O
VdTsr3	O
can	O
be	O
divided	O
into	O
two	O
domains	O
(	O
Figure	O
4A	O
).	O
	O
The	O
N	O
-	O
terminal	O
domain	O
(	O
aa	O
1	O
–	O
92	O
)	O
has	O
a	O
mixed	O
α	O
/	O
β	O
-	O
structure	O
centered	O
around	O
a	O
five	O
-	O
stranded	O
all	O
-	O
parallel	O
β	O
-	O
sheet	O
(	O
Figure	O
4B	O
)	O
with	O
the	O
strand	O
order	O
β5	O
↑-	O
β3	O
↑-	O
β4	O
↑-	O
β1	O
↑-	O
β2	O
↑.	O
The	O
loops	O
connecting	O
β1	O
and	O
β2	O
,	O
β3	O
and	O
β4	O
and	O
β4	O
and	O
β5	O
include	O
α	O
-	O
helices	O
α1	O
,	O
α2	O
and	O
α3	O
,	O
respectively	O
.	O
	O
Thus	O
,	O
the	O
VdTsr3	O
structure	O
contains	O
a	O
deep	O
trefoil	O
knot	O
.	O
	O
The	O
structure	O
of	O
SsTsr3	O
in	O
the	O
apo	O
state	O
is	O
very	O
similar	O
to	O
that	O
of	O
VdTsr3	O
(	O
Figure	O
4C	O
)	O
with	O
an	O
RMSD	O
for	O
equivalent	O
Cα	O
atoms	O
of	O
1	O
.	O
1	O
Å	O
.	O
The	O
only	O
significant	O
difference	O
in	O
the	O
global	O
structure	O
of	O
the	O
two	O
proteins	O
is	O
the	O
presence	O
of	O
an	O
extended	O
α	O
-	O
helix	O
α8	O
and	O
the	O
absence	O
of	O
α	O
-	O
helix	O
α9	O
in	O
SsTsr3	O
.	O
	O
β	O
-	O
strands	O
are	O
colored	O
in	O
crimson	O
whereas	O
α	O
-	O
helices	O
in	O
the	O
N	O
-	O
terminal	O
domain	O
are	O
colored	O
light	O
blue	O
and	O
α	O
-	O
helices	O
in	O
the	O
C	O
-	O
terminal	O
domain	O
are	O
colored	O
dark	O
blue	O
.	O
	O
The	O
locations	O
of	O
the	O
α	O
-	O
helix	O
α8	O
which	O
is	O
longer	O
in	O
SsTsr3	O
and	O
of	O
α	O
-	O
helix	O
α9	O
which	O
is	O
only	O
present	O
in	O
VdTsr3	O
are	O
indicated	O
.	O
(	O
D	O
)	O
Secondary	O
structure	O
cartoon	O
(	O
left	O
)	O
of	O
S	O
.	O
pombe	O
Trm10	O
(	O
pdb4jwf	O
)—	O
the	O
SPOUT	O
-	O
class	O
RNA	O
methyltransferase	O
structurally	O
most	O
similar	O
to	O
Tsr3	O
and	O
superposition	O
of	O
the	O
VdTsr3	O
and	O
Trm10	O
X	O
-	O
ray	O
structures	O
(	O
right	O
).	O
(	O
E	O
)	O
Analytical	O
gel	O
filtration	O
profiles	O
for	O
VdTsr3	O
(	O
red	O
)	O
and	O
SsTsr3	O
(	O
blue	O
)	O
show	O
that	O
both	O
proteins	O
are	O
monomeric	O
in	O
solution	O
.	O
	O
Vd	O
,	O
Vulcanisaeta	O
distributa	O
;	O
Ss	O
,	O
Sulfolobus	O
solfataricus	O
.	O
	O
However	O
,	O
no	O
structural	O
similarity	O
to	O
an	O
RLI	O
-	O
domain	O
was	O
detectable	O
.	O
	O
This	O
is	O
in	O
accordance	O
with	O
the	O
functional	O
analysis	O
of	O
alanine	O
replacement	O
mutations	O
of	O
cysteine	O
residues	O
in	O
ScTsr3	O
(	O
Supplementary	O
Figure	O
S3	O
).	O
	O
The	O
β	O
-	O
strand	O
topology	O
and	O
the	O
deep	O
C	O
-	O
terminal	O
trefoil	O
knot	O
of	O
archaeal	O
Tsr3	O
are	O
the	O
structural	O
hallmarks	O
of	O
the	O
SPOUT	O
-	O
class	O
RNA	O
-	O
methyltransferase	O
fold	O
.	O
	O
The	O
closest	O
structural	O
homolog	O
identified	O
in	O
a	O
DALI	O
search	O
is	O
the	O
tRNA	O
methyltransferase	O
Trm10	O
(	O
DALI	O
Z	O
-	O
score	O
6	O
.	O
8	O
)	O
which	O
methylates	O
the	O
N1	O
nitrogen	O
of	O
G9	O
/	O
A9	O
in	O
many	O
archaeal	O
and	O
eukaryotic	O
tRNAs	O
by	O
using	O
SAM	O
as	O
the	O
methyl	O
group	O
donor	O
.	O
	O
Interestingly	O
,	O
Nep1	O
—	O
the	O
enzyme	O
preceding	O
Tsr3	O
in	O
the	O
biosynthetic	O
pathway	O
for	O
the	O
synthesis	O
of	O
m1acp3Ψ	O
—	O
also	O
belongs	O
to	O
the	O
SPOUT	O
-	O
class	O
of	O
RNA	O
methyltransferases	O
.	O
	O
However	O
,	O
the	O
structural	O
similarities	O
between	O
Nep1	O
and	O
Tsr3	O
(	O
DALI	O
Z	O
-	O
score	O
4	O
.	O
4	O
)	O
are	O
less	O
pronounced	O
than	O
between	O
Tsr3	O
and	O
Trm10	O
.	O
	O
Most	O
SPOUT	O
-	O
class	O
RNA	O
-	O
methyltransferases	O
are	O
homodimers	O
.	O
	O
Gel	O
filtration	O
experiments	O
with	O
both	O
VdTsr3	O
and	O
SsTsr3	O
(	O
Figure	O
4E	O
)	O
showed	O
that	O
both	O
proteins	O
are	O
monomeric	O
in	O
solution	O
thereby	O
extending	O
the	O
structural	O
similarities	O
to	O
Trm10	O
.	O
	O
So	O
far	O
,	O
structural	O
information	O
is	O
only	O
available	O
for	O
one	O
other	O
enzyme	O
that	O
transfers	O
the	O
acp	O
group	O
from	O
SAM	O
to	O
an	O
RNA	O
nucleotide	O
.	O
	O
This	O
enzyme	O
,	O
Tyw2	O
,	O
is	O
part	O
of	O
the	O
biosynthesis	O
pathway	O
of	O
wybutosine	O
nucleotides	O
in	O
tRNAs	O
.	O
	O
Instead	O
,	O
Tyw2	O
has	O
a	O
fold	O
typical	O
for	O
the	O
class	O
-	O
I	O
-	O
or	O
Rossmann	O
-	O
fold	O
class	O
of	O
methyltransferases	O
(	O
Supplementary	O
Figure	O
S5B	O
).	O
	O
Cofactor	O
binding	O
of	O
Tsr3	O
	O
The	O
SAM	O
-	O
binding	O
site	O
of	O
Tsr3	O
is	O
located	O
in	O
a	O
deep	O
crevice	O
between	O
the	O
N	O
-	O
and	O
C	O
-	O
terminal	O
domains	O
in	O
the	O
vicinity	O
of	O
the	O
trefoil	O
knot	O
as	O
typical	O
for	O
SPOUT	O
-	O
class	O
RNA	O
-	O
methyltransferases	O
(	O
Figure	O
4A	O
).	O
	O
The	O
adenine	O
base	O
of	O
the	O
cofactor	O
is	O
recognized	O
by	O
hydrogen	O
bonds	O
between	O
its	O
N1	O
nitrogen	O
and	O
the	O
backbone	O
amide	O
of	O
L93	O
directly	O
preceding	O
β5	O
as	O
well	O
as	O
between	O
its	O
N6	O
-	O
amino	O
group	O
and	O
the	O
backbone	O
carbonyl	O
group	O
of	O
Y108	O
located	O
in	O
the	O
loop	O
connecting	O
β5	O
in	O
the	O
N	O
-	O
terminal	O
and	O
α4	O
in	O
the	O
C	O
-	O
terminal	O
domain	O
(	O
Figure	O
5A	O
).	O
	O
Furthermore	O
,	O
the	O
adenine	O
base	O
of	O
SAM	O
is	O
involved	O
in	O
hydrophobic	O
packing	O
interactions	O
with	O
the	O
side	O
chains	O
of	O
L45	O
(	O
β3	O
),	O
P47	O
and	O
W73	O
(	O
α3	O
)	O
in	O
the	O
N	O
-	O
terminal	O
domain	O
as	O
well	O
as	O
with	O
L93	O
,	O
L110	O
(	O
both	O
in	O
the	O
loop	O
connecting	O
β5	O
and	O
α4	O
)	O
and	O
A115	O
(	O
α5	O
)	O
in	O
the	O
C	O
-	O
terminal	O
domain	O
.	O
	O
The	O
acp	O
side	O
chain	O
of	O
SAM	O
is	O
fixed	O
in	O
position	O
by	O
hydrogen	O
bonding	O
of	O
its	O
carboxylate	O
group	O
to	O
the	O
backbone	O
amide	O
and	O
the	O
side	O
chain	O
hydroxyl	O
group	O
of	O
T19	O
in	O
α1	O
as	O
well	O
as	O
the	O
backbone	O
amide	O
group	O
of	O
T112	O
in	O
α4	O
(	O
C	O
-	O
terminal	O
domain	O
).	O
	O
Most	O
importantly	O
,	O
the	O
methyl	O
group	O
of	O
SAM	O
is	O
buried	O
in	O
a	O
hydrophobic	O
pocket	O
formed	O
by	O
the	O
sidechains	O
of	O
W73	O
and	O
A76	O
both	O
located	O
in	O
α3	O
(	O
Figure	O
5A	O
and	O
B	O
).	O
	O
SAM	O
-	O
binding	O
by	O
Tsr3	O
.	O
	O
Nitrogen	O
atoms	O
are	O
dark	O
blue	O
,	O
oxygen	O
atoms	O
red	O
,	O
sulfur	O
atoms	O
orange	O
,	O
carbon	O
atoms	O
of	O
the	O
protein	O
light	O
blue	O
and	O
carbon	O
atoms	O
of	O
SAM	O
yellow	O
.	O
	O
(	O
B	O
)	O
Solvent	O
accessibility	O
of	O
the	O
acp	O
group	O
of	O
SAM	O
bound	O
to	O
VdTsr3	O
.	O
	O
The	O
solvent	O
accessible	O
surface	O
of	O
the	O
protein	O
is	O
shown	O
in	O
semitransparent	O
gray	O
whereas	O
SAM	O
is	O
show	O
in	O
a	O
stick	O
representation	O
.	O
	O
A	O
red	O
arrow	O
indicates	O
the	O
reactive	O
CH2	O
-	O
moiety	O
of	O
the	O
acp	O
group	O
.	O
(	O
C	O
)	O
Solvent	O
accessibility	O
of	O
the	O
SAM	O
methyl	O
group	O
for	O
SAM	O
bound	O
to	O
the	O
RNA	O
methyltransferase	O
Trm10	O
.	O
	O
Tryptophan	O
fluorescence	O
quenching	O
curves	O
upon	O
addition	O
of	O
SAM	O
(	O
blue	O
),	O
5	O
′-	O
methyl	O
-	O
thioadenosine	O
(	O
red	O
)	O
and	O
SAH	O
(	O
black	O
).	O
	O
Radioactively	O
labeled	O
SAM	O
is	O
retained	O
on	O
a	O
filter	O
in	O
the	O
presence	O
of	O
SsTsr3	O
.	O
	O
Addition	O
of	O
unlabeled	O
SAM	O
competes	O
with	O
the	O
binding	O
of	O
labeled	O
SAM	O
.	O
	O
A	O
W66A	B-mutant
-	O
mutant	O
of	O
SsTsr3	O
(	O
W73	O
in	O
VdTsr3	O
)	O
does	O
not	O
bind	O
SAM	O
.	O
	O
(	O
F	O
)	O
Primer	O
extension	O
(	O
upper	O
left	O
)	O
shows	O
a	O
strongly	O
reduced	O
acp	O
modification	O
of	O
yeast	O
18S	O
rRNA	O
in	O
Δtsr3	B-mutant
cells	O
expressing	O
Tsr3	B-mutant
-	I-mutant
S62D	I-mutant
,	O
-	B-mutant
E111A	I-mutant
or	O
–	B-mutant
W114A	I-mutant
.	O
	O
3xHA	O
tagged	O
Tsr3	O
mutants	O
are	O
expressed	O
comparable	O
to	O
the	O
wild	O
type	O
as	O
shown	O
by	O
western	O
blot	O
(	O
lower	O
left	O
).	O
	O
Binding	O
affinities	O
for	O
SAM	O
and	O
its	O
analogs	O
5	O
′-	O
methylthioadenosin	O
and	O
SAH	O
to	O
SsTsr3	O
were	O
measured	O
using	O
tryptophan	O
fluorescence	O
quenching	O
.	O
	O
VdTsr3	O
could	O
not	O
be	O
used	O
in	O
these	O
experiments	O
since	O
we	O
could	O
not	O
purify	O
it	O
in	O
a	O
stable	O
SAM	O
-	O
free	O
form	O
.	O
	O
S	O
-	O
adenosylhomocysteine	O
which	O
lacks	O
the	O
methyl	O
group	O
of	O
SAM	O
binds	O
with	O
significantly	O
lower	O
affinity	O
(	O
KD	O
=	O
55	O
.	O
5	O
μM	O
)	O
(	O
Figure	O
5D	O
).	O
	O
On	O
the	O
other	O
hand	O
,	O
the	O
loss	O
of	O
hydrogen	O
bonds	O
between	O
the	O
acp	O
sidechain	O
carboxylate	O
group	O
and	O
the	O
protein	O
appears	O
to	O
be	O
thermodynamically	O
less	O
important	O
but	O
these	O
hydrogen	O
bonds	O
might	O
play	O
a	O
crucial	O
role	O
for	O
the	O
proper	O
orientation	O
of	O
the	O
cofactor	O
side	O
chain	O
in	O
the	O
substrate	O
binding	O
pocket	O
.	O
	O
Accordingly	O
,	O
a	O
W66A	B-mutant
-	O
mutation	O
(	O
W73	O
in	O
VdTsr3	O
)	O
of	O
SsTsr3	O
significantly	O
diminished	O
SAM	O
-	O
binding	O
in	O
a	O
filter	O
binding	O
assay	O
compared	O
to	O
the	O
wild	O
type	O
(	O
Figure	O
5E	O
).	O
	O
Nevertheless	O
,	O
a	O
mutation	O
of	O
the	O
equivalent	O
position	O
S62	O
of	O
ScTsr3	O
to	O
D	O
,	O
but	O
not	O
to	O
A	O
,	O
resulted	O
in	O
reduced	O
acp	O
modification	O
in	O
vivo	O
,	O
as	O
shown	O
by	O
primer	O
extension	O
analysis	O
(	O
Figure	O
5F	O
).	O
	O
The	O
acp	O
-	O
transfer	O
reaction	O
catalyzed	O
by	O
Tsr3	O
most	O
likely	O
requires	O
the	O
presence	O
of	O
a	O
catalytic	O
base	O
in	O
order	O
to	O
abstract	O
a	O
proton	O
from	O
the	O
N3	O
imino	O
group	O
of	O
the	O
modified	O
pseudouridine	O
.	O
	O
The	O
side	O
chain	O
of	O
D70	O
(	O
VdTsr3	O
)	O
located	O
in	O
β4	O
is	O
∼	O
5	O
Å	O
away	O
from	O
the	O
SAM	O
sulfur	O
atom	O
.	O
	O
This	O
residue	O
is	O
conserved	O
as	O
D	O
or	O
E	O
both	O
in	O
archaeal	O
and	O
eukaryotic	O
Tsr3	O
homologs	O
.	O
	O
However	O
,	O
the	O
mutation	O
of	O
the	O
corresponding	O
residue	O
of	O
ScTsr3	O
(	O
E111A	B-mutant
)	O
leads	O
to	O
a	O
significant	O
decrease	O
of	O
the	O
acp	O
transferase	O
activity	O
in	O
vivo	O
(	O
Figure	O
5F	O
).	O
	O
RNA	O
-	O
binding	O
of	O
Tsr3	O
	O
Furthermore	O
,	O
a	O
negatively	O
charged	O
MES	O
-	O
ion	O
is	O
found	O
in	O
the	O
crystal	O
structure	O
of	O
VdTsr3	O
complexed	O
to	O
the	O
side	O
chain	O
of	O
K22	O
in	O
helix	O
α1	O
.	O
	O
Helix	O
α1	O
contains	O
two	O
more	O
positively	O
charged	O
amino	O
acids	O
K17	O
and	O
R25	O
as	O
does	O
the	O
loop	O
preceding	O
it	O
(	O
R9	O
).	O
	O
A	O
second	O
cluster	O
of	O
positively	O
charged	O
residues	O
is	O
found	O
in	O
or	O
near	O
helix	O
α3	O
(	O
K74	O
,	O
R75	O
,	O
K82	O
,	O
R85	O
and	O
K87	O
).	O
	O
Some	O
of	O
these	O
amino	O
acids	O
are	O
conserved	O
between	O
archaeal	O
and	O
eukaryotic	O
Tsr3	O
(	O
Supplementary	O
Figure	O
S1A	O
).	O
	O
A	O
triple	O
mutation	O
of	O
the	O
conserved	O
positively	O
charged	O
residues	O
R60	O
,	O
K65	O
and	O
R131	O
to	O
A	O
in	O
ScTsr3	O
resulted	O
in	O
a	O
protein	O
with	O
a	O
significantly	O
impaired	O
acp	O
transferase	O
activity	O
in	O
vivo	O
(	O
Figure	O
6D	O
)	O
in	O
line	O
with	O
an	O
important	O
functional	O
role	O
for	O
these	O
positively	O
charged	O
residues	O
.	O
	O
(	O
A	O
)	O
Electrostatic	O
charge	O
distribution	O
on	O
the	O
surface	O
of	O
VdTsr3	O
.	O
	O
SAM	O
is	O
shown	O
in	O
a	O
stick	O
representation	O
.	O
	O
Conserved	O
basic	O
amino	O
acids	O
are	O
labeled	O
.	O
(	O
B	O
)	O
Comparison	O
of	O
the	O
secondary	O
structures	O
of	O
helix	O
31	O
from	O
the	O
small	O
ribosomal	O
subunit	O
rRNAs	O
in	O
S	O
.	O
cerevisiae	O
and	O
S	O
.	O
solfataricus	O
with	O
the	O
location	O
of	O
the	O
hypermodified	O
nucleotide	O
indicated	O
in	O
red	O
.	O
	O
For	O
S	O
.	O
solfataricus	O
the	O
chemical	O
identity	O
of	O
the	O
hypermodified	O
nucleotide	O
is	O
not	O
known	O
but	O
the	O
existence	O
of	O
NEP1	O
and	O
TSR3	O
homologs	O
suggest	O
that	O
it	O
is	O
indeed	O
N1	O
-	O
methyl	O
-	O
N3	O
-	O
acp	O
-	O
pseudouridine	O
.	O
	O
(	O
C	O
)	O
Binding	O
of	O
SsTsr3	O
to	O
RNA	O
.	O
	O
5	O
′-	O
fluoresceine	O
labeled	O
RNA	O
oligonucleotides	O
corresponding	O
either	O
to	O
the	O
native	O
(	O
20mer	O
–	O
see	O
inset	O
)	O
or	O
a	O
stabilized	O
(	O
20mer_GC	O
-	O
inset	O
)	O
helix	O
31	O
of	O
the	O
small	O
ribosomal	O
subunit	O
rRNA	O
from	O
S	O
.	O
solfataricus	O
were	O
titrated	O
with	O
increasing	O
amounts	O
of	O
SsTsr3	O
and	O
the	O
changes	O
in	O
the	O
fluoresceine	O
fluorescence	O
anisotropy	O
were	O
measured	O
and	O
fitted	O
to	O
a	O
binding	O
curve	O
(	O
20mer	O
–	O
red	O
,	O
20mer_GC	O
–	O
blue	O
).	O
	O
Oligo	O
-	O
U9	O
-	O
RNA	O
was	O
used	O
for	O
comparison	O
(	O
black	O
).	O
	O
The	O
20mer_GC	O
RNA	O
was	O
also	O
titrated	O
with	O
SsTsr3	O
in	O
the	O
presence	O
of	O
2	O
mM	O
SAM	O
(	O
purple	O
).	O
(	O
D	O
)	O
Mutants	O
of	O
ScTsr3	O
R60	O
,	O
K65	O
or	O
R131	O
(	O
equivalent	O
to	O
K17	O
,	O
K22	O
and	O
R91	O
in	O
VdTsr3	O
)	O
expressed	O
in	O
Δtsr3	B-mutant
yeast	O
cells	O
show	O
a	O
primer	O
extension	O
stop	O
comparable	O
to	O
the	O
wild	O
type	O
.	O
	O
Combination	O
of	O
the	O
three	O
point	O
mutations	O
(	O
R60A	B-mutant
/	O
K65A	B-mutant
/	O
R131A	B-mutant
)	O
leads	O
to	O
a	O
strongly	O
reduced	O
acp	O
modification	O
of	O
18S	O
rRNA	O
.	O
	O
SsTsr3	O
in	O
the	O
apo	O
state	O
bound	O
a	O
20mer	O
RNA	O
corresponding	O
to	O
helix	O
31	O
of	O
S	O
.	O
solfataricus	O
16S	O
rRNA	O
(	O
Figure	O
6B	O
)	O
with	O
a	O
KD	O
of	O
1	O
.	O
9	O
μM	O
and	O
to	O
a	O
version	O
of	O
this	O
hairpin	O
stabilized	O
by	O
additional	O
GC	O
base	O
pairs	O
(	O
20mer	O
-	O
GC	O
)	O
with	O
a	O
KD	O
of	O
0	O
.	O
6	O
μM	O
(	O
Figure	O
6C	O
).	O
	O
U1191	O
is	O
the	O
only	O
hypermodified	O
base	O
in	O
the	O
yeast	O
18S	O
rRNA	O
and	O
is	O
strongly	O
conserved	O
in	O
eukaryotes	O
.	O
	O
The	O
formation	O
of	O
1	O
-	O
methyl	O
-	O
3	O
-(	O
3	O
-	O
amino	O
-	O
3	O
-	O
carboxypropyl	O
)-	O
pseudouridine	O
(	O
m1acp3Ψ	O
)	O
is	O
very	O
complex	O
requiring	O
three	O
successive	O
modification	O
reactions	O
involving	O
one	O
H	O
/	O
ACA	O
snoRNP	O
(	O
snR35	O
)	O
and	O
two	O
protein	O
enzymes	O
(	O
Nep1	O
/	O
Emg1	O
and	O
Tsr3	O
).	O
	O
The	O
m1acp3Ψ	O
base	O
is	O
located	O
at	O
the	O
tip	O
of	O
helix	O
31	O
on	O
the	O
18S	O
rRNA	O
(	O
Supplementary	O
Figure	O
S1B	O
)	O
which	O
,	O
together	O
with	O
helices	O
18	O
,	O
24	O
,	O
34	O
and	O
44	O
,	O
contribute	O
to	O
building	O
the	O
decoding	O
center	O
of	O
the	O
small	O
ribosomal	O
subunit	O
.	O
	O
As	O
shown	O
here	O
TSR3	O
encodes	O
the	O
transferase	O
catalyzing	O
the	O
acp	O
modification	O
as	O
the	O
last	O
step	O
in	O
the	O
biosynthesis	O
of	O
m1acp3Ψ	O
in	O
yeast	O
and	O
human	O
cells	O
.	O
	O
Unexpectedly	O
,	O
archaeal	O
Tsr3	O
has	O
a	O
structure	O
similar	O
to	O
SPOUT	O
-	O
class	O
RNA	O
methyltransferases	O
,	O
and	O
it	O
is	O
the	O
first	O
example	O
for	O
an	O
enzyme	O
of	O
this	O
class	O
transferring	O
an	O
acp	O
group	O
,	O
due	O
to	O
a	O
modified	O
SAM	O
-	O
binding	O
pocket	O
that	O
exposes	O
the	O
acp	O
instead	O
of	O
the	O
methyl	O
group	O
of	O
SAM	O
to	O
its	O
RNA	O
substrate	O
.	O
	O
Similar	O
to	O
the	O
structurally	O
unrelated	O
Rossmann	O
-	O
fold	O
Tyw2	O
acp	O
transferase	O
,	O
the	O
SAM	O
methyl	O
group	O
of	O
Tsr3	O
is	O
bound	O
in	O
an	O
inaccessible	O
hydrophobic	O
pocket	O
whereas	O
the	O
acp	O
side	O
chain	O
becomes	O
accessible	O
for	O
a	O
nucleophilic	O
attack	O
by	O
the	O
N3	O
of	O
pseudouridine	O
.	O
	O
In	O
contrast	O
,	O
in	O
the	O
structurally	O
closely	O
related	O
RNA	O
methyltransferase	O
Trm10	O
the	O
methyl	O
group	O
of	O
the	O
cofactor	O
SAM	O
is	O
accessible	O
whereas	O
its	O
acp	O
side	O
chain	O
is	O
buried	O
inside	O
the	O
protein	O
.	O
	O
Thus	O
,	O
additional	O
examples	O
for	O
acp	O
transferase	O
enzymes	O
might	O
be	O
found	O
with	O
similarities	O
to	O
other	O
structural	O
classes	O
of	O
methyltransferases	O
.	O
	O
This	O
suggests	O
that	O
Tsr3	O
is	O
not	O
stably	O
incorporated	O
into	O
pre	O
-	O
ribosomal	O
particles	O
and	O
that	O
its	O
binding	O
to	O
the	O
nascent	O
ribosomal	O
subunit	O
possibly	O
requires	O
additional	O
interactions	O
with	O
other	O
pre	O
-	O
ribosomal	O
components	O
.	O
	O
Consistently	O
,	O
in	O
sucrose	O
gradient	O
analysis	O
,	O
Tsr3	O
was	O
found	O
in	O
low	O
-	O
molecular	O
weight	O
fractions	O
rather	O
than	O
with	O
pre	O
-	O
ribosome	O
containing	O
high	O
-	O
molecular	O
weight	O
fractions	O
.	O
	O
In	O
contrast	O
to	O
several	O
enzymes	O
that	O
catalyze	O
base	O
specific	O
modifications	O
in	O
rRNAs	O
Tsr3	O
is	O
not	O
an	O
essential	O
protein	O
.	O
	O
Typically	O
,	O
other	O
small	O
subunit	O
rRNA	O
methyltransferases	O
as	O
Dim1	O
,	O
Bud23	O
and	O
Nep1	O
/	O
Emg1	O
carry	O
dual	O
functions	O
,	O
in	O
ribosome	O
biogenesis	O
and	O
rRNA	O
modification	O
,	O
and	O
it	O
is	O
their	O
involvement	O
in	O
pre	O
-	O
RNA	O
processing	O
that	O
is	O
essential	O
rather	O
than	O
their	O
RNA	O
-	O
methylating	O
activity	O
(,	O
discussed	O
in	O
7	O
).	O
	O
In	O
contrast	O
,	O
for	O
several	O
Tsr3	O
mutants	O
(	O
SAM	O
-	O
binding	O
and	O
cysteine	O
mutations	O
)	O
we	O
found	O
a	O
systematic	O
correlation	O
between	O
the	O
loss	O
of	O
acp	O
modification	O
and	O
the	O
efficiency	O
of	O
18S	O
rRNA	O
maturation	O
.	O
	O
Apart	O
from	O
most	O
of	O
the	O
ribosomal	O
proteins	O
,	O
cytoplasmic	O
pre	O
-	O
40S	O
particles	O
contain	O
20S	O
rRNA	O
and	O
at	O
least	O
seven	O
non	O
-	O
ribosomal	O
proteins	O
including	O
the	O
D	O
-	O
site	O
endonuclease	O
Nob1	O
as	O
well	O
as	O
Tsr1	O
,	O
a	O
putative	O
GTPase	O
and	O
Rio2	O
which	O
block	O
the	O
mRNA	O
channel	O
and	O
the	O
initiator	O
tRNA	O
binding	O
site	O
,	O
respectively	O
,	O
thus	O
preventing	O
translation	O
initiation	O
.	O
	O
Finally	O
,	O
termination	O
factor	O
Rli1	O
,	O
an	O
ATPase	O
,	O
promotes	O
the	O
dissociation	O
of	O
assembly	O
factors	O
and	O
the	O
80S	O
-	O
like	O
complex	O
dissociates	O
and	O
releases	O
the	O
mature	O
40S	O
subunit	O
.	O
	O
Interestingly	O
,	O
differences	O
in	O
the	O
level	O
of	O
acp	O
modification	O
were	O
demonstrated	O
for	O
different	O
steps	O
of	O
the	O
cytoplasmic	O
pre	O
-	O
40S	O
subunit	O
maturation	O
after	O
analyzing	O
purified	O
20S	O
pre	O
-	O
rRNAs	O
using	O
different	O
purification	O
bait	O
proteins	O
.	O
	O
Early	O
cytoplasmic	O
pre	O
-	O
40S	O
subunits	O
still	O
containing	O
the	O
ribosome	O
assembly	O
factors	O
Tsr1	O
,	O
Ltv1	O
,	O
Enp1	O
and	O
Rio2	O
were	O
not	O
or	O
only	O
partially	O
acp	O
modified	O
.	O
	O
In	O
contrast	O
,	O
late	O
pre	O
-	O
40S	O
subunits	O
containing	O
Nob1	O
and	O
Rio1	O
or	O
already	O
associated	O
with	O
60S	O
subunits	O
in	O
80S	O
-	O
like	O
particles	O
showed	O
acp	O
modification	O
levels	O
comparable	O
to	O
mature	O
40S	O
subunits	O
.	O
	O
These	O
data	O
and	O
the	O
finding	O
that	O
a	O
missing	O
acp	O
modification	O
hinders	O
pre	O
-	O
20S	O
rRNA	O
processing	O
,	O
suggest	O
that	O
the	O
acp	O
modification	O
together	O
with	O
the	O
release	O
of	O
Rio2	O
promotes	O
the	O
formation	O
of	O
the	O
decoding	O
site	O
and	O
thus	O
D	O
-	O
site	O
cleavage	O
by	O
Nob1	O
.	O
	O
The	O
interrelation	O
between	O
acp	O
modification	O
and	O
Rio2	O
release	O
is	O
also	O
supported	O
by	O
CRAC	O
analysis	O
showing	O
that	O
Rio2	O
binds	O
to	O
helix	O
31	O
next	O
to	O
the	O
Ψ1191	O
residue	O
that	O
receives	O
the	O
acp	O
modification	O
.	O
	O
Therefore	O
,	O
Rio2	O
either	O
blocks	O
the	O
access	O
of	O
Tsr3	O
to	O
helix	O
31	O
,	O
and	O
acp	O
modification	O
can	O
only	O
occur	O
after	O
Rio2	O
is	O
released	O
,	O
or	O
the	O
acp	O
modification	O
of	O
m1Ψ1191	O
and	O
putative	O
subsequent	O
conformational	O
changes	O
of	O
20S	O
rRNA	O
weaken	O
the	O
binding	O
of	O
Rio2	O
to	O
helix	O
31	O
and	O
support	O
its	O
release	O
from	O
the	O
pre	O
-	O
rRNA	O
.	O
	O
In	O
summary	O
,	O
by	O
identifying	O
Tsr3	O
as	O
the	O
enzyme	O
responsible	O
for	O
introducing	O
the	O
acp	O
group	O
to	O
the	O
hypermodified	O
m1acp3Ψ	O
nucleotide	O
at	O
position	O
1191	O
(	O
yeast	O
)/	O
1248	O
(	O
humans	O
)	O
of	O
18S	O
rRNA	O
we	O
added	O
one	O
of	O
the	O
last	O
remaining	O
pieces	O
to	O
the	O
puzzle	O
of	O
eukaryotic	O
small	O
ribosomal	O
subunit	O
rRNA	O
modifications	O
.	O
	O
Structural	O
insights	O
into	O
the	O
regulatory	O
mechanism	O
of	O
the	O
Pseudomonas	O
aeruginosa	O
YfiBNR	O
system	O
	O
YfiBNR	O
is	O
a	O
recently	O
identified	O
bis	O
-(	O
3	O
’-	O
5	O
’)-	O
cyclic	O
dimeric	O
GMP	O
(	O
c	O
-	O
di	O
-	O
GMP	O
)	O
signaling	O
system	O
in	O
opportunistic	O
pathogens	O
.	O
	O
In	O
response	O
to	O
cell	O
stress	O
,	O
YfiB	O
in	O
the	O
outer	O
membrane	O
can	O
sequester	O
the	O
periplasmic	O
protein	O
YfiR	O
,	O
releasing	O
its	O
inhibition	O
of	O
YfiN	O
on	O
the	O
inner	O
membrane	O
and	O
thus	O
provoking	O
the	O
diguanylate	O
cyclase	O
activity	O
of	O
YfiN	O
to	O
induce	O
c	O
-	O
di	O
-	O
GMP	O
production	O
.	O
	O
Here	O
,	O
we	O
report	O
the	O
crystal	O
structures	O
of	O
YfiB	O
alone	O
and	O
of	O
an	O
active	O
mutant	O
YfiBL43P	B-mutant
complexed	O
with	O
YfiR	O
with	O
2	O
:	O
2	O
stoichiometry	O
.	O
	O
Structural	O
analyses	O
revealed	O
that	O
in	O
contrast	O
to	O
the	O
compact	O
conformation	O
of	O
the	O
dimeric	O
YfiB	O
alone	O
,	O
YfiBL43P	B-mutant
adopts	O
a	O
stretched	O
conformation	O
allowing	O
activated	O
YfiB	O
to	O
penetrate	O
the	O
peptidoglycan	O
(	O
PG	O
)	O
layer	O
and	O
access	O
YfiR	O
.	O
YfiBL43P	B-mutant
shows	O
a	O
more	O
compact	O
PG	O
-	O
binding	O
pocket	O
and	O
much	O
higher	O
PG	O
binding	O
affinity	O
than	O
wild	O
-	O
type	O
YfiB	O
,	O
suggesting	O
a	O
tight	O
correlation	O
between	O
PG	O
binding	O
and	O
YfiB	O
activation	O
.	O
	O
Based	O
on	O
the	O
structural	O
and	O
biochemical	O
data	O
,	O
we	O
propose	O
an	O
updated	O
regulatory	O
model	O
of	O
the	O
YfiBNR	O
system	O
.	O
	O
Bis	O
-(	O
3	O
’-	O
5	O
’)-	O
cyclic	O
dimeric	O
GMP	O
(	O
c	O
-	O
di	O
-	O
GMP	O
)	O
is	O
a	O
ubiquitous	O
second	O
messenger	O
that	O
bacteria	O
use	O
to	O
facilitate	O
behavioral	O
adaptations	O
to	O
their	O
ever	O
-	O
changing	O
environment	O
.	O
	O
Intriguingly	O
,	O
studies	O
in	O
diverse	O
species	O
have	O
revealed	O
that	O
a	O
single	O
bacterium	O
can	O
have	O
dozens	O
of	O
DGCs	O
and	O
PDEs	O
(	O
Hickman	O
et	O
al	O
.,;	O
Kirillina	O
et	O
al	O
.,;	O
Kulasakara	O
et	O
al	O
.,;	O
Tamayo	O
et	O
al	O
.,).	O
	O
In	O
Pseudomonas	O
aeruginosa	O
in	O
particular	O
,	O
42	O
genes	O
containing	O
putative	O
DGCs	O
and	O
/	O
or	O
PDEs	O
were	O
identified	O
(	O
Kulasakara	O
et	O
al	O
.,).	O
	O
The	O
functional	O
role	O
of	O
a	O
number	O
of	O
downstream	O
effectors	O
of	O
c	O
-	O
di	O
-	O
GMP	O
has	O
been	O
characterized	O
as	O
affecting	O
exopolysaccharide	O
(	O
EPS	O
)	O
production	O
,	O
transcription	O
,	O
motility	O
,	O
and	O
surface	O
attachment	O
(	O
Caly	O
et	O
al	O
.,;	O
Camilli	O
and	O
Bassler	O
,;	O
Ha	O
and	O
O	O
’	O
Toole	O
,;	O
Pesavento	O
and	O
Hengge	O
,).	O
	O
However	O
,	O
due	O
to	O
the	O
intricacy	O
of	O
c	O
-	O
di	O
-	O
GMP	O
signaling	O
networks	O
and	O
the	O
diversity	O
of	O
experimental	O
cues	O
,	O
the	O
detailed	O
mechanisms	O
by	O
which	O
these	O
signaling	O
pathways	O
specifically	O
sense	O
and	O
integrate	O
different	O
inputs	O
remain	O
largely	O
elusive	O
.	O
	O
Biofilm	O
formation	O
protects	O
pathogenic	O
bacteria	O
from	O
antibiotic	O
treatment	O
,	O
and	O
c	O
-	O
di	O
-	O
GMP	O
-	O
regulated	O
biofilm	O
formation	O
has	O
been	O
extensively	O
studied	O
in	O
P	O
.	O
aeruginosa	O
(	O
Evans	O
,;	O
Kirisits	O
et	O
al	O
.,;	O
Malone	O
,;	O
Reinhardt	O
et	O
al	O
.,).	O
	O
In	O
the	O
lungs	O
of	O
cystic	O
fibrosis	O
(	O
CF	O
)	O
patients	O
,	O
adherent	O
biofilm	O
formation	O
and	O
the	O
appearance	O
of	O
small	O
colony	O
variant	O
(	O
SCV	O
)	O
morphologies	O
of	O
P	O
.	O
aeruginosa	O
correlate	O
with	O
prolonged	O
persistence	O
of	O
infection	O
and	O
poor	O
lung	O
function	O
(	O
Govan	O
and	O
Deretic	O
,;	O
Haussler	O
et	O
al	O
.,;	O
Haussler	O
et	O
al	O
.,;	O
Parsek	O
and	O
Singh	O
,;	O
Smith	O
et	O
al	O
.,).	O
	O
Recently	O
,	O
Malone	O
and	O
coworkers	O
identified	O
the	O
tripartite	O
c	O
-	O
di	O
-	O
GMP	O
signaling	O
module	O
system	O
YfiBNR	O
(	O
also	O
known	O
as	O
AwsXRO	O
(	O
Beaumont	O
et	O
al	O
.,;	O
Giddens	O
et	O
al	O
.,)	O
or	O
Tbp	O
(	O
Ueda	O
and	O
Wood	O
,))	O
by	O
genetic	O
screening	O
for	O
mutants	O
that	O
displayed	O
SCV	O
phenotypes	O
in	O
P	O
.	O
aeruginosa	O
PAO1	O
(	O
Malone	O
et	O
al	O
.,;	O
Malone	O
et	O
al	O
.,).	O
	O
YfiB	O
is	O
an	O
OmpA	O
/	O
Pal	O
-	O
like	O
outer	O
-	O
membrane	O
lipoprotein	O
(	O
Parsons	O
et	O
al	O
.,)	O
that	O
can	O
activate	O
YfiN	O
by	O
sequestering	O
YfiR	O
(	O
Malone	O
et	O
al	O
.,)	O
in	O
an	O
unknown	O
manner	O
.	O
	O
It	O
is	O
also	O
proposed	O
that	O
the	O
sequestration	O
of	O
YfiR	O
by	O
YfiB	O
can	O
be	O
induced	O
by	O
certain	O
YfiB	O
-	O
mediated	O
cell	O
wall	O
stress	O
,	O
and	O
mutagenesis	O
studies	O
revealed	O
a	O
number	O
of	O
activation	O
residues	O
of	O
YfiB	O
that	O
were	O
located	O
in	O
close	O
proximity	O
to	O
the	O
predicted	O
first	O
helix	O
of	O
the	O
periplasmic	O
domain	O
(	O
Malone	O
et	O
al	O
.,).	O
	O
In	O
the	O
present	O
study	O
,	O
we	O
solved	O
the	O
crystal	O
structures	O
of	O
an	O
N	O
-	O
terminal	O
truncated	O
form	O
of	O
YfiB	O
(	O
34	O
–	O
168	O
)	O
and	O
YfiR	O
in	O
complex	O
with	O
an	O
active	O
mutant	O
YfiBL43P	B-mutant
.	O
	O
Most	O
recently	O
,	O
Li	O
and	O
coworkers	O
reported	O
the	O
crystal	O
structures	O
of	O
YfiB	O
(	O
27	O
–	O
168	O
)	O
alone	O
and	O
YfiRC71S	B-mutant
in	O
complex	O
with	O
YfiB	O
(	O
59	O
–	O
168	O
)	O
(	O
Li	O
et	O
al	O
.,).	O
	O
Compared	O
with	O
the	O
reported	O
complex	O
structure	O
,	O
YfiBL43P	B-mutant
in	O
our	O
YfiB	O
-	O
YfiR	O
complex	O
structure	O
has	O
additional	O
visible	O
N	O
-	O
terminal	O
residues	O
44	O
–	O
58	O
that	O
are	O
shown	O
to	O
play	O
essential	O
roles	O
in	O
YfiB	O
activation	O
and	O
biofilm	O
formation	O
.	O
	O
Therefore	O
,	O
we	O
are	O
able	O
to	O
visualize	O
the	O
detailed	O
allosteric	O
arrangement	O
of	O
the	O
N	O
-	O
terminal	O
structure	O
of	O
YfiB	O
and	O
its	O
important	O
role	O
in	O
YfiB	O
-	O
YfiR	O
interaction	O
.	O
	O
In	O
addition	O
,	O
we	O
found	O
that	O
the	O
YfiBL43P	B-mutant
shows	O
a	O
much	O
higher	O
PG	O
-	O
binding	O
affinity	O
than	O
wild	O
-	O
type	O
YfiB	O
,	O
most	O
likely	O
due	O
to	O
its	O
more	O
compact	O
PG	O
-	O
binding	O
pocket	O
.	O
	O
Moreover	O
,	O
we	O
found	O
that	O
Vitamin	O
B6	O
(	O
VB6	O
)	O
or	O
L	O
-	O
Trp	O
can	O
bind	O
YfiR	O
with	O
an	O
affinity	O
in	O
the	O
ten	O
millimolar	O
range	O
.	O
	O
Together	O
with	O
functional	O
data	O
,	O
these	O
results	O
provide	O
new	O
mechanistic	O
insights	O
into	O
how	O
activated	O
YfiB	O
sequesters	O
YfiR	O
and	O
releases	O
the	O
suppression	O
of	O
YfiN	O
.	O
These	O
findings	O
may	O
facilitate	O
the	O
development	O
and	O
optimization	O
of	O
anti	O
-	O
biofilm	O
drugs	O
for	O
the	O
treatment	O
of	O
chronic	O
infections	O
.	O
	O
Overall	O
structure	O
of	O
YfiB	O
	O
We	O
obtained	O
two	O
crystal	O
forms	O
of	O
YfiB	O
(	O
residues	O
34	O
–	O
168	O
,	O
lacking	O
the	O
signal	O
peptide	O
from	O
residues	O
1	O
–	O
26	O
and	O
periplasmic	O
residues	O
27	O
–	O
33	O
),	O
crystal	O
forms	O
I	O
and	O
II	O
,	O
belonging	O
to	O
space	O
groups	O
P21	O
and	O
P41	O
,	O
respectively	O
.	O
	O
Overall	O
structure	O
of	O
YfiB	O
.	O
(	O
A	O
)	O
The	O
overall	O
structure	O
of	O
the	O
YfiB	O
monomer	O
.	O
(	O
B	O
)	O
A	O
topology	O
diagram	O
of	O
the	O
YfiB	O
monomer	O
.	O
(	O
C	O
and	O
D	O
)	O
The	O
analytical	O
ultracentrifugation	O
experiment	O
results	O
for	O
the	O
wild	O
-	O
type	O
YfiB	O
and	O
YfiBL43P	B-mutant
	O
Two	O
dimeric	O
types	O
of	O
YfiB	O
dimer	O
.	O
(	O
A	O
–	O
C	O
)	O
The	O
“	O
head	O
to	O
head	O
”	O
dimer	O
.	O
	O
(	O
A	O
)	O
and	O
(	O
E	O
)	O
indicate	O
the	O
front	O
views	O
of	O
the	O
two	O
dimers	O
,	O
(	O
B	O
)	O
and	O
(	O
F	O
)	O
indicate	O
the	O
top	O
views	O
of	O
the	O
two	O
dimers	O
,	O
and	O
(	O
C	O
)	O
and	O
(	O
D	O
)	O
indicate	O
the	O
details	O
of	O
the	O
two	O
dimeric	O
interfaces	O
	O
In	O
addition	O
,	O
there	O
is	O
a	O
short	O
helix	O
turn	O
connecting	O
the	O
β4	O
strand	O
and	O
α4	O
helix	O
(	O
Fig	O
.	O
1A	O
and	O
1B	O
).	O
	O
Each	O
crystal	O
form	O
contains	O
three	O
different	O
dimeric	O
types	O
of	O
YfiB	O
,	O
two	O
of	O
which	O
are	O
present	O
in	O
both	O
,	O
suggesting	O
that	O
the	O
rest	O
of	O
the	O
dimeric	O
types	O
may	O
result	O
from	O
crystal	O
packing	O
.	O
	O
Here	O
,	O
we	O
refer	O
to	O
the	O
two	O
dimeric	O
types	O
as	O
“	O
head	O
to	O
head	O
”	O
and	O
“	O
back	O
to	O
back	O
”	O
according	O
to	O
the	O
interacting	O
mode	O
(	O
Fig	O
.	O
2A	O
and	O
2E	O
),	O
with	O
the	O
total	O
buried	O
surface	O
areas	O
being	O
316	O
.	O
8	O
Å2	O
and	O
554	O
.	O
3	O
Å2	O
,	O
respectively	O
.	O
	O
The	O
“	O
head	O
to	O
head	O
”	O
dimer	O
exhibits	O
a	O
clamp	O
shape	O
.	O
	O
The	O
dimeric	O
interaction	O
is	O
mainly	O
hydrophilic	O
,	O
occurring	O
among	O
the	O
main	O
-	O
chain	O
and	O
side	O
-	O
chain	O
atoms	O
of	O
N68	O
,	O
L69	O
,	O
D70	O
and	O
R71	O
on	O
the	O
α2	O
-	O
α3	O
loops	O
and	O
R116	O
and	O
S120	O
on	O
the	O
α4	O
helices	O
of	O
both	O
molecules	O
,	O
resulting	O
in	O
a	O
complex	O
hydrogen	O
bond	O
network	O
(	O
Fig	O
.	O
2D	O
–	O
F	O
).	O
	O
The	O
YfiB	O
-	O
YfiR	O
interaction	O
	O
The	O
YfiBL43P	B-mutant
molecules	O
are	O
shown	O
in	O
cyan	O
and	O
yellow	O
.	O
	O
The	O
YfiR	O
molecules	O
are	O
shown	O
in	O
green	O
and	O
magenta	O
.	O
	O
To	O
illustrate	O
the	O
differences	O
between	O
apo	O
YfiB	O
and	O
YfiR	O
-	O
bound	O
YfiBL43P	B-mutant
,	O
the	O
apo	O
YfiB	O
is	O
shown	O
in	O
pink	O
,	O
except	O
residues	O
34	O
–	O
70	O
are	O
shown	O
in	O
red	O
,	O
whereas	O
the	O
YfiR	O
-	O
bound	O
YfiBL43P	B-mutant
is	O
shown	O
in	O
cyan	O
,	O
except	O
residues	O
44	O
–	O
70	O
are	O
shown	O
in	O
blue	O
.	O
(	O
C	O
)	O
Close	O
-	O
up	O
view	O
of	O
the	O
differences	O
between	O
apo	O
YfiB	O
and	O
YfiR	O
-	O
bound	O
YfiBL43P	B-mutant
.	O
	O
The	O
key	O
residues	O
in	O
apo	O
YfiB	O
are	O
shown	O
in	O
red	O
and	O
those	O
in	O
YfiBL43P	B-mutant
are	O
shown	O
in	O
blue	O
.	O
(	O
D	O
)	O
Close	O
-	O
up	O
views	O
showing	O
interactions	O
in	O
regions	O
I	O
and	O
II	O
.	O
	O
YfiBL43P	B-mutant
and	O
YfiR	O
are	O
shown	O
in	O
cyan	O
and	O
green	O
,	O
respectively	O
.	O
(	O
E	O
and	O
F	O
)	O
The	O
conserved	O
surface	O
in	O
YfiR	O
contributes	O
to	O
the	O
interaction	O
with	O
YfiB	O
.	O
(	O
G	O
)	O
The	O
residues	O
of	O
YfiR	O
responsible	O
for	O
interacting	O
with	O
YfiB	O
are	O
shown	O
in	O
green	O
sticks	O
,	O
and	O
the	O
proposed	O
YfiN	O
-	O
interacting	O
residues	O
are	O
shown	O
in	O
yellow	O
sticks	O
.	O
	O
The	O
red	O
sticks	O
,	O
which	O
represent	O
the	O
YfiB	O
-	O
interacting	O
residues	O
,	O
are	O
also	O
responsible	O
for	O
the	O
proposed	O
interactions	O
with	O
YfiN	O
	O
To	O
gain	O
structural	O
insights	O
into	O
the	O
YfiB	O
-	O
YfiR	O
interaction	O
,	O
we	O
co	O
-	O
expressed	O
YfiB	O
(	O
residues	O
34	O
–	O
168	O
)	O
and	O
YfiR	O
(	O
residues	O
35	O
–	O
190	O
,	O
lacking	O
the	O
signal	O
peptide	O
),	O
but	O
failed	O
to	O
obtain	O
the	O
complex	O
,	O
in	O
accordance	O
with	O
a	O
previous	O
report	O
in	O
which	O
no	O
stable	O
complex	O
of	O
YfiB	O
-	O
YfiR	O
was	O
observed	O
(	O
Malone	O
et	O
al	O
.,).	O
	O
It	O
is	O
likely	O
that	O
these	O
residues	O
may	O
be	O
involved	O
in	O
the	O
conformational	O
changes	O
of	O
YfiB	O
that	O
are	O
related	O
to	O
YfiR	O
sequestration	O
(	O
Fig	O
.	O
3C	O
).	O
	O
Therefore	O
,	O
we	O
constructed	O
two	O
such	O
single	O
mutants	O
of	O
YfiB	O
(	O
YfiBL43P	B-mutant
and	O
YfiBF48S	B-mutant
).	O
	O
As	O
expected	O
,	O
both	O
mutants	O
form	O
a	O
stable	O
complex	O
with	O
YfiR	O
.	O
Finally	O
,	O
we	O
crystalized	O
YfiR	O
in	O
complex	O
with	O
the	O
YfiBL43P	B-mutant
mutant	O
and	O
solved	O
the	O
structure	O
at	O
1	O
.	O
78	O
Å	O
resolution	O
by	O
molecular	O
replacement	O
using	O
YfiR	O
and	O
YfiB	O
as	O
models	O
.	O
	O
The	O
YfiR	O
dimer	O
in	O
the	O
complex	O
is	O
identical	O
to	O
the	O
non	O
-	O
oxidized	O
YfiR	O
dimer	O
alone	O
(	O
Yang	O
et	O
al	O
.,),	O
with	O
only	O
Cys145	O
-	O
Cys152	O
of	O
the	O
two	O
disulfide	O
bonds	O
well	O
formed	O
,	O
suggesting	O
Cys71	O
-	O
Cys110	O
disulfide	O
bond	O
formation	O
is	O
not	O
essential	O
for	O
forming	O
YfiB	O
-	O
YfiR	O
complex	O
.	O
	O
The	O
N	O
-	O
terminal	O
structural	O
conformation	O
of	O
YfiBL43P	B-mutant
,	O
from	O
the	O
foremost	O
N	O
-	O
terminus	O
to	O
residue	O
D70	O
,	O
is	O
significantly	O
altered	O
compared	O
with	O
that	O
of	O
the	O
apo	O
YfiB	O
.	O
The	O
majority	O
of	O
the	O
α1	O
helix	O
(	O
residues	O
34	O
–	O
43	O
)	O
is	O
invisible	O
on	O
the	O
electron	O
density	O
map	O
,	O
and	O
the	O
α2	O
helix	O
and	O
β1	O
and	O
β2	O
strands	O
are	O
rearranged	O
to	O
form	O
a	O
long	O
loop	O
containing	O
two	O
short	O
α	O
-	O
helix	O
turns	O
(	O
Fig	O
.	O
3B	O
and	O
3C	O
),	O
thus	O
embracing	O
the	O
YfiR	O
dimer	O
.	O
	O
The	O
observed	O
changes	O
in	O
conformation	O
of	O
YfiB	O
and	O
the	O
results	O
of	O
mutagenesis	O
suggest	O
a	O
mechanism	O
by	O
which	O
YfiB	O
sequesters	O
YfiR	O
.	O
	O
The	O
YfiB	O
-	O
YfiR	O
interface	O
can	O
be	O
divided	O
into	O
two	O
regions	O
(	O
Fig	O
.	O
3A	O
and	O
3D	O
).	O
	O
Region	O
I	O
is	O
formed	O
by	O
numerous	O
main	O
-	O
chain	O
and	O
side	O
-	O
chain	O
hydrophilic	O
interactions	O
between	O
residues	O
E45	O
,	O
G47	O
and	O
E53	O
from	O
the	O
N	O
-	O
terminal	O
extended	O
loop	O
of	O
YfiB	O
and	O
residues	O
S57	O
,	O
R60	O
,	O
A89	O
and	O
H177	O
from	O
YfiR	O
(	O
Fig	O
.	O
3D	O
-	O
I	O
(	O
i	O
)).	O
	O
In	O
region	O
II	O
,	O
the	O
side	O
chains	O
of	O
R96	O
,	O
E98	O
and	O
E157	O
from	O
YfiB	O
interact	O
with	O
the	O
side	O
chains	O
of	O
E163	O
,	O
S146	O
and	O
R171	O
from	O
YfiR	O
,	O
respectively	O
.	O
	O
Additionally	O
,	O
the	O
main	O
chains	O
of	O
I163	O
and	O
V165	O
from	O
YfiB	O
form	O
hydrogen	O
bonds	O
with	O
the	O
main	O
chains	O
of	O
L166	O
and	O
A164	O
from	O
YfiR	O
,	O
respectively	O
,	O
and	O
the	O
main	O
chain	O
of	O
P166	O
from	O
YfiB	O
interacts	O
with	O
the	O
side	O
chain	O
of	O
R185	O
from	O
YfiR	O
(	O
Fig	O
.	O
3D	O
-	O
II	O
).	O
	O
Based	O
on	O
the	O
observations	O
that	O
two	O
separated	O
YfiBL43P	B-mutant
molecules	O
form	O
a	O
2	O
:	O
2	O
complex	O
structure	O
with	O
YfiR	O
dimer	O
,	O
we	O
performed	O
an	O
analytical	O
ultracentrifugation	O
experiment	O
to	O
check	O
the	O
oligomeric	O
states	O
of	O
wild	O
-	O
type	O
YfiB	O
and	O
YfiBL43P	B-mutant
.	O
	O
The	O
results	O
showed	O
that	O
wild	O
-	O
type	O
YfiB	O
exists	O
in	O
both	O
monomeric	O
and	O
dimeric	O
states	O
in	O
solution	O
,	O
while	O
YfiBL43P	B-mutant
primarily	O
adopts	O
the	O
monomer	O
state	O
in	O
solution	O
(	O
Fig	O
.	O
1C	O
–	O
D	O
).	O
	O
For	O
simplicity	O
,	O
we	O
only	O
discuss	O
the	O
“	O
head	O
to	O
head	O
”	O
dimer	O
in	O
the	O
following	O
text	O
.	O
	O
The	O
PG	O
-	O
binding	O
site	O
in	O
YfiB	O
.	O
(	O
A	O
)	O
Structural	O
superposition	O
of	O
the	O
PG	O
-	O
binding	O
sites	O
of	O
the	O
H	O
.	O
influenzae	O
Pal	O
/	O
PG	O
-	O
P	O
complex	O
and	O
YfiR	O
-	O
bound	O
YfiBL43P	B-mutant
complexed	O
with	O
sulfate	O
ions	O
.	O
	O
(	O
B	O
)	O
Close	O
-	O
up	O
view	O
showing	O
the	O
key	O
residues	O
of	O
Pal	O
interacting	O
with	O
the	O
m	O
-	O
Dap5	O
ε	O
-	O
carboxylate	O
group	O
of	O
PG	O
-	O
P	O
.	O
Pal	O
is	O
shown	O
in	O
wheat	O
and	O
PG	O
-	O
P	O
is	O
in	O
magenta	O
.	O
	O
(	O
C	O
)	O
Close	O
-	O
up	O
view	O
showing	O
the	O
key	O
residues	O
of	O
YfiR	O
-	O
bound	O
YfiBL43P	B-mutant
interacting	O
with	O
a	O
sulfate	O
ion	O
.	O
	O
YfiR	O
-	O
bound	O
YfiBL43P	B-mutant
is	O
shown	O
in	O
cyan	O
;	O
the	O
sulfate	O
ion	O
,	O
in	O
green	O
;	O
and	O
the	O
water	O
molecule	O
,	O
in	O
yellow	O
.	O
(	O
D	O
)	O
Structural	O
superposition	O
of	O
the	O
PG	O
-	O
binding	O
sites	O
of	O
apo	O
YfiB	O
and	O
YfiR	O
-	O
bound	O
YfiBL43P	B-mutant
,	O
the	O
key	O
residues	O
are	O
shown	O
in	O
stick	O
.	O
	O
Apo	O
YfiB	O
is	O
shown	O
in	O
yellow	O
and	O
YfiR	O
-	O
bound	O
YfiBL43P	B-mutant
in	O
cyan	O
.	O
(	O
E	O
and	O
F	O
)	O
MST	O
data	O
and	O
analysis	O
for	O
binding	O
affinities	O
of	O
(	O
E	O
)	O
YfiB	O
wild	O
-	O
type	O
and	O
(	O
F	O
)	O
YfiBL43P	B-mutant
with	O
PG	O
.	O
(	O
G	O
)	O
The	O
sequence	O
alignment	O
of	O
P	O
.	O
aeruginosa	O
and	O
E	O
.	O
coli	O
sources	O
of	O
YfiB	O
,	O
Pal	O
and	O
the	O
periplasmic	O
domain	O
of	O
OmpA	O
	O
PG	O
-	O
associated	O
lipoprotein	O
(	O
Pal	O
)	O
is	O
highly	O
conserved	O
in	O
Gram	O
-	O
negative	O
bacteria	O
and	O
anchors	O
to	O
the	O
outer	O
membrane	O
through	O
an	O
N	O
-	O
terminal	O
lipid	O
attachment	O
and	O
to	O
PG	O
layer	O
through	O
its	O
periplasmic	O
domain	O
,	O
which	O
is	O
implicated	O
in	O
maintaining	O
outer	O
membrane	O
integrity	O
.	O
	O
Previous	O
homology	O
modeling	O
studies	O
suggested	O
that	O
YfiB	O
contains	O
a	O
Pal	O
-	O
like	O
PG	O
-	O
binding	O
site	O
(	O
Parsons	O
et	O
al	O
.,),	O
and	O
the	O
mutation	O
of	O
two	O
residues	O
at	O
this	O
site	O
,	O
D102	O
and	O
G105	O
,	O
reduces	O
the	O
ability	O
for	O
biofilm	O
formation	O
and	O
surface	O
attachment	O
(	O
Malone	O
et	O
al	O
.,).	O
	O
Structural	O
superposition	O
between	O
YfiBL43P	B-mutant
and	O
Haemophilus	O
influenzae	O
Pal	O
complexed	O
with	O
biosynthetic	O
peptidoglycan	O
precursor	O
(	O
PG	O
-	O
P	O
),	O
UDP	O
-	O
N	O
-	O
acetylmuramyl	O
-	O
L	O
-	O
Ala	O
-	O
α	O
-	O
D	O
-	O
Glu	O
-	O
m	O
-	O
Dap	O
-	O
D	O
-	O
Ala	O
-	O
D	O
-	O
Ala	O
(	O
m	O
-	O
Dap	O
is	O
meso	O
-	O
diaminopimelate	O
)	O
(	O
PDB	O
code	O
:	O
2aiz	O
)	O
(	O
Parsons	O
et	O
al	O
.,),	O
revealed	O
that	O
the	O
sulfate	O
ion	O
is	O
located	O
at	O
the	O
position	O
of	O
the	O
m	O
-	O
Dap5	O
ϵ	O
-	O
carboxylate	O
group	O
in	O
the	O
Pal	O
/	O
PG	O
-	O
P	O
complex	O
(	O
Fig	O
.	O
4A	O
).	O
	O
Similarly	O
,	O
in	O
the	O
YfiR	O
-	O
bound	O
YfiBL43P	B-mutant
structure	O
,	O
the	O
sulfate	O
ion	O
interacts	O
with	O
the	O
side	O
-	O
chain	O
atoms	O
of	O
D102	O
(	O
corresponding	O
to	O
D71	O
in	O
Pal	O
)	O
and	O
R117	O
(	O
corresponding	O
to	O
R86	O
in	O
Pal	O
)	O
and	O
the	O
main	O
-	O
chain	O
amide	O
of	O
N68	O
(	O
corresponding	O
to	O
D37	O
in	O
Pal	O
).	O
	O
In	O
addition	O
,	O
sequence	O
alignment	O
of	O
YfiB	O
with	O
Pal	O
and	O
the	O
periplasmic	O
domain	O
of	O
OmpA	O
(	O
proteins	O
containing	O
PG	O
-	O
binding	O
site	O
)	O
showed	O
that	O
N68	O
and	O
D102	O
are	O
highly	O
conserved	O
(	O
Fig	O
.	O
4G	O
,	O
blue	O
stars	O
),	O
suggesting	O
that	O
these	O
residues	O
contribute	O
to	O
the	O
PG	O
-	O
binding	O
ability	O
of	O
YfiB	O
.	O
	O
Interestingly	O
,	O
superposition	O
of	O
apo	O
YfiB	O
with	O
YfiR	O
-	O
bound	O
YfiBL43P	B-mutant
revealed	O
that	O
the	O
PG	O
-	O
binding	O
region	O
is	O
largely	O
altered	O
mainly	O
due	O
to	O
different	O
conformation	O
of	O
the	O
N68	O
containing	O
loop	O
.	O
	O
Compared	O
to	O
YfiBL43P	B-mutant
,	O
the	O
N68	O
-	O
containing	O
loop	O
of	O
the	O
apo	O
YfiB	O
flips	O
away	O
about	O
7	O
Å	O
,	O
and	O
D102	O
and	O
R117	O
swing	O
slightly	O
outward	O
;	O
thus	O
,	O
the	O
PG	O
-	O
binding	O
pocket	O
is	O
enlarged	O
with	O
no	O
sulfate	O
ion	O
or	O
water	O
bound	O
(	O
Fig	O
.	O
4D	O
).	O
	O
Therefore	O
,	O
we	O
proposed	O
that	O
the	O
PG	O
-	O
binding	O
ability	O
of	O
inactive	O
YfiB	O
might	O
be	O
weaker	O
than	O
that	O
of	O
active	O
YfiB	O
.	O
To	O
validate	O
this	O
,	O
we	O
performed	O
a	O
microscale	O
thermophoresis	O
(	O
MST	O
)	O
assay	O
to	O
measure	O
the	O
binding	O
affinities	O
of	O
PG	O
to	O
wild	O
-	O
type	O
YfiB	O
and	O
YfiBL43P	B-mutant
,	O
respectively	O
.	O
	O
As	O
the	O
experiment	O
is	O
performed	O
in	O
the	O
absence	O
of	O
YfiR	O
,	O
it	O
suggests	O
that	O
an	O
increase	O
in	O
the	O
PG	O
-	O
binding	O
affinity	O
of	O
YfiB	O
is	O
not	O
a	O
result	O
of	O
YfiB	O
-	O
YfiR	O
interaction	O
and	O
is	O
highly	O
coupled	O
to	O
the	O
activation	O
of	O
YfiB	O
characterized	O
by	O
a	O
stretched	O
N	O
-	O
terminal	O
conformation	O
.	O
	O
The	O
conserved	O
surface	O
in	O
YfiR	O
is	O
functional	O
for	O
binding	O
YfiB	O
and	O
YfiN	O
	O
Interestingly	O
,	O
the	O
majority	O
of	O
this	O
conserved	O
surface	O
contributes	O
to	O
the	O
interaction	O
with	O
YfiB	O
(	O
Fig	O
.	O
3E	O
and	O
3F	O
).	O
	O
Malone	O
JG	O
et	O
al	O
.	O
have	O
reported	O
that	O
F151	O
,	O
E163	O
,	O
I169	O
and	O
Q187	O
,	O
located	O
near	O
the	O
C	O
-	O
terminus	O
of	O
YfiR	O
,	O
comprise	O
a	O
putative	O
YfiN	O
binding	O
site	O
(	O
Malone	O
et	O
al	O
.,).	O
	O
Interestingly	O
,	O
these	O
residues	O
are	O
part	O
of	O
the	O
conserved	O
surface	O
of	O
YfiR	O
(	O
Fig	O
.	O
3G	O
).	O
	O
F151	O
,	O
E163	O
and	O
I169	O
form	O
a	O
hydrophobic	O
core	O
while	O
,	O
Q187	O
is	O
located	O
at	O
the	O
end	O
of	O
the	O
α6	O
helix	O
.	O
	O
Collectively	O
,	O
a	O
part	O
of	O
the	O
YfiB	O
-	O
YfiR	O
interface	O
overlaps	O
with	O
the	O
proposed	O
YfiR	O
-	O
YfiN	O
interface	O
,	O
suggesting	O
alteration	O
in	O
the	O
association	O
-	O
disassociation	O
equilibrium	O
of	O
YfiR	O
-	O
YfiN	O
and	O
hence	O
the	O
ability	O
of	O
YfiB	O
to	O
sequester	O
YfiR	O
.	O
	O
YfiR	O
binds	O
small	O
molecules	O
	O
Previous	O
studies	O
indicated	O
that	O
YfiR	O
constitutes	O
a	O
YfiB	O
-	O
independent	O
sensing	O
device	O
that	O
can	O
activate	O
YfiN	O
in	O
response	O
to	O
the	O
redox	O
status	O
of	O
the	O
periplasm	O
,	O
and	O
we	O
have	O
reported	O
YfiR	O
structures	O
in	O
both	O
the	O
non	O
-	O
oxidized	O
and	O
the	O
oxidized	O
states	O
earlier	O
,	O
revealing	O
that	O
the	O
Cys145	O
-	O
Cys152	O
disulfide	O
bond	O
plays	O
an	O
essential	O
role	O
in	O
maintaining	O
the	O
correct	O
folding	O
of	O
YfiR	O
(	O
Yang	O
et	O
al	O
.,).	O
	O
However	O
,	O
whether	O
YfiR	O
is	O
involved	O
in	O
other	O
regulatory	O
mechanisms	O
is	O
still	O
an	O
open	O
question	O
.	O
	O
Overall	O
Structures	O
of	O
VB6	O
-	O
bound	O
and	O
Trp	O
-	O
bound	O
YfiR	O
.	O
(	O
A	O
)	O
Superposition	O
of	O
the	O
overall	O
structures	O
of	O
VB6	O
-	O
bound	O
and	O
Trp	O
-	O
bound	O
YfiR	O
.	O
(	O
B	O
)	O
Close	O
-	O
up	O
views	O
showing	O
the	O
key	O
residues	O
of	O
YfiR	O
that	O
bind	O
VB6	O
and	O
L	O
-	O
Trp	O
.	O
	O
The	O
electron	O
densities	O
of	O
VB6	O
and	O
Trp	O
are	O
countered	O
at	O
3	O
.	O
0σ	O
and	O
2	O
.	O
3σ	O
,	O
respectively	O
,	O
in	O
|	O
Fo	O
|-|	O
Fc	O
|	O
maps	O
.	O
(	O
C	O
)	O
Superposition	O
of	O
the	O
hydrophobic	O
pocket	O
of	O
YfiR	O
with	O
VB6	O
,	O
L	O
-	O
Trp	O
and	O
F57	O
of	O
YfiB	O
	O
For	O
this	O
purpose	O
,	O
we	O
co	O
-	O
crystallized	O
YfiR	O
or	O
soaked	O
YfiR	O
crystals	O
with	O
different	O
small	O
molecules	O
,	O
including	O
L	O
-	O
Trp	O
and	O
B	O
-	O
group	O
vitamins	O
.	O
	O
Fortunately	O
,	O
we	O
found	O
obvious	O
small	O
-	O
molecule	O
density	O
in	O
the	O
VB6	O
-	O
bound	O
and	O
Trp	O
-	O
bound	O
YfiR	O
crystal	O
structures	O
(	O
Fig	O
.	O
5A	O
and	O
5B	O
),	O
and	O
in	O
both	O
structures	O
,	O
the	O
YfiR	O
dimers	O
resemble	O
the	O
oxidized	O
YfiR	O
structure	O
in	O
which	O
both	O
two	O
disulfide	O
bonds	O
are	O
well	O
formed	O
(	O
Yang	O
et	O
al	O
.,).	O
	O
Functional	O
analysis	O
of	O
VB6	O
and	O
L	O
-	O
Trp	O
.	O
(	O
A	O
and	O
B	O
)	O
The	O
effect	O
of	O
increasing	O
concentrations	O
of	O
VB6	O
or	O
L	O
-	O
Trp	O
on	O
YfiBL43P	B-mutant
-	O
induced	O
attachment	O
(	O
bars	O
).	O
	O
(	O
C	O
and	O
D	O
)	O
BIAcore	O
data	O
and	O
analysis	O
for	O
binding	O
affinities	O
of	O
(	O
C	O
)	O
VB6	O
and	O
(	O
D	O
)	O
L	O
-	O
Trp	O
with	O
YfiR	O
.	O
(	O
E	O
–	O
G	O
)	O
ITC	O
data	O
and	O
analysis	O
for	O
titration	O
of	O
(	O
E	O
)	O
YfiB	O
wild	O
-	O
type	O
,	O
(	O
F	O
)	O
YfiBL43P	O
,	O
and	O
(	O
G	O
)	O
YfiBL43P	B-mutant
/	O
F57A	B-mutant
into	O
YfiR	O
	O
Interestingly	O
,	O
VB6	O
and	O
L	O
-	O
Trp	O
were	O
found	O
to	O
occupy	O
the	O
same	O
hydrophobic	O
pocket	O
,	O
formed	O
by	O
L166	O
/	O
I169	O
/	O
V176	O
/	O
P178	O
/	O
L181	O
of	O
YfiR	O
,	O
which	O
is	O
also	O
a	O
binding	O
pocket	O
for	O
F57	O
of	O
YfiB	O
,	O
as	O
observed	O
in	O
the	O
YfiB	O
-	O
YfiR	O
complex	O
(	O
Fig	O
.	O
5C	O
).	O
	O
The	O
results	O
showed	O
Kd	O
values	O
of	O
1	O
.	O
4	O
×	O
10	O
−	O
7	O
mol	O
/	O
L	O
and	O
5	O
.	O
3	O
×	O
10	O
−	O
7	O
mol	O
/	O
L	O
for	O
YfiBL43P	B-mutant
and	O
YfiBL43P	B-mutant
/	O
F57A	B-mutant
,	O
respectively	O
,	O
revealing	O
that	O
the	O
YfiBL43P	B-mutant
/	O
F57A	B-mutant
mutant	O
caused	O
a	O
3	O
.	O
8	O
-	O
fold	O
reduction	O
in	O
the	O
binding	O
affinity	O
compared	O
with	O
the	O
YfiBL43P	B-mutant
mutant	O
(	O
Fig	O
.	O
6F	O
and	O
6G	O
).	O
	O
In	O
parallel	O
,	O
to	O
better	O
understand	O
the	O
putative	O
functional	O
role	O
of	O
VB6	O
and	O
L	O
-	O
Trp	O
,	O
yfiB	O
was	O
deleted	O
in	O
a	O
PAO1	O
wild	O
-	O
type	O
strain	O
,	O
and	O
a	O
construct	O
expressing	O
the	O
YfiBL43P	B-mutant
mutant	O
was	O
transformed	O
into	O
the	O
PAO1	O
ΔyfiB	B-mutant
strain	O
to	O
trigger	O
YfiBL43P	B-mutant
-	O
induced	O
biofilm	O
formation	O
.	O
	O
Growth	O
and	O
surface	O
attachment	O
assays	O
were	O
carried	O
out	O
for	O
the	O
yfiB	B-mutant
-	I-mutant
L43P	I-mutant
strain	O
in	O
the	O
presence	O
of	O
increasing	O
concentrations	O
of	O
VB6	O
or	O
L	O
-	O
Trp	O
.	O
	O
As	O
shown	O
in	O
Fig	O
.	O
6A	O
and	O
6B	O
,	O
the	O
over	O
-	O
expression	O
of	O
YfiBL43P	B-mutant
induced	O
strong	O
surface	O
attachment	O
and	O
much	O
slower	O
growth	O
of	O
the	O
yfiB	B-mutant
-	I-mutant
L43P	I-mutant
strain	O
,	O
and	O
as	O
expected	O
,	O
a	O
certain	O
amount	O
of	O
VB6	O
or	O
L	O
-	O
Trp	O
(	O
4	O
–	O
6	O
mmol	O
/	O
L	O
for	O
VB6	O
and	O
6	O
–	O
10	O
mmol	O
/	O
L	O
for	O
L	O
-	O
Trp	O
)	O
could	O
reduce	O
the	O
surface	O
attachment	O
.	O
	O
Interestingly	O
,	O
at	O
a	O
concentration	O
higher	O
than	O
8	O
mmol	O
/	O
L	O
,	O
VB6	O
lost	O
its	O
ability	O
to	O
inhibit	O
biofilm	O
formation	O
,	O
implying	O
that	O
the	O
VB6	O
-	O
involving	O
regulatory	O
mechanism	O
is	O
highly	O
complicated	O
and	O
remains	O
to	O
be	O
further	O
investigated	O
.	O
	O
Of	O
note	O
,	O
both	O
VB6	O
and	O
L	O
-	O
Trp	O
have	O
been	O
reported	O
to	O
correlate	O
with	O
biofilm	O
formation	O
in	O
certain	O
Gram	O
-	O
negative	O
bacteria	O
(	O
Grubman	O
et	O
al	O
.,;	O
Shimazaki	O
et	O
al	O
.,).	O
	O
In	O
Helicobacter	O
pylori	O
in	O
particular	O
,	O
VB6	O
biosynthetic	O
enzymes	O
act	O
as	O
novel	O
virulence	O
factors	O
,	O
and	O
VB6	O
is	O
required	O
for	O
full	O
motility	O
and	O
virulence	O
(	O
Grubman	O
et	O
al	O
.,).	O
	O
In	O
E	O
.	O
coli	O
,	O
mutants	O
with	O
decreased	O
tryptophan	O
synthesis	O
show	O
greater	O
biofilm	O
formation	O
,	O
and	O
matured	O
biofilm	O
is	O
degraded	O
by	O
L	O
-	O
tryptophan	O
addition	O
(	O
Shimazaki	O
et	O
al	O
.,).	O
	O
To	O
answer	O
the	O
question	O
whether	O
competition	O
of	O
VB6	O
or	O
L	O
-	O
Trp	O
for	O
the	O
YfiB	O
F57	O
-	O
binding	O
pocket	O
of	O
YfiR	O
plays	O
an	O
essential	O
role	O
in	O
inhibiting	O
biofilm	O
formation	O
,	O
we	O
measured	O
the	O
binding	O
affinities	O
of	O
VB6	O
and	O
L	O
-	O
Trp	O
for	O
YfiR	O
via	O
BIAcore	O
experiments	O
.	O
	O
The	O
results	O
showed	O
relatively	O
weak	O
Kd	O
values	O
of	O
35	O
.	O
2	O
mmol	O
/	O
L	O
and	O
76	O
.	O
9	O
mmol	O
/	O
L	O
for	O
VB6	O
and	O
L	O
-	O
Trp	O
,	O
respectively	O
(	O
Fig	O
.	O
6C	O
and	O
6D	O
).	O
	O
Based	O
on	O
our	O
results	O
,	O
we	O
concluded	O
that	O
VB6	O
or	O
L	O
-	O
Trp	O
can	O
bind	O
to	O
YfiR	O
,	O
however	O
,	O
VB6	O
or	O
L	O
-	O
Trp	O
alone	O
may	O
have	O
little	O
effects	O
in	O
interrupting	O
the	O
YfiB	O
-	O
YfiR	O
interaction	O
,	O
the	O
mechanism	O
by	O
which	O
VB6	O
or	O
L	O
-	O
Trp	O
inhibits	O
biofilm	O
formation	O
remains	O
unclear	O
and	O
requires	O
further	O
investigation	O
.	O
	O
In	O
addition	O
to	O
the	O
preceding	O
8	O
aa	O
loop	O
(	O
from	O
the	O
lipid	O
acceptor	O
Cys26	O
to	O
Gly34	O
),	O
the	O
full	O
length	O
of	O
the	O
periplasmic	O
portion	O
of	O
apo	O
YfiB	O
can	O
reach	O
approximately	O
60	O
Å	O
.	O
It	O
was	O
reported	O
that	O
the	O
distance	O
between	O
the	O
outer	O
membrane	O
and	O
the	O
cell	O
wall	O
is	O
approximately	O
50	O
Å	O
and	O
that	O
the	O
thickness	O
of	O
the	O
PG	O
layer	O
is	O
approximately	O
70	O
Å	O
(	O
Matias	O
et	O
al	O
.,).	O
	O
Thus	O
,	O
YfiB	O
alone	O
represents	O
an	O
inactive	O
form	O
that	O
may	O
only	O
partially	O
insert	O
into	O
the	O
PG	O
matrix	O
.	O
	O
By	O
contrast	O
,	O
YfiR	O
-	O
bound	O
YfiBL43P	B-mutant
(	O
residues	O
44	O
–	O
168	O
)	O
has	O
a	O
stretched	O
conformation	O
of	O
approximately	O
55	O
Å	O
in	O
length	O
.	O
	O
In	O
addition	O
to	O
the	O
17	O
preceding	O
intracellular	O
residues	O
(	O
from	O
the	O
lipid	O
acceptor	O
Cys26	O
to	O
Leu43	O
),	O
the	O
length	O
of	O
the	O
intracellular	O
portion	O
of	O
active	O
YfiB	O
may	O
extend	O
over	O
100	O
Å	O
,	O
assuming	O
a	O
fully	O
stretched	O
conformation	O
.	O
	O
Regulatory	O
model	O
of	O
the	O
YfiBNR	O
tripartite	O
system	O
.	O
	O
The	O
periplasmic	O
domain	O
of	O
YfiB	O
and	O
the	O
YfiB	O
-	O
YfiR	O
complex	O
are	O
depicted	O
according	O
to	O
the	O
crystal	O
structures	O
.	O
	O
The	O
lipid	O
acceptor	O
Cys26	O
is	O
indicated	O
as	O
blue	O
ball	O
.	O
	O
The	O
loop	O
connecting	O
Cys26	O
and	O
Gly34	O
of	O
YfiB	O
is	O
modeled	O
.	O
	O
The	O
PAS	O
domain	O
of	O
YfiN	O
is	O
shown	O
as	O
pink	O
oval	O
.	O
	O
These	O
results	O
,	O
together	O
with	O
our	O
observation	O
that	O
activated	O
YfiB	O
has	O
a	O
much	O
higher	O
cell	O
wall	O
binding	O
affinity	O
,	O
and	O
previous	O
mutagenesis	O
data	O
showing	O
that	O
(	O
1	O
)	O
both	O
PG	O
binding	O
and	O
membrane	O
anchoring	O
are	O
required	O
for	O
YfiB	O
activity	O
and	O
(	O
2	O
)	O
activating	O
mutations	O
possessing	O
an	O
altered	O
N	O
-	O
terminal	O
loop	O
length	O
are	O
dominant	O
over	O
the	O
loss	O
of	O
PG	O
binding	O
(	O
Malone	O
et	O
al	O
.,),	O
suggest	O
an	O
updated	O
regulatory	O
model	O
of	O
the	O
YfiBNR	O
system	O
(	O
Fig	O
.	O
7	O
).	O
	O
This	O
allows	O
the	O
C	O
-	O
terminal	O
portion	O
of	O
the	O
membrane	O
-	O
anchored	O
YfiB	O
to	O
reach	O
,	O
bind	O
and	O
penetrate	O
the	O
cell	O
wall	O
and	O
sequester	O
the	O
YfiR	O
dimer	O
.	O
	O
The	O
mechanism	O
by	O
which	O
activated	O
YfiB	O
relieves	O
the	O
repression	O
of	O
YfiN	O
may	O
be	O
applicable	O
to	O
the	O
YfiBNR	O
system	O
in	O
other	O
bacteria	O
and	O
to	O
analogous	O
outside	O
-	O
in	O
signaling	O
for	O
c	O
-	O
di	O
-	O
GMP	O
production	O
,	O
which	O
in	O
turn	O
may	O
be	O
relevant	O
to	O
the	O
development	O
of	O
drugs	O
that	O
can	O
circumvent	O
complicated	O
antibiotic	O
resistance	O
.	O
	O
X	O
-	O
ray	O
Crystallographic	O
Structures	O
of	O
a	O
Trimer	O
,	O
Dodecamer	O
,	O
and	O
Annular	O
Pore	O
Formed	O
by	O
an	O
Aβ17	O
–	O
36	O
β	O
-	O
Hairpin	O
	O
This	O
paper	O
presents	O
the	O
X	O
-	O
ray	O
crystallographic	O
structures	O
of	O
oligomers	O
formed	O
by	O
a	O
20	O
-	O
residue	O
peptide	O
segment	O
derived	O
from	O
Aβ	O
.	O
	O
The	O
development	O
of	O
a	O
peptide	O
in	O
which	O
Aβ17	O
–	O
36	O
is	O
stabilized	O
as	O
a	O
β	O
-	O
hairpin	O
is	O
described	O
,	O
and	O
the	O
X	O
-	O
ray	O
crystallographic	O
structures	O
of	O
oligomers	O
it	O
forms	O
are	O
reported	O
.	O
	O
Two	O
covalent	O
constraints	O
act	O
in	O
tandem	O
to	O
stabilize	O
the	O
Aβ17	O
–	O
36	O
peptide	O
in	O
a	O
hairpin	O
conformation	O
:	O
a	O
δ	O
-	O
linked	O
ornithine	O
turn	O
connecting	O
positions	O
17	O
and	O
36	O
to	O
create	O
a	O
macrocycle	O
and	O
an	O
intramolecular	O
disulfide	O
linkage	O
between	O
positions	O
24	O
and	O
29	O
.	O
	O
An	O
N	O
-	O
methyl	O
group	O
at	O
position	O
33	O
blocks	O
uncontrolled	O
aggregation	O
.	O
	O
The	O
peptide	O
readily	O
crystallizes	O
as	O
a	O
folded	O
β	O
-	O
hairpin	O
,	O
which	O
assembles	O
hierarchically	O
in	O
the	O
crystal	O
lattice	O
.	O
	O
Three	O
β	O
-	O
hairpin	O
monomers	O
assemble	O
to	O
form	O
a	O
triangular	O
trimer	O
,	O
four	O
trimers	O
assemble	O
in	O
a	O
tetrahedral	O
arrangement	O
to	O
form	O
a	O
dodecamer	O
,	O
and	O
five	O
dodecamers	O
pack	O
together	O
to	O
form	O
an	O
annular	O
pore	O
.	O
	O
This	O
hierarchical	O
assembly	O
provides	O
a	O
model	O
,	O
in	O
which	O
full	O
-	O
length	O
Aβ	O
transitions	O
from	O
an	O
unfolded	O
monomer	O
to	O
a	O
folded	O
β	O
-	O
hairpin	O
,	O
which	O
assembles	O
to	O
form	O
oligomers	O
that	O
further	O
pack	O
to	O
form	O
an	O
annular	O
pore	O
.	O
	O
High	O
-	O
resolution	O
structures	O
of	O
oligomers	O
formed	O
by	O
the	O
β	O
-	O
amyloid	O
peptide	O
Aβ	O
are	O
desperately	O
needed	O
to	O
understand	O
the	O
molecular	O
basis	O
of	O
Alzheimer	O
’	O
s	O
disease	O
and	O
ultimately	O
develop	O
preventions	O
or	O
treatments	O
.	O
	O
In	O
Alzheimer	O
’	O
s	O
disease	O
,	O
monomeric	O
Aβ	O
aggregates	O
to	O
form	O
soluble	O
low	O
molecular	O
weight	O
oligomers	O
,	O
such	O
as	O
dimers	O
,	O
trimers	O
,	O
tetramers	O
,	O
hexamers	O
,	O
nonamers	O
,	O
and	O
dodecamers	O
,	O
as	O
well	O
as	O
high	O
molecular	O
weight	O
aggregates	O
,	O
such	O
as	O
annular	O
protofibrils	O
.	O
	O
Mouse	O
models	O
for	O
Alzheimer	O
’	O
s	O
disease	O
have	O
helped	O
shape	O
our	O
current	O
understanding	O
about	O
the	O
Aβ	O
oligomerization	O
that	O
precedes	O
neurodegeneration	O
.	O
	O
A	O
56	O
kDa	O
soluble	O
oligomer	O
identified	O
by	O
SDS	O
-	O
PAGE	O
was	O
found	O
to	O
be	O
especially	O
important	O
within	O
this	O
mixture	O
.	O
	O
This	O
oligomer	O
was	O
termed	O
Aβ	O
*	O
56	O
and	O
appears	O
to	O
be	O
a	O
dodecamer	O
of	O
Aβ	O
.	O
	O
Purified	O
Aβ	O
*	O
56	O
injected	O
intercranially	O
into	O
healthy	O
rats	O
was	O
found	O
to	O
impair	O
memory	O
,	O
providing	O
evidence	O
that	O
this	O
Aβ	O
oligomer	O
may	O
cause	O
memory	O
loss	O
in	O
Alzheimer	O
’	O
s	O
disease	O
.	O
	O
Treatment	O
of	O
the	O
mixture	O
of	O
low	O
molecular	O
weight	O
oligomers	O
with	O
hexafluoroisopropanol	O
resulted	O
in	O
the	O
dissociation	O
of	O
the	O
putative	O
dodecamers	O
,	O
nonamers	O
,	O
and	O
hexamers	O
into	O
trimers	O
and	O
monomers	O
,	O
suggesting	O
that	O
trimers	O
may	O
be	O
the	O
building	O
block	O
of	O
the	O
dodecamers	O
,	O
nonamers	O
,	O
and	O
hexamers	O
.	O
	O
Recently	O
,	O
Aβ	O
trimers	O
and	O
Aβ	O
*	O
56	O
were	O
identified	O
in	O
the	O
brains	O
of	O
cognitively	O
normal	O
humans	O
and	O
were	O
found	O
to	O
increase	O
with	O
age	O
.	O
	O
APFs	O
were	O
first	O
discovered	O
in	O
vitro	O
using	O
chemically	O
synthesized	O
Aβ	O
that	O
aggregated	O
into	O
porelike	O
structures	O
that	O
could	O
be	O
observed	O
by	O
atomic	O
force	O
microscopy	O
(	O
AFM	O
)	O
and	O
transmission	O
electron	O
microscopy	O
(	O
TEM	O
).	O
	O
The	O
sizes	O
of	O
APFs	O
prepared	O
in	O
vitro	O
vary	O
among	O
different	O
studies	O
.	O
	O
Quist	O
et	O
al	O
.	O
observed	O
APFs	O
with	O
an	O
outer	O
diameter	O
of	O
16	O
nm	O
embedded	O
in	O
a	O
lipid	O
bilayer	O
.	O
	O
Although	O
the	O
APFs	O
in	O
these	O
studies	O
differ	O
in	O
size	O
,	O
they	O
share	O
a	O
similar	O
annular	O
morphology	O
and	O
appear	O
to	O
be	O
composed	O
of	O
smaller	O
oligomers	O
.	O
	O
APFs	O
have	O
also	O
been	O
observed	O
in	O
the	O
brains	O
of	O
APP23	O
transgenic	O
mice	O
by	O
immunofluorescence	O
with	O
an	O
anti	O
-	O
APF	O
antibody	O
and	O
were	O
found	O
to	O
accumulate	O
in	O
neuronal	O
processes	O
and	O
synapses	O
.	O
	O
In	O
a	O
subsequent	O
study	O
,	O
APFs	O
were	O
isolated	O
from	O
the	O
brains	O
of	O
Alzheimer	O
’	O
s	O
patients	O
by	O
immunoprecipitation	O
with	O
an	O
anti	O
-	O
APF	O
antibody	O
.	O
	O
Dimers	O
of	O
Aβ	O
have	O
also	O
been	O
isolated	O
from	O
the	O
brains	O
of	O
Alzheimer	O
’	O
s	O
patients	O
.−	O
Aβ	O
dimers	O
inhibit	O
long	O
-	O
term	O
potentiation	O
in	O
mice	O
and	O
promote	O
hyperphosphorylation	O
of	O
the	O
microtubule	O
-	O
associated	O
protein	O
tau	O
,	O
leading	O
to	O
neuritic	O
damage	O
.	O
	O
Aβ	O
dimers	O
have	O
only	O
been	O
isolated	O
from	O
human	O
or	O
transgenic	O
mouse	O
brains	O
that	O
contain	O
the	O
pathognomonic	O
fibrillar	O
Aβ	O
plaques	O
associated	O
with	O
Alzheimer	O
’	O
s	O
disease	O
.	O
	O
Furthermore	O
,	O
the	O
endogenous	O
rise	O
of	O
Aβ	O
dimers	O
in	O
the	O
brains	O
of	O
Tg2576	O
and	O
J20	O
transgenic	O
mice	O
coincides	O
with	O
the	O
deposition	O
of	O
Aβ	O
plaques	O
.	O
	O
The	O
approach	O
of	O
isolating	O
and	O
characterizing	O
Aβ	O
oligomers	O
has	O
not	O
provided	O
any	O
high	O
-	O
resolution	O
structures	O
of	O
Aβ	O
oligomers	O
.	O
	O
Techniques	O
such	O
as	O
SDS	O
-	O
PAGE	O
,	O
TEM	O
,	O
and	O
AFM	O
have	O
only	O
provided	O
information	O
about	O
the	O
molecular	O
weights	O
,	O
sizes	O
,	O
morphologies	O
,	O
and	O
stoichiometry	O
of	O
Aβ	O
oligomers	O
.	O
	O
High	O
-	O
resolution	O
structural	O
studies	O
of	O
Aβ	O
have	O
primarily	O
focused	O
on	O
Aβ	O
fibrils	O
and	O
Aβ	O
monomers	O
.	O
	O
Solid	O
-	O
state	O
NMR	O
spectroscopy	O
studies	O
of	O
Aβ	O
fibrils	O
revealed	O
that	O
Aβ	O
fibrils	O
are	O
generally	O
composed	O
of	O
extended	O
networks	O
of	O
in	O
-	O
register	O
parallel	O
β	O
-	O
sheets	O
.−	O
X	O
-	O
ray	O
crystallographic	O
studies	O
using	O
fragments	O
of	O
Aβ	O
have	O
provided	O
additional	O
information	O
about	O
how	O
Aβ	O
fibrils	O
pack	O
.	O
	O
Solution	O
-	O
phase	O
NMR	O
and	O
solid	O
-	O
state	O
NMR	O
have	O
been	O
used	O
to	O
study	O
the	O
structures	O
of	O
the	O
Aβ	O
monomers	O
within	O
oligomeric	O
assemblies	O
.−	O
A	O
major	O
finding	O
from	O
these	O
studies	O
is	O
that	O
oligomeric	O
assemblies	O
of	O
Aβ	O
are	O
primarily	O
composed	O
of	O
antiparallel	O
β	O
-	O
sheets	O
.	O
	O
The	O
structure	O
revealed	O
that	O
monomeric	O
Aβ	O
forms	O
a	O
β	O
-	O
hairpin	O
when	O
bound	O
to	O
the	O
affibody	O
.	O
	O
Locking	O
Aβ	O
into	O
a	O
β	O
-	O
hairpin	O
structure	O
resulted	O
in	O
the	O
formation	O
Aβ	O
oligomers	O
,	O
which	O
were	O
observed	O
by	O
size	O
exclusion	O
chromatography	O
(	O
SEC	O
)	O
and	O
SDS	O
-	O
PAGE	O
.	O
	O
The	O
oligomers	O
with	O
a	O
molecular	O
weight	O
of	O
∼	O
100	O
kDa	O
that	O
were	O
isolated	O
by	O
SEC	O
were	O
toxic	O
toward	O
neuronally	O
derived	O
SH	O
-	O
SY5Y	O
cells	O
.	O
	O
This	O
study	O
provides	O
evidence	O
for	O
the	O
role	O
of	O
β	O
-	O
hairpin	O
structure	O
in	O
Aβ	O
oligomerization	O
and	O
neurotoxicity	O
.	O
	O
Inspired	O
by	O
these	O
β	O
-	O
hairpin	O
structures	O
,	O
our	O
laboratory	O
developed	O
a	O
macrocyclic	O
β	O
-	O
sheet	O
peptide	O
derived	O
from	O
Aβ17	O
–	O
36	O
designed	O
to	O
mimic	O
an	O
Aβ	O
β	O
-	O
hairpin	O
and	O
reported	O
its	O
X	O
-	O
ray	O
crystallographic	O
structure	O
.	O
	O
This	O
peptide	O
(	O
peptide	B-mutant
1	I-mutant
)	O
consists	O
of	O
two	O
β	O
-	O
strands	O
comprising	O
Aβ17	O
–	O
23	O
and	O
Aβ30	O
–	O
36	O
covalently	O
linked	O
by	O
two	O
δ	O
-	O
linked	O
ornithine	O
(	O
δOrn	O
)	O
β	O
-	O
turn	O
mimics	O
.	O
	O
The	O
δOrn	O
that	O
connects	O
residues	O
D23	O
and	O
A30	O
replaces	O
the	O
Aβ24	O
–	O
29	O
loop	O
.	O
	O
The	O
δOrn	O
that	O
connects	O
residues	O
L17	O
and	O
V36	O
enforces	O
β	O
-	O
hairpin	O
structure	O
.	O
	O
We	O
incorporated	O
an	O
N	O
-	O
methyl	O
group	O
at	O
position	O
G33	O
to	O
prevent	O
uncontrolled	O
aggregation	O
and	O
precipitation	O
of	O
the	O
peptide	O
.	O
	O
To	O
improve	O
the	O
solubility	O
of	O
the	O
peptide	O
we	O
replaced	O
M35	O
with	O
the	O
hydrophilic	O
isostere	O
of	O
methionine	O
,	O
ornithine	O
(	O
α	O
-	O
linked	O
)	O
(	O
Figure	O
1B	O
).	O
	O
The	O
X	O
-	O
ray	O
crystallographic	O
structure	O
of	O
peptide	B-mutant
1	I-mutant
reveals	O
that	O
it	O
folds	O
to	O
form	O
a	O
β	O
-	O
hairpin	O
that	O
assembles	O
to	O
form	O
trimers	O
and	O
that	O
the	O
trimers	O
further	O
assemble	O
to	O
form	O
hexamers	O
and	O
dodecamers	O
.	O
	O
(	O
B	O
)	O
Chemical	O
structure	O
of	O
peptide	B-mutant
1	I-mutant
illustrating	O
Aβ17	O
–	O
23	O
and	O
Aβ30	O
–	O
36	O
,	O
M35Orn	O
,	O
the	O
N	O
-	O
methyl	O
group	O
,	O
and	O
the	O
δ	O
-	O
linked	O
ornithine	O
turns	O
.	O
(	O
C	O
)	O
Chemical	O
structure	O
of	O
peptide	B-mutant
2	I-mutant
illustrating	O
Aβ17	O
–	O
36	O
,	O
the	O
N	O
-	O
methyl	O
group	O
,	O
the	O
disulfide	O
bond	O
across	O
positions	O
24	O
and	O
29	O
,	O
and	O
the	O
δ	O
-	O
linked	O
ornithine	O
turn	O
.	O
	O
Our	O
design	O
of	O
peptide	B-mutant
1	I-mutant
omitted	O
the	O
Aβ24	O
–	O
29	O
loop	O
.	O
	O
To	O
visualize	O
the	O
Aβ24	O
–	O
29	O
loop	O
,	O
we	O
performed	O
replica	O
-	O
exchange	O
molecular	O
dynamics	O
(	O
REMD	O
)	O
simulations	O
on	O
Aβ17	O
–	O
36	O
using	O
the	O
X	O
-	O
ray	O
crystallographic	O
coordinates	O
of	O
Aβ17	O
–	O
23	O
and	O
Aβ30	O
–	O
36	O
from	O
peptide	B-mutant
1	I-mutant
.	O
	O
These	O
studies	O
provided	O
a	O
working	O
model	O
for	O
a	O
trimer	O
of	O
Aβ17	O
–	O
36	O
β	O
-	O
hairpins	O
and	O
demonstrated	O
that	O
the	O
trimer	O
should	O
be	O
capable	O
of	O
accommodating	O
the	O
Aβ24	O
–	O
29	O
loop	O
.	O
	O
In	O
the	O
current	O
study	O
we	O
set	O
out	O
to	O
restore	O
the	O
Aβ24	O
–	O
29	O
loop	O
,	O
reintroduce	O
the	O
methionine	O
residue	O
at	O
position	O
35	O
,	O
and	O
determine	O
the	O
X	O
-	O
ray	O
crystallographic	O
structures	O
of	O
oligomers	O
that	O
form	O
.	O
	O
We	O
designed	O
peptide	B-mutant
2	I-mutant
as	O
a	O
homologue	O
of	O
peptide	B-mutant
1	I-mutant
that	O
embodies	O
these	O
ideas	O
.	O
	O
Here	O
,	O
we	O
describe	O
the	O
development	O
of	O
peptide	B-mutant
2	I-mutant
and	O
report	O
the	O
X	O
-	O
ray	O
crystallographic	O
structures	O
of	O
the	O
trimer	O
,	O
dodecamer	O
,	O
and	O
annular	O
pore	O
observed	O
within	O
the	O
crystal	O
structure	O
.	O
	O
Development	O
of	O
Peptide	B-mutant
2	I-mutant
	O
We	O
developed	O
peptide	B-mutant
2	I-mutant
from	O
peptide	B-mutant
1	I-mutant
by	O
an	O
iterative	O
process	O
,	O
in	O
which	O
we	O
first	O
attempted	O
to	O
restore	O
the	O
Aβ24	O
–	O
29	O
loop	O
without	O
a	O
disulfide	O
linkage	O
.	O
	O
We	O
envisioned	O
peptide	B-mutant
3	I-mutant
as	O
a	O
homologue	O
of	O
peptide	B-mutant
1	I-mutant
with	O
the	O
Aβ24	O
–	O
29	O
loop	O
in	O
place	O
of	O
the	O
δOrn	O
that	O
connects	O
D23	O
and	O
A30	O
and	O
p	O
-	O
iodophenylalanine	O
(	O
FI	O
)	O
in	O
place	O
of	O
F19	O
.	O
	O
We	O
routinely	O
use	O
p	O
-	O
iodophenylalanine	O
to	O
determine	O
the	O
X	O
-	O
ray	O
crystallographic	O
phases	O
.	O
	O
We	O
postulate	O
that	O
the	O
loss	O
of	O
the	O
δOrn	O
constraint	O
leads	O
to	O
conformational	O
heterogeneity	O
that	O
prevents	O
peptide	B-mutant
3	I-mutant
from	O
crystallizing	O
.	O
	O
We	O
designed	O
peptide	B-mutant
4	I-mutant
to	O
embody	O
this	O
idea	O
,	O
mutating	O
Val24	O
and	O
Gly29	O
to	O
cysteine	O
and	O
forming	O
an	O
interstrand	O
disulfide	O
linkage	O
.	O
	O
Residues	O
V24	O
and	O
G29	O
form	O
a	O
non	O
-	O
hydrogen	O
-	O
bonded	O
pair	O
,	O
which	O
can	O
readily	O
accommodate	O
disulfide	O
linkages	O
in	O
antiparallel	O
β	O
-	O
sheets	O
.	O
	O
Disulfide	O
bonds	O
across	O
non	O
-	O
hydrogen	O
-	O
bonded	O
pairs	O
stabilize	O
β	O
-	O
hairpins	O
,	O
while	O
disulfide	O
bonds	O
across	O
hydrogen	O
-	O
bonded	O
pairs	O
do	O
not	O
.	O
	O
We	O
were	O
gratified	O
to	O
find	O
that	O
peptide	B-mutant
4	I-mutant
afforded	O
crystals	O
suitable	O
for	O
X	O
-	O
ray	O
crystallography	O
.	O
	O
As	O
the	O
next	O
step	O
in	O
the	O
iterative	O
process	O
,	O
we	O
determined	O
the	O
X	O
-	O
ray	O
crystallographic	O
structure	O
of	O
this	O
peptide	O
(	O
PDB	O
5HOW	O
).	O
	O
After	O
determining	O
the	O
X	O
-	O
ray	O
crystallographic	O
structure	O
of	O
peptide	B-mutant
4	I-mutant
we	O
reintroduced	O
the	O
native	O
phenylalanine	O
at	O
position	O
19	O
and	O
the	O
methionine	O
at	O
position	O
35	O
to	O
afford	O
peptide	B-mutant
2	I-mutant
.	O
	O
We	O
completed	O
the	O
iterative	O
process	O
—	O
from	O
1	O
to	O
3	O
to	O
4	O
to	O
2	O
—	O
by	O
successfully	O
determining	O
the	O
X	O
-	O
ray	O
crystallographic	O
structure	O
of	O
peptide	B-mutant
2	I-mutant
(	O
PDB	O
5HOX	O
and	O
5HOY	O
).	O
	O
The	O
following	O
sections	O
describe	O
the	O
synthesis	O
of	O
peptides	B-mutant
2	I-mutant
–	I-mutant
4	I-mutant
and	O
the	O
X	O
-	O
ray	O
crystallographic	O
structure	O
of	O
peptide	B-mutant
2	I-mutant
.	O
	O
Synthesis	O
of	O
Peptides	B-mutant
2	I-mutant
–	I-mutant
4	I-mutant
	O
We	O
synthesized	O
peptides	B-mutant
2	I-mutant
–	I-mutant
4	I-mutant
by	O
similar	O
procedures	O
to	O
those	O
we	O
have	O
developed	O
for	O
other	O
macrocyclic	O
peptides	O
.	O
	O
We	O
used	O
acid	O
-	O
stable	O
Acm	O
-	O
protected	O
cysteine	O
residues	O
at	O
positions	O
24	O
and	O
29	O
and	O
removed	O
the	O
Acm	O
groups	O
by	O
oxidation	O
with	O
I2	O
in	O
aqueous	O
acetic	O
acid	O
to	O
afford	O
the	O
disulfide	O
linkage	O
.	O
	O
Peptides	B-mutant
2	I-mutant
–	I-mutant
4	I-mutant
were	O
purified	O
by	O
RP	O
-	O
HPLC	O
.	O
	O
We	O
screened	O
crystallization	O
conditions	O
for	O
peptide	B-mutant
4	I-mutant
in	O
a	O
96	O
-	O
well	O
-	O
plate	O
format	O
using	O
three	O
different	O
Hampton	O
Research	O
crystallization	O
kits	O
(	O
Crystal	O
Screen	O
,	O
Index	O
,	O
and	O
PEG	O
/	O
Ion	O
)	O
with	O
three	O
ratios	O
of	O
peptide	O
and	O
mother	O
liquor	O
per	O
condition	O
(	O
864	O
experiments	O
).	O
	O
Peptide	B-mutant
4	I-mutant
afforded	O
crystals	O
in	O
a	O
single	O
set	O
of	O
conditions	O
containing	O
HEPES	O
buffer	O
and	O
Jeffamine	O
M	O
-	O
600	O
—	O
the	O
same	O
crystallization	O
conditions	O
that	O
afforded	O
crystals	O
of	O
peptide	B-mutant
1	I-mutant
.	O
	O
Peptide	B-mutant
2	I-mutant
also	O
afforded	O
crystals	O
in	O
these	O
conditions	O
.	O
	O
We	O
further	O
optimized	O
these	O
conditions	O
to	O
rapidly	O
(∼	O
72	O
h	O
)	O
yield	O
crystals	O
suitable	O
for	O
X	O
-	O
ray	O
crystallography	O
.	O
	O
The	O
optimized	O
conditions	O
consist	O
of	O
0	O
.	O
1	O
M	O
HEPES	O
at	O
pH	O
6	O
.	O
4	O
with	O
31	O
%	O
Jeffamine	O
M	O
-	O
600	O
for	O
peptide	B-mutant
4	I-mutant
and	O
0	O
.	O
1	O
M	O
HEPES	O
pH	O
7	O
.	O
1	O
with	O
29	O
%	O
Jeffamine	O
M	O
-	O
600	O
for	O
peptide	B-mutant
2	I-mutant
.	O
	O
Crystal	O
diffraction	O
data	O
for	O
peptide	B-mutant
2	I-mutant
were	O
also	O
collected	O
at	O
the	O
Advanced	O
Light	O
Source	O
at	O
Lawrence	O
Berkeley	O
National	O
Laboratory	O
with	O
a	O
synchrotron	O
source	O
at	O
1	O
.	O
00	O
Å	O
wavelength	O
to	O
achieve	O
higher	O
resolution	O
.	O
	O
Phases	O
for	O
peptide	B-mutant
4	I-mutant
were	O
determined	O
by	O
single	O
-	O
wavelength	O
anomalous	O
diffraction	O
(	O
SAD	O
)	O
phasing	O
by	O
using	O
the	O
coordinates	O
of	O
the	O
iodine	O
anomalous	O
signal	O
from	O
p	O
-	O
iodophenylalanine	O
.	O
	O
Phases	O
for	O
peptide	B-mutant
2	I-mutant
were	O
determined	O
by	O
isomorphous	O
replacement	O
of	O
peptide	B-mutant
4	I-mutant
.	O
	O
The	O
structures	O
of	O
peptides	B-mutant
2	I-mutant
and	I-mutant
4	I-mutant
were	O
solved	O
and	O
refined	O
in	O
the	O
P6122	O
space	O
group	O
.	O
	O
The	O
asymmetric	O
unit	O
of	O
each	O
peptide	O
consists	O
of	O
six	O
monomers	O
,	O
arranged	O
as	O
two	O
trimers	O
.	O
	O
Peptides	B-mutant
2	I-mutant
and	I-mutant
4	I-mutant
form	O
morphologically	O
identical	O
structures	O
and	O
assemblies	O
in	O
the	O
crystal	O
lattice	O
.	O
	O
The	O
X	O
-	O
ray	O
crystallographic	O
structure	O
of	O
peptide	B-mutant
2	I-mutant
reveals	O
that	O
it	O
folds	O
to	O
form	O
a	O
twisted	O
β	O
-	O
hairpin	O
comprising	O
two	O
β	O
-	O
strands	O
connected	O
by	O
a	O
loop	O
(	O
Figure	O
2A	O
).	O
	O
Eight	O
residues	O
make	O
up	O
each	O
surface	O
of	O
the	O
β	O
-	O
hairpin	O
:	O
L17	O
,	O
F19	O
,	O
A21	O
,	O
D23	O
,	O
A30	O
,	O
I32	O
,	O
L34	O
,	O
and	O
V36	O
make	O
up	O
one	O
surface	O
;	O
V18	O
,	O
F20	O
,	O
E22	O
,	O
C24	O
,	O
C29	O
,	O
I31	O
,	O
G33	O
,	O
and	O
M35	O
make	O
up	O
the	O
other	O
surface	O
.	O
	O
The	O
β	O
-	O
strands	O
of	O
the	O
monomers	O
in	O
the	O
asymmetric	O
unit	O
are	O
virtually	O
identical	O
,	O
differing	O
primarily	O
in	O
rotamers	O
of	O
F20	O
,	O
E22	O
,	O
C24	O
,	O
C29	O
,	O
I31	O
,	O
and	O
M35	O
(	O
Figure	O
S1	O
).	O
	O
The	O
disulfide	O
linkages	O
suffered	O
radiation	O
damage	O
under	O
synchrotron	O
radiation	O
.	O
	O
We	O
refined	O
three	O
of	O
the	O
β	O
-	O
hairpins	O
with	O
intact	O
disulfide	O
linkages	O
and	O
three	O
with	O
thiols	O
to	O
represent	O
cleaved	O
disulfide	O
linkages	O
in	O
the	O
synchrotron	O
data	O
set	O
(	O
PDB	O
5HOX	O
).	O
	O
No	O
evidence	O
for	O
cleavage	O
of	O
the	O
disulfides	O
was	O
observed	O
in	O
the	O
refinement	O
of	O
the	O
data	O
set	O
collected	O
on	O
the	O
X	O
-	O
ray	O
diffractometer	O
,	O
and	O
we	O
refined	O
all	O
disulfide	O
linkages	O
as	O
intact	O
(	O
PDB	O
5HOY	O
).	O
	O
X	O
-	O
ray	O
crystallographic	O
structure	O
of	O
peptide	B-mutant
2	I-mutant
(	O
PDB	O
5HOX	O
,	O
synchrotron	O
data	O
set	O
).	O
(	O
A	O
)	O
X	O
-	O
ray	O
crystallographic	O
structure	O
of	O
a	O
representative	O
β	O
-	O
hairpin	O
monomer	O
formed	O
by	O
peptide	B-mutant
2	I-mutant
.	O
(	O
B	O
)	O
Overlay	O
of	O
the	O
six	O
β	O
-	O
hairpin	O
monomers	O
in	O
the	O
asymmetric	O
unit	O
.	O
	O
The	O
β	O
-	O
hairpins	O
are	O
shown	O
as	O
cartoons	O
to	O
illustrate	O
the	O
differences	O
in	O
the	O
Aβ25	O
–	O
28	O
loops	O
.	O
	O
The	O
Aβ25	O
–	O
28	O
loops	O
of	O
the	O
six	O
monomers	O
within	O
the	O
asymmetric	O
unit	O
vary	O
substantially	O
in	O
backbone	O
geometry	O
and	O
side	O
chain	O
rotamers	O
(	O
Figures	O
2B	O
and	O
S1	O
).	O
	O
The	O
electron	O
density	O
for	O
the	O
loops	O
is	O
weak	O
and	O
diffuse	O
compared	O
to	O
the	O
electron	O
density	O
for	O
the	O
β	O
-	O
strands	O
.	O
	O
Trimer	O
	O
Peptide	B-mutant
2	I-mutant
forms	O
a	O
trimer	O
,	O
much	O
like	O
that	O
which	O
we	O
observed	O
previously	O
for	O
peptide	B-mutant
1	I-mutant
,	O
in	O
which	O
three	O
β	O
-	O
hairpins	O
assemble	O
to	O
form	O
an	O
equilateral	O
triangle	O
(	O
Figure	O
3A	O
).	O
	O
The	O
disulfide	O
bonds	O
between	O
residues	O
24	O
and	O
29	O
are	O
adjacent	O
to	O
the	O
structural	O
core	O
of	O
the	O
trimer	O
and	O
do	O
not	O
make	O
any	O
substantial	O
intermolecular	O
contacts	O
.	O
	O
Two	O
crystallographically	O
distinct	O
trimers	O
comprise	O
the	O
peptide	O
portion	O
of	O
the	O
asymmetric	O
unit	O
.	O
	O
The	O
two	O
trimers	O
are	O
almost	O
identical	O
in	O
structure	O
,	O
differing	O
slightly	O
among	O
side	O
chain	O
rotamers	O
and	O
loop	O
conformations	O
.	O
	O
X	O
-	O
ray	O
crystallographic	O
structure	O
of	O
the	O
trimer	O
formed	O
by	O
peptide	B-mutant
2	I-mutant
.	O
(	O
A	O
)	O
Triangular	O
trimer	O
.	O
	O
The	O
three	O
water	O
molecules	O
in	O
the	O
center	O
hole	O
of	O
the	O
trimer	O
are	O
shown	O
as	O
spheres	O
.	O
(	O
B	O
)	O
Detailed	O
view	O
of	O
the	O
intermolecular	O
hydrogen	O
bonds	O
between	O
the	O
main	O
chains	O
of	O
V18	O
and	O
E22	O
and	O
δOrn	O
and	O
C24	O
,	O
at	O
the	O
three	O
corners	O
of	O
the	O
triangular	O
trimer	O
.	O
(	O
C	O
)	O
The	O
F19	O
face	O
of	O
the	O
trimer	O
,	O
with	O
key	O
side	O
chains	O
shown	O
as	O
spheres	O
.	O
(	O
D	O
)	O
The	O
F20	O
face	O
of	O
the	O
trimer	O
,	O
with	O
key	O
side	O
chains	O
as	O
spheres	O
.	O
	O
A	O
network	O
of	O
18	O
intermolecular	O
hydrogen	O
bonds	O
helps	O
stabilize	O
the	O
trimer	O
.	O
	O
Three	O
ordered	O
water	O
molecules	O
fill	O
the	O
hole	O
in	O
the	O
center	O
of	O
the	O
trimer	O
,	O
hydrogen	O
bonding	O
to	O
each	O
other	O
and	O
to	O
the	O
main	O
chain	O
of	O
F20	O
(	O
Figure	O
3A	O
).	O
	O
Hydrophobic	O
contacts	O
between	O
residues	O
at	O
the	O
three	O
corners	O
of	O
the	O
trimer	O
,	O
where	O
the	O
β	O
-	O
hairpins	O
meet	O
,	O
further	O
stabilize	O
the	O
trimer	O
.	O
	O
At	O
each	O
corner	O
,	O
the	O
side	O
chains	O
of	O
residues	O
L17	O
,	O
F19	O
,	O
and	O
V36	O
of	O
one	O
β	O
-	O
hairpin	O
pack	O
against	O
the	O
side	O
chains	O
of	O
residues	O
A21	O
,	O
I32	O
,	O
L34	O
,	O
and	O
also	O
D23	O
of	O
the	O
adjacent	O
β	O
-	O
hairpin	O
to	O
create	O
a	O
hydrophobic	O
cluster	O
(	O
Figure	O
3C	O
).	O
The	O
three	O
hydrophobic	O
clusters	O
create	O
a	O
large	O
hydrophobic	O
surface	O
on	O
one	O
face	O
of	O
the	O
trimer	O
.	O
	O
In	O
subsequent	O
discussion	O
,	O
we	O
designate	O
the	O
former	O
surface	O
the	O
“	O
F19	O
face	O
”	O
and	O
the	O
latter	O
surface	O
the	O
“	O
F20	O
face	O
”.	O
	O
Four	O
trimers	O
assemble	O
to	O
form	O
a	O
dodecamer	O
.	O
	O
Each	O
of	O
the	O
12	O
β	O
-	O
hairpins	O
constitutes	O
an	O
edge	O
of	O
the	O
octahedron	O
,	O
and	O
the	O
triangular	O
trimers	O
occupy	O
four	O
of	O
the	O
eight	O
faces	O
of	O
the	O
octahedron	O
.	O
	O
Figure	O
4A	O
illustrates	O
the	O
octahedral	O
shape	O
of	O
the	O
dodecamer	O
.	O
	O
Figure	O
4B	O
illustrates	O
the	O
tetrahedral	O
arrangement	O
of	O
the	O
four	O
trimers	O
.	O
	O
X	O
-	O
ray	O
crystallographic	O
structure	O
of	O
the	O
dodecamer	O
formed	O
by	O
peptide	B-mutant
2	I-mutant
.	O
(	O
A	O
)	O
View	O
of	O
the	O
dodecamer	O
that	O
illustrates	O
the	O
octahedral	O
shape	O
.	O
(	O
B	O
)	O
View	O
of	O
the	O
dodecamer	O
that	O
illustrates	O
the	O
tetrahedral	O
arrangement	O
of	O
the	O
four	O
trimers	O
that	O
comprise	O
the	O
dodecamer	O
.	O
(	O
C	O
)	O
View	O
of	O
two	O
trimer	O
subunits	O
from	O
inside	O
the	O
cavity	O
of	O
the	O
dodecamer	O
.	O
	O
The	O
F19	O
faces	O
of	O
the	O
trimers	O
line	O
the	O
interior	O
of	O
the	O
dodecamer	O
.	O
	O
At	O
the	O
six	O
vertices	O
,	O
hydrophobic	O
packing	O
between	O
the	O
side	O
chains	O
of	O
L17	O
,	O
L34	O
,	O
and	O
V36	O
helps	O
stabilize	O
the	O
dodecamer	O
(	O
Figures	O
4C	O
and	O
D	O
).	O
	O
Salt	O
bridges	O
between	O
the	O
side	O
chains	O
of	O
D23	O
and	O
δOrn	O
at	O
the	O
vertices	O
further	O
stabilize	O
the	O
dodecamer	O
.	O
	O
Each	O
of	O
the	O
six	O
vertices	O
includes	O
two	O
Aβ25	O
–	O
28	O
loops	O
that	O
extend	O
past	O
the	O
core	O
of	O
the	O
dodecamer	O
without	O
making	O
any	O
substantial	O
intermolecular	O
contacts	O
.	O
	O
The	O
exterior	O
of	O
the	O
dodecamer	O
displays	O
four	O
F20	O
faces	O
(	O
Figure	O
S3	O
).	O
	O
Although	O
the	O
asymmetric	O
unit	O
comprises	O
half	O
a	O
dodecamer	O
,	O
the	O
crystal	O
lattice	O
may	O
be	O
thought	O
of	O
as	O
being	O
built	O
of	O
dodecamers	O
.	O
	O
The	O
shape	O
and	O
length	O
of	O
the	O
electron	O
density	O
is	O
consistent	O
with	O
the	O
structure	O
of	O
Jeffamine	O
M	O
-	O
600	O
,	O
which	O
is	O
an	O
essential	O
component	O
of	O
the	O
crystallization	O
conditions	O
.	O
	O
The	O
Jeffamine	O
M	O
-	O
600	O
appears	O
to	O
stabilize	O
the	O
dodecamer	O
by	O
occupying	O
the	O
central	O
cavity	O
and	O
making	O
hydrophobic	O
contacts	O
with	O
residues	O
lining	O
the	O
cavity	O
(	O
Figure	O
S3	O
).	O
	O
In	O
a	O
dodecamer	O
formed	O
by	O
full	O
-	O
length	O
Aβ	O
,	O
the	O
hydrophobic	O
C	O
-	O
terminal	O
residues	O
(	O
Aβ37	O
–	O
40	O
or	O
Aβ37	O
–	O
42	O
)	O
might	O
play	O
a	O
similar	O
role	O
in	O
filling	O
the	O
dodecamer	O
and	O
thus	O
create	O
a	O
packed	O
hydrophobic	O
core	O
within	O
the	O
central	O
cavity	O
of	O
the	O
dodecamer	O
.	O
	O
Five	O
dodecamers	O
assemble	O
to	O
form	O
an	O
annular	O
porelike	O
structure	O
(	O
Figure	O
5A	O
).	O
	O
Two	O
morphologically	O
distinct	O
interactions	O
between	O
trimers	O
occur	O
at	O
the	O
interfaces	O
of	O
the	O
five	O
dodecamers	O
:	O
one	O
in	O
which	O
the	O
trimers	O
are	O
eclipsed	O
(	O
Figure	O
5B	O
),	O
and	O
one	O
in	O
which	O
the	O
trimers	O
are	O
staggered	O
(	O
Figure	O
5C	O
).	O
	O
Hydrophobic	O
packing	O
between	O
the	O
side	O
chains	O
of	O
F20	O
,	O
I31	O
,	O
and	O
E22	O
stabilizes	O
these	O
interfaces	O
(	O
Figure	O
5D	O
and	O
E	O
).	O
	O
The	O
annular	O
pore	O
contains	O
three	O
eclipsed	O
interfaces	O
and	O
two	O
staggered	O
interfaces	O
.	O
	O
The	O
staggered	O
interfaces	O
occur	O
between	O
dodecamers	O
2	O
and	O
3	O
and	O
4	O
and	O
5	O
.	O
	O
The	O
annular	O
pore	O
is	O
not	O
completely	O
flat	O
,	O
instead	O
,	O
adopting	O
a	O
slightly	O
puckered	O
shape	O
,	O
which	O
accommodates	O
the	O
eclipsed	O
and	O
staggered	O
interfaces	O
.	O
	O
Ten	O
Aβ25	O
–	O
28	O
loops	O
from	O
the	O
vertices	O
of	O
the	O
five	O
dodecamers	O
line	O
the	O
hole	O
in	O
the	O
center	O
of	O
the	O
pore	O
.	O
	O
The	O
hydrophilic	O
side	O
chains	O
of	O
S26	O
,	O
N27	O
,	O
and	O
K28	O
decorate	O
the	O
hole	O
.	O
	O
X	O
-	O
ray	O
crystallographic	O
structure	O
of	O
the	O
annular	O
pore	O
formed	O
by	O
peptide	B-mutant
2	I-mutant
.	O
(	O
A	O
)	O
Annular	O
porelike	O
structure	O
illustrating	O
the	O
relationship	O
of	O
the	O
five	O
dodecamers	O
that	O
form	O
the	O
pore	O
(	O
top	O
view	O
).	O
	O
(	O
B	O
)	O
Eclipsed	O
interface	O
between	O
dodecamers	O
1	O
and	O
2	O
(	O
side	O
view	O
).	O
	O
The	O
same	O
staggered	O
interface	O
also	O
occurs	O
between	O
dodecamers	O
4	O
and	O
5	O
.	O
(	O
D	O
)	O
Eclipsed	O
interface	O
between	O
dodecamers	O
1	O
and	O
5	O
(	O
top	O
view	O
).	O
	O
The	O
annular	O
pore	O
is	O
comparable	O
in	O
size	O
to	O
other	O
large	O
protein	O
assemblies	O
.	O
	O
The	O
diameter	O
of	O
the	O
hole	O
in	O
the	O
center	O
of	O
the	O
pore	O
is	O
∼	O
2	O
nm	O
.	O
	O
The	O
thickness	O
of	O
the	O
pore	O
is	O
∼	O
5	O
nm	O
,	O
which	O
is	O
comparable	O
to	O
that	O
of	O
a	O
lipid	O
bilayer	O
membrane	O
.	O
	O
Rather	O
,	O
the	O
crystal	O
lattice	O
is	O
composed	O
of	O
conjoined	O
annular	O
pores	O
in	O
which	O
all	O
four	O
F20	O
faces	O
on	O
the	O
surface	O
of	O
each	O
dodecamer	O
contact	O
F20	O
faces	O
on	O
other	O
dodecamers	O
(	O
Figure	O
S4	O
).	O
	O
The	O
crystal	O
lattice	O
shows	O
how	O
the	O
dodecamers	O
can	O
further	O
assemble	O
to	O
form	O
larger	O
structures	O
.	O
	O
Each	O
dodecamer	O
may	O
be	O
thought	O
of	O
as	O
a	O
tetravalent	O
building	O
block	O
with	O
the	O
potential	O
to	O
assemble	O
on	O
all	O
four	O
faces	O
to	O
form	O
higher	O
-	O
order	O
supramolecular	O
assemblies	O
.	O
	O
The	O
X	O
-	O
ray	O
crystallographic	O
study	O
of	O
peptide	B-mutant
2	I-mutant
described	O
here	O
provides	O
high	O
-	O
resolution	O
structures	O
of	O
oligomers	O
formed	O
by	O
an	O
Aβ17	O
–	O
36	O
β	O
-	O
hairpin	O
.	O
	O
Model	O
for	O
the	O
hierarchical	O
assembly	O
of	O
an	O
Aβ	O
β	O
-	O
hairpin	O
into	O
a	O
trimer	O
,	O
dodecamer	O
,	O
and	O
annular	O
pore	O
based	O
on	O
the	O
crystallographic	O
assembly	O
of	O
peptide	B-mutant
2	I-mutant
.	O
	O
Three	O
β	O
-	O
hairpin	O
monomers	O
assemble	O
to	O
form	O
a	O
triangular	O
trimer	O
.	O
	O
The	O
molecular	O
weights	O
shown	O
correspond	O
to	O
an	O
Aβ42	O
monomer	O
(∼	O
4	O
.	O
5	O
kDa	O
),	O
an	O
Aβ42	O
trimer	O
(∼	O
13	O
.	O
5	O
kDa	O
),	O
an	O
Aβ42	O
dodecamer	O
(∼	O
54	O
kDa	O
),	O
and	O
an	O
Aβ42	O
annular	O
pore	O
composed	O
of	O
five	O
dodecamers	O
(∼	O
270	O
kDa	O
).	O
	O
Two	O
general	O
types	O
of	O
endogenous	O
Aβ	O
oligomers	O
have	O
been	O
observed	O
:	O
Aβ	O
oligomers	O
that	O
occur	O
on	O
a	O
pathway	O
to	O
fibrils	O
,	O
or	O
“	O
fibrillar	O
oligomers	O
”,	O
and	O
Aβ	O
oligomers	O
that	O
evade	O
a	O
fibrillar	O
fate	O
,	O
or	O
“	O
nonfibrillar	O
oligomers	O
”.−	O
Fibrillar	O
oligomers	O
accumulate	O
in	O
Alzheimer	O
’	O
s	O
disease	O
later	O
than	O
nonfibrillar	O
oligomers	O
and	O
coincide	O
with	O
the	O
deposition	O
of	O
plaques	O
.	O
	O
Nonfibrillar	O
oligomers	O
accumulate	O
early	O
in	O
Alzheimer	O
’	O
s	O
disease	O
before	O
plaque	O
deposition	O
.	O
	O
Fibrillar	O
oligomers	O
are	O
recognized	O
by	O
the	O
OC	O
antibody	O
but	O
not	O
the	O
A11	O
antibody	O
,	O
whereas	O
nonfibrillar	O
oligomers	O
are	O
recognized	O
by	O
the	O
A11	O
antibody	O
but	O
not	O
the	O
OC	O
antibody	O
.	O
	O
Larson	O
and	O
Lesné	O
proposed	O
a	O
model	O
for	O
the	O
endogenous	O
production	O
of	O
nonfibrillar	O
oligomers	O
that	O
explains	O
these	O
observations	O
.	O
	O
In	O
this	O
model	O
,	O
folded	O
Aβ	O
monomer	O
assembles	O
into	O
a	O
trimer	O
,	O
the	O
trimer	O
further	O
assembles	O
into	O
hexamers	O
and	O
dodecamers	O
,	O
and	O
the	O
dodecamers	O
further	O
assemble	O
to	O
form	O
annular	O
protofibrils	O
.	O
	O
The	O
crystallographically	O
observed	O
annular	O
pore	O
formed	O
by	O
peptide	B-mutant
2	I-mutant
is	O
morphologically	O
similar	O
to	O
the	O
APFs	O
formed	O
by	O
full	O
-	O
length	O
Aβ	O
.	O
	O
The	O
annular	O
pore	O
formed	O
by	O
peptide	B-mutant
2	I-mutant
is	O
comparable	O
in	O
size	O
to	O
the	O
APFs	O
prepared	O
in	O
vitro	O
or	O
isolated	O
from	O
Alzheimer	O
’	O
s	O
brains	O
(	O
Figure	O
7	O
and	O
Table	O
1	O
).	O
	O
The	O
varying	O
sizes	O
of	O
APFs	O
formed	O
by	O
full	O
-	O
length	O
Aβ	O
might	O
result	O
from	O
differences	O
in	O
the	O
number	O
of	O
oligomer	O
subunits	O
comprising	O
each	O
APF	O
.	O
	O
Although	O
the	O
annular	O
pore	O
formed	O
by	O
peptide	B-mutant
2	I-mutant
contains	O
five	O
dodecamer	O
subunits	O
,	O
pores	O
containing	O
fewer	O
or	O
more	O
subunits	O
can	O
easily	O
be	O
envisioned	O
.	O
	O
The	O
dodecamers	O
that	O
comprise	O
the	O
annular	O
pore	O
exhibit	O
two	O
modes	O
of	O
assembly	O
—	O
eclipsed	O
interactions	O
and	O
staggered	O
interactions	O
between	O
the	O
F20	O
faces	O
of	O
trimers	O
within	O
dodecamers	O
.	O
	O
Surface	O
views	O
of	O
the	O
annular	O
pore	O
formed	O
by	O
peptide	B-mutant
2	I-mutant
.	O
(	O
A	O
)	O
Top	O
view	O
.	O
	O
annular	O
pore	O
source	O
outer	O
diameter	O
inner	O
diameter	O
observation	O
method	O
peptide	O
2	O
∼	O
11	O
–	O
12	O
nm	O
∼	O
2	O
nm	O
X	O
-	O
ray	O
crystallography	O
synthetic	O
Aβ	O
7	O
–	O
10	O
nm	O
1	O
.	O
5	O
–	O
2	O
nm	O
TEM	O
synthetic	O
Aβ	O
16	O
nm	O
not	O
reported	O
AFM	O
synthetic	O
Aβ	O
8	O
–	O
25	O
nm	O
not	O
reported	O
TEM	O
Alzheimer	O
’	O
s	O
brain	O
11	O
–	O
14	O
nm	O
2	O
.	O
5	O
–	O
4	O
nm	O
TEM	O
	O
Dot	O
blot	O
analysis	O
shows	O
that	O
peptide	B-mutant
2	I-mutant
is	O
reactive	O
toward	O
the	O
A11	O
antibody	O
(	O
Figure	O
S5	O
).	O
	O
This	O
reactivity	O
suggests	O
that	O
peptide	B-mutant
2	I-mutant
forms	O
oligomers	O
in	O
solution	O
that	O
share	O
structural	O
similarities	O
to	O
the	O
nonfibrillar	O
oligomers	O
formed	O
by	O
full	O
-	O
length	O
Aβ	O
.	O
	O
Further	O
studies	O
are	O
needed	O
to	O
elucidate	O
the	O
species	O
that	O
peptide	B-mutant
2	I-mutant
forms	O
in	O
solution	O
and	O
to	O
study	O
their	O
biological	O
properties	O
.	O
	O
The	O
difficulty	O
in	O
studying	O
the	O
oligomers	O
formed	O
in	O
solution	O
may	O
reflect	O
the	O
propensity	O
of	O
the	O
dodecamer	O
to	O
assemble	O
on	O
all	O
four	O
F20	O
faces	O
.	O
	O
The	O
X	O
-	O
ray	O
crystallographic	O
structure	O
and	O
A11	O
reactivity	O
of	O
peptide	B-mutant
2	I-mutant
support	O
the	O
model	O
proposed	O
by	O
Larsen	O
and	O
Lesné	O
and	O
suggest	O
that	O
β	O
-	O
hairpins	O
constitute	O
a	O
fundamental	O
building	O
block	O
for	O
nonfibrillar	O
oligomers	O
.	O
	O
What	O
makes	O
β	O
-	O
hairpins	O
special	O
is	O
that	O
three	O
β	O
-	O
hairpins	O
can	O
nestle	O
together	O
to	O
form	O
trimers	O
,	O
stabilized	O
by	O
a	O
network	O
of	O
hydrogen	O
bonds	O
and	O
hydrophobic	O
interactions	O
.	O
	O
The	O
foldon	O
domain	O
of	O
bacteriophage	O
T4	O
fibritin	O
is	O
composed	O
of	O
three	O
β	O
-	O
hairpins	O
that	O
assemble	O
into	O
a	O
triangular	O
trimer	O
similar	O
to	O
the	O
triangular	O
trimer	O
formed	O
by	O
peptide	B-mutant
2	I-mutant
.	O
	O
Additionally	O
,	O
our	O
research	O
group	O
has	O
observed	O
a	O
similar	O
assembly	O
of	O
a	O
β	O
-	O
hairpin	O
peptide	O
derived	O
from	O
β2	O
-	O
microglobulin	O
.	O
	O
Although	O
we	O
began	O
these	O
studies	O
with	O
a	O
relatively	O
simple	O
hypothesis	O
—	O
that	O
the	O
trimers	O
and	O
dodecamers	O
formed	O
by	O
peptide	B-mutant
1	I-mutant
could	O
accommodate	O
the	O
Aβ24	O
–	O
29	O
loop	O
—	O
an	O
even	O
more	O
exciting	O
finding	O
has	O
emerged	O
—	O
that	O
the	O
dodecamers	O
can	O
assemble	O
to	O
form	O
annular	O
pores	O
.	O
	O
This	O
finding	O
could	O
not	O
have	O
been	O
anticipated	O
from	O
the	O
X	O
-	O
ray	O
crystallographic	O
structure	O
of	O
peptide	B-mutant
1	I-mutant
and	O
reveals	O
a	O
new	O
level	O
of	O
hierarchical	O
assembly	O
that	O
recapitulates	O
micrographic	O
observations	O
of	O
annular	O
protofibrils	O
.	O
	O
The	O
crystallographically	O
observed	O
dodecamer	O
,	O
in	O
turn	O
,	O
recapitulates	O
the	O
observation	O
of	O
Aβ	O
*	O
56	O
,	O
which	O
appears	O
to	O
be	O
a	O
dodecamer	O
of	O
Aβ	O
.	O
	O
The	O
crystallographically	O
observed	O
trimer	O
recapitulates	O
the	O
Aβ	O
trimers	O
that	O
are	O
observed	O
even	O
before	O
the	O
onset	O
of	O
symptoms	O
in	O
Alzheimer	O
’	O
s	O
disease	O
.	O
	O
Predictive	O
features	O
of	O
ligand	O
‐	O
specific	O
signaling	O
through	O
the	O
estrogen	O
receptor	O
	O
Some	O
estrogen	O
receptor	O
‐	O
α	O
(	O
ERα	O
)‐	O
targeted	O
breast	O
cancer	O
therapies	O
such	O
as	O
tamoxifen	O
have	O
tissue	O
‐	O
selective	O
or	O
cell	O
‐	O
specific	O
activities	O
,	O
while	O
others	O
have	O
similar	O
activities	O
in	O
different	O
cell	O
types	O
.	O
	O
To	O
identify	O
biophysical	O
determinants	O
of	O
cell	O
‐	O
specific	O
signaling	O
and	O
breast	O
cancer	O
cell	O
proliferation	O
,	O
we	O
synthesized	O
241	O
ERα	O
ligands	O
based	O
on	O
19	O
chemical	O
scaffolds	O
,	O
and	O
compared	O
ligand	O
response	O
using	O
quantitative	O
bioassays	O
for	O
canonical	O
ERα	O
activities	O
and	O
X	O
‐	O
ray	O
crystallography	O
.	O
	O
For	O
some	O
ligand	O
series	O
,	O
a	O
single	O
inter	O
‐	O
atomic	O
distance	O
in	O
the	O
ligand	O
‐	O
binding	O
domain	O
predicted	O
their	O
proliferative	O
effects	O
.	O
	O
Thus	O
,	O
incorporating	O
systems	O
structural	O
analyses	O
with	O
quantitative	O
chemical	O
biology	O
reveals	O
how	O
ligands	O
can	O
achieve	O
distinct	O
allosteric	O
signaling	O
outcomes	O
through	O
ERα	O
.	O
	O
Many	O
drugs	O
are	O
small	O
‐	O
molecule	O
ligands	O
of	O
allosteric	O
signaling	O
proteins	O
,	O
including	O
G	O
protein	O
‐	O
coupled	O
receptors	O
(	O
GPCRs	O
)	O
and	O
nuclear	O
receptors	O
such	O
as	O
ERα	O
.	O
	O
Small	O
‐	O
molecule	O
ligands	O
control	O
receptor	O
activity	O
by	O
modulating	O
recruitment	O
of	O
effector	O
enzymes	O
to	O
distal	O
regions	O
of	O
the	O
receptor	O
,	O
relative	O
to	O
the	O
ligand	O
‐	O
binding	O
site	O
.	O
	O
Allosteric	O
control	O
of	O
ERα	O
activity	O
	O
Chemical	O
structures	O
of	O
some	O
common	O
ERα	O
ligands	O
.	O
	O
Schematic	O
illustration	O
of	O
the	O
canonical	O
ERα	O
signaling	O
pathway	O
.	O
	O
ERα	O
contains	O
structurally	O
conserved	O
globular	O
domains	O
of	O
the	O
nuclear	O
receptor	O
superfamily	O
,	O
including	O
a	O
DNA	O
‐	O
binding	O
domain	O
(	O
DBD	O
)	O
that	O
is	O
connected	O
by	O
a	O
flexible	O
hinge	O
region	O
to	O
the	O
ligand	O
‐	O
binding	O
domain	O
(	O
LBD	O
),	O
as	O
well	O
as	O
unstructured	O
AB	O
and	O
F	O
domains	O
at	O
its	O
amino	O
and	O
carboxyl	O
termini	O
,	O
respectively	O
(	O
Fig	O
1B	O
).	O
	O
The	O
LBD	O
contains	O
a	O
ligand	O
‐	O
dependent	O
coactivator	O
‐	O
binding	O
site	O
called	O
activation	O
function	O
‐	O
2	O
(	O
AF	O
‐	O
2	O
).	O
	O
AF	O
‐	O
1	O
and	O
AF	O
‐	O
2	O
bind	O
distinct	O
but	O
overlapping	O
sets	O
of	O
coregulators	O
(	O
Webb	O
et	O
al	O
,	O
1998	O
;	O
Endoh	O
et	O
al	O
,	O
1999	O
;	O
Delage	O
‐	O
Mourroux	O
et	O
al	O
,	O
2000	O
;	O
Yi	O
et	O
al	O
,	O
2015	O
).	O
	O
AF	O
‐	O
2	O
binds	O
the	O
signature	O
LxxLL	O
motif	O
peptides	O
of	O
coactivators	O
such	O
as	O
NCOA1	O
/	O
2	O
/	O
3	O
(	O
also	O
known	O
as	O
SRC	O
‐	O
1	O
/	O
2	O
/	O
3	O
).	O
	O
Yet	O
,	O
it	O
is	O
unknown	O
how	O
different	O
ERα	O
ligands	O
control	O
AF	O
‐	O
1	O
through	O
the	O
LBD	O
,	O
and	O
whether	O
this	O
inter	O
‐	O
domain	O
communication	O
is	O
required	O
for	O
cell	O
‐	O
specific	O
signaling	O
or	O
anti	O
‐	O
proliferative	O
responses	O
.	O
	O
In	O
the	O
canonical	O
model	O
of	O
the	O
ERα	O
signaling	O
pathway	O
(	O
Fig	O
1C	O
),	O
E2	O
‐	O
bound	O
ERα	O
forms	O
a	O
homodimer	O
that	O
binds	O
DNA	O
at	O
estrogen	O
‐	O
response	O
elements	O
(	O
EREs	O
),	O
recruits	O
NCOA1	O
/	O
2	O
/	O
3	O
(	O
Metivier	O
et	O
al	O
,	O
2003	O
;	O
Johnson	O
&	O
O	O
'	O
Malley	O
,	O
2012	O
),	O
and	O
activates	O
the	O
GREB1	O
gene	O
,	O
which	O
is	O
required	O
for	O
proliferation	O
of	O
ERα	O
‐	O
positive	O
breast	O
cancer	O
cells	O
(	O
Ghosh	O
et	O
al	O
,	O
2000	O
;	O
Rae	O
et	O
al	O
,	O
2005	O
;	O
Deschenes	O
et	O
al	O
,	O
2007	O
;	O
Liu	O
et	O
al	O
,	O
2012	O
;	O
Srinivasan	O
et	O
al	O
,	O
2013	O
).	O
	O
Our	O
long	O
‐	O
term	O
goal	O
is	O
to	O
be	O
able	O
to	O
predict	O
proliferative	O
or	O
anti	O
‐	O
proliferative	O
activity	O
of	O
a	O
ligand	O
in	O
different	O
tissues	O
from	O
its	O
crystal	O
structure	O
by	O
identifying	O
different	O
structural	O
perturbations	O
that	O
lead	O
to	O
specific	O
signaling	O
outcomes	O
.	O
	O
We	O
also	O
determined	O
the	O
structures	O
of	O
76	O
distinct	O
ERα	O
LBD	O
complexes	O
bound	O
to	O
different	O
ligand	O
types	O
,	O
which	O
allowed	O
us	O
to	O
understand	O
how	O
diverse	O
ligand	O
scaffolds	O
distort	O
the	O
active	O
conformation	O
of	O
the	O
ERα	O
LBD	O
.	O
	O
Our	O
findings	O
here	O
indicate	O
that	O
specific	O
structural	O
perturbations	O
can	O
be	O
tied	O
to	O
ligand	O
‐	O
selective	O
domain	O
usage	O
and	O
signaling	O
patterns	O
,	O
thus	O
providing	O
a	O
framework	O
for	O
structure	O
‐	O
based	O
design	O
of	O
improved	O
breast	O
cancer	O
therapeutics	O
,	O
and	O
understanding	O
the	O
different	O
phenotypic	O
effects	O
of	O
environmental	O
estrogens	O
.	O
	O
High	O
‐	O
throughput	O
screens	O
for	O
ERα	O
ligand	O
profiling	O
	O
Summary	O
of	O
ligand	O
screening	O
assays	O
used	O
to	O
measure	O
ER	O
‐	O
mediated	O
activities	O
.	O
	O
ERE	O
,	O
estrogen	O
‐	O
response	O
element	O
;	O
Luc	O
,	O
luciferase	O
reporter	O
gene	O
;	O
M2H	O
,	O
mammalian	O
2	O
‐	O
hybrid	O
;	O
UAS	O
,	O
upstream	O
‐	O
activating	O
sequence	O
.	O
	O
Strength	O
of	O
AF	O
‐	O
1	O
signaling	O
does	O
not	O
determine	O
cell	O
‐	O
specific	O
signaling	O
	O
To	O
compare	O
ERα	O
signaling	O
induced	O
by	O
diverse	O
ligand	O
types	O
,	O
we	O
synthesized	O
and	O
assayed	O
a	O
library	O
of	O
241	O
ERα	O
ligands	O
containing	O
19	O
distinct	O
molecular	O
scaffolds	O
.	O
	O
These	O
include	O
15	O
indirect	O
modulator	O
series	O
,	O
which	O
lack	O
a	O
SERM	O
‐	O
like	O
side	O
chain	O
and	O
modulate	O
coactivator	O
binding	O
indirectly	O
from	O
the	O
ligand	O
‐	O
binding	O
pocket	O
(	O
Fig	O
2A	O
–	O
E	O
;	O
Dataset	O
EV1	O
)	O
(	O
Zheng	O
et	O
al	O
,	O
2012	O
)	O
(	O
Zhu	O
et	O
al	O
,	O
2012	O
)	O
(	O
Muthyala	O
et	O
al	O
,	O
2003	O
;	O
Seo	O
et	O
al	O
,	O
2006	O
)	O
(	O
Srinivasan	O
et	O
al	O
,	O
2013	O
)	O
(	O
Wang	O
et	O
al	O
,	O
2012	O
)	O
(	O
Liao	O
et	O
al	O
,	O
2014	O
)	O
(	O
Min	O
et	O
al	O
,	O
2013	O
).	O
	O
We	O
also	O
generated	O
four	O
direct	O
modulator	O
series	O
with	O
side	O
chains	O
designed	O
to	O
directly	O
dislocate	O
h12	O
and	O
thereby	O
completely	O
occlude	O
the	O
AF	O
‐	O
2	O
surface	O
(	O
Fig	O
2C	O
and	O
E	O
;	O
Dataset	O
EV1	O
)	O
(	O
Kieser	O
et	O
al	O
,	O
2010	O
).	O
	O
Ligand	O
profiling	O
using	O
our	O
quantitative	O
bioassays	O
revealed	O
a	O
wide	O
range	O
of	O
ligand	O
‐	O
induced	O
GREB1	O
expression	O
,	O
reporter	O
gene	O
activities	O
,	O
ERα	O
‐	O
coactivator	O
interactions	O
,	O
and	O
proliferative	O
effects	O
on	O
MCF	O
‐	O
7	O
breast	O
cancer	O
cells	O
(	O
Figs	O
EV1	O
and	O
EV2A	O
–	O
J	O
).	O
	O
This	O
wide	O
variance	O
enabled	O
us	O
to	O
probe	O
specific	O
features	O
of	O
ERα	O
signaling	O
using	O
ligand	O
class	O
analyses	O
,	O
and	O
identify	O
signaling	O
patterns	O
shared	O
by	O
specific	O
ligand	O
series	O
or	O
scaffolds	O
.	O
	O
Classes	O
of	O
compounds	O
in	O
the	O
ERα	O
ligand	O
library	O
	O
Structural	O
details	O
of	O
the	O
ERα	O
LBD	O
bound	O
to	O
the	O
indicated	O
ligands	O
.	O
	O
Unlike	O
E2	O
(	O
PDB	O
1GWR	O
),	O
TAM	O
is	O
a	O
direct	O
modulator	O
with	O
a	O
BSC	O
that	O
dislocates	O
h12	O
to	O
block	O
the	O
NCOA2	O
‐	O
binding	O
site	O
(	O
PDB	O
3ERT	O
).	O
	O
ERα	O
ligands	O
induced	O
a	O
range	O
of	O
agonist	O
activity	O
profiles	O
	O
To	O
this	O
end	O
,	O
we	O
compared	O
the	O
average	O
ligand	O
‐	O
induced	O
GREB1	O
mRNA	O
levels	O
in	O
MCF	O
‐	O
7	O
cells	O
and	O
3	O
×	O
ERE	O
‐	O
Luc	O
reporter	O
gene	O
activity	O
in	O
Ishikawa	O
endometrial	O
cancer	O
cells	O
(	O
E	O
‐	O
Luc	O
)	O
or	O
in	O
HepG2	O
cells	O
transfected	O
with	O
wild	O
‐	O
type	O
ERα	O
(	O
L	O
‐	O
Luc	O
ERα	O
‐	O
WT	O
)	O
(	O
Figs	O
3A	O
and	O
EV2A	O
–	O
C	O
).	O
	O
Direct	O
modulators	O
showed	O
significant	O
differences	O
in	O
average	O
activity	O
between	O
cell	O
types	O
except	O
OBHS	O
‐	O
ASC	O
analogs	O
,	O
which	O
had	O
similar	O
low	O
agonist	O
activities	O
in	O
the	O
three	O
cell	O
types	O
.	O
	O
While	O
it	O
was	O
known	O
that	O
direct	O
modulators	O
such	O
as	O
tamoxifen	O
drive	O
cell	O
‐	O
specific	O
signaling	O
,	O
these	O
experiments	O
reveal	O
that	O
indirect	O
modulators	O
also	O
drive	O
cell	O
‐	O
specific	O
signaling	O
,	O
since	O
eight	O
of	O
fourteen	O
classes	O
showed	O
significant	O
differences	O
in	O
average	O
activity	O
(	O
Figs	O
3A	O
and	O
EV2A	O
–	O
C	O
).	O
	O
(	O
A	O
)	O
Ligand	O
‐	O
specific	O
ERα	O
activities	O
in	O
HepG2	O
,	O
Ishikawa	O
and	O
MCF	O
‐	O
7	O
cells	O
.	O
	O
The	O
ligand	O
‐	O
induced	O
L	O
‐	O
Luc	O
ERα	O
‐	O
WT	O
and	O
E	O
‐	O
Luc	O
activities	O
and	O
GREB1	O
mRNA	O
levels	O
are	O
shown	O
by	O
scaffold	O
(	O
mean	O
+	O
SD	O
).	O
	O
Significant	O
sensitivity	O
to	O
AB	O
domain	O
deletion	O
was	O
determined	O
by	O
Student	O
'	O
s	O
t	O
‐	O
test	O
(	O
n	O
=	O
number	O
of	O
ligands	O
per	O
scaffold	O
in	O
Fig	O
2	O
).	O
	O
Correlation	O
and	O
regression	O
analyses	O
in	O
a	O
large	O
test	O
set	O
.	O
	O
In	O
cluster	O
2	O
,	O
only	O
one	O
of	O
these	O
comparisons	O
revealed	O
a	O
significant	O
positive	O
correlation	O
,	O
while	O
none	O
was	O
significant	O
in	O
cluster	O
3	O
.	O
+,	O
statistically	O
significant	O
correlations	O
gained	O
by	O
deletion	O
of	O
the	O
AB	O
or	O
F	O
domains	O
.	O
	O
To	O
test	O
this	O
idea	O
,	O
we	O
compared	O
the	O
average	O
L	O
‐	O
Luc	O
activities	O
of	O
each	O
scaffold	O
in	O
HepG2	O
cells	O
co	O
‐	O
transfected	O
with	O
wild	O
‐	O
type	O
ERα	O
or	O
with	O
ERα	O
lacking	O
the	O
AB	O
domain	O
(	O
Figs	O
1B	O
and	O
EV1	O
).	O
	O
While	O
E2	O
showed	O
similar	O
L	O
‐	O
Luc	O
ERα	O
‐	O
WT	O
and	O
ERα	B-mutant
‐	I-mutant
ΔAB	I-mutant
activities	O
,	O
tamoxifen	O
showed	O
complete	O
loss	O
of	O
activity	O
without	O
the	O
AB	O
domain	O
(	O
Fig	O
EV1B	O
).	O
	O
These	O
“	O
AF	O
‐	O
1	O
‐	O
sensitive	O
”	O
activities	O
were	O
exhibited	O
by	O
both	O
direct	O
and	O
indirect	O
modulators	O
,	O
and	O
were	O
not	O
limited	O
to	O
scaffolds	O
that	O
showed	O
cell	O
‐	O
specific	O
signaling	O
(	O
Fig	O
3A	O
and	O
B	O
).	O
	O
For	O
each	O
ligand	O
class	O
or	O
scaffold	O
,	O
we	O
calculated	O
the	O
Pearson	O
'	O
s	O
correlation	O
coefficient	O
,	O
r	O
,	O
for	O
pairwise	O
comparison	O
of	O
activity	O
profiles	O
in	O
breast	O
(	O
GREB1	O
),	O
liver	O
(	O
L	O
‐	O
Luc	O
),	O
and	O
endometrial	O
cells	O
(	O
E	O
‐	O
Luc	O
).	O
	O
We	O
also	O
calculated	O
the	O
coefficient	O
of	O
determination	O
,	O
r	O
2	O
,	O
which	O
describes	O
the	O
percentage	O
of	O
variance	O
in	O
a	O
dependent	O
variable	O
such	O
as	O
proliferation	O
that	O
can	O
be	O
predicted	O
by	O
an	O
independent	O
variable	O
such	O
as	O
GREB1	O
expression	O
.	O
	O
We	O
present	O
both	O
calculations	O
as	O
r	O
2	O
to	O
readily	O
compare	O
signaling	O
specificities	O
using	O
a	O
heat	O
map	O
on	O
which	O
the	O
red	O
–	O
yellow	O
palette	O
indicates	O
significant	O
positive	O
correlations	O
(	O
P	O
≤	O
0	O
.	O
05	O
,	O
F	O
‐	O
test	O
for	O
nonzero	O
slope	O
),	O
while	O
the	O
blue	O
palette	O
denotes	O
negative	O
correlations	O
(	O
Fig	O
3C	O
–	O
F	O
).	O
	O
Correlation	O
analysis	O
of	O
OBHS	O
versus	O
OBHS	O
‐	O
BSC	O
activity	O
across	O
cell	O
types	O
.	O
	O
Correlation	O
analysis	O
of	O
L	O
‐	O
Luc	O
ERα	B-mutant
‐	I-mutant
ΔAB	I-mutant
activity	O
versus	O
endogenous	O
ERα	O
activity	O
of	O
OBHS	O
analogs	O
.	O
	O
Correlation	O
analysis	O
of	O
L	O
‐	O
Luc	O
ERα	B-mutant
‐	I-mutant
ΔF	I-mutant
activity	O
versus	O
endogenous	O
ERα	O
activities	O
of	O
OBHS	O
analogs	O
.	O
	O
Correlation	O
analysis	O
of	O
MCF	O
‐	O
7	O
cell	O
proliferation	O
versus	O
NCOA2	O
/	O
3	O
recruitment	O
or	O
GREB1	O
levels	O
observed	O
in	O
response	O
to	O
(	O
G	O
)	O
OBHS	O
‐	O
N	O
and	O
(	O
H	O
)	O
OBHS	O
‐	O
BSC	O
analogs	O
.	O
	O
Scaffolds	O
in	O
cluster	O
1	O
exhibited	O
strongly	O
correlated	O
GREB1	O
levels	O
,	O
E	O
‐	O
Luc	O
and	O
L	O
‐	O
Luc	O
activity	O
profiles	O
across	O
the	O
three	O
cell	O
types	O
(	O
Fig	O
3C	O
lanes	O
1	O
–	O
4	O
),	O
suggesting	O
these	O
ligands	O
use	O
similar	O
ERα	O
signaling	O
pathways	O
in	O
the	O
breast	O
,	O
endometrial	O
,	O
and	O
liver	O
cell	O
types	O
.	O
	O
This	O
cluster	O
includes	O
WAY	O
‐	O
C	O
,	O
OBHS	O
,	O
OBHS	O
‐	O
N	O
,	O
and	O
triaryl	O
‐	O
ethylene	O
analogs	O
,	O
all	O
of	O
which	O
are	O
indirect	O
modulators	O
.	O
	O
For	O
example	O
,	O
3	O
,	O
4	O
‐	O
DTP	O
,	O
furan	O
,	O
and	O
S	O
‐	O
OBHS	O
‐	O
2	O
drove	O
positively	O
correlated	O
GREB1	O
levels	O
and	O
E	O
‐	O
Luc	O
but	O
not	O
L	O
‐	O
Luc	O
ERα	O
‐	O
WT	O
activity	O
(	O
Fig	O
3C	O
lanes	O
5	O
–	O
7	O
).	O
	O
In	O
contrast	O
,	O
WAY	O
dimer	O
and	O
WAY	O
‐	O
D	O
analogs	O
drove	O
positively	O
correlated	O
GREB1	O
levels	O
and	O
L	O
‐	O
Luc	O
ERα	O
‐	O
WT	O
but	O
not	O
E	O
‐	O
Luc	O
activity	O
(	O
Fig	O
3C	O
lanes	O
8	O
and	O
9	O
).	O
	O
This	O
cluster	O
includes	O
two	O
direct	O
modulator	O
scaffolds	O
(	O
OBHS	O
‐	O
ASC	O
and	O
OBHS	O
‐	O
BSC	O
),	O
and	O
five	O
indirect	O
modulator	O
scaffolds	O
(	O
A	O
‐	O
CD	O
,	O
cyclofenil	O
,	O
3	O
,	O
4	O
‐	O
DTPD	O
,	O
imine	O
,	O
and	O
imidazopyridine	O
).	O
	O
These	O
results	O
suggest	O
that	O
addition	O
of	O
an	O
extended	O
side	O
chain	O
to	O
an	O
ERα	O
ligand	O
scaffold	O
is	O
sufficient	O
to	O
induce	O
cell	O
‐	O
specific	O
signaling	O
,	O
where	O
the	O
relative	O
activity	O
profiles	O
of	O
the	O
individual	O
ligands	O
change	O
between	O
cell	O
types	O
.	O
	O
This	O
is	O
demonstrated	O
by	O
directly	O
comparing	O
the	O
signaling	O
specificities	O
of	O
matched	O
OBHS	O
(	O
indirect	O
modulator	O
,	O
cluster	O
1	O
)	O
and	O
OBHS	O
‐	O
BSC	O
analogs	O
(	O
direct	O
modulator	O
,	O
cluster	O
3	O
),	O
which	O
differ	O
only	O
in	O
the	O
basic	O
side	O
chain	O
(	O
Fig	O
2E	O
).	O
	O
The	O
activities	O
of	O
OBHS	O
analogs	O
were	O
positively	O
correlated	O
across	O
the	O
three	O
cell	O
types	O
,	O
but	O
the	O
side	O
chain	O
of	O
OBHS	O
‐	O
BSC	O
analogs	O
was	O
sufficient	O
to	O
abolish	O
these	O
correlations	O
(	O
Figs	O
3C	O
lanes	O
1	O
and	O
19	O
,	O
and	O
EV3A	O
–	O
C	O
).	O
	O
Modulation	O
of	O
signaling	O
specificity	O
by	O
AF	O
‐	O
1	O
	O
To	O
evaluate	O
the	O
role	O
of	O
AF	O
‐	O
1	O
and	O
the	O
F	O
domain	O
in	O
ERα	O
signaling	O
specificity	O
,	O
we	O
compared	O
activity	O
of	O
truncated	O
ERα	O
constructs	O
in	O
HepG2	O
liver	O
cells	O
with	O
endogenous	O
ERα	O
activity	O
in	O
the	O
other	O
cell	O
types	O
.	O
	O
The	O
positive	O
correlation	O
between	O
the	O
L	O
‐	O
Luc	O
and	O
E	O
‐	O
Luc	O
activities	O
or	O
GREB1	O
levels	O
induced	O
by	O
scaffolds	O
in	O
cluster	O
1	O
was	O
generally	O
retained	O
without	O
the	O
AB	O
domain	O
,	O
or	O
the	O
F	O
domain	O
(	O
Fig	O
3D	O
lanes	O
1	O
–	O
4	O
).	O
	O
This	O
demonstrates	O
that	O
the	O
signaling	O
specificities	O
underlying	O
these	O
positive	O
correlations	O
are	O
not	O
modified	O
by	O
AF	O
‐	O
1	O
.	O
	O
OBHS	O
analogs	O
showed	O
an	O
average	O
L	O
‐	O
Luc	O
ERα	B-mutant
‐	I-mutant
ΔAB	I-mutant
activity	O
of	O
3	O
.	O
2	O
%	O
±	O
3	O
(	O
mean	O
+	O
SEM	O
)	O
relative	O
to	O
E2	O
.	O
	O
Despite	O
this	O
nearly	O
complete	O
lack	O
of	O
activity	O
,	O
the	O
pattern	O
of	O
L	O
‐	O
Luc	O
ERα	B-mutant
‐	I-mutant
ΔAB	I-mutant
activity	O
was	O
still	O
highly	O
correlated	O
with	O
the	O
E	O
‐	O
Luc	O
activity	O
and	O
GREB1	O
expression	O
(	O
Fig	O
EV3D	O
and	O
E	O
),	O
demonstrating	O
that	O
very	O
small	O
AF	O
‐	O
2	O
activities	O
can	O
be	O
amplified	O
by	O
AF	O
‐	O
1	O
to	O
produce	O
robust	O
signals	O
.	O
	O
Similarly	O
,	O
deletion	O
of	O
the	O
F	O
domain	O
did	O
not	O
abolish	O
correlations	O
between	O
the	O
L	O
‐	O
Luc	O
and	O
E	O
‐	O
Luc	O
or	O
GREB1	O
levels	O
induced	O
by	O
OBHS	O
analogs	O
(	O
Fig	O
EV3F	O
).	O
	O
These	O
similar	O
patterns	O
of	O
ligand	O
activity	O
in	O
the	O
wild	O
‐	O
type	O
and	O
deletion	O
mutants	O
suggest	O
that	O
AF	O
‐	O
1	O
and	O
the	O
F	O
domain	O
purely	O
amplify	O
the	O
AF	O
‐	O
2	O
activities	O
of	O
ligands	O
in	O
cluster	O
1	O
.	O
	O
Comparing	O
Fig	O
3C	O
and	O
D	O
,	O
the	O
+	O
and	O
−	O
signs	O
indicate	O
where	O
the	O
deletion	O
mutant	O
assays	O
led	O
to	O
a	O
gain	O
or	O
loss	O
of	O
statically	O
significant	O
correlation	O
,	O
respectively	O
.	O
	O
Thus	O
,	O
ligands	O
in	O
cluster	O
2	O
rely	O
on	O
AF	O
‐	O
1	O
for	O
both	O
activity	O
(	O
Fig	O
3B	O
)	O
and	O
signaling	O
specificity	O
(	O
Fig	O
3D	O
).	O
	O
Ligand	O
‐	O
specific	O
control	O
of	O
GREB1	O
expression	O
	O
In	O
cluster	O
1	O
,	O
the	O
recruitment	O
of	O
NCOA1	O
and	O
NCOA2	O
was	O
highest	O
for	O
WAY	O
‐	O
C	O
,	O
followed	O
by	O
triaryl	O
‐	O
ethylene	O
,	O
OBHS	O
‐	O
N	O
,	O
and	O
OBHS	O
series	O
,	O
while	O
for	O
NCOA3	O
,	O
OBHS	O
‐	O
N	O
compounds	O
induced	O
the	O
most	O
recruitment	O
and	O
OBHS	O
ligands	O
were	O
inverse	O
agonists	O
(	O
Fig	O
EV2F	O
–	O
H	O
).	O
	O
The	O
average	O
induction	O
of	O
GREB1	O
by	O
cluster	O
1	O
ligands	O
showed	O
greater	O
variance	O
,	O
with	O
a	O
range	O
between	O
~	O
25	O
and	O
~	O
75	O
%	O
for	O
OBHS	O
and	O
a	O
range	O
from	O
full	O
agonist	O
to	O
inverse	O
agonist	O
for	O
the	O
others	O
in	O
cluster	O
1	O
(	O
Fig	O
EV2A	O
).	O
	O
GREB1	O
levels	O
induced	O
by	O
OBHS	O
analogs	O
were	O
determined	O
by	O
recruitment	O
of	O
NCOA1	O
but	O
not	O
NCOA2	O
/	O
3	O
(	O
Fig	O
3E	O
lane	O
1	O
),	O
suggesting	O
that	O
there	O
may	O
be	O
alternate	O
or	O
preferential	O
use	O
of	O
these	O
coactivators	O
by	O
different	O
classes	O
.	O
	O
However	O
,	O
in	O
cluster	O
1	O
,	O
NCOA1	O
/	O
2	O
/	O
3	O
recruitment	O
generally	O
predicted	O
GREB1	O
levels	O
(	O
Fig	O
3E	O
lanes	O
1	O
–	O
4	O
),	O
consistent	O
with	O
the	O
canonical	O
signaling	O
model	O
(	O
Fig	O
1D	O
).	O
	O
Direct	O
modulators	O
showed	O
low	O
NCOA1	O
/	O
2	O
/	O
3	O
recruitment	O
(	O
Fig	O
EV2F	O
–	O
H	O
),	O
but	O
only	O
OBHS	O
‐	O
ASC	O
analogs	O
had	O
NCOA2	O
recruitment	O
profiles	O
that	O
predicted	O
a	O
full	O
range	O
of	O
effects	O
on	O
GREB1	O
levels	O
(	O
Figs	O
3E	O
lanes	O
9	O
,	O
11	O
,	O
18	O
–	O
19	O
,	O
and	O
EV2A	O
).	O
	O
The	O
indirect	O
modulators	O
in	O
clusters	O
2	O
and	O
3	O
stimulated	O
NCOA1	O
/	O
2	O
/	O
3	O
recruitment	O
and	O
GREB1	O
expression	O
with	O
substantial	O
variance	O
(	O
Figs	O
3A	O
and	O
EV2F	O
–	O
H	O
).	O
	O
These	O
results	O
suggest	O
that	O
compounds	O
that	O
show	O
cell	O
‐	O
specific	O
signaling	O
do	O
not	O
activate	O
GREB1	O
,	O
or	O
use	O
coactivators	O
other	O
than	O
NCOA1	O
/	O
2	O
/	O
3	O
to	O
control	O
GREB1	O
expression	O
(	O
Fig	O
1E	O
).	O
	O
In	O
cluster	O
1	O
,	O
E	O
‐	O
Luc	O
and	O
L	O
‐	O
Luc	O
activities	O
,	O
NCOA1	O
/	O
2	O
/	O
3	O
recruitment	O
,	O
and	O
GREB1	O
levels	O
generally	O
predicted	O
the	O
proliferative	O
response	O
(	O
Fig	O
3F	O
lanes	O
2	O
–	O
4	O
).	O
	O
Despite	O
this	O
phenotypic	O
variance	O
,	O
proliferation	O
was	O
not	O
generally	O
predicted	O
by	O
correlated	O
NCOA1	O
/	O
2	O
/	O
3	O
recruitment	O
and	O
GREB1	O
induction	O
(	O
Figs	O
3F	O
lanes	O
5	O
–	O
19	O
,	O
and	O
EV3H	O
).	O
	O
For	O
ligands	O
that	O
show	O
cell	O
‐	O
specific	O
signaling	O
,	O
ERα	O
‐	O
mediated	O
recruitment	O
of	O
other	O
coregulators	O
and	O
activation	O
of	O
other	O
target	O
genes	O
likely	O
determine	O
their	O
proliferative	O
effects	O
on	O
MCF	O
‐	O
7	O
cells	O
.	O
	O
NCOA3	O
occupancy	O
at	O
GREB1	O
did	O
not	O
predict	O
the	O
proliferative	O
response	O
	O
We	O
also	O
questioned	O
whether	O
promoter	O
occupancy	O
by	O
coactivators	O
is	O
statistically	O
robust	O
and	O
reproducible	O
for	O
ligand	O
class	O
analysis	O
using	O
a	O
chromatin	O
immunoprecipitation	O
(	O
ChIP	O
)‐	O
based	O
quantitative	O
assay	O
,	O
and	O
whether	O
it	O
has	O
a	O
better	O
predictive	O
power	O
than	O
the	O
M2H	O
assay	O
.	O
	O
ERα	O
and	O
NCOA3	O
cycle	O
on	O
and	O
off	O
the	O
GREB1	O
promoter	O
(	O
Nwachukwu	O
et	O
al	O
,	O
2014	O
).	O
	O
At	O
this	O
time	O
point	O
,	O
other	O
WAY	O
‐	O
C	O
analogs	O
also	O
induced	O
recruitment	O
of	O
NCOA3	O
at	O
this	O
site	O
to	O
varying	O
degrees	O
(	O
Fig	O
4B	O
).	O
	O
The	O
Z	O
’	O
for	O
this	O
assay	O
was	O
0	O
.	O
6	O
,	O
showing	O
statistical	O
robustness	O
(	O
see	O
Materials	O
and	O
Methods	O
).	O
	O
We	O
prepared	O
biological	O
replicates	O
with	O
different	O
cell	O
passage	O
numbers	O
and	O
separately	O
prepared	O
samples	O
,	O
which	O
showed	O
r	O
2	O
of	O
0	O
.	O
81	O
,	O
demonstrating	O
high	O
reproducibility	O
(	O
Fig	O
4C	O
).	O
	O
Kinetic	O
ChIP	O
assay	O
examining	O
recruitment	O
of	O
NCOA3	O
to	O
the	O
GREB1	O
gene	O
in	O
MCF	O
‐	O
7	O
cells	O
stimulated	O
with	O
E2	O
or	O
the	O
indicated	O
WAY	O
‐	O
C	O
analog	O
.	O
	O
NCOA3	O
occupancy	O
at	O
GREB1	O
was	O
compared	O
by	O
ChIP	O
assay	O
45	O
min	O
after	O
stimulation	O
with	O
vehicle	O
,	O
E2	O
,	O
or	O
the	O
WAY	O
‐	O
C	O
analogs	O
.	O
	O
In	O
panel	O
(	O
B	O
),	O
the	O
average	O
recruitment	O
of	O
two	O
biological	O
replicates	O
are	O
shown	O
as	O
mean	O
+	O
SEM	O
,	O
and	O
the	O
Z	O
‐	O
score	O
is	O
indicated	O
.	O
	O
In	O
panel	O
(	O
C	O
),	O
correlation	O
analysis	O
was	O
performed	O
for	O
two	O
biological	O
replicates	O
.	O
	O
Linear	O
regression	O
analyses	O
comparing	O
the	O
ability	O
of	O
NCOA3	O
recruitment	O
,	O
measured	O
by	O
ChIP	O
or	O
M2H	O
,	O
to	O
predict	O
other	O
agonist	O
activities	O
of	O
WAY	O
‐	O
C	O
analogs	O
.	O
*	O
Significant	O
positive	O
correlation	O
(	O
F	O
‐	O
test	O
for	O
nonzero	O
slope	O
,	O
P	O
‐	O
value	O
).	O
	O
The	O
M2H	O
assay	O
for	O
NCOA3	O
recruitment	O
broadly	O
correlated	O
with	O
the	O
other	O
assays	O
,	O
and	O
was	O
predictive	O
for	O
GREB1	O
expression	O
and	O
cell	O
proliferation	O
(	O
Fig	O
3E	O
).	O
	O
Thus	O
,	O
the	O
simplified	O
coactivator	O
‐	O
binding	O
assay	O
showed	O
much	O
greater	O
predictive	O
power	O
than	O
the	O
ChIP	O
assay	O
for	O
ligand	O
‐	O
specific	O
effects	O
on	O
GREB1	O
expression	O
and	O
cell	O
proliferation	O
.	O
	O
ERβ	O
activity	O
is	O
not	O
an	O
independent	O
predictor	O
of	O
cell	O
‐	O
specific	O
activity	O
	O
One	O
difference	O
between	O
MCF	O
‐	O
7	O
breast	O
cancer	O
cells	O
and	O
Ishikawa	O
endometrial	O
cancer	O
cells	O
is	O
the	O
contribution	O
of	O
ERβ	O
to	O
estrogenic	O
response	O
,	O
as	O
Ishikawa	O
cells	O
may	O
express	O
ERβ	O
(	O
Bhat	O
&	O
Pezzuto	O
,	O
2001	O
).	O
	O
When	O
overexpressed	O
in	O
MCF	O
‐	O
7	O
cells	O
,	O
ERβ	O
alters	O
E2	O
‐	O
induced	O
expression	O
of	O
only	O
a	O
subset	O
of	O
ERα	O
‐	O
target	O
genes	O
(	O
Wu	O
et	O
al	O
,	O
2011	O
),	O
raising	O
the	O
possibility	O
that	O
ligand	O
‐	O
induced	O
ERβ	O
activity	O
may	O
contribute	O
to	O
E	O
‐	O
Luc	O
activities	O
,	O
and	O
thus	O
underlie	O
the	O
lack	O
of	O
correlation	O
between	O
the	O
E	O
‐	O
Luc	O
and	O
L	O
‐	O
Luc	O
ERα	O
‐	O
WT	O
activities	O
or	O
GREB1	O
levels	O
induced	O
by	O
cell	O
‐	O
specific	O
modulators	O
in	O
cluster	O
2	O
and	O
cluster	O
3	O
(	O
Fig	O
3C	O
).	O
	O
To	O
test	O
this	O
idea	O
,	O
we	O
determined	O
the	O
L	O
‐	O
Luc	O
ERβ	O
activity	O
profiles	O
of	O
the	O
ligands	O
(	O
Fig	O
EV1	O
).	O
	O
For	O
most	O
scaffolds	O
,	O
L	O
‐	O
Luc	O
ERβ	O
and	O
E	O
‐	O
Luc	O
activities	O
were	O
not	O
correlated	O
,	O
except	O
for	O
2	O
,	O
5	O
‐	O
DTP	O
and	O
cyclofenil	O
analogs	O
,	O
which	O
showed	O
moderate	O
but	O
significant	O
correlations	O
(	O
Fig	O
EV4A	O
).	O
	O
Nevertheless	O
,	O
the	O
E	O
‐	O
Luc	O
activities	O
of	O
both	O
2	O
,	O
5	O
‐	O
DTP	O
and	O
cyclofenil	O
analogs	O
were	O
better	O
predicted	O
by	O
their	O
L	O
‐	O
Luc	O
ERα	O
‐	O
WT	O
than	O
L	O
‐	O
Luc	O
ERβ	O
activities	O
(	O
Fig	O
EV4A	O
and	O
B	O
).	O
	O
ERβ	O
activity	O
in	O
HepG2	O
cells	O
rarely	O
correlates	O
with	O
E	O
‐	O
Luc	O
activity	O
.	O
	O
Data	O
information	O
:	O
The	O
r	O
2	O
and	O
P	O
values	O
for	O
the	O
indicated	O
correlations	O
are	O
shown	O
in	O
both	O
panels	O
.	O
*	O
Significant	O
positive	O
correlation	O
(	O
F	O
‐	O
test	O
for	O
nonzero	O
slope	O
,	O
P	O
‐	O
value	O
)	O
	O
To	O
overcome	O
barriers	O
to	O
crystallization	O
of	O
ERα	O
LBD	O
complexes	O
,	O
we	O
developed	O
a	O
conformation	O
‐	O
trapping	O
X	O
‐	O
ray	O
crystallography	O
approach	O
using	O
the	O
ERα	B-mutant
‐	I-mutant
Y537S	I-mutant
mutation	O
(	O
Nettles	O
et	O
al	O
,	O
2008	O
;	O
Bruning	O
et	O
al	O
,	O
2010	O
;	O
Srinivasan	O
et	O
al	O
,	O
2013	O
).	O
	O
Eleven	O
of	O
these	O
structures	O
have	O
been	O
published	O
,	O
while	O
65	O
are	O
new	O
,	O
including	O
the	O
DES	O
‐	O
bound	O
ERα	B-mutant
‐	I-mutant
Y537S	I-mutant
LBD	O
.	O
	O
We	O
present	O
57	O
of	O
these	O
new	O
structures	O
here	O
(	O
Dataset	O
EV2	O
),	O
while	O
the	O
remaining	O
eight	O
new	O
structures	O
bound	O
to	O
OBHS	O
‐	O
N	O
analogs	O
will	O
be	O
published	O
elsewhere	O
(	O
S	O
.	O
Srinivasan	O
et	O
al	O
,	O
in	O
preparation	O
).	O
	O
Examining	O
many	O
closely	O
related	O
structures	O
allows	O
us	O
to	O
visualize	O
subtle	O
structural	O
differences	O
,	O
in	O
effect	O
using	O
X	O
‐	O
ray	O
crystallography	O
as	O
a	O
systems	O
biology	O
tool	O
.	O
	O
Based	O
on	O
our	O
original	O
OBHS	O
structure	O
,	O
the	O
OBHS	O
,	O
OBHS	O
‐	O
N	O
,	O
and	O
triaryl	O
‐	O
ethylene	O
compounds	O
were	O
modified	O
with	O
h11	O
‐	O
directed	O
pendant	O
groups	O
(	O
Zheng	O
et	O
al	O
,	O
2012	O
;	O
Zhu	O
et	O
al	O
,	O
2012	O
;	O
Liao	O
et	O
al	O
,	O
2014	O
).	O
	O
Superposing	O
the	O
LBDs	O
based	O
on	O
the	O
class	O
of	O
bound	O
ligands	O
provides	O
an	O
ensemble	O
view	O
of	O
the	O
structural	O
variance	O
and	O
clarifies	O
what	O
part	O
of	O
the	O
ligand	O
‐	O
binding	O
pocket	O
is	O
differentially	O
perturbed	O
or	O
targeted	O
.	O
	O
For	O
the	O
triaryl	O
‐	O
ethylene	O
analogs	O
,	O
the	O
displacement	O
of	O
h11	O
was	O
in	O
a	O
perpendicular	O
direction	O
,	O
away	O
from	O
Ile424	O
in	O
h8	O
and	O
toward	O
h12	O
.	O
	O
Structure	O
‐	O
class	O
analysis	O
of	O
triaryl	O
‐	O
ethylene	O
analogs	O
.	O
	O
Triaryl	O
‐	O
ethylene	O
analogs	O
bound	O
to	O
the	O
superposed	O
crystal	O
structures	O
of	O
the	O
ERα	O
LBD	O
are	O
shown	O
.	O
	O
Arrows	O
indicate	O
chemical	O
variance	O
in	O
the	O
orientation	O
of	O
the	O
different	O
h11	O
‐	O
directed	O
ligand	O
side	O
groups	O
(	O
PDB	O
5DK9	O
,	O
5DKB	O
,	O
5DKE	O
,	O
5DKG	O
,	O
5DKS	O
,	O
5DL4	O
,	O
5DLR	O
,	O
5DMC	O
,	O
5DMF	O
and	O
5DP0	O
).	O
	O
Triaryl	O
‐	O
ethylene	O
analogs	O
induce	O
variance	O
of	O
ERα	O
conformations	O
at	O
the	O
C	O
‐	O
terminal	O
region	O
of	O
h11	O
.	O
	O
Panel	O
(	O
B	O
)	O
shows	O
the	O
crystal	O
structure	O
of	O
a	O
triaryl	O
‐	O
ethylene	O
analog	O
‐	O
bound	O
ERα	O
LBD	O
(	O
PDB	O
5DLR	O
).	O
	O
This	O
region	O
was	O
expanded	O
in	O
panel	O
(	O
C	O
),	O
where	O
the	O
10	O
triaryl	O
‐	O
ethylene	O
analog	O
‐	O
bound	O
ERα	O
LBD	O
structures	O
(	O
see	O
Datasets	O
EV1	O
and	O
EV2	O
)	O
were	O
superposed	O
to	O
show	O
variations	O
in	O
the	O
h11	O
C	O
‐	O
terminus	O
(	O
PDB	O
5DK9	O
,	O
5DKB	O
,	O
5DKE	O
,	O
5DKG	O
,	O
5DKS	O
,	O
5DL4	O
,	O
5DLR	O
,	O
5DMC	O
,	O
5DMF	O
,	O
and	O
5DP0	O
).	O
	O
ERα	O
LBDs	O
in	O
complex	O
with	O
diethylstilbestrol	O
(	O
DES	O
)	O
or	O
a	O
triaryl	O
‐	O
ethylene	O
analog	O
were	O
superposed	O
to	O
show	O
that	O
the	O
ligand	O
‐	O
induced	O
difference	O
in	O
h11	O
conformation	O
is	O
transmitted	O
to	O
the	O
C	O
‐	O
terminus	O
of	O
h12	O
(	O
PDB	O
4ZN7	O
,	O
5DMC	O
).	O
	O
Inter	O
‐	O
atomic	O
distances	O
predict	O
the	O
proliferative	O
effects	O
of	O
specific	O
ligand	O
series	O
.	O
	O
Ile424	O
–	O
His524	O
distance	O
measured	O
in	O
the	O
crystal	O
structures	O
correlates	O
with	O
the	O
proliferative	O
effect	O
of	O
triaryl	O
‐	O
ethylene	O
analogs	O
in	O
MCF	O
‐	O
7	O
cells	O
.	O
	O
In	O
contrast	O
,	O
the	O
Leu354	O
–	O
Leu525	O
distance	O
correlates	O
with	O
the	O
proliferative	O
effects	O
of	O
OBHS	O
‐	O
N	O
analogs	O
in	O
MCF	O
‐	O
7	O
cells	O
.	O
	O
Structure	O
‐	O
class	O
analysis	O
of	O
WAY	O
‐	O
C	O
analogs	O
.	O
	O
WAY	O
‐	O
C	O
side	O
groups	O
subtly	O
nudge	O
h12	O
Leu540	O
.	O
	O
ERα	O
LBD	O
structures	O
bound	O
to	O
4	O
distinct	O
WAY	O
‐	O
C	O
analogs	O
were	O
superposed	O
(	O
PDB	O
4	O
IU7	O
,	O
4IV4	O
,	O
4IVW	O
,	O
4IW6	O
)	O
(	O
see	O
Datasets	O
EV1	O
and	O
EV2	O
).	O
	O
Structure	O
‐	O
class	O
analysis	O
of	O
indirect	O
modulators	O
	O
Structure	O
‐	O
class	O
analysis	O
of	O
indirect	O
modulators	O
in	O
cluster	O
1	O
.	O
	O
Crystal	O
structures	O
of	O
the	O
ERα	O
LBD	O
bound	O
to	O
OBHS	O
and	O
OBHS	O
‐	O
N	O
analogs	O
were	O
superposed	O
.	O
	O
Arrows	O
indicate	O
chemical	O
variance	O
in	O
the	O
orientation	O
of	O
the	O
different	O
h11	O
‐	O
directed	O
ligand	O
side	O
groups	O
.	O
	O
Panel	O
(	O
B	O
)	O
shows	O
the	O
ligand	O
‐	O
induced	O
conformational	O
variation	O
at	O
the	O
C	O
‐	O
terminal	O
region	O
of	O
h11	O
(	O
OBHS	O
:	O
PDB	O
4ZN9	O
,	O
4ZNH	O
,	O
4ZNS	O
,	O
4ZNT	O
,	O
4ZNU	O
,	O
4ZNV	O
,	O
and	O
4ZNW	O
;	O
OBHS	O
‐	O
N	O
:	O
PDB	O
4ZUB	O
,	O
4ZUC	O
,	O
4ZWH	O
,	O
4ZWK	O
,	O
5BNU	O
,	O
5BP6	O
,	O
5BPR	O
,	O
and	O
5BQ4	O
).	O
	O
Structure	O
‐	O
class	O
analysis	O
of	O
indirect	O
modulators	O
in	O
clusters	O
2	O
and	O
3	O
.	O
	O
Crystal	O
structures	O
of	O
the	O
ERα	O
LBD	O
bound	O
to	O
ligands	O
with	O
cell	O
‐	O
specific	O
activities	O
were	O
superposed	O
.	O
	O
The	O
bound	O
ligands	O
are	O
shown	O
,	O
and	O
arrows	O
indicate	O
considerable	O
variation	O
in	O
the	O
orientation	O
of	O
the	O
different	O
h3	O
‐,	O
h8	O
‐,	O
h11	O
‐,	O
or	O
h12	O
‐	O
directed	O
ligand	O
side	O
groups	O
.	O
	O
As	O
visualized	O
in	O
four	O
LBD	O
structures	O
(	O
Srinivasan	O
et	O
al	O
,	O
2013	O
),	O
WAY	O
‐	O
C	O
analogs	O
were	O
designed	O
with	O
small	O
substitutions	O
that	O
slightly	O
nudge	O
h12	O
Leu540	O
,	O
without	O
exiting	O
the	O
ligand	O
‐	O
binding	O
pocket	O
(	O
Fig	O
5G	O
and	O
H	O
).	O
	O
Therefore	O
,	O
changing	O
h12	O
dynamics	O
maintains	O
the	O
canonical	O
signaling	O
pathway	O
defined	O
by	O
E2	O
(	O
Fig	O
1D	O
)	O
to	O
support	O
AF	O
‐	O
2	O
‐	O
driven	O
signaling	O
and	O
recruit	O
NCOA1	O
/	O
2	O
/	O
3	O
for	O
GREB1	O
‐	O
stimulated	O
proliferation	O
.	O
	O
Ligands	O
with	O
cell	O
‐	O
specific	O
activity	O
alter	O
the	O
shape	O
of	O
the	O
AF	O
‐	O
2	O
surface	O
	O
These	O
structures	O
demonstrated	O
that	O
cell	O
‐	O
specific	O
activity	O
derived	O
from	O
altering	O
the	O
shape	O
of	O
the	O
AF	O
‐	O
2	O
surface	O
without	O
an	O
extended	O
side	O
chain	O
.	O
	O
For	O
instance	O
,	O
S	O
‐	O
OBHS	O
‐	O
2	O
and	O
S	O
‐	O
OBHS	O
‐	O
3	O
analogs	O
(	O
Fig	O
2	O
)	O
had	O
similar	O
ERα	O
activity	O
profiles	O
in	O
the	O
different	O
cell	O
types	O
(	O
Fig	O
EV2A	O
–	O
C	O
),	O
but	O
the	O
2	O
‐	O
versus	O
3	O
‐	O
methyl	O
substituted	O
phenol	O
rings	O
altered	O
the	O
correlated	O
signaling	O
patterns	O
in	O
different	O
cell	O
types	O
(	O
Fig	O
3B	O
lanes	O
7	O
and	O
12	O
).	O
	O
This	O
difference	O
in	O
ligand	O
positioning	O
altered	O
the	O
AF	O
‐	O
2	O
surface	O
via	O
a	O
shift	O
in	O
the	O
N	O
‐	O
terminus	O
of	O
h12	O
,	O
which	O
directly	O
contacts	O
the	O
coactivator	O
.	O
	O
This	O
effect	O
is	O
evident	O
in	O
a	O
single	O
structure	O
due	O
to	O
its	O
1	O
Å	O
magnitude	O
(	O
Fig	O
6A	O
and	O
B	O
).	O
	O
The	O
shifts	O
in	O
h12	O
residues	O
Asp538	O
and	O
Leu539	O
led	O
to	O
rotation	O
of	O
the	O
coactivator	O
peptide	O
(	O
Fig	O
6C	O
).	O
	O
Thus	O
,	O
cell	O
‐	O
specific	O
activity	O
can	O
stem	O
from	O
perturbation	O
of	O
the	O
AF	O
‐	O
2	O
surface	O
without	O
an	O
extended	O
side	O
chain	O
,	O
which	O
presumably	O
alters	O
the	O
receptor	O
–	O
coregulator	O
interaction	O
profile	O
.	O
	O
S	O
‐	O
OBHS	O
‐	O
2	O
/	O
3	O
analogs	O
subtly	O
distort	O
the	O
AF	O
‐	O
2	O
surface	O
.	O
	O
Panel	O
(	O
A	O
)	O
shows	O
the	O
crystal	O
structure	O
of	O
an	O
S	O
‐	O
OBHS	O
‐	O
3	O
‐	O
bound	O
ERα	O
LBD	O
(	O
PDB	O
5DUH	O
).	O
	O
The	O
h3	O
–	O
h12	O
interface	O
(	O
circled	O
)	O
at	O
AF	O
‐	O
2	O
(	O
pink	O
)	O
was	O
expanded	O
in	O
panels	O
(	O
B	O
,	O
C	O
).	O
	O
The	O
S	O
‐	O
OBHS	O
‐	O
2	O
/	O
3	O
‐	O
bound	O
ERα	O
LBDs	O
were	O
superposed	O
to	O
show	O
shifts	O
in	O
h3	O
(	O
panel	O
B	O
)	O
and	O
the	O
NCOA2	O
peptide	O
docked	O
at	O
the	O
AF	O
‐	O
2	O
surface	O
(	O
panel	O
C	O
).	O
	O
Crystal	O
structures	O
show	O
that	O
2	O
,	O
5	O
‐	O
DTP	O
analogs	O
shift	O
h3	O
and	O
h11	O
further	O
apart	O
compared	O
to	O
an	O
A	O
‐	O
CD	O
‐	O
ring	O
estrogen	O
(	O
PDB	O
4PPS	O
,	O
5DRM	O
,	O
5DRJ	O
).	O
	O
Average	O
(	O
mean	O
+	O
SEM	O
)	O
α	O
‐	O
carbon	O
distance	O
measured	O
from	O
h3	O
Thr347	O
to	O
h11	O
Leu525	O
of	O
A	O
‐	O
CD	O
‐,	O
2	O
,	O
5	O
‐	O
DTP	O
‐,	O
and	O
3	O
,	O
4	O
‐	O
DTPD	O
‐	O
bound	O
ERα	O
LBDs	O
.	O
	O
*	O
Two	O
‐	O
tailed	O
Student	O
'	O
s	O
t	O
‐	O
test	O
,	O
P	O
=	O
0	O
.	O
002	O
(	O
PDB	O
A	O
‐	O
CD	O
:	O
5DI7	O
,	O
5DID	O
,	O
5DIE	O
,	O
5DIG	O
,	O
and	O
4PPS	O
;	O
2	O
,	O
5	O
‐	O
DTP	O
:	O
4IWC	O
,	O
5DRM	O
,	O
and	O
5DRJ	O
;	O
3	O
,	O
4	O
‐	O
DTPD	O
:	O
5DTV	O
and	O
5DU5	O
).	O
	O
Crystal	O
structures	O
show	O
that	O
a	O
3	O
,	O
4	O
‐	O
DTPD	O
analog	O
shifts	O
h3	O
(	O
F	O
)	O
and	O
the	O
NCOA2	O
(	O
G	O
)	O
peptide	O
compared	O
to	O
an	O
A	O
‐	O
CD	O
‐	O
ring	O
estrogen	O
(	O
PDB	O
4PPS	O
,	O
5DTV	O
).	O
	O
Hierarchical	O
clustering	O
of	O
ligand	O
‐	O
specific	O
binding	O
of	O
154	O
interacting	O
peptides	O
to	O
the	O
ERα	O
LBD	O
was	O
performed	O
in	O
triplicate	O
by	O
MARCoNI	O
analysis	O
.	O
	O
These	O
compounds	O
bind	O
the	O
LBD	O
in	O
an	O
unusual	O
fashion	O
because	O
they	O
have	O
a	O
phenol	O
‐	O
to	O
‐	O
phenol	O
length	O
of	O
~	O
12	O
Å	O
,	O
which	O
is	O
longer	O
than	O
steroids	O
and	O
other	O
prototypical	O
ERα	O
agonists	O
that	O
are	O
~	O
10	O
Å	O
in	O
length	O
.	O
	O
One	O
phenol	O
pushed	O
further	O
toward	O
h3	O
(	O
Fig	O
6D	O
),	O
while	O
the	O
other	O
phenol	O
pushed	O
toward	O
the	O
C	O
‐	O
terminus	O
of	O
h11	O
to	O
a	O
greater	O
extent	O
than	O
A	O
‐	O
CD	O
‐	O
ring	O
estrogens	O
(	O
Nwachukwu	O
et	O
al	O
,	O
2014	O
),	O
which	O
are	O
close	O
structural	O
analogs	O
of	O
E2	O
that	O
lack	O
a	O
B	O
‐	O
ring	O
(	O
Fig	O
2	O
).	O
	O
To	O
quantify	O
this	O
difference	O
,	O
we	O
compared	O
the	O
distance	O
between	O
α	O
‐	O
carbons	O
at	O
h3	O
Thr347	O
and	O
h11	O
Leu525	O
in	O
the	O
set	O
of	O
structures	O
containing	O
2	O
,	O
5	O
‐	O
DTP	O
analogs	O
(	O
n	O
=	O
3	O
)	O
or	O
A	O
‐	O
CD	O
‐	O
ring	O
analogs	O
(	O
n	O
=	O
5	O
)	O
(	O
Fig	O
6E	O
).	O
	O
We	O
observed	O
a	O
difference	O
of	O
0	O
.	O
4	O
Å	O
that	O
was	O
significant	O
(	O
two	O
‐	O
tailed	O
Student	O
'	O
s	O
t	O
‐	O
test	O
,	O
P	O
=	O
0	O
.	O
002	O
)	O
due	O
to	O
the	O
very	O
tight	O
clustering	O
of	O
the	O
2	O
,	O
5	O
‐	O
DTP	O
‐	O
induced	O
LBD	O
conformation	O
.	O
	O
The	O
2	O
,	O
5	O
‐	O
DTP	O
and	O
3	O
,	O
4	O
‐	O
DTP	O
scaffolds	O
are	O
isomeric	O
,	O
but	O
with	O
aryl	O
groups	O
at	O
obtuse	O
and	O
acute	O
angles	O
,	O
respectively	O
(	O
Fig	O
2	O
).	O
	O
The	O
crystal	O
structure	O
of	O
ERα	O
in	O
complex	O
with	O
a	O
3	O
,	O
4	O
‐	O
DTP	O
is	O
unknown	O
;	O
however	O
,	O
we	O
solved	O
two	O
crystal	O
structures	O
of	O
ERα	O
bound	O
to	O
3	O
,	O
4	O
‐	O
DTPD	O
analogs	O
and	O
one	O
structure	O
containing	O
a	O
furan	O
ligand	O
—	O
all	O
of	O
which	O
have	O
a	O
3	O
,	O
4	O
‐	O
diaryl	O
configuration	O
(	O
Fig	O
2	O
;	O
Datasets	O
EV1	O
and	O
EV2	O
).	O
	O
In	O
these	O
structures	O
,	O
the	O
A	O
‐	O
ring	O
mimetic	O
of	O
the	O
3	O
,	O
4	O
‐	O
DTPD	O
scaffold	O
bound	O
h3	O
Glu353	O
as	O
expected	O
,	O
but	O
the	O
other	O
phenol	O
wrapped	O
around	O
h3	O
to	O
form	O
a	O
hydrogen	O
bond	O
with	O
Thr347	O
,	O
indicating	O
a	O
change	O
in	O
binding	O
epitopes	O
in	O
the	O
ERα	O
ligand	O
‐	O
binding	O
pocket	O
(	O
Fig	O
6F	O
).	O
	O
The	O
3	O
,	O
4	O
‐	O
DTPD	O
analogs	O
also	O
induced	O
a	O
shift	O
in	O
h3	O
positioning	O
,	O
which	O
translated	O
again	O
into	O
a	O
shift	O
in	O
the	O
bound	O
coactivator	O
peptide	O
(	O
Fig	O
6F	O
).	O
	O
To	O
test	O
whether	O
the	O
AF	O
‐	O
2	O
surface	O
shows	O
changes	O
in	O
shape	O
in	O
solution	O
,	O
we	O
used	O
the	O
microarray	O
assay	O
for	O
real	O
‐	O
time	O
coregulator	O
–	O
nuclear	O
receptor	O
interaction	O
(	O
MARCoNI	O
)	O
analysis	O
(	O
Aarts	O
et	O
al	O
,	O
2013	O
).	O
	O
Here	O
,	O
the	O
ligand	O
‐	O
dependent	O
interactions	O
of	O
the	O
ERα	O
LBD	O
with	O
over	O
150	O
distinct	O
LxxLL	O
motif	O
peptides	O
were	O
assayed	O
to	O
define	O
structural	O
fingerprints	O
for	O
the	O
AF	O
‐	O
2	O
surface	O
,	O
in	O
a	O
manner	O
similar	O
to	O
the	O
use	O
of	O
phage	O
display	O
peptides	O
as	O
structural	O
probes	O
(	O
Connor	O
et	O
al	O
,	O
2001	O
).	O
	O
Despite	O
the	O
similar	O
average	O
activities	O
of	O
these	O
ligand	O
classes	O
(	O
Fig	O
3A	O
and	O
B	O
),	O
2	O
,	O
5	O
‐	O
DTP	O
and	O
3	O
,	O
4	O
‐	O
DTP	O
analogs	O
displayed	O
remarkably	O
different	O
peptide	O
recruitment	O
patterns	O
(	O
Fig	O
6H	O
),	O
consistent	O
with	O
the	O
structural	O
analyses	O
.	O
	O
However	O
,	O
there	O
was	O
a	O
unique	O
cluster	O
of	O
peptides	O
that	O
were	O
recruited	O
by	O
E2	O
but	O
not	O
the	O
2	O
,	O
5	O
‐	O
DTP	O
analogs	O
.	O
	O
In	O
contrast	O
,	O
3	O
,	O
4	O
‐	O
DTP	O
analogs	O
dismissed	O
most	O
of	O
the	O
peptides	O
from	O
the	O
AF	O
‐	O
2	O
surface	O
(	O
Fig	O
6H	O
).	O
	O
Thus	O
,	O
the	O
isomeric	O
attachment	O
of	O
diaryl	O
groups	O
to	O
the	O
thiophene	O
core	O
changed	O
the	O
AF	O
‐	O
2	O
surface	O
from	O
inside	O
the	O
ligand	O
‐	O
binding	O
pocket	O
,	O
as	O
predicted	O
by	O
the	O
crystal	O
structures	O
.	O
	O
Together	O
,	O
these	O
findings	O
suggest	O
that	O
without	O
an	O
extended	O
side	O
chain	O
,	O
cell	O
‐	O
specific	O
activity	O
stems	O
from	O
different	O
coregulator	O
recruitment	O
profiles	O
,	O
due	O
to	O
unique	O
ligand	O
‐	O
induced	O
conformations	O
of	O
the	O
AF	O
‐	O
2	O
surface	O
,	O
in	O
addition	O
to	O
differential	O
usage	O
of	O
AF	O
‐	O
1	O
.	O
	O
Indirect	O
modulators	O
in	O
cluster	O
1	O
avoid	O
this	O
by	O
perturbing	O
the	O
h11	O
–	O
h12	O
interface	O
,	O
and	O
modulating	O
the	O
dynamics	O
of	O
h12	O
without	O
changing	O
the	O
shape	O
of	O
AF	O
‐	O
2	O
when	O
stabilized	O
.	O
	O
Our	O
goal	O
was	O
to	O
identify	O
a	O
minimal	O
set	O
of	O
predictors	O
that	O
would	O
link	O
specific	O
structural	O
perturbations	O
to	O
ERα	O
signaling	O
pathways	O
that	O
control	O
cell	O
‐	O
specific	O
signaling	O
and	O
proliferation	O
.	O
	O
We	O
found	O
a	O
very	O
strong	O
set	O
of	O
predictors	O
,	O
where	O
ligands	O
in	O
cluster	O
1	O
,	O
defined	O
by	O
similar	O
signaling	O
across	O
cell	O
types	O
,	O
showed	O
indirect	O
modulation	O
of	O
h12	O
dynamics	O
via	O
the	O
h11	O
–	O
12	O
interface	O
or	O
slight	O
contact	O
with	O
h12	O
.	O
	O
This	O
perturbation	O
determined	O
proliferation	O
that	O
correlated	O
strongly	O
with	O
AF	O
‐	O
2	O
activity	O
,	O
recruitment	O
of	O
NCOA1	O
/	O
2	O
/	O
3	O
family	O
members	O
,	O
and	O
induction	O
of	O
the	O
GREB1	O
gene	O
,	O
consistent	O
with	O
the	O
canonical	O
ERα	O
signaling	O
pathway	O
(	O
Fig	O
1D	O
).	O
	O
For	O
ligands	O
in	O
cluster	O
1	O
,	O
deletion	O
of	O
AF	O
‐	O
1	O
reduced	O
activity	O
to	O
varying	O
degrees	O
,	O
but	O
did	O
not	O
change	O
the	O
underlying	O
signaling	O
patterns	O
established	O
through	O
AF	O
‐	O
2	O
.	O
	O
Compared	O
to	O
cluster	O
1	O
,	O
the	O
structural	O
rules	O
are	O
less	O
clear	O
in	O
clusters	O
2	O
and	O
3	O
,	O
but	O
a	O
number	O
of	O
indirect	O
modulator	O
classes	O
perturbed	O
the	O
LBD	O
conformation	O
at	O
the	O
intersection	O
of	O
h3	O
,	O
the	O
h12	O
N	O
‐	O
terminus	O
,	O
and	O
the	O
AF	O
‐	O
2	O
surface	O
.	O
	O
Ligands	O
in	O
these	O
classes	O
altered	O
the	O
shape	O
of	O
AF	O
‐	O
2	O
to	O
affect	O
coregulator	O
preferences	O
.	O
	O
For	O
direct	O
and	O
indirect	O
modulators	O
in	O
cluster	O
2	O
or	O
3	O
,	O
the	O
canonical	O
ERα	O
signaling	O
pathway	O
involving	O
recruitment	O
of	O
NCOA1	O
/	O
2	O
/	O
3	O
and	O
induction	O
of	O
GREB1	O
did	O
not	O
generally	O
predict	O
their	O
proliferative	O
effects	O
,	O
indicating	O
an	O
alternate	O
causal	O
model	O
(	O
Fig	O
1E	O
).	O
	O
These	O
principles	O
outlined	O
above	O
provide	O
a	O
structural	O
basis	O
for	O
how	O
the	O
ligand	O
–	O
receptor	O
interface	O
leads	O
to	O
different	O
signaling	O
specificities	O
through	O
AF	O
‐	O
1	O
and	O
AF	O
‐	O
2	O
.	O
	O
It	O
is	O
noteworthy	O
that	O
regulation	O
of	O
h12	O
dynamics	O
indirectly	O
through	O
h11	O
can	O
virtually	O
abolish	O
AF	O
‐	O
2	O
activity	O
,	O
and	O
yet	O
still	O
drive	O
robust	O
transcriptional	O
activity	O
through	O
AF	O
‐	O
1	O
,	O
as	O
demonstrated	O
with	O
the	O
OBHS	O
series	O
.	O
	O
This	O
finding	O
can	O
be	O
explained	O
by	O
the	O
fact	O
that	O
NCOA1	O
/	O
2	O
/	O
3	O
contain	O
distinct	O
binding	O
sites	O
for	O
interaction	O
with	O
AF	O
‐	O
1	O
and	O
AF	O
‐	O
2	O
(	O
McInerney	O
et	O
al	O
,	O
1996	O
;	O
Webb	O
et	O
al	O
,	O
1998	O
),	O
which	O
allows	O
ligands	O
to	O
nucleate	O
ERα	O
–	O
NCOA1	O
/	O
2	O
/	O
3	O
interaction	O
through	O
AF	O
‐	O
2	O
,	O
and	O
reinforce	O
this	O
interaction	O
with	O
additional	O
binding	O
to	O
AF	O
‐	O
1	O
.	O
	O
Completely	O
blocking	O
AF	O
‐	O
2	O
with	O
an	O
extended	O
side	O
chain	O
or	O
altering	O
the	O
shape	O
of	O
AF	O
‐	O
2	O
changes	O
the	O
preference	O
away	O
from	O
NCOA1	O
/	O
2	O
/	O
3	O
for	O
determining	O
GREB1	O
levels	O
and	O
proliferation	O
of	O
breast	O
cancer	O
cells	O
.	O
	O
AF	O
‐	O
2	O
blockade	O
also	O
allows	O
AF	O
‐	O
1	O
to	O
function	O
independently	O
,	O
which	O
is	O
important	O
since	O
AF	O
‐	O
1	O
drives	O
tissue	O
‐	O
selective	O
effects	O
in	O
vivo	O
.	O
	O
This	O
was	O
demonstrated	O
with	O
AF	O
‐	O
1	O
knockout	O
mice	O
that	O
show	O
E2	O
‐	O
dependent	O
vascular	O
protection	O
,	O
but	O
not	O
uterine	O
proliferation	O
,	O
thus	O
highlighting	O
the	O
role	O
of	O
AF	O
‐	O
1	O
in	O
tissue	O
‐	O
selective	O
or	O
cell	O
‐	O
specific	O
signaling	O
(	O
Billon	O
‐	O
Gales	O
et	O
al	O
,	O
2009	O
;	O
Abot	O
et	O
al	O
,	O
2013	O
).	O
	O
Here	O
,	O
we	O
examined	O
many	O
LBD	O
structures	O
and	O
tested	O
several	O
variables	O
that	O
were	O
not	O
predictive	O
,	O
including	O
ERβ	O
activity	O
,	O
the	O
strength	O
of	O
AF	O
‐	O
1	O
signaling	O
,	O
and	O
NCOA3	O
occupancy	O
at	O
the	O
GREB1	O
gene	O
.	O
	O
Similarly	O
,	O
we	O
visualized	O
structures	O
to	O
identify	O
patterns	O
.	O
	O
For	O
example	O
,	O
phage	O
display	O
was	O
used	O
to	O
identify	O
the	O
androgen	O
receptor	O
interactome	O
,	O
which	O
was	O
cloned	O
into	O
an	O
M2H	O
library	O
and	O
used	O
to	O
identify	O
clusters	O
of	O
ligand	O
‐	O
selective	O
interactions	O
(	O
Norris	O
et	O
al	O
,	O
2009	O
).	O
	O
We	O
have	O
identified	O
atomic	O
vectors	O
for	O
the	O
OBHS	O
‐	O
N	O
and	O
triaryl	O
‐	O
ethylene	O
classes	O
that	O
predict	O
ligand	O
response	O
(	O
Fig	O
5E	O
and	O
F	O
).	O
	O
Indeed	O
,	O
the	O
most	O
anti	O
‐	O
proliferative	O
compound	O
in	O
the	O
OBHS	O
‐	O
N	O
series	O
had	O
a	O
fulvestrant	O
‐	O
like	O
profile	O
across	O
a	O
battery	O
of	O
assays	O
(	O
S	O
.	O
Srinivasan	O
et	O
al	O
,	O
in	O
preparation	O
).	O
	O
Secondly	O
,	O
our	O
finding	O
that	O
WAY	O
‐	O
C	O
compounds	O
do	O
not	O
rely	O
of	O
AF	O
‐	O
1	O
for	O
signaling	O
efficacy	O
may	O
derive	O
from	O
the	O
slight	O
contacts	O
with	O
h12	O
observed	O
in	O
crystal	O
structures	O
(	O
Figs	O
3B	O
and	O
5H	O
),	O
unlike	O
other	O
compounds	O
in	O
cluster	O
1	O
that	O
dislocate	O
h11	O
and	O
rely	O
on	O
AF	O
‐	O
1	O
for	O
signaling	O
efficacy	O
(	O
Figs	O
3B	O
and	O
5C	O
,	O
and	O
EV5B	O
).	O
	O
Investigation	O
of	O
the	O
Interaction	O
between	O
Cdc42	O
and	O
Its	O
Effector	O
TOCA1	O
	O
Transducer	O
of	O
Cdc42	O
-	O
dependent	O
actin	O
assembly	O
protein	O
1	O
(	O
TOCA1	O
)	O
is	O
an	O
effector	O
of	O
the	O
Rho	O
family	O
small	O
G	O
protein	O
Cdc42	O
.	O
	O
It	O
contains	O
a	O
membrane	O
-	O
deforming	O
F	O
-	O
BAR	O
domain	O
as	O
well	O
as	O
a	O
Src	O
homology	O
3	O
(	O
SH3	O
)	O
domain	O
and	O
a	O
G	O
protein	O
-	O
binding	O
homology	O
region	O
1	O
(	O
HR1	O
)	O
domain	O
.	O
	O
TOCA1	O
binding	O
to	O
Cdc42	O
leads	O
to	O
actin	O
rearrangements	O
,	O
which	O
are	O
thought	O
to	O
be	O
involved	O
in	O
processes	O
such	O
as	O
endocytosis	O
,	O
filopodia	O
formation	O
,	O
and	O
cell	O
migration	O
.	O
	O
We	O
have	O
also	O
investigated	O
the	O
binding	O
of	O
the	O
TOCA	O
HR1	O
domain	O
to	O
Cdc42	O
and	O
the	O
potential	O
ternary	O
complex	O
between	O
Cdc42	O
and	O
the	O
G	O
protein	O
-	O
binding	O
regions	O
of	O
TOCA1	O
and	O
a	O
member	O
of	O
the	O
Wiskott	O
-	O
Aldrich	O
syndrome	O
protein	O
family	O
,	O
N	O
-	O
WASP	O
.	O
	O
TOCA1	O
binds	O
Cdc42	O
with	O
micromolar	O
affinity	O
,	O
in	O
contrast	O
to	O
the	O
nanomolar	O
affinity	O
of	O
the	O
N	O
-	O
WASP	O
G	O
protein	O
-	O
binding	O
region	O
for	O
Cdc42	O
.	O
	O
The	O
Ras	O
superfamily	O
of	O
small	O
GTPases	O
comprises	O
over	O
150	O
members	O
that	O
regulate	O
a	O
multitude	O
of	O
cellular	O
processes	O
in	O
eukaryotes	O
.	O
	O
All	O
members	O
share	O
a	O
well	O
defined	O
core	O
structure	O
of	O
∼	O
20	O
kDa	O
known	O
as	O
the	O
G	O
domain	O
,	O
which	O
is	O
responsible	O
for	O
guanine	O
nucleotide	O
binding	O
.	O
	O
The	O
guanine	O
nucleotide	O
exchange	O
factors	O
mediate	O
formation	O
of	O
the	O
active	O
state	O
by	O
promoting	O
the	O
dissociation	O
of	O
GDP	O
,	O
allowing	O
GTP	O
to	O
bind	O
.	O
	O
The	O
GTPase	O
-	O
activating	O
proteins	O
stimulate	O
the	O
rate	O
of	O
intrinsic	O
GTP	O
hydrolysis	O
,	O
mediating	O
the	O
return	O
to	O
the	O
inactive	O
state	O
(	O
reviewed	O
in	O
Ref	O
.).	O
	O
The	O
overall	O
conformation	O
of	O
small	O
G	O
proteins	O
in	O
the	O
active	O
and	O
inactive	O
states	O
is	O
similar	O
,	O
but	O
they	O
differ	O
significantly	O
in	O
two	O
main	O
regions	O
known	O
as	O
switch	O
I	O
and	O
switch	O
II	O
.	O
	O
The	O
structures	O
of	O
more	O
than	O
60	O
small	O
G	O
protein	O
·	O
effector	O
complexes	O
have	O
been	O
solved	O
,	O
and	O
,	O
not	O
surprisingly	O
,	O
the	O
switch	O
regions	O
have	O
been	O
implicated	O
in	O
a	O
large	O
proportion	O
of	O
the	O
G	O
protein	O
-	O
effector	O
interactions	O
(	O
reviewed	O
in	O
Ref	O
.).	O
	O
However	O
,	O
because	O
each	O
of	O
the	O
150	O
members	O
of	O
the	O
superfamily	O
interacts	O
with	O
multiple	O
effectors	O
,	O
there	O
are	O
still	O
a	O
huge	O
number	O
of	O
known	O
G	O
protein	O
-	O
effector	O
interactions	O
that	O
have	O
not	O
yet	O
been	O
studied	O
structurally	O
.	O
	O
The	O
Rho	O
family	O
comprises	O
20	O
members	O
,	O
of	O
which	O
three	O
,	O
RhoA	O
,	O
Rac1	O
,	O
and	O
Cdc42	O
,	O
have	O
been	O
relatively	O
well	O
studied	O
.	O
	O
N	O
-	O
WASP	O
exists	O
in	O
an	O
autoinhibited	O
conformation	O
,	O
which	O
is	O
released	O
upon	O
PI	O
(	O
4	O
,	O
5	O
)	O
P2	O
and	O
Cdc42	O
binding	O
or	O
by	O
other	O
factors	O
,	O
such	O
as	O
phosphorylation	O
.	O
	O
The	O
importance	O
of	O
TOCA1	O
in	O
actin	O
polymerization	O
has	O
been	O
demonstrated	O
in	O
a	O
range	O
of	O
in	O
vitro	O
and	O
in	O
vivo	O
studies	O
,	O
but	O
the	O
exact	O
role	O
of	O
TOCA1	O
in	O
the	O
many	O
pathways	O
involving	O
actin	O
assembly	O
remains	O
unclear	O
.	O
	O
The	O
most	O
widely	O
studied	O
role	O
of	O
TOCA1	O
is	O
in	O
membrane	O
invagination	O
and	O
endocytosis	O
,	O
although	O
it	O
has	O
also	O
been	O
implicated	O
in	O
filopodia	O
formation	O
,	O
neurite	O
elongation	O
,	O
transcriptional	O
reprogramming	O
via	O
nuclear	O
actin	O
,	O
and	O
interaction	O
with	O
ZO	O
-	O
1	O
at	O
tight	O
junctions	O
.	O
	O
TOCA1	O
comprises	O
an	O
N	O
-	O
terminal	O
F	O
-	O
BAR	O
domain	O
,	O
a	O
central	O
homology	O
region	O
1	O
(	O
HR1	O
)	O
domain	O
,	O
and	O
a	O
C	O
-	O
terminal	O
SH3	O
domain	O
.	O
	O
The	O
F	O
-	O
BAR	O
domain	O
is	O
a	O
known	O
dimerization	O
,	O
membrane	O
-	O
binding	O
,	O
and	O
membrane	O
-	O
deforming	O
module	O
found	O
in	O
a	O
number	O
of	O
cell	O
signaling	O
proteins	O
.	O
	O
Other	O
HR1	O
domains	O
studied	O
so	O
far	O
,	O
including	O
those	O
from	O
the	O
PRK	O
family	O
,	O
have	O
been	O
found	O
to	O
bind	O
their	O
cognate	O
Rho	O
family	O
G	O
protein	O
-	O
binding	O
partner	O
with	O
high	O
specificity	O
and	O
affinities	O
in	O
the	O
nanomolar	O
range	O
.	O
	O
The	O
structures	O
of	O
the	O
PRK1	O
HR1a	O
domain	O
in	O
complex	O
with	O
RhoA	O
and	O
the	O
HR1b	O
domain	O
in	O
complex	O
with	O
Rac1	O
show	O
that	O
the	O
HR1	O
domain	O
comprises	O
an	O
anti	O
-	O
parallel	O
coiled	O
-	O
coil	O
that	O
interacts	O
with	O
its	O
G	O
protein	O
binding	O
partner	O
via	O
both	O
helices	O
.	O
	O
The	O
coiled	O
-	O
coil	O
fold	O
is	O
shared	O
by	O
the	O
HR1	O
domain	O
of	O
the	O
TOCA	O
family	O
protein	O
,	O
CIP4	O
,	O
and	O
,	O
based	O
on	O
sequence	O
homology	O
,	O
by	O
TOCA1	O
itself	O
.	O
	O
These	O
HR1	O
domains	O
,	O
however	O
,	O
show	O
specificity	O
for	O
Cdc42	O
,	O
rather	O
than	O
RhoA	O
or	O
Rac1	O
.	O
	O
How	O
different	O
HR1	O
domain	O
proteins	O
distinguish	O
their	O
specific	O
G	O
protein	O
partners	O
remains	O
only	O
partially	O
understood	O
,	O
and	O
structural	O
characterization	O
of	O
a	O
novel	O
G	O
protein	O
-	O
HR1	O
domain	O
interaction	O
would	O
add	O
to	O
the	O
growing	O
body	O
of	O
information	O
pertaining	O
to	O
these	O
protein	O
complexes	O
.	O
	O
Furthermore	O
,	O
the	O
biological	O
function	O
of	O
the	O
interaction	O
between	O
TOCA1	O
and	O
Cdc42	O
remains	O
poorly	O
understood	O
,	O
and	O
so	O
far	O
there	O
has	O
been	O
no	O
biophysical	O
or	O
structural	O
insight	O
.	O
	O
There	O
is	O
some	O
evidence	O
for	O
a	O
ternary	O
complex	O
between	O
Cdc42	O
,	O
N	O
-	O
WASP	O
,	O
and	O
TOCA1	O
,	O
but	O
there	O
was	O
no	O
direct	O
demonstration	O
of	O
simultaneous	O
contacts	O
between	O
the	O
two	O
effectors	O
and	O
a	O
single	O
molecule	O
of	O
Cdc42	O
.	O
	O
Nonetheless	O
,	O
the	O
substantial	O
difference	O
between	O
the	O
structures	O
of	O
the	O
G	O
protein	O
-	O
binding	O
regions	O
of	O
the	O
two	O
effectors	O
is	O
intriguing	O
and	O
implies	O
that	O
they	O
bind	O
to	O
Cdc42	O
quite	O
differently	O
,	O
providing	O
motivation	O
for	O
investigating	O
the	O
possibility	O
that	O
Cdc42	O
can	O
bind	O
both	O
effectors	O
concurrently	O
.	O
	O
WASP	O
interacts	O
with	O
Cdc42	O
via	O
a	O
conserved	O
,	O
unstructured	O
binding	O
motif	O
known	O
as	O
the	O
Cdc42	O
-	O
and	O
Rac	O
-	O
interactive	O
binding	O
region	O
(	O
CRIB	O
),	O
which	O
forms	O
an	O
intermolecular	O
β	O
-	O
sheet	O
,	O
expanding	O
the	O
anti	O
-	O
parallel	O
β2	O
and	O
β3	O
strands	O
of	O
Cdc42	O
.	O
	O
In	O
contrast	O
,	O
the	O
TOCA	O
family	O
proteins	O
are	O
thought	O
to	O
interact	O
via	O
the	O
HR1	O
domain	O
,	O
which	O
may	O
form	O
a	O
triple	O
coiled	O
-	O
coil	O
with	O
switch	O
II	O
of	O
Rac1	O
,	O
like	O
the	O
HR1b	O
domain	O
of	O
PRK1	O
.	O
	O
Here	O
,	O
we	O
present	O
the	O
solution	O
NMR	O
structure	O
of	O
the	O
HR1	O
domain	O
of	O
TOCA1	O
,	O
providing	O
the	O
first	O
structural	O
data	O
for	O
this	O
protein	O
.	O
	O
We	O
also	O
present	O
data	O
pertaining	O
to	O
binding	O
of	O
the	O
TOCA	O
HR1	O
domain	O
to	O
Cdc42	O
,	O
which	O
is	O
the	O
first	O
biophysical	O
description	O
of	O
an	O
HR1	O
domain	O
binding	O
this	O
particular	O
Rho	O
family	O
small	O
G	O
protein	O
.	O
	O
Finally	O
,	O
we	O
investigate	O
the	O
potential	O
ternary	O
complex	O
between	O
Cdc42	O
and	O
the	O
G	O
protein	O
-	O
binding	O
regions	O
of	O
TOCA1	O
and	O
N	O
-	O
WASP	O
,	O
contributing	O
to	O
our	O
understanding	O
of	O
G	O
protein	O
-	O
effector	O
interactions	O
as	O
well	O
as	O
the	O
roles	O
of	O
Cdc42	O
,	O
N	O
-	O
WASP	O
,	O
and	O
TOCA1	O
in	O
the	O
pathways	O
that	O
govern	O
actin	O
dynamics	O
.	O
	O
TOCA1	O
was	O
identified	O
in	O
Xenopus	O
extracts	O
as	O
a	O
protein	O
necessary	O
for	O
Cdc42	O
-	O
dependent	O
actin	O
assembly	O
and	O
was	O
shown	O
to	O
bind	O
to	O
Cdc42	O
·	O
GTPγS	O
but	O
not	O
to	O
Cdc42	O
·	O
GDP	O
or	O
to	O
Rac1	O
and	O
RhoA	O
.	O
Given	O
its	O
homology	O
to	O
other	O
Rho	O
family	O
binding	O
modules	O
,	O
it	O
is	O
likely	O
that	O
the	O
HR1	O
domain	O
of	O
TOCA1	O
is	O
sufficient	O
to	O
bind	O
Cdc42	O
.	O
	O
The	O
C	O
.	O
elegans	O
TOCA1	O
orthologues	O
also	O
bind	O
to	O
Cdc42	O
via	O
their	O
consensus	O
HR1	O
domain	O
.	O
	O
The	O
HR1	O
domains	O
from	O
the	O
PRK	O
family	O
bind	O
their	O
G	O
protein	O
partners	O
with	O
a	O
high	O
affinity	O
,	O
exhibiting	O
a	O
range	O
of	O
submicromolar	O
dissociation	O
constants	O
(	O
Kd	O
)	O
as	O
low	O
as	O
26	O
nm	O
.	O
	O
A	O
Kd	O
in	O
the	O
nanomolar	O
range	O
was	O
therefore	O
expected	O
for	O
the	O
interaction	O
of	O
the	O
TOCA1	O
HR1	O
domain	O
with	O
Cdc42	O
.	O
	O
We	O
generated	O
an	O
X	O
.	O
tropicalis	O
TOCA1	O
HR1	O
domain	O
construct	O
encompassing	O
residues	O
330	O
–	O
426	O
.	O
	O
This	O
region	O
comprises	O
the	O
complete	O
HR1	O
domain	O
based	O
on	O
secondary	O
structure	O
predictions	O
and	O
sequence	O
alignments	O
with	O
another	O
TOCA	O
family	O
member	O
,	O
CIP4	O
,	O
whose	O
structure	O
has	O
been	O
determined	O
.	O
	O
The	O
interaction	O
between	O
[	O
3H	O
]	O
GTP	O
·	O
Cdc42	O
and	O
a	O
C	O
-	O
terminally	O
His	O
-	O
tagged	O
TOCA1	O
HR1	O
domain	O
construct	O
was	O
investigated	O
using	O
SPA	O
.	O
	O
The	O
binding	O
isotherm	O
for	O
the	O
interaction	O
is	O
shown	O
in	O
Fig	O
.	O
1A	O
,	O
together	O
with	O
the	O
Cdc42	O
-	O
PAK	O
interaction	O
as	O
a	O
positive	O
control	O
.	O
	O
The	O
binding	O
of	O
TOCA1	O
HR1	O
to	O
Cdc42	O
was	O
unexpectedly	O
weak	O
,	O
with	O
a	O
Kd	O
of	O
>	O
1	O
μm	O
.	O
	O
It	O
was	O
not	O
possible	O
to	O
estimate	O
the	O
Kd	O
more	O
accurately	O
using	O
direct	O
SPA	O
experiments	O
,	O
because	O
saturation	O
could	O
not	O
be	O
reached	O
due	O
to	O
nonspecific	O
signal	O
at	O
higher	O
protein	O
concentrations	O
.	O
	O
The	O
TOCA1	O
HR1	O
-	O
Cdc42	O
interaction	O
is	O
low	O
affinity	O
.	O
	O
The	O
SPA	O
signal	O
was	O
corrected	O
by	O
subtraction	O
of	O
control	O
data	O
with	O
no	O
GST	B-mutant
-	I-mutant
PAK	I-mutant
or	O
HR1	B-mutant
-	I-mutant
His6	I-mutant
.	O
	O
The	O
Kd	O
values	O
derived	O
for	O
the	O
TOCA1	O
HR1	O
with	O
Cdc42Δ7	B-mutant
and	O
full	O
-	O
length	O
Cdc42	O
were	O
6	O
.	O
05	O
±	O
1	O
.	O
96	O
and	O
5	O
.	O
39	O
±	O
1	O
.	O
69	O
μm	O
,	O
respectively	O
.	O
	O
It	O
was	O
possible	O
that	O
the	O
low	O
affinity	O
observed	O
was	O
due	O
to	O
negative	O
effects	O
of	O
immobilization	O
of	O
the	O
HR1	O
domain	O
,	O
so	O
other	O
methods	O
were	O
employed	O
,	O
which	O
utilized	O
untagged	O
proteins	O
.	O
	O
Isothermal	O
titration	O
calorimetry	O
was	O
carried	O
out	O
,	O
but	O
no	O
heat	O
changes	O
were	O
observed	O
at	O
a	O
range	O
of	O
concentrations	O
and	O
temperatures	O
(	O
data	O
not	O
shown	O
),	O
suggesting	O
that	O
the	O
interaction	O
is	O
predominantly	O
entropically	O
driven	O
.	O
	O
A	O
complex	O
of	O
a	O
GST	O
fusion	O
of	O
the	O
GBD	O
of	O
ACK	O
,	O
which	O
binds	O
with	O
a	O
high	O
affinity	O
to	O
Cdc42	O
,	O
with	O
radiolabeled	O
[	O
3H	O
]	O
GTP	O
·	O
Cdc42	O
was	O
preformed	O
,	O
and	O
the	O
effect	O
of	O
increasing	O
concentrations	O
of	O
untagged	O
TOCA1	O
HR1	O
domain	O
was	O
examined	O
.	O
	O
Competition	O
of	O
GST	B-mutant
-	I-mutant
ACK	I-mutant
GBD	O
bound	O
to	O
[	O
3H	O
]	O
GTP	O
·	O
Cdc42	O
by	O
free	O
ACK	O
GBD	O
was	O
used	O
as	O
a	O
control	O
and	O
to	O
establish	O
the	O
value	O
of	O
background	O
counts	O
when	O
Cdc42	O
is	O
fully	O
displaced	O
.	O
	O
The	O
data	O
were	O
fitted	O
to	O
a	O
binding	O
isotherm	O
describing	O
competition	O
.	O
	O
The	O
Cdc42	O
construct	O
used	O
in	O
the	O
binding	O
assays	O
has	O
seven	O
residues	O
deleted	O
from	O
the	O
C	O
terminus	O
to	O
facilitate	O
purification	O
.	O
	O
As	O
the	O
observed	O
affinity	O
between	O
TOCA1	O
HR1	O
and	O
Cdc42	O
was	O
much	O
lower	O
than	O
expected	O
,	O
we	O
reasoned	O
that	O
the	O
C	O
terminus	O
of	O
Cdc42	O
might	O
be	O
necessary	O
for	O
a	O
high	O
affinity	O
interaction	O
.	O
	O
Thus	O
,	O
the	O
C	O
-	O
terminal	O
region	O
of	O
Cdc42	O
is	O
not	O
required	O
for	O
maximal	O
binding	O
of	O
TOCA1	O
HR1	O
.	O
	O
Indeed	O
,	O
GST	O
pull	O
-	O
downs	O
performed	O
with	O
in	O
vitro	O
translated	O
human	O
TOCA1	O
fragments	O
had	O
suggested	O
that	O
residues	O
N	O
-	O
terminal	O
to	O
the	O
HR1	O
domain	O
may	O
be	O
required	O
to	O
stabilize	O
the	O
HR1	O
domain	O
structure	O
.	O
	O
TOCA1	O
dimerizes	O
via	O
its	O
F	O
-	O
BAR	O
domain	O
,	O
which	O
could	O
also	O
affect	O
Cdc42	O
binding	O
,	O
for	O
example	O
by	O
presenting	O
two	O
HR1	O
domains	O
for	O
Cdc42	O
interactions	O
.	O
	O
Various	O
TOCA1	O
fragments	O
(	O
Fig	O
.	O
2A	O
)	O
were	O
therefore	O
assessed	O
for	O
binding	O
to	O
full	O
-	O
length	O
Cdc42	O
by	O
direct	O
SPA	O
.	O
	O
The	O
isolated	O
F	O
-	O
BAR	O
domain	O
showed	O
no	O
binding	O
to	O
full	O
-	O
length	O
Cdc42	O
(	O
Fig	O
.	O
2B	O
).	O
	O
The	O
HR1	B-mutant
-	I-mutant
SH3	I-mutant
protein	O
could	O
not	O
be	O
purified	O
to	O
homogeneity	O
as	O
a	O
fusion	O
protein	O
,	O
so	O
it	O
was	O
assayed	O
in	O
competition	O
assays	O
after	O
cleavage	O
of	O
the	O
His	O
tag	O
.	O
	O
Taken	O
together	O
,	O
these	O
data	O
suggest	O
that	O
the	O
TOCA1	O
HR1	O
domain	O
is	O
sufficient	O
for	O
maximal	O
binding	O
and	O
that	O
this	O
binding	O
is	O
of	O
a	O
relatively	O
low	O
affinity	O
compared	O
with	O
many	O
other	O
Cdc42	O
·	O
effector	O
complexes	O
.	O
	O
The	O
Cdc42	O
-	O
HR1	O
interaction	O
is	O
of	O
low	O
affinity	O
in	O
the	O
context	O
of	O
full	O
-	O
length	O
protein	O
and	O
in	O
TOCA1	O
paralogues	O
.	O
	O
A	O
,	O
diagram	O
illustrating	O
the	O
TOCA1	O
constructs	O
assayed	O
for	O
Cdc42	O
binding	O
.	O
	O
The	O
SPA	O
signal	O
was	O
corrected	O
by	O
subtraction	O
of	O
control	O
data	O
with	O
no	O
fusion	O
protein	O
.	O
	O
C	O
–	O
E	O
,	O
representative	O
examples	O
of	O
competition	O
SPA	O
experiments	O
carried	O
out	O
with	O
the	O
indicated	O
concentrations	O
of	O
the	O
TOCA1	O
HR1	B-mutant
-	I-mutant
SH3	I-mutant
construct	O
titrated	O
into	O
30	O
nm	O
GST	B-mutant
-	I-mutant
ACK	I-mutant
and	O
30	O
nm	O
Cdc42Δ7Q61L	O
·[	O
3H	O
]	O
GTP	O
(	O
C	O
)	O
or	O
HR1CIP4	O
(	O
D	O
)	O
or	O
HR1FBP17	O
(	O
E	O
)	O
titrated	O
into	O
30	O
nm	O
GST	B-mutant
-	I-mutant
ACK	I-mutant
and	O
30	O
nm	O
Cdc42FLQ61L	O
·[	O
3H	O
]	O
GTP	O
.	O
	O
The	O
low	O
affinity	O
of	O
the	O
TOCA1	O
HR1	O
-	O
Cdc42	O
interaction	O
raised	O
the	O
question	O
of	O
whether	O
the	O
other	O
known	O
Cdc42	O
-	O
binding	O
TOCA	O
family	O
proteins	O
,	O
FBP17	O
and	O
CIP4	O
,	O
also	O
bind	O
weakly	O
.	O
	O
The	O
HR1	O
domains	O
from	O
FBP17	O
and	O
CIP4	O
were	O
purified	O
and	O
assayed	O
for	O
Cdc42	O
binding	O
in	O
competition	O
SPAs	O
,	O
analogous	O
to	O
those	O
carried	O
out	O
with	O
the	O
TOCA1	O
HR1	O
domain	O
.	O
	O
The	O
affinities	O
of	O
both	O
the	O
FBP17	O
and	O
CIP4	O
HR1	O
domains	O
were	O
also	O
in	O
the	O
low	O
micromolar	O
range	O
(	O
10	O
and	O
5	O
μm	O
,	O
respectively	O
)	O
(	O
Fig	O
.	O
2	O
,	O
D	O
and	O
E	O
),	O
suggesting	O
that	O
low	O
affinity	O
interactions	O
with	O
Cdc42	O
are	O
a	O
common	O
feature	O
within	O
the	O
TOCA	O
family	O
.	O
	O
Structure	O
of	O
the	O
TOCA1	O
HR1	O
Domain	O
	O
Because	O
the	O
TOCA1	O
HR1	O
domain	O
was	O
sufficient	O
for	O
maximal	O
Cdc42	O
-	O
binding	O
,	O
we	O
used	O
this	O
construct	O
for	O
structural	O
studies	O
.	O
	O
Initial	O
experiments	O
were	O
performed	O
with	O
TOCA1	O
residues	O
324	O
–	O
426	O
,	O
but	O
we	O
observed	O
that	O
the	O
N	O
terminus	O
was	O
cleaved	O
during	O
purification	O
to	O
yield	O
a	O
new	O
N	O
terminus	O
at	O
residue	O
330	O
(	O
data	O
not	O
shown	O
).	O
	O
We	O
therefore	O
engineered	O
a	O
construct	O
comprising	O
residues	O
330	O
–	O
426	O
to	O
produce	O
the	O
minimal	O
,	O
stable	O
HR1	O
domain	O
.	O
	O
2	O
,	O
778	O
non	O
-	O
degenerate	O
NOE	O
restraints	O
were	O
used	O
in	O
initial	O
structure	O
calculations	O
(	O
1	O
,	O
791	O
unambiguous	O
and	O
987	O
ambiguous	O
),	O
derived	O
from	O
three	O
-	O
dimensional	O
15N	O
-	O
separated	O
NOESY	O
and	O
13C	O
-	O
separated	O
NOESY	O
experiments	O
.	O
	O
100	O
structures	O
were	O
calculated	O
in	O
the	O
final	O
iteration	O
;	O
the	O
50	O
lowest	O
energy	O
structures	O
were	O
water	O
-	O
refined	O
;	O
and	O
of	O
these	O
,	O
the	O
35	O
lowest	O
energy	O
structures	O
were	O
analyzed	O
.	O
	O
Table	O
1	O
indicates	O
that	O
the	O
HR1	O
domain	O
structure	O
is	O
well	O
defined	O
by	O
the	O
NMR	O
data	O
.	O
	O
a	O
<	O
SA	O
>,	O
the	O
average	O
root	O
mean	O
square	O
deviations	O
for	O
the	O
ensemble	O
±	O
S	O
.	O
D	O
.	O
	O
b	O
<	O
SA	O
>	O
c	O
,	O
values	O
for	O
the	O
structure	O
that	O
is	O
closest	O
to	O
the	O
mean	O
.	O
	O
The	O
structure	O
closest	O
to	O
the	O
mean	O
is	O
shown	O
in	O
Fig	O
.	O
3A	O
.	O
	O
The	O
two	O
α	O
-	O
helices	O
of	O
the	O
HR1	O
domain	O
interact	O
to	O
form	O
an	O
anti	O
-	O
parallel	O
coiled	O
-	O
coil	O
with	O
a	O
slight	O
left	O
-	O
handed	O
twist	O
,	O
reminiscent	O
of	O
the	O
HR1	O
domains	O
of	O
CIP4	O
(	O
PDB	O
code	O
2KE4	O
)	O
and	O
PRK1	O
(	O
PDB	O
codes	O
1CXZ	O
and	O
1URF	O
).	O
	O
A	O
,	O
the	O
backbone	O
trace	O
of	O
the	O
35	O
lowest	O
energy	O
structures	O
of	O
the	O
HR1	O
domain	O
overlaid	O
with	O
the	O
structure	O
closest	O
to	O
the	O
mean	O
is	O
shown	O
alongside	O
a	O
schematic	O
representation	O
of	O
the	O
structure	O
closest	O
to	O
the	O
mean	O
.	O
	O
Flexible	O
regions	O
at	O
the	O
N	O
and	O
C	O
termini	O
(	O
residues	O
330	O
–	O
333	O
and	O
421	O
–	O
426	O
)	O
are	O
omitted	O
for	O
clarity	O
.	O
	O
The	O
secondary	O
structure	O
was	O
deduced	O
using	O
Stride	O
,	O
based	O
on	O
the	O
Ramachandran	O
angles	O
,	O
and	O
is	O
indicated	O
as	O
follows	O
:	O
gray	O
,	O
turn	O
;	O
yellow	O
,	O
α	O
-	O
helix	O
;	O
blue	O
,	O
310	O
helix	O
;	O
white	O
,	O
coil	O
.	O
	O
C	O
,	O
a	O
close	O
-	O
up	O
of	O
the	O
N	O
-	O
terminal	O
region	O
of	O
TOCA1	O
HR1	O
,	O
indicating	O
some	O
of	O
the	O
NOEs	O
defining	O
its	O
position	O
with	O
respect	O
to	O
the	O
two	O
α	O
-	O
helices	O
.	O
	O
Dotted	O
lines	O
,	O
NOE	O
restraints	O
.	O
	O
D	O
,	O
a	O
close	O
-	O
up	O
of	O
the	O
interhelix	O
loop	O
region	O
showing	O
some	O
of	O
the	O
contacts	O
between	O
the	O
loop	O
and	O
helix	O
1	O
.	O
	O
In	O
the	O
HR1a	O
domain	O
of	O
PRK1	O
,	O
a	O
region	O
N	O
-	O
terminal	O
to	O
helix	O
1	O
forms	O
a	O
short	O
α	O
-	O
helix	O
,	O
which	O
packs	O
against	O
both	O
helices	O
of	O
the	O
HR1	O
domain	O
.	O
	O
This	O
region	O
of	O
TOCA1	O
HR1	O
(	O
residues	O
334	O
–	O
340	O
)	O
is	O
well	O
defined	O
in	O
the	O
family	O
of	O
structures	O
(	O
Fig	O
.	O
3A	O
)	O
but	O
does	O
not	O
form	O
an	O
α	O
-	O
helix	O
.	O
	O
It	O
instead	O
forms	O
a	O
series	O
of	O
turns	O
,	O
defined	O
by	O
NOE	O
restraints	O
observed	O
between	O
residues	O
separated	O
by	O
one	O
(	O
residues	O
332	O
–	O
334	O
,	O
333	O
–	O
335	O
,	O
etc	O
.)	O
or	O
two	O
(	O
residues	O
337	O
–	O
340	O
)	O
residues	O
in	O
the	O
sequence	O
and	O
the	O
φ	O
and	O
ψ	O
angles	O
,	O
assessed	O
using	O
Stride	O
.	O
	O
These	O
turns	O
cause	O
the	O
chain	O
to	O
reverse	O
direction	O
,	O
allowing	O
the	O
N	O
-	O
terminal	O
segment	O
(	O
residues	O
334	O
–	O
340	O
)	O
to	O
contact	O
both	O
helices	O
of	O
the	O
HR1	O
domain	O
.	O
	O
Long	O
range	O
NOEs	O
were	O
observed	O
linking	O
Leu	O
-	O
334	O
,	O
Glu	O
-	O
335	O
,	O
and	O
Asp	O
-	O
336	O
with	O
Trp	O
-	O
413	O
of	O
helix	O
2	O
,	O
Leu	O
-	O
334	O
with	O
Lys	O
-	O
409	O
of	O
helix	O
2	O
,	O
and	O
Phe	O
-	O
337	O
and	O
Ser	O
-	O
338	O
with	O
Arg	O
-	O
345	O
,	O
Arg	O
-	O
348	O
,	O
and	O
Leu	O
-	O
349	O
of	O
helix	O
1	O
.	O
	O
The	O
two	O
α	O
-	O
helices	O
of	O
TOCA1	O
HR1	O
are	O
separated	O
by	O
a	O
long	O
loop	O
of	O
10	O
residues	O
(	O
residues	O
380	O
–	O
389	O
)	O
that	O
contains	O
two	O
short	O
310	O
helices	O
(	O
residues	O
381	O
–	O
383	O
and	O
386	O
–	O
389	O
).	O
	O
Interestingly	O
,	O
side	O
chains	O
of	O
residues	O
within	O
the	O
loop	O
region	O
point	O
back	O
toward	O
helix	O
1	O
;	O
for	O
example	O
,	O
there	O
are	O
numerous	O
distinct	O
NOEs	O
between	O
the	O
side	O
chains	O
of	O
Asn	O
-	O
380	O
and	O
Met	O
-	O
383	O
of	O
the	O
loop	O
region	O
and	O
Tyr	O
-	O
377	O
and	O
Val	O
-	O
376	O
of	O
helix	O
1	O
(	O
Fig	O
.	O
3D	O
).	O
	O
The	O
backbone	O
NH	O
and	O
CHα	O
groups	O
of	O
Gly	O
-	O
384	O
and	O
Asp	O
-	O
385	O
also	O
show	O
NOEs	O
with	O
the	O
side	O
chain	O
of	O
Tyr	O
-	O
377	O
.	O
	O
Mapping	O
the	O
TOCA1	O
and	O
Cdc42	O
Binding	O
Interfaces	O
	O
The	O
HR1TOCA1	O
-	O
Cdc42	O
interface	O
was	O
investigated	O
using	O
NMR	O
spectroscopy	O
.	O
	O
A	O
comparison	O
of	O
the	O
15N	O
HSQC	O
spectra	O
of	O
free	O
HR1	O
and	O
HR1	O
in	O
the	O
presence	O
of	O
excess	O
Cdc42	O
shows	O
that	O
although	O
some	O
peaks	O
were	O
shifted	O
,	O
several	O
were	O
much	O
broader	O
in	O
the	O
complex	O
,	O
and	O
a	O
considerable	O
subset	O
had	O
disappeared	O
(	O
Fig	O
.	O
4A	O
).	O
	O
This	O
behavior	O
cannot	O
be	O
explained	O
by	O
the	O
increase	O
in	O
molecular	O
mass	O
(	O
from	O
12	O
to	O
33	O
kDa	O
)	O
when	O
Cdc42	O
binds	O
and	O
is	O
more	O
likely	O
to	O
be	O
due	O
to	O
conformational	O
exchange	O
.	O
	O
Overall	O
chemical	O
shift	O
perturbations	O
(	O
CSPs	O
)	O
were	O
calculated	O
for	O
each	O
residue	O
,	O
whereas	O
those	O
that	O
had	O
disappeared	O
were	O
assigned	O
a	O
shift	O
change	O
of	O
0	O
.	O
2	O
(	O
Fig	O
.	O
4B	O
).	O
	O
A	O
peak	O
that	O
disappeared	O
or	O
had	O
a	O
CSP	O
above	O
the	O
mean	O
CSP	O
for	O
the	O
spectrum	O
was	O
considered	O
to	O
be	O
significantly	O
affected	O
.	O
	O
Mapping	O
the	O
binding	O
surface	O
of	O
Cdc42	O
onto	O
the	O
TOCA1	O
HR1	O
domain	O
.	O
	O
A	O
,	O
the	O
15N	O
HSQC	O
of	O
200	O
μm	O
TOCA1	O
HR1	O
domain	O
is	O
shown	O
in	O
the	O
free	O
form	O
(	O
black	O
)	O
and	O
in	O
the	O
presence	O
of	O
a	O
4	O
-	O
fold	O
molar	O
excess	O
of	O
Cdc42Δ7Q61L	O
·	O
GMPPNP	O
(	O
red	O
).	O
	O
The	O
mean	O
CSP	O
is	O
marked	O
with	O
a	O
red	O
line	O
.	O
	O
Those	O
that	O
were	O
not	O
traceable	O
due	O
to	O
spectral	O
overlap	O
were	O
assigned	O
a	O
CSP	O
of	O
zero	O
and	O
are	O
marked	O
with	O
an	O
asterisk	O
below	O
the	O
bar	O
.	O
	O
C	O
,	O
a	O
schematic	O
representation	O
of	O
the	O
HR1	O
domain	O
.	O
	O
Residues	O
with	O
significantly	O
affected	O
backbone	O
and	O
side	O
chain	O
groups	O
that	O
are	O
solvent	O
-	O
accessible	O
are	O
colored	O
red	O
.	O
	O
A	O
close	O
-	O
up	O
of	O
the	O
binding	O
region	O
is	O
shown	O
,	O
with	O
affected	O
side	O
chain	O
heavy	O
atoms	O
shown	O
as	O
sticks	O
.	O
	O
D	O
,	O
the	O
G	O
protein	O
-	O
binding	O
region	O
is	O
marked	O
in	O
red	O
onto	O
structures	O
of	O
the	O
HR1	O
domains	O
as	O
indicated	O
.	O
	O
15N	O
HSQC	O
shift	O
mapping	O
experiments	O
report	O
on	O
changes	O
to	O
amide	O
groups	O
,	O
which	O
are	O
mainly	O
inaccessible	O
because	O
they	O
are	O
buried	O
inside	O
the	O
helices	O
and	O
are	O
involved	O
in	O
hydrogen	O
bonds	O
.	O
	O
Therefore	O
,	O
13C	O
HSQC	O
and	O
methyl	O
-	O
selective	O
SOFAST	O
-	O
HMQC	O
experiments	O
were	O
also	O
recorded	O
on	O
15N	O
,	O
13C	O
-	O
labeled	O
TOCA1	O
HR1	O
to	O
yield	O
more	O
information	O
on	O
side	O
chain	O
involvement	O
.	O
	O
TOCA1	O
residues	O
whose	O
signals	O
were	O
affected	O
by	O
Cdc42	O
binding	O
were	O
mapped	O
onto	O
the	O
structure	O
of	O
TOCA1	O
HR1	O
(	O
Fig	O
.	O
4C	O
).	O
	O
The	O
changes	O
were	O
localized	O
to	O
one	O
end	O
of	O
the	O
coiled	O
-	O
coil	O
,	O
and	O
the	O
binding	O
site	O
appeared	O
to	O
include	O
residues	O
from	O
both	O
α	O
-	O
helices	O
and	O
the	O
loop	O
region	O
that	O
joins	O
them	O
.	O
	O
The	O
residues	O
in	O
the	O
interhelical	O
loop	O
and	O
helix	O
1	O
that	O
contact	O
each	O
other	O
(	O
Fig	O
.	O
3D	O
)	O
show	O
shift	O
changes	O
in	O
their	O
backbone	O
NH	O
and	O
side	O
chains	O
in	O
the	O
presence	O
of	O
Cdc42	O
.	O
	O
For	O
example	O
,	O
the	O
side	O
chain	O
of	O
Asn	O
-	O
380	O
and	O
the	O
backbones	O
of	O
Val	O
-	O
376	O
and	O
Tyr	O
-	O
377	O
were	O
significantly	O
affected	O
but	O
are	O
all	O
buried	O
in	O
the	O
free	O
TOCA1	O
HR1	O
structure	O
,	O
indicating	O
that	O
local	O
conformational	O
changes	O
in	O
the	O
loop	O
may	O
facilitate	O
complex	O
formation	O
.	O
	O
The	O
chemical	O
shift	O
mapping	O
data	O
indicate	O
that	O
the	O
G	O
protein	O
-	O
binding	O
region	O
of	O
the	O
TOCA1	O
HR1	O
domain	O
is	O
broadly	O
similar	O
to	O
that	O
of	O
the	O
CIP4	O
and	O
PRK1	O
HR1	O
domains	O
(	O
Figs	O
.	O
3B	O
and	O
4D	O
).	O
	O
As	O
was	O
the	O
case	O
when	O
labeled	O
HR1	O
was	O
observed	O
,	O
several	O
peaks	O
were	O
shifted	O
in	O
the	O
complex	O
,	O
but	O
many	O
disappeared	O
,	O
indicating	O
exchange	O
on	O
an	O
unfavorable	O
,	O
millisecond	O
time	O
scale	O
(	O
Fig	O
.	O
5A	O
).	O
	O
Detailed	O
side	O
chain	O
data	O
could	O
not	O
be	O
obtained	O
for	O
all	O
residues	O
due	O
to	O
spectral	O
overlap	O
,	O
but	O
constant	O
time	O
13C	O
HSQC	O
and	O
methyl	O
-	O
selective	O
SOFAST	O
-	O
HMQC	O
experiments	O
provided	O
further	O
information	O
on	O
certain	O
well	O
resolved	O
side	O
chains	O
(	O
marked	O
with	O
green	O
asterisks	O
in	O
Fig	O
.	O
5B	O
).	O
	O
Mapping	O
the	O
binding	O
surface	O
of	O
the	O
HR1	O
domain	O
onto	O
Cdc42	O
.	O
	O
B	O
,	O
CSPs	O
are	O
shown	O
for	O
backbone	O
NH	O
groups	O
.	O
	O
Residues	O
with	O
disappeared	O
peaks	O
in	O
13C	O
HSQC	O
experiments	O
are	O
marked	O
on	O
the	O
chart	O
with	O
green	O
asterisks	O
.	O
	O
Residues	O
with	O
either	O
side	O
chain	O
or	O
backbone	O
groups	O
affected	O
are	O
colored	O
blue	O
if	O
buried	O
and	O
yellow	O
if	O
solvent	O
-	O
accessible	O
.	O
	O
The	O
flexible	O
switch	O
regions	O
are	O
circled	O
.	O
	O
As	O
many	O
of	O
the	O
peaks	O
disappeared	O
,	O
the	O
mean	O
chemical	O
shift	O
change	O
was	O
relatively	O
low	O
,	O
so	O
a	O
threshold	O
of	O
the	O
mean	O
plus	O
one	O
S	O
.	O
D	O
.	O
value	O
was	O
used	O
to	O
define	O
a	O
significant	O
CSP	O
.	O
	O
Parts	O
of	O
the	O
switch	O
regions	O
(	O
Fig	O
.	O
5	O
,	O
B	O
and	O
C	O
)	O
are	O
invisible	O
in	O
NMR	O
spectra	O
recorded	O
on	O
free	O
Cdc42	O
due	O
to	O
conformational	O
exchange	O
.	O
	O
These	O
switch	O
regions	O
become	O
visible	O
in	O
Cdc42	O
and	O
other	O
small	O
G	O
protein	O
·	O
effector	O
complexes	O
due	O
to	O
a	O
decrease	O
in	O
conformational	O
freedom	O
upon	O
complex	O
formation	O
.	O
	O
The	O
switch	O
regions	O
of	O
Cdc42	O
did	O
not	O
,	O
however	O
,	O
become	O
visible	O
in	O
the	O
presence	O
of	O
the	O
TOCA1	O
HR1	O
domain	O
.	O
	O
Indeed	O
,	O
Ser	O
-	O
30	O
of	O
switch	O
I	O
and	O
Arg	O
-	O
66	O
,	O
Arg	O
-	O
68	O
,	O
Leu	O
-	O
70	O
,	O
and	O
Ser	O
-	O
71	O
of	O
switch	O
II	O
are	O
visible	O
in	O
free	O
Cdc42	O
but	O
disappear	O
in	O
the	O
presence	O
of	O
the	O
HR1	O
domain	O
.	O
	O
Nevertheless	O
,	O
mapping	O
of	O
the	O
affected	O
residues	O
onto	O
the	O
NMR	O
structure	O
of	O
free	O
Cdc42Δ7Q61L	O
·	O
GMPPNP	O
(	O
Fig	O
.	O
5C	O
)	O
8	O
shows	O
that	O
,	O
although	O
they	O
are	O
relatively	O
widespread	O
compared	O
with	O
changes	O
in	O
the	O
HR1	O
domain	O
,	O
in	O
general	O
,	O
they	O
are	O
on	O
the	O
face	O
of	O
the	O
protein	O
that	O
includes	O
the	O
switches	O
.	O
	O
Although	O
the	O
binding	O
interface	O
may	O
be	O
overestimated	O
,	O
this	O
suggests	O
that	O
the	O
switch	O
regions	O
are	O
involved	O
in	O
binding	O
to	O
TOCA1	O
.	O
	O
The	O
Cdc42	O
·	O
HR1TOCA1	O
complex	O
was	O
not	O
amenable	O
to	O
full	O
structural	O
analysis	O
due	O
to	O
the	O
weak	O
interaction	O
and	O
the	O
extensive	O
exchange	O
broadening	O
seen	O
in	O
the	O
NMR	O
experiments	O
.	O
	O
The	O
orientation	O
of	O
the	O
HR1	O
domain	O
with	O
respect	O
to	O
Cdc42	O
cannot	O
be	O
definitively	O
concluded	O
in	O
the	O
absence	O
of	O
unambiguous	O
distance	O
restraints	O
;	O
hence	O
,	O
HADDOCK	O
produced	O
a	O
set	O
of	O
models	O
in	O
which	O
the	O
HR1	O
domain	O
contacts	O
the	O
same	O
surface	O
on	O
Cdc42	O
but	O
is	O
in	O
various	O
orientations	O
with	O
respect	O
to	O
Cdc42	O
.	O
	O
The	O
cluster	O
with	O
the	O
lowest	O
root	O
mean	O
square	O
deviation	O
from	O
the	O
lowest	O
energy	O
structure	O
is	O
assumed	O
to	O
be	O
the	O
best	O
model	O
.	O
	O
By	O
these	O
criteria	O
,	O
in	O
the	O
best	O
model	O
,	O
the	O
HR1	O
domain	O
is	O
in	O
a	O
similar	O
orientation	O
to	O
the	O
HR1a	O
domain	O
of	O
PRK1	O
bound	O
to	O
RhoA	O
and	O
the	O
HR1b	O
domain	O
bound	O
to	O
Rac1	O
.	O
	O
A	O
representative	O
model	O
from	O
this	O
cluster	O
is	O
shown	O
in	O
Fig	O
.	O
6A	O
alongside	O
the	O
Rac1	O
-	O
HR1b	O
structure	O
(	O
PDB	O
code	O
2RMK	O
)	O
in	O
Fig	O
.	O
6B	O
.	O
	O
Model	O
of	O
Cdc42	O
·	O
HR1	O
complex	O
.	O
	O
A	O
,	O
a	O
representative	O
model	O
of	O
the	O
Cdc42	O
·	O
HR1	O
complex	O
from	O
the	O
cluster	O
closest	O
to	O
the	O
lowest	O
energy	O
model	O
produced	O
using	O
HADDOCK	O
.	O
	O
Residues	O
of	O
Cdc42	O
that	O
are	O
affected	O
in	O
the	O
presence	O
of	O
the	O
HR1	O
domain	O
but	O
are	O
not	O
in	O
close	O
proximity	O
to	O
it	O
are	O
colored	O
in	O
red	O
and	O
labeled	O
.	O
	O
B	O
,	O
structure	O
of	O
Rac1	O
in	O
complex	O
with	O
the	O
HR1b	O
domain	O
of	O
PRK1	O
(	O
PDB	O
code	O
2RMK	O
).	O
	O
C	O
,	O
sequence	O
alignment	O
of	O
RhoA	O
,	O
Cdc42	O
and	O
Rac1	O
.	O
	O
Contact	O
residues	O
of	O
RhoA	O
and	O
Rac1	O
to	O
PRK1	O
HR1a	O
and	O
HR1b	O
,	O
respectively	O
,	O
are	O
colored	O
cyan	O
.	O
	O
Residues	O
of	O
Cdc42	O
that	O
disappear	O
or	O
show	O
chemical	O
shift	O
changes	O
in	O
the	O
presence	O
of	O
TOCA1	O
are	O
colored	O
cyan	O
if	O
also	O
identified	O
as	O
contacts	O
in	O
RhoA	O
and	O
Rac1	O
and	O
yellow	O
if	O
they	O
are	O
not	O
.	O
	O
A	O
sequence	O
alignment	O
of	O
RhoA	O
,	O
Cdc42	O
,	O
and	O
Rac1	O
is	O
shown	O
in	O
Fig	O
.	O
6C	O
.	O
	O
The	O
RhoA	O
and	O
Rac1	O
contact	O
residues	O
in	O
the	O
switch	O
regions	O
are	O
invisible	O
in	O
the	O
spectra	O
of	O
Cdc42	O
,	O
but	O
they	O
are	O
generally	O
conserved	O
between	O
all	O
three	O
G	O
proteins	O
.	O
	O
Several	O
Cdc42	O
residues	O
identified	O
by	O
chemical	O
shift	O
mapping	O
are	O
not	O
in	O
close	O
contact	O
in	O
the	O
Cdc42	O
·	O
TOCA1	O
model	O
(	O
Fig	O
.	O
6A	O
).	O
	O
Other	O
residues	O
that	O
are	O
affected	O
in	O
the	O
Cdc42	O
·	O
TOCA1	O
complex	O
but	O
that	O
do	O
not	O
correspond	O
to	O
contact	O
residues	O
of	O
RhoA	O
or	O
Rac1	O
(	O
Fig	O
.	O
6C	O
)	O
include	O
Gln	O
-	O
2Cdc42	O
,	O
Lys	O
-	O
16Cdc42	O
,	O
Thr	O
-	O
52Cdc42	O
,	O
and	O
Arg	O
-	O
68Cdc42	O
.	O
	O
From	O
the	O
known	O
interactions	O
and	O
effects	O
of	O
the	O
proteins	O
in	O
biological	O
systems	O
,	O
it	O
has	O
been	O
suggested	O
that	O
TOCA1	O
and	O
N	O
-	O
WASP	O
could	O
bind	O
Cdc42	O
simultaneously	O
.	O
	O
Studies	O
in	O
CHO	O
cells	O
indicated	O
that	O
a	O
Cdc42	O
·	O
N	O
-	O
WASP	O
·	O
TOCA1	O
complex	O
existed	O
because	O
FRET	O
was	O
observed	O
between	O
RFP	O
-	O
TOCA1	O
and	O
GFP	O
-	O
N	O
-	O
WASP	O
,	O
and	O
the	O
efficiency	O
was	O
decreased	O
when	O
an	O
N	O
-	O
WASP	O
mutant	O
was	O
used	O
that	O
no	O
longer	O
binds	O
Cdc42	O
.	O
	O
An	O
overlay	O
of	O
the	O
HADDOCK	O
model	O
of	O
the	O
Cdc42	O
·	O
HR1TOCA1	O
complex	O
and	O
the	O
structure	O
of	O
Cdc42	O
in	O
complex	O
with	O
the	O
GBD	O
of	O
the	O
N	O
-	O
WASP	O
homologue	O
,	O
WASP	O
(	O
PDB	O
code	O
1CEE	O
),	O
shows	O
that	O
the	O
HR1	O
and	O
GBD	O
binding	O
sites	O
only	O
partly	O
overlap	O
,	O
and	O
,	O
therefore	O
,	O
a	O
ternary	O
complex	O
remained	O
possible	O
(	O
Fig	O
.	O
7A	O
).	O
	O
Interestingly	O
,	O
the	O
presence	O
of	O
the	O
TOCA1	O
HR1	O
would	O
not	O
prevent	O
the	O
core	O
CRIB	O
of	O
WASP	O
from	O
binding	O
to	O
Cdc42	O
,	O
although	O
the	O
regions	O
C	O
-	O
terminal	O
to	O
the	O
CRIB	O
that	O
are	O
required	O
for	O
high	O
affinity	O
binding	O
of	O
WASP	O
would	O
interfere	O
sterically	O
with	O
the	O
TOCA1	O
HR1	O
.	O
	O
A	O
basic	O
region	O
in	O
WASP	O
including	O
three	O
lysines	O
(	O
residues	O
230	O
–	O
232	O
),	O
N	O
-	O
terminal	O
to	O
the	O
core	O
CRIB	O
,	O
has	O
been	O
implicated	O
in	O
an	O
electrostatic	O
steering	O
mechanism	O
,	O
and	O
these	O
residues	O
would	O
be	O
free	O
to	O
bind	O
in	O
the	O
presence	O
of	O
TOCA1	O
HR1	O
(	O
Fig	O
.	O
7A	O
).	O
	O
The	O
N	O
-	O
WASP	O
GBD	O
displaces	O
the	O
TOCA1	O
HR1	O
domain	O
.	O
	O
The	O
core	O
CRIB	O
region	O
of	O
WASP	O
is	O
shown	O
in	O
red	O
,	O
whereas	O
its	O
basic	O
region	O
is	O
shown	O
in	O
orange	O
and	O
the	O
C	O
-	O
terminal	O
region	O
required	O
for	O
maximal	O
affinity	O
is	O
shown	O
in	O
cyan	O
.	O
	O
A	O
semitransparent	O
surface	O
representation	O
of	O
Cdc42	O
and	O
WASP	O
is	O
shown	O
overlaid	O
with	O
the	O
schematic	O
.	O
	O
C	O
,	O
Selected	O
regions	O
of	O
the	O
15N	O
HSQC	O
of	O
145	O
μm	O
Cdc42Δ7Q61L	O
·	O
GMPPNP	O
with	O
the	O
indicated	O
ratios	O
of	O
the	O
TOCA1	O
HR1	O
domain	O
,	O
the	O
N	O
-	O
WASP	O
GBD	O
,	O
or	O
both	O
,	O
showing	O
that	O
the	O
TOCA	O
HR1	O
domain	O
does	O
not	O
displace	O
the	O
N	O
-	O
WASP	O
GBD	O
.	O
	O
D	O
,	O
selected	O
regions	O
of	O
the	O
15N	O
HSQC	O
of	O
600	O
μm	O
TOCA1	O
HR1	O
domain	O
in	O
complex	O
with	O
Cdc42	O
in	O
the	O
absence	O
and	O
presence	O
of	O
the	O
N	O
-	O
WASP	O
GBD	O
,	O
showing	O
displacement	O
of	O
Cdc42	O
from	O
the	O
HR1	O
domain	O
by	O
N	O
-	O
WASP	O
.	O
	O
An	O
N	O
-	O
WASP	O
GBD	O
construct	O
was	O
produced	O
,	O
and	O
its	O
affinity	O
for	O
Cdc42	O
was	O
measured	O
by	O
competition	O
SPA	O
(	O
Fig	O
.	O
7B	O
).	O
	O
The	O
Kd	O
that	O
was	O
determined	O
(	O
37	O
nm	O
)	O
is	O
consistent	O
with	O
the	O
previously	O
reported	O
affinity	O
.	O
	O
Unlabeled	O
HR1TOCA1	O
was	O
then	O
added	O
to	O
the	O
Cdc42	O
·	O
N	O
-	O
WASP	O
complex	O
,	O
and	O
no	O
changes	O
were	O
seen	O
,	O
suggesting	O
that	O
the	O
N	O
-	O
WASP	O
GBD	O
was	O
not	O
displaced	O
even	O
in	O
the	O
presence	O
of	O
a	O
5	O
-	O
fold	O
excess	O
of	O
HR1TOCA1	O
.	O
	O
These	O
experiments	O
were	O
recorded	O
at	O
sufficiently	O
high	O
protein	O
concentrations	O
(	O
145	O
μm	O
Cdc42	O
,	O
145	O
μm	O
N	O
-	O
WASP	O
GBD	O
,	O
725	O
μm	O
TOCA1	O
HR1	O
domain	O
)	O
to	O
be	O
far	O
in	O
excess	O
of	O
the	O
Kd	O
values	O
of	O
the	O
individual	O
interactions	O
(	O
TOCA1	O
Kd	O
≈	O
5	O
μm	O
,	O
N	O
-	O
WASP	O
Kd	O
=	O
37	O
nm	O
).	O
	O
Furthermore	O
,	O
15N	O
-	O
TOCA1	O
HR1	O
was	O
monitored	O
in	O
the	O
presence	O
of	O
unlabeled	O
Cdc42Δ7Q61L	O
·	O
GMPPNP	O
(	O
1	O
:	O
1	O
)	O
before	O
and	O
after	O
the	O
addition	O
of	O
0	O
.	O
25	O
and	O
1	O
.	O
0	O
eq	O
of	O
unlabeled	O
N	O
-	O
WASP	O
GBD	O
.	O
	O
When	O
in	O
fast	O
exchange	O
,	O
the	O
NMR	O
signal	O
represents	O
a	O
population	O
-	O
weighted	O
average	O
between	O
free	O
and	O
bound	O
states	O
,	O
so	O
the	O
intermediate	O
spectrum	O
indicates	O
that	O
the	O
population	O
comprises	O
a	O
mixture	O
of	O
free	O
and	O
bound	O
HR1	O
domain	O
.	O
	O
Again	O
,	O
the	O
experiments	O
were	O
recorded	O
on	O
protein	O
samples	O
far	O
in	O
excess	O
of	O
the	O
individual	O
Kd	O
values	O
(	O
600	O
μm	O
each	O
protein	O
).	O
	O
These	O
data	O
indicate	O
that	O
the	O
HR1	O
domain	O
is	O
displaced	O
from	O
Cdc42	O
by	O
N	O
-	O
WASP	O
and	O
that	O
a	O
ternary	O
complex	O
comprising	O
TOCA1	O
HR1	O
,	O
N	O
-	O
WASP	O
GBD	O
,	O
and	O
Cdc42	O
is	O
not	O
formed	O
.	O
	O
To	O
extend	O
these	O
studies	O
to	O
a	O
more	O
complex	O
system	O
and	O
to	O
assess	O
the	O
ability	O
of	O
TOCA1	O
HR1	O
to	O
compete	O
with	O
full	O
-	O
length	O
N	O
-	O
WASP	O
,	O
pyrene	O
actin	O
assays	O
were	O
employed	O
.	O
	O
These	O
assays	O
,	O
described	O
in	O
detail	O
elsewhere	O
,	O
were	O
carried	O
out	O
using	O
pyrene	O
actin	O
-	O
supplemented	O
Xenopus	O
extracts	O
into	O
which	O
exogenous	O
TOCA1	O
HR1	O
domain	O
or	O
N	O
-	O
WASP	O
GBD	O
was	O
added	O
,	O
to	O
assess	O
their	O
effects	O
on	O
actin	O
polymerization	O
.	O
	O
Actin	O
polymerization	O
triggered	O
by	O
the	O
addition	O
of	O
PI	O
(	O
4	O
,	O
5	O
)	O
P2	O
-	O
containing	O
liposomes	O
has	O
previously	O
been	O
shown	O
to	O
depend	O
on	O
TOCA1	O
and	O
N	O
-	O
WASP	O
.	O
	O
The	O
addition	O
of	O
the	O
isolated	O
N	O
-	O
WASP	O
GBD	O
significantly	O
inhibited	O
the	O
polymerization	O
of	O
actin	O
at	O
concentrations	O
as	O
low	O
as	O
100	O
nm	O
and	O
completely	O
abolished	O
polymerization	O
at	O
higher	O
concentrations	O
(	O
Fig	O
.	O
8	O
).	O
	O
The	O
addition	O
of	O
the	O
TOCA1	O
HR1	O
domain	O
to	O
100	O
μm	O
had	O
no	O
significant	O
effect	O
on	O
the	O
rate	O
of	O
actin	O
polymerization	O
or	O
maximum	O
fluorescence	O
.	O
	O
Actin	O
polymerization	O
downstream	O
of	O
Cdc42	O
·	O
N	O
-	O
WASP	O
·	O
TOCA1	O
is	O
inhibited	O
by	O
excess	O
N	O
-	O
WASP	O
GBD	O
but	O
not	O
by	O
the	O
TOCA1	O
HR1	O
domain	O
.	O
	O
The	O
Cdc42	O
-	O
TOCA1	O
Interaction	O
	O
A	O
single	O
binding	O
interface	O
on	O
both	O
the	O
HR1	O
domain	O
and	O
Cdc42	O
can	O
be	O
concluded	O
from	O
the	O
data	O
presented	O
here	O
.	O
	O
Furthermore	O
,	O
the	O
interfaces	O
are	O
comparable	O
with	O
those	O
of	O
other	O
G	O
protein	O
-	O
HR1	O
interactions	O
(	O
Fig	O
.	O
4	O
),	O
and	O
the	O
lowest	O
energy	O
model	O
produced	O
in	O
rigid	O
body	O
docking	O
resembles	O
previously	O
studied	O
G	O
protein	O
·	O
HR1	O
complexes	O
(	O
Fig	O
.	O
6	O
).	O
	O
It	O
seems	O
,	O
therefore	O
,	O
that	O
the	O
interaction	O
,	O
despite	O
its	O
relatively	O
low	O
affinity	O
,	O
is	O
specific	O
and	O
sterically	O
similar	O
to	O
other	O
HR1	O
domain	O
-	O
G	O
protein	O
interactions	O
.	O
	O
A	O
short	O
region	O
N	O
-	O
terminal	O
to	O
the	O
coiled	O
-	O
coil	O
exhibits	O
a	O
series	O
of	O
turns	O
and	O
contacts	O
residues	O
of	O
both	O
helices	O
of	O
the	O
coiled	O
-	O
coil	O
(	O
Fig	O
.	O
3	O
).	O
	O
The	O
corresponding	O
sequence	O
in	O
CIP4	O
also	O
includes	O
a	O
series	O
of	O
turns	O
but	O
is	O
flexible	O
,	O
whereas	O
in	O
the	O
HR1a	O
domain	O
of	O
PRK1	O
,	O
the	O
equivalent	O
region	O
adopts	O
an	O
α	O
-	O
helical	O
structure	O
that	O
packs	O
against	O
the	O
coiled	O
-	O
coil	O
.	O
	O
The	O
contacts	O
between	O
the	O
N	O
-	O
terminal	O
region	O
and	O
the	O
coiled	O
-	O
coil	O
are	O
predominantly	O
hydrophobic	O
in	O
both	O
cases	O
,	O
but	O
sequence	O
-	O
specific	O
contacts	O
do	O
not	O
appear	O
to	O
be	O
conserved	O
.	O
	O
This	O
region	O
is	O
distant	O
from	O
the	O
G	O
protein	O
-	O
binding	O
interface	O
of	O
the	O
HR1	O
domains	O
,	O
so	O
the	O
structural	O
differences	O
may	O
relate	O
to	O
the	O
structure	O
and	O
regulation	O
of	O
these	O
domains	O
rather	O
than	O
their	O
G	O
protein	O
interactions	O
.	O
	O
TOCA1	O
and	O
CIP4	O
both	O
bind	O
weakly	O
to	O
Cdc42	O
,	O
whereas	O
the	O
HR1a	O
domain	O
of	O
PRK1	O
binds	O
tightly	O
to	O
RhoA	O
and	O
Rac1	O
,	O
and	O
the	O
HR1b	O
domain	O
binds	O
to	O
Rac1	O
.	O
	O
The	O
structural	O
features	O
shared	O
by	O
TOCA1	O
and	O
CIP4	O
may	O
therefore	O
be	O
related	O
to	O
Cdc42	O
binding	O
specificity	O
and	O
the	O
low	O
affinities	O
.	O
	O
In	O
free	O
TOCA1	O
,	O
the	O
side	O
chains	O
of	O
the	O
interhelical	O
region	O
make	O
extensive	O
contacts	O
with	O
residues	O
in	O
helix	O
1	O
.	O
	O
The	O
lowest	O
energy	O
model	O
produced	O
by	O
HADDOCK	O
using	O
ambiguous	O
interaction	O
restraints	O
from	O
the	O
titration	O
data	O
resembled	O
the	O
NMR	O
structures	O
of	O
RhoA	O
and	O
Rac1	O
in	O
complex	O
with	O
their	O
HR1	O
domain	O
partners	O
.	O
	O
Phe	O
-	O
56Cdc42	O
,	O
which	O
is	O
a	O
Trp	O
in	O
both	O
Rac1	O
and	O
RhoA	O
(	O
Fig	O
.	O
6C	O
),	O
is	O
thought	O
to	O
pack	O
behind	O
switch	O
I	O
when	O
Cdc42	O
interacts	O
with	O
ACK	O
,	O
maintaining	O
the	O
switch	O
in	O
a	O
binding	O
-	O
competent	O
orientation	O
.	O
	O
This	O
residue	O
has	O
also	O
been	O
identified	O
as	O
important	O
for	O
Cdc42	O
-	O
WASP	O
binding	O
.	O
	O
Phe	O
-	O
56Cdc42	O
is	O
therefore	O
likely	O
to	O
be	O
involved	O
in	O
the	O
Cdc42	O
-	O
TOCA1	O
interaction	O
,	O
probably	O
by	O
stabilizing	O
the	O
position	O
of	O
switch	O
I	O
.	O
	O
Gln	O
-	O
2Cdc42	O
,	O
which	O
has	O
also	O
been	O
identified	O
as	O
a	O
contact	O
residue	O
in	O
the	O
Cdc42	O
·	O
ACK	O
complex	O
,	O
contacts	O
Val	O
-	O
376TOCA1	O
and	O
Asn	O
-	O
380TOCA1	O
in	O
the	O
model	O
and	O
disrupts	O
the	O
contacts	O
between	O
the	O
interhelical	O
loop	O
and	O
the	O
first	O
helix	O
of	O
the	O
TOCA1	O
coiled	O
-	O
coil	O
.	O
	O
Thr	O
-	O
52Cdc42	O
,	O
which	O
has	O
also	O
been	O
identified	O
as	O
making	O
minor	O
contacts	O
with	O
ACK	O
,	O
falls	O
near	O
the	O
side	O
chains	O
of	O
HR1TOCA1	O
helix	O
1	O
,	O
particularly	O
Lys	O
-	O
372TOCA1	O
,	O
whereas	O
the	O
equivalent	O
position	O
in	O
Rac1	O
is	O
Asn	O
-	O
52Rac1	O
.	O
	O
N52T	B-mutant
is	O
one	O
of	O
a	O
combination	O
of	O
seven	O
residues	O
found	O
to	O
confer	O
ACK	O
binding	O
on	O
Rac1	O
and	O
so	O
may	O
represent	O
a	O
specific	O
Cdc42	O
-	O
effector	O
contact	O
residue	O
.	O
	O
The	O
position	O
equivalent	O
to	O
Lys	O
-	O
372TOCA1	O
in	O
PRK1	O
is	O
Glu	O
-	O
58HR1a	O
or	O
Gln	O
-	O
151HR1b	O
.	O
	O
Thr	O
-	O
52Cdc42	O
-	O
Lys	O
-	O
372TOCA1	O
may	O
therefore	O
represent	O
a	O
specific	O
Cdc42	O
-	O
HR1TOCA1	O
contact	O
.	O
	O
Arg	O
-	O
68Cdc42	O
of	O
switch	O
II	O
is	O
positioned	O
close	O
to	O
Glu	O
-	O
395TOCA1	O
(	O
Fig	O
.	O
6D	O
),	O
suggesting	O
a	O
direct	O
electrostatic	O
contact	O
between	O
switch	O
II	O
of	O
Cdc42	O
and	O
helix	O
2	O
of	O
the	O
HR1	O
domain	O
.	O
	O
The	O
importance	O
of	O
this	O
residue	O
in	O
the	O
Cdc42	O
-	O
TOCA1	O
interaction	O
remains	O
unclear	O
,	O
although	O
its	O
mutation	O
reduces	O
binding	O
to	O
RhoGAP	O
,	O
suggesting	O
that	O
it	O
can	O
be	O
involved	O
in	O
Cdc42	O
interactions	O
.	O
	O
The	O
solution	O
structure	O
of	O
the	O
TOCA1	O
HR1	O
domain	O
presented	O
here	O
,	O
along	O
with	O
the	O
model	O
of	O
the	O
HR1TOCA1	O
·	O
Cdc42	O
complex	O
is	O
consistent	O
with	O
a	O
conserved	O
mode	O
of	O
binding	O
across	O
the	O
known	O
HR1	O
domain	O
-	O
Rho	O
family	O
interactions	O
,	O
despite	O
their	O
differing	O
affinities	O
.	O
	O
We	O
have	O
previously	O
postulated	O
that	O
the	O
inherent	O
flexibility	O
of	O
HR1	O
domains	O
contributes	O
to	O
their	O
ability	O
to	O
bind	O
to	O
different	O
Rho	O
family	O
G	O
proteins	O
,	O
with	O
Rho	O
-	O
binding	O
HR1	O
domains	O
displaying	O
increased	O
flexibility	O
,	O
reflected	O
in	O
their	O
lower	O
melting	O
temperatures	O
(	O
Tm	O
)	O
and	O
Rac	O
binders	O
being	O
more	O
rigid	O
.	O
	O
The	O
Tm	O
of	O
the	O
TOCA1	O
HR1	O
domain	O
is	O
61	O
.	O
9	O
°	O
C	O
(	O
data	O
not	O
shown	O
),	O
which	O
is	O
the	O
highest	O
Tm	O
that	O
we	O
have	O
measured	O
for	O
an	O
HR1	O
domain	O
thus	O
far	O
.	O
	O
An	O
investigation	O
into	O
the	O
local	O
motions	O
,	O
particularly	O
in	O
the	O
G	O
protein	O
-	O
binding	O
regions	O
,	O
may	O
offer	O
further	O
insight	O
into	O
the	O
differential	O
specificities	O
and	O
affinities	O
of	O
G	O
protein	O
-	O
HR1	O
domain	O
interactions	O
.	O
	O
The	O
low	O
affinity	O
of	O
the	O
Cdc42	O
-	O
HR1TOCA1	O
interaction	O
is	O
consistent	O
with	O
a	O
tightly	O
spatially	O
and	O
temporally	O
regulated	O
pathway	O
,	O
requiring	O
combinatorial	O
signals	O
leading	O
to	O
a	O
series	O
of	O
coincident	O
weak	O
interactions	O
that	O
elicit	O
full	O
activation	O
.	O
	O
The	O
HR1	O
domains	O
from	O
other	O
TOCA	O
family	O
members	O
,	O
CIP4	O
and	O
FBP17	O
,	O
also	O
bind	O
at	O
low	O
micromolar	O
affinities	O
to	O
Cdc42	O
,	O
so	O
the	O
low	O
affinity	O
interaction	O
appears	O
to	O
be	O
commonplace	O
among	O
this	O
family	O
of	O
HR1	O
domain	O
proteins	O
,	O
in	O
contrast	O
to	O
the	O
PRK	O
family	O
.	O
	O
Evidence	O
suggests	O
that	O
the	O
TOCA	O
family	O
of	O
proteins	O
are	O
recruited	O
to	O
the	O
membrane	O
via	O
an	O
interaction	O
between	O
their	O
F	O
-	O
BAR	O
domain	O
and	O
specific	O
signaling	O
lipids	O
.	O
	O
For	O
example	O
,	O
electrostatic	O
interactions	O
between	O
the	O
F	O
-	O
BAR	O
domain	O
and	O
the	O
membrane	O
are	O
required	O
for	O
TOCA1	O
recruitment	O
to	O
membrane	O
vesicles	O
and	O
tubules	O
,	O
and	O
TOCA1	O
-	O
dependent	O
actin	O
polymerization	O
is	O
known	O
to	O
depend	O
specifically	O
on	O
PI	O
(	O
4	O
,	O
5	O
)	O
P2	O
.	O
	O
Once	O
at	O
the	O
membrane	O
,	O
high	O
local	O
concentrations	O
of	O
TOCA1	O
could	O
exceed	O
the	O
Kd	O
of	O
F	O
-	O
BAR	O
dimerization	O
(	O
likely	O
to	O
be	O
comparable	O
with	O
that	O
of	O
the	O
FCHo2	O
F	O
-	O
BAR	O
domain	O
(	O
2	O
.	O
5	O
μm	O
))	O
and	O
that	O
of	O
the	O
Cdc42	O
-	O
HR1TOCA1	O
interaction	O
.	O
	O
Cdc42	O
-	O
HR1TOCA1	O
binding	O
would	O
then	O
be	O
favorable	O
,	O
as	O
long	O
as	O
coincident	O
activation	O
of	O
Cdc42	O
had	O
occurred	O
,	O
leading	O
to	O
stabilization	O
of	O
TOCA1	O
at	O
the	O
membrane	O
and	O
downstream	O
activation	O
of	O
N	O
-	O
WASP	O
.	O
	O
It	O
has	O
been	O
postulated	O
that	O
WASP	O
and	O
N	O
-	O
WASP	O
exist	O
in	O
equilibrium	O
between	O
folded	O
(	O
inactive	O
)	O
and	O
unfolded	O
(	O
active	O
)	O
forms	O
,	O
and	O
the	O
affinity	O
of	O
Cdc42	O
for	O
the	O
unfolded	O
WASP	O
proteins	O
is	O
significantly	O
enhanced	O
.	O
	O
The	O
unfolded	O
,	O
high	O
affinity	O
state	O
of	O
WASP	O
is	O
represented	O
by	O
a	O
short	O
peptide	O
,	O
the	O
GBD	O
,	O
which	O
binds	O
with	O
a	O
low	O
nanomolar	O
affinity	O
to	O
Cdc42	O
.	O
	O
In	O
contrast	O
,	O
the	O
best	O
estimate	O
of	O
the	O
affinity	O
of	O
full	O
-	O
length	O
WASP	O
for	O
Cdc42	O
is	O
low	O
micromolar	O
.	O
	O
In	O
the	O
inactive	O
state	O
of	O
WASP	O
,	O
the	O
actin	O
-	O
and	O
Arp2	O
/	O
3	O
-	O
binding	O
VCA	O
domain	O
contacts	O
the	O
GBD	O
,	O
competing	O
for	O
Cdc42	O
binding	O
.	O
	O
The	O
high	O
affinity	O
of	O
Cdc42	O
for	O
the	O
unfolded	O
,	O
active	O
form	O
pushes	O
the	O
equilibrium	O
in	O
favor	O
of	O
(	O
N	O
-)	O
WASP	O
activation	O
.	O
	O
Binding	O
of	O
PI	O
(	O
4	O
,	O
5	O
)	O
P2	O
to	O
the	O
basic	O
region	O
just	O
N	O
-	O
terminal	O
to	O
the	O
GBD	O
further	O
favors	O
the	O
active	O
conformation	O
.	O
	O
Cdc42	O
is	O
activated	O
in	O
response	O
to	O
co	O
-	O
incident	O
signals	O
and	O
can	O
then	O
bind	O
to	O
TOCA1	O
,	O
further	O
stabilizing	O
TOCA1	O
at	O
the	O
membrane	O
.	O
	O
The	O
recruitment	O
of	O
N	O
-	O
WASP	O
alone	O
and	O
of	O
the	O
N	O
-	O
WASP	O
·	O
WIP	O
complex	O
by	O
TOCA1	O
and	O
FBP17	O
has	O
been	O
demonstrated	O
.	O
	O
It	O
may	O
therefore	O
be	O
envisaged	O
that	O
WIP	O
and	O
TOCA1	O
exert	O
opposing	O
allosteric	O
effects	O
on	O
N	O
-	O
WASP	O
,	O
with	O
TOCA1	O
favoring	O
the	O
unfolded	O
,	O
active	O
conformation	O
of	O
N	O
-	O
WASP	O
and	O
increasing	O
its	O
affinity	O
for	O
Cdc42	O
.	O
	O
TOCA1	O
may	O
also	O
activate	O
N	O
-	O
WASP	O
by	O
effective	O
oligomerization	O
because	O
clustering	O
of	O
TOCA1	O
at	O
the	O
membrane	O
following	O
coincident	O
interactions	O
with	O
PI	O
(	O
4	O
,	O
5	O
)	O
P2	O
and	O
Cdc42	O
would	O
in	O
turn	O
lead	O
to	O
clustering	O
of	O
N	O
-	O
WASP	O
,	O
in	O
addition	O
to	O
pushing	O
the	O
equilibrium	O
toward	O
the	O
unfolded	O
,	O
active	O
state	O
.	O
	O
In	O
such	O
an	O
array	O
of	O
molecules	O
localized	O
to	O
a	O
discrete	O
region	O
of	O
the	O
membrane	O
,	O
it	O
is	O
plausible	O
that	O
WASP	O
could	O
bind	O
to	O
a	O
second	O
Cdc42	O
molecule	O
rather	O
than	O
displacing	O
TOCA1	O
from	O
its	O
cognate	O
Cdc42	O
.	O
	O
Our	O
NMR	O
and	O
affinity	O
data	O
,	O
however	O
,	O
are	O
consistent	O
with	O
displacement	O
of	O
the	O
TOCA1	O
HR1	O
by	O
the	O
N	O
-	O
WASP	O
GBD	O
.	O
	O
Furthermore	O
,	O
TOCA1	O
is	O
required	O
for	O
Cdc42	O
-	O
mediated	O
activation	O
of	O
N	O
-	O
WASP	O
·	O
WIP	O
,	O
implying	O
that	O
it	O
may	O
not	O
be	O
possible	O
for	O
Cdc42	O
to	O
bind	O
and	O
activate	O
N	O
-	O
WASP	O
prior	O
to	O
TOCA1	O
-	O
Cdc42	O
binding	O
.	O
	O
In	O
light	O
of	O
this	O
,	O
we	O
favor	O
an	O
“	O
effector	O
handover	O
”	O
scheme	O
whereby	O
TOCA1	O
interacts	O
with	O
Cdc42	O
prior	O
to	O
N	O
-	O
WASP	O
activation	O
,	O
after	O
which	O
N	O
-	O
WASP	O
displaces	O
TOCA1	O
from	O
its	O
bound	O
Cdc42	O
in	O
order	O
to	O
be	O
fully	O
activated	O
rather	O
than	O
binding	O
a	O
second	O
Cdc42	O
molecule	O
.	O
	O
The	O
concomitant	O
release	O
of	O
TOCA1	O
from	O
Cdc42	O
while	O
still	O
bound	O
to	O
N	O
-	O
WASP	O
presumably	O
enhances	O
the	O
ability	O
of	O
TOCA1	O
to	O
further	O
activate	O
N	O
-	O
WASP	O
·	O
WIP	O
-	O
induced	O
actin	O
polymerization	O
.	O
	O
Hence	O
,	O
actin	O
polymerization	O
cannot	O
occur	O
until	O
F	O
-	O
BAR	O
domains	O
are	O
poised	O
for	O
membrane	O
distortion	O
.	O
	O
Our	O
model	O
of	O
the	O
Cdc42	O
·	O
HR1TOCA1	O
complex	O
indicates	O
a	O
mechanism	O
by	O
which	O
such	O
a	O
handover	O
could	O
take	O
place	O
(	O
Fig	O
.	O
9	O
)	O
because	O
it	O
shows	O
that	O
the	O
effector	O
binding	O
sites	O
only	O
partially	O
overlap	O
on	O
Cdc42	O
.	O
	O
The	O
lysine	O
residues	O
thought	O
to	O
be	O
involved	O
in	O
an	O
electrostatic	O
steering	O
mechanism	O
in	O
WASP	O
-	O
Cdc42	O
binding	O
are	O
conserved	O
in	O
N	O
-	O
WASP	O
and	O
would	O
be	O
able	O
to	O
interact	O
with	O
Cdc42	O
even	O
when	O
the	O
TOCA1	O
HR1	O
domain	O
is	O
already	O
bound	O
.	O
	O
It	O
has	O
been	O
postulated	O
that	O
the	O
initial	O
interactions	O
between	O
this	O
basic	O
region	O
and	O
Cdc42	O
could	O
stabilize	O
the	O
active	O
conformation	O
of	O
WASP	O
,	O
leading	O
to	O
high	O
affinity	O
binding	O
between	O
the	O
core	O
CRIB	O
and	O
Cdc42	O
.	O
	O
The	O
region	O
C	O
-	O
terminal	O
to	O
the	O
core	O
CRIB	O
,	O
required	O
for	O
maximal	O
affinity	O
binding	O
,	O
would	O
then	O
fully	O
displace	O
the	O
TOCA1	O
HR1	O
.	O
	O
A	O
simplified	O
model	O
of	O
the	O
early	O
stages	O
of	O
Cdc42	O
·	O
N	O
-	O
WASP	O
·	O
TOCA1	O
-	O
dependent	O
actin	O
polymerization	O
.	O
	O
Step	O
2	O
,	O
N	O
-	O
WASP	O
exists	O
in	O
an	O
inactive	O
,	O
folded	O
conformation	O
.	O
	O
The	O
TOCA1	O
SH3	O
domain	O
interacts	O
with	O
N	O
-	O
WASP	O
,	O
causing	O
an	O
activatory	O
allosteric	O
effect	O
.	O
	O
Step	O
3	O
,	O
electrostatic	O
interactions	O
between	O
Cdc42	O
and	O
the	O
basic	O
region	O
upstream	O
of	O
the	O
CRIB	O
initiate	O
Cdc42	O
·	O
N	O
-	O
WASP	O
binding	O
.	O
	O
The	O
VCA	O
domain	O
is	O
released	O
for	O
downstream	O
interactions	O
,	O
and	O
actin	O
polymerization	O
proceeds	O
.	O
	O
In	O
conclusion	O
,	O
the	O
data	O
presented	O
here	O
show	O
that	O
the	O
TOCA1	O
HR1	O
domain	O
is	O
sufficient	O
for	O
Cdc42	O
binding	O
in	O
vitro	O
and	O
that	O
the	O
interaction	O
is	O
of	O
micromolar	O
affinity	O
,	O
lower	O
than	O
that	O
of	O
other	O
G	O
protein	O
-	O
HR1	O
domain	O
interactions	O
.	O
	O
The	O
analogous	O
HR1	O
domains	O
from	O
other	O
TOCA1	O
family	O
members	O
,	O
FBP17	O
and	O
CIP4	O
,	O
also	O
exhibit	O
micromolar	O
affinity	O
for	O
Cdc42	O
.	O
	O
A	O
role	O
for	O
the	O
TOCA1	O
-,	O
FBP17	O
-,	O
and	O
CIP4	O
-	O
Cdc42	O
interactions	O
in	O
the	O
recruitment	O
of	O
these	O
proteins	O
to	O
the	O
membrane	O
therefore	O
appears	O
unlikely	O
.	O
	O
Instead	O
,	O
our	O
findings	O
agree	O
with	O
earlier	O
suggestions	O
that	O
the	O
F	O
-	O
BAR	O
domain	O
is	O
responsible	O
for	O
membrane	O
recruitment	O
.	O
	O
The	O
role	O
of	O
the	O
Cdc42	O
-	O
TOCA1	O
interaction	O
remains	O
somewhat	O
elusive	O
,	O
but	O
it	O
may	O
serve	O
to	O
position	O
activated	O
Cdc42	O
and	O
N	O
-	O
WASP	O
to	O
allow	O
full	O
activation	O
of	O
N	O
-	O
WASP	O
and	O
as	O
such	O
serve	O
to	O
couple	O
F	O
-	O
BAR	O
-	O
mediated	O
membrane	O
deformation	O
with	O
N	O
-	O
WASP	O
activation	O
.	O
	O
Our	O
data	O
are	O
therefore	O
easily	O
reconciled	O
with	O
the	O
dynamic	O
instability	O
models	O
described	O
in	O
relation	O
to	O
the	O
formation	O
of	O
endocytic	O
vesicles	O
and	O
with	O
the	O
current	O
data	O
pertaining	O
to	O
the	O
complex	O
activation	O
of	O
WASP	O
/	O
N	O
-	O
WASP	O
pathways	O
by	O
allosteric	O
and	O
oligomeric	O
effects	O
.	O
	O
We	O
therefore	O
postulate	O
an	O
effector	O
handover	O
mechanism	O
based	O
on	O
current	O
evidence	O
surrounding	O
WASP	O
/	O
N	O
-	O
WASP	O
activation	O
and	O
our	O
model	O
of	O
the	O
Cdc42	O
·	O
HR1TOCA1	O
complex	O
.	O
	O
The	O
displacement	O
of	O
the	O
TOCA1	O
HR1	O
domain	O
from	O
Cdc42	O
by	O
N	O
-	O
WASP	O
may	O
represent	O
a	O
unidirectional	O
step	O
in	O
the	O
pathway	O
of	O
Cdc42	O
·	O
N	O
-	O
WASP	O
·	O
TOCA1	O
-	O
dependent	O
actin	O
assembly	O
.	O
	O
The	O
dynamic	O
organization	O
of	O
fungal	O
acetyl	O
-	O
CoA	O
carboxylase	O
	O
Eukaryotic	O
ACCs	O
are	O
single	O
-	O
chain	O
multienzymes	O
characterized	O
by	O
a	O
large	O
,	O
non	O
-	O
catalytic	O
central	O
domain	O
(	O
CD	O
),	O
whose	O
role	O
in	O
ACC	O
regulation	O
remains	O
poorly	O
characterized	O
.	O
	O
Here	O
we	O
report	O
the	O
crystal	O
structure	O
of	O
the	O
yeast	O
ACC	O
CD	O
,	O
revealing	O
a	O
unique	O
four	O
-	O
domain	O
organization	O
.	O
	O
A	O
regulatory	O
loop	O
,	O
which	O
is	O
phosphorylated	O
at	O
the	O
key	O
functional	O
phosphorylation	O
site	O
of	O
fungal	O
ACC	O
,	O
wedges	O
into	O
a	O
crevice	O
between	O
two	O
domains	O
of	O
CD	O
.	O
	O
Acetyl	O
-	O
CoA	O
carboxylases	O
are	O
central	O
regulatory	O
hubs	O
of	O
fatty	O
acid	O
metabolism	O
and	O
are	O
important	O
targets	O
for	O
drug	O
development	O
in	O
obesity	O
and	O
cancer	O
.	O
	O
Here	O
,	O
the	O
authors	O
demonstrate	O
that	O
the	O
regulation	O
of	O
these	O
highly	O
dynamic	O
enzymes	O
in	O
fungi	O
is	O
governed	O
by	O
a	O
mechanism	O
based	O
on	O
phosphorylation	O
-	O
dependent	O
conformational	O
variability	O
.	O
	O
Biotin	O
-	O
dependent	O
acetyl	O
-	O
CoA	O
carboxylases	O
(	O
ACCs	O
)	O
are	O
essential	O
enzymes	O
that	O
catalyse	O
the	O
ATP	O
-	O
dependent	O
carboxylation	O
of	O
acetyl	O
-	O
CoA	O
to	O
malonyl	O
-	O
CoA	O
.	O
This	O
reaction	O
provides	O
the	O
committed	O
activated	O
substrate	O
for	O
the	O
biosynthesis	O
of	O
fatty	O
acids	O
via	O
fatty	O
-	O
acid	O
synthase	O
.	O
	O
By	O
catalysing	O
this	O
rate	O
-	O
limiting	O
step	O
in	O
fatty	O
-	O
acid	O
biosynthesis	O
,	O
ACC	O
plays	O
a	O
key	O
role	O
in	O
anabolic	O
metabolism	O
.	O
	O
ACC	O
inhibition	O
and	O
knock	O
-	O
out	O
studies	O
show	O
the	O
potential	O
of	O
targeting	O
ACC	O
for	O
treatment	O
of	O
the	O
metabolic	O
syndrome	O
.	O
	O
Furthermore	O
,	O
elevated	O
ACC	O
activity	O
is	O
observed	O
in	O
malignant	O
tumours	O
.	O
	O
A	O
direct	O
link	O
between	O
ACC	O
and	O
cancer	O
is	O
provided	O
by	O
cancer	O
-	O
associated	O
mutations	B-mutant
in	O
the	O
breast	O
cancer	O
susceptibility	O
gene	O
1	O
(	O
BRCA1	O
),	O
which	O
relieve	O
inhibitory	O
interactions	O
of	O
BRCA1	O
with	O
ACC	O
.	O
	O
Thus	O
,	O
ACC	O
is	O
a	O
relevant	O
drug	O
target	O
for	O
type	O
2	O
diabetes	O
and	O
cancer	O
.	O
	O
Microbial	O
ACCs	O
are	O
also	O
the	O
principal	O
target	O
of	O
antifungal	O
and	O
antibiotic	O
compounds	O
,	O
such	O
as	O
Soraphen	O
A	O
.	O
	O
The	O
principal	O
functional	O
protein	O
components	O
of	O
ACCs	O
have	O
been	O
described	O
already	O
in	O
the	O
late	O
1960s	O
for	O
Escherichia	O
coli	O
(	O
E	O
.	O
coli	O
)	O
ACC	O
:	O
Biotin	O
carboxylase	O
(	O
BC	O
)	O
catalyses	O
the	O
ATP	O
-	O
dependent	O
carboxylation	O
of	O
a	O
biotin	O
moiety	O
,	O
which	O
is	O
covalently	O
linked	O
to	O
the	O
biotin	O
carboxyl	O
carrier	O
protein	O
(	O
BCCP	O
).	O
	O
Carboxyltransferase	O
(	O
CT	O
)	O
transfers	O
the	O
activated	O
carboxyl	O
group	O
from	O
carboxybiotin	O
to	O
acetyl	O
-	O
CoA	O
to	O
yield	O
malonyl	O
-	O
CoA	O
.	O
Prokaryotic	O
ACCs	O
are	O
transient	O
assemblies	O
of	O
individual	O
BC	O
,	O
CT	O
and	O
BCCP	O
subunits	O
.	O
	O
Eukaryotic	O
ACCs	O
,	O
instead	O
,	O
are	O
multienzymes	O
,	O
which	O
integrate	O
all	O
functional	O
components	O
into	O
a	O
single	O
polypeptide	O
chain	O
of	O
∼	O
2	O
,	O
300	O
amino	O
acids	O
.	O
	O
Human	O
ACC	O
occurs	O
in	O
two	O
closely	O
related	O
isoforms	O
,	O
ACC1	O
and	O
2	O
,	O
located	O
in	O
the	O
cytosol	O
and	O
at	O
the	O
outer	O
mitochondrial	O
membrane	O
,	O
respectively	O
.	O
	O
The	O
CD	O
comprises	O
one	O
-	O
third	O
of	O
the	O
protein	O
and	O
is	O
a	O
unique	O
feature	O
of	O
eukaryotic	O
ACCs	O
without	O
homologues	O
in	O
other	O
proteins	O
.	O
	O
The	O
BT	O
domain	O
has	O
been	O
visualized	O
in	O
bacterial	O
carboxylases	O
,	O
where	O
it	O
mediates	O
contacts	O
between	O
α	O
-	O
and	O
β	O
-	O
subunits	O
.	O
	O
Structural	O
studies	O
on	O
the	O
functional	O
architecture	O
of	O
intact	O
ACCs	O
have	O
been	O
hindered	O
by	O
their	O
huge	O
size	O
and	O
pronounced	O
dynamics	O
,	O
as	O
well	O
as	O
the	O
transient	O
assembly	O
mode	O
of	O
bacterial	O
ACCs	O
.	O
	O
However	O
,	O
crystal	O
structures	O
of	O
individual	O
components	O
or	O
domains	O
from	O
prokaryotic	O
and	O
eukaryotic	O
ACCs	O
,	O
respectively	O
,	O
have	O
been	O
solved	O
.	O
	O
The	O
structure	O
determination	O
of	O
the	O
holoenzymes	O
of	O
bacterial	O
biotin	O
-	O
dependent	O
carboxylases	O
,	O
which	O
lack	O
the	O
characteristic	O
CD	O
,	O
such	O
as	O
the	O
pyruvate	O
carboxylase	O
(	O
PC	O
),	O
propionyl	O
-	O
CoA	O
carboxylase	O
,	O
3	O
-	O
methyl	O
-	O
crotonyl	O
-	O
CoA	O
carboxylase	O
and	O
a	O
long	O
-	O
chain	O
acyl	O
-	O
CoA	O
carboxylase	O
revealed	O
strikingly	O
divergent	O
architectures	O
despite	O
a	O
general	O
conservation	O
of	O
all	O
functional	O
components	O
.	O
	O
In	O
these	O
structures	O
,	O
the	O
BC	O
and	O
CT	O
active	O
sites	O
are	O
at	O
distances	O
between	O
40	O
and	O
80	O
Å	O
,	O
such	O
that	O
substrate	O
transfer	O
could	O
be	O
mediated	O
solely	O
by	O
the	O
mobility	O
of	O
the	O
flexibly	O
tethered	O
BCCP	O
.	O
	O
Human	O
ACC1	O
is	O
regulated	O
allosterically	O
,	O
via	O
specific	O
protein	O
–	O
protein	O
interactions	O
,	O
and	O
by	O
reversible	O
phosphorylation	O
.	O
	O
Dynamic	O
polymerization	O
of	O
human	O
ACC1	O
is	O
linked	O
to	O
increased	O
activity	O
and	O
is	O
regulated	O
allosterically	O
by	O
the	O
activator	O
citrate	O
and	O
the	O
inhibitor	O
palmitate	O
,	O
or	O
by	O
binding	O
of	O
the	O
small	O
protein	O
MIG	O
-	O
12	O
(	O
ref	O
.).	O
	O
Human	O
ACC1	O
is	O
further	O
regulated	O
by	O
specific	O
phosphorylation	O
-	O
dependent	O
binding	O
of	O
BRCA1	O
to	O
Ser1263	O
in	O
the	O
CD	O
.	O
	O
Furthermore	O
,	O
phosphorylation	O
by	O
AMP	O
-	O
activated	O
protein	O
kinase	O
(	O
AMPK	O
)	O
and	O
cAMP	O
-	O
dependent	O
protein	O
kinase	O
(	O
PKA	O
)	O
leads	O
to	O
a	O
decrease	O
in	O
ACC1	O
activity	O
.	O
	O
AMPK	O
phosphorylates	O
ACC1	O
in	O
vitro	O
at	O
Ser80	O
,	O
Ser1201	O
and	O
Ser1216	O
and	O
PKA	O
at	O
Ser78	O
and	O
Ser1201	O
.	O
	O
However	O
,	O
regulatory	O
effects	O
on	O
ACC1	O
activity	O
are	O
mainly	O
mediated	O
by	O
phosphorylation	O
of	O
Ser80	O
and	O
Ser1201	O
(	O
refs	O
).	O
	O
Phosphorylated	O
Ser80	O
,	O
which	O
is	O
highly	O
conserved	O
only	O
in	O
higher	O
eukaryotes	O
,	O
presumably	O
binds	O
into	O
the	O
Soraphen	O
A	O
-	O
binding	O
pocket	O
.	O
	O
The	O
regulatory	O
Ser1201	O
shows	O
only	O
moderate	O
conservation	O
across	O
higher	O
eukaryotes	O
,	O
while	O
the	O
phosphorylated	O
Ser1216	O
is	O
highly	O
conserved	O
across	O
all	O
eukaryotes	O
.	O
	O
However	O
,	O
no	O
effect	O
of	O
Ser1216	O
phosphorylation	O
on	O
ACC	O
activity	O
has	O
been	O
reported	O
in	O
higher	O
eukaryotes	O
.	O
	O
For	O
fungal	O
ACC	O
,	O
neither	O
spontaneous	O
nor	O
inducible	O
polymerization	O
has	O
been	O
detected	O
despite	O
considerable	O
sequence	O
conservation	O
to	O
human	O
ACC1	O
.	O
	O
The	O
BRCA1	O
-	O
interacting	O
phosphoserine	O
position	O
is	O
not	O
conserved	O
in	O
fungal	O
ACC	O
,	O
and	O
no	O
other	O
phospho	O
-	O
dependent	O
protein	O
–	O
protein	O
interactions	O
of	O
fungal	O
ACC	O
have	O
been	O
described	O
.	O
	O
In	O
yeast	O
ACC	O
,	O
phosphorylation	O
sites	O
have	O
been	O
identified	O
at	O
Ser2	O
,	O
Ser735	O
,	O
Ser1148	O
,	O
Ser1157	O
and	O
Ser1162	O
(	O
ref	O
.).	O
	O
Despite	O
the	O
outstanding	O
relevance	O
of	O
ACC	O
in	O
primary	O
metabolism	O
and	O
disease	O
,	O
the	O
dynamic	O
organization	O
and	O
regulation	O
of	O
the	O
giant	O
eukaryotic	O
,	O
and	O
in	O
particular	O
fungal	O
ACC	O
,	O
remain	O
poorly	O
characterized	O
.	O
	O
Here	O
we	O
provide	O
the	O
structure	O
of	O
Saccharomyces	O
cerevisiae	O
(	O
Sce	O
)	O
ACC	O
CD	O
,	O
intermediate	O
-	O
and	O
low	O
-	O
resolution	O
structures	O
of	O
human	O
(	O
Hsa	O
)	O
ACC	O
CD	O
and	O
larger	B-mutant
fragments	I-mutant
of	O
fungal	O
ACC	O
from	O
Chaetomium	O
thermophilum	O
(	O
Cth	O
;	O
Fig	O
.	O
1a	O
).	O
	O
The	O
overall	O
extent	O
of	O
the	O
SceCD	O
is	O
70	O
by	O
75	O
Å	O
(	O
Fig	O
.	O
1b	O
and	O
Supplementary	O
Fig	O
.	O
1a	O
,	O
b	O
),	O
and	O
the	O
attachment	O
points	O
of	O
the	O
N	O
-	O
terminal	O
26	O
-	O
residue	O
linker	O
to	O
the	O
BCCP	O
domain	O
and	O
the	O
C	O
-	O
terminal	O
CT	O
domain	O
are	O
separated	O
by	O
46	O
Å	O
(	O
the	O
N	O
-	O
and	O
C	O
termini	O
are	O
indicated	O
with	O
spheres	O
in	O
Fig	O
.	O
1b	O
).	O
	O
CDN	O
adopts	O
a	O
letter	O
C	O
shape	O
,	O
where	O
one	O
of	O
the	O
ends	O
is	O
a	O
regular	O
four	O
-	O
helix	O
bundle	O
(	O
Nα3	O
-	O
6	O
),	O
the	O
other	O
end	O
is	O
a	O
helical	O
hairpin	O
(	O
Nα8	O
,	O
9	O
)	O
and	O
the	O
bridging	O
region	O
comprises	O
six	O
helices	O
(	O
Nα1	O
,	O
2	O
,	O
7	O
,	O
10	O
–	O
12	O
).	O
	O
CDL	O
is	O
composed	O
of	O
a	O
small	O
,	O
irregular	O
four	O
-	O
helix	O
bundle	O
(	O
Lα1	O
–	O
4	O
)	O
and	O
tightly	O
interacts	O
with	O
the	O
open	O
face	O
of	O
CDC1	O
via	O
an	O
interface	O
of	O
1	O
,	O
300	O
Å2	O
involving	O
helices	O
Lα3	O
and	O
Lα4	O
.	O
	O
CDC2	O
is	O
extended	O
at	O
its	O
C	O
terminus	O
by	O
an	O
additional	O
β	O
-	O
strand	O
and	O
an	O
irregular	O
β	O
-	O
hairpin	O
.	O
	O
A	O
regulatory	O
loop	O
mediates	O
interdomain	O
interactions	O
	O
In	O
insect	O
-	O
cell	O
-	O
expressed	O
full	O
-	O
length	O
SceACC	O
,	O
the	O
highly	O
conserved	O
Ser1157	O
is	O
the	O
only	O
fully	O
occupied	O
phosphorylation	O
site	O
with	O
functional	O
relevance	O
in	O
S	O
.	O
cerevisiae	O
.	O
	O
Additional	O
phosphorylation	O
was	O
detected	O
for	O
Ser2101	O
and	O
Tyr2179	O
;	O
however	O
,	O
these	O
sites	O
are	O
neither	O
conserved	O
across	O
fungal	O
ACC	O
nor	O
natively	O
phosphorylated	O
in	O
yeast	O
.	O
	O
MS	O
analysis	O
of	O
dissolved	O
crystals	O
confirmed	O
the	O
phosphorylated	O
state	O
of	O
Ser1157	O
also	O
in	O
SceCD	O
crystals	O
.	O
	O
In	O
the	O
SceCD	O
crystal	O
structure	O
,	O
the	O
phosphorylated	O
Ser1157	O
resides	O
in	O
a	O
regulatory	O
36	O
-	O
amino	O
-	O
acid	O
loop	O
between	O
strands	O
β2	O
and	O
β3	O
of	O
CDC1	O
(	O
Fig	O
.	O
1b	O
,	O
d	O
),	O
which	O
contains	O
two	O
additional	O
less	O
-	O
conserved	O
phosphorylation	O
sites	O
(	O
Ser1148	O
and	O
Ser1162	O
)	O
confirmed	O
in	O
yeast	O
,	O
but	O
not	O
occupied	O
here	O
.	O
	O
This	O
regulatory	O
loop	O
wedges	O
between	O
the	O
CDC1	O
and	O
CDC2	O
domains	O
and	O
provides	O
the	O
largest	O
contribution	O
to	O
the	O
interdomain	O
interface	O
.	O
	O
The	O
N	O
-	O
terminal	O
region	O
of	O
the	O
regulatory	O
loop	O
also	O
directly	O
contacts	O
the	O
C	O
-	O
terminal	O
region	O
of	O
CDC2	O
leading	O
into	O
CT	O
.	O
	O
Phosphoserine	O
1157	O
is	O
tightly	O
bound	O
by	O
two	O
highly	O
conserved	O
arginines	O
(	O
Arg1173	O
and	O
Arg1260	O
)	O
of	O
CDC1	O
(	O
Fig	O
.	O
1d	O
).	O
	O
Already	O
the	O
binding	O
of	O
phosphorylated	O
Ser1157	O
apparently	O
stabilizes	O
the	O
regulatory	O
loop	O
conformation	O
;	O
the	O
accessory	O
phosphorylation	O
sites	O
Ser1148	O
and	O
Ser1162	O
in	O
the	O
same	O
loop	O
may	O
further	O
modulate	O
the	O
strength	O
of	O
interaction	O
between	O
the	O
regulatory	O
loop	O
and	O
the	O
CDC1	O
and	O
CDC2	O
domains	O
.	O
	O
Phosphorylation	O
of	O
the	O
regulatory	O
loop	O
thus	O
determines	O
interdomain	O
interactions	O
of	O
CDC1	O
and	O
CDC2	O
,	O
suggesting	O
that	O
it	O
may	O
exert	O
its	O
regulatory	O
function	O
by	O
modifying	O
the	O
overall	O
structure	O
and	O
dynamics	O
of	O
the	O
CD	O
.	O
	O
The	O
functional	O
role	O
of	O
Ser1157	O
was	O
confirmed	O
by	O
an	O
activity	O
assay	O
based	O
on	O
the	O
incorporation	O
of	O
radioactive	O
carbonate	O
into	O
acid	O
non	O
-	O
volatile	O
material	O
.	O
	O
The	O
values	O
obtained	O
for	O
dephosphorylated	O
SceACC	O
are	O
comparable	O
to	O
earlier	O
measurements	O
of	O
non	O
-	O
phosphorylated	O
yeast	O
ACC	O
expressed	O
in	O
E	O
.	O
coli	O
.	O
	O
The	O
variable	O
CD	O
is	O
conserved	O
between	O
yeast	O
and	O
human	O
	O
To	O
compare	O
the	O
organization	O
of	O
fungal	O
and	O
human	O
ACC	O
CD	O
,	O
we	O
determined	O
the	O
structure	O
of	O
a	O
human	O
ACC1	B-mutant
fragment	I-mutant
that	O
comprises	O
the	O
BT	O
and	O
CD	O
domains	O
(	O
HsaBT	B-mutant
-	I-mutant
CD	I-mutant
),	O
but	O
lacks	O
the	O
mobile	O
BCCP	O
in	O
between	O
(	O
Fig	O
.	O
1a	O
).	O
	O
An	O
experimentally	O
phased	O
map	O
was	O
obtained	O
at	O
3	O
.	O
7	O
Å	O
resolution	O
for	O
a	O
cadmium	O
-	O
derivatized	O
crystal	O
and	O
was	O
interpreted	O
by	O
a	O
poly	O
-	O
alanine	O
model	O
(	O
Fig	O
.	O
1e	O
and	O
Table	O
1	O
).	O
	O
With	O
CDL	O
/	O
CDC1	O
superposed	O
,	O
CDN	O
in	O
HsaBT	B-mutant
-	I-mutant
CD	I-mutant
is	O
rotated	O
by	O
160	O
°	O
around	O
a	O
hinge	O
at	O
the	O
connection	O
of	O
CDN	O
/	O
CDL	O
(	O
Supplementary	O
Fig	O
.	O
1d	O
).	O
	O
This	O
rotation	O
displaces	O
the	O
N	O
terminus	O
of	O
CDN	O
in	O
HsaBT	B-mutant
-	I-mutant
CD	I-mutant
by	O
51	O
Å	O
compared	O
with	O
SceCD	O
,	O
resulting	O
in	O
a	O
separation	O
of	O
the	O
attachment	O
points	O
of	O
the	O
N	O
-	O
terminal	O
linker	O
to	O
the	O
BCCP	O
domain	O
and	O
the	O
C	O
-	O
terminal	O
CT	O
domain	O
by	O
67	O
Å	O
(	O
the	O
attachment	O
points	O
are	O
indicated	O
with	O
spheres	O
in	O
Fig	O
.	O
1e	O
).	O
	O
The	O
BT	O
domain	O
of	O
HsaBT	B-mutant
-	I-mutant
CD	I-mutant
consists	O
of	O
a	O
helix	O
that	O
is	O
surrounded	O
at	O
its	O
N	O
terminus	O
by	O
an	O
antiparallel	O
eight	O
-	O
stranded	O
β	O
-	O
barrel	O
.	O
	O
On	O
the	O
basis	O
of	O
MS	O
analysis	O
of	O
insect	O
-	O
cell	O
-	O
expressed	O
human	O
full	O
-	O
length	O
ACC	O
,	O
Ser80	O
shows	O
the	O
highest	O
degree	O
of	O
phosphorylation	O
(	O
90	O
%).	O
	O
Ser29	O
and	O
Ser1263	O
,	O
implicated	O
in	O
insulin	O
-	O
dependent	O
phosphorylation	O
and	O
BRCA1	O
binding	O
,	O
respectively	O
,	O
are	O
phosphorylated	O
at	O
intermediate	O
levels	O
(	O
40	O
%).	O
	O
The	O
highly	O
conserved	O
Ser1216	O
(	O
corresponding	O
to	O
S	O
.	O
cerevisiae	O
Ser1157	O
),	O
as	O
well	O
as	O
Ser1201	O
,	O
both	O
in	O
the	O
regulatory	O
loop	O
discussed	O
above	O
,	O
are	O
not	O
phosphorylated	O
.	O
	O
However	O
,	O
residual	O
phosphorylation	O
levels	O
were	O
detected	O
for	O
Ser1204	O
(	O
7	O
%)	O
and	O
Ser1218	O
(	O
7	O
%)	O
in	O
the	O
same	O
loop	O
.	O
	O
MS	O
analysis	O
of	O
the	O
HsaBT	B-mutant
-	I-mutant
CD	I-mutant
crystallization	O
sample	O
reveals	O
partial	O
proteolytic	O
digestion	O
of	O
the	O
regulatory	O
loop	O
.	O
	O
Accordingly	O
,	O
most	O
of	O
this	O
loop	O
is	O
not	O
represented	O
in	O
the	O
HsaBT	B-mutant
-	I-mutant
CD	I-mutant
crystal	O
structure	O
.	O
	O
Besides	O
the	O
regulatory	O
loop	O
,	O
also	O
the	O
phosphopeptide	O
target	O
region	O
for	O
BRCA1	O
interaction	O
is	O
not	O
resolved	O
presumably	O
because	O
of	O
pronounced	O
flexibility	O
.	O
	O
At	O
the	O
level	O
of	O
isolated	O
yeast	O
and	O
human	O
CD	O
,	O
the	O
structural	O
analysis	O
indicates	O
the	O
presence	O
of	O
at	O
least	O
two	O
hinges	O
,	O
one	O
with	O
large	O
-	O
scale	O
flexibility	O
at	O
the	O
CDN	O
/	O
CDL	O
connection	O
,	O
and	O
one	O
with	O
tunable	O
plasticity	O
between	O
CDL	O
/	O
CDC1	O
and	O
CDC2	O
,	O
plausibly	O
affected	O
by	O
phosphorylation	O
in	O
the	O
regulatory	O
loop	O
region	O
.	O
	O
The	O
integration	O
of	O
CD	O
into	O
the	O
fungal	O
ACC	O
multienzyme	O
	O
Using	O
molecular	O
replacement	O
based	O
on	O
fungal	O
ACC	O
CD	O
and	O
CT	O
models	O
,	O
we	O
obtained	O
structures	O
of	O
a	O
variant	B-mutant
comprising	O
CthCT	O
and	O
CDC1	O
/	O
CDC2	O
in	O
two	O
crystal	O
forms	O
at	O
resolutions	O
of	O
3	O
.	O
6	O
and	O
4	O
.	O
5	O
Å	O
(	O
CthCD	B-mutant
-	I-mutant
CTCter1	I-mutant
/	I-mutant
2	I-mutant
),	O
respectively	O
,	O
as	O
well	O
as	O
of	O
a	O
CthCT	O
linked	O
to	O
the	O
entire	O
CD	O
at	O
7	O
.	O
2	O
Å	O
resolution	O
(	O
CthCD	B-mutant
-	I-mutant
CT	I-mutant
;	O
Figs	O
1a	O
and	O
2	O
,	O
Table	O
1	O
).	O
	O
For	O
CthΔBCCP	B-mutant
,	O
crystals	O
diffracting	O
to	O
8	O
.	O
4	O
Å	O
resolution	O
were	O
obtained	O
.	O
	O
Owing	O
to	O
the	O
limited	O
resolution	O
the	O
discussion	O
of	O
structures	O
of	O
CthCD	B-mutant
-	I-mutant
CT	I-mutant
and	O
CthΔBCCP	B-mutant
is	O
restricted	O
to	O
the	O
analysis	O
of	O
domain	O
localization	O
.	O
	O
Still	O
,	O
these	O
structures	O
contribute	O
considerably	O
to	O
the	O
visualization	O
of	O
an	O
intrinsically	O
dynamic	O
fungal	O
ACC	O
.	O
	O
In	O
all	O
these	O
crystal	O
structures	O
,	O
the	O
CT	O
domains	O
build	O
a	O
canonical	O
head	O
-	O
to	O
-	O
tail	O
dimer	O
,	O
with	O
active	O
sites	O
formed	O
by	O
contributions	O
from	O
both	O
protomers	O
(	O
Fig	O
.	O
2	O
and	O
Supplementary	O
Fig	O
.	O
3a	O
).	O
	O
The	O
connection	O
of	O
CD	O
and	O
CT	O
is	O
provided	O
by	O
a	O
10	O
-	O
residue	O
peptide	O
stretch	O
,	O
which	O
links	O
the	O
N	O
terminus	O
of	O
CT	O
to	O
the	O
irregular	O
β	O
-	O
hairpin	O
/	O
β	O
-	O
strand	O
extension	O
of	O
CDC2	O
(	O
Supplementary	O
Fig	O
.	O
3b	O
).	O
	O
CD	O
/	O
CT	O
contacts	O
are	O
only	O
formed	O
in	O
direct	O
vicinity	O
of	O
the	O
covalent	O
linkage	O
and	O
involve	O
the	O
β	O
-	O
hairpin	O
extension	O
of	O
CDC2	O
as	O
well	O
as	O
the	O
loop	O
between	O
strands	O
β2	O
/	O
β3	O
of	O
the	O
CT	O
N	O
-	O
lobe	O
,	O
which	O
contains	O
a	O
conserved	O
RxxGxN	O
motif	O
.	O
	O
The	O
neighbouring	O
loop	O
on	O
the	O
CT	O
side	O
(	O
between	O
CT	O
β1	O
/	O
β2	O
)	O
is	O
displaced	O
by	O
2	O
.	O
5	O
Å	O
compared	O
to	O
isolated	O
CT	O
structures	O
(	O
Supplementary	O
Fig	O
.	O
3c	O
).	O
	O
On	O
the	O
basis	O
of	O
an	O
interface	O
area	O
of	O
∼	O
600	O
Å2	O
and	O
its	O
edge	O
-	O
to	O
-	O
edge	O
connection	O
characteristics	O
,	O
the	O
interface	O
between	O
CT	O
and	O
CD	O
might	O
be	O
classified	O
as	O
conformationally	O
variable	O
.	O
	O
The	O
CDC2	O
/	O
CT	O
interface	O
acts	O
as	O
a	O
true	O
hinge	O
with	O
observed	O
rotation	O
up	O
to	O
16	O
°,	O
which	O
results	O
in	O
a	O
translocation	O
of	O
the	O
distal	O
end	O
of	O
CDC2	O
by	O
8	O
Å	O
.	O
	O
The	O
interface	O
between	O
CDC2	O
and	O
CDL	O
/	O
CDC1	O
,	O
which	O
is	O
mediated	O
by	O
the	O
phosphorylated	O
regulatory	O
loop	O
in	O
the	O
SceCD	O
structure	O
,	O
is	O
less	O
variable	O
than	O
the	O
CD	O
–	O
CT	O
junction	O
,	O
and	O
permits	O
only	O
limited	O
rotation	O
and	O
tilting	O
(	O
Fig	O
.	O
3b	O
).	O
	O
Analysis	O
of	O
the	O
impact	O
of	O
phosphorylation	O
on	O
the	O
interface	O
between	O
CDC2	O
and	O
CDL	O
/	O
CDC1	O
in	O
CthACC	B-mutant
variant	I-mutant
structures	O
is	O
precluded	O
by	O
the	O
limited	O
crystallographic	O
resolution	O
.	O
	O
However	O
,	O
MS	O
analysis	O
of	O
CthCD	B-mutant
-	I-mutant
CT	I-mutant
and	O
CthΔBCCP	B-mutant
constructs	O
revealed	O
between	O
60	O
and	O
70	O
%	O
phosphorylation	O
of	O
Ser1170	O
(	O
corresponding	O
to	O
SceACC	O
Ser1157	O
).	O
	O
The	O
CDN	O
domain	O
positioning	O
relative	O
to	O
CDL	O
/	O
CDC1	O
is	O
highly	O
variable	O
with	O
three	O
main	O
orientations	O
observed	O
in	O
the	O
structures	O
of	O
SceCD	O
and	O
the	O
larger	B-mutant
CthACC	I-mutant
fragments	I-mutant
:	O
CDN	O
tilts	O
,	O
resulting	O
in	O
a	O
displacement	O
of	O
its	O
N	O
terminus	O
by	O
23	O
Å	O
(	O
Fig	O
.	O
4a	O
,	O
observed	O
in	O
both	O
protomers	O
of	O
CthCD	B-mutant
-	I-mutant
CT	I-mutant
and	O
one	O
protomer	O
of	O
CthΔBCCP	B-mutant
,	O
denoted	O
as	O
CthCD	B-mutant
-	I-mutant
CT1	I-mutant
/	I-mutant
2	I-mutant
and	O
CthΔBCCP1	B-mutant
,	O
respectively	O
).	O
	O
In	O
addition	O
,	O
CDN	O
can	O
rotate	O
around	O
hinges	O
in	O
the	O
connection	O
between	O
CDN	O
/	O
CDL	O
by	O
70	O
°	O
(	O
Fig	O
.	O
4b	O
,	O
observed	O
in	O
the	O
second	O
protomer	O
of	O
CthΔBCCP	B-mutant
,	O
denoted	O
as	O
CthΔBCCP2	B-mutant
)	O
and	O
160	O
°	O
(	O
Fig	O
.	O
4c	O
,	O
observed	O
in	O
SceCD	O
)	O
leading	O
to	O
displacement	O
of	O
the	O
anchor	O
site	O
for	O
the	O
BCCP	O
linker	O
by	O
up	O
to	O
33	O
and	O
40	O
Å	O
,	O
respectively	O
.	O
	O
On	O
the	O
basis	O
of	O
the	O
occurrence	O
of	O
related	O
conformational	O
changes	O
between	O
fungal	O
and	O
human	O
ACC	B-mutant
fragments	I-mutant
,	O
the	O
observed	O
set	O
of	O
conformations	O
may	O
well	O
represent	O
general	O
states	O
present	O
in	O
all	O
eukaryotic	O
ACCs	O
.	O
	O
Large	O
-	O
scale	O
conformational	O
variability	O
of	O
fungal	O
ACC	O
	O
To	O
obtain	O
a	O
comprehensive	O
view	O
of	O
fungal	O
ACC	O
dynamics	O
in	O
solution	O
,	O
we	O
employed	O
SAXS	O
and	O
EM	O
.	O
	O
The	O
smooth	O
appearance	O
of	O
scattering	O
curves	O
and	O
derived	O
distance	O
distributions	O
might	O
indicate	O
substantial	O
interdomain	O
flexibility	O
(	O
Supplementary	O
Fig	O
.	O
2a	O
–	O
c	O
).	O
	O
Direct	O
observation	O
of	O
individual	O
full	O
-	O
length	O
CthACC	O
particles	O
,	O
according	O
to	O
MS	O
results	O
predominantly	O
in	O
a	O
phosphorylated	O
low	O
-	O
activity	O
state	O
,	O
in	O
negative	O
stain	O
EM	O
reveals	O
a	O
large	O
set	O
of	O
conformations	O
from	O
rod	O
-	O
like	O
extended	O
to	O
U	O
-	O
shaped	O
particles	O
.	O
	O
Class	O
averages	O
,	O
obtained	O
by	O
maximum	O
-	O
likelihood	O
-	O
based	O
two	O
-	O
dimensional	O
(	O
2D	O
)	O
classification	O
,	O
are	O
focused	O
on	O
the	O
dimeric	O
CT	O
domain	O
and	O
the	O
full	O
BC	B-mutant
–	I-mutant
BCCP	I-mutant
–	I-mutant
CD	I-mutant
domain	O
of	O
only	O
one	O
protomer	O
,	O
due	O
to	O
the	O
non	O
-	O
coordinated	O
motions	O
of	O
the	O
lateral	O
BC	O
/	O
CD	O
regions	O
relative	O
to	O
the	O
CT	O
dimer	O
.	O
	O
They	O
identify	O
the	O
connections	O
between	O
CDN	O
/	O
CDL	O
and	O
between	O
CDC2	O
/	O
CT	O
as	O
major	O
contributors	O
to	O
conformational	O
heterogeneity	O
(	O
Supplementary	O
Fig	O
.	O
4a	O
,	O
b	O
).	O
	O
The	O
BC	O
domain	O
is	O
not	O
completely	O
disordered	O
,	O
but	O
laterally	O
attached	O
to	O
BT	O
/	O
CDN	O
in	O
a	O
generally	O
conserved	O
position	O
,	O
albeit	O
with	O
increased	O
flexibility	O
.	O
	O
Furthermore	O
,	O
based	O
on	O
an	O
average	O
length	O
of	O
the	O
BCCP	O
–	O
CD	O
linker	O
in	O
fungal	O
ACC	O
of	O
26	O
amino	O
acids	O
,	O
mobility	O
of	O
the	O
BCCP	O
alone	O
would	O
not	O
be	O
sufficient	O
to	O
bridge	O
the	O
active	O
sites	O
of	O
BC	O
and	O
CT	O
.	O
	O
The	O
most	O
relevant	O
candidate	O
site	O
for	O
mediating	O
such	O
additional	O
flexibility	O
and	O
permitting	O
an	O
extended	O
set	O
of	O
conformations	O
is	O
the	O
CDC1	O
/	O
CDC2	O
interface	O
,	O
which	O
is	O
rigidified	O
by	O
the	O
Ser1157	O
-	O
phosphorylated	O
regulatory	O
loop	O
,	O
as	O
depicted	O
in	O
the	O
SceCD	O
crystal	O
structure	O
.	O
	O
Altogether	O
,	O
the	O
architecture	O
of	O
fungal	O
ACC	O
is	O
based	O
on	O
the	O
central	O
dimeric	O
CT	O
domain	O
(	O
Fig	O
.	O
4d	O
).	O
	O
The	O
CD	O
has	O
no	O
direct	O
role	O
in	O
substrate	O
recognition	O
or	O
catalysis	O
but	O
contributes	O
to	O
the	O
regulation	O
of	O
all	O
eukaryotic	O
ACCs	O
.	O
	O
In	O
higher	O
eukaryotic	O
ACCs	O
,	O
regulation	O
via	O
phosphorylation	O
is	O
achieved	O
by	O
combining	O
the	O
effects	O
of	O
phosphorylation	O
at	O
Ser80	O
,	O
Ser1201	O
and	O
Ser1263	O
.	O
	O
In	O
fungal	O
ACC	O
,	O
however	O
,	O
Ser1157	O
in	O
the	O
regulatory	O
loop	O
of	O
the	O
CD	O
is	O
the	O
only	O
phosphorylation	O
site	O
that	O
has	O
been	O
demonstrated	O
to	O
be	O
both	O
phosphorylated	O
in	O
vivo	O
and	O
involved	O
in	O
the	O
regulation	O
of	O
ACC	O
activity	O
.	O
	O
In	O
its	O
phosphorylated	O
state	O
,	O
the	O
regulatory	O
loop	O
containing	O
Ser1157	O
wedges	O
between	O
CDC1	O
/	O
CDC2	O
and	O
presumably	O
limits	O
the	O
conformational	O
freedom	O
at	O
this	O
interdomain	O
interface	O
.	O
	O
However	O
,	O
flexibility	O
at	O
this	O
hinge	O
may	O
be	O
required	O
for	O
full	O
ACC	O
activity	O
,	O
as	O
the	O
distances	O
between	O
the	O
BCCP	O
anchor	O
points	O
and	O
the	O
active	O
sites	O
of	O
BC	O
and	O
CT	O
observed	O
here	O
are	O
such	O
large	O
that	O
mobility	O
of	O
the	O
BCCP	O
alone	O
is	O
not	O
sufficient	O
for	O
substrate	O
transfer	O
.	O
	O
The	O
current	O
data	O
thus	O
suggest	O
that	O
regulation	O
of	O
fungal	O
ACC	O
is	O
mediated	O
by	O
controlling	O
the	O
dynamics	O
of	O
the	O
unique	O
CD	O
,	O
rather	O
than	O
directly	O
affecting	O
catalytic	O
turnover	O
at	O
the	O
active	O
sites	O
of	O
BC	O
and	O
CT	O
.	O
	O
A	O
comparison	O
between	O
fungal	O
and	O
human	O
ACC	O
will	O
help	O
to	O
further	O
discriminate	O
mechanistic	O
differences	O
that	O
contribute	O
to	O
the	O
extended	O
control	O
and	O
polymerization	O
of	O
human	O
ACC	O
.	O
	O
Most	O
recently	O
,	O
a	O
crystal	O
structure	O
of	O
near	O
full	O
-	O
length	O
non	O
-	O
phosphorylated	O
ACC	O
from	O
S	O
.	O
cerevisae	O
(	O
lacking	O
only	O
21	O
N	O
-	O
terminal	O
amino	O
acids	O
,	O
here	O
denoted	O
as	O
flACC	O
)	O
was	O
published	O
by	O
Wei	O
and	O
Tong	O
.	O
	O
In	O
flACC	O
,	O
the	O
ACC	O
dimer	O
obeys	O
twofold	O
symmetry	O
and	O
assembles	O
in	O
a	O
triangular	O
architecture	O
with	O
dimeric	O
BC	O
domains	O
(	O
Supplementary	O
Fig	O
.	O
5a	O
).	O
	O
In	O
their	O
study	O
,	O
mutational	O
data	O
indicate	O
a	O
requirement	O
for	O
BC	O
dimerization	O
for	O
catalytic	O
activity	O
.	O
	O
The	O
transition	O
from	O
the	O
elongated	O
open	O
shape	O
,	O
observed	O
in	O
our	O
experiments	O
,	O
towards	O
a	O
compact	O
triangular	O
shape	O
is	O
based	O
on	O
an	O
intricate	O
interplay	O
of	O
several	O
hinge	O
-	O
bending	O
motions	O
in	O
the	O
CD	O
(	O
Fig	O
.	O
4d	O
).	O
	O
Comparison	O
of	O
flACC	O
with	O
our	O
CthΔBCCP	B-mutant
structure	O
reveals	O
the	O
CDC2	O
/	O
CT	O
hinge	O
as	O
a	O
major	O
contributor	O
to	O
conformational	O
flexibility	O
(	O
Supplementary	O
Fig	O
.	O
5b	O
,	O
c	O
).	O
	O
In	O
flACC	O
,	O
CDC2	O
rotates	O
∼	O
120	O
°	O
with	O
respect	O
to	O
the	O
CT	O
domain	O
.	O
	O
On	O
the	O
basis	O
of	O
a	O
superposition	O
of	O
CDC2	O
,	O
CDC1	O
of	O
the	O
phosphorylated	O
SceCD	O
is	O
rotated	O
by	O
30	O
°	O
relative	O
to	O
CDC1	O
of	O
the	O
non	O
-	O
phosphorylated	O
flACC	O
(	O
Supplementary	O
Fig	O
.	O
5d	O
),	O
similar	O
to	O
what	O
we	O
have	O
observed	O
for	O
the	O
non	O
-	O
phosphorylated	O
HsaBT	B-mutant
-	I-mutant
CD	I-mutant
(	O
Supplementary	O
Fig	O
.	O
1d	O
).	O
	O
When	O
inspecting	O
all	O
individual	O
protomer	O
and	O
fragment	B-mutant
structures	O
in	O
their	O
study	O
,	O
Wei	O
and	O
Tong	O
also	O
identify	O
the	O
CDN	O
/	O
CDC1	O
connection	O
as	O
a	O
highly	O
flexible	O
hinge	O
,	O
in	O
agreement	O
with	O
our	O
observations	O
.	O
	O
Only	O
in	O
three	O
out	O
of	O
eight	O
observed	O
protomers	O
a	O
short	O
peptide	O
stretch	O
(	O
including	O
Ser1157	O
)	O
was	O
modelled	O
.	O
	O
In	O
those	O
instances	O
the	O
Ser1157	O
residue	O
is	O
located	O
at	O
a	O
distance	O
of	O
14	O
–	O
20	O
Å	O
away	O
from	O
the	O
location	O
of	O
the	O
phosphorylated	O
serine	O
observed	O
here	O
,	O
based	O
on	O
superposition	O
of	O
either	O
CDC1	O
or	O
CDC2	O
.	O
	O
Applying	O
the	O
conformation	O
of	O
the	O
CDC1	O
/	O
CDC2	O
hinge	O
observed	O
in	O
SceCD	O
on	O
flACC	O
leads	O
to	O
CDN	O
sterically	O
clashing	O
with	O
CDC2	O
and	O
BT	O
/	O
CDN	O
clashing	O
with	O
CT	O
(	O
Supplementary	O
Fig	O
.	O
6a	O
,	O
b	O
).	O
	O
In	O
addition	O
,	O
EM	O
micrographs	O
of	O
phosphorylated	O
and	O
dephosphorylated	O
SceACC	O
display	O
for	O
both	O
samples	O
mainly	O
elongated	O
and	O
U	O
-	O
shaped	O
conformations	O
and	O
reveal	O
no	O
apparent	O
differences	O
in	O
particle	O
shape	O
distributions	O
(	O
Supplementary	O
Fig	O
.	O
7	O
).	O
	O
This	O
implicates	O
that	O
the	O
triangular	O
shape	O
with	O
dimeric	O
BC	O
domains	O
has	O
a	O
low	O
population	O
also	O
in	O
the	O
active	O
form	O
,	O
even	O
though	O
a	O
biasing	O
influence	O
of	O
grid	O
preparation	O
cannot	O
be	O
excluded	O
completely	O
.	O
	O
Large	O
-	O
scale	O
conformational	O
variability	O
has	O
also	O
been	O
observed	O
in	O
most	O
other	O
carrier	O
protein	O
-	O
based	O
multienzymes	O
,	O
including	O
polyketide	O
and	O
fatty	O
-	O
acid	O
synthases	O
(	O
with	O
the	O
exception	O
of	O
fungal	O
-	O
type	O
fatty	O
-	O
acid	O
synthases	O
),	O
non	O
-	O
ribosomal	O
peptide	O
synthetases	O
and	O
the	O
pyruvate	O
dehydrogenase	O
complexes	O
,	O
although	O
based	O
on	O
completely	O
different	O
architectures	O
.	O
	O
Together	O
,	O
this	O
structural	O
information	O
suggests	O
that	O
variable	O
carrier	O
protein	O
tethering	O
is	O
not	O
sufficient	O
for	O
efficient	O
substrate	O
transfer	O
and	O
catalysis	O
in	O
any	O
of	O
these	O
systems	O
.	O
	O
The	O
determination	O
of	O
a	O
set	O
of	O
crystal	O
structures	O
of	O
SceACC	O
in	O
two	O
states	O
,	O
unphosphorylated	O
and	O
phosphorylated	O
at	O
the	O
major	O
regulatory	O
site	O
Ser1157	O
,	O
provides	O
a	O
unique	O
depiction	O
of	O
multienzyme	O
regulation	O
by	O
post	O
-	O
translational	O
modification	O
(	O
Fig	O
.	O
4d	O
).	O
	O
It	O
disfavours	O
the	O
adoption	O
of	O
a	O
rare	O
,	O
compact	O
conformation	O
,	O
in	O
which	O
intramolecular	O
dimerization	O
of	O
the	O
BC	O
domains	O
results	O
in	O
catalytic	O
turnover	O
.	O
	O
The	O
regulation	O
of	O
activity	O
thus	O
results	O
from	O
restrained	O
large	O
-	O
scale	O
conformational	O
dynamics	O
rather	O
than	O
a	O
direct	O
or	O
indirect	O
influence	O
on	O
active	O
site	O
structure	O
.	O
	O
To	O
our	O
best	O
knowledge	O
,	O
ACC	O
is	O
the	O
first	O
multienzyme	O
for	O
which	O
such	O
a	O
phosphorylation	O
-	O
dependent	O
mechanical	O
control	O
mechanism	O
has	O
been	O
visualized	O
.	O
	O
However	O
,	O
the	O
example	O
of	O
ACC	O
now	O
demonstrates	O
the	O
possibility	O
of	O
regulating	O
activity	O
by	O
controlled	O
dynamics	O
of	O
non	O
-	O
enzymatic	O
linker	O
regions	O
also	O
in	O
other	O
families	O
of	O
carrier	O
-	O
dependent	O
multienzymes	O
.	O
	O
The	O
phosphorylated	O
central	O
domain	O
of	O
yeast	O
ACC	O
.	O
	O
(	O
a	O
)	O
Schematic	O
overview	O
of	O
the	O
domain	O
organization	O
of	O
eukaryotic	O
ACCs	O
.	O
	O
Crystallized	O
constructs	O
are	O
indicated	O
.	O
	O
CDN	O
is	O
linked	O
by	O
a	O
four	O
-	O
helix	O
bundle	O
(	O
CDL	O
)	O
to	O
two	O
α	O
–	O
β	O
-	O
fold	O
domains	O
(	O
CDC1	O
and	O
CDC2	O
).	O
	O
The	O
regulatory	O
loop	O
is	O
shown	O
as	O
bold	O
cartoon	O
,	O
and	O
the	O
phosphorylated	O
Ser1157	O
is	O
marked	O
by	O
a	O
red	O
triangle	O
.	O
	O
(	O
e	O
)	O
Structural	O
overview	O
of	O
HsaBT	B-mutant
-	I-mutant
CD	I-mutant
.	O
	O
The	O
attachment	O
points	O
to	O
the	O
N	O
-	O
terminal	O
BCCP	O
domain	O
and	O
the	O
C	O
-	O
terminal	O
CT	O
domain	O
are	O
indicated	O
with	O
spheres	O
.	O
	O
Architecture	O
of	O
the	O
CD	O
–	O
CT	O
core	O
of	O
fungal	O
ACC	O
.	O
	O
One	O
protomer	O
is	O
shown	O
in	O
colour	O
and	O
one	O
in	O
grey	O
.	O
	O
Individual	O
domains	O
are	O
labelled	O
;	O
the	O
active	O
site	O
of	O
CT	O
and	O
the	O
position	O
of	O
the	O
conserved	O
regulatory	O
phosphoserine	O
site	O
based	O
on	O
SceCD	O
are	O
indicated	O
by	O
an	O
asterisk	O
and	O
a	O
triangle	O
,	O
respectively	O
.	O
	O
Variability	O
of	O
the	O
connections	O
of	O
CDC2	O
to	O
CT	O
and	O
CDC1	O
in	O
fungal	O
ACC	O
.	O
	O
(	O
a	O
)	O
Hinge	O
properties	O
of	O
the	O
CDC2	O
–	O
CT	O
connection	O
analysed	O
by	O
a	O
CT	O
-	O
based	O
superposition	O
of	O
eight	O
instances	O
of	O
the	O
CDC2	B-mutant
-	I-mutant
CT	I-mutant
segment	I-mutant
.	O
	O
For	O
clarity	O
,	O
only	O
one	O
protomer	O
of	O
CthCD	B-mutant
-	I-mutant
CTCter1	I-mutant
is	O
shown	O
in	O
full	O
colour	O
as	O
reference	O
.	O
	O
For	O
other	O
instances	O
,	O
CDC2	O
domains	O
are	O
shown	O
in	O
transparent	O
tube	O
representation	O
with	O
only	O
one	O
helix	O
each	O
highlighted	O
.	O
	O
Representation	O
as	O
in	O
a	O
,	O
but	O
the	O
CDC1	O
and	O
CDC2	O
are	O
superposed	O
based	O
on	O
CDC2	O
.	O
	O
One	O
protomer	O
of	O
CthΔBCCP	B-mutant
is	O
shown	O
in	O
colour	O
,	O
the	O
CDL	O
domains	O
are	O
omitted	O
for	O
clarity	O
and	O
the	O
position	O
of	O
the	O
phosphorylated	O
serine	O
based	O
on	O
SceCD	O
is	O
indicated	O
with	O
a	O
red	O
triangle	O
.	O
	O
(	O
a	O
–	O
c	O
)	O
Large	O
-	O
scale	O
conformational	O
variability	O
of	O
the	O
CDN	O
domain	O
relative	O
to	O
the	O
CDL	O
/	O
CDC1	O
domain	O
.	O
	O
Domains	O
other	O
than	O
CDN	O
and	O
CDL	O
/	O
CDC1	O
are	O
omitted	O
for	O
clarity	O
.	O
	O
(	O
d	O
)	O
Schematic	O
model	O
of	O
fungal	O
ACC	O
showing	O
the	O
intrinsic	O
,	O
regulated	O
flexibility	O
of	O
CD	O
in	O
the	O
phosphorylated	O
inhibited	O
or	O
the	O
non	O
-	O
phosphorylated	O
activated	O
state	O
.	O
	O
Flexibility	O
of	O
the	O
CDC2	O
/	O
CT	O
and	O
CDN	O
/	O
CDL	O
hinges	O
is	O
illustrated	O
by	O
arrows	O
.	O
	O
The	O
Ser1157	O
phosphorylation	O
site	O
and	O
the	O
regulatory	O
loop	O
are	O
schematically	O
indicated	O
in	O
magenta	O
.	O
	O
The	O
genome	O
of	O
the	O
Synechococcus	O
elongatus	O
strain	O
PCC	O
7942	O
encodes	O
a	O
putative	O
sugar	O
kinase	O
(	O
SePSK	O
),	O
which	O
shares	O
44	O
.	O
9	O
%	O
sequence	O
identity	O
with	O
the	O
xylulose	O
kinase	O
-	O
1	O
(	O
AtXK	O
-	O
1	O
)	O
from	O
Arabidopsis	O
thaliana	O
.	O
	O
Sequence	O
alignment	O
suggests	O
that	O
both	O
kinases	O
belong	O
to	O
the	O
ribulokinase	O
-	O
like	O
carbohydrate	O
kinases	O
,	O
a	O
sub	O
-	O
family	O
of	O
FGGY	O
family	O
carbohydrate	O
kinases	O
.	O
	O
In	O
addition	O
,	O
our	O
enzymatic	O
assays	O
suggested	O
that	O
SePSK	O
has	O
the	O
capability	O
to	O
phosphorylate	O
D	O
-	O
ribulose	O
.	O
	O
In	O
order	O
to	O
understand	O
the	O
catalytic	O
mechanism	O
of	O
SePSK	O
,	O
we	O
solved	O
the	O
structure	O
of	O
SePSK	O
in	O
complex	O
with	O
D	O
-	O
ribulose	O
and	O
found	O
two	O
potential	O
substrate	O
binding	O
pockets	O
in	O
SePSK	O
.	O
	O
Using	O
mutation	O
and	O
activity	O
analysis	O
,	O
we	O
further	O
verified	O
the	O
key	O
residues	O
important	O
for	O
its	O
catalytic	O
activity	O
.	O
	O
Moreover	O
,	O
our	O
structural	O
comparison	O
with	O
other	O
family	O
members	O
suggests	O
that	O
there	O
are	O
major	O
conformational	O
changes	O
in	O
SePSK	O
upon	O
substrate	O
binding	O
,	O
facilitating	O
the	O
catalytic	O
process	O
.	O
	O
Together	O
,	O
these	O
results	O
provide	O
important	O
information	O
for	O
a	O
more	O
detailed	O
understanding	O
of	O
the	O
cofactor	O
and	O
substrate	O
binding	O
mode	O
as	O
well	O
as	O
the	O
catalytic	O
mechanism	O
of	O
SePSK	O
,	O
and	O
possible	O
similarities	O
with	O
its	O
plant	O
homologue	O
AtXK	O
-	O
1	O
.	O
	O
Carbohydrates	O
are	O
essential	O
cellular	O
compounds	O
involved	O
in	O
the	O
metabolic	O
processes	O
present	O
in	O
all	O
organisms	O
.	O
	O
The	O
FGGY	O
family	O
carbohydrate	O
kinases	O
contain	O
different	O
types	O
of	O
sugar	O
kinases	O
,	O
all	O
of	O
which	O
possess	O
different	O
catalytic	O
substrates	O
with	O
preferences	O
for	O
short	O
-	O
chained	O
sugar	O
substrates	O
,	O
ranging	O
from	O
triose	O
to	O
heptose	O
.	O
	O
These	O
sugar	O
substrates	O
include	O
L	O
-	O
ribulose	O
,	O
erythritol	O
,	O
L	O
-	O
fuculose	O
,	O
D	O
-	O
glycerol	O
,	O
D	O
-	O
gluconate	O
,	O
L	O
-	O
xylulose	O
,	O
D	O
-	O
ribulose	O
,	O
L	O
-	O
rhamnulose	O
and	O
D	O
-	O
xylulose	O
.	O
	O
Structures	O
reported	O
in	O
the	O
Protein	O
Data	O
Bank	O
of	O
the	O
FGGY	O
family	O
carbohydrate	O
kinases	O
exhibit	O
a	O
similar	O
overall	O
architecture	O
containing	O
two	O
protein	O
domains	O
,	O
one	O
of	O
which	O
is	O
responsible	O
for	O
the	O
binding	O
of	O
substrate	O
,	O
while	O
the	O
second	O
is	O
used	O
for	O
binding	O
cofactor	O
ATP	O
.	O
	O
While	O
the	O
binding	O
pockets	O
for	O
substrates	O
are	O
at	O
the	O
same	O
position	O
,	O
each	O
FGGY	O
family	O
carbohydrate	O
kinases	O
uses	O
different	O
substrate	O
-	O
binding	O
residues	O
,	O
resulting	O
in	O
high	O
substrate	O
specificity	O
.	O
	O
Synpcc7942_2462	O
from	O
the	O
cyanobacteria	O
Synechococcus	O
elongatus	O
PCC	O
7942	O
encodes	O
a	O
putative	O
sugar	O
kinase	O
(	O
SePSK	O
),	O
and	O
this	O
kinase	O
contains	O
426	O
amino	O
acids	O
.	O
	O
The	O
At2g21370	O
gene	O
product	O
from	O
Arabidopsis	O
thaliana	O
,	O
xylulose	O
kinase	O
-	O
1	O
(	O
AtXK	O
-	O
1	O
),	O
whose	O
mature	O
form	O
contains	O
436	O
amino	O
acids	O
,	O
is	O
located	O
in	O
the	O
chloroplast	O
(	O
ChloroP	O
1	O
.	O
1	O
Server	O
).	O
	O
Members	O
of	O
this	O
sub	O
-	O
family	O
are	O
responsible	O
for	O
the	O
phosphorylation	O
of	O
sugars	O
similar	O
to	O
L	O
-	O
ribulose	O
and	O
D	O
-	O
ribulose	O
.	O
	O
The	O
sequence	O
and	O
the	O
substrate	O
specificity	O
of	O
ribulokinase	O
-	O
like	O
carbohydrate	O
kinases	O
are	O
different	O
,	O
but	O
they	O
share	O
the	O
common	O
folding	O
feature	O
with	O
two	O
domains	O
.	O
	O
Domain	O
I	O
exhibits	O
a	O
ribonuclease	O
H	O
-	O
like	O
folding	O
pattern	O
,	O
and	O
is	O
responsible	O
for	O
the	O
substrate	O
binding	O
,	O
while	O
domain	O
II	O
possesses	O
an	O
actin	O
-	O
like	O
ATPase	O
domain	O
that	O
binds	O
cofactor	O
ATP	O
.	O
	O
Two	O
possible	O
xylulose	O
kinases	O
(	O
xylulose	O
kinase	O
-	O
1	O
:	O
XK	O
-	O
1	O
and	O
xylulose	O
kinase	O
-	O
2	O
:	O
XK	O
-	O
2	O
)	O
from	O
Arabidopsis	O
thaliana	O
were	O
previously	O
proposed	O
.	O
	O
It	O
was	O
shown	O
that	O
XK	O
-	O
2	O
(	O
At5g49650	O
)	O
located	O
in	O
the	O
cytosol	O
is	O
indeed	O
xylulose	O
kinase	O
.	O
	O
SePSK	O
from	O
Synechococcus	O
elongatus	O
strain	O
PCC	O
7942	O
is	O
the	O
homolog	O
of	O
AtXK	O
-	O
1	O
,	O
though	O
its	O
physiological	O
function	O
and	O
substrates	O
remain	O
unclear	O
.	O
	O
In	O
order	O
to	O
obtain	O
functional	O
and	O
structural	O
information	O
about	O
these	O
two	O
proteins	O
,	O
here	O
we	O
reported	O
the	O
crystal	O
structures	O
of	O
SePSK	O
and	O
AtXK	O
-	O
1	O
.	O
	O
Overall	O
structures	O
of	O
apo	O
-	O
SePSK	O
and	O
apo	O
-	O
AtXK	O
-	O
1	O
	O
The	O
attempt	O
to	O
solve	O
the	O
SePSK	O
structure	O
by	O
molecular	O
replacement	O
method	O
failed	O
with	O
ribulokinase	O
from	O
Bacillus	O
halodurans	O
(	O
PDB	O
code	O
:	O
3QDK	O
,	O
15	O
.	O
7	O
%	O
sequence	O
identity	O
)	O
as	O
an	O
initial	O
model	O
.	O
	O
We	O
therefore	O
used	O
single	O
isomorphous	O
replacement	O
anomalous	O
scattering	O
method	O
(	O
SIRAS	O
)	O
for	O
successful	O
solution	O
of	O
the	O
apo	O
-	O
SePSK	O
structure	O
at	O
a	O
resolution	O
of	O
2	O
.	O
3	O
Å	O
.	O
Subsequently	O
,	O
the	O
apo	O
-	O
SePSK	O
structure	O
was	O
used	O
as	O
molecular	O
replacement	O
model	O
to	O
solve	O
all	O
other	O
structures	O
identified	O
in	O
this	O
study	O
.	O
	O
Our	O
structural	O
analysis	O
showed	O
that	O
apo	O
-	O
SePSK	O
consists	O
of	O
one	O
SePSK	O
protein	O
molecule	O
in	O
an	O
asymmetric	O
unit	O
.	O
	O
The	O
amino	O
-	O
acid	O
residues	O
were	O
traced	O
from	O
Val2	O
to	O
His419	O
,	O
except	O
for	O
the	O
Met1	O
residue	O
and	O
the	O
seven	O
residues	O
at	O
the	O
C	O
-	O
termini	O
.	O
	O
Domain	O
I	O
consists	O
of	O
non	O
-	O
contiguous	O
portions	O
of	O
the	O
polypeptide	O
chains	O
(	O
aa	O
.	O
	O
402	O
–	O
419	O
),	O
exhibiting	O
11	O
α	O
-	O
helices	O
and	O
11	O
β	O
-	O
sheets	O
.	O
	O
In	O
addition	O
,	O
four	O
β	O
-	O
sheets	O
(	O
β7	O
,	O
β10	O
,	O
β12	O
and	O
β16	O
)	O
and	O
five	O
α	O
-	O
helices	O
(	O
α8	O
,	O
α9	O
,	O
α13	O
,	O
α14	O
and	O
α15	O
)	O
flank	O
the	O
left	O
side	O
of	O
the	O
core	O
region	O
.	O
	O
Domain	O
II	O
is	O
comprised	O
of	O
aa	O
.	O
	O
229	O
–	O
401	O
and	O
classified	O
into	O
B2	O
(	O
β31	O
/	O
β29	O
/	O
β22	O
/	O
β23	O
/	O
β25	O
/	O
β24	O
)	O
and	O
A3	O
(	O
α26	O
/	O
α27	O
/	O
α28	O
/	O
α30	O
)	O
(	O
Fig	O
1A	O
and	O
S1	O
Fig	O
).	O
	O
In	O
the	O
SePSK	O
structure	O
,	O
B1	O
and	O
B2	O
are	O
sandwiched	O
by	O
A1	O
,	O
A2	O
and	O
A3	O
,	O
and	O
the	O
whole	O
structure	O
shows	O
the	O
A1	O
/	O
B1	O
/	O
A2	O
/	O
B2	O
/	O
A3	O
(	O
α	O
/	O
β	O
/	O
α	O
/	O
β	O
/	O
α	O
)	O
folding	O
pattern	O
,	O
which	O
is	O
in	O
common	O
with	O
other	O
members	O
of	O
FGGY	O
family	O
carbohydrate	O
kinases	O
(	O
S2	O
Fig	O
).	O
	O
The	O
overall	O
folding	O
of	O
SePSK	O
resembles	O
a	O
clip	O
,	O
with	O
A2	O
of	O
domain	O
I	O
acting	O
as	O
a	O
hinge	O
region	O
.	O
	O
Overall	O
structures	O
of	O
SePSK	O
and	O
AtXK	O
-	O
1	O
.	O
	O
(	O
A	O
)	O
Three	O
-	O
dimensional	O
structure	O
of	O
apo	O
-	O
SePSK	O
.	O
	O
The	O
secondary	O
structural	O
elements	O
are	O
indicated	O
(	O
α	O
-	O
helix	O
:	O
cyan	O
,	O
β	O
-	O
sheet	O
:	O
yellow	O
).	O
	O
(	O
B	O
)	O
Three	O
-	O
dimensional	O
structure	O
of	O
apo	O
-	O
AtXK	O
-	O
1	O
.	O
	O
Apo	O
-	O
AtXK	O
-	O
1	O
exhibits	O
a	O
folding	O
pattern	O
similar	O
to	O
that	O
of	O
SePSK	O
in	O
line	O
with	O
their	O
high	O
sequence	O
identity	O
(	O
Fig	O
1B	O
and	O
S1	O
Fig	O
).	O
	O
However	O
,	O
superposition	O
of	O
structures	O
of	O
AtXK	O
-	O
1	O
and	O
SePSK	O
shows	O
some	O
differences	O
,	O
especially	O
at	O
the	O
loop	O
regions	O
.	O
	O
A	O
considerable	O
difference	O
is	O
found	O
in	O
the	O
loop3	O
linking	O
β3	O
and	O
α4	O
,	O
which	O
is	O
stretched	O
out	O
in	O
the	O
AtXK	O
-	O
1	O
structure	O
,	O
while	O
in	O
the	O
SePSK	O
structure	O
,	O
it	O
is	O
bent	O
back	O
towards	O
the	O
inner	O
part	O
.	O
	O
The	O
corresponding	O
residues	O
between	O
these	O
two	O
structures	O
(	O
SePSK	O
-	O
Lys35	O
and	O
AtXK	O
-	O
1	O
-	O
Lys48	O
)	O
have	O
a	O
distance	O
of	O
15	O
.	O
4	O
Å	O
(	O
S3	O
Fig	O
).	O
	O
Activity	O
assays	O
of	O
SePSK	O
and	O
AtXK	O
-	O
1	O
	O
In	O
order	O
to	O
understand	O
the	O
function	O
of	O
these	O
two	O
kinases	O
,	O
we	O
performed	O
structural	O
comparison	O
using	O
Dali	O
server	O
.	O
	O
We	O
first	O
tested	O
whether	O
both	O
enzymes	O
possessed	O
ATP	O
hydrolysis	O
activity	O
in	O
the	O
absence	O
of	O
substrates	O
.	O
	O
This	O
finding	O
is	O
in	O
agreement	O
with	O
a	O
previous	O
result	O
showing	O
that	O
xylulose	O
kinase	O
(	O
PDB	O
code	O
:	O
2ITM	O
)	O
possessed	O
ATP	O
hydrolysis	O
activity	O
without	O
adding	O
substrate	O
.	O
	O
As	O
shown	O
in	O
Fig	O
2B	O
,	O
the	O
ATP	O
hydrolysis	O
activity	O
of	O
SePSK	O
greatly	O
increased	O
upon	O
adding	O
D	O
-	O
ribulose	O
than	O
adding	O
other	O
potential	O
substrates	O
,	O
suggesting	O
that	O
it	O
has	O
D	O
-	O
ribulose	O
kinase	O
activity	O
.	O
	O
In	O
contrary	O
,	O
limited	O
increasing	O
of	O
ATP	O
hydrolysis	O
activity	O
was	O
detected	O
for	O
AtXK	O
-	O
1	O
upon	O
addition	O
of	O
D	O
-	O
ribulose	O
(	O
Fig	O
2C	O
),	O
despite	O
its	O
structural	O
similarity	O
with	O
SePSK	O
.	O
	O
The	O
enzymatic	O
activity	O
assays	O
of	O
SePSK	O
and	O
AtXK	O
-	O
1	O
.	O
	O
The	O
substrates	O
are	O
DR	O
(	O
D	O
-	O
ribulose	O
),	O
LR	O
(	O
L	O
-	O
ribulose	O
),	O
DX	O
(	O
D	O
-	O
xylulose	O
),	O
LX	O
(	O
L	O
-	O
xylulose	O
)	O
and	O
GLY	O
(	O
Glycerol	O
).	O
(	O
C	O
)	O
The	O
ATP	O
hydrolysis	O
activity	O
of	O
SePSK	O
and	O
AtXK	O
-	O
1	O
with	O
or	O
without	O
D	O
-	O
ribulose	O
.	O
(	O
D	O
)	O
The	O
ATP	O
hydrolysis	O
activity	O
of	O
wild	O
-	O
type	O
(	O
WT	O
)	O
and	O
single	O
-	O
site	O
mutants	O
of	O
SePSK	O
.	O
	O
Three	O
single	O
-	O
site	O
mutants	O
of	O
SePSK	O
are	O
D8A	B-mutant
-	O
SePSK	O
,	O
T11A	B-mutant
-	O
SePSK	O
and	O
D221A	B-mutant
-	O
SePSK	O
.	O
	O
The	O
ATP	O
hydrolysis	O
activity	O
measured	O
via	O
luminescent	O
ADP	O
-	O
Glo	O
assay	O
(	O
Promega	O
).	O
	O
To	O
understand	O
the	O
catalytic	O
mechanism	O
of	O
SePSK	O
,	O
we	O
performed	O
structural	O
comparisons	O
among	O
xylulose	O
kinase	O
,	O
glycerol	O
kinase	O
,	O
ribulose	O
kinase	O
and	O
SePSK	O
.	O
	O
Our	O
results	O
suggested	O
that	O
three	O
conserved	O
residues	O
(	O
D8	O
,	O
T11	O
and	O
D221	O
of	O
SePSK	O
)	O
play	O
an	O
important	O
role	O
in	O
SePSK	O
function	O
.	O
	O
Mutations	O
of	O
the	O
corresponding	O
residue	O
in	O
xylulose	O
kinase	O
and	O
glycerol	O
kinase	O
from	O
Escherichia	O
coli	O
greatly	O
reduced	O
their	O
activity	O
.	O
	O
To	O
identify	O
the	O
function	O
of	O
these	O
three	O
residues	O
of	O
SePSK	O
,	O
we	O
constructed	O
D8A	B-mutant
,	O
T11A	B-mutant
and	O
D221A	B-mutant
mutants	O
.	O
	O
Using	O
enzymatic	O
activity	O
assays	O
,	O
we	O
found	O
that	O
all	O
of	O
these	O
mutants	O
exhibit	O
much	O
lower	O
activity	O
of	O
ATP	O
hydrolysis	O
after	O
adding	O
D	O
-	O
ribulose	O
than	O
that	O
of	O
wild	O
type	O
,	O
indicating	O
the	O
possibility	O
that	O
these	O
three	O
residues	O
are	O
involved	O
in	O
the	O
catalytic	O
process	O
of	O
phosphorylation	O
D	O
-	O
ribulose	O
and	O
are	O
vital	O
for	O
the	O
function	O
of	O
SePSK	O
(	O
Fig	O
2D	O
).	O
	O
In	O
both	O
structures	O
,	O
a	O
strong	O
electron	O
density	O
was	O
found	O
in	O
the	O
conserved	O
ATP	O
binding	O
pocket	O
,	O
but	O
can	O
only	O
be	O
fitted	O
with	O
an	O
ADP	O
molecule	O
(	O
S4	O
Fig	O
).	O
	O
The	O
extremely	O
weak	O
electron	O
densities	O
of	O
ATP	O
γ	O
-	O
phosphate	O
in	O
both	O
structures	O
suggest	O
that	O
the	O
γ	O
-	O
phosphate	O
group	O
of	O
ATP	O
is	O
either	O
flexible	O
or	O
hydrolyzed	O
by	O
SePSK	O
and	O
AtXK	O
-	O
1	O
.	O
	O
To	O
avoid	O
hydrolysis	O
of	O
ATP	O
,	O
we	O
soaked	O
the	O
crystals	O
of	O
apo	O
-	O
SePSK	O
and	O
apo	O
-	O
AtXK	O
-	O
1	O
into	O
the	O
reservoir	O
adding	O
AMP	O
-	O
PNP	O
.	O
	O
However	O
,	O
we	O
found	O
that	O
the	O
electron	O
densities	O
of	O
γ	O
-	O
phosphate	O
group	O
of	O
AMP	O
-	O
PNP	O
(	O
AMP	O
-	O
PNP	O
γ	O
-	O
phosphate	O
)	O
are	O
still	O
weak	O
in	O
the	O
AMP	O
-	O
PNP	O
-	O
SePSK	O
and	O
AMP	O
-	O
PNP	O
-	O
AtXK	O
-	O
1	O
structures	O
,	O
suggesting	O
high	O
flexibility	O
of	O
ATP	O
-	O
γ	O
-	O
phosphate	O
.	O
	O
The	O
γ	O
-	O
phosphate	O
group	O
of	O
ATP	O
is	O
transferred	O
to	O
the	O
sugar	O
substrate	O
during	O
the	O
reaction	O
process	O
,	O
so	O
this	O
flexibility	O
might	O
be	O
important	O
for	O
the	O
ability	O
of	O
these	O
kinases	O
.	O
	O
The	O
overall	O
structures	O
as	O
well	O
as	O
the	O
coordination	O
modes	O
of	O
ADP	O
and	O
AMP	O
-	O
PNP	O
in	O
the	O
AMP	O
-	O
PNP	O
-	O
AtXK	O
-	O
1	O
,	O
ADP	O
-	O
AtXK	O
-	O
1	O
,	O
ADP	O
-	O
SePSK	O
and	O
AMP	O
-	O
PNP	O
-	O
SePSK	O
structures	O
are	O
nearly	O
identical	O
(	O
S5	O
Fig	O
),	O
therefore	O
the	O
structure	O
of	O
AMP	O
-	O
PNP	O
-	O
SePSK	O
is	O
used	O
here	O
to	O
describe	O
the	O
structural	O
details	O
and	O
to	O
compare	O
with	O
those	O
of	O
other	O
family	O
members	O
.	O
	O
As	O
shown	O
in	O
Fig	O
3A	O
,	O
one	O
SePSK	O
protein	O
molecule	O
is	O
in	O
an	O
asymmetric	O
unit	O
with	O
one	O
AMP	O
-	O
PNP	O
molecule	O
.	O
	O
The	O
AMP	O
-	O
PNP	O
is	O
bound	O
at	O
the	O
domain	O
II	O
,	O
where	O
it	O
fits	O
well	O
inside	O
a	O
positively	O
charged	O
groove	O
.	O
	O
The	O
AMP	O
-	O
PNP	O
binding	O
pocket	O
consists	O
of	O
four	O
α	O
-	O
helices	O
(	O
α26	O
,	O
α28	O
,	O
α27	O
and	O
α30	O
)	O
and	O
forms	O
a	O
shape	O
resembling	O
a	O
half	O
-	O
fist	O
(	O
Fig	O
3A	O
and	O
3B	O
).	O
	O
The	O
head	O
group	O
of	O
the	O
AMP	O
-	O
PNP	O
is	O
embedded	O
in	O
a	O
pocket	O
surrounded	O
by	O
Trp383	O
,	O
Asn380	O
,	O
Gly376	O
and	O
Gly377	O
.	O
	O
The	O
purine	O
ring	O
of	O
AMP	O
-	O
PNP	O
is	O
positioned	O
in	O
parallel	O
to	O
the	O
indole	O
ring	O
of	O
Trp383	O
.	O
	O
In	O
addition	O
,	O
it	O
is	O
hydrogen	O
-	O
bonded	O
with	O
the	O
side	O
chain	O
amide	O
of	O
Asn380	O
(	O
Fig	O
3B	O
).	O
	O
Together	O
,	O
this	O
structure	O
clearly	O
shows	O
that	O
the	O
AMP	O
-	O
PNP	O
-	O
β	O
-	O
phosphate	O
is	O
sticking	O
out	O
of	O
the	O
ATP	O
binding	O
pocket	O
,	O
thus	O
the	O
γ	O
-	O
phosphate	O
group	O
is	O
at	O
the	O
empty	O
space	O
between	O
domain	O
I	O
and	O
domain	O
II	O
and	O
is	O
unconstrained	O
in	O
its	O
movement	O
by	O
the	O
protein	O
.	O
	O
The	O
SePSK	O
structure	O
is	O
shown	O
in	O
the	O
electrostatic	O
potential	O
surface	O
mode	O
.	O
	O
The	O
head	O
of	O
AMP	O
-	O
PNP	O
is	O
sandwiched	O
by	O
four	O
residues	O
(	O
Leu293	O
,	O
Gly376	O
,	O
Gly377	O
and	O
Trp383	O
).	O
	O
The	O
results	O
from	O
our	O
activity	O
assays	O
suggested	O
that	O
SePSK	O
has	O
D	O
-	O
ribulose	O
kinase	O
activity	O
.	O
	O
As	O
shown	O
in	O
S6	O
Fig	O
,	O
two	O
residual	O
electron	O
densities	O
are	O
visible	O
in	O
domain	O
I	O
,	O
which	O
can	O
be	O
interpreted	O
as	O
two	O
D	O
-	O
ribulose	O
molecules	O
with	O
reasonable	O
fit	O
.	O
	O
7	O
.	O
1	O
Å	O
(	O
RBL1	O
-	O
C4	O
and	O
RBL2	O
-	O
C1	O
).	O
	O
RBL1	O
is	O
located	O
in	O
the	O
pocket	O
consisting	O
of	O
α21	O
and	O
the	O
loop	O
between	O
β6	O
and	O
β7	O
.	O
	O
The	O
O4	O
and	O
O5	O
of	O
RBL1	O
are	O
coordinated	O
with	O
the	O
side	O
chain	O
carboxyl	O
group	O
of	O
Asp221	O
.	O
	O
Furthermore	O
,	O
the	O
O2	O
of	O
RBL1	O
interacts	O
with	O
the	O
main	O
chain	O
amide	O
nitrogen	O
of	O
Ser72	O
(	O
Fig	O
4B	O
).	O
	O
This	O
pocket	O
is	O
at	O
a	O
similar	O
position	O
of	O
substrate	O
binding	O
site	O
of	O
other	O
sugar	O
kinase	O
,	O
such	O
as	O
L	O
-	O
ribulokinase	O
(	O
PDB	O
code	O
:	O
3QDK	O
)	O
(	O
S7	O
Fig	O
).	O
	O
However	O
,	O
structural	O
comparison	O
shows	O
that	O
the	O
substrate	O
ligating	O
residues	O
between	O
the	O
two	O
structures	O
are	O
not	O
strictly	O
conserved	O
.	O
	O
Based	O
on	O
the	O
structures	O
,	O
the	O
ligating	O
residues	O
of	O
RBL1	O
in	O
RBL	O
-	O
SePSK	O
structure	O
are	O
Ser72	O
,	O
Asp221	O
and	O
Ser222	O
,	O
and	O
the	O
interacting	O
residues	O
of	O
L	O
-	O
ribulose	O
with	O
L	O
-	O
ribulokinase	O
are	O
Ala96	O
,	O
Lys208	O
,	O
Asp274	O
and	O
Glu329	O
(	O
S7	O
Fig	O
).	O
	O
Glu329	O
in	O
3QDK	O
has	O
no	O
counterpart	O
in	O
RBL	O
-	O
SePSK	O
structure	O
.	O
	O
In	O
addition	O
,	O
although	O
Lys208	O
of	O
L	O
-	O
ribulokinase	O
has	O
the	O
corresponding	O
residue	O
(	O
Lys163	O
)	O
in	O
RBL	O
-	O
SePSK	O
structure	O
,	O
the	O
hydrogen	O
bond	O
of	O
Lys163	O
is	O
broken	O
because	O
of	O
the	O
conformational	O
change	O
of	O
two	O
α	O
-	O
helices	O
(	O
α9	O
and	O
α13	O
)	O
of	O
SePSK	O
.	O
	O
The	O
binding	O
of	O
D	O
-	O
ribulose	O
(	O
RBL	O
)	O
with	O
SePSK	O
.	O
	O
(	O
A	O
)	O
The	O
electrostatic	O
potential	O
surface	O
map	O
of	O
RBL	O
-	O
SePSK	O
and	O
a	O
zoom	O
-	O
in	O
view	O
of	O
RBL	O
binding	O
site	O
.	O
	O
Single	O
-	O
cycle	O
kinetic	O
data	O
are	O
reflecting	O
the	O
interaction	O
of	O
SePSK	O
and	O
D8A	B-mutant
-	O
SePSK	O
with	O
D	O
-	O
ribulose	O
.	O
	O
It	O
shows	O
two	O
experimental	O
sensorgrams	O
after	O
minus	O
the	O
empty	O
sensorgrams	O
.	O
	O
The	O
original	O
data	O
is	O
shown	O
as	O
black	O
curve	O
,	O
and	O
the	O
fitted	O
data	O
is	O
shown	O
as	O
different	O
color	O
(	O
wild	O
type	O
SePSK	O
:	O
red	O
curve	O
,	O
D8A	B-mutant
-	O
SePSK	O
:	O
green	O
curve	O
).	O
	O
Dissociation	O
rate	O
constant	O
of	O
wild	O
type	O
and	O
D8A	B-mutant
-	O
SePSK	O
are	O
3	O
ms	O
-	O
1	O
and	O
9	O
ms	O
-	O
1	O
,	O
respectively	O
.	O
	O
The	O
binding	O
pocket	O
of	O
RBL2	O
with	O
relatively	O
weak	O
electron	O
density	O
is	O
near	O
the	O
N	O
-	O
terminal	O
region	O
of	O
SePSK	O
and	O
is	O
negatively	O
charged	O
.	O
	O
The	O
side	O
chain	O
of	O
Asp8	O
interacts	O
strongly	O
with	O
O3	O
and	O
O4	O
of	O
RBL2	O
.	O
	O
The	O
backbone	O
amide	O
nitrogens	O
of	O
Gly13	O
and	O
Arg15	O
also	O
keep	O
hydrogen	O
bonds	O
with	O
RBL2	O
(	O
Fig	O
4B	O
).	O
	O
This	O
break	O
is	O
probably	O
induced	O
by	O
the	O
conformational	O
change	O
of	O
the	O
two	O
β	O
-	O
sheets	O
(	O
β1	O
and	O
β2	O
),	O
with	O
the	O
result	O
that	O
the	O
linking	O
loop	O
(	O
loop	O
1	O
)	O
is	O
located	O
further	O
away	O
from	O
the	O
RBL2	O
binding	O
site	O
.	O
	O
Our	O
SePSK	O
structure	O
shows	O
that	O
the	O
Asp8	O
residue	O
forms	O
strong	O
hydrogen	O
bond	O
with	O
RBL2	O
(	O
Fig	O
4B	O
).	O
	O
In	O
addition	O
,	O
our	O
enzymatic	O
assays	O
indicated	O
that	O
Asp8	O
is	O
important	O
for	O
the	O
activity	O
of	O
SePSK	O
(	O
Fig	O
2D	O
).	O
	O
To	O
further	O
verified	O
this	O
result	O
,	O
we	O
measured	O
the	O
binding	O
affinity	O
for	O
D	O
-	O
ribulose	O
of	O
both	O
wild	O
type	O
(	O
WT	O
)	O
and	O
D8A	B-mutant
mutant	O
of	O
SePSK	O
using	O
a	O
surface	O
plasmon	O
resonance	O
method	O
.	O
	O
Dissociation	O
rate	O
constant	O
(	O
Kd	O
)	O
of	O
wild	O
type	O
and	O
D8A	B-mutant
-	O
SePSK	O
are	O
3	O
ms	O
-	O
1	O
and	O
9	O
ms	O
-	O
1	O
,	O
respectively	O
.	O
	O
The	O
results	O
implied	O
that	O
the	O
second	O
RBL	O
binding	O
site	O
plays	O
a	O
role	O
in	O
the	O
D	O
-	O
ribulose	O
kinase	O
function	O
of	O
SePSK	O
.	O
	O
However	O
,	O
considering	O
the	O
high	O
concentration	O
of	O
D	O
-	O
ribulose	O
used	O
for	O
crystal	O
soaking	O
,	O
as	O
well	O
as	O
the	O
relatively	O
weak	O
electron	O
density	O
of	O
RBL2	O
,	O
it	O
is	O
also	O
possible	O
that	O
the	O
second	O
binding	O
site	O
of	O
D	O
-	O
ribulose	O
in	O
SePSK	O
is	O
an	O
artifact	O
.	O
	O
Simulated	O
conformational	O
change	O
of	O
SePSK	O
during	O
the	O
catalytic	O
process	O
	O
It	O
was	O
reported	O
earlier	O
that	O
the	O
crossing	O
angle	O
between	O
the	O
domain	O
I	O
and	O
domain	O
II	O
in	O
FGGY	O
family	O
carbohydrate	O
kinases	O
is	O
different	O
.	O
	O
In	O
addition	O
,	O
this	O
difference	O
may	O
be	O
caused	O
by	O
the	O
binding	O
of	O
substrates	O
and	O
/	O
or	O
ATP	O
.	O
	O
As	O
reported	O
previously	O
,	O
members	O
of	O
the	O
sugar	O
kinase	O
family	O
undergo	O
a	O
conformational	O
change	O
to	O
narrow	O
the	O
crossing	O
angle	O
between	O
two	O
domains	O
and	O
reduce	O
the	O
distance	O
between	O
substrate	O
and	O
ATP	O
in	O
order	O
to	O
facilitate	O
the	O
catalytic	O
reaction	O
of	O
phosphorylation	O
of	O
sugar	O
substrates	O
.	O
	O
After	O
comparing	O
structures	O
of	O
apo	O
-	O
SePSK	O
,	O
RBL	O
-	O
SePSK	O
and	O
AMP	O
-	O
PNP	O
-	O
SePSK	O
,	O
we	O
noticed	O
that	O
these	O
structures	O
presented	O
here	O
are	O
similar	O
.	O
	O
Superposing	O
the	O
structures	O
of	O
RBL	O
-	O
SePSK	O
and	O
AMP	O
-	O
PNP	O
-	O
SePSK	O
,	O
the	O
results	O
show	O
that	O
the	O
nearest	O
distance	O
between	O
AMP	O
-	O
PNP	O
γ	O
-	O
phosphate	O
and	O
RBL1	O
/	O
RBL2	O
is	O
7	O
.	O
5	O
Å	O
(	O
RBL1	O
-	O
O5	O
)/	O
6	O
.	O
7	O
Å	O
(	O
RBL2	O
-	O
O1	O
)	O
(	O
S8	O
Fig	O
).	O
	O
This	O
distance	O
is	O
too	O
long	O
to	O
transfer	O
the	O
γ	O
-	O
phosphate	O
group	O
from	O
ATP	O
to	O
the	O
substrate	O
.	O
	O
Since	O
the	O
two	O
domains	O
of	O
SePSK	O
are	O
widely	O
separated	O
in	O
this	O
structure	O
,	O
we	O
hypothesize	O
that	O
our	O
structures	O
of	O
SePSK	O
represent	O
its	O
open	O
form	O
,	O
and	O
that	O
a	O
conformational	O
rearrangement	O
must	O
occur	O
to	O
switch	O
to	O
the	O
closed	O
state	O
in	O
order	O
to	O
facilitate	O
the	O
catalytic	O
process	O
of	O
phosphorylation	O
of	O
sugar	O
substrates	O
.	O
	O
For	O
studying	O
such	O
potential	O
conformational	O
change	O
,	O
a	O
simulation	O
on	O
the	O
Hingeprot	O
Server	O
was	O
performed	O
to	O
predict	O
the	O
movement	O
of	O
different	O
SePSK	O
domains	O
.	O
	O
The	O
results	O
showed	O
that	O
domain	O
I	O
and	O
domain	O
II	O
are	O
closer	O
to	O
each	O
other	O
with	O
Ala228	O
and	O
Thr401	O
in	O
A2	O
as	O
Hinge	O
-	O
residues	O
.	O
	O
Based	O
on	O
the	O
above	O
results	O
,	O
SePSK	O
is	O
divided	O
into	O
two	O
rigid	O
parts	O
.	O
	O
The	O
domain	O
I	O
of	O
RBL	O
-	O
SePSK	O
(	O
aa	O
.	O
1	O
–	O
228	O
,	O
aa	O
.	O
402	O
–	O
421	O
)	O
and	O
the	O
domain	O
II	O
of	O
AMP	O
-	O
PNP	O
-	O
SePSK	O
(	O
aa	O
.	O
229	O
–	O
401	O
)	O
were	O
superposed	O
with	O
structures	O
,	O
including	O
apo	O
-	O
AtXK	O
-	O
1	O
,	O
apo	O
-	O
SePSK	O
,	O
xylulose	O
kinase	O
from	O
Lactobacillus	O
acidophilus	O
(	O
PDB	O
code	O
:	O
3LL3	O
)	O
and	O
the	O
S58W	B-mutant
mutant	O
form	O
of	O
glycerol	O
kinase	O
from	O
Escherichia	O
coli	O
(	O
PDB	O
code	O
:	O
1GLJ	O
).	O
	O
The	O
results	O
of	O
superposition	O
displayed	O
different	O
crossing	O
angle	O
between	O
these	O
two	O
domains	O
.	O
	O
After	O
superposition	O
,	O
the	O
distances	O
of	O
AMP	O
-	O
PNP	O
γ	O
-	O
phosphate	O
and	O
the	O
fifth	O
hydroxyl	O
group	O
of	O
RBL1	O
are	O
7	O
.	O
9	O
Å	O
(	O
superposed	O
with	O
AtXK	O
-	O
1	O
),	O
7	O
.	O
4	O
Å	O
(	O
superposed	O
with	O
SePSK	O
),	O
6	O
.	O
6	O
Å	O
(	O
superposed	O
with	O
3LL3	O
)	O
and	O
6	O
.	O
1	O
Å	O
(	O
superposed	O
with	O
1GLJ	O
).	O
	O
Meanwhile	O
,	O
the	O
distances	O
of	O
AMP	O
-	O
PNP	O
γ	O
-	O
phosphate	O
and	O
the	O
first	O
hydroxyl	O
group	O
of	O
RBL2	O
are	O
7	O
.	O
2	O
Å	O
(	O
superposed	O
with	O
AtXK	O
-	O
1	O
),	O
6	O
.	O
7	O
Å	O
(	O
superposed	O
with	O
SePSK	O
),	O
3	O
.	O
7	O
Å	O
(	O
superposed	O
with	O
3LL3	O
),	O
until	O
AMP	O
-	O
PNP	O
γ	O
-	O
phosphate	O
fully	O
contacts	O
RBL2	O
after	O
superposition	O
with	O
1GLJ	O
(	O
Fig	O
5	O
).	O
	O
Simulated	O
conformational	O
change	O
of	O
SePSK	O
during	O
the	O
catalytic	O
process	O
.	O
	O
The	O
structures	O
are	O
shown	O
as	O
cartoon	O
and	O
the	O
ligands	O
are	O
shown	O
as	O
sticks	O
.	O
	O
Domain	O
I	O
from	O
D	O
-	O
ribulose	O
-	O
SePSK	O
(	O
green	O
)	O
and	O
Domain	O
II	O
from	O
AMP	O
-	O
PNP	O
-	O
SePSK	O
(	O
cyan	O
)	O
are	O
superposed	O
with	O
apo	O
-	O
AtXK	O
-	O
1	O
(	O
1st	O
),	O
apo	O
-	O
SePSK	O
(	O
2nd	O
),	O
3LL3	O
(	O
3rd	O
)	O
and	O
1GLJ	O
(	O
4th	O
),	O
respectively	O
.	O
	O
The	O
numbers	O
near	O
the	O
black	O
dashed	O
lines	O
show	O
the	O
distances	O
(	O
Å	O
)	O
between	O
two	O
nearest	O
atoms	O
of	O
RBL	O
and	O
AMP	O
-	O
PNP	O
.	O
	O
Our	O
results	O
provide	O
the	O
detailed	O
information	O
about	O
the	O
interaction	O
of	O
SePSK	O
with	O
ATP	O
and	O
substrates	O
.	O
	O
Moreover	O
,	O
structural	O
superposition	O
results	O
enable	O
us	O
to	O
visualize	O
the	O
conformational	O
change	O
of	O
SePSK	O
during	O
the	O
catalytic	O
process	O
.	O
	O
In	O
conclusion	O
,	O
our	O
results	O
provide	O
important	O
information	O
for	O
a	O
more	O
detailed	O
understanding	O
of	O
the	O
mechanisms	O
of	O
SePSK	O
and	O
other	O
members	O
of	O
FGGY	O
family	O
carbohydrate	O
kinases	O
.	O
	O
Structural	O
insights	O
into	O
the	O
Escherichia	O
coli	O
lysine	O
decarboxylases	O
and	O
molecular	O
determinants	O
of	O
interaction	O
with	O
the	O
AAA	O
+	O
ATPase	O
RavA	O
	O
The	O
inducible	O
lysine	O
decarboxylase	O
LdcI	O
is	O
an	O
important	O
enterobacterial	O
acid	O
stress	O
response	O
enzyme	O
whereas	O
LdcC	O
is	O
its	O
close	O
paralogue	O
thought	O
to	O
play	O
mainly	O
a	O
metabolic	O
role	O
.	O
	O
Previously	O
,	O
we	O
proposed	O
a	O
pseudoatomic	O
model	O
of	O
the	O
LdcI	O
-	O
RavA	O
cage	O
based	O
on	O
its	O
cryo	O
-	O
electron	O
microscopy	O
map	O
and	O
crystal	O
structures	O
of	O
an	O
inactive	O
LdcI	O
decamer	O
and	O
a	O
RavA	O
monomer	O
.	O
	O
Comparison	O
with	O
each	O
other	O
and	O
with	O
available	O
structures	O
uncovers	O
differences	O
between	O
LdcI	O
and	O
LdcC	O
explaining	O
why	O
only	O
the	O
acid	O
stress	O
response	O
enzyme	O
is	O
capable	O
of	O
binding	O
RavA	O
.	O
We	O
identify	O
interdomain	O
movements	O
associated	O
with	O
the	O
pH	O
-	O
dependent	O
enzyme	O
activation	O
and	O
with	O
the	O
RavA	O
binding	O
.	O
	O
Enterobacterial	O
inducible	O
decarboxylases	O
of	O
basic	O
amino	O
acids	O
lysine	O
,	O
arginine	O
and	O
ornithine	O
have	O
a	O
common	O
evolutionary	O
origin	O
and	O
belong	O
to	O
the	O
α	O
-	O
family	O
of	O
pyridoxal	O
-	O
5	O
′-	O
phosphate	O
(	O
PLP	O
)-	O
dependent	O
enzymes	O
.	O
	O
Each	O
decarboxylase	O
is	O
induced	O
by	O
an	O
excess	O
of	O
the	O
target	O
amino	O
acid	O
and	O
a	O
specific	O
range	O
of	O
extracellular	O
pH	O
,	O
and	O
works	O
in	O
conjunction	O
with	O
a	O
cognate	O
inner	O
membrane	O
antiporter	O
.	O
	O
Decarboxylation	O
of	O
the	O
amino	O
acid	O
into	O
a	O
polyamine	O
is	O
catalysed	O
by	O
a	O
PLP	O
cofactor	O
in	O
a	O
multistep	O
reaction	O
that	O
consumes	O
a	O
cytoplasmic	O
proton	O
and	O
produces	O
a	O
CO2	O
molecule	O
passively	O
diffusing	O
out	O
of	O
the	O
cell	O
,	O
while	O
the	O
polyamine	O
is	O
excreted	O
by	O
the	O
antiporter	O
in	O
exchange	O
for	O
a	O
new	O
amino	O
acid	O
substrate	O
.	O
	O
Consequently	O
,	O
these	O
enzymes	O
buffer	O
both	O
the	O
bacterial	O
cytoplasm	O
and	O
the	O
local	O
extracellular	O
environment	O
.	O
	O
Inducible	O
enterobacterial	O
amino	O
acid	O
decarboxylases	O
have	O
been	O
intensively	O
studied	O
since	O
the	O
early	O
1940	O
because	O
the	O
ability	O
of	O
bacteria	O
to	O
withstand	O
acid	O
stress	O
can	O
be	O
linked	O
to	O
their	O
pathogenicity	O
in	O
humans	O
.	O
	O
Furthermore	O
,	O
both	O
LdcI	O
and	O
the	O
biosynthetic	O
lysine	O
decarboxylase	O
LdcC	O
of	O
uropathogenic	O
Escherichia	O
coli	O
(	O
UPEC	O
)	O
appear	O
to	O
play	O
an	O
important	O
role	O
in	O
increased	O
resistance	O
of	O
this	O
pathogen	O
to	O
nitrosative	O
stress	O
produced	O
by	O
nitric	O
oxide	O
and	O
other	O
damaging	O
reactive	O
nitrogen	O
intermediates	O
accumulating	O
during	O
the	O
course	O
of	O
urinary	O
tract	O
infections	O
(	O
UTI	O
).	O
	O
This	O
effect	O
is	O
attributed	O
to	O
cadaverine	O
,	O
the	O
diamine	O
produced	O
by	O
decarboxylation	O
of	O
lysine	O
by	O
LdcI	O
and	O
LdcC	O
,	O
that	O
was	O
shown	O
to	O
enhance	O
UPEC	O
colonisation	O
of	O
the	O
bladder	O
.	O
	O
In	O
addition	O
,	O
the	O
biosynthetic	O
E	O
.	O
coli	O
lysine	O
decarboxylase	O
LdcC	O
,	O
long	O
thought	O
to	O
be	O
constitutively	O
expressed	O
in	O
low	O
amounts	O
,	O
was	O
demonstrated	O
to	O
be	O
strongly	O
upregulated	O
by	O
fluoroquinolones	O
via	O
their	O
induction	O
of	O
RpoS	O
.	O
A	O
direct	O
correlation	O
between	O
the	O
level	O
of	O
cadaverine	O
and	O
the	O
resistance	O
of	O
E	O
.	O
coli	O
to	O
these	O
antibiotics	O
commonly	O
used	O
as	O
a	O
first	O
-	O
line	O
treatment	O
of	O
UTI	O
could	O
be	O
established	O
.	O
	O
Both	O
acid	O
pH	O
and	O
cadaverine	O
induce	O
closure	O
of	O
outer	O
membrane	O
porins	O
thereby	O
contributing	O
to	O
bacterial	O
protection	O
from	O
acid	O
stress	O
,	O
but	O
also	O
from	O
certain	O
antibiotics	O
,	O
by	O
reduction	O
in	O
membrane	O
permeability	O
.	O
	O
Each	O
monomer	O
is	O
composed	O
of	O
three	O
domains	O
–	O
an	O
N	O
-	O
terminal	O
wing	O
domain	O
(	O
residues	O
1	O
–	O
129	O
),	O
a	O
PLP	O
-	O
binding	O
core	O
domain	O
(	O
residues	O
130	O
–	O
563	O
),	O
and	O
a	O
C	O
-	O
terminal	O
domain	O
(	O
CTD	O
,	O
residues	O
564	O
–	O
715	O
).	O
	O
Ten	O
years	O
ago	O
we	O
showed	O
that	O
the	O
E	O
.	O
coli	O
AAA	O
+	O
ATPase	O
RavA	O
,	O
involved	O
in	O
multiple	O
stress	O
response	O
pathways	O
,	O
tightly	O
interacted	O
with	O
LdcI	O
but	O
was	O
not	O
capable	O
of	O
binding	O
to	O
LdcC	O
.	O
We	O
described	O
how	O
two	O
double	O
pentameric	O
rings	O
of	O
the	O
LdcI	O
tightly	O
associate	O
with	O
five	O
hexameric	O
rings	O
of	O
RavA	O
to	O
form	O
a	O
unique	O
cage	O
-	O
like	O
architecture	O
that	O
enables	O
the	O
bacterium	O
to	O
withstand	O
acid	O
stress	O
even	O
under	O
conditions	O
of	O
nutrient	O
deprivation	O
eliciting	O
stringent	O
response	O
.	O
	O
Furthermore	O
,	O
we	O
recently	O
solved	O
the	O
structure	O
of	O
the	O
E	O
.	O
coli	O
LdcI	O
-	O
RavA	O
complex	O
by	O
cryo	O
-	O
electron	O
microscopy	O
(	O
cryoEM	O
)	O
and	O
combined	O
it	O
with	O
the	O
crystal	O
structures	O
of	O
the	O
individual	O
proteins	O
.	O
	O
In	O
spite	O
of	O
this	O
wealth	O
of	O
structural	O
information	O
,	O
the	O
fact	O
that	O
LdcC	O
does	O
not	O
interact	O
with	O
RavA	O
,	O
although	O
the	O
two	O
lysine	O
decarboxylases	O
are	O
69	O
%	O
identical	O
and	O
84	O
%	O
similar	O
,	O
and	O
the	O
physiological	O
significance	O
of	O
the	O
absence	O
of	O
this	O
interaction	O
remained	O
unexplored	O
.	O
	O
To	O
solve	O
this	O
discrepancy	O
,	O
in	O
the	O
present	O
work	O
we	O
provided	O
a	O
three	O
-	O
dimensional	O
(	O
3D	O
)	O
cryoEM	O
reconstruction	O
of	O
LdcC	O
and	O
compared	O
it	O
with	O
the	O
available	O
LdcI	O
and	O
LdcI	O
-	O
RavA	O
structures	O
.	O
	O
Given	O
that	O
the	O
LdcI	O
crystal	O
structures	O
were	O
obtained	O
at	O
high	O
pH	O
where	O
the	O
enzyme	O
is	O
inactive	O
(	O
LdcIi	O
,	O
pH	O
8	O
.	O
5	O
),	O
whereas	O
the	O
cryoEM	O
reconstructions	O
of	O
LdcI	O
-	O
RavA	O
and	O
LdcI	O
-	O
LARA	O
were	O
done	O
at	O
acidic	O
pH	O
optimal	O
for	O
the	O
enzymatic	O
activity	O
,	O
for	O
a	O
meaningful	O
comparison	O
,	O
we	O
also	O
produced	O
a	O
3D	O
reconstruction	O
of	O
the	O
LdcI	O
at	O
active	O
pH	O
(	O
LdcIa	O
,	O
pH	O
6	O
.	O
2	O
).	O
	O
Remarkably	O
,	O
this	O
analysis	O
revealed	O
that	O
several	O
specific	O
residues	O
in	O
the	O
above	O
-	O
mentioned	O
β	O
-	O
sheet	O
,	O
independently	O
of	O
the	O
rest	O
of	O
the	O
protein	O
sequence	O
,	O
are	O
sufficient	O
to	O
define	O
if	O
a	O
particular	O
lysine	O
decarboxylase	O
should	O
be	O
classified	O
as	O
an	O
“	O
LdcC	O
-	O
like	O
”	O
or	O
an	O
“	O
LdcI	O
-	O
like	O
”.	O
	O
This	O
fascinating	O
parallelism	O
between	O
the	O
propensity	O
for	O
RavA	O
binding	O
and	O
the	O
genetic	O
environment	O
of	O
an	O
enterobacterial	O
lysine	O
decarboxylase	O
,	O
as	O
well	O
as	O
the	O
high	O
degree	O
of	O
conservation	O
of	O
this	O
small	O
structural	O
motif	O
,	O
emphasize	O
the	O
functional	O
importance	O
of	O
the	O
interaction	O
between	O
biodegradative	O
enterobacterial	O
lysine	O
decarboxylases	O
and	O
the	O
AAA	O
+	O
ATPase	O
RavA	O
.	O
	O
CryoEM	O
3D	O
reconstructions	O
of	O
LdcC	O
,	O
LdcIa	O
and	O
LdcI	O
-	O
LARA	O
	O
In	O
the	O
frame	O
of	O
this	O
work	O
,	O
we	O
produced	O
two	O
novel	O
subnanometer	O
resolution	O
cryoEM	O
reconstructions	O
of	O
the	O
E	O
.	O
coli	O
lysine	O
decarboxylases	O
at	O
pH	O
optimal	O
for	O
their	O
enzymatic	O
activity	O
–	O
a	O
5	O
.	O
5	O
Å	O
resolution	O
cryoEM	O
map	O
of	O
the	O
LdcC	O
(	O
pH	O
7	O
.	O
5	O
)	O
for	O
which	O
no	O
3D	O
structural	O
information	O
has	O
been	O
previously	O
available	O
(	O
Figs	O
1A	O
,	O
B	O
and	O
S1	O
),	O
and	O
a	O
6	O
.	O
1	O
Å	O
resolution	O
cryoEM	O
map	O
of	O
the	O
LdcIa	O
,	O
(	O
pH	O
6	O
.	O
2	O
)	O
(	O
Figs	O
1C	O
,	O
D	O
and	O
S2	O
).	O
	O
The	O
wing	O
domains	O
as	O
a	O
stable	O
anchor	O
at	O
the	O
center	O
of	O
the	O
double	O
-	O
ring	O
	O
As	O
a	O
first	O
step	O
of	O
a	O
comparative	O
analysis	O
,	O
we	O
superimposed	O
the	O
three	O
cryoEM	O
reconstructions	O
(	O
LdcIa	O
,	O
LdcI	O
-	O
LARA	O
and	O
LdcC	O
)	O
and	O
the	O
crystal	O
structure	O
of	O
the	O
LdcIi	O
decamer	O
(	O
Fig	O
.	O
2	O
and	O
Movie	O
S1	O
).	O
	O
This	O
superposition	O
reveals	O
that	O
the	O
densities	O
lining	O
the	O
central	O
hole	O
of	O
the	O
toroid	O
are	O
roughly	O
at	O
the	O
same	O
location	O
,	O
while	O
the	O
rest	O
of	O
the	O
structure	O
exhibits	O
noticeable	O
changes	O
.	O
	O
Specifically	O
,	O
at	O
the	O
center	O
of	O
the	O
double	O
-	O
ring	O
the	O
wing	O
domains	O
of	O
the	O
subunits	O
provide	O
the	O
conserved	O
basis	O
for	O
the	O
assembly	O
with	O
the	O
lowest	O
root	O
mean	O
square	O
deviation	O
(	O
RMSD	O
)	O
(	O
between	O
1	O
.	O
4	O
and	O
2	O
Å	O
for	O
the	O
Cα	O
atoms	O
only	O
),	O
whereas	O
the	O
peripheral	O
CTDs	O
containing	O
the	O
RavA	O
binding	O
interface	O
manifest	O
the	O
highest	O
RMSD	O
(	O
up	O
to	O
4	O
.	O
2	O
Å	O
)	O
(	O
Table	O
S2	O
).	O
	O
In	O
addition	O
,	O
the	O
wing	O
domains	O
of	O
all	O
structures	O
are	O
very	O
similar	O
,	O
with	O
the	O
RMSD	O
after	O
optimal	O
rigid	O
body	O
alignment	O
(	O
RMSDmin	O
)	O
less	O
than	O
1	O
.	O
1	O
Å	O
.	O
Thus	O
,	O
taking	O
the	O
limited	O
resolution	O
of	O
the	O
cryoEM	O
maps	O
into	O
account	O
,	O
we	O
consider	O
that	O
the	O
wing	O
domains	O
of	O
all	O
the	O
four	O
structures	O
are	O
essentially	O
identical	O
and	O
that	O
in	O
the	O
present	O
study	O
the	O
RMSD	O
of	O
less	O
than	O
2	O
Å	O
can	O
serve	O
as	O
a	O
baseline	O
below	O
which	O
differences	O
may	O
be	O
assumed	O
as	O
insignificant	O
.	O
	O
This	O
preservation	O
of	O
the	O
central	O
part	O
of	O
the	O
double	O
-	O
ring	O
assembly	O
may	O
help	O
the	O
enzymes	O
to	O
maintain	O
their	O
decameric	O
state	O
upon	O
activation	O
and	O
incorporation	O
into	O
the	O
LdcI	O
-	O
RavA	O
cage	O
.	O
	O
The	O
core	O
domain	O
and	O
the	O
active	O
site	O
rearrangements	O
upon	O
pH	O
-	O
dependent	O
enzyme	O
activation	O
and	O
LARA	O
binding	O
	O
The	O
decameric	O
enzyme	O
is	O
built	O
of	O
five	O
dimers	O
associating	O
into	O
a	O
5	O
-	O
fold	O
symmetrical	O
double	O
-	O
ring	O
(	O
two	O
monomers	O
making	O
a	O
dimer	O
are	O
delineated	O
in	O
Fig	O
.	O
1	O
).	O
	O
As	O
common	O
for	O
the	O
α	O
family	O
of	O
the	O
PLP	O
-	O
dependent	O
decarboxylases	O
,	O
dimerization	O
is	O
required	O
for	O
the	O
enzymatic	O
activity	O
because	O
the	O
active	O
site	O
is	O
buried	O
in	O
the	O
dimer	O
interface	O
(	O
Fig	O
.	O
3A	O
,	O
B	O
).	O
	O
This	O
interface	O
is	O
formed	O
essentially	O
by	O
the	O
core	O
domains	O
with	O
some	O
contribution	O
of	O
the	O
CTDs	O
.	O
	O
Zooming	O
in	O
the	O
variations	O
in	O
the	O
PLP	O
-	O
SD	O
shows	O
that	O
most	O
of	O
the	O
structural	O
changes	O
concern	O
displacements	O
in	O
the	O
active	O
site	O
(	O
Fig	O
.	O
3C	O
–	O
F	O
).	O
	O
The	O
most	O
conspicuous	O
differences	O
between	O
the	O
PLP	O
-	O
SDs	O
can	O
be	O
linked	O
to	O
the	O
pH	O
-	O
dependent	O
activation	O
of	O
the	O
enzymes	O
.	O
	O
Between	O
these	O
two	O
extremes	O
,	O
the	O
PLP	O
-	O
SDs	O
of	O
LdcIa	O
and	O
LdcC	O
are	O
similar	O
both	O
in	O
the	O
context	O
of	O
the	O
decamer	O
(	O
Fig	O
.	O
3F	O
)	O
and	O
in	O
terms	O
of	O
RMSDmin	O
=	O
0	O
.	O
9	O
Å	O
,	O
which	O
probably	O
reflects	O
the	O
fact	O
that	O
,	O
at	O
the	O
optimal	O
pH	O
,	O
these	O
lysine	O
decarboxylases	O
have	O
a	O
similar	O
enzymatic	O
activity	O
.	O
	O
In	O
addition	O
,	O
our	O
earlier	O
biochemical	O
observation	O
that	O
the	O
enzymatic	O
activity	O
of	O
LdcIa	O
is	O
unaffected	O
by	O
RavA	O
binding	O
is	O
consistent	O
with	O
the	O
relatively	O
small	O
changes	O
undergone	O
by	O
the	O
active	O
site	O
upon	O
transition	O
from	O
LdcIa	O
to	O
LdcI	O
-	O
LARA	O
.	O
	O
Worthy	O
of	O
note	O
,	O
our	O
previous	O
comparison	O
of	O
the	O
crystal	O
structure	O
of	O
LdcIi	O
with	O
that	O
of	O
the	O
inducible	O
arginine	O
decarboxylase	O
AdiA	O
revealed	O
high	O
conservation	O
of	O
the	O
PLP	O
-	O
coordinating	O
residues	O
and	O
identified	O
a	O
patch	O
of	O
negatively	O
charged	O
residues	O
lining	O
the	O
active	O
site	O
channel	O
as	O
a	O
potential	O
binding	O
site	O
for	O
the	O
target	O
amino	O
acid	O
substrate	O
(	O
Figs	O
S3	O
and	O
S4	O
in	O
ref	O
.).	O
	O
Rearrangements	O
of	O
the	O
ppGpp	O
binding	O
pocket	O
upon	O
pH	O
-	O
dependent	O
enzyme	O
activation	O
and	O
LARA	O
binding	O
	O
An	O
inhibitor	O
of	O
the	O
LdcI	O
and	O
LdcC	O
activity	O
,	O
the	O
stringent	O
response	O
alarmone	O
ppGpp	O
,	O
is	O
known	O
to	O
bind	O
at	O
the	O
interface	O
between	O
neighboring	O
monomers	O
within	O
each	O
ring	O
(	O
Fig	O
.	O
S4	O
).	O
	O
The	O
ppGpp	O
binding	O
pocket	O
is	O
made	O
up	O
by	O
residues	O
from	O
all	O
domains	O
and	O
is	O
located	O
approximately	O
30	O
Å	O
away	O
from	O
the	O
PLP	O
moiety	O
.	O
	O
Whereas	O
the	O
crystal	O
structure	O
of	O
the	O
ppGpp	O
-	O
LdcIi	O
was	O
solved	O
to	O
2	O
Å	O
resolution	O
,	O
only	O
a	O
4	O
.	O
1	O
Å	O
resolution	O
structure	O
of	O
the	O
ppGpp	O
-	O
free	O
LdcIi	O
could	O
be	O
obtained	O
.	O
	O
Thus	O
,	O
we	O
speculated	O
that	O
inhibition	O
of	O
LdcI	O
by	O
ppGpp	O
would	O
be	O
accompanied	O
by	O
a	O
transduction	O
of	O
subtle	O
structural	O
changes	O
at	O
the	O
level	O
of	O
individual	O
amino	O
acid	O
side	O
chains	O
between	O
the	O
ppGpp	O
binding	O
pocket	O
and	O
the	O
active	O
site	O
of	O
the	O
enzyme	O
.	O
	O
All	O
our	O
current	O
cryoEM	O
reconstructions	O
of	O
the	O
lysine	O
decarboxylases	O
were	O
obtained	O
in	O
the	O
absence	O
of	O
ppGpp	O
in	O
order	O
to	O
be	O
closer	O
to	O
the	O
active	O
state	O
of	O
the	O
enzymes	O
under	O
study	O
.	O
	O
The	O
fact	O
that	O
interaction	O
with	O
RavA	O
reduces	O
the	O
ppGpp	O
affinity	O
for	O
LdcI	O
despite	O
the	O
long	O
distance	O
of	O
~	O
30	O
Å	O
between	O
the	O
LARA	O
domain	O
binding	O
site	O
and	O
the	O
closest	O
ppGpp	O
binding	O
pocket	O
(	O
Fig	O
.	O
S5	O
)	O
seems	O
to	O
favor	O
an	O
allosteric	O
regulation	O
mechanism	O
.	O
	O
Interestingly	O
,	O
although	O
a	O
number	O
of	O
ppGpp	O
binding	O
residues	O
are	O
strictly	O
conserved	O
between	O
LdcI	O
and	O
AdiA	O
that	O
also	O
forms	O
decamers	O
at	O
low	O
pH	O
optimal	O
for	O
its	O
arginine	O
decarboxylase	O
activity	O
,	O
no	O
ppGpp	O
regulation	O
of	O
AdiA	O
could	O
be	O
demonstrated	O
.	O
	O
Indeed	O
,	O
all	O
CTDs	O
have	O
very	O
similar	O
structures	O
(	O
RMSDmin	O
<	O
1	O
Å	O
).	O
	O
The	O
LdcIi	O
monomer	O
is	O
the	O
most	O
compact	O
,	O
whereas	O
LdcIa	O
and	O
especially	O
LdcI	O
-	O
LARA	O
gradually	O
extend	O
their	O
CTDs	O
towards	O
the	O
LARA	O
domain	O
of	O
RavA	O
(	O
Figs	O
2	O
and	O
4	O
).	O
	O
These	O
small	O
but	O
noticeable	O
swinging	O
and	O
stretching	O
(	O
up	O
to	O
~	O
4	O
Å	O
)	O
may	O
be	O
related	O
to	O
the	O
incorporation	O
of	O
the	O
LdcI	O
decamer	O
into	O
the	O
LdcI	O
-	O
RavA	O
cage	O
.	O
	O
The	O
C	O
-	O
terminal	O
β	O
-	O
sheet	O
of	O
a	O
lysine	O
decarboxylase	O
as	O
a	O
major	O
determinant	O
of	O
the	O
interaction	O
with	O
RavA	O
	O
In	O
our	O
previous	O
contribution	O
,	O
based	O
on	O
the	O
fit	O
of	O
the	O
LdcIi	O
and	O
the	O
LARA	O
crystal	O
structures	O
into	O
the	O
LdcI	O
-	O
LARA	O
cryoEM	O
density	O
,	O
we	O
predicted	O
that	O
the	O
LdcI	O
-	O
RavA	O
interaction	O
should	O
involve	O
the	O
C	O
-	O
terminal	O
two	O
-	O
stranded	O
β	O
-	O
sheet	O
of	O
the	O
LdcI	O
.	O
Our	O
present	O
cryoEM	O
maps	O
and	O
pseudoatomic	O
models	O
provide	O
first	O
structure	O
-	O
based	O
insights	O
into	O
the	O
differences	O
between	O
the	O
inducible	O
and	O
the	O
constitutive	O
lysine	O
decarboxylases	O
.	O
	O
Therefore	O
,	O
we	O
wanted	O
to	O
check	O
the	O
influence	O
of	O
the	O
primary	O
sequence	O
of	O
the	O
two	O
proteins	O
in	O
this	O
region	O
on	O
their	O
ability	O
to	O
interact	O
with	O
RavA	O
.	O
To	O
this	O
end	O
,	O
we	O
swapped	O
the	O
relevant	O
β	O
-	O
sheets	O
of	O
the	O
two	O
proteins	O
and	O
produced	O
their	O
chimeras	B-mutant
,	O
namely	O
LdcIC	B-mutant
(	O
i	O
.	O
e	O
.	O
LdcI	O
with	O
the	O
C	O
-	O
terminal	O
β	O
-	O
sheet	O
of	O
LdcC	O
)	O
and	O
LdcCI	B-mutant
(	O
i	O
.	O
e	O
.	O
LdcC	O
with	O
the	O
C	O
-	O
terminal	O
β	O
-	O
sheet	O
of	O
LdcI	O
)	O
(	O
Fig	O
.	O
5A	O
–	O
C	O
).	O
	O
Both	B-mutant
constructs	I-mutant
could	O
be	O
purified	O
and	O
could	O
form	O
decamers	O
visually	O
indistinguishable	O
from	O
the	O
wild	O
-	O
type	O
proteins	O
.	O
	O
As	O
expected	O
,	O
binding	O
of	O
LdcI	O
to	O
RavA	O
was	O
completely	O
abolished	O
by	O
this	O
procedure	O
and	O
no	O
LdcIC	O
-	O
RavA	O
complex	O
could	O
be	O
detected	O
.	O
	O
On	O
the	O
negative	O
stain	O
EM	O
grid	O
,	O
the	O
chimeric	O
cages	O
appeared	O
less	O
rigid	O
than	O
the	O
native	O
LdcI	O
-	O
RavA	O
,	O
which	O
probably	O
means	O
that	O
the	O
environment	O
of	O
the	O
β	O
-	O
sheet	O
contributes	O
to	O
the	O
efficiency	O
of	O
the	O
interaction	O
and	O
the	O
stability	O
of	O
the	O
entire	O
architecture	O
(	O
Fig	O
.	O
5D	O
–	O
F	O
).	O
	O
The	O
C	O
-	O
terminal	O
β	O
-	O
sheet	O
of	O
a	O
lysine	O
decarboxylase	O
is	O
a	O
highly	O
conserved	O
signature	O
allowing	O
to	O
distinguish	O
between	O
LdcI	O
and	O
LdcC	O
	O
Alignment	O
of	O
the	O
primary	O
sequences	O
of	O
the	O
E	O
.	O
coli	O
LdcI	O
and	O
LdcC	O
shows	O
that	O
some	O
amino	O
acid	O
residues	O
of	O
the	O
C	O
-	O
terminal	O
β	O
-	O
sheet	O
are	O
the	O
same	O
in	O
the	O
two	O
proteins	O
,	O
whereas	O
others	O
are	O
notably	O
different	O
in	O
chemical	O
nature	O
.	O
	O
Importantly	O
,	O
most	O
of	O
the	O
amino	O
acid	O
differences	O
between	O
the	O
two	O
enzymes	O
are	O
located	O
in	O
this	O
very	O
region	O
.	O
	O
Thus	O
,	O
to	O
advance	O
beyond	O
our	O
experimental	O
confirmation	O
of	O
the	O
C	O
-	O
terminal	O
β	O
-	O
sheet	O
as	O
a	O
major	O
determinant	O
of	O
the	O
capacity	O
of	O
a	O
particular	O
lysine	O
decarboxylase	O
to	O
form	O
a	O
cage	O
with	O
RavA	O
,	O
we	O
set	O
out	O
to	O
investigate	O
whether	O
certain	O
residues	O
in	O
this	O
β	O
-	O
sheet	O
are	O
conserved	O
in	O
lysine	O
decarboxylases	O
of	O
different	O
enterobacteria	O
that	O
have	O
the	O
ravA	O
-	O
viaA	O
operon	O
in	O
their	O
genome	O
.	O
	O
We	O
inspected	O
the	O
genetic	O
environment	O
of	O
lysine	O
decarboxylases	O
from	O
22	O
enterobacterial	O
species	O
referenced	O
in	O
the	O
NCBI	O
database	O
,	O
corrected	O
the	O
gene	O
annotation	O
where	O
necessary	O
(	O
Tables	O
S3	O
and	O
S4	O
),	O
and	O
performed	O
multiple	O
sequence	O
alignment	O
coupled	O
to	O
a	O
phylogenetic	O
analysis	O
(	O
see	O
Methods	O
).	O
	O
First	O
of	O
all	O
,	O
consensus	O
sequence	O
for	O
the	O
entire	O
lysine	O
decarboxylase	O
family	O
was	O
derived	O
.	O
	O
Second	O
,	O
the	O
phylogenetic	O
analysis	O
clearly	O
split	O
the	O
lysine	O
decarboxylases	O
into	O
two	O
groups	O
(	O
Fig	O
.	O
6A	O
).	O
	O
All	O
lysine	O
decarboxylases	O
predicted	O
to	O
be	O
“	O
LdcI	O
-	O
like	O
”	O
or	O
biodegradable	O
based	O
on	O
their	O
genetic	O
environment	O
,	O
as	O
for	O
example	O
their	O
organization	O
in	O
an	O
operon	O
with	O
a	O
gene	O
encoding	O
the	O
CadB	O
antiporter	O
(	O
see	O
Methods	O
),	O
were	O
found	O
in	O
one	O
group	O
,	O
whereas	O
all	O
enzymes	O
predicted	O
as	O
“	O
LdcC	O
-	O
like	O
”	O
or	O
biosynthetic	O
partitioned	O
into	O
another	O
group	O
.	O
	O
Thus	O
,	O
consensus	O
sequences	O
could	O
also	O
be	O
determined	O
for	O
each	O
of	O
the	O
two	O
groups	O
(	O
Figs	O
6B	O
,	O
C	O
and	O
S7	O
).	O
	O
Inspection	O
of	O
these	O
consensus	O
sequences	O
revealed	O
important	O
differences	O
between	O
the	O
groups	O
regarding	O
charge	O
,	O
size	O
and	O
hydrophobicity	O
of	O
several	O
residues	O
precisely	O
at	O
the	O
level	O
of	O
the	O
C	O
-	O
terminal	O
β	O
-	O
sheet	O
that	O
is	O
responsible	O
for	O
the	O
interaction	O
with	O
RavA	O
(	O
Fig	O
.	O
6B	O
–	O
D	O
).	O
	O
For	O
example	O
,	O
in	O
our	O
previous	O
study	O
,	O
site	O
-	O
directed	O
mutations	O
identified	O
Y697	O
as	O
critically	O
required	O
for	O
the	O
RavA	O
binding	O
.	O
	O
Our	O
current	O
analysis	O
shows	O
that	O
Y697	O
is	O
strictly	O
conserved	O
in	O
the	O
“	O
LdcI	O
-	O
like	O
”	O
group	O
whereas	O
the	O
“	O
LdcC	O
-	O
like	O
”	O
enzymes	O
always	O
have	O
a	O
lysine	O
in	O
this	O
position	O
;	O
it	O
also	O
uncovers	O
several	O
other	O
residues	O
potentially	O
essential	O
for	O
the	O
interaction	O
with	O
RavA	O
which	O
can	O
now	O
be	O
addressed	O
by	O
site	O
-	O
directed	O
mutagenesis	O
.	O
	O
The	O
third	O
and	O
most	O
remarkable	O
finding	O
was	O
that	O
exactly	O
the	O
same	O
separation	O
into	O
“	O
LdcI	O
-	O
like	O
”	O
and	O
“	O
LdcC	O
”-	O
like	O
groups	O
can	O
be	O
obtained	O
based	O
on	O
a	O
comparison	O
of	O
the	O
C	O
-	O
terminal	O
β	O
-	O
sheets	O
only	O
,	O
without	O
taking	O
the	O
rest	O
of	O
the	O
primary	O
sequence	O
into	O
account	O
.	O
	O
Therefore	O
the	O
C	O
-	O
terminal	O
β	O
-	O
sheet	O
emerges	O
as	O
being	O
a	O
highly	O
conserved	O
signature	O
sequence	O
,	O
sufficient	O
to	O
unambiguously	O
discriminate	O
between	O
the	O
“	O
LdcI	O
-	O
like	O
”	O
and	O
“	O
LdcC	O
-	O
like	O
”	O
enterobacterial	O
lysine	O
decarboxylases	O
independently	O
of	O
any	O
other	O
information	O
(	O
Figs	O
6	O
and	O
S7	O
).	O
	O
Thus	O
,	O
enterobacteria	O
identified	O
here	O
(	O
Fig	O
.	O
6	O
,	O
Table	O
S4	O
)	O
appear	O
to	O
exert	O
evolutionary	O
pressure	O
on	O
the	O
biodegradative	O
lysine	O
decarboxylase	O
towards	O
the	O
RavA	O
binding	O
.	O
	O
The	O
conformational	O
rearrangements	O
of	O
LdcI	O
upon	O
enzyme	O
activation	O
and	O
RavA	O
binding	O
revealed	O
in	O
this	O
work	O
,	O
and	O
our	O
amazing	O
finding	O
that	O
the	O
molecular	O
determinant	O
of	O
the	O
LdcI	O
-	O
RavA	O
interaction	O
is	O
the	O
one	O
that	O
straightforwardly	O
determines	O
if	O
a	O
particular	O
enterobacterial	O
lysine	O
decarboxylase	O
belongs	O
to	O
“	O
LdcI	O
-	O
like	O
”	O
or	O
“	O
LdcC	O
-	O
like	O
”	O
proteins	O
,	O
should	O
give	O
a	O
new	O
impetus	O
to	O
functional	O
studies	O
of	O
the	O
unique	O
LdcI	O
-	O
RavA	O
cage	O
.	O
	O
Together	O
with	O
the	O
apo	O
-	O
LdcI	O
and	O
ppGpp	O
-	O
LdcIi	O
crystal	O
structures	O
,	O
our	O
cryoEM	O
reconstructions	O
provide	O
a	O
structural	O
framework	O
for	O
future	O
studies	O
of	O
structure	O
-	O
function	O
relationships	O
of	O
lysine	O
decarboxylases	O
from	O
other	O
enterobacteria	O
and	O
even	O
of	O
their	O
homologues	O
outside	O
Enterobacteriaceae	O
.	O
For	O
example	O
,	O
the	O
lysine	O
decarboxylase	O
of	O
Eikenella	O
corrodens	O
is	O
thought	O
to	O
play	O
a	O
major	O
role	O
in	O
the	O
periodontal	O
disease	O
and	O
its	O
inhibitors	O
were	O
shown	O
to	O
retard	O
gingivitis	O
development	O
.	O
	O
Finally	O
,	O
cadaverine	O
being	O
an	O
important	O
platform	O
chemical	O
for	O
the	O
production	O
of	O
industrial	O
polymers	O
such	O
as	O
nylon	O
,	O
structural	O
information	O
is	O
valuable	O
for	O
optimisation	O
of	O
bacterial	O
lysine	O
decarboxylases	O
used	O
for	O
its	O
production	O
in	O
biotechnology	O
.	O
	O
3D	O
cryoEM	O
reconstructions	O
of	O
LdcC	O
,	O
LdcI	O
-	O
LARA	O
and	O
LdcIa	O
.	O
	O
In	O
the	O
rest	O
of	O
the	O
protomers	O
,	O
the	O
wing	O
,	O
core	O
and	O
C	O
-	O
terminal	O
domains	O
are	O
colored	O
from	O
light	O
to	O
dark	O
in	O
shades	O
of	O
green	O
for	O
LdcC	O
(	O
A	O
),	O
pink	O
for	O
LdcIa	O
(	O
C	O
)	O
and	O
blue	O
for	O
LdcI	O
in	O
LdcI	O
-	O
LARA	O
(	O
E	O
).	O
	O
In	O
(	O
E	O
),	O
the	O
LARA	O
domain	O
density	O
is	O
shown	O
in	O
dark	O
grey	O
.	O
	O
Two	O
monomers	O
making	O
a	O
dimer	O
are	O
delineated	O
.	O
	O
Scale	O
bar	O
50	O
Å	O
.	O
(	O
B	O
,	O
D	O
,	O
F	O
)	O
One	O
protomer	O
from	O
the	O
cryoEM	O
map	O
of	O
the	O
LdcC	O
(	O
B	O
),	O
LdcIa	O
(	O
D	O
)	O
and	O
LdcI	O
-	O
LARA	O
(	O
F	O
)	O
in	O
light	O
grey	O
with	O
the	O
pseudoatomic	O
model	O
represented	O
as	O
cartoons	O
and	O
colored	O
as	O
the	O
densities	O
in	O
(	O
A	O
,	O
C	O
,	O
E	O
).	O
	O
Superposition	O
of	O
the	O
pseudoatomic	O
models	O
of	O
LdcC	O
,	O
LdcI	O
from	O
LdcI	O
-	O
LARA	O
and	O
LdcIa	O
colored	O
as	O
in	O
Fig	O
.	O
1	O
,	O
and	O
the	O
crystal	O
structure	O
of	O
LdcIi	O
in	O
shades	O
of	O
yellow	O
.	O
	O
The	O
dashed	O
circle	O
indicates	O
the	O
central	O
region	O
that	O
remains	O
virtually	O
unchanged	O
between	O
all	O
the	O
structures	O
,	O
while	O
the	O
periphery	O
undergoes	O
visible	O
movements	O
.	O
	O
(	O
A	O
)	O
LdcIi	O
crystal	O
structure	O
,	O
with	O
one	O
ring	O
represented	O
as	O
a	O
grey	O
surface	O
and	O
the	O
second	O
as	O
a	O
cartoon	O
.	O
	O
A	O
monomer	O
with	O
its	O
PLP	O
cofactor	O
is	O
delineated	O
.	O
	O
The	O
PLP	O
moieties	O
of	O
the	O
cartoon	O
ring	O
are	O
shown	O
in	O
red	O
.	O
	O
(	O
B	O
)	O
The	O
LdcIi	O
dimer	O
extracted	O
from	O
the	O
crystal	O
structure	O
of	O
the	O
decamer	O
.	O
	O
One	O
monomer	O
is	O
colored	O
in	O
shades	O
of	O
yellow	O
as	O
in	O
Figs	O
1	O
and	O
2	O
,	O
while	O
the	O
monomer	O
related	O
by	O
C2	O
symmetry	O
is	O
grey	O
.	O
	O
The	O
PLP	O
is	O
red	O
.	O
	O
Stretching	O
of	O
the	O
LdcI	O
monomer	O
upon	O
pH	O
-	O
dependent	O
enzyme	O
activation	O
and	O
LARA	O
binding	O
.	O
	O
(	O
A	O
–	O
C	O
)	O
A	O
slice	O
through	O
the	O
pseudoatomic	O
models	O
of	O
the	O
LdcI	O
monomers	O
extracted	O
from	O
the	O
superimposed	O
decamers	O
(	O
Fig	O
.	O
2	O
)	O
The	O
rectangle	O
indicates	O
the	O
regions	O
enlarged	O
in	O
(	O
D	O
–	O
F	O
).	O
	O
(	O
A	O
)	O
compares	O
LdcIi	O
(	O
yellow	O
)	O
and	O
LdcIa	O
(	O
pink	O
),	O
(	O
B	O
)	O
compares	O
LdcIa	O
(	O
pink	O
)	O
and	O
LdcI	O
-	O
LARA	O
(	O
blue	O
),	O
and	O
(	O
C	O
)	O
compares	O
LdcIi	O
(	O
yellow	O
),	O
LdcIa	O
(	O
pink	O
)	O
and	O
LdcI	O
-	O
LARA	O
(	O
blue	O
)	O
simultaneously	O
in	O
order	O
to	O
show	O
the	O
progressive	O
stretching	O
described	O
in	O
the	O
text	O
.	O
	O
The	O
cryoEM	O
density	O
of	O
the	O
LARA	O
domain	O
is	O
represented	O
as	O
a	O
grey	O
surface	O
to	O
show	O
the	O
position	O
of	O
the	O
binding	O
site	O
and	O
the	O
direction	O
of	O
the	O
movement	O
.	O
	O
(	O
D	O
–	O
F	O
)	O
Inserts	O
zooming	O
at	O
the	O
CTD	O
part	O
in	O
proximity	O
of	O
the	O
LARA	O
binding	O
site	O
.	O
	O
Analysis	O
of	O
the	O
LdcIC	B-mutant
and	O
LdcCI	B-mutant
chimeras	B-mutant
.	O
	O
Sequence	O
analysis	O
of	O
enterobacterial	O
lysine	O
decarboxylases	O
.	O
	O
(	O
B	O
)	O
Analysis	O
of	O
consensus	O
“	O
LdcI	O
-	O
like	O
”	O
and	O
“	O
LdcC	O
-	O
like	O
”	O
sequences	O
around	O
the	O
first	O
and	O
second	O
C	O
-	O
terminal	O
β	O
-	O
strands	O
.	O
	O
(	O
C	O
)	O
Signature	O
sequences	O
of	O
LdcI	O
and	O
LdcC	O
in	O
the	O
C	O
-	O
terminal	O
β	O
-	O
sheet	O
.	O
	O
Polarity	O
differences	O
are	O
highlighted	O
.	O
(	O
D	O
)	O
Position	O
and	O
nature	O
of	O
these	O
differences	O
at	O
the	O
surface	O
of	O
the	O
respective	O
cryoEM	O
maps	O
with	O
the	O
color	O
code	O
as	O
in	O
B	O
.	O
See	O
also	O
Fig	O
.	O
S7	O
and	O
Tables	O
S3	O
and	O
S4	O
.	O
	O
Mep2	O
activity	O
is	O
tightly	O
regulated	O
by	O
phosphorylation	O
,	O
but	O
how	O
this	O
is	O
achieved	O
at	O
the	O
molecular	O
level	O
is	O
not	O
clear	O
.	O
	O
Here	O
we	O
report	O
X	O
-	O
ray	O
crystal	O
structures	O
of	O
the	O
Mep2	O
orthologues	O
from	O
Saccharomyces	O
cerevisiae	O
and	O
Candida	O
albicans	O
and	O
show	O
that	O
under	O
nitrogen	O
-	O
sufficient	O
conditions	O
the	O
transporters	O
are	O
not	O
phosphorylated	O
and	O
present	O
in	O
closed	O
,	O
inactive	O
conformations	O
.	O
	O
The	O
phosphorylation	O
site	O
in	O
the	O
CTR	O
is	O
solvent	O
accessible	O
and	O
located	O
in	O
a	O
negatively	O
charged	O
pocket	O
∼	O
30	O
Å	O
away	O
from	O
the	O
channel	O
exit	O
.	O
	O
The	O
crystal	O
structure	O
of	O
phosphorylation	O
-	O
mimicking	O
Mep2	B-mutant
variants	I-mutant
from	O
C	O
.	O
albicans	O
show	O
large	O
conformational	O
changes	O
in	O
a	O
conserved	O
and	O
functionally	O
important	O
region	O
of	O
the	O
CTR	O
.	O
	O
The	O
results	O
allow	O
us	O
to	O
propose	O
a	O
model	O
for	O
regulation	O
of	O
eukaryotic	O
ammonium	O
transport	O
by	O
phosphorylation	O
.	O
	O
Mep2	O
proteins	O
are	O
tightly	O
regulated	O
fungal	O
ammonium	O
transporters	O
.	O
	O
Transceptors	O
are	O
membrane	O
proteins	O
that	O
function	O
not	O
only	O
as	O
transporters	O
but	O
also	O
as	O
receptors	O
/	O
sensors	O
during	O
nutrient	O
sensing	O
to	O
activate	O
downstream	O
signalling	O
pathways	O
.	O
	O
One	O
of	O
the	O
most	O
important	O
unresolved	O
questions	O
in	O
the	O
field	O
is	O
how	O
the	O
transceptors	O
couple	O
to	O
downstream	O
signalling	O
pathways	O
.	O
	O
One	O
hypothesis	O
is	O
that	O
downstream	O
signalling	O
is	O
dependent	O
on	O
a	O
specific	O
conformation	O
of	O
the	O
transporter	O
.	O
	O
Mep2	O
(	O
methylammonium	O
(	O
MA	O
)	O
permease	O
)	O
proteins	O
are	O
ammonium	O
transceptors	O
that	O
are	O
ubiquitous	O
in	O
fungi	O
.	O
	O
They	O
belong	O
to	O
the	O
Amt	O
/	O
Mep	O
/	O
Rh	O
family	O
of	O
transporters	O
that	O
are	O
present	O
in	O
all	O
kingdoms	O
of	O
life	O
and	O
they	O
take	O
up	O
ammonium	O
from	O
the	O
extracellular	O
environment	O
.	O
	O
Fungi	O
typically	O
have	O
more	O
than	O
one	O
Mep	O
paralogue	O
,	O
for	O
example	O
,	O
Mep1	O
-	O
3	O
in	O
S	O
.	O
cerevisiae	O
.	O
	O
Under	O
conditions	O
of	O
nitrogen	O
limitation	O
,	O
Mep2	O
initiates	O
a	O
signalling	O
cascade	O
that	O
results	O
in	O
a	O
switch	O
from	O
the	O
yeast	O
form	O
to	O
filamentous	O
(	O
pseudohyphal	O
)	O
growth	O
that	O
may	O
be	O
required	O
for	O
fungal	O
pathogenicity	O
.	O
	O
As	O
is	O
the	O
case	O
for	O
other	O
transceptors	O
,	O
it	O
is	O
not	O
clear	O
how	O
Mep2	O
interacts	O
with	O
downstream	O
signalling	O
partners	O
,	O
but	O
the	O
protein	O
kinase	O
A	O
and	O
mitogen	O
-	O
activated	O
protein	O
kinase	O
pathways	O
have	O
been	O
proposed	O
as	O
downstream	O
effectors	O
of	O
Mep2	O
(	O
refs	O
).	O
	O
Compared	O
with	O
Mep1	O
and	O
Mep3	O
,	O
Mep2	O
is	O
highly	O
expressed	O
and	O
functions	O
as	O
a	O
low	O
-	O
capacity	O
,	O
high	O
-	O
affinity	O
transporter	O
in	O
the	O
uptake	O
of	O
MA	O
.	O
	O
In	O
addition	O
,	O
Mep2	O
is	O
also	O
important	O
for	O
uptake	O
of	O
ammonium	O
produced	O
by	O
growth	O
on	O
other	O
nitrogen	O
sources	O
.	O
	O
By	O
contrast	O
,	O
several	O
bacterial	O
Amt	O
orthologues	O
have	O
been	O
characterized	O
in	O
detail	O
via	O
high	O
-	O
resolution	O
crystal	O
structures	O
and	O
a	O
number	O
of	O
molecular	O
dynamics	O
(	O
MD	O
)	O
studies	O
.	O
	O
All	O
the	O
solved	O
structures	O
including	O
that	O
of	O
RhCG	O
are	O
very	O
similar	O
,	O
establishing	O
the	O
basic	O
architecture	O
of	O
ammonium	O
transporters	O
.	O
	O
The	O
proteins	O
form	O
stable	O
trimers	O
,	O
with	O
each	O
monomer	O
having	O
11	O
transmembrane	O
(	O
TM	O
)	O
helices	O
and	O
a	O
central	O
channel	O
for	O
the	O
transport	O
of	O
ammonium	O
.	O
	O
All	O
structures	O
show	O
the	O
transporters	O
in	O
open	O
conformations	O
.	O
	O
Where	O
earlier	O
studies	O
favoured	O
the	O
transport	O
of	O
ammonia	O
gas	O
,	O
recent	O
data	O
and	O
theoretical	O
considerations	O
suggest	O
that	O
Amt	O
/	O
Mep	O
proteins	O
are	O
instead	O
active	O
,	O
electrogenic	O
transporters	O
of	O
either	O
NH4	O
+	O
(	O
uniport	O
)	O
or	O
NH3	O
/	O
H	O
+	O
(	O
symport	O
).	O
	O
A	O
highly	O
conserved	O
pair	O
of	O
channel	O
-	O
lining	O
histidine	O
residues	O
dubbed	O
the	O
twin	O
-	O
His	O
motif	O
may	O
serve	O
as	O
a	O
proton	O
relay	O
system	O
while	O
NH3	O
moves	O
through	O
the	O
channel	O
during	O
NH3	O
/	O
H	O
+	O
symport	O
.	O
	O
Ammonium	O
transport	O
is	O
tightly	O
regulated	O
.	O
	O
In	O
animals	O
,	O
this	O
is	O
due	O
to	O
toxicity	O
of	O
elevated	O
intracellular	O
ammonium	O
levels	O
,	O
whereas	O
for	O
microorganisms	O
ammonium	O
is	O
a	O
preferred	O
nitrogen	O
source	O
.	O
	O
By	O
binding	O
tightly	O
to	O
Amt	O
proteins	O
without	O
inducing	O
a	O
conformational	O
change	O
in	O
the	O
transporter	O
,	O
GlnK	O
sterically	O
blocks	O
ammonium	O
conductance	O
when	O
nitrogen	O
levels	O
are	O
sufficient	O
.	O
	O
Importantly	O
,	O
eukaryotes	O
do	O
not	O
have	O
GlnK	O
orthologues	O
and	O
have	O
a	O
different	O
mechanism	O
for	O
regulation	O
of	O
ammonium	O
transport	O
activity	O
.	O
	O
In	O
plants	O
,	O
transporter	O
phosphorylation	O
and	O
dephosphorylation	O
are	O
known	O
to	O
regulate	O
activity	O
.	O
	O
In	O
S	O
.	O
cerevisiae	O
,	O
phosphorylation	O
of	O
Ser457	O
within	O
the	O
C	O
-	O
terminal	O
region	O
(	O
CTR	O
)	O
in	O
the	O
cytoplasm	O
was	O
recently	O
proposed	O
to	O
cause	O
Mep2	O
opening	O
,	O
possibly	O
via	O
inducing	O
a	O
conformational	O
change	O
.	O
	O
The	O
structures	O
are	O
similar	O
to	O
each	O
other	O
but	O
show	O
considerable	O
differences	O
to	O
all	O
other	O
ammonium	O
transporter	O
structures	O
.	O
	O
The	O
most	O
striking	O
difference	O
is	O
the	O
fact	O
that	O
the	O
Mep2	O
proteins	O
have	O
closed	O
conformations	O
.	O
	O
The	O
putative	O
phosphorylation	O
site	O
is	O
solvent	O
accessible	O
and	O
located	O
in	O
a	O
negatively	O
charged	O
pocket	O
∼	O
30	O
Å	O
away	O
from	O
the	O
channel	O
exit	O
.	O
	O
The	O
channels	O
of	O
phosphorylation	O
-	O
mimicking	O
mutants	O
of	O
C	O
.	O
albicans	O
Mep2	O
are	O
still	O
closed	O
but	O
show	O
large	O
conformational	O
changes	O
within	O
a	O
conserved	O
part	O
of	O
the	O
CTR	O
.	O
	O
Together	O
with	O
a	O
structure	O
of	O
a	O
C	O
-	O
terminal	O
Mep2	B-mutant
variant	I-mutant
lacking	O
the	O
segment	O
containing	O
the	O
phosphorylation	O
site	O
,	O
the	O
results	O
allow	O
us	O
to	O
propose	O
a	O
structural	O
model	O
for	O
phosphorylation	O
-	O
based	O
regulation	O
of	O
eukaryotic	O
ammonium	O
transport	O
.	O
	O
General	O
architecture	O
of	O
Mep2	O
ammonium	O
transceptors	O
	O
The	O
Mep2	O
protein	O
of	O
S	O
.	O
cerevisiae	O
(	O
ScMep2	O
)	O
was	O
overexpressed	O
in	O
S	O
.	O
cerevisiae	O
in	O
high	O
yields	O
,	O
enabling	O
structure	O
determination	O
by	O
X	O
-	O
ray	O
crystallography	O
using	O
data	O
to	O
3	O
.	O
2	O
Å	O
resolution	O
by	O
molecular	O
replacement	O
(	O
MR	O
)	O
with	O
the	O
archaebacterial	O
Amt	O
-	O
1	O
structure	O
(	O
see	O
Methods	O
section	O
).	O
	O
Given	O
that	O
the	O
modest	O
resolution	O
of	O
the	O
structure	O
and	O
the	O
limited	O
detergent	O
stability	O
of	O
ScMep2	O
would	O
likely	O
complicate	O
structure	O
–	O
function	O
studies	O
,	O
several	O
other	O
fungal	O
Mep2	O
orthologues	O
were	O
subsequently	O
overexpressed	O
and	O
screened	O
for	O
diffraction	O
-	O
quality	O
crystals	O
.	O
	O
Of	O
these	O
,	O
Mep2	O
from	O
C	O
.	O
albicans	O
(	O
CaMep2	O
)	O
showed	O
superior	O
stability	O
in	O
relatively	O
harsh	O
detergents	O
such	O
as	O
nonyl	O
-	O
glucoside	O
,	O
allowing	O
structure	O
determination	O
in	O
two	O
different	O
crystal	O
forms	O
to	O
high	O
resolution	O
(	O
up	O
to	O
1	O
.	O
5	O
Å	O
).	O
	O
Electron	O
density	O
is	O
visible	O
for	O
the	O
entire	O
polypeptide	O
chains	O
,	O
with	O
the	O
exception	O
of	O
the	O
C	O
-	O
terminal	O
43	O
(	O
ScMep2	O
)	O
and	O
25	O
residues	O
(	O
CaMep2	O
),	O
which	O
are	O
poorly	O
conserved	O
and	O
presumably	O
disordered	O
.	O
	O
Both	O
Mep2	O
proteins	O
show	O
the	O
archetypal	O
trimeric	O
assemblies	O
in	O
which	O
each	O
monomer	O
consists	O
of	O
11	O
TM	O
helices	O
surrounding	O
a	O
central	O
pore	O
.	O
	O
Important	O
functional	O
features	O
such	O
as	O
the	O
extracellular	O
ammonium	O
binding	O
site	O
,	O
the	O
Phe	O
gate	O
and	O
the	O
twin	O
-	O
His	O
motif	O
within	O
the	O
hydrophobic	O
channel	O
are	O
all	O
very	O
similar	O
to	O
those	O
present	O
in	O
the	O
bacterial	O
transporters	O
and	O
RhCG	O
.	O
	O
In	O
the	O
remainder	O
of	O
the	O
manuscript	O
,	O
we	O
will	O
specifically	O
discuss	O
CaMep2	O
due	O
to	O
the	O
superior	O
resolution	O
of	O
the	O
structure	O
.	O
	O
While	O
the	O
overall	O
architecture	O
of	O
Mep2	O
is	O
similar	O
to	O
that	O
of	O
the	O
prokaryotic	O
transporters	O
(	O
Cα	O
r	O
.	O
m	O
.	O
s	O
.	O
d	O
.	O
with	O
Amt	O
-	O
1	O
=	O
1	O
.	O
4	O
Å	O
for	O
361	O
residues	O
),	O
there	O
are	O
large	O
differences	O
within	O
the	O
N	O
terminus	O
,	O
intracellular	O
loops	O
(	O
ICLs	O
)	O
ICL1	O
and	O
ICL3	O
,	O
and	O
the	O
CTR	O
.	O
	O
Moreover	O
,	O
the	O
N	O
terminus	O
of	O
one	O
monomer	O
interacts	O
with	O
the	O
extended	O
extracellular	O
loop	O
ECL5	O
of	O
a	O
neighbouring	O
monomer	O
.	O
	O
However	O
,	O
given	O
that	O
an	O
N	O
-	O
terminal	O
deletion	O
mutant	O
(	O
2	B-mutant
-	I-mutant
27Δ	I-mutant
)	O
grows	O
as	O
well	O
as	O
wild	O
-	O
type	O
(	O
WT	O
)	O
Mep2	O
on	O
minimal	O
ammonium	O
medium	O
(	O
Fig	O
.	O
3	O
and	O
Supplementary	O
Fig	O
.	O
1	O
),	O
the	O
importance	O
of	O
the	O
N	O
terminus	O
for	O
Mep2	O
activity	O
is	O
not	O
clear	O
.	O
	O
Mep2	O
channels	O
are	O
closed	O
by	O
a	O
two	O
-	O
tier	O
channel	O
block	O
	O
In	O
the	O
vicinity	O
of	O
the	O
Mep2	O
channel	O
exit	O
,	O
the	O
cytoplasmic	O
end	O
of	O
TM2	O
has	O
unwound	O
,	O
generating	O
a	O
longer	O
ICL1	O
even	O
though	O
there	O
are	O
no	O
insertions	O
in	O
this	O
region	O
compared	O
to	O
the	O
bacterial	O
proteins	O
(	O
Figs	O
2	O
and	O
4	O
).	O
	O
The	O
largest	O
backbone	O
movements	O
of	O
equivalent	O
residues	O
within	O
ICL1	O
are	O
∼	O
10	O
Å	O
,	O
markedly	O
affecting	O
the	O
conserved	O
basic	O
RxK	O
motif	O
(	O
Fig	O
.	O
4	O
).	O
	O
In	O
addition	O
to	O
changing	O
the	O
RxK	O
motif	O
,	O
the	O
movement	O
of	O
ICL1	O
has	O
another	O
,	O
crucial	O
functional	O
consequence	O
.	O
	O
At	O
the	O
C	O
-	O
terminal	O
end	O
of	O
TM1	O
,	O
the	O
side	O
-	O
chain	O
hydroxyl	O
group	O
of	O
the	O
relatively	O
conserved	O
Tyr49	O
(	O
Tyr53	O
in	O
ScMep2	O
)	O
makes	O
a	O
strong	O
hydrogen	O
bond	O
with	O
the	O
ɛ2	O
nitrogen	O
atom	O
of	O
the	O
absolutely	O
conserved	O
His342	O
of	O
the	O
twin	O
-	O
His	O
motif	O
(	O
His348	O
in	O
ScMep2	O
),	O
closing	O
the	O
channel	O
(	O
Figs	O
4	O
and	O
5	O
).	O
	O
In	O
bacterial	O
Amt	O
proteins	O
,	O
this	O
Tyr	O
side	O
chain	O
is	O
rotated	O
∼	O
4	O
Å	O
away	O
as	O
a	O
result	O
of	O
the	O
different	O
conformation	O
of	O
TM1	O
,	O
leaving	O
the	O
channel	O
open	O
and	O
the	O
histidine	O
available	O
for	O
its	O
putative	O
role	O
in	O
substrate	O
transport	O
(	O
Supplementary	O
Fig	O
.	O
2	O
).	O
	O
Finally	O
,	O
the	O
important	O
ICL3	O
linking	O
the	O
pseudo	O
-	O
symmetrical	O
halves	O
(	O
TM1	O
-	O
5	O
and	O
TM6	O
-	O
10	O
)	O
of	O
the	O
transporter	O
is	O
also	O
shifted	O
up	O
to	O
∼	O
10	O
Å	O
and	O
forms	O
an	O
additional	O
barrier	O
that	O
closes	O
the	O
channel	O
on	O
the	O
cytoplasmic	O
side	O
(	O
Fig	O
.	O
5	O
).	O
	O
This	O
two	O
-	O
tier	O
channel	O
block	O
likely	O
ensures	O
that	O
very	O
little	O
ammonium	O
transport	O
will	O
take	O
place	O
under	O
nitrogen	O
-	O
sufficient	O
conditions	O
.	O
	O
The	O
closed	O
state	O
of	O
the	O
channel	O
might	O
also	O
explain	O
why	O
no	O
density	O
,	O
which	O
could	O
correspond	O
to	O
ammonium	O
(	O
or	O
water	O
),	O
is	O
observed	O
in	O
the	O
hydrophobic	O
part	O
of	O
the	O
Mep2	O
channel	O
close	O
to	O
the	O
twin	O
-	O
His	O
motif	O
.	O
	O
The	O
final	O
region	O
in	O
Mep2	O
that	O
shows	O
large	O
differences	O
compared	O
with	O
the	O
bacterial	O
transporters	O
is	O
the	O
CTR	O
.	O
	O
By	O
contrast	O
,	O
in	O
the	O
structures	O
of	O
bacterial	O
proteins	O
,	O
the	O
CTR	O
is	O
docked	O
tightly	O
onto	O
the	O
N	O
-	O
terminal	O
half	O
of	O
the	O
transporters	O
(	O
corresponding	O
to	O
TM1	O
-	O
5	O
),	O
resulting	O
in	O
a	O
more	O
compact	O
structure	O
.	O
	O
This	O
is	O
illustrated	O
by	O
the	O
positions	O
of	O
the	O
five	O
universally	O
conserved	O
residues	O
within	O
the	O
CTR	O
,	O
that	O
is	O
,	O
Arg415	O
(	O
370	O
),	O
Glu421	O
(	O
376	O
),	O
Gly424	O
(	O
379	O
),	O
Asp426	O
(	O
381	O
)	O
and	O
Tyr	O
435	O
(	O
390	O
)	O
in	O
CaMep2	O
(	O
Amt	O
-	O
1	O
)	O
(	O
Fig	O
.	O
2	O
).	O
	O
On	O
one	O
side	O
,	O
the	O
Tyr390	O
hydroxyl	O
in	O
Amt	O
-	O
1	O
is	O
hydrogen	O
bonded	O
with	O
the	O
side	O
chain	O
of	O
the	O
conserved	O
His185	O
at	O
the	O
C	O
-	O
terminal	O
end	O
of	O
loop	O
ICL3	O
.	O
	O
Similar	O
interactions	O
were	O
also	O
modelled	O
in	O
the	O
active	O
,	O
non	O
-	O
phosphorylated	O
plant	O
AtAmt	O
-	O
1	O
;	O
1	O
structure	O
(	O
for	O
example	O
,	O
Y467	O
-	O
H239	O
and	O
D458	O
-	O
K71	O
).	O
	O
The	O
result	O
of	O
these	O
interactions	O
is	O
that	O
the	O
CTR	O
‘	O
hugs	O
'	O
the	O
N	O
-	O
terminal	O
half	O
of	O
the	O
transporters	O
(	O
Fig	O
.	O
4	O
).	O
	O
Also	O
noteworthy	O
is	O
Asp381	O
,	O
the	O
side	O
chain	O
of	O
which	O
interacts	O
strongly	O
with	O
the	O
positive	O
dipole	O
on	O
the	O
N	O
-	O
terminal	O
end	O
of	O
TM2	O
.	O
	O
In	O
the	O
Mep2	O
structures	O
,	O
none	O
of	O
the	O
interactions	O
mentioned	O
above	O
are	O
present	O
.	O
	O
Recently	O
Boeckstaens	O
et	O
al	O
.	O
provided	O
evidence	O
that	O
Ser457	O
in	O
ScMep2	O
(	O
corresponding	O
to	O
Ser453	O
in	O
CaMep2	O
)	O
is	O
phosphorylated	O
by	O
the	O
TORC1	O
effector	O
kinase	O
Npr1	O
under	O
nitrogen	O
-	O
limiting	O
conditions	O
.	O
	O
Conversely	O
,	O
the	O
phosphorylation	O
-	O
mimicking	O
S457D	B-mutant
variant	O
is	O
active	O
both	O
in	O
the	O
triple	B-mutant
mepΔ	I-mutant
background	O
and	O
in	O
a	O
triple	B-mutant
mepΔ	I-mutant
npr1Δ	I-mutant
strain	O
(	O
Fig	O
.	O
3	O
).	O
	O
Mutation	O
of	O
other	O
potential	O
phosphorylation	O
sites	O
in	O
the	O
CTR	O
did	O
not	O
support	O
growth	O
in	O
the	O
npr1Δ	B-mutant
background	O
.	O
	O
Collectively	O
,	O
these	O
data	O
suggest	O
that	O
phosphorylation	O
of	O
Ser457	O
opens	O
the	O
Mep2	O
channel	O
to	O
allow	O
ammonium	O
uptake	O
.	O
	O
Ser457	O
is	O
located	O
in	O
a	O
part	O
of	O
the	O
CTR	O
that	O
is	O
conserved	O
in	O
a	O
subgroup	O
of	O
Mep2	O
proteins	O
,	O
but	O
which	O
is	O
not	O
present	O
in	O
bacterial	O
proteins	O
(	O
Fig	O
.	O
2	O
).	O
	O
This	O
segment	O
(	O
residues	O
450	O
–	O
457	O
in	O
ScMep2	O
and	O
446	O
–	O
453	O
in	O
CaMep2	O
)	O
was	O
dubbed	O
an	O
autoinhibitory	O
(	O
AI	O
)	O
region	O
based	O
on	O
the	O
fact	O
that	O
its	O
removal	O
generates	O
an	O
active	O
transporter	O
in	O
the	O
absence	O
of	O
Npr1	O
(	O
Fig	O
.	O
3	O
).	O
	O
The	O
AI	O
region	O
packs	O
against	O
the	O
cytoplasmic	O
ends	O
of	O
TM2	O
and	O
TM4	O
,	O
physically	O
linking	O
the	O
main	O
body	O
of	O
the	O
transporter	O
with	O
the	O
CTR	O
via	O
main	O
chain	O
interactions	O
and	O
side	O
-	O
chain	O
interactions	O
of	O
Val447	O
,	O
Asp449	O
,	O
Pro450	O
and	O
Arg452	O
(	O
Fig	O
.	O
6	O
).	O
	O
Strikingly	O
,	O
the	O
Npr1	O
target	O
serine	O
residue	O
is	O
located	O
at	O
the	O
periphery	O
of	O
the	O
trimer	O
,	O
far	O
away	O
(∼	O
30	O
Å	O
)	O
from	O
any	O
channel	O
exit	O
(	O
Fig	O
.	O
6	O
).	O
	O
Despite	O
its	O
location	O
at	O
the	O
periphery	O
of	O
the	O
trimer	O
,	O
the	O
electron	O
density	O
for	O
the	O
serine	O
is	O
well	O
defined	O
in	O
both	O
Mep2	O
structures	O
and	O
corresponds	O
to	O
the	O
non	O
-	O
phosphorylated	O
state	O
(	O
Fig	O
.	O
6	O
).	O
	O
For	O
ScMep2	O
,	O
Ser457	O
is	O
the	O
most	O
C	O
-	O
terminal	O
residue	O
for	O
which	O
electron	O
density	O
is	O
visible	O
,	O
indicating	O
that	O
the	O
region	O
beyond	O
Ser457	O
is	O
disordered	O
.	O
	O
In	O
CaMep2	O
,	O
the	O
visible	O
part	O
of	O
the	O
sequence	O
extends	O
for	O
two	O
residues	O
beyond	O
Ser453	O
(	O
Fig	O
.	O
6	O
).	O
	O
The	O
disordered	O
part	O
of	O
the	O
CTR	O
is	O
not	O
conserved	O
in	O
ammonium	O
transporters	O
(	O
Fig	O
.	O
2	O
),	O
suggesting	O
that	O
it	O
is	O
not	O
important	O
for	O
transport	O
.	O
	O
Interestingly	O
,	O
a	O
ScMep2	O
457Δ	B-mutant
truncation	O
mutant	O
in	O
which	O
a	O
His	O
-	O
tag	O
directly	O
follows	O
Ser457	O
is	O
highly	O
expressed	O
but	O
has	O
low	O
activity	O
(	O
Fig	O
.	O
3	O
and	O
Supplementary	O
Fig	O
.	O
1b	O
),	O
suggesting	O
that	O
the	O
His	O
-	O
tag	O
interferes	O
with	O
phosphorylation	O
by	O
Npr1	O
.	O
	O
The	O
same	O
mutant	B-mutant
lacking	O
the	O
His	O
-	O
tag	O
has	O
WT	O
properties	O
(	O
Supplementary	O
Fig	O
.	O
1b	O
),	O
confirming	O
that	O
the	O
region	O
following	O
the	O
phosphorylation	O
site	O
is	O
dispensable	O
for	O
function	O
.	O
	O
Given	O
that	O
Ser457	O
/	O
453	O
is	O
far	O
from	O
any	O
channel	O
exit	O
(	O
Fig	O
.	O
6	O
),	O
the	O
crucial	O
question	O
is	O
how	O
phosphorylation	O
opens	O
the	O
Mep2	O
channel	O
to	O
generate	O
an	O
active	O
transporter	O
.	O
	O
Boeckstaens	O
et	O
al	O
.	O
proposed	O
that	O
phosphorylation	O
does	O
not	O
affect	O
channel	O
activity	O
directly	O
,	O
but	O
instead	O
relieves	O
inhibition	O
by	O
the	O
AI	O
region	O
.	O
	O
The	O
data	O
behind	O
this	O
hypothesis	O
is	O
the	O
observation	O
that	O
a	O
ScMep2	O
449	B-mutant
-	I-mutant
485Δ	I-mutant
deletion	O
mutant	O
lacking	O
the	O
AI	O
region	O
is	O
highly	O
active	O
in	O
MA	O
uptake	O
both	O
in	O
the	O
triple	B-mutant
mepΔ	I-mutant
and	O
triple	B-mutant
mepΔ	I-mutant
npr1Δ	I-mutant
backgrounds	O
,	O
implying	O
that	O
this	O
Mep2	B-mutant
variant	I-mutant
has	O
a	O
constitutively	O
open	O
channel	O
.	O
	O
This	O
is	O
not	O
unexpected	O
given	O
the	O
fact	O
that	O
the	O
AI	O
region	O
bridges	O
the	O
CTR	O
and	O
the	O
main	O
body	O
of	O
Mep2	O
(	O
Fig	O
.	O
6	O
).	O
	O
Interestingly	O
,	O
however	O
,	O
the	O
Tyr49	O
-	O
His342	O
hydrogen	O
bond	O
that	O
closes	O
the	O
channel	O
in	O
the	O
WT	O
protein	O
is	O
still	O
present	O
(	O
Fig	O
.	O
7	O
and	O
Supplementary	O
Fig	O
.	O
2	O
).	O
	O
The	O
second	O
possibility	O
is	O
that	O
the	O
Tyr	O
–	O
His	O
hydrogen	O
bond	O
has	O
to	O
be	O
disrupted	O
by	O
the	O
incoming	O
substrate	O
to	O
open	O
the	O
channel	O
.	O
	O
The	O
importance	O
of	O
the	O
Tyr	O
–	O
His	O
hydrogen	O
bond	O
is	O
underscored	O
by	O
the	O
fact	O
that	O
its	O
removal	O
in	O
the	O
ScMep2	O
Y53A	B-mutant
mutant	O
results	O
in	O
a	O
constitutively	O
active	O
transporter	O
(	O
Fig	O
.	O
3	O
).	O
	O
Phosphorylation	O
causes	O
a	O
conformational	O
change	O
in	O
the	O
CTR	O
	O
Do	O
the	O
Mep2	O
structures	O
provide	O
any	O
clues	O
regarding	O
the	O
potential	O
effect	O
of	O
phosphorylation	O
?	O
	O
The	O
side	O
-	O
chain	O
hydroxyl	O
of	O
Ser457	O
/	O
453	O
is	O
located	O
in	O
a	O
well	O
-	O
defined	O
electronegative	O
pocket	O
that	O
is	O
solvent	O
accessible	O
(	O
Fig	O
.	O
6	O
).	O
	O
The	O
closest	O
atoms	O
to	O
the	O
serine	O
hydroxyl	O
group	O
are	O
the	O
backbone	O
carbonyl	O
atoms	O
of	O
Asp419	O
,	O
Glu420	O
and	O
Glu421	O
,	O
which	O
are	O
3	O
–	O
4	O
Å	O
away	O
.	O
	O
We	O
therefore	O
predict	O
that	O
phosphorylation	O
of	O
Ser453	O
will	O
result	O
in	O
steric	O
clashes	O
as	O
well	O
as	O
electrostatic	O
repulsion	O
,	O
which	O
in	O
turn	O
might	O
cause	O
substantial	O
conformational	O
changes	O
within	O
the	O
CTR	O
.	O
	O
Unexpectedly	O
,	O
the	O
AI	O
segment	O
containing	O
the	O
mutated	O
residues	O
has	O
only	O
undergone	O
a	O
slight	O
shift	O
compared	O
with	O
the	O
WT	O
protein	O
(	O
Fig	O
.	O
8	O
and	O
Supplementary	O
Fig	O
.	O
3	O
).	O
	O
By	O
contrast	O
,	O
the	O
conserved	O
part	O
of	O
the	O
CTR	O
has	O
undergone	O
a	O
large	O
conformational	O
change	O
involving	O
formation	O
of	O
a	O
12	O
-	O
residue	O
-	O
long	O
α	O
-	O
helix	O
from	O
Leu427	O
to	O
Asp438	O
.	O
	O
This	O
is	O
the	O
first	O
time	O
a	O
large	O
conformational	O
change	O
has	O
been	O
observed	O
in	O
an	O
ammonium	O
transporter	O
as	O
a	O
result	O
of	O
a	O
mutation	O
,	O
and	O
confirms	O
previous	O
hypotheses	O
that	O
phosphorylation	O
causes	O
structural	O
changes	O
in	O
the	O
CTR	O
.	O
	O
To	O
exclude	O
the	O
possibility	O
that	O
the	O
additional	O
R452D	B-mutant
mutation	O
is	O
responsible	O
for	O
the	O
observed	O
changes	O
,	O
we	O
also	O
determined	O
the	O
structure	O
of	O
the	O
‘	O
single	B-mutant
D	I-mutant
'	O
S453D	B-mutant
mutant	O
.	O
	O
As	O
shown	O
in	O
Supplementary	O
Fig	O
.	O
4	O
,	O
the	O
consequence	O
of	O
the	O
single	B-mutant
D	I-mutant
mutation	O
is	O
very	O
similar	O
to	O
that	O
of	O
the	O
DD	B-mutant
substitution	I-mutant
,	O
with	O
conformational	O
changes	O
and	O
increased	O
dynamics	O
confined	O
to	O
the	O
conserved	O
part	O
of	O
the	O
CTR	O
(	O
Supplementary	O
Fig	O
.	O
4	O
).	O
	O
To	O
supplement	O
the	O
crystal	O
structures	O
,	O
we	O
also	O
performed	O
modelling	O
and	O
MD	O
studies	O
of	O
WT	O
CaMep2	O
,	O
the	O
DD	B-mutant
mutant	I-mutant
and	O
phosphorylated	O
protein	O
(	O
S453J	B-mutant
).	O
	O
In	O
the	O
WT	O
structure	O
,	O
the	O
acidic	O
residues	O
Asp419	O
,	O
Glu420	O
and	O
Glu421	O
are	O
within	O
hydrogen	O
bonding	O
distance	O
of	O
Ser453	O
.	O
	O
After	O
200	O
ns	O
of	O
MD	O
simulation	O
,	O
the	O
interactions	O
between	O
these	O
residues	O
and	O
Ser453	O
remain	O
intact	O
.	O
	O
In	O
particular	O
,	O
persistent	O
hydrogen	O
bonds	O
are	O
observed	O
between	O
the	O
Ser453	O
hydroxyl	O
group	O
and	O
the	O
acidic	O
group	O
of	O
Glu420	O
,	O
and	O
also	O
between	O
the	O
amine	O
group	O
of	O
Ser453	O
and	O
the	O
backbone	O
carbonyl	O
of	O
Glu420	O
(	O
Supplementary	O
Fig	O
.	O
5	O
).	O
	O
The	O
DD	B-mutant
mutant	I-mutant
is	O
also	O
stable	O
during	O
the	O
simulations	O
,	O
but	O
the	O
average	O
backbone	O
r	O
.	O
m	O
.	O
s	O
.	O
d	O
of	O
∼	O
3	O
.	O
6	O
Å	O
suggests	O
slightly	O
more	O
conformational	O
flexibility	O
than	O
WT	O
.	O
	O
As	O
the	O
simulation	O
proceeds	O
,	O
the	O
side	O
chains	O
of	O
the	O
acidic	O
residues	O
move	O
away	O
from	O
Asp452	O
and	O
Asp453	O
,	O
presumably	O
to	O
avoid	O
electrostatic	O
repulsion	O
.	O
	O
The	O
protein	O
is	O
structurally	O
stable	O
throughout	O
the	O
simulation	O
with	O
little	O
deviation	O
in	O
the	O
other	O
parts	O
of	O
the	O
protein	O
.	O
	O
Finally	O
,	O
the	O
S453J	B-mutant
mutant	O
is	O
also	O
stable	O
throughout	O
the	O
200	O
-	O
ns	O
simulation	O
and	O
has	O
an	O
average	O
backbone	O
deviation	O
of	O
∼	O
3	O
.	O
8	O
Å	O
,	O
which	O
is	O
similar	O
to	O
the	O
DD	B-mutant
mutant	I-mutant
.	O
	O
The	O
short	O
helix	O
formed	O
by	O
residues	O
Leu427	O
to	O
Asp438	O
unravels	O
during	O
the	O
simulations	O
to	O
a	O
disordered	O
state	O
.	O
	O
Thus	O
,	O
the	O
MD	O
simulations	O
support	O
the	O
notion	O
from	O
the	O
crystal	O
structures	O
that	O
phosphorylation	O
generates	O
conformational	O
changes	O
in	O
the	O
conserved	O
part	O
of	O
the	O
CTR	O
.	O
	O
However	O
,	O
the	O
conformational	O
changes	O
for	O
the	O
phosphomimetic	B-mutant
mutants	I-mutant
in	O
the	O
crystals	O
are	O
confined	O
to	O
the	O
CTR	O
(	O
Fig	O
.	O
8	O
),	O
and	O
the	O
channels	O
are	O
still	O
closed	O
(	O
Supplementary	O
Fig	O
.	O
2	O
).	O
	O
One	O
possible	O
explanation	O
is	O
that	O
the	O
mutants	B-mutant
do	O
not	O
accurately	O
mimic	O
a	O
phosphoserine	O
,	O
but	O
the	O
observation	O
that	O
the	O
S453D	B-mutant
and	O
DD	B-mutant
mutants	I-mutant
are	O
fully	O
active	O
in	O
the	O
absence	O
of	O
Npr1	O
suggests	O
that	O
the	O
mutations	O
do	O
mimic	O
the	O
effect	O
of	O
phosphorylation	O
(	O
Fig	O
.	O
3	O
).	O
	O
The	O
fact	O
that	O
the	O
S453D	B-mutant
structure	O
was	O
obtained	O
in	O
the	O
presence	O
of	O
10	O
mM	O
ammonium	O
ions	O
suggests	O
that	O
the	O
crystallization	O
process	O
favours	O
closed	O
states	O
of	O
the	O
Mep2	O
channels	O
.	O
	O
Knowledge	O
about	O
ammonium	O
transporter	O
structure	O
has	O
been	O
obtained	O
from	O
experimental	O
and	O
theoretical	O
studies	O
on	O
bacterial	O
family	O
members	O
.	O
	O
In	O
addition	O
,	O
a	O
number	O
of	O
biochemical	O
and	O
genetic	O
studies	O
are	O
available	O
for	O
bacterial	O
,	O
fungal	O
and	O
plant	O
proteins	O
.	O
	O
These	O
efforts	O
have	O
advanced	O
our	O
knowledge	O
considerably	O
but	O
have	O
not	O
yet	O
yielded	O
atomic	O
-	O
level	O
answers	O
to	O
several	O
important	O
mechanistic	O
questions	O
,	O
including	O
how	O
ammonium	O
transport	O
is	O
regulated	O
in	O
eukaryotes	O
and	O
the	O
mechanism	O
of	O
ammonium	O
signalling	O
.	O
	O
In	O
Arabidopsis	O
thaliana	O
Amt	O
-	O
1	O
;	O
1	O
,	O
phosphorylation	O
of	O
the	O
CTR	O
residue	O
T460	O
under	O
conditions	O
of	O
high	O
ammonium	O
inhibits	O
transport	O
activity	O
,	O
that	O
is	O
,	O
the	O
default	O
(	O
non	O
-	O
phosphorylated	O
)	O
state	O
of	O
the	O
plant	O
transporter	O
is	O
open	O
.	O
	O
Interestingly	O
,	O
phosphomimetic	B-mutant
mutations	I-mutant
introduced	O
into	O
one	O
monomer	O
inactivate	O
the	O
entire	O
trimer	O
,	O
indicating	O
that	O
(	O
i	O
)	O
heterotrimerization	O
occurs	O
and	O
(	O
ii	O
)	O
the	O
CTR	O
mediates	O
allosteric	O
regulation	O
of	O
ammonium	O
transport	O
activity	O
via	O
phosphorylation	O
.	O
	O
Owing	O
to	O
the	O
lack	O
of	O
structural	O
information	O
for	O
plant	O
AMTs	O
,	O
the	O
details	O
of	O
channel	O
closure	O
and	O
inter	O
-	O
monomer	O
crosstalk	O
are	O
not	O
yet	O
clear	O
.	O
	O
Contrasting	O
with	O
the	O
plant	O
transporters	O
,	O
the	O
inactive	O
states	O
of	O
Mep2	O
proteins	O
under	O
conditions	O
of	O
high	O
ammonium	O
are	O
non	O
-	O
phosphorylated	O
,	O
with	O
channels	O
that	O
are	O
closed	O
on	O
the	O
cytoplasmic	O
side	O
.	O
	O
In	O
fungi	O
,	O
preventing	O
ammonium	O
entry	O
via	O
channel	O
closure	O
in	O
ammonium	O
transporters	O
would	O
be	O
one	O
way	O
to	O
alleviate	O
ammonium	O
toxicity	O
,	O
in	O
addition	O
to	O
ammonium	O
excretion	O
via	O
Ato	O
transporters	O
and	O
amino	O
-	O
acid	O
secretion	O
.	O
	O
In	O
addition	O
,	O
ICL1	O
has	O
shifted	O
inwards	O
to	O
contribute	O
to	O
the	O
channel	O
closure	O
by	O
engaging	O
His2	O
from	O
the	O
twin	O
-	O
His	O
motif	O
via	O
hydrogen	O
bonding	O
with	O
a	O
highly	O
conserved	O
tyrosine	O
hydroxyl	O
group	O
.	O
	O
Upon	O
phosphorylation	O
by	O
the	O
Npr1	O
kinase	O
in	O
response	O
to	O
nitrogen	O
limitation	O
,	O
the	O
region	O
around	O
the	O
conserved	O
ExxGxD	O
motif	O
undergoes	O
a	O
conformational	O
change	O
that	O
opens	O
the	O
channel	O
(	O
Fig	O
.	O
9	O
).	O
	O
Importantly	O
,	O
the	O
structural	O
similarities	O
in	O
the	O
TM	O
parts	O
of	O
Mep2	O
and	O
AfAmt	O
-	O
1	O
(	O
Fig	O
.	O
5a	O
)	O
suggest	O
that	O
channel	O
opening	O
/	O
closure	O
does	O
not	O
require	O
substantial	O
changes	O
in	O
the	O
residues	O
lining	O
the	O
channel	O
.	O
	O
How	O
exactly	O
the	O
channel	O
opens	O
and	O
whether	O
opening	O
is	O
intra	O
-	O
monomeric	O
are	O
still	O
open	O
questions	O
;	O
it	O
is	O
possible	O
that	O
the	O
change	O
in	O
the	O
CTR	O
may	O
disrupt	O
its	O
interactions	O
with	O
ICL3	O
of	O
the	O
neighbouring	O
monomer	O
(	O
Fig	O
.	O
9b	O
),	O
which	O
could	O
result	O
in	O
opening	O
of	O
the	O
neighbouring	O
channel	O
via	O
inward	O
movement	O
of	O
its	O
ICL3	O
.	O
	O
Whether	O
or	O
not	O
Mep2	O
channel	O
opening	O
requires	O
,	O
in	O
addition	O
to	O
phosphorylation	O
,	O
disruption	O
of	O
the	O
Tyr	O
–	O
His2	O
interaction	O
by	O
the	O
ammonium	O
substrate	O
is	O
not	O
yet	O
clear	O
.	O
	O
Is	O
our	O
model	O
for	O
opening	O
and	O
closing	O
of	O
Mep2	O
channels	O
valid	O
for	O
other	O
eukaryotic	O
ammonium	O
transporters	O
?	O
Our	O
structural	O
data	O
support	O
previous	O
studies	O
and	O
clarify	O
the	O
central	O
role	O
of	O
the	O
CTR	O
and	O
cytoplasmic	O
loops	O
in	O
the	O
transition	O
between	O
closed	O
and	O
open	O
states	O
.	O
	O
In	O
addition	O
,	O
the	O
AI	O
region	O
of	O
the	O
CTR	O
containing	O
the	O
Npr1	O
kinase	O
site	O
is	O
conserved	O
in	O
only	O
a	O
subset	O
of	O
fungal	O
transporters	O
,	O
suggesting	O
that	O
the	O
details	O
of	O
the	O
structural	O
changes	O
underpinning	O
regulation	O
vary	O
.	O
	O
Nevertheless	O
,	O
given	O
the	O
central	O
role	O
of	O
absolutely	O
conserved	O
residues	O
within	O
the	O
ICL1	O
-	O
ICL3	O
-	O
CTR	O
interaction	O
network	O
(	O
Fig	O
.	O
4	O
),	O
we	O
propose	O
that	O
the	O
structural	O
basics	O
of	O
fungal	O
ammonium	O
transporter	O
activation	O
are	O
conserved	O
.	O
	O
The	O
fact	O
that	O
Mep2	O
orthologues	O
of	O
distantly	O
related	O
fungi	O
are	O
fully	O
functional	O
in	O
ammonium	O
transport	O
and	O
signalling	O
in	O
S	O
.	O
cerevisiae	O
supports	O
this	O
notion	O
.	O
	O
With	O
regards	O
to	O
plant	O
AMTs	O
,	O
it	O
has	O
been	O
proposed	O
that	O
phosphorylation	O
at	O
T460	O
generates	O
conformational	O
changes	O
that	O
would	O
close	O
the	O
neighbouring	O
pore	O
via	O
the	O
C	O
terminus	O
.	O
	O
This	O
assumption	O
was	O
based	O
partly	O
on	O
a	O
homology	O
model	O
for	O
Amt	O
-	O
1	O
;	O
1	O
based	O
on	O
the	O
(	O
open	O
)	O
archaebacterial	O
AfAmt	O
-	O
1	O
structure	O
,	O
which	O
suggested	O
that	O
the	O
C	O
terminus	O
of	O
Amt	O
-	O
1	O
;	O
1	O
would	O
extend	O
further	O
to	O
the	O
neighbouring	O
monomer	O
.	O
	O
Our	O
Mep2	O
structures	O
show	O
that	O
this	O
assumption	O
may	O
not	O
be	O
correct	O
(	O
Fig	O
.	O
4	O
and	O
Supplementary	O
Fig	O
.	O
6	O
).	O
	O
Based	O
on	O
the	O
available	O
structural	O
information	O
,	O
we	O
consider	O
it	O
more	O
likely	O
that	O
phosphorylation	O
-	O
mediated	O
pore	O
closure	O
in	O
Amt	O
-	O
1	O
;	O
1	O
is	O
intra	O
-	O
monomeric	O
,	O
via	O
disruption	O
of	O
the	O
interactions	O
between	O
the	O
CTR	O
and	O
ICL1	O
/	O
ICL3	O
(	O
for	O
example	O
,	O
Y467	O
-	O
H239	O
and	O
D458	O
-	O
K71	O
).	O
	O
There	O
is	O
generally	O
no	O
equivalent	O
for	O
CaMep2	O
Tyr49	O
in	O
plant	O
AMTs	O
,	O
indicating	O
that	O
a	O
Tyr	O
–	O
His2	O
hydrogen	O
bond	O
as	O
observed	O
in	O
Mep2	O
may	O
not	O
contribute	O
to	O
the	O
closed	O
state	O
in	O
plant	O
transporters	O
.	O
	O
We	O
propose	O
that	O
intra	O
-	O
monomeric	O
CTR	O
-	O
ICL1	O
/	O
ICL3	O
interactions	O
lie	O
at	O
the	O
basis	O
of	O
regulation	O
of	O
both	O
fungal	O
and	O
plant	O
ammonium	O
transporters	O
;	O
close	O
interactions	O
generate	O
open	O
channels	O
,	O
whereas	O
the	O
lack	O
of	O
‘	O
intra	O
-'	O
interactions	O
leads	O
to	O
inactive	O
states	O
.	O
	O
The	O
need	O
to	O
regulate	O
in	O
opposite	O
ways	O
may	O
be	O
the	O
reason	O
why	O
the	O
phosphorylation	O
sites	O
are	O
in	O
different	O
parts	O
of	O
the	O
CTR	O
,	O
that	O
is	O
,	O
centrally	O
located	O
close	O
to	O
the	O
ExxGxD	O
motif	O
in	O
AMTs	O
and	O
peripherally	O
in	O
Mep2	O
.	O
	O
In	O
this	O
way	O
,	O
phosphorylation	O
can	O
either	O
lead	O
to	O
channel	O
closing	O
(	O
in	O
the	O
case	O
of	O
AMTs	O
)	O
or	O
channel	O
opening	O
in	O
the	O
case	O
of	O
Mep2	O
.	O
	O
Our	O
model	O
also	O
provides	O
an	O
explanation	O
for	O
the	O
observation	O
that	O
certain	B-mutant
mutations	I-mutant
within	O
the	O
CTR	O
completely	O
abolish	O
transport	O
activity	O
.	O
	O
An	O
example	O
of	O
an	O
inactivating	O
residue	O
is	O
the	O
glycine	O
of	O
the	O
ExxGxD	O
motif	O
of	O
the	O
CTR	O
.	O
	O
Mutation	O
of	O
this	O
residue	O
(	O
G393	O
in	O
EcAmtB	O
;	O
G456	O
in	O
AtAmt	O
-	O
1	O
;	O
1	O
)	O
inactivates	O
transporters	O
as	O
diverse	O
as	O
Escherichia	O
coli	O
AmtB	O
and	O
A	O
.	O
thaliana	O
Amt	O
-	O
1	O
;	O
1	O
(	O
refs	O
).	O
	O
Such	O
mutations	O
likely	O
cause	O
structural	O
changes	O
in	O
the	O
CTR	O
that	O
prevent	O
close	O
contacts	O
between	O
the	O
CTR	O
and	O
ICL1	O
/	O
ICL3	O
,	O
thereby	O
stabilizing	O
a	O
closed	O
state	O
that	O
may	O
be	O
similar	O
to	O
that	O
observed	O
in	O
Mep2	O
.	O
	O
Regulation	O
and	O
modulation	O
of	O
membrane	O
transport	O
by	O
phosphorylation	O
is	O
known	O
to	O
occur	O
in	O
,	O
for	O
example	O
,	O
aquaporins	O
and	O
urea	O
transporters	O
,	O
and	O
is	O
likely	O
to	O
be	O
a	O
common	O
theme	O
for	O
eukaryotic	O
channels	O
and	O
transporters	O
.	O
	O
Recently	O
,	O
phosphorylation	O
was	O
also	O
shown	O
to	O
modulate	O
substrate	O
affinity	O
in	O
nitrate	O
transporters	O
.	O
	O
With	O
respect	O
to	O
ammonium	O
transport	O
,	O
phosphorylation	O
has	O
thus	O
far	O
only	O
been	O
shown	O
for	O
A	O
.	O
thaliana	O
AMTs	O
and	O
for	O
S	O
.	O
cerevisiae	O
Mep2	O
(	O
refs	O
).	O
	O
However	O
,	O
the	O
absence	O
of	O
GlnK	O
proteins	O
in	O
eukaryotes	O
suggests	O
that	O
phosphorylation	O
-	O
based	O
regulation	O
of	O
ammonium	O
transport	O
may	O
be	O
widespread	O
.	O
	O
With	O
respect	O
to	O
Mep2	O
-	O
mediated	O
signalling	O
to	O
induce	O
pseudohyphal	O
growth	O
,	O
two	O
models	O
have	O
been	O
put	O
forward	O
as	O
to	O
how	O
this	O
occurs	O
and	O
why	O
it	O
is	O
specific	O
to	O
Mep2	O
proteins	O
.	O
	O
In	O
one	O
model	O
,	O
signalling	O
is	O
proposed	O
to	O
depend	O
on	O
the	O
nature	O
of	O
the	O
transported	O
substrate	O
,	O
which	O
might	O
be	O
different	O
in	O
certain	O
subfamilies	O
of	O
ammonium	O
transporters	O
(	O
for	O
example	O
,	O
Mep1	O
/	O
Mep3	O
versus	O
Mep2	O
).	O
	O
In	O
the	O
other	O
model	O
,	O
signalling	O
is	O
thought	O
to	O
require	O
a	O
distinct	O
conformation	O
of	O
the	O
Mep2	O
transporter	O
occurring	O
during	O
the	O
transport	O
cycle	O
.	O
	O
While	O
the	O
current	O
study	O
does	O
not	O
specifically	O
address	O
the	O
mechanism	O
of	O
signalling	O
underlying	O
pseudohyphal	O
growth	O
,	O
our	O
structures	O
do	O
show	O
that	O
Mep2	O
proteins	O
can	O
assume	O
different	O
conformations	O
.	O
	O
It	O
is	O
clear	O
that	O
ammonium	O
transport	O
across	O
biomembranes	O
remains	O
a	O
fascinating	O
and	O
challenging	O
field	O
in	O
large	O
part	O
due	O
to	O
the	O
unique	O
properties	O
of	O
the	O
substrate	O
.	O
	O
Our	O
Mep2	O
structural	O
work	O
now	O
provides	O
a	O
foundation	O
for	O
future	O
studies	O
to	O
uncover	O
the	O
details	O
of	O
the	O
structural	O
changes	O
that	O
occur	O
during	O
eukaryotic	O
ammonium	O
transport	O
and	O
signaling	O
,	O
and	O
to	O
assess	O
the	O
possibility	O
to	O
utilize	O
small	O
molecules	O
to	O
shut	O
down	O
ammonium	O
sensing	O
and	O
downstream	O
signalling	O
pathways	O
in	O
pathogenic	O
fungi	O
.	O
	O
X	O
-	O
ray	O
crystal	O
structures	O
of	O
Mep2	O
transceptors	O
.	O
	O
The	O
region	O
showing	O
ICL1	O
(	O
blue	O
),	O
ICL3	O
(	O
green	O
)	O
and	O
the	O
CTR	O
(	O
red	O
)	O
is	O
boxed	O
for	O
comparison	O
.	O
	O
(	O
b	O
)	O
CaMep2	O
trimer	O
viewed	O
from	O
the	O
intracellular	O
side	O
(	O
right	O
).	O
	O
The	O
CTR	O
is	O
boxed	O
.	O
	O
(	O
c	O
)	O
Overlay	O
of	O
ScMep2	O
(	O
grey	O
)	O
and	O
CaMep2	O
(	O
rainbow	O
),	O
illustrating	O
the	O
differences	O
in	O
the	O
CTRs	O
.	O
	O
Sequence	O
conservation	O
in	O
ammonium	O
transporters	O
.	O
	O
ClustalW	O
alignment	O
of	O
CaMep2	O
,	O
ScMep2	O
,	O
A	O
.	O
fulgidus	O
Amt	O
-	O
1	O
,	O
E	O
.	O
coli	O
AmtB	O
and	O
A	O
.	O
thaliana	O
Amt	O
-	O
1	O
;	O
1	O
.	O
	O
The	O
conserved	O
RxK	O
motif	O
in	O
ICL1	O
is	O
boxed	O
in	O
blue	O
,	O
the	O
ER	O
motif	O
in	O
ICL2	O
in	O
cyan	O
,	O
the	O
conserved	O
ExxGxD	O
motif	O
of	O
the	O
CTR	O
in	O
red	O
and	O
the	O
AI	O
region	O
in	O
yellow	O
.	O
	O
Coloured	O
residues	O
are	O
functionally	O
important	O
and	O
correspond	O
to	O
those	O
of	O
the	O
Phe	O
gate	O
(	O
blue	O
),	O
the	O
binding	O
site	O
Trp	O
residue	O
(	O
magenta	O
)	O
and	O
the	O
twin	O
-	O
His	O
motif	O
(	O
red	O
).	O
	O
The	O
Npr1	O
kinase	O
site	O
in	O
the	O
AI	O
region	O
is	O
highlighted	O
pink	O
.	O
	O
The	O
grey	O
sequences	O
at	O
the	O
C	O
termini	O
of	O
CaMep2	O
and	O
ScMep2	O
are	O
not	O
visible	O
in	O
the	O
structures	O
and	O
are	O
likely	O
disordered	O
.	O
	O
The	O
quantified	O
cell	O
density	O
reflects	O
logarithmic	O
growth	O
after	O
24	O
h	O
.	O
Error	O
bars	O
are	O
the	O
s	O
.	O
d	O
.	O
for	O
three	O
replicates	O
of	O
each	O
strain	O
(	O
b	O
)	O
The	O
strains	O
used	O
in	O
a	O
were	O
also	O
serially	O
diluted	O
and	O
spotted	O
onto	O
minimal	O
agar	O
plates	O
containing	O
glutamate	O
(	O
0	O
.	O
1	O
%)	O
or	O
ammonium	O
sulphate	O
(	O
1	O
mM	O
),	O
and	O
grown	O
for	O
3	O
days	O
at	O
30	O
°	O
C	O
.	O
	O
Structural	O
differences	O
between	O
Mep2	O
and	O
bacterial	O
ammonium	O
transporters	O
.	O
	O
(	O
a	O
)	O
ICL1	O
in	O
AfAmt	O
-	O
1	O
(	O
light	O
blue	O
)	O
and	O
CaMep2	O
(	O
dark	O
blue	O
),	O
showing	O
unwinding	O
and	O
inward	O
movement	O
in	O
the	O
fungal	O
protein	O
.	O
(	O
b	O
)	O
Stereo	O
diagram	O
viewed	O
from	O
the	O
cytosol	O
of	O
ICL1	O
,	O
ICL3	O
(	O
green	O
)	O
and	O
the	O
CTR	O
(	O
red	O
)	O
in	O
AfAmt	O
-	O
1	O
(	O
light	O
colours	O
)	O
and	O
CaMep2	O
(	O
dark	O
colours	O
).	O
	O
The	O
side	O
chains	O
of	O
residues	O
in	O
the	O
RxK	O
motif	O
as	O
well	O
as	O
those	O
of	O
Tyr49	O
and	O
His342	O
are	O
labelled	O
.	O
	O
(	O
c	O
)	O
Conserved	O
residues	O
in	O
ICL1	O
-	O
3	O
and	O
the	O
CTR	O
.	O
	O
Views	O
from	O
the	O
cytosol	O
for	O
CaMep2	O
(	O
left	O
)	O
and	O
AfAmt	O
-	O
1	O
,	O
highlighting	O
the	O
large	O
differences	O
in	O
conformation	O
of	O
the	O
conserved	O
residues	O
in	O
ICL1	O
(	O
RxK	O
motif	O
;	O
blue	O
),	O
ICL2	O
(	O
ER	O
motif	O
;	O
cyan	O
),	O
ICL3	O
(	O
green	O
)	O
and	O
the	O
CTR	O
(	O
red	O
).	O
	O
The	O
labelled	O
residues	O
are	O
analogous	O
within	O
both	O
structures	O
.	O
	O
In	O
b	O
and	O
c	O
,	O
the	O
centre	O
of	O
the	O
trimer	O
is	O
at	O
top	O
.	O
	O
(	O
a	O
)	O
Stereo	O
superposition	O
of	O
AfAmt	O
-	O
1	O
and	O
CaMep2	O
showing	O
the	O
residues	O
of	O
the	O
Phe	O
gate	O
,	O
His2	O
of	O
the	O
twin	O
-	O
His	O
motif	O
and	O
the	O
tyrosine	O
residue	O
Y49	O
in	O
TM1	O
that	O
forms	O
a	O
hydrogen	O
bond	O
with	O
His2	O
in	O
CaMep2	O
.	O
(	O
b	O
)	O
Surface	O
views	O
from	O
the	O
side	O
in	O
rainbow	O
colouring	O
,	O
showing	O
the	O
two	O
-	O
tier	O
channel	O
block	O
(	O
indicated	O
by	O
the	O
arrows	O
)	O
in	O
CaMep2	O
.	O
	O
The	O
Npr1	O
kinase	O
target	O
Ser453	O
is	O
dephosphorylated	O
and	O
located	O
in	O
an	O
electronegative	O
pocket	O
.	O
	O
(	O
a	O
)	O
Stereoviews	O
of	O
CaMep2	O
showing	O
2Fo	O
–	O
Fc	O
electron	O
density	O
(	O
contoured	O
at	O
1	O
.	O
0	O
σ	O
)	O
for	O
CTR	O
residues	O
Asp419	O
-	O
Met422	O
and	O
for	O
Tyr446	O
-	O
Thr455	O
of	O
the	O
AI	O
region	O
.	O
	O
The	O
phosphorylation	O
target	O
residue	O
Ser453	O
is	O
labelled	O
in	O
bold	O
.	O
	O
(	O
c	O
)	O
Cytoplasmic	O
view	O
of	O
the	O
Mep2	O
trimer	O
indicating	O
the	O
large	O
distance	O
between	O
Ser453	O
and	O
the	O
channel	O
exits	O
(	O
circles	O
;	O
Ile52	O
lining	O
the	O
channel	O
exit	O
is	O
shown	O
).	O
	O
Effect	O
of	O
removal	O
of	O
the	O
AI	O
region	O
on	O
Mep2	O
structure	O
.	O
	O
(	O
a	O
)	O
Side	O
views	O
for	O
WT	O
CaMep2	O
(	O
left	O
)	O
and	O
the	O
truncation	O
mutant	O
442Δ	B-mutant
(	O
right	O
).	O
	O
The	O
latter	O
is	O
shown	O
as	O
a	O
putty	O
model	O
according	O
to	O
B	O
-	O
factors	O
to	O
illustrate	O
the	O
disorder	O
in	O
the	O
protein	O
on	O
the	O
cytoplasmic	O
side	O
.	O
	O
Phosphorylation	O
causes	O
conformational	O
changes	O
in	O
the	O
CTR	O
.	O
	O
(	O
a	O
)	O
Cytoplasmic	O
view	O
of	O
the	O
DD	B-mutant
mutant	I-mutant
trimer	O
,	O
with	O
WT	O
CaMep2	O
superposed	O
in	O
grey	O
for	O
one	O
of	O
the	O
monomers	O
.	O
	O
The	O
AI	O
region	O
is	O
coloured	O
magenta	O
.	O
	O
(	O
b	O
)	O
Monomer	O
side	O
-	O
view	O
superposition	O
of	O
WT	O
CaMep2	O
and	O
the	O
DD	B-mutant
mutant	I-mutant
,	O
showing	O
the	O
conformational	O
change	O
and	O
disorder	O
around	O
the	O
ExxGxD	O
motif	O
.	O
	O
Side	O
chains	O
for	O
residues	O
452	O
and	O
453	O
are	O
shown	O
as	O
stick	O
models	O
.	O
	O
(	O
a	O
)	O
In	O
the	O
closed	O
,	O
non	O
-	O
phosphorylated	O
state	O
(	O
i	O
),	O
the	O
CTR	O
(	O
magenta	O
)	O
and	O
ICL3	O
(	O
green	O
)	O
are	O
far	O
apart	O
with	O
the	O
latter	O
blocking	O
the	O
intracellular	O
channel	O
exit	O
(	O
indicated	O
with	O
a	O
hatched	O
circle	O
).	O
	O
The	O
open	O
-	O
channel	O
Mep2	O
structure	O
is	O
represented	O
by	O
archaebacterial	O
Amt	O
-	O
1	O
and	O
shown	O
in	O
lighter	O
colours	O
consistent	O
with	O
Fig	O
.	O
4	O
.	O
	O
As	O
discussed	O
in	O
the	O
text	O
,	O
similar	O
structural	O
arrangements	O
may	O
occur	O
in	O
plant	O
AMTs	O
.	O
	O
In	O
this	O
case	O
however	O
,	O
the	O
open	O
channel	O
corresponds	O
to	O
the	O
non	O
-	O
phosphorylated	O
state	O
;	O
phosphorylation	O
breaks	O
the	O
CTR	O
–	O
ICL3	O
interactions	O
leading	O
to	O
channel	O
closure	O
.	O
(	O
b	O
)	O
Model	O
based	O
on	O
AMT	O
transporter	O
analogy	O
showing	O
how	O
phosphorylation	O
of	O
a	O
Mep2	O
monomer	O
might	O
allosterically	O
open	O
channels	O
in	O
the	O
entire	O
trimer	O
via	O
disruption	O
of	O
the	O
interactions	O
between	O
the	O
CTR	O
and	O
ICL3	O
of	O
a	O
neighbouring	O
monomer	O
(	O
arrow	O
).	O
	O
Structural	O
diversity	O
in	O
a	O
human	O
antibody	O
germline	O
library	O
	O
All	O
four	O
heavy	O
chains	O
of	O
the	O
antigen	O
-	O
binding	O
fragments	O
(	O
Fabs	O
)	O
have	O
the	O
same	O
complementarity	O
-	O
determining	O
region	O
(	O
CDR	O
)	O
H3	O
that	O
was	O
reported	O
in	O
an	O
earlier	O
Fab	O
structure	O
.	O
	O
The	O
structure	O
analyses	O
include	O
comparisons	O
of	O
the	O
overall	O
structures	O
,	O
canonical	O
structures	O
of	O
the	O
CDRs	O
and	O
the	O
VH	O
:	O
VL	O
packing	O
interactions	O
.	O
	O
The	O
CDR	O
conformations	O
for	O
the	O
most	O
part	O
are	O
tightly	O
clustered	O
,	O
especially	O
for	O
the	O
ones	O
with	O
shorter	O
lengths	O
.	O
	O
The	O
longer	O
CDRs	O
with	O
tandem	O
glycines	O
or	O
serines	O
have	O
more	O
conformational	O
diversity	O
than	O
the	O
others	O
.	O
	O
One	O
conclusion	O
is	O
that	O
the	O
CDR	O
H3	O
conformations	O
are	O
influenced	O
by	O
both	O
their	O
amino	O
acid	O
sequence	O
and	O
their	O
structural	O
environment	O
determined	O
by	O
the	O
heavy	O
and	O
light	O
chain	O
pairing	O
.	O
	O
The	O
stem	O
regions	O
of	O
14	O
of	O
the	O
variant	O
pairs	O
are	O
in	O
the	O
‘	O
kinked	O
’	O
conformation	O
,	O
and	O
only	O
2	O
are	O
in	O
the	O
extended	O
conformation	O
.	O
	O
The	O
packing	O
of	O
the	O
VH	O
and	O
VL	O
domains	O
is	O
consistent	O
with	O
our	O
knowledge	O
of	O
antibody	O
structure	O
,	O
and	O
the	O
tilt	O
angles	O
between	O
these	O
domains	O
cover	O
a	O
range	O
of	O
11	O
degrees	O
.	O
	O
Two	O
of	O
16	O
structures	O
showed	O
particularly	O
large	O
variations	O
in	O
the	O
tilt	O
angles	O
when	O
compared	O
with	O
the	O
other	O
pairings	O
.	O
	O
At	O
present	O
,	O
therapeutic	O
antibodies	O
are	O
the	O
largest	O
class	O
of	O
biotherapeutic	O
proteins	O
that	O
are	O
in	O
clinical	O
trials	O
.	O
	O
The	O
use	O
of	O
monoclonal	O
antibodies	O
as	O
therapeutics	O
began	O
in	O
the	O
early	O
1980s	O
,	O
and	O
their	O
composition	O
has	O
transitioned	O
from	O
murine	O
antibodies	O
to	O
generally	O
less	O
immunogenic	O
humanized	O
and	O
human	O
antibodies	O
.	O
	O
The	O
technologies	O
currently	O
used	O
to	O
obtain	O
human	O
antibodies	O
include	O
transgenic	O
mice	O
containing	O
human	O
antibody	O
repertoires	O
,	O
cloning	O
directly	O
from	O
human	O
B	O
cells	O
,	O
and	O
in	O
vitro	O
selection	O
from	O
antibody	O
libraries	O
using	O
various	O
display	O
technologies	O
.	O
	O
All	O
engineering	O
efforts	O
are	O
guided	O
by	O
our	O
understanding	O
of	O
the	O
atomic	O
structures	O
of	O
antibodies	O
.	O
	O
In	O
such	O
efforts	O
,	O
the	O
crystal	O
structure	O
of	O
the	O
specific	O
antibody	O
may	O
not	O
be	O
available	O
,	O
but	O
modeling	O
can	O
be	O
used	O
to	O
guide	O
the	O
engineering	O
efforts	O
.	O
	O
Today	O
'	O
s	O
antibody	O
modeling	O
approaches	O
,	O
which	O
normally	O
focus	O
on	O
the	O
variable	O
region	O
,	O
are	O
being	O
developed	O
by	O
the	O
application	O
of	O
structural	O
principles	O
and	O
insights	O
that	O
are	O
evolving	O
as	O
our	O
knowledge	O
of	O
antibody	O
structures	O
continues	O
to	O
expand	O
.	O
	O
Our	O
current	O
structural	O
knowledge	O
of	O
antibodies	O
is	O
based	O
on	O
a	O
multitude	O
of	O
studies	O
that	O
used	O
many	O
techniques	O
to	O
gain	O
insight	O
into	O
the	O
functional	O
and	O
structural	O
properties	O
of	O
this	O
class	O
of	O
macromolecule	O
.	O
	O
Five	O
different	O
antibody	O
isotypes	O
occur	O
,	O
IgG	O
,	O
IgD	O
,	O
IgE	O
,	O
IgA	O
and	O
IgM	O
,	O
and	O
each	O
isotype	O
has	O
a	O
unique	O
role	O
in	O
the	O
adaptive	O
immune	O
system	O
.	O
	O
IgG	O
,	O
IgD	O
and	O
IgE	O
isotypes	O
are	O
composed	O
of	O
2	O
heavy	O
chains	O
(	O
HCs	O
)	O
and	O
2	O
light	O
chains	O
(	O
LCs	O
)	O
linked	O
through	O
disulfide	O
bonds	O
,	O
while	O
IgA	O
and	O
IgM	O
are	O
double	O
and	O
quintuple	O
versions	O
of	O
antibodies	O
,	O
respectively	O
.	O
	O
These	O
multimeric	O
forms	O
are	O
linked	O
with	O
an	O
additional	O
J	O
chain	O
.	O
	O
The	O
LCs	O
that	O
associate	O
with	O
the	O
HCs	O
are	O
divided	O
into	O
2	O
functionally	O
indistinguishable	O
classes	O
,	O
κ	O
and	O
λ	O
.	O
	O
Both	O
κ	O
and	O
λ	O
polypeptide	O
chains	O
are	O
composed	O
of	O
a	O
single	O
V	O
domain	O
and	O
a	O
single	O
C	O
domain	O
.	O
	O
The	O
heavy	O
and	O
light	O
chains	O
are	O
composed	O
of	O
structural	O
domains	O
that	O
have	O
∼	O
110	O
amino	O
acid	O
residues	O
.	O
	O
All	O
immunoglobulin	O
chains	O
have	O
an	O
N	O
-	O
terminal	O
V	O
domain	O
followed	O
by	O
1	O
to	O
4	O
C	O
domains	O
,	O
depending	O
upon	O
the	O
chain	O
type	O
.	O
	O
This	O
site	O
,	O
which	O
interacts	O
with	O
the	O
antigen	O
(	O
or	O
target	O
),	O
is	O
the	O
focus	O
of	O
current	O
antibody	O
modeling	O
efforts	O
.	O
	O
This	O
interaction	O
site	O
is	O
composed	O
of	O
6	O
complementarity	O
-	O
determining	O
regions	O
(	O
CDRs	O
)	O
that	O
were	O
identified	O
in	O
early	O
antibody	O
amino	O
acid	O
sequence	O
analyses	O
to	O
be	O
hypervariable	O
in	O
nature	O
,	O
and	O
thus	O
are	O
responsible	O
for	O
the	O
sequence	O
and	O
structural	O
diversity	O
of	O
our	O
antibody	O
repertoire	O
.	O
	O
However	O
,	O
an	O
initial	O
structural	O
analysis	O
of	O
the	O
combining	O
sites	O
of	O
the	O
small	O
set	O
of	O
structures	O
of	O
immunoglobulin	O
fragments	O
available	O
in	O
the	O
1980s	O
found	O
that	O
5	O
of	O
the	O
6	O
hypervariable	O
loops	O
or	O
CDRs	O
had	O
canonical	O
structures	O
(	O
a	O
limited	O
set	O
of	O
main	O
-	O
chain	O
conformations	O
).	O
	O
A	O
CDR	O
canonical	O
structure	O
is	O
defined	O
by	O
its	O
length	O
and	O
conserved	O
residues	O
located	O
in	O
the	O
hypervariable	O
loop	O
and	O
framework	O
residues	O
(	O
V	O
-	O
region	O
residues	O
that	O
are	O
not	O
part	O
of	O
the	O
CDRs	O
).	O
	O
Furthermore	O
,	O
studies	O
of	O
antibody	O
sequences	O
revealed	O
that	O
the	O
total	O
number	O
of	O
canonical	O
structures	O
are	O
limited	O
for	O
each	O
CDR	O
,	O
indicating	O
possibly	O
that	O
antigen	O
recognition	O
may	O
be	O
affected	O
by	O
structural	O
restrictions	O
at	O
the	O
antigen	O
-	O
binding	O
site	O
.	O
	O
Additional	O
efforts	O
have	O
led	O
to	O
our	O
current	O
understanding	O
that	O
the	O
LC	O
CDRs	O
L1	O
,	O
L2	O
,	O
and	O
L3	O
have	O
preferred	O
sets	O
of	O
canonical	O
structures	O
based	O
on	O
length	O
and	O
amino	O
acid	O
sequence	O
composition	O
.	O
	O
This	O
was	O
also	O
found	O
to	O
be	O
the	O
case	O
for	O
the	O
H1	O
and	O
H2	O
CDRs	O
.	O
	O
Classification	O
schemes	O
for	O
the	O
canonical	O
structures	O
of	O
these	O
5	O
CDRs	O
have	O
emerged	O
and	O
evolved	O
as	O
the	O
number	O
of	O
depositions	O
in	O
the	O
Protein	O
Data	O
Bank	O
of	O
Fab	O
fragments	O
of	O
antibodies	O
grow	O
.	O
	O
Recently	O
,	O
a	O
comprehensive	O
CDR	O
classification	O
scheme	O
was	O
reported	O
identifying	O
72	O
clusters	O
of	O
conformations	O
observed	O
in	O
antibody	O
structures	O
.	O
	O
The	O
knowledge	O
and	O
predictability	O
of	O
these	O
CDR	O
canonical	O
structures	O
have	O
greatly	O
advanced	O
antibody	O
modeling	O
efforts	O
.	O
	O
The	O
cataloging	O
and	O
development	O
of	O
the	O
rules	O
for	O
predicting	O
the	O
conformation	O
of	O
the	O
anchor	O
region	O
of	O
CDR	O
H3	O
continue	O
to	O
be	O
refined	O
,	O
producing	O
new	O
insight	O
into	O
the	O
CDR	O
H3	O
conformations	O
and	O
new	O
tools	O
for	O
antibody	O
engineering	O
.	O
	O
Recent	O
antibody	O
modeling	O
assessments	O
show	O
continued	O
improvement	O
in	O
the	O
quality	O
of	O
the	O
models	O
being	O
generated	O
by	O
a	O
variety	O
of	O
modeling	O
methods	O
.	O
	O
Although	O
antibody	O
modeling	O
is	O
improving	O
,	O
the	O
latest	O
assessment	O
revealed	O
a	O
number	O
of	O
challenges	O
that	O
need	O
to	O
be	O
overcome	O
to	O
provide	O
accurate	O
3	O
-	O
dimensional	O
models	O
of	O
antibody	O
V	O
regions	O
,	O
including	O
accuracies	O
in	O
the	O
modeling	O
of	O
CDR	O
H3	O
.	O
	O
The	O
need	O
for	O
improvement	O
in	O
this	O
area	O
was	O
also	O
highlighted	O
in	O
a	O
recent	O
study	O
reporting	O
an	O
approach	O
and	O
results	O
that	O
may	O
influence	O
future	O
antibody	O
modeling	O
efforts	O
.	O
	O
One	O
important	O
finding	O
of	O
the	O
antibody	O
modeling	O
assessments	O
was	O
that	O
errors	O
in	O
the	O
structural	O
templates	O
that	O
are	O
used	O
as	O
the	O
basis	O
for	O
homology	O
models	O
can	O
propagate	O
into	O
the	O
final	O
models	O
,	O
producing	O
inaccuracies	O
that	O
may	O
negatively	O
influence	O
the	O
predictive	O
nature	O
of	O
the	O
V	O
region	O
model	O
.	O
	O
This	O
Fab	O
library	O
is	O
composed	O
of	O
3	O
HC	O
germlines	O
,	O
IGHV1	B-mutant
-	I-mutant
69	I-mutant
(	O
H1	B-mutant
-	I-mutant
69	I-mutant
),	O
IGHV3	B-mutant
-	I-mutant
23	I-mutant
(	O
H3	B-mutant
-	I-mutant
23	I-mutant
)	O
and	O
IGHV5	B-mutant
-	I-mutant
51	I-mutant
(	O
H5	B-mutant
-	I-mutant
51	I-mutant
),	O
and	O
4	O
LC	O
germlines	O
(	O
all	O
κ	O
),	O
IGKV1	B-mutant
-	I-mutant
39	I-mutant
(	O
L1	B-mutant
-	I-mutant
39	I-mutant
),	O
IGKV3	B-mutant
-	I-mutant
11	I-mutant
(	O
L3	B-mutant
-	I-mutant
11	I-mutant
),	O
IGKV3	B-mutant
-	I-mutant
20	I-mutant
(	O
L3	B-mutant
-	I-mutant
20	I-mutant
)	O
and	O
IGKV4	B-mutant
-	I-mutant
1	I-mutant
(	O
L4	B-mutant
-	I-mutant
1	I-mutant
).	O
	O
Selection	O
of	O
these	O
genes	O
was	O
based	O
on	O
the	O
high	O
frequency	O
of	O
their	O
use	O
and	O
their	O
cognate	O
canonical	O
structures	O
that	O
were	O
found	O
binding	O
to	O
peptides	O
and	O
proteins	O
,	O
as	O
well	O
as	O
their	O
ability	O
to	O
be	O
expressed	O
in	O
bacteria	O
and	O
displayed	O
on	O
filamentous	O
phage	O
.	O
	O
The	O
implementation	O
of	O
the	O
library	O
involves	O
the	O
diversification	O
of	O
the	O
human	O
germline	O
genes	O
to	O
mimic	O
that	O
found	O
in	O
natural	O
human	O
libraries	O
.	O
	O
The	O
crystal	O
structure	O
determinations	O
and	O
structural	O
analyses	O
of	O
all	O
germline	O
Fabs	O
in	O
the	O
library	O
described	O
above	O
along	O
with	O
the	O
structures	O
of	O
a	O
fourth	O
HC	O
germline	O
,	O
IGHV3	B-mutant
-	I-mutant
53	I-mutant
(	O
H3	B-mutant
-	I-mutant
53	I-mutant
),	O
paired	O
with	O
the	O
4	O
LCs	O
of	O
the	O
library	O
have	O
been	O
carried	O
out	O
to	O
support	O
antibody	O
therapeutic	O
development	O
.	O
	O
All	O
16	O
HCs	O
of	O
the	O
Fabs	O
have	O
the	O
same	O
CDR	O
H3	O
that	O
was	O
reported	O
in	O
an	O
earlier	O
Fab	O
structure	O
.	O
	O
This	O
is	O
the	O
first	O
systematic	O
study	O
of	O
the	O
same	O
VH	O
and	O
VL	O
structures	O
in	O
the	O
context	O
of	O
different	O
pairings	O
.	O
	O
The	O
structure	O
analyses	O
include	O
comparisons	O
of	O
the	O
overall	O
structures	O
,	O
canonical	O
structures	O
of	O
the	O
L1	O
,	O
L2	O
,	O
L3	O
,	O
H1	O
and	O
H2	O
CDRs	O
,	O
the	O
structures	O
of	O
all	O
CDR	O
H3s	O
,	O
and	O
the	O
VH	O
:	O
VL	O
packing	O
interactions	O
.	O
	O
The	O
structures	O
and	O
their	O
analyses	O
provide	O
a	O
foundation	O
for	O
future	O
antibody	O
engineering	O
and	O
structure	O
determination	O
efforts	O
.	O
	O
Crystal	O
structures	O
	O
Crystal	O
data	O
,	O
X	O
-	O
ray	O
data	O
,	O
and	O
refinement	O
statistics	O
.	O
	O
(	O
Continued	O
)	O
Crystal	O
data	O
,	O
X	O
-	O
ray	O
data	O
,	O
and	O
refinement	O
statistics	O
.	O
	O
The	O
crystal	O
structures	O
of	O
a	O
germline	O
library	O
composed	O
of	O
16	O
Fabs	O
generated	O
by	O
combining	O
4	O
HCs	O
(	O
H1	B-mutant
-	I-mutant
69	I-mutant
,	O
H3	B-mutant
-	I-mutant
23	I-mutant
,	O
H3	B-mutant
-	I-mutant
53	I-mutant
and	O
H5	B-mutant
-	I-mutant
51	I-mutant
)	O
and	O
4	O
LCs	O
(	O
L1	B-mutant
-	I-mutant
39	I-mutant
,	O
L3	B-mutant
-	I-mutant
11	I-mutant
,	O
L3	B-mutant
-	I-mutant
20	I-mutant
and	O
L4	B-mutant
-	I-mutant
1	I-mutant
)	O
have	O
been	O
determined	O
.	O
	O
The	O
Fab	O
heavy	O
and	O
light	O
chain	O
sequences	O
for	O
the	O
variants	O
numbered	O
according	O
to	O
Chothia	O
are	O
shown	O
in	O
Fig	O
.	O
S1	O
.	O
	O
The	O
four	O
different	O
HCs	O
all	O
have	O
the	O
same	O
CDR	O
H3	O
sequence	O
,	O
ARYDGIYGELDF	O
.	O
	O
These	O
include	O
(	O
1	O
)	O
H3	O
-	O
23	O
:	O
L3	O
-	O
11	O
and	O
H3	O
-	O
23	O
:	O
L4	O
-	O
1	O
in	O
P212121	O
,	O
(	O
2	O
)	O
H3	O
-	O
53	O
:	O
L1	O
-	O
39	O
,	O
H3	O
-	O
53	O
:	O
L3	O
-	O
11	O
and	O
H3	O
-	O
53	O
:	O
L3	O
-	O
20	O
in	O
P6522	O
,	O
and	O
(	O
3	O
)	O
H5	O
-	O
51	O
:	O
L1	O
-	O
39	O
,	O
H5	O
-	O
51	O
:	O
L3	O
-	O
11	O
and	O
H5	O
-	O
51	O
:	O
L3	O
-	O
20	O
in	O
P212121	O
.	O
	O
The	O
similarity	O
in	O
the	O
crystal	O
forms	O
is	O
attributed	O
in	O
part	O
to	O
cross	O
-	O
seeding	O
using	O
the	O
microseed	O
matrix	O
screening	O
for	O
groups	O
2	O
and	O
3	O
.	O
	O
The	O
number	O
of	O
Fab	O
molecules	O
in	O
the	O
crystallographic	O
asymmetric	O
unit	O
varies	O
from	O
1	O
(	O
for	O
12	O
Fabs	O
)	O
to	O
2	O
(	O
for	O
4	O
Fabs	O
).	O
	O
Invariably	O
,	O
the	O
HCs	O
have	O
more	O
disorder	O
than	O
the	O
LCs	O
.	O
	O
For	O
the	O
LC	O
,	O
the	O
disorder	O
is	O
observed	O
at	O
2	O
of	O
the	O
C	O
-	O
terminal	O
residues	O
with	O
few	O
exceptions	O
.	O
	O
The	O
C	O
-	O
terminal	O
residues	O
including	O
the	O
6xHis	O
tags	O
are	O
disordered	O
in	O
all	O
16	O
structures	O
.	O
	O
In	O
addition	O
to	O
these	O
,	O
2	O
primary	O
disordered	O
stretches	O
of	O
residues	O
are	O
observed	O
in	O
a	O
number	O
of	O
structures	O
(	O
Table	O
S1	O
).	O
	O
One	O
involves	O
the	O
loop	O
connecting	O
the	O
first	O
2	O
β	O
-	O
strands	O
of	O
the	O
constant	O
domain	O
(	O
in	O
all	O
Fabs	O
except	O
H3	O
-	O
23	O
:	O
L1	O
-	O
39	O
,	O
H3	O
-	O
23	O
:	O
L3	O
-	O
11	O
and	O
H3	O
-	O
53	O
:	O
L1	O
-	O
39	O
).	O
	O
The	O
other	O
is	O
located	O
in	O
CDR	O
H3	O
(	O
in	O
H5	O
-	O
51	O
:	O
L3	O
-	O
11	O
,	O
H5	O
-	O
51	O
:	O
L3	O
-	O
20	O
and	O
in	O
one	O
of	O
2	O
copies	O
of	O
H3	O
-	O
23	O
:	O
L4	O
-	O
1	O
).	O
	O
CDR	O
canonical	O
structures	O
	O
Several	O
CDR	O
definitions	O
have	O
evolved	O
over	O
decades	O
of	O
antibody	O
research	O
.	O
	O
Depending	O
on	O
the	O
focus	O
of	O
the	O
study	O
,	O
the	O
CDR	O
boundaries	O
differ	O
slightly	O
between	O
various	O
definitions	O
.	O
	O
In	O
this	O
work	O
,	O
we	O
use	O
the	O
CDR	O
definition	O
of	O
North	O
et	O
al	O
.,	O
which	O
is	O
similar	O
to	O
that	O
of	O
Martin	O
with	O
the	O
following	O
exceptions	O
:	O
1	O
)	O
CDRs	O
H1	O
and	O
H3	O
begin	O
immediately	O
after	O
the	O
Cys	O
;	O
and	O
2	O
)	O
CDR	O
L2	O
includes	O
an	O
additional	O
residue	O
at	O
the	O
N	O
-	O
terminal	O
side	O
,	O
typically	O
Tyr	O
.	O
	O
CDR	O
H1	O
	O
The	O
superposition	O
of	O
CDR	O
H1	O
backbones	O
for	O
all	O
HC	O
:	O
LC	O
pairs	O
with	O
heavy	O
chains	O
:	O
(	O
A	O
)	O
H1	B-mutant
-	I-mutant
69	I-mutant
,	O
(	O
B	O
)	O
H3	B-mutant
-	I-mutant
23	I-mutant
,	O
(	O
C	O
)	O
H3	B-mutant
-	I-mutant
53	I-mutant
and	O
(	O
D	O
)	O
H5	B-mutant
-	I-mutant
51	I-mutant
.	O
	O
Assignments	O
for	O
2	O
copies	O
of	O
the	O
Fab	O
in	O
the	O
asymmetric	O
unit	O
are	O
given	O
for	O
5	O
structures	O
.	O
	O
No	O
assignment	O
(	O
NA	O
)	O
for	O
CDRs	O
with	O
missing	O
residues	O
.	O
	O
The	O
four	O
HCs	O
feature	O
CDR	O
H1	O
of	O
the	O
same	O
length	O
,	O
and	O
their	O
sequences	O
are	O
highly	O
similar	O
(	O
Table	O
2	O
).	O
	O
The	O
CDR	O
H1	O
backbone	O
conformations	O
for	O
all	O
variants	O
for	O
each	O
of	O
the	O
HCs	O
are	O
shown	O
in	O
Fig	O
.	O
1	O
.	O
	O
Some	O
deviation	O
is	O
observed	O
for	O
H3	B-mutant
-	I-mutant
53	I-mutant
,	O
mostly	O
due	O
to	O
H3	O
-	O
53	O
:	O
L4	O
-	O
1	O
,	O
which	O
exhibits	O
a	O
significant	O
degree	O
of	O
disorder	O
in	O
CDR	O
H1	O
.	O
	O
The	O
electron	O
density	O
for	O
the	O
backbone	O
is	O
weak	O
and	O
discontinuous	O
,	O
and	O
completely	O
missing	O
for	O
several	O
side	O
chains	O
.	O
	O
The	O
CDR	O
H1	O
structures	O
with	O
H1	B-mutant
-	I-mutant
69	I-mutant
shown	O
in	O
Fig	O
.	O
1A	O
are	O
quite	O
variable	O
,	O
both	O
for	O
the	O
structures	O
with	O
different	O
LCs	O
and	O
for	O
the	O
copies	O
of	O
the	O
same	O
Fab	O
in	O
the	O
asymmetric	O
unit	O
,	O
H1	O
-	O
69	O
:	O
L3	O
-	O
11	O
and	O
H1	O
-	O
69	O
:	O
L3	O
-	O
20	O
.	O
	O
In	O
total	O
,	O
6	O
independent	O
Fab	O
structures	O
produce	O
5	O
different	O
canonical	O
structures	O
,	O
namely	O
H1	B-mutant
-	I-mutant
13	I-mutant
-	I-mutant
1	I-mutant
,	O
H1	B-mutant
-	I-mutant
13	I-mutant
-	I-mutant
3	I-mutant
,	O
H1	B-mutant
-	I-mutant
13	I-mutant
-	I-mutant
4	I-mutant
,	O
H1	B-mutant
-	I-mutant
13	I-mutant
-	I-mutant
6	I-mutant
and	O
H1	B-mutant
-	I-mutant
13	I-mutant
-	I-mutant
10	I-mutant
.	O
	O
A	O
major	O
difference	O
of	O
H1	B-mutant
-	I-mutant
69	I-mutant
from	O
the	O
other	O
germlines	O
in	O
the	O
experimental	O
data	O
set	O
is	O
the	O
presence	O
of	O
Gly	O
instead	O
of	O
Phe	O
or	O
Tyr	O
at	O
position	O
27	O
(	O
residue	O
5	O
of	O
13	O
in	O
CDR	O
H1	O
).	O
	O
Glycine	O
introduces	O
the	O
possibility	O
of	O
a	O
higher	O
degree	O
of	O
conformational	O
flexibility	O
that	O
undoubtedly	O
translates	O
to	O
the	O
differences	O
observed	O
,	O
and	O
contributes	O
to	O
the	O
elevated	O
thermal	O
parameters	O
for	O
the	O
atoms	O
in	O
the	O
amino	O
acid	O
residues	O
in	O
this	O
region	O
.	O
	O
The	O
superposition	O
of	O
CDR	O
H2	O
backbones	O
for	O
all	O
HC	O
:	O
LC	O
pairs	O
with	O
heavy	O
chains	O
:	O
(	O
A	O
)	O
H1	B-mutant
-	I-mutant
69	I-mutant
,	O
(	O
B	O
)	O
H3	B-mutant
-	I-mutant
23	I-mutant
,	O
(	O
C	O
)	O
H3	B-mutant
-	I-mutant
53	I-mutant
and	O
(	O
D	O
)	O
H5	B-mutant
-	I-mutant
51	I-mutant
.	O
	O
The	O
canonical	O
structures	O
of	O
CDR	O
H2	O
have	O
fairly	O
consistent	O
conformations	O
(	O
Table	O
2	O
,	O
Fig	O
.	O
2	O
).	O
	O
The	O
conformations	O
for	O
all	O
of	O
these	O
CDR	O
H2s	O
are	O
tightly	O
clustered	O
(	O
Fig	O
.	O
2	O
).	O
	O
In	O
one	O
case	O
,	O
in	O
the	O
second	O
Fab	O
of	O
H1	O
-	O
69	O
:	O
L3	O
-	O
20	O
,	O
CDR	O
H2	O
is	O
partially	O
disordered	O
(	O
Δ55	B-mutant
-	I-mutant
60	I-mutant
).	O
	O
Although	O
three	O
of	O
the	O
germlines	O
have	O
CDR	O
H2	O
of	O
the	O
same	O
length	O
,	O
10	O
residues	O
,	O
they	O
adopt	O
2	O
distinctively	O
different	O
conformations	O
depending	O
mostly	O
on	O
the	O
residue	O
at	O
position	O
71	O
from	O
the	O
so	O
-	O
called	O
CDR	O
H4	O
.	O
	O
Arg71	O
in	O
H3	B-mutant
-	I-mutant
23	I-mutant
fills	O
the	O
space	O
between	O
CDRs	O
H2	O
and	O
H4	O
,	O
and	O
defines	O
the	O
conformation	O
of	O
the	O
tip	O
of	O
CDR	O
H2	O
so	O
that	O
residue	O
54	O
points	O
away	O
from	O
the	O
antigen	O
binding	O
site	O
.	O
	O
Conformations	O
of	O
CDR	O
H2	O
in	O
H1	B-mutant
-	I-mutant
69	I-mutant
and	O
H5	B-mutant
-	I-mutant
51	I-mutant
,	O
both	O
of	O
which	O
have	O
canonical	O
structure	O
H2	B-mutant
-	I-mutant
10	I-mutant
-	I-mutant
1	I-mutant
,	O
show	O
little	O
deviation	O
within	O
each	O
set	O
of	O
4	O
structures	O
.	O
	O
CDR	O
L1	O
	O
The	O
superposition	O
of	O
CDR	O
L1	O
backbones	O
for	O
all	O
HC	O
:	O
LC	O
pairs	O
with	O
light	O
chains	O
:	O
(	O
A	O
)	O
L1	B-mutant
-	I-mutant
39	I-mutant
,	O
(	O
B	O
)	O
L3	B-mutant
-	I-mutant
11	I-mutant
,	O
(	O
C	O
)	O
L3	B-mutant
-	I-mutant
20	I-mutant
and	O
(	O
D	O
)	O
L4	B-mutant
-	I-mutant
1	I-mutant
.	O
	O
Of	O
these	O
LCs	O
,	O
L1	B-mutant
-	I-mutant
39	I-mutant
and	O
L3	B-mutant
-	I-mutant
11	I-mutant
have	O
the	O
same	O
canonical	O
structure	O
,	O
L1	B-mutant
-	I-mutant
11	I-mutant
-	I-mutant
1	I-mutant
,	O
and	O
superimpose	O
very	O
well	O
(	O
Fig	O
.	O
3A	O
,	O
B	O
).	O
	O
L4	B-mutant
-	I-mutant
1	I-mutant
has	O
the	O
longest	O
CDR	O
L1	O
,	O
composed	O
of	O
17	O
amino	O
acid	O
residues	O
(	O
Fig	O
.	O
3D	O
).	O
	O
Despite	O
this	O
,	O
the	O
conformations	O
are	O
tightly	O
clustered	O
(	O
rmsd	O
is	O
0	O
.	O
20	O
Å	O
).	O
	O
The	O
backbone	O
conformations	O
of	O
the	O
stem	O
regions	O
superimpose	O
well	O
.	O
	O
Some	O
changes	O
in	O
conformation	O
occur	O
between	O
residues	O
30a	O
and	O
30f	O
(	O
residues	O
8	O
and	O
13	O
of	O
17	O
in	O
CDR	O
L1	O
).	O
	O
This	O
is	O
the	O
tip	O
of	O
the	O
loop	O
region	O
,	O
which	O
appears	O
to	O
have	O
similar	O
conformations	O
that	O
fan	O
out	O
the	O
structures	O
because	O
of	O
the	O
slight	O
differences	O
in	O
torsion	O
angles	O
in	O
the	O
backbone	O
near	O
Tyr30a	O
and	O
Lys30f	O
.	O
	O
The	O
conformation	O
of	O
CDR	O
L1	O
in	O
these	O
2	O
Fabs	O
is	O
slightly	O
different	O
,	O
and	O
both	O
conformations	O
fall	O
somewhere	O
between	O
L1	B-mutant
-	I-mutant
12	I-mutant
-	I-mutant
1	I-mutant
and	O
L1	B-mutant
-	I-mutant
12	I-mutant
-	I-mutant
2	I-mutant
.	O
	O
This	O
reflects	O
the	O
lack	O
of	O
accuracy	O
in	O
the	O
structure	O
due	O
to	O
low	O
resolution	O
of	O
the	O
X	O
-	O
ray	O
data	O
(	O
3	O
.	O
3	O
Å	O
).	O
	O
CDR	O
L2	O
	O
The	O
superposition	O
of	O
CDR	O
L2	O
backbones	O
for	O
all	O
HC	O
:	O
LC	O
pairs	O
with	O
light	O
chains	O
:	O
(	O
A	O
)	O
L1	B-mutant
-	I-mutant
39	I-mutant
,	O
(	O
B	O
)	O
L3	B-mutant
-	I-mutant
11	I-mutant
,	O
(	O
C	O
)	O
L3	B-mutant
-	I-mutant
20	I-mutant
and	O
(	O
D	O
)	O
L4	B-mutant
-	I-mutant
1	I-mutant
.	O
	O
All	O
four	O
LCs	O
have	O
CDR	O
L2	O
of	O
the	O
same	O
length	O
and	O
canonical	O
structure	O
,	O
L2	B-mutant
-	I-mutant
8	I-mutant
-	I-mutant
1	I-mutant
(	O
Table	O
2	O
).	O
	O
The	O
CDR	O
L2	O
conformations	O
for	O
each	O
of	O
the	O
LCs	O
paired	O
with	O
the	O
4	O
HCs	O
are	O
clustered	O
more	O
tightly	O
than	O
any	O
of	O
the	O
other	O
CDRs	O
(	O
rmsd	O
values	O
are	O
in	O
the	O
range	O
0	O
.	O
09	O
-	O
0	O
.	O
16	O
Å	O
),	O
and	O
all	O
4	O
sets	O
have	O
virtually	O
the	O
same	O
conformation	O
despite	O
the	O
sequence	O
diversity	O
of	O
the	O
loop	O
.	O
	O
CDR	O
L3	O
	O
The	O
superposition	O
of	O
CDR	O
L3	O
backbones	O
for	O
all	O
HC	O
:	O
LC	O
pairs	O
with	O
light	O
chains	O
:	O
(	O
A	O
)	O
L1	B-mutant
-	I-mutant
39	I-mutant
,	O
(	O
B	O
)	O
L3	B-mutant
-	I-mutant
11	I-mutant
,	O
(	O
C	O
)	O
L3	B-mutant
-	I-mutant
20	I-mutant
and	O
(	O
D	O
)	O
L4	B-mutant
-	I-mutant
1	I-mutant
.	O
	O
The	O
conformations	O
of	O
CDR	O
L3	O
for	O
L1	B-mutant
-	I-mutant
39	I-mutant
,	O
L3	B-mutant
-	I-mutant
11	I-mutant
,	O
and	O
particularly	O
for	O
L320	O
,	O
are	O
not	O
as	O
tightly	O
clustered	O
as	O
those	O
of	O
L4	B-mutant
-	I-mutant
1	I-mutant
(	O
Fig	O
.	O
5	O
).	O
	O
The	O
slight	O
conformational	O
variability	O
occurs	O
in	O
the	O
region	O
of	O
amino	O
acid	O
residues	O
90	O
-	O
92	O
,	O
which	O
is	O
in	O
contact	O
with	O
CDR	O
H3	O
.	O
	O
The	O
loop	O
and	O
the	O
2	O
β	O
-	O
strands	O
of	O
the	O
CDR	O
H3	O
in	O
this	O
‘	O
parent	O
’	O
structure	O
are	O
stabilized	O
by	O
H	O
-	O
bonds	O
between	O
the	O
carbonyl	O
oxygen	O
and	O
peptide	O
nitrogen	O
atoms	O
in	O
the	O
2	O
strands	O
.	O
	O
An	O
interesting	O
feature	O
of	O
these	O
CDR	O
H3	O
structures	O
is	O
the	O
presence	O
of	O
a	O
water	O
molecule	O
that	O
interacts	O
with	O
the	O
peptide	O
nitrogens	O
and	O
carbonyl	O
oxygens	O
near	O
the	O
bridging	O
loop	O
connecting	O
the	O
2	O
β	O
-	O
strands	O
.	O
	O
This	O
water	O
is	O
present	O
in	O
both	O
the	O
bound	O
(	O
4DN4	O
)	O
and	O
unbound	O
(	O
4DN3	O
)	O
forms	O
of	O
CNTO	O
888	O
.	O
	O
The	O
stem	O
region	O
of	O
CDR	O
H3	O
in	O
the	O
parental	O
Fab	O
is	O
in	O
a	O
‘	O
kinked	O
’	O
conformation	O
,	O
in	O
which	O
the	O
indole	O
nitrogen	O
of	O
Trp103	O
forms	O
a	O
hydrogen	O
bond	O
with	O
the	O
carbonyl	O
oxygen	O
of	O
Leu100b	O
.	O
	O
The	O
carboxyl	O
group	O
of	O
Asp101	O
forms	O
a	O
salt	O
bridge	O
with	O
Arg94	O
.	O
	O
Ribbon	O
representations	O
of	O
(	O
A	O
)	O
the	O
superposition	O
of	O
all	O
CDR	O
H3s	O
of	O
the	O
structures	O
with	O
complete	O
backbone	O
traces	O
.	O
(	O
B	O
)	O
The	O
CDR	O
H3s	O
rotated	O
90	O
°	O
about	O
the	O
y	O
axis	O
of	O
the	O
page	O
.	O
	O
The	O
structure	O
of	O
each	O
CDR	O
H3	O
is	O
represented	O
with	O
a	O
different	O
color	O
.	O
	O
Three	O
of	O
the	O
21	O
Fab	O
structures	O
(	O
including	O
multiple	O
copies	O
in	O
the	O
asymmetric	O
unit	O
),	O
H5	O
-	O
51	O
:	O
L3	O
-	O
11	O
,	O
H551	O
:	O
L3	O
-	O
20	O
and	O
H3	O
-	O
23	O
:	O
L4	O
-	O
1	O
(	O
one	O
of	O
the	O
2	O
Fabs	O
),	O
have	O
missing	O
(	O
disordered	O
)	O
residues	O
at	O
the	O
apex	O
of	O
the	O
CDR	O
loop	O
.	O
	O
A	O
comparison	O
of	O
representatives	O
of	O
the	O
“	O
kinked	O
”	O
and	O
“	O
extended	O
”	O
structures	O
.	O
	O
(	O
A	O
)	O
The	O
“	O
kinked	O
”	O
CDR	O
H3	O
of	O
H1	O
-	O
69	O
:	O
L3	O
-	O
11	O
with	O
purple	O
carbon	O
atoms	O
and	O
yellow	O
dashed	O
lines	O
connecting	O
the	O
H	O
-	O
bond	O
pairs	O
for	O
Leu100b	O
O	O
and	O
Trp103	O
NE1	O
,	O
Arg94	O
NE	O
and	O
Asp101	O
OD1	O
,	O
and	O
Arg94	O
NH2	O
and	O
Asp101	O
OD2	O
.	O
	O
In	O
10	O
of	O
the	O
18	O
Fab	O
structures	O
,	O
H1	O
-	O
69	O
:	O
L1	O
-	O
39	O
,	O
H1	O
-	O
69	O
:	O
L3	O
-	O
11	O
(	O
2	O
Fabs	O
),	O
H1	O
-	O
69	O
:	O
L4	O
-	O
1	O
,	O
H3	O
-	O
23	O
:	O
L3	O
-	O
11	O
(	O
2	O
Fabs	O
),	O
H3	O
-	O
23	O
:	O
L3	O
-	O
20	O
,	O
H3	O
-	O
53	O
:	O
L3	O
-	O
11	O
,	O
H3	O
-	O
53	O
:	O
L3	O
-	O
20	O
and	O
H5	O
-	O
51	O
:	O
L1	O
-	O
39	O
,	O
the	O
CDRs	O
have	O
similar	O
conformations	O
to	O
that	O
found	O
in	O
4DN3	O
.	O
	O
The	O
bases	O
of	O
these	O
structures	O
have	O
the	O
‘	O
kinked	O
’	O
conformation	O
with	O
the	O
H	O
-	O
bond	O
between	O
Trp103	O
and	O
Leu100b	O
.	O
	O
The	O
largest	O
backbone	O
conformational	O
deviation	O
for	O
the	O
set	O
is	O
at	O
Tyr99	O
,	O
where	O
the	O
C	O
=	O
O	O
is	O
rotated	O
by	O
90	O
°	O
relative	O
to	O
that	O
observed	O
in	O
4DN3	O
.	O
	O
Also	O
,	O
it	O
is	O
worth	O
noting	O
that	O
only	O
one	O
of	O
these	O
structures	O
,	O
H1	O
-	O
69	O
:	O
L4	O
-	O
1	O
,	O
has	O
the	O
conserved	O
water	O
molecule	O
in	O
CDR	O
H3	O
observed	O
in	O
the	O
4DN3	O
and	O
4DN4	O
structures	O
.	O
	O
The	O
CDR	O
H3	O
for	O
this	O
structure	O
is	O
shown	O
in	O
Fig	O
.	O
S3	O
.	O
	O
The	O
stem	O
regions	O
in	O
these	O
3	O
cases	O
are	O
in	O
the	O
‘	O
kinked	O
’	O
conformation	O
consistent	O
with	O
that	O
observed	O
for	O
4DN3	O
.	O
	O
The	O
five	O
remaining	O
Fabs	O
,	O
H5	O
-	O
51	O
:	O
L4	O
-	O
1	O
(	O
2	O
copies	O
),	O
H1	O
-	O
69	O
:	O
L3	O
-	O
20	O
(	O
2	O
copies	O
)	O
and	O
H3	O
-	O
53	O
:	O
L4	O
-	O
1	O
,	O
have	O
3	O
different	O
CDR	O
H3	O
conformations	O
(	O
Fig	O
.	O
S4	O
).	O
	O
The	O
stem	O
regions	O
of	O
CDR	O
H3	O
for	O
the	O
H5	O
-	O
51	O
:	O
L4	O
-	O
1	O
Fabs	O
are	O
in	O
the	O
‘	O
kinked	O
’	O
conformation	O
while	O
,	O
surprisingly	O
,	O
those	O
of	O
the	O
H1	O
-	O
69	O
:	O
L3	O
-	O
20	O
pair	O
and	O
H3	O
-	O
53	O
:	O
L4	O
-	O
1	O
are	O
in	O
the	O
‘	O
extended	O
’	O
conformation	O
(	O
Fig	O
.	O
7B	O
).	O
	O
The	O
two	O
domains	O
pack	O
together	O
such	O
that	O
the	O
5	O
-	O
stranded	O
β	O
-	O
sheets	O
,	O
which	O
have	O
hydrophobic	O
surfaces	O
,	O
interact	O
with	O
each	O
other	O
bringing	O
the	O
CDRs	O
from	O
both	O
the	O
VH	O
and	O
VL	O
domains	O
into	O
close	O
proximity	O
.	O
	O
The	O
domain	O
packing	O
of	O
the	O
variants	O
was	O
assessed	O
by	O
computing	O
the	O
domain	O
interface	O
interactions	O
,	O
the	O
VH	O
:	O
VL	O
tilt	O
angles	O
,	O
the	O
buried	O
surface	O
area	O
and	O
surface	O
complementarity	O
.	O
	O
VH	O
:	O
VL	O
interface	O
amino	O
acid	O
residue	O
interactions	O
	O
The	O
conserved	O
VH	O
:	O
VL	O
interactions	O
as	O
viewed	O
along	O
the	O
VH	O
/	O
VL	O
axis	O
.	O
	O
The	O
VH	O
residues	O
are	O
in	O
blue	O
,	O
the	O
VL	O
residues	O
are	O
in	O
orange	O
.	O
	O
The	O
VH	O
:	O
VL	O
interface	O
is	O
pseudosymmetric	O
,	O
and	O
involves	O
2	O
stretches	O
of	O
the	O
polypeptide	O
chain	O
from	O
each	O
domain	O
,	O
namely	O
CDR3	O
and	O
the	O
framework	O
region	O
between	O
CDRs	O
1	O
and	O
2	O
.	O
	O
These	O
stretches	O
form	O
antiparallel	O
β	O
-	O
hairpins	O
within	O
the	O
internal	O
5	O
-	O
stranded	O
β	O
-	O
sheet	O
.	O
	O
There	O
are	O
a	O
few	O
principal	O
inter	O
-	O
domain	O
interactions	O
that	O
are	O
conserved	O
not	O
only	O
in	O
the	O
experimental	O
set	O
of	O
16	O
Fabs	O
,	O
but	O
in	O
all	O
human	O
antibodies	O
.	O
	O
These	O
core	O
interactions	O
provide	O
enough	O
stability	O
to	O
the	O
VH	O
:	O
VL	O
dimer	O
so	O
that	O
additional	O
VH	O
-	O
VL	O
contacts	O
can	O
tolerate	O
amino	O
acid	O
sequence	O
variations	O
in	O
CDRs	O
H3	O
and	O
L3	O
that	O
form	O
part	O
of	O
the	O
VH	O
:	O
VL	O
interface	O
.	O
	O
In	O
total	O
,	O
about	O
20	O
residues	O
are	O
involved	O
in	O
the	O
VH	O
:	O
VL	O
interactions	O
on	O
each	O
side	O
(	O
Fig	O
.	O
S5	O
).	O
	O
Half	O
of	O
them	O
are	O
in	O
the	O
framework	O
regions	O
and	O
those	O
residues	O
(	O
except	O
residue	O
61	O
in	O
HC	O
,	O
which	O
is	O
actually	O
in	O
CDR2	O
in	O
Kabat	O
'	O
s	O
definition	O
)	O
are	O
conserved	O
in	O
the	O
set	O
of	O
16	O
Fabs	O
.	O
	O
One	O
notable	O
exception	O
is	O
H	O
-	O
Trp47	O
,	O
which	O
exhibits	O
2	O
conformations	O
of	O
the	O
indole	O
ring	O
.	O
	O
Interestingly	O
,	O
these	O
are	O
the	O
only	O
2	O
structures	O
with	O
residues	O
missing	O
in	O
CDR	O
H3	O
because	O
of	O
disorder	O
,	O
although	O
both	O
structures	O
are	O
determined	O
at	O
high	O
resolution	O
and	O
the	O
rest	O
of	O
the	O
structure	O
is	O
well	O
defined	O
.	O
	O
Apparently	O
,	O
residues	O
flanking	O
CDR	O
H3	O
in	O
the	O
2	O
VH	O
:	O
VL	O
pairings	O
are	O
inconsistent	O
with	O
any	O
stable	O
conformation	O
of	O
CDR	O
H3	O
,	O
which	O
translates	O
into	O
a	O
less	O
restricted	O
conformational	O
space	O
for	O
some	O
of	O
them	O
,	O
including	O
H	O
-	O
Trp47	O
.	O
	O
VH	O
:	O
VL	O
tilt	O
angles	O
	O
The	O
relative	O
orientation	O
of	O
VH	O
and	O
VL	O
has	O
been	O
measured	O
in	O
a	O
number	O
of	O
different	O
ways	O
.	O
	O
The	O
four	O
LCs	O
all	O
are	O
classified	O
as	O
Type	O
A	O
because	O
they	O
have	O
a	O
proline	O
at	O
position	O
44	O
,	O
and	O
the	O
results	O
for	O
each	O
orientation	O
parameter	O
are	O
within	O
the	O
range	O
of	O
values	O
of	O
this	O
type	O
reported	O
by	O
Dunbar	O
and	O
co	O
-	O
workers	O
.	O
	O
The	O
only	O
exception	O
is	O
HC1	O
,	O
which	O
is	O
shifted	O
toward	O
smaller	O
angles	O
with	O
the	O
mean	O
value	O
of	O
70	O
.	O
8	O
°	O
as	O
compared	O
to	O
the	O
distribution	O
centered	O
at	O
72	O
°	O
for	O
the	O
entire	O
PDB	O
.	O
	O
This	O
probably	O
reflects	O
the	O
invariance	O
of	O
CDR	O
H3	O
in	O
the	O
current	O
set	O
as	O
opposed	O
to	O
the	O
CDR	O
H3	O
diversity	O
in	O
the	O
PDB	O
.	O
	O
The	O
second	O
approach	O
used	O
for	O
comparing	O
tilt	O
angles	O
involved	O
computing	O
the	O
difference	O
in	O
the	O
tilt	O
angles	O
between	O
all	O
pairs	O
of	O
structures	O
.	O
	O
The	O
differences	O
between	O
independent	O
Fabs	O
in	O
the	O
same	O
structure	O
are	O
4	O
.	O
9	O
°	O
for	O
H1	O
-	O
69	O
:	O
L3	O
-	O
20	O
,	O
1	O
.	O
6	O
°	O
for	O
H1	O
-	O
69	O
:	O
L3	O
-	O
11	O
,	O
1	O
.	O
4	O
°	O
for	O
H3	O
-	O
23	O
:	O
L4	O
-	O
1	O
,	O
3	O
.	O
3	O
°	O
for	O
H3	O
-	O
23	O
:	O
L3	O
-	O
11	O
,	O
and	O
2	O
.	O
5	O
°	O
for	O
H5	O
-	O
51	O
:	O
L4	O
-	O
1	O
.	O
	O
With	O
the	O
exception	O
of	O
H1	O
-	O
69	O
:	O
L3	O
-	O
20	O
,	O
the	O
angles	O
are	O
within	O
the	O
range	O
of	O
2	O
-	O
3	O
°	O
as	O
are	O
observed	O
in	O
the	O
identical	O
structures	O
in	O
the	O
PDB	O
.	O
	O
In	O
H1	O
-	O
69	O
:	O
L3	O
-	O
20	O
,	O
one	O
of	O
the	O
Fabs	O
is	O
substantially	O
disordered	O
so	O
that	O
part	O
of	O
CDR	O
H2	O
(	O
the	O
outer	O
β	O
-	O
strand	O
,	O
residues	O
55	O
-	O
60	O
)	O
is	O
completely	O
missing	O
.	O
	O
This	O
kind	O
of	O
disorder	O
may	O
compromise	O
the	O
integrity	O
of	O
the	O
VH	O
domain	O
and	O
its	O
interaction	O
with	O
the	O
VL	O
.	O
	O
Indeed	O
,	O
this	O
Fab	O
has	O
the	O
largest	O
twist	O
angle	O
HC2	O
within	O
the	O
experimental	O
set	O
that	O
exceeds	O
the	O
mean	O
value	O
by	O
2	O
.	O
5	O
standard	O
deviations	O
(	O
Table	O
S2	O
).	O
	O
Differences	O
in	O
VH	O
:	O
VL	O
tilt	O
angles	O
.	O
	O
The	O
differences	O
in	O
the	O
tilt	O
angle	O
are	O
shown	O
for	O
all	O
pairs	O
of	O
V	O
regions	O
in	O
Table	O
3	O
.	O
	O
The	O
smallest	O
differences	O
in	O
the	O
tilt	O
angle	O
are	O
between	O
the	O
Fabs	O
in	O
isomorphous	O
crystal	O
forms	O
.	O
	O
VH	O
:	O
VL	O
surface	O
areas	O
and	O
surface	O
complementarity	O
.	O
	O
The	O
results	O
of	O
the	O
PISA	O
contact	O
surface	O
calculation	O
and	O
surface	O
complementarity	O
calculation	O
are	O
shown	O
in	O
Table	O
4	O
.	O
	O
The	O
interface	O
areas	O
are	O
calculated	O
as	O
the	O
average	O
of	O
the	O
VH	O
and	O
VL	O
contact	O
surfaces	O
.	O
	O
Six	O
of	O
the	O
16	O
structures	O
have	O
CDR	O
H3	O
side	O
chains	O
or	O
complete	O
residues	O
missing	O
,	O
and	O
therefore	O
their	O
interfaces	O
are	O
much	O
smaller	O
than	O
in	O
the	O
other	O
10	O
structures	O
with	O
complete	O
CDRs	O
(	O
the	O
results	O
are	O
provided	O
for	O
all	O
Fabs	O
for	O
completeness	O
).	O
	O
Among	O
the	O
complete	O
structures	O
,	O
the	O
interface	O
areas	O
range	O
from	O
684	O
to	O
836	O
Å2	O
.	O
	O
Interestingly	O
,	O
the	O
2	O
structures	O
that	O
have	O
the	O
largest	O
tilt	O
angle	O
differences	O
with	O
the	O
other	O
variants	O
,	O
H3	O
-	O
23	O
:	O
L3	O
-	O
20	O
and	O
H1	O
-	O
69	O
:	O
L3	O
-	O
20	O
,	O
have	O
the	O
smallest	O
VH	O
:	O
VL	O
interfaces	O
,	O
684	O
and	O
725	O
Å2	O
,	O
respectively	O
.	O
	O
H3	O
-	O
23	O
:	O
L3	O
-	O
20	O
is	O
also	O
unique	O
in	O
that	O
it	O
has	O
the	O
lowest	O
value	O
(	O
0	O
.	O
676	O
)	O
of	O
surface	O
complementarity	O
.	O
	O
Melting	O
temperatures	O
for	O
the	O
16	O
Fabs	O
.	O
	O
Colors	O
:	O
blue	O
(	O
Tm	O
<	O
70	O
°	O
C	O
),	O
green	O
(	O
70	O
°	O
C	O
<	O
Tm	O
<	O
73	O
°	O
C	O
),	O
yellow	O
(	O
73	O
°	O
C	O
<	O
Tm	O
<	O
78	O
°	O
C	O
),	O
orange	O
(	O
Tm	O
>	O
78	O
°	O
C	O
).	O
	O
Melting	O
temperatures	O
(	O
Tm	O
)	O
were	O
measured	O
for	O
all	O
Fabs	O
using	O
differential	O
scanning	O
calorimetry	O
(	O
Table	O
5	O
).	O
	O
It	O
appears	O
that	O
for	O
each	O
given	O
LC	O
,	O
the	O
Fabs	O
with	O
germlines	O
H1	B-mutant
-	I-mutant
69	I-mutant
and	O
H3	B-mutant
-	I-mutant
23	I-mutant
are	O
substantially	O
more	O
stable	O
than	O
those	O
with	O
germlines	O
H3	B-mutant
-	I-mutant
53	I-mutant
and	O
H5	B-mutant
-	I-mutant
51	I-mutant
.	O
	O
In	O
addition	O
,	O
L1	B-mutant
-	I-mutant
39	I-mutant
provides	O
a	O
much	O
higher	O
degree	O
of	O
stabilization	O
than	O
the	O
other	O
3	O
LC	O
germlines	O
when	O
combined	O
with	O
any	O
of	O
the	O
HCs	O
.	O
	O
As	O
a	O
result	O
,	O
the	O
Tm	O
for	O
pairs	O
H1	O
-	O
69	O
:	O
L1	O
-	O
39	O
and	O
H3	O
-	O
23	O
:	O
L1	O
-	O
39	O
is	O
12	O
-	O
13	O
°	O
higher	O
than	O
for	O
pairs	O
H3	O
-	O
53	O
:	O
L3	O
-	O
20	O
,	O
H3	O
-	O
53	O
:	O
L4	O
-	O
1	O
,	O
H5	O
-	O
51	O
:	O
L3	O
-	O
20	O
and	O
H5	O
-	O
51	O
:	O
L4	O
-	O
1	O
.	O
	O
These	O
findings	O
correlate	O
well	O
with	O
the	O
degree	O
of	O
conformational	O
disorder	O
observed	O
in	O
the	O
crystal	O
structures	O
.	O
	O
No	O
electron	O
density	O
is	O
observed	O
for	O
a	O
number	O
of	O
side	O
chains	O
in	O
CDRs	O
H3	O
and	O
L3	O
in	O
all	O
Fabs	O
with	O
germline	O
H3	B-mutant
-	I-mutant
53	I-mutant
,	O
which	O
indicates	O
loose	O
packing	O
of	O
the	O
variable	O
domains	O
.	O
	O
All	O
those	O
molecules	O
are	O
relatively	O
unstable	O
,	O
as	O
is	O
reflected	O
in	O
their	O
low	O
Tms	O
.	O
	O
The	O
structural	O
data	O
set	O
taken	O
as	O
a	O
whole	O
provides	O
insight	O
into	O
how	O
the	O
backbone	O
conformations	O
of	O
the	O
CDRs	O
of	O
a	O
specific	O
heavy	O
or	O
light	O
chain	O
vary	O
when	O
it	O
is	O
paired	O
with	O
4	O
different	O
light	O
or	O
heavy	O
chains	O
,	O
respectively	O
.	O
	O
A	O
large	O
variability	O
in	O
the	O
CDR	O
conformations	O
for	O
the	O
sets	O
of	O
HCs	O
and	O
LCs	O
is	O
observed	O
.	O
	O
In	O
some	O
cases	O
the	O
CDR	O
conformations	O
for	O
all	O
members	O
of	O
a	O
set	O
are	O
virtually	O
identical	O
,	O
for	O
others	O
subtle	O
changes	O
occur	O
in	O
a	O
few	O
members	O
of	O
a	O
set	O
,	O
and	O
in	O
some	O
cases	O
larger	O
deviations	O
are	O
observed	O
within	O
a	O
set	O
.	O
	O
The	O
five	O
variants	O
that	O
crystallized	O
with	O
2	O
copies	O
of	O
the	O
Fab	O
in	O
the	O
asymmetric	O
unit	O
serve	O
somewhat	O
as	O
controls	O
for	O
the	O
influence	O
of	O
crystal	O
packing	O
on	O
the	O
conformations	O
of	O
the	O
CDRs	O
.	O
	O
In	O
four	O
of	O
the	O
5	O
structures	O
the	O
CDR	O
conformations	O
are	O
consistent	O
.	O
	O
In	O
only	O
one	O
case	O
,	O
that	O
of	O
H1	O
-	O
69	O
:	O
L3	O
-	O
20	O
(	O
the	O
lowest	O
resolution	O
structure	O
),	O
do	O
we	O
see	O
differences	O
in	O
the	O
conformations	O
of	O
the	O
2	O
copies	O
of	O
CDRs	O
H1	O
and	O
L1	O
.	O
	O
This	O
variability	O
is	O
likely	O
a	O
result	O
of	O
2	O
factors	O
,	O
crystal	O
packing	O
interactions	O
and	O
internal	O
instability	O
of	O
the	O
variable	O
domain	O
.	O
	O
For	O
the	O
CDRs	O
with	O
canonical	O
structures	O
,	O
the	O
largest	O
changes	O
in	O
conformation	O
occur	O
for	O
CDR	O
H1	O
of	O
H1	B-mutant
-	I-mutant
69	I-mutant
and	O
H3	B-mutant
-	I-mutant
53	I-mutant
.	O
	O
The	O
other	O
2	O
HCs	O
,	O
H3	B-mutant
-	I-mutant
23	I-mutant
and	O
H5	B-mutant
-	I-mutant
51	I-mutant
,	O
have	O
canonical	O
structures	O
that	O
are	O
remarkably	O
well	O
conserved	O
(	O
Fig	O
.	O
1	O
).	O
	O
Of	O
the	O
4	O
HCs	O
,	O
H1	B-mutant
-	I-mutant
69	I-mutant
has	O
the	O
greatest	O
number	O
of	O
canonical	O
structure	O
assignments	O
(	O
Table	O
2	O
).	O
	O
H1	B-mutant
-	I-mutant
69	I-mutant
is	O
unique	O
in	O
having	O
a	O
pair	O
of	O
glycine	O
residues	O
at	O
positions	O
26	O
and	O
27	O
,	O
which	O
provide	O
more	O
conformational	O
freedom	O
in	O
CDR	O
H1	O
.	O
	O
Besides	O
IGHV1	B-mutant
-	I-mutant
69	I-mutant
,	O
only	O
the	O
germlines	O
of	O
the	O
VH4	O
family	O
possess	O
double	O
glycines	O
in	O
CDR	O
H1	O
,	O
and	O
it	O
will	O
be	O
interesting	O
to	O
see	O
if	O
they	O
are	O
also	O
conformationally	O
unstable	O
.	O
	O
As	O
mentioned	O
in	O
the	O
Results	O
section	O
,	O
this	O
data	O
set	O
is	O
composed	O
of	O
21	O
Fabs	O
,	O
since	O
5	O
of	O
the	O
16	O
variants	O
have	O
2	O
Fab	O
copies	O
in	O
the	O
asymmetric	O
unit	O
.	O
	O
For	O
the	O
18	O
Fabs	O
with	O
complete	O
backbone	O
atoms	O
for	O
CDR	O
H3	O
,	O
10	O
have	O
conformations	O
similar	O
to	O
that	O
of	O
the	O
parent	O
,	O
while	O
the	O
others	O
have	O
significantly	O
different	O
conformations	O
(	O
Fig	O
.	O
6	O
).	O
	O
More	O
than	O
half	O
of	O
the	O
variants	O
retain	O
the	O
conformation	O
of	O
the	O
parent	O
despite	O
having	O
differences	O
in	O
the	O
VH	O
:	O
VL	O
pairing	O
.	O
	O
This	O
subset	O
includes	O
2	O
structures	O
with	O
2	O
copies	O
of	O
the	O
Fab	O
in	O
the	O
asymmetric	O
unit	O
,	O
all	O
of	O
which	O
are	O
nearly	O
identical	O
in	O
conformation	O
.	O
	O
The	O
remaining	O
8	O
structures	O
exhibit	O
“	O
non	O
-	O
parental	O
”	O
conformations	O
,	O
indicating	O
that	O
the	O
VH	O
and	O
VL	O
context	O
can	O
also	O
be	O
a	O
dominating	O
factor	O
influencing	O
CDR	O
H3	O
.	O
	O
Interestingly	O
,	O
as	O
described	O
earlier	O
,	O
these	O
2	O
pairs	O
differ	O
in	O
the	O
stem	O
regions	O
with	O
the	O
H1	O
-	O
69	O
:	O
L3	O
-	O
20	O
pair	O
in	O
the	O
‘	O
extended	O
’	O
conformation	O
and	O
H5	O
-	O
51	O
:	O
L4	O
-	O
1	O
pair	O
in	O
the	O
‘	O
kinked	O
’	O
conformation	O
.	O
	O
The	O
CDR	O
H3	O
conformational	O
analysis	O
shows	O
that	O
,	O
for	O
each	O
set	O
of	O
variants	O
of	O
one	O
HC	O
paired	O
with	O
the	O
4	O
different	O
LCs	O
,	O
both	O
“	O
parental	O
”	O
and	O
“	O
non	O
-	O
parental	O
”	O
conformations	O
are	O
observed	O
.	O
	O
Thus	O
,	O
no	O
patterns	O
of	O
conformational	O
preference	O
for	O
a	O
particular	O
HC	O
or	O
LC	O
emerge	O
to	O
shed	O
any	O
direct	O
light	O
on	O
what	O
drives	O
the	O
conformational	O
differences	O
.	O
	O
This	O
finding	O
supports	O
the	O
hypothesis	O
of	O
Weitzner	O
et	O
al	O
.	O
that	O
the	O
H3	O
conformation	O
is	O
controlled	O
both	O
by	O
its	O
sequence	O
and	O
its	O
environment	O
.	O
	O
In	O
looking	O
at	O
a	O
possible	O
correlation	O
between	O
the	O
tilt	O
angle	O
and	O
the	O
conformation	O
of	O
CDR	O
H3	O
,	O
no	O
clear	O
trends	O
are	O
observed	O
.	O
	O
Two	O
variants	O
,	O
H1	O
-	O
69	O
:	O
L3	O
-	O
20	O
and	O
H3	O
-	O
23	O
:	O
L3	O
-	O
20	O
,	O
have	O
the	O
largest	O
differences	O
in	O
the	O
tilt	O
angles	O
compared	O
to	O
other	O
variants	O
as	O
seen	O
in	O
Table	O
3	O
.	O
	O
The	O
absolute	O
VH	O
:	O
VL	O
orientation	O
parameters	O
for	O
the	O
2	O
Fabs	O
(	O
Table	O
S2	O
)	O
show	O
significant	O
deviation	O
in	O
HL	O
,	O
LC1	O
and	O
HC2	O
values	O
(	O
2	O
-	O
3	O
standard	O
deviations	O
from	O
the	O
mean	O
).	O
	O
One	O
of	O
the	O
variants	O
,	O
H3	O
-	O
23	O
:	O
L3	O
-	O
20	O
,	O
has	O
the	O
CDR	O
H3	O
conformation	O
similar	O
to	O
the	O
parent	O
,	O
but	O
the	O
other	O
,	O
H1	O
-	O
69	O
:	O
L3	O
-	O
20	O
,	O
is	O
different	O
.	O
	O
These	O
smaller	O
interfaces	O
may	O
perhaps	O
translate	O
to	O
a	O
significant	O
deviation	O
in	O
how	O
VH	O
is	O
oriented	O
relative	O
to	O
VL	O
than	O
the	O
other	O
variants	O
.	O
	O
These	O
deviations	O
from	O
the	O
other	O
variants	O
can	O
also	O
be	O
seen	O
to	O
some	O
extent	O
in	O
VH	O
:	O
VL	O
orientation	O
parameters	O
in	O
Table	O
S2	O
,	O
as	O
well	O
as	O
in	O
the	O
smaller	O
number	O
of	O
residues	O
involved	O
in	O
the	O
VH	O
:	O
VL	O
interfaces	O
of	O
these	O
2	O
variants	O
(	O
Fig	O
.	O
S5	O
).	O
	O
These	O
differences	O
undoubtedly	O
influence	O
the	O
conformation	O
of	O
the	O
CDRs	O
,	O
in	O
particular	O
CDR	O
H1	O
(	O
Fig	O
.	O
1A	O
)	O
and	O
CDR	O
L1	O
(	O
Fig	O
.	O
3C	O
),	O
especially	O
with	O
the	O
tandem	O
glycines	O
and	O
multiple	O
serines	O
present	O
,	O
respectively	O
.	O
	O
As	O
indicated	O
by	O
the	O
melting	O
temperatures	O
,	O
germlines	O
H1	B-mutant
-	I-mutant
69	I-mutant
and	O
H3	B-mutant
-	I-mutant
23	I-mutant
for	O
HC	O
and	O
germline	O
L1	B-mutant
-	I-mutant
39	I-mutant
for	O
LC	O
produce	O
more	O
stable	O
Fabs	O
compared	O
to	O
the	O
other	O
germlines	O
in	O
the	O
experimental	O
set	O
.	O
	O
One	O
possible	O
explanation	O
of	O
the	O
clear	O
preference	O
of	O
LC	O
germline	O
L1	B-mutant
-	I-mutant
39	I-mutant
is	O
that	O
CDR	O
L3	O
has	O
smaller	O
residues	O
at	O
positions	O
91	O
and	O
94	O
,	O
allowing	O
for	O
more	O
room	O
to	O
accommodate	O
CDR	O
H3	O
.	O
	O
Various	O
combinations	O
of	O
germline	O
sequences	O
for	O
VL	O
and	O
VH	O
impose	O
certain	O
constraints	O
on	O
CDR	O
H3	O
,	O
which	O
has	O
to	O
adapt	O
to	O
the	O
environment	O
.	O
	O
At	O
the	O
other	O
end	O
of	O
the	O
stability	O
range	O
is	O
LC	O
germline	O
L3	B-mutant
-	I-mutant
20	I-mutant
,	O
which	O
yields	O
antibodies	O
with	O
the	O
lowest	O
Tms	O
.	O
	O
While	O
pairings	O
with	O
H3	B-mutant
-	I-mutant
53	I-mutant
and	O
H5	B-mutant
-	I-mutant
51	I-mutant
may	O
be	O
safely	O
called	O
a	O
mismatch	O
,	O
those	O
with	O
H1	B-mutant
-	I-mutant
69	I-mutant
and	O
H3	B-mutant
-	I-mutant
23	I-mutant
have	O
Tms	O
about	O
5	O
-	O
6	O
°	O
higher	O
.	O
	O
It	O
is	O
possible	O
that	O
by	O
adopting	O
extreme	O
tilt	O
angles	O
the	O
structure	O
modulates	O
CDR	O
H3	O
and	O
its	O
environment	O
,	O
which	O
apparently	O
cannot	O
be	O
achieved	O
solely	O
by	O
conformational	O
rearrangement	O
of	O
the	O
CDR	O
.	O
	O
Overall	O
,	O
the	O
stability	O
of	O
the	O
Fab	O
,	O
as	O
measured	O
by	O
Tm	O
,	O
is	O
a	O
result	O
of	O
the	O
mutual	O
adjustment	O
of	O
the	O
HC	O
and	O
LC	O
variable	O
domains	O
and	O
adjustment	O
of	O
CDR	O
H3	O
to	O
the	O
VH	O
:	O
VL	O
cleft	O
.	O
	O
The	O
final	O
conformation	O
represents	O
an	O
energetic	O
minimum	O
;	O
however	O
,	O
in	O
most	O
cases	O
it	O
is	O
very	O
shallow	O
,	O
so	O
that	O
a	O
single	O
mutation	O
can	O
cause	O
a	O
dramatic	O
rearrangement	O
of	O
the	O
structure	O
.	O
	O
In	O
summary	O
,	O
the	O
analysis	O
of	O
this	O
structural	O
library	O
of	O
germline	O
variants	O
composed	O
of	O
all	O
pairs	O
of	O
4	O
HCs	O
and	O
4LCs	O
,	O
all	O
with	O
the	O
same	O
CDR	O
H3	O
,	O
offers	O
some	O
unique	O
insights	O
into	O
antibody	O
structure	O
and	O
how	O
pairing	O
and	O
sequence	O
may	O
influence	O
,	O
or	O
not	O
,	O
the	O
canonical	O
structures	O
of	O
the	O
L1	O
,	O
L2	O
,	O
L3	O
,	O
H1	O
and	O
H2	O
CDRs	O
.	O
	O
These	O
data	O
reveal	O
the	O
difficulty	O
of	O
modeling	O
CDR	O
H3	O
accurately	O
,	O
as	O
shown	O
again	O
in	O
Antibody	O
Modeling	O
Assessment	O
II	O
.	O
	O
Furthermore	O
,	O
antibody	O
CDRs	O
,	O
H3	O
in	O
particular	O
,	O
may	O
go	O
through	O
conformational	O
changes	O
upon	O
binding	O
their	O
targets	O
,	O
making	O
structural	O
prediction	O
for	O
docking	O
purposes	O
an	O
even	O
more	O
difficult	O
task	O
.	O
	O
For	O
those	O
applications	O
where	O
accurate	O
CDR	O
structures	O
are	O
essential	O
,	O
such	O
as	O
docking	O
,	O
the	O
results	O
in	O
this	O
work	O
demonstrate	O
the	O
importance	O
of	O
experimental	O
structures	O
.	O
	O
With	O
the	O
recent	O
advances	O
in	O
expression	O
and	O
crystallization	O
methods	O
,	O
Fab	O
structures	O
can	O
be	O
obtained	O
rapidly	O
.	O
	O
The	O
set	O
of	O
16	O
germline	O
Fab	O
structures	O
offers	O
a	O
unique	O
dataset	O
to	O
facilitate	O
software	O
development	O
for	O
antibody	O
modeling	O
.	O
	O
The	O
results	O
essentially	O
support	O
the	O
underlying	O
idea	O
of	O
canonical	O
structures	O
,	O
indicating	O
that	O
most	O
CDRs	O
with	O
germline	O
sequences	O
tend	O
to	O
adopt	O
predefined	O
conformations	O
.	O
	O
This	O
would	O
insure	O
more	O
structural	O
diversity	O
,	O
leading	O
to	O
a	O
more	O
diverse	O
panel	O
of	O
antibodies	O
that	O
would	O
bind	O
to	O
a	O
broad	O
spectrum	O
of	O
targets	O
.	O
	O
Challenges	O
in	O
determining	O
the	O
structures	O
of	O
heterogeneous	O
and	O
dynamic	O
protein	O
complexes	O
have	O
greatly	O
hampered	O
past	O
efforts	O
to	O
obtain	O
a	O
mechanistic	O
understanding	O
of	O
many	O
important	O
biological	O
processes	O
.	O
	O
One	O
such	O
process	O
is	O
chaperone	O
-	O
assisted	O
protein	O
folding	O
,	O
where	O
obtaining	O
structural	O
ensembles	O
of	O
chaperone	O
:	O
substrate	O
complexes	O
would	O
ultimately	O
reveal	O
how	O
chaperones	O
help	O
proteins	O
fold	O
into	O
their	O
native	O
state	O
.	O
	O
To	O
address	O
this	O
problem	O
,	O
we	O
devised	O
a	O
novel	O
structural	O
biology	O
approach	O
based	O
on	O
X	O
-	O
ray	O
crystallography	O
,	O
termed	O
Residual	O
Electron	O
and	O
Anomalous	O
Density	O
(	O
READ	O
).	O
	O
READ	O
enabled	O
us	O
to	O
visualize	O
even	O
sparsely	O
populated	O
conformations	O
of	O
the	O
substrate	O
protein	O
immunity	O
protein	O
7	O
(	O
Im7	O
)	O
in	O
complex	O
with	O
the	O
E	O
.	O
coli	O
chaperone	O
Spy	O
.	O
	O
The	O
ensemble	O
shows	O
that	O
Spy	O
-	O
associated	O
Im7	O
samples	O
conformations	O
ranging	O
from	O
unfolded	O
to	O
partially	O
folded	O
and	O
native	O
-	O
like	O
states	O
,	O
and	O
reveals	O
how	O
a	O
substrate	O
can	O
explore	O
its	O
folding	O
landscape	O
while	O
bound	O
to	O
a	O
chaperone	O
.	O
	O
High	O
-	O
resolution	O
structural	O
models	O
of	O
protein	O
-	O
protein	O
interactions	O
are	O
critical	O
for	O
obtaining	O
mechanistic	O
insights	O
into	O
biological	O
processes	O
.	O
	O
However	O
,	O
many	O
protein	O
-	O
protein	O
interactions	O
are	O
highly	O
dynamic	O
,	O
making	O
it	O
difficult	O
to	O
obtain	O
high	O
-	O
resolution	O
data	O
.	O
	O
Particularly	O
challenging	O
are	O
interactions	O
of	O
intrinsically	O
or	O
conditionally	O
disordered	O
sections	O
of	O
proteins	O
with	O
their	O
partner	O
proteins	O
.	O
	O
It	O
is	O
clear	O
that	O
molecular	O
chaperones	O
aid	O
in	O
protein	O
folding	O
.	O
	O
Structural	O
characterization	O
of	O
chaperone	O
-	O
assisted	O
protein	O
folding	O
likely	O
would	O
help	O
bring	O
clarity	O
to	O
this	O
question	O
.	O
	O
Structural	O
models	O
of	O
chaperone	O
-	O
substrate	O
complexes	O
have	O
recently	O
begun	O
to	O
provide	O
information	O
as	O
to	O
how	O
a	O
chaperone	O
can	O
recognize	O
its	O
substrate	O
.	O
	O
However	O
,	O
the	O
impact	O
that	O
chaperones	O
have	O
on	O
their	O
substrates	O
,	O
and	O
how	O
these	O
interactions	O
affect	O
the	O
folding	O
process	O
remain	O
largely	O
unknown	O
.	O
	O
For	O
most	O
chaperones	O
,	O
it	O
is	O
still	O
unclear	O
whether	O
the	O
chaperone	O
actively	O
participates	O
in	O
and	O
affects	O
the	O
folding	O
of	O
the	O
substrate	O
proteins	O
,	O
or	O
merely	O
provides	O
a	O
suitable	O
microenvironment	O
enabling	O
the	O
substrate	O
to	O
fold	O
on	O
its	O
own	O
.	O
	O
This	O
is	O
a	O
truly	O
fundamental	O
question	O
in	O
the	O
chaperone	O
field	O
,	O
and	O
one	O
that	O
has	O
eluded	O
the	O
community	O
largely	O
because	O
of	O
the	O
highly	O
dynamic	O
nature	O
of	O
the	O
chaperone	O
-	O
substrate	O
complexes	O
.	O
	O
To	O
address	O
this	O
question	O
,	O
we	O
investigated	O
the	O
ATP	O
-	O
independent	O
Escherichia	O
coli	O
periplasmic	O
chaperone	O
Spy	O
.	O
	O
Spy	O
prevents	O
protein	O
aggregation	O
and	O
aids	O
in	O
protein	O
folding	O
under	O
various	O
stress	O
conditions	O
,	O
including	O
treatment	O
with	O
tannin	O
and	O
butanol	O
.	O
	O
We	O
originally	O
discovered	O
Spy	O
by	O
its	O
ability	O
to	O
stabilize	O
the	O
protein	O
-	O
folding	O
model	O
Im7	O
in	O
vivo	O
and	O
recently	O
demonstrated	O
that	O
Im7	O
folds	O
while	O
associated	O
with	O
Spy	O
.	O
	O
The	O
crystal	O
structure	O
of	O
Spy	O
revealed	O
that	O
it	O
forms	O
a	O
thin	O
α	O
-	O
helical	O
homodimeric	O
cradle	O
.	O
	O
Crosslinking	O
and	O
genetic	O
experiments	O
suggested	O
that	O
Spy	O
interacts	O
with	O
substrates	O
somewhere	O
on	O
its	O
concave	O
side	O
.	O
	O
By	O
using	O
a	O
novel	O
X	O
-	O
ray	O
crystallography	O
-	O
based	O
approach	O
to	O
model	O
disorder	O
in	O
crystal	O
structures	O
,	O
we	O
have	O
now	O
determined	O
the	O
high	O
-	O
resolution	O
ensemble	O
of	O
the	O
dynamic	O
Spy	O
:	O
Im7	O
complex	O
.	O
	O
This	O
work	O
provides	O
a	O
detailed	O
view	O
of	O
chaperone	O
-	O
mediated	O
protein	O
folding	O
and	O
shows	O
how	O
substrates	O
like	O
Im7	O
find	O
their	O
native	O
fold	O
while	O
bound	O
to	O
their	O
chaperones	O
.	O
	O
Crystallizing	O
the	O
Spy	O
:	O
Im7	O
complex	O
	O
We	O
reasoned	O
that	O
to	O
obtain	O
crystals	O
of	O
complexes	O
between	O
Spy	O
(	O
domain	O
boundaries	O
in	O
Supplementary	O
Fig	O
.	O
1	O
)	O
and	O
its	O
substrate	O
proteins	O
,	O
our	O
best	O
approach	O
was	O
to	O
identify	O
crystallization	O
conditions	O
that	O
yielded	O
Spy	O
crystals	O
in	O
the	O
presence	O
of	O
protein	O
substrates	O
but	O
not	O
in	O
their	O
absence	O
.	O
	O
We	O
therefore	O
screened	O
crystallization	O
conditions	O
for	O
Spy	O
with	O
four	O
different	O
substrate	O
proteins	O
:	O
a	O
fragment	O
of	O
the	O
largely	O
unfolded	O
bovine	O
α	O
-	O
casein	O
protein	O
,	O
wild	O
-	O
type	O
(	O
WT	O
)	O
E	O
.	O
coli	O
Im7	O
,	O
an	O
unfolded	O
variant	O
of	O
Im7	O
(	O
L18A	B-mutant
L19A	B-mutant
L37A	B-mutant
),	O
and	O
the	O
N	O
-	O
terminal	O
half	O
of	O
Im7	O
(	O
Im76	B-mutant
-	I-mutant
45	I-mutant
),	O
which	O
encompasses	O
the	O
entire	O
Spy	O
-	O
binding	O
portion	O
of	O
Im7	O
.	O
	O
We	O
found	O
conditions	O
in	O
which	O
all	O
four	O
substrates	O
co	O
-	O
crystallized	O
with	O
Spy	O
,	O
but	O
in	O
which	O
Spy	O
alone	O
did	O
not	O
yield	O
crystals	O
.	O
	O
Subsequent	O
crystal	O
washing	O
and	O
dissolution	O
experiments	O
confirmed	O
the	O
presence	O
of	O
the	O
substrates	O
in	O
the	O
co	O
-	O
crystals	O
(	O
Supplementary	O
Fig	O
.	O
2	O
).	O
	O
The	O
crystals	O
diffracted	O
to	O
~	O
1	O
.	O
8	O
Å	O
resolution	O
.	O
	O
We	O
used	O
Spy	O
:	O
Im76	O
-	O
45	O
selenomethionine	O
crystals	O
for	O
phasing	O
with	O
single	O
-	O
wavelength	O
anomalous	O
diffraction	O
(	O
SAD	O
)	O
experiments	O
,	O
and	O
used	O
this	O
solution	O
to	O
build	O
the	O
well	O
-	O
ordered	O
Spy	O
portions	O
of	O
all	O
four	O
complexes	O
.	O
	O
Even	O
the	O
minimal	O
binding	O
portion	O
of	O
Im7	O
(	O
Im76	B-mutant
-	I-mutant
45	I-mutant
)	O
showed	O
highly	O
dispersed	O
electron	O
density	O
(	O
Fig	O
.	O
1a	O
).	O
	O
We	O
hypothesized	O
that	O
the	O
fragmented	O
density	O
was	O
due	O
to	O
multiple	O
,	O
partially	O
occupied	O
conformations	O
of	O
the	O
substrate	O
bound	O
within	O
the	O
crystal	O
.	O
	O
Such	O
residual	O
density	O
is	O
typically	O
not	O
considered	O
usable	O
by	O
traditional	O
X	O
-	O
ray	O
crystallography	O
methods	O
.	O
	O
Thus	O
,	O
we	O
developed	O
a	O
new	O
approach	O
to	O
interpret	O
the	O
chaperone	O
-	O
bound	O
substrate	O
in	O
multiple	O
conformations	O
.	O
	O
READ	O
:	O
a	O
strategy	O
to	O
visualize	O
heterogeneous	O
and	O
dynamic	O
biomolecules	O
	O
We	O
split	O
this	O
approach	O
into	O
five	O
steps	O
:	O
(	O
1	O
)	O
By	O
using	O
a	O
well	O
-	O
diffracting	O
Spy	O
:	O
substrate	O
co	O
-	O
crystal	O
,	O
we	O
first	O
determined	O
the	O
structure	O
of	O
the	O
folded	O
domain	O
of	O
Spy	O
and	O
obtained	O
high	O
quality	O
residual	O
electron	O
density	O
within	O
the	O
dynamic	O
regions	O
of	O
the	O
substrate	O
.	O
	O
(	O
2	O
)	O
We	O
then	O
labeled	O
individual	O
residues	O
in	O
the	O
flexible	O
regions	O
of	O
the	O
substrate	O
with	O
the	O
strong	O
anomalous	O
scatterer	O
iodine	O
,	O
which	O
serves	O
to	O
locate	O
these	O
residues	O
in	O
three	O
-	O
dimensional	O
space	O
using	O
their	O
anomalous	O
density	O
.	O
	O
(	O
3	O
)	O
We	O
performed	O
molecular	O
dynamics	O
(	O
MD	O
)	O
simulations	O
to	O
generate	O
a	O
pool	O
of	O
energetically	O
reasonable	O
conformations	O
of	O
the	O
dynamic	O
complex	O
and	O
(	O
4	O
)	O
applied	O
a	O
sample	O
-	O
and	O
-	O
select	O
algorithm	O
to	O
determine	O
the	O
minimal	O
set	O
of	O
substrate	O
conformations	O
that	O
fit	O
both	O
the	O
residual	O
and	O
anomalous	O
density	O
.	O
	O
The	O
electron	O
density	O
then	O
allowed	O
us	O
to	O
connect	O
the	O
labeled	O
residues	O
of	O
the	O
substrate	O
by	O
confining	O
the	O
protein	O
chain	O
within	O
regions	O
of	O
detectable	O
density	O
.	O
	O
In	O
this	O
way	O
,	O
the	O
two	O
forms	O
of	O
data	O
together	O
were	O
able	O
to	O
describe	O
multiple	O
conformations	O
of	O
the	O
substrate	O
within	O
the	O
crystal	O
.	O
	O
However	O
,	O
we	O
believe	O
that	O
READ	O
will	O
prove	O
generally	O
applicable	O
to	O
visualizing	O
heterogeneous	O
and	O
dynamic	O
complexes	O
that	O
have	O
previously	O
escaped	O
detailed	O
structural	O
analysis	O
.	O
	O
Collecting	O
READ	O
data	O
for	O
the	O
Spy	O
:	O
Im76	O
-	O
45	O
complex	O
	O
To	O
apply	O
the	O
READ	O
technique	O
to	O
the	O
folding	O
mechanism	O
employed	O
by	O
the	O
chaperone	O
Spy	O
,	O
we	O
selected	O
Im76	B-mutant
-	I-mutant
45	I-mutant
for	O
further	O
investigation	O
because	O
NMR	O
data	O
suggested	O
that	O
Im76	B-mutant
-	I-mutant
45	I-mutant
could	O
recapitulate	O
unfolded	O
,	O
partially	O
folded	O
,	O
and	O
native	O
-	O
like	O
states	O
of	O
Im7	O
(	O
Supplementary	O
Fig	O
.	O
3	O
).	O
	O
Moreover	O
,	O
binding	O
experiments	O
indicated	O
that	O
Im76	B-mutant
-	I-mutant
45	I-mutant
comprises	O
the	O
entire	O
Spy	O
-	O
binding	O
region	O
.	O
	O
To	O
introduce	O
the	O
anomalous	O
scatterer	O
iodine	O
,	O
we	O
replaced	O
eight	O
Im76	B-mutant
-	I-mutant
45	I-mutant
residues	O
with	O
the	O
non	O
-	O
canonical	O
amino	O
acid	O
4	O
-	O
iodophenylalanine	O
(	O
pI	O
-	O
Phe	O
).	O
	O
Consistent	O
with	O
our	O
electron	O
density	O
map	O
,	O
we	O
found	O
that	O
the	O
majority	O
of	O
anomalous	O
signals	O
emerged	O
in	O
the	O
cradle	O
of	O
Spy	O
,	O
implying	O
that	O
this	O
is	O
the	O
likely	O
Im7	O
substrate	O
binding	O
site	O
.	O
	O
Consistent	O
with	O
the	O
fragmented	O
density	O
,	O
however	O
,	O
we	O
observed	O
multiple	O
iodine	O
positions	O
for	O
seven	O
of	O
the	O
eight	O
substituted	O
residues	O
.	O
	O
READ	O
sample	O
-	O
and	O
-	O
select	O
procedure	O
	O
During	O
each	O
round	O
of	O
the	O
selection	O
,	O
a	O
genetic	O
algorithm	O
alters	O
the	O
ensemble	O
and	O
its	O
agreement	O
to	O
the	O
experimental	O
data	O
is	O
re	O
-	O
evaluated	O
.	O
	O
If	O
successful	O
,	O
the	O
selection	O
identifies	O
the	O
smallest	O
group	O
of	O
specific	O
conformations	O
that	O
best	O
fits	O
the	O
residual	O
electron	O
density	O
and	O
anomalous	O
signals	O
.	O
	O
This	O
strategy	O
allows	O
a	O
wide	O
range	O
of	O
substrate	O
conformations	O
to	O
interact	O
with	O
the	O
chaperone	O
.	O
	O
From	O
the	O
MD	O
simulations	O
,	O
we	O
extracted	O
~	O
10	O
,	O
000	O
diverse	O
Spy	O
:	O
Im76	O
-	O
45	O
complexes	O
to	O
be	O
used	O
by	O
the	O
ensuing	O
selection	O
.	O
	O
Each	O
complex	O
within	O
this	O
pool	O
comprises	O
one	O
Spy	O
dimer	O
bound	O
to	O
a	O
single	O
Im76	B-mutant
-	I-mutant
45	I-mutant
substrate	O
.	O
	O
This	O
pool	O
was	O
then	O
used	O
by	O
the	O
selection	O
algorithm	O
to	O
identify	O
the	O
minimal	O
ensemble	O
that	O
best	O
satisfies	O
both	O
the	O
residual	O
electron	O
and	O
anomalous	O
crystallographic	O
data	O
.	O
	O
Simultaneously	O
,	O
it	O
uses	O
the	O
residual	O
electron	O
density	O
to	O
help	O
choose	O
ensembles	O
.	O
	O
This	O
process	O
provided	O
us	O
with	O
a	O
target	O
map	O
that	O
the	O
ensuing	O
selection	O
tried	O
to	O
recapitulate	O
.	O
	O
To	O
reduce	O
the	O
extent	O
of	O
3D	O
space	O
to	O
be	O
explored	O
,	O
this	O
compressed	O
map	O
was	O
created	O
by	O
only	O
using	O
density	O
from	O
regions	O
of	O
space	O
significantly	O
sampled	O
by	O
Im76	B-mutant
-	I-mutant
45	I-mutant
in	O
the	O
Spy	O
:	O
Im76	O
-	O
45	O
MD	O
simulations	O
.	O
	O
For	O
each	O
of	O
the	O
~	O
10	O
,	O
000	O
complexes	O
in	O
the	O
coarse	O
-	O
grained	O
MD	O
pool	O
,	O
the	O
electron	O
density	O
at	O
the	O
Cα	O
positions	O
of	O
Im76	B-mutant
-	I-mutant
45	I-mutant
was	O
extracted	O
and	O
used	O
to	O
construct	O
an	O
electron	O
density	O
map	O
(	O
Online	O
Methods	O
).	O
	O
These	O
individual	O
electron	O
density	O
maps	O
from	O
the	O
separate	O
conformers	O
could	O
then	O
be	O
combined	O
(	O
Fig	O
.	O
2	O
)	O
and	O
compared	O
to	O
the	O
averaged	O
experimental	O
electron	O
density	O
map	O
as	O
part	O
of	O
the	O
selection	O
algorithm	O
.	O
	O
This	O
approach	O
allowed	O
us	O
to	O
simultaneously	O
use	O
both	O
the	O
iodine	O
anomalous	O
signals	O
and	O
the	O
residual	O
electron	O
density	O
in	O
the	O
selection	O
procedure	O
.	O
	O
The	O
selection	O
resulted	O
in	O
small	O
ensembles	O
from	O
the	O
MD	O
pool	O
that	O
best	O
fit	O
the	O
READ	O
data	O
(	O
Fig	O
.	O
1c	O
,	O
d	O
).	O
	O
Before	O
analyzing	O
the	O
details	O
of	O
the	O
Spy	O
:	O
Im76	O
-	O
45	O
complex	O
,	O
we	O
first	O
engaged	O
in	O
a	O
series	O
of	O
validation	O
tests	O
to	O
verify	O
the	O
ensemble	O
and	O
selection	O
procedure	O
(	O
Supplementary	O
Note	O
1	O
,	O
Figures	O
1c	O
,	O
d	O
,	O
Supplemental	O
Figures	O
5	O
-	O
7	O
).	O
	O
Of	O
note	O
,	O
the	O
final	O
six	O
-	O
membered	O
ensemble	O
was	O
the	O
largest	O
ensemble	O
that	O
could	O
simultaneously	O
decrease	O
the	O
RFree	O
and	O
pass	O
the	O
10	O
-	O
fold	O
cross	O
-	O
validation	O
test	O
.	O
	O
This	O
ensemble	O
depicts	O
the	O
conformations	O
that	O
the	O
substrate	O
Im76	B-mutant
-	I-mutant
45	I-mutant
adopts	O
while	O
bound	O
to	O
the	O
chaperone	O
Spy	O
(	O
Fig	O
.	O
3	O
Supplementary	O
Movie	O
1	O
,	O
and	O
Table	O
1	O
).	O
	O
Our	O
results	O
showed	O
that	O
by	O
using	O
this	O
novel	O
READ	O
approach	O
,	O
we	O
were	O
able	O
to	O
obtain	O
structural	O
information	O
about	O
the	O
dynamic	O
interaction	O
of	O
a	O
chaperone	O
with	O
its	O
substrate	O
protein	O
.	O
	O
We	O
were	O
particularly	O
interested	O
in	O
finding	O
answers	O
to	O
one	O
of	O
the	O
most	O
fundamental	O
questions	O
in	O
chaperone	O
biology	O
—	O
how	O
does	O
chaperone	O
binding	O
affect	O
substrate	O
structure	O
and	O
vice	O
versa	O
.	O
	O
By	O
analyzing	O
the	O
individual	O
structures	O
of	O
the	O
six	O
-	O
member	O
ensemble	O
of	O
Im76	B-mutant
-	I-mutant
45	I-mutant
bound	O
to	O
Spy	O
,	O
we	O
observed	O
that	O
Im76	B-mutant
-	I-mutant
45	I-mutant
takes	O
on	O
several	O
different	O
conformations	O
while	O
bound	O
.	O
	O
We	O
found	O
these	O
conformations	O
to	O
be	O
highly	O
heterogeneous	O
and	O
to	O
include	O
unfolded	O
,	O
partially	O
folded	O
,	O
and	O
native	O
-	O
like	O
states	O
(	O
Fig	O
.	O
3	O
).	O
	O
We	O
found	O
that	O
the	O
primary	O
interaction	O
sites	O
on	O
Spy	O
reside	O
at	O
the	O
N	O
and	O
C	O
termini	O
(	O
Arg122	O
,	O
Thr124	O
,	O
and	O
Phe29	O
)	O
as	O
well	O
as	O
on	O
the	O
concave	O
face	O
of	O
the	O
chaperone	O
(	O
Arg61	O
,	O
Arg43	O
,	O
Lys47	O
,	O
His96	O
,	O
and	O
Met46	O
).	O
	O
Surprisingly	O
,	O
we	O
noted	O
that	O
in	O
the	O
ensemble	O
,	O
Im76	B-mutant
-	I-mutant
45	I-mutant
interacts	O
with	O
only	O
38	O
%	O
of	O
the	O
hydrophobic	O
residues	O
in	O
the	O
Spy	O
cradle	O
,	O
but	O
interacts	O
with	O
61	O
%	O
of	O
the	O
hydrophilic	O
residues	O
in	O
the	O
cradle	O
.	O
	O
With	O
respect	O
to	O
the	O
substrate	O
,	O
we	O
observed	O
that	O
nearly	O
every	O
residue	O
in	O
Im76	B-mutant
-	I-mutant
45	I-mutant
is	O
in	O
contact	O
with	O
Spy	O
(	O
Fig	O
.	O
4a	O
).	O
	O
However	O
,	O
we	O
did	O
notice	O
that	O
despite	O
this	O
uniformity	O
,	O
regions	O
of	O
Im76	B-mutant
-	I-mutant
45	I-mutant
preferentially	O
interact	O
with	O
different	O
regions	O
in	O
Spy	O
(	O
Fig	O
.	O
4b	O
).	O
	O
For	O
example	O
,	O
the	O
N	O
-	O
terminal	O
half	O
of	O
Im76	B-mutant
-	I-mutant
45	I-mutant
binds	O
more	O
consistently	O
in	O
the	O
Spy	O
cradle	O
,	O
whereas	O
the	O
C	O
-	O
terminal	O
half	O
predominantly	O
binds	O
to	O
the	O
outer	O
edges	O
of	O
Spy	O
’	O
s	O
concave	O
surface	O
.	O
	O
Not	O
unexpectedly	O
,	O
we	O
found	O
that	O
as	O
Im76	B-mutant
-	I-mutant
45	I-mutant
progresses	O
from	O
the	O
unfolded	O
to	O
the	O
native	O
state	O
,	O
its	O
interactions	O
with	O
Spy	O
shift	O
accordingly	O
.	O
	O
Whereas	O
the	O
least	O
-	O
folded	O
Im76	B-mutant
-	I-mutant
45	I-mutant
pose	O
in	O
the	O
ensemble	O
forms	O
the	O
most	O
hydrophobic	O
contacts	O
with	O
Spy	O
(	O
Fig	O
.	O
3	O
),	O
the	O
two	O
most	O
-	O
folded	O
conformations	O
form	O
the	O
fewest	O
hydrophobic	O
contacts	O
(	O
Fig	O
.	O
3	O
).	O
	O
Although	O
we	O
do	O
not	O
yet	O
have	O
time	O
resolution	O
data	O
of	O
these	O
various	O
snapshots	O
of	O
Im76	B-mutant
-	I-mutant
45	I-mutant
,	O
this	O
ensemble	O
illustrates	O
how	O
a	O
substrate	O
samples	O
its	O
folding	O
landscape	O
while	O
bound	O
to	O
a	O
chaperone	O
.	O
	O
Spy	O
changes	O
conformation	O
upon	O
substrate	O
binding	O
	O
Comparing	O
the	O
structure	O
of	O
Spy	O
in	O
its	O
substrate	O
-	O
bound	O
and	O
apo	O
states	O
revealed	O
that	O
the	O
Spy	O
dimer	O
also	O
undergoes	O
significant	O
conformational	O
changes	O
upon	O
substrate	O
binding	O
(	O
Fig	O
.	O
5a	O
and	O
Supplementary	O
Movie	O
2	O
).	O
	O
Upon	O
substrate	O
binding	O
,	O
the	O
Spy	O
dimer	O
twists	O
9	O
°	O
about	O
its	O
center	O
relative	O
to	O
its	O
apo	O
form	O
.	O
	O
This	O
twist	O
yields	O
asymmetry	O
and	O
results	O
in	O
substantially	O
different	O
interaction	O
patterns	O
in	O
the	O
two	O
Spy	O
monomers	O
(	O
Fig	O
.	O
4b	O
).	O
	O
This	O
increased	O
disorder	O
might	O
explain	O
how	O
Spy	O
is	O
able	O
to	O
recognize	O
and	O
bind	O
different	O
substrates	O
and	O
/	O
or	O
differing	O
conformations	O
of	O
the	O
same	O
substrate	O
.	O
	O
The	O
RMSD	O
between	O
the	O
well	O
-	O
folded	O
sections	O
of	O
Spy	O
in	O
the	O
four	O
chaperone	O
-	O
substrate	O
complexes	O
was	O
very	O
small	O
,	O
less	O
than	O
0	O
.	O
3	O
Å	O
.	O
Combined	O
with	O
competition	O
experiments	O
showing	O
that	O
the	O
substrates	O
compete	O
in	O
solution	O
for	O
Spy	O
binding	O
(	O
Fig	O
.	O
5c	O
and	O
Supplementary	O
Fig	O
.	O
8	O
),	O
we	O
conclude	O
that	O
all	O
the	O
tested	O
substrates	O
share	O
the	O
same	O
overall	O
Spy	O
binding	O
site	O
.	O
	O
To	O
shed	O
light	O
on	O
how	O
chaperones	O
interact	O
with	O
their	O
substrates	O
,	O
we	O
developed	O
a	O
novel	O
structural	O
biology	O
method	O
(	O
READ	O
)	O
and	O
applied	O
it	O
to	O
determine	O
a	O
conformational	O
ensemble	O
of	O
the	O
chaperone	O
Spy	O
bound	O
to	O
substrate	O
.	O
	O
As	O
a	O
substrate	O
,	O
we	O
used	O
Im76	B-mutant
-	I-mutant
45	I-mutant
,	O
the	O
chaperone	O
-	O
interacting	O
portion	O
of	O
the	O
protein	O
-	O
folding	O
model	O
protein	O
Im7	O
.	O
	O
In	O
the	O
chaperone	O
-	O
bound	O
ensemble	O
,	O
Im76	B-mutant
-	I-mutant
45	I-mutant
samples	O
unfolded	O
,	O
partially	O
folded	O
,	O
and	O
native	O
-	O
like	O
states	O
.	O
	O
The	O
ensemble	O
provides	O
an	O
unprecedented	O
description	O
of	O
the	O
conformations	O
that	O
a	O
substrate	O
assumes	O
while	O
exploring	O
its	O
chaperone	O
-	O
associated	O
folding	O
landscape	O
.	O
	O
We	O
recently	O
showed	O
that	O
Im7	O
can	O
fold	O
while	O
remaining	O
continuously	O
bound	O
to	O
Spy	O
.	O
	O
The	O
structures	O
of	O
our	O
ensemble	O
agree	O
well	O
with	O
lower	O
-	O
resolution	O
crosslinking	O
data	O
,	O
which	O
indicate	O
that	O
chaperone	O
-	O
substrate	O
interactions	O
primarily	O
occur	O
on	O
the	O
concave	O
surface	O
of	O
Spy	O
.	O
	O
The	O
ensemble	O
suggests	O
a	O
model	O
in	O
which	O
Spy	O
provides	O
an	O
amphipathic	O
surface	O
that	O
allows	O
substrate	O
proteins	O
to	O
assume	O
different	O
conformations	O
while	O
bound	O
to	O
the	O
chaperone	O
.	O
	O
This	O
model	O
is	O
consistent	O
with	O
previous	O
studies	O
postulating	O
that	O
the	O
flexible	O
binding	O
of	O
chaperones	O
allows	O
for	O
substrate	O
protein	O
folding	O
.	O
	O
The	O
amphipathic	O
concave	O
surface	O
of	O
Spy	O
likely	O
facilitates	O
this	O
flexible	O
binding	O
and	O
may	O
be	O
a	O
crucial	O
feature	O
for	O
Spy	O
and	O
potentially	O
other	O
chaperones	O
,	O
allowing	O
them	O
to	O
bind	O
multiple	O
conformations	O
of	O
many	O
different	O
substrates	O
.	O
	O
In	O
contrast	O
to	O
Spy	O
’	O
s	O
binding	O
hotspots	O
,	O
Im76	B-mutant
-	I-mutant
45	I-mutant
displays	O
substantially	O
less	O
specificity	O
in	O
its	O
binding	O
sites	O
.	O
	O
This	O
trend	O
suggests	O
that	O
complex	O
formation	O
between	O
an	O
ATP	O
-	O
independent	O
chaperone	O
and	O
its	O
unfolded	O
substrate	O
may	O
initially	O
involve	O
hydrophobic	O
interactions	O
,	O
effectively	O
shielding	O
the	O
exposed	O
aggregation	O
-	O
sensitive	O
hydrophobic	O
regions	O
in	O
the	O
substrate	O
.	O
	O
Notably	O
,	O
the	O
most	O
frequent	O
contacts	O
between	O
Spy	O
and	O
Im76	B-mutant
-	I-mutant
45	I-mutant
are	O
charge	O
-	O
charge	O
interactions	O
.	O
	O
The	O
negatively	O
charged	O
Im7	O
residues	O
Glu21	O
,	O
Asp32	O
,	O
and	O
Asp35	O
reside	O
on	O
the	O
surface	O
of	O
Im7	O
and	O
form	O
interactions	O
with	O
Spy	O
’	O
s	O
positively	O
charged	O
cradle	O
in	O
both	O
the	O
unfolded	O
and	O
native	O
-	O
like	O
states	O
.	O
	O
Interaction	O
with	O
Spy	O
’	O
s	O
positively	O
-	O
charged	O
residues	O
likely	O
relieves	O
the	O
charge	O
repulsion	O
between	O
Asp32	O
and	O
Asp35	O
,	O
promoting	O
their	O
compaction	O
into	O
a	O
helical	O
conformation	O
.	O
	O
As	O
inter	O
-	O
molecular	O
hydrophobic	O
interactions	O
between	O
Spy	O
and	O
the	O
substrate	O
become	O
progressively	O
replaced	O
by	O
intra	O
-	O
molecular	O
interactions	O
within	O
the	O
substrate	O
,	O
the	O
affinity	O
between	O
chaperone	O
and	O
substrates	O
could	O
decrease	O
,	O
eventually	O
leading	O
to	O
release	O
of	O
the	O
folded	O
client	O
protein	O
.	O
	O
Recently	O
,	O
we	O
employed	O
a	O
genetic	O
selection	O
system	O
to	O
improve	O
the	O
chaperone	O
activity	O
of	O
Spy	O
.	O
	O
This	O
selection	O
resulted	O
in	O
“	O
Super	O
Spy	O
”	O
variants	O
that	O
were	O
more	O
effective	O
at	O
both	O
preventing	O
aggregation	O
and	O
promoting	O
protein	O
folding	O
.	O
	O
Our	O
ensemble	O
revealed	O
that	O
two	O
of	O
the	O
Super	O
Spy	O
mutations	O
(	O
H96L	B-mutant
and	O
Q100L	B-mutant
)	O
form	O
part	O
of	O
the	O
chaperone	O
contact	O
surface	O
that	O
binds	O
to	O
Im76	B-mutant
-	I-mutant
45	I-mutant
(	O
Fig	O
.	O
4a	O
).	O
	O
By	O
sampling	O
multiple	O
conformations	O
,	O
this	O
linker	O
region	O
may	O
allow	O
diverse	O
substrate	O
conformations	O
to	O
be	O
accommodated	O
.	O
	O
Other	O
Super	O
Spy	O
mutations	O
(	O
F115I	B-mutant
and	O
F115L	B-mutant
)	O
caused	O
increased	O
flexibility	O
but	O
not	O
tighter	O
substrate	O
binding	O
.	O
	O
This	O
residue	O
does	O
not	O
directly	O
contact	O
Im76	B-mutant
-	I-mutant
45	I-mutant
in	O
our	O
READ	O
-	O
derived	O
ensemble	O
.	O
	O
This	O
interaction	O
presumably	O
reduces	O
the	O
mobility	O
of	O
the	O
C	O
-	O
terminal	O
helix	O
.	O
	O
The	O
F115I	B-mutant
/	O
L	B-mutant
substitutions	O
would	O
replace	O
these	O
hydrogen	O
bonds	O
with	O
hydrophobic	O
interactions	O
that	O
have	O
little	O
angular	O
dependence	O
.	O
	O
As	O
a	O
result	O
,	O
the	O
C	O
-	O
terminus	O
,	O
and	O
possibly	O
also	O
the	O
flexible	O
linker	O
,	O
is	O
likely	O
to	O
become	O
more	O
flexible	O
and	O
thus	O
more	O
accommodating	O
of	O
different	O
conformations	O
of	O
substrates	O
.	O
	O
Our	O
study	O
indicates	O
that	O
the	O
chaperone	O
Spy	O
employs	O
a	O
simple	O
surface	O
binding	O
approach	O
that	O
allows	O
the	O
substrate	O
to	O
explore	O
various	O
conformations	O
and	O
form	O
transiently	O
favorable	O
interactions	O
while	O
being	O
protected	O
from	O
aggregation	O
.	O
	O
We	O
speculate	O
that	O
many	O
other	O
chaperones	O
could	O
utilize	O
a	O
similar	O
strategy	O
.	O
	O
ATP	O
and	O
co	O
-	O
chaperone	O
dependencies	O
may	O
have	O
emerged	O
later	O
through	O
evolution	O
to	O
better	O
modulate	O
and	O
control	O
chaperone	O
action	O
.	O
	O
In	O
addition	O
to	O
insights	O
into	O
chaperone	O
function	O
,	O
this	O
work	O
presents	O
a	O
new	O
method	O
for	O
determining	O
heterogeneous	O
structural	O
ensembles	O
via	O
a	O
hybrid	O
methodology	O
of	O
X	O
-	O
ray	O
crystallography	O
and	O
computational	O
modeling	O
.	O
	O
Crystallographic	O
data	O
and	O
ensemble	O
selection	O
.	O
(	O
a	O
)	O
2mFo	O
−	O
DFc	O
omit	O
map	O
of	O
residual	O
Im76	B-mutant
-	I-mutant
45	I-mutant
and	O
flexible	O
linker	O
electron	O
density	O
contoured	O
at	O
0	O
.	O
5	O
σ	O
.	O
	O
This	O
is	O
the	O
residual	O
density	O
that	O
is	O
used	O
in	O
the	O
READ	O
selection	O
.	O
	O
(	O
b	O
)	O
Composites	O
of	O
iodine	O
positions	O
detected	O
from	O
anomalous	O
signals	O
using	O
pI	O
-	O
Phe	O
substitutions	O
,	O
colored	O
and	O
numbered	O
by	O
sequence	O
.	O
	O
Multiple	O
iodine	O
positions	O
were	O
detected	O
for	O
most	O
residues	O
.	O
	O
Agreement	O
to	O
the	O
residual	O
Im76	B-mutant
-	I-mutant
45	I-mutant
electron	O
density	O
(	O
c	O
)	O
and	O
anomalous	O
iodine	O
signals	O
(	O
d	O
)	O
for	O
ensembles	O
of	O
varying	O
size	O
generated	O
by	O
randomly	O
choosing	O
from	O
the	O
MD	O
pool	O
(	O
blue	O
)	O
and	O
from	O
the	O
selection	O
procedure	O
(	O
black	O
).	O
	O
The	O
cost	O
function	O
,	O
χ2	O
,	O
decreases	O
as	O
the	O
agreement	O
to	O
the	O
experimental	O
data	O
increases	O
and	O
is	O
defined	O
in	O
the	O
Online	O
Methods	O
.	O
	O
Spy	O
:	O
Im76	O
-	O
45	O
ensemble	O
,	O
arranged	O
by	O
RMSD	O
to	O
native	O
state	O
of	O
Im76	B-mutant
-	I-mutant
45	I-mutant
.	O
Although	O
the	O
six	O
-	O
membered	O
ensemble	O
from	O
the	O
READ	O
selection	O
should	O
be	O
considered	O
only	O
as	O
an	O
ensemble	O
,	O
for	O
clarity	O
,	O
the	O
individual	O
conformers	O
are	O
shown	O
separately	O
here	O
.	O
	O
Atoms	O
that	O
were	O
either	O
not	O
directly	O
selected	O
in	O
the	O
READ	O
procedure	O
,	O
or	O
whose	O
position	O
could	O
not	O
be	O
justified	O
based	O
on	O
agreement	O
with	O
the	O
residual	O
electron	O
density	O
were	O
removed	O
,	O
leading	O
to	O
non	O
-	O
contiguous	O
sections	O
.	O
	O
Dashed	O
lines	O
connect	O
non	O
-	O
contiguous	O
segments	O
of	O
the	O
Im76	B-mutant
-	I-mutant
45	I-mutant
substrate	O
.	O
	O
Residues	O
of	O
the	O
Spy	O
flexible	O
linker	O
region	O
that	O
fit	O
the	O
residual	O
electron	O
density	O
are	O
shown	O
as	O
larger	O
gray	O
spheres	O
.	O
	O
Shown	O
below	O
each	O
ensemble	O
member	O
is	O
the	O
RMSD	O
of	O
each	O
conformer	O
to	O
the	O
native	O
state	O
of	O
Im76	B-mutant
-	I-mutant
45	I-mutant
,	O
as	O
well	O
as	O
the	O
percentage	O
of	O
contacts	O
between	O
Im76	B-mutant
-	I-mutant
45	I-mutant
and	O
Spy	O
that	O
are	O
hydrophobic	O
.	O
	O
Contact	O
maps	O
of	O
Spy	O
:	O
Im76	O
-	O
45	O
complex	O
.	O
	O
For	O
clarity	O
,	O
Im76	B-mutant
-	I-mutant
45	I-mutant
is	O
represented	O
with	O
a	O
single	O
conformation	O
.	O
	O
As	O
the	O
residues	O
involved	O
in	O
contacts	O
are	O
more	O
evenly	O
distributed	O
in	O
Im76	B-mutant
-	I-mutant
45	I-mutant
compared	O
to	O
Spy	O
,	O
its	O
contact	O
map	O
was	O
amplified	O
.	O
(	O
b	O
)	O
Detailed	O
contact	O
maps	O
of	O
Spy	O
:	O
Im76	O
-	O
45	O
.	O
	O
Contacts	O
to	O
the	O
two	O
Spy	O
monomers	O
are	O
depicted	O
separately	O
.	O
	O
Note	O
that	O
the	O
flexible	O
linker	O
region	O
of	O
Spy	O
(	O
residues	O
47	O
–	O
57	O
)	O
is	O
not	O
represented	O
in	O
the	O
2D	O
contact	O
maps	O
.	O
	O
Spy	O
conformation	O
changes	O
upon	O
substrate	O
binding	O
.	O
	O
Flexibility	O
of	O
Spy	O
linker	O
region	O
and	O
effect	O
of	O
Super	O
Spy	O
mutants	O
.	O
(	O
a	O
)	O
The	O
Spy	O
linker	O
region	O
adopts	O
one	O
dominant	O
conformation	O
in	O
its	O
apo	O
state	O
(	O
PDB	O
ID	O
3039	O
,	O
gray	O
),	O
but	O
expands	O
and	O
adopts	O
multiple	O
conformations	O
in	O
bound	O
states	O
(	O
green	O
).	O
	O
The	O
Super	O
Spy	O
mutants	O
F115L	B-mutant
,	O
F115I	B-mutant
,	O
and	O
L32P	B-mutant
are	O
proposed	O
to	O
gain	O
activity	O
by	O
increasing	O
the	O
flexibility	O
or	O
size	O
of	O
this	O
linker	O
region	O
.	O
	O
L32	O
,	O
F115	O
,	O
and	O
Y104	O
are	O
rendered	O
in	O
purple	O
to	O
illustrate	O
residues	O
that	O
are	O
most	O
affected	O
by	O
Super	O
Spy	O
mutations	O
;	O
CH	O
⋯	O
π	O
hydrogen	O
bonds	O
are	O
depicted	O
by	O
orange	O
dashes	O
.	O
	O
Mechanism	O
of	O
extracellular	O
ion	O
exchange	O
and	O
binding	O
-	O
site	O
occlusion	O
in	O
the	O
sodium	O
-	O
calcium	O
exchanger	O
	O
Na	O
+/	O
Ca2	O
+	O
exchangers	O
utilize	O
the	O
Na	O
+	O
electrochemical	O
gradient	O
across	O
the	O
plasma	O
membrane	O
to	O
extrude	O
intracellular	O
Ca2	O
+,	O
and	O
play	O
a	O
central	O
role	O
in	O
Ca2	O
+	O
homeostasis	O
.	O
	O
Here	O
,	O
we	O
elucidate	O
their	O
mechanisms	O
of	O
extracellular	O
ion	O
recognition	O
and	O
exchange	O
through	O
a	O
structural	O
analysis	O
of	O
the	O
exchanger	O
from	O
Methanococcus	O
jannaschii	O
(	O
NCX_Mj	O
)	O
bound	O
to	O
Na	O
+,	O
Ca2	O
+	O
or	O
Sr2	O
+	O
in	O
various	O
occupancies	O
and	O
in	O
an	O
apo	O
state	O
.	O
	O
This	O
analysis	O
defines	O
the	O
binding	O
mode	O
and	O
relative	O
affinity	O
of	O
these	O
ions	O
,	O
establishes	O
the	O
structural	O
basis	O
for	O
the	O
anticipated	O
3Na	O
+:	O
1Ca2	O
+	O
exchange	O
stoichiometry	O
,	O
and	O
reveals	O
the	O
conformational	O
changes	O
at	O
the	O
onset	O
of	O
the	O
alternating	O
-	O
access	O
transport	O
mechanism	O
.	O
	O
An	O
independent	O
analysis	O
of	O
the	O
dynamics	O
and	O
conformational	O
free	O
-	O
energy	O
landscape	O
of	O
NCX_Mj	O
in	O
different	O
ion	O
-	O
occupancy	O
states	O
,	O
based	O
on	O
enhanced	O
-	O
sampling	O
molecular	O
-	O
dynamics	O
simulations	O
,	O
demonstrates	O
that	O
the	O
crystal	O
structures	O
reflect	O
mechanistically	O
relevant	O
,	O
interconverting	O
conformations	O
.	O
	O
These	O
calculations	O
also	O
reveal	O
the	O
mechanism	O
by	O
which	O
the	O
outward	O
-	O
to	O
-	O
inward	O
transition	O
is	O
controlled	O
by	O
the	O
ion	O
-	O
occupancy	O
state	O
,	O
thereby	O
explaining	O
the	O
emergence	O
of	O
strictly	O
-	O
coupled	O
Na	O
+/	O
Ca2	O
+	O
antiport	O
.	O
	O
Na	O
+/	O
Ca2	O
+	O
exchangers	O
(	O
NCX	O
)	O
play	O
physiologically	O
essential	O
roles	O
in	O
Ca2	O
+	O
signaling	O
and	O
homeostasis	O
.	O
	O
NCX	O
catalyzes	O
the	O
uphill	O
extrusion	O
of	O
intracellular	O
Ca2	O
+	O
across	O
the	O
cell	O
membrane	O
,	O
by	O
coupling	O
this	O
process	O
to	O
the	O
downhill	O
permeation	O
of	O
Na	O
+	O
into	O
the	O
cell	O
,	O
with	O
a	O
3	O
Na	O
+	O
to	O
1	O
Ca2	O
+	O
stoichiometry	O
.	O
	O
The	O
basic	O
functional	O
unit	O
for	O
ion	O
transport	O
in	O
NCX	O
consists	O
of	O
ten	O
membrane	O
-	O
spanning	O
segments	O
,	O
comprising	O
two	O
homologous	O
halves	O
.	O
	O
Each	O
of	O
these	O
halves	O
contains	O
a	O
highly	O
conserved	O
region	O
,	O
referred	O
to	O
as	O
α	O
-	O
repeat	O
,	O
known	O
to	O
be	O
important	O
for	O
ion	O
binding	O
and	O
translocation	O
;	O
in	O
eukaryotic	O
NCX	O
,	O
the	O
two	O
halves	O
are	O
connected	O
by	O
a	O
large	O
intracellular	O
regulatory	O
domain	O
,	O
which	O
is	O
absent	O
in	O
microbial	O
NCX	O
(	O
Supplementary	O
Fig	O
.	O
1	O
).	O
	O
Our	O
recent	O
atomic	O
-	O
resolution	O
structure	O
of	O
NCX_Mj	O
from	O
Methanococcus	O
jannaschii	O
provided	O
the	O
first	O
view	O
of	O
the	O
basic	O
functional	O
unit	O
of	O
an	O
NCX	O
protein	O
.	O
	O
This	O
structure	O
shows	O
the	O
exchanger	O
in	O
an	O
outward	O
-	O
facing	O
conformation	O
and	O
reveals	O
four	O
putative	O
ion	O
-	O
binding	O
sites	O
,	O
denominated	O
internal	O
(	O
Sint	O
),	O
external	O
(	O
Sext	O
),	O
Ca2	O
+-	O
binding	O
(	O
SCa	O
)	O
and	O
middle	O
(	O
Smid	O
),	O
clustered	O
in	O
the	O
center	O
of	O
the	O
protein	O
and	O
occluded	O
from	O
the	O
solvent	O
(	O
Fig	O
.	O
1a	O
-	O
b	O
).	O
	O
These	O
structures	O
reveal	O
the	O
mode	O
of	O
recognition	O
of	O
these	O
ions	O
,	O
their	O
relative	O
affinities	O
,	O
and	O
the	O
mechanism	O
of	O
extracellular	O
ion	O
exchange	O
,	O
for	O
a	O
well	O
-	O
defined	O
,	O
functional	O
conformation	O
in	O
a	O
membrane	O
-	O
like	O
environment	O
.	O
	O
An	O
independent	O
analysis	O
based	O
on	O
molecular	O
-	O
dynamics	O
simulations	O
demonstrates	O
that	O
the	O
structures	O
capture	O
mechanistically	O
relevant	O
states	O
.	O
	O
These	O
calculations	O
also	O
reveal	O
how	O
the	O
ion	O
occupancy	O
state	O
of	O
the	O
outward	O
-	O
facing	O
exchanger	O
determines	O
the	O
feasibility	O
of	O
the	O
transition	O
to	O
the	O
inward	O
-	O
facing	O
conformation	O
,	O
thereby	O
addressing	O
a	O
key	O
outstanding	O
question	O
in	O
secondary	O
-	O
active	O
transport	O
,	O
namely	O
how	O
the	O
transported	O
substrates	O
control	O
the	O
alternating	O
-	O
access	O
mechanism	O
.	O
	O
Crystals	O
were	O
grown	O
in	O
150	O
mM	O
NaCl	O
using	O
the	O
lipidic	O
cubic	O
phase	O
(	O
LCP	O
)	O
technique	O
.	O
	O
The	O
crystallization	O
solutions	O
around	O
the	O
LCP	O
droplets	O
were	O
then	O
slowly	O
replaced	O
by	O
solutions	O
containing	O
different	O
concentrations	O
of	O
NaCl	O
and	O
EGTA	O
(	O
Methods	O
).	O
	O
Occupancy	O
refinement	O
indicated	O
two	O
Na	O
+	O
ions	O
bind	O
to	O
Sint	O
and	O
SCa	O
at	O
low	O
Na	O
+	O
concentrations	O
(	O
Fig	O
.	O
1c	O
),	O
with	O
a	O
slight	O
preference	O
for	O
Sint	O
(	O
Table	O
1	O
).	O
	O
Binding	O
of	O
a	O
third	O
Na	O
+	O
to	O
Sext	O
occurs	O
at	O
higher	O
concentrations	O
,	O
as	O
no	O
density	O
was	O
observed	O
there	O
at	O
10	O
mM	O
Na	O
+	O
or	O
lower	O
(	O
Fig	O
.	O
1c	O
);	O
Sext	O
is	O
however	O
partially	O
occupied	O
at	O
20	O
mM	O
Na	O
+,	O
and	O
fully	O
occupied	O
at	O
150	O
mM	O
(	O
Fig	O
.	O
1c	O
).	O
	O
The	O
Na	O
+	O
occupation	O
at	O
SCa	O
,	O
compounded	O
with	O
the	O
expected	O
3Na	O
+:	O
1Ca2	O
+	O
stoichiometry	O
,	O
implies	O
our	O
previous	O
assignment	O
of	O
the	O
Smid	O
site	O
must	O
be	O
re	O
-	O
evaluated	O
.	O
	O
Second	O
,	O
the	O
protein	O
coordination	O
geometry	O
at	O
Smid	O
is	O
clearly	O
suboptimal	O
for	O
Na	O
+	O
(	O
Supplementary	O
Fig	O
.	O
1d	O
).	O
	O
The	O
water	O
molecule	O
at	O
Smid	O
forms	O
hydrogen	O
-	O
bonds	O
with	O
the	O
highly	O
conserved	O
Glu54	O
and	O
Glu213	O
(	O
Supplementary	O
Fig	O
.	O
1d	O
),	O
stabilizing	O
their	O
orientation	O
to	O
properly	O
coordinate	O
multiple	O
Na	O
+	O
ions	O
at	O
Sext	O
,	O
SCa	O
and	O
Sint	O
.	O
	O
It	O
can	O
be	O
inferred	O
from	O
this	O
assignment	O
that	O
Glu54	O
and	O
Glu213	O
are	O
ionized	O
,	O
while	O
Asp240	O
,	O
which	O
flanks	O
Smid	O
(	O
and	O
is	O
replaced	O
by	O
Asn	O
in	O
eukaryotic	O
NCX	O
)	O
would	O
be	O
protonated	O
,	O
as	O
indicated	O
by	O
the	O
abovementioned	O
simulation	O
study	O
.	O
	O
Na	O
+-	O
dependent	O
conformational	O
change	O
	O
The	O
NCX_Mj	O
structures	O
in	O
various	O
Na	O
+	O
concentrations	O
also	O
reveal	O
that	O
Na	O
+	O
binding	O
to	O
Sext	O
is	O
coupled	O
to	O
a	O
subtle	O
but	O
important	O
conformational	O
change	O
(	O
Fig	O
.	O
2	O
).	O
	O
However	O
,	O
when	O
Sext	O
becomes	O
empty	O
at	O
low	O
Na	O
+	O
concentrations	O
,	O
TM7a	O
and	O
TM7b	O
become	O
a	O
continuous	O
straight	O
helix	O
(	O
Fig	O
.	O
2a	O
),	O
and	O
the	O
carbonyl	O
group	O
of	O
Ala206	O
retracts	O
away	O
(	O
Fig	O
.	O
2b	O
-	O
d	O
).	O
	O
TM7a	O
also	O
forms	O
hydrophobic	O
contacts	O
with	O
the	O
C	O
-	O
terminal	O
half	O
of	O
TM6	O
.	O
	O
These	O
contacts	O
are	O
absent	O
in	O
the	O
structure	O
with	O
Na	O
+	O
at	O
Sext	O
,	O
in	O
which	O
there	O
is	O
an	O
open	O
gap	O
between	O
the	O
two	O
helices	O
(	O
Fig	O
.	O
2b	O
).	O
	O
This	O
difference	O
is	O
noteworthy	O
because	O
TM6	O
and	O
TM1	O
are	O
believed	O
to	O
undergo	O
a	O
sliding	O
motion	O
,	O
relative	O
to	O
the	O
rest	O
of	O
the	O
protein	O
,	O
when	O
the	O
transporter	O
switches	O
to	O
the	O
inward	O
-	O
facing	O
conformation	O
.	O
	O
The	O
straightening	O
of	O
TM7ab	O
also	O
opens	O
up	O
a	O
passageway	O
from	O
the	O
external	O
solution	O
to	O
Sext	O
and	O
Smid	O
,	O
while	O
SCa	O
and	O
Sint	O
remain	O
occluded	O
(	O
Fig	O
.	O
2d	O
).	O
	O
Thus	O
,	O
the	O
structures	O
at	O
high	O
and	O
low	O
Na	O
+	O
concentrations	O
represent	O
the	O
outward	O
-	O
facing	O
occluded	O
and	O
partially	O
open	O
states	O
,	O
respectively	O
.	O
	O
This	O
conformational	O
change	O
is	O
dependent	O
on	O
the	O
Na	O
+	O
occupancy	O
of	O
Sext	O
and	O
occurs	O
when	O
Na	O
+	O
already	O
occupies	O
Sint	O
and	O
SCa	O
.	O
	O
Our	O
crystallographic	O
titration	O
experiment	O
indicates	O
that	O
the	O
K1	O
/	O
2	O
of	O
this	O
Na	O
+-	O
driven	O
conformational	O
transition	O
is	O
~	O
20	O
mM	O
.	O
At	O
this	O
concentration	O
,	O
Sext	O
is	O
partially	O
occupied	O
and	O
the	O
NCX_Mj	O
crystal	O
is	O
a	O
mixture	O
of	O
both	O
the	O
occluded	O
and	O
partially	O
open	O
conformations	O
.	O
	O
The	O
finding	O
that	O
the	O
Na	O
+	O
occupancy	O
change	O
from	O
2	O
to	O
3	O
ions	O
coincides	O
with	O
a	O
conformational	O
change	O
of	O
the	O
transporter	O
also	O
provides	O
a	O
rationale	O
to	O
the	O
Hill	O
coefficient	O
of	O
the	O
Na	O
+-	O
dependent	O
activation	O
process	O
in	O
eukaryotic	O
NCX	O
.	O
	O
To	O
determine	O
how	O
Ca2	O
+	O
binds	O
to	O
NCX_Mj	O
and	O
competes	O
with	O
Na	O
+,	O
we	O
first	O
titrated	O
the	O
crystals	O
with	O
Sr2	O
+	O
(	O
Methods	O
).	O
	O
Sr2	O
+	O
is	O
transported	O
by	O
NCX	O
similarly	O
to	O
Ca2	O
+	O
,	O
and	O
is	O
distinguishable	O
from	O
Na	O
+	O
by	O
its	O
greater	O
electron	O
-	O
density	O
intensity	O
.	O
	O
Protein	O
crystals	O
soaked	O
with	O
10	O
mM	O
Sr2	O
+	O
and	O
2	O
.	O
5	O
mM	O
Na	O
+	O
revealed	O
a	O
strong	O
electron	O
-	O
density	O
peak	O
at	O
site	O
SCa	O
,	O
indicating	O
binding	O
of	O
a	O
single	O
Sr2	O
+	O
ion	O
(	O
Fig	O
.	O
3a	O
).	O
	O
The	O
Sr2	O
+-	O
loaded	O
NCX_Mj	O
structure	O
adopts	O
the	O
partially	O
open	O
conformation	O
observed	O
at	O
low	O
Na	O
+	O
concentrations	O
.	O
	O
Binding	O
of	O
Sr2	O
+,	O
however	O
,	O
excludes	O
Na	O
+	O
entirely	O
.	O
	O
Crystal	O
titrations	O
with	O
decreasing	O
Sr2	O
+	O
or	O
increasing	O
Na	O
+	O
demonstrated	O
that	O
Sr2	O
+	O
binds	O
to	O
the	O
outward	O
-	O
facing	O
NCX_Mj	O
with	O
low	O
affinity	O
,	O
and	O
that	O
it	O
can	O
be	O
out	O
-	O
competed	O
by	O
Na	O
+	O
even	O
at	O
low	O
concentrations	O
(	O
Supplementary	O
Note	O
1	O
and	O
Supplementary	O
Fig	O
.	O
2a	O
-	O
b	O
).	O
	O
Thus	O
,	O
in	O
100	O
mM	O
Na	O
+	O
and	O
10	O
mM	O
Sr2	O
+,	O
Na	O
+	O
completely	O
replaced	O
Sr2	O
+	O
(	O
Fig	O
.	O
3a	O
)	O
and	O
reverted	O
NCX_Mj	O
to	O
the	O
Na	O
+-	O
loaded	O
,	O
fully	O
occluded	O
state	O
.	O
	O
Similar	O
titration	O
experiments	O
showed	O
that	O
Ca2	O
+	O
and	O
Sr2	O
+	O
binding	O
to	O
NCX_Mj	O
are	O
not	O
exactly	O
alike	O
The	O
electron	O
density	O
distribution	O
from	O
crystals	O
soaked	O
in	O
high	O
Ca2	O
+	O
and	O
low	O
Na	O
+,	O
indicates	O
that	O
Ca2	O
+	O
can	O
bind	O
to	O
Smid	O
as	O
well	O
as	O
SCa	O
,	O
with	O
a	O
preference	O
for	O
SCa	O
(	O
Fig	O
.	O
3b	O
).	O
	O
The	O
partial	O
Ca2	O
+	O
occupancy	O
at	O
Smid	O
is	O
likely	O
caused	O
by	O
Asp240	O
,	O
which	O
flanks	O
this	O
site	O
and	O
can	O
in	O
principle	O
coordinate	O
Ca2	O
+.	O
	O
Previous	O
functional	O
and	O
computational	O
studies	O
,	O
however	O
,	O
indicate	O
Asp240	O
becomes	O
protonated	O
during	O
transport	O
.	O
	O
SCa	O
is	O
therefore	O
the	O
functional	O
Ca2	O
+	O
site	O
.	O
	O
Specifically	O
,	O
our	O
crystallographic	O
titration	O
assay	O
indicates	O
Ca2	O
+	O
binds	O
with	O
sub	O
-	O
millimolar	O
affinity	O
,	O
in	O
good	O
agreement	O
with	O
the	O
external	O
apparent	O
Ca2	O
+	O
affinities	O
deduced	O
functionally	O
for	O
cardiac	O
NCX	O
(	O
Km	O
~	O
0	O
.	O
32	O
mM	O
)	O
and	O
NCX_Mj	O
(	O
Km	O
~	O
0	O
.	O
175	O
mM	O
).	O
	O
Taken	O
together	O
,	O
these	O
crystal	O
titration	O
experiments	O
demonstrate	O
that	O
the	O
four	O
binding	O
sites	O
in	O
outward	O
-	O
facing	O
NCX_Mj	O
exhibit	O
different	O
specificity	O
:	O
Sint	O
and	O
Sext	O
are	O
Na	O
+	O
specific	O
whereas	O
SCa	O
,	O
previously	O
hypothesized	O
to	O
be	O
Ca2	O
+	O
specific	O
,	O
can	O
also	O
bind	O
Na	O
+,	O
confirming	O
our	O
earlier	O
simulation	O
study	O
,	O
as	O
well	O
as	O
Sr2	O
+;	O
Smid	O
can	O
also	O
transiently	O
accommodate	O
Ca2	O
+	O
but	O
during	O
transport	O
Smid	O
is	O
most	O
likely	O
occupied	O
by	O
water	O
.	O
	O
The	O
ion	O
-	O
binding	O
sites	O
in	O
NCX_Mj	O
can	O
therefore	O
accommodate	O
up	O
to	O
three	O
Na	O
+	O
ions	O
or	O
a	O
single	O
divalent	O
ion	O
,	O
and	O
occupancy	O
by	O
Na	O
+	O
and	O
Ca2	O
+	O
(	O
or	O
Sr2	O
+)	O
are	O
mutually	O
exclusive	O
,	O
as	O
was	O
deduced	O
for	O
eukaryotic	O
exchangers	O
.	O
	O
A	O
structure	O
of	O
NCX_Mj	O
without	O
Na	O
+	O
or	O
Ca2	O
+	O
bound	O
	O
An	O
apo	O
state	O
of	O
outward	O
-	O
facing	O
NCX_Mj	O
is	O
likely	O
to	O
exist	O
transiently	O
in	O
physiological	O
conditions	O
,	O
despite	O
the	O
high	O
amounts	O
of	O
extracellular	O
Na	O
+	O
(~	O
150	O
mM	O
)	O
and	O
Ca2	O
+	O
(~	O
2	O
mM	O
).	O
	O
This	O
structure	O
is	O
similar	O
to	O
the	O
partially	O
open	O
structure	O
with	O
two	O
Na	O
+	O
or	O
either	O
one	O
Ca2	O
+	O
or	O
one	O
Sr2	O
+	O
ion	O
,	O
with	O
two	O
noticeable	O
differences	O
.	O
	O
First	O
,	O
TM7ab	O
along	O
with	O
the	O
extracellular	O
half	O
of	O
the	O
TM6	O
and	O
TM1	O
swing	O
further	O
away	O
from	O
the	O
protein	O
core	O
(	O
Fig	O
.	O
3c	O
),	O
resulting	O
in	O
a	O
slightly	O
wider	O
passageway	O
into	O
the	O
binding	O
sites	O
.	O
	O
Second	O
,	O
Glu54	O
and	O
Glu213	O
side	O
chains	O
rotate	O
away	O
from	O
the	O
binding	O
sites	O
and	O
appear	O
to	O
form	O
hydrogen	O
-	O
bonds	O
with	O
residues	O
involved	O
in	O
ion	O
coordination	O
in	O
the	O
fully	O
Na	O
+-	O
loaded	O
structure	O
(	O
Fig	O
.	O
3d	O
).	O
	O
This	O
apo	O
structure	O
might	O
therefore	O
represent	O
the	O
unloaded	O
,	O
open	O
state	O
of	O
outward	O
-	O
facing	O
NCX_Mj	O
.	O
	O
Alternatively	O
,	O
this	O
structure	O
might	O
capture	O
a	O
fully	O
protonated	O
state	O
of	O
the	O
transporter	O
,	O
to	O
which	O
Na	O
+	O
and	O
Ca2	O
+	O
cannot	O
bind	O
.	O
	O
Indeed	O
,	O
transport	O
assays	O
of	O
NCX_Mj	O
have	O
shown	O
that	O
even	O
in	O
the	O
presence	O
of	O
Na	O
+	O
or	O
Ca2	O
+,	O
low	O
pH	O
inactivates	O
the	O
transport	O
cycle	O
.	O
	O
Ion	O
occupancy	O
determines	O
the	O
free	O
-	O
energy	O
landscape	O
of	O
NCX_Mj	O
	O
NCX	O
must	O
be	O
loaded	O
either	O
with	O
3	O
Na	O
+	O
or	O
1	O
Ca2	O
+,	O
and	O
therefore	O
functions	O
as	O
an	O
antiporter	O
;	O
symporters	O
,	O
by	O
contrast	O
,	O
undergo	O
the	O
alternating	O
-	O
access	O
transition	O
only	O
when	O
all	O
substrates	O
and	O
coupling	O
ions	O
are	O
concurrently	O
bound	O
,	O
or	O
in	O
the	O
apo	O
state	O
.	O
	O
A	O
series	O
of	O
exploratory	O
MD	O
simulations	O
was	O
initially	O
carried	O
out	O
to	O
examine	O
what	O
features	O
of	O
the	O
NCX_Mj	O
structure	O
might	O
depend	O
on	O
the	O
ion	O
-	O
binding	O
sites	O
occupancy	O
.	O
	O
Specifically	O
,	O
we	O
first	O
simulated	O
the	O
outward	O
-	O
occluded	O
form	O
,	O
in	O
the	O
ion	O
configuration	O
we	O
previously	O
predicted	O
,	O
now	O
confirmed	O
by	O
the	O
high	O
-	O
Na	O
+	O
crystal	O
structure	O
described	O
above	O
(	O
Fig	O
.	O
1b	O
).	O
	O
That	O
is	O
,	O
Na	O
+	O
ions	O
occupy	O
Sext	O
,	O
SCa	O
,	O
and	O
Sint	O
,	O
while	O
D240	O
is	O
protonated	O
and	O
a	O
water	O
molecule	O
occupies	O
Smid	O
.	O
	O
The	O
Na	O
+	O
ion	O
at	O
Sext	O
was	O
then	O
relocated	O
from	O
the	O
site	O
to	O
the	O
bulk	O
solution	O
(	O
Methods	O
),	O
and	O
this	O
system	O
was	O
then	O
allowed	O
to	O
evolve	O
freely	O
in	O
time	O
.	O
	O
The	O
Na	O
+	O
ions	O
at	O
SCa	O
and	O
Sint	O
were	O
displaced	O
subsequently	O
,	O
and	O
an	O
analogous	O
simulation	O
was	O
then	O
carried	O
out	O
.	O
	O
The	O
most	O
notable	O
change	O
upon	O
displacement	O
of	O
Na	O
+	O
from	O
Sext	O
was	O
the	O
straightening	O
of	O
TM7ab	O
(	O
Fig	O
.	O
4a	O
).	O
	O
When	O
3	O
Na	O
+	O
ions	O
are	O
bound	O
,	O
TM7ab	O
primarily	O
folds	O
as	O
two	O
distinct	O
,	O
non	O
-	O
collinear	O
α	O
-	O
helical	O
fragments	O
,	O
owing	O
to	O
the	O
loss	O
of	O
the	O
backbone	O
carbonyl	O
-	O
amide	O
hydrogen	O
-	O
bonds	O
between	O
F202	O
and	O
A206	O
,	O
and	O
T203	O
and	O
F207	O
(	O
Fig	O
.	O
4b	O
).	O
	O
With	O
Sext	O
empty	O
,	O
however	O
,	O
TM7ab	O
forms	O
a	O
canonical	O
α	O
-	O
helix	O
(	O
Fig	O
.	O
4a	O
-	O
b	O
,	O
4g	O
),	O
thereby	O
creating	O
an	O
opening	O
between	O
TM3	O
and	O
TM7	O
,	O
which	O
in	O
turn	O
allows	O
water	O
molecules	O
from	O
the	O
external	O
solution	O
to	O
reach	O
into	O
Sext	O
(	O
Fig	O
.	O
4e	O
,	O
4h	O
-	O
i	O
),	O
i	O
.	O
e	O
.	O
the	O
transporter	O
is	O
no	O
longer	O
occluded	O
.	O
	O
Displacement	O
of	O
Na	O
+	O
from	O
SCa	O
and	O
Sint	O
induces	O
further	O
changes	O
(	O
Fig	O
.	O
4c	O
).	O
	O
Together	O
,	O
these	O
changes	O
open	O
a	O
second	O
aqueous	O
channel	O
leading	O
directly	O
into	O
SCa	O
and	O
Sint	O
(	O
Fig	O
.	O
4f	O
,	O
Fig	O
.	O
4h	O
-	O
i	O
).	O
	O
The	O
transporter	O
thus	O
becomes	O
fully	O
outward	O
-	O
open	O
.	O
	O
As	O
above	O
,	O
we	O
initially	O
examined	O
three	O
occupancy	O
states	O
,	O
namely	O
with	O
Na	O
+	O
in	O
Sext	O
,	O
SCa	O
and	O
Sint	O
,	O
with	O
Na	O
+	O
only	O
at	O
SCa	O
and	O
Sint	O
,	O
and	O
without	O
Na	O
+.	O
	O
These	O
calculations	O
demonstrate	O
that	O
the	O
Na	O
+	O
occupancy	O
state	O
of	O
the	O
transporter	O
has	O
a	O
profound	O
effect	O
on	O
its	O
conformational	O
free	O
-	O
energy	O
landscape	O
.	O
	O
At	O
a	O
small	O
energetic	O
cost	O
,	O
however	O
,	O
the	O
transporter	O
can	O
adopt	O
a	O
metastable	O
‘	O
half	O
-	O
open	O
’	O
conformation	O
in	O
which	O
TM7ab	O
is	O
completely	O
straight	O
and	O
Sext	O
is	O
open	O
to	O
the	O
exterior	O
(	O
Fig	O
.	O
5a	O
,	O
5b	O
).	O
	O
This	O
semi	O
-	O
open	O
conformation	O
is	O
nearly	O
identical	O
to	O
that	O
found	O
to	O
be	O
the	O
most	O
probable	O
when	O
Na	O
+	O
occupies	O
only	O
SCa	O
and	O
Sint	O
(	O
2	O
×	O
Na	O
+,	O
Fig	O
.	O
5a	O
),	O
demonstrating	O
that	O
binding	O
(	O
or	O
release	O
)	O
of	O
Na	O
+	O
to	O
Sext	O
occurs	O
in	O
this	O
metastable	O
conformation	O
.	O
	O
Crucially	O
,	O
though	O
,	O
the	O
free	O
-	O
energy	O
landscape	O
for	O
this	O
partially	O
occupied	O
state	O
demonstrates	O
that	O
the	O
occluded	O
conformation	O
is	O
no	O
longer	O
energetically	O
feasible	O
(	O
Fig	O
.	O
5a	O
).	O
	O
Displacement	O
of	O
the	O
two	O
remaining	O
Na	O
+	O
ions	O
from	O
SCa	O
and	O
Sint	O
further	O
reshapes	O
the	O
free	O
-	O
energy	O
landscape	O
of	O
the	O
transporter	O
(	O
No	O
ions	O
,	O
Fig	O
.	O
5a	O
),	O
which	O
now	O
can	O
only	O
adopt	O
a	O
fully	O
open	O
state	O
featuring	O
the	O
two	O
aqueous	O
channels	O
(	O
Fig	O
.	O
5b	O
-	O
c	O
).	O
	O
The	O
transition	O
to	O
the	O
occluded	O
state	O
in	O
this	O
apo	O
state	O
is	O
again	O
energetically	O
unfeasible	O
.	O
	O
From	O
a	O
mechanistic	O
standpoint	O
,	O
it	O
is	O
satisfying	O
to	O
observe	O
how	O
the	O
open	O
and	O
semi	O
-	O
open	O
states	O
are	O
each	O
compatible	O
with	O
two	O
different	O
Na	O
+	O
occupancies	O
,	O
explaining	O
how	O
sequential	O
Na	O
+	O
binding	O
to	O
energetically	O
accessible	O
conformations	O
(	O
prior	O
to	O
those	O
binding	O
events	O
)	O
progressively	O
reshape	O
the	O
free	O
-	O
energy	O
landscape	O
of	O
the	O
transporter	O
;	O
by	O
contrast	O
,	O
the	O
occluded	O
conformation	O
is	O
forbidden	O
unless	O
the	O
Na	O
+	O
occupancy	O
is	O
complete	O
.	O
	O
To	O
assess	O
this	O
hypothesis	O
,	O
we	O
carried	O
out	O
enhanced	O
-	O
sampling	O
simulations	O
for	O
the	O
Ca2	O
+	O
and	O
H	O
+-	O
bound	O
states	O
of	O
outward	O
-	O
facing	O
NCX_Mj	O
analogous	O
to	O
those	O
described	O
above	O
for	O
Na	O
+	O
(	O
see	O
Supplementary	O
Note	O
2	O
and	O
Supplementary	O
Fig	O
.	O
3	O
-	O
4	O
for	O
details	O
on	O
how	O
the	O
structures	O
of	O
the	O
Ca2	O
+-	O
bound	O
state	O
was	O
predicted	O
).	O
	O
The	O
calculated	O
free	O
-	O
energy	O
landscape	O
for	O
Ca2	O
+-	O
bound	O
NCX_Mj	O
confirms	O
the	O
hypothesis	O
outlined	O
above	O
(	O
1	O
×	O
Ca2	O
+,	O
Fig	O
.	O
6a	O
):	O
consistent	O
with	O
the	O
fact	O
that	O
NCX_Mj	O
transports	O
a	O
single	O
Ca2	O
+,	O
the	O
occluded	O
,	O
dehydrated	O
conformation	O
is	O
one	O
of	O
the	O
major	O
energetic	O
minima	O
,	O
but	O
clearly	O
the	O
exchanger	O
can	O
also	O
adopt	O
the	O
semi	O
-	O
open	O
and	O
open	O
states	O
that	O
would	O
be	O
required	O
for	O
Ca2	O
+	O
release	O
and	O
Na	O
+	O
entry	O
,	O
via	O
either	O
of	O
the	O
aqueous	O
access	O
channels	O
that	O
lead	O
to	O
Sext	O
and	O
SCa	O
(	O
Fig	O
.	O
6b	O
-	O
c	O
).	O
	O
By	O
contrast	O
,	O
protonation	O
of	O
Glu54	O
and	O
Glu213	O
makes	O
the	O
occluded	O
conformation	O
energetically	O
unfeasible	O
,	O
consistent	O
with	O
the	O
fact	O
that	O
NCX_Mj	O
does	O
not	O
transport	O
protons	O
;	O
in	O
this	O
H	O
+-	O
bound	O
state	O
,	O
though	O
,	O
the	O
exchanger	O
can	O
adopt	O
the	O
semi	O
-	O
open	O
conformation	O
captured	O
in	O
the	O
low	O
pH	O
,	O
apo	O
crystal	O
structure	O
(	O
2	O
×	O
H	O
+,	O
Fig	O
.	O
6a	O
-	O
c	O
).	O
	O
Taken	O
together	O
,	O
this	O
systematic	O
computational	O
analysis	O
of	O
outward	O
-	O
facing	O
NCX_Mj	O
clearly	O
demonstrates	O
that	O
the	O
alternating	O
-	O
access	O
and	O
ion	O
-	O
recognition	O
mechanisms	O
in	O
this	O
Na	O
+/	O
Ca2	O
+	O
exchanger	O
are	O
coupled	O
through	O
the	O
influence	O
that	O
the	O
bound	O
ions	O
have	O
on	O
the	O
free	O
-	O
energy	O
landscape	O
of	O
the	O
protein	O
,	O
which	O
in	O
turn	O
determines	O
whether	O
or	O
not	O
the	O
occluded	O
conformation	O
is	O
energetically	O
feasible	O
.	O
	O
The	O
alternating	O
-	O
access	O
hypothesis	O
implicitly	O
dictates	O
that	O
the	O
switch	O
between	O
outward	O
-	O
and	O
inward	O
-	O
open	O
conformations	O
of	O
a	O
given	O
secondary	O
-	O
active	O
transporter	O
must	O
not	O
occur	O
unless	O
the	O
appropriate	O
type	O
and	O
number	O
of	O
substrates	O
are	O
recognized	O
.	O
	O
Yet	O
,	O
when	O
multiple	O
species	O
are	O
to	O
be	O
co	O
-	O
translocated	O
,	O
by	O
either	O
an	O
antiporter	O
or	O
a	O
symporter	O
,	O
partial	O
occupancies	O
must	O
not	O
be	O
conducive	O
to	O
the	O
alternating	O
-	O
access	O
switch	O
.	O
	O
Here	O
,	O
we	O
have	O
provided	O
novel	O
insights	O
into	O
this	O
intriguing	O
mechanism	O
of	O
conformational	O
control	O
through	O
structural	O
studies	O
and	O
quantitative	O
molecular	O
simulations	O
of	O
a	O
Na	O
+/	O
Ca2	O
+	O
exchanger	O
.	O
	O
Specifically	O
,	O
our	O
studies	O
of	O
NCX_Mj	O
reveal	O
the	O
mechanism	O
of	O
forward	O
ion	O
exchange	O
(	O
Fig	O
.	O
7	O
).	O
	O
Here	O
,	O
we	O
demonstrate	O
that	O
conformational	O
changes	O
in	O
the	O
extracellular	O
region	O
of	O
the	O
TM2	O
-	O
TM3	O
and	O
TM7	O
-	O
TM8	O
bundle	O
precede	O
and	O
are	O
necessary	O
for	O
the	O
transition	O
,	O
and	O
are	O
associated	O
with	O
ion	O
recognition	O
and	O
/	O
or	O
release	O
.	O
	O
Interestingly	O
,	O
the	O
bending	O
of	O
TM7	O
associated	O
with	O
the	O
occlusion	O
of	O
the	O
ion	O
-	O
binding	O
sites	O
also	O
unlocks	O
its	O
interaction	O
with	O
TM6	O
,	O
and	O
thus	O
enables	O
TM6	O
and	O
TM1	O
to	O
freely	O
slide	O
to	O
the	O
inward	O
-	O
facing	O
conformation	O
.	O
	O
The	O
crystal	O
structures	O
of	O
NCX_Mj	O
reported	O
here	O
,	O
with	O
either	O
Na	O
+,	O
Ca2	O
+,	O
Sr2	O
+	O
or	O
H	O
+	O
bound	O
,	O
capture	O
the	O
exchanger	O
in	O
different	O
conformational	O
states	O
.	O
	O
These	O
states	O
can	O
only	O
represent	O
a	O
subset	O
among	O
all	O
possible	O
,	O
but	O
they	O
ought	O
to	O
reflect	O
inherent	O
preferences	O
of	O
the	O
transporter	O
,	O
modulated	O
by	O
the	O
experimental	O
conditions	O
.	O
	O
For	O
example	O
,	O
in	O
the	O
crystal	O
of	O
NCX_Mj	O
in	O
LCP	O
,	O
the	O
extracellular	O
half	O
of	O
the	O
gating	O
helices	O
(	O
TM6	O
and	O
TM1	O
)	O
form	O
a	O
lattice	O
contact	O
,	O
which	O
might	O
ultimately	O
restrict	O
the	O
degree	O
of	O
opening	O
of	O
the	O
ion	O
-	O
binding	O
sites	O
in	O
some	O
cases	O
(	O
e	O
.	O
g	O
.	O
in	O
the	O
apo	O
,	O
low	O
pH	O
structure	O
).	O
	O
Nonetheless	O
,	O
the	O
calculated	O
free	O
-	O
energy	O
landscapes	O
,	O
derived	O
without	O
knowledge	O
of	O
the	O
experimental	O
data	O
,	O
reassuringly	O
confirm	O
that	O
the	O
crystallized	O
structures	O
correspond	O
to	O
mechanistically	O
relevant	O
,	O
interconverting	O
states	O
.	O
	O
The	O
simulations	O
also	O
demonstrate	O
how	O
this	O
landscape	O
is	O
drastically	O
re	O
-	O
shaped	O
upon	O
each	O
ion	O
-	O
binding	O
event	O
.	O
	O
We	O
posit	O
that	O
a	O
similar	O
principle	O
might	O
govern	O
the	O
alternating	O
-	O
access	O
mechanism	O
in	O
other	O
transporters	O
;	O
that	O
is	O
,	O
we	O
anticipate	O
that	O
for	O
both	O
symporters	O
and	O
antiporters	O
,	O
it	O
is	O
the	O
feasibility	O
of	O
the	O
occluded	O
state	O
,	O
encoded	O
in	O
the	O
protein	O
conformational	O
free	O
-	O
energy	O
landscape	O
and	O
its	O
dependence	O
on	O
substrate	O
binding	O
,	O
that	O
ultimately	O
explains	O
their	O
specific	O
coupling	O
mechanisms	O
.	O
	O
In	O
multiple	O
ways	O
,	O
our	O
findings	O
provide	O
an	O
explanation	O
for	O
,	O
existing	O
functional	O
,	O
biochemical	O
and	O
biophysical	O
data	O
for	O
both	O
NCX_Mj	O
and	O
its	O
eukaryotic	O
homologues	O
.	O
	O
The	O
crystallographic	O
data	O
also	O
provides	O
the	O
long	O
-	O
sought	O
structural	O
basis	O
for	O
the	O
‘	O
two	O
-	O
site	O
’	O
model	O
proposed	O
to	O
describe	O
competitive	O
cation	O
binding	O
in	O
eukaryotic	O
NCX	O
,	O
underscoring	O
the	O
relevance	O
of	O
these	O
studies	O
of	O
NCX_Mj	O
as	O
a	O
prototypical	O
Na	O
+/	O
Ca2	O
+	O
exchanger	O
.	O
	O
Interestingly	O
,	O
binding	O
of	O
Ca2	O
+	O
to	O
Smid	O
appears	O
to	O
be	O
also	O
possible	O
,	O
but	O
available	O
evidence	O
indicates	O
that	O
this	O
event	O
transiently	O
blocks	O
the	O
exchange	O
cycle	O
.	O
	O
Indeed	O
,	O
structures	O
of	O
NCX_Mj	O
bound	O
to	O
Cd2	O
+	O
or	O
Mn2	O
+,	O
both	O
of	O
which	O
inhibit	O
transport	O
,	O
show	O
these	O
ions	O
at	O
Smid	O
;	O
by	O
contrast	O
,	O
Sr2	O
+	O
binds	O
only	O
to	O
SCa	O
,	O
and	O
accordingly	O
,	O
is	O
transported	O
by	O
NCX	O
similarly	O
to	O
calcium	O
.	O
	O
In	O
addition	O
,	O
the	O
increased	O
compactness	O
of	O
the	O
protein	O
tertiary	O
structure	O
in	O
the	O
occluded	O
state	O
would	O
also	O
slow	O
down	O
the	O
dynamics	O
of	O
the	O
secondary	O
-	O
structure	O
elements	O
,	O
and	O
thus	O
further	O
reduce	O
the	O
HDX	O
rate	O
.	O
	O
As	O
the	O
calculated	O
free	O
-	O
energy	O
landscapes	O
show	O
,	O
Na	O
+	O
and	O
Ca2	O
+	O
induce	O
the	O
occlusion	O
of	O
the	O
transporter	O
in	O
a	O
comparable	O
manner	O
,	O
and	O
yet	O
the	O
ion	O
-	O
bound	O
states	O
retain	O
the	O
ability	O
to	O
explore	O
conformations	O
that	O
are	O
partially	O
or	O
fully	O
open	O
to	O
the	O
extracellular	O
solution	O
,	O
precisely	O
so	O
as	O
to	O
be	O
able	O
to	O
unload	O
and	O
re	O
-	O
load	O
the	O
substrates	O
.	O
	O
Na	O
+	O
binding	O
to	O
outward	O
-	O
facing	O
NCX_Mj	O
.	O
	O
(	O
a	O
)	O
Overall	O
structure	O
of	O
native	O
outward	O
-	O
facing	O
NCX_Mj	O
from	O
crystals	O
grown	O
in	O
150	O
mM	O
Na	O
+.	O
	O
Colored	O
spheres	O
represent	O
the	O
bound	O
Na	O
+	O
(	O
green	O
)	O
and	O
water	O
(	O
red	O
).	O
	O
(	O
b	O
)	O
Structural	O
details	O
and	O
definition	O
of	O
the	O
four	O
central	O
binding	O
sites	O
.	O
	O
The	O
electron	O
density	O
(	O
grey	O
mesh	O
,	O
1	O
.	O
9	O
Å	O
Fo	O
-	O
Fc	O
ion	O
omit	O
map	O
contoured	O
at	O
4σ	O
)	O
at	O
Smid	O
was	O
modeled	O
as	O
water	O
(	O
red	O
sphere	O
)	O
and	O
those	O
at	O
Sext	O
,	O
SCa	O
and	O
Sint	O
as	O
Na	O
+	O
ions	O
(	O
green	O
spheres	O
).	O
	O
Further	O
details	O
are	O
shown	O
in	O
Supplementary	O
Fig	O
.	O
1	O
.	O
(	O
c	O
)	O
Concentration	O
-	O
dependent	O
change	O
in	O
Na	O
+	O
occupancy	O
(	O
see	O
also	O
Table	O
1	O
).	O
	O
All	O
Fo	O
–	O
Fc	O
ion	O
-	O
omit	O
maps	O
are	O
calculated	O
to	O
2	O
.	O
4	O
Å	O
and	O
contoured	O
at	O
3σ	O
for	O
comparison	O
.	O
	O
The	O
displacement	O
of	O
A206	O
reflects	O
the	O
[	O
Na	O
+]-	O
dependent	O
conformational	O
change	O
from	O
the	O
partially	O
open	O
to	O
the	O
occluded	O
state	O
(	O
observed	O
at	O
low	O
and	O
high	O
Na	O
+	O
concentrations	O
,	O
respectively	O
).	O
	O
At	O
20	O
mM	O
Na	O
+,	O
both	O
conformations	O
co	O
-	O
exist	O
.	O
	O
Na	O
+-	O
occupancy	O
dependent	O
conformational	O
change	O
in	O
NCX_Mj	O
.	O
	O
(	O
a	O
)	O
Superimposition	O
of	O
the	O
NCX_Mj	O
crystal	O
structures	O
obtained	O
in	O
high	O
(	O
100	O
mM	O
,	O
cyan	O
cylinders	O
)	O
and	O
low	O
(	O
10	O
mM	O
,	O
brown	O
cylinders	O
)	O
Na	O
+	O
concentrations	O
.	O
	O
(	O
b	O
)	O
Close	O
-	O
up	O
view	O
of	O
the	O
interface	O
between	O
TM6	O
and	O
TM7ab	O
in	O
the	O
NCX_Mj	O
structures	O
obtained	O
at	O
high	O
and	O
low	O
Na	O
+	O
concentrations	O
(	O
top	O
and	O
lower	O
panels	O
,	O
respectively	O
).	O
	O
Residues	O
forming	O
van	O
-	O
der	O
-	O
Waals	O
contacts	O
in	O
the	O
structure	O
at	O
low	O
Na	O
+	O
concentration	O
are	O
shown	O
in	O
detail	O
.	O
	O
T50	O
and	O
T209	O
(	O
labeled	O
in	O
red	O
)	O
coordinate	O
Sr2	O
+	O
through	O
their	O
backbone	O
carbonyl	O
-	O
oxygen	O
atoms	O
.	O
	O
High	O
Na	O
+	O
concentration	O
(	O
100	O
mM	O
)	O
completely	O
displaces	O
Sr2	O
+	O
and	O
reverts	O
NCX_Mj	O
to	O
the	O
occluded	O
state	O
(	O
right	O
panel	O
).	O
	O
The	O
contour	O
level	O
of	O
the	O
Fo	O
–	O
Fc	O
omit	O
map	O
of	O
the	O
structure	O
at	O
high	O
Na	O
+	O
concentration	O
was	O
lowered	O
(	O
to	O
4σ	O
)	O
so	O
as	O
to	O
visualize	O
the	O
density	O
from	O
Na	O
+	O
ions	O
and	O
H2O	O
.	O
	O
The	O
side	O
chains	O
of	O
E54	O
and	O
E213	O
from	O
the	O
low	O
Na	O
+	O
structure	O
are	O
also	O
shown	O
(	O
light	O
brown	O
)	O
for	O
comparison	O
.	O
	O
White	O
spheres	O
indicate	O
the	O
location	O
Sint	O
,	O
Smid	O
SCa	O
.	O
(	O
e	O
)	O
Extracellular	O
solvent	O
accessibility	O
of	O
the	O
ion	O
-	O
binding	O
sites	O
in	O
apo	O
NCX_Mj	O
.	O
	O
Spontaneous	O
changes	O
in	O
the	O
structure	O
of	O
outward	O
-	O
occluded	O
,	O
fully	O
Na	O
+-	O
occupied	O
NCX_Mj	O
(	O
PDB	O
code	O
3V5U	O
)	O
upon	O
sequential	O
displacement	O
of	O
Na	O
+.	O
	O
(	O
c	O
)	O
Representative	O
simulation	O
snapshots	O
(	O
same	O
as	O
above	O
)	O
with	O
Na	O
+	O
bound	O
at	O
SCa	O
and	O
Sint	O
(	O
marine	O
cartoons	O
,	O
yellow	O
spheres	O
)	O
and	O
without	O
any	O
Na	O
+	O
bound	O
(	O
grey	O
cartoons	O
).	O
	O
(	O
d	O
)	O
Close	O
-	O
up	O
of	O
the	O
ion	O
-	O
binding	O
region	O
in	O
the	O
fully	O
Na	O
+-	O
occupied	O
state	O
.	O
	O
(	O
f	O
)	O
Close	O
-	O
up	O
of	O
the	O
ion	O
-	O
binding	O
region	O
in	O
the	O
Na	O
+-	O
free	O
state	O
.	O
(	O
g	O
-	O
i	O
)	O
Analytical	O
descriptors	O
of	O
the	O
changes	O
just	O
described	O
,	O
calculated	O
from	O
the	O
simulations	O
of	O
each	O
Na	O
+-	O
occupancy	O
state	O
depicted	O
in	O
panels	O
(	O
a	O
-	O
f	O
).	O
	O
These	O
descriptors	O
were	O
employed	O
as	O
collective	O
variables	O
in	O
the	O
Bias	O
-	O
Exchange	O
Metadynamics	O
simulations	O
(	O
Methods	O
).	O
	O
Thermodynamic	O
basis	O
for	O
the	O
proposed	O
mechanism	O
of	O
substrate	O
control	O
of	O
the	O
alternating	O
-	O
access	O
transition	O
of	O
NCX	O
.	O
(	O
a	O
)	O
Calculated	O
conformational	O
free	O
-	O
energy	O
landscapes	O
for	O
outward	O
-	O
facing	O
NCX_Mj	O
,	O
for	O
two	O
different	O
Na	O
+-	O
occupancy	O
states	O
,	O
and	O
for	O
a	O
state	O
with	O
no	O
ions	O
bound	O
.	O
	O
Black	O
circles	O
map	O
the	O
X	O
-	O
ray	O
structures	O
of	O
NCX_Mj	O
obtained	O
at	O
high	O
and	O
low	O
Na	O
+	O
concentration	O
,	O
as	O
well	O
as	O
that	O
at	O
low	O
pH	O
,	O
reported	O
in	O
this	O
study	O
.	O
	O
The	O
two	O
inverted	O
-	O
topology	O
repeats	O
in	O
the	O
transporter	O
structure	O
(	O
transparent	O
cartoons	O
)	O
are	O
colored	O
differently	O
(	O
TM1	O
-	O
5	O
,	O
orange	O
;	O
TM6	O
-	O
10	O
,	O
marine	O
).	O
	O
(	O
c	O
)	O
Close	O
-	O
up	O
views	O
of	O
the	O
ion	O
-	O
binding	O
region	O
in	O
the	O
same	O
conformational	O
free	O
-	O
energy	O
minima	O
.	O
	O
Key	O
residues	O
involved	O
in	O
Na	O
+	O
and	O
water	O
coordination	O
(	O
W	O
)	O
are	O
highlighted	O
(	O
sticks	O
,	O
black	O
lines	O
).	O
	O
The	O
water	O
-	O
density	O
maps	O
in	O
(	O
b	O
)	O
is	O
shown	O
here	O
as	O
a	O
grey	O
mesh	O
.	O
	O
Note	O
D240	O
is	O
protonated	O
,	O
while	O
E54	O
and	O
E213	O
are	O
ionized	O
.	O
	O
Thermodynamic	O
basis	O
for	O
the	O
proposed	O
mechanism	O
of	O
substrate	O
control	O
of	O
the	O
alternating	O
-	O
access	O
transition	O
of	O
NCX	O
.	O
(	O
a	O
)	O
Calculated	O
free	O
-	O
energy	O
landscapes	O
for	O
outward	O
-	O
facing	O
NCX_Mj	O
,	O
for	O
the	O
Ca2	O
+	O
and	O
the	O
fully	O
protonated	O
state	O
.	O
	O
The	O
free	O
energy	O
is	O
plotted	O
as	O
in	O
Fig	O
.	O
5	O
.	O
	O
For	O
Ca2	O
+,	O
a	O
map	O
is	O
shown	O
in	O
which	O
a	O
correction	O
for	O
the	O
charge	O
-	O
transfer	O
between	O
the	O
ion	O
and	O
the	O
protein	O
is	O
introduced	O
,	O
alongside	O
the	O
uncorrected	O
map	O
(	O
see	O
Supplementary	O
Notes	O
3	O
-	O
4	O
and	O
Supplementary	O
Fig	O
.	O
5	O
-	O
6	O
).	O
	O
The	O
uncorrected	O
map	O
overstabilizes	O
the	O
open	O
state	O
relative	O
to	O
the	O
semi	O
-	O
open	O
and	O
occluded	O
because	O
it	O
also	O
overestimates	O
the	O
cost	O
of	O
dehydration	O
of	O
the	O
ion	O
,	O
once	O
it	O
is	O
bound	O
to	O
the	O
protein	O
(	O
this	O
effect	O
is	O
negligible	O
for	O
Na	O
+).	O
	O
The	O
Ca2	O
+	O
ion	O
is	O
shown	O
as	O
a	O
red	O
sphere	O
;	O
the	O
protein	O
is	O
shown	O
as	O
in	O
Fig	O
.	O
5	O
.	O
(	O
c	O
)	O
Close	O
-	O
up	O
views	O
of	O
the	O
ion	O
-	O
binding	O
region	O
in	O
the	O
same	O
conformational	O
free	O
-	O
energy	O
minima	O
.	O
	O
Key	O
residues	O
involved	O
in	O
Ca2	O
+	O
and	O
water	O
coordination	O
(	O
W	O
)	O
are	O
highlighted	O
(	O
sticks	O
,	O
black	O
lines	O
).	O
	O
In	O
the	O
occluded	O
state	O
with	O
Ca2	O
+	O
bound	O
,	O
helix	O
TM7ab	O
bends	O
in	O
the	O
same	O
way	O
as	O
in	O
the	O
fully	O
occupied	O
Na	O
+	O
state	O
,	O
as	O
the	O
carbonyl	O
of	O
Ala206	O
forms	O
a	O
hydrogen	O
-	O
bonding	O
interaction	O
with	O
Ser210	O
.	O
	O
Structural	O
mechanism	O
of	O
extracellular	O
forward	O
ion	O
exchange	O
in	O
NCX	O
.	O
	O
The	O
carbonyl	O
groups	O
of	O
Ala47	O
(	O
on	O
TM2b	O
)	O
and	O
Ala206	O
(	O
on	O
TM7b	O
),	O
and	O
the	O
side	O
chains	O
of	O
Glu54	O
(	O
on	O
TM2c	O
)	O
and	O
Glu213	O
(	O
on	O
TM7c	O
)	O
are	O
highlighted	O
;	O
these	O
are	O
four	O
of	O
the	O
key	O
residues	O
for	O
ion	O
chelation	O
and	O
conformational	O
changes	O
.	O
	O
The	O
green	O
open	O
cylinders	O
represent	O
the	O
gating	O
helices	O
TM1	O
and	O
TM6	O
.	O
	O
Asterisks	O
mark	O
the	O
states	O
whose	O
crystal	O
structures	O
have	O
been	O
determined	O
in	O
this	O
study	O
.	O
	O
These	O
states	O
and	O
their	O
connectivity	O
can	O
also	O
be	O
deduced	O
from	O
the	O
calculated	O
free	O
-	O
energy	O
landscapes	O
,	O
which	O
also	O
reveal	O
a	O
Ca2	O
+-	O
loaded	O
outward	O
-	O
facing	O
occluded	O
state	O
,	O
and	O
an	O
unloaded	O
,	O
fully	O
open	O
state	O
.	O
	O
How	O
the	O
essential	O
pre	O
-	O
mRNA	O
splicing	O
factor	O
U2AF65	O
recognizes	O
the	O
polypyrimidine	O
(	O
Py	O
)	O
signals	O
of	O
the	O
major	O
class	O
of	O
3	O
′	O
splice	O
sites	O
in	O
human	O
gene	O
transcripts	O
remains	O
incompletely	O
understood	O
.	O
	O
We	O
determined	O
four	O
structures	O
of	O
an	O
extended	O
U2AF65	O
–	O
RNA	O
-	O
binding	O
domain	O
bound	O
to	O
Py	O
-	O
tract	O
oligonucleotides	O
at	O
resolutions	O
between	O
2	O
.	O
0	O
and	O
1	O
.	O
5	O
Å	O
.	O
These	O
structures	O
together	O
with	O
RNA	O
binding	O
and	O
splicing	O
assays	O
reveal	O
unforeseen	O
roles	O
for	O
U2AF65	O
inter	O
-	O
domain	O
residues	O
in	O
recognizing	O
a	O
contiguous	O
,	O
nine	O
-	O
nucleotide	O
Py	O
tract	O
.	O
	O
The	O
U2AF65	O
linker	O
residues	O
between	O
the	O
dual	O
RNA	O
recognition	O
motifs	O
(	O
RRMs	O
)	O
recognize	O
the	O
central	O
nucleotide	O
,	O
whereas	O
the	O
N	O
-	O
and	O
C	O
-	O
terminal	O
RRM	O
extensions	O
recognize	O
the	O
3	O
′	O
terminus	O
and	O
third	O
nucleotide	O
.	O
	O
Single	O
-	O
molecule	O
FRET	O
experiments	O
suggest	O
that	O
conformational	O
selection	O
and	O
induced	O
fit	O
of	O
the	O
U2AF65	O
RRMs	O
are	O
complementary	O
mechanisms	O
for	O
Py	O
-	O
tract	O
association	O
.	O
	O
Altogether	O
,	O
these	O
results	O
advance	O
the	O
mechanistic	O
understanding	O
of	O
molecular	O
recognition	O
for	O
a	O
major	O
class	O
of	O
splice	O
site	O
signals	O
.	O
	O
The	O
pre	O
-	O
mRNA	O
splicing	O
factor	O
U2AF65	O
recognizes	O
3	O
′	O
splice	O
sites	O
in	O
human	O
gene	O
transcripts	O
,	O
but	O
the	O
details	O
are	O
not	O
fully	O
understood	O
.	O
	O
The	O
differential	O
skipping	O
or	O
inclusion	O
of	O
alternatively	O
spliced	O
pre	O
-	O
mRNA	O
regions	O
is	O
a	O
major	O
source	O
of	O
diversity	O
for	O
nearly	O
all	O
human	O
gene	O
transcripts	O
.	O
	O
At	O
the	O
3	O
′	O
splice	O
site	O
of	O
the	O
major	O
intron	O
class	O
,	O
these	O
include	O
a	O
polypyrimidine	O
(	O
Py	O
)	O
tract	O
comprising	O
primarily	O
Us	O
or	O
Cs	O
,	O
which	O
is	O
preceded	O
by	O
a	O
branch	O
point	O
sequence	O
(	O
BPS	O
)	O
that	O
ultimately	O
serves	O
as	O
the	O
nucleophile	O
in	O
the	O
splicing	O
reaction	O
and	O
an	O
AG	O
-	O
dinucleotide	O
at	O
the	O
3	O
′	O
splice	O
site	O
junction	O
.	O
	O
Disease	O
-	O
causing	O
mutations	O
often	O
compromise	O
pre	O
-	O
mRNA	O
splicing	O
(	O
reviewed	O
in	O
refs	O
),	O
yet	O
a	O
priori	O
predictions	O
of	O
splice	O
sites	O
and	O
the	O
consequences	O
of	O
their	O
mutations	O
are	O
challenged	O
by	O
the	O
brevity	O
and	O
degeneracy	O
of	O
known	O
splice	O
site	O
sequences	O
.	O
	O
High	O
-	O
resolution	O
structures	O
of	O
intact	O
splicing	O
factor	O
–	O
RNA	O
complexes	O
would	O
offer	O
key	O
insights	O
regarding	O
the	O
juxtaposition	O
of	O
the	O
distinct	O
splice	O
site	O
consensus	O
sequences	O
and	O
their	O
relationship	O
to	O
disease	O
-	O
causing	O
point	O
mutations	O
.	O
	O
The	O
early	O
-	O
stage	O
pre	O
-	O
mRNA	O
splicing	O
factor	O
U2AF65	O
is	O
essential	O
for	O
viability	O
in	O
vertebrates	O
and	O
other	O
model	O
organisms	O
(	O
for	O
example	O
,	O
ref	O
.).	O
	O
A	O
tightly	O
controlled	O
assembly	O
among	O
U2AF65	O
,	O
the	O
pre	O
-	O
mRNA	O
,	O
and	O
partner	O
proteins	O
sequentially	O
identifies	O
the	O
3	O
′	O
splice	O
site	O
and	O
promotes	O
association	O
of	O
the	O
spliceosome	O
,	O
which	O
ultimately	O
accomplishes	O
the	O
task	O
of	O
splicing	O
.	O
	O
Subsequently	O
U2AF65	O
recruits	O
the	O
U2	O
small	O
nuclear	O
ribonucleoprotein	O
particle	O
(	O
snRNP	O
)	O
and	O
ultimately	O
dissociates	O
from	O
the	O
active	O
spliceosome	O
.	O
	O
Biochemical	O
characterizations	O
of	O
U2AF65	O
demonstrated	O
that	O
tandem	O
RNA	O
recognition	O
motifs	O
(	O
RRM1	O
and	O
RRM2	O
)	O
recognize	O
the	O
Py	O
tract	O
(	O
Fig	O
.	O
1a	O
).	O
	O
Milestone	O
crystal	O
structures	O
of	O
the	O
core	O
U2AF65	O
RRM1	O
and	O
RRM2	O
connected	O
by	O
a	O
shortened	O
inter	O
-	O
RRM	O
linker	O
(	O
dU2AF651	B-mutant
,	I-mutant
2	I-mutant
)	O
detailed	O
a	O
subset	O
of	O
nucleotide	O
interactions	O
with	O
the	O
individual	O
U2AF65	O
RRMs	O
.	O
	O
A	O
subsequent	O
NMR	O
structure	O
characterized	O
the	O
side	O
-	O
by	O
-	O
side	O
arrangement	O
of	O
the	O
minimal	O
U2AF65	O
RRM1	O
and	O
RRM2	O
connected	O
by	O
a	O
linker	O
of	O
natural	O
length	O
(	O
U2AF651	B-mutant
,	I-mutant
2	I-mutant
),	O
yet	O
depended	O
on	O
the	O
dU2AF651	B-mutant
,	I-mutant
2	I-mutant
crystal	O
structures	O
for	O
RNA	O
interactions	O
and	O
an	O
ab	O
initio	O
model	O
for	O
the	O
inter	O
-	O
RRM	O
linker	O
conformation	O
.	O
	O
Here	O
,	O
we	O
use	O
X	O
-	O
ray	O
crystallography	O
and	O
biochemical	O
studies	O
to	O
reveal	O
new	O
roles	O
in	O
Py	O
-	O
tract	O
recognition	O
for	O
the	O
inter	O
-	O
RRM	O
linker	O
and	O
key	O
residues	O
surrounding	O
the	O
core	O
U2AF65	O
RRMs	O
.	O
	O
Cognate	O
U2AF65	O
–	O
Py	O
-	O
tract	O
recognition	O
requires	O
RRM	O
extensions	O
	O
The	O
RNA	O
affinity	O
of	O
the	O
minimal	O
U2AF651	B-mutant
,	I-mutant
2	I-mutant
domain	O
comprising	O
the	O
core	O
RRM1	O
–	O
RRM2	O
folds	O
(	O
U2AF651	B-mutant
,	I-mutant
2	I-mutant
,	O
residues	O
148	O
–	O
336	O
)	O
is	O
relatively	O
weak	O
compared	O
with	O
full	O
-	O
length	O
U2AF65	O
(	O
Fig	O
.	O
1a	O
,	O
b	O
;	O
Supplementary	O
Fig	O
.	O
1	O
).	O
	O
Historically	O
,	O
this	O
difference	O
was	O
attributed	O
to	O
the	O
U2AF65	O
arginine	O
–	O
serine	O
rich	O
domain	O
,	O
which	O
contacts	O
pre	O
-	O
mRNA	O
–	O
U2	O
snRNA	O
duplexes	O
outside	O
of	O
the	O
Py	O
tract	O
.	O
	O
We	O
noticed	O
that	O
the	O
RNA	O
-	O
binding	O
affinity	O
of	O
the	O
U2AF651	B-mutant
,	I-mutant
2	I-mutant
domain	O
was	O
greatly	O
enhanced	O
by	O
the	O
addition	O
of	O
seven	O
and	O
six	O
residues	O
at	O
the	O
respective	O
N	O
and	O
C	O
termini	O
of	O
the	O
minimal	O
RRM1	O
and	O
RRM2	O
(	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
,	O
residues	O
141	O
–	O
342	O
;	O
Fig	O
.	O
1a	O
).	O
	O
In	O
a	O
fluorescence	O
anisotropy	O
assay	O
for	O
binding	O
a	O
representative	O
Py	O
tract	O
derived	O
from	O
the	O
well	O
-	O
characterized	O
splice	O
site	O
of	O
the	O
adenovirus	O
major	O
late	O
promoter	O
(	O
AdML	O
),	O
the	O
RNA	O
affinity	O
of	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
increased	O
by	O
100	O
-	O
fold	O
relative	O
to	O
U2AF651	B-mutant
,	I-mutant
2	I-mutant
to	O
comparable	O
levels	O
as	O
full	O
-	O
length	O
U2AF65	O
(	O
Fig	O
.	O
1b	O
;	O
Supplementary	O
Fig	O
.	O
1a	O
–	O
d	O
).	O
	O
U2AF65	O
-	O
bound	O
Py	O
tract	O
comprises	O
nine	O
contiguous	O
nucleotides	O
	O
To	O
investigate	O
the	O
structural	O
basis	O
for	O
cognate	O
U2AF65	O
recognition	O
of	O
a	O
contiguous	O
Py	O
tract	O
,	O
we	O
determined	O
four	O
crystal	O
structures	O
of	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
bound	O
to	O
Py	O
-	O
tract	O
oligonucleotides	O
(	O
Fig	O
.	O
2a	O
;	O
Table	O
1	O
).	O
	O
The	O
protein	O
and	O
oligonucleotide	O
conformations	O
are	O
nearly	O
identical	O
among	O
the	O
four	O
new	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
structures	O
(	O
Supplementary	O
Fig	O
.	O
2a	O
).	O
	O
The	O
extensions	O
at	O
the	O
N	O
terminus	O
of	O
RRM1	O
and	O
C	O
terminus	O
of	O
RRM2	O
adopt	O
well	O
-	O
ordered	O
α	O
-	O
helices	O
.	O
	O
The	O
discovery	O
of	O
nine	O
U2AF65	O
-	O
binding	O
sites	O
for	O
contiguous	O
Py	O
-	O
tract	O
nucleotides	O
was	O
unexpected	O
.	O
	O
The	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
structures	O
characterize	O
ribose	O
(	O
r	O
)	O
nucleotides	O
at	O
all	O
of	O
the	O
binding	O
sites	O
except	O
the	O
seventh	O
and	O
eighth	O
deoxy	O
-(	O
d	O
)	O
U	O
,	O
which	O
are	O
likely	O
to	O
lack	O
2	O
′-	O
hydroxyl	O
contacts	O
based	O
on	O
the	O
RNA	O
-	O
bound	O
dU2AF651	B-mutant
,	I-mutant
2	I-mutant
structure	O
.	O
	O
In	O
striking	O
departures	O
from	O
prior	O
partial	O
views	O
,	O
the	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
structures	O
reveal	O
three	O
unanticipated	O
nucleotide	O
-	O
binding	O
sites	O
at	O
the	O
centre	O
of	O
the	O
Py	O
tract	O
,	O
as	O
well	O
as	O
numerous	O
new	O
interactions	O
that	O
underlie	O
cognate	O
recognition	O
of	O
the	O
Py	O
tract	O
(	O
Fig	O
.	O
3a	O
–	O
h	O
).	O
	O
U2AF65	O
inter	O
-	O
RRM	O
linker	O
interacts	O
with	O
the	O
Py	O
tract	O
	O
The	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
RRM2	O
,	O
the	O
inter	O
-	O
RRM	O
linker	O
and	O
RRM1	O
concomitantly	O
recognize	O
the	O
three	O
central	O
nucleotides	O
of	O
the	O
Py	O
tract	O
,	O
which	O
are	O
likely	O
to	O
coordinate	O
the	O
conformational	O
arrangement	O
of	O
these	O
disparate	O
portions	O
of	O
the	O
protein	O
.	O
	O
Residues	O
in	O
the	O
C	O
-	O
terminal	O
region	O
of	O
the	O
U2AF65	O
inter	O
-	O
RRM	O
linker	O
comprise	O
a	O
centrally	O
located	O
binding	O
site	O
for	O
the	O
fifth	O
nucleotide	O
on	O
the	O
RRM2	O
surface	O
and	O
abutting	O
the	O
RRM1	O
/	O
RRM2	O
interface	O
(	O
Fig	O
.	O
3d	O
).	O
	O
The	O
backbone	O
amide	O
of	O
the	O
linker	O
V254	O
and	O
the	O
carbonyl	O
of	O
T252	O
engage	O
in	O
hydrogen	O
bonds	O
with	O
the	O
rU5	O
-	O
O4	O
and	O
-	O
N3H	O
atoms	O
.	O
	O
The	O
C	O
-	O
terminal	O
region	O
of	O
the	O
inter	O
-	O
RRM	O
linker	O
also	O
participates	O
in	O
the	O
preceding	O
rU4	O
-	O
binding	O
site	O
,	O
where	O
the	O
V254	O
backbone	O
carbonyl	O
and	O
D256	O
carboxylate	O
position	O
the	O
K260	O
side	O
chain	O
to	O
hydrogen	O
bond	O
with	O
the	O
rU4	O
-	O
O4	O
(	O
Fig	O
.	O
3c	O
).	O
	O
At	O
the	O
opposite	O
side	O
of	O
the	O
central	O
fifth	O
nucleotide	O
,	O
the	O
sixth	O
rU6	O
nucleotide	O
is	O
located	O
at	O
the	O
inter	O
-	O
RRM1	O
/	O
RRM2	O
interface	O
(	O
Fig	O
.	O
3e	O
;	O
Supplementary	O
Movie	O
1	O
).	O
	O
The	O
rU6	O
base	O
edge	O
is	O
relatively	O
solvent	O
exposed	O
;	O
accordingly	O
,	O
the	O
rU6	O
hydrogen	O
bonds	O
with	O
U2AF65	O
are	O
water	O
mediated	O
apart	O
from	O
a	O
single	O
direct	O
interaction	O
by	O
the	O
RRM1	O
-	O
N196	O
side	O
chain	O
.	O
	O
Mutagenesis	O
of	O
either	O
V254	O
in	O
the	O
U2AF65	O
inter	O
-	O
RRM	O
linker	O
to	O
proline	O
or	O
RRM1	O
–	O
R227	O
to	O
alanine	O
,	O
which	O
remove	O
the	O
hydrogen	O
bond	O
with	O
the	O
fifth	O
uracil	O
-	O
O4	O
or	O
-	O
O2	O
,	O
reduced	O
the	O
affinities	O
of	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
for	O
the	O
representative	O
AdML	O
Py	O
tract	O
by	O
four	O
-	O
or	O
five	O
-	O
fold	O
,	O
respectively	O
.	O
	O
The	O
energetic	O
penalties	O
due	O
to	O
these	O
mutations	O
(	O
ΔΔG	O
0	O
.	O
8	O
–	O
0	O
.	O
9	O
kcal	O
mol	O
−	O
1	O
)	O
are	O
consistent	O
with	O
the	O
loss	O
of	O
each	O
hydrogen	O
bond	O
with	O
the	O
rU5	O
base	O
and	O
support	O
the	O
relevance	O
of	O
the	O
central	O
nucleotide	O
interactions	O
observed	O
in	O
the	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
structures	O
.	O
	O
Rather	O
than	O
interacting	O
with	O
a	O
new	O
5	O
′-	O
terminal	O
nucleotide	O
as	O
we	O
had	O
hypothesized	O
,	O
the	O
C	O
-	O
terminal	O
α	O
-	O
helix	O
of	O
RRM2	O
instead	O
folds	O
across	O
one	O
surface	O
of	O
rU3	O
in	O
the	O
third	O
binding	O
site	O
(	O
Fig	O
.	O
3b	O
).	O
	O
There	O
,	O
a	O
salt	O
bridge	O
between	O
the	O
K340	O
side	O
chain	O
and	O
nucleotide	O
phosphate	O
,	O
as	O
well	O
as	O
G338	O
-	O
base	O
stacking	O
and	O
a	O
hydrogen	O
bond	O
between	O
the	O
backbone	O
amide	O
of	O
G338	O
and	O
the	O
rU3	O
-	O
O4	O
,	O
secure	O
the	O
RRM2	O
extension	O
.	O
	O
At	O
the	O
N	O
terminus	O
,	O
the	O
α	O
-	O
helical	O
extension	O
of	O
U2AF65	O
RRM1	O
positions	O
the	O
Q147	O
side	O
chain	O
to	O
bridge	O
the	O
eighth	O
and	O
ninth	O
nucleotides	O
at	O
the	O
3	O
′	O
terminus	O
of	O
the	O
Py	O
tract	O
(	O
Fig	O
.	O
3f	O
–	O
h	O
).	O
	O
The	O
Q147	O
residue	O
participates	O
in	O
hydrogen	O
bonds	O
with	O
the	O
-	O
N3H	O
of	O
the	O
eighth	O
uracil	O
and	O
-	O
O2	O
of	O
the	O
ninth	O
pyrimidine	O
.	O
	O
Versatile	O
primary	O
sequence	O
of	O
the	O
U2AF65	O
inter	O
-	O
RRM	O
linker	O
	O
The	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
structures	O
reveal	O
that	O
the	O
inter	O
-	O
RRM	O
linker	O
mediates	O
an	O
extensive	O
interface	O
with	O
the	O
second	O
α	O
-	O
helix	O
of	O
RRM1	O
,	O
the	O
β2	O
/	O
β3	O
strands	O
of	O
RRM2	O
and	O
the	O
N	O
-	O
terminal	O
α	O
-	O
helical	O
extension	O
of	O
RRM1	O
.	O
	O
Altogether	O
,	O
the	O
U2AF65	O
inter	O
-	O
RRM	O
linker	O
residues	O
(	O
R228	O
–	O
K260	O
)	O
bury	O
2	O
,	O
800	O
Å2	O
of	O
surface	O
area	O
in	O
the	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
holo	O
-	O
protein	O
,	O
suggestive	O
of	O
a	O
cognate	O
interface	O
compared	O
with	O
1	O
,	O
900	O
Å2	O
for	O
a	O
typical	O
protein	O
–	O
protein	O
complex	O
.	O
	O
The	O
path	O
of	O
the	O
linker	O
initiates	O
at	O
P229	O
following	O
the	O
core	O
RRM1	O
β	O
-	O
strand	O
,	O
in	O
a	O
kink	O
that	O
is	O
positioned	O
by	O
intra	O
-	O
molecular	O
stacking	O
among	O
the	O
consecutive	O
R228	O
,	O
Y232	O
and	O
P234	O
side	O
chains	O
(	O
Fig	O
.	O
4a	O
,	O
lower	O
right	O
).	O
	O
A	O
second	O
kink	O
at	O
P236	O
,	O
coupled	O
with	O
respective	O
packing	O
of	O
the	O
L235	O
and	O
M238	O
side	O
chains	O
on	O
the	O
N	O
-	O
terminal	O
α	O
-	O
helical	O
RRM1	O
extension	O
and	O
the	O
core	O
RRM1	O
α2	O
-	O
helix	O
,	O
reverses	O
the	O
direction	O
of	O
the	O
inter	O
-	O
RRM	O
linker	O
towards	O
the	O
RRM1	O
/	O
RRM2	O
interface	O
and	O
away	O
from	O
the	O
RNA	O
-	O
binding	O
site	O
.	O
	O
In	O
the	O
neighbouring	O
apical	O
region	O
of	O
the	O
linker	O
,	O
the	O
V244	O
and	O
V246	O
side	O
chains	O
pack	O
in	O
a	O
hydrophobic	O
pocket	O
between	O
two	O
α	O
-	O
helices	O
of	O
the	O
core	O
RRM1	O
.	O
	O
The	O
adjacent	O
V249	O
and	O
V250	O
are	O
notable	O
for	O
their	O
respective	O
interactions	O
that	O
connect	O
RRM1	O
and	O
RRM2	O
at	O
this	O
distal	O
interface	O
from	O
the	O
RNA	O
-	O
binding	O
site	O
(	O
Fig	O
.	O
4a	O
,	O
top	O
).	O
	O
A	O
third	O
kink	O
stacks	O
P247	O
and	O
G248	O
with	O
Y245	O
and	O
re	O
-	O
orients	O
the	O
C	O
-	O
terminal	O
region	O
of	O
the	O
linker	O
towards	O
the	O
RRM2	O
and	O
bound	O
RNA	O
.	O
	O
Few	O
direct	O
contacts	O
are	O
made	O
between	O
the	O
remaining	O
residues	O
of	O
the	O
linker	O
and	O
the	O
U2AF65	O
RRM2	O
;	O
instead	O
,	O
the	O
C	O
-	O
terminal	O
conformation	O
of	O
the	O
linker	O
appears	O
primarily	O
RNA	O
mediated	O
(	O
Fig	O
.	O
3c	O
,	O
d	O
).	O
	O
We	O
investigated	O
whether	O
the	O
observed	O
contacts	O
between	O
the	O
RRMs	O
and	O
linker	O
were	O
critical	O
for	O
RNA	O
binding	O
by	O
structure	O
-	O
guided	O
mutagenesis	O
(	O
Fig	O
.	O
4b	O
).	O
	O
We	O
titrated	O
these	O
mutant	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
proteins	O
into	O
fluorescein	O
-	O
labelled	O
AdML	O
Py	O
-	O
tract	O
RNA	O
and	O
fit	O
the	O
fluorescence	O
anisotropy	O
changes	O
to	O
obtain	O
the	O
apparent	O
equilibrium	O
affinities	O
(	O
Supplementary	O
Fig	O
.	O
4d	O
–	O
h	O
).	O
	O
First	O
,	O
we	O
replaced	O
V249	O
and	O
V250	O
at	O
the	O
RRM1	O
/	O
RRM2	O
interface	O
and	O
V254	O
at	O
the	O
bound	O
RNA	O
site	O
with	O
glycine	O
(	O
3Gly	B-mutant
).	O
	O
A	O
more	O
conservative	O
substitution	O
of	O
these	O
five	O
residues	O
(	O
251	O
–	O
255	O
)	O
with	O
an	O
unrelated	O
sequence	O
capable	O
of	O
backbone	O
-	O
mediated	O
hydrogen	O
bonds	O
(	O
STVVP	B-mutant
>	I-mutant
NLALA	I-mutant
)	O
confirmed	O
the	O
subtle	O
impact	O
of	O
this	O
versatile	O
inter	O
-	O
RRM	O
sequence	O
on	O
affinity	O
for	O
the	O
AdML	O
Py	O
tract	O
.	O
	O
Finally	O
,	O
to	O
ensure	O
that	O
these	O
selective	O
mutations	O
were	O
sufficient	O
to	O
disrupt	O
the	O
linker	O
/	O
RRM	O
contacts	O
,	O
we	O
substituted	O
glycine	O
for	O
the	O
majority	O
of	O
buried	O
hydrophobic	O
residues	O
in	O
the	O
inter	O
-	O
RRM	O
linker	O
(	O
including	O
M144	O
,	O
L235	O
,	O
M238	O
,	O
V244	O
,	O
V246	O
,	O
V249	O
,	O
V250	O
,	O
S251	O
,	O
T252	O
,	O
V253	O
,	O
V254	O
,	O
P255	O
;	O
called	O
12Gly	B-mutant
).	O
	O
Despite	O
12	O
concurrent	O
mutations	O
,	O
the	O
AdML	O
RNA	O
affinity	O
of	O
the	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
-	I-mutant
12Gly	I-mutant
variant	O
was	O
reduced	O
by	O
only	O
three	O
-	O
fold	O
relative	O
to	O
the	O
unmodified	O
protein	O
(	O
Fig	O
.	O
4b	O
),	O
which	O
is	O
less	O
than	O
the	O
penalty	O
of	O
the	O
V254P	B-mutant
mutation	O
that	O
disrupts	O
the	O
rU5	O
hydrogen	O
bond	O
(	O
Fig	O
.	O
3d	O
,	O
i	O
).	O
	O
To	O
test	O
the	O
interplay	O
of	O
the	O
U2AF65	O
inter	O
-	O
RRM	O
linker	O
with	O
its	O
N	O
-	O
and	O
C	O
-	O
terminal	O
RRM	O
extensions	O
,	O
we	O
constructed	O
an	O
internal	O
linker	O
deletion	O
of	O
20	O
-	O
residues	O
within	O
the	O
extended	O
RNA	O
-	O
binding	O
domain	O
(	O
dU2AF651	B-mutant
,	I-mutant
2L	I-mutant
).	O
	O
We	O
found	O
that	O
the	O
affinity	O
of	O
dU2AF651	B-mutant
,	I-mutant
2L	I-mutant
for	O
the	O
AdML	O
RNA	O
was	O
significantly	O
reduced	O
relative	O
to	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
(	O
four	O
-	O
fold	O
,	O
Figs	O
1b	O
and	O
4b	O
;	O
Supplementary	O
Fig	O
.	O
4i	O
).	O
	O
Yet	O
,	O
it	O
is	O
well	O
known	O
that	O
the	O
linker	O
deletion	O
in	O
the	O
context	O
of	O
the	O
minimal	O
RRM1	O
–	O
RRM2	O
boundaries	O
has	O
no	O
detectable	O
effect	O
on	O
the	O
RNA	O
affinities	O
of	O
dU2AF651	B-mutant
,	I-mutant
2	I-mutant
compared	O
with	O
U2AF651	B-mutant
,	I-mutant
2	I-mutant
(	O
refs	O
;	O
Figs	O
1b	O
and	O
4b	O
;	O
Supplementary	O
Fig	O
.	O
4j	O
).	O
	O
The	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
structures	O
suggest	O
that	O
an	O
extended	O
conformation	O
of	O
the	O
truncated	O
dU2AF651	B-mutant
,	I-mutant
2	I-mutant
inter	O
-	O
RRM	O
linker	O
would	O
suffice	O
to	O
connect	O
the	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
RRM1	O
C	O
terminus	O
to	O
the	O
N	O
terminus	O
of	O
RRM2	O
(	O
24	O
Å	O
distance	O
between	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
R227	O
-	O
Cα	O
–	O
H259	O
-	O
Cα	O
atoms	O
),	O
which	O
agrees	O
with	O
the	O
greater	O
RNA	O
affinities	O
of	O
dU2AF651	B-mutant
,	I-mutant
2	I-mutant
and	O
U2AF651	B-mutant
,	I-mutant
2	I-mutant
dual	O
RRMs	O
compared	O
with	O
the	O
individual	O
U2AF65	O
RRMs	O
.	O
	O
However	O
,	O
stretching	O
of	O
the	O
truncated	O
dU2AF651	B-mutant
,	I-mutant
2L	I-mutant
linker	O
to	O
connect	O
the	O
RRM	O
termini	O
is	O
expected	O
to	O
disrupt	O
its	O
nucleotide	O
interactions	O
.	O
	O
Likewise	O
,	O
deletion	O
of	O
the	O
N	O
-	O
terminal	O
RRM1	O
extension	O
in	O
the	O
shortened	O
constructs	O
would	O
remove	O
packing	O
interactions	O
that	O
position	O
the	O
linker	O
in	O
a	O
kinked	O
turn	O
following	O
P229	O
(	O
Fig	O
.	O
4a	O
),	O
consistent	O
with	O
the	O
lower	O
RNA	O
affinities	O
of	O
dU2AF651	B-mutant
,	I-mutant
2L	I-mutant
,	O
dU2AF651	B-mutant
,	I-mutant
2	I-mutant
and	O
U2AF651	B-mutant
,	I-mutant
2	I-mutant
compared	O
with	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
.	O
	O
Notably	O
,	O
the	O
Q147A	B-mutant
/	O
V254P	B-mutant
/	O
R227A	B-mutant
mutation	O
reduced	O
the	O
RNA	O
affinity	O
of	O
the	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
-	I-mutant
3Mut	I-mutant
protein	O
by	O
30	O
-	O
fold	O
more	O
than	O
would	O
be	O
expected	O
based	O
on	O
simple	O
addition	O
of	O
the	O
ΔΔG	O
'	O
s	O
for	O
the	O
single	O
mutations	O
.	O
	O
This	O
difference	O
indicates	O
that	O
the	O
linearly	O
distant	O
regions	O
of	O
the	O
U2AF65	O
primary	O
sequence	O
,	O
including	O
Q147	O
in	O
the	O
N	O
-	O
terminal	O
RRM1	O
extension	O
and	O
R227	O
/	O
V254	O
in	O
the	O
N	O
-/	O
C	O
-	O
terminal	O
linker	O
regions	O
at	O
the	O
fifth	O
nucleotide	O
site	O
,	O
cooperatively	O
recognize	O
the	O
Py	O
tract	O
.	O
	O
Altogether	O
,	O
we	O
conclude	O
that	O
the	O
conformation	O
of	O
the	O
U2AF65	O
inter	O
-	O
RRM	O
linker	O
is	O
key	O
for	O
recognizing	O
RNA	O
and	O
is	O
positioned	O
by	O
the	O
RRM	O
extension	O
but	O
otherwise	O
relatively	O
independent	O
of	O
the	O
side	O
chain	O
composition	O
.	O
	O
The	O
non	O
-	O
additive	O
effects	O
of	O
the	O
Q147A	B-mutant
/	O
V254P	B-mutant
/	O
R227A	B-mutant
triple	O
mutation	O
,	O
coupled	O
with	O
the	O
context	O
-	O
dependent	O
penalties	O
of	O
an	O
internal	O
U2AF65	O
linker	O
deletion	O
,	O
highlights	O
the	O
importance	O
of	O
the	O
structural	O
interplay	O
among	O
the	O
U2AF65	O
linker	O
and	O
the	O
N	O
-	O
and	O
C	O
-	O
terminal	O
extensions	O
flanking	O
the	O
core	O
RRMs	O
.	O
	O
As	O
a	O
representative	O
splicing	O
substrate	O
,	O
we	O
utilized	O
a	O
well	O
-	O
characterized	O
minigene	O
splicing	O
reporter	O
(	O
called	O
pyPY	O
)	O
comprising	O
a	O
weak	O
(	O
that	O
is	O
,	O
degenerate	O
,	O
py	O
)	O
and	O
strong	O
(	O
that	O
is	O
,	O
U	O
-	O
rich	O
,	O
PY	O
)	O
polypyrimidine	O
tracts	O
preceding	O
two	O
alternative	O
splice	O
sites	O
(	O
Fig	O
.	O
5a	O
).	O
	O
When	O
co	O
-	O
transfected	O
with	O
an	O
expression	O
plasmid	O
for	O
wild	O
-	O
type	O
U2AF65	O
,	O
use	O
of	O
the	O
py	O
splice	O
site	O
significantly	O
increases	O
(	O
by	O
more	O
than	O
five	O
-	O
fold	O
)	O
and	O
as	O
documented	O
converts	O
a	O
fraction	O
of	O
the	O
unspliced	O
to	O
spliced	O
transcript	O
.	O
	O
We	O
introduced	O
the	O
triple	O
mutation	O
(	O
V254P	B-mutant
/	O
R227A	B-mutant
/	O
Q147A	B-mutant
)	O
that	O
significantly	O
reduced	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
association	O
with	O
the	O
Py	O
tract	O
(	O
Fig	O
.	O
4b	O
)	O
in	O
the	O
context	O
of	O
full	O
-	O
length	O
U2AF65	O
(	O
U2AF65	B-mutant
-	I-mutant
3Mut	I-mutant
).	O
	O
Co	O
-	O
transfection	O
of	O
the	O
U2AF65	B-mutant
-	I-mutant
3Mut	I-mutant
with	O
the	O
pyPY	O
splicing	O
substrate	O
significantly	O
reduced	O
splicing	O
of	O
the	O
weak	O
‘	O
py	O
'	O
splice	O
site	O
relative	O
to	O
wild	O
-	O
type	O
U2AF65	O
(	O
Fig	O
.	O
5b	O
,	O
c	O
).	O
	O
We	O
conclude	O
that	O
the	O
Py	O
-	O
tract	O
interactions	O
with	O
these	O
residues	O
of	O
the	O
U2AF65	O
inter	O
-	O
RRM	O
linker	O
and	O
RRM	O
extensions	O
are	O
important	O
for	O
splicing	O
as	O
well	O
as	O
for	O
binding	O
a	O
representative	O
of	O
the	O
major	O
U2	O
-	O
class	O
of	O
splice	O
sites	O
.	O
	O
Sparse	O
inter	O
-	O
RRM	O
contacts	O
underlie	O
apo	O
-	O
U2AF65	O
dynamics	O
	O
The	O
direct	O
interface	O
between	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
RRM1	O
and	O
RRM2	O
is	O
minor	O
,	O
burying	O
265	O
Å2	O
of	O
solvent	O
accessible	O
surface	O
area	O
compared	O
with	O
570	O
Å2	O
on	O
average	O
for	O
a	O
crystal	O
packing	O
interface	O
.	O
	O
A	O
handful	O
of	O
inter	O
-	O
RRM	O
hydrogen	O
bonds	O
are	O
apparent	O
between	O
the	O
side	O
chains	O
of	O
RRM1	O
-	O
N155	O
and	O
RRM2	O
-	O
K292	O
,	O
RRM1	O
-	O
N155	O
and	O
RRM2	O
-	O
D272	O
as	O
well	O
as	O
the	O
backbone	O
atoms	O
of	O
RRM1	O
-	O
G221	O
and	O
RRM2	O
-	O
D273	O
(	O
Fig	O
.	O
4c	O
).	O
	O
This	O
minor	O
U2AF65	O
RRM1	O
/	O
RRM2	O
interface	O
,	O
coupled	O
with	O
the	O
versatile	O
sequence	O
of	O
the	O
inter	O
-	O
RRM	O
linker	O
,	O
highlighted	O
the	O
potential	O
role	O
for	O
inter	O
-	O
RRM	O
conformational	O
dynamics	O
in	O
U2AF65	O
-	O
splice	O
site	O
recognition	O
.	O
	O
Paramagnetic	O
resonance	O
enhancement	O
(	O
PRE	O
)	O
measurements	O
previously	O
had	O
suggested	O
a	O
predominant	O
back	O
-	O
to	O
-	O
back	O
,	O
or	O
‘	O
closed	O
'	O
conformation	O
of	O
the	O
apo	O
-	O
U2AF651	B-mutant
,	I-mutant
2	I-mutant
RRM1	O
and	O
RRM2	O
in	O
equilibrium	O
with	O
a	O
minor	O
‘	O
open	O
'	O
conformation	O
resembling	O
the	O
RNA	O
-	O
bound	O
inter	O
-	O
RRM	O
arrangement	O
.	O
	O
Yet	O
,	O
small	O
-	O
angle	O
X	O
-	O
ray	O
scattering	O
(	O
SAXS	O
)	O
data	O
indicated	O
that	O
both	O
the	O
minimal	O
U2AF651	B-mutant
,	I-mutant
2	I-mutant
and	O
longer	O
constructs	O
comprise	O
a	O
highly	O
diverse	O
continuum	O
of	O
conformations	O
in	O
the	O
absence	O
of	O
RNA	O
that	O
includes	O
the	O
‘	O
closed	O
'	O
and	O
‘	O
open	O
'	O
conformations	O
.	O
	O
The	O
inter	O
-	O
RRM	O
dynamics	O
of	O
U2AF65	O
were	O
followed	O
using	O
FRET	O
between	O
fluorophores	O
attached	O
to	O
RRM1	O
and	O
RRM2	O
(	O
Fig	O
.	O
6a	O
,	O
b	O
,	O
Methods	O
).	O
	O
The	O
FRET	O
efficiencies	O
of	O
either	O
of	O
these	O
structurally	O
characterized	O
conformations	O
also	O
are	O
expected	O
to	O
be	O
significantly	O
greater	O
than	O
elongated	O
U2AF65	O
conformations	O
that	O
lack	O
inter	O
-	O
RRM	O
contacts	O
.	O
	O
We	O
first	O
characterized	O
the	O
conformational	O
dynamics	O
spectrum	O
of	O
U2AF65	O
in	O
the	O
absence	O
of	O
RNA	O
(	O
Fig	O
.	O
6c	O
,	O
d	O
;	O
Supplementary	O
Fig	O
.	O
7a	O
,	O
b	O
).	O
	O
Virtually	O
no	O
fluorescent	O
molecules	O
were	O
detected	O
in	O
the	O
absence	O
of	O
biotin	O
-	O
NTA	O
/	O
Ni	O
+	O
2	O
,	O
which	O
demonstrates	O
the	O
absence	O
of	O
detectable	O
non	O
-	O
specific	O
binding	O
of	O
U2AF651	B-mutant
,	I-mutant
2LFRET	I-mutant
to	O
the	O
slide	O
.	O
	O
The	O
FRET	O
distribution	O
histogram	O
built	O
from	O
more	O
than	O
a	O
thousand	O
traces	O
of	O
U2AF651	B-mutant
,	I-mutant
2LFRET	I-mutant
(	O
Cy3	O
/	O
Cy5	O
)	O
in	O
the	O
absence	O
of	O
ligand	O
showed	O
an	O
extremely	O
broad	O
distribution	O
centred	O
at	O
a	O
FRET	O
efficiency	O
of	O
∼	O
0	O
.	O
4	O
(	O
Fig	O
.	O
6d	O
).	O
	O
Approximately	O
40	O
%	O
of	O
the	O
smFRET	O
traces	O
showed	O
apparent	O
transitions	O
between	O
multiple	O
FRET	O
values	O
(	O
for	O
example	O
,	O
Fig	O
.	O
6c	O
).	O
	O
Despite	O
the	O
large	O
width	O
of	O
the	O
FRET	O
-	O
distribution	O
histogram	O
,	O
the	O
majority	O
(	O
80	O
%)	O
of	O
traces	O
that	O
showed	O
fluctuations	O
sampled	O
only	O
two	O
distinct	O
FRET	O
states	O
(	O
for	O
example	O
,	O
Supplementary	O
Fig	O
.	O
7a	O
).	O
	O
Approximately	O
70	O
%	O
of	O
observed	O
fluctuations	O
were	O
interchanges	O
between	O
the	O
∼	O
0	O
.	O
65	O
and	O
∼	O
0	O
.	O
45	O
FRET	O
values	O
(	O
Supplementary	O
Fig	O
.	O
7b	O
).	O
	O
We	O
cannot	O
exclude	O
a	O
possibility	O
that	O
tethering	O
of	O
U2AF651	B-mutant
,	I-mutant
2LFRET	I-mutant
(	O
Cy3	O
/	O
Cy5	O
)	O
to	O
the	O
microscope	O
slide	O
introduces	O
structural	O
heterogeneity	O
into	O
the	O
protein	O
and	O
,	O
thus	O
,	O
contributes	O
to	O
the	O
breadth	O
of	O
the	O
FRET	O
distribution	O
histogram	O
.	O
	O
We	O
conclude	O
that	O
weak	O
contacts	O
between	O
the	O
U2AF65	O
RRM1	O
and	O
RRM2	O
permit	O
dissociation	O
of	O
these	O
RRMs	O
in	O
the	O
absence	O
of	O
RNA	O
.	O
	O
U2AF65	O
conformational	O
selection	O
and	O
induced	O
fit	O
by	O
bound	O
RNA	O
	O
Addition	O
of	O
the	O
AdML	O
RNA	O
to	O
tethered	O
U2AF651	B-mutant
,	I-mutant
2LFRET	I-mutant
(	O
Cy3	O
/	O
Cy5	O
)	O
selectively	O
increases	O
a	O
fraction	O
of	O
molecules	O
showing	O
an	O
∼	O
0	O
.	O
45	O
apparent	O
FRET	O
efficiency	O
,	O
suggesting	O
that	O
RNA	O
binding	O
stabilizes	O
a	O
single	O
conformation	O
,	O
which	O
corresponds	O
to	O
the	O
0	O
.	O
45	O
FRET	O
state	O
(	O
Fig	O
.	O
6e	O
,	O
f	O
).	O
	O
We	O
tethered	O
the	O
AdML	O
RNA	O
to	O
the	O
slide	O
via	O
a	O
biotinylated	O
oligonucleotide	O
DNA	O
handle	O
and	O
added	O
U2AF651	B-mutant
,	I-mutant
2LFRET	I-mutant
(	O
Cy3	O
/	O
Cy5	O
)	O
in	O
the	O
absence	O
of	O
biotin	O
-	O
NTA	O
resin	O
(	O
Fig	O
.	O
6g	O
,	O
h	O
;	O
Supplementary	O
Fig	O
.	O
7c	O
–	O
g	O
).	O
	O
A	O
0	O
.	O
45	O
FRET	O
value	O
was	O
again	O
predominant	O
,	O
indicating	O
a	O
similar	O
RNA	O
-	O
bound	O
conformation	O
and	O
structural	O
dynamics	O
for	O
the	O
untethered	O
and	O
tethered	O
U2AF651	B-mutant
,	I-mutant
2LFRET	I-mutant
(	O
Cy3	O
/	O
Cy5	O
).	O
	O
We	O
introduced	O
an	O
rArA	O
purine	O
dinucleotide	O
within	O
a	O
variant	O
of	O
the	O
AdML	O
Py	O
tract	O
(	O
detailed	O
in	O
Methods	O
).	O
	O
Insertion	O
of	O
adenine	O
nucleotides	O
decreased	O
binding	O
affinity	O
of	O
U2AF65	O
to	O
RNA	O
by	O
approximately	O
five	O
-	O
fold	O
.	O
	O
Nevertheless	O
,	O
in	O
the	O
presence	O
of	O
saturating	O
concentrations	O
of	O
rArA	O
-	O
interrupted	O
RNA	O
slide	O
-	O
tethered	O
U2AF651	B-mutant
,	I-mutant
2LFRET	I-mutant
(	O
Cy3	O
/	O
Cy5	O
)	O
showed	O
a	O
prevalent	O
∼	O
0	O
.	O
45	O
apparent	O
FRET	O
value	O
(	O
Fig	O
.	O
6i	O
,	O
j	O
),	O
which	O
was	O
also	O
predominant	O
in	O
the	O
presence	O
of	O
continuous	O
Py	O
tract	O
.	O
	O
It	O
should	O
be	O
noted	O
that	O
inferring	O
distances	O
from	O
FRET	O
values	O
is	O
prone	O
to	O
significant	O
error	O
because	O
of	O
uncertainties	O
in	O
the	O
determination	O
of	O
fluorophore	O
orientation	O
factor	O
κ2	O
and	O
Förster	O
radius	O
R0	O
,	O
the	O
parameters	O
used	O
in	O
distance	O
calculations	O
.	O
	O
The	O
remaining	O
∼	O
30	O
%	O
of	O
traces	O
for	O
U2AF651	B-mutant
,	I-mutant
2LFRET	I-mutant
(	O
Cy3	O
/	O
Cy5	O
)	O
bound	O
to	O
the	O
slide	O
-	O
tethered	O
RNA	O
showed	O
fluctuations	O
between	O
distinct	O
FRET	O
values	O
.	O
	O
The	O
majority	O
of	O
traces	O
that	O
show	O
fluctuations	O
began	O
at	O
high	O
(	O
0	O
.	O
65	O
–	O
0	O
.	O
8	O
)	O
FRET	O
value	O
and	O
transitioned	O
to	O
a	O
∼	O
0	O
.	O
45	O
FRET	O
value	O
(	O
Supplementary	O
Fig	O
.	O
7c	O
–	O
g	O
).	O
	O
Although	O
a	O
compact	O
conformation	O
(	O
or	O
multiple	O
conformations	O
)	O
of	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
corresponding	O
to	O
∼	O
0	O
.	O
7	O
–	O
0	O
.	O
8	O
FRET	O
values	O
can	O
bind	O
RNA	O
,	O
on	O
RNA	O
binding	O
,	O
these	O
compact	O
conformations	O
of	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
transition	O
into	O
a	O
more	O
stable	O
structural	O
state	O
that	O
corresponds	O
to	O
∼	O
0	O
.	O
45	O
FRET	O
value	O
and	O
is	O
likely	O
similar	O
to	O
the	O
side	O
-	O
by	O
-	O
side	O
inter	O
-	O
RRM	O
-	O
arrangement	O
of	O
the	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
crystal	O
structures	O
.	O
	O
Thus	O
,	O
the	O
sequence	O
of	O
structural	O
rearrangements	O
of	O
U2AF65	O
observed	O
in	O
smFRET	O
traces	O
(	O
Supplementary	O
Fig	O
.	O
7c	O
–	O
g	O
)	O
suggests	O
that	O
a	O
‘	O
conformational	O
selection	O
'	O
mechanism	O
of	O
Py	O
-	O
tract	O
recognition	O
(	O
that	O
is	O
,	O
RNA	O
ligand	O
stabilization	O
of	O
a	O
pre	O
-	O
configured	O
U2AF65	O
conformation	O
)	O
is	O
complemented	O
by	O
‘	O
induced	O
fit	O
'	O
(	O
that	O
is	O
,	O
RNA	O
-	O
induced	O
rearrangement	O
of	O
the	O
U2AF65	O
RRMs	O
to	O
achieve	O
the	O
final	O
‘	O
side	O
-	O
by	O
-	O
side	O
'	O
conformation	O
),	O
as	O
discussed	O
below	O
.	O
	O
Several	O
observations	O
indicate	O
that	O
the	O
numerous	O
intramolecular	O
contacts	O
,	O
here	O
revealed	O
among	O
the	O
inter	O
-	O
RRM	O
linker	O
and	O
RRM1	O
,	O
RRM2	O
,	O
and	O
the	O
N	O
-	O
terminal	O
RRM1	O
extension	O
,	O
synergistically	O
coordinate	O
U2AF65	O
–	O
Py	O
-	O
tract	O
recognition	O
.	O
	O
Likewise	O
,	O
deletion	O
of	O
20	O
inter	O
-	O
RRM	O
linker	O
residues	O
significantly	O
reduces	O
U2AF65	O
–	O
RNA	O
binding	O
only	O
when	O
introduced	O
in	O
the	O
context	O
of	O
the	O
longer	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
construct	O
comprising	O
the	O
RRM	O
extensions	O
,	O
which	O
in	O
turn	O
position	O
the	O
linker	O
for	O
RNA	O
interactions	O
.	O
	O
Notably	O
,	O
a	O
triple	O
mutation	O
of	O
three	O
residues	O
(	O
V254P	B-mutant
,	O
Q147A	B-mutant
and	O
R227A	B-mutant
)	O
in	O
the	O
respective	O
inter	O
-	O
RRM	O
linker	O
,	O
N	O
-	O
and	O
C	O
-	O
terminal	O
extensions	O
non	O
-	O
additively	O
reduce	O
RNA	O
binding	O
by	O
150	O
-	O
fold	O
.	O
	O
Altogether	O
,	O
these	O
data	O
indicate	O
that	O
interactions	O
among	O
the	O
U2AF65	O
RRM1	O
/	O
RRM2	O
,	O
inter	O
-	O
RRM	O
linker	O
,	O
N	O
-	O
and	O
C	O
-	O
terminal	O
extensions	O
are	O
mutually	O
inter	O
-	O
dependent	O
for	O
cognate	O
Py	O
-	O
tract	O
recognition	O
.	O
	O
The	O
implications	O
of	O
this	O
finding	O
for	O
U2AF65	O
conservation	O
and	O
Py	O
-	O
tract	O
recognition	O
are	O
detailed	O
in	O
the	O
Supplementary	O
Discussion	O
.	O
	O
Recently	O
,	O
high	O
-	O
throughput	O
sequencing	O
studies	O
have	O
shown	O
that	O
somatic	O
mutations	O
in	O
pre	O
-	O
mRNA	O
splicing	O
factors	O
occur	O
in	O
the	O
majority	O
of	O
patients	O
with	O
myelodysplastic	O
syndrome	O
(	O
MDS	O
).	O
	O
MDS	O
-	O
relevant	O
mutations	O
are	O
common	O
in	O
the	O
small	O
U2AF	O
subunit	O
(	O
U2AF35	O
,	O
or	O
U2AF1	O
),	O
yet	O
such	O
mutations	O
are	O
rare	O
in	O
the	O
large	O
U2AF65	O
subunit	O
(	O
also	O
called	O
U2AF2	O
)—	O
possibly	O
due	O
to	O
the	O
selective	O
versus	O
nearly	O
universal	O
requirements	O
of	O
these	O
factors	O
for	O
splicing	O
.	O
	O
A	O
confirmed	O
somatic	O
mutation	O
of	O
U2AF65	O
in	O
patients	O
with	O
MDS	O
,	O
L187V	B-mutant
,	O
is	O
located	O
on	O
a	O
solvent	O
-	O
exposed	O
surface	O
of	O
RRM1	O
that	O
is	O
distinct	O
from	O
the	O
RNA	O
interface	O
(	O
Fig	O
.	O
7a	O
).	O
	O
This	O
L187	O
surface	O
is	O
oriented	O
towards	O
the	O
N	O
terminus	O
of	O
the	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
construct	O
,	O
where	O
it	O
is	O
expected	O
to	O
abut	O
the	O
U2AF35	O
-	O
binding	O
site	O
in	O
the	O
context	O
of	O
the	O
full	O
-	O
length	O
U2AF	O
heterodimer	O
.	O
	O
Likewise	O
,	O
an	O
unconfirmed	O
M144I	B-mutant
mutation	O
reported	O
by	O
the	O
same	O
group	O
corresponds	O
to	O
the	O
N	O
-	O
terminal	O
residue	O
of	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
,	O
which	O
is	O
separated	O
by	O
only	O
∼	O
20	O
residues	O
from	O
the	O
U2AF35	O
-	O
binding	O
site	O
.	O
	O
Our	O
smFRET	O
results	O
agree	O
with	O
prior	O
NMR	O
/	O
PRE	O
evidence	O
for	O
multi	O
-	O
domain	O
conformational	O
selection	O
as	O
one	O
mechanistic	O
basis	O
for	O
U2AF65	O
–	O
RNA	O
association	O
(	O
Fig	O
.	O
7b	O
).	O
	O
An	O
∼	O
0	O
.	O
45	O
FRET	O
value	O
is	O
likely	O
to	O
correspond	O
to	O
the	O
U2AF65	O
conformation	O
visualized	O
in	O
our	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
crystal	O
structures	O
,	O
in	O
which	O
the	O
RRM1	O
and	O
RRM2	O
bind	O
side	O
-	O
by	O
-	O
side	O
to	O
the	O
Py	O
-	O
tract	O
oligonucleotide	O
.	O
	O
The	O
lesser	O
0	O
.	O
65	O
–	O
0	O
.	O
8	O
and	O
0	O
.	O
2	O
–	O
0	O
.	O
3	O
FRET	O
values	O
in	O
the	O
untethered	O
U2AF651	B-mutant
,	I-mutant
2LFRET	I-mutant
(	O
Cy3	O
/	O
Cy5	O
)	O
experiment	O
could	O
correspond	O
to	O
respective	O
variants	O
of	O
the	O
‘	O
closed	O
',	O
back	O
-	O
to	O
-	O
back	O
U2AF65	O
conformations	O
characterized	O
by	O
NMR	O
/	O
PRE	O
data	O
,	O
or	O
to	O
extended	O
U2AF65	O
conformations	O
,	O
in	O
which	O
the	O
intramolecular	O
RRM1	O
/	O
RRM2	O
interactions	O
have	O
dissociated	O
the	O
protein	O
is	O
bound	O
to	O
RNA	O
via	O
single	O
RRMs	O
.	O
	O
Notably	O
,	O
our	O
smFRET	O
results	O
reveal	O
that	O
U2AF65	O
–	O
Py	O
-	O
tract	O
recognition	O
can	O
be	O
characterized	O
by	O
an	O
‘	O
extended	O
conformational	O
selection	O
'	O
model	O
(	O
Fig	O
.	O
7b	O
).	O
	O
Here	O
,	O
the	O
majority	O
of	O
changes	O
in	O
smFRET	O
traces	O
for	O
U2AF651	B-mutant
,	I-mutant
2LFRET	I-mutant
(	O
Cy3	O
/	O
Cy5	O
)	O
bound	O
to	O
slide	O
-	O
tethered	O
RNA	O
began	O
at	O
high	O
(	O
0	O
.	O
65	O
–	O
0	O
.	O
8	O
)	O
FRET	O
value	O
and	O
transition	O
to	O
the	O
predominant	O
0	O
.	O
45	O
FRET	O
value	O
(	O
Supplementary	O
Fig	O
.	O
7c	O
–	O
g	O
).	O
	O
As	O
such	O
,	O
the	O
smFRET	O
approach	O
reconciles	O
prior	O
inconsistencies	O
between	O
two	O
major	O
conformations	O
that	O
were	O
detected	O
by	O
NMR	O
/	O
PRE	O
experiments	O
and	O
a	O
broad	O
ensemble	O
of	O
diverse	O
inter	O
-	O
RRM	O
arrangements	O
that	O
fit	O
the	O
SAXS	O
data	O
for	O
the	O
apo	O
-	O
protein	O
.	O
	O
Similar	O
interdisciplinary	O
structural	O
approaches	O
are	O
likely	O
to	O
illuminate	O
whether	O
similar	O
mechanistic	O
bases	O
for	O
RNA	O
binding	O
are	O
widespread	O
among	O
other	O
members	O
of	O
the	O
vast	O
multi	O
-	O
RRM	O
family	O
.	O
	O
Based	O
on	O
(	O
i	O
)	O
similar	O
RNA	O
affinities	O
of	O
U2AF65	O
and	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
,	O
(	O
ii	O
)	O
indistinguishable	O
conformations	O
among	O
four	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
structures	O
in	O
two	O
different	O
crystal	O
packing	O
arrangements	O
and	O
(	O
iii	O
)	O
penalties	O
of	O
structure	O
-	O
guided	O
mutations	O
in	O
RNA	O
binding	O
and	O
splicing	O
assays	O
,	O
we	O
suggest	O
that	O
the	O
extended	O
inter	O
-	O
RRM	O
regions	O
of	O
the	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
structures	O
underlie	O
cognate	O
Py	O
-	O
tract	O
recognition	O
by	O
the	O
full	O
-	O
length	O
U2AF65	O
protein	O
.	O
	O
Further	O
research	O
will	O
be	O
needed	O
to	O
understand	O
the	O
roles	O
of	O
SF1	O
and	O
U2AF35	O
subunits	O
in	O
the	O
conformational	O
equilibria	O
underlying	O
U2AF65	O
association	O
with	O
Py	O
tracts	O
.	O
	O
Moreover	O
,	O
structural	O
differences	O
among	O
U2AF65	O
homologues	O
and	O
paralogues	O
may	O
regulate	O
splice	O
site	O
selection	O
.	O
	O
Ultimately	O
,	O
these	O
guidelines	O
will	O
assist	O
the	O
identification	O
of	O
3	O
′	O
splice	O
sites	O
and	O
the	O
relationship	O
of	O
disease	O
-	O
causing	O
mutations	O
to	O
penalties	O
for	O
U2AF65	O
association	O
.	O
	O
(	O
a	O
)	O
Domain	O
organization	O
of	O
full	O
-	O
length	O
(	O
fl	O
)	O
U2AF65	O
and	O
constructs	O
used	O
for	O
RNA	O
binding	O
and	O
structural	O
experiments	O
.	O
	O
An	O
internal	O
deletion	O
(	O
d	B-mutant
,	O
Δ	B-mutant
)	O
of	O
residues	O
238	O
–	O
257	O
removes	O
a	O
portion	O
of	O
the	O
inter	O
-	O
RRM	O
linker	O
from	O
the	O
dU2AF651	B-mutant
,	I-mutant
2	I-mutant
and	O
dU2AF651	B-mutant
,	I-mutant
2L	I-mutant
constructs	O
.	O
	O
The	O
flU2AF65	O
protein	O
includes	O
a	O
heterodimerization	O
domain	O
of	O
the	O
U2AF35	O
subunit	O
to	O
promote	O
solubility	O
and	O
folding	O
.	O
	O
The	O
KD	O
'	O
s	O
for	O
binding	O
5	O
′-	O
CCUUUUCCCCCCC	O
-	O
3	O
′	O
are	O
:	O
flU2AF65	O
,	O
41	O
±	O
2	O
nM	O
;	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
,	O
31	O
±	O
3	O
nM	O
.	O
The	O
KD	O
'	O
s	O
for	O
binding	O
5	O
′-	O
CCCCCCCUUUUCC	O
-	O
3	O
′	O
are	O
:	O
flU2AF65	O
,	O
414	O
±	O
12	O
nM	O
;	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
,	O
417	O
±	O
10	O
nM	O
.	O
Bar	O
graphs	O
are	O
hatched	O
to	O
match	O
the	O
constructs	O
shown	O
in	O
a	O
.	O
The	O
average	O
apparent	O
equilibrium	O
affinity	O
(	O
KA	O
)	O
and	O
s	O
.	O
e	O
.	O
m	O
.	O
for	O
three	O
independent	O
titrations	O
are	O
plotted	O
.	O
	O
RRM	O
,	O
RNA	O
recognition	O
motif	O
;	O
RS	O
,	O
arginine	O
-	O
serine	O
rich	O
;	O
UHM	O
,	O
U2AF	O
homology	O
motif	O
;	O
ULM	O
,	O
U2AF	O
ligand	O
motif	O
.	O
	O
(	O
a	O
)	O
Alignment	O
of	O
oligonucleotide	O
sequences	O
that	O
were	O
co	O
-	O
crystallized	O
in	O
the	O
indicated	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
structures	O
.	O
	O
The	O
regions	O
of	O
RRM1	O
,	O
RRM2	O
and	O
linker	O
contacts	O
are	O
indicated	O
above	O
by	O
respective	O
black	O
and	O
blue	O
arrows	O
from	O
N	O
-	O
to	O
C	O
-	O
terminus	O
.	O
	O
The	O
prior	O
dU2AF651	B-mutant
,	I-mutant
2	I-mutant
nucleotide	O
-	O
binding	O
sites	O
are	O
given	O
in	O
parentheses	O
(	O
site	O
4	O
'	O
interacts	O
with	O
dU2AF65	B-mutant
RRM1	O
and	O
RRM2	O
by	O
crystallographic	O
symmetry	O
).	O
	O
Crystallographic	O
statistics	O
are	O
given	O
in	O
Table	O
1	O
and	O
the	O
overall	O
conformations	O
of	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
and	O
prior	O
dU2AF651	B-mutant
,	I-mutant
2	I-mutant
/	O
U2AF651	B-mutant
,	I-mutant
2	I-mutant
structures	O
are	O
compared	O
in	O
Supplementary	O
Fig	O
.	O
2	O
.	O
	O
Representative	O
views	O
of	O
the	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
interactions	O
with	O
each	O
new	O
nucleotide	O
of	O
the	O
bound	O
Py	O
tract	O
.	O
	O
New	O
residues	O
of	O
the	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
structures	O
are	O
coloured	O
a	O
darker	O
shade	O
of	O
blue	O
,	O
apart	O
from	O
residues	O
that	O
were	O
tested	O
by	O
site	O
-	O
directed	O
mutagenesis	O
,	O
which	O
are	O
coloured	O
yellow	O
.	O
	O
The	O
nucleotide	O
-	O
binding	O
sites	O
of	O
the	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
and	O
prior	O
dU2AF651	B-mutant
,	I-mutant
2	I-mutant
structure	O
are	O
compared	O
in	O
Supplementary	O
Fig	O
.	O
3a	O
–	O
h	O
.	O
The	O
first	O
and	O
seventh	O
U2AF651	O
,	O
2L	O
-	O
binding	O
sites	O
are	O
unchanged	O
from	O
the	O
prior	O
dU2AF651	O
,	O
2	O
–	O
RNA	O
structure	O
and	O
are	O
portrayed	O
in	O
Supplementary	O
Fig	O
.	O
3a	O
,	O
f	O
.	O
The	O
four	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
structures	O
are	O
similar	O
with	O
the	O
exception	O
of	O
pH	O
-	O
dependent	O
variations	O
at	O
the	O
ninth	O
site	O
that	O
are	O
detailed	O
in	O
Supplementary	O
Fig	O
.	O
3i	O
,	O
j	O
.	O
The	O
representative	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
structure	O
shown	O
has	O
the	O
highest	O
resolution	O
and	O
/	O
or	O
ribose	O
nucleotide	O
at	O
the	O
given	O
site	O
:	O
(	O
a	O
)	O
rU2	O
of	O
structure	O
iv	O
;	O
(	O
b	O
)	O
rU3	O
of	O
structure	O
iii	O
;	O
(	O
c	O
)	O
rU4	O
of	O
structure	O
i	O
;	O
(	O
d	O
)	O
rU5	O
of	O
structure	O
iii	O
;	O
(	O
e	O
)	O
rU6	O
of	O
structure	O
ii	O
;	O
(	O
f	O
)	O
dU8	O
of	O
structure	O
iii	O
;	O
(	O
g	O
)	O
dU9	O
of	O
structure	O
iii	O
;	O
(	O
h	O
)	O
rC9	O
of	O
structure	O
iv	O
.	O
	O
(	O
i	O
)	O
Bar	O
graph	O
of	O
apparent	O
equilibrium	O
affinities	O
(	O
KA	O
)	O
of	O
the	O
wild	O
type	O
(	O
blue	O
)	O
and	O
the	O
indicated	O
mutant	O
(	O
yellow	O
)	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
proteins	O
binding	O
the	O
AdML	O
Py	O
tract	O
(	O
5	O
′-	O
CCCUUUUUUUUCC	O
-	O
3	O
′).	O
	O
The	O
apparent	O
equilibrium	O
dissociation	O
constants	O
(	O
KD	O
)	O
of	O
the	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
mutant	O
proteins	O
are	O
:	O
wild	O
type	O
(	O
WT	O
),	O
35	O
±	O
6	O
nM	O
;	O
R227A	B-mutant
,	O
166	O
±	O
2	O
nM	O
;	O
V254P	B-mutant
,	O
137	O
±	O
10	O
nM	O
;	O
Q147A	B-mutant
,	O
171	O
±	O
21	O
nM	O
.	O
The	O
average	O
KA	O
and	O
s	O
.	O
e	O
.	O
m	O
.	O
for	O
three	O
independent	O
titrations	O
are	O
plotted	O
.	O
	O
The	O
average	O
fitted	O
fluorescence	O
anisotropy	O
RNA	O
-	O
binding	O
curves	O
are	O
shown	O
in	O
Supplementary	O
Fig	O
.	O
4a	O
–	O
c	O
.	O
	O
(	O
a	O
)	O
Contacts	O
of	O
the	O
U2AF65	O
inter	O
-	O
RRM	O
linker	O
with	O
the	O
RRMs	O
.	O
	O
A	O
semi	O
-	O
transparent	O
space	O
-	O
filling	O
surface	O
is	O
shown	O
for	O
the	O
RRM1	O
(	O
green	O
)	O
and	O
RRM2	O
(	O
light	O
blue	O
).	O
	O
Residues	O
V249	O
,	O
V250	O
,	O
V254	O
(	O
yellow	O
)	O
are	O
mutated	O
to	O
V249G	B-mutant
/	O
V250G	B-mutant
/	O
V254G	B-mutant
in	O
the	O
3Gly	B-mutant
mutant	I-mutant
;	O
residues	O
S251	O
,	O
T252	O
,	O
V253	O
,	O
P255	O
(	O
red	O
)	O
along	O
with	O
V254	O
are	O
mutated	O
to	O
S251G	B-mutant
/	O
T252G	B-mutant
/	O
V253G	B-mutant
/	O
V254G	B-mutant
/	O
P255G	B-mutant
in	O
the	O
5Gly	B-mutant
mutant	I-mutant
or	O
to	O
S251N	B-mutant
/	O
T252L	B-mutant
/	O
V253A	B-mutant
/	O
V254L	B-mutant
/	O
P255A	B-mutant
in	O
the	O
NLALA	B-mutant
mutant	I-mutant
;	O
residues	O
M144	O
,	O
L235	O
,	O
M238	O
,	O
V244	O
,	O
V246	O
(	O
orange	O
)	O
along	O
with	O
V249	O
,	O
V250	O
,	O
S251	O
,	O
T252	O
,	O
V253	O
,	O
V254	O
,	O
P255	O
are	O
mutated	O
to	O
M144G	B-mutant
/	O
L235G	B-mutant
/	O
M238G	B-mutant
/	O
V244G	B-mutant
/	O
V246G	B-mutant
/	O
V249G	B-mutant
/	O
V250G	B-mutant
/	O
S251G	B-mutant
/	O
T252G	B-mutant
/	O
V253G	B-mutant
/	O
V254G	B-mutant
/	O
P255G	B-mutant
in	O
the	O
12Gly	B-mutant
mutant	I-mutant
.	O
	O
Other	O
linker	O
residues	O
are	O
coloured	O
either	O
dark	O
blue	O
for	O
new	O
residues	O
in	O
the	O
U2AF651	B-mutant
,	I-mutant
2L	I-mutant
structure	O
or	O
light	O
blue	O
for	O
the	O
remaining	O
inter	O
-	O
RRM	O
residues	O
.	O
	O
The	O
central	O
panel	O
shows	O
an	O
overall	O
view	O
with	O
stick	O
diagrams	O
for	O
mutated	O
residues	O
;	O
boxed	O
regions	O
are	O
expanded	O
to	O
show	O
the	O
C	O
-	O
terminal	O
(	O
bottom	O
left	O
)	O
and	O
central	O
linker	O
regions	O
(	O
top	O
)	O
at	O
the	O
inter	O
-	O
RRM	O
interfaces	O
,	O
and	O
N	O
-	O
terminal	O
linker	O
region	O
contacts	O
with	O
RRM1	O
(	O
bottom	O
right	O
).	O
	O
The	O
fitted	O
fluorescence	O
anisotropy	O
RNA	O
-	O
binding	O
curves	O
are	O
shown	O
in	O
Supplementary	O
Fig	O
.	O
4d	O
–	O
j	O
.	O
(	O
c	O
)	O
Close	O
view	O
of	O
the	O
U2AF65	O
RRM1	O
/	O
RRM2	O
interface	O
following	O
a	O
two	O
-	O
fold	O
rotation	O
about	O
the	O
x	O
-	O
axis	O
relative	O
to	O
a	O
.	O
	O
U2AF65	O
inter	O
-	O
domain	O
residues	O
are	O
important	O
for	O
splicing	O
a	O
representative	O
pre	O
-	O
mRNA	O
substrate	O
in	O
human	O
cells	O
.	O
	O
(	O
a	O
)	O
Schematic	O
diagram	O
of	O
the	O
pyPY	O
reporter	O
minigene	O
construct	O
comprising	O
two	O
alternative	O
splice	O
sites	O
preceded	O
by	O
either	O
the	O
weak	O
IgM	O
Py	O
tract	O
(	O
py	O
)	O
or	O
the	O
strong	O
AdML	O
Py	O
tract	O
(	O
PY	O
)	O
(	O
sequences	O
inset	O
).	O
	O
(	O
b	O
)	O
Representative	O
RT	O
-	O
PCR	O
of	O
pyPY	O
transcripts	O
from	O
HEK293T	O
cells	O
co	O
-	O
transfected	O
with	O
constructs	O
encoding	O
the	O
pyPY	O
minigene	O
and	O
either	O
wild	O
-	O
type	O
(	O
WT	O
)	O
U2AF65	O
or	O
a	O
triple	O
U2AF65	O
mutant	O
(	O
3Mut	B-mutant
)	O
of	O
Q147A	B-mutant
,	O
R227A	B-mutant
and	O
V254P	B-mutant
residues	O
.	O
(	O
c	O
)	O
A	O
bar	O
graph	O
of	O
the	O
average	O
percentage	O
of	O
the	O
py	O
-	O
spliced	O
mRNA	O
relative	O
to	O
total	O
detected	O
pyPY	O
transcripts	O
(	O
spliced	O
and	O
unspliced	O
)	O
for	O
the	O
corresponding	O
gel	O
lanes	O
(	O
black	O
,	O
no	O
U2AF65	O
added	O
;	O
white	O
,	O
WT	O
U2AF65	O
;	O
grey	O
,	O
3Mut	B-mutant
U2AF65	O
).	O
	O
Protein	O
overexpression	O
and	O
qRT	O
-	O
PCR	O
results	O
are	O
shown	O
in	O
Supplementary	O
Fig	O
.	O
5	O
.	O
	O
The	O
U2AF651	B-mutant
,	I-mutant
2LFRET	I-mutant
proteins	O
were	O
doubly	O
labelled	O
at	O
A181C	B-mutant
/	O
Q324C	B-mutant
such	O
that	O
a	O
mixture	O
of	O
Cy3	O
/	O
Cy5	O
fluorophores	O
are	O
expected	O
to	O
be	O
present	O
at	O
each	O
site	O
.	O
	O
(	O
c	O
–	O
f	O
,	O
i	O
,	O
j	O
)	O
The	O
U2AF651	B-mutant
,	I-mutant
2LFRET	I-mutant
(	O
Cy3	O
/	O
Cy5	O
)	O
protein	O
was	O
immobilized	O
on	O
the	O
microscope	O
slide	O
via	O
biotin	O
-	O
NTA	O
/	O
Ni	O
+	O
2	O
(	O
orange	O
line	O
)	O
on	O
a	O
neutravidin	O
(	O
black	O
X	O
)-	O
biotin	O
-	O
PEG	O
(	O
orange	O
triangle	O
)-	O
treated	O
surface	O
and	O
imaged	O
either	O
in	O
the	O
absence	O
of	O
ligands	O
(	O
c	O
,	O
d	O
),	O
in	O
the	O
presence	O
of	O
5	O
μM	O
AdML	O
Py	O
-	O
tract	O
RNA	O
(	O
5	O
′-	O
CCUUUUUUUUCC	O
-	O
3	O
′)	O
(	O
e	O
,	O
f	O
),	O
or	O
in	O
the	O
presence	O
of	O
10	O
μM	O
adenosine	O
-	O
interrupted	O
variant	O
RNA	O
(	O
5	O
′-	O
CUUUUUAAUUUCCA	O
-	O
3	O
′)	O
(	O
i	O
,	O
j	O
).	O
	O
The	O
untethered	O
U2AF651	B-mutant
,	I-mutant
2LFRET	I-mutant
(	O
Cy3	O
/	O
Cy5	O
)	O
protein	O
(	O
1	O
nM	O
)	O
was	O
added	O
to	O
AdML	O
RNA	O
–	O
polyethylene	O
-	O
glycol	O
-	O
linker	O
–	O
DNA	O
oligonucleotide	O
(	O
10	O
nM	O
),	O
which	O
was	O
immobilized	O
on	O
the	O
microscope	O
slide	O
by	O
annealing	O
with	O
a	O
complementary	O
biotinyl	O
-	O
DNA	O
oligonucleotide	O
(	O
black	O
vertical	O
line	O
).	O
	O
Typical	O
single	O
-	O
molecule	O
FRET	O
traces	O
(	O
c	O
,	O
e	O
,	O
g	O
,	O
i	O
)	O
show	O
fluorescence	O
intensities	O
from	O
Cy3	O
(	O
green	O
)	O
and	O
Cy5	O
(	O
red	O
)	O
and	O
the	O
calculated	O
apparent	O
FRET	O
efficiency	O
(	O
blue	O
).	O
	O
N	O
is	O
the	O
number	O
of	O
single	O
-	O
molecule	O
traces	O
compiled	O
for	O
each	O
histogram	O
.	O
	O
Schematic	O
models	O
of	O
U2AF65	O
recognizing	O
the	O
Py	O
tract	O
.	O
	O
(	O
a	O
)	O
Diagram	O
of	O
the	O
U2AF65	O
,	O
SF1	O
and	O
U2AF35	O
splicing	O
factors	O
bound	O
to	O
the	O
consensus	O
elements	O
of	O
the	O
3	O
′	O
splice	O
site	O
.	O
	O
MDS	O
-	O
relevant	O
mutated	O
residues	O
of	O
U2AF65	O
are	O
shown	O
as	O
yellow	O
spheres	O
(	O
L187	O
and	O
M144	O
).	O
	O
Alternatively	O
,	O
a	O
conformation	O
of	O
U2AF65	O
corresponding	O
to	O
∼	O
0	O
.	O
45	O
FRET	O
value	O
can	O
directly	O
bind	O
to	O
RNA	O
;	O
RNA	O
binding	O
stabilizes	O
the	O
‘	O
open	O
',	O
side	O
-	O
by	O
-	O
side	O
conformation	O
and	O
thus	O
shifts	O
the	O
U2AF65	O
population	O
towards	O
the	O
∼	O
0	O
.	O
45	O
FRET	O
value	O
.	O
	O
RRM1	O
,	O
green	O
;	O
RRM2	O
,	O
pale	O
blue	O
;	O
RRM	O
extensions	O
/	O
linker	O
,	O
blue	O
.	O
	O
Systematic	O
analysis	O
of	O
radiation	O
damage	O
within	O
a	O
protein	O
–	O
RNA	O
complex	O
over	O
a	O
large	O
dose	O
range	O
(	O
1	O
.	O
3	O
–	O
25	O
MGy	O
)	O
reveals	O
significant	O
differential	O
susceptibility	O
of	O
RNA	O
and	O
protein	O
.	O
	O
A	O
new	O
method	O
of	O
difference	O
electron	O
-	O
density	O
quantification	O
is	O
presented	O
.	O
	O
Radiation	O
damage	O
during	O
macromolecular	O
X	O
-	O
ray	O
crystallographic	O
data	O
collection	O
is	O
still	O
the	O
main	O
impediment	O
for	O
many	O
macromolecular	O
structure	O
determinations	O
.	O
	O
Although	O
this	O
has	O
been	O
well	O
characterized	O
within	O
protein	O
crystals	O
,	O
far	O
less	O
is	O
known	O
about	O
specific	O
damage	O
effects	O
within	O
the	O
larger	O
class	O
of	O
nucleoprotein	O
complexes	O
.	O
	O
Here	O
,	O
a	O
methodology	O
has	O
been	O
developed	O
whereby	O
per	O
-	O
atom	O
density	O
changes	O
could	O
be	O
quantified	O
with	O
increasing	O
dose	O
over	O
a	O
wide	O
(	O
1	O
.	O
3	O
–	O
25	O
.	O
0	O
MGy	O
)	O
range	O
and	O
at	O
higher	O
resolution	O
(	O
1	O
.	O
98	O
Å	O
)	O
than	O
the	O
previous	O
systematic	O
specific	O
damage	O
study	O
on	O
a	O
protein	O
–	O
DNA	O
complex	O
.	O
	O
Specific	O
damage	O
manifestations	O
were	O
determined	O
within	O
the	O
large	O
trp	O
RNA	O
-	O
binding	O
attenuation	O
protein	O
(	O
TRAP	O
)	O
bound	O
to	O
a	O
single	O
-	O
stranded	O
RNA	O
that	O
forms	O
a	O
belt	O
around	O
the	O
protein	O
.	O
	O
The	O
11	O
-	O
fold	O
symmetry	O
within	O
each	O
TRAP	O
ring	O
permitted	O
statistically	O
significant	O
analysis	O
of	O
the	O
Glu	O
and	O
Asp	O
damage	O
patterns	O
,	O
with	O
RNA	O
binding	O
unexpectedly	O
being	O
observed	O
to	O
protect	O
these	O
otherwise	O
highly	O
sensitive	O
residues	O
within	O
the	O
11	O
RNA	O
-	O
binding	O
pockets	O
distributed	O
around	O
the	O
outside	O
of	O
the	O
protein	O
molecule	O
.	O
	O
Additionally	O
,	O
the	O
method	O
enabled	O
a	O
quantification	O
of	O
the	O
reduction	O
in	O
radiation	O
-	O
induced	O
Lys	O
and	O
Phe	O
disordering	O
upon	O
RNA	O
binding	O
directly	O
from	O
the	O
electron	O
density	O
.	O
	O
Significant	O
progress	O
has	O
been	O
made	O
in	O
recent	O
years	O
in	O
understanding	O
the	O
inevitable	O
manifestations	O
of	O
X	O
-	O
ray	O
-	O
induced	O
RD	O
within	O
protein	O
crystals	O
,	O
and	O
there	O
is	O
now	O
a	O
body	O
of	O
literature	O
on	O
possible	O
strategies	O
to	O
mitigate	O
the	O
effects	O
of	O
RD	O
(	O
e	O
.	O
g	O
.	O
Zeldin	O
,	O
Brockhauser	O
et	O
al	O
.,	O
2013	O
;	O
Bourenkov	O
&	O
Popov	O
,	O
2010	O
).	O
	O
However	O
,	O
there	O
is	O
still	O
no	O
general	O
consensus	O
within	O
the	O
field	O
on	O
how	O
to	O
minimize	O
RD	O
during	O
MX	O
data	O
collection	O
,	O
and	O
debates	O
on	O
the	O
dependence	O
of	O
RD	O
progression	O
on	O
incident	O
X	O
-	O
ray	O
energy	O
(	O
Shimizu	O
et	O
al	O
.,	O
2007	O
;	O
Liebschner	O
et	O
al	O
.,	O
2015	O
)	O
and	O
the	O
efficacy	O
of	O
radical	O
scavengers	O
(	O
Allan	O
et	O
al	O
.,	O
2013	O
)	O
have	O
yet	O
to	O
be	O
resolved	O
.	O
	O
Global	O
radiation	O
damage	O
is	O
observed	O
within	O
reciprocal	O
space	O
as	O
the	O
overall	O
decay	O
of	O
the	O
summed	O
intensity	O
of	O
reflections	O
detected	O
within	O
the	O
diffraction	O
pattern	O
as	O
dose	O
increases	O
(	O
Garman	O
,	O
2010	O
;	O
Murray	O
&	O
Garman	O
,	O
2002	O
).	O
	O
At	O
100	O
K	O
,	O
an	O
experimental	O
dose	O
limit	O
of	O
30	O
MGy	O
has	O
been	O
recommended	O
as	O
an	O
upper	O
limit	O
beyond	O
which	O
the	O
biological	O
information	O
derived	O
from	O
any	O
macromolecular	O
crystal	O
may	O
be	O
compromised	O
(	O
Owen	O
et	O
al	O
.,	O
2006	O
).	O
	O
Indeed	O
,	O
the	O
C	O
—	O
Se	O
bond	O
in	O
selenomethionine	O
,	O
the	O
stability	O
of	O
which	O
is	O
key	O
for	O
the	O
success	O
of	O
experimental	O
phasing	O
methods	O
,	O
can	O
be	O
cleaved	O
at	O
a	O
dose	O
as	O
low	O
as	O
2	O
MGy	O
for	O
a	O
crystal	O
maintained	O
at	O
100	O
K	O
(	O
Holton	O
,	O
2007	O
).	O
	O
Active	O
-	O
site	O
residues	O
appear	O
to	O
be	O
particularly	O
susceptible	O
,	O
particularly	O
for	O
photosensitive	O
proteins	O
and	O
in	O
instances	O
where	O
chemical	O
strain	O
is	O
an	O
intrinsic	O
feature	O
of	O
the	O
reaction	O
mechanism	O
.	O
	O
For	O
instance	O
,	O
structure	O
determination	O
of	O
the	O
purple	O
membrane	O
protein	O
bacterio	O
­	O
rhodopsin	O
required	O
careful	O
corrections	O
for	O
radiation	O
-	O
induced	O
structural	O
changes	O
before	O
the	O
correct	O
photosensitive	O
intermediate	O
states	O
could	O
be	O
isolated	O
(	O
Matsui	O
et	O
al	O
.,	O
2002	O
).	O
	O
The	O
significant	O
chemical	O
strain	O
required	O
for	O
catalysis	O
within	O
the	O
active	O
site	O
of	O
phosphoserine	O
aminotransferase	O
has	O
been	O
observed	O
to	O
diminish	O
during	O
X	O
-	O
ray	O
exposure	O
(	O
Dubnovitsky	O
et	O
al	O
.,	O
2005	O
).	O
	O
Since	O
the	O
majority	O
of	O
SRD	O
studies	O
to	O
date	O
have	O
focused	O
on	O
proteins	O
,	O
much	O
less	O
is	O
known	O
about	O
the	O
effects	O
of	O
X	O
-	O
ray	O
irradiation	O
on	O
the	O
wider	O
class	O
of	O
crystalline	O
nucleoprotein	O
complexes	O
or	O
how	O
to	O
correct	O
for	O
such	O
radiation	O
-	O
induced	O
structural	O
changes	O
.	O
	O
It	O
is	O
essential	O
to	O
understand	O
how	O
these	O
increasingly	O
complex	O
macromolecular	O
structures	O
are	O
affected	O
by	O
the	O
radiation	O
used	O
to	O
solve	O
them	O
.	O
	O
Nucleoproteins	O
also	O
represent	O
one	O
of	O
the	O
main	O
targets	O
of	O
radiotherapy	O
,	O
and	O
an	O
insight	O
into	O
the	O
damage	O
mechanisms	O
induced	O
by	O
X	O
-	O
ray	O
irradiation	O
could	O
inform	O
innovative	O
treatments	O
.	O
	O
Investigations	O
on	O
sub	O
-	O
ionization	O
-	O
level	O
LEEs	O
(	O
0	O
–	O
15	O
eV	O
)	O
interacting	O
with	O
both	O
dried	O
and	O
aqueous	O
oligonucleotides	O
(	O
Alizadeh	O
&	O
Sanche	O
,	O
2014	O
;	O
Simons	O
,	O
2006	O
)	O
concluded	O
that	O
resonant	O
electron	O
attachment	O
to	O
DNA	O
bases	O
and	O
the	O
sugar	O
-	O
phosphate	O
backbone	O
could	O
lead	O
to	O
the	O
preferential	O
cleavage	O
of	O
strong	O
(∼	O
4	O
eV	O
,	O
385	O
kJ	O
mol	O
−	O
1	O
)	O
sugar	O
-	O
phosphate	O
C	O
—	O
O	O
covalent	O
bonds	O
within	O
the	O
DNA	O
backbone	O
and	O
then	O
base	O
-	O
sugar	O
N1	O
—	O
C	O
bonds	O
,	O
eventually	O
leading	O
to	O
single	O
-	O
strand	O
breakages	O
(	O
SSBs	O
;	O
Ptasińska	O
&	O
Sanche	O
,	O
2007	O
).	O
	O
Electrons	O
have	O
been	O
shown	O
to	O
be	O
mobile	O
at	O
77	O
K	O
by	O
electron	O
spin	O
resonance	O
spectroscopy	O
studies	O
(	O
Symons	O
,	O
1997	O
;	O
Jones	O
et	O
al	O
.,	O
1987	O
),	O
with	O
rapid	O
electron	O
quantum	O
tunnelling	O
and	O
positive	O
hole	O
migration	O
along	O
the	O
protein	O
backbone	O
and	O
through	O
stacked	O
DNA	O
bases	O
indicated	O
as	O
a	O
dominant	O
mechanism	O
by	O
which	O
oxidative	O
and	O
reductive	O
damage	O
localizes	O
at	O
distances	O
from	O
initial	O
ionization	O
sites	O
at	O
100	O
K	O
(	O
O	O
’	O
Neill	O
et	O
al	O
.,	O
2002	O
).	O
	O
Only	O
at	O
doses	O
above	O
20	O
MGy	O
were	O
precursors	O
of	O
phosphodiester	O
-	O
bond	O
cleavage	O
observed	O
within	O
AT	O
-	O
rich	O
regions	O
of	O
the	O
35	O
-	O
mer	O
DNA	O
.	O
	O
For	O
crystalline	O
complexes	O
such	O
as	O
C	O
.	O
Esp1396I	O
,	O
whether	O
the	O
protein	O
is	O
intrinsically	O
more	O
susceptible	O
to	O
X	O
-	O
ray	O
-	O
induced	O
damage	O
or	O
whether	O
the	O
protein	O
scavenges	O
electrons	O
to	O
protect	O
the	O
DNA	O
remains	O
unclear	O
in	O
the	O
absence	O
of	O
a	O
non	O
-	O
nucleic	O
acid	O
-	O
bound	O
protein	O
control	O
obtained	O
under	O
exactly	O
the	O
same	O
crystallization	O
and	O
data	O
-	O
collection	O
conditions	O
.	O
	O
To	O
monitor	O
the	O
effects	O
of	O
nucleic	O
acid	O
binding	O
on	O
protein	O
damage	O
susceptibility	O
,	O
a	O
crystal	O
containing	O
two	O
protein	O
molecules	O
per	O
asymmetric	O
unit	O
,	O
only	O
one	O
of	O
which	O
was	O
bound	O
to	O
RNA	O
,	O
is	O
reported	O
here	O
(	O
Fig	O
.	O
1	O
▸).	O
	O
TRAP	O
consists	O
of	O
11	O
identical	O
subunits	O
assembled	O
into	O
a	O
ring	O
with	O
11	O
-	O
fold	O
rotational	O
symmetry	O
.	O
	O
In	O
this	O
structure	O
,	O
the	O
bases	O
of	O
the	O
G1	O
-	O
A2	O
-	O
G3	O
nucleotides	O
form	O
direct	O
hydrogen	O
bonds	O
to	O
the	O
protein	O
,	O
unlike	O
the	O
U4	O
-	O
U5	O
nucleotides	O
,	O
which	O
appear	O
to	O
be	O
more	O
flexible	O
.	O
	O
Ten	O
successive	O
1	O
.	O
98	O
Å	O
resolution	O
MX	O
data	O
sets	O
were	O
collected	O
from	O
the	O
same	O
TRAP	O
–	O
RNA	O
crystal	O
to	O
analyse	O
X	O
-	O
ray	O
-	O
induced	O
structural	O
changes	O
over	O
a	O
large	O
dose	O
range	O
(	O
d	O
1	O
=	O
1	O
.	O
3	O
MGy	O
to	O
d	O
10	O
=	O
25	O
.	O
0	O
MGy	O
).	O
	O
To	O
avoid	O
the	O
previous	O
necessity	O
for	O
visual	O
inspection	O
of	O
electron	O
-	O
density	O
maps	O
to	O
detect	O
SRD	O
sites	O
,	O
a	O
computational	O
approach	O
was	O
designed	O
to	O
quantify	O
the	O
electron	O
-	O
density	O
change	O
for	O
each	O
refined	O
atom	O
with	O
increasing	O
dose	O
,	O
thus	O
providing	O
a	O
rapid	O
systematic	O
method	O
for	O
SRD	O
study	O
on	O
such	O
large	O
multimeric	O
complexes	O
.	O
	O
By	O
employing	O
the	O
high	O
11	O
-	O
fold	O
structural	O
symmetry	O
within	O
each	O
TRAP	O
macromolecule	O
,	O
this	O
approach	O
permitted	O
a	O
thorough	O
statistical	O
quantification	O
of	O
the	O
RD	O
effects	O
of	O
RNA	O
binding	O
to	O
TRAP	O
.	O
	O
Per	O
-	O
atom	O
quantification	O
of	O
electron	O
density	O
	O
To	O
quantify	O
the	O
exact	O
effects	O
of	O
nucleic	O
acid	O
binding	O
to	O
a	O
protein	O
on	O
SRD	O
susceptibility	O
,	O
a	O
high	O
-	O
throughput	O
and	O
automated	O
pipeline	O
was	O
created	O
to	O
systematically	O
calculate	O
the	O
electron	O
-	O
density	O
change	O
for	O
every	O
refined	O
atom	O
within	O
the	O
TRAP	O
–	O
RNA	O
structure	O
as	O
a	O
function	O
of	O
dose	O
.	O
	O
This	O
provides	O
an	O
atom	O
-	O
specific	O
quantification	O
of	O
density	O
–	O
dose	O
dynamics	O
,	O
which	O
was	O
previously	O
lacking	O
within	O
the	O
field	O
.	O
	O
However	O
,	O
these	O
σ	O
levels	O
depend	O
on	O
the	O
standard	O
deviation	O
values	O
of	O
the	O
map	O
,	O
which	O
can	O
deviate	O
between	O
data	O
sets	O
,	O
and	O
are	O
thus	O
unsuitable	O
for	O
quantitative	O
comparison	O
of	O
density	O
between	O
different	O
dose	O
data	O
sets	O
.	O
	O
Instead	O
,	O
we	O
use	O
here	O
a	O
maximum	O
density	O
-	O
loss	O
metric	O
(	O
D	O
loss	O
),	O
which	O
is	O
the	O
per	O
-	O
atom	O
equivalent	O
of	O
the	O
magnitude	O
of	O
these	O
negative	O
Fourier	O
difference	O
map	O
peaks	O
in	O
units	O
of	O
e	O
Å	O
−	O
3	O
.	O
	O
Large	O
positive	O
D	O
loss	O
values	O
indicate	O
radiation	O
-	O
induced	O
atomic	O
disordering	O
reproducibly	O
throughout	O
the	O
unit	O
cells	O
with	O
respect	O
to	O
the	O
initial	O
low	O
-	O
dose	O
data	O
set	O
.	O
	O
For	O
each	O
TRAP	O
–	O
RNA	O
data	O
set	O
,	O
the	O
D	O
loss	O
metric	O
successfully	O
identified	O
the	O
recognized	O
forms	O
of	O
protein	O
SRD	O
(	O
Fig	O
.	O
2	O
▸	O
a	O
),	O
with	O
clear	O
Glu	O
and	O
Asp	O
side	O
-	O
chain	O
decarboxylation	O
even	O
in	O
the	O
first	O
difference	O
map	O
calculated	O
(	O
3	O
.	O
9	O
MGy	O
;	O
Fig	O
.	O
3	O
▸	O
a	O
).	O
	O
The	O
main	O
sequence	O
of	O
TRAP	O
does	O
not	O
contain	O
any	O
Trp	O
and	O
Cys	O
residues	O
(	O
and	O
thus	O
contains	O
no	O
disulfide	O
bonds	O
).	O
	O
The	O
substrate	O
Trp	O
amino	O
-	O
acid	O
ligands	O
also	O
exhibited	O
disordering	O
of	O
the	O
free	O
terminal	O
carboxyl	O
groups	O
at	O
higher	O
doses	O
(	O
Fig	O
.	O
2	O
▸	O
a	O
);	O
however	O
,	O
no	O
clear	O
Fourier	O
difference	O
peaks	O
could	O
be	O
observed	O
visually	O
.	O
	O
Even	O
for	O
radiation	O
-	O
insensitive	O
residues	O
(	O
e	O
.	O
g	O
.	O
Gly	O
)	O
the	O
average	O
D	O
loss	O
increases	O
with	O
dose	O
:	O
this	O
is	O
the	O
effect	O
of	O
global	O
radiation	O
damage	O
,	O
since	O
as	O
dose	O
increases	O
the	O
electron	O
density	O
associated	O
with	O
each	O
refined	O
atom	O
becomes	O
weaker	O
as	O
the	O
atomic	O
occupancy	O
decreases	O
(	O
Fig	O
.	O
2	O
▸	O
b	O
).	O
	O
Only	O
Glu	O
and	O
Asp	O
residues	O
exhibit	O
a	O
rate	O
of	O
D	O
loss	O
increase	O
that	O
consistently	O
exceeds	O
the	O
average	O
decay	O
(	O
Fig	O
.	O
2	O
▸	O
b	O
,	O
dashed	O
line	O
)	O
at	O
each	O
dose	O
.	O
	O
The	O
rate	O
of	O
D	O
loss	O
(	O
attributed	O
to	O
side	O
-	O
chain	O
decarboxylation	O
)	O
was	O
consistently	O
larger	O
for	O
Glu	O
compared	O
with	O
Asp	O
residues	O
over	O
the	O
large	O
dose	O
range	O
(	O
Fig	O
.	O
2	O
▸	O
b	O
and	O
Supplementary	O
Fig	O
.	O
S3	O
);	O
this	O
observation	O
is	O
consistent	O
with	O
our	O
calculations	O
on	O
model	O
systems	O
(	O
see	O
above	O
)	O
that	O
suggest	O
that	O
,	O
without	O
considering	O
differential	O
hydrogen	O
-	O
bonding	O
environments	O
,	O
CO2	O
loss	O
is	O
more	O
exothermic	O
by	O
around	O
8	O
kJ	O
mol	O
−	O
1	O
from	O
oxidized	O
Glu	O
residues	O
than	O
from	O
their	O
Asp	O
counterparts	O
.	O
	O
RNA	O
is	O
less	O
susceptible	O
to	O
electron	O
-	O
density	O
loss	O
than	O
protein	O
within	O
the	O
TRAP	O
–	O
RNA	O
complex	O
	O
Only	O
at	O
the	O
highest	O
doses	O
investigated	O
(>	O
20	O
MGy	O
)	O
was	O
density	O
loss	O
observed	O
at	O
the	O
RNA	O
phosphate	O
and	O
C	O
—	O
O	O
bonds	O
of	O
the	O
phosphodiester	O
backbone	O
.	O
	O
However	O
,	O
the	O
median	O
D	O
loss	O
was	O
lower	O
by	O
a	O
factor	O
of	O
>	O
2	O
for	O
RNA	O
P	O
atoms	O
than	O
for	O
Glu	O
and	O
Asp	O
side	O
-	O
chain	O
groups	O
at	O
25	O
.	O
0	O
MGy	O
(	O
Supplementary	O
Fig	O
.	O
S4	O
),	O
and	O
furthermore	O
could	O
not	O
be	O
numerically	O
distinguished	O
from	O
Gly	O
Cα	O
atoms	O
within	O
TRAP	O
,	O
which	O
are	O
not	O
radiation	O
-	O
sensitive	O
at	O
the	O
doses	O
tested	O
here	O
(	O
Supplementary	O
Fig	O
.	O
S3	O
).	O
	O
RNA	O
binding	O
protects	O
radiation	O
-	O
sensitive	O
residues	O
	O
For	O
the	O
large	O
number	O
of	O
acidic	O
residues	O
per	O
TRAP	O
ring	O
(	O
four	O
Asp	O
and	O
six	O
Glu	O
residues	O
per	O
protein	O
monomer	O
),	O
a	O
strong	O
dependence	O
of	O
decarboxylation	O
susceptibility	O
on	O
local	O
environment	O
was	O
observed	O
(	O
Fig	O
.	O
4	O
▸).	O
	O
For	O
each	O
Glu	O
Cδ	O
or	O
Asp	O
Cγ	O
atom	O
,	O
D	O
loss	O
provided	O
a	O
direct	O
measure	O
of	O
the	O
rate	O
of	O
side	O
-	O
chain	O
carboxyl	O
-	O
group	O
disordering	O
and	O
subsequent	O
decarboxylation	O
.	O
	O
For	O
acidic	O
residues	O
with	O
no	O
differing	O
interactions	O
between	O
nonbound	O
and	O
bound	O
TRAP	O
(	O
Fig	O
.	O
4	O
▸	O
a	O
),	O
similar	O
damage	O
was	O
apparent	O
between	O
the	O
two	O
rings	O
within	O
the	O
asymmetric	O
unit	O
,	O
as	O
expected	O
.	O
	O
However	O
,	O
TRAP	O
residues	O
directly	O
on	O
the	O
RNA	O
-	O
binding	O
interfaces	O
exhibited	O
greater	O
damage	O
accumulation	O
in	O
nonbound	O
TRAP	O
(	O
Fig	O
.	O
4	O
▸	O
b	O
),	O
and	O
for	O
residues	O
at	O
the	O
ring	O
–	O
ring	O
interfaces	O
(	O
where	O
crystal	O
contacts	O
were	O
detected	O
)	O
bound	O
TRAP	O
exhibited	O
enhanced	O
SRD	O
accumulation	O
(	O
Fig	O
.	O
4	O
▸	O
c	O
).	O
	O
Hotelling	O
’	O
s	O
T	O
-	O
squared	O
test	O
(	O
the	O
multivariate	O
counterpart	O
of	O
Student	O
’	O
s	O
t	O
-	O
test	O
)	O
was	O
used	O
to	O
reject	O
the	O
null	O
hypothesis	O
that	O
the	O
means	O
of	O
the	O
D	O
loss	O
metric	O
were	O
equal	O
for	O
the	O
bound	O
and	O
nonbound	O
groups	O
in	O
Fig	O
.	O
5	O
▸.	O
	O
A	O
significant	O
reduction	O
in	O
D	O
loss	O
is	O
seen	O
for	O
Glu36	O
in	O
RNA	O
-	O
bound	O
compared	O
with	O
nonbound	O
TRAP	O
,	O
indicative	O
of	O
a	O
lower	O
rate	O
of	O
side	O
-	O
chain	O
decarboxylation	O
(	O
Fig	O
.	O
5	O
▸	O
a	O
;	O
p	O
=	O
6	O
.	O
06	O
×	O
10	O
−	O
5	O
).	O
	O
In	O
our	O
analysis	O
,	O
Asp39	O
in	O
the	O
TRAP	O
–(	O
GAGUU	O
)	O
10GAG	O
structure	O
appears	O
to	O
exhibit	O
two	O
distinct	O
hydrogen	O
bonds	O
to	O
the	O
G1	O
base	O
within	O
each	O
of	O
the	O
11	O
TRAP	O
–	O
RNA	O
interfaces	O
,	O
as	O
does	O
Glu36	O
to	O
G3	O
;	O
however	O
,	O
the	O
reduction	O
in	O
density	O
disordering	O
upon	O
RNA	O
binding	O
is	O
far	O
less	O
significant	O
for	O
Asp39	O
than	O
for	O
Glu36	O
(	O
Fig	O
.	O
5	O
▸	O
b	O
,	O
p	O
=	O
0	O
.	O
0925	O
).	O
	O
RNA	O
binding	O
reduces	O
radiation	O
-	O
induced	O
disorder	O
on	O
the	O
atomic	O
scale	O
	O
One	O
oxygen	O
(	O
O	O
∊	O
1	O
)	O
of	O
Glu42	O
appears	O
to	O
form	O
a	O
hydrogen	O
bond	O
to	O
a	O
nearby	O
water	O
within	O
each	O
TRAP	O
RNA	O
-	O
binding	O
pocket	O
,	O
with	O
the	O
other	O
(	O
O	O
∊	O
2	O
)	O
being	O
involved	O
in	O
a	O
salt	O
-	O
bridge	O
interaction	O
with	O
Arg58	O
(	O
Hopcroft	O
et	O
al	O
.,	O
2002	O
;	O
Antson	O
et	O
al	O
.,	O
1999	O
).	O
	O
A	O
significant	O
difference	O
was	O
observed	O
between	O
the	O
D	O
loss	O
dynamics	O
for	O
the	O
nonbound	O
/	O
bound	O
Glu42	O
O	O
∊	O
1	O
atoms	O
(	O
Fig	O
.	O
5	O
▸	O
c	O
;	O
p	O
=	O
0	O
.	O
007	O
)	O
but	O
not	O
for	O
the	O
Glu42	O
O	O
∊	O
2	O
atoms	O
(	O
Fig	O
.	O
5	O
▸	O
d	O
;	O
p	O
=	O
0	O
.	O
239	O
),	O
indicating	O
that	O
the	O
stabilizing	O
strength	O
of	O
this	O
salt	O
-	O
bridge	O
interaction	O
was	O
conserved	O
upon	O
RNA	O
binding	O
and	O
that	O
the	O
water	O
-	O
mediated	O
hydrogen	O
bond	O
had	O
a	O
greater	O
relative	O
susceptibility	O
to	O
atomic	O
disordering	O
in	O
the	O
absence	O
of	O
RNA	O
.	O
	O
The	O
density	O
-	O
change	O
dynamics	O
were	O
statistically	O
indistinguishable	O
between	O
bound	O
and	O
nonbound	O
TRAP	O
for	O
each	O
Glu42	O
carboxyl	O
group	O
Cδ	O
atom	O
(	O
p	O
=	O
0	O
.	O
435	O
),	O
indicating	O
that	O
upon	O
RNA	O
binding	O
the	O
conserved	O
salt	O
-	O
bridge	O
interaction	O
ultimately	O
dictated	O
the	O
overall	O
Glu42	O
decarboxylation	O
rate	O
.	O
	O
The	O
RNA	O
-	O
stabilizing	O
effect	O
was	O
not	O
restricted	O
to	O
radiation	O
-	O
sensitive	O
acidic	O
residues	O
.	O
	O
The	O
side	O
chain	O
of	O
Phe32	O
stacks	O
against	O
the	O
G3	O
base	O
within	O
the	O
11	O
TRAP	O
RNA	O
-	O
binding	O
interfaces	O
(	O
Antson	O
et	O
al	O
.,	O
1999	O
).	O
	O
The	O
D	O
loss	O
for	O
Lys37	O
side	O
-	O
chain	O
atoms	O
was	O
also	O
reduced	O
when	O
stacked	O
against	O
the	O
G1	O
base	O
(	O
Fig	O
.	O
5	O
▸	O
f	O
;	O
p	O
=	O
0	O
.	O
0243	O
for	O
Lys37	O
C	O
∊	O
atoms	O
).	O
	O
Here	O
,	O
MX	O
radiation	O
-	O
induced	O
specific	O
structural	O
changes	O
within	O
the	O
large	O
TRAP	O
–	O
RNA	O
assembly	O
over	O
a	O
large	O
dose	O
range	O
(	O
1	O
.	O
3	O
–	O
25	O
.	O
0	O
MGy	O
)	O
have	O
been	O
analysed	O
using	O
a	O
high	O
-	O
throughput	O
quantitative	O
approach	O
,	O
providing	O
a	O
measure	O
of	O
the	O
electron	O
-	O
density	O
distribution	O
for	O
each	O
refined	O
atom	O
with	O
increasing	O
dose	O
,	O
D	O
loss	O
.	O
	O
Compared	O
with	O
previous	O
studies	O
,	O
the	O
results	O
provide	O
a	O
further	O
step	O
in	O
the	O
detailed	O
characterization	O
of	O
SRD	O
effects	O
in	O
MX	O
.	O
	O
Here	O
,	O
it	O
provided	O
the	O
precision	O
required	O
to	O
quantify	O
the	O
role	O
of	O
RNA	O
in	O
the	O
damage	O
susceptibilities	O
of	O
equivalent	O
atoms	O
between	O
RNA	O
-	O
bound	O
and	O
nonbound	O
TRAP	O
,	O
but	O
it	O
is	O
applicable	O
to	O
any	O
MX	O
SRD	O
study	O
.	O
	O
RNA	O
backbone	O
disordering	O
thus	O
appears	O
to	O
be	O
the	O
main	O
radiation	O
-	O
induced	O
effect	O
in	O
RNA	O
,	O
with	O
the	O
protein	O
–	O
base	O
interactions	O
maintained	O
even	O
at	O
high	O
doses	O
(>	O
20	O
MGy	O
).	O
	O
The	O
U4	O
phosphate	O
exhibited	O
marginally	O
larger	O
D	O
loss	O
values	O
above	O
20	O
MGy	O
than	O
G1	O
,	O
A2	O
and	O
G3	O
(	O
Supplementary	O
Fig	O
.	O
S4	O
).	O
	O
Since	O
U4	O
is	O
the	O
only	O
refined	O
nucleotide	O
not	O
to	O
exhibit	O
significant	O
base	O
–	O
protein	O
interactions	O
around	O
TRAP	O
(	O
with	O
a	O
water	O
-	O
mediated	O
hydrogen	O
bond	O
detected	O
in	O
only	O
three	O
of	O
the	O
11	O
subunits	O
and	O
a	O
single	O
Arg58	O
hydrogen	O
bond	O
suggested	O
in	O
a	O
further	O
four	O
subunits	O
),	O
this	O
increased	O
U4	O
D	O
loss	O
can	O
be	O
explained	O
owing	O
to	O
its	O
greater	O
flexibility	O
.	O
	O
At	O
25	O
.	O
0	O
MGy	O
,	O
the	O
magnitude	O
of	O
the	O
RNA	O
backbone	O
D	O
loss	O
was	O
of	O
the	O
same	O
order	O
as	O
for	O
the	O
radiation	O
-	O
insensitive	O
Gly	O
Cα	O
atoms	O
and	O
on	O
average	O
less	O
than	O
half	O
that	O
of	O
the	O
acidic	O
residues	O
of	O
the	O
protein	O
(	O
Supplementary	O
Fig	O
.	O
S3	O
).	O
	O
Consequently	O
,	O
no	O
clear	O
single	O
-	O
strand	O
breaks	O
could	O
be	O
located	O
,	O
and	O
since	O
RNA	O
-	O
binding	O
within	O
the	O
current	O
TRAP	O
–(	O
GAGUU	O
)	O
10GAG	O
complex	O
is	O
mediated	O
predominantly	O
through	O
base	O
–	O
protein	O
interactions	O
,	O
the	O
biological	O
integrity	O
of	O
the	O
RNA	O
complex	O
was	O
dictated	O
by	O
the	O
rate	O
at	O
which	O
protein	O
decarboxylation	O
occurred	O
.	O
	O
RNA	O
interacting	O
with	O
TRAP	O
was	O
shown	O
to	O
offer	O
significant	O
protection	O
against	O
radiation	O
-	O
induced	O
structural	O
changes	O
.	O
	O
However	O
,	O
compared	O
with	O
Asp39	O
,	O
Glu36	O
is	O
strikingly	O
less	O
decarboxylated	O
when	O
bound	O
to	O
RNA	O
(	O
Fig	O
.	O
4	O
▸).	O
	O
For	O
Glu36	O
and	O
Asp39	O
,	O
no	O
direct	O
quantitative	O
correlation	O
could	O
be	O
established	O
between	O
hydrogen	O
-	O
bond	O
length	O
and	O
D	O
loss	O
(	O
linear	O
R	O
2	O
of	O
<	O
0	O
.	O
23	O
for	O
all	O
doses	O
;	O
Supplementary	O
Fig	O
.	O
S5	O
).	O
	O
Thus	O
,	O
another	O
factor	O
must	O
be	O
responsible	O
for	O
this	O
clear	O
reduction	O
in	O
Glu36	O
CO2	O
decarboxyl	O
­	O
ation	O
in	O
RNA	O
-	O
bound	O
TRAP	O
.	O
	O
The	O
Glu36	O
carboxyl	O
side	O
chain	O
also	O
potentially	O
forms	O
hydrogen	O
bonds	O
to	O
His34	O
and	O
Lys56	O
,	O
but	O
since	O
these	O
interactions	O
are	O
conserved	O
irrespective	O
of	O
G3	O
nucleotide	O
binding	O
,	O
this	O
cannot	O
directly	O
account	O
for	O
the	O
stabilization	O
effect	O
on	O
Glu36	O
in	O
RNA	O
-	O
bound	O
TRAP	O
.	O
	O
When	O
bound	O
to	O
RNA	O
,	O
the	O
average	O
solvent	O
-	O
accessible	O
area	O
of	O
the	O
Glu36	O
side	O
-	O
chain	O
O	O
atoms	O
is	O
reduced	O
from	O
∼	O
15	O
to	O
0	O
Å2	O
.	O
	O
We	O
propose	O
that	O
with	O
no	O
solvent	O
accessibility	O
Glu36	O
decarboxylation	O
is	O
inhibited	O
,	O
since	O
the	O
CO2	O
-	O
formation	O
rate	O
K	O
2	O
is	O
greatly	O
reduced	O
,	O
and	O
suggest	O
that	O
steric	O
hindrance	O
prevents	O
each	O
radicalized	O
Glu36	O
CO2	O
group	O
from	O
achieving	O
the	O
planar	O
conformation	O
required	O
for	O
complete	O
dissociation	O
from	O
TRAP	O
.	O
	O
The	O
electron	O
-	O
recombination	O
rate	O
K	O
−	O
1	O
remains	O
high	O
,	O
however	O
,	O
owing	O
to	O
rapid	O
electron	O
migration	O
through	O
the	O
protein	O
–	O
RNA	O
complex	O
to	O
refill	O
the	O
Glu36	O
positive	O
hole	O
(	O
the	O
precursor	O
for	O
Glu	O
decarboxylation	O
).	O
	O
Upon	O
RNA	O
binding	O
,	O
the	O
Asp39	O
side	O
-	O
chain	O
carboxyl	O
group	O
solvent	O
-	O
accessible	O
area	O
changes	O
from	O
∼	O
75	O
to	O
35	O
Å2	O
,	O
still	O
allowing	O
a	O
high	O
CO2	O
-	O
formation	O
rate	O
K	O
2	O
.	O
	O
By	O
comparing	O
equivalent	O
acidic	O
residues	O
with	O
and	O
without	O
RNA	O
,	O
we	O
have	O
now	O
deconvoluted	O
the	O
role	O
of	O
solvent	O
accessibility	O
from	O
other	O
local	O
protein	O
environment	O
factors	O
,	O
and	O
thus	O
propose	O
a	O
suitable	O
mechanism	O
by	O
which	O
exceptionally	O
low	O
solvent	O
accessibility	O
can	O
reduce	O
the	O
rate	O
of	O
decarboxylation	O
.	O
	O
Apart	O
from	O
these	O
RNA	O
-	O
binding	O
interfaces	O
,	O
RNA	O
binding	O
was	O
seen	O
to	O
enhance	O
decarboxylation	O
for	O
residues	O
Glu50	O
,	O
Glu71	O
and	O
Glu73	O
,	O
all	O
of	O
which	O
are	O
involved	O
in	O
crystal	O
contacts	O
between	O
TRAP	O
rings	O
(	O
Fig	O
.	O
4	O
▸	O
c	O
).	O
	O
However	O
,	O
for	O
each	O
of	O
these	O
residues	O
the	O
exact	O
crystal	O
contacts	O
are	O
not	O
preserved	O
between	O
bound	O
and	O
nonbound	O
TRAP	O
or	O
even	O
between	O
monomers	O
within	O
one	O
TRAP	O
ring	O
.	O
	O
For	O
example	O
,	O
in	O
bound	O
TRAP	O
,	O
Glu73	O
hydrogen	O
-	O
bonds	O
to	O
a	O
nearby	O
lysine	O
on	O
each	O
of	O
the	O
11	O
subunits	O
,	O
whereas	O
in	O
nonbound	O
TRAP	O
no	O
such	O
interaction	O
exists	O
and	O
Glu73	O
interacts	O
with	O
a	O
variable	O
number	O
of	O
refined	O
waters	O
in	O
each	O
subunit	O
.	O
	O
Radiation	O
-	O
induced	O
side	O
-	O
chain	O
conformational	O
changes	O
have	O
been	O
poorly	O
characterized	O
in	O
previous	O
SRD	O
investigations	O
owing	O
to	O
their	O
strong	O
dependence	O
on	O
packing	O
density	O
and	O
geometric	O
strain	O
.	O
	O
Such	O
structural	O
changes	O
are	O
known	O
to	O
have	O
significant	O
roles	O
within	O
enzymatic	O
pathways	O
,	O
and	O
experimenters	O
must	O
be	O
aware	O
of	O
these	O
possible	O
confounding	O
factors	O
when	O
assigning	O
true	O
functional	O
mechanisms	O
using	O
MX	O
.	O
	O
Our	O
results	O
show	O
that	O
RNA	O
binding	O
to	O
TRAP	O
physically	O
stabilizes	O
non	O
-	O
acidic	O
residues	O
within	O
the	O
TRAP	O
macromolecule	O
,	O
most	O
notably	O
Lys37	O
and	O
Phe32	O
,	O
which	O
stack	O
against	O
the	O
G1	O
and	O
G3	O
bases	O
,	O
respectively	O
.	O
	O
It	O
has	O
been	O
suggested	O
(	O
Burmeister	O
,	O
2000	O
)	O
that	O
Tyr	O
residues	O
can	O
lose	O
their	O
aromatic	O
–	O
OH	O
group	O
owing	O
to	O
radiation	O
-	O
induced	O
effects	O
;	O
however	O
,	O
no	O
energetically	O
favourable	O
pathway	O
for	O
–	O
OH	O
cleavage	O
exists	O
and	O
this	O
has	O
not	O
been	O
detected	O
in	O
aqueous	O
radiation	O
-	O
chemistry	O
studies	O
.	O
	O
In	O
TRAP	O
,	O
D	O
loss	O
increased	O
at	O
a	O
similar	O
rate	O
for	O
both	O
the	O
Tyr	O
O	O
atoms	O
and	O
aromatic	O
ring	O
atoms	O
,	O
suggesting	O
that	O
full	O
ring	O
conformational	O
disordering	O
is	O
more	O
likely	O
.	O
	O
Indeed	O
,	O
no	O
convincing	O
reproducible	O
Fourier	O
difference	O
peaks	O
above	O
the	O
background	O
map	O
noise	O
were	O
observed	O
around	O
any	O
Tyr	O
terminal	O
–	O
OH	O
groups	O
.	O
	O
The	O
RNA	O
-	O
stabilization	O
effects	O
on	O
protein	O
are	O
observed	O
at	O
short	O
ranges	O
and	O
are	O
restricted	O
to	O
within	O
the	O
RNA	O
-	O
binding	O
interfaces	O
around	O
the	O
TRAP	O
ring	O
.	O
	O
An	O
increase	O
in	O
the	O
dose	O
at	O
which	O
functionally	O
important	O
residues	O
remain	O
intact	O
has	O
biological	O
ramifications	O
for	O
understanding	O
the	O
mechanisms	O
at	O
which	O
ionizing	O
radiation	O
damage	O
is	O
mitigated	O
within	O
naturally	O
forming	O
DNA	O
–	O
protein	O
and	O
RNA	O
–	O
protein	O
complexes	O
.	O
	O
In	O
these	O
studies	O
,	O
the	O
main	O
damaging	O
species	O
is	O
predicted	O
to	O
be	O
the	O
oxidizing	O
hydroxyl	O
radical	O
produced	O
through	O
solvent	O
irradiation	O
,	O
which	O
is	O
known	O
to	O
add	O
to	O
double	O
covalent	O
bonds	O
within	O
both	O
DNA	O
and	O
RNA	O
bases	O
to	O
induce	O
strand	O
breaks	O
and	O
base	O
modification	O
(	O
Spotheim	O
-	O
Maurizot	O
&	O
Davídková	O
,	O
2011	O
;	O
Chance	O
et	O
al	O
.,	O
1997	O
).	O
	O
It	O
was	O
suggested	O
that	O
physical	O
screening	O
of	O
DNA	O
by	O
protein	O
shielded	O
the	O
DNA	O
–	O
protein	O
interaction	O
sites	O
from	O
radical	O
damage	O
,	O
yielding	O
an	O
extended	O
life	O
-	O
dose	O
for	O
the	O
nucleoprotein	O
complex	O
compared	O
with	O
separate	O
protein	O
and	O
DNA	O
constituents	O
at	O
RT	O
.	O
	O
The	O
results	O
presented	O
here	O
suggest	O
that	O
biologically	O
relevant	O
nucleoprotein	O
complexes	O
also	O
exhibit	O
prolonged	O
life	O
-	O
doses	O
under	O
the	O
effect	O
of	O
LEE	O
-	O
induced	O
structural	O
changes	O
,	O
involving	O
direct	O
physical	O
protection	O
of	O
key	O
RNA	O
-	O
binding	O
residues	O
.	O
	O
Such	O
reduced	O
radiation	O
-	O
sensitivity	O
in	O
this	O
case	O
ensures	O
that	O
the	O
interacting	O
protein	O
remains	O
bound	O
long	O
enough	O
to	O
the	O
RNA	O
to	O
complete	O
its	O
function	O
,	O
even	O
whilst	O
exposed	O
to	O
ionizing	O
radiation	O
.	O
	O
Within	O
the	O
nonbound	O
TRAP	O
macromolecule	O
,	O
the	O
acidic	O
residues	O
within	O
the	O
unoccupied	O
RNA	O
-	O
binding	O
interfaces	O
(	O
Asp39	O
,	O
Glu36	O
,	O
Glu42	O
)	O
are	O
notably	O
amongst	O
the	O
most	O
susceptible	O
residues	O
within	O
the	O
asymmetric	O
unit	O
(	O
Fig	O
.	O
4	O
▸).	O
	O
When	O
exposed	O
to	O
X	O
-	O
rays	O
,	O
these	O
residues	O
will	O
be	O
preferentially	O
damaged	O
by	O
X	O
-	O
rays	O
and	O
subsequently	O
reduce	O
the	O
affinity	O
with	O
which	O
TRAP	O
binds	O
to	O
RNA	O
.	O
	O
Within	O
the	O
cellular	O
environment	O
,	O
this	O
mechanism	O
could	O
reduce	O
the	O
risk	O
that	O
radiation	O
-	O
damaged	O
proteins	O
might	O
bind	O
to	O
RNA	O
,	O
thus	O
avoiding	O
the	O
detrimental	O
introduction	O
of	O
incorrect	O
DNA	O
-	O
repair	O
,	O
transcriptional	O
and	O
base	O
-	O
modification	O
pathways	O
.	O
	O
The	O
TRAP	O
–(	O
GAGUU	O
)	O
10GAG	O
complex	O
asymmetric	O
unit	O
(	O
PDB	O
entry	O
1gtf	O
;	O
Hopcroft	O
et	O
al	O
.,	O
2002	O
).	O
	O
Bound	O
tryptophan	O
ligands	O
are	O
represented	O
as	O
coloured	O
spheres	O
.	O
	O
(	O
a	O
)	O
Electron	O
-	O
density	O
loss	O
sites	O
as	O
indicated	O
by	O
D	O
loss	O
in	O
the	O
TRAP	O
–	O
RNA	O
complex	O
crystal	O
by	O
residue	O
/	O
nucleotide	O
type	O
for	O
five	O
doses	O
[	O
sites	O
determined	O
above	O
the	O
4	O
×	O
average	O
D	O
loss	O
threshold	O
,	O
calculated	O
over	O
the	O
TRAP	O
–	O
RNA	O
structure	O
for	O
the	O
first	O
difference	O
map	O
:	O
F	O
obs	O
(	O
d	O
2	O
)	O
−	O
F	O
obs	O
(	O
d	O
1	O
)].	O
	O
Only	O
a	O
subset	O
of	O
key	O
TRAP	O
residue	O
types	O
are	O
included	O
.	O
	O
The	O
average	O
D	O
loss	O
(	O
calculated	O
over	O
the	O
whole	O
TRAP	O
asymmetric	O
unit	O
)	O
is	O
shown	O
at	O
each	O
dose	O
(	O
dashed	O
line	O
).	O
	O
In	O
(	O
a	O
)	O
clear	O
difference	O
density	O
is	O
observed	O
around	O
the	O
Glu42	O
carboxyl	O
side	O
chain	O
in	O
chain	O
H	O
,	O
within	O
the	O
lowest	O
dose	O
difference	O
map	O
at	O
d	O
2	O
=	O
3	O
.	O
9	O
MGy	O
.	O
	O
Radiation	O
-	O
induced	O
protein	O
disordering	O
is	O
evident	O
across	O
the	O
large	O
dose	O
range	O
(	O
b	O
,	O
c	O
);	O
in	O
comparison	O
,	O
no	O
clear	O
deterioration	O
of	O
the	O
RNA	O
density	O
was	O
observed	O
.	O
	O
D	O
loss	O
calculated	O
for	O
all	O
side	O
-	O
chain	O
carboxyl	O
group	O
Glu	O
Cδ	O
and	O
Asp	O
Cγ	O
atoms	O
within	O
the	O
TRAP	O
–	O
RNA	O
complex	O
for	O
a	O
dose	O
of	O
19	O
.	O
3	O
MGy	O
(	O
d	O
8	O
).	O
	O
D	O
loss	O
against	O
dose	O
for	O
(	O
a	O
)	O
Glu36	O
Cδ	O
,	O
(	O
b	O
)	O
Asp39	O
Cγ	O
,	O
(	O
c	O
)	O
Glu42	O
O	O
∊	O
1	O
,	O
(	O
d	O
)	O
Glu42	O
O	O
∊	O
2	O
,	O
(	O
e	O
)	O
Phe32	O
Cζ	O
and	O
(	O
f	O
)	O
Lys37	O
C	O
∊	O
atoms	O
.	O
	O
95	O
%	O
CI	O
are	O
included	O
for	O
each	O
set	O
of	O
11	O
equivalent	O
atoms	O
grouped	O
as	O
bound	O
/	O
nonbound	O
.	O
	O
RNA	O
-	O
binding	O
interface	O
interactions	O
are	O
shown	O
for	O
TRAP	O
chain	O
N	O
,	O
with	O
the	O
F	O
obs	O
(	O
d	O
7	O
)	O
−	O
F	O
obs	O
(	O
d	O
1	O
)	O
Fourier	O
difference	O
map	O
(	O
dose	O
16	O
.	O
7	O
MGy	O
)	O
overlaid	O
and	O
contoured	O
at	O
a	O
±	O
4σ	O
level	O
.	O
	O
A	O
conserved	O
motif	O
in	O
JNK	O
/	O
p38	O
-	O
specific	O
MAPK	O
phosphatases	O
as	O
a	O
determinant	O
for	O
JNK1	O
recognition	O
and	O
inactivation	O
	O
Mitogen	O
-	O
activated	O
protein	O
kinases	O
(	O
MAPKs	O
),	O
important	O
in	O
a	O
large	O
array	O
of	O
signalling	O
pathways	O
,	O
are	O
tightly	O
controlled	O
by	O
a	O
cascade	O
of	O
protein	O
kinases	O
and	O
by	O
MAPK	O
phosphatases	O
(	O
MKPs	O
).	O
	O
MAPK	O
signalling	O
efficiency	O
and	O
specificity	O
is	O
modulated	O
by	O
protein	O
–	O
protein	O
interactions	O
between	O
individual	O
MAPKs	O
and	O
the	O
docking	O
motifs	O
in	O
cognate	O
binding	O
partners	O
.	O
	O
Two	O
types	O
of	O
docking	O
interactions	O
have	O
been	O
identified	O
:	O
D	O
-	O
motif	O
-	O
mediated	O
interaction	O
and	O
FXF	O
-	O
docking	O
interaction	O
.	O
	O
Here	O
we	O
report	O
the	O
crystal	O
structure	O
of	O
JNK1	O
bound	O
to	O
the	O
catalytic	O
domain	O
of	O
MKP7	O
at	O
2	O
.	O
4	O
-	O
Å	O
resolution	O
,	O
providing	O
high	O
-	O
resolution	O
structural	O
insight	O
into	O
the	O
FXF	O
-	O
docking	O
interaction	O
.	O
	O
The	O
285FNFL288	O
segment	O
in	O
MKP7	O
directly	O
binds	O
to	O
a	O
hydrophobic	O
site	O
on	O
JNK1	O
that	O
is	O
near	O
the	O
MAPK	O
insertion	O
and	O
helix	O
αG	O
.	O
Biochemical	O
studies	O
further	O
reveal	O
that	O
this	O
highly	O
conserved	O
structural	O
motif	O
is	O
present	O
in	O
all	O
members	O
of	O
the	O
MKP	O
family	O
,	O
and	O
the	O
interaction	O
mode	O
is	O
universal	O
and	O
critical	O
for	O
the	O
MKP	O
-	O
MAPK	O
recognition	O
and	O
biological	O
function	O
.	O
	O
The	O
important	O
MAPK	O
family	O
of	O
signalling	O
proteins	O
is	O
controlled	O
by	O
MAPK	O
phosphatases	O
(	O
MKPs	O
).	O
	O
Here	O
,	O
the	O
authors	O
report	O
the	O
structure	O
of	O
MKP7	O
bound	O
to	O
JNK1	O
and	O
characterise	O
the	O
conserved	O
MKP	O
-	O
MAPK	O
interaction	O
.	O
	O
The	O
mitogen	O
-	O
activated	O
protein	O
kinases	O
(	O
MAPKs	O
)	O
are	O
central	O
components	O
of	O
the	O
signal	O
-	O
transduction	O
pathways	O
,	O
which	O
mediate	O
the	O
cellular	O
response	O
to	O
a	O
variety	O
of	O
extracellular	O
stimuli	O
,	O
ranging	O
from	O
growth	O
factors	O
to	O
environmental	O
stresses	O
.	O
	O
The	O
MAPK	O
signalling	O
pathways	O
are	O
evolutionally	O
highly	O
conserved	O
.	O
	O
The	O
basic	O
assembly	O
of	O
MAPK	O
pathways	O
is	O
a	O
three	O
-	O
tier	O
kinase	O
module	O
that	O
establishes	O
a	O
sequential	O
activation	O
cascade	O
:	O
a	O
MAPK	O
kinase	O
kinase	O
activates	O
a	O
MAPK	O
kinase	O
,	O
which	O
in	O
turn	O
activates	O
a	O
MAPK	O
.	O
	O
The	O
MAPKs	O
are	O
activated	O
by	O
MAPK	O
kinases	O
that	O
phosphorylate	O
the	O
MAPKs	O
at	O
conserved	O
threonine	O
and	O
tyrosine	O
residues	O
within	O
their	O
activation	O
loop	O
.	O
	O
After	O
activation	O
,	O
each	O
MAPK	O
phosphorylates	O
a	O
distinct	O
set	O
of	O
protein	O
substrates	O
,	O
which	O
act	O
as	O
the	O
critical	O
effectors	O
that	O
enable	O
cells	O
to	O
mount	O
the	O
appropriate	O
responses	O
to	O
varied	O
stimuli	O
.	O
	O
MAPKs	O
lie	O
at	O
the	O
bottom	O
of	O
conserved	O
three	O
-	O
component	O
phosphorylation	O
cascades	O
and	O
utilize	O
docking	O
interactions	O
to	O
link	O
module	O
components	O
and	O
bind	O
substrates	O
.	O
	O
Two	O
types	O
of	O
docking	O
motifs	O
have	O
been	O
identified	O
in	O
MAPK	O
substrates	O
and	O
cognate	O
proteins	O
:	O
kinase	O
-	O
interacting	O
motif	O
(	O
D	O
-	O
motif	O
)	O
and	O
FXF	O
-	O
motif	O
(	O
also	O
called	O
DEF	O
motif	O
,	O
docking	O
site	O
for	O
ERK	O
FXF	O
).	O
	O
The	O
best	O
-	O
studied	O
docking	O
interactions	O
are	O
those	O
between	O
MAP	O
kinases	O
and	O
‘	O
D	O
-	O
motifs	O
',	O
which	O
consists	O
of	O
two	O
or	O
more	O
basic	O
residues	O
followed	O
by	O
a	O
short	O
linker	O
and	O
a	O
cluster	O
of	O
hydrophobic	O
residues	O
.	O
	O
The	O
D	O
-	O
motif	O
-	O
docking	O
site	O
(	O
D	O
-	O
site	O
)	O
in	O
MAPKs	O
is	O
situated	O
in	O
a	O
noncatalytic	O
region	O
opposite	O
of	O
the	O
kinase	O
catalytic	O
pocket	O
and	O
is	O
comprised	O
of	O
a	O
highly	O
acidic	O
patch	O
and	O
a	O
hydrophobic	O
groove	O
.	O
	O
D	O
-	O
motifs	O
are	O
found	O
in	O
many	O
MAPK	O
-	O
interacting	O
proteins	O
,	O
including	O
substrates	O
,	O
activating	O
kinases	O
and	O
inactivating	O
phosphatases	O
,	O
as	O
well	O
as	O
scaffolding	O
proteins	O
.	O
	O
A	O
second	O
docking	O
motif	O
for	O
MAPKs	O
consists	O
of	O
two	O
Phe	O
residues	O
separated	O
by	O
one	O
residue	O
(	O
FXF	O
-	O
motif	O
).	O
	O
This	O
motif	O
has	O
been	O
observed	O
in	O
several	O
MAPK	O
substrates	O
.	O
	O
The	O
FXF	O
-	O
motif	O
-	O
binding	O
site	O
of	O
ERK2	O
has	O
been	O
mapped	O
to	O
a	O
hydrophobic	O
pocket	O
formed	O
between	O
the	O
P	O
+	O
1	O
site	O
,	O
αG	O
helix	O
and	O
the	O
MAPK	O
insert	O
.	O
	O
However	O
,	O
the	O
generality	O
and	O
mechanism	O
of	O
the	O
FXF	O
-	O
mediated	O
interaction	O
is	O
unclear	O
.	O
	O
The	O
physiological	O
outcome	O
of	O
MAPK	O
signalling	O
depends	O
on	O
both	O
the	O
magnitude	O
and	O
the	O
duration	O
of	O
kinase	O
activation	O
.	O
	O
Downregulation	O
of	O
MAPK	O
activity	O
can	O
be	O
achieved	O
through	O
direct	O
dephosphorylation	O
of	O
the	O
phospho	O
-	O
threonine	O
and	O
/	O
or	O
tyrosine	O
residues	O
by	O
various	O
serine	O
/	O
threonine	O
phosphatases	O
,	O
tyrosine	O
phosphatases	O
and	O
dual	O
-	O
specificity	O
phosphatases	O
(	O
DUSPs	O
)	O
termed	O
MKPs	O
.	O
	O
Dysregulated	O
expression	O
of	O
MKPs	O
has	O
been	O
associated	O
with	O
pathogenesis	O
of	O
various	O
diseases	O
,	O
and	O
understanding	O
their	O
precise	O
recognition	O
mechanism	O
presents	O
an	O
important	O
challenge	O
and	O
opportunity	O
for	O
drug	O
development	O
.	O
	O
Here	O
,	O
we	O
present	O
the	O
crystal	O
structure	O
of	O
JNK1	O
in	O
complex	O
with	O
the	O
catalytic	O
domain	O
of	O
MKP7	O
.	O
	O
In	O
the	O
JNK1	O
–	O
MKP7	O
complex	O
,	O
a	O
hydrophobic	O
motif	O
(	O
285FNFL288	O
)	O
that	O
initiates	O
the	O
helix	O
α5	O
in	O
the	O
MKP7	O
catalytic	O
domain	O
directly	O
binds	O
to	O
the	O
FXF	O
-	O
motif	O
-	O
binding	O
site	O
on	O
JNK1	O
,	O
providing	O
the	O
structural	O
insight	O
into	O
the	O
classic	O
FXF	O
-	O
type	O
docking	O
interaction	O
.	O
	O
Thus	O
,	O
our	O
study	O
reveals	O
a	O
hitherto	O
unrecognized	O
interaction	O
mode	O
for	O
encoding	O
complex	O
target	O
specificity	O
among	O
MAPK	O
isoforms	O
.	O
	O
Interaction	O
of	O
JNK1	O
with	O
the	O
MKP7	O
catalytic	O
domain	O
	O
DUSPs	O
belong	O
to	O
the	O
protein	O
-	O
tyrosine	O
phosphatases	O
(	O
PTPase	O
)	O
superfamily	O
,	O
which	O
is	O
defined	O
by	O
the	O
PTPase	O
-	O
signature	O
motif	O
CXXGXXR	O
.	O
	O
MKPs	O
represent	O
a	O
distinct	O
subfamily	O
within	O
a	O
larger	O
group	O
of	O
DUSPs	O
.	O
	O
All	O
MKPs	O
contain	O
a	O
highly	O
conserved	O
C	O
-	O
terminal	O
catalytic	O
domain	O
(	O
CD	O
)	O
and	O
an	O
N	O
-	O
terminal	O
kinase	O
-	O
binding	O
domain	O
(	O
KBD	O
).	O
	O
The	O
KBD	O
is	O
homologous	O
to	O
the	O
rhodanese	O
family	O
and	O
contains	O
an	O
intervening	O
cluster	O
of	O
basic	O
amino	O
acids	O
,	O
which	O
has	O
been	O
suggested	O
to	O
be	O
important	O
for	O
interacting	O
with	O
the	O
target	O
MAPKs	O
.	O
	O
Biochemical	O
and	O
structural	O
studies	O
have	O
revealed	O
that	O
the	O
KBD	O
of	O
MKPs	O
is	O
critical	O
for	O
MKP3	O
docking	O
to	O
ERK2	O
,	O
and	O
MKP5	O
binding	O
to	O
p38α	O
,	O
although	O
their	O
binding	O
mechanisms	O
are	O
completely	O
different	O
.	O
	O
However	O
,	O
it	O
is	O
unknown	O
if	O
other	O
MAPKs	O
can	O
interact	O
with	O
the	O
KBD	O
of	O
their	O
cognate	O
phosphatases	O
in	O
the	O
same	O
manner	O
as	O
observed	O
for	O
recognition	O
of	O
ERK2	O
and	O
p38α	O
by	O
their	O
MKPs	O
,	O
or	O
whether	O
they	O
recognize	O
distinct	O
docking	O
motifs	O
of	O
MKPs	O
.	O
	O
MKP7	O
,	O
the	O
biggest	O
molecule	O
in	O
the	O
MKP	O
family	O
,	O
selectively	O
inactivates	O
JNK	O
and	O
p38	O
following	O
stress	O
activation	O
.	O
	O
To	O
quantitatively	O
assess	O
the	O
contribution	O
of	O
the	O
N	O
-	O
terminal	O
domain	O
to	O
the	O
MKP7	O
-	O
catalysed	O
JNK1	O
dephosphorylation	O
,	O
we	O
first	O
measured	O
the	O
kinetic	O
parameters	O
of	O
the	O
C	O
-	O
terminal	O
truncation	O
of	O
MKP7	O
(	O
MKP7ΔC304	B-mutant
,	O
residues	O
5	O
–	O
303	O
)	O
and	O
MKP7	O
-	O
CD	O
(	O
residues	O
156	O
–	O
301	O
)	O
towards	O
phosphorylated	O
JNK1	O
(	O
pJNK1	O
).	O
	O
The	O
kcat	O
and	O
Km	O
of	O
the	O
MKP7	O
-	O
CD	O
(	O
0	O
.	O
028	O
s	O
−	O
1	O
and	O
0	O
.	O
26	O
μM	O
)	O
so	O
determined	O
were	O
nearly	O
identical	O
to	O
those	O
of	O
MKP7ΔC304	B-mutant
(	O
0	O
.	O
029	O
s	O
−	O
1	O
and	O
0	O
.	O
27	O
μM	O
),	O
indicating	O
that	O
the	O
MKP7	O
-	O
KBD	O
has	O
no	O
effect	O
on	O
enzyme	O
catalysis	O
.	O
	O
We	O
next	O
examined	O
the	O
interaction	O
of	O
JNK1	O
with	O
the	O
CD	O
and	O
KBD	O
of	O
MKP7	O
by	O
gel	O
filtration	O
analysis	O
.	O
	O
When	O
3	O
molar	O
equivalents	O
of	O
CD	O
were	O
mixed	O
with	O
1	O
molar	O
equivalent	O
of	O
JNK1	O
,	O
a	O
significant	O
amount	O
of	O
CD	O
co	O
-	O
migrated	O
with	O
JNK1	O
to	O
earlier	O
fractions	O
,	O
and	O
the	O
excess	O
amount	O
of	O
CD	O
was	O
eluted	O
from	O
the	O
size	O
exclusion	O
column	O
as	O
a	O
monomer	O
,	O
indicating	O
stable	O
complex	O
formation	O
(	O
Fig	O
.	O
2c	O
).	O
	O
In	O
contrast	O
,	O
no	O
KBD	O
–	O
JNK1	O
complex	O
was	O
detected	O
when	O
3	O
molar	O
equivalents	O
of	O
KBD	O
were	O
mixed	O
with	O
1	O
molar	O
equivalent	O
of	O
JNK1	O
.	O
	O
As	O
shown	O
in	O
Fig	O
.	O
2d	O
,	O
the	O
CD	O
of	O
MKP7	O
can	O
be	O
pulled	O
down	O
by	O
JNK1	O
,	O
while	O
the	O
KBD	O
failed	O
to	O
bind	O
to	O
the	O
counterpart	O
protein	O
.	O
	O
Taken	O
together	O
,	O
our	O
data	O
indicate	O
that	O
the	O
CD	O
of	O
MKP7	O
,	O
but	O
not	O
the	O
KBD	O
domain	O
,	O
is	O
responsible	O
for	O
JNK	O
substrate	O
-	O
binding	O
and	O
enzymatic	O
specificity	O
.	O
	O
Crystal	O
structure	O
of	O
JNK1	O
in	O
complex	O
with	O
the	O
MKP7	O
-	O
CD	O
	O
To	O
understand	O
the	O
molecular	O
basis	O
of	O
JNK1	O
recognition	O
by	O
MKP7	O
,	O
we	O
determined	O
the	O
crystal	O
structure	O
of	O
unphosphorylated	O
JNK1	O
in	O
complex	O
with	O
the	O
MKP7	O
-	O
CD	O
(	O
Fig	O
.	O
3a	O
,	O
Supplementary	O
Fig	O
.	O
1a	O
and	O
Table	O
1	O
).	O
	O
In	O
the	O
complex	O
,	O
JNK1	O
has	O
its	O
characteristic	O
bilobal	O
structure	O
comprising	O
an	O
N	O
-	O
terminal	O
lobe	O
rich	O
in	O
β	O
-	O
sheet	O
and	O
a	O
C	O
-	O
terminal	O
lobe	O
that	O
is	O
mostly	O
α	O
-	O
helical	O
.	O
	O
One	O
side	O
of	O
the	O
β	O
-	O
sheet	O
is	O
covered	O
with	O
two	O
α	O
-	O
helices	O
and	O
the	O
other	O
is	O
covered	O
with	O
four	O
α	O
-	O
helices	O
(	O
Fig	O
.	O
3b	O
).	O
	O
MKP7	O
-	O
CD	O
is	O
positioned	O
onto	O
the	O
JNK1	O
molecule	O
so	O
that	O
the	O
active	O
site	O
of	O
the	O
phosphatase	O
faces	O
towards	O
the	O
activation	O
segment	O
.	O
	O
In	O
an	O
alignment	O
of	O
the	O
structure	O
of	O
MKP7	O
-	O
CD	O
with	O
that	O
of	O
VHR	O
,	O
an	O
atypical	O
‘	O
MKP	O
'	O
consisting	O
of	O
only	O
a	O
catalytic	O
domain	O
,	O
119	O
of	O
147	O
MKP7	O
-	O
CD	O
residues	O
could	O
be	O
superimposed	O
with	O
a	O
r	O
.	O
m	O
.	O
s	O
.	O
d	O
.	O
(	O
root	O
mean	O
squared	O
deviation	O
)	O
of	O
1	O
.	O
05	O
Å	O
(	O
Fig	O
.	O
3c	O
).	O
	O
The	O
most	O
striking	O
difference	O
is	O
that	O
helix	O
α0	O
and	O
loop	O
α0	O
–	O
β1	O
of	O
VHR	O
are	O
absent	O
in	O
MKP7	O
-	O
CD	O
.	O
	O
Another	O
region	O
that	O
cannot	O
be	O
aligned	O
with	O
VHR	O
is	O
found	O
in	O
loop	O
β3	O
–	O
β4	O
.	O
	O
This	O
loop	O
is	O
shortened	O
by	O
nine	O
residues	O
in	O
MKP7	O
-	O
CD	O
compared	O
with	O
that	O
in	O
VHR	O
.	O
	O
Asp213	O
in	O
MKP7	O
also	O
adopts	O
a	O
position	O
similar	O
to	O
that	O
of	O
Asp92	O
in	O
VHR	O
(	O
Supplementary	O
Fig	O
.	O
1c	O
),	O
indicating	O
that	O
Asp213	O
is	O
likely	O
to	O
function	O
as	O
the	O
general	O
acid	O
in	O
MKP7	O
.	O
	O
We	O
also	O
observed	O
the	O
binding	O
of	O
a	O
chloride	O
ion	O
in	O
the	O
active	O
site	O
of	O
MKP7	O
-	O
CD	O
.	O
	O
It	O
is	O
located	O
3	O
.	O
36	O
Å	O
from	O
the	O
Cys244	O
side	O
chain	O
and	O
makes	O
electrostatic	O
interactions	O
with	O
the	O
dipole	O
moment	O
of	O
helix	O
α3	O
and	O
with	O
several	O
main	O
-	O
chain	O
amide	O
groups	O
.	O
	O
Thus	O
this	O
chloride	O
ion	O
is	O
a	O
mimic	O
for	O
the	O
phosphate	O
group	O
of	O
the	O
substrate	O
,	O
as	O
revealed	O
by	O
a	O
comparison	O
with	O
the	O
structure	O
of	O
PTP1B	O
in	O
complex	O
with	O
phosphotyrosine	O
(	O
Supplementary	O
Fig	O
.	O
1d	O
).	O
	O
Although	O
the	O
catalytically	O
important	O
residues	O
in	O
MKP7	O
-	O
CD	O
are	O
well	O
aligned	O
with	O
those	O
in	O
VHR	O
,	O
the	O
residues	O
in	O
the	O
P	O
-	O
loop	O
of	O
MKP7	O
are	O
smaller	O
and	O
have	O
a	O
more	O
hydrophobic	O
character	O
than	O
those	O
of	O
VHR	O
(	O
Cys124	O
-	O
Arg125	O
-	O
Glu126	O
-	O
Gly127	O
-	O
Tyr128	O
-	O
Gly129	O
-	O
Arg130	O
;	O
Fig	O
.	O
3b	O
,	O
c	O
).	O
	O
The	O
difference	O
in	O
the	O
polarity	O
/	O
hydrophobicity	O
of	O
the	O
surface	O
may	O
also	O
point	O
to	O
the	O
origin	O
of	O
the	O
differences	O
in	O
the	O
substrate	O
-	O
recognition	O
mechanism	O
for	O
these	O
two	O
phosphatases	O
(	O
Supplementary	O
Fig	O
.	O
1e	O
,	O
f	O
).	O
	O
As	O
a	O
result	O
,	O
the	O
buried	O
solvent	O
-	O
accessible	O
surface	O
area	O
is	O
∼	O
1	O
,	O
315	O
Å	O
.	O
In	O
the	O
C	O
-	O
terminal	O
domain	O
,	O
JNK1	O
has	O
an	O
insertion	O
after	O
the	O
helix	O
αG	O
.	O
This	O
insertion	O
consists	O
of	O
two	O
helices	O
(	O
α1L14	O
and	O
α2L14	O
)	O
that	O
are	O
common	O
to	O
all	O
members	O
of	O
the	O
MAPK	O
family	O
.	O
	O
The	O
interactive	O
surface	O
in	O
JNK1	O
,	O
formed	O
by	O
the	O
helices	O
αG	O
and	O
α2L14	O
,	O
displays	O
a	O
hydrophobic	O
region	O
,	O
centred	O
at	O
Trp234	O
(	O
Fig	O
.	O
3d	O
).	O
	O
The	O
MKP7	O
-	O
docking	O
region	O
includes	O
two	O
helices	O
,	O
α4	O
and	O
α5	O
,	O
and	O
the	O
general	O
acid	O
loop	O
.	O
	O
The	O
aromatic	O
ring	O
of	O
Phe285	O
on	O
MKP7	O
α5	O
-	O
helix	O
is	O
nestled	O
in	O
a	O
hydrophobic	O
pocket	O
on	O
JNK1	O
,	O
formed	O
by	O
side	O
chains	O
of	O
Ile197	O
,	O
Leu198	O
,	O
Ile231	O
,	O
Trp234	O
,	O
Val256	O
,	O
Tyr259	O
,	O
Val260	O
and	O
the	O
aliphatic	O
portion	O
of	O
His230	O
(	O
Fig	O
.	O
3d	O
,	O
f	O
and	O
Supplementary	O
Fig	O
.	O
1g	O
).	O
	O
In	O
addition	O
,	O
there	O
are	O
hydrogen	O
bonds	O
between	O
Ser282	O
and	O
Asn286	O
of	O
MKP7	O
and	O
His230	O
and	O
Thr255	O
of	O
JNK1	O
,	O
and	O
the	O
main	O
chain	O
of	O
Phe215	O
in	O
the	O
general	O
acid	O
loop	O
of	O
MKP7	O
is	O
hydrogen	O
-	O
bonded	O
to	O
the	O
side	O
chain	O
of	O
Gln253	O
in	O
JNK1	O
.	O
	O
The	O
second	O
interactive	O
area	O
involves	O
the	O
α4	O
helix	O
of	O
MKP7	O
and	O
charged	O
/	O
polar	O
residues	O
of	O
JNK1	O
(	O
Fig	O
.	O
3e	O
).	O
	O
Mutational	O
analysis	O
of	O
the	O
JNK1	O
–	O
MKP7	O
docking	O
interface	O
	O
To	O
assess	O
the	O
importance	O
of	O
the	O
aforementioned	O
interactions	O
,	O
we	O
generated	O
a	O
series	O
of	O
point	O
mutations	O
on	O
the	O
MKP7	O
-	O
CD	O
and	O
examined	O
their	O
effect	O
on	O
the	O
MKP7	O
-	O
catalysed	O
JNK1	O
dephosphorylation	O
(	O
Fig	O
.	O
4a	O
).	O
	O
When	O
the	O
hydrophobic	O
residues	O
Phe285	O
and	O
Phe287	O
on	O
the	O
α5	O
of	O
MKP7	O
-	O
CD	O
were	O
replaced	O
by	O
Asp	O
or	O
Ala	O
,	O
their	O
phosphatase	O
activities	O
for	O
JNK1	O
dephosphorylation	O
decreased	O
∼	O
10	O
-	O
fold	O
.	O
	O
In	O
comparison	O
,	O
replacement	O
of	O
the	O
other	O
residues	O
(	O
Phe215	O
,	O
Asp268	O
,	O
Lys275	O
,	O
Ser282	O
,	O
Asn286	O
and	O
Leu292	O
)	O
with	O
an	O
Ala	O
or	O
Asp	O
individually	O
led	O
to	O
a	O
modest	O
decrease	O
in	O
catalytic	O
efficiencies	O
,	O
suggesting	O
that	O
this	O
position	O
may	O
only	O
affect	O
some	O
selectivity	O
of	O
MKP	O
.	O
	O
Interestingly	O
,	O
mutation	O
of	O
Phe287	O
results	O
in	O
a	O
considerable	O
loss	O
of	O
activity	O
against	O
pJNK1	O
without	O
altering	O
the	O
affinity	O
of	O
MKP7	O
-	O
CD	O
for	O
JNK1	O
(	O
Supplementary	O
Fig	O
.	O
2a	O
).	O
	O
We	O
also	O
generated	O
a	O
series	O
of	O
point	O
mutations	O
in	O
the	O
JNK1	O
and	O
assessed	O
the	O
effect	O
on	O
JNK1	O
binding	O
using	O
the	O
GST	O
pull	O
-	O
down	O
assay	O
(	O
Fig	O
.	O
4c	O
).	O
	O
Substitution	O
at	O
Asp229	O
,	O
Trp234	O
,	O
Thr255	O
,	O
Val256	O
,	O
Tyr259	O
and	O
Val260	O
significantly	O
reduced	O
the	O
binding	O
affinity	O
of	O
MKP7	O
-	O
CD	O
for	O
JNK	O
.	O
	O
Taken	O
together	O
,	O
these	O
results	O
are	O
consistent	O
with	O
the	O
present	O
crystallographic	O
model	O
,	O
which	O
reveal	O
the	O
hydrophobic	O
contacts	O
between	O
the	O
MKP7	O
catalytic	O
domain	O
and	O
JNK1	O
have	O
a	O
predominant	O
role	O
in	O
the	O
enzyme	O
–	O
substrate	O
interaction	O
,	O
and	O
hydrophobic	O
residue	O
Phe285	O
in	O
the	O
MKP7	O
-	O
CD	O
is	O
a	O
key	O
residue	O
for	O
its	O
high	O
-	O
affinity	O
binding	O
to	O
JNK1	O
.	O
	O
This	O
allosteric	O
activation	O
of	O
MKP3	O
has	O
been	O
well	O
-	O
documented	O
in	O
vitro	O
using	O
pNPP	O
,	O
a	O
small	O
-	O
molecule	O
phosphotyrosine	O
analogue	O
of	O
its	O
normal	O
substrate	O
.	O
	O
We	O
then	O
assayed	O
pNPPase	O
activities	O
of	O
MKP7ΔC304	B-mutant
and	O
MKP7	O
-	O
CD	O
in	O
the	O
presence	O
of	O
JNK1	O
.	O
	O
We	O
therefore	O
examined	O
the	O
effects	O
of	O
the	O
MKP7	O
-	O
CD	O
mutants	O
on	O
their	O
pNPPase	O
activities	O
.	O
	O
As	O
shown	O
in	O
Fig	O
.	O
4f	O
,	O
all	O
the	O
mutants	O
,	O
except	O
F287D	B-mutant
/	I-mutant
A	I-mutant
,	O
showed	O
little	O
or	O
no	O
activity	O
change	O
compared	O
with	O
the	O
wild	O
-	O
type	O
MKP7	O
-	O
CD	O
.	O
	O
In	O
the	O
JNK1	O
/	O
MKP7	O
-	O
CD	O
complex	O
structure	O
,	O
Phe287	O
of	O
MKP7	O
does	O
not	O
make	O
contacts	O
with	O
JNK1	O
substrate	O
.	O
	O
It	O
penetrates	O
into	O
a	O
pocket	O
formed	O
by	O
residues	O
from	O
the	O
P	O
-	O
loop	O
and	O
general	O
acid	O
loop	O
and	O
forms	O
hydrophobic	O
contacts	O
with	O
the	O
aliphatic	O
portions	O
of	O
side	O
chains	O
of	O
Arg250	O
,	O
Glu217	O
and	O
Ile219	O
,	O
suggesting	O
that	O
Phe287	O
in	O
MKP7	O
would	O
play	O
a	O
similar	O
role	O
to	O
that	O
of	O
its	O
structural	O
counterpart	O
in	O
the	O
PTPs	O
(	O
Gln266	O
in	O
PTP1B	O
)	O
and	O
VHR	O
(	O
Phe166	O
in	O
VHR	O
)	O
in	O
the	O
precise	O
alignment	O
of	O
active	O
-	O
site	O
residues	O
in	O
MKP7	O
with	O
respect	O
to	O
the	O
substrate	O
for	O
efficient	O
catalysis	O
(	O
Supplementary	O
Fig	O
.	O
2c	O
).	O
	O
Kinase	O
-	O
associated	O
phosphatase	O
(	O
KAP	O
),	O
a	O
member	O
of	O
the	O
DUSP	O
family	O
,	O
plays	O
a	O
crucial	O
role	O
in	O
cell	O
cycle	O
regulation	O
by	O
dephosphorylating	O
the	O
pThr160	O
residue	O
of	O
CDK2	O
(	O
cyclin	O
-	O
dependent	O
kinase	O
2	O
).	O
	O
The	O
crystal	O
structure	O
of	O
the	O
CDK2	O
/	O
KAP	O
complex	O
has	O
been	O
determined	O
at	O
3	O
.	O
0	O
Å	O
(	O
Fig	O
.	O
5a	O
).	O
	O
The	O
interface	O
between	O
these	O
two	O
proteins	O
consists	O
of	O
three	O
discontinuous	O
contact	O
regions	O
.	O
	O
There	O
is	O
a	O
hydrogen	O
bond	O
between	O
the	O
main	O
-	O
chain	O
nitrogen	O
of	O
Ile183	O
(	O
KAP	O
)	O
and	O
side	O
chain	O
oxygen	O
of	O
Glu208	O
(	O
CDK2	O
),	O
and	O
salt	O
bridges	O
between	O
Lys184	O
of	O
KAP	O
and	O
Asp235	O
of	O
CDK2	O
.	O
	O
The	O
substitution	O
of	O
the	O
two	O
hydrophobic	O
residues	O
with	O
charged	O
/	O
polar	O
residues	O
(	O
F285I	B-mutant
/	O
N286K	B-mutant
)	O
seriously	O
disrupts	O
the	O
hydrophobic	O
interaction	O
required	O
for	O
MKP7	O
binding	O
on	O
JNK1	O
(	O
Fig	O
.	O
4a	O
).	O
	O
F	O
-	O
site	O
interaction	O
is	O
crucial	O
for	O
JNK1	O
inactivation	O
in	O
vivo	O
	O
JNK	O
is	O
activated	O
following	O
cellular	O
exposure	O
to	O
a	O
number	O
of	O
acute	O
stimuli	O
such	O
as	O
anisomycin	O
,	O
H2O2	O
,	O
ultraviolet	O
light	O
,	O
sorbitol	O
,	O
DNA	O
-	O
damaging	O
agents	O
and	O
several	O
strong	O
apoptosis	O
inducers	O
(	O
etoposide	O
,	O
cisplatin	O
and	O
taxol	O
).	O
	O
To	O
assess	O
the	O
effects	O
of	O
MKP7	O
and	O
its	O
mutants	O
on	O
the	O
activation	O
of	O
endogenous	O
JNK	O
in	O
vivo	O
,	O
HEK293T	O
cells	O
were	O
transfected	O
with	O
blank	O
vector	O
or	O
with	O
HA	O
-	O
tagged	O
constructs	O
for	O
full	O
-	O
length	O
MKP7	O
,	O
MKP7ΔC304	B-mutant
and	O
MKP7	O
-	O
CD	O
or	O
MKP7	O
mutants	O
,	O
and	O
stimulated	O
with	O
ultraviolet	O
or	O
etoposide	O
treatment	O
.	O
	O
As	O
shown	O
in	O
Fig	O
.	O
6a	O
–	O
c	O
,	O
immunobloting	O
showed	O
similar	O
expression	O
levels	O
for	O
the	O
different	O
MKP7	O
constructs	O
in	O
all	O
the	O
cells	O
.	O
	O
Overexpressed	O
full	O
-	O
length	O
MKP7	O
,	O
MKP7ΔC304	B-mutant
and	O
MKP7	O
-	O
CD	O
significantly	O
reduced	O
the	O
endogenous	O
level	O
of	O
phosphorylated	O
JNK	O
compared	O
with	O
vector	O
-	O
transfected	O
cells	O
.	O
	O
We	O
next	O
tested	O
in	O
vivo	O
interactions	O
between	O
JNK1	O
mutants	O
and	O
full	O
-	O
length	O
MKP7	O
by	O
coimmunoprecipitation	O
experiments	O
under	O
unstimulated	O
conditions	O
.	O
	O
When	O
co	O
-	O
expressed	O
in	O
HEK293T	O
cells	O
,	O
wild	O
-	O
type	O
(	O
HA	O
)-	O
JNK1	O
was	O
readily	O
precipitated	O
with	O
(	O
Myc	O
)-	O
MKP7	O
(	O
Fig	O
.	O
6d	O
),	O
indicating	O
that	O
MKP7	O
binds	O
dephosphorylated	O
JNK1	O
protein	O
in	O
vivo	O
.	O
	O
Activation	O
of	O
the	O
JNK	O
signalling	O
pathway	O
is	O
frequently	O
associated	O
with	O
apoptotic	O
cell	O
death	O
,	O
and	O
inhibition	O
of	O
JNK	O
can	O
prevent	O
apoptotic	O
death	O
of	O
multiple	O
cells	O
.	O
	O
To	O
examine	O
whether	O
the	O
inhibition	O
of	O
JNK	O
activity	O
by	O
MKP7	O
would	O
provide	O
protections	O
against	O
the	O
apoptosis	O
,	O
we	O
analysed	O
the	O
rate	O
of	O
apoptosis	O
in	O
ultraviolet	O
-	O
irradiated	O
cells	O
transfected	O
with	O
MKP7	O
(	O
wild	O
type	O
or	O
mutants	O
)	O
by	O
flow	O
cytometry	O
.	O
	O
The	O
results	O
showed	O
similar	O
apoptotic	O
rates	O
between	O
cells	O
transfected	O
with	O
blank	O
vector	O
or	O
with	O
MKP7	O
(	O
wild	O
type	O
or	O
mutants	O
)	O
under	O
unstimulated	O
conditions	O
(	O
Supplementary	O
Fig	O
.	O
3b	O
),	O
while	O
ultraviolet	O
-	O
irradiation	O
significantly	O
increased	O
apoptotic	O
rate	O
in	O
cells	O
transfected	O
with	O
blank	O
vector	O
(	O
Fig	O
.	O
6e	O
).	O
	O
Expressions	O
of	O
wild	O
-	O
type	O
MKP7	O
,	O
MKP7ΔC304	B-mutant
and	O
MKP7	O
-	O
CD	O
significantly	O
decreased	O
the	O
proportion	O
of	O
apoptotic	O
cells	O
after	O
ultraviolet	O
treatment	O
.	O
	O
In	O
contrast	O
,	O
cells	O
transfected	O
with	O
the	O
MKP7	O
FXF	O
-	O
motif	O
mutants	O
(	O
F285D	B-mutant
,	O
F287D	B-mutant
and	O
L288D	B-mutant
)	O
showed	O
little	O
protective	O
effect	O
after	O
ultraviolet	O
treatment	O
and	O
similar	O
levels	O
of	O
apoptosis	O
rates	O
were	O
detected	O
to	O
cells	O
transfected	O
with	O
control	O
vectors	O
(	O
Fig	O
.	O
6e	O
,	O
f	O
).	O
	O
Taken	O
together	O
,	O
our	O
results	O
suggested	O
that	O
FXF	O
-	O
motif	O
-	O
mediated	O
,	O
rather	O
than	O
KBD	O
-	O
mediated	O
,	O
interaction	O
is	O
essential	O
for	O
MKP7	O
to	O
block	O
ultraviolet	O
-	O
induced	O
apoptosis	O
.	O
	O
A	O
similar	O
docking	O
mechanism	O
for	O
JNK1	O
recognition	O
by	O
MKP5	O
	O
MKP5	O
is	O
unique	O
among	O
the	O
MKPs	O
in	O
possessing	O
an	O
additional	O
domain	O
of	O
unknown	O
function	O
at	O
the	O
N	O
-	O
terminus	O
(	O
Fig	O
.	O
7a	O
).	O
	O
Deletion	O
of	O
the	O
KBD	O
in	O
MKP5	O
leads	O
to	O
a	O
280	O
-	O
fold	O
increase	O
in	O
Km	O
for	O
p38α	O
substrate	O
.	O
	O
The	O
crystal	O
structure	O
of	O
human	O
MKP5	O
-	O
CD	O
has	O
been	O
determined	O
.	O
	O
To	O
address	O
this	O
issue	O
,	O
we	O
first	O
examined	O
the	O
docking	O
ability	O
of	O
JNK1	O
to	O
the	O
KBD	O
and	O
CD	O
of	O
MKP5	O
using	O
gel	O
filtration	O
analysis	O
and	O
pull	O
-	O
down	O
assays	O
.	O
	O
It	O
can	O
be	O
seen	O
from	O
gel	O
filtration	O
experiments	O
that	O
JNK1	O
can	O
forms	O
a	O
stable	O
heterodimer	O
with	O
MKP5	O
-	O
CD	O
in	O
solution	O
,	O
but	O
no	O
detectable	O
interaction	O
was	O
found	O
with	O
the	O
KBD	O
domain	O
(	O
Fig	O
.	O
7d	O
).	O
	O
The	O
catalytic	O
domain	O
of	O
MKP5	O
,	O
but	O
not	O
its	O
KBD	O
,	O
was	O
able	O
to	O
pull	O
-	O
down	O
a	O
detectable	O
amount	O
of	O
JNK1	O
(	O
Fig	O
.	O
7e	O
),	O
implicating	O
a	O
different	O
substrate	O
-	O
recognition	O
mechanisms	O
for	O
p38	O
and	O
JNK	O
MAPKs	O
.	O
	O
To	O
further	O
test	O
our	O
hypothesis	O
,	O
we	O
generated	O
forms	O
of	O
MKP5	O
-	O
CD	O
bearing	O
mutations	O
corresponding	O
to	O
the	O
changes	O
we	O
made	O
on	O
MKP7	O
-	O
CD	O
on	O
the	O
basis	O
of	O
sequence	O
and	O
structural	O
alignment	O
and	O
examined	O
their	O
effects	O
on	O
the	O
phosphatase	O
activity	O
.	O
	O
Taken	O
together	O
,	O
our	O
results	O
suggest	O
that	O
MKP5	O
binds	O
JNK1	O
in	O
a	O
docking	O
mode	O
similar	O
to	O
that	O
in	O
the	O
JNK1	O
–	O
MKP7	O
complex	O
,	O
and	O
the	O
detailed	O
interaction	O
model	O
can	O
be	O
generated	O
using	O
molecular	O
dynamics	O
simulation	O
based	O
on	O
the	O
structure	O
of	O
JNK1	O
–	O
MKP7	O
-	O
CD	O
complex	O
(	O
Supplementary	O
Fig	O
.	O
4b	O
,	O
c	O
).	O
	O
In	O
this	O
model	O
,	O
the	O
MKP5	O
-	O
CD	O
adopts	O
a	O
conformation	O
nearly	O
identical	O
to	O
that	O
in	O
its	O
unbound	O
form	O
,	O
suggesting	O
that	O
the	O
conformation	O
of	O
the	O
catalytic	O
domain	O
undergoes	O
little	O
change	O
,	O
if	O
any	O
at	O
all	O
,	O
upon	O
JNK1	O
binding	O
.	O
	O
In	O
particular	O
,	O
Leu449	O
of	O
MKP5	O
,	O
which	O
is	O
equivalent	O
to	O
the	O
key	O
residue	O
Phe285	O
of	O
MKP7	O
,	O
buried	O
deeply	O
within	O
the	O
hydrophobic	O
pocket	O
of	O
JNK1	O
in	O
the	O
same	O
way	O
as	O
Phe285	O
in	O
the	O
JNK1	O
–	O
MKP7	O
-	O
CD	O
complex	O
(	O
Supplementary	O
Fig	O
.	O
4d	O
).	O
	O
Despite	O
the	O
strong	O
similarities	O
between	O
JNK1	O
–	O
MKP5	O
-	O
CD	O
and	O
JNK1	O
–	O
MKP7	O
-	O
CD	O
,	O
however	O
,	O
there	O
are	O
differences	O
.	O
	O
The	O
JNK1	O
–	O
MKP7	O
-	O
CD	O
interaction	O
is	O
better	O
and	O
more	O
extensive	O
.	O
	O
Asp268	O
of	O
MKP7	O
-	O
CD	O
forms	O
salt	O
bridge	O
with	O
JNK1	O
Arg263	O
,	O
whereas	O
the	O
corresponding	O
residue	O
Thr432	O
in	O
MKP5	O
-	O
CD	O
may	O
not	O
interact	O
with	O
JNK1	O
.	O
	O
Crystal	O
structures	O
of	O
ERK2	O
bound	O
with	O
the	O
D	O
-	O
motif	O
sequences	O
derived	O
from	O
MKP3	O
and	O
HePTP	O
have	O
been	O
reported	O
.	O
	O
These	O
structures	O
revealed	O
that	O
linear	O
docking	O
motifs	O
in	O
interacting	O
proteins	O
bind	O
to	O
a	O
common	O
docking	O
site	O
on	O
MAPKs	O
outside	O
the	O
kinase	O
active	O
site	O
.	O
	O
The	O
particular	O
amino	O
acids	O
and	O
their	O
spacing	O
within	O
D	O
-	O
motif	O
sequences	O
and	O
amino	O
acid	O
composition	O
of	O
the	O
docking	O
sites	O
on	O
MAPKs	O
appear	O
to	O
determine	O
the	O
specificity	O
of	O
D	O
-	O
motifs	O
for	O
individual	O
MAPKs	O
.	O
	O
Recently	O
,	O
the	O
crystal	O
structure	O
of	O
a	O
complex	O
between	O
the	O
KBD	O
of	O
MKP5	O
and	O
p38α	O
has	O
been	O
obtained	O
.	O
	O
This	O
complex	O
has	O
revealed	O
a	O
distinct	O
interaction	O
mode	O
for	O
MKP5	O
.	O
	O
The	O
KBD	O
of	O
MKP5	O
binds	O
to	O
p38α	O
in	O
the	O
opposite	O
polypeptide	O
direction	O
compared	O
with	O
how	O
the	O
D	O
-	O
motif	O
of	O
MKP3	O
binds	O
to	O
ERK2	O
.	O
	O
In	O
contrast	O
to	O
the	O
canonical	O
D	O
-	O
motif	O
-	O
binding	O
mode	O
,	O
separate	O
helices	O
,	O
α2	O
and	O
α3	O
′,	O
in	O
the	O
KBD	O
of	O
MKP5	O
engage	O
the	O
p38α	O
-	O
docking	O
site	O
.	O
	O
Further	O
structural	O
and	O
biochemical	O
studies	O
indicate	O
that	O
KBD	O
of	O
MKP7	O
may	O
interact	O
with	O
p38α	O
in	O
a	O
similar	O
manner	O
to	O
that	O
of	O
MKP5	O
.	O
	O
Taken	O
together	O
,	O
these	O
results	O
suggest	O
that	O
MKP7	O
utilizes	O
a	O
bipartite	O
recognition	O
mechanism	O
to	O
achieve	O
the	O
efficiency	O
and	O
fidelity	O
of	O
p38α	O
signalling	O
.	O
	O
In	O
addition	O
to	O
the	O
canonical	O
D	O
-	O
site	O
,	O
the	O
MAPK	O
ERK2	O
contains	O
a	O
second	O
binding	O
site	O
utilized	O
by	O
transcription	O
factor	O
substrates	O
and	O
phosphatases	O
,	O
the	O
FXF	O
-	O
motif	O
-	O
binding	O
site	O
(	O
also	O
called	O
F	O
-	O
site	O
),	O
that	O
is	O
exposed	O
in	O
active	O
ERK2	O
and	O
the	O
D	O
-	O
motif	O
peptide	O
-	O
induced	O
conformation	O
of	O
MAPKs	O
.	O
	O
MKP3	O
is	O
highly	O
specific	O
in	O
dephosphorylating	O
and	O
inactivating	O
ERK2	O
,	O
and	O
the	O
phosphatase	O
activity	O
of	O
the	O
MKP3	O
-	O
catalysed	O
pNPP	O
reaction	O
can	O
be	O
markedly	O
increased	O
in	O
the	O
presence	O
of	O
ERK2	O
(	O
refs	O
).	O
	O
Sequence	O
alignment	O
of	O
all	O
MKPs	O
reveals	O
a	O
high	O
degree	O
of	O
conservation	O
of	O
residues	O
surrounding	O
the	O
interacting	O
region	O
observed	O
in	O
JNK1	O
–	O
MKP7	O
-	O
CD	O
complex	O
(	O
Supplementary	O
Fig	O
.	O
5	O
).	O
	O
A	O
comprehensive	O
examination	O
of	O
the	O
molecular	O
basis	O
of	O
the	O
specific	O
ERK2	O
recognition	O
by	O
MKP3	O
is	O
underway	O
.	O
	O
The	O
FXF	O
-	O
motif	O
-	O
mediated	O
interaction	O
is	O
critical	O
for	O
both	O
pERK2	O
inactivation	O
and	O
ERK2	O
-	O
induced	O
MKP3	O
activation	O
(	O
manuscript	O
in	O
preparation	O
).	O
	O
In	O
summary	O
,	O
we	O
have	O
resolved	O
the	O
structure	O
of	O
JNK1	O
in	O
complex	O
with	O
the	O
catalytic	O
domain	O
of	O
MKP7	O
.	O
	O
Results	O
from	O
biochemical	O
characterization	O
of	O
the	O
Phe285	O
and	O
Phe287	O
MKP7	O
mutants	O
combined	O
with	O
structural	O
information	O
support	O
the	O
conclusion	O
that	O
the	O
two	O
Phe	O
residues	O
serve	O
different	O
roles	O
in	O
the	O
catalytic	O
reaction	O
.	O
	O
This	O
newly	O
identified	O
FXF	O
-	O
type	O
motif	O
is	O
present	O
in	O
all	O
MKPs	O
,	O
except	O
that	O
the	O
residue	O
at	O
the	O
first	O
position	O
in	O
MKP5	O
is	O
an	O
equivalent	O
hydrophobic	O
leucine	O
residue	O
(	O
see	O
also	O
Fig	O
.	O
7f	O
,	O
g	O
),	O
suggesting	O
that	O
these	O
two	O
Phe	O
residues	O
would	O
play	O
a	O
similar	O
role	O
in	O
facilitating	O
substrate	O
recognition	O
and	O
catalysis	O
,	O
respectively	O
.	O
	O
An	O
important	O
feature	O
of	O
MKP	O
–	O
JNK1	O
interactions	O
is	O
that	O
MKP7	O
or	O
MKP5	O
only	O
interact	O
with	O
the	O
F	O
-	O
site	O
of	O
JNK1	O
.	O
	O
One	O
possible	O
explanation	O
is	O
that	O
JNK1	O
needs	O
to	O
use	O
the	O
D	O
-	O
site	O
to	O
interact	O
with	O
JIP	O
-	O
1	O
,	O
a	O
scaffold	O
protein	O
for	O
JNK	O
signalling	O
.	O
	O
The	O
N	O
-	O
terminal	O
JNK	O
-	O
binding	O
domain	O
of	O
JIP	O
-	O
1	O
interacts	O
with	O
the	O
D	O
-	O
site	O
on	O
JNK	O
and	O
this	O
interaction	O
is	O
required	O
for	O
JIP	O
-	O
1	O
-	O
mediated	O
enhancement	O
of	O
JNK	O
activation	O
.	O
	O
In	O
addition	O
,	O
JIP	O
-	O
1	O
can	O
also	O
associate	O
with	O
MKP7	O
via	O
the	O
C	O
-	O
terminal	O
region	O
of	O
MKP7	O
(	O
ref	O
.).	O
	O
Thus	O
,	O
our	O
biochemical	O
and	O
structural	O
data	O
allow	O
us	O
to	O
present	O
a	O
model	O
for	O
the	O
JNK1	O
–	O
JIP	O
-	O
1	O
–	O
MKP7	O
ternary	O
complex	O
and	O
provide	O
an	O
important	O
insight	O
into	O
the	O
assembly	O
and	O
function	O
of	O
JNK	O
signalling	O
modules	O
(	O
Supplementary	O
Fig	O
.	O
6	O
).	O
	O
The	O
cDNAs	O
of	O
human	O
MKP7	O
and	O
MKP5	O
were	O
kindly	O
provided	O
by	O
Dr	O
Mathijs	O
Baens	O
(	O
University	O
of	O
Leuven	O
)	O
and	O
Dr	O
Eisuke	O
Nishida	O
(	O
Kyoto	O
University	O
),	O
respectively	O
.	O
	O
The	O
cDNAs	O
of	O
human	O
ASK1	O
,	O
MKK4	O
,	O
MKK7	O
and	O
JNK1	O
were	O
kindly	O
provided	O
by	O
Dr	O
Zhenguo	O
Wu	O
(	O
Hong	O
Kong	O
University	O
of	O
Science	O
and	O
Technology	O
).	O
	O
The	O
human	O
full	O
-	O
length	O
JNK1	O
,	O
MKK4	O
,	O
MKK7	O
and	O
the	O
kinase	O
domain	O
of	O
ASK1	O
(	O
659	O
–	O
951	O
)	O
were	O
cloned	O
into	O
pGEX4T	O
-	O
2	O
,	O
pET15b	O
and	O
/	O
or	O
pET21b	O
vectors	O
to	O
produce	O
a	O
GST	O
-	O
or	O
His	O
-	O
tagged	O
protein	O
.	O
	O
The	O
crystal	O
structure	O
of	O
unphosphorylated	O
JNK1	O
in	O
complex	O
with	O
the	O
catalytic	O
domain	O
of	O
MKP7	O
was	O
refined	O
to	O
2	O
.	O
4	O
Å	O
resolution	O
.	O
	O
Domain	O
structures	O
of	O
ten	O
human	O
MKPs	O
and	O
the	O
atypical	O
VHR	O
.	O
	O
On	O
the	O
basis	O
of	O
sequence	O
similarity	O
,	O
protein	O
structure	O
,	O
substrate	O
specificity	O
and	O
subcellular	O
localization	O
,	O
the	O
ten	O
members	O
of	O
MKP	O
family	O
can	O
be	O
divided	O
into	O
three	O
groups	O
.	O
	O
The	O
first	O
subfamily	O
comprises	O
MKP1	O
,	O
MKP2	O
,	O
PAC1	O
and	O
hVH3	O
,	O
which	O
are	O
inducible	O
nuclear	O
phosphatases	O
and	O
can	O
dephosphorylate	O
ERK	O
(	O
and	O
JNK	O
,	O
p38	O
)	O
MAPKs	O
.	O
	O
The	O
second	O
subfamily	O
contains	O
MKP3	O
,	O
MKP4	O
and	O
MKPX	O
,	O
which	O
are	O
cytoplasmic	O
ERK	O
-	O
specific	O
MKPs	O
.	O
	O
The	O
third	O
subfamily	O
comprises	O
MKP5	O
,	O
MKP7	O
and	O
hVH5	O
,	O
which	O
were	O
located	O
in	O
both	O
nucleus	O
and	O
cytoplasm	O
,	O
and	O
selectively	O
inactivate	O
JNK	O
and	O
p38	O
.	O
	O
All	O
MKPs	O
contain	O
both	O
the	O
CD	O
and	O
KBD	O
domains	O
,	O
whereas	O
VHR	O
,	O
an	O
atypical	O
MKP	O
,	O
only	O
contains	O
a	O
highly	O
conserved	O
catalytic	O
domain	O
.	O
	O
In	O
addition	O
to	O
the	O
CD	O
and	O
KBD	O
,	O
MKP7	O
contains	O
a	O
unique	O
long	O
C	O
-	O
terminal	O
region	O
that	O
contains	O
NES	O
,	O
NLS	O
and	O
PEST	O
motifs	O
,	O
which	O
has	O
no	O
effect	O
on	O
the	O
binding	O
ability	O
and	O
phosphatase	O
activity	O
of	O
MKP7	O
toward	O
MAPKs	O
.	O
	O
NES	O
,	O
nuclear	O
export	O
signal	O
;	O
NLS	O
,	O
nuclear	O
localization	O
signal	O
;	O
PEST	O
,	O
C	O
-	O
terminal	O
sequence	O
rich	O
in	O
prolines	O
,	O
glutamates	O
,	O
serines	O
and	O
threonines	O
.	O
	O
MKP7	O
-	O
CD	O
is	O
crucial	O
for	O
JNK1	O
binding	O
and	O
enzyme	O
catalysis	O
.	O
	O
The	O
KBD	O
and	O
CD	O
of	O
MKP7	O
are	O
shown	O
in	O
green	O
and	O
blue	O
,	O
and	O
the	O
N	O
-	O
lobe	O
and	O
C	O
-	O
lobe	O
of	O
JNK1	O
are	O
coloured	O
in	O
lemon	O
and	O
yellow	O
,	O
respectively	O
.	O
	O
The	O
error	O
bars	O
represent	O
s	O
.	O
e	O
.	O
m	O
.	O
(	O
c	O
)	O
Gel	O
filtration	O
analysis	O
for	O
interaction	O
of	O
JNK1	O
with	O
MKP7	O
-	O
CD	O
and	O
MKP7	O
-	O
KBD	O
.	O
	O
(	O
d	O
)	O
GST	O
-	O
mediated	O
pull	O
-	O
down	O
assay	O
for	O
interaction	O
of	O
JNK1	O
with	O
MKP7	O
-	O
CD	O
and	O
MKP7	O
-	O
KBD	O
.	O
	O
The	O
protein	O
amounts	O
of	O
MKP7	O
used	O
are	O
shown	O
at	O
the	O
bottom	O
.	O
	O
Structure	O
of	O
JNK1	O
in	O
complex	O
with	O
MKP7	O
-	O
CD	O
.	O
	O
(	O
a	O
)	O
Ribbon	O
diagram	O
of	O
JNK1	O
–	O
MKP7	O
-	O
CD	O
complex	O
in	O
two	O
views	O
related	O
by	O
a	O
45	O
°	O
rotation	O
around	O
a	O
vertical	O
axis	O
.	O
(	O
b	O
)	O
Structure	O
of	O
MKP7	O
-	O
CD	O
with	O
its	O
active	O
site	O
highlight	O
in	O
cyan	O
.	O
	O
The	O
JNK1	O
is	O
shown	O
in	O
surface	O
representation	O
coloured	O
according	O
to	O
electrostatic	O
potential	O
(	O
positive	O
,	O
blue	O
;	O
negative	O
,	O
red	O
).	O
	O
(	O
e	O
)	O
Interaction	O
networks	O
mainly	O
involving	O
helices	O
α4	O
and	O
α5	O
from	O
MKP7	O
-	O
CD	O
,	O
and	O
αG	O
and	O
α2L14	O
of	O
JNK1	O
.	O
	O
(	O
f	O
)	O
The	O
2Fo	O
−	O
Fc	O
omit	O
map	O
(	O
contoured	O
at	O
1	O
.	O
5σ	O
)	O
clearly	O
shows	O
electron	O
density	O
for	O
the	O
285FNFL288	O
segment	O
of	O
MKP7	O
-	O
CD	O
.	O
	O
(	O
a	O
)	O
Effects	O
of	O
mutations	O
in	O
MKP7	O
-	O
CD	O
on	O
the	O
JNK1	O
dephosphorylation	O
(	O
mean	O
±	O
s	O
.	O
e	O
.	O
m	O
.,	O
n	O
=	O
3	O
).	O
	O
Residues	O
involved	O
in	O
hydrophobic	O
and	O
hydrophilic	O
contacts	O
are	O
coloured	O
in	O
red	O
and	O
blue	O
,	O
respectively	O
.	O
(	O
b	O
)	O
Gel	O
filtration	O
analysis	O
for	O
interaction	O
of	O
JNK1	O
with	O
MKP7	O
-	O
CD	O
mutant	O
F285D	B-mutant
.	O
	O
Mutant	O
F285D	B-mutant
and	O
JNK1	O
were	O
eluted	O
as	O
monomers	O
,	O
with	O
the	O
molecular	O
masses	O
of	O
∼	O
17	O
and	O
44	O
kDa	O
,	O
respectively	O
.	O
	O
The	O
top	O
panel	O
shows	O
the	O
relative	O
affinities	O
of	O
MKP7	O
-	O
CD	O
to	O
JNK1	O
mutants	O
,	O
with	O
the	O
affinity	O
of	O
wild	O
-	O
type	O
JNK1	O
defined	O
as	O
100	O
%,	O
the	O
middle	O
panel	O
is	O
the	O
electrophoretic	O
pattern	O
of	O
MKP7	O
-	O
CD	O
and	O
JNK1	O
mutants	O
after	O
GST	O
pull	O
-	O
down	O
assays	O
.	O
	O
The	O
protein	O
amounts	O
of	O
MKP7	O
-	O
CD	O
used	O
are	O
shown	O
at	O
the	O
bottom	O
.	O
(	O
d	O
)	O
Circular	O
dichroism	O
spectra	O
for	O
MKP7	O
-	O
CD	O
wild	O
type	O
and	O
mutants	O
.	O
	O
(	O
f	O
)	O
Effects	O
of	O
mutations	O
in	O
MKP7	O
-	O
CD	O
on	O
the	O
pNPP	O
hydrolysis	O
reaction	O
(	O
mean	O
±	O
s	O
.	O
e	O
.	O
m	O
.,	O
n	O
=	O
3	O
).	O
	O
Comparison	O
of	O
CDK2	O
-	O
KAP	O
and	O
JNK1	O
–	O
MKP7	O
-	O
CD	O
.	O
	O
(	O
a	O
)	O
Superposition	O
of	O
the	O
complex	O
structures	O
of	O
CDK2	O
-	O
KAP	O
(	O
PDB	O
1FQ1	O
)	O
and	O
JNK1	O
–	O
MKP7	O
-	O
CD	O
.	O
	O
The	O
interactions	O
between	O
these	O
two	O
proteins	O
consist	O
of	O
three	O
discontinuous	O
contact	O
regions	O
,	O
centred	O
at	O
the	O
multiple	O
hydrogen	O
bonds	O
between	O
the	O
pThr160	O
of	O
CDK2	O
and	O
the	O
active	O
site	O
of	O
KAP	O
(	O
region	O
I	O
).	O
	O
(	O
b	O
)	O
Interactions	O
networks	O
at	O
the	O
auxiliary	O
region	O
II	O
mainly	O
involving	O
helix	O
α7	O
from	O
KAP	O
and	O
the	O
αG	O
helix	O
and	O
following	O
L14	O
loop	O
of	O
CDK2	O
.	O
	O
Residues	O
of	O
KAP	O
and	O
CDK2	O
are	O
highlighted	O
as	O
green	O
and	O
red	O
sticks	O
,	O
respectively	O
.	O
	O
(	O
d	O
)	O
Sequence	O
alignment	O
of	O
the	O
F	O
-	O
site	O
regions	O
on	O
MAPKs	O
.	O
	O
Residues	O
of	O
JNK1	O
involved	O
in	O
recognition	O
of	O
MKP7	O
are	O
indicated	O
by	O
orange	O
asterisks	O
,	O
and	O
those	O
forming	O
the	O
F	O
-	O
site	O
are	O
highlighted	O
in	O
yellow	O
.	O
	O
FXF	O
-	O
motif	O
is	O
critical	O
for	O
controlling	O
the	O
phosphorylation	O
of	O
JNK	O
and	O
ultraviolet	O
-	O
induced	O
apoptosis	O
.	O
	O
(	O
a	O
–	O
c	O
)	O
FXF	O
-	O
motif	O
is	O
essential	O
for	O
the	O
dephosphorylation	O
of	O
JNK	O
by	O
MKP7	O
.	O
	O
HEK293T	O
cells	O
were	O
infected	O
with	O
lentiviruses	O
expressing	O
MKP7	O
and	O
its	O
mutants	O
(	O
1	O
.	O
0	O
μg	O
).	O
	O
After	O
36	O
h	O
infection	O
,	O
cells	O
were	O
untreated	O
in	O
a	O
,	O
stimulated	O
with	O
30	O
μM	O
etoposide	O
for	O
3	O
h	O
in	O
b	O
or	O
irradiated	O
with	O
25	O
J	O
m	O
−	O
2	O
ultraviolet	O
light	O
at	O
30	O
min	O
before	O
lysis	O
in	O
c	O
.	O
Whole	O
-	O
cell	O
extracts	O
were	O
then	O
immunoblotted	O
with	O
antibody	O
indicated	O
.	O
	O
HEK293T	O
cells	O
were	O
co	O
-	O
transfected	O
with	O
MKP7	O
full	O
-	O
length	O
(	O
1	O
.	O
0	O
μg	O
)	O
and	O
JNK1	O
(	O
wild	O
type	O
or	O
mutants	O
as	O
indicated	O
,	O
1	O
.	O
0	O
μg	O
).	O
	O
Whole	O
-	O
cell	O
extracts	O
were	O
then	O
immunoprecipitated	O
with	O
antibody	O
against	O
Myc	O
for	O
MKP7	O
;	O
immunobloting	O
was	O
carried	O
out	O
with	O
antibodies	O
indicated	O
.	O
	O
IP	O
,	O
immunoprecipitation	O
;	O
TCL	O
,	O
total	O
cell	O
lysate	O
.	O
	O
HeLa	O
cells	O
were	O
infected	O
with	O
lentiviruses	O
expressing	O
MKP7	O
full	O
-	O
length	O
and	O
its	O
mutants	O
.	O
	O
Cells	O
were	O
then	O
subjected	O
to	O
flow	O
cytometry	O
analysis	O
.	O
	O
Apoptotic	O
cells	O
were	O
determined	O
by	O
Annexin	O
-	O
V	O
-	O
APC	O
/	O
PI	O
staining	O
.	O
	O
Staining	O
with	O
both	O
Annexin	O
-	O
V	O
and	O
PI	O
indicate	O
apoptosis	O
(	O
upper	O
right	O
quadrant	O
).	O
	O
MKP5	O
-	O
CD	O
is	O
crucial	O
for	O
JNK1	O
binding	O
and	O
enzyme	O
catalysis	O
.	O
	O
(	O
a	O
)	O
Domain	O
organization	O
of	O
human	O
MKP5	O
.	O
	O
The	O
KBD	O
and	O
CD	O
of	O
MKP5	O
are	O
shown	O
in	O
brown	O
and	O
grey	O
,	O
respectively	O
.	O
(	O
b	O
)	O
Plots	O
of	O
initial	O
velocity	O
of	O
the	O
MKP5	O
-	O
catalysed	O
reaction	O
versus	O
phospho	O
-	O
JNK1	O
concentration	O
.	O
	O
The	O
corresponding	O
residues	O
on	O
MKP5	O
are	O
depicted	O
as	O
orange	O
sticks	O
,	O
and	O
MKP5	O
residues	O
numbers	O
are	O
in	O
parentheses	O
.	O
	O
(	O
h	O
)	O
Pull	O
-	O
down	O
assays	O
of	O
MKP5	O
-	O
CD	O
by	O
GST	O
-	O
tagged	O
JNK1	O
mutants	O
.	O
	O
Plants	O
constantly	O
renew	O
during	O
their	O
life	O
cycle	O
and	O
thus	O
require	O
to	O
shed	O
senescent	O
and	O
damaged	O
organs	O
.	O
	O
Floral	O
abscission	O
is	O
controlled	O
by	O
the	O
leucine	O
-	O
rich	O
repeat	O
receptor	O
kinase	O
(	O
LRR	O
-	O
RK	O
)	O
HAESA	O
and	O
the	O
peptide	O
hormone	O
IDA	O
.	O
	O
Here	O
we	O
show	O
that	O
IDA	O
is	O
sensed	O
directly	O
by	O
the	O
HAESA	O
ectodomain	O
.	O
	O
Crystal	O
structures	O
of	O
HAESA	O
in	O
complex	O
with	O
IDA	O
reveal	O
a	O
hormone	O
binding	O
pocket	O
that	O
accommodates	O
an	O
active	O
dodecamer	O
peptide	O
.	O
	O
A	O
central	O
hydroxyproline	O
residue	O
anchors	O
IDA	O
to	O
the	O
receptor	O
.	O
	O
This	O
sequence	O
pattern	O
is	O
conserved	O
among	O
diverse	O
plant	O
peptides	O
,	O
suggesting	O
that	O
plant	O
peptide	O
hormone	O
receptors	O
may	O
share	O
a	O
common	O
ligand	O
binding	O
mode	O
and	O
activation	O
mechanism	O
.	O
	O
However	O
,	O
the	O
molecular	O
details	O
of	O
how	O
IDA	O
triggers	O
organ	O
shedding	O
are	O
not	O
clear	O
.	O
	O
The	O
shedding	O
of	O
floral	O
organs	O
(	O
or	O
leaves	O
)	O
can	O
be	O
easily	O
studied	O
in	O
a	O
model	O
plant	O
called	O
Arabidopsis	O
.	O
	O
Santiago	O
et	O
al	O
.	O
used	O
protein	O
biochemistry	O
,	O
structural	O
biology	O
and	O
genetics	O
to	O
uncover	O
how	O
the	O
IDA	O
hormone	O
activates	O
HAESA	O
.	O
	O
The	O
experiments	O
show	O
that	O
IDA	O
binds	O
directly	O
to	O
a	O
canyon	O
shaped	O
pocket	O
in	O
HAESA	O
that	O
extends	O
out	O
from	O
the	O
surface	O
of	O
the	O
cell	O
.	O
	O
IDA	O
binding	O
to	O
HAESA	O
allows	O
another	O
receptor	O
protein	O
called	O
SERK1	O
to	O
bind	O
to	O
HAESA	O
,	O
which	O
results	O
in	O
the	O
release	O
of	O
signals	O
inside	O
the	O
cell	O
that	O
trigger	O
the	O
shedding	O
of	O
organs	O
.	O
	O
The	O
next	O
step	O
following	O
on	O
from	O
this	O
work	O
is	O
to	O
understand	O
what	O
signals	O
are	O
produced	O
when	O
IDA	O
activates	O
HAESA	O
.	O
	O
Another	O
challenge	O
will	O
be	O
to	O
find	O
out	O
where	O
IDA	O
is	O
produced	O
in	O
the	O
plant	O
and	O
what	O
causes	O
it	O
to	O
accumulate	O
in	O
specific	O
places	O
in	O
preparation	O
for	O
organ	O
shedding	O
.	O
	O
(	O
A	O
)	O
SDS	O
PAGE	O
analysis	O
of	O
the	O
purified	O
Arabidopsis	O
thaliana	O
HAESA	O
ectodomain	O
(	O
residues	O
20	O
–	O
620	O
)	O
obtained	O
by	O
secreted	O
expression	O
in	O
insect	O
cells	O
.	O
	O
The	O
calculated	O
molecular	O
mass	O
is	O
65	O
.	O
7	O
kDa	O
,	O
the	O
actual	O
molecular	O
mass	O
obtained	O
by	O
mass	O
spectrometry	O
is	O
74	O
,	O
896	O
Da	O
,	O
accounting	O
for	O
the	O
N	O
-	O
glycans	O
.	O
(	O
B	O
)	O
Ribbon	O
diagrams	O
showing	O
front	O
(	O
left	O
panel	O
)	O
and	O
side	O
views	O
(	O
right	O
panel	O
)	O
of	O
the	O
isolated	O
HAESA	O
LRR	O
domain	O
.	O
	O
The	O
N	O
-	O
(	O
residues	O
20	O
–	O
88	O
)	O
and	O
C	O
-	O
terminal	O
(	O
residues	O
593	O
–	O
615	O
)	O
capping	O
domains	O
are	O
shown	O
in	O
yellow	O
,	O
the	O
central	O
21	O
LRR	O
motifs	O
are	O
in	O
blue	O
and	O
disulphide	O
bonds	O
are	O
highlighted	O
in	O
green	O
(	O
in	O
bonds	O
representation	O
).	O
(	O
C	O
)	O
Structure	O
based	O
sequence	O
alignment	O
of	O
the	O
21	O
leucine	O
-	O
rich	O
repeats	O
in	O
HAESA	O
with	O
the	O
plant	O
LRR	O
consensus	O
sequence	O
shown	O
for	O
comparison	O
.	O
	O
Conserved	O
hydrophobic	O
residues	O
are	O
shaded	O
in	O
gray	O
,	O
N	O
-	O
glycosylation	O
sites	O
visible	O
in	O
our	O
structures	O
are	O
highlighted	O
in	O
blue	O
,	O
cysteine	O
residues	O
involved	O
in	O
disulphide	O
bridge	O
formation	O
in	O
green	O
.	O
(	O
D	O
)	O
Asn	O
-	O
linked	O
glycans	O
mask	O
the	O
N	O
-	O
terminal	O
portion	O
of	O
the	O
HAESA	O
ectodomain	O
.	O
	O
Oligomannose	O
core	O
structures	O
(	O
containing	O
two	O
N	O
-	O
actylglucosamines	O
and	O
three	O
terminal	O
mannose	O
units	O
)	O
as	O
found	O
in	O
Trichoplusia	O
ni	O
cells	O
and	O
in	O
plants	O
were	O
modeled	O
onto	O
the	O
seven	O
glycosylation	O
sites	O
observed	O
in	O
our	O
HAESA	O
structures	O
,	O
to	O
visualize	O
the	O
surface	O
areas	O
potentially	O
not	O
masked	O
by	O
carbohydrate	O
.	O
	O
Hydrophobic	O
contacts	O
and	O
a	O
hydrogen	O
-	O
bond	O
network	O
mediate	O
the	O
interaction	O
between	O
HAESA	O
and	O
the	O
peptide	O
hormone	O
IDA	O
.	O
	O
(	O
A	O
)	O
Details	O
of	O
the	O
IDA	O
binding	O
pocket	O
.	O
	O
HAESA	O
is	O
shown	O
in	O
blue	O
(	O
ribbon	O
diagram	O
),	O
the	O
C	O
-	O
terminal	O
Arg	O
-	O
His	O
-	O
Asn	O
motif	O
(	O
left	O
panel	O
),	O
the	O
central	O
Hyp	O
anchor	O
(	O
center	O
)	O
and	O
the	O
N	O
-	O
terminal	O
Pro	O
-	O
rich	O
motif	O
in	O
IDA	O
(	O
right	O
panel	O
)	O
are	O
shown	O
in	O
yellow	O
(	O
in	O
bonds	O
representation	O
).	O
	O
HAESA	O
interface	O
residues	O
are	O
shown	O
as	O
sticks	O
,	O
selected	O
hydrogen	O
bond	O
interactions	O
are	O
denoted	O
as	O
dotted	O
lines	O
(	O
in	O
magenta	O
).	O
(	O
B	O
)	O
View	O
of	O
the	O
complete	O
IDA	O
(	O
in	O
bonds	O
representation	O
,	O
in	O
yellow	O
)	O
binding	O
pocket	O
in	O
HAESA	O
(	O
surface	O
view	O
,	O
in	O
blue	O
).	O
	O
Orientation	O
as	O
in	O
(	O
A	O
).	O
(	O
C	O
)	O
Structure	O
based	O
sequence	O
alignment	O
of	O
leucine	O
-	O
rich	O
repeats	O
in	O
HAESA	O
with	O
the	O
plant	O
LRR	O
consensus	O
sequence	O
shown	O
for	O
comparison	O
.	O
	O
The	O
IDA	O
binding	O
pocket	O
covers	O
LRRs	O
2	O
–	O
14	O
and	O
all	O
residues	O
originate	O
from	O
the	O
inner	O
surface	O
of	O
the	O
HAESA	O
superhelix	O
.	O
	O
The	O
IDA	O
-	O
HAESA	O
and	O
SERK1	O
-	O
HAESA	O
complex	O
interfaces	O
are	O
conserved	O
among	O
HAESA	O
and	O
HAESA	O
-	O
like	O
proteins	O
from	O
different	O
plant	O
species	O
.	O
	O
Structure	O
-	O
based	O
sequence	O
alignment	O
of	O
the	O
HAESA	O
family	O
members	O
:	O
Arabidopsis	O
thaliana	O
HAESA	O
(	O
Uniprot	O
(	O
http	O
://	O
www	O
.	O
uniprot	O
.	O
org	O
)	O
ID	O
P47735	O
),	O
Arabidopsis	O
thaliana	O
HSL2	O
(	O
Uniprot	O
ID	O
C0LGX3	O
),	O
Capsella	O
rubella	O
HAESA	O
(	O
Uniprot	O
ID	O
R0F2U6	O
),	O
Citrus	O
clementina	O
HSL2	O
(	O
Uniprot	O
ID	O
V4U227	O
),	O
Vitis	O
vinifera	O
HAESA	O
(	O
Uniprot	O
ID	O
F6HM39	O
).	O
	O
HAESA	O
residues	O
interacting	O
with	O
the	O
IDA	O
peptide	O
and	O
/	O
or	O
the	O
SERK1	O
co	O
-	O
receptor	O
kinase	O
ectodomain	O
are	O
highlighted	O
in	O
blue	O
and	O
orange	O
,	O
respectively	O
.	O
	O
The	O
peptide	O
hormone	O
IDA	O
binds	O
to	O
the	O
HAESA	O
LRR	O
ectodomain	O
.	O
	O
(	O
A	O
)	O
Multiple	O
sequence	O
alignment	O
of	O
selected	O
IDA	O
family	O
members	O
.	O
	O
The	O
conserved	O
PIP	O
motif	O
is	O
highlighted	O
in	O
yellow	O
,	O
the	O
central	O
Hyp	O
in	O
blue	O
.	O
	O
The	O
PKGV	O
motif	O
present	O
in	O
our	O
N	O
-	O
terminally	O
extended	O
IDA	O
peptide	O
is	O
highlighted	O
in	O
red	O
.	O
(	O
B	O
)	O
Isothermal	O
titration	O
calorimetry	O
of	O
the	O
HAESA	O
ectodomain	O
vs	O
.	O
IDA	O
and	O
including	O
the	O
synthetic	O
peptide	O
sequence	O
.	O
	O
(	O
C	O
)	O
Structure	O
of	O
the	O
HAESA	O
–	O
IDA	O
complex	O
with	O
HAESA	O
shown	O
in	O
blue	O
(	O
ribbon	O
diagram	O
).	O
	O
The	O
peptide	O
binding	O
pocket	O
covers	O
HAESA	O
LRRs	O
2	O
–	O
14	O
.	O
(	O
D	O
)	O
Close	O
-	O
up	O
view	O
of	O
the	O
entire	O
IDA	O
(	O
in	O
yellow	O
)	O
peptide	O
binding	O
site	O
in	O
HAESA	O
(	O
in	O
blue	O
).	O
	O
Details	O
of	O
the	O
interactions	O
between	O
the	O
central	O
Hyp	O
anchor	O
in	O
IDA	O
and	O
the	O
C	O
-	O
terminal	O
Arg	O
-	O
His	O
-	O
Asn	O
motif	O
with	O
HAESA	O
are	O
highlighted	O
in	O
(	O
E	O
)	O
and	O
(	O
F	O
),	O
respectively	O
.	O
	O
During	O
their	O
growth	O
,	O
development	O
and	O
reproduction	O
plants	O
use	O
cell	O
separation	O
processes	O
to	O
detach	O
no	O
-	O
longer	O
required	O
,	O
damaged	O
or	O
senescent	O
organs	O
.	O
	O
The	O
LRR	O
-	O
RKs	O
HAESA	O
(	O
greek	O
:	O
to	O
adhere	O
to	O
)	O
and	O
HAESA	O
-	O
LIKE	O
2	O
(	O
HSL2	O
)	O
redundantly	O
control	O
floral	O
abscission	O
.	O
	O
Loss	O
-	O
of	O
-	O
function	O
of	O
the	O
secreted	O
small	O
protein	O
INFLORESCENCE	O
DEFICIENT	O
IN	O
ABSCISSION	O
(	O
IDA	O
)	O
causes	O
floral	O
organs	O
to	O
remain	O
attached	O
while	O
its	O
over	O
-	O
expression	O
leads	O
to	O
premature	O
shedding	O
.	O
	O
Full	O
-	O
length	O
IDA	O
is	O
proteolytically	O
processed	O
and	O
a	O
conserved	O
stretch	O
of	O
20	O
amino	O
-	O
acids	O
(	O
termed	O
EPIP	O
)	O
can	O
rescue	O
the	O
IDA	O
loss	O
-	O
of	O
-	O
function	O
phenotype	O
(	O
Figure	O
1A	O
).	O
	O
It	O
has	O
been	O
demonstrated	O
that	O
a	O
dodecamer	O
peptide	O
within	O
EPIP	O
is	O
able	O
to	O
activate	O
HAESA	O
and	O
HSL2	O
in	O
transient	O
assays	O
in	O
tobacco	O
cells	O
.	O
	O
IDA	O
directly	O
binds	O
to	O
the	O
LRR	O
domain	O
of	O
HAESA	O
	O
Active	O
IDA	O
-	O
family	O
peptide	O
hormones	O
are	O
hydroxyprolinated	O
dodecamers	O
.	O
	O
Note	O
that	O
Pro58IDA	O
and	O
Leu67IDA	O
are	O
the	O
first	O
residues	O
defined	O
by	O
electron	O
density	O
when	O
bound	O
to	O
the	O
HAESA	O
ectodomain	O
.	O
(	O
D	O
)	O
Table	O
summaries	O
for	O
equilibrium	O
dissociation	O
constants	O
(	O
Kd	O
),	O
binding	O
enthalpies	O
(	O
ΔH	O
),	O
binding	O
entropies	O
(	O
ΔS	O
)	O
and	O
stoichoimetries	O
(	O
N	O
)	O
for	O
different	O
IDA	O
peptides	O
binding	O
to	O
the	O
HAESA	O
ectodomain	O
(	O
±	O
fitting	O
errors	O
;	O
n	O
.	O
d	O
.	O
	O
no	O
detectable	O
binding	O
).	O
(	O
E	O
)	O
Structural	O
superposition	O
of	O
the	O
active	O
IDA	O
(	O
in	O
bonds	O
representation	O
,	O
in	O
gray	O
)	O
and	O
IDL1	O
peptide	O
(	O
in	O
yellow	O
)	O
hormones	O
bound	O
to	O
the	O
HAESA	O
ectodomain	O
.	O
	O
Root	O
mean	O
square	O
deviation	O
(	O
r	O
.	O
m	O
.	O
s	O
.	O
d	O
.)	O
is	O
1	O
.	O
0	O
Å	O
comparing	O
100	O
corresponding	O
atoms	O
.	O
	O
The	O
receptor	O
kinase	O
SERK1	O
acts	O
as	O
a	O
HAESA	O
co	O
-	O
receptor	O
and	O
promotes	O
high	O
-	O
affinity	O
IDA	O
sensing	O
.	O
	O
(	O
A	O
)	O
Petal	O
break	O
-	O
strength	O
assays	O
measure	O
the	O
force	O
(	O
expressed	O
in	O
gram	O
equivalents	O
)	O
required	O
to	O
remove	O
the	O
petals	O
from	O
the	O
flower	O
of	O
serk	O
mutant	O
plants	O
compared	O
to	O
haesa	O
/	O
hsl2	O
mutant	O
and	O
Col	O
-	O
0	O
wild	O
-	O
type	O
flowers	O
.	O
	O
Petal	O
break	O
-	O
strength	O
was	O
found	O
significantly	O
increased	O
in	O
almost	O
all	O
positions	O
(	O
indicated	O
with	O
a	O
*)	O
for	O
haesa	O
/	O
hsl2	O
and	O
serk1	O
-	O
1	O
mutant	O
plants	O
with	O
respect	O
to	O
the	O
Col	O
-	O
0	O
control	O
.	O
	O
The	O
HAESA	O
LRR	O
domain	O
elutes	O
as	O
a	O
monomer	O
(	O
black	O
dotted	O
line	O
),	O
as	O
does	O
the	O
isolated	O
SERK1	O
ectodomain	O
(	O
blue	O
dotted	O
line	O
).	O
	O
A	O
HAESA	O
–	O
IDA	O
–	O
SERK1	O
complex	O
elutes	O
as	O
an	O
apparent	O
heterodimer	O
(	O
red	O
line	O
),	O
while	O
a	O
mixture	O
of	O
HAESA	O
and	O
SERK1	O
yields	O
two	O
isolated	O
peaks	O
that	O
correspond	O
to	O
monomeric	O
HAESA	O
and	O
SERK1	O
,	O
respectively	O
(	O
black	O
line	O
).	O
	O
no	O
detectable	O
binding	O
)	O
(	O
D	O
)	O
Analytical	O
size	O
-	O
exclusion	O
chromatography	O
in	O
the	O
presence	O
of	O
the	O
IDA	O
Hyp64	O
→	O
Pro	O
mutant	O
peptide	O
reveals	O
no	O
complex	O
formation	O
between	O
HAESA	O
and	O
SERK1	O
ectodomains	O
.	O
	O
(	O
E	O
)	O
In	O
vitro	O
kinase	O
assays	O
of	O
the	O
HAESA	O
and	O
SERK1	O
kinase	O
domains	O
.	O
	O
Wild	O
-	O
type	O
HAESA	O
and	O
SERK1	O
kinase	O
domains	O
(	O
KDs	O
)	O
exhibit	O
auto	O
-	O
phosphorylation	O
activities	O
(	O
lanes	O
1	O
+	O
3	O
).	O
	O
Transphosphorylation	O
activity	O
from	O
the	O
active	O
kinase	O
to	O
the	O
mutated	O
form	O
can	O
be	O
observed	O
in	O
both	O
directions	O
(	O
lanes	O
5	O
+	O
6	O
).	O
	O
A	O
Hyp	O
-	O
modified	O
dodecamer	O
comprising	O
the	O
highly	O
conserved	O
PIP	O
motif	O
in	O
IDA	O
(	O
Figure	O
1A	O
)	O
interacts	O
with	O
HAESA	O
with	O
1	O
:	O
1	O
stoichiometry	O
(	O
N	O
)	O
and	O
with	O
a	O
dissociation	O
constant	O
(	O
Kd	O
)	O
of	O
~	O
20	O
μM	O
(	O
Figure	O
1B	O
).	O
	O
The	O
central	O
Hyp64IDA	O
is	O
buried	O
in	O
a	O
specific	O
pocket	O
formed	O
by	O
HAESA	O
LRRs	O
8	O
–	O
10	O
,	O
with	O
its	O
hydroxyl	O
group	O
establishing	O
hydrogen	O
bonds	O
with	O
the	O
strictly	O
conserved	O
Glu266HAESA	O
and	O
with	O
a	O
water	O
molecule	O
,	O
which	O
in	O
turn	O
is	O
coordinated	O
by	O
the	O
main	O
chain	O
oxygens	O
of	O
Phe289HAESA	O
and	O
Ser311HAESA	O
(	O
Figure	O
1E	O
;	O
Figure	O
1	O
—	O
figure	O
supplement	O
3	O
).	O
	O
The	O
restricted	O
size	O
of	O
the	O
Hyp	O
pocket	O
suggests	O
that	O
IDA	O
does	O
not	O
require	O
arabinosylation	O
of	O
Hyp64IDA	O
for	O
activity	O
in	O
vivo	O
,	O
a	O
modification	O
that	O
has	O
been	O
reported	O
for	O
Hyp	O
residues	O
in	O
plant	O
CLE	O
peptide	O
hormones	O
.	O
	O
The	O
C	O
-	O
terminal	O
Arg	O
-	O
His	O
-	O
Asn	O
motif	O
in	O
IDA	O
maps	O
to	O
a	O
cavity	O
formed	O
by	O
HAESA	O
LRRs	O
11	O
–	O
14	O
(	O
Figure	O
1D	O
,	O
F	O
).	O
	O
The	O
COO	O
-	O
group	O
of	O
Asn69IDA	O
is	O
in	O
direct	O
contact	O
with	O
Arg407HAESA	O
and	O
Arg409HAESA	O
and	O
HAESA	O
cannot	O
bind	O
a	O
C	O
-	O
terminally	O
extended	O
IDA	B-mutant
-	I-mutant
SFVN	I-mutant
peptide	O
(	O
Figures	O
1D	O
,	O
F	O
,	O
2D	O
).	O
	O
This	O
suggests	O
that	O
the	O
conserved	O
Asn69IDA	O
may	O
constitute	O
the	O
very	O
C	O
-	O
terminus	O
of	O
the	O
mature	O
IDA	O
peptide	O
in	O
planta	O
and	O
that	O
active	O
IDA	O
is	O
generated	O
by	O
proteolytic	O
processing	O
from	O
a	O
longer	O
pre	O
-	O
protein	O
.	O
	O
Mutation	O
of	O
Arg417HSL2	O
(	O
which	O
corresponds	O
to	O
Arg409HAESA	O
)	O
causes	O
a	O
loss	O
-	O
of	O
-	O
function	O
phenotype	O
in	O
HSL2	O
,	O
which	O
indicates	O
that	O
the	O
peptide	O
binding	O
pockets	O
in	O
different	O
HAESA	O
receptors	O
have	O
common	O
structural	O
and	O
sequence	O
features	O
.	O
	O
Indeed	O
,	O
we	O
find	O
many	O
of	O
the	O
residues	O
contributing	O
to	O
the	O
formation	O
of	O
the	O
IDA	O
binding	O
surface	O
in	O
HAESA	O
to	O
be	O
conserved	O
in	O
HSL2	O
and	O
in	O
other	O
HAESA	O
-	O
type	O
receptors	O
in	O
different	O
plant	O
species	O
(	O
Figure	O
1	O
—	O
figure	O
supplement	O
3	O
).	O
	O
A	O
N	O
-	O
terminal	O
Pro	O
-	O
rich	O
motif	O
in	O
IDA	O
makes	O
contacts	O
with	O
LRRs	O
2	O
–	O
6	O
of	O
the	O
receptor	O
(	O
Figure	O
1D	O
,	O
Figure	O
1	O
—	O
figure	O
supplement	O
2A	O
–	O
C	O
).	O
	O
HAESA	O
specifically	O
senses	O
IDA	O
-	O
family	O
dodecamer	O
peptides	O
	O
We	O
next	O
investigated	O
whether	O
HAESA	O
binds	O
N	O
-	O
terminally	O
extended	O
versions	O
of	O
IDA	O
.	O
	O
We	O
obtained	O
a	O
structure	O
of	O
HAESA	O
in	O
complex	O
with	O
a	O
PKGV	B-mutant
-	I-mutant
IDA	I-mutant
peptide	O
at	O
1	O
.	O
94	O
Å	O
resolution	O
(	O
Table	O
2	O
).	O
	O
In	O
this	O
structure	O
,	O
no	O
additional	O
electron	O
density	O
accounts	O
for	O
the	O
PKGV	O
motif	O
at	O
the	O
IDA	O
N	O
-	O
terminus	O
(	O
Figure	O
2A	O
,	O
B	O
).	O
	O
IDL1	O
,	O
which	O
can	O
rescue	O
IDA	O
loss	O
-	O
of	O
-	O
function	O
mutants	O
when	O
introduced	O
in	O
abscission	O
zone	O
cells	O
,	O
can	O
also	O
be	O
sensed	O
by	O
HAESA	O
,	O
albeit	O
with	O
lower	O
affinity	O
(	O
Figure	O
2D	O
).	O
	O
A	O
2	O
.	O
56	O
Å	O
co	O
-	O
crystal	O
structure	O
with	O
IDL1	O
reveals	O
that	O
different	O
IDA	O
family	O
members	O
use	O
a	O
common	O
binding	O
mode	O
to	O
interact	O
with	O
HAESA	O
-	O
type	O
receptors	O
(	O
Figure	O
2A	O
–	O
C	O
,	O
E	O
,	O
Table	O
2	O
).	O
	O
We	O
do	O
not	O
detect	O
interaction	O
between	O
HAESA	O
and	O
a	O
synthetic	O
peptide	O
missing	O
the	O
C	O
-	O
terminal	O
Asn69IDA	O
(	O
ΔN69	B-mutant
),	O
highlighting	O
the	O
importance	O
of	O
the	O
polar	O
interactions	O
between	O
the	O
IDA	O
carboxy	O
-	O
terminus	O
and	O
Arg407HAESA	O
/	O
Arg409HAESA	O
(	O
Figures	O
1F	O
,	O
2D	O
).	O
	O
The	O
co	O
-	O
receptor	O
kinase	O
SERK1	O
allows	O
for	O
high	O
-	O
affinity	O
IDA	O
sensing	O
	O
It	O
has	O
been	O
recently	O
reported	O
that	O
SOMATIC	O
EMBRYOGENESIS	O
RECEPTOR	O
KINASES	O
(	O
SERKs	O
)	O
are	O
positive	O
regulators	O
of	O
floral	O
abscission	O
and	O
can	O
interact	O
with	O
HAESA	O
and	O
HSL2	O
in	O
an	O
IDA	O
-	O
dependent	O
manner	O
.	O
	O
As	O
all	O
five	O
SERK	O
family	O
members	O
appear	O
to	O
be	O
expressed	O
in	O
the	O
Arabidopsis	O
abscission	O
zone	O
,	O
we	O
quantified	O
their	O
relative	O
contribution	O
to	O
floral	O
abscission	O
in	O
Arabidopsis	O
using	O
a	O
petal	O
break	O
-	O
strength	O
assay	O
.	O
	O
We	O
found	O
that	O
the	O
force	O
required	O
to	O
remove	O
the	O
petals	O
of	O
serk1	O
-	O
1	O
mutants	O
is	O
significantly	O
higher	O
than	O
that	O
needed	O
for	O
wild	O
-	O
type	O
plants	O
,	O
as	O
previously	O
observed	O
for	O
haesa	O
/	O
hsl2	O
mutants	O
,	O
and	O
that	O
floral	O
abscission	O
is	O
delayed	O
in	O
serk1	O
-	O
1	O
(	O
Figure	O
3A	O
).	O
	O
Possibly	O
because	O
SERKs	O
have	O
additional	O
roles	O
in	O
plant	O
development	O
such	O
as	O
in	O
pollen	O
formation	O
and	O
brassinosteroid	O
signaling	O
,	O
we	O
found	O
that	O
higher	O
-	O
order	O
SERK	O
mutants	O
exhibit	O
pleiotropic	O
phenotypes	O
in	O
the	O
flower	O
,	O
rendering	O
their	O
analysis	O
and	O
comparison	O
by	O
quantitative	O
petal	O
break	O
-	O
strength	O
assays	O
difficult	O
.	O
	O
We	O
next	O
quantified	O
the	O
contribution	O
of	O
SERK1	O
to	O
IDA	O
recognition	O
by	O
HAESA	O
.	O
	O
We	O
next	O
titrated	O
SERK1	O
into	O
a	O
solution	O
containing	O
only	O
the	O
HAESA	O
ectodomain	O
.	O
	O
In	O
this	O
case	O
,	O
there	O
was	O
no	O
detectable	O
interaction	O
between	O
receptor	O
and	O
co	O
-	O
receptor	O
,	O
while	O
in	O
the	O
presence	O
of	O
IDA	O
,	O
SERK1	O
strongly	O
binds	O
HAESA	O
with	O
a	O
dissociation	O
constant	O
in	O
the	O
mid	O
-	O
nanomolar	O
range	O
(	O
Figure	O
3C	O
).	O
	O
This	O
suggests	O
that	O
IDA	O
itself	O
promotes	O
receptor	O
–	O
co	O
-	O
receptor	O
association	O
,	O
as	O
previously	O
described	O
for	O
the	O
steroid	O
hormone	O
brassinolide	O
and	O
for	O
other	O
LRR	O
-	O
RK	O
complexes	O
.	O
	O
Consistently	O
,	O
the	O
HAESA	O
kinase	O
domain	O
can	O
transphosphorylate	O
SERK1	O
and	O
vice	O
versa	O
in	O
in	O
vitro	O
transphosphorylation	O
assays	O
(	O
Figure	O
3E	O
).	O
	O
SERK1	O
senses	O
a	O
conserved	O
motif	O
in	O
IDA	O
family	O
peptides	O
	O
(	O
A	O
)	O
Overview	O
of	O
the	O
ternary	O
complex	O
with	O
HAESA	O
in	O
blue	O
(	O
surface	O
representation	O
),	O
IDA	O
in	O
yellow	O
(	O
bonds	O
representation	O
)	O
and	O
SERK1	O
in	O
orange	O
(	O
surface	O
view	O
).	O
(	O
B	O
)	O
The	O
HAESA	O
ectodomain	O
undergoes	O
a	O
conformational	O
change	O
upon	O
SERK1	O
co	O
-	O
receptor	O
binding	O
.	O
	O
Shown	O
are	O
Cα	O
traces	O
of	O
a	O
structural	O
superposition	O
of	O
the	O
unbound	O
(	O
yellow	O
)	O
and	O
SERK1	O
-	O
bound	O
(	O
blue	O
)	O
HAESA	O
ectodomains	O
(	O
r	O
.	O
m	O
.	O
s	O
.	O
d	O
.	O
is	O
1	O
.	O
5	O
Å	O
between	O
572	O
corresponding	O
Cα	O
atoms	O
).	O
	O
SERK1	O
(	O
in	O
orange	O
)	O
and	O
IDA	O
(	O
in	O
red	O
)	O
are	O
shown	O
alongside	O
.	O
	O
The	O
N	O
-	O
terminal	O
capping	O
domain	O
of	O
SERK1	O
(	O
in	O
orange	O
)	O
directly	O
contacts	O
the	O
C	O
-	O
terminal	O
part	O
of	O
IDA	O
(	O
in	O
yellow	O
,	O
in	O
bonds	O
representation	O
)	O
and	O
the	O
receptor	O
HAESA	O
(	O
in	O
blue	O
).	O
	O
Polar	O
contacts	O
of	O
SERK1	O
with	O
IDA	O
are	O
shown	O
in	O
magenta	O
,	O
with	O
the	O
HAESA	O
LRR	O
domain	O
in	O
gray	O
.	O
(	O
D	O
)	O
Details	O
of	O
the	O
zipper	O
-	O
like	O
SERK1	O
-	O
HAESA	O
interface	O
.	O
	O
Ribbon	O
diagrams	O
of	O
HAESA	O
(	O
in	O
blue	O
)	O
and	O
SERK1	O
(	O
in	O
orange	O
)	O
are	O
shown	O
with	O
selected	O
interface	O
residues	O
(	O
in	O
bonds	O
representation	O
).	O
	O
To	O
understand	O
in	O
molecular	O
terms	O
how	O
SERK1	O
contributes	O
to	O
high	O
-	O
affinity	O
IDA	O
recognition	O
,	O
we	O
solved	O
a	O
2	O
.	O
43	O
Å	O
crystal	O
structure	O
of	O
the	O
ternary	O
HAESA	O
–	O
IDA	O
–	O
SERK1	O
complex	O
(	O
Figure	O
4A	O
,	O
Table	O
2	O
).	O
	O
HAESA	O
LRRs	O
16	O
–	O
21	O
and	O
its	O
C	O
-	O
terminal	O
capping	O
domain	O
undergo	O
a	O
conformational	O
change	O
upon	O
SERK1	O
binding	O
(	O
Figure	O
4B	O
).	O
	O
SERK1	O
loop	O
residues	O
establish	O
multiple	O
hydrophobic	O
and	O
polar	O
contacts	O
with	O
Lys66IDA	O
and	O
the	O
C	O
-	O
terminal	O
Arg	O
-	O
His	O
-	O
Asn	O
motif	O
in	O
IDA	O
(	O
Figure	O
4C	O
).	O
	O
SERK1	O
LRRs	O
1	O
–	O
5	O
and	O
its	O
C	O
-	O
terminal	O
capping	O
domain	O
form	O
an	O
additional	O
zipper	O
-	O
like	O
interface	O
with	O
residues	O
originating	O
from	O
HAESA	O
LRRs	O
15	O
–	O
21	O
and	O
from	O
the	O
HAESA	O
C	O
-	O
terminal	O
cap	O
(	O
Figure	O
4D	O
).	O
	O
SERK1	O
binds	O
HAESA	O
using	O
these	O
two	O
distinct	O
interaction	O
surfaces	O
(	O
Figure	O
1	O
—	O
figure	O
supplement	O
3	O
),	O
with	O
the	O
N	O
-	O
cap	O
of	O
the	O
SERK1	O
LRR	O
domain	O
partially	O
covering	O
the	O
IDA	O
peptide	O
binding	O
cleft	O
.	O
	O
The	O
IDA	O
C	O
-	O
terminal	O
motif	O
is	O
required	O
for	O
HAESA	O
-	O
SERK1	O
complex	O
formation	O
and	O
for	O
IDA	O
bioactivity	O
.	O
	O
Purified	O
HAESA	O
and	O
SERK1	O
are	O
~	O
75	O
and	O
~	O
28	O
kDa	O
,	O
respectively	O
.	O
	O
Left	O
panel	O
:	O
IDA	B-mutant
K66A	I-mutant
/	I-mutant
R67A	I-mutant
;	O
center	O
:	O
IDA	B-mutant
ΔN69	I-mutant
,	O
right	O
panel	O
:	O
SDS	O
-	O
PAGE	O
of	O
peak	O
fractions	O
.	O
	O
Note	O
that	O
the	O
HAESA	O
and	O
SERK1	O
input	O
lanes	O
have	O
already	O
been	O
shown	O
in	O
Figure	O
3D	O
.	O
(	O
B	O
)	O
Isothermal	O
titration	O
thermographs	O
of	O
wild	O
-	O
type	O
and	O
mutant	O
IDA	O
peptides	O
titrated	O
into	O
a	O
HAESA	O
-	O
SERK1	O
mixture	O
in	O
the	O
cell	O
.	O
	O
Table	O
summaries	O
for	O
calorimetric	O
binding	O
constants	O
and	O
stoichoimetries	O
for	O
different	O
IDA	O
peptides	O
binding	O
to	O
the	O
HAESA	O
–	O
SERK1	O
ectodomain	O
mixture	O
(	O
±	O
fitting	O
errors	O
;	O
n	O
.	O
d	O
.	O
	O
(	O
C	O
)	O
Quantitative	O
petal	O
break	O
-	O
strength	O
assay	O
for	O
Col	O
-	O
0	O
wild	O
-	O
type	O
flowers	O
and	O
35S	O
::	O
IDA	O
wild	O
-	O
type	O
and	O
35S	O
::	O
IDA	B-mutant
K66A	I-mutant
/	I-mutant
R67A	I-mutant
mutant	O
flowers	O
.	O
	O
35S	O
::	O
IDA	O
plants	O
showed	O
significantly	O
increased	O
abscission	O
compared	O
to	O
Col	O
-	O
0	O
controls	O
in	O
inflorescence	O
positions	O
2	O
and	O
3	O
(	O
a	O
).	O
	O
Up	O
to	O
inflorescence	O
position	O
4	O
,	O
petal	O
break	O
in	O
35S	O
::	O
IDA	B-mutant
K66A	I-mutant
/	I-mutant
R67A	I-mutant
mutant	O
plants	O
was	O
significantly	O
increased	O
compared	O
to	O
both	O
Col	O
-	O
0	O
control	O
plants	O
(	O
b	O
)	O
and	O
35S	O
::	O
IDA	O
plants	O
(	O
c	O
)	O
(	O
D	O
)	O
Normalized	O
expression	O
levels	O
(	O
relative	O
expression	O
±	O
standard	O
error	O
;	O
ida	O
:	O
-	O
0	O
.	O
02	O
±	O
0	O
.	O
001	O
;	O
Col	O
-	O
0	O
:	O
1	O
±	O
0	O
.	O
11	O
;	O
35S	O
::	O
IDA	O
124	O
±	O
0	O
.	O
75	O
;	O
35S	O
::	O
IDA	B-mutant
K66A	I-mutant
/	I-mutant
R67A	I-mutant
:	O
159	O
±	O
0	O
.	O
58	O
)	O
of	O
IDA	O
wild	O
-	O
type	O
and	O
mutant	O
transcripts	O
in	O
the	O
35S	O
promoter	O
over	O
-	O
expression	O
lines	O
analyzed	O
in	O
(	O
C	O
).	O
(	O
E	O
)	O
Magnified	O
view	O
of	O
representative	O
abscission	O
zones	O
from	O
35S	O
::	O
IDA	O
,	O
Col	O
-	O
0	O
wild	O
-	O
type	O
and	O
35S	O
::	O
IDA	B-mutant
K66A	I-mutant
/	I-mutant
R67A	I-mutant
double	O
-	O
mutant	O
T3	O
transgenic	O
lines	O
.	O
	O
The	O
four	O
C	O
-	O
terminal	O
residues	O
in	O
IDA	O
(	O
Lys66IDA	O
-	O
Asn69IDA	O
)	O
are	O
conserved	O
among	O
IDA	O
family	O
members	O
and	O
are	O
in	O
direct	O
contact	O
with	O
SERK1	O
(	O
Figures	O
1A	O
,	O
4C	O
).	O
	O
We	O
thus	O
assessed	O
their	O
contribution	O
to	O
HAESA	O
–	O
SERK1	O
complex	O
formation	O
.	O
	O
Deletion	O
of	O
the	O
buried	O
Asn69IDA	O
completely	O
inhibits	O
receptor	O
–	O
co	O
-	O
receptor	O
complex	O
formation	O
and	O
HSL2	O
activation	O
(	O
Figure	O
5A	O
,	O
B	O
).	O
	O
A	O
synthetic	O
Lys66IDA	B-mutant
/	I-mutant
Arg67IDA	I-mutant
→	I-mutant
Ala	I-mutant
mutant	O
peptide	O
(	O
IDA	B-mutant
K66A	I-mutant
/	I-mutant
R66A	I-mutant
)	O
showed	O
a	O
10	O
fold	O
reduced	O
binding	O
affinity	O
when	O
titrated	O
in	O
a	O
HAESA	O
/	O
SERK1	O
protein	O
solution	O
(	O
Figures	O
5A	O
,	O
B	O
,	O
2D	O
).	O
	O
Comparison	O
of	O
35S	O
::	O
IDA	O
wild	O
-	O
type	O
and	O
mutant	O
plants	O
further	O
indicates	O
that	O
mutation	O
of	O
Lys66IDA	B-mutant
/	I-mutant
Arg67IDA	I-mutant
→	I-mutant
Ala	I-mutant
may	O
cause	O
a	O
weak	O
dominant	O
negative	O
effect	O
(	O
Figure	O
5C	O
–	O
E	O
).	O
	O
In	O
agreement	O
with	O
our	O
structures	O
and	O
biochemical	O
assays	O
,	O
this	O
experiment	O
suggests	O
a	O
role	O
of	O
the	O
conserved	O
IDA	O
C	O
-	O
terminus	O
in	O
the	O
control	O
of	O
floral	O
abscission	O
.	O
	O
In	O
contrast	O
to	O
animal	O
LRR	O
receptors	O
,	O
plant	O
LRR	O
-	O
RKs	O
harbor	O
spiral	O
-	O
shaped	O
ectodomains	O
and	O
thus	O
they	O
require	O
shape	O
-	O
complementary	O
co	O
-	O
receptor	O
proteins	O
for	O
receptor	O
activation	O
.	O
	O
SERK1	O
has	O
been	O
previously	O
reported	O
as	O
a	O
positive	O
regulator	O
in	O
plant	O
embryogenesis	O
,	O
male	O
sporogenesis	O
,	O
brassinosteroid	O
signaling	O
and	O
in	O
phytosulfokine	O
perception	O
.	O
	O
Recent	O
findings	O
by	O
and	O
our	O
mechanistic	O
studies	O
now	O
also	O
support	O
a	O
positive	O
role	O
for	O
SERK1	O
in	O
floral	O
abscission	O
.	O
	O
As	O
serk1	O
-	O
1	O
mutant	O
plants	O
show	O
intermediate	O
abscission	O
phenotypes	O
when	O
compared	O
to	O
haesa	O
/	O
hsl2	O
mutants	O
,	O
SERK1	O
likely	O
acts	O
redundantly	O
with	O
other	O
SERKs	O
in	O
the	O
abscission	O
zone	O
(	O
Figure	O
3A	O
).	O
	O
It	O
has	O
been	O
previously	O
suggested	O
that	O
SERK1	O
can	O
inhibit	O
cell	O
separation	O
.	O
	O
However	O
our	O
results	O
show	O
that	O
SERK1	O
also	O
can	O
activate	O
this	O
process	O
upon	O
IDA	O
sensing	O
,	O
indicating	O
that	O
SERKs	O
may	O
fulfill	O
several	O
different	O
functions	O
in	O
the	O
course	O
of	O
the	O
abscission	O
process	O
.	O
	O
While	O
the	O
sequence	O
of	O
the	O
mature	O
IDA	O
peptide	O
has	O
not	O
been	O
experimentally	O
determined	O
in	O
planta	O
,	O
our	O
HAESA	O
-	O
IDA	O
complex	O
structures	O
and	O
calorimetry	O
assays	O
suggest	O
that	O
active	O
IDLs	O
are	O
hydroxyprolinated	O
dodecamers	O
.	O
	O
Our	O
comparative	O
structural	O
and	O
biochemical	O
analysis	O
further	O
suggests	O
that	O
IDLs	O
share	O
a	O
common	O
receptor	O
binding	O
mode	O
,	O
but	O
may	O
preferably	O
bind	O
to	O
HAESA	O
,	O
HSL1	O
or	O
HSL2	O
in	O
different	O
plant	O
tissues	O
and	O
organs	O
.	O
	O
The	O
fact	O
that	O
SERK1	O
specifically	O
interacts	O
with	O
the	O
very	O
C	O
-	O
terminus	O
of	O
IDLs	O
may	O
allow	O
for	O
the	O
rational	O
design	O
of	O
peptide	O
hormone	O
antagonists	O
,	O
as	O
previously	O
demonstrated	O
for	O
the	O
brassinosteroid	O
pathway	O
.	O
	O
Importantly	O
,	O
our	O
calorimetry	O
assays	O
reveal	O
that	O
the	O
SERK1	O
ectodomain	O
binds	O
HAESA	O
with	O
nanomolar	O
affinity	O
,	O
but	O
only	O
in	O
the	O
presence	O
of	O
IDA	O
(	O
Figure	O
3C	O
).	O
	O
This	O
ligand	O
-	O
induced	O
formation	O
of	O
a	O
receptor	O
–	O
co	O
-	O
receptor	O
complex	O
may	O
allow	O
the	O
HAESA	O
and	O
SERK1	O
kinase	O
domains	O
to	O
efficiently	O
trans	O
-	O
phosphorylate	O
and	O
activate	O
each	O
other	O
in	O
the	O
cytoplasm	O
.	O
	O
SERK1	O
uses	O
partially	O
overlapping	O
surface	O
areas	O
to	O
activate	O
different	O
plant	O
signaling	O
receptors	O
.	O
	O
(	O
A	O
)	O
Structural	O
comparison	O
of	O
plant	O
steroid	O
and	O
peptide	O
hormone	O
membrane	O
signaling	O
complexes	O
.	O
	O
Left	O
panel	O
:	O
Ribbon	O
diagram	O
of	O
HAESA	O
(	O
in	O
blue	O
),	O
SERK1	O
(	O
in	O
orange	O
)	O
and	O
IDA	O
(	O
in	O
bonds	O
and	O
surface	O
represention	O
).	O
	O
Right	O
panel	O
:	O
Ribbon	O
diagram	O
of	O
the	O
plant	O
steroid	O
receptor	O
BRI1	O
(	O
in	O
blue	O
)	O
bound	O
to	O
brassinolide	O
(	O
in	O
gray	O
,	O
in	O
bonds	O
representation	O
)	O
and	O
to	O
SERK1	O
,	O
shown	O
in	O
the	O
same	O
orientation	O
(	O
PDB	O
-	O
ID	O
.	O
4lsx	O
).	O
	O
A	O
ribbon	O
diagram	O
of	O
SERK1	O
in	O
the	O
same	O
orientation	O
is	O
shown	O
alongside	O
.	O
	O
Residues	O
interacting	O
with	O
the	O
HAESA	O
or	O
BRI1	O
LRR	O
domains	O
are	O
shown	O
in	O
orange	O
or	O
magenta	O
,	O
respectively	O
.	O
	O
Comparison	O
of	O
our	O
HAESA	O
–	O
IDA	O
–	O
SERK1	O
structure	O
with	O
the	O
brassinosteroid	O
receptor	O
signaling	O
complex	O
,	O
where	O
SERK1	O
also	O
acts	O
as	O
co	O
-	O
receptor	O
,	O
reveals	O
an	O
overall	O
conserved	O
mode	O
of	O
SERK1	O
binding	O
,	O
while	O
the	O
ligand	O
binding	O
pockets	O
map	O
to	O
very	O
different	O
areas	O
in	O
the	O
corresponding	O
receptors	O
(	O
LRRs	O
2	O
–	O
14	O
;	O
HAESA	O
;	O
LRRs	O
21	O
–	O
25	O
,	O
BRI1	O
)	O
and	O
may	O
involve	O
an	O
island	O
domain	O
(	O
BRI1	O
)	O
or	O
not	O
(	O
HAESA	O
)	O
(	O
Figure	O
6A	O
).	O
	O
Several	O
residues	O
in	O
the	O
SERK1	O
N	O
-	O
terminal	O
capping	O
domain	O
(	O
Thr59SERK1	O
,	O
Phe61SERK1	O
)	O
and	O
the	O
LRR	O
inner	O
surface	O
(	O
Asp75SERK1	O
,	O
Tyr101SERK1	O
,	O
SER121SERK1	O
,	O
Phe145SERK1	O
)	O
contribute	O
to	O
the	O
formation	O
of	O
both	O
complexes	O
(	O
Figures	O
4C	O
,	O
D	O
,	O
6B	O
).	O
	O
These	O
residues	O
are	O
not	O
involved	O
in	O
the	O
sensing	O
of	O
the	O
steroid	O
hormone	O
brassinolide	O
.	O
	O
In	O
both	O
cases	O
however	O
,	O
the	O
co	O
-	O
receptor	O
completes	O
the	O
hormone	O
binding	O
pocket	O
.	O
	O
This	O
fact	O
together	O
with	O
the	O
largely	O
overlapping	O
SERK1	O
binding	O
surfaces	O
in	O
HAESA	O
and	O
BRI1	O
allows	O
us	O
to	O
speculate	O
that	O
SERK1	O
may	O
promote	O
high	O
-	O
affinity	O
peptide	O
hormone	O
and	O
brassinosteroid	O
sensing	O
by	O
simply	O
slowing	O
down	O
dissociation	O
of	O
the	O
ligand	O
from	O
its	O
cognate	O
receptor	O
.	O
	O
The	O
conserved	O
(	O
Arg	O
)-	O
His	O
-	O
Asn	O
motif	O
is	O
highlighted	O
in	O
red	O
,	O
the	O
central	O
Hyp	O
residue	O
in	O
IDLs	O
and	O
CLEs	O
is	O
marked	O
in	O
blue	O
.	O
	O
Our	O
experiments	O
reveal	O
that	O
SERK1	O
recognizes	O
a	O
C	O
-	O
terminal	O
Arg	O
-	O
His	O
-	O
Asn	O
motif	O
in	O
IDA	O
.	O
	O
Among	O
these	O
are	O
the	O
CLE	O
peptides	O
regulating	O
stem	O
cell	O
maintenance	O
in	O
the	O
shoot	O
and	O
the	O
root	O
.	O
	O
Diverse	O
plant	O
peptide	O
hormones	O
may	O
thus	O
also	O
bind	O
their	O
LRR	O
-	O
RK	O
receptors	O
in	O
an	O
extended	O
conformation	O
along	O
the	O
inner	O
surface	O
of	O
the	O
LRR	O
domain	O
and	O
may	O
also	O
use	O
small	O
,	O
shape	O
-	O
complementary	O
co	O
-	O
receptors	O
for	O
high	O
-	O
affinity	O
ligand	O
binding	O
and	O
receptor	O
activation	O
.	O
	O
Ensemble	O
cryo	O
-	O
EM	O
uncovers	O
inchworm	O
-	O
like	O
translocation	O
of	O
a	O
viral	O
IRES	O
through	O
the	O
ribosome	O
	O
Internal	O
ribosome	O
entry	O
sites	O
(	O
IRESs	O
)	O
mediate	O
cap	O
-	O
independent	O
translation	O
of	O
viral	O
mRNAs	O
.	O
	O
The	O
IRES	O
rearranges	O
from	O
extended	O
to	O
bent	O
to	O
extended	O
conformations	O
.	O
	O
This	O
inchworm	O
-	O
like	O
movement	O
is	O
coupled	O
with	O
ribosomal	O
inter	O
-	O
subunit	O
rotation	O
and	O
40S	O
head	O
swivel	O
.	O
	O
eEF2	O
,	O
attached	O
to	O
the	O
60S	O
subunit	O
,	O
slides	O
along	O
the	O
rotating	O
40S	O
subunit	O
to	O
enter	O
the	O
A	O
site	O
.	O
	O
Its	O
diphthamide	O
-	O
bearing	O
tip	O
at	O
domain	O
IV	O
separates	O
the	O
tRNA	O
-	O
mRNA	O
-	O
like	O
pseudoknot	O
I	O
(	O
PKI	O
)	O
of	O
the	O
IRES	O
from	O
the	O
decoding	O
center	O
.	O
	O
This	O
unlocks	O
40S	O
domains	O
,	O
facilitating	O
head	O
swivel	O
and	O
biasing	O
IRES	O
translocation	O
via	O
hitherto	O
-	O
elusive	O
intermediates	O
with	O
PKI	O
captured	O
between	O
the	O
A	O
and	O
P	O
sites	O
.	O
	O
Virus	O
propagation	O
relies	O
on	O
the	O
host	O
translational	O
apparatus	O
.	O
	O
To	O
efficiently	O
compete	O
with	O
host	O
mRNAs	O
and	O
engage	O
in	O
translation	O
under	O
stress	O
,	O
some	O
viral	O
mRNAs	O
undergo	O
cap	O
-	O
independent	O
translation	O
.	O
	O
An	O
IRES	O
is	O
located	O
at	O
the	O
5	O
’	O
untranslated	O
region	O
of	O
the	O
viral	O
mRNA	O
,	O
preceding	O
an	O
open	O
reading	O
frame	O
(	O
ORF	O
).	O
	O
Subsequent	O
binding	O
of	O
an	O
elongator	O
aminoacyl	O
-	O
tRNA	O
to	O
the	O
ribosomal	O
A	O
site	O
transitions	O
the	O
initiation	O
complex	O
into	O
the	O
elongation	O
cycle	O
of	O
translation	O
.	O
	O
Upon	O
peptide	O
bond	O
formation	O
,	O
the	O
two	O
tRNAs	O
and	O
their	O
respective	O
mRNA	O
codons	O
translocate	O
from	O
the	O
A	O
and	O
P	O
to	O
P	O
and	O
E	O
(	O
exit	O
)	O
sites	O
,	O
freeing	O
the	O
A	O
site	O
for	O
the	O
next	O
elongator	O
tRNA	O
.	O
	O
The	O
IGR	O
IRES	O
mRNAs	O
do	O
not	O
contain	O
an	O
AUG	O
start	O
codon	O
.	O
	O
The	O
IGR	O
-	O
IRES	O
-	O
driven	O
initiation	O
does	O
not	O
involve	O
initiator	O
tRNAMet	O
and	O
initiation	O
factors	O
.	O
	O
As	O
such	O
,	O
this	O
group	O
of	O
IRESs	O
represents	O
the	O
most	O
streamlined	O
mechanism	O
of	O
eukaryotic	O
translation	O
initiation	O
.	O
	O
Early	O
electron	O
cryo	O
-	O
microscopy	O
(	O
cryo	O
-	O
EM	O
)	O
studies	O
have	O
found	O
that	O
the	O
CrPV	O
IRES	O
packs	O
in	O
the	O
ribosome	O
intersubunit	O
space	O
.	O
	O
Recent	O
cryo	O
-	O
EM	O
structures	O
of	O
ribosome	O
-	O
bound	O
TSV	O
IRES	O
and	O
CrPV	O
IRES	O
revealed	O
that	O
IGR	O
IRESs	O
position	O
the	O
ORF	O
by	O
mimicking	O
a	O
translating	O
ribosome	O
bound	O
with	O
tRNA	O
and	O
mRNA	O
.	O
	O
The	O
~	O
200	O
-	O
nt	O
IRES	O
RNAs	O
span	O
from	O
the	O
A	O
site	O
beyond	O
the	O
E	O
site	O
.	O
	O
A	O
conserved	O
tRNA	O
-	O
mRNA	O
–	O
like	O
structural	O
element	O
of	O
pseudoknot	O
I	O
(	O
PKI	O
)	O
interacts	O
with	O
the	O
decoding	O
center	O
in	O
the	O
A	O
site	O
of	O
the	O
40S	O
subunit	O
.	O
	O
The	O
downstream	O
initiation	O
codon	O
—	O
coding	O
for	O
alanine	O
—	O
is	O
placed	O
in	O
the	O
mRNA	O
tunnel	O
,	O
preceding	O
the	O
decoding	O
center	O
.	O
	O
PKI	O
of	O
IGR	O
IRESs	O
therefore	O
mimics	O
an	O
A	O
-	O
site	O
elongator	O
tRNA	O
interacting	O
with	O
an	O
mRNA	O
sense	O
codon	O
,	O
but	O
not	O
a	O
P	O
-	O
site	O
initiator	O
tRNAMet	O
and	O
the	O
AUG	O
start	O
codon	O
.	O
	O
A	O
cryo	O
-	O
EM	O
structure	O
of	O
the	O
ribosome	O
bound	O
with	O
a	O
CrPV	O
IRES	O
and	O
release	O
factor	O
eRF1	O
occupying	O
the	O
A	O
site	O
provided	O
insight	O
into	O
the	O
post	O
-	O
translocation	O
state	O
.	O
	O
In	O
this	O
structure	O
,	O
PKI	O
is	O
positioned	O
in	O
the	O
P	O
site	O
and	O
the	O
first	O
mRNA	O
codon	O
is	O
located	O
in	O
the	O
A	O
site	O
.	O
	O
How	O
the	O
large	O
IRES	O
RNA	O
translocates	O
within	O
the	O
ribosome	O
,	O
allowing	O
PKI	O
translocation	O
from	O
the	O
A	O
to	O
P	O
site	O
is	O
not	O
known	O
.	O
	O
The	O
structural	O
similarity	O
of	O
PKI	O
and	O
the	O
tRNA	O
anticodon	O
stem	O
loop	O
(	O
ASL	O
)	O
bound	O
to	O
a	O
codon	O
suggests	O
that	O
their	O
mechanisms	O
of	O
translocation	O
are	O
similar	O
to	O
some	O
extent	O
.	O
	O
Translocation	O
of	O
the	O
IRES	O
or	O
tRNA	O
-	O
mRNA	O
requires	O
eukaryotic	O
elongation	O
factor	O
2	O
(	O
eEF2	O
),	O
a	O
structural	O
and	O
functional	O
homolog	O
of	O
the	O
well	O
-	O
studied	O
bacterial	O
EF	O
-	O
G	O
.	O
Pre	O
-	O
translocation	O
tRNA	O
-	O
bound	O
ribosomes	O
contain	O
a	O
peptidyl	O
-	O
and	O
deacyl	O
-	O
tRNA	O
,	O
both	O
base	O
-	O
paired	O
to	O
mRNA	O
codons	O
in	O
the	O
A	O
and	O
P	O
sites	O
(	O
termed	O
2tRNA	O
•	O
mRNA	O
complex	O
).	O
	O
Intersubunit	O
rotation	O
occurs	O
spontaneously	O
upon	O
peptidyl	O
transfer	O
,	O
and	O
is	O
coupled	O
with	O
formation	O
of	O
hybrid	O
tRNA	O
states	O
.	O
	O
In	O
the	O
rotated	O
pre	O
-	O
translocation	O
ribosome	O
,	O
the	O
peptidyl	O
-	O
tRNA	O
binds	O
the	O
A	O
site	O
of	O
the	O
small	O
subunit	O
with	O
its	O
ASL	O
and	O
the	O
P	O
site	O
of	O
the	O
large	O
subunit	O
with	O
the	O
CCA	O
3	O
’	O
end	O
(	O
A	O
/	O
P	O
hybrid	O
state	O
).	O
	O
The	O
ribosome	O
can	O
undergo	O
spontaneous	O
,	O
thermally	O
-	O
driven	O
forward	O
-	O
reverse	O
rotation	O
that	O
shifts	O
the	O
two	O
tRNAs	O
between	O
the	O
hybrid	O
and	O
'	O
classical	O
'	O
states	O
while	O
the	O
anticodon	O
stem	O
loops	O
remain	O
non	O
-	O
translocated	O
.	O
	O
EF	O
-	O
G	O
is	O
thought	O
to	O
'	O
unlock	O
'	O
the	O
pre	O
-	O
translocation	O
ribosome	O
,	O
allowing	O
movement	O
of	O
the	O
2tRNA	O
•	O
mRNA	O
complex	O
,	O
however	O
the	O
structural	O
details	O
of	O
this	O
unlocking	O
are	O
not	O
known	O
.	O
	O
The	O
second	O
large	O
-	O
scale	O
rearrangement	O
involves	O
rotation	O
,	O
or	O
swiveling	O
,	O
of	O
the	O
head	O
of	O
the	O
small	O
subunit	O
relative	O
to	O
the	O
body	O
.	O
	O
The	O
head	O
can	O
rotate	O
by	O
up	O
to	O
~	O
20	O
°	O
around	O
the	O
axis	O
nearly	O
orthogonal	O
to	O
that	O
of	O
intersubunit	O
rotation	O
,	O
in	O
the	O
absence	O
of	O
tRNA	O
or	O
in	O
the	O
presence	O
of	O
a	O
single	O
P	O
/	O
E	O
tRNA	O
and	O
eEF2	O
or	O
EF	O
-	O
G	O
.	O
Förster	O
resonance	O
energy	O
transfer	O
(	O
FRET	O
)	O
data	O
suggest	O
that	O
head	O
swivel	O
of	O
the	O
rotated	O
small	O
subunit	O
facilitates	O
EF	O
-	O
G	O
-	O
mediated	O
movement	O
of	O
2tRNA	O
•	O
mRNA	O
.	O
	O
The	O
structural	O
role	O
of	O
head	O
swivel	O
is	O
not	O
fully	O
understood	O
.	O
	O
Whether	O
and	O
how	O
the	O
head	O
swivel	O
mediates	O
tRNA	O
transition	O
from	O
the	O
A	O
to	O
P	O
site	O
remains	O
unknown	O
.	O
	O
Comparison	O
of	O
70S	O
•	O
2tRNA	O
•	O
mRNA	O
and	O
80S	O
•	O
IRES	O
translocation	O
complexes	O
.	O
	O
Nucleotides	O
C1054	O
,	O
G966	O
and	O
G693	O
of	O
16S	O
rRNA	O
are	O
shown	O
in	O
black	O
to	O
denote	O
the	O
A	O
,	O
P	O
and	O
E	O
sites	O
,	O
respectively	O
.	O
	O
The	O
extents	O
of	O
the	O
30S	O
subunit	O
rotation	O
and	O
head	O
swivel	O
relative	O
to	O
their	O
positions	O
in	O
the	O
post	O
-	O
translocation	O
structure	O
are	O
shown	O
with	O
arrows	O
.	O
	O
References	O
and	O
PDB	O
codes	O
of	O
the	O
structures	O
are	O
shown	O
.	O
	O
(	O
b	O
)	O
Structures	O
of	O
the	O
80S	O
•	O
IRES	O
complexes	O
in	O
the	O
absence	O
and	O
presence	O
of	O
eEF2	O
(	O
this	O
work	O
).	O
	O
The	O
large	O
ribosomal	O
subunit	O
is	O
shown	O
in	O
cyan	O
;	O
the	O
small	O
subunit	O
in	O
light	O
yellow	O
(	O
head	O
)	O
and	O
wheat	O
-	O
yellow	O
(	O
body	O
);	O
the	O
TSV	O
IRES	O
in	O
red	O
,	O
eEF2	O
in	O
green	O
.	O
	O
Unresolved	O
regions	O
of	O
the	O
IRES	O
in	O
densities	O
for	O
Structures	O
III	O
and	O
V	O
are	O
shown	O
in	O
gray	O
.	O
	O
The	O
extents	O
of	O
the	O
40S	O
subunit	O
rotation	O
and	O
head	O
swivel	O
relative	O
to	O
their	O
positions	O
in	O
the	O
post	O
-	O
translocation	O
structure	O
are	O
shown	O
with	O
arrows	O
.	O
	O
3D	O
classification	O
using	O
a	O
3D	O
mask	O
around	O
the	O
40S	O
head	O
,	O
TSV	O
IRES	O
and	O
eEF2	O
,	O
of	O
the	O
4x	O
binned	O
stack	O
was	O
used	O
to	O
identify	O
particles	O
containing	O
both	O
the	O
IRES	O
and	O
eEF2	O
.	O
	O
Cryo	O
-	O
EM	O
density	O
of	O
Structures	O
I	O
-	O
V	O
.	O
	O
In	O
panels	O
(	O
a	O
-	O
e	O
),	O
the	O
maps	O
are	O
segmented	O
and	O
colored	O
as	O
in	O
Figure	O
1	O
.	O
	O
The	O
maps	O
in	O
all	O
panels	O
were	O
B	O
-	O
softened	O
by	O
applying	O
a	O
B	O
-	O
factor	O
of	O
30	O
Å2	O
.	O
	O
(	O
a	O
-	O
e	O
)	O
Cryo	O
-	O
EM	O
map	O
of	O
Structures	O
I	O
,	O
II	O
,	O
III	O
,	O
IV	O
and	O
V	O
.	O
(	O
f	O
-	O
j	O
)	O
Local	O
resolution	O
of	O
unfiltered	O
and	O
unmasked	O
cryo	O
-	O
EM	O
reconstructions	O
,	O
assessed	O
using	O
Blocres	O
from	O
the	O
BSoft	O
package	O
,	O
for	O
Structures	O
I	O
,	O
II	O
,	O
III	O
,	O
IV	O
and	O
V	O
.	O
(	O
k	O
-	O
o	O
)	O
Cryo	O
-	O
EM	O
density	O
for	O
the	O
TSV	O
IRES	O
(	O
red	O
model	O
)	O
and	O
eEF2	O
(	O
green	O
model	O
)	O
in	O
Structures	O
I	O
,	O
II	O
,	O
III	O
,	O
IV	O
and	O
V	O
.	O
(	O
p	O
)	O
Fourier	O
shell	O
correlation	O
(	O
FSC	O
)	O
curves	O
for	O
Structures	O
I	O
-	O
V	O
.	O
The	O
horizontal	O
axis	O
is	O
labeled	O
with	O
spatial	O
frequency	O
Å	O
-	O
1	O
and	O
with	O
Å	O
.	O
The	O
resolutions	O
stated	O
in	O
the	O
text	O
correspond	O
to	O
an	O
FSC	O
threshold	O
value	O
of	O
0	O
.	O
143	O
,	O
shown	O
as	O
a	O
dotted	O
line	O
,	O
for	O
the	O
FREALIGN	O
-	O
derived	O
FSC	O
('	O
Part_FSC	O
').	O
	O
Nucleotides	O
C1274	O
,	O
U1191	O
of	O
the	O
40S	O
head	O
and	O
G904	O
of	O
the	O
platform	O
(	O
C1054	O
,	O
G966	O
and	O
G693	O
in	O
E	O
.	O
coli	O
16S	O
rRNA	O
)	O
are	O
shown	O
in	O
black	O
to	O
denote	O
the	O
A	O
,	O
P	O
and	O
E	O
sites	O
,	O
respectively	O
.	O
	O
(	O
b	O
)	O
Schematic	O
representation	O
of	O
the	O
structures	O
shown	O
in	O
panel	O
a	O
,	O
denoting	O
the	O
conformations	O
of	O
the	O
small	O
subunit	O
relative	O
to	O
the	O
large	O
subunit	O
.	O
	O
A	O
,	O
P	O
and	O
E	O
sites	O
are	O
shown	O
as	O
rectangles	O
.	O
	O
We	O
used	O
cryo	O
-	O
EM	O
to	O
visualize	O
80S	O
•	O
TSV	O
IRES	O
complexes	O
formed	O
in	O
the	O
presence	O
of	O
eEF2	O
•	O
GTP	O
and	O
the	O
translation	O
inhibitor	O
sordarin	O
,	O
which	O
stabilizes	O
eEF2	O
on	O
the	O
ribosome	O
.	O
	O
Although	O
the	O
mechanism	O
of	O
sordarin	O
action	O
is	O
not	O
fully	O
understood	O
,	O
the	O
inhibitor	O
does	O
not	O
affect	O
the	O
conformation	O
of	O
eEF2	O
•	O
GDPNP	O
on	O
the	O
ribosome	O
,	O
rendering	O
it	O
an	O
excellent	O
tool	O
in	O
translocation	O
studies	O
.	O
	O
This	O
ensemble	O
of	O
structures	O
allowed	O
us	O
to	O
reconstruct	O
a	O
sequence	O
of	O
steps	O
in	O
IRES	O
translocation	O
induced	O
by	O
eEF2	O
.	O
	O
We	O
used	O
single	O
-	O
particle	O
cryo	O
-	O
EM	O
and	O
maximum	O
-	O
likelihood	O
image	O
classification	O
in	O
FREALIGN	O
to	O
obtain	O
three	O
-	O
dimensional	O
density	O
maps	O
from	O
a	O
single	O
specimen	O
.	O
	O
The	O
translocation	O
complex	O
was	O
formed	O
using	O
S	O
.	O
cerevisiae	O
80S	O
ribosomes	O
,	O
Taura	O
syndrome	O
virus	O
IRES	O
,	O
and	O
S	O
.	O
cerevisiae	O
eEF2	O
in	O
the	O
presence	O
of	O
GTP	O
and	O
the	O
eEF2	O
-	O
binding	O
translation	O
inhibitor	O
sordarin	O
.	O
	O
Unsupervised	O
cryo	O
-	O
EM	O
data	O
classification	O
was	O
combined	O
with	O
the	O
use	O
of	O
three	O
-	O
dimensional	O
and	O
two	O
-	O
dimensional	O
masking	O
around	O
the	O
ribosomal	O
A	O
site	O
(	O
Figure	O
1	O
—	O
figure	O
supplement	O
2	O
).	O
	O
Large	O
-	O
scale	O
rearrangements	O
in	O
Structures	O
I	O
through	O
V	O
,	O
coupled	O
with	O
the	O
movement	O
of	O
PKI	O
from	O
the	O
A	O
to	O
P	O
site	O
and	O
eEF2	O
entry	O
into	O
the	O
A	O
site	O
.	O
	O
(	O
a	O
)	O
Rotational	O
states	O
of	O
the	O
40S	O
subunit	O
in	O
the	O
80S	O
•	O
IRES	O
structure	O
(	O
INIT	O
;	O
PDB	O
3J6Y	O
)	O
and	O
in	O
80S	O
•	O
IRES	O
•	O
eEF2	O
Structures	O
I	O
,	O
II	O
,	O
III	O
,	O
IV	O
and	O
V	O
(	O
this	O
work	O
).	O
	O
For	O
each	O
structure	O
,	O
the	O
triangle	O
outlines	O
the	O
contours	O
of	O
the	O
40S	O
body	O
;	O
the	O
lower	O
angle	O
illustrates	O
the	O
extent	O
of	O
intersubunit	O
(	O
body	O
)	O
rotation	O
.	O
	O
The	O
sizes	O
of	O
the	O
arrows	O
correspond	O
to	O
the	O
extent	O
of	O
the	O
head	O
swivel	O
(	O
yellow	O
)	O
and	O
subunit	O
rotation	O
(	O
black	O
).	O
	O
(	O
b	O
)	O
Solvent	O
view	O
(	O
opposite	O
from	O
that	O
shown	O
in	O
(	O
a	O
))	O
of	O
the	O
40S	O
subunit	O
in	O
the	O
80S	O
•	O
IRES	O
structure	O
(	O
INIT	O
;	O
PDB	O
3J6Y	O
)	O
and	O
in	O
80S	O
•	O
IRES	O
•	O
eEF2	O
Structures	O
I	O
,	O
II	O
,	O
III	O
,	O
IV	O
and	O
V	O
(	O
this	O
work	O
).	O
	O
The	O
structures	O
are	O
colored	O
as	O
in	O
Figure	O
1	O
.	O
	O
(	O
a	O
)	O
Comparison	O
of	O
the	O
40S	O
-	O
subunit	O
rotational	O
states	O
in	O
Structures	O
I	O
through	O
V	O
,	O
sampling	O
a	O
~	O
10	O
°	O
range	O
between	O
Structure	O
I	O
(	O
fully	O
rotated	O
)	O
and	O
Structure	O
V	O
(	O
non	O
-	O
rotated	O
).	O
	O
The	O
superpositions	O
of	O
Structures	O
I	O
-	O
V	O
were	O
performed	O
by	O
structural	O
alignments	O
of	O
the	O
25S	O
ribosomal	O
RNAs	O
.	O
	O
(	O
b	O
)	O
Bar	O
graph	O
of	O
the	O
angles	O
characterizing	O
the	O
40S	O
rotational	O
and	O
40S	O
head	O
swiveling	O
states	O
in	O
Structures	O
I	O
through	O
V	O
.	O
Measurements	O
for	O
the	O
two	O
80S	O
•	O
IRES	O
(	O
INIT	O
)	O
structures	O
are	O
included	O
for	O
comparison	O
.	O
	O
(	O
d	O
)	O
Comparison	O
of	O
conformations	O
of	O
the	O
L1	O
and	O
P	O
stalks	O
of	O
the	O
large	O
subunit	O
in	O
Structures	O
I	O
through	O
V	O
with	O
those	O
in	O
the	O
80S	O
•	O
IRES	O
and	O
tRNA	O
-	O
bound	O
80S	O
structures	O
.	O
	O
Superpositions	O
were	O
performed	O
by	O
structural	O
alignments	O
of	O
25S	O
ribosomal	O
RNAs	O
.	O
	O
(	O
e	O
)	O
Bar	O
graph	O
of	O
the	O
positions	O
of	O
PKI	O
and	O
domain	O
IV	O
of	O
eEF2	O
relative	O
to	O
the	O
P	O
site	O
residues	O
of	O
the	O
head	O
(	O
U1191	O
)	O
and	O
body	O
(	O
C1637	O
)	O
in	O
Structures	O
I	O
through	O
V	O
.	O
(	O
f	O
and	O
g	O
)	O
Close	O
-	O
up	O
view	O
of	O
rearrangements	O
in	O
the	O
A	O
and	O
P	O
sites	O
from	O
the	O
initiation	O
state	O
(	O
INIT	O
:	O
PDB	O
ID	O
3J6Y	O
)	O
to	O
the	O
post	O
-	O
translocation	O
Structure	O
V	O
.	O
The	O
fragment	O
shown	O
within	O
a	O
rectangle	O
in	O
panel	O
f	O
is	O
magnified	O
in	O
panel	O
g	O
.	O
Nucleotides	O
of	O
the	O
40S	O
body	O
are	O
shown	O
in	O
orange	O
,	O
40S	O
head	O
in	O
yellow	O
.	O
	O
Our	O
structures	O
represent	O
hitherto	O
uncharacterized	O
translocation	O
complexes	O
of	O
the	O
TSV	O
IRES	O
captured	O
within	O
globally	O
distinct	O
80S	O
conformations	O
(	O
Figures	O
1b	O
and	O
2	O
).	O
	O
We	O
numbered	O
the	O
structures	O
from	O
I	O
to	O
V	O
,	O
according	O
to	O
the	O
position	O
of	O
the	O
tRNA	O
-	O
mRNA	O
-	O
like	O
PKI	O
on	O
the	O
40S	O
subunit	O
(	O
Figure	O
2	O
—	O
source	O
data	O
1	O
).	O
	O
Specifically	O
,	O
PKI	O
is	O
partially	O
withdrawn	O
from	O
the	O
A	O
site	O
in	O
Structure	O
I	O
,	O
and	O
fully	O
translocated	O
to	O
the	O
P	O
site	O
in	O
Structure	O
V	O
(	O
Figure	O
4	O
;	O
see	O
also	O
Figure	O
3	O
—	O
figure	O
supplement	O
1	O
).	O
	O
Thus	O
Structures	O
I	O
to	O
IV	O
represent	O
different	O
positions	O
of	O
PKI	O
between	O
the	O
A	O
and	O
P	O
sites	O
(	O
Figure	O
2	O
—	O
source	O
data	O
1	O
),	O
suggesting	O
that	O
these	O
structures	O
describe	O
intermediate	O
states	O
of	O
translocation	O
.	O
	O
Structure	O
V	O
corresponds	O
to	O
the	O
post	O
-	O
translocation	O
state	O
.	O
	O
Changes	O
in	O
ribosome	O
conformation	O
and	O
eEF2	O
positions	O
are	O
coupled	O
with	O
IRES	O
movement	O
through	O
the	O
ribosome	O
	O
Using	O
the	O
post	O
-	O
translocation	O
S	O
.	O
cerevisiae	O
80S	O
ribosome	O
bound	O
with	O
the	O
P	O
and	O
E	O
site	O
tRNAs	O
as	O
a	O
reference	O
(	O
80S	O
•	O
2tRNA	O
•	O
mRNA	O
),	O
in	O
which	O
both	O
the	O
subunit	O
rotation	O
and	O
the	O
head	O
-	O
body	O
swivel	O
are	O
0	O
°,	O
we	O
found	O
that	O
the	O
ribosome	O
adopts	O
four	O
globally	O
distinct	O
conformations	O
in	O
Structures	O
I	O
through	O
V	O
(	O
Figure	O
1b	O
;	O
see	O
also	O
Figure	O
1	O
—	O
figure	O
supplement	O
1	O
and	O
Figure	O
2	O
—	O
source	O
data	O
1	O
).	O
	O
Structure	O
I	O
comprises	O
the	O
most	O
rotated	O
ribosome	O
conformation	O
(~	O
10	O
°),	O
characteristic	O
of	O
pre	O
-	O
translocation	O
hybrid	O
-	O
tRNA	O
states	O
.	O
	O
From	O
Structure	O
I	O
to	O
V	O
,	O
the	O
body	O
of	O
the	O
small	O
subunit	O
undergoes	O
backward	O
(	O
reverse	O
)	O
rotation	O
(	O
Figure	O
2b	O
;	O
see	O
also	O
Figure	O
1	O
—	O
figure	O
supplement	O
2	O
and	O
Figure	O
2	O
—	O
figure	O
supplement	O
1	O
).	O
	O
Structure	O
V	O
is	O
in	O
a	O
nearly	O
non	O
-	O
rotated	O
conformation	O
(	O
0	O
.	O
5	O
°),	O
very	O
similar	O
to	O
that	O
of	O
post	O
-	O
translocation	O
ribosome	O
-	O
tRNA	O
complexes	O
.	O
	O
Thus	O
,	O
intersubunit	O
rotation	O
of	O
~	O
9	O
°	O
from	O
Structure	O
I	O
to	O
V	O
covers	O
a	O
nearly	O
complete	O
range	O
of	O
relative	O
subunit	O
positions	O
,	O
similar	O
to	O
what	O
was	O
reported	O
for	O
tRNA	O
-	O
bound	O
yeast	O
,	O
bacterial	O
and	O
mammalian	O
ribosomes	O
.	O
	O
The	O
pattern	O
of	O
40S	O
head	O
swivel	O
between	O
the	O
structures	O
is	O
different	O
from	O
that	O
of	O
intersubunit	O
rotation	O
(	O
Figures	O
2c	O
and	O
d	O
;	O
see	O
also	O
Figure	O
2	O
—	O
source	O
data	O
1	O
).	O
	O
As	O
with	O
the	O
intersubunit	O
rotation	O
,	O
the	O
small	O
head	O
swivel	O
(~	O
1	O
°)	O
in	O
the	O
non	O
-	O
rotated	O
Structure	O
V	O
is	O
closest	O
to	O
that	O
in	O
the	O
80S	O
•	O
2tRNA	O
•	O
mRNA	O
post	O
-	O
translocation	O
ribosome	O
.	O
	O
However	O
in	O
the	O
pre	O
-	O
translocation	O
intermediates	O
(	O
from	O
Structure	O
I	O
to	O
IV	O
),	O
the	O
beak	O
of	O
the	O
head	O
domain	O
first	O
turns	O
toward	O
the	O
large	O
subunit	O
and	O
then	O
backs	O
off	O
(	O
Figure	O
2	O
—	O
figure	O
supplement	O
1	O
).	O
	O
The	O
head	O
samples	O
a	O
mid	O
-	O
swiveled	O
position	O
in	O
Structure	O
I	O
(	O
12	O
°),	O
then	O
a	O
highly	O
-	O
swiveled	O
position	O
in	O
Structures	O
II	O
and	O
III	O
(	O
17	O
°)	O
and	O
a	O
less	O
swiveled	O
position	O
in	O
Structure	O
IV	O
(	O
14	O
°).	O
	O
The	O
maximum	O
head	O
swivel	O
is	O
observed	O
in	O
the	O
mid	O
-	O
rotated	O
complexes	O
II	O
and	O
III	O
,	O
in	O
which	O
PKI	O
transitions	O
from	O
the	O
A	O
to	O
P	O
site	O
,	O
while	O
eEF2	O
occupies	O
the	O
A	O
site	O
partially	O
.	O
	O
By	O
comparison	O
,	O
the	O
similarly	O
mid	O
-	O
rotated	O
(	O
4	O
°)	O
80S	O
•	O
TSV	O
IRES	O
initiation	O
complex	O
,	O
in	O
the	O
absence	O
of	O
eEF2	O
,	O
adopts	O
a	O
mid	O
-	O
swiveled	O
position	O
(~	O
10	O
°)	O
(	O
Figure	O
2c	O
).	O
	O
These	O
observations	O
suggest	O
that	O
eEF2	O
is	O
necessary	O
for	O
inducing	O
or	O
stabilizing	O
the	O
large	O
head	O
swivel	O
of	O
the	O
40S	O
subunit	O
characteristic	O
for	O
IRES	O
translocation	O
intermediates	O
.	O
	O
(	O
a	O
)	O
Positions	O
of	O
the	O
IRES	O
and	O
eEF2	O
in	O
the	O
initiation	O
,	O
pre	O
-	O
translocation	O
(	O
I	O
)	O
and	O
post	O
-	O
translocation	O
(	O
V	O
)	O
states	O
,	O
relative	O
to	O
the	O
body	O
of	O
the	O
40S	O
subunit	O
(	O
not	O
shown	O
)	O
(	O
b	O
)	O
Positions	O
of	O
the	O
IRES	O
and	O
eEF2	O
in	O
the	O
initiation	O
state	O
(	O
INIT	O
)	O
and	O
intermediate	O
steps	O
of	O
translocation	O
(	O
II	O
,	O
III	O
and	O
IV	O
),	O
relative	O
to	O
the	O
body	O
of	O
the	O
40S	O
subunit	O
(	O
not	O
shown	O
).	O
	O
Superpositions	O
were	O
obtained	O
by	O
structural	O
alignments	O
of	O
the	O
18S	O
rRNAs	O
excluding	O
the	O
head	O
domains	O
(	O
nt	O
1150	O
–	O
1620	O
).	O
	O
Positions	O
of	O
the	O
IRES	O
relative	O
to	O
proteins	O
uS7	O
,	O
uS11	O
and	O
eS25	O
.	O
	O
(	O
a	O
)	O
Intra	O
-	O
IRES	O
rearrangements	O
from	O
the	O
80S	O
*	O
IRES	O
initiation	O
structure	O
(	O
INIT	O
;	O
PDB	O
3J6Y	O
,)	O
to	O
Structures	O
I	O
through	O
V	O
.	O
For	O
each	O
structure	O
(	O
shown	O
in	O
red	O
),	O
the	O
conformation	O
from	O
a	O
preceding	O
structure	O
is	O
shown	O
in	O
light	O
red	O
for	O
comparison	O
.	O
	O
Superpositions	O
were	O
obtained	O
by	O
structural	O
alignments	O
of	O
18S	O
rRNA	O
.	O
	O
(	O
b	O
)	O
Positions	O
of	O
the	O
IRES	O
and	O
eEF2	O
relative	O
to	O
those	O
of	O
classical	O
P	O
-	O
and	O
E	O
-	O
site	O
tRNAs	O
in	O
the	O
80S	O
•	O
tRNA	O
complex	O
.	O
(	O
c	O
)	O
Positions	O
of	O
the	O
IRES	O
relative	O
to	O
proteins	O
uS11	O
(	O
40S	O
platform	O
)	O
and	O
uS7	O
and	O
eS25	O
(	O
40S	O
head	O
),	O
which	O
interact	O
with	O
the	O
5	O
′	O
domain	O
of	O
the	O
IRES	O
in	O
the	O
initiation	O
state	O
(	O
left	O
panel	O
).	O
	O
Positions	O
of	O
the	O
L1stalk	O
,	O
tRNA	O
and	O
TSV	O
IRES	O
relative	O
to	O
proteins	O
uS7	O
and	O
eS25	O
,	O
in	O
80S	O
•	O
tRNA	O
structures	O
and	O
80S	O
•	O
IRES	O
structures	O
I	O
and	O
V	O
(	O
this	O
work	O
).	O
	O
The	O
view	O
shows	O
the	O
vicinity	O
of	O
the	O
ribosomal	O
E	O
site	O
.	O
	O
Interactions	O
of	O
the	O
stem	O
loops	O
4	O
and	O
5	O
of	O
the	O
TSV	O
with	O
proteins	O
uS7	O
and	O
eS25	O
.	O
	O
2668	O
–	O
2687	O
)	O
and	O
protein	O
uL5	O
(	O
collectively	O
labeled	O
as	O
central	O
protuberance	O
,	O
CP	O
,	O
in	O
the	O
upper	O
-	O
row	O
first	O
figure	O
,	O
and	O
individually	O
labeled	O
in	O
the	O
lower	O
-	O
row	O
first	O
figure	O
).	O
	O
Structures	O
of	O
80S	O
•	O
IRES	O
complexes	O
in	O
the	O
absence	O
of	O
eEF2	O
(	O
INIT	O
;	O
PDB	O
3J6Y	O
,)	O
and	O
in	O
the	O
presence	O
of	O
eEF2	O
(	O
this	O
work	O
)	O
are	O
shown	O
in	O
the	O
upper	O
row	O
and	O
labeled	O
.	O
	O
Structures	O
of	O
the	O
80S	O
complexes	O
with	O
tRNAs	O
are	O
shown	O
in	O
the	O
lower	O
row	O
in	O
a	O
view	O
similar	O
to	O
that	O
for	O
the	O
80S	O
•	O
IRES	O
complex	O
.	O
	O
(	O
a	O
)	O
Secondary	O
structure	O
of	O
the	O
TSV	O
IRES	O
.	O
	O
The	O
TSV	O
IRES	O
comprises	O
two	O
domains	O
:	O
the	O
5	O
'	O
domain	O
(	O
blue	O
)	O
and	O
the	O
PKI	O
domain	O
(	O
red	O
).	O
	O
The	O
open	O
reading	O
frame	O
(	O
gray	O
)	O
is	O
immediately	O
following	O
pseudoknot	O
I	O
(	O
PKI	O
).	O
	O
(	O
b	O
)	O
Three	O
-	O
dimensional	O
structure	O
of	O
the	O
TSV	O
IRES	O
(	O
Structure	O
II	O
).	O
	O
(	O
c	O
)	O
Positions	O
of	O
the	O
IRES	O
and	O
eEF2	O
on	O
the	O
small	O
subunit	O
in	O
Structures	O
I	O
to	O
V	O
.	O
The	O
initiation	O
-	O
state	O
IRES	O
is	O
shown	O
in	O
gray	O
.	O
	O
The	O
insert	O
shows	O
density	O
for	O
interaction	O
of	O
diphthamide	O
699	O
(	O
eEF2	O
;	O
green	O
)	O
with	O
the	O
codon	O
-	O
anticodon	O
-	O
like	O
helix	O
(	O
PKI	O
;	O
red	O
)	O
in	O
Structure	O
V	O
.	O
(	O
d	O
and	O
e	O
)	O
Density	O
of	O
the	O
P	O
site	O
in	O
Structure	O
V	O
shows	O
that	O
interactions	O
of	O
PKI	O
with	O
the	O
18S	O
rRNA	O
nucleotides	O
(	O
c	O
)	O
are	O
nearly	O
identical	O
to	O
those	O
in	O
the	O
P	O
site	O
of	O
the	O
2tRNA	O
•	O
mRNA	O
-	O
bound	O
70S	O
ribosome	O
(	O
d	O
).	O
	O
In	O
each	O
structure	O
,	O
the	O
TSV	O
IRES	O
adopts	O
a	O
distinct	O
conformation	O
in	O
the	O
intersubunit	O
space	O
of	O
the	O
ribosome	O
(	O
Figures	O
3	O
and	O
4	O
).	O
	O
The	O
IRES	O
(	O
nt	O
6758	O
–	O
6952	O
)	O
consists	O
of	O
two	O
globular	O
parts	O
(	O
Figure	O
3a	O
):	O
the	O
5	O
’-	O
region	O
(	O
domains	O
I	O
and	O
II	O
,	O
nt	O
6758	O
–	O
6888	O
)	O
and	O
the	O
PKI	O
domain	O
(	O
domain	O
III	O
,	O
nt	O
6889	O
–	O
6952	O
).	O
	O
We	O
collectively	O
term	O
domains	O
I	O
and	O
II	O
the	O
5	O
’	O
domain	O
.	O
	O
The	O
PKI	O
domain	O
comprises	O
PKI	O
and	O
stem	O
loop	O
3	O
(	O
SL3	O
),	O
which	O
stacks	O
on	O
top	O
of	O
the	O
stem	O
of	O
PKI	O
.	O
	O
The	O
6953GCU	O
triplet	O
immediately	O
following	O
the	O
PKI	O
domain	O
is	O
the	O
first	O
codon	O
of	O
the	O
open	O
reading	O
frame	O
.	O
	O
In	O
the	O
eEF2	O
-	O
free	O
80S	O
•	O
IRES	O
initiation	O
complex	O
(	O
INIT	O
),	O
the	O
bulk	O
of	O
the	O
5	O
’-	O
domain	O
(	O
nt	O
.	O
	O
6758	O
–	O
6888	O
)	O
binds	O
near	O
the	O
E	O
site	O
,	O
contacting	O
the	O
ribosome	O
mostly	O
by	O
means	O
of	O
three	O
protruding	O
structural	O
elements	O
:	O
the	O
L1	O
.	O
1	O
region	O
and	O
stem	O
loops	O
4	O
and	O
5	O
(	O
SL4	O
and	O
SL5	O
).	O
	O
In	O
Structures	O
I	O
to	O
IV	O
,	O
these	O
contacts	O
remain	O
as	O
in	O
the	O
initiation	O
complex	O
(	O
Figure	O
1a	O
).	O
	O
Specifically	O
,	O
the	O
L1	O
.	O
1	O
region	O
interacts	O
with	O
the	O
L1	O
stalk	O
of	O
the	O
large	O
subunit	O
,	O
while	O
SL4	O
and	O
SL5	O
bind	O
at	O
the	O
side	O
of	O
the	O
40S	O
head	O
and	O
interact	O
with	O
proteins	O
uS7	O
,	O
uS11	O
and	O
eS25	O
(	O
Figure	O
3	O
—	O
figure	O
supplement	O
2	O
and	O
Figure	O
3	O
—	O
figure	O
supplement	O
3	O
;	O
ribosomal	O
proteins	O
are	O
termed	O
according	O
to	O
).	O
	O
In	O
Structures	O
I	O
-	O
IV	O
,	O
the	O
minor	O
groove	O
of	O
SL4	O
(	O
at	O
nt	O
6840	O
–	O
6846	O
)	O
binds	O
next	O
to	O
an	O
α	O
-	O
helix	O
of	O
uS7	O
,	O
which	O
is	O
rich	O
in	O
positively	O
charged	O
residues	O
(	O
K212	O
,	O
K213	O
,	O
R219	O
and	O
K222	O
).	O
	O
The	O
tip	O
of	O
SL4	O
binds	O
in	O
the	O
vicinity	O
of	O
R157	O
in	O
the	O
β	O
-	O
hairpin	O
of	O
uS7	O
and	O
of	O
Y58	O
in	O
uS11	O
.	O
	O
In	O
Structure	O
V	O
,	O
however	O
,	O
the	O
density	O
for	O
SL5	O
is	O
missing	O
suggesting	O
that	O
SL5	O
is	O
mobile	O
,	O
while	O
weak	O
SL4	O
density	O
suggests	O
that	O
SL4	O
is	O
shifted	O
along	O
the	O
surface	O
of	O
uS7	O
,	O
~	O
20	O
Å	O
away	O
from	O
its	O
initial	O
position	O
(	O
Figure	O
3	O
—	O
figure	O
supplement	O
2c	O
).	O
	O
Inchworm	O
-	O
like	O
translocation	O
of	O
the	O
TSV	O
IRES	O
.	O
	O
Conformations	O
and	O
positions	O
of	O
the	O
IRES	O
in	O
the	O
initiation	O
state	O
and	O
in	O
Structures	O
I	O
-	O
V	O
are	O
shown	O
relative	O
to	O
those	O
of	O
the	O
A	O
-,	O
P	O
-	O
and	O
E	O
-	O
site	O
tRNAs	O
.	O
	O
Distances	O
between	O
nucleotides	O
6848	O
and	O
6913	O
in	O
SL4	O
and	O
PKI	O
,	O
respectively	O
,	O
are	O
shown	O
(	O
see	O
also	O
Figure	O
2	O
—	O
source	O
data	O
1	O
).	O
	O
The	O
shape	O
of	O
the	O
IRES	O
changes	O
considerably	O
from	O
the	O
initiation	O
state	O
to	O
Structures	O
I	O
through	O
V	O
,	O
from	O
an	O
extended	O
to	O
compact	O
to	O
extended	O
conformation	O
(	O
Figure	O
4	O
;	O
see	O
also	O
Figure	O
3	O
—	O
figure	O
supplement	O
2a	O
).	O
	O
Because	O
in	O
Structures	O
I	O
to	O
IV	O
the	O
PKI	O
domain	O
shifts	O
toward	O
the	O
P	O
site	O
,	O
while	O
the	O
5	O
’	O
remains	O
unchanged	O
near	O
the	O
E	O
site	O
,	O
the	O
distance	O
between	O
the	O
domains	O
shortens	O
(	O
Figure	O
4	O
).	O
	O
In	O
the	O
80S	O
•	O
IRES	O
initiation	O
state	O
,	O
the	O
A	O
-	O
site	O
-	O
bound	O
PKI	O
is	O
separated	O
from	O
SL4	O
by	O
almost	O
50	O
Å	O
(	O
Figure	O
4	O
).	O
	O
In	O
Structures	O
I	O
and	O
II	O
,	O
the	O
PKI	O
is	O
partially	O
retracted	O
from	O
the	O
A	O
site	O
and	O
the	O
distance	O
from	O
SL4	O
shortens	O
to	O
~	O
35	O
Å	O
.	O
As	O
PKI	O
moves	O
toward	O
the	O
P	O
site	O
in	O
Structures	O
III	O
and	O
IV	O
,	O
the	O
PKI	O
domain	O
approaches	O
to	O
within	O
~	O
25	O
Å	O
of	O
SL4	O
.	O
	O
Because	O
the	O
5	O
’-	O
domain	O
in	O
the	O
following	O
structure	O
(	O
V	O
)	O
moves	O
by	O
~	O
20	O
Å	O
along	O
the	O
40S	O
head	O
,	O
the	O
IRES	O
returns	O
to	O
an	O
extended	O
conformation	O
(~	O
45	O
Å	O
)	O
that	O
is	O
similar	O
to	O
that	O
in	O
the	O
80S	O
•	O
IRES	O
initiation	O
complex	O
.	O
	O
2668	O
–	O
2687	O
)	O
and	O
protein	O
uL5	O
(	O
Figure	O
3	O
—	O
figure	O
supplement	O
6	O
).	O
	O
This	O
position	O
of	O
SL3	O
is	O
~	O
25	O
Å	O
away	O
from	O
that	O
in	O
the	O
80S	O
•	O
IRES	O
initiation	O
state	O
,	O
in	O
which	O
PKI	O
and	O
SL3	O
closely	O
mimic	O
the	O
ASL	O
and	O
elbow	O
of	O
the	O
A	O
-	O
site	O
tRNA	O
,	O
respectively	O
.	O
	O
In	O
the	O
highly	O
bent	O
Structures	O
III	O
and	O
IV	O
,	O
the	O
hinge	O
region	O
interacts	O
with	O
the	O
universally	O
conserved	O
uL5	O
and	O
the	O
C	O
-	O
terminal	O
tail	O
of	O
eL42	O
(	O
Figure	O
3	O
—	O
figure	O
supplement	O
7	O
).	O
	O
However	O
,	O
in	O
the	O
extended	O
conformations	O
,	O
these	O
parts	O
of	O
the	O
IRES	O
and	O
the	O
60S	O
subunit	O
are	O
separated	O
by	O
more	O
than	O
10	O
Å	O
,	O
suggesting	O
that	O
an	O
interaction	O
between	O
them	O
stabilizes	O
the	O
bent	O
conformations	O
but	O
not	O
the	O
extended	O
ones	O
.	O
	O
This	O
loop	O
is	O
poorly	O
resolved	O
in	O
Structures	O
I	O
through	O
IV	O
,	O
suggesting	O
conformational	O
flexibility	O
in	O
agreement	O
with	O
structural	O
studies	O
of	O
the	O
isolated	O
PKI	O
and	O
biochemical	O
studies	O
of	O
unbound	O
IRESs	O
.	O
	O
In	O
Structure	O
V	O
,	O
loop	O
3	O
is	O
bound	O
in	O
the	O
40S	O
E	O
site	O
and	O
the	O
backbone	O
of	O
loop	O
3	O
near	O
the	O
codon	O
-	O
like	O
part	O
of	O
PKI	O
(	O
at	O
nt	O
.	O
	O
6945	O
–	O
6946	O
)	O
interacts	O
with	O
R148	O
and	O
R157	O
in	O
β	O
-	O
hairpin	O
of	O
uS7	O
.	O
	O
This	O
interpretation	O
is	O
consistent	O
with	O
the	O
recent	O
observation	O
that	O
alterations	O
in	O
loop	O
3	O
of	O
the	O
CrPV	O
IRES	O
result	O
in	O
decreased	O
efficiency	O
of	O
translocation	O
.	O
	O
eEF2	O
structures	O
	O
Elements	O
of	O
the	O
80S	O
ribosome	O
that	O
contact	O
eEF2	O
in	O
Structures	O
I	O
through	O
V	O
.	O
	O
The	O
switch	O
loop	O
I	O
in	O
Structure	O
I	O
is	O
shown	O
in	O
blue	O
.	O
	O
The	O
putative	O
position	O
of	O
the	O
switch	O
loop	O
I	O
,	O
unresolved	O
in	O
the	O
density	O
of	O
Structure	O
II	O
,	O
is	O
shown	O
with	O
a	O
dashed	O
line	O
.	O
	O
Colors	O
for	O
the	O
ribosome	O
and	O
eEF2	O
are	O
as	O
in	O
Figure	O
1	O
.	O
	O
Conformations	O
and	O
interactions	O
of	O
eEF2	O
.	O
	O
(	O
a	O
)	O
Conformations	O
of	O
eEF2	O
in	O
Structures	O
I	O
-	O
V	O
and	O
domain	O
organization	O
of	O
eEF2	O
are	O
shown	O
.	O
	O
Roman	O
numerals	O
denote	O
eEF2	O
domains	O
.	O
	O
Superposition	O
was	O
obtained	O
by	O
structural	O
alignment	O
of	O
domains	O
I	O
and	O
II	O
.	O
	O
(	O
b	O
)	O
Elements	O
of	O
the	O
80S	O
ribosome	O
in	O
Structures	O
I	O
and	O
V	O
that	O
contact	O
eEF2	O
.	O
	O
eEF2	O
is	O
shown	O
in	O
green	O
,	O
IRES	O
RNA	O
in	O
red	O
,	O
40S	O
subunit	O
elements	O
in	O
orange	O
,	O
60S	O
in	O
cyan	O
/	O
teal	O
.	O
	O
(	O
d	O
)	O
Interactions	O
of	O
the	O
GTPase	O
domains	O
with	O
the	O
40S	O
and	O
60S	O
subunits	O
in	O
Structure	O
I	O
(	O
colored	O
in	O
green	O
/	O
blue	O
,	O
eEF2	O
;	O
orange	O
,	O
40S	O
;	O
cyan	O
/	O
teal	O
,	O
60S	O
)	O
and	O
in	O
Structure	O
II	O
(	O
gray	O
).	O
	O
Switch	O
loop	O
I	O
(	O
SWI	O
)	O
in	O
Structure	O
I	O
is	O
in	O
blue	O
;	O
dashed	O
line	O
shows	O
the	O
putative	O
location	O
of	O
unresolved	O
switch	O
loop	O
I	O
in	O
Structure	O
II	O
.	O
	O
(	O
h	O
)	O
Cryo	O
-	O
EM	O
density	O
showing	O
the	O
sordarin	O
-	O
binding	O
pocket	O
of	O
eEF2	O
(	O
Structure	O
II	O
).	O
	O
Sordarin	O
is	O
shown	O
in	O
pink	O
with	O
oxygen	O
atoms	O
in	O
red	O
.	O
	O
Elongation	O
factor	O
eEF2	O
in	O
all	O
five	O
structures	O
is	O
bound	O
with	O
GDP	O
and	O
sordarin	O
(	O
Figure	O
5	O
).	O
	O
The	O
elongation	O
factor	O
consists	O
of	O
three	O
dynamic	O
superdomains	O
:	O
an	O
N	O
-	O
terminal	O
globular	O
superdomain	O
formed	O
by	O
the	O
G	O
(	O
GTPase	O
)	O
domain	O
(	O
domain	O
I	O
)	O
and	O
domain	O
II	O
;	O
a	O
linker	O
domain	O
III	O
;	O
and	O
a	O
C	O
-	O
terminal	O
superdomain	O
comprising	O
domains	O
IV	O
and	O
V	O
(	O
Figure	O
5a	O
).	O
	O
Domain	O
IV	O
extends	O
from	O
the	O
main	O
body	O
and	O
is	O
critical	O
for	O
translocation	O
catalyzed	O
by	O
eEF2	O
or	O
EF	O
-	O
G	O
.	O
ADP	O
-	O
ribosylation	O
of	O
eEF2	O
at	O
the	O
tip	O
of	O
domain	O
IV	O
or	O
deletion	O
of	O
domain	O
IV	O
from	O
EF	O
-	O
G	O
abrogate	O
translocation	O
.	O
	O
In	O
post	O
-	O
translocation	O
-	O
like	O
80S	O
•	O
tRNA	O
•	O
eEF2	O
complexes	O
,	O
domain	O
IV	O
binds	O
in	O
the	O
40S	O
A	O
site	O
,	O
suggesting	O
direct	O
involvement	O
of	O
domain	O
IV	O
in	O
translocation	O
of	O
tRNA	O
from	O
the	O
A	O
to	O
P	O
site	O
.	O
	O
GDP	O
in	O
our	O
structures	O
is	O
bound	O
in	O
the	O
GTPase	O
center	O
(	O
Figures	O
5d	O
,	O
e	O
and	O
f	O
)	O
and	O
sordarin	O
is	O
sandwiched	O
between	O
the	O
β	O
-	O
platforms	O
of	O
domains	O
III	O
and	O
V	O
(	O
Figures	O
5g	O
and	O
h	O
),	O
as	O
in	O
the	O
structure	O
of	O
free	O
eEF2	O
•	O
sordarin	O
complex	O
.	O
	O
The	O
global	O
conformations	O
of	O
eEF2	O
(	O
Figure	O
5a	O
)	O
are	O
similar	O
in	O
these	O
structures	O
(	O
all	O
-	O
atom	O
RMSD	O
≤	O
2	O
Å	O
),	O
but	O
the	O
positions	O
of	O
eEF2	O
relative	O
to	O
the	O
40S	O
subunit	O
differ	O
substantially	O
as	O
a	O
result	O
of	O
40S	O
subunit	O
rotation	O
(	O
Figure	O
2	O
—	O
source	O
data	O
1	O
).	O
	O
From	O
Structure	O
I	O
to	O
V	O
,	O
eEF2	O
is	O
rigidly	O
attached	O
to	O
the	O
GTPase	O
-	O
associated	O
center	O
of	O
the	O
60S	O
subunit	O
.	O
	O
The	O
GTPase	O
-	O
associated	O
center	O
comprises	O
the	O
P	O
stalk	O
(	O
L11	O
and	O
L7	O
/	O
L12	O
stalk	O
in	O
bacteria	O
)	O
and	O
the	O
sarcin	O
-	O
ricin	O
loop	O
(	O
SRL	O
,	O
nt	O
3012	O
–	O
3042	O
).	O
	O
The	O
tips	O
of	O
25S	O
rRNA	O
helices	O
43	O
and	O
44	O
of	O
the	O
P	O
stalk	O
(	O
nucleotides	O
G1242	O
and	O
A1270	O
,	O
respectively	O
)	O
stack	O
on	O
V754	O
and	O
Y744	O
of	O
domain	O
V	O
.	O
An	O
αββ	O
motif	O
of	O
the	O
eukaryote	O
-	O
specific	O
protein	O
P0	O
(	O
aa	O
126	O
–	O
154	O
)	O
packs	O
in	O
the	O
crevice	O
between	O
the	O
long	O
α	O
-	O
helix	O
D	O
(	O
aa	O
172	O
–	O
188	O
)	O
of	O
the	O
GTPase	O
domain	O
and	O
the	O
β	O
-	O
sheet	O
region	O
(	O
aa	O
246	O
–	O
263	O
)	O
of	O
the	O
GTPase	O
domain	O
insert	O
(	O
or	O
G	O
’	O
insert	O
)	O
of	O
eEF2	O
(	O
secondary	O
-	O
structure	O
nomenclatures	O
for	O
eEF2	O
and	O
EF	O
-	O
G	O
are	O
the	O
same	O
).	O
	O
Although	O
the	O
P	O
/	O
L11	O
stalk	O
is	O
known	O
to	O
be	O
dynamic	O
,	O
its	O
position	O
remains	O
unchanged	O
from	O
Structure	O
I	O
to	O
V	O
:	O
all	O
-	O
atom	O
root	O
-	O
mean	O
-	O
square	O
differences	O
for	O
the	O
25S	O
rRNA	O
of	O
the	O
P	O
stalk	O
(	O
nt	O
1223	O
–	O
1286	O
)	O
are	O
within	O
2	O
.	O
5	O
Å	O
.	O
However	O
,	O
with	O
respect	O
to	O
its	O
position	O
in	O
the	O
80S	O
•	O
IRES	O
complex	O
in	O
the	O
absence	O
of	O
eEF2	O
and	O
in	O
the	O
80S	O
•	O
2tRNA	O
•	O
mRNA	O
complex	O
,	O
the	O
P	O
stalk	O
is	O
shifted	O
by	O
~	O
13	O
Å	O
toward	O
the	O
A	O
site	O
(	O
Figure	O
2d	O
).	O
	O
While	O
the	O
overall	O
mode	O
of	O
this	O
interaction	O
is	O
similar	O
to	O
that	O
seen	O
in	O
70S	O
•	O
EF	O
-	O
G	O
crystal	O
structures	O
,	O
there	O
is	O
an	O
important	O
local	O
difference	O
between	O
Structure	O
I	O
and	O
Structures	O
II	O
-	O
V	O
in	O
switch	O
loop	O
I	O
,	O
as	O
discussed	O
below	O
.	O
	O
Structures	O
I	O
through	O
V	O
are	O
shown	O
.	O
	O
Electrostatic	O
surface	O
of	O
eEF2	O
is	O
shown	O
;	O
negatively	O
and	O
positively	O
charged	O
regions	O
are	O
shown	O
in	O
red	O
and	O
blue	O
,	O
respectively	O
.	O
	O
The	O
view	O
was	O
obtained	O
by	O
structural	O
alignment	O
of	O
the	O
18S	O
rRNAs	O
.	O
	O
The	O
view	O
was	O
obtained	O
by	O
superpositions	O
of	O
the	O
body	O
domains	O
of	O
18S	O
rRNAs	O
.	O
	O
(	O
c	O
)	O
Rearrangements	O
,	O
from	O
Structure	O
I	O
through	O
V	O
,	O
of	O
a	O
positively	O
charged	O
cluster	O
of	O
eEF2	O
(	O
K613	O
,	O
R617	O
and	O
R631	O
)	O
positioned	O
over	O
the	O
phosphate	O
backbone	O
of	O
18S	O
helices	O
33	O
and	O
34	O
,	O
suggesting	O
a	O
role	O
of	O
electrostatic	O
interactions	O
in	O
eEF2	O
diffusion	O
over	O
the	O
40S	O
surface	O
.	O
	O
(	O
d	O
)	O
Shift	O
of	O
the	O
tip	O
of	O
domain	O
III	O
of	O
eEF2	O
,	O
interacting	O
with	O
uS12	O
upon	O
reverse	O
subunit	O
rotation	O
from	O
Structure	O
I	O
to	O
Structure	O
V	O
.	O
Structure	O
I	O
colored	O
as	O
in	O
Figure	O
1	O
,	O
except	O
uS12	O
,	O
which	O
is	O
in	O
purple	O
;	O
Structure	O
V	O
is	O
in	O
gray	O
.	O
	O
There	O
are	O
two	O
modest	O
but	O
noticeable	O
domain	O
rearrangements	O
between	O
Structures	O
I	O
and	O
V	O
.	O
Unlike	O
in	O
free	O
eEF2	O
,	O
which	O
can	O
sample	O
large	O
movements	O
of	O
at	O
least	O
50	O
Å	O
of	O
the	O
C	O
-	O
terminal	O
superdomain	O
relative	O
to	O
the	O
N	O
-	O
terminal	O
superdomain	O
(	O
Figure	O
5c	O
),	O
eEF2	O
undergoes	O
moderate	O
repositioning	O
of	O
domain	O
IV	O
(~	O
3	O
Å	O
;	O
Figure	O
5a	O
)	O
and	O
domain	O
III	O
(~	O
5	O
Å	O
;	O
Figure	O
6d	O
).	O
	O
This	O
limited	O
flexibility	O
of	O
the	O
ribosome	O
-	O
bound	O
eEF2	O
is	O
likely	O
the	O
result	O
of	O
simultaneous	O
fixation	O
of	O
eEF2	O
superdomains	O
,	O
via	O
domains	O
I	O
and	O
V	O
,	O
by	O
the	O
GTPase	O
-	O
associated	O
center	O
of	O
the	O
large	O
subunit	O
.	O
	O
Domain	O
IV	O
of	O
eEF2	O
binds	O
at	O
the	O
40S	O
A	O
site	O
in	O
Structures	O
I	O
to	O
V	O
but	O
the	O
mode	O
of	O
interaction	O
differs	O
in	O
each	O
complex	O
(	O
Figure	O
6	O
).	O
	O
eEF2	O
settles	O
into	O
the	O
A	O
site	O
from	O
Structure	O
I	O
to	O
V	O
,	O
as	O
the	O
tip	O
of	O
domain	O
IV	O
shifts	O
by	O
~	O
10	O
Å	O
relative	O
to	O
the	O
body	O
and	O
by	O
~	O
20	O
Å	O
relative	O
to	O
the	O
swiveling	O
head	O
.	O
	O
Modest	O
intra	O
-	O
eEF2	O
shifts	O
of	O
domain	O
IV	O
between	O
Structures	O
I	O
to	O
V	O
outline	O
a	O
stochastic	O
trajectory	O
(	O
Figure	O
5a	O
),	O
consistent	O
with	O
local	O
adjustments	O
of	O
the	O
domain	O
in	O
the	O
A	O
site	O
.	O
	O
At	O
the	O
central	O
region	O
of	O
eEF2	O
,	O
domains	O
II	O
and	O
III	O
contact	O
the	O
40S	O
body	O
(	O
mainly	O
at	O
nucleotides	O
48	O
–	O
52	O
and	O
429	O
–	O
432	O
of	O
18S	O
rRNA	O
helix	O
5	O
and	O
uS12	O
).	O
	O
Comparison	O
of	O
eEF2	O
conformations	O
reveals	O
that	O
in	O
Structure	O
V	O
,	O
domain	O
III	O
is	O
displaced	O
as	O
a	O
result	O
of	O
interaction	O
with	O
uS12	O
,	O
as	O
discussed	O
below	O
.	O
	O
In	O
summary	O
,	O
between	O
Structures	O
I	O
and	O
V	O
,	O
a	O
step	O
-	O
wise	O
translocation	O
of	O
PKI	O
by	O
~	O
15	O
Å	O
from	O
the	O
A	O
to	O
P	O
site	O
-	O
within	O
the	O
40S	O
subunit	O
–	O
occurs	O
simultaneously	O
with	O
the	O
~	O
11	O
Å	O
side	O
-	O
way	O
entry	O
of	O
domain	O
IV	O
into	O
the	O
A	O
site	O
coupled	O
with	O
~	O
3	O
to	O
5	O
Å	O
inter	O
-	O
domain	O
rearrangements	O
in	O
eEF2	O
.	O
	O
These	O
shifts	O
occur	O
during	O
the	O
reverse	O
rotation	O
of	O
the	O
40S	O
body	O
coupled	O
with	O
the	O
forward	O
-	O
then	O
-	O
reverse	O
head	O
swivel	O
.	O
	O
To	O
elucidate	O
the	O
detailed	O
structural	O
mechanism	O
of	O
IRES	O
translocation	O
and	O
the	O
roles	O
of	O
eEF2	O
and	O
ribosome	O
rearrangements	O
,	O
we	O
describe	O
in	O
the	O
following	O
sections	O
the	O
interactions	O
of	O
PKI	O
and	O
eEF2	O
with	O
the	O
ribosomal	O
A	O
and	O
P	O
sites	O
in	O
Structures	O
I	O
through	O
V	O
(	O
Figure	O
2g	O
;	O
see	O
also	O
Figure	O
1	O
—	O
figure	O
supplement	O
1	O
).	O
	O
Structure	O
I	O
represents	O
a	O
pre	O
-	O
translocation	O
IRES	O
and	O
initial	O
entry	O
of	O
eEF2	O
in	O
a	O
GTP	O
-	O
like	O
state	O
	O
In	O
the	O
fully	O
rotated	O
Structure	O
I	O
,	O
PKI	O
is	O
shifted	O
toward	O
the	O
P	O
site	O
by	O
~	O
3	O
Å	O
relative	O
to	O
its	O
position	O
in	O
the	O
initiation	O
complex	O
but	O
maintains	O
interactions	O
with	O
the	O
partially	O
swiveled	O
head	O
.	O
	O
The	O
C1274	O
:	O
G6953	O
base	O
pair	O
provides	O
a	O
stacking	O
platform	O
for	O
the	O
codon	O
-	O
anticodon	O
–	O
like	O
helix	O
of	O
PKI	O
.	O
	O
We	O
therefore	O
define	O
C1274	O
as	O
the	O
foundation	O
of	O
the	O
'	O
head	O
A	O
site	O
'.	O
	O
Accordingly	O
,	O
we	O
use	O
U1191	O
(	O
G966	O
in	O
E	O
.	O
coli	O
)	O
and	O
C1637	O
(	O
C1400	O
in	O
E	O
.	O
coli	O
)	O
as	O
the	O
reference	O
points	O
of	O
the	O
'	O
head	O
P	O
site	O
'	O
and	O
'	O
body	O
P	O
site	O
'	O
(	O
Figure	O
2g	O
),	O
respectively	O
,	O
because	O
these	O
nucleotides	O
form	O
a	O
stacking	O
foundation	O
for	O
the	O
fully	O
translocated	O
mRNA	O
-	O
tRNA	O
helix	O
in	O
tRNA	O
-	O
bound	O
structures	O
and	O
in	O
our	O
post	O
-	O
translocation	O
Structure	O
V	O
discussed	O
below	O
.	O
	O
Interactions	O
of	O
the	O
residues	O
at	O
the	O
eEF2	O
tip	O
with	O
the	O
decoding	O
center	O
of	O
the	O
IRES	O
-	O
bound	O
ribosome	O
.	O
	O
Key	O
elements	O
of	O
the	O
decoding	O
center	O
of	O
the	O
'	O
locked	O
'	O
initiation	O
structure	O
,	O
'	O
unlocked	O
'	O
Structure	O
I	O
,	O
and	O
post	O
-	O
translocation	O
Structure	O
V	O
(	O
this	O
work	O
)	O
are	O
shown	O
.	O
	O
The	O
histidine	O
-	O
diphthamide	O
tip	O
of	O
eEF2	O
is	O
shown	O
in	O
green	O
.	O
	O
Nucleotides	O
of	O
the	O
18S	O
rRNA	O
body	O
are	O
in	O
orange	O
and	O
head	O
in	O
yellow	O
;	O
25S	O
rRNA	O
nucleotide	O
A2256	O
is	O
blue	O
.	O
	O
A	O
and	O
P	O
sites	O
are	O
schematically	O
demarcated	O
by	O
dotted	O
lines	O
.	O
	O
The	O
interaction	O
of	O
PKI	O
with	O
the	O
40S	O
body	O
is	O
substantially	O
rearranged	O
relative	O
to	O
that	O
in	O
the	O
initiation	O
state	O
.	O
	O
In	O
Structure	O
I	O
,	O
PKI	O
does	O
not	O
contact	O
these	O
nucleotides	O
(	O
Figures	O
2g	O
and	O
7	O
).	O
	O
The	O
position	O
of	O
eEF2	O
on	O
the	O
40S	O
subunit	O
of	O
Structure	O
I	O
is	O
markedly	O
distinct	O
from	O
those	O
in	O
Structures	O
II	O
to	O
V	O
.	O
The	O
translocase	O
interacts	O
with	O
the	O
40S	O
body	O
but	O
does	O
not	O
contact	O
the	O
head	O
(	O
Figures	O
5b	O
and	O
6a	O
;	O
Figure	O
5	O
—	O
figure	O
supplement	O
1	O
).	O
	O
The	O
tip	O
of	O
domain	O
IV	O
is	O
wedged	O
between	O
PKI	O
and	O
decoding	O
-	O
center	O
nucleotides	O
A1755	O
and	O
A1756	O
,	O
which	O
are	O
bulged	O
out	O
of	O
h44	O
.	O
	O
This	O
tip	O
contains	O
the	O
histidine	O
-	O
diphthamide	O
triad	O
(	O
H583	O
,	O
H694	O
and	O
Diph699	O
),	O
which	O
interacts	O
with	O
the	O
codon	O
-	O
anticodon	O
-	O
like	O
helix	O
of	O
PKI	O
and	O
A1756	O
(	O
Figure	O
7	O
).	O
	O
Diphthamide	O
is	O
a	O
unique	O
posttranslational	O
modification	O
conserved	O
in	O
archaeal	O
and	O
eukaryotic	O
EF2	O
(	O
at	O
residue	O
699	O
in	O
S	O
.	O
cerevisiae	O
)	O
and	O
involves	O
addition	O
of	O
a	O
~	O
7	O
-	O
Å	O
long	O
3	O
-	O
carboxyamido	O
-	O
3	O
-(	O
trimethylamino	O
)-	O
propyl	O
moiety	O
to	O
the	O
histidine	O
imidazole	O
ring	O
at	O
CE1	O
.	O
	O
The	O
opposite	O
surface	O
of	O
the	O
tail	O
is	O
oriented	O
toward	O
the	O
minor	O
-	O
groove	O
side	O
of	O
the	O
second	O
base	O
pair	O
of	O
the	O
codon	O
-	O
anticodon	O
helix	O
(	O
G6906	O
:	O
C6951	O
).	O
	O
The	O
splitting	O
of	O
the	O
interaction	O
of	O
A1755	O
-	O
A1756	O
and	O
PKI	O
is	O
achieved	O
by	O
providing	O
the	O
histidine	O
-	O
diphthamine	O
tip	O
as	O
a	O
binding	O
partner	O
for	O
both	O
A1756	O
and	O
the	O
minor	O
groove	O
of	O
the	O
codon	O
-	O
anticodon	O
helix	O
(	O
Figure	O
7	O
).	O
	O
Unlike	O
in	O
Structures	O
II	O
to	O
V	O
,	O
the	O
conformation	O
of	O
the	O
eEF2	O
GTPase	O
center	O
in	O
Structure	O
I	O
resembles	O
that	O
of	O
a	O
GTP	O
-	O
bound	O
translocase	O
(	O
Figure	O
5e	O
).	O
	O
Switch	O
loop	O
II	O
(	O
aa	O
105	O
–	O
110	O
),	O
which	O
carries	O
the	O
catalytic	O
H108	O
(	O
H92	O
in	O
E	O
.	O
coli	O
EF	O
-	O
G	O
;	O
is	O
well	O
resolved	O
in	O
all	O
five	O
structures	O
.	O
	O
The	O
N	O
-	O
terminal	O
part	O
of	O
the	O
loop	O
(	O
aa	O
50	O
–	O
60	O
)	O
is	O
sandwiched	O
between	O
the	O
tip	O
of	O
helix	O
14	O
(	O
415CAAA418	O
)	O
of	O
the	O
18S	O
rRNA	O
of	O
the	O
40S	O
subunit	O
and	O
helix	O
A	O
(	O
aa	O
32	O
–	O
42	O
)	O
of	O
eEF2	O
(	O
Figure	O
5d	O
).	O
	O
Bulged	O
A416	O
interacts	O
with	O
the	O
switch	O
loop	O
in	O
the	O
vicinity	O
of	O
D53	O
.	O
	O
In	O
Structure	O
II	O
,	O
relative	O
to	O
Structure	O
I	O
,	O
PKI	O
is	O
further	O
shifted	O
along	O
the	O
40S	O
body	O
,	O
traversing	O
~	O
4	O
Å	O
toward	O
the	O
P	O
site	O
(	O
Figures	O
2e	O
,	O
f	O
,	O
and	O
g	O
),	O
while	O
stacking	O
on	O
C1274	O
at	O
the	O
head	O
A	O
site	O
.	O
	O
Thus	O
,	O
the	O
intermediate	O
position	O
of	O
PKI	O
is	O
possible	O
due	O
to	O
a	O
large	O
swivel	O
of	O
the	O
head	O
relative	O
to	O
the	O
body	O
,	O
which	O
brings	O
the	O
head	O
A	O
site	O
close	O
to	O
the	O
body	O
P	O
site	O
.	O
	O
The	O
head	O
interface	O
of	O
domain	O
IV	O
interacts	O
with	O
the	O
40S	O
head	O
(	O
Figure	O
6a	O
).	O
	O
Here	O
,	O
a	O
positively	O
charged	O
surface	O
of	O
eEF2	O
,	O
formed	O
by	O
K613	O
,	O
R617	O
and	O
R631	O
contacts	O
the	O
phosphate	O
backbone	O
of	O
helix	O
33	O
(	O
Figures	O
6c	O
;	O
see	O
also	O
Figure	O
6	O
—	O
figure	O
supplement	O
1	O
).	O
	O
Consistent	O
with	O
the	O
similar	O
head	O
swivels	O
in	O
Structure	O
III	O
and	O
Structure	O
II	O
,	O
relative	O
positions	O
of	O
the	O
40S	O
head	O
A	O
site	O
and	O
body	O
P	O
site	O
remain	O
as	O
in	O
Structure	O
II	O
.	O
	O
The	O
map	O
allows	O
placement	O
of	O
PKI	O
at	O
the	O
body	O
P	O
site	O
(	O
Figure	O
1	O
—	O
figure	O
supplement	O
3	O
).	O
	O
Thus	O
,	O
in	O
Structure	O
III	O
,	O
PKI	O
has	O
translocated	O
along	O
the	O
40S	O
body	O
,	O
but	O
the	O
head	O
remains	O
fully	O
swiveled	O
so	O
that	O
PKI	O
is	O
between	O
the	O
head	O
A	O
and	O
P	O
sites	O
.	O
	O
Lower	O
resolution	O
of	O
the	O
map	O
in	O
this	O
region	O
suggests	O
that	O
PKI	O
is	O
somewhat	O
destabilized	O
in	O
the	O
vicinity	O
of	O
the	O
body	O
P	O
site	O
in	O
the	O
absence	O
of	O
stacking	O
with	O
the	O
foundations	O
of	O
the	O
head	O
A	O
site	O
(	O
C1274	O
)	O
or	O
P	O
site	O
(	O
U1191	O
).	O
	O
The	O
position	O
of	O
eEF2	O
is	O
similar	O
to	O
that	O
in	O
Structure	O
II	O
.	O
	O
Structure	O
IV	O
represents	O
a	O
highly	O
bent	O
IRES	O
with	O
PKI	O
partially	O
accommodated	O
in	O
the	O
P	O
site	O
	O
Unwinding	O
of	O
the	O
head	O
moves	O
the	O
head	O
P	O
-	O
site	O
residue	O
U1191	O
and	O
body	O
P	O
-	O
site	O
residue	O
C1637	O
closer	O
together	O
,	O
resulting	O
in	O
a	O
partially	O
restored	O
40S	O
P	O
site	O
.	O
	O
Whereas	O
C1637	O
forms	O
a	O
stacking	O
platform	O
for	O
the	O
last	O
base	O
pair	O
of	O
PKI	O
,	O
U1191	O
does	O
not	O
yet	O
stack	O
on	O
PKI	O
because	O
the	O
head	O
remains	O
partially	O
swiveled	O
.	O
	O
This	O
renders	O
PKI	O
partially	O
accommodated	O
in	O
the	O
P	O
site	O
(	O
Figure	O
2g	O
).	O
	O
This	O
results	O
in	O
rearrangements	O
of	O
eEF2	O
interactions	O
with	O
the	O
head	O
,	O
allowing	O
eEF2	O
to	O
advance	O
further	O
into	O
the	O
A	O
site	O
.	O
	O
To	O
this	O
end	O
,	O
the	O
head	O
-	O
interacting	O
interface	O
of	O
domain	O
IV	O
slides	O
along	O
the	O
surface	O
of	O
the	O
head	O
by	O
5	O
Å	O
.	O
Helix	O
A	O
of	O
domain	O
IV	O
is	O
positioned	O
next	O
to	O
the	O
backbone	O
of	O
h34	O
,	O
with	O
positively	O
charged	O
residues	O
K613	O
,	O
R617	O
and	O
R631	O
rearranged	O
from	O
the	O
backbone	O
of	O
h33	O
(	O
Figure	O
6c	O
;	O
see	O
also	O
Figure	O
6	O
—	O
figure	O
supplement	O
1	O
).	O
	O
Structure	O
V	O
represents	O
an	O
extended	O
IRES	O
with	O
PKI	O
fully	O
accommodated	O
in	O
the	O
P	O
site	O
and	O
domain	O
IV	O
of	O
eEF2	O
in	O
the	O
A	O
site	O
	O
In	O
the	O
nearly	O
non	O
-	O
rotated	O
and	O
non	O
-	O
swiveled	O
ribosome	O
conformation	O
in	O
Structure	O
V	O
closely	O
resembling	O
that	O
of	O
the	O
post	O
-	O
translocation	O
80S	O
•	O
2tRNA	O
•	O
mRNA	O
complex	O
,	O
PKI	O
is	O
fully	O
accommodated	O
in	O
the	O
P	O
site	O
.	O
	O
The	O
codon	O
-	O
anticodon	O
–	O
like	O
helix	O
is	O
stacked	O
on	O
P	O
-	O
site	O
residues	O
U1191	O
and	O
C1637	O
(	O
Figure	O
3d	O
),	O
analogous	O
to	O
stacking	O
of	O
the	O
tRNA	O
-	O
mRNA	O
helix	O
(	O
Figure	O
3e	O
).	O
	O
A	O
notable	O
conformational	O
change	O
in	O
eEF2	O
from	O
that	O
in	O
the	O
preceding	O
Structures	O
is	O
visible	O
in	O
the	O
position	O
of	O
domain	O
III	O
,	O
which	O
contacts	O
uS12	O
(	O
Figure	O
6d	O
).	O
	O
In	O
this	O
position	O
,	O
uS12	O
forms	O
extensive	O
interactions	O
with	O
eEF2	O
domains	O
II	O
and	O
III	O
.	O
	O
Specifically	O
,	O
the	O
C	O
-	O
terminal	O
tail	O
of	O
uS12	O
packs	O
against	O
the	O
β	O
-	O
barrel	O
of	O
domain	O
II	O
,	O
while	O
the	O
β	O
-	O
barrel	O
of	O
uS12	O
packs	O
against	O
helix	O
A	O
of	O
domain	O
III	O
.	O
	O
Domain	O
IV	O
of	O
eEF2	O
is	O
fully	O
accommodated	O
in	O
the	O
A	O
site	O
.	O
	O
The	O
first	O
codon	O
of	O
the	O
open	O
reading	O
frame	O
is	O
also	O
positioned	O
in	O
the	O
A	O
site	O
,	O
with	O
bases	O
exposed	O
toward	O
eEF2	O
(	O
Figure	O
7	O
),	O
resembling	O
the	O
conformations	O
of	O
the	O
A	O
-	O
site	O
codons	O
in	O
EF	O
-	O
G	O
-	O
bound	O
70S	O
complexes	O
.	O
	O
As	O
in	O
the	O
preceding	O
Structures	O
,	O
the	O
histidine	O
-	O
diphthamide	O
tip	O
is	O
bound	O
in	O
the	O
minor	O
groove	O
of	O
the	O
P	O
-	O
site	O
codon	O
-	O
anticodon	O
helix	O
.	O
	O
Diph699	O
slightly	O
rearranges	O
,	O
relative	O
to	O
that	O
in	O
Structure	O
I	O
(	O
Figure	O
7	O
),	O
and	O
interacts	O
with	O
four	O
out	O
of	O
six	O
codon	O
-	O
anticodon	O
nucleotides	O
.	O
	O
The	O
amide	O
at	O
the	O
diphthamide	O
end	O
interacts	O
with	O
N2	O
of	O
G6906	O
and	O
O2	O
and	O
O2	O
’	O
of	O
C6951	O
(	O
corresponding	O
to	O
nt	O
2	O
of	O
the	O
codon	O
).	O
	O
The	O
trimethylamino	O
-	O
group	O
is	O
positioned	O
over	O
the	O
ribose	O
of	O
C6952	O
(	O
codon	O
nt	O
3	O
).	O
	O
IRES	O
translocation	O
mechanism	O
	O
Animation	O
showing	O
the	O
transition	O
from	O
the	O
initiation	O
80S	O
•	O
TSV	O
IRES	O
structures	O
(	O
Koh	O
et	O
al	O
.,	O
2014	O
)	O
to	O
eEF2	O
-	O
bound	O
Structures	O
I	O
through	O
V	O
(	O
this	O
work	O
).	O
	O
In	O
scenes	O
1	O
,	O
2	O
and	O
3	O
,	O
nucleotides	O
C1274	O
,	O
U1191	O
of	O
the	O
40S	O
head	O
and	O
G904	O
of	O
the	O
40S	O
platform	O
are	O
shown	O
in	O
black	O
to	O
denote	O
the	O
A	O
,	O
P	O
and	O
E	O
sites	O
,	O
respectively	O
.	O
	O
In	O
scene	O
4	O
,	O
C1274	O
and	O
U1191	O
are	O
labeled	O
and	O
shown	O
in	O
yellow	O
;	O
G577	O
,	O
A1755	O
and	O
A1756	O
of	O
the	O
40S	O
body	O
A	O
site	O
and	O
C1637	O
of	O
the	O
body	O
P	O
site	O
are	O
labeled	O
and	O
shown	O
in	O
orange	O
.	O
	O
In	O
this	O
work	O
we	O
have	O
captured	O
the	O
structures	O
of	O
the	O
TSV	O
IRES	O
,	O
whose	O
PKI	O
samples	O
positions	O
between	O
the	O
A	O
and	O
P	O
sites	O
(	O
Structures	O
I	O
–	O
IV	O
),	O
as	O
well	O
as	O
in	O
the	O
P	O
site	O
(	O
Structure	O
V	O
).	O
	O
We	O
propose	O
that	O
together	O
with	O
the	O
previously	O
reported	O
initiation	O
state	O
,	O
these	O
structures	O
represent	O
the	O
trajectory	O
of	O
eEF2	O
-	O
induced	O
IRES	O
translocation	O
(	O
shown	O
as	O
an	O
animation	O
in	O
http	O
://	O
labs	O
.	O
umassmed	O
.	O
edu	O
/	O
korostelevlab	O
/	O
msc	O
/	O
iresmovie	O
.	O
gif	O
and	O
Video	O
1	O
).	O
	O
Furthermore	O
,	O
they	O
provide	O
insight	O
into	O
the	O
mechanism	O
of	O
eEF2	O
•	O
GTP	O
association	O
with	O
the	O
pre	O
-	O
translocation	O
ribosome	O
and	O
eEF2	O
•	O
GDP	O
dissociation	O
from	O
the	O
post	O
-	O
translocation	O
ribosome	O
,	O
also	O
delineating	O
the	O
mechanism	O
of	O
translation	O
inhibition	O
by	O
the	O
antifungal	O
drug	O
sordarin	O
.	O
	O
In	O
summary	O
,	O
the	O
reported	O
ensemble	O
of	O
structures	O
substantially	O
enhances	O
our	O
understanding	O
of	O
the	O
translocation	O
mechanism	O
,	O
including	O
that	O
of	O
tRNAs	O
as	O
discussed	O
below	O
.	O
	O
Translocation	O
of	O
the	O
TSV	O
IRES	O
on	O
the	O
40S	O
subunit	O
globally	O
resembles	O
a	O
step	O
of	O
an	O
inchworm	O
(	O
Figure	O
4	O
;	O
see	O
also	O
Figure	O
3	O
—	O
figure	O
supplement	O
2	O
).	O
	O
At	O
the	O
start	O
(	O
initiation	O
state	O
),	O
the	O
IRES	O
adopts	O
an	O
extended	O
conformation	O
(	O
extended	O
inchworm	O
).	O
	O
PKI	O
,	O
representing	O
the	O
hind	O
end	O
,	O
is	O
bound	O
in	O
the	O
A	O
site	O
.	O
	O
This	O
shortens	O
the	O
distance	O
between	O
PKI	O
and	O
SL4	O
by	O
up	O
to	O
20	O
Å	O
relative	O
to	O
the	O
initiating	O
IRES	O
structure	O
,	O
resulting	O
in	O
a	O
bent	O
IRES	O
conformation	O
(	O
bent	O
inchworm	O
).	O
	O
Finally	O
(	O
Structures	O
IV	O
to	O
V	O
),	O
as	O
the	O
hind	O
end	O
is	O
accommodated	O
in	O
the	O
P	O
site	O
,	O
the	O
front	O
'	O
legs	O
'	O
advance	O
by	O
departing	O
from	O
their	O
initial	O
binding	O
sites	O
.	O
	O
This	O
converts	O
the	O
IRES	O
into	O
an	O
extended	O
conformation	O
,	O
rendering	O
the	O
inchworm	O
prepared	O
for	O
the	O
next	O
translocation	O
step	O
.	O
	O
In	O
the	O
post	O
-	O
translocation	O
CrPV	O
IRES	O
structure	O
,	O
the	O
5	O
’-	O
domain	O
similarly	O
protrudes	O
between	O
the	O
subunits	O
and	O
interacts	O
with	O
the	O
L1	O
stalk	O
,	O
as	O
in	O
the	O
initiation	O
state	O
for	O
this	O
IRES	O
.	O
	O
This	O
underlines	O
structural	O
similarity	O
for	O
the	O
TSV	O
and	O
CrPV	O
IRES	O
translocation	O
mechanisms	O
.	O
	O
One	O
of	O
the	O
mechanistic	O
scenarios	O
(	O
discussed	O
in	O
)	O
involves	O
binding	O
of	O
the	O
first	O
aminoacyl	O
-	O
tRNA	O
to	O
the	O
post	O
-	O
translocated	O
IRES	O
mRNA	O
frame	O
shifted	O
by	O
one	O
nucleotide	O
(	O
predominantly	O
a	O
+	O
1	O
frame	O
shift	O
).	O
	O
It	O
is	O
likely	O
that	O
alternative	O
frame	O
setting	O
occurs	O
following	O
eEF2	O
release	O
and	O
that	O
this	O
depends	O
on	O
transient	O
displacement	O
of	O
the	O
start	O
codon	O
in	O
the	O
decoding	O
center	O
,	O
allowing	O
binding	O
of	O
the	O
corresponding	O
amino	O
acyl	O
-	O
tRNA	O
to	O
an	O
off	O
-	O
frame	O
codon	O
.	O
	O
Further	O
structural	O
studies	O
involving	O
80S	O
•	O
IRES	O
•	O
tRNA	O
complexes	O
are	O
necessary	O
to	O
understand	O
the	O
mechanisms	O
underlying	O
alternative	O
reading	O
frame	O
selection	O
.	O
	O
The	O
presence	O
of	O
several	O
translocation	O
complexes	O
in	O
a	O
single	O
sample	O
suggests	O
that	O
the	O
structures	O
represent	O
equilibrium	O
states	O
of	O
forward	O
and	O
reverse	O
translocation	O
of	O
the	O
IRES	O
,	O
which	O
interconvert	O
among	O
each	O
other	O
.	O
	O
Specifically	O
,	O
biochemical	O
toe	O
-	O
printing	O
studies	O
in	O
the	O
presence	O
of	O
eEF2	O
•	O
GTP	O
identified	O
IRES	O
in	O
a	O
non	O
-	O
translocated	O
position	O
unless	O
eEF1a	O
•	O
aa	O
-	O
tRNA	O
is	O
also	O
present	O
.	O
	O
These	O
findings	O
indicate	O
that	O
IRES	O
translocation	O
by	O
eEF2	O
is	O
futile	O
:	O
the	O
IRES	O
returns	O
to	O
the	O
A	O
site	O
upon	O
releasing	O
eEF2	O
•	O
GDP	O
unless	O
an	O
amino	O
-	O
acyl	O
tRNA	O
enters	O
the	O
A	O
site	O
and	O
blocks	O
IRES	O
back	O
-	O
translocation	O
.	O
	O
This	O
contrasts	O
with	O
the	O
post	O
-	O
translocated	O
2tRNA	O
•	O
mRNA	O
complex	O
,	O
in	O
which	O
the	O
classical	O
P	O
and	O
E	O
-	O
site	O
tRNAs	O
are	O
stabilized	O
in	O
the	O
non	O
-	O
rotated	O
ribosome	O
after	O
translocase	O
release	O
.	O
	O
Thus	O
,	O
the	O
meta	O
-	O
stability	O
of	O
the	O
post	O
-	O
translocation	O
IRES	O
is	O
likely	O
due	O
to	O
the	O
absence	O
of	O
stabilizing	O
structural	O
features	O
present	O
in	O
the	O
2tRNA	O
•	O
mRNA	O
complex	O
.	O
	O
Furthermore	O
,	O
interactions	O
of	O
SL4	O
and	O
SL5	O
with	O
the	O
40S	O
subunit	O
likely	O
contribute	O
to	O
stabilization	O
of	O
pre	O
-	O
translocation	O
structures	O
.	O
	O
Our	O
structures	O
delineate	O
the	O
mechanistic	O
functions	O
for	O
intersubunit	O
rotation	O
and	O
head	O
swivel	O
in	O
translocation	O
.	O
	O
Various	O
degrees	O
of	O
intersubunit	O
rotation	O
have	O
been	O
observed	O
in	O
cryo	O
-	O
EM	O
studies	O
of	O
the	O
80S	O
•	O
IRES	O
initiation	O
complexes	O
.	O
	O
This	O
suggests	O
that	O
the	O
subunits	O
are	O
capable	O
of	O
spontaneous	O
rotation	O
,	O
as	O
is	O
the	O
case	O
for	O
tRNA	O
-	O
bound	O
pre	O
-	O
translocation	O
complexes	O
.	O
	O
The	O
pre	O
-	O
translocation	O
Structure	O
I	O
with	O
eEF2	O
least	O
advanced	O
into	O
the	O
A	O
site	O
adopts	O
a	O
fully	O
rotated	O
conformation	O
.	O
	O
Reverse	O
intersubunit	O
rotation	O
from	O
Structure	O
I	O
to	O
V	O
shifts	O
the	O
translocation	O
tunnel	O
(	O
the	O
tunnel	O
between	O
the	O
A	O
,	O
P	O
and	O
E	O
sites	O
)	O
toward	O
eEF2	O
,	O
which	O
is	O
rigidly	O
attached	O
to	O
the	O
60S	O
subunit	O
.	O
	O
This	O
allows	O
eEF2	O
to	O
move	O
into	O
the	O
A	O
site	O
.	O
	O
As	O
such	O
,	O
reverse	O
intersubunit	O
rotation	O
facilitates	O
full	O
docking	O
of	O
eEF2	O
in	O
the	O
A	O
site	O
.	O
	O
The	O
head	O
swivel	O
allows	O
gradual	O
translocation	O
of	O
PKI	O
to	O
the	O
P	O
site	O
,	O
first	O
with	O
respect	O
to	O
the	O
body	O
and	O
then	O
to	O
the	O
head	O
.	O
	O
The	O
fully	O
swiveled	O
conformations	O
of	O
Structures	O
II	O
and	O
III	O
represent	O
the	O
mid	O
-	O
point	O
of	O
translocation	O
,	O
in	O
which	O
PKI	O
relocates	O
between	O
the	O
head	O
A	O
site	O
and	O
body	O
P	O
site	O
.	O
	O
We	O
note	O
that	O
such	O
mid	O
-	O
states	O
have	O
not	O
been	O
observed	O
for	O
2tRNA	O
•	O
mRNA	O
,	O
but	O
their	O
formation	O
can	O
explain	O
the	O
formation	O
of	O
subsequent	O
pe	O
/	O
E	O
hybrid	O
and	O
ap	O
/	O
P	O
chimeric	O
structures	O
(	O
Figure	O
1	O
—	O
figure	O
supplement	O
1	O
).	O
	O
Reverse	O
swivel	O
from	O
Structure	O
III	O
to	O
V	O
brings	O
the	O
head	O
to	O
the	O
non	O
-	O
swiveled	O
position	O
,	O
restoring	O
the	O
A	O
and	O
P	O
sites	O
on	O
the	O
small	O
subunit	O
.	O
	O
The	O
functions	O
of	O
eEF2	O
in	O
translocation	O
	O
The	O
structures	O
,	O
therefore	O
,	O
offer	O
a	O
unique	O
opportunity	O
to	O
address	O
the	O
role	O
of	O
the	O
elongation	O
factors	O
during	O
translocation	O
.	O
	O
As	O
discussed	O
above	O
,	O
the	O
first	O
role	O
is	O
to	O
directly	O
shift	O
PKI	O
out	O
of	O
the	O
A	O
site	O
upon	O
spontaneous	O
reverse	O
intersubunit	O
rotation	O
.	O
	O
The	O
bulky	O
ADP	O
-	O
ribosyl	O
moiety	O
at	O
this	O
position	O
would	O
disrupt	O
the	O
interaction	O
,	O
rendering	O
eEF2	O
unable	O
to	O
bind	O
to	O
the	O
A	O
site	O
and	O
/	O
or	O
stalled	O
on	O
ribosomes	O
in	O
a	O
non	O
-	O
productive	O
conformation	O
.	O
	O
As	O
eEF2	O
shifts	O
PKI	O
toward	O
the	O
P	O
site	O
in	O
the	O
course	O
of	O
reverse	O
intersubunit	O
rotation	O
,	O
the	O
60S	O
-	O
attached	O
translocase	O
migrates	O
along	O
the	O
surface	O
of	O
the	O
40S	O
subunit	O
,	O
guided	O
by	O
electrostatic	O
interactions	O
.	O
	O
Positively	O
-	O
charged	O
patches	O
of	O
domains	O
II	O
and	O
III	O
(	O
R391	O
,	O
K394	O
,	O
R433	O
,	O
R510	O
)	O
and	O
IV	O
(	O
K613	O
,	O
R617	O
,	O
R609	O
,	O
R631	O
,	O
K651	O
)	O
slide	O
over	O
rRNA	O
of	O
the	O
40S	O
body	O
(	O
h5	O
)	O
and	O
head	O
(	O
h18	O
and	O
h33	O
/	O
h34	O
),	O
respectively	O
.	O
	O
The	O
Structures	O
reveal	O
hopping	O
of	O
the	O
positive	O
clusters	O
over	O
rRNA	O
helices	O
.	O
	O
For	O
example	O
,	O
between	O
Structures	O
II	O
and	O
V	O
,	O
the	O
K613	O
/	O
R617	O
/	O
R631	O
cluster	O
of	O
domain	O
IV	O
hops	O
by	O
~	O
19	O
Å	O
(	O
for	O
Cα	O
of	O
R617	O
)	O
from	O
the	O
phosphate	O
backbone	O
of	O
h33	O
(	O
at	O
nt	O
1261	O
–	O
1264	O
)	O
to	O
that	O
of	O
the	O
neighboring	O
h34	O
(	O
at	O
nt	O
1442	O
–	O
1445	O
).	O
	O
Thus	O
,	O
sliding	O
of	O
eEF2	O
involves	O
reorganization	O
of	O
electrostatic	O
,	O
perhaps	O
isoenergetic	O
interactions	O
,	O
echoing	O
those	O
implied	O
in	O
extraordinarily	O
fast	O
ribosome	O
inactivation	O
rates	O
by	O
the	O
small	O
-	O
protein	O
ribotoxins	O
and	O
in	O
fast	O
protein	O
association	O
and	O
diffusion	O
along	O
DNA	O
.	O
	O
Comparison	O
of	O
our	O
structures	O
with	O
the	O
80S	O
•	O
IRES	O
initiation	O
structure	O
reveals	O
the	O
structural	O
basis	O
for	O
the	O
second	O
key	O
function	O
of	O
the	O
translocase	O
:	O
'	O
unlocking	O
'	O
of	O
intrasubunit	O
rearrangements	O
that	O
are	O
required	O
for	O
step	O
-	O
wise	O
translocation	O
of	O
PKI	O
on	O
the	O
small	O
subunit	O
.	O
	O
Whereas	O
intersubunit	O
rotation	O
of	O
the	O
pre	O
-	O
translocation	O
complex	O
occurs	O
spontaneously	O
,	O
the	O
head	O
swivel	O
is	O
induced	O
by	O
the	O
eEF2	O
/	O
EF	O
-	O
G	O
translocase	O
,	O
consistent	O
with	O
requirement	O
of	O
eEF2	O
for	O
unlocking	O
.	O
	O
Structural	O
studies	O
revealed	O
large	O
head	O
swivels	O
in	O
various	O
70S	O
•	O
tRNA	O
•	O
EF	O
-	O
G	O
and	O
80S	O
•	O
tRNA	O
•	O
eEF2	O
complexes	O
,	O
but	O
not	O
in	O
'	O
locked	O
'	O
complexes	O
with	O
the	O
A	O
site	O
occupied	O
by	O
the	O
tRNA	O
in	O
the	O
absence	O
of	O
the	O
translocase	O
.	O
	O
Our	O
structures	O
suggest	O
that	O
eEF2	O
induces	O
head	O
swivel	O
by	O
'	O
unlocking	O
'	O
the	O
head	O
-	O
body	O
interactions	O
(	O
Figure	O
7	O
).	O
	O
Binding	O
of	O
the	O
ASL	O
to	O
the	O
A	O
site	O
is	O
known	O
from	O
structural	O
studies	O
of	O
bacterial	O
ribosomes	O
to	O
result	O
in	O
'	O
domain	O
closure	O
'	O
of	O
the	O
small	O
subunit	O
,	O
i	O
.	O
e	O
.	O
closer	O
association	O
of	O
the	O
head	O
,	O
shoulder	O
and	O
body	O
domains	O
.	O
	O
The	O
domain	O
closure	O
'	O
locks	O
'	O
cognate	O
tRNA	O
in	O
the	O
A	O
site	O
via	O
stacking	O
on	O
the	O
head	O
A	O
site	O
(	O
C1274	O
in	O
S	O
.	O
cerevisiae	O
or	O
C1054	O
in	O
E	O
.	O
coli	O
)	O
and	O
interactions	O
with	O
the	O
body	O
A	O
-	O
site	O
nucleotides	O
A1755	O
and	O
A1756	O
(	O
A1492	O
and	O
A1493	O
in	O
E	O
.	O
coli	O
).	O
	O
This	O
'	O
locked	O
'	O
state	O
is	O
identical	O
to	O
that	O
observed	O
for	O
PKI	O
in	O
the	O
80S	O
•	O
IRES	O
initiation	O
structures	O
in	O
the	O
absence	O
of	O
eEF2	O
.	O
	O
Structure	O
I	O
demonstrates	O
that	O
at	O
an	O
early	O
pre	O
-	O
translocation	O
step	O
,	O
the	O
histidine	O
-	O
diphthamide	O
tip	O
of	O
eEF2	O
is	O
wedged	O
between	O
A1755	O
and	O
A1756	O
and	O
PKI	O
.	O
	O
Destabilization	O
of	O
the	O
head	O
-	O
bound	O
PKI	O
at	O
the	O
body	O
A	O
site	O
thus	O
allows	O
mobility	O
of	O
the	O
head	O
relative	O
to	O
the	O
body	O
.	O
	O
The	O
histidine	O
-	O
diphthamide	O
-	O
induced	O
disengagement	O
of	O
PKI	O
from	O
A1755	O
and	O
A1756	O
therefore	O
provides	O
the	O
structural	O
definition	O
for	O
the	O
'	O
unlocking	O
'	O
mode	O
of	O
eEF2	O
action	O
.	O
	O
In	O
summary	O
,	O
our	O
structures	O
are	O
consistent	O
with	O
a	O
model	O
of	O
eEF2	O
-	O
induced	O
translocation	O
in	O
which	O
both	O
PKI	O
and	O
eEF2	O
passively	O
migrate	O
into	O
the	O
P	O
and	O
A	O
site	O
,	O
respectively	O
,	O
during	O
spontaneous	O
40S	O
body	O
rotation	O
and	O
head	O
swivel	O
,	O
the	O
latter	O
being	O
allowed	O
by	O
'	O
unlocking	O
'	O
of	O
the	O
A	O
site	O
by	O
eEF2	O
.	O
	O
Observation	O
of	O
different	O
PKI	O
conformations	O
sampling	O
a	O
range	O
of	O
positions	O
between	O
the	O
A	O
and	O
P	O
sites	O
in	O
the	O
presence	O
of	O
eEF2	O
•	O
GDP	O
implies	O
that	O
thermal	O
fluctuations	O
of	O
the	O
40S	O
head	O
domain	O
are	O
sufficient	O
for	O
translocation	O
along	O
the	O
energetically	O
flat	O
trajectory	O
.	O
	O
Insights	O
into	O
eEF2	O
association	O
with	O
and	O
dissociation	O
from	O
the	O
ribosome	O
	O
The	O
conformational	O
rearrangements	O
in	O
eEF2	O
from	O
Structure	O
I	O
through	O
Structure	O
V	O
provide	O
insights	O
into	O
the	O
mechanisms	O
of	O
eEF2	O
association	O
with	O
the	O
pre	O
-	O
translocation	O
ribosome	O
and	O
dissociation	O
from	O
the	O
post	O
-	O
translocation	O
ribosome	O
.	O
	O
In	O
all	O
five	O
structures	O
,	O
the	O
GTPase	O
domain	O
is	O
attached	O
to	O
the	O
P	O
stalk	O
and	O
the	O
sarcin	O
-	O
ricin	O
loop	O
.	O
	O
Here	O
,	O
switch	O
loop	O
I	O
interacts	O
with	O
helix	O
14	O
(	O
415CAAA418	O
)	O
of	O
the	O
18S	O
rRNA	O
.	O
	O
This	O
stabilization	O
renders	O
the	O
GTPase	O
center	O
to	O
adopt	O
a	O
GTP	O
-	O
bound	O
conformation	O
,	O
similar	O
to	O
those	O
observed	O
in	O
other	O
translational	O
GTPases	O
in	O
the	O
presence	O
of	O
GTP	O
analogs	O
and	O
in	O
the	O
80S	O
•	O
eEF2	O
complex	O
bound	O
with	O
a	O
transition	O
-	O
state	O
mimic	O
GDP	O
•	O
AlF4	O
–.	O
The	O
switch	O
loop	O
contacts	O
the	O
base	O
of	O
A416	O
(	O
invariable	O
A344	O
in	O
E	O
.	O
coli	O
and	O
A463	O
in	O
H	O
.	O
sapiens	O
).	O
	O
This	O
structural	O
basis	O
rationalizes	O
the	O
observation	O
of	O
transient	O
stabilization	O
of	O
the	O
rotated	O
70S	O
ribosome	O
upon	O
EF	O
-	O
G	O
•	O
GTP	O
binding	O
and	O
prior	O
to	O
translocation	O
.	O
	O
The	O
least	O
rotated	O
conformation	O
of	O
the	O
post	O
-	O
translocation	O
Structure	O
V	O
suggests	O
conformational	O
changes	O
that	O
may	O
trigger	O
eEF2	O
release	O
from	O
the	O
ribosome	O
at	O
the	O
end	O
of	O
translocation	O
.	O
	O
The	O
most	O
pronounced	O
inter	O
-	O
domain	O
rearrangement	O
in	O
eEF2	O
involves	O
movement	O
of	O
domain	O
III	O
.	O
	O
In	O
the	O
rotated	O
or	O
mid	O
-	O
rotated	O
Structures	O
I	O
through	O
III	O
,	O
this	O
domain	O
remains	O
rigidly	O
associated	O
with	O
domain	O
V	O
and	O
the	O
N	O
-	O
terminal	O
superdomain	O
and	O
does	O
not	O
undergo	O
noticeable	O
rearrangements	O
.	O
	O
In	O
Structure	O
V	O
,	O
however	O
,	O
the	O
tip	O
of	O
helix	O
A	O
of	O
domain	O
III	O
is	O
displaced	O
toward	O
domain	O
I	O
by	O
~	O
5	O
Å	O
relative	O
to	O
that	O
in	O
mid	O
-	O
rotated	O
or	O
fully	O
rotated	O
structures	O
.	O
	O
Sordarin	O
stabilizes	O
GDP	O
-	O
bound	O
eEF2	O
on	O
the	O
ribosome	O
	O
Based	O
on	O
biochemical	O
experiments	O
,	O
two	O
alternative	O
mechanisms	O
of	O
action	O
were	O
proposed	O
:	O
sordarin	O
either	O
prevents	O
eEF2	O
departure	O
by	O
inhibiting	O
GTP	O
hydrolysis	O
or	O
acts	O
after	O
GTP	O
hydrolysis	O
.	O
	O
Our	O
structures	O
therefore	O
indicate	O
that	O
sordarin	O
stalls	O
eEF2	O
on	O
the	O
ribosome	O
in	O
the	O
GDP	O
-	O
bound	O
form	O
,	O
i	O
.	O
e	O
.	O
following	O
GTP	O
hydrolysis	O
and	O
phosphate	O
release	O
.	O
	O
The	O
mechanism	O
of	O
stalling	O
is	O
suggested	O
by	O
comparison	O
of	O
pre	O
-	O
translocation	O
and	O
post	O
-	O
translocation	O
structures	O
in	O
our	O
ensemble	O
.	O
	O
In	O
the	O
nearly	O
non	O
-	O
rotated	O
post	O
-	O
translocation	O
Structure	O
V	O
,	O
the	O
tip	O
of	O
domain	O
III	O
is	O
shifted	O
,	O
however	O
the	O
interface	O
between	O
domains	O
III	O
and	O
V	O
remains	O
unchanged	O
,	O
suggesting	O
strong	O
stabilization	O
of	O
this	O
interface	O
by	O
sordarin	O
.	O
	O
We	O
note	O
that	O
Structure	O
V	O
is	O
slightly	O
more	O
rotated	O
than	O
the	O
80S	O
•	O
2tRNA	O
•	O
mRNA	O
complex	O
in	O
the	O
absence	O
of	O
eEF2	O
•	O
sordarin	O
,	O
implying	O
that	O
sordarin	O
interferes	O
with	O
the	O
final	O
stages	O
of	O
reverse	O
rotation	O
of	O
the	O
post	O
-	O
translocation	O
ribosome	O
.	O
	O
We	O
propose	O
that	O
sordarin	O
acts	O
to	O
prevent	O
full	O
reverse	O
rotation	O
and	O
release	O
of	O
eEF2	O
•	O
GDP	O
by	O
stabilizing	O
the	O
interdomain	O
interface	O
and	O
thus	O
blocking	O
uS12	O
-	O
induced	O
disengagement	O
of	O
domain	O
III	O
from	O
domain	O
V	O
.	O
	O
Because	O
translocation	O
of	O
tRNA	O
must	O
involve	O
large	O
-	O
scale	O
dynamics	O
,	O
this	O
step	O
has	O
long	O
been	O
regarded	O
as	O
the	O
most	O
puzzling	O
step	O
of	O
translation	O
.	O
	O
The	O
structural	O
understanding	O
of	O
ribosome	O
and	O
tRNA	O
dynamics	O
has	O
been	O
greatly	O
aided	O
by	O
a	O
wealth	O
of	O
X	O
-	O
ray	O
and	O
cryo	O
-	O
EM	O
structures	O
(	O
reviewed	O
in	O
).	O
	O
However	O
,	O
visualization	O
of	O
the	O
eEF2	O
/	O
EF	O
-	O
G	O
-	O
induced	O
translocation	O
is	O
confined	O
to	O
very	O
early	O
pre	O
-	O
EF	O
-	O
G	O
-	O
entry	O
states	O
and	O
late	O
(	O
almost	O
translocated	O
or	O
fully	O
translocated	O
)	O
states	O
,	O
leaving	O
most	O
of	O
the	O
path	O
from	O
the	O
A	O
to	O
the	O
P	O
site	O
uncharacterized	O
(	O
Figure	O
1	O
—	O
figure	O
supplement	O
1	O
).	O
	O
Our	O
study	O
provides	O
new	O
insights	O
into	O
the	O
structural	O
understanding	O
of	O
tRNA	O
translocation	O
.	O
	O
This	O
is	O
evident	O
from	O
the	O
fact	O
that	O
ribosome	O
rearrangements	O
in	O
translocation	O
are	O
inherent	O
to	O
the	O
ribosome	O
and	O
likely	O
occur	O
in	O
similar	O
ways	O
in	O
both	O
cases	O
.	O
	O
For	O
example	O
,	O
fluorescence	O
and	O
biochemical	O
studies	O
revealed	O
that	O
the	O
early	O
pre	O
-	O
translocation	O
EF	O
-	O
G	O
-	O
bound	O
ribosomes	O
are	O
fully	O
rotated	O
and	O
translocation	O
of	O
the	O
tRNA	O
-	O
mRNA	O
complex	O
occurs	O
during	O
reverse	O
rotation	O
of	O
the	O
small	O
subunit	O
,	O
coupled	O
with	O
head	O
swivel	O
.	O
	O
The	O
sequence	O
of	O
ribosome	O
rearrangements	O
during	O
IRES	O
translocation	O
also	O
agrees	O
with	O
that	O
inferred	O
from	O
70S	O
•	O
EF	O
-	O
G	O
structures	O
,	O
including	O
those	O
in	O
which	O
the	O
A	O
-	O
to	O
-	O
P	O
-	O
site	O
translocating	O
tRNA	O
was	O
not	O
present	O
.	O
	O
Specifically	O
,	O
an	O
earlier	O
translocation	O
intermediate	O
ribosome	O
(	O
TIpre	O
)	O
was	O
proposed	O
to	O
adopt	O
a	O
rotated	O
(	O
7	O
–	O
9	O
°)	O
body	O
and	O
a	O
partly	O
rotated	O
head	O
(	O
5	O
–	O
7	O
.	O
5	O
°),	O
in	O
agreement	O
with	O
the	O
conformation	O
of	O
our	O
Structure	O
I	O
.	O
The	O
most	O
swiveled	O
head	O
(	O
18	O
–	O
21	O
°)	O
was	O
observed	O
in	O
a	O
mid	O
-	O
rotated	O
ribosome	O
(	O
3	O
–	O
5	O
°)	O
of	O
a	O
later	O
translocation	O
intermediate	O
TIpost	O
,	O
similar	O
to	O
the	O
conformation	O
of	O
our	O
Structure	O
III	O
.	O
	O
Overall	O
,	O
these	O
correlations	O
suggest	O
that	O
the	O
intermediate	O
locations	O
of	O
the	O
elusive	O
A	O
-	O
to	O
-	O
P	O
-	O
site	O
translocating	O
tRNA	O
are	O
similar	O
to	O
those	O
of	O
PKI	O
in	O
our	O
structures	O
.	O
	O
Second	O
,	O
the	O
structures	O
clarify	O
the	O
structural	O
basis	O
of	O
the	O
often	O
-	O
used	O
but	O
structurally	O
undefined	O
terms	O
'	O
locking	O
'	O
and	O
'	O
unlocking	O
'	O
with	O
respect	O
to	O
the	O
pre	O
-	O
translocation	O
complex	O
(	O
Figure	O
6f	O
).	O
	O
These	O
interactions	O
are	O
maintained	O
for	O
the	O
classical	O
-	O
and	O
hybrid	O
-	O
state	O
tRNAs	O
in	O
the	O
spontaneously	O
sampled	O
non	O
-	O
rotated	O
and	O
rotated	O
ribosomes	O
,	O
respectively	O
.	O
	O
Unlocking	O
involves	O
separation	O
of	O
the	O
codon	O
-	O
anticodon	O
helix	O
from	O
the	O
decoding	O
center	O
residues	O
by	O
the	O
protruding	O
tip	O
of	O
eEF2	O
/	O
EF	O
-	O
G	O
(	O
Figure	O
7	O
),	O
occurring	O
in	O
the	O
fully	O
rotated	O
ribosome	O
at	O
an	O
early	O
pre	O
-	O
translocation	O
step	O
.	O
	O
Third	O
,	O
our	O
findings	O
uncover	O
a	O
new	O
role	O
of	O
the	O
head	O
swivel	O
.	O
	O
Previous	O
studies	O
showed	O
that	O
this	O
movement	O
widens	O
the	O
constriction	O
('	O
gate	O
')	O
between	O
the	O
P	O
and	O
E	O
sites	O
,	O
thus	O
allowing	O
the	O
P	O
-	O
tRNA	O
passage	O
to	O
the	O
E	O
site	O
.	O
	O
In	O
addition	O
to	O
the	O
'	O
gate	O
-	O
opening	O
'	O
role	O
,	O
we	O
now	O
show	O
that	O
the	O
head	O
swivel	O
brings	O
the	O
head	O
A	O
site	O
to	O
the	O
body	O
P	O
site	O
,	O
allowing	O
a	O
step	O
-	O
wise	O
conveying	O
of	O
the	O
codon	O
-	O
anticodon	O
helix	O
between	O
the	O
A	O
and	O
P	O
sites	O
.	O
	O
Our	O
findings	O
implicate	O
,	O
however	O
,	O
that	O
the	O
energy	O
landscape	O
is	O
not	O
completely	O
flat	O
and	O
contains	O
local	O
minima	O
for	O
transient	O
positions	O
of	O
the	O
codon	O
-	O
anticodon	O
helix	O
between	O
the	O
A	O
and	O
P	O
sites	O
.	O
	O
The	O
shift	O
of	O
the	O
PKI	O
with	O
respect	O
to	O
the	O
body	O
occurs	O
during	O
forward	O
head	O
swivel	O
in	O
two	O
major	O
sub	O
-	O
steps	O
of	O
~	O
4	O
Å	O
each	O
(	O
initiation	O
complex	O
to	O
I	O
,	O
and	O
I	O
to	O
II	O
),	O
after	O
which	O
PKI	O
undergoes	O
small	O
shifts	O
to	O
settle	O
in	O
the	O
body	O
P	O
site	O
in	O
Structures	O
III	O
,	O
IV	O
and	O
V	O
(	O
Figure	O
2	O
—	O
source	O
data	O
1	O
).	O
	O
Movement	O
of	O
PKI	O
relative	O
to	O
the	O
head	O
occurs	O
during	O
the	O
subsequent	O
reverse	O
swivel	O
in	O
three	O
3	O
–	O
7	O
Å	O
sub	O
-	O
steps	O
(	O
II	O
to	O
III	O
to	O
IV	O
to	O
V	O
).	O
	O
Translation	O
of	O
viral	O
mRNA	O
	O
Our	O
work	O
sheds	O
light	O
on	O
the	O
dynamic	O
mechanism	O
of	O
cap	O
-	O
independent	O
translation	O
by	O
IGR	O
IRESs	O
,	O
tightly	O
coupled	O
with	O
the	O
universally	O
conserved	O
dynamic	O
properties	O
of	O
the	O
ribosome	O
.	O
	O
Like	O
in	O
the	O
2tRNA	O
•	O
mRNA	O
translocating	O
complex	O
in	O
which	O
the	O
two	O
tRNAs	O
move	O
independently	O
of	O
each	O
other	O
,	O
the	O
PKI	O
domain	O
moves	O
relative	O
to	O
the	O
5	O
´-	O
domain	O
,	O
causing	O
the	O
IRES	O
to	O
undergo	O
an	O
inchworm	O
-	O
walk	O
translocation	O
.	O
	O
A	O
large	O
structural	O
difference	O
between	O
the	O
IRES	O
and	O
the	O
2tRNA	O
•	O
mRNA	O
complex	O
exists	O
,	O
however	O
,	O
in	O
that	O
the	O
IRES	O
lacks	O
three	O
out	O
of	O
six	O
tRNA	O
-	O
like	O
domains	O
involved	O
in	O
tRNA	O
translocation	O
.	O
	O
Although	O
structurally	O
handicapped	O
,	O
the	O
TSV	O
IRES	O
manages	O
to	O
translocate	O
by	O
employing	O
ribosome	O
dynamics	O
that	O
are	O
remarkably	O
similar	O
to	O
that	O
in	O
2tRNA	O
•	O
mRNA	O
translocation	O
.	O
	O
This	O
property	O
is	O
rendered	O
by	O
the	O
relative	O
mobility	O
of	O
the	O
three	O
major	O
building	O
blocks	O
,	O
the	O
60S	O
subunit	O
and	O
the	O
40S	O
head	O
and	O
body	O
,	O
assisted	O
by	O
ligand	O
-	O
interacting	O
extensions	O
including	O
the	O
L1	O
stalk	O
and	O
the	O
P	O
stalk	O
.	O
	O
Viral	O
mRNAs	O
have	O
evolved	O
to	O
adopt	O
an	O
atypical	O
structure	O
to	O
employ	O
the	O
inherent	O
ribosome	O
dynamics	O
,	O
to	O
be	O
able	O
to	O
hijack	O
the	O
host	O
translational	O
machinery	O
in	O
a	O
simple	O
fashion	O
.	O
	O
Ensemble	O
cryo	O
-	O
EM	O
	O
High	O
-	O
resolution	O
crystal	O
structures	O
,	O
on	O
the	O
other	O
hand	O
,	O
can	O
provide	O
static	O
images	O
of	O
an	O
assembly	O
,	O
and	O
the	O
structural	O
dynamics	O
can	O
only	O
be	O
inferred	O
by	O
comparing	O
structures	O
that	O
are	O
usually	O
obtained	O
in	O
different	O
experiments	O
and	O
under	O
different	O
,	O
often	O
non	O
-	O
native	O
,	O
conditions	O
.	O
	O
Cryo	O
-	O
EM	O
offers	O
the	O
possibility	O
of	O
obtaining	O
integrated	O
information	O
of	O
both	O
structure	O
and	O
dynamics	O
as	O
demonstrated	O
in	O
lower	O
-	O
resolution	O
studies	O
of	O
bacterial	O
ribosome	O
complexes	O
.	O
	O
This	O
is	O
presumably	O
one	O
of	O
the	O
reasons	O
why	O
most	O
recent	O
studies	O
of	O
ribosome	O
complexes	O
have	O
focused	O
on	O
a	O
single	O
high	O
-	O
resolution	O
structure	O
despite	O
the	O
non	O
-	O
uniform	O
local	O
resolution	O
of	O
the	O
maps	O
that	O
likely	O
reflects	O
structural	O
heterogeneity	O
.	O
	O
The	O
computational	O
efficiency	O
of	O
FREALIGN	O
has	O
allowed	O
us	O
to	O
classify	O
a	O
relatively	O
large	O
dataset	O
(	O
1	O
.	O
1	O
million	O
particles	O
)	O
into	O
15	O
classes	O
(	O
Figure	O
1	O
—	O
figure	O
supplement	O
2	O
)	O
and	O
obtain	O
eight	O
near	O
-	O
atomic	O
-	O
resolution	O
structures	O
from	O
it	O
.	O
	O
Therefore	O
,	O
cryo	O
-	O
EM	O
has	O
the	O
potential	O
to	O
become	O
a	O
standard	O
tool	O
for	O
uncovering	O
detailed	O
dynamic	O
pathways	O
of	O
complex	O
macromolecular	O
machines	O
.	O
	O
A	O
unified	O
mechanism	O
for	O
proteolysis	O
and	O
autocatalytic	O
activation	O
in	O
the	O
20S	O
proteasome	O
	O
Biogenesis	O
of	O
the	O
20S	O
proteasome	O
is	O
tightly	O
regulated	O
.	O
	O
However	O
,	O
the	O
trigger	O
for	O
the	O
self	O
-	O
activation	O
and	O
the	O
reason	O
for	O
the	O
strict	O
conservation	O
of	O
threonine	O
as	O
the	O
active	O
site	O
nucleophile	O
remain	O
enigmatic	O
.	O
	O
Here	O
we	O
use	O
mutagenesis	O
,	O
X	O
-	O
ray	O
crystallography	O
and	O
biochemical	O
assays	O
to	O
suggest	O
that	O
Lys33	O
initiates	O
nucleophilic	O
attack	O
of	O
the	O
propeptide	O
by	O
deprotonating	O
the	O
Thr1	O
hydroxyl	O
group	O
and	O
that	O
both	O
residues	O
together	O
with	O
Asp17	O
are	O
part	O
of	O
a	O
catalytic	O
triad	O
.	O
	O
Substitution	O
of	O
Thr1	O
by	O
Cys	O
disrupts	O
the	O
interaction	O
with	O
Lys33	O
and	O
inactivates	O
the	O
proteasome	O
.	O
	O
Although	O
a	O
Thr1Ser	B-mutant
mutant	O
is	O
active	O
,	O
it	O
is	O
less	O
efficient	O
compared	O
with	O
wild	O
type	O
because	O
of	O
the	O
unfavourable	O
orientation	O
of	O
Ser1	O
towards	O
incoming	O
substrates	O
.	O
	O
The	O
proteasome	O
,	O
an	O
essential	O
molecular	O
machine	O
,	O
is	O
a	O
threonine	O
protease	O
,	O
but	O
the	O
evolution	O
and	O
the	O
components	O
of	O
its	O
proteolytic	O
centre	O
are	O
unclear	O
.	O
	O
Here	O
,	O
the	O
authors	O
use	O
structural	O
biology	O
and	O
biochemistry	O
to	O
investigate	O
the	O
role	O
of	O
proteasome	O
active	O
site	O
residues	O
on	O
maturation	O
and	O
activity	O
.	O
	O
The	O
20S	O
proteasome	O
core	O
particle	O
(	O
CP	O
)	O
is	O
the	O
key	O
non	O
-	O
lysosomal	O
protease	O
of	O
eukaryotic	O
cells	O
.	O
	O
Its	O
seven	O
different	O
α	O
and	O
seven	O
different	O
β	O
subunits	O
assemble	O
into	O
four	O
heptameric	O
rings	O
that	O
are	O
stacked	O
on	O
each	O
other	O
to	O
form	O
a	O
hollow	O
cylinder	O
.	O
	O
While	O
the	O
inactive	O
α	O
subunits	O
build	O
the	O
two	O
outer	O
rings	O
,	O
the	O
β	O
subunits	O
form	O
the	O
inner	O
rings	O
.	O
	O
Only	O
three	O
out	O
of	O
the	O
seven	O
different	O
β	O
subunits	O
,	O
namely	O
β1	O
,	O
β2	O
and	O
β5	O
,	O
bear	O
N	O
-	O
terminal	O
proteolytic	O
active	O
centres	O
,	O
and	O
before	O
CP	O
maturation	O
these	O
are	O
protected	O
by	O
propeptides	O
.	O
	O
In	O
the	O
last	O
stage	O
of	O
CP	O
biogenesis	O
,	O
the	O
prosegments	O
are	O
autocatalytically	O
removed	O
through	O
nucleophilic	O
attack	O
by	O
the	O
active	O
site	O
residue	O
Thr1	O
on	O
the	O
preceding	O
peptide	O
bond	O
involving	O
Gly	O
(-	O
1	O
).	O
	O
Release	O
of	O
the	O
propeptides	O
creates	O
a	O
functionally	O
active	O
CP	O
that	O
cleaves	O
proteins	O
into	O
short	O
peptides	O
.	O
	O
Although	O
the	O
chemical	O
nature	O
of	O
the	O
substrate	O
-	O
binding	O
channel	O
and	O
hence	O
substrate	O
preferences	O
are	O
unique	O
to	O
each	O
of	O
the	O
distinct	O
active	O
β	O
subunits	O
,	O
all	O
active	O
sites	O
employ	O
an	O
identical	O
reaction	O
mechanism	O
to	O
hydrolyse	O
peptide	O
bonds	O
.	O
	O
Nucleophilic	O
attack	O
of	O
Thr1Oγ	O
on	O
the	O
carbonyl	O
carbon	O
atom	O
of	O
the	O
scissile	O
peptide	O
bond	O
creates	O
a	O
first	O
cleavage	O
product	O
and	O
a	O
covalent	O
acyl	O
-	O
enzyme	O
intermediate	O
.	O
	O
Hydrolysis	O
of	O
this	O
complex	O
by	O
the	O
addition	O
of	O
a	O
nucleophilic	O
water	O
molecule	O
regenerates	O
the	O
enzyme	O
and	O
releases	O
the	O
second	O
peptide	O
fragment	O
.	O
	O
The	O
proteasome	O
belongs	O
to	O
the	O
family	O
of	O
N	O
-	O
terminal	O
nucleophilic	O
(	O
Ntn	O
)	O
hydrolases	O
,	O
and	O
the	O
free	O
N	O
-	O
terminal	O
amine	O
group	O
of	O
Thr1	O
was	O
proposed	O
to	O
deprotonate	O
the	O
Thr1	O
hydroxyl	O
group	O
to	O
generate	O
a	O
nucleophilic	O
Thr1Oγ	O
for	O
peptide	O
-	O
bond	O
cleavage	O
.	O
	O
An	O
alternative	O
candidate	O
for	O
deprotonating	O
the	O
Thr1	O
hydroxyl	O
group	O
is	O
the	O
side	O
chain	O
of	O
Lys33	O
as	O
it	O
is	O
within	O
hydrogen	O
-	O
bonding	O
distance	O
to	O
Thr1OH	O
(	O
2	O
.	O
7	O
Å	O
).	O
	O
In	O
principle	O
it	O
could	O
function	O
as	O
the	O
general	O
base	O
during	O
both	O
autocatalytic	O
removal	O
of	O
the	O
propeptide	O
and	O
protein	O
substrate	O
cleavage	O
.	O
	O
Here	O
we	O
provide	O
experimental	O
evidences	O
for	O
this	O
distinct	O
view	O
of	O
the	O
proteasome	O
active	O
-	O
site	O
mechanism	O
.	O
	O
Furthermore	O
,	O
we	O
determine	O
the	O
advantages	O
of	O
Thr	O
over	O
Cys	O
or	O
Ser	O
as	O
the	O
active	O
-	O
site	O
nucleophile	O
using	O
X	O
-	O
ray	O
crystallography	O
together	O
with	O
activity	O
and	O
inhibition	O
assays	O
.	O
	O
Inactivation	O
of	O
the	O
active	O
site	O
Thr1	O
by	O
mutation	O
to	O
Ala	O
has	O
been	O
used	O
to	O
study	O
substrate	O
specificity	O
and	O
the	O
hierarchy	O
of	O
the	O
proteasome	O
active	O
sites	O
.	O
	O
Yeast	O
strains	O
carrying	O
the	O
single	O
mutations	O
β1	B-mutant
-	I-mutant
T1A	I-mutant
or	O
β2	B-mutant
-	I-mutant
T1A	I-mutant
,	O
or	O
both	O
,	O
are	O
viable	O
,	O
even	O
though	O
one	O
or	O
two	O
of	O
the	O
three	O
distinct	O
catalytic	O
β	O
subunits	O
are	O
disabled	O
and	O
carry	O
remnants	O
of	O
their	O
N	O
-	O
terminal	O
propeptides	O
(	O
Table	O
1	O
).	O
	O
By	O
contrast	O
,	O
the	O
T1A	B-mutant
mutation	O
in	O
subunit	O
β5	O
has	O
been	O
reported	O
to	O
be	O
lethal	O
or	O
nearly	O
so	O
.	O
	O
Viability	O
is	O
restored	O
if	O
the	O
β5	B-mutant
-	I-mutant
T1A	I-mutant
subunit	O
has	O
its	O
propeptide	O
(	O
pp	O
)	O
deleted	O
but	O
expressed	O
separately	O
in	O
trans	O
(	O
β5	B-mutant
-	I-mutant
T1A	I-mutant
pp	O
trans	O
),	O
although	O
substantial	O
phenotypic	O
impairment	O
remains	O
(	O
Table	O
1	O
).	O
	O
The	O
extremely	O
weak	O
growth	O
of	O
the	O
β5	B-mutant
-	I-mutant
T1A	I-mutant
mutant	O
pp	O
cis	O
described	O
by	O
Chen	O
and	O
Hochstrasser	O
compared	O
with	O
the	O
inviability	O
reported	O
by	O
Heinemeyer	O
et	O
al	O
.	O
prompted	O
us	O
to	O
analyse	O
this	O
discrepancy	O
.	O
	O
We	O
also	O
identified	O
an	O
additional	O
point	O
mutation	O
K81R	B-mutant
in	O
subunit	O
β5	O
that	O
was	O
present	O
in	O
the	O
allele	O
used	O
in	O
ref	O
..	O
This	O
single	O
amino	O
-	O
acid	O
exchange	O
is	O
located	O
at	O
the	O
interface	O
of	O
the	O
subunits	O
α4	O
,	O
β4	O
and	O
β5	O
(	O
Supplementary	O
Fig	O
.	O
1b	O
)	O
and	O
might	O
weakly	O
promote	O
CP	O
assembly	O
by	O
enhancing	O
inter	O
-	O
subunit	O
contacts	O
.	O
	O
The	O
slightly	O
better	O
growth	O
of	O
the	O
β5	B-mutant
-	I-mutant
T1A	I-mutant
-	I-mutant
K81R	I-mutant
mutant	O
allowed	O
us	O
to	O
solve	O
the	O
crystal	O
structure	O
of	O
a	O
yeast	O
proteasome	O
(	O
yCP	O
)	O
with	O
the	O
β5	B-mutant
-	I-mutant
T1A	I-mutant
mutation	O
,	O
which	O
is	O
discussed	O
in	O
the	O
following	O
section	O
(	O
for	O
details	O
see	O
Supplementary	O
Note	O
1	O
).	O
	O
Propeptide	O
conformation	O
and	O
triggering	O
of	O
autolysis	O
	O
In	O
the	O
final	O
steps	O
of	O
proteasome	O
biogenesis	O
,	O
the	O
propeptides	O
are	O
autocatalytically	O
cleaved	O
from	O
the	O
mature	O
β	O
-	O
subunit	O
domains	O
.	O
	O
Furthermore	O
,	O
it	O
was	O
observed	O
that	O
the	O
prosegment	O
forms	O
an	O
antiparallel	O
β	O
-	O
sheet	O
in	O
the	O
active	O
site	O
,	O
and	O
that	O
Gly	O
(-	O
1	O
)	O
adopts	O
a	O
γ	O
-	O
turn	O
conformation	O
,	O
which	O
by	O
definition	O
is	O
characterized	O
by	O
a	O
hydrogen	O
bond	O
between	O
Leu	O
(-	O
2	O
)	O
O	O
and	O
Thr1NH	O
(	O
ref	O
.).	O
	O
In	O
subunit	O
β1	O
,	O
we	O
found	O
that	O
Gly	O
(-	O
1	O
)	O
indeed	O
forms	O
a	O
sharp	O
turn	O
,	O
which	O
relaxes	O
on	O
prosegment	O
cleavage	O
(	O
Fig	O
.	O
1a	O
and	O
Supplementary	O
Fig	O
.	O
2a	O
).	O
	O
However	O
,	O
the	O
γ	O
-	O
turn	O
conformation	O
and	O
the	O
associated	O
hydrogen	O
bond	O
initially	O
proposed	O
is	O
for	O
geometric	O
and	O
chemical	O
reasons	O
inappropriate	O
and	O
would	O
not	O
perfectly	O
position	O
the	O
carbonyl	O
carbon	O
atom	O
of	O
Gly	O
(-	O
1	O
)	O
for	O
nucleophilic	O
attack	O
by	O
Thr1	O
.	O
	O
Regarding	O
the	O
β2	O
propeptide	O
,	O
Thr	O
(-	O
2	O
)	O
occupies	O
the	O
S1	O
pocket	O
but	O
is	O
less	O
deeply	O
anchored	O
compared	O
with	O
Leu	O
(-	O
2	O
)	O
in	O
β1	O
,	O
which	O
might	O
be	O
due	O
to	O
the	O
rather	O
large	O
β2	O
-	O
S1	O
pocket	O
created	O
by	O
Gly45	O
.	O
	O
Next	O
,	O
we	O
examined	O
the	O
position	O
of	O
the	O
β5	O
propeptide	O
in	O
the	O
β5	B-mutant
-	I-mutant
T1A	I-mutant
-	I-mutant
K81R	I-mutant
mutant	O
.	O
	O
Surprisingly	O
,	O
Gly	O
(-	O
1	O
)	O
is	O
completely	O
extended	O
and	O
forces	O
the	O
histidine	O
side	O
chain	O
at	O
position	O
(-	O
2	O
)	O
to	O
occupy	O
the	O
S2	O
instead	O
of	O
the	O
S1	O
pocket	O
,	O
thereby	O
disrupting	O
the	O
antiparallel	O
β	O
-	O
sheet	O
.	O
	O
Nonetheless	O
,	O
the	O
carbonyl	O
carbon	O
of	O
Gly	O
(-	O
1	O
)	O
would	O
be	O
ideally	O
placed	O
for	O
nucleophilic	O
attack	O
by	O
Thr1Oγ	O
(	O
Fig	O
.	O
1c	O
and	O
Supplementary	O
Fig	O
.	O
2c	O
,	O
d	O
).	O
	O
As	O
the	O
K81R	B-mutant
mutation	O
is	O
located	O
far	O
from	O
the	O
active	O
site	O
(	O
Thr1Cα	O
–	O
Arg81Cα	O
:	O
24	O
Å	O
),	O
any	O
influence	O
on	O
propeptide	O
conformation	O
can	O
be	O
excluded	O
.	O
	O
Processing	O
of	O
β	O
-	O
subunit	O
precursors	O
requires	O
deprotonation	O
of	O
Thr1OH	O
;	O
however	O
,	O
the	O
general	O
base	O
initiating	O
autolysis	O
is	O
unknown	O
.	O
	O
Remarkably	O
,	O
eukaryotic	O
proteasomal	O
β5	O
subunits	O
bear	O
a	O
His	O
residue	O
in	O
position	O
(-	O
2	O
)	O
of	O
the	O
propeptide	O
(	O
Supplementary	O
Fig	O
.	O
3a	O
).	O
	O
Gly	O
(-	O
1	O
)	O
and	O
Phe	O
/	O
Lys	O
(-	O
2	O
)	O
were	O
visualized	O
at	O
low	O
occupancy	O
,	O
while	O
Ala	O
/	O
Asn	O
(-	O
2	O
)	O
could	O
not	O
be	O
assigned	O
.	O
	O
This	O
observation	O
indicates	O
a	O
mixture	O
of	O
processed	O
and	O
unprocessed	O
β5	O
subunits	O
and	O
partially	O
impaired	O
autolysis	O
,	O
thereby	O
excluding	O
any	O
essential	O
role	O
of	O
residue	O
(-	O
2	O
)	O
as	O
the	O
general	O
base	O
.	O
	O
Leu	O
(-	O
2	O
)	O
is	O
encoded	O
in	O
the	O
yeast	O
β1	O
subunit	O
precursor	O
(	O
Supplementary	O
Fig	O
.	O
3a	O
);	O
Thr	O
(-	O
2	O
)	O
is	O
generally	O
part	O
of	O
β2	O
-	O
propeptides	O
(	O
Supplementary	O
Fig	O
.	O
3a	O
);	O
and	O
Ala	O
(-	O
2	O
)	O
was	O
expected	O
to	O
fit	O
the	O
β5	O
-	O
S1	O
pocket	O
without	O
inducing	O
conformational	O
changes	O
of	O
Met45	O
,	O
allowing	O
it	O
to	O
accommodate	O
‘	O
β1	O
-	O
like	O
'	O
propeptide	O
positioning	O
.	O
	O
As	O
expected	O
from	O
β5	B-mutant
-	I-mutant
T1A	I-mutant
mutants	O
,	O
the	O
yeasts	O
show	O
severe	O
growth	O
phenotypes	O
,	O
with	O
minor	O
variations	O
(	O
Supplementary	O
Fig	O
.	O
4a	O
and	O
Table	O
1	O
).	O
	O
We	O
determined	O
crystal	O
structures	O
of	O
the	O
β5	B-mutant
-	I-mutant
H	I-mutant
(-	I-mutant
2	I-mutant
)	I-mutant
L	I-mutant
-	I-mutant
T1A	I-mutant
,	O
β5	B-mutant
-	I-mutant
H	I-mutant
(-	I-mutant
2	I-mutant
)	I-mutant
T	I-mutant
-	I-mutant
T1A	I-mutant
and	O
the	O
β5	B-mutant
-	I-mutant
H	I-mutant
(-	I-mutant
2	I-mutant
)	I-mutant
A	I-mutant
-	I-mutant
T1A	I-mutant
-	I-mutant
K81R	I-mutant
mutants	O
(	O
Supplementary	O
Table	O
1	O
).	O
	O
For	O
the	O
β5	B-mutant
-	I-mutant
H	I-mutant
(-	I-mutant
2	I-mutant
)	I-mutant
A	I-mutant
-	I-mutant
T1A	I-mutant
-	I-mutant
K81R	I-mutant
variant	O
,	O
only	O
the	O
residues	O
Gly	O
(-	O
1	O
)	O
and	O
Ala	O
(-	O
2	O
)	O
could	O
be	O
visualized	O
,	O
indicating	O
that	O
Ala	O
(-	O
2	O
)	O
leads	O
to	O
insufficient	O
stabilization	O
of	O
the	O
propeptide	O
in	O
the	O
substrate	O
-	O
binding	O
channel	O
(	O
Supplementary	O
Fig	O
.	O
4d	O
).	O
	O
Nevertheless	O
,	O
both	O
Leu	O
(-	O
2	O
)	O
and	O
Thr	O
(-	O
2	O
)	O
were	O
found	O
to	O
occupy	O
the	O
S1	O
specificity	O
pocket	O
formed	O
by	O
Met45	O
(	O
Fig	O
.	O
2a	O
,	O
b	O
and	O
Supplementary	O
Fig	O
.	O
4f	O
–	O
h	O
).	O
	O
Since	O
Gly	O
(-	O
1	O
)	O
adopts	O
the	O
same	O
position	O
in	O
both	O
wild	O
-	O
type	O
(	O
WT	O
)	O
and	O
mutant	O
β5	O
propeptides	O
,	O
and	O
since	O
in	O
all	O
cases	O
its	O
carbonyl	O
carbon	O
is	O
perfectly	O
placed	O
for	O
nucleophilic	O
attack	O
by	O
Thr1Oγ	O
(	O
Fig	O
.	O
2b	O
),	O
we	O
propose	O
that	O
neither	O
binding	O
of	O
residue	O
(-	O
2	O
)	O
to	O
the	O
S1	O
pocket	O
nor	O
formation	O
of	O
the	O
antiparallel	O
β	O
-	O
sheet	O
is	O
essential	O
for	O
autolysis	O
of	O
the	O
propeptide	O
.	O
	O
Next	O
,	O
we	O
determined	O
the	O
crystal	O
structure	O
of	O
a	O
chimeric	O
yCP	O
having	O
the	O
yeast	O
β1	O
-	O
propeptide	O
replaced	O
by	O
its	O
β5	O
counterpart	O
.	O
	O
Although	O
we	O
observed	O
fragments	O
of	O
2FO	O
–	O
FC	O
electron	O
density	O
in	O
the	O
β1	O
active	O
site	O
,	O
the	O
data	O
were	O
not	O
interpretable	O
.	O
	O
As	O
proven	O
by	O
the	O
β2	B-mutant
-	I-mutant
T1A	I-mutant
crystal	O
structures	O
,	O
Thr	O
(-	O
2	O
)	O
hydrogen	O
bonds	O
to	O
Gly	O
(-	O
1	O
)	O
O	O
.	O
Although	O
this	O
interaction	O
was	O
not	O
observed	O
for	O
the	O
β5	B-mutant
-	I-mutant
H	I-mutant
(-	I-mutant
2	I-mutant
)	I-mutant
T	I-mutant
-	I-mutant
T1A	I-mutant
mutant	O
(	O
Fig	O
.	O
2c	O
and	O
Supplementary	O
Fig	O
.	O
4c	O
,	O
i	O
),	O
exchange	O
of	O
Thr	O
(-	O
2	O
)	O
by	O
Val	O
in	O
β2	O
,	O
a	O
conservative	O
mutation	O
regarding	O
size	O
but	O
drastic	O
with	O
respect	O
to	O
polarity	O
,	O
was	O
found	O
to	O
inhibit	O
maturation	O
of	O
this	O
subunit	O
(	O
Fig	O
.	O
2d	O
and	O
Supplementary	O
Fig	O
.	O
4e	O
,	O
j	O
).	O
	O
In	O
particular	O
,	O
Val	O
(-	O
2	O
)	O
is	O
displaced	O
from	O
the	O
S1	O
site	O
and	O
Gly	O
(-	O
1	O
)	O
is	O
severely	O
shifted	O
(	O
movement	O
of	O
the	O
carbonyl	O
oxygen	O
atom	O
of	O
3	O
.	O
8	O
Å	O
),	O
thereby	O
preventing	O
nucleophilic	O
attack	O
of	O
Thr1	O
(	O
Fig	O
.	O
2d	O
and	O
Supplementary	O
Fig	O
.	O
4j	O
,	O
k	O
).	O
	O
These	O
results	O
further	O
confirm	O
that	O
correct	O
positioning	O
of	O
the	O
active	O
-	O
site	O
residues	O
and	O
Gly	O
(-	O
1	O
)	O
is	O
decisive	O
for	O
the	O
maturation	O
of	O
the	O
proteasome	O
.	O
	O
The	O
active	O
site	O
of	O
the	O
proteasome	O
	O
Proton	O
shuttling	O
from	O
the	O
proteasomal	O
active	O
site	O
Thr1OH	O
to	O
Thr1NH2	O
via	O
a	O
nucleophilic	O
water	O
molecule	O
was	O
suggested	O
to	O
initiate	O
peptide	O
-	O
bond	O
hydrolysis	O
.	O
	O
A	O
proposed	O
catalytic	O
tetrad	O
model	O
involving	O
Thr1OH	O
,	O
Thr1NH2	O
,	O
Lys33NH2	O
and	O
Asp17Oδ	O
,	O
as	O
well	O
as	O
a	O
nucleophilic	O
water	O
molecule	O
as	O
the	O
proton	O
shuttle	O
appeared	O
to	O
accommodate	O
all	O
possible	O
views	O
of	O
the	O
proteasomal	O
active	O
site	O
.	O
	O
Twenty	O
years	O
later	O
,	O
with	O
a	O
plethora	O
of	O
yCP	O
X	O
-	O
ray	O
structures	O
in	O
hand	O
,	O
we	O
decided	O
to	O
re	O
-	O
analyse	O
the	O
active	O
site	O
of	O
the	O
proteasome	O
and	O
to	O
resolve	O
the	O
uncertainty	O
regarding	O
the	O
nature	O
of	O
the	O
general	O
base	O
.	O
	O
Mutation	O
of	O
β5	O
-	O
Lys33	O
to	O
Ala	O
causes	O
a	O
strongly	O
deleterious	O
phenotype	O
,	O
and	O
previous	O
structural	O
and	O
biochemical	O
analyses	O
confirmed	O
that	O
this	O
is	O
caused	O
by	O
failure	O
of	O
propeptide	O
cleavage	O
,	O
and	O
consequently	O
,	O
lack	O
of	O
ChT	O
-	O
L	O
activity	O
(	O
Fig	O
.	O
4a	O
,	O
Supplementary	O
Fig	O
.	O
3b	O
and	O
Table	O
1	O
;	O
for	O
details	O
see	O
Supplementary	O
Note	O
1	O
).	O
	O
This	O
discrepancy	O
in	O
growth	O
was	O
traced	O
to	O
an	O
additional	O
point	O
mutation	O
L	B-mutant
(-	I-mutant
49	I-mutant
)	I-mutant
S	I-mutant
in	O
the	O
β5	O
-	O
propeptide	O
of	O
the	O
β5	B-mutant
-	I-mutant
K33A	I-mutant
mutant	O
(	O
see	O
also	O
Supplementary	O
Note	O
1	O
).	O
	O
This	O
structural	O
alteration	O
destroys	O
active	O
-	O
site	O
integrity	O
and	O
abolishes	O
catalytic	O
activity	O
of	O
the	O
β5	O
active	O
site	O
(	O
Supplementary	O
Fig	O
.	O
5a	O
).	O
	O
Additional	O
proof	O
for	O
the	O
key	O
function	O
of	O
Lys33	O
was	O
obtained	O
from	O
the	O
β5	B-mutant
-	I-mutant
K33A	I-mutant
mutant	O
,	O
with	O
the	O
propeptide	O
expressed	O
separately	O
from	O
the	O
main	O
subunit	O
(	O
pp	O
trans	O
).	O
	O
The	O
Thr1	O
N	O
terminus	O
of	O
this	O
mutant	O
is	O
not	O
blocked	O
by	O
the	O
propeptide	O
,	O
yet	O
its	O
catalytic	O
activity	O
is	O
reduced	O
by	O
∼	O
83	O
%	O
(	O
Supplementary	O
Fig	O
.	O
6b	O
).	O
	O
Consistent	O
with	O
this	O
,	O
the	O
crystal	O
structure	O
of	O
the	O
β5	B-mutant
-	I-mutant
K33A	I-mutant
pp	O
trans	O
mutant	O
in	O
complex	O
with	O
carfilzomib	O
only	O
showed	O
partial	O
occupancy	O
of	O
the	O
ligand	O
at	O
the	O
β5	O
active	O
sites	O
(	O
Supplementary	O
Fig	O
.	O
5b	O
and	O
Supplementary	O
Table	O
1	O
).	O
	O
Since	O
no	O
acetylation	O
of	O
the	O
Thr1	O
N	O
terminus	O
was	O
observed	O
for	O
the	O
β5	B-mutant
-	I-mutant
K33A	I-mutant
pp	O
trans	O
apo	O
crystal	O
structure	O
,	O
the	O
reduced	O
reactivity	O
towards	O
substrates	O
and	O
inhibitors	O
indicates	O
that	O
Lys33NH2	O
,	O
rather	O
than	O
Thr1NH2	O
,	O
deprotonates	O
and	O
activates	O
Thr1OH	O
.	O
	O
Furthermore	O
,	O
the	O
crystal	O
structure	O
of	O
the	O
β5	B-mutant
-	I-mutant
K33A	I-mutant
pp	O
trans	O
mutant	O
without	O
inhibitor	O
revealed	O
that	O
Thr1Oγ	O
strongly	O
coordinates	O
a	O
well	O
-	O
defined	O
water	O
molecule	O
(∼	O
2	O
Å	O
;	O
Fig	O
.	O
3c	O
and	O
Supplementary	O
Fig	O
.	O
5c	O
,	O
d	O
).	O
	O
Remarkably	O
,	O
the	O
solvent	O
molecule	O
occupies	O
the	O
position	O
normally	O
taken	O
by	O
Lys33NH2	O
in	O
the	O
WT	O
proteasome	O
structure	O
(	O
Fig	O
.	O
3c	O
),	O
further	O
corroborating	O
the	O
essential	O
role	O
of	O
Lys33	O
as	O
the	O
general	O
base	O
for	O
autolysis	O
and	O
proteolysis	O
.	O
	O
Conservative	O
substitution	O
of	O
Lys33	O
by	O
Arg	O
delays	O
autolysis	O
of	O
the	O
β5	O
precursor	O
and	O
impairs	O
yeast	O
growth	O
(	O
for	O
details	O
see	O
Supplementary	O
Note	O
1	O
).	O
	O
While	O
Thr1	O
occupies	O
the	O
same	O
position	O
as	O
in	O
WT	O
yCPs	O
,	O
Arg33	O
is	O
unable	O
to	O
hydrogen	O
bond	O
to	O
Asp17	O
,	O
thereby	O
inactivating	O
the	O
β5	O
active	O
site	O
(	O
Supplementary	O
Fig	O
.	O
5e	O
).	O
	O
The	O
conservative	O
mutation	O
of	O
Asp17	O
to	O
Asn	O
in	O
subunit	O
β5	O
of	O
the	O
yCP	O
also	O
provokes	O
a	O
severe	O
growth	O
defect	O
(	O
Supplementary	O
Note	O
1	O
,	O
Supplementary	O
Fig	O
.	O
6a	O
and	O
Table	O
1	O
).	O
	O
Notably	O
,	O
only	O
with	O
the	O
additional	O
point	O
mutation	O
L	B-mutant
(-	I-mutant
49	I-mutant
)	I-mutant
S	I-mutant
present	O
in	O
the	O
β5	O
propeptide	O
could	O
we	O
purify	O
a	O
small	O
amount	O
of	O
the	O
β5	B-mutant
-	I-mutant
D17N	I-mutant
mutant	O
yCP	O
.	O
	O
As	O
determined	O
by	O
crystallographic	O
analysis	O
,	O
this	O
mutant	O
β5	O
subunit	O
was	O
partially	O
processed	O
(	O
Table	O
1	O
)	O
but	O
displayed	O
impaired	O
reactivity	O
towards	O
the	O
proteasome	O
inhibitor	O
carfilzomib	O
compared	O
with	O
the	O
subunits	O
β1	O
and	O
β2	O
,	O
and	O
with	O
WT	O
β5	O
(	O
Supplementary	O
Fig	O
.	O
7a	O
).	O
	O
Even	O
though	O
the	O
β5	B-mutant
-	I-mutant
D17N	I-mutant
pp	O
trans	O
yCP	O
crystal	O
structure	O
appeared	O
identical	O
to	O
the	O
WT	O
yCP	O
(	O
Supplementary	O
Fig	O
.	O
7b	O
),	O
the	O
co	O
-	O
crystal	O
structure	O
with	O
the	O
α	O
′,	O
β	O
′	O
epoxyketone	O
inhibitor	O
carfilzomib	O
visualized	O
only	O
partial	O
occupancy	O
of	O
the	O
ligand	O
in	O
the	O
β5	O
active	O
site	O
(	O
Supplementary	O
Fig	O
.	O
7a	O
).	O
	O
Autolysis	O
and	O
residual	O
catalytic	O
activity	O
of	O
the	O
β5	B-mutant
-	I-mutant
D17N	I-mutant
mutants	O
may	O
originate	O
from	O
the	O
carbonyl	O
group	O
of	O
Asn17	O
,	O
which	O
albeit	O
to	O
a	O
lower	O
degree	O
still	O
can	O
polarize	O
Lys33	O
for	O
the	O
activation	O
of	O
Thr1	O
.	O
	O
In	O
agreement	O
,	O
an	O
E17A	B-mutant
mutant	O
in	O
the	O
proteasomal	O
β	O
-	O
subunit	O
of	O
the	O
archaeon	O
Thermoplasma	O
acidophilum	O
prevents	O
autolysis	O
and	O
catalysis	O
.	O
	O
On	O
the	O
basis	O
of	O
these	O
results	O
,	O
we	O
propose	O
that	O
CPs	O
from	O
all	O
domains	O
of	O
life	O
use	O
a	O
catalytic	O
triad	O
consisting	O
of	O
Thr1	O
,	O
Lys33	O
and	O
Asp	O
/	O
Glu17	O
for	O
both	O
autocatalytic	O
precursor	O
processing	O
and	O
proteolysis	O
(	O
Fig	O
.	O
3d	O
).	O
	O
This	O
model	O
is	O
also	O
consistent	O
with	O
the	O
fact	O
that	O
no	O
defined	O
water	O
molecule	O
is	O
observed	O
in	O
the	O
mature	O
WT	O
proteasomal	O
active	O
site	O
that	O
could	O
shuttle	O
the	O
proton	O
from	O
Thr1Oγ	O
to	O
Thr1NH2	O
.	O
	O
To	O
explore	O
this	O
active	O
-	O
site	O
model	O
further	O
,	O
we	O
exchanged	O
the	O
conserved	O
Asp166	O
residue	O
for	O
Asn	O
in	O
the	O
yeast	O
β5	O
subunit	O
.	O
	O
X	O
-	O
ray	O
data	O
on	O
the	O
β5	B-mutant
-	I-mutant
D166N	I-mutant
mutant	O
indicate	O
that	O
the	O
β5	O
propeptide	O
is	O
hydrolysed	O
,	O
but	O
due	O
to	O
reorientation	O
of	O
Ser129OH	O
,	O
the	O
interaction	O
with	O
Asn166Oδ	O
is	O
disrupted	O
(	O
Supplementary	O
Fig	O
.	O
8a	O
).	O
	O
Soaking	O
the	O
β5	B-mutant
-	I-mutant
D166N	I-mutant
crystals	O
with	O
carfilzomib	O
and	O
MG132	O
resulted	O
in	O
covalent	O
modification	O
of	O
Thr1	O
at	O
high	O
occupancy	O
(	O
Supplementary	O
Fig	O
.	O
8c	O
).	O
	O
In	O
the	O
MG132	O
-	O
bound	O
state	O
,	O
Thr1N	O
is	O
unmodified	O
,	O
and	O
we	O
again	O
observe	O
that	O
Ser129	O
is	O
hydrogen	O
-	O
bonded	O
to	O
a	O
water	O
molecule	O
instead	O
of	O
Asn166	O
.	O
	O
Substitution	O
of	O
the	O
active	O
-	O
site	O
Thr1	O
by	O
Cys	O
	O
Mutation	O
of	O
Thr1	O
to	O
Cys	O
inactivates	O
the	O
20S	O
proteasome	O
from	O
the	O
archaeon	O
T	O
.	O
acidophilum	O
.	O
	O
In	O
yeast	O
,	O
this	O
mutation	O
causes	O
a	O
strong	O
growth	O
defect	O
(	O
Fig	O
.	O
4a	O
and	O
Table	O
1	O
),	O
although	O
the	O
propeptide	O
is	O
hydrolysed	O
,	O
as	O
shown	O
here	O
by	O
its	O
X	O
-	O
ray	O
structure	O
.	O
	O
In	O
one	O
of	O
the	O
two	O
β5	O
subunits	O
,	O
however	O
,	O
we	O
found	O
the	O
cleaved	O
propeptide	O
still	O
bound	O
in	O
the	O
substrate	O
-	O
binding	O
channel	O
(	O
Fig	O
.	O
4c	O
).	O
	O
His	O
(-	O
2	O
)	O
occupies	O
the	O
S2	O
pocket	O
like	O
observed	O
for	O
the	O
β5	B-mutant
-	I-mutant
T1A	I-mutant
-	I-mutant
K81R	I-mutant
mutant	O
,	O
but	O
in	O
contrast	O
to	O
the	O
latter	O
,	O
the	O
propeptide	O
in	O
the	O
T1C	B-mutant
mutant	O
adopts	O
an	O
antiparallel	O
β	O
-	O
sheet	O
conformation	O
as	O
known	O
from	O
inhibitors	O
like	O
MG132	O
(	O
Fig	O
.	O
4c	O
–	O
e	O
and	O
Supplementary	O
Fig	O
.	O
9b	O
).	O
	O
On	O
the	O
basis	O
of	O
the	O
phenotype	O
of	O
the	O
T1C	B-mutant
mutant	O
and	O
the	O
propeptide	O
remnant	O
identified	O
in	O
its	O
active	O
site	O
,	O
we	O
suppose	O
that	O
autolysis	O
is	O
retarded	O
and	O
may	O
not	O
have	O
been	O
completed	O
before	O
crystallization	O
.	O
	O
Despite	O
propeptide	O
hydrolysis	O
,	O
the	O
β5	B-mutant
-	I-mutant
T1C	I-mutant
active	O
site	O
is	O
catalytically	O
inactive	O
(	O
Fig	O
.	O
4b	O
and	O
Supplementary	O
Fig	O
.	O
9a	O
).	O
	O
Moreover	O
,	O
the	O
structural	O
data	O
reveal	O
that	O
the	O
thiol	O
group	O
of	O
Cys1	O
is	O
rotated	O
by	O
74	O
°	O
with	O
respect	O
to	O
the	O
hydroxyl	O
side	O
chain	O
of	O
Thr1	O
(	O
Fig	O
.	O
4f	O
and	O
Supplementary	O
Fig	O
.	O
9b	O
).	O
	O
Consequently	O
,	O
the	O
hydrogen	O
bond	O
bridging	O
the	O
active	O
-	O
site	O
nucleophile	O
and	O
Lys33	O
in	O
WT	O
CPs	O
is	O
broken	O
with	O
Cys1	O
.	O
	O
Notably	O
,	O
the	O
2FO	O
–	O
FC	O
electron	O
-	O
density	O
map	O
of	O
the	O
T1C	B-mutant
mutant	O
also	O
indicates	O
that	O
Lys33NH2	O
is	O
disordered	O
.	O
	O
The	O
benefit	O
of	O
Thr	O
over	O
Ser	O
as	O
the	O
active	O
-	O
site	O
nucleophile	O
	O
All	O
proteasomes	O
strictly	O
employ	O
threonine	O
as	O
the	O
active	O
-	O
site	O
residue	O
instead	O
of	O
serine	O
.	O
	O
To	O
investigate	O
the	O
reason	O
for	O
this	O
singularity	O
,	O
we	O
analysed	O
a	O
β5	B-mutant
-	I-mutant
T1S	I-mutant
mutant	O
,	O
which	O
is	O
viable	O
but	O
suffers	O
from	O
growth	O
defects	O
(	O
Fig	O
.	O
4a	O
and	O
Table	O
1	O
).	O
	O
By	O
contrast	O
,	O
turnover	O
of	O
the	O
substrate	O
Z	O
-	O
GGL	O
-	O
pNA	O
,	O
used	O
to	O
monitor	O
ChT	O
-	O
L	O
activity	O
in	O
situ	O
but	O
in	O
a	O
less	O
quantitative	O
fashion	O
,	O
is	O
not	O
detectably	O
impaired	O
(	O
Supplementary	O
Fig	O
.	O
9a	O
).	O
	O
Crystal	O
structure	O
analysis	O
of	O
the	O
β5	B-mutant
-	I-mutant
T1S	I-mutant
mutant	O
confirmed	O
precursor	O
processing	O
(	O
Fig	O
.	O
4g	O
),	O
and	O
ligand	O
-	O
complex	O
structures	O
with	O
bortezomib	O
and	O
carfilzomib	O
unambiguously	O
corroborated	O
the	O
reactivity	O
of	O
Ser1	O
(	O
Fig	O
.	O
5	O
).	O
	O
However	O
,	O
the	O
apo	O
crystal	O
structure	O
revealed	O
that	O
Ser1Oγ	O
is	O
turned	O
away	O
from	O
the	O
substrate	O
-	O
binding	O
channel	O
(	O
Fig	O
.	O
4g	O
).	O
	O
Because	O
both	O
conformations	O
of	O
Ser1Oγ	O
are	O
hydrogen	O
-	O
bonded	O
to	O
Lys33NH2	O
(	O
Fig	O
.	O
4h	O
),	O
the	O
relay	O
system	O
is	O
capable	O
of	O
hydrolysing	O
peptide	O
substrates	O
,	O
albeit	O
at	O
lower	O
rates	O
compared	O
with	O
Thr1	O
.	O
	O
The	O
active	O
-	O
site	O
residue	O
Thr1	O
is	O
fixed	O
in	O
its	O
position	O
,	O
as	O
its	O
methyl	O
group	O
is	O
engaged	O
in	O
hydrophobic	O
interactions	O
with	O
Thr3	O
and	O
Ala46	O
(	O
Fig	O
.	O
4h	O
).	O
	O
Consequently	O
,	O
the	O
hydroxyl	O
group	O
of	O
Thr1	O
requires	O
no	O
reorientation	O
before	O
substrate	O
cleavage	O
and	O
is	O
thus	O
more	O
catalytically	O
efficient	O
than	O
Ser1	O
.	O
	O
In	O
vitro	O
,	O
the	O
mutant	O
proteasome	O
is	O
less	O
susceptible	O
to	O
proteasome	O
inhibition	O
by	O
bortezomib	O
(	O
3	O
.	O
7	O
-	O
fold	O
)	O
and	O
carfilzomib	O
(	O
1	O
.	O
8	O
-	O
fold	O
;	O
Fig	O
.	O
5	O
).	O
	O
Nevertheless	O
,	O
inhibitor	O
complex	O
structures	O
indicate	O
identical	O
binding	O
modes	O
compared	O
with	O
the	O
WT	O
yCP	O
structures	O
,	O
with	O
the	O
same	O
inhibitors	O
.	O
	O
Notably	O
,	O
the	O
affinity	O
of	O
the	O
tetrapeptide	O
carfilzomib	O
is	O
less	O
impaired	O
,	O
as	O
it	O
is	O
better	O
stabilized	O
in	O
the	O
substrate	O
-	O
binding	O
channel	O
than	O
the	O
dipeptide	O
bortezomib	O
,	O
which	O
lacks	O
a	O
defined	O
P3	O
site	O
and	O
has	O
only	O
a	O
few	O
interactions	O
with	O
the	O
surrounding	O
protein	O
.	O
	O
Considered	O
together	O
,	O
these	O
results	O
provide	O
a	O
plausible	O
explanation	O
for	O
the	O
invariance	O
of	O
threonine	O
as	O
the	O
active	O
-	O
site	O
nucleophile	O
in	O
proteasomes	O
in	O
all	O
three	O
domains	O
of	O
life	O
,	O
as	O
well	O
as	O
in	O
proteasome	O
-	O
like	O
proteases	O
such	O
as	O
HslV	O
(	O
ref	O
.).	O
	O
The	O
β	O
-	O
subunit	O
propeptides	O
,	O
particularly	O
that	O
of	O
β5	O
,	O
are	O
key	O
factors	O
that	O
help	O
drive	O
proper	O
assembly	O
of	O
the	O
CP	O
complex	O
.	O
	O
By	O
contrast	O
,	O
the	O
prosegments	O
of	O
β	O
subunits	O
are	O
dispensable	O
for	O
archaeal	O
proteasome	O
assembly	O
,	O
at	O
least	O
when	O
heterologously	O
expressed	O
in	O
Escherichia	O
coli	O
.	O
	O
In	O
eukaryotes	O
,	O
deletion	O
of	O
or	O
failure	O
to	O
cleave	O
the	O
β1	O
and	O
β2	O
propeptides	O
is	O
well	O
tolerated	O
.	O
	O
These	O
observations	O
highlight	O
the	O
unique	O
function	O
and	O
importance	O
of	O
the	O
β5	O
propeptide	O
as	O
well	O
as	O
the	O
β5	O
active	O
site	O
for	O
maturation	O
and	O
function	O
of	O
the	O
eukaryotic	O
CP	O
.	O
	O
Here	O
we	O
have	O
described	O
the	O
atomic	O
structures	O
of	O
various	O
β5	B-mutant
-	I-mutant
T1A	I-mutant
mutants	O
,	O
which	O
allowed	O
for	O
the	O
first	O
time	O
visualization	O
of	O
the	O
residual	O
β5	O
propeptide	O
.	O
	O
From	O
these	O
data	O
we	O
conclude	O
that	O
only	O
the	O
positioning	O
of	O
Gly	O
(-	O
1	O
)	O
and	O
Thr1	O
as	O
well	O
as	O
the	O
integrity	O
of	O
the	O
proteasomal	O
active	O
site	O
are	O
required	O
for	O
autolysis	O
.	O
	O
In	O
this	O
regard	O
,	O
inappropriate	O
N	O
-	O
acetylation	O
of	O
the	O
Thr1	O
N	O
terminus	O
cannot	O
be	O
removed	O
by	O
Thr1Oγ	O
due	O
to	O
the	O
rotational	O
freedom	O
and	O
flexibility	O
of	O
the	O
acetyl	O
group	O
.	O
	O
The	O
propeptide	O
needs	O
some	O
anchoring	O
in	O
the	O
substrate	O
-	O
binding	O
channel	O
to	O
properly	O
position	O
Gly	O
(-	O
1	O
),	O
but	O
this	O
seems	O
to	O
be	O
independent	O
of	O
the	O
orientation	O
of	O
residue	O
(-	O
2	O
).	O
	O
Autolytic	O
activation	O
of	O
the	O
CP	O
constitutes	O
one	O
of	O
the	O
final	O
steps	O
of	O
proteasome	O
biogenesis	O
,	O
but	O
the	O
trigger	O
for	O
propeptide	O
cleavage	O
had	O
remained	O
enigmatic	O
.	O
	O
Thus	O
,	O
specific	O
protein	O
surroundings	O
can	O
significantly	O
alter	O
the	O
chemical	O
properties	O
of	O
amino	O
acids	O
such	O
as	O
Lys	O
to	O
function	O
as	O
an	O
acid	O
–	O
base	O
catalyst	O
.	O
	O
Consistent	O
with	O
this	O
model	O
,	O
the	O
positively	O
charged	O
Thr1	O
N	O
terminus	O
is	O
engaged	O
in	O
hydrogen	O
bonds	O
with	O
inhibitory	O
compounds	O
like	O
fellutamide	O
B	O
(	O
ref	O
.),	O
α	O
-	O
ketoamides	O
,	O
homobelactosin	O
C	O
(	O
ref	O
.)	O
and	O
salinosporamide	O
A	O
(	O
ref	O
.).	O
	O
Furthermore	O
,	O
opening	O
of	O
the	O
β	O
-	O
lactone	O
compound	O
omuralide	O
by	O
Thr1	O
creates	O
a	O
C3	O
-	O
hydroxyl	O
group	O
,	O
whose	O
proton	O
originates	O
from	O
Thr1NH3	O
+.	O
	O
The	O
resulting	O
uncharged	O
Thr1NH2	O
is	O
hydrogen	O
-	O
bridged	O
to	O
the	O
C3	O
-	O
OH	O
group	O
.	O
	O
In	O
agreement	O
,	O
acetylation	O
of	O
the	O
Thr1	O
N	O
terminus	O
irreversibly	O
blocks	O
hydrolytic	O
activity	O
,	O
and	O
binding	O
of	O
substrates	O
is	O
prevented	O
for	O
steric	O
reasons	O
.	O
	O
By	O
acting	O
as	O
a	O
proton	O
donor	O
during	O
catalysis	O
,	O
the	O
Thr1	O
N	O
terminus	O
may	O
also	O
favour	O
cleavage	O
of	O
substrate	O
peptide	O
bonds	O
(	O
Fig	O
.	O
3d	O
).	O
	O
During	O
autolysis	O
the	O
Thr1	O
N	O
terminus	O
is	O
engaged	O
in	O
a	O
hydroxyoxazolidine	O
ring	O
intermediate	O
(	O
Fig	O
.	O
3d	O
),	O
which	O
is	O
unstable	O
and	O
short	O
-	O
lived	O
.	O
	O
The	O
mutation	O
D166N	B-mutant
lowers	O
the	O
pKa	O
of	O
Thr1N	O
,	O
which	O
is	O
thus	O
more	O
likely	O
to	O
exist	O
in	O
the	O
uncharged	O
deprotonated	O
state	O
(	O
Thr1NH2	O
).	O
	O
Hence	O
,	O
the	O
proteasome	O
can	O
be	O
viewed	O
as	O
having	O
a	O
second	O
triad	O
that	O
is	O
essential	O
for	O
efficient	O
proteolysis	O
.	O
	O
However	O
,	O
owing	O
to	O
Cys	O
being	O
a	O
strong	O
nucleophile	O
,	O
the	O
propeptide	O
can	O
still	O
be	O
cleaved	O
off	O
over	O
time	O
.	O
	O
While	O
only	O
one	O
single	O
turnover	O
is	O
necessary	O
for	O
autolysis	O
,	O
continuous	O
enzymatic	O
activity	O
is	O
required	O
for	O
significant	O
and	O
detectable	O
substrate	O
hydrolysis	O
.	O
	O
Notably	O
,	O
in	O
the	O
Ntn	O
hydrolase	O
penicillin	O
acylase	O
,	O
substitution	O
of	O
the	O
catalytic	O
N	O
-	O
terminal	O
Ser	O
residue	O
by	O
Cys	O
also	O
inactivates	O
the	O
enzyme	O
but	O
still	O
enables	O
precursor	O
processing	O
.	O
	O
To	O
investigate	O
why	O
the	O
CP	O
specifically	O
employs	O
threonine	O
as	O
its	O
active	O
-	O
site	O
residue	O
,	O
we	O
used	O
a	O
β5	B-mutant
-	I-mutant
T1S	I-mutant
mutant	O
of	O
the	O
yCP	O
and	O
characterized	O
it	O
biochemically	O
and	O
structurally	O
.	O
	O
Activity	O
assays	O
with	O
the	O
β5	B-mutant
-	I-mutant
T1S	I-mutant
mutant	O
revealed	O
reduced	O
turnover	O
of	O
Suc	O
-	O
LLVY	O
-	O
AMC	O
.	O
	O
We	O
also	O
observed	O
slightly	O
lower	O
affinity	O
of	O
the	O
β5	B-mutant
-	I-mutant
T1S	I-mutant
mutant	O
yCP	O
for	O
the	O
Food	O
and	O
Drug	O
Administration	O
-	O
approved	O
proteasome	O
inhibitors	O
bortezomib	O
and	O
carfilzomib	O
.	O
	O
In	O
contrast	O
to	O
Thr1	O
,	O
the	O
hydroxyl	O
group	O
of	O
Ser1	O
occupies	O
the	O
position	O
of	O
the	O
Thr1	O
methyl	O
side	O
chain	O
in	O
the	O
WT	O
enzyme	O
,	O
which	O
requires	O
its	O
reorientation	O
relative	O
to	O
the	O
substrate	O
to	O
allow	O
cleavage	O
(	O
Fig	O
.	O
4g	O
,	O
h	O
).	O
	O
Similarly	O
,	O
although	O
the	O
serine	O
mutant	O
is	O
active	O
,	O
threonine	O
is	O
more	O
efficient	O
in	O
the	O
context	O
of	O
the	O
proteasome	O
active	O
site	O
.	O
	O
The	O
greater	O
suitability	O
of	O
threonine	O
for	O
the	O
proteasome	O
active	O
site	O
,	O
which	O
has	O
been	O
noted	O
in	O
biochemical	O
as	O
well	O
as	O
in	O
kinetic	O
studies	O
,	O
constitutes	O
a	O
likely	O
reason	O
for	O
the	O
conservation	O
of	O
the	O
Thr1	O
residue	O
in	O
all	O
proteasomes	O
from	O
bacteria	O
to	O
eukaryotes	O
.	O
	O
Conformation	O
of	O
proteasomal	O
propeptides	O
.	O
	O
(	O
a	O
)	O
Structural	O
superposition	O
of	O
the	O
β1	B-mutant
-	I-mutant
T1A	I-mutant
propeptide	O
and	O
the	O
matured	O
WT	O
β1	O
active	O
-	O
site	O
Thr1	O
.	O
	O
Only	O
the	O
residues	O
(-	O
5	O
)	O
to	O
(-	O
1	O
)	O
of	O
the	O
β1	B-mutant
-	I-mutant
T1A	I-mutant
propeptide	O
are	O
displayed	O
.	O
	O
(	O
b	O
)	O
Structural	O
superposition	O
of	O
the	O
β1	B-mutant
-	I-mutant
T1A	I-mutant
propeptide	O
and	O
the	O
β2	B-mutant
-	I-mutant
T1A	I-mutant
propeptide	O
highlights	O
subtle	O
differences	O
in	O
their	O
conformations	O
,	O
but	O
illustrates	O
that	O
Ala1	O
and	O
Gly	O
(-	O
1	O
)	O
match	O
well	O
.	O
	O
Thr	O
(-	O
2	O
)	O
OH	O
is	O
hydrogen	O
-	O
bonded	O
to	O
Gly	O
(-	O
1	O
)	O
O	O
(∼	O
2	O
.	O
8	O
Å	O
;	O
black	O
dashed	O
line	O
).	O
	O
(	O
c	O
)	O
Structural	O
superposition	O
of	O
the	O
β1	B-mutant
-	I-mutant
T1A	I-mutant
,	O
the	O
β2	B-mutant
-	I-mutant
T1A	I-mutant
and	O
the	O
β5	B-mutant
-	I-mutant
T1A	I-mutant
-	I-mutant
K81R	I-mutant
propeptide	O
remnants	O
depict	O
their	O
differences	O
in	O
conformation	O
.	O
	O
Nonetheless	O
,	O
in	O
all	O
mutants	O
the	O
carbonyl	O
carbon	O
atom	O
of	O
Gly	O
(-	O
1	O
)	O
is	O
ideally	O
placed	O
for	O
the	O
nucleophilic	O
attack	O
by	O
Thr1Oγ	O
.	O
	O
The	O
hydrogen	O
bond	O
between	O
Thr	O
(-	O
2	O
)	O
OH	O
and	O
Gly	O
(-	O
1	O
)	O
O	O
(∼	O
2	O
.	O
8	O
Å	O
)	O
is	O
indicated	O
by	O
a	O
black	O
dashed	O
line	O
.	O
	O
Mutations	O
of	O
residue	O
(-	O
2	O
)	O
and	O
their	O
influence	O
on	O
propeptide	O
conformation	O
and	O
autolysis	O
.	O
	O
(	O
a	O
)	O
Structural	O
superposition	O
of	O
the	O
β1	B-mutant
-	I-mutant
T1A	I-mutant
propeptide	O
and	O
the	O
β5	B-mutant
-	I-mutant
H	I-mutant
(-	I-mutant
2	I-mutant
)	I-mutant
L	I-mutant
-	I-mutant
T1A	I-mutant
mutant	O
propeptide	O
.	O
	O
The	O
(-	O
2	O
)	O
residues	O
of	O
both	O
prosegments	O
point	O
into	O
the	O
S1	O
pocket	O
.	O
	O
(	O
c	O
)	O
Structural	O
superposition	O
of	O
the	O
β2	B-mutant
-	I-mutant
T1A	I-mutant
propeptide	O
and	O
the	O
β5	B-mutant
-	I-mutant
H	I-mutant
(-	I-mutant
2	I-mutant
)	I-mutant
T	I-mutant
-	I-mutant
T1A	I-mutant
mutant	O
propeptide	O
.	O
	O
Notably	O
,	O
Val	O
(-	O
2	O
)	O
of	O
the	O
latter	O
does	O
not	O
occupy	O
the	O
S1	O
pocket	O
,	O
thereby	O
changing	O
the	O
orientation	O
of	O
Gly	O
(-	O
1	O
)	O
and	O
preventing	O
nucleophilic	O
attack	O
of	O
Thr1Oγ	O
on	O
the	O
carbonyl	O
carbon	O
atom	O
of	O
Gly	O
(-	O
1	O
).	O
	O
Architecture	O
and	O
proposed	O
reaction	O
mechanism	O
of	O
the	O
proteasomal	O
active	O
site	O
.	O
	O
(	O
a	O
)	O
Hydrogen	O
-	O
bonding	O
network	O
at	O
the	O
mature	O
WT	O
β5	O
proteasomal	O
active	O
site	O
(	O
dotted	O
lines	O
).	O
	O
Thr1OH	O
is	O
hydrogen	O
-	O
bonded	O
to	O
Lys33NH2	O
(	O
2	O
.	O
7	O
Å	O
),	O
which	O
in	O
turn	O
interacts	O
with	O
Asp17Oδ	O
.	O
	O
(	O
c	O
)	O
Structural	O
superposition	O
of	O
the	O
WT	O
β5	O
and	O
the	O
β5	B-mutant
-	I-mutant
K33A	I-mutant
pp	O
trans	O
mutant	O
active	O
site	O
.	O
	O
Similarly	O
to	O
Lys33	O
,	O
the	O
water	O
molecule	O
hydrogen	O
bonds	O
to	O
Arg19O	O
,	O
Asp17Oδ	O
and	O
Thr1OH	O
.	O
	O
(	O
d	O
)	O
Proposed	O
chemical	O
reaction	O
mechanism	O
for	O
autocatalytic	O
precursor	O
processing	O
and	O
proteolysis	O
in	O
the	O
proteasome	O
.	O
	O
The	O
active	O
-	O
site	O
Thr1	O
is	O
depicted	O
in	O
blue	O
,	O
the	O
propeptide	O
segment	O
and	O
the	O
peptide	O
substrate	O
are	O
coloured	O
in	O
green	O
,	O
whereas	O
the	O
scissile	O
peptide	O
bond	O
is	O
highlighted	O
in	O
red	O
.	O
	O
Autolysis	O
(	O
left	O
set	O
of	O
structures	O
)	O
is	O
initiated	O
by	O
deprotonation	O
of	O
Thr1OH	O
via	O
Lys33NH2	O
and	O
the	O
formation	O
of	O
a	O
tetrahedral	O
transition	O
state	O
.	O
	O
Collapse	O
of	O
the	O
transition	O
state	O
frees	O
the	O
Thr1	O
N	O
terminus	O
(	O
by	O
completing	O
an	O
N	O
-	O
to	O
-	O
O	O
acyl	O
shift	O
of	O
the	O
propeptide	O
),	O
which	O
is	O
subsequently	O
protonated	O
by	O
Asp166OH	O
via	O
Ser129OH	O
.	O
	O
The	O
resulting	O
deprotonated	O
Thr1NH2	O
finally	O
activates	O
a	O
water	O
molecule	O
for	O
hydrolysis	O
of	O
the	O
acyl	O
-	O
enzyme	O
.	O
	O
(	O
b	O
)	O
Purified	O
WT	O
and	O
mutant	O
proteasomes	O
were	O
tested	O
for	O
their	O
chymotrypsin	O
-	O
like	O
activity	O
(	O
β5	O
)	O
using	O
the	O
substrate	O
Suc	O
-	O
LLVY	O
-	O
AMC	O
.	O
	O
The	O
prosegment	O
is	O
cleaved	O
but	O
still	O
bound	O
in	O
the	O
substrate	O
-	O
binding	O
channel	O
.	O
	O
Notably	O
,	O
His	O
(-	O
2	O
)	O
does	O
not	O
occupy	O
the	O
S1	O
pocket	O
formed	O
by	O
Met45	O
,	O
similar	O
to	O
what	O
was	O
observed	O
for	O
the	O
β5	B-mutant
-	I-mutant
T1A	I-mutant
-	I-mutant
K81R	I-mutant
mutant	O
.	O
	O
(	O
d	O
)	O
Structural	O
superposition	O
of	O
the	O
β5	B-mutant
-	I-mutant
T1A	I-mutant
-	I-mutant
K81R	I-mutant
and	O
the	O
β5	B-mutant
-	I-mutant
T1C	I-mutant
mutant	O
subunits	O
onto	O
the	O
WT	O
β5	O
subunit	O
.	O
(	O
e	O
)	O
Structural	O
superposition	O
of	O
the	O
β5	B-mutant
-	I-mutant
T1C	I-mutant
propeptide	O
onto	O
the	O
β1	B-mutant
-	I-mutant
T1A	I-mutant
active	O
site	O
(	O
blue	O
)	O
and	O
the	O
WT	O
β5	O
active	O
site	O
in	O
complex	O
with	O
the	O
proteasome	O
inhibitor	O
MG132	O
(	O
ref	O
.).	O
	O
The	O
inhibitor	O
as	O
well	O
as	O
the	O
propeptides	O
adopt	O
similar	O
conformations	O
in	O
the	O
substrate	O
-	O
binding	O
channel	O
.	O
	O
(	O
f	O
)	O
Structural	O
superposition	O
of	O
the	O
WT	O
β5	O
and	O
β5	B-mutant
-	I-mutant
T1C	I-mutant
mutant	O
active	O
sites	O
illustrates	O
the	O
different	O
orientations	O
of	O
the	O
hydroxyl	O
group	O
of	O
Thr1	O
and	O
the	O
thiol	O
side	O
chain	O
of	O
Cys1	O
.	O
	O
(	O
g	O
)	O
Structural	O
superposition	O
of	O
the	O
WT	O
β5	O
and	O
β5	B-mutant
-	I-mutant
T1S	I-mutant
mutant	O
active	O
sites	O
reveals	O
different	O
orientations	O
of	O
the	O
hydroxyl	O
groups	O
of	O
Thr1	O
and	O
Ser1	O
,	O
respectively	O
.	O
	O
The	O
2FO	O
–	O
FC	O
electron	O
-	O
density	O
map	O
for	O
Ser1	O
(	O
blue	O
mesh	O
contoured	O
at	O
1σ	O
)	O
is	O
illustrated	O
.	O
	O
(	O
h	O
)	O
The	O
methyl	O
group	O
of	O
Thr1	O
is	O
anchored	O
by	O
hydrophobic	O
interactions	O
with	O
Ala46Cβ	O
and	O
Thr3Cγ	O
.	O
	O
Inhibition	O
of	O
WT	O
and	O
mutant	O
β5	B-mutant
-	I-mutant
T1S	I-mutant
proteasomes	O
by	O
bortezomib	O
and	O
carfilzomib	O
.	O
	O
Purified	O
yeast	O
proteasomes	O
were	O
tested	O
for	O
the	O
susceptibility	O
of	O
their	O
ChT	O
-	O
L	O
(	O
β5	O
)	O
activity	O
to	O
inhibition	O
by	O
bortezomib	O
and	O
carfilzomib	O
using	O
the	O
substrate	O
Suc	O
-	O
LLVY	O
-	O
AMC	O
.	O
	O
IC50	O
values	O
were	O
determined	O
in	O
triplicate	O
;	O
s	O
.	O
d	O
.'	O
s	O
are	O
indicated	O
by	O
error	O
bars	O
.	O
	O
Note	O
that	O
IC50	O
values	O
depend	O
on	O
time	O
and	O
enzyme	O
concentration	O
.	O
	O
Proteasomes	O
(	O
final	O
concentration	O
:	O
66	O
nM	O
)	O
were	O
incubated	O
with	O
inhibitor	O
for	O
45	O
min	O
before	O
substrate	O
addition	O
(	O
final	O
concentration	O
:	O
200	O
μM	O
).	O
	O
Structures	O
of	O
the	O
β5	B-mutant
-	I-mutant
T1S	I-mutant
mutant	O
in	O
complex	O
with	O
both	O
ligands	O
(	O
green	O
)	O
prove	O
the	O
reactivity	O
of	O
Ser1	O
(	O
right	O
panel	O
).	O
	O
The	O
WT	O
proteasome	O
:	O
inhibitor	O
complex	O
structures	O
(	O
inhibitor	O
in	O
grey	O
;	O
Thr1	O
in	O
black	O
)	O
are	O
superimposed	O
and	O
demonstrate	O
that	O
mutation	O
of	O
Thr1	O
to	O
Ser	O
does	O
not	O
affect	O
the	O
binding	O
mode	O
of	O
bortezomib	O
or	O
carfilzomib	O
.	O
	O
The	O
discovery	O
of	O
new	O
histone	O
modifications	O
is	O
unfolding	O
at	O
startling	O
rates	O
,	O
however	O
,	O
the	O
identification	O
of	O
effectors	O
capable	O
of	O
interpreting	O
these	O
modifications	O
has	O
lagged	O
behind	O
.	O
	O
Here	O
we	O
report	O
the	O
YEATS	O
domain	O
as	O
an	O
effective	O
reader	O
of	O
histone	O
lysine	O
crotonylation	O
–	O
an	O
epigenetic	O
signature	O
associated	O
with	O
active	O
transcription	O
.	O
	O
We	O
show	O
that	O
the	O
Taf14	O
YEATS	O
domain	O
engages	O
crotonyllysine	O
via	O
a	O
unique	O
π	O
-	O
π	O
-	O
π	O
-	O
stacking	O
mechanism	O
and	O
that	O
other	O
YEATS	O
domains	O
have	O
crotonyllysine	O
binding	O
activity	O
.	O
	O
Crotonylation	O
of	O
lysine	O
residues	O
(	O
crotonyllysine	O
,	O
Kcr	O
)	O
has	O
emerged	O
as	O
one	O
of	O
the	O
fundamental	O
histone	O
post	O
-	O
translational	O
modifications	O
(	O
PTMs	O
)	O
found	O
in	O
mammalian	O
chromatin	O
.	O
	O
The	O
crotonyllysine	O
mark	O
on	O
histone	O
H3K18	O
is	O
produced	O
by	O
p300	O
,	O
a	O
histone	O
acetyltransferase	O
also	O
responsible	O
for	O
acetylation	O
of	O
histones	O
.	O
	O
While	O
a	O
number	O
of	O
acetyllysine	O
readers	O
have	O
been	O
identified	O
and	O
characterized	O
,	O
a	O
specific	O
reader	O
of	O
the	O
crotonyllysine	O
mark	O
remains	O
unknown	O
(	O
reviewed	O
in	O
).	O
	O
The	O
family	O
of	O
acetyllysine	O
readers	O
has	O
been	O
expanded	O
with	O
the	O
discovery	O
that	O
the	O
YEATS	O
(	O
Yaf9	O
,	O
ENL	O
,	O
AF9	O
,	O
Taf14	O
,	O
Sas5	O
)	O
domains	O
of	O
human	O
AF9	O
and	O
yeast	O
Taf14	O
are	O
capable	O
of	O
recognizing	O
the	O
histone	O
mark	O
H3K9ac	O
.	O
	O
The	O
acetyllysine	O
binding	O
function	O
of	O
the	O
AF9	O
YEATS	O
domain	O
is	O
essential	O
for	O
the	O
recruitment	O
of	O
the	O
histone	O
methyltransferase	O
DOT1L	O
to	O
H3K9ac	O
-	O
containing	O
chromatin	O
and	O
for	O
DOT1L	O
-	O
mediated	O
H3K79	O
methylation	O
and	O
transcription	O
.	O
	O
Similarly	O
,	O
activation	O
of	O
a	O
subset	O
of	O
genes	O
and	O
DNA	O
damage	O
repair	O
in	O
yeast	O
require	O
the	O
acetyllysine	O
binding	O
activity	O
of	O
the	O
Taf14	O
YEATS	O
domain	O
.	O
	O
Consistent	O
with	O
its	O
role	O
in	O
gene	O
regulation	O
,	O
Taf14	O
was	O
identified	O
as	O
a	O
core	O
component	O
of	O
the	O
transcription	O
factor	O
complexes	O
TFIID	O
and	O
TFIIF	O
.	O
	O
We	O
found	O
that	O
H3K9cr	O
is	O
present	O
in	O
yeast	O
and	O
is	O
dynamically	O
regulated	O
.	O
	O
To	O
elucidate	O
the	O
molecular	O
basis	O
for	O
recognition	O
of	O
the	O
H3K9cr	O
mark	O
,	O
we	O
obtained	O
a	O
crystal	O
structure	O
of	O
the	O
Taf14	O
YEATS	O
domain	O
in	O
complex	O
with	O
H3K9cr5	O
-	O
13	O
(	O
residues	O
5	O
–	O
13	O
of	O
H3	O
)	O
peptide	O
(	O
Fig	O
.	O
1	O
,	O
Supplementary	O
Results	O
,	O
Supplementary	O
Fig	O
.	O
1	O
and	O
Supplementary	O
Table	O
1	O
).	O
	O
The	O
Taf14	O
YEATS	O
domain	O
adopts	O
an	O
immunoglobin	O
-	O
like	O
β	O
sandwich	O
fold	O
containing	O
eight	O
anti	O
-	O
parallel	O
β	O
strands	O
linked	O
by	O
short	O
loops	O
that	O
form	O
a	O
binding	O
site	O
for	O
H3K9cr	O
(	O
Fig	O
.	O
1b	O
).	O
	O
The	O
H3K9cr	O
peptide	O
lays	O
in	O
an	O
extended	O
conformation	O
in	O
an	O
orientation	O
orthogonal	O
to	O
the	O
β	O
strands	O
and	O
is	O
stabilized	O
through	O
an	O
extensive	O
network	O
of	O
direct	O
and	O
water	O
-	O
mediated	O
hydrogen	O
bonds	O
and	O
a	O
salt	O
bridge	O
(	O
Fig	O
.	O
1c	O
).	O
	O
The	O
fully	O
extended	O
side	O
chain	O
of	O
K9cr	O
transverses	O
the	O
narrow	O
tunnel	O
,	O
crossing	O
the	O
β	O
sandwich	O
at	O
right	O
angle	O
in	O
a	O
corkscrew	O
-	O
like	O
manner	O
(	O
Fig	O
.	O
1b	O
and	O
Supplementary	O
Figure	O
1b	O
).	O
	O
The	O
planar	O
crotonyl	O
group	O
is	O
inserted	O
between	O
Trp81	O
and	O
Phe62	O
of	O
the	O
protein	O
,	O
the	O
aromatic	O
rings	O
of	O
which	O
are	O
positioned	O
strictly	O
parallel	O
to	O
each	O
other	O
and	O
at	O
equal	O
distance	O
from	O
the	O
crotonyl	O
group	O
,	O
yielding	O
a	O
novel	O
aromatic	O
-	O
amide	O
/	O
aliphatic	O
-	O
aromatic	O
π	O
-	O
π	O
-	O
π	O
-	O
stacking	O
system	O
that	O
,	O
to	O
our	O
knowledge	O
,	O
has	O
not	O
been	O
reported	O
previously	O
for	O
any	O
protein	O
-	O
protein	O
interaction	O
(	O
Fig	O
.	O
1d	O
and	O
Supplementary	O
Fig	O
.	O
1c	O
).	O
	O
The	O
side	O
chain	O
of	O
Trp81	O
appears	O
to	O
adopt	O
two	O
conformations	O
,	O
one	O
of	O
which	O
provides	O
maximum	O
π	O
-	O
stacking	O
with	O
the	O
alkene	O
functional	O
group	O
while	O
the	O
other	O
rotamer	O
affords	O
maximum	O
π	O
-	O
stacking	O
with	O
the	O
amide	O
π	O
electrons	O
(	O
Supplementary	O
Fig	O
.	O
1c	O
).	O
	O
The	O
dual	O
conformation	O
of	O
Trp81	O
is	O
likely	O
due	O
to	O
the	O
conjugated	O
nature	O
of	O
the	O
C	O
=	O
C	O
and	O
C	O
=	O
O	O
π	O
-	O
orbitals	O
within	O
the	O
crotonyl	O
functional	O
group	O
.	O
	O
In	O
addition	O
to	O
π	O
-	O
π	O
-	O
π	O
stacking	O
,	O
the	O
crotonyl	O
group	O
is	O
stabilized	O
by	O
a	O
set	O
of	O
hydrogen	O
bonds	O
and	O
electrostatic	O
interactions	O
.	O
	O
This	O
provides	O
the	O
capability	O
for	O
the	O
alkene	O
moiety	O
to	O
form	O
electrostatic	O
contacts	O
,	O
as	O
Cα	O
and	O
Cβ	O
lay	O
within	O
electrostatic	O
interaction	O
distances	O
of	O
the	O
carbonyl	O
oxygen	O
of	O
Gln79	O
and	O
of	O
the	O
hydroxyl	O
group	O
of	O
Thr61	O
,	O
respectively	O
.	O
	O
The	O
hydroxyl	O
group	O
of	O
Thr61	O
also	O
participates	O
in	O
a	O
hydrogen	O
bond	O
with	O
the	O
amide	O
nitrogen	O
of	O
the	O
K9cr	O
side	O
chain	O
(	O
Fig	O
.	O
1d	O
).	O
	O
The	O
fixed	O
position	O
of	O
the	O
Thr61	O
hydroxyl	O
group	O
,	O
which	O
facilitates	O
interactions	O
with	O
both	O
the	O
amide	O
and	O
Cα	O
of	O
K9cr	O
,	O
is	O
achieved	O
through	O
a	O
hydrogen	O
bond	O
with	O
imidazole	O
ring	O
of	O
His59	O
.	O
	O
Extra	O
stabilization	O
of	O
K9cr	O
is	O
attained	O
by	O
a	O
hydrogen	O
bond	O
formed	O
between	O
its	O
carbonyl	O
oxygen	O
and	O
the	O
backbone	O
nitrogen	O
of	O
Trp81	O
,	O
as	O
well	O
as	O
a	O
water	O
-	O
mediated	O
hydrogen	O
bond	O
with	O
the	O
backbone	O
carbonyl	O
group	O
of	O
Gly82	O
(	O
Fig	O
1d	O
).	O
	O
Binding	O
of	O
the	O
Taf14	O
YEATS	O
domain	O
to	O
H3K9cr	O
is	O
robust	O
.	O
	O
This	O
value	O
is	O
in	O
the	O
range	O
of	O
binding	O
affinities	O
exhibited	O
by	O
the	O
majority	O
of	O
histone	O
readers	O
,	O
thus	O
attesting	O
to	O
the	O
physiological	O
relevance	O
of	O
the	O
H3K9cr	O
recognition	O
by	O
Taf14	O
.	O
	O
Both	O
H3K9cr	O
and	O
H3K9ac	O
were	O
detected	O
in	O
yeast	O
histones	O
;	O
to	O
our	O
knowledge	O
,	O
this	O
is	O
the	O
first	O
report	O
of	O
H3K9cr	O
occurring	O
in	O
yeast	O
.	O
	O
As	O
shown	O
in	O
Figure	O
2a	O
,	O
b	O
and	O
Supplementary	O
Fig	O
.	O
3e	O
,	O
H3K9cr	O
levels	O
were	O
abolished	O
or	O
reduced	O
considerably	O
in	O
the	O
HAT	O
deletion	O
strains	O
,	O
whereas	O
they	O
were	O
dramatically	O
increased	O
in	O
the	O
HDAC	O
deletion	O
strains	O
.	O
	O
We	O
have	O
previously	O
shown	O
that	O
among	O
acetylated	O
histone	O
marks	O
,	O
the	O
Taf14	O
YEATS	O
domain	O
prefers	O
acetylated	O
H3K9	O
(	O
also	O
see	O
Supplementary	O
Fig	O
.	O
3b	O
),	O
however	O
it	O
binds	O
to	O
H3K9cr	O
tighter	O
.	O
	O
Binding	O
of	O
H3K9cr	O
induced	O
resonance	O
changes	O
in	O
slow	O
exchange	O
regime	O
on	O
the	O
NMR	O
time	O
scale	O
,	O
indicative	O
of	O
strong	O
interaction	O
.	O
	O
Furthermore	O
,	O
crosspeaks	O
of	O
Gly80	O
and	O
Trp81	O
of	O
the	O
YEATS	O
domain	O
were	O
uniquely	O
perturbed	O
by	O
H3K9cr	O
and	O
H3K9ac	O
,	O
indicating	O
a	O
different	O
chemical	O
environment	O
in	O
the	O
respective	O
crotonyllysine	O
and	O
acetyllysine	O
binding	O
pockets	O
(	O
Supplementary	O
Fig	O
.	O
4a	O
).	O
	O
These	O
differences	O
support	O
our	O
model	O
that	O
Trp81	O
adopts	O
two	O
conformations	O
upon	O
complex	O
formation	O
with	O
the	O
H3K9cr	O
mark	O
as	O
compared	O
to	O
H3K9ac	O
(	O
Supplementary	O
Figs	O
.	O
1c	O
,	O
d	O
and	O
4c	O
).	O
	O
One	O
of	O
the	O
conformations	O
,	O
characterized	O
by	O
the	O
π	O
stacking	O
involving	O
two	O
aromatic	O
residues	O
and	O
the	O
alkene	O
group	O
,	O
is	O
observed	O
only	O
in	O
the	O
YEATS	O
-	O
H3K9cr	O
complex	O
.	O
	O
To	O
establish	O
whether	O
the	O
Taf14	O
YEATS	O
domain	O
is	O
able	O
to	O
recognize	O
other	O
recently	O
identified	O
acyllysine	O
marks	O
,	O
we	O
performed	O
solution	O
pull	O
-	O
down	O
assays	O
using	O
H3	O
peptides	O
acetylated	O
,	O
propionylated	O
,	O
butyrylated	O
,	O
and	O
crotonylated	O
at	O
lysine	O
9	O
(	O
residues	O
1	O
–	O
20	O
of	O
H3	O
).	O
	O
As	O
shown	O
in	O
Figure	O
2d	O
and	O
Supplementary	O
Fig	O
.	O
5a	O
,	O
the	O
Taf14	O
YEATS	O
domain	O
binds	O
more	O
strongly	O
to	O
H3K9cr1	O
-	O
20	O
,	O
as	O
compared	O
to	O
other	O
acylated	O
histone	O
peptides	O
.	O
	O
Addition	O
of	O
H3K9ac1	O
-	O
20	O
,	O
H3K9pr1	O
-	O
20	O
,	O
and	O
H3K9bu1	O
-	O
20	O
peptides	O
caused	O
chemical	O
shift	O
perturbations	O
in	O
the	O
Taf14	O
YEATS	O
domain	O
in	O
intermediate	O
exchange	O
regime	O
,	O
implying	O
that	O
these	O
interactions	O
are	O
weaker	O
compared	O
to	O
the	O
interaction	O
with	O
the	O
H3K9cr1	O
-	O
20	O
peptide	O
(	O
Supplementary	O
Fig	O
.	O
5b	O
).	O
	O
We	O
concluded	O
that	O
H3K9cr	O
is	O
the	O
preferred	O
target	O
of	O
this	O
domain	O
.	O
	O
From	O
comparative	O
structural	O
analysis	O
of	O
the	O
YEATS	O
complexes	O
,	O
Gly80	O
emerged	O
as	O
candidate	O
residue	O
potentially	O
responsible	O
for	O
the	O
preference	O
for	O
crotonyllysine	O
.	O
	O
In	O
attempt	O
to	O
generate	O
a	O
mutant	O
capable	O
of	O
accommodating	O
a	O
short	O
acetyl	O
moiety	O
but	O
discriminating	O
against	O
a	O
longer	O
,	O
planar	O
crotonyl	O
moiety	O
,	O
we	O
mutated	O
Gly80	O
to	O
more	O
bulky	O
residues	O
,	O
however	O
all	O
mutants	O
of	O
Gly80	O
lost	O
their	O
binding	O
activities	O
towards	O
either	O
acylated	O
peptide	O
,	O
suggesting	O
that	O
Gly80	O
is	O
absolutely	O
required	O
for	O
the	O
interaction	O
.	O
	O
In	O
contrast	O
,	O
mutation	O
of	O
Val24	O
,	O
a	O
residue	O
located	O
on	O
another	O
side	O
of	O
Trp81	O
,	O
had	O
no	O
effect	O
on	O
binding	O
(	O
Fig	O
.	O
2d	O
and	O
Supplementary	O
Fig	O
.	O
5a	O
,	O
c	O
).	O
	O
We	O
found	O
that	O
all	O
YEATS	O
domains	O
tested	O
are	O
capable	O
of	O
binding	O
to	O
crotonyllysine	O
peptides	O
,	O
though	O
they	O
display	O
variable	O
preferences	O
for	O
the	O
acyl	O
moieties	O
.	O
	O
While	O
YEATS2	O
and	O
ENL	O
showed	O
selectivity	O
for	O
the	O
crotonylated	O
peptides	O
,	O
GAS41	O
and	O
AF9	O
bound	O
acylated	O
peptides	O
almost	O
equally	O
well	O
.	O
	O
Unlike	O
the	O
YEATS	O
domain	O
,	O
a	O
known	O
acetyllysine	O
reader	O
,	O
bromodomain	O
,	O
does	O
not	O
recognize	O
crotonyllysine	O
.	O
	O
We	O
assayed	O
a	O
large	O
set	O
of	O
BDs	O
in	O
pull	O
-	O
down	O
experiments	O
and	O
found	O
that	O
this	O
module	O
is	O
highly	O
specific	O
for	O
acetyllysine	O
and	O
propionyllysine	O
containing	O
peptides	O
(	O
Supplementary	O
Fig	O
.	O
7	O
).	O
	O
However	O
,	O
bromodomains	O
did	O
not	O
interact	O
(	O
or	O
associated	O
very	O
weakly	O
)	O
with	O
longer	O
acyl	O
modifications	O
,	O
including	O
crotonyllysine	O
,	O
as	O
in	O
the	O
case	O
of	O
BDs	O
of	O
TAF1	O
and	O
BRD2	O
,	O
supporting	O
recent	O
reports	O
.	O
	O
In	O
conclusion	O
,	O
we	O
have	O
identified	O
the	O
YEATS	O
domain	O
of	O
Taf14	O
as	O
the	O
first	O
reader	O
of	O
histone	O
crotonylation	O
.	O
	O
As	O
we	O
previously	O
showed	O
the	O
importance	O
of	O
acyllysine	O
binding	O
by	O
the	O
Taf14	O
YEATS	O
domain	O
for	O
the	O
DNA	O
damage	O
response	O
and	O
gene	O
transcription	O
,	O
it	O
will	O
be	O
essential	O
in	O
the	O
future	O
to	O
define	O
the	O
physiological	O
role	O
of	O
crotonyllysine	O
recognition	O
and	O
to	O
differentiate	O
the	O
activities	O
of	O
Taf14	O
that	O
are	O
due	O
to	O
binding	O
to	O
crotonyllysine	O
and	O
acetyllysine	O
modifications	O
.	O
	O
Furthermore	O
,	O
the	O
functional	O
significance	O
of	O
crotonyllysine	O
recognition	O
by	O
other	O
YEATS	O
proteins	O
will	O
be	O
of	O
great	O
importance	O
to	O
elucidate	O
and	O
compare	O
.	O
	O
The	O
structural	O
mechanism	O
for	O
the	O
recognition	O
of	O
H3K9cr	O
	O
(	O
a	O
)	O
Chemical	O
structure	O
of	O
crotonyllysine	O
.	O
(	O
b	O
)	O
The	O
crystal	O
structure	O
of	O
the	O
Taf14	O
YEATS	O
domain	O
(	O
wheat	O
)	O
in	O
complex	O
with	O
the	O
H3K9cr5	O
-	O
13	O
peptide	O
(	O
green	O
).	O
(	O
c	O
)	O
H3K9cr	O
is	O
stabilized	O
via	O
an	O
extensive	O
network	O
of	O
intermolecular	O
electrostatic	O
and	O
polar	O
interactions	O
with	O
the	O
Taf14	O
YEATS	O
domain	O
.	O
	O
(	O
d	O
)	O
The	O
π	O
-	O
π	O
-	O
π	O
stacking	O
mechanism	O
involving	O
the	O
alkene	O
moiety	O
of	O
crotonyllysine	O
.	O
	O
(	O
a	O
,	O
b	O
)	O
Western	O
blot	O
analysis	O
comparing	O
the	O
levels	O
of	O
H3K9cr	O
and	O
H3K9ac	O
in	O
wild	O
type	O
(	O
WT	O
),	O
HAT	O
deletion	O
,	O
or	O
HDAC	O
deletion	O
yeast	O
strains	O
.	O
	O
(	O
c	O
)	O
Superimposed	O
1H	O
,	O
15N	O
HSQC	O
spectra	O
of	O
Taf14	O
YEATS	O
recorded	O
as	O
H3K9cr5	O
-	O
13	O
and	O
H3K9ac5	O
-	O
13	O
peptides	O
were	O
titrated	O
in	O
.	O
	O
Spectra	O
are	O
color	O
coded	O
according	O
to	O
the	O
protein	O
:	O
peptide	O
molar	O
ratio	O
.	O
	O
(	O
d	O
)	O
Western	O
blot	O
analyses	O
of	O
peptide	O
pull	O
-	O
down	O
assays	O
using	O
wild	O
-	O
type	O
and	O
mutated	O
Taf14	O
YEATS	O
domains	O
and	O
indicated	O
peptides	O
.	O
	O
In	O
this	O
data	O
article	O
,	O
we	O
report	O
the	O
solution	O
NMR	O
-	O
derived	O
structure	O
of	O
the	O
Tom1	O
GAT	O
domain	O
.	O
	O
The	O
estimated	O
protein	O
structure	O
exhibits	O
a	O
bundle	O
of	O
three	O
helical	O
elements	O
.	O
	O
We	O
compare	O
the	O
Tom1	O
GAT	O
structure	O
with	O
those	O
structures	O
corresponding	O
to	O
the	O
Tollip	O
TBD	O
-	O
and	O
ubiquitin	O
-	O
bound	O
states	O
.	O
	O
NMR	O
data	O
was	O
recorded	O
using	O
a	O
Bruker	O
800	O
MHz	O
Data	O
format	O
PDB	O
format	O
text	O
file	O
.	O
	O
Analyzed	O
by	O
CS	O
-	O
Rosetta	O
,	O
Protein	O
Structure	O
Validation	O
Server	O
(	O
PSVS	O
),	O
NMRPipe	O
,	O
NMRDraw	O
,	O
and	O
PyMol	O
Experimental	O
factors	O
Recombinant	O
human	O
Tom1	O
GAT	O
domain	O
was	O
purified	O
to	O
homogeneity	O
before	O
use	O
Experimental	O
features	O
Solution	O
structure	O
of	O
Tom1	O
GAT	O
was	O
determined	O
from	O
NMR	O
chemical	O
shift	O
data	O
Data	O
source	O
location	O
Virginia	O
and	O
Colorado	O
,	O
United	O
States	O
.	O
	O
Tom1	O
GAT	O
structural	O
data	O
is	O
publicly	O
available	O
in	O
the	O
RCSB	O
Protein	O
Data	O
Bank	O
(	O
http	O
://	O
www	O
.	O
rscb	O
.	O
org	O
/)	O
under	O
the	O
accession	O
number	O
PDB	O
:	O
2n9d	O
	O
Tom1	O
GAT	O
can	O
adopt	O
distinct	O
conformations	O
upon	O
ligand	O
binding	O
.	O
	O
Unlike	O
ubiquitin	O
binding	O
,	O
data	O
suggest	O
that	O
conformational	O
changes	O
of	O
the	O
Tom1	O
GAT	O
α	O
-	O
helices	O
1	O
and	O
2	O
occur	O
upon	O
Tollip	O
TBD	O
binding	O
(	O
Fig	O
.	O
3A	O
,	O
B	O
).	O
	O
Representative	O
far	O
-	O
UV	O
CD	O
spectrum	O
of	O
the	O
His	O
-	O
Tom1	O
GAT	O
domain	O
.	O
	O
(	O
A	O
)	O
Stereo	O
view	O
displaying	O
the	O
best	O
-	O
fit	O
backbone	O
superposition	O
of	O
the	O
refined	O
structures	O
for	O
the	O
Tom1	O
GAT	O
domain	O
.	O
	O
(	O
A	O
)	O
Two	O
views	O
of	O
the	O
superimposed	O
structures	O
of	O
the	O
Tom1	O
GAT	O
domain	O
in	O
the	O
free	O
state	O
(	O
gray	O
)	O
with	O
that	O
in	O
the	O
Tollip	O
TBD	O
-	O
bound	O
state	O
(	O
red	O
).	O
(	O
B	O
)	O
Two	O
views	O
of	O
the	O
superimposed	O
structures	O
of	O
the	O
Tom1	O
GAT	O
domain	O
(	O
gray	O
)	O
with	O
that	O
in	O
the	O
Ub	O
-	O
bound	O
state	O
(	O
green	O
).	O
	O
NMR	O
and	O
refinement	O
statistics	O
for	O
the	O
Tom1	O
GAT	O
domain	O
.	O
	O
Haem	O
-	O
dependent	O
dimerization	O
of	O
PGRMC1	O
/	O
Sigma	O
-	O
2	O
receptor	O
facilitates	O
cancer	O
proliferation	O
and	O
chemoresistance	O
	O
Progesterone	O
-	O
receptor	O
membrane	O
component	O
1	O
(	O
PGRMC1	O
/	O
Sigma	O
-	O
2	O
receptor	O
)	O
is	O
a	O
haem	O
-	O
containing	O
protein	O
that	O
interacts	O
with	O
epidermal	O
growth	O
factor	O
receptor	O
(	O
EGFR	O
)	O
and	O
cytochromes	O
P450	O
to	O
regulate	O
cancer	O
proliferation	O
and	O
chemoresistance	O
;	O
its	O
structural	O
basis	O
remains	O
unknown	O
.	O
	O
Here	O
crystallographic	O
analyses	O
of	O
the	O
PGRMC1	O
cytosolic	O
domain	O
at	O
1	O
.	O
95	O
Å	O
resolution	O
reveal	O
that	O
it	O
forms	O
a	O
stable	O
dimer	O
through	O
stacking	O
interactions	O
of	O
two	O
protruding	O
haem	O
molecules	O
.	O
	O
The	O
haem	O
iron	O
is	O
five	O
-	O
coordinated	O
by	O
Tyr113	O
,	O
and	O
the	O
open	O
surface	O
of	O
the	O
haem	O
mediates	O
dimerization	O
.	O
	O
Carbon	O
monoxide	O
(	O
CO	O
)	O
interferes	O
with	O
PGRMC1	O
dimerization	O
by	O
binding	O
to	O
the	O
sixth	O
coordination	O
site	O
of	O
the	O
haem	O
.	O
	O
Haem	O
-	O
mediated	O
PGRMC1	O
dimerization	O
is	O
required	O
for	O
interactions	O
with	O
EGFR	O
and	O
cytochromes	O
P450	O
,	O
cancer	O
proliferation	O
and	O
chemoresistance	O
against	O
anti	O
-	O
cancer	O
drugs	O
;	O
these	O
events	O
are	O
attenuated	O
by	O
either	O
CO	O
or	O
haem	O
deprivation	O
in	O
cancer	O
cells	O
.	O
	O
This	O
study	O
demonstrates	O
protein	O
dimerization	O
via	O
haem	O
–	O
haem	O
stacking	O
,	O
which	O
has	O
not	O
been	O
seen	O
in	O
eukaryotes	O
,	O
and	O
provides	O
insights	O
into	O
its	O
functional	O
significance	O
in	O
cancer	O
.	O
	O
PGRMC1	O
binds	O
to	O
EGFR	O
and	O
cytochromes	O
P450	O
,	O
and	O
is	O
known	O
to	O
be	O
involved	O
in	O
cancer	O
proliferation	O
and	O
in	O
drug	O
resistance	O
.	O
	O
Here	O
,	O
the	O
authors	O
determine	O
the	O
structure	O
of	O
the	O
cytosolic	O
domain	O
of	O
PGRMC1	O
,	O
which	O
forms	O
a	O
dimer	O
via	O
haem	O
–	O
haem	O
stacking	O
,	O
and	O
propose	O
how	O
this	O
interaction	O
could	O
be	O
involved	O
in	O
its	O
function	O
.	O
	O
Increased	O
dietary	O
intake	O
of	O
haem	O
is	O
a	O
risk	O
factor	O
for	O
several	O
types	O
of	O
cancer	O
.	O
	O
Previous	O
studies	O
showed	O
that	O
deprivation	O
of	O
iron	O
or	O
haem	O
suppresses	O
tumourigenesis	O
.	O
	O
On	O
the	O
other	O
hand	O
,	O
carbon	O
monoxide	O
(	O
CO	O
),	O
the	O
gaseous	O
mediator	O
generated	O
by	O
oxidative	O
degradation	O
of	O
haem	O
via	O
haem	O
oxygenase	O
(	O
HO	O
),	O
inhibits	O
tumour	O
growth	O
.	O
	O
To	O
gain	O
insight	O
into	O
the	O
underlying	O
mechanisms	O
,	O
we	O
took	O
chemical	O
biological	O
approaches	O
using	O
affinity	O
nanobeads	O
carrying	O
haem	O
and	O
identified	O
progesterone	O
-	O
receptor	O
membrane	O
component	O
1	O
(	O
PGRMC1	O
)	O
as	O
a	O
haem	O
-	O
binding	O
protein	O
from	O
mouse	O
liver	O
extracts	O
(	O
Supplementary	O
Fig	O
.	O
1	O
).	O
	O
PGRMC1	O
is	O
anchored	O
to	O
the	O
cell	O
membrane	O
through	O
the	O
N	O
-	O
terminal	O
transmembrane	O
helix	O
and	O
interacts	O
with	O
epidermal	O
growth	O
factor	O
receptor	O
(	O
EGFR	O
)	O
and	O
cytochromes	O
P450	O
(	O
ref	O
).	O
	O
While	O
PGRMC1	O
is	O
implicated	O
in	O
cell	O
proliferation	O
and	O
cholesterol	O
biosynthesis	O
,	O
the	O
structural	O
basis	O
on	O
which	O
PGRMC1	O
exerts	O
its	O
function	O
remains	O
largely	O
unknown	O
.	O
	O
Here	O
we	O
show	O
that	O
PGRMC1	O
exhibits	O
a	O
unique	O
haem	O
-	O
dependent	O
dimerization	O
.	O
	O
The	O
dimer	O
binds	O
to	O
EGFR	O
and	O
cytochromes	O
P450	O
to	O
enhance	O
tumour	O
cell	O
proliferation	O
and	O
chemoresistance	O
.	O
	O
The	O
dimer	O
is	O
dissociated	O
to	O
monomers	O
by	O
physiological	O
levels	O
of	O
CO	O
,	O
suggesting	O
that	O
PGRMC1	O
serves	O
as	O
a	O
CO	O
-	O
sensitive	O
molecular	O
switch	O
regulating	O
cancer	O
cell	O
proliferation	O
.	O
	O
X	O
-	O
ray	O
crystal	O
structure	O
of	O
PGRMC1	O
	O
We	O
solved	O
the	O
crystal	O
structure	O
of	O
the	O
haem	O
-	O
bound	O
PGRMC1	O
cytosolic	O
domain	O
(	O
a	O
.	O
a	O
.	O
72	O
–	O
195	O
)	O
at	O
1	O
.	O
95	O
Å	O
resolution	O
(	O
Supplementary	O
Fig	O
.	O
2	O
).	O
	O
In	O
the	O
presence	O
of	O
haem	O
,	O
PGRMC1	O
forms	O
a	O
dimeric	O
structure	O
largely	O
through	O
hydrophobic	O
interactions	O
between	O
the	O
haem	O
moieties	O
of	O
two	O
monomers	O
(	O
Fig	O
.	O
1a	O
,	O
Table	O
1	O
and	O
Supplementary	O
Fig	O
.	O
3	O
;	O
a	O
stereo	O
-	O
structural	O
image	O
is	O
shown	O
in	O
Supplementary	O
Fig	O
4	O
).	O
	O
While	O
the	O
overall	O
fold	O
of	O
PGRMC1	O
is	O
similar	O
to	O
that	O
of	O
canonical	O
cytochrome	O
b5	O
,	O
their	O
haem	O
irons	O
are	O
coordinated	O
differently	O
.	O
	O
In	O
cytochrome	O
b5	O
,	O
the	O
haem	O
iron	O
is	O
six	O
-	O
coordinated	O
by	O
two	O
axial	O
histidine	O
residues	O
.	O
	O
A	O
homologous	O
helix	O
that	O
holds	O
haem	O
in	O
cytochrome	O
b5	O
is	O
longer	O
,	O
shifts	O
away	O
from	O
haem	O
,	O
and	O
does	O
not	O
form	O
a	O
coordinate	O
bond	O
in	O
PGRMC1	O
(	O
Fig	O
.	O
1c	O
).	O
	O
Contrary	O
to	O
our	O
finding	O
,	O
Kaluka	O
et	O
al	O
.	O
recently	O
reported	O
that	O
Tyr164	O
of	O
PGRMC1	O
is	O
the	O
axial	O
ligand	O
of	O
haem	O
because	O
mutation	O
of	O
this	O
residue	O
impairs	O
haem	O
binding	O
.	O
	O
Our	O
structural	O
data	O
revealed	O
that	O
Tyr164	O
and	O
a	O
few	O
other	O
residues	O
such	O
as	O
Tyr107	O
and	O
Lys163	O
are	O
in	O
fact	O
hydrogen	O
-	O
bonded	O
to	O
haem	O
propionates	O
.	O
	O
This	O
is	O
consistent	O
with	O
observations	O
by	O
Min	O
et	O
al	O
.	O
that	O
Tyr	O
107	O
and	O
Tyr113	O
of	O
PGRMC1	O
are	O
involved	O
in	O
binding	O
with	O
haem	O
.	O
	O
These	O
amino	O
acid	O
residues	O
are	O
conserved	O
among	O
MAPR	O
family	O
members	O
(	O
Supplementary	O
Fig	O
.	O
5a	O
),	O
suggesting	O
that	O
these	O
proteins	O
share	O
the	O
ability	O
to	O
exhibit	O
haem	O
-	O
dependent	O
dimerization	O
.	O
	O
PGRMC1	O
exhibits	O
haem	O
-	O
dependent	O
dimerization	O
in	O
solution	O
	O
In	O
the	O
PGRMC1	O
crystal	O
,	O
two	O
different	O
types	O
of	O
crystal	O
contacts	O
(	O
chain	O
A	O
–	O
A	O
″	O
and	O
A	O
–	O
B	O
)	O
were	O
observed	O
in	O
addition	O
to	O
the	O
haem	O
-	O
mediated	O
dimer	O
(	O
chain	O
A	O
–	O
A	O
′)	O
(	O
Supplementary	O
Figs	O
3	O
and	O
6a	O
).	O
	O
To	O
confirm	O
that	O
haem	O
-	O
assisted	O
dimerization	O
of	O
PGRMC1	O
occurs	O
in	O
solution	O
,	O
we	O
analysed	O
the	O
structure	O
of	O
apo	O
-	O
and	O
haem	O
-	O
bound	O
PGMRC1	O
by	O
two	O
-	O
dimensional	O
nuclear	O
magnetic	O
resonance	O
(	O
NMR	O
)	O
using	O
heteronuclear	O
single	O
-	O
quantum	O
coherence	O
and	O
transverse	O
relaxation	O
-	O
optimized	O
spectroscopy	O
(	O
Supplementary	O
Figs	O
6b	O
and	O
7	O
).	O
	O
NMR	O
signals	O
from	O
some	O
amino	O
acid	O
residues	O
of	O
PGRMC1	O
disappeared	O
due	O
to	O
the	O
paramagnetic	O
relaxation	O
effect	O
of	O
haem	O
(	O
Supplementary	O
Figs	O
6b	O
);	O
these	O
residues	O
were	O
located	O
in	O
the	O
haem	O
-	O
binding	O
region	O
.	O
	O
When	O
chemical	O
shifts	O
of	O
apo	O
-	O
and	O
haem	O
-	O
bound	O
forms	O
of	O
PGMRC1	O
were	O
compared	O
,	O
some	O
amino	O
acid	O
residues	O
close	O
to	O
those	O
which	O
disappeared	O
because	O
of	O
the	O
paramagnetic	O
relaxation	O
effect	O
of	O
haem	O
exhibit	O
notable	O
chemical	O
shifts	O
(	O
Supplementary	O
Fig	O
.	O
6a	O
,	O
b	O
;	O
dark	O
yellow	O
).	O
	O
We	O
also	O
attempted	O
to	O
predict	O
the	O
secondary	O
structure	O
of	O
PGRMC1	O
through	O
NMR	O
data	O
by	O
calculating	O
with	O
TALOS	O
+	O
program	O
(	O
Supplementary	O
Fig	O
.	O
8	O
);	O
the	O
prediction	O
suggested	O
that	O
the	O
overall	O
secondary	O
structure	O
is	O
comparable	O
between	O
apo	O
-	O
and	O
haem	O
-	O
bound	O
forms	O
of	O
PGRMC1	O
in	O
solution	O
.	O
	O
We	O
analysed	O
the	O
haem	O
-	O
dependent	O
dimerization	O
of	O
the	O
PGRMC1	O
cytosolic	O
domain	O
(	O
a	O
.	O
a	O
.	O
44	O
–	O
195	O
)	O
in	O
solution	O
(	O
Fig	O
.	O
2	O
and	O
Table	O
2	O
).	O
	O
Mass	O
spectrometry	O
(	O
MS	O
)	O
analyses	O
under	O
non	O
-	O
denaturing	O
condition	O
demonstrated	O
that	O
the	O
apo	O
-	O
monomer	O
PGRMC1	O
resulted	O
in	O
dimerization	O
by	O
binding	O
with	O
haem	O
(	O
Fig	O
.	O
2a	O
).	O
	O
This	O
observation	O
led	O
us	O
to	O
examine	O
whether	O
or	O
not	O
the	O
disulfide	O
bond	O
contributes	O
to	O
PGRMC1	O
dimerization	O
.	O
	O
MS	O
analyses	O
under	O
non	O
-	O
denaturing	O
conditions	O
clearly	O
showed	O
that	O
the	O
Cys129Ser	B-mutant
(	O
C129S	B-mutant
)	O
mutant	O
is	O
dimerized	O
in	O
the	O
presence	O
of	O
haem	O
,	O
indicating	O
that	O
the	O
haem	O
-	O
mediated	O
dimerization	O
of	O
PGRMC1	O
occurs	O
independently	O
of	O
the	O
disulfide	O
bond	O
formation	O
via	O
Cys129	O
(	O
Fig	O
.	O
2a	O
).	O
	O
Supporting	O
this	O
,	O
MS	O
analyses	O
under	O
denaturing	O
conditions	O
showed	O
that	O
haem	O
-	O
mediated	O
PGRMC1	O
dimer	O
is	O
completely	O
dissociated	O
into	O
monomer	O
,	O
indicating	O
that	O
dimerization	O
of	O
this	O
kind	O
is	O
not	O
mediated	O
by	O
any	O
covalent	O
bond	O
such	O
as	O
disulfide	O
bond	O
(	O
Supplementary	O
Fig	O
.	O
9	O
).	O
	O
We	O
also	O
analysed	O
the	O
haem	O
-	O
dependent	O
dimerization	O
of	O
PGRMC1	O
by	O
diffusion	O
-	O
ordered	O
NMR	O
spectroscopy	O
(	O
DOSY	O
)	O
analyses	O
(	O
Table	O
2	O
,	O
Supplementary	O
Fig	O
.	O
10	O
).	O
	O
The	O
results	O
suggested	O
that	O
the	O
hydrodynamic	O
radius	O
of	O
haem	O
-	O
bound	O
PGRMC1	O
is	O
larger	O
than	O
that	O
of	O
apo	O
-	O
PGRMC1	O
.	O
	O
To	O
further	O
evaluate	O
changes	O
in	O
molecular	O
weights	O
in	O
dimerization	O
of	O
PGRMC1	O
,	O
sedimentation	O
velocity	O
analytical	O
ultracentrifugation	O
(	O
SV	O
-	O
AUC	O
)	O
analysis	O
was	O
carried	O
out	O
.	O
	O
Whereas	O
the	O
wild	O
-	O
type	O
(	O
wt	O
)	O
apo	O
-	O
PGRMC1	O
appeared	O
at	O
a	O
1	O
.	O
9	O
S	O
peak	O
as	O
monomer	O
,	O
the	O
haem	O
-	O
binding	O
PGRMC1	O
was	O
converted	O
into	O
dimer	O
at	O
a	O
3	O
.	O
1	O
S	O
peak	O
(	O
Fig	O
.	O
2b	O
).	O
	O
Similarly	O
,	O
the	O
C129S	B-mutant
mutant	O
of	O
PGRMC1	O
converted	O
from	O
monomer	O
to	O
dimer	O
by	O
binding	O
to	O
haem	O
(	O
Fig	O
.	O
2b	O
).	O
	O
SV	O
-	O
AUC	O
analyses	O
also	O
allowed	O
us	O
to	O
examine	O
the	O
stability	O
of	O
haem	O
/	O
PGRMC1	O
dimer	O
.	O
	O
To	O
this	O
end	O
,	O
we	O
used	O
different	O
concentrations	O
(	O
3	O
.	O
5	O
–	O
147	O
μmol	O
l	O
−	O
1	O
)	O
of	O
haem	O
-	O
bound	O
PGRMC1	O
protein	O
(	O
a	O
.	O
a	O
.	O
72	O
–	O
195	O
),	O
which	O
were	O
identical	O
to	O
that	O
used	O
in	O
the	O
crystallographic	O
analysis	O
.	O
	O
The	O
sedimentation	O
coefficients	O
calculated	O
on	O
the	O
basis	O
of	O
the	O
crystal	O
structure	O
were	O
1	O
.	O
71	O
S	O
for	O
monomer	O
and	O
2	O
.	O
56	O
S	O
for	O
dimer	O
(	O
Supplementary	O
Fig	O
.	O
11	O
,	O
upper	O
panel	O
).	O
	O
The	O
results	O
showed	O
that	O
the	O
PGRMC1	O
dimer	O
is	O
not	O
dissociated	O
into	O
monomer	O
at	O
all	O
concentrations	O
examined	O
(	O
Supplementary	O
Fig	O
.	O
11	O
,	O
lower	O
panel	O
),	O
suggesting	O
that	O
the	O
Kd	O
value	O
of	O
haem	O
-	O
mediated	O
dimer	O
of	O
PGRMC1	O
is	O
under	O
3	O
.	O
5	O
μmol	O
l	O
−	O
1	O
.	O
	O
We	O
also	O
showed	O
by	O
haem	O
titration	O
experiments	O
that	O
haem	O
binding	O
to	O
PGRMC1	O
was	O
of	O
low	O
affinity	O
with	O
a	O
Kd	O
value	O
of	O
50	O
nmol	O
l	O
−	O
1	O
;	O
this	O
is	O
comparable	O
with	O
that	O
of	O
iron	O
regulatory	O
protein	O
2	O
,	O
which	O
is	O
known	O
to	O
be	O
regulated	O
by	O
intracellular	O
levels	O
of	O
haem	O
(	O
Fig	O
.	O
2c	O
and	O
Supplementary	O
Table	O
1	O
).	O
	O
These	O
results	O
raised	O
the	O
possibility	O
that	O
the	O
function	O
of	O
PGRMC1	O
is	O
regulated	O
by	O
intracellular	O
haem	O
concentrations	O
.	O
	O
CO	O
inhibits	O
haem	O
-	O
dependent	O
dimerization	O
of	O
PGRMC1	O
	O
Crystallographic	O
analyses	O
revealed	O
that	O
Tyr113	O
of	O
PGRMC1	O
is	O
an	O
axial	O
ligand	O
for	O
haem	O
and	O
contributes	O
to	O
haem	O
-	O
dependent	O
dimerization	O
(	O
Fig	O
.	O
1a	O
).	O
	O
Analysis	O
of	O
UV	O
-	O
visible	O
spectra	O
revealed	O
that	O
the	O
heme	O
of	O
PGRMC1	O
is	O
reducible	O
from	O
ferric	O
to	O
ferrous	O
state	O
,	O
thus	O
allowing	O
CO	O
binding	O
(	O
Fig	O
.	O
3a	O
).	O
	O
Analysis	O
of	O
the	O
ferric	O
form	O
of	O
PGRMC1	O
using	O
resonance	O
Raman	O
spectroscopy	O
(	O
Supplementary	O
Fig	O
.	O
13	O
)	O
showed	O
that	O
the	O
relative	O
intensity	O
of	O
oxidation	O
and	O
spin	O
state	O
marker	O
bands	O
(	O
ν4	O
and	O
ν3	O
)	O
is	O
close	O
to	O
1	O
.	O
0	O
,	O
which	O
is	O
consistent	O
with	O
it	O
being	O
a	O
haem	O
protein	O
with	O
a	O
proximal	O
Tyr	O
coordination	O
.	O
	O
A	O
specific	O
Raman	O
shift	O
peaking	O
at	O
vFe	O
–	O
CO	O
=	O
500	O
cm	O
−	O
1	O
demonstrated	O
that	O
the	O
CO	O
-	O
bound	O
haem	O
of	O
PGRMC1	O
is	O
six	O
-	O
coordinated	O
(	O
Supplementary	O
Fig	O
.	O
13	O
).	O
	O
Since	O
PGRMC1	O
dimerization	O
involves	O
the	O
open	O
surface	O
of	O
haem	O
on	O
the	O
opposite	O
side	O
of	O
the	O
axial	O
Tyr113	O
,	O
no	O
space	O
for	O
CO	O
binding	O
is	O
available	O
in	O
the	O
dimeric	O
structure	O
(	O
Fig	O
.	O
3b	O
).	O
	O
This	O
prompted	O
us	O
to	O
ask	O
if	O
CO	O
binding	O
to	O
haem	O
causes	O
dissociation	O
of	O
the	O
PGRMC1	O
dimer	O
.	O
	O
Analysis	O
by	O
gel	O
filtration	O
chromatography	O
revealed	O
that	O
the	O
relative	O
molecular	O
sizes	O
of	O
the	O
wild	O
-	O
type	O
and	O
the	O
C129S	B-mutant
mutant	O
of	O
PGRMC1	O
are	O
increased	O
by	O
adding	O
haem	O
to	O
apo	O
-	O
PGRMC1	O
regardless	O
of	O
the	O
oxidation	O
state	O
of	O
the	O
iron	O
(	O
Fig	O
.	O
3c	O
),	O
which	O
is	O
in	O
agreement	O
with	O
the	O
results	O
in	O
Table	O
1	O
.	O
	O
CO	O
application	O
to	O
ferrous	O
PGRMC1	O
abolished	O
the	O
haem	O
-	O
dependent	O
increase	O
in	O
its	O
molecular	O
size	O
.	O
	O
Under	O
this	O
reducing	O
condition	O
in	O
the	O
presence	O
of	O
dithionite	O
,	O
analyses	O
of	O
UV	O
-	O
visible	O
spectra	O
indicated	O
that	O
CO	O
-	O
binding	O
with	O
haem	O
-	O
PGRMC1	O
is	O
stable	O
,	O
showing	O
only	O
20	O
%	O
reduction	O
of	O
the	O
absorbance	O
at	O
412	O
nm	O
within	O
2	O
h	O
(	O
Supplementary	O
Fig	O
.	O
14	O
).	O
	O
Furthermore	O
,	O
the	O
Tyr113Phe	B-mutant
(	O
Y113F	B-mutant
)	O
mutant	O
of	O
PGRMC1	O
was	O
not	O
responsive	O
to	O
haem	O
.	O
	O
The	O
peak	O
corresponding	O
to	O
the	O
haem	O
/	O
PGRMC1	O
dimer	O
was	O
detected	O
under	O
reducing	O
conditions	O
in	O
the	O
presence	O
of	O
dithionite	O
(	O
Supplementary	O
Fig	O
.	O
15	O
,	O
middle	O
panel	O
).	O
	O
Under	O
these	O
circumstances	O
,	O
CO	O
application	O
induced	O
dissociation	O
of	O
the	O
haem	O
-	O
mediated	O
dimers	O
of	O
PGRMC1	O
to	O
generate	O
a	O
peak	O
of	O
monomers	O
(	O
Supplementary	O
Fig	O
.	O
15	O
,	O
lower	O
panel	O
).	O
	O
These	O
observations	O
raised	O
the	O
transition	O
model	O
for	O
structural	O
regulation	O
of	O
PGRMC1	O
in	O
response	O
to	O
haem	O
(	O
Fig	O
.	O
3d	O
).	O
	O
As	O
mentioned	O
above	O
,	O
apo	O
-	O
PGRMC1	O
exists	O
as	O
monomer	O
.	O
	O
CO	O
induces	O
the	O
dissociation	O
of	O
the	O
haem	O
-	O
mediated	O
dimer	O
of	O
PGRMC1	O
by	O
interfering	O
with	O
the	O
haem	O
-	O
stacking	O
interface	O
via	O
formation	O
of	O
the	O
six	O
-	O
coordinated	O
CO	O
-	O
haem	O
-	O
PGRMC1	O
complex	O
.	O
	O
Such	O
a	O
dynamic	O
structural	O
regulation	O
led	O
us	O
to	O
further	O
examine	O
the	O
regulation	O
of	O
PGRMC1	O
functions	O
in	O
cancer	O
cells	O
.	O
	O
PGRMC1	O
dimerization	O
is	O
required	O
for	O
binding	O
to	O
EGFR	O
	O
As	O
shown	O
in	O
Fig	O
.	O
4a	O
,	O
the	O
cytosolic	O
domain	O
of	O
wild	O
-	O
type	O
PGRMC1	O
,	O
but	O
not	O
the	O
Y113F	B-mutant
mutant	O
,	O
interacted	O
with	O
purified	O
EGFR	O
in	O
a	O
haem	O
-	O
dependent	O
manner	O
.	O
	O
This	O
interaction	O
was	O
disrupted	O
by	O
the	O
ruthenium	O
-	O
based	O
CO	O
-	O
releasing	O
molecule	O
,	O
CORM3	O
,	O
but	O
not	O
by	O
RuCl3	O
as	O
a	O
control	O
reagent	O
(	O
Fig	O
.	O
4b	O
).	O
	O
We	O
further	O
analysed	O
the	O
intracellular	O
interaction	O
between	O
PGRMC1	O
and	O
EGFR	O
.	O
	O
FLAG	O
-	O
tagged	O
PGRMC1	O
ectopically	O
expressed	O
in	O
human	O
colon	O
cancer	O
HCT116	O
cells	O
was	O
immunoprecipitated	O
with	O
anti	O
-	O
FLAG	O
antibody	O
,	O
and	O
co	O
-	O
immunoprecipitated	O
EGFR	O
and	O
endogenous	O
PGRMC1	O
binding	O
to	O
FLAG	O
-	O
PGRMC1	O
were	O
detected	O
by	O
Western	O
blotting	O
(	O
Fig	O
.	O
4c	O
).	O
	O
The	O
C129S	B-mutant
mutant	O
of	O
PGRMC1	O
also	O
interacted	O
with	O
endogenous	O
PGRMC1	O
and	O
EGFR	O
(	O
Supplementary	O
Fig	O
.	O
16	O
).	O
	O
Whereas	O
FLAG	O
-	O
tagged	O
wild	O
-	O
type	O
PGRMC1	O
interacted	O
with	O
endogenous	O
PGRMC1	O
and	O
EGFR	O
,	O
the	O
Y113F	B-mutant
mutant	O
did	O
not	O
.	O
	O
As	O
expected	O
,	O
SA	O
significantly	O
reduced	O
PGRMC1	O
dimerization	O
and	O
its	O
interaction	O
with	O
EGFR	O
(	O
Fig	O
.	O
4e	O
),	O
indicating	O
that	O
haem	O
-	O
mediated	O
dimerization	O
of	O
PGMRC1	O
is	O
critical	O
for	O
its	O
binding	O
to	O
EGFR	O
.	O
	O
PGRMC1	O
dimer	O
facilitates	O
EGFR	O
-	O
mediated	O
cancer	O
growth	O
	O
Next	O
,	O
we	O
investigated	O
the	O
functional	O
significance	O
of	O
PGRMC1	O
dimerization	O
in	O
EGFR	O
signaling	O
.	O
	O
EGF	O
-	O
induced	O
phosphorylations	O
of	O
EGFR	O
and	O
its	O
downstream	O
targets	O
AKT	O
and	O
ERK	O
were	O
decreased	O
by	O
PGRMC1	O
knockdown	O
(	O
PGRMC1	B-mutant
-	I-mutant
KD	I-mutant
)	O
(	O
Fig	O
.	O
4f	O
).	O
	O
Similarly	O
,	O
EGFR	O
signaling	O
was	O
suppressed	O
by	O
treatment	O
of	O
HCT116	O
cells	O
with	O
SA	O
(	O
Fig	O
.	O
4g	O
)	O
or	O
CORM3	O
(	O
Fig	O
.	O
4h	O
).	O
	O
These	O
results	O
suggested	O
that	O
haem	O
-	O
mediated	O
dimerization	O
of	O
PGRMC1	O
is	O
critical	O
for	O
EGFR	O
signaling	O
.	O
	O
To	O
further	O
investigate	O
the	O
role	O
of	O
the	O
dimerized	O
form	O
of	O
PGRMC1	O
in	O
cancer	O
proliferation	O
,	O
we	O
performed	O
PGRMC1	O
knockdown	O
-	O
rescue	O
experiments	O
using	O
FLAG	O
-	O
tagged	O
wild	O
-	O
type	O
and	O
Y113F	B-mutant
PGRMC1	O
expression	O
vectors	O
,	O
in	O
which	O
silent	O
mutations	O
were	O
introduced	O
into	O
the	O
nucleotide	O
sequence	O
targeted	O
by	O
shRNA	O
(	O
Fig	O
.	O
5a	O
).	O
	O
Chemosensitivity	O
enhancement	O
by	O
two	O
different	O
shRNAs	O
to	O
PGRMC1	O
was	O
seen	O
also	O
in	O
HCT116	O
cells	O
and	O
human	O
hepatoma	O
HuH7	O
cells	O
(	O
Supplementary	O
Fig	O
.	O
17	O
).	O
	O
Ten	O
days	O
after	O
intra	O
-	O
splenic	O
implantation	O
of	O
HCT116	O
cells	O
that	O
were	O
genetically	O
tagged	O
with	O
a	O
fluorescent	O
protein	O
Venus	O
,	O
the	O
group	O
implanted	O
with	O
PGRMC1	B-mutant
-	I-mutant
KD	I-mutant
cells	O
showed	O
a	O
significant	O
decrease	O
of	O
liver	O
metastasis	O
in	O
comparison	O
with	O
the	O
control	O
group	O
(	O
Fig	O
.	O
5d	O
).	O
	O
Interaction	O
of	O
PGRMC1	O
dimer	O
with	O
cytochromes	O
P450	O
	O
Since	O
PGRMC1	O
has	O
been	O
shown	O
to	O
interact	O
with	O
cytochromes	O
P450	O
(	O
ref	O
),	O
we	O
investigated	O
whether	O
the	O
haem	O
-	O
mediated	O
dimerization	O
of	O
PGRMC1	O
is	O
necessary	O
for	O
their	O
interactions	O
.	O
	O
Moreover	O
,	O
the	O
interaction	O
of	O
PGRMC1	O
with	O
CYP1A2	O
was	O
blocked	O
by	O
CORM3	O
under	O
reducing	O
conditions	O
(	O
Fig	O
.	O
6c	O
),	O
indicating	O
that	O
PGRMC1	O
dimerization	O
is	O
necessary	O
for	O
its	O
interaction	O
with	O
cytochromes	O
P450	O
.	O
	O
Doxorubicin	O
is	O
an	O
anti	O
-	O
cancer	O
reagent	O
that	O
is	O
metabolized	O
into	O
inactive	O
doxorubicinol	O
by	O
CYP2D6	O
and	O
CYP3A4	O
(	O
Fig	O
.	O
6d	O
).	O
	O
PGRMC1	B-mutant
-	I-mutant
KD	I-mutant
significantly	O
suppressed	O
the	O
conversion	O
of	O
doxorubicin	O
to	O
doxorubicinol	O
(	O
Fig	O
.	O
6d	O
)	O
and	O
increased	O
sensitivity	O
to	O
doxorubicin	O
(	O
Fig	O
.	O
6e	O
).	O
	O
This	O
effect	O
was	O
reversed	O
by	O
co	O
-	O
expression	O
of	O
the	O
wild	O
-	O
type	O
PGRMC1	O
but	O
not	O
of	O
the	O
Y113F	B-mutant
mutant	O
,	O
suggesting	O
that	O
PGRMC1	O
enhances	O
doxorubicin	O
resistance	O
of	O
cancer	O
cells	O
by	O
facilitating	O
its	O
degradation	O
via	O
cytochromes	O
P450	O
.	O
	O
To	O
gain	O
further	O
insight	O
into	O
the	O
interaction	O
between	O
PGRMC1	O
and	O
cytochromes	O
P450	O
,	O
surface	O
plasmon	O
resonance	O
analyses	O
were	O
conducted	O
using	O
recombinant	O
CYP51	O
and	O
PGRMC1	O
.	O
	O
This	O
was	O
based	O
on	O
a	O
previous	O
study	O
showing	O
that	O
PGRMC1	O
binds	O
to	O
CYP51	O
and	O
enhances	O
cholesterol	O
biosynthesis	O
by	O
CYP51	O
(	O
refs	O
).	O
	O
CYP51	O
interacted	O
with	O
PGRMC1	O
in	O
a	O
concentration	O
-	O
dependent	O
manner	O
in	O
the	O
presence	O
of	O
haem	O
,	O
but	O
not	O
in	O
its	O
absence	O
(	O
Supplementary	O
Fig	O
.	O
19	O
),	O
suggesting	O
the	O
requirement	O
for	O
the	O
haem	O
-	O
dependent	O
dimerization	O
of	O
PGRMC1	O
.	O
	O
This	O
is	O
the	O
report	O
showing	O
crystallographic	O
evidence	O
that	O
indicates	O
roles	O
of	O
the	O
direct	O
haem	O
–	O
haem	O
stacking	O
in	O
haem	O
-	O
mediated	O
dimerization	O
in	O
eukaryotes	O
,	O
although	O
a	O
few	O
examples	O
are	O
known	O
in	O
bacteria	O
.	O
	O
Recently	O
,	O
Peluso	O
et	O
al	O
.	O
reported	O
that	O
PGRMC1	O
binds	O
to	O
PGRMC2	O
,	O
suggesting	O
that	O
MAPR	O
family	O
members	O
may	O
also	O
undergo	O
haem	O
-	O
mediated	O
heterodimerization	O
.	O
	O
While	O
the	O
effects	O
of	O
PGRMC1	O
on	O
cholesterol	O
synthesis	O
mediated	O
by	O
CYP51	O
have	O
been	O
well	O
documented	O
in	O
yeast	O
and	O
human	O
cells	O
,	O
it	O
has	O
not	O
been	O
clear	O
whether	O
drug	O
-	O
metabolizing	O
CYP	O
activities	O
are	O
regulated	O
by	O
PGRMC1	O
.	O
	O
Szczesna	O
-	O
Skorupa	O
and	O
Kemper	O
reported	O
that	O
PGRMC1	O
exhibited	O
an	O
inhibitory	O
effect	O
on	O
CYP3A4	O
drug	O
metabolizing	O
activity	O
by	O
competitively	O
binding	O
with	O
cytochrome	O
P450	O
reductase	O
(	O
CPR	O
)	O
in	O
HEK293	O
or	O
HepG2	O
cells	O
.	O
	O
Several	O
other	O
groups	O
showed	O
that	O
PGRMC1	O
enhanced	O
chemoresistance	O
in	O
several	O
cancer	O
cells	O
such	O
as	O
uterine	O
sarcoma	O
,	O
breast	O
cancer	O
,	O
endometrial	O
tumour	O
and	O
ovarian	O
cancer	O
;	O
however	O
,	O
no	O
evidence	O
of	O
PGRMC1	O
-	O
dependent	O
regulation	O
of	O
CYP	O
activity	O
was	O
provided	O
.	O
	O
Our	O
results	O
showed	O
that	O
PGRMC1	O
contributes	O
to	O
enhancement	O
of	O
the	O
doxorubicin	O
metabolism	O
,	O
which	O
is	O
mediated	O
by	O
CYP2D6	O
or	O
CYP3A4	O
in	O
human	O
colon	O
cancer	O
HCT116	O
cells	O
(	O
Fig	O
.	O
6d	O
).	O
	O
While	O
the	O
effects	O
of	O
structural	O
diversity	O
of	O
CYP	O
family	O
proteins	O
and	O
interactions	O
with	O
different	O
xenobiotic	O
substrates	O
should	O
further	O
be	O
examined	O
,	O
the	O
current	O
results	O
suggest	O
that	O
the	O
interaction	O
of	O
drug	O
-	O
metabolizing	O
CYPs	O
with	O
the	O
haem	O
-	O
mediated	O
dimer	O
of	O
PGRMC1	O
plays	O
a	O
crucial	O
role	O
in	O
regulating	O
their	O
activities	O
.	O
	O
We	O
showed	O
that	O
haem	O
-	O
mediated	O
dimerization	O
of	O
PGRMC1	O
enhances	O
proliferation	O
and	O
chemoresistance	O
of	O
cancer	O
cells	O
through	O
binding	O
to	O
and	O
regulating	O
EGFR	O
and	O
cytochromes	O
P450	O
(	O
illustrated	O
in	O
Fig	O
.	O
7	O
).	O
	O
Since	O
the	O
haem	O
-	O
binding	O
affinity	O
of	O
PGRMC1	O
is	O
lower	O
than	O
those	O
of	O
constitutive	O
haem	O
-	O
binding	O
proteins	O
such	O
as	O
myoglobin	O
,	O
PGMRC1	O
is	O
probably	O
interconverted	O
between	O
apo	O
-	O
monomer	O
and	O
haem	O
-	O
bound	O
dimer	O
forms	O
in	O
response	O
to	O
changes	O
in	O
the	O
intracellular	O
haem	O
concentration	O
.	O
	O
Considering	O
microenvironments	O
in	O
and	O
around	O
malignant	O
tumours	O
,	O
the	O
haem	O
concentration	O
in	O
cancer	O
cells	O
is	O
likely	O
to	O
be	O
elevated	O
through	O
multiple	O
mechanisms	O
,	O
such	O
as	O
(	O
i	O
)	O
an	O
increased	O
intake	O
of	O
haem	O
,	O
(	O
ii	O
)	O
mutation	O
of	O
enzymes	O
in	O
TCA	O
cycle	O
(	O
for	O
example	O
,	O
fumarate	O
hydratase	O
)	O
that	O
increases	O
the	O
level	O
of	O
succinyl	O
CoA	O
,	O
a	O
substrate	O
for	O
haem	O
biosynthesis	O
and	O
(	O
iii	O
)	O
metastasis	O
to	O
haem	O
-	O
rich	O
organs	O
such	O
as	O
liver	O
,	O
brain	O
and	O
bone	O
marrow	O
.	O
	O
On	O
the	O
other	O
hand	O
,	O
excessive	O
haem	O
induces	O
HO	O
-	O
1	O
,	O
the	O
enzyme	O
that	O
oxidatively	O
degrades	O
haem	O
and	O
generates	O
CO	O
.	O
	O
Thus	O
,	O
HO	O
-	O
1	O
induction	O
in	O
cancer	O
cells	O
may	O
inhibit	O
the	O
haem	O
-	O
mediated	O
dimerization	O
of	O
PGRMC1	O
through	O
the	O
production	O
of	O
CO	O
and	O
thereby	O
suppress	O
tumour	O
progression	O
.	O
	O
This	O
idea	O
is	O
consistent	O
with	O
the	O
observation	O
that	O
HO	O
-	O
1	O
induction	O
or	O
CO	O
inhibits	O
tumour	O
growth	O
.	O
	O
Furthermore	O
,	O
Sigma	O
-	O
2	O
ligand	O
-	O
binding	O
is	O
decreased	O
in	O
transgenic	O
amyloid	O
beta	O
deposition	O
model	O
APP	O
/	O
PS1	O
female	O
mice	O
.	O
	O
These	O
results	O
suggest	O
a	O
possible	O
involvement	O
of	O
PGRMC1	O
in	O
Alzheimer	O
'	O
s	O
disease	O
.	O
	O
The	O
roles	O
of	O
haem	O
-	O
dependent	O
dimerization	O
of	O
PGRMC1	O
in	O
the	O
functional	O
regulation	O
of	O
its	O
target	O
proteins	O
deserve	O
further	O
studies	O
to	O
find	O
evidence	O
that	O
therapeutic	O
interventions	O
to	O
interfere	O
with	O
the	O
function	O
of	O
the	O
dimer	O
may	O
control	O
varied	O
disease	O
conditions	O
.	O
	O
Two	O
PGRMC1	O
subunits	O
(	O
blue	O
and	O
green	O
ribbons	O
)	O
dimerize	O
via	O
stacking	O
of	O
the	O
haem	O
molecules	O
.	O
	O
(	O
b	O
)	O
Haem	O
coordination	O
of	O
PGRMC1	O
with	O
Tyr113	O
.	O
	O
Comparison	O
of	O
PGRMC1	O
(	O
blue	O
)	O
and	O
cytochrome	O
b5	O
(	O
yellow	O
,	O
ID	O
:	O
3NER	O
).	O
(	O
c	O
)	O
PGRMC1	O
has	O
a	O
longer	O
helix	O
(	O
a	O
.	O
a	O
.	O
147	O
–	O
163	O
),	O
which	O
is	O
shifted	O
away	O
from	O
the	O
haem	O
(	O
arrow	O
).	O
	O
PGRCM1	O
is	O
dimerized	O
by	O
binding	O
with	O
haem	O
.	O
	O
(	O
a	O
)	O
Mass	O
spectrometric	O
analyses	O
of	O
the	O
wild	O
-	O
type	O
(	O
wt	O
)	O
PGRMC1	O
or	O
the	O
C129S	B-mutant
mutant	O
in	O
the	O
presence	O
or	O
absence	O
of	O
haem	O
under	O
non	O
-	O
denaturing	O
condition	O
.	O
	O
Both	O
proteins	O
had	O
identical	O
lengths	O
(	O
a	O
.	O
a	O
.	O
44	O
–	O
195	O
).	O
	O
SV	O
-	O
AUC	O
experiments	O
were	O
performed	O
with	O
1	O
.	O
5	O
mg	O
ml	O
−	O
1	O
of	O
PGRMC1	O
proteins	O
.	O
	O
The	O
major	O
peak	O
with	O
sedimentation	O
coefficient	O
S20	O
,	O
w	O
of	O
1	O
.	O
9	O
∼	O
2	O
.	O
0	O
S	O
(	O
monomer	O
)	O
or	O
3	O
.	O
1	O
S	O
(	O
dimer	O
)	O
was	O
detected	O
.	O
	O
(	O
c	O
)	O
Difference	O
absorption	O
spectra	O
of	O
PGRMC1	O
(	O
a	O
.	O
a	O
.	O
44	O
–	O
195	O
)	O
titrated	O
with	O
haem	O
(	O
left	O
panel	O
).	O
	O
The	O
titration	O
curve	O
of	O
haem	O
to	O
PGRMC1	O
(	O
right	O
panel	O
).	O
	O
The	O
absorbance	O
difference	O
at	O
400	O
nm	O
is	O
plotted	O
against	O
the	O
haem	O
concentration	O
.	O
	O
Carbon	O
monoxide	O
inhibits	O
haem	O
-	O
dependent	O
PGRMC1	O
dimerization	O
.	O
	O
(	O
a	O
)	O
UV	O
-	O
visible	O
absorption	O
spectra	O
of	O
PGRMC1	O
(	O
a	O
.	O
a	O
.	O
44	O
–	O
195	O
).	O
	O
Measurements	O
were	O
performed	O
in	O
the	O
presence	O
of	O
the	O
oxidized	O
form	O
of	O
haem	O
(	O
ferric	O
),	O
the	O
reduced	O
form	O
of	O
haem	O
(	O
ferrous	O
)	O
and	O
the	O
reduced	O
form	O
of	O
haem	O
plus	O
CO	O
gas	O
(	O
ferrous	O
+	O
CO	O
).	O
	O
(	O
b	O
)	O
Close	O
-	O
up	O
view	O
of	O
haem	O
stacking	O
.	O
	O
(	O
c	O
)	O
Gel	O
-	O
filtration	O
chromatography	O
analyses	O
of	O
PGRMC1	O
(	O
a	O
.	O
a	O
.	O
44	O
–	O
195	O
)	O
wild	O
-	O
type	O
(	O
wt	O
)	O
and	O
the	O
Y113F	B-mutant
or	O
C129S	B-mutant
mutant	O
in	O
the	O
presence	O
or	O
absence	O
of	O
haem	O
,	O
dithionite	O
and	O
/	O
or	O
CO	O
.	O
(	O
d	O
)	O
Transition	O
model	O
for	O
structural	O
regulation	O
of	O
PGRMC1	O
in	O
response	O
to	O
haem	O
and	O
CO	O
.	O
	O
Haem	O
-	O
dependent	O
dimerization	O
of	O
PGRMC1	O
is	O
necessary	O
for	O
tumour	O
proliferation	O
mediated	O
by	O
EGFR	O
signalling	O
.	O
	O
Input	O
and	O
bound	O
proteins	O
were	O
detected	O
by	O
Western	O
blotting	O
.	O
	O
(	O
b	O
)	O
In	O
vitro	O
binding	O
assay	O
was	O
performed	O
as	O
in	O
(	O
a	O
)	O
using	O
haem	O
-	O
bound	O
FLAG	O
-	O
PGRMC1	O
wt	O
(	O
a	O
.	O
a	O
.	O
44	O
–	O
195	O
)	O
and	O
purified	O
EGFR	O
with	O
or	O
without	O
treatment	O
of	O
RuCl3	O
and	O
CORM3	O
.	O
	O
(	O
c	O
)	O
FLAG	O
-	O
PGRMC1	O
wt	O
or	O
Y113F	B-mutant
(	O
full	O
length	O
)	O
was	O
over	O
-	O
expressed	O
in	O
HCT116	O
cells	O
and	O
immunoprecipitated	O
with	O
anti	O
-	O
FLAG	O
antibody	O
-	O
conjugated	O
beads	O
.	O
	O
Co	O
-	O
immunoprecipitated	O
proteins	O
(	O
FLAG	O
-	O
PGRMC1	O
,	O
endogenous	O
PGRMC1	O
and	O
EGFR	O
)	O
were	O
detected	O
with	O
Western	O
blotting	O
by	O
using	O
anti	O
-	O
PGRMC1	O
or	O
anti	O
-	O
EGFR	O
antibody	O
.	O
	O
(	O
d	O
)	O
HCT116	O
cells	O
were	O
treated	O
with	O
or	O
without	O
250	O
μmol	O
l	O
−	O
1	O
of	O
succinylacetone	O
(	O
SA	O
)	O
for	O
48	O
h	O
.	O
The	O
intracellular	O
haem	O
was	O
extracted	O
and	O
quantified	O
by	O
reverse	O
-	O
phase	O
HPLC	O
.	O
	O
of	O
four	O
separate	O
experiments	O
.	O
**	O
P	O
<	O
0	O
.	O
01	O
using	O
unpaired	O
Student	O
'	O
s	O
t	O
-	O
test	O
.	O
(	O
e	O
)	O
Co	O
-	O
immunoprecipitation	O
assay	O
was	O
performed	O
as	O
in	O
(	O
c	O
)	O
with	O
or	O
without	O
SA	O
treatment	O
in	O
HCT116	O
cells	O
.	O
	O
Haem	O
-	O
dependent	O
dimerization	O
of	O
PGRMC1	O
accelerates	O
tumour	O
growth	O
through	O
the	O
EGFR	O
signaling	O
pathway	O
.	O
	O
(	O
a	O
)	O
Nucleotide	O
sequences	O
of	O
PGRMC1	O
targeted	O
by	O
shRNA	O
and	O
of	O
the	O
shRNA	O
-	O
resistant	O
full	O
length	O
PGRMC1	O
expression	O
vector	O
.	O
	O
of	O
four	O
separate	O
experiments	O
.	O
*	O
P	O
<	O
0	O
.	O
01	O
using	O
ANOVA	O
with	O
Fischer	O
'	O
s	O
LSD	O
test	O
.	O
	O
(	O
c	O
)	O
Spheroid	O
formation	O
in	O
control	O
and	O
PGRMC1	B-mutant
-	I-mutant
KD	I-mutant
HCT116	O
cells	O
.	O
	O
Scale	O
bar	O
:	O
0	O
.	O
1	O
mm	O
.	O
(	O
d	O
)	O
Tumour	O
-	O
bearing	O
livers	O
of	O
NOG	O
mice	O
at	O
10	O
days	O
after	O
intrasplenic	O
injection	O
of	O
HCT116	O
(	O
control	O
)	O
or	O
PGRMC1	B-mutant
-	I-mutant
KD	I-mutant
cells	O
.	O
	O
of	O
10	O
separate	O
experiments	O
.	O
*	O
P	O
<	O
0	O
.	O
05	O
using	O
unpaired	O
Student	O
'	O
s	O
t	O
-	O
test	O
.	O
	O
(	O
d	O
)	O
Schematic	O
illustration	O
of	O
doxorubicin	O
metabolism	O
is	O
shown	O
on	O
the	O
left	O
.	O
	O
Doxorubicin	O
was	O
incubated	O
with	O
HCT116	O
cells	O
expressing	O
control	O
shRNA	O
or	O
shPGRMC1	O
(	O
PGRMC1	B-mutant
-	I-mutant
KD	I-mutant
),	O
and	O
the	O
doxorubicinol	O
/	O
doxorubicin	O
ratios	O
in	O
cell	O
pellets	O
were	O
determined	O
using	O
LC	O
-	O
MS	O
.	O
	O
of	O
four	O
separate	O
experiments	O
.	O
**	O
P	O
<	O
0	O
.	O
01	O
versus	O
control	O
using	O
unpaired	O
Student	O
'	O
s	O
t	O
-	O
test	O
.	O
(	O
e	O
)	O
Indicated	O
amounts	O
of	O
doxorubicin	O
were	O
added	O
to	O
HCT116	O
(	O
control	O
)	O
cells	O
,	O
PGRMC1	B-mutant
-	I-mutant
KD	I-mutant
cells	O
,	O
or	O
PGRMC1	B-mutant
-	I-mutant
KD	I-mutant
cells	O
expressing	O
shRNA	O
-	O
resistant	O
full	O
-	O
length	O
PGRMC1	O
wt	O
or	O
Y113F	B-mutant
,	O
and	O
cell	O
viability	O
was	O
examined	O
by	O
MTT	O
assay	O
.	O
	O
Apo	O
-	O
PGRMC1	O
exists	O
as	O
an	O
inactive	O
monomer	O
.	O
	O
On	O
binding	O
to	O
haem	O
,	O
PGRMC1	O
forms	O
a	O
dimer	O
through	O
stacking	O
interactions	O
between	O
the	O
haem	O
moieties	O
,	O
which	O
enables	O
PGRMC1	O
to	O
interact	O
with	O
EGFR	O
and	O
cytochromes	O
P450	O
,	O
leading	O
to	O
an	O
enhanced	O
proliferation	O
and	O
chemoresistance	O
of	O
cancer	O
cells	O
.	O
	O
CO	O
interferes	O
with	O
the	O
stacking	O
interactions	O
of	O
the	O
haems	O
and	O
thereby	O
inhibits	O
PGRMC1	O
functions	O
.	O
	O
PGRMC1	O
proteins	O
exhibit	O
haem	O
-	O
dependent	O
dimerization	O
in	O
solution	O
.	O
	O
Apo	O
form	O
Haem	O
-	O
bound	O
form	O
Mass	O
(	O
Da	O
)	O
Mass	O
(	O
Da	O
)	O
aPGRMC1	O
wt	O
(	O
a	O
.	O
a	O
.	O
44	O
–	O
195	O
)	O
ESI	O
-	O
MS	O
—	O
17	O
,	O
844	O
.	O
14	O
—	O
36	O
,	O
920	O
.	O
19	O
Theoretical	O
17	O
,	O
843	O
.	O
65	O
36	O
,	O
918	O
.	O
06	O
Hydrodynamic	O
radius	O
10	O
−	O
9	O
(	O
m	O
)	O
MW	O
(	O
kDa	O
)	O
Hydrodynamic	O
radius	O
10	O
−	O
9	O
(	O
m	O
)	O
MW	O
(	O
kDa	O
)	O
DOSY	O
2	O
.	O
04	O
–	O
2	O
.	O
15	O
20	O
2	O
.	O
94	O
–	O
3	O
.	O
02	O
42	O
S20	O
,	O
w	O
(	O
S	O
)	O
MW	O
(	O
kDa	O
)	O
S20	O
,	O
w	O
(	O
S	O
)	O
MW	O
(	O
kDa	O
)	O
SV	O
-	O
AUC	O
1	O
.	O
9	O
17	O
.	O
6	O
3	O
.	O
1	O
35	O
.	O
5	O
bPGRMC1	O
C129S	B-mutant
(	O
a	O
.	O
a	O
.	O
44	O
–	O
195	O
)	O
ESI	O
-	O
MS	O
—	O
17	O
,	O
827	O
.	O
91	O
—	O
36	O
,	O
887	O
.	O
07	O
Theoretical	O
17	O
,	O
827	O
.	O
59	O
36	O
,	O
885	O
.	O
6	O
S20	O
,	O
w	O
(	O
S	O
)	O
MW	O
(	O
kDa	O
)	O
S20	O
,	O
w	O
(	O
S	O
)	O
MW	O
(	O
kDa	O
)	O
SV	O
-	O
AUC	O
2	O
.	O
0	O
18	O
.	O
1	O
3	O
.	O
1	O
35	O
.	O
8	O
	O
The	O
protein	O
sizes	O
of	O
the	O
wt	O
and	O
C129S	B-mutant
PGRMC1	O
cytosolic	O
domains	O
(	O
a	O
.	O
a	O
.	O
44	O
–	O
195	O
)	O
in	O
the	O
presence	O
or	O
absence	O
of	O
haem	O
were	O
estimated	O
by	O
ESI	O
-	O
MS	O
,	O
DOSY	O
and	O
SV	O
-	O
AUC	O
.	O
	O
However	O
,	O
the	O
mechanisms	O
that	O
allow	O
the	O
clonal	O
T	O
cell	O
antigen	O
receptor	O
(	O
TCR	O
)	O
to	O
functionally	O
engage	O
multiple	O
peptide	O
–	O
major	O
histocompatibility	O
complexes	O
(	O
pMHC	O
)	O
are	O
unclear	O
.	O
	O
Here	O
,	O
we	O
studied	O
multiligand	O
discrimination	O
by	O
a	O
human	O
,	O
preproinsulin	O
reactive	O
,	O
MHC	O
class	O
-	O
I	O
–	O
restricted	O
CD8	O
+	O
T	O
cell	O
clone	O
(	O
1E6	O
)	O
that	O
can	O
recognize	O
over	O
1	O
million	O
different	O
peptides	O
.	O
	O
Evaluation	O
of	O
these	O
structures	O
demonstrated	O
that	O
binding	O
was	O
stabilized	O
through	O
a	O
conserved	O
lock	O
-	O
and	O
-	O
key	O
–	O
like	O
minimal	O
binding	O
footprint	O
that	O
enables	O
1E6	O
TCR	O
to	O
tolerate	O
vast	O
numbers	O
of	O
substitutions	O
outside	O
of	O
this	O
so	O
-	O
called	O
hotspot	O
.	O
	O
T	O
cells	O
perform	O
an	O
essential	O
role	O
in	O
adaptive	O
immunity	O
by	O
interrogating	O
the	O
host	O
proteome	O
for	O
anomalies	O
,	O
classically	O
by	O
recognizing	O
peptides	O
bound	O
in	O
major	O
histocompatibility	O
(	O
MHC	O
)	O
molecules	O
at	O
the	O
cell	O
surface	O
.	O
	O
Recent	O
data	O
supports	O
the	O
notion	O
that	O
,	O
to	O
perform	O
this	O
role	O
,	O
the	O
highly	O
variable	O
αβ	O
T	O
cell	O
antigen	O
receptor	O
(	O
TCR	O
)	O
must	O
be	O
able	O
to	O
recognize	O
thousands	O
,	O
if	O
not	O
millions	O
,	O
of	O
different	O
peptide	O
ligands	O
.	O
	O
Several	O
mechanisms	O
,	O
by	O
which	O
TCRs	O
could	O
bind	O
to	O
a	O
large	O
number	O
of	O
different	O
peptide	O
-	O
MHC	O
(	O
pMHC	O
),	O
have	O
been	O
proposed	O
.	O
	O
Structures	O
of	O
unligated	O
and	O
ligated	O
TCRs	O
have	O
shown	O
that	O
the	O
TCR	O
complementarity	O
determining	O
region	O
(	O
CDR	O
)	O
loops	O
can	O
be	O
flexible	O
,	O
perhaps	O
enabling	O
peptide	O
binding	O
using	O
different	O
loop	O
conformations	O
.	O
	O
Other	O
studies	O
,	O
mainly	O
in	O
the	O
murine	O
system	O
,	O
have	O
demonstrated	O
that	O
the	O
same	O
TCR	O
can	O
interact	O
with	O
different	O
pMHCs	O
using	O
a	O
common	O
or	O
divergent	O
modality	O
.	O
	O
Recent	O
studies	O
in	O
model	O
murine	O
systems	O
demonstrate	O
that	O
TCR	O
cross	O
-	O
reactivity	O
can	O
be	O
governed	O
by	O
recognition	O
of	O
a	O
conserved	O
region	O
in	O
the	O
peptide	O
that	O
allows	O
tolerance	O
of	O
peptide	O
sequence	O
variation	O
outside	O
of	O
this	O
hotspot	O
.	O
	O
CD8	O
+	O
T	O
cells	O
that	O
recognize	O
HLA	O
-	O
A	O
*	O
0201	O
–	O
ALWGPDPAAA	O
have	O
been	O
shown	O
to	O
populate	O
insulitic	O
lesions	O
in	O
patients	O
with	O
type	O
1	O
diabetes	O
(	O
T1D	O
).	O
	O
We	O
demonstrated	O
that	O
the	O
TCR	O
from	O
the	O
1E6	O
T	O
cell	O
clone	O
bound	O
to	O
HLA	O
-	O
A	O
*	O
0201	O
–	O
ALWGPDPAAA	O
using	O
a	O
limited	O
footprint	O
and	O
very	O
weak	O
binding	O
affinity	O
.	O
	O
Here	O
,	O
we	O
solved	O
the	O
structure	O
of	O
the	O
1E6	O
TCR	O
with	O
7	O
altered	O
peptide	O
ligands	O
(	O
APLs	O
)	O
determined	O
by	O
our	O
previously	O
published	O
combinatorial	O
peptide	O
library	O
(	O
CPL	O
)	O
screening	O
,	O
2	O
of	O
which	O
mapped	O
within	O
human	O
pathogens	O
.	O
	O
We	O
also	O
solved	O
the	O
structure	O
of	O
each	O
unligated	O
APL	O
to	O
investigate	O
whether	O
structural	O
changes	O
occurred	O
before	O
or	O
after	O
binding	O
—	O
which	O
,	O
combined	O
with	O
an	O
in	O
-	O
depth	O
cellular	O
and	O
biophysical	O
analysis	O
of	O
the	O
1E6	O
interaction	O
with	O
each	O
APL	O
,	O
demonstrated	O
the	O
molecular	O
mechanism	O
mediating	O
the	O
high	O
level	O
of	O
cross	O
-	O
reactivity	O
exhibited	O
by	O
this	O
preproinsulin	O
-	O
reactive	O
human	O
CD8	O
+	O
T	O
cell	O
clone	O
.	O
	O
The	O
1E6	O
T	O
cell	O
clone	O
recognizes	O
APLs	O
across	O
a	O
large	O
dynamic	O
range	O
.	O
	O
We	O
have	O
previously	O
demonstrated	O
that	O
the	O
1E6	O
T	O
cell	O
clone	O
can	O
recognize	O
over	O
1	O
million	O
different	O
peptides	O
with	O
a	O
potency	O
comparable	O
with	O
,	O
or	O
better	O
than	O
,	O
the	O
cognate	O
preproinsulin	O
peptide	O
ALWGPDPAAA	O
.	O
	O
From	O
this	O
large	O
functional	O
scan	O
,	O
we	O
selected	O
7	O
different	O
APLs	O
that	O
activated	O
the	O
1E6	O
T	O
cell	O
clone	O
across	O
a	O
wide	O
(	O
4	O
-	O
log	O
)	O
functional	O
range	O
(	O
Table	O
1	O
).	O
	O
Competitive	O
functional	O
testing	O
revealed	O
that	O
the	O
preproinsulin	O
-	O
derived	O
sequence	O
ALWGPDPAAA	O
was	O
one	O
of	O
the	O
least	O
potent	O
targets	O
for	O
1E6	O
,	O
with	O
only	O
the	O
MVWGPDPLYV	O
and	O
YLGGPDFPTI	O
demonstrating	O
a	O
similar	O
low	O
-	O
activity	O
profile	O
in	O
MIP	O
-	O
1β	O
secretion	O
and	O
target	O
killing	O
assays	O
(	O
Figure	O
1	O
,	O
A	O
and	O
B	O
).	O
	O
At	O
the	O
other	O
end	O
of	O
the	O
spectrum	O
,	O
the	O
RQFGPDFPTI	O
peptide	O
stimulated	O
MIP	O
-	O
1β	O
release	O
and	O
killing	O
by	O
1E6	O
at	O
an	O
exogenous	O
peptide	O
concentration	O
2	O
–	O
3	O
logs	O
lower	O
compared	O
with	O
ALWGPDPAAA	O
.	O
	O
The	O
pattern	O
of	O
peptide	O
potency	O
was	O
closely	O
mirrored	O
by	O
pMHC	O
tetramer	O
staining	O
experiments	O
(	O
Figure	O
1C	O
and	O
plots	O
shown	O
in	O
Supplemental	O
Figure	O
1	O
;	O
supplemental	O
material	O
available	O
online	O
with	O
this	O
article	O
;	O
doi	O
:	O
10	O
.	O
1172	O
/	O
JCI85679DS1	O
).	O
	O
Here	O
,	O
the	O
A2	O
-	O
RQFGPDFPTI	O
tetramer	O
stained	O
1E6	O
with	O
the	O
greatest	O
MFI	O
,	O
gradually	O
decreasing	O
to	O
the	O
weakest	O
tetramers	O
:	O
A2	O
-	O
MVWGPDPLYV	O
and	O
-	O
YLGGPDFPTI	O
.	O
	O
To	O
parallel	O
the	O
functional	O
analysis	O
,	O
we	O
also	O
performed	O
thermal	O
melt	O
(	O
Tm	O
)	O
experiments	O
using	O
synchrotron	O
radiation	O
circular	O
dichroism	O
(	O
SRCD	O
)	O
to	O
investigate	O
the	O
stability	O
of	O
each	O
APL	O
(	O
Figure	O
1D	O
).	O
	O
This	O
pattern	O
of	O
stability	O
did	O
not	O
correlate	O
with	O
the	O
T	O
cell	O
activation	O
or	O
tetramer	O
staining	O
experiments	O
,	O
indicating	O
that	O
peptide	O
binding	O
to	O
the	O
MHC	O
do	O
not	O
explain	O
ligand	O
potency	O
.	O
	O
The	O
1E6	O
TCR	O
can	O
bind	O
peptides	O
with	O
strong	O
antipathogen	O
-	O
like	O
affinities	O
.	O
	O
We	O
,	O
and	O
others	O
,	O
have	O
previously	O
demonstrated	O
that	O
antipathogenic	O
TCRs	O
tend	O
to	O
bind	O
with	O
stronger	O
affinity	O
compared	O
with	O
self	O
-	O
reactive	O
TCRs	O
,	O
likely	O
a	O
consequence	O
of	O
the	O
deletion	O
of	O
T	O
cells	O
with	O
high	O
-	O
affinity	O
self	O
-	O
reactive	O
TCR	O
during	O
thymic	O
selection	O
.	O
	O
In	O
accordance	O
with	O
this	O
trend	O
,	O
the	O
1E6	O
TCR	O
bound	O
the	O
natural	O
preproinsulin	O
peptide	O
,	O
ALWGPDPAAA	O
,	O
with	O
the	O
weakest	O
affinity	O
currently	O
published	O
for	O
a	O
human	O
CD8	O
+	O
T	O
cell	O
–	O
derived	O
TCR	O
with	O
a	O
biologically	O
relevant	O
ligand	O
(	O
KD	O
>	O
200	O
μM	O
;	O
KD	O
,	O
equilibrium	O
binding	O
constant	O
).	O
	O
Surface	O
plasmon	O
resonance	O
(	O
SPR	O
)	O
analysis	O
of	O
the	O
1E6	O
TCR	O
–	O
pMHC	O
interaction	O
for	O
all	O
7	O
APLs	O
(	O
Figure	O
2	O
,	O
A	O
–	O
H	O
)	O
demonstrated	O
that	O
stronger	O
binding	O
affinity	O
(	O
represented	O
as	O
ΔG	O
°,	O
kcal	O
/	O
mol	O
)	O
correlated	O
well	O
with	O
the	O
EC50	O
values	O
(	O
peptide	O
concentration	O
required	O
to	O
reach	O
half	O
-	O
maximal	O
1E6	O
T	O
cell	O
killing	O
)	O
for	O
each	O
ligand	O
,	O
demonstrated	O
by	O
a	O
Pearson	O
’	O
s	O
correlation	O
analysis	O
value	O
of	O
0	O
.	O
8	O
(	O
P	O
=	O
0	O
.	O
01	O
)	O
(	O
Figure	O
2I	O
).	O
	O
It	O
should	O
be	O
noted	O
that	O
this	O
correlation	O
,	O
although	O
consistent	O
with	O
the	O
T	O
cell	O
killing	O
experiments	O
,	O
uses	O
only	O
approximate	O
affinities	O
calculated	O
for	O
the	O
2	O
weakest	O
ligands	O
.	O
	O
Third	O
,	O
the	O
1E6	O
TCR	O
bound	O
to	O
A2	O
-	O
RQFGPDWIVA	O
peptide	O
,	O
within	O
the	O
C	O
.	O
asparagiforme	O
proteome	O
,	O
with	O
approximately	O
4	O
-	O
fold	O
stronger	O
affinity	O
than	O
A2	O
-	O
ALWGPDPAAA	O
,	O
demonstrating	O
the	O
potential	O
for	O
a	O
pathogen	O
-	O
derived	O
antigen	O
to	O
initiate	O
a	O
response	O
to	O
the	O
self	O
-	O
derived	O
sequence	O
.	O
	O
Finally	O
,	O
these	O
data	O
demonstrate	O
the	O
largest	O
range	O
of	O
binding	O
affinities	O
reported	O
for	O
a	O
natural	O
,	O
endogenous	O
human	O
TCR	O
of	O
more	O
than	O
3	O
logs	O
of	O
magnitude	O
(	O
A2	O
-	O
MVWGPDPLYV	O
vs	O
.	O
A2	O
-	O
RQFGPDFPTI	O
).	O
	O
In	O
agreement	O
with	O
SPR	O
experiments	O
,	O
the	O
range	O
of	O
2D	O
affinities	O
we	O
detected	O
differed	O
by	O
around	O
3	O
logs	O
,	O
with	O
the	O
A2	O
-	O
MVWGPDPLYV	O
generating	O
the	O
weakest	O
2D	O
affinity	O
(	O
2	O
.	O
6	O
×	O
10	O
–	O
5	O
AcKa	O
μm4	O
)	O
and	O
A2	O
-	O
RQFGPDFPTI	O
the	O
strongest	O
(	O
4	O
.	O
5	O
×	O
10	O
–	O
2	O
AcKa	O
μm4	O
)	O
(	O
Figure	O
2J	O
).	O
	O
Of	O
note	O
,	O
these	O
data	O
demonstrate	O
a	O
close	O
agreement	O
between	O
the	O
3D	O
affinity	O
values	O
generated	O
using	O
SPR	O
and	O
2D	O
affinity	O
values	O
generated	O
using	O
adhesion	O
frequency	O
assays	O
.	O
	O
The	O
1E6	O
TCR	O
uses	O
a	O
consensus	O
binding	O
mode	O
to	O
engage	O
multiple	O
APLs	O
.	O
	O
Our	O
previous	O
structure	O
of	O
the	O
1E6	O
-	O
A2	O
-	O
ALWGPDPAAA	O
complex	O
demonstrated	O
a	O
limited	O
binding	O
footprint	O
between	O
the	O
TCR	O
and	O
pMHC	O
.	O
	O
In	O
order	O
to	O
examine	O
the	O
mechanism	O
by	O
which	O
the	O
1E6	O
TCR	O
engaged	O
a	O
wide	O
range	O
of	O
peptides	O
with	O
divergent	O
binding	O
affinities	O
,	O
we	O
solved	O
the	O
structure	O
of	O
the	O
1E6	O
TCR	O
in	O
complex	O
with	O
all	O
7	O
APLs	O
used	O
in	O
Figure	O
2	O
.	O
	O
All	O
structures	O
were	O
solved	O
in	O
space	O
group	O
P1	O
to	O
2	O
–	O
3	O
Å	O
resolution	O
with	O
crystallographic	O
Rwork	O
/	O
Rfree	O
ratios	O
within	O
accepted	O
limits	O
as	O
shown	O
in	O
the	O
theoretically	O
expected	O
distribution	O
(	O
ref	O
.	O
and	O
Supplemental	O
Table	O
1	O
).	O
	O
The	O
1E6	O
TCR	O
used	O
a	O
very	O
similar	O
overall	O
binding	O
modality	O
to	O
engage	O
all	O
of	O
the	O
APLs	O
,	O
with	O
root	O
mean	O
square	O
deviation	O
ranging	O
between	O
0	O
.	O
81	O
and	O
1	O
.	O
12	O
Å2	O
(	O
compared	O
with	O
1E6	O
-	O
A2	O
-	O
ALWGPDPAAA	O
).	O
	O
The	O
relatively	O
broad	O
range	O
of	O
buried	O
surface	O
areas	O
(	O
1	O
,	O
670	O
–	O
1	O
,	O
920	O
Å2	O
)	O
did	O
not	O
correlate	O
well	O
with	O
TCR	O
binding	O
affinity	O
(	O
Pearson	O
’	O
s	O
correlation	O
=	O
0	O
.	O
45	O
,	O
P	O
=	O
0	O
.	O
2	O
).	O
	O
The	O
surface	O
complementarity	O
values	O
(	O
0	O
.	O
52	O
–	O
0	O
.	O
7	O
)	O
correlated	O
slightly	O
with	O
affinity	O
(	O
Pearson	O
’	O
s	O
correlation	O
=	O
0	O
.	O
7	O
,	O
P	O
=	O
0	O
.	O
05	O
)	O
but	O
could	O
not	O
explain	O
all	O
differences	O
in	O
binding	O
(	O
Figure	O
3A	O
and	O
Table	O
2	O
).	O
	O
The	O
TCR	O
CDR	O
loops	O
were	O
in	O
a	O
very	O
similar	O
position	O
in	O
all	O
complexes	O
,	O
apart	O
from	O
some	O
slight	O
deviations	O
in	O
the	O
TCR	O
β	O
-	O
chain	O
(	O
Figure	O
3B	O
);	O
the	O
peptides	O
were	O
all	O
presented	O
in	O
a	O
similar	O
conformation	O
(	O
Figure	O
3C	O
);	O
and	O
there	O
was	O
minimal	O
variation	O
in	O
crossing	O
angles	O
of	O
the	O
TCR	O
(	O
42	O
.	O
3	O
°–	O
45	O
.	O
6	O
°)	O
(	O
Figure	O
3D	O
).	O
	O
However	O
,	O
subtle	O
differences	O
in	O
the	O
respective	O
interfaces	O
were	O
apparent	O
(	O
discussed	O
below	O
)	O
and	O
resulted	O
in	O
altered	O
binding	O
affinities	O
of	O
the	O
respective	O
complexes	O
.	O
	O
Interactions	O
between	O
the	O
1E6	O
TCR	O
and	O
different	O
APLs	O
are	O
focused	O
around	O
a	O
conserved	O
GPD	O
peptide	O
motif	O
.	O
	O
We	O
next	O
performed	O
an	O
in	O
-	O
depth	O
atomic	O
analysis	O
of	O
the	O
contacts	O
between	O
the	O
1E6	O
TCR	O
and	O
each	O
APL	O
to	O
determine	O
the	O
structural	O
basis	O
for	O
the	O
altered	O
T	O
cell	O
peptide	O
sensitivities	O
and	O
TCR	O
binding	O
affinities	O
(	O
Table	O
2	O
).	O
	O
Concomitant	O
with	O
our	O
global	O
analysis	O
of	O
1E6	O
TCR	O
binding	O
to	O
the	O
APLs	O
,	O
we	O
observed	O
a	O
common	O
interaction	O
element	O
,	O
consistent	O
with	O
our	O
previous	O
findings	O
,	O
that	O
utilized	O
TCR	O
residues	O
Tyr97α	O
and	O
Trp97β	O
,	O
forming	O
an	O
aromatic	O
cap	O
over	O
a	O
central	O
GPD	O
motif	O
that	O
was	O
present	O
in	O
all	O
of	O
the	O
APLs	O
(	O
Figure	O
4	O
).	O
	O
Interactions	O
between	O
these	O
2	O
TCR	O
and	O
3	O
peptide	O
residues	O
accounted	O
for	O
41	O
%–	O
50	O
%	O
of	O
the	O
total	O
contacts	O
across	O
all	O
complexes	O
(	O
Table	O
2	O
),	O
demonstrating	O
the	O
conserved	O
peptide	O
centric	O
binding	O
mode	O
utilized	O
by	O
the	O
1E6	O
TCR	O
.	O
	O
This	O
fixed	O
anchoring	O
between	O
the	O
2	O
molecules	O
was	O
important	O
for	O
stabilization	O
of	O
the	O
TCR	O
-	O
pMHC	O
complex	O
,	O
as	O
—	O
although	O
other	O
peptides	O
without	O
the	O
‘	O
GDP	O
’	O
motif	O
were	O
tested	O
and	O
shown	O
to	O
activate	O
the	O
1E6	O
T	O
cell	O
clone	O
—	O
we	O
were	O
unable	O
to	O
measure	O
robust	O
affinities	O
using	O
SPR	O
(	O
data	O
not	O
shown	O
).	O
	O
These	O
data	O
support	O
the	O
requirement	O
for	O
a	O
conserved	O
interaction	O
between	O
the	O
1E6	O
TCR	O
and	O
the	O
GPD	O
motif	O
,	O
as	O
we	O
observed	O
in	O
our	O
previously	O
published	O
1E6	O
-	O
A2	O
-	O
ALWGPDPAAA	O
structure	O
.	O
	O
For	O
example	O
,	O
the	O
1E6	O
TCR	O
made	O
only	O
47	O
peptide	O
contacts	O
with	O
A2	O
-	O
MVWGPDPLYV	O
(	O
KD	O
=	O
~	O
600	O
μM	O
)	O
compared	O
with	O
63	O
and	O
57	O
contacts	O
with	O
A2	O
-	O
YQFGPDFPIA	O
(	O
KD	O
=	O
7	O
.	O
4	O
μM	O
)	O
and	O
A2	O
-	O
RQFGPDFPTI	O
(	O
KD	O
=	O
0	O
.	O
5	O
μM	O
),	O
respectively	O
.	O
	O
For	O
example	O
,	O
the	O
1E6	O
TCR	O
made	O
64	O
peptide	O
contacts	O
with	O
A2	O
-	O
YLGGPDFPTI	O
(	O
KD	O
=	O
~	O
400	O
μM	O
)	O
compared	O
with	O
43	O
contacts	O
with	O
A2	O
-	O
RQWGPDPAAV	O
(	O
KD	O
=	O
7	O
.	O
8	O
μM	O
).	O
	O
The	O
most	O
important	O
peptide	O
modification	O
in	O
terms	O
of	O
generating	O
new	O
contacts	O
was	O
peptide	O
position	O
1	O
.	O
	O
The	O
stronger	O
ligands	O
all	O
encoded	O
larger	O
side	O
chains	O
(	O
Arg	O
or	O
Tyr	O
)	O
at	O
peptide	O
position	O
1	O
(	O
Figure	O
5	O
,	O
E	O
–	O
H	O
),	O
enabling	O
interactions	O
with	O
1E6	O
that	O
were	O
not	O
present	O
in	O
the	O
weaker	O
APLs	O
that	O
lacked	O
large	O
side	O
chains	O
in	O
this	O
position	O
(	O
Figure	O
5	O
,	O
A	O
,	O
C	O
,	O
and	O
D	O
).	O
	O
We	O
have	O
previously	O
shown	O
that	O
the	O
1E6	O
TCR	O
uses	O
a	O
rigid	O
lock	O
-	O
and	O
-	O
key	O
mechanism	O
during	O
binding	O
to	O
A2	O
-	O
ALWGPDPAAA	O
.	O
	O
In	O
order	O
to	O
determine	O
whether	O
any	O
of	O
the	O
APLs	O
required	O
an	O
induced	O
fit	O
mechanism	O
during	O
binding	O
that	O
could	O
explain	O
the	O
difference	O
in	O
free	O
binding	O
energy	O
(	O
ΔG	O
)	O
between	O
each	O
complex	O
(	O
Table	O
2	O
),	O
we	O
solved	O
the	O
unligated	O
structures	O
of	O
all	O
7	O
APLs	O
(	O
the	O
A2	O
-	O
ALWGPDPAAA	O
structure	O
has	O
been	O
previously	O
published	O
and	O
was	O
used	O
in	O
this	O
comparison	O
,	O
ref	O
.)	O
(	O
Figure	O
6	O
and	O
Supplemental	O
Table	O
2	O
).	O
	O
This	O
movement	O
could	O
result	O
in	O
an	O
entropic	O
penalty	O
contributing	O
to	O
the	O
weak	O
TCR	O
binding	O
affinity	O
we	O
observed	O
for	O
this	O
ligand	O
.	O
	O
Additional	O
small	O
movements	O
in	O
the	O
Cα	O
backbone	O
of	O
the	O
peptide	O
around	O
peptide	O
residue	O
Asp6	O
were	O
apparent	O
in	O
the	O
A2	O
-	O
YLGGPDFPTI	O
(	O
KD	O
=	O
~	O
400	O
μM	O
),	O
A2	O
-	O
ALWGPDPAAA	O
(	O
KD	O
=	O
~	O
208	O
μM	O
),	O
and	O
A2	O
-	O
RQFGPDWIVA	O
(	O
KD	O
=	O
44	O
.	O
4	O
μM	O
)	O
structures	O
(	O
Figure	O
6	O
,	O
B	O
,	O
C	O
,	O
and	O
E	O
).	O
	O
The	O
unligated	O
structures	O
of	O
A2	O
-	O
AQWGPDAAA	O
,	O
A2	O
-	O
RQWGPDPAAV	O
,	O
A2	O
-	O
YQFGPDFPIA	O
,	O
and	O
A2	O
-	O
RQFGPDFPTI	O
were	O
virtually	O
identical	O
when	O
in	O
complex	O
with	O
1E6	O
(	O
Figure	O
6	O
,	O
D	O
and	O
F	O
–	O
H	O
).	O
	O
Apart	O
from	O
the	O
case	O
of	O
A2	O
-	O
AQWGPDAAA	O
(	O
KD	O
=	O
61	O
.	O
9	O
μM	O
),	O
these	O
observations	O
support	O
the	O
conclusion	O
that	O
the	O
higher	O
-	O
affinity	O
ligands	O
required	O
less	O
conformational	O
melding	O
during	O
binding	O
,	O
which	O
could	O
be	O
energetically	O
beneficial	O
(	O
lower	O
entopic	O
cost	O
)	O
during	O
ligation	O
with	O
the	O
1E6	O
TCR	O
.	O
	O
Peptide	O
modifications	O
alter	O
the	O
interaction	O
between	O
the	O
1E6	O
TCR	O
and	O
the	O
MHC	O
surface	O
.	O
	O
In	O
addition	O
to	O
changes	O
between	O
the	O
TCR	O
and	O
peptide	O
component	O
,	O
we	O
also	O
observed	O
that	O
different	O
APLs	O
had	O
different	O
knock	O
-	O
on	O
effects	O
between	O
the	O
TCR	O
and	O
MHC	O
.	O
	O
MHC	O
residue	O
Arg65	O
that	O
forms	O
part	O
of	O
the	O
MHC	O
restriction	O
triad	O
(	O
Arg65	O
,	O
Ala69	O
,	O
and	O
Gln155	O
)	O
played	O
a	O
central	O
role	O
in	O
TCR	O
-	O
MHC	O
contacts	O
,	O
with	O
Gln155	O
playing	O
a	O
less	O
important	O
role	O
and	O
Ala69	O
playing	O
no	O
role	O
in	O
binding	O
at	O
the	O
interface	O
(	O
Figure	O
7	O
).	O
	O
Generally	O
,	O
the	O
weaker	O
-	O
affinity	O
APLs	O
made	O
fewer	O
contacts	O
with	O
the	O
MHC	O
surface	O
(	O
27	O
–	O
29	O
interactions	O
)	O
compared	O
with	O
the	O
stronger	O
-	O
affinity	O
APLs	O
(	O
29	O
–	O
35	O
contacts	O
),	O
consistent	O
with	O
a	O
better	O
Pearson	O
’	O
s	O
correlation	O
value	O
(	O
0	O
.	O
55	O
)	O
compared	O
with	O
TCR	O
-	O
peptide	O
interactions	O
versus	O
affinity	O
(	O
0	O
.	O
045	O
).	O
	O
For	O
instance	O
,	O
contacts	O
were	O
made	O
between	O
TCR	O
residue	O
Val53β	O
and	O
MHC	O
residue	O
Gln72	O
in	O
all	O
APLs	O
except	O
for	O
in	O
the	O
weakest	O
affinity	O
ligand	O
pair	O
,	O
1E6	O
-	O
A2	O
-	O
MVWGPDPLYV	O
,	O
in	O
which	O
a	O
subtle	O
change	O
in	O
TCR	O
conformation	O
—	O
probably	O
mediated	O
by	O
different	O
peptide	O
contacts	O
—	O
abrogated	O
this	O
interaction	O
(	O
Figure	O
7A	O
).	O
	O
An	O
energetic	O
switch	O
from	O
unfavorable	O
to	O
favorable	O
entropy	O
(	O
order	O
-	O
to	O
-	O
disorder	O
)	O
correlates	O
with	O
antigen	O
potency	O
.	O
	O
Our	O
analysis	O
of	O
the	O
contact	O
network	O
provided	O
some	O
clues	O
that	O
could	O
explain	O
the	O
different	O
antigen	O
potencies	O
and	O
binding	O
affinities	O
between	O
the	O
1E6	O
TCR	O
and	O
the	O
different	O
APLs	O
.	O
	O
For	O
example	O
,	O
the	O
1E6	O
TCR	O
bound	O
to	O
A2	O
-	O
RQWGPDPAAV	O
with	O
the	O
third	O
strongest	O
affinity	O
(	O
KD	O
=	O
7	O
.	O
8	O
μM	O
)	O
but	O
made	O
fewer	O
contacts	O
than	O
with	O
A2	O
-	O
ALWGPDPAAA	O
(	O
KD	O
=	O
~	O
208	O
μM	O
)	O
(	O
Table	O
2	O
).	O
	O
The	O
weak	O
binding	O
affinity	O
between	O
1E6	O
and	O
A2	O
-	O
MVWGPDPLYV	O
and	O
A2	O
-	O
YLGGPDFPTI	O
generated	O
thermodynamic	O
data	O
that	O
were	O
not	O
robust	O
enough	O
to	O
gain	O
insight	O
into	O
the	O
enthalpic	O
(	O
ΔH	O
°)	O
and	O
entropic	O
(	O
TΔS	O
°)	O
changes	O
that	O
contributed	O
to	O
the	O
different	O
binding	O
affinities	O
/	O
potencies	O
for	O
each	O
APL	O
.	O
	O
For	O
instance	O
,	O
the	O
A2	O
-	O
ALWGPDPAAA	O
,	O
A2	O
-	O
AQWGPDAAA	O
,	O
and	O
A2	O
-	O
RQFGPDWIVA	O
(	O
KD	O
=	O
~	O
208	O
μM	O
,	O
KD	O
=	O
61	O
.	O
9	O
μM	O
,	O
and	O
KD	O
=	O
44	O
.	O
4	O
μM	O
,	O
respectively	O
)	O
were	O
all	O
entropically	O
unfavorable	O
(	O
TΔS	O
°	O
=	O
–	O
2	O
.	O
9	O
to	O
–	O
5	O
.	O
6	O
kcal	O
/	O
mol	O
),	O
indicating	O
a	O
net	O
change	O
from	O
disorder	O
to	O
order	O
.	O
	O
Conversely	O
,	O
the	O
stronger	O
-	O
affinity	O
ligands	O
A2	O
-	O
RQWGPDPAAV	O
(	O
KD	O
=	O
7	O
.	O
8	O
μM	O
),	O
A2	O
-	O
YQFGPDFPIA	O
(	O
KD	O
=	O
7	O
.	O
4	O
μM	O
),	O
and	O
A2	O
-	O
RQFGPDFPTI	O
(	O
KD	O
=	O
0	O
.	O
5	O
μM	O
)	O
exhibited	O
favorable	O
entropy	O
(	O
TΔS	O
°	O
=	O
2	O
.	O
2	O
to	O
14	O
.	O
9	O
kcal	O
/	O
mol	O
),	O
indicating	O
an	O
order	O
-	O
to	O
-	O
disorder	O
change	O
during	O
binding	O
,	O
possibly	O
through	O
the	O
expulsion	O
of	O
ordered	O
water	O
molecules	O
.	O
	O
The	O
potential	O
requirement	O
for	O
a	O
larger	O
degree	O
of	O
induced	O
fit	O
during	O
binding	O
to	O
these	O
weaker	O
-	O
affinity	O
ligands	O
is	O
consistent	O
with	O
the	O
larger	O
entropic	O
penalties	O
observed	O
for	O
these	O
interactions	O
.	O
	O
Potential	O
epitopes	O
for	O
1E6	O
TCR	O
occur	O
commonly	O
in	O
the	O
viral	O
proteome	O
.	O
	O
Three	O
hundred	O
forty	O
-	O
two	O
of	O
these	O
decamers	O
conformed	O
to	O
the	O
motif	O
xxxGPDxxxx	O
.	O
	O
Of	O
these	O
,	O
53	O
peptides	O
contained	O
the	O
motif	O
xOxGPDxxxO	O
,	O
where	O
O	O
is	O
one	O
of	O
the	O
hydrophobic	O
amino	O
acid	O
residues	O
A	O
,	O
V	O
,	O
I	O
,	O
L	O
,	O
M	O
,	O
Y	O
,	O
F	O
,	O
and	O
W	O
that	O
might	O
allow	O
binding	O
to	O
HLA	O
-	O
A	O
*	O
0201	O
(	O
Supplemental	O
Table	O
4	O
).	O
	O
Thus	O
,	O
there	O
are	O
many	O
pathogen	O
-	O
encoded	O
peptides	O
that	O
could	O
act	O
as	O
agonists	O
for	O
the	O
1E6	O
T	O
cell	O
beyond	O
the	O
MVWGPDPLYV	O
and	O
RQFGPDWIVA	O
sequences	O
studied	O
here	O
.	O
	O
Extension	O
of	O
these	O
analyses	O
to	O
include	O
the	O
larger	O
genomes	O
of	O
bacterial	O
pathogens	O
would	O
be	O
expected	O
to	O
considerably	O
increase	O
these	O
numbers	O
.	O
	O
The	O
binding	O
affinity	O
of	O
the	O
1E6	O
TCR	O
interaction	O
with	O
A2	O
-	O
RQFGPDWIVA	O
is	O
considerably	O
higher	O
than	O
with	O
the	O
disease	O
-	O
implicated	O
A2	O
-	O
ALWGPDPAAA	O
sequence	O
(	O
KD	O
=	O
44	O
.	O
4	O
μM	O
and	O
KD	O
>	O
200	O
μM	O
,	O
respectively	O
),	O
highlighting	O
how	O
a	O
pathogen	O
-	O
derived	O
sequence	O
might	O
be	O
capable	O
of	O
priming	O
a	O
1E6	O
-	O
like	O
T	O
cell	O
.	O
	O
T	O
cell	O
antigen	O
discrimination	O
is	O
governed	O
by	O
an	O
interaction	O
between	O
the	O
clonally	O
expressed	O
TCR	O
and	O
pMHC	O
,	O
mediated	O
by	O
the	O
chemical	O
characteristics	O
of	O
the	O
interacting	O
molecules	O
.	O
	O
It	O
has	O
recently	O
become	O
clear	O
that	O
TCR	O
cross	O
-	O
reactivity	O
with	O
large	O
numbers	O
of	O
different	O
pMHC	O
ligands	O
is	O
essential	O
to	O
plug	O
holes	O
in	O
T	O
cell	O
immune	O
coverage	O
that	O
pathogens	O
could	O
exploit	O
.	O
	O
Flexibility	O
at	O
the	O
interface	O
between	O
the	O
TCR	O
and	O
pMHC	O
,	O
demonstrated	O
in	O
various	O
studies	O
,	O
has	O
been	O
suggested	O
as	O
a	O
mechanism	O
mediating	O
T	O
cell	O
cross	O
-	O
reactivity	O
with	O
multiple	O
distinct	O
epitopes	O
.	O
	O
Focused	O
binding	O
around	O
a	O
minimal	O
peptide	O
motif	O
has	O
also	O
been	O
implicated	O
as	O
an	O
alternative	O
mechanism	O
enabling	O
TCR	O
cross	O
-	O
reactivity	O
.	O
	O
Notably	O
among	O
these	O
studies	O
,	O
Garcia	O
and	O
colleagues	O
recently	O
used	O
the	O
alloreactive	O
murine	O
TCR	O
-	O
MHC	O
pair	O
of	O
the	O
42F3	O
TCR	O
and	O
H2	O
-	O
Ld	O
to	O
demonstrate	O
recognition	O
of	O
a	O
large	O
number	O
of	O
different	O
peptides	O
via	O
conserved	O
hotspot	O
contacts	O
with	O
prominent	O
up	O
-	O
facing	O
peptide	O
residues	O
.	O
	O
First	O
,	O
we	O
currently	O
know	O
nothing	O
about	O
how	O
human	O
MHCI	O
–	O
restricted	O
TCRs	O
mediate	O
cross	O
-	O
reactivity	O
in	O
the	O
context	O
of	O
a	O
clinically	O
relevant	O
model	O
of	O
autoimmunity	O
,	O
thought	O
to	O
be	O
a	O
major	O
pathway	O
of	O
disease	O
initiation	O
in	O
several	O
autoimmune	O
diseases	O
.	O
	O
Second	O
,	O
molecular	O
studies	O
have	O
not	O
yet	O
revealed	O
a	O
broad	O
set	O
of	O
rules	O
that	O
determine	O
TCR	O
cross	O
-	O
reactivity	O
because	O
,	O
with	O
the	O
exception	O
of	O
the	O
allo	O
–	O
TCR	O
-	O
MHC	O
pair	O
of	O
the	O
42F3	O
TCR	O
and	O
H2	O
-	O
Ld	O
that	O
did	O
not	O
encounter	O
each	O
other	O
during	O
T	O
cell	O
development	O
,	O
studies	O
have	O
been	O
limited	O
to	O
structures	O
of	O
a	O
TCR	O
with	O
only	O
2	O
or	O
3	O
different	O
ligands	O
.	O
	O
Here	O
,	O
we	O
investigated	O
a	O
highly	O
cross	O
-	O
reactive	O
MHCI	O
-	O
restricted	O
TCR	O
isolated	O
from	O
a	O
patient	O
with	O
T1D	O
that	O
recognizes	O
an	O
HLA	O
-	O
A	O
*	O
0201	O
–	O
restricted	O
preproinsulin	O
signal	O
peptide	O
(	O
ALWGPDPAAA15	O
–	O
24	O
).	O
	O
Human	O
CD8	O
+	O
T	O
cell	O
clones	O
expressing	O
TCRs	O
with	O
this	O
specificity	O
mediate	O
the	O
destruction	O
of	O
β	O
cells	O
,	O
have	O
been	O
found	O
in	O
islets	O
early	O
in	O
infection	O
,	O
and	O
are	O
proposed	O
to	O
be	O
a	O
major	O
driver	O
of	O
disease	O
.	O
	O
We	O
solved	O
the	O
structure	O
of	O
the	O
1E6	O
TCR	O
with	O
7	O
APLs	O
to	O
enable	O
a	O
comprehensive	O
analysis	O
of	O
the	O
molecular	O
basis	O
of	O
TCR	O
degeneracy	O
.	O
	O
Overall	O
,	O
the	O
difference	O
in	O
antigen	O
potency	O
correlated	O
well	O
with	O
the	O
binding	O
energy	O
(	O
ΔG	O
°	O
kcal	O
/	O
mol	O
)	O
of	O
the	O
1E6	O
TCR	O
for	O
the	O
different	O
epitopes	O
,	O
which	O
ranged	O
from	O
values	O
of	O
ΔG	O
°	O
=	O
~–	O
4	O
.	O
4	O
to	O
–	O
8	O
.	O
6	O
kcal	O
/	O
mol	O
(	O
calculated	O
from	O
3D	O
affinity	O
data	O
)	O
or	O
2D	O
affinity	O
values	O
of	O
AcKa	O
=	O
2	O
.	O
5	O
×	O
10	O
–	O
5	O
to	O
4	O
.	O
4	O
×	O
10	O
–	O
2	O
μm4	O
.	O
	O
The	O
weaker	O
end	O
of	O
this	O
spectrum	O
extends	O
our	O
understanding	O
of	O
the	O
limits	O
in	O
which	O
T	O
cells	O
can	O
functionally	O
operate	O
in	O
terms	O
of	O
TCR	O
3D	O
binding	O
affinity	O
and	O
is	O
in	O
line	O
with	O
the	O
types	O
of	O
very	O
low	O
affinity	O
,	O
yet	O
fully	O
functional	O
self	O
-	O
reactive	O
CD8	O
+	O
T	O
cells	O
we	O
have	O
observed	O
in	O
tumor	O
-	O
infiltrating	O
lymphocytes	O
.	O
	O
Previous	O
studies	O
of	O
autoreactive	O
TCRs	O
have	O
shown	O
that	O
their	O
binding	O
mode	O
is	O
generally	O
atypical	O
,	O
either	O
due	O
to	O
an	O
unusual	O
binding	O
manner	O
,	O
weak	O
TCR	O
binding	O
affinity	O
,	O
an	O
unstable	O
pMHC	O
,	O
or	O
a	O
combination	O
of	O
these	O
factors	O
.	O
	O
Our	O
data	O
demonstrate	O
the	O
potential	O
for	O
an	O
autoreactive	O
TCR	O
to	O
bind	O
with	O
a	O
conventional	O
binding	O
mode	O
to	O
a	O
stable	O
pMHC	O
with	O
antipathogen	O
-	O
like	O
affinity	O
(	O
KD	O
=	O
0	O
.	O
5	O
μM	O
)	O
depending	O
on	O
the	O
peptide	O
sequence	O
.	O
	O
Our	O
structural	O
analysis	O
revealed	O
that	O
the	O
1E6	O
TCR	O
bound	O
with	O
a	O
conserved	O
conformation	O
across	O
all	O
APLs	O
investigated	O
.	O
	O
This	O
binding	O
orientation	O
was	O
mediated	O
through	O
a	O
focused	O
interaction	O
with	O
TCR	O
residues	O
Tyr97α	O
and	O
Trp97β	O
that	O
formed	O
an	O
aromatic	O
cap	O
over	O
a	O
central	O
‘	O
GDP	O
’	O
motif	O
that	O
was	O
common	O
to	O
all	O
APLs	O
.	O
	O
This	O
hotspot	O
binding	O
,	O
defined	O
as	O
a	O
localized	O
cluster	O
of	O
interactions	O
that	O
dominate	O
binding	O
energy	O
during	O
protein	O
-	O
protein	O
interactions	O
,	O
has	O
been	O
previously	O
shown	O
to	O
contribute	O
to	O
TCR	O
recognition	O
of	O
MHC	O
as	O
a	O
mechanism	O
that	O
tunes	O
T	O
cell	O
cross	O
-	O
reactivity	O
by	O
providing	O
fixed	O
anchor	O
points	O
that	O
enable	O
TCRs	O
to	O
tolerate	O
a	O
variable	O
peptide	O
cargo	O
.	O
	O
Alternatively	O
,	O
interactions	O
between	O
the	O
TCR	O
and	O
peptide	O
have	O
been	O
shown	O
to	O
dominate	O
the	O
energetic	O
landscape	O
during	O
ligand	O
engagement	O
,	O
ensuring	O
that	O
T	O
cells	O
retain	O
peptide	O
specificity	O
.	O
	O
The	O
binding	O
mechanism	O
utilized	O
by	O
the	O
1E6	O
TCR	O
during	O
pMHC	O
recognition	O
is	O
consistent	O
with	O
both	O
of	O
these	O
models	O
.	O
	O
Ligand	O
engagement	O
is	O
dominated	O
by	O
peptide	O
interactions	O
,	O
but	O
hotspot	O
-	O
like	O
interactions	O
with	O
the	O
central	O
GPD	O
motif	O
enable	O
the	O
1E6	O
TCR	O
to	O
tolerate	O
peptide	O
residues	O
that	O
vary	O
outside	O
of	O
this	O
region	O
,	O
explaining	O
how	O
T	O
cells	O
expressing	O
this	O
TCR	O
may	O
cross	O
-	O
react	O
with	O
a	O
large	O
number	O
of	O
different	O
peptides	O
.	O
	O
In	O
both	O
of	O
these	O
examples	O
,	O
self	O
-	O
recognition	O
is	O
mediated	O
by	O
TCR	O
residues	O
with	O
aromatic	O
side	O
chains	O
.	O
	O
Combined	O
with	O
evidence	O
demonstrating	O
that	O
aromatic	O
side	O
chains	O
are	O
conserved	O
in	O
the	O
CDR2	O
loops	O
of	O
TCRs	O
from	O
many	O
species	O
,	O
we	O
speculate	O
that	O
these	O
aromatic	O
residues	O
could	O
impart	O
a	O
level	O
of	O
“	O
stickiness	O
”	O
to	O
TCRs	O
,	O
which	O
might	O
be	O
enriched	O
in	O
an	O
autoimmune	O
setting	O
when	O
the	O
TCR	O
often	O
binds	O
in	O
a	O
nonoptimal	O
fashion	O
.	O
	O
For	O
example	O
,	O
all	O
of	O
the	O
stronger	O
ligands	O
encoded	O
larger	O
side	O
chains	O
(	O
Arg	O
or	O
Tyr	O
)	O
at	O
peptide	O
position	O
1	O
that	O
enabled	O
new	O
interactions	O
with	O
1E6	O
not	O
present	O
with	O
the	O
Ala	O
at	O
this	O
position	O
in	O
the	O
natural	O
preproinsulin	O
peptide	O
.	O
	O
We	O
have	O
recently	O
demonstrated	O
how	O
a	O
suboptimal	O
position	O
2	O
anchor	O
in	O
a	O
melanoma	O
-	O
derived	O
antigen	O
can	O
improve	O
TCR	O
binding	O
through	O
a	O
similar	O
mechanism	O
.	O
	O
Early	O
thermodynamic	O
analysis	O
of	O
TCR	O
-	O
pMHC	O
interactions	O
suggested	O
a	O
common	O
energetic	O
signature	O
,	O
driven	O
by	O
favorable	O
enthalpy	O
(	O
generally	O
mediated	O
through	O
an	O
increase	O
in	O
electrostatic	O
interactions	O
)	O
and	O
unfavorable	O
entropy	O
(	O
changes	O
from	O
disorder	O
to	O
order	O
).	O
	O
However	O
,	O
more	O
recent	O
data	O
have	O
shown	O
that	O
TCRs	O
can	O
utilize	O
a	O
range	O
of	O
energetic	O
strategies	O
during	O
pMHC	O
binding	O
,	O
currently	O
with	O
no	O
obvious	O
pattern	O
in	O
terms	O
of	O
TCR	O
affinity	O
,	O
binding	O
mechanism	O
,	O
or	O
specificity	O
(	O
pathogen	O
,	O
cancer	O
,	O
or	O
self	O
-	O
ligands	O
).	O
	O
The	O
weaker	O
APL	O
ligands	O
were	O
characterized	O
by	O
favorable	O
enthalpy	O
and	O
unfavorable	O
entropy	O
,	O
whereas	O
the	O
stronger	O
ligands	O
progressively	O
shifted	O
to	O
favorable	O
entropy	O
.	O
	O
Thus	O
,	O
the	O
enhanced	O
antigen	O
potency	O
was	O
probably	O
mediated	O
through	O
a	O
shift	O
from	O
an	O
induced	O
fit	O
to	O
a	O
lock	O
-	O
and	O
-	O
key	O
interaction	O
between	O
the	O
stronger	O
ligands	O
(	O
less	O
requirement	O
for	O
energetically	O
unfavorable	O
disorder	O
-	O
to	O
-	O
order	O
changes	O
),	O
resulting	O
in	O
a	O
more	O
energetically	O
favorable	O
ΔG	O
value	O
.	O
	O
Importantly	O
,	O
the	O
preproinsulin	O
-	O
derived	O
epitope	O
was	O
one	O
of	O
the	O
least	O
potent	O
peptides	O
,	O
demonstrating	O
that	O
the	O
1E6	O
T	O
cell	O
clone	O
had	O
the	O
ability	O
to	O
respond	O
to	O
different	O
peptide	O
sequences	O
with	O
far	O
greater	O
potency	O
.	O
	O
The	O
RQFGPDWIVA	O
peptide	O
,	O
which	O
was	O
substantially	O
more	O
potent	O
than	O
the	O
preproinsulin	O
peptide	O
,	O
is	O
within	O
the	O
proteome	O
of	O
a	O
common	O
human	O
pathogen	O
(	O
C	O
.	O
asparagiforme	O
),	O
demonstrating	O
the	O
potential	O
for	O
an	O
encounter	O
between	O
a	O
naive	O
1E6	O
-	O
like	O
T	O
cell	O
and	O
a	O
foreign	O
peptide	O
with	O
a	O
more	O
potent	O
ligand	O
that	O
might	O
then	O
break	O
self	O
-	O
tolerance	O
.	O
	O
Further	O
experiments	O
will	O
be	O
required	O
to	O
determine	O
whether	O
any	O
naturally	O
presented	O
,	O
human	O
pathogen	O
–	O
derived	O
peptides	O
act	O
as	O
active	O
ligands	O
for	O
1E6	O
,	O
but	O
our	O
work	O
presented	O
here	O
demonstrates	O
that	O
it	O
is	O
at	O
least	O
feasible	O
for	O
an	O
autoimmune	O
TCR	O
to	O
bind	O
to	O
a	O
different	O
peptide	O
sequence	O
that	O
could	O
be	O
present	O
in	O
a	O
pathogen	O
proteome	O
with	O
substantially	O
higher	O
affinity	O
and	O
potency	O
than	O
the	O
interaction	O
it	O
might	O
use	O
to	O
attack	O
self	O
-	O
tissue	O
.	O
	O
In	O
summary	O
,	O
this	O
investigation	O
into	O
the	O
molecular	O
basis	O
of	O
T	O
cell	O
cross	O
-	O
reactivity	O
using	O
a	O
clinically	O
relevant	O
cytotoxic	O
CD8	O
+	O
T	O
cell	O
clone	O
that	O
kills	O
human	O
pancreatic	O
β	O
cells	O
provides	O
answers	O
to	O
a	O
number	O
of	O
previously	O
outstanding	O
questions	O
.	O
	O
First	O
,	O
our	O
data	O
shows	O
that	O
a	O
single	O
TCR	O
has	O
the	O
potential	O
to	O
functionally	O
(	O
assessed	O
through	O
T	O
cell	O
activation	O
)	O
bind	O
to	O
different	O
ligands	O
with	O
affinities	O
ranging	O
across	O
3	O
orders	O
of	O
magnitude	O
.	O
	O
Second	O
,	O
this	O
is	O
the	O
first	O
example	O
in	O
which	O
ligands	O
have	O
been	O
identified	O
and	O
characterized	O
for	O
a	O
human	O
autoreactive	O
TCR	O
that	O
are	O
substantially	O
more	O
potent	O
than	O
the	O
natural	O
self	O
-	O
ligand	O
,	O
demonstrating	O
the	O
potential	O
for	O
a	O
pathogenic	O
ligand	O
to	O
break	O
self	O
-	O
tolerance	O
and	O
prime	O
self	O
-	O
reactive	O
T	O
cells	O
.	O
	O
Third	O
,	O
this	O
first	O
structural	O
analysis	O
of	O
a	O
cross	O
-	O
reactive	O
human	O
MHCI	O
–	O
restricted	O
autoimmune	O
TCR	O
showed	O
that	O
degeneracy	O
was	O
mediated	O
through	O
TCR	O
-	O
pMHC	O
anchoring	O
by	O
a	O
conserved	O
minimal	O
binding	O
peptide	O
motif	O
.	O
	O
Finally	O
,	O
TCR	O
ligand	O
discrimination	O
was	O
characterized	O
by	O
an	O
energetic	O
shift	O
from	O
an	O
enthalpically	O
to	O
entropically	O
driven	O
interaction	O
.	O
	O
Our	O
demonstration	O
of	O
the	O
molecular	O
mechanism	O
governing	O
cross	O
-	O
reactivity	O
by	O
this	O
preproinsulin	O
reactive	O
human	O
CD8	O
+	O
T	O
cell	O
clone	O
supports	O
the	O
notion	O
first	O
put	O
forward	O
by	O
Wucherpfennig	O
and	O
Strominger	O
that	O
molecular	O
mimicry	O
could	O
mediate	O
autoimmunity	O
and	O
has	O
far	O
-	O
reaching	O
implications	O
for	O
the	O
complex	O
nature	O
of	O
T	O
cell	O
antigen	O
discrimination	O
.	O
	O
The	O
1E6	O
T	O
cell	O
clone	O
reacts	O
with	O
a	O
broad	O
sensitivity	O
range	O
to	O
APLs	O
.	O
	O
(	O
A	O
and	O
B	O
)	O
The	O
1E6	O
T	O
cell	O
clone	O
was	O
tested	O
in	O
a	O
peptide	O
dilution	O
assay	O
,	O
in	O
triplicate	O
,	O
with	O
MVWGPDPLYV	O
(	O
gray	O
),	O
YLGGPDFPTI	O
(	O
red	O
),	O
ALWGPDPAAA	O
(	O
blue	O
),	O
AQWGPDPAAA	O
(	O
green	O
),	O
RQFGPDWIVA	O
(	O
dark	O
blue	O
),	O
RQWGPDPAAV	O
(	O
purple	O
),	O
YQFGPDFPTA	O
(	O
yellow	O
),	O
and	O
RQFGPDFPTI	O
(	O
cyan	O
)	O
peptides	O
presented	O
by	O
HLA	O
-	O
A	O
*	O
0201	O
–	O
expressing	O
C1R	O
cells	O
for	O
release	O
of	O
MIP	O
-	O
1β	O
(	O
A	O
)	O
and	O
killing	O
(	O
B	O
).	O
	O
Tm	O
values	O
were	O
calculated	O
using	O
a	O
Boltzmann	O
fit	O
to	O
each	O
set	O
of	O
data	O
.	O
	O
3D	O
and	O
2D	O
binding	O
analysis	O
of	O
the	O
1E6	O
TCR	O
with	O
A2	O
-	O
ALW	O
and	O
the	O
APLs	O
.	O
	O
(	O
A	O
–	O
H	O
)	O
Binding	O
affinity	O
of	O
the	O
1E6	O
TCR	O
interaction	O
at	O
25	O
°	O
C	O
using	O
SPR	O
.	O
	O
Eight	O
serial	O
dilutions	O
of	O
the	O
1E6	O
TCR	O
were	O
measured	O
(	O
shown	O
in	O
the	O
inset	O
);	O
representative	O
data	O
from	O
3	O
independent	O
experiments	O
are	O
plotted	O
.	O
	O
The	O
equilibrium	O
binding	O
constant	O
(	O
KD	O
)	O
values	O
were	O
calculated	O
using	O
a	O
nonlinear	O
curve	O
fit	O
(	O
y	O
=	O
[	O
P1x	O
]/[	O
P2	O
+	O
X	O
]).	O
	O
In	O
order	O
to	O
calculate	O
each	O
response	O
,	O
the	O
1E6	O
TCR	O
was	O
also	O
injected	O
over	O
a	O
control	O
sample	O
(	O
HLA	O
-	O
A	O
*	O
0201	O
–	O
ILAKFLHWL	O
)	O
that	O
was	O
deducted	O
from	O
the	O
experimental	O
data	O
.	O
	O
(	O
A	O
)	O
1E6	O
-	O
A2	O
-	O
MVWGPDPLYV	O
(	O
approximate	O
value	O
);	O
(	O
B	O
)	O
1E6	O
-	O
A2	O
-	O
YLGGPDFPTI	O
(	O
approximate	O
value	O
);	O
(	O
C	O
)	O
1E6	O
-	O
A2	O
-	O
ALWGPDPAAA	O
;	O
(	O
D	O
)	O
1E6	O
-	O
A2	O
-	O
AQWGPDPAAA	O
;	O
(	O
E	O
)	O
1E6	O
-	O
A2	O
-	O
RQFGPDWIVA	O
;	O
(	O
F	O
)	O
1E6	O
-	O
A2	O
-	O
RQWGPDPAAV	O
;	O
(	O
G	O
)	O
1E6	O
-	O
A2	O
-	O
YQFGPDFPTA	O
;	O
and	O
(	O
H	O
)	O
1E6	O
-	O
A2	O
-	O
RQFGPDFPTI	O
.	O
(	O
I	O
)	O
ΔG	O
values	O
,	O
calculated	O
from	O
SPR	O
experiments	O
,	O
plotted	O
against	O
1	O
/	O
EC50	O
(	O
the	O
reciprocal	O
peptide	O
concentration	O
required	O
to	O
reach	O
half	O
-	O
maximal	O
1E6	O
T	O
cell	O
killing	O
)	O
showing	O
Pearson	O
’	O
s	O
coefficient	O
analysis	O
(	O
r	O
)	O
and	O
P	O
value	O
(	O
including	O
approximate	O
values	O
from	O
A	O
and	O
B	O
).	O
	O
(	O
J	O
)	O
Effective	O
2D	O
affinity	O
(	O
AcKa	O
)	O
calculated	O
using	O
adhesion	O
frequency	O
assays	O
,	O
using	O
at	O
least	O
5	O
cell	O
pairs	O
,	O
and	O
calculated	O
as	O
an	O
average	O
of	O
100	O
cell	O
cell	O
contacts	O
.	O
	O
The	O
1E6	O
TCR	O
and	O
each	O
peptide	O
are	O
colored	O
according	O
to	O
the	O
APL	O
used	O
in	O
the	O
complex	O
as	O
in	O
Figure	O
1	O
.	O
(	O
B	O
)	O
Position	O
of	O
the	O
1E6	O
TCR	O
CDR	O
loops	O
(	O
multicolored	O
lines	O
)	O
in	O
each	O
complex	O
.	O
	O
The	O
ALWGPDPAAA	O
peptide	O
(	O
green	O
sticks	O
)	O
is	O
shown	O
in	O
the	O
HLA	O
-	O
A	O
*	O
0201	O
binding	O
groove	O
(	O
gray	O
surface	O
).	O
(	O
C	O
)	O
The	O
Cα	O
backbone	O
conformation	O
of	O
each	O
APL	O
(	O
multicolored	O
illustration	O
)	O
in	O
the	O
context	O
of	O
the	O
HLA	O
-	O
A	O
*	O
0201	O
α1	O
helices	O
(	O
gray	O
illustration	O
).	O
(	O
D	O
)	O
Crossing	O
angle	O
of	O
the	O
1E6	O
TCR	O
(	O
multicolored	O
lines	O
)	O
calculated	O
using	O
previously	O
published	O
parameters	O
in	O
the	O
context	O
of	O
the	O
ALWGPDPAAA	O
peptide	O
(	O
green	O
sticks	O
)	O
bound	O
in	O
the	O
HLA	O
-	O
A	O
*	O
0201	O
binding	O
groove	O
(	O
gray	O
surface	O
).	O
	O
A	O
conserved	O
interaction	O
with	O
a	O
GPD	O
motif	O
underpins	O
the	O
1E6	O
TCR	O
interaction	O
with	O
the	O
APLs	O
.	O
	O
The	O
rest	O
of	O
the	O
peptide	O
,	O
and	O
the	O
MHCα1	O
helix	O
,	O
are	O
shown	O
as	O
a	O
gray	O
illustration	O
.	O
	O
The	O
1E6	O
TCR	O
makes	O
distinct	O
peptide	O
contacts	O
with	O
peripheral	O
APL	O
residues	O
.	O
	O
Boxes	O
show	O
total	O
contacts	O
between	O
the	O
1E6	O
TCR	O
and	O
each	O
peptide	O
ligand	O
.	O
	O
(	O
D	O
)	O
Interaction	O
between	O
the	O
1E6	O
TCR	O
(	O
green	O
illustration	O
and	O
sticks	O
)	O
and	O
A2	O
-	O
AQWGPDPAAA	O
(	O
green	O
illustration	O
and	O
sticks	O
).	O
(	O
E	O
)	O
Interaction	O
between	O
the	O
1E6	O
TCR	O
(	O
dark	O
blue	O
illustration	O
and	O
sticks	O
)	O
and	O
A2	O
-	O
RQFGPDWIVA	O
(	O
dark	O
blue	O
illustration	O
and	O
sticks	O
).	O
(	O
F	O
)	O
Interaction	O
between	O
the	O
1E6	O
TCR	O
(	O
purple	O
illustration	O
and	O
sticks	O
)	O
and	O
A2	O
-	O
RQWGPDPAAV	O
(	O
purple	O
illustration	O
and	O
sticks	O
).	O
(	O
G	O
)	O
Interaction	O
between	O
the	O
1E6	O
TCR	O
(	O
yellow	O
illustration	O
and	O
sticks	O
)	O
and	O
A2	O
-	O
YQFGPDFPTA	O
(	O
yellow	O
illustration	O
and	O
sticks	O
).	O
(	O
H	O
)	O
Interaction	O
between	O
the	O
1E6	O
TCR	O
(	O
cyan	O
illustration	O
and	O
sticks	O
)	O
and	O
A2	O
-	O
RQFGPDFPTI	O
(	O
cyan	O
illustration	O
and	O
sticks	O
).	O
	O
Comparison	O
of	O
ligated	O
and	O
unligated	O
APLs	O
.	O
	O
Superposition	O
of	O
each	O
APL	O
in	O
unligated	O
form	O
and	O
ligated	O
to	O
the	O
1E6	O
TCR	O
.	O
	O
All	O
unligated	O
pMHCs	O
are	O
shown	O
as	O
light	O
green	O
illustrations	O
.	O
	O
Peptide	O
sequences	O
are	O
shown	O
underneath	O
each	O
structure	O
aligned	O
with	O
the	O
peptide	O
structure	O
.	O
	O
A	O
large	O
conformational	O
shift	O
was	O
observed	O
for	O
Tyr8	O
in	O
the	O
ligated	O
versus	O
unligated	O
states	O
(	O
black	O
circle	O
).	O
(	O
B	O
)	O
A2	O
-	O
YLGGPDFPTI	O
(	O
red	O
sticks	O
).	O
(	O
C	O
)	O
A2	O
-	O
ALWGPDPAAA	O
(	O
blue	O
sticks	O
)	O
reproduced	O
from	O
previous	O
published	O
data	O
.	O
(	O
D	O
)	O
A2	O
-	O
AQWGPDPAAA	O
(	O
green	O
sticks	O
).	O
(	O
E	O
)	O
A2	O
-	O
RQFGPDWIVA	O
(	O
dark	O
blue	O
sticks	O
).	O
(	O
F	O
)	O
A2	O
-	O
RQWGPDPAAV	O
(	O
purple	O
sticks	O
).	O
(	O
G	O
)	O
A2	O
-	O
YQFGPDFPTA	O
(	O
yellow	O
sticks	O
).	O
(	O
H	O
)	O
A2	O
-	O
RQFGPDFPTI	O
(	O
cyan	O
sticks	O
).	O
	O
Interactions	O
between	O
the	O
1E6	O
TCR	O
and	O
the	O
MHC	O
α1	O
helix	O
residues	O
Arg65	O
,	O
Lys66	O
,	O
and	O
Gln72	O
.	O
	O
Hydrogen	O
bonds	O
are	O
shown	O
as	O
red	O
dotted	O
lines	O
;	O
vdW	O
contacts	O
are	O
shown	O
as	O
black	O
dotted	O
lines	O
.	O
	O
MHCα1	O
helix	O
are	O
shown	O
in	O
gray	O
illustrations	O
.	O
	O
Thermodynamic	O
analysis	O
of	O
the	O
1E6	O
TCR	O
with	O
A2	O
-	O
ALWGPDPAAA	O
and	O
the	O
APLs	O
.	O
	O
Eight	O
serial	O
dilutions	O
of	O
the	O
1E6	O
TCR	O
were	O
injected	O
,	O
in	O
duplicate	O
,	O
over	O
each	O
immobilized	O
APL	O
and	O
A2	O
-	O
ALW	O
at	O
5	O
°	O
C	O
,	O
13	O
°	O
C	O
,	O
18	O
°	O
C	O
,	O
25	O
°	O
C	O
,	O
30	O
°	O
C	O
,	O
and	O
37	O
°	O
C	O
.	O
	O
The	O
equilibrium	O
binding	O
constant	O
(	O
KD	O
)	O
values	O
were	O
calculated	O
using	O
a	O
nonlinear	O
curve	O
fit	O
(	O
y	O
=	O
[	O
P1x	O
]/[	O
P2	O
+	O
X	O
]),	O
and	O
thermodynamic	O
parameters	O
were	O
calculated	O
according	O
to	O
the	O
Gibbs	O
-	O
Helmholtz	O
equation	O
(	O
ΔG	O
°	O
=	O
ΔH	O
−	O
TΔS	O
°).	O
	O
The	O
binding	O
free	O
energies	O
,	O
ΔG	O
°	O
(	O
ΔG	O
°	O
=	O
RTlnKD	O
),	O
were	O
plotted	O
against	O
temperature	O
(	O
K	O
)	O
using	O
nonlinear	O
regression	O
to	O
fit	O
the	O
3	O
-	O
parameters	O
van	O
’	O
t	O
Hoff	O
equation	O
(	O
RT	O
ln	O
KD	O
=	O
ΔH	O
°	O
–	O
TΔS	O
°	O
+	O
ΔCp	O
°[	O
T	O
-	O
T0	O
]	O
–	O
TΔCp	O
°	O
ln	O
[	O
T	O
/	O
T0	O
]	O
with	O
T0	O
=	O
298	O
K	O
).	O
	O
(	O
A	O
)	O
1E6	O
-	O
A2	O
-	O
ALWGPDPAAA	O
;	O
(	O
B	O
)	O
1E6	O
-	O
A2	O
-	O
AQWGPDPAAA	O
;	O
(	O
C	O
)	O
1E6	O
-	O
A2	O
-	O
RQFGPDWIVA	O
;	O
(	O
D	O
)	O
1E6	O
-	O
A2	O
-	O
RQWGPDPAAV	O
,	O
(	O
E	O
)	O
1E6	O
-	O
A2	O
-	O
YQFGPDFPTA	O
;	O
and	O
(	O
F	O
)	O
1E6	O
-	O
A2	O
-	O
RQFGPDFPTI	O
.	O
	O
1E6	O
TCR	O
-	O
pMHC	O
contacts	O
,	O
affinity	O
measurements	O
and	O
thermodynamics	O
	O