Structural insights into the inhibition mechanism of human sterol O-acyltransferase 1 by a competitive inhibitor

Article abstract of DOI:10.1038/s41467-020-16288-4

Sterol O-acyltransferase 1 (SOAT1) is an endoplasmic reticulum (ER) resident, multi-transmembrane enzyme that belongs to the membrane-bound O-acyltransferase (MBOAT) family. It catalyzes the esterification of cholesterol to generate cholesteryl esters for cholesterol storage. SOAT1 is a target to treat several human diseases. However, its structure and mechanism remain elusive since its discovery. Here, we report the structure of human SOAT1 (hSOAT1) determined by cryo-EM. hSOAT1 is a tetramer consisted of a dimer of dimer. The structure of hSOAT1 dimer at 3.5 ? resolution reveals that a small molecule inhibitor CI-976 binds inside the catalytic chamber and blocks the accessibility of the active site residues H460, N421 and W420. Our results pave the way for future mechanistic study and rational drug design targeting hSOAT1 and other mammalian MBOAT family members.

Full text of DOI:10.1038/s41467-020-16288-4

A
R
T
I
C
L
E
https://doi.org/10.1038/s41467-020-16288-4
OPEN
S t r u ct u r a l i n s i g h ts i n t o t h e i n hi b i ti o n m e c h a n i s m
o f h u m a n s t e r ol O- a c y l t r a n s f e r a s e
1 b y a c o m p e t i t i v e i n h i b i to r
1
1
2
1
1
3
C h en g c h e n g G u a n , Y a ng e N i u , S i- C o ng C he n , Y u n l u K a n g , J i n g - X i a n g W u , K o j i N i sh i ,
✉
1 , 4 , 5
3
3
2
,
4
,
5
C
a
t
h
e
r
i
n
e
C
.
Y
.
C
h
a
n
g
,
T
a
-
Y
u
a
n
C
h
a
n
g
,
T
u
o
p
i
n
g
L
u
o
&
L
e
i
C
h
e
n
S
t
e
r
o
l O
-
a
c
y
l
t
r
a
n
s
f
e
r
a
s
e
1
(
S
O
A
o
T
1
)
i
s
a
n
e
n
d
o
p
l
a
a
t
s
m
i
c
r
e
t
i
c
u
l
u
m
(
E
R
)
r
e
s
i
d
e
n
t
,
m
u
l
t
i
-
t
r
a
n
s
m
e
I
m
b
r
a
n
e
e
n
z
y
m
e
t
h
a
t
r
b
e
l
n
g
s
t
o
f
t
h
e
m
e
m
b
r
n
o
u
e-
b
o
u
n
d
O
-
a
c
y
l
t
r
a
n
s
f
e
r
a
s
e
(
M
B
O
A
T
)
f
a
m
i
l
y
.
t
c
a
t
a
l
y
z
e
s
t
h
e
e
s
t
e
i
fi
c
a
t
i
o
n
o
c
h
o
l
e
s
te
r
o
l
g
e
n
e
ra
t
e
c
h
o
l
es
t
e
r
y
l
e
s
t
e
r
s
f
o
r
c
h
o
-
l
e
s
t
e
r
o
l
s
t
o
r
a
g
e
.
S
O
A
T
1
i
s
a
t
a
r
g
e
t
t
o
t
r
e
a
t
s
e
v
e
r
a
l
h
m
a
n
d
e
t
i
s
e
a
s
e
s
.
H
o
w
e
v
e
r
,
i
t
s
s
t
r
u
c
t
u
r
e
a
n
d
m
e
c
h
a
n
is
m
r
e
m
a
i
n
e
l
u
s
i
v
e
s
i
n
c
e
i
t
s
d
i
s
c
o
v
e
r
y
.
H
1
e
re
,
w
r
e
p
o
r
t
t
h
e
s
t
r
u
c
t
u
r
e
o
f
h
um
a
n
S
O
A
T
1
(
h
S
O
A
T
1
)
d
e
t
e
r
m
i
n
e
d
b
y
c
r
y
o
-
E
M
.
h
S
O
A
T
r
i
s
a
t
e
r
a
m
e
r
c
o
n
s
i
s
t
e
d
o
f
a
d
i
m
e
r
o
f
e
e
y
d
i
m
e
r
.
T
r
h
e
s
t
r
u
c
t
u
r
e
o
f
h
S
O
A
T
1
d
i
m
e
r
a
t
3
.
5
Å
e
s
o
l
u
t
i
o
n
r
e
v
e
a
l
s
t
h
a
t
a
s
m
a
l
l
m
o
l
e
c
u
l
i
n
h
i
b
i
t
o
C
I
-
9
7
6
b
in
d
s
i
n
s
i
d
e
t
h
e
c
a
ta
l
y
t
i
c
c
h
a
m
b
e
r
a
n
d
b
l
o
c
k
s
t
h
e
a
c
c
e
s
s
i
b
i
l
i
t
y
o
f
t
h
e
a
c
t
i
v
d
r
s
i
t
e
d
r
e
s
i
d
u
e
s
H
4
6
0
,
N
4
2
1
a
n
d
W
4
2
0
.
O
u
r
r
e
s
u
l
t
s
p
a
v
r
e
t
h
e
w
a
y
f
o
r
f
u
t
u
r
e
m
e
c
h
a
n
i
s
t
i
c
s
t
u
a
n
r
at
i
o
n
a
l
d
r
u
g
d
e
s
i
g
n
t
a
r
g
e
t
i
n
g
h
S
O
A
T
1
a
n
d
o
t
h
e
m
a
m
m
a
l
i
a
n
M
B
O
A
T
f
a
m
i
l
y
m
e
m
b
e
s
.
1 State Key Laboratory of Membrane Biology, Institute of Molecular Medicine, Peking University, Beijing Key Laboratory of Cardiometabolic Molecular
Medicine, 100871 Beijing, China. 2 Key Laboratory of Bioorganic Chemistry and Molecular Engineering, Ministry of Education and Beijing National Laboratory
for Molecular Science, College of Chemistry and Molecular Engineering, Peking University, 100871 Beijing, China. 3 Department of Biochemistry and Cell
Biology, Geisel School of Medicine at Dartmouth, Hanover, NH 03755, USA. 4 Peking-Tsinghua Center for Life Sciences, Peking University, 100871
✉
Beijing, China. 5 Academy for Advanced Interdisciplinary Studies, Peking University, 100871 Beijing, China. email: chenlei2016@pku.edu.cn
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
|
(
2
0
2
0
)
1
1
:
2
4
7
8
|
h
t
t
p
s
:
/
/
d
o
i
.
o
r
g
/
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
2
0
-
1
6
2
8
8
-
4
|
w
w
w
.
n
a
t
u
r
e
.
c
o
m
/
n
a
t
u
r
e
c
o
m
m
u
n
i
c
a
t
i
o
n
s
1

A
R
T
I
C
L
E
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
|
h
t
t
p
s
:
/
/
d
o
i
.
o
r
g
/
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
2
0
-
1
6
2
8
8
-
4
holesterol is an essential lipid molecule in the cell mem- hSOAT1 became predominantly dimeric with the deletion of the
branes of all vertebrate. It is important for maintaining the N-terminal tetramerization domain21. We compared the recom-
fluidity and integrity of the membrane and is the pre- binant SOAT1 enzyme activities in HEK293F suspension cell
C
cursor for the biosynthesis of other crucial endogenous molecules, after infection by the recombinant BacMam viruses, versus those
such as steroid hormones and bile acids. In addition, cholesterol expressed in the stable transfectant of AC29 cell line22, or the
can modulate the activity of many membrane proteins such as adherent HEK293 cell23. We used the crude cell extracts solubi-
3
GPCR1 and ion channels2. The concentration of cellular free lized by CHAPS as the enzyme source22, added the H-oleoyl-
cholesterol is highly regulated3. Excessive intracellular cholesterol CoA and cholesterol present in taurocholate/phospholipid/cho-
may form cholesteryl esters, which are catalyzed by the enzyme, lesterol mixture as the substrates. The crude cell extracts from
sterol O-acyltransferase (SOAT), also called acyl-coenzyme A: uninfected/untransfected HEK293 cells, or from AC29 cells (a
cholesterol acyltransferase (ACAT). SOAT catalyzes the reaction CHO cell mutant devoid of endogenous SOAT1 protein24 were
between long chain fatty acyl-CoA and intracellular cholesterol to used as controls). The results (Supplementary Fig. 1b) showed
form the more hydrophobic cholesteryl ester, which is then stored that AC29 cells expressed essentially no SOAT1 activity; both the
in lipid droplets within the cell or transported in secreted lipo- HEK293F cells and the adherent HEK293 monolayers expressed
protein particles to other tissues that need cholesterol. In addition very low (endogenous) SOAT1 activity. Both the stable trans-
to cholesterol, SOAT can use multiple sterols as substrates and fectant of AC29 cells and the stable transfectant of HEK293 cells
activators4. Because of its functional importance, SOAT1 is a expressed markedly higher SOAT1 activities, with the adherent
potential drug target for Alzheimer’s disease5, atherosclerosis6 HEK293 cells producing much higher recombinant SOAT1
and several types of cancers7–10
.
activity than AC29 cells. These results are consistent with what
Previous studies have shown that SOAT1 is an ER-localized were reported previously. Strikingly, HEK293F cells infected with
multi-transmembrane protein that is evolutionary conserved SOAT1 tetramer virus or with SOAT1 dimeric virus expressed
from yeast to humans11. There are two SOAT enzymes in SOAT activities at levels even much higher than those found in
mammals: SOAT1 and SOAT2, which have a protein sequence the stable transfectant of HEK293 cells, demonstrating that the
identity of 48% in human (258 out of 537 residues aligned). recombinant BacMam viruses infected 293-F cells is an excellent
SOAT1 is ubiquitously expressed in many types of cells12; while method for producing large amount of catalytically active
SOAT2 is mainly expressed in the small intestine and liver13. Due recombinant SOAT1 (Supplementary Fig. 1b).
to the pathophysiological importance of SOAT, many SOAT
In order to measure the cholesterol-activated O-acyltransferase
inhibitors of various structural types have been made. The small activity of the purified proteins, we developed the in vitro
molecule CI-976 is a prototypical member of SOAT inhibitors hSOAT1 assay using fluorescence-labeled NBD-cholesterol and
that belong to the fatty acyl amide analog family. Studies showed oleoyl-CoA as substrates based on the previously reported SOAT1
that feeding CI-976 reduced atherosclerotic plaques as well as whole cell assay25 (Fig. 1a and Supplementary Fig. 1c–h). The
reduced plasma cholesterol levels in animals fed with high cho- results show that the activity of hSOAT1 tetramer is linear within
lesterol diet14. CI-976 exhibits competitive inhibition against fatty the first 15 min (Supplementary Fig. 1h), and both the purified
acyl-CoA15,16
.
tetrameric and dimeric hSOAT1 proteins showed cholesterol-
SOAT1 is the founding member of the membrane-bound O- activated O-acyltransferase activity in detergent micelles (Fig. 1b,
acyltransferase (MBOAT) enzyme family, which transfers the acyl Supplementary Fig. 1d–h). SOATs are allosteric enzymes that can
chain onto various substrates, including lipids, peptides and small be activated by cholesterol4 and it is predicted that SOATs have
proteins. There are 11 MBOAT family members in humans17, two functional distinct cholesterol binding sites. One site is the
which participate in many physiological processes, such as the last substrate-binding site and the other is the allosteric activating site
step of triglyceride biosynthesis catalyzed by acyl-CoA: dia- that provides the feedback regulation mechanism regarding
cylglycerol acyltransferase 1 (DGAT1), the maturation of cholesterol concentration in the ER4. Our results suggest the
hedgehog morphogen catalyzed by hedgehog acyltransferase NBD-cholesterol can act as a substrate while not the activator
(HHAT), and the acylation of peptide hormone ghrelin catalyzed (Fig. 1b). Therefore, this assay provides the opportunities to study
by ghrelin O-acyltransferase (GOAT). Recently, the X-ray crystal cholesterol activation effect. As expected, the hSOAT1 enzyme
structure of DltB from Streptococcus thermophilus was deter- activity is inhibited by CI-976 in a dose-dependent manner
mined and is the only available structure of an MBOAT family (Fig. 1c). Moreover, we measured the thermostability of hSOAT1
member18. The low-sequence conservation between DltB and dimer by fluorescence-detection size-exclusion chromatography
SOAT (14.6% identity) hinders the accurate modeling of the (FSEC)26 (Fig. 1d) and found the solubilized hSOAT1 dimer
human SOAT1 structure. Therefore, despite the important phy- dissociated into putative monomer upon heat treatment while the
siological functions of human SOAT enzymes, their architecture presence of CI-976 significantly enhanced the stability of
and mechanism remain elusive due to a lack of high-resolution hSOAT1 dimer (Fig. 1e).
structures. In this study, several human SOAT1 (hSOAT1)
To investigate the structure of hSOAT1, we prepared the
structures were determined. These structures reveal the archi- sample of hSOAT1 in the presence of CI-976 in detergent
tecture of hSOAT1 and the binding site of the competitive inhi- micelles for cryo-EM analysis. The 3D classification showed the
bitor CI-976, and provide a structural basis to understand the hSOAT1 sample had severe conformational heterogeneity
inhibition mechanism of these enzymes by small molecules.
(Supplementary Fig. 2). One of the 3D classes can be further
refined to 12 Å and the map allow the visualization of the general
shape of hSOAT1, in which the cytosolic N terminal oligomer-
ization domain lays above the transmembrane domain. More-
Results
Structure determination. N-terminal GFP-tagged full-length over, the central slice of the transmembrane domain density map
(1–550) and N-terminal truncated hSOAT1 (66–550), solubi- indicated the hSOAT1 is a tetramer composed of dimer of dimer
lized in detergent micelles migrated before and after mouse TPC1 (Supplementary Fig. 2d), which is consistent with previous
channel, which is a well-characterized dimeric channel with a biochemical data21. To further stabilize the structure of
molecular weight of 189 kDa (ref.19) (Supplementary Fig. 1a). transmembrane domain, we reconstituted hSOAT1 into a lipid
This is consistent with previous studies that hSOAT1 is a tetra- nanodisc (Supplementary Fig. 3a and b) for cryo-EM analysis
mer with a molecular weight of around 260 kDa (ref.20) and that (Supplementary Fig. 4). The top view of 2D class averages of the
2
N
A
T
U
R
E
C
O
M
M
U
NI
C
A
T
I
O
N
S
|
(
2
0
2
0
)
1
1
:
2
4
7
8
|
h
t
t
p
s
:
/
/
d
o
i
.
o
r
g
/
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
2
0
-
1
6
2
8
8
-
4
|
w
w
w
.
n
a
t
u
r
e
.
c
o
m
/
n
a
t
u
r
e
c
o
m
m
u
n
i
c
a
t
i
o
n
s

N
A
T
U
R
E
C
O
M
M
UN
I
CA
T
I
O
N
S
|
h
t
t
p
s
:
//d
o
i
.
o
r
g
/
1
0
.
1
0
38
/
s
4
1
4
6
7
-
0
2
0
-
1
6
2
8
8
-
4
A
R
T
I
C
L
E
a
O
H
CoA
H
S
H
H
O
hSOAT1
+
H
H
+
CoA-SH
H
H
O
HO
Oleoyl-CoA
Cholesterol
Cholesteryl-oleate
b
c
O
H
N
0.4
0.3
0.2
0.1
0.0
O
O
O
CI-976
IC50 = 6.95 μM (6.49,7.74)
100
80
60
40
20
0
–3
–2
–1
0
1
2
3
Log(concentration μM)
d
e
Vehicle 4 °C
2.0
2500
2000
1500
1000
500
Dimer
Vehicle 45 °C
****
CI-976 4 °C
CI-976 45 °C
1.5
1.0
0.5
0.0
Monomer
0
1.5
2.0
Elution volume (ml)
2.5
Vehicle
CI-976
Fig. 1 The enzymatic reaction catalyzed by hSOAT1. a
C
h
r
e
e
m
i
t
c
i
a
o
l
n
s
.
t
T
r
u
h
c
e
t
u
h
r
y
e
d
s
r
o
f
t
h
g
e
r
s
o
u
u
b
p
s
t
t
h
r
a
a
t
t
e
a
s
c
a
c
n
e
d
p
p
r
a
o
c
d
y
u
l
c
g
t
s
o
u
f
p
h
i
S
s
O
h
A
T
h
1
e
n
t
z
e
y
d
m
i
e
n
a
r
e
d
s
. b
h
o
T
w
h
n
e
.
T
h
e
v
r
a
e
d
o
d
n
a
e
s
h
e
e
c
d
t
l
i
n
e
i
n
d
i
c
a
t
e
s
t
h
e
b
o
n
d
t
h
a
t
i
s
b
r
o
k
e
n
d
u
r
i
n
g
a
c
y
l
-
t
r
a
n
s
f
e
r
a
c
o
x
y
l
t
s
r
o
i
g
l
i
g
h
r
e
a
c
t
i
t
i
f
f
o
d
f
e
c
h
a
o
l
o
r
h
e
n
s
e
s
s
t
s
t
e
r
o
l
(
C
3
o
H
L
o
t
e
)
l
o
n
t
a
n
h
e
y
e
s
d
t
e
e
i
r
p
a
5
ifi
e
c
n
a
d
o
b
t
e
i
o
n
t
a
n
t
.
r
s
D
e
e
a
c
p
t
i
l
o
e
r
m
n
o
. c
f
N
C
B
h
n
e fi
D
-
c
i
s
h
c
m
n
o
a
l
l
e
)
e
s
a
. d
s
t
t
r
n
e
u
s
r
c
o
t
l
u
c
a
t
o
n
s
a
f
d
e
l
y
C
a
t
z
e
d
9
d
b
y
h
S
O
d
A
T
s
, n =
e
1
e
t
-
e
d
t
e
3
-
r
p
a
m
e
d
l
c
r
e
o
a
n
n
t
c
n
d
d
h
i
i
m
e
r
o
e
x
(
n
p
c
D
c
a
u
t
r
d
a
v
e
a
e
r
o
e
f
s
h
s
h
h
S
o
O
p
w
A
n
a
1
a
s
t
n
a
m
e
a
n
y
a
m
u
(
n
s
e
s
t
a
C
n
d
9
a
7
r
d
6
v
i
t
i
, n =
d
s
b
i
o
g
i
c
l
l
i
n
a
m
s
)
e
m
r
e
I
-
7
6
a
n
d
o
e
n
i
n
b
i
t
i
T
,
e
t
r
r
b
E
b
y
I-
(
T
h
e
fi
a
t
a
a
p
i
n
i
s
a
n
a
r
t
i fi
c
C
l
p
i
n
a
d
t
a
f
a
e
s
h
o
w
a
s
t
a
e
r
a
d
t
e
e fl
v
i
a
t
i
o
o
n
s
b
e
i
o
g
i
a
l
l
y
i
n
d
e
en
nt
a
m
l
e
s
d
m
F
e
r
s
i
n
p
a
re
n
t
e
r
e
t
h
r
a
g
e
f
o
r
I
0
o
t
i
n
r
o
c
u
r
v
t
t
i
g
R
e
p
r
n
i
v
u
r
e
s
c
n
c
e
d
t
e
t
i
o
s
i
z
e
-
l
u
s
i
o
n
c
r
o
m
a
t
o
g
r
p
h
S
C
)
t
r
a
c
e
s
o
f
N
t
.
M
e
T
S
r
h
O
m
e
i
n
e
l
a
l
G
F
n
(
P
-
p
e
t
o
h
a
s
g
i
g
e
o
,
d
n
g
S
s
O
A
S
)
T
O
o
1
d
i
m
1
0 µ
e
d
r
a
t
e
4
°
n
C
d
7
(
p
6
d
u
a
t
s
a
y
h
e
v
n
d
e
)
l
m
i
n
o
D
e
n
a
s
o
t
)
o
m
a
r
e
4
5
r
s
°
e
h
C
(
b
n
s
e
o
l
a
l
e
i
d
d
l
b
m
i
n
y
e
e
a
a
s
r
n
)
r
s
i
o
n
w
t
s
h
. e
a
e
p
T
r
h
r
e
e
d
s
e
p
d
n
e
e
c
a
v
e
k
o
h
f
e
o
1
0
h
s
0 µ
M
C
o
b
I
-
f
i
9
t
o
7
h
6
e
g
(
d
c
y
a
e
y
n
)
o
m
r
o
D
M
m
S
e
e
O
a
n
l
t
o
h
a
n
e
m
e
(
v
e
s
e
s
,
h
e
i
n
c
c
l
e
e
,
g
r
a
y
)
l
u
t
i
o
t
i
o
f
A
T
i
m
r
a
t
i
r
a
l
a
i
g
t
r
a
t
i
o
3
i
m
r
/
n
o
r
i
p
r
o
f
D
a
o
n
e
v
i
c
l
e
r
a
y
r
1
0
M
C
I
-
9
(
c
a
(
a
r
e
o
w
s
s
t
n
d
a
i
a
t
i
n
,
n
=
l
o
i
c
a
l
l
i
n
d
e
p
e
n
d
n
t
s
p
l
e
*
*
*
*
p
<
0
.
0
0
0
1
;
t
w
o
-
t
a
i
l
e
d
p
a
i
r
e
d
t
-
t
e
s
t
s
)
.
S
o
u
r
c
e
d
a
t
a
a
r
e
p
r
o
v
i
d
e
d
a
s
a
S
o
u
r
c
e
D
a
t
a
fi
l
e
.
nanodisc sample had markedly enhanced features and confirmed multiple rounds of 2D and 3D classification, two 3D classes with
the transmembrane region of hSOAT1 in nanodisc to be a dimer discernable transmembrane helix densities were isolated. Sub-
of dimer (Supplementary Fig. 4b). However, some 2D class sequent refinements generated reconstructions at resolutions of
averages in top views showed that one distinct dimer was adjacent 8.2 Å (for the oval-shaped tetramer) and 7.6 Å (for the rhombic-
to another blurry but still distinguishable dimer (Supplementary shaped tetramer) (Supplementary Fig. 4c, d). The oval-shaped
Fig. 4b), suggesting the dimer-dimer interface is mobile. Through structure occupies a 3D space of 170 Å×120 Å×50 Å (Fig. 2) and
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
|
(
2
0
2
0
)
1
1
:
2
4
7
8
|
h
t
t
p
s
:
/
/
d
o
i
.
o
r
g
/
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
2
0
-
1
6
2
8
8
-
4
|
w
w
w
.
n
a
t
u
r
e
.
c
o
m
/
n
a
t
u
r
e
c
o
m
m
u
n
i
c
a
t
i
o
n
s
3

A R T I C L E
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
|
h
t
t
p
s
:
/
/
d
o
i
.
o
r
g
/
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
2
0
-
1
6
2
8
8
-
4
a
b
Cytoplasmic
ER luminal
90°
50
Å
c
d
185 Å
Cytoplasmic
ER luminal
90°
110
Å
50
Å
Fig. 2 Cryo-EM maps of the human SOAT1 tetramer. a
C
r
y
e
A
o
,
-
r
T
E
e
1
i
M
d
e
t
a
o
n
s
e
e
n
i
t
y
.
m
D
a
e
h
p
p
n
o
o
f
i
n
t
e
i
h
s
n
s
e
o
a
o
f
c
e
v
a
e
r
i
l
-
s
M
h
a
S
n
p
P
m
e
d
a
o
h
S
l
l
d
O
A
T
s
o
a
1
t
e
t
n
e
r
r
a
w
r
a
n
.
m
o
E
e
d
a
e
r
i
c
i
n
s
t
o
p
e
u
o
v
c
n
f
i
e
o
t
h
w
a
n
d
o
b
d
i
s
n
i
d
g
d
h
e
v
y
i
e
w
s
d
w
.
h
m
S
T
s
e
O
w
e
a
A
o
m
s
T
s
u
b
u
n
i
t
s
i
n
o
n
e
h
S
O
A
T
1
d
im
e
r
a
r
e
c
o
lo
re
d
i
n
g
r
e
e
n
a
n
d
b
l
u
s
p
e
c
i
v
l
y
s
i
t
t
h
n
d
i
p
i
d
i
n
v
s
c
a
r
l
o
r
e
r
a
i
t
i
i
1
-
n
t
a
r
. c
a
n
s
T
p
h
a
e
r
e
d
n
e
c
n
y
s
. b
T
m
h
e
a
d
p
o
o
m
f
a
i
n
e
a
r
r
h
r
o
a
m
n
g
b
e
i
m
e
n
a
t
o
f
o
d
v
h
a
S
l
O
-
s
A
h
a
T
p
1
e
t
d
e
h
t
S
O
m
t
e
t
r
m
w
r
i
s
s
w
a
v
t
e
o
o
. d
d
e
i
n
t
m
p
i
i
a
h
m
s
u
b
t
i
r
i
s
c
m
l
o
r
e
t
a
h
e
a
h
i
t
y
t
h
c
-
sh
p
e
r
a
e
r
s
s
h
i
n
t
o
a
n
d
i
d
w
T
h
e
o
n
a
n
g
e
n
o
i
c
-
s
p
e
t
e
t
r
a
m
e
r
i
n
t
o
p
v
i
e
w
.
E
a
c
h
s
u
b
u
n
i
t
i
s
c
o
lo
re
d
i
n
t
h
e
s
a
m
e
w
a
y
a
s
i
n
c
.
the shape is similar to 3D Class 1 observed in detergent micelles previously reported that this loop is important for hSOAT1 activity
(Supplementary Fig. 2), and the rhombic-shaped structure and stability28,29. Indeed, the CT loop interacts with both the
occupies 185 Å×110 Å×50 Å (Fig. 2), similar to the 3D Class 2 luminal M7-M8 loop and the M5-M6 loop (Fig. 3d) at the ER
observed in detergent micelles (Supplementary Fig. 2). The N luminal side.
terminal tetramerization domain is invisible in both maps
The TM2-9 of hSOAT1 share a similar structural fold with H4-
probably due to its flexibility. These two structures reveal that H16 of DltB18, with core RMSD of 3.2 Å for 227 structurally
the dimer-dimer interfaces of oval-shaped and rhombic-shaped aligned residues, despite of their low-sequence identity (Fig. 3e, f),
hSOAT1 are distinct (Fig. 2b, d), correlating with the 3D suggesting a common evolutionary origin of the MBOAT family
heterogeneity observed in detergent micelles. This suggests the enzymes.
interfaces between dimers are unstable and dynamic in nature,
Dimer interface. In contrast to the monomeric DltB protein18,
the functional building block of hSOAT1 is a dimer, with a
packing interface area of 6,520 Å2. The two hSOAT1 protomers
interact through the M1, M6, M6-αD loop and M9 helices in a
symmetric way (Fig. 4a, b). In the inner leaflet of the ER mem-
brane, M144, A147, L148 and L151 on M1 of one subunit interact
with I370 and F378 on M6, and V501, W504 and F508 of M9 of
the other subunit (Fig. 4c, d). In the outer leaflet of ER mem-
brane, V158 and V159 on M1 interact with V363 on M6 of the
other subunit (Fig. 4e). The residues that form the dimer interface
are mostly hydrophobic and interact with each other in a shape-
complementary manner. To interrogate the function of the dimer
in hSOAT1, we mutated several residues on the dimer interface to
larger residues, including T140R, A147F, L151W and V159W, to
disrupt dimer formation. FSEC results showed theses mutants
have reduced dimers fractions while increased monomer fractions
to different extent, with A147F and T140R having the most
pronounced effect (Fig. 4f). We purified T140R and A147F
mutants in hSOAT1 dimer construct and found they only exhibit
little residual enzyme activities, at levels similar to that of the
catalytic dead H460N mutant (Fig. 4g and Supplementary Fig. 1i).
These results indicate the local dimeric architecture might be
important for SOAT1 activity, correlating with the observation
that disruption of dimer by Triton X-100 or octyl glucoside was
associated with loss of activity30.
which in turn hampered high-resolution structure determination.
To overcome the conformational heterogeneity in the tetrameric
hSOAT1 sample, the functional dimer construct (hSOAT1
66–550) was expressed and purified. The protein was recon-
stituted into the nanodisc in the presence of inhibitor CI-976
(Supplementary Fig. 3c, d). Subsequent 3D reconstruction
generated a 3.5 Å cryo-EM map, which allowed the model to be
built de novo (Fig. 3a, Supplementary Figs. 5 and 6, and
Supplementary Table 1).
Structure of hSOAT1 dimer. The hSOAT1 dimer has a sym-
metric rubber raft shape and occupies a 135 Å×70 Å×50 Å three-
dimensional space (Fig. 3a). Each hSOAT1 subunit is composed of
nine transmembrane helices (Fig. 3b), which is consistent with
previous prediction based on the biochemical data27. The first 52
amino acids (66–117) of the dimeric hSOAT1 were invisible in the
cryo-EM map, presumably due to their high flexibility (Fig. 3c, d).
Notably, this region has the low-sequence conservation between
SOATs from different species and paralogues. The cytosolic pre-αA
loop runs parallel to the membrane. The amphipathic αA floats on
the putative lipid bilayer with hydrophobic di-leucine motifs facing
the membrane and connects to M1 via a near 90° turn (Fig. 3d). M1
is linked to the long tilted M2-M4 tri-helix bundle by an ER-
luminal loop and a short helix αB (Fig. 3d). The six transmembrane
helices M4-9 form a funnel-shaped central cavity, which is capped
by the cytosolic helices αC, αD, and αE on the top. At the end of
M9, a C-terminal ER luminal loop (CT loop) is cross-linked by the Reaction chamber of hSOAT1. Many acyltransferases utilize
disulfide bond formed between C528 and C546 (Fig. 3d). It was histidine as the catalytic base31. Previous studies suggest that the
4
N
A
T
U
R
E
C
O
M
M
U
NI
C
A
TI
O
N
S
|
( 2 0 2 0 ) 1 1 : 2 47 8 | h t t p s : / / d o i . o r g / 1 0 . 1 0 3 8 / s 41 4 6 7 - 0 2 0 - 1 6 2 8 8 - 4 | w ww . n a t u r e . c o m / n a t u re c o m m u n i c a t i o n s

N
A
T
UR
E
C
O
M
M
UN
I
CA
T
I
O
NS
|
h
t
t
p
s
:
//d
o
i
.
o
r
g
/
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
2
0
-
1
6
2
8
8
-
4
A R T I C L E
a
b
Cytoplasmic
ER luminal
2
5
1
90°
4
3
6
7
9
8
c
d
Cytoplasmic
Cytoplasmic
αA
90°
1
2
3
4
s
s
ER luminal
Intracellular
ER luminal
e
f
DltC
90°
Extracecullar
DltB
Fig. 3 The structure of hSOAT1 dimer. a
C
r
y
o
-
c
E
o
e
s
M
m
a
p
o
g
f
h
S
y
O
w
e
A
T
h
1
d
i
m
m
i
o
e
-
f
r
t
t
e
i
n
n
e
c
s
s
i
p
d
a
e
r
v
i
e
c
e
h
w
a
n
T
d
o
n
a
h
t
p
e
s
S
o
p
v
h
y
v
i
w
i
e
w
o
s
.
f
T
t
w
o
s
u
o
e
s
b
s
u
s
b
b
n
i
t
e
e
s
c
o
f
t
n
h
o
n
S
e
f
d
t
t
A
i
h
a
m
e
e
t
r
r
a
r
s
e
e
m
l
r
c
e
o
m
l
o
r
b
e
r
n
w
d
a
u
i
n
m
c
n
e
b
o
w
g
d
e
r
o
r
e
m
s
a
e
n
a
n
T
n
a
h
h
d
b
l
ue
.
D
en
si
t
y
c
o
r
r
e
s
p
o
n
d
i
n
g
t
o
t
h
e
n
a
n
o
d
i
s
c
i
s
lo
r
e
d
i
n
r
a
i
t
s
e
r
a
e
n
y
.
b
i
e
h
e
c
r
-
s
t
i
o
a
n
a
i
t
e
e
t
s
o
h
e
u
h
a
c
p
p
r
u
r
a
b
o
l
u
x
m
i
m
o
a
d
i
t
e
s
e
l
e
o
v
f
g
e
h
r
l
i
S
y
n
O
.
d
A
C
i
c
T
a
1
9
t
e
d
7
d
b
e
m
y
r
o
t
a
l
h
r
e
e
e
c
d
s
a
h
e
s
o
h
e
d
l
i
n
s
i
n
a
.
w
e
T
h
a
e
i
d
n
t
i
t
i
i
e
s
h
t
r
a
n
s
m
m
o
o
b
w
n
r
a
e
l
ce
f
r
o
m
t
h
l
u
s
u
b
u
i
r
e
l
a
b
e
d
i
b
s
n
o
a
. c
n
s
tr
tu
l
s
i
m
w
n
s
i
n
i
d
e
a
v
s
i
e
n
d
t
o
p
v
. d
e
w
.
H
l
i
e
s
a
o
r
g
e
y
s
o
n
e
c
O
l
i
n
d
1
e
r
s
.
O
n
i
e
u
h
u
n
c
i
t
o
f
h
s
O
e
T
1
i
d
s
i
n
a
i
n
l
or
d
t
t
e
r
s
n
i
t
i
n
a
I
-
6
u
l
i
s
h
o
w
n
r
d
s
p
h
e
r
e
s
T
h
e
t
o
p
o
l
f
A
T
s
u
b
u
n
t
.
T
e
o
l
o
r
a
r
u
s
e
i
n
t
h
e
m
e
a
y
a
i
n
c
.
e
g
T
h
. f
e
c
T
r
y
h
s
e
t
a
s
l
u
s
p
t
e
r
u
c
o
t
u
s
r
e
i
o
f
D
l
t
D
B
-
D
B
l
t
(
C
g
c
o
y
m
)
p
o
l
n
e
t
x
o
i
n
o
s
i
d
h
e
S
v
O
i
e
A
w
T
a
1
n
p
d
r
t
o
o
p
m
v
e
i
e
r
w
(
o
P
D
r
B
I
)
D
.
:
6
B
U
G
)
.
T
h
e
D
l
t
B
s
u
b
u
n
i
t
i
s
i
n
r
a
i
n
b
o
w
c
o
l
o
r
w
h
i
l
e
t
h
e
D
l
t
C
s
ub
u
n
i
t
i
s
i
n
r
a
y
r
p
i
t
o
n
o
f
lt
r
a
n
e
o
t
(
c
l
o
e
d
conserved H460 on M7 is crucial for hSOAT1 activity and is the cysteines caused loss in hSOAT1 enzyme activity (by greater than
putative catalytic residue27. Our structures show that the side 90%) without severely lowering the cellular hSOAT1 protein
chain of H460 points towards the interior of a central cavity expression in transfected cells (Fig. 5g). Because the hydrophilic
cradled by M4−M9 (Fig. 5), suggesting this central cavity is the side chains of S414 and R418 on αD face the cytosolic side of
chamber where the acyl-transfer reaction takes place. In accor- SOAT1, S414 and R418 might play important roles for hSOAT1
dance with this, previous studies have also identified several activity, such as binding of highly hydrophilic CoA group of fatty
residues in the central cavity that are important for hSOAT1 acyl-CoA substrate. These results further emphasize the impor-
function. Mutations of residues on M7 and M8 affect the catalytic tant role of the cytosolic lid of the reaction chamber in the
activity32,33. C467 at the end of M7 is the major target site for p- enzymatic reaction.
chloromercuribenzene sulfonic acid-mediated SOAT1 inactiva-
tion34. Intriguingly, most residues aligning the interior of reaction
chamber are highly conserved (Supplementary Fig. 7b). These
data also indicate that the local environment in the central cavity
is important for the catalytic reaction.
Inhibitor and ligands binding sites. A strong extra non-protein
density was found in the central cavity and the size and shape of
the density matched that of the inhibitor CI-976, which was
included during cryo-EM sample preparation (Supplementary
Fig. 8a). By comparing the current map with the maps without
CI-976 (as described later) (Supplementary Fig. 8), we proposed
that this non-protein material represents the CI-976 molecule
(Fig. 5a–d). The large trimethoxyphenyl head of CI-976 is
sandwiched between the catalytic residue H460 on M7, and
residues N421 as well as W420 on the αD-αE loop (Fig. 5b), all of
which are crucial for the catalytic activity of hSOAT1. The head
of CI-976 also interact with Y416 and Y417 on αD (Fig. 5b). The
elongated dodecanamide tail of CI-976 extends in the cavity and
interacts with L377 on M6 and L515 on M9 (Fig. 5b). In accor-
dance with our assignment of CI-976, mutations of CI-976
interaction residues, N421A, H460A and H460N greatly dimin-
ished the enhancement effect of CI-976 on the thermostability of
The reaction chamber is covered by a lid formed by M4-αC
loop, αC, M6-αD loop, αD and αE on the cytosolic side (Fig. 5a).
Previous studies suggested that part of the cytosolic lid, residues
403–409 on M6-αD loop, may be involved in the binding of fatty
acyl-CoA35. To further explore the role of the residues in this
region, we performed alanine or cysteine mutagenesis of
conserved residues within this region (amino acids 406–422
from M6-αD loop to αD, Fig. 5f and (Supplementary Fig. 7), and
analyzed the effects of mutations after transient expressions of
each of these mutant plasmid DNAs in a CHO cell clone AC29,
that is devoid of endogenous SOAT activity, but regains enzyme
activity upon transient expression of hSOAT plasmid DNA36.
The results showed that mutating any of the following residues
W407, S414, Y416, Y417, R418, W420, and N421 to alanines or
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
|
(
2
0
2
0
)
1
1
:
2
4
7
8
|
h
t
t
p
s
:
/
/
d
o
i
.
o
r
g
/
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
2
0
-
1
6
2
8
8
-
4
|
w
w
w
.
n
a
t
u
r
e
.
c
o
m
/
n
a
t
u
r
e
c
o
m
m
u
n
i
c
a
t
i
o
n
s
5

A
R
T
I
C
L
E
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
|
h
t
t
p
s
:
/
/
d
o
i
.
o
r
g
/
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
2
0
-
1
6
2
8
8
-
4
a
b
M1
M6
M9
90°
cd
M9
M6
M1
e
M9
M6
M1
c
d
e
W408
N409
N409
W408
T140
M144
T140
M144
W504
W504
180°
F378
V501
F378
V158
V501
A147
A147
L148
F508
T505
F508
T505
V363
M6
V159
L148
L151
L151
I370
I370
M1
M1
hSOAT1-dimer-WT
hSOAT1-dimer-T140R
hSOAT1-dimer-A147F
hSOAT1-dimer-L151W
M6
M9
M9
M6
M1
f
g
0.8
0.5
0.4
0.2
0.0
Dimer
5000
2500
0
hSOAT1-dimer-V159W
Monomer
GFP
1.5
2.0
2.5
3.0
Elution volume (ml)
Fig. 4 The dimer interface of hSOAT1 dimer. a, b
T
o
p
b
r
n
v
o
i
e
n
g
f
o
w
a
n
C
A
m
d
l
s
i
d
e
0
t
e
v
p
r
n
i
e
v
t
. g
w
o
w
f
o
n
t
e
h
w
c
h
e
c
S
d
o
i
O
A
e
p
s
T
r
a
o
1
d
i
m
e
r
a
r
e
b
C
d
n
s
h
e
o
o
d
w
b
-
n
y
u
i
n
s
v
t
s
u
r
f
a
c
e
e
t
a
r
s
e
p
r
e
.
s
e
n
t
c
s
t
e
t
a
r
t
a
i
o
c
n
w
i
t
r
h
e
u
h
s
e
m
i
s
-
t
r
a
n
s
p
a
r
e
n
cy
.
M
1
,
M
6
a
n
d
M
9
h
e
l
ic
e
s
a
r
e
s
h
o
w
n
a
s
r
i
b
s
.
c
o
1
s
8
-
u
i
e
o
f
i
m
i
n
t
e
r
f
a
c
e
o
x
o
l
i
d
l
i
n
i
n b
I
n
t
i
n
g
s
i
d
u
e
w
e
b
a
r
e
y
s
s
d
h
a
o
w
e
n
d
i
n
s
e
t
s
i
c
i
k
s
.
R
e
s
F
i
S
d
E
u
C
e
s
t
m
u
e
t
a
s
t
o
e
d
f
i
h
n
f
w
e
1
r
e
d
s
h
o
e
w
n
i
n
o
i
r
a
n
e
.
d
°
o
s
u
a
t
e
d
v
e
i
e
a
m
r
f
a
e
d
t
o
c
.
e
l
s
e
p
u
n
i
e
w
t
o
m
f
h
e
i
n
t
e
r
a
u
i
n
g
r
B
e
s
i
d
-
e
s
b
o
x
e
e
d
s
h
l
i
n
n
b
.
f
r
a
c
S
O
A
T
i
m
r
w
i
t
h
t
e
r
r
a
c
e
u
t
a
i
o
T
h
t
i
v
t
i
e
h
a
S
r
O
e
A
s
T
1
i
m
e
m
r
m
a
n
s
s
e
s
u
r
e
d
i
n
g
N
D
c
o
l
e
s
t
r
o
l
s
u
b
s
t
r
a
t
e
,
i
n
t
h
e
p
r
e
s
e
n
c
e
o
f
c
h
o
l
e
s
t
e
r
o
l
.
S
o
u
r
c
e
d
a
t
a
a
e
p
v
i
d
e
d
a
s
a
S
o
r
c
e
D
a
t
a
fi
l
e
(
D
a
t
h
o
w
a
s
e
a
s
t
a
n
d
a
r
d
d
e
v
i
a
t
i
o
n
s
,
n
=
3
b
i
o
l
o
g
i
c
a
l
l
y
i
n
d
e
p
e
n
d
e
n
t
s
a
m
p
l
e
s
)
.
hSOAT1 dimer (Fig. 5e and Supplementary Fig. 9). The binding
There are several extra non-protein densities observed in the
position of CI-976, right in the catalytic center, suggests that it cryo-EM map. One density (density A) is close to the dimer
inhibits the enzyme by precluding the loading of substrate into interface. It is surrounded by L129, L132 and L133 on the
the catalytic center. This is consistent with the competitive hydrophobic side of αA, F145 on M1, C333 on M5, F382 on M6
behavior of CI-976 (ref.15). Moreover, it has been reported that and W408 on M6-αD loop (Supplementary Fig. 8c). The shape of
certain residues on M9 are responsible for the selectivity of this density is similar to cholesterol, suggesting this ligand might
subtype-specific SOAT inhibitors, such as pyripyropene A37. This be a tightly-bound sterol-like endogenous molecule or the head
further suggests that this large catalytic chamber can accom- group of the digitonin detergent that was carried on during
modate inhibitors with diverse chemical structures.
membrane protein extraction and purification. Previous work
6
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
|
(
2
0
2
0
)
1
1
:
2
4
7
8
|
h
t
t
p
s
:
/
/
d
o
i
.
o
r
g
/
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
2
0
-
1
6
2
8
8
-
4
|
w
w
w
.
n
a
t
u
r
e
.
c
o
m
/
n
a
t
u
r
e
c
o
m
m
u
n
i
c
a
t
i
o
n
s

N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
|
h
t
t
p
s
:
/
/
d
o
i
.
o
r
g
/
1
0
.
1
0
38
/
s
4
1
4
6
7
-
0
2
0
-
1
6
2
8
8
-
4
A R T I C L E
a
b
c
d
M4-αC loop
CI-976
CI-976
M6-αD loop
N421
Y417
M7
M4
M2
Y416
V424
M8
Cytoplasmic
M5
W420
L377
αE
αD
αC
90°
M3
M6
H460
CI-976
M9
CI-976
M6
M1
H460
L515
ER luminal
e
f
g
Cytoplasmic
2.5
2.0
1.5
1.0
0.5
0.0
p = 0.01
p = 0.0027
2.0
R418
N409
S414
D406
M6-αD loop
****
1.5
1.0
0.5
0.0
S412
αD
V422
W408
S410
Y417
Y416
W407
N421
W420
hSOAT1
Vinculin
50 kD
117 kD
h
i
j
Cytosolic lid
CI-976
90°
–59.681 kTe–1
59.681 kTe–1
Fig. 5 The catalytic chamber and the CI-976 binding site in hSOAT1. a
S
i
d
s
e
h
c
d
a
S
v
o
a
v
n
O
0
i
i
e
w
n
s
o
l
c
f
f
e
H
c
t
o
m
4
,
i
(
n
o
6
d
n
D
H
f
e
n
0
e
d
a
h
s
S
p
w
i
O
h
e
A
T
e
s
f
c
s
1
. b
s
u
C
w
6
b
u
n
s
i
w
i
t
e
i
-
b
s
n
u
l
c
p
a
v
r
t
o
w
e
o
n
o
h
r
f
e
e
C
s
p
r
9
e
s
7
t
e
6
h
n
b
e
e
m
e
t
a
t
i
d
o
i
m
o
o
t
s
w
n
n
e
s
.
g
T
s
h
e
e
t
s
r
.
u
a
R
r
e
k
3
s
r
n
e
s
s
m
e
u
e
n
g
l
m
b
r
t
a
h
a
e
t
a
n
a
p
n
e
t
s
t
h
e
t
e
l
e
r
i
c
e
c
f
s
t
u
a
w
r
e
c
h
r
o
C
s
l
o
I
a
r
-
e
9
d
7
e
i
6
n
n
t
h
e
e
s
a
o
m
w
e
n
w
a
Å
a
s
2
y
s
b
a
s
c
C
i
k
n
s
o
F
. c
i
g
E
f
.
3
le
r
c
c
v
,
t
i
d
.
T
n
h
d
e
e
a
i
n
n
h
t
n
i
y
. d
b
i
o
t
o
f
A
r
C
C
I
9
-
9
7
7
6
6
a
t
i
s
w
a
i
n
e
r
l
o
i
e
I
-
i
n
i
t
i
d
e
s
i
n
r
a
i
t
a
r
s
h
t
i
r
o
s
i
I
-
n
a
e
D
; p =
t
d
t
n
t
l
y
t
i
r
e
h
o
n
n
u
e
m
i
i
s
o
v
s
d
d
s
a
T
i
d
p
r
t
e
s
r
a
m
o
i
c
o
n
i
r
t
o
l
e
v
a
l
.
M
m
r
w
e
r
t
h
e
h
r
p
e
d
a
t
−
5
0
y
o
t
o
s
u
a
l
i
z
t
i
o
9
0
°
r
o
e
h
M
i
c
e
e
w
m
o
n
s
t
y
o
s
e
H
4
0
l
n
c
a
d
o
m
t
v
t
e
.
e
h
m
a
a
t
h
t
a
b
l
i
t
y
n
h
h
b
-
c
e
e
f
f
t
e
h
e
c
e
p
M
t
o
p
n
f
r
d
C
I
e
-
9
7
e
a
o
6
o
n
C
v
I
s
d α
a
r
i
7
*
D
o
6
*
u
s
w
h
i
<
S
O
t
A
h
0
T
e
0
1
p
d
e
i
a
o
a
m
k
r
e
r
h
N
s
m
u
h
t
t
a
n
t
s
.
o
T
i
0
h
n
e
t
t
h
0
s
h
e
2
.
e
r
p
m
o
s
r
s
e
H
u
t
n
a
c
4
e
b
e
6
s
i
l
i
o
t
y
f
e
n
e
x
w
a
a
l
c
u
a
t
e
b
y
d
n
i
i
n
g
a
d
p
d
n
e
d
r
/
n
a
e
m
e
p
e
a
e
i
h
t
a
i
o
i
i
n
n
e
s
n
c
o
f
-
9
t
h
e
i
g
r
a
t
i
r
e
a
l
o
0
n
e
f
t
a
a
r
h
s
1
o
p
e
w
e
n
a
v
m
s
e
m
e
a
t
r
i
e
d
d t
v
-
i
o
n
s
,
n
=
i
o
o
g
i
c
l
l
y
f
.
d
e
e
n
t
s
m
p
p
a
l
e
,
*
*p
w
0
.
0
1
f
4
2
1
A
; p =
n
.
0
7
R
f
o
0
A
.
m
0
1
0
o
r
46
r
0
S
N
O
r
e
t
i
l
y
,
a
t
w
-
i
t
a
i
r
l
e
e
r
e
t
t
e
s
)
.
f
C
l
l
o
o
e
e
u
p
v
i
e
w
e
o
n
t
h
e
6
-
α
D
l
o
n
i
t
h
s
i
d
e
c
h
i
n
s
h
o
w
i
n
s
t
i
c
k
e
s
i
d
t
h
a
a
r
e
p
o
r
t
a
n
t
o
A
T
z
y
e
c
t
i
t
y
o
t
d
i
n
g
r
e
c
o
d
i
n
g
r
e
g
E
n
z
y
m
a
t
i
c
a
c
t
i
v
i
t
i
e
s
o
f
v
a
r
i
o
u
s
h
S
O
A
T
1
m
u
t
a
n
t
s
.
(
F
o
r
D
4
0
6
A
,
n
=
4
.
F
o
r
W
4
0
7
A
,
n
=
4
.
F
o
r
W
4
0
8
A
,
n
=
5
.
F
o
r
N
4
0
9
A
,
n
=
8
.
F
o
r
S
4
1
0
C
,
n
=
6
.
F
o
r
S
4
d
r
1
e
2
p
S
o
A
e
4
r
,
n
=
3
t
.
s
F
a
o
m
0
r
p
.
S
4
s
7
e
a
v
1
.
.
r
t
i
4
)
O
e
C
D
t
p
,
n
=
6
.
F
s
o
h
n
s
c
u
r
o
t
Y
w
s
a
4
n
h
1
a
a
6
s
d
u
t
a
A
m
p
,
n
=
4
s
.
F
s
o
t
s
r
a
Y
n
4
1
7
A
e
0
,
n
=
4
.
A
.
n
F
o
r
R
o
e
4
1
a
p
v
8
A
,
n
=
4
a
.
i
F
o
d
l
h
r
W
4
s
h
2
t
S
T
0
w
O
1
A
,
n
=
3
d
u
.
t
t
F
o
o
c
t
r
a
s
N
4
u
n
e
2
1
t
A
e
,
n
=
4
v
n
.
F
o
u
s
s
r
e
V
4
F
c
2
o
u
t
2
r
A
,
n
=
5
, p =
h
s
b
i
o
l
0
o
.
b
c
g
9
e
e
i
4
c
a
3
w
a
l
4
l
y
.
.
i
n
n
d
e
n
l
e
a
t
a
a
r
e
u
e
a
n
d
a
r
d
r
r
o
r
s
.
t
w
-
t
i
l
e
d
u
n
p
r
e
t
t
e
a
s
u
s
e
l
c
l
a
t
h
e
p
a
l
s
.
S
4
1
0
C
F
o
1
2
A
,
p
=
0
1
h
e
r
m
t
a
v
a
l
u
e
s
l
e
s
t
h
a
n
T
e
.
0
0
0
1
a
a
T
d
y
h
s
v
e
x
r
e
s
s
i
o
n
l
e
v
e
s
S
n
o
f
A
m
c
T
1
m
a
n
a
d
s
i
e
n
d
o
g
e
o
p
6
u
v
e
i
n
n
l
i
n
n
e
a
r
e
s
o
w
n
l
o
T
h
e
i
g
i
n
a
l
b
l
o
t
s
w
r
o
v
i
d
e
d
a
S
a
l
o
S
o
r
c
e
D
a
t
a
fi
l
e
.
h
,
i
h
e
t
o
p
i
d
e
i
e
w
w
s
o
f
o
n
e
b
O
A
o
n
o
m
e
t
r
h
i
n
t
h
u
t
r
f
a
c
e
r
e
r
e
t
a
i
o
.
T
h
e
u
n
r
f
a
s
r
e
r
c
o
l
u
o
r
e
d
b
y
t
e
a
l
e
c
t
r
o
p
s
t
i
c
p
o
t
e
n
t
i
a
r
l
c
u
l
a
D
e
d
b
y
P
y
m
o
l
.
j
T
h
c
u
t
-
a
w
i
e
w
s
h
o
i
n
g
t
h
e
i
d
i
n
g
p
o
k
e
t
o
f
e
i
n
h
i
b
o
r
C
I
-
9
7
i
n
s
i
d
e
t
h
r
e
a
c
t
i
o
c
h
a
m
b
e
.
S
o
r
c
e
d
a
a
r
e
r
o
d
e
d
a
s
a
c
e
t
a
fi
l
e
.
showed that SOAT1 exhibits only low affinity binding towards sample preparation procedure. Another elongated density
cholesterol, either as substrate or as activator, with dissociation (density B) is inside the central cavity and surrounded by F258
constant at sub-millimolar concentration23. Therefore, we spec- and R262 on M4, F384 and W388 on M6, P304 on αC-M5 linker
ulate that the molecule in density A may be neither a cholesterol and V424 on αE (Supplementary Fig. 8e). One additional density
substrate nor a cholesterol activator, but a sterol-like molecule, (density C) is on the ER luminal side of the central cavity and
such as the head group of digitonin, bound to the enzyme in the surrounded by Y176 on αB, S519, W522 and Y523 on M9, L468
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
|
(
2
0
2
0
)
1
1
:
2
4
7
8
|
h
t
t
p
s
:
/
/
d
o
i
.
o
r
g
/
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
2
0
-
1
6
2
8
8
-
4
|
w
w
w
.
n
a
t
u
r
e
.
c
o
m
/
n
a
t
u
r
e
c
o
m
m
u
n
i
c
a
t
i
o
n
s
7

A
R
T
I
C
L
E
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
|
h
t
t
p
s
:
/
/
d
o
i
.
o
r
g
/
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
2
0
-
1
6
2
8
8
-
4
on M7-M8 linker, P250 on M3−M4 linker (Supplementary induce a conformational change to open the reaction chamber
Fig. 8g). The exact identities of these densities A-C and their roles and cholesterol binding at the allosteric activator site would
on the hSOAT1 function remain elusive (Supplementary Fig. 8i). induce a further conformational change for the sterol-dependent
In order to gain more mechanistic insights into the catalytic fully activation of hSOAT1 (ref.4) (Supplementary Fig. 13).
mechanism of hSOAT1 and to trap the catalytic reaction in a Notably, a hSOAT1 tetramer structure in complex with oleoyl-
transition state, we designed and synthesized a compound that CoA was solved by another group and reported on BioRxiv
might mimics the catalytic transition intermediate. Inspired by recently40 and there is indeed a conformational change of the
the previous work on GOAT38, we hypothesized that the catalytic reaction chamber to accommodate oleoyl-CoA substrate.
reaction intermediate of hSOAT1 might be a ternary complex of
The structures of human SOAT1 presented here provide a
sterol, acyl-CoA and the enzyme. Therefore, the covalent linkage high-resolution glimpse of the architecture and domain organi-
of sterol and acyl-CoA would yield a competitive inhibitor with a zation of this important enzyme and shed light on the structure of
higher affinity than each individual substrate alone. Pregnenolone other closely related MBOAT family proteins, such as DGAT1.
was previously reported to be a substrate of hSOAT1 with a lower This work not only paves the way towards a better mechanistic
Km and better solubility than cholesterol39. In the current study, understanding of SOAT1-catalyzed reaction, but also provides a
CoA group was chemically covalently linked with stearoyl- template for structure-based inhibitor design to target several
pregnenolone to generate a bi-substrate analog for hSOAT1 human diseases. Our current work shows that SOAT1 has 9
(BiSAS) (Supplementary Fig. 10a). Indeed, BiSAS inhibited the transmembrane domains (TMDs), with the active site His 460
purified hSOAT1 enzyme in the in vitro NBD-cholesterol based located at the luminal half of M7. This is consistent with previous
assay (Supplementary Fig. 10b). A cryo-EM sample of hSOAT1 biochemical work27, and supports the idea that the cholesteryl
dimer was prepared in the presence of BiSAS and cholesterol esters produced by SOAT1 may first appear within the ER
(Supplementary Fig. 3e, f) and the cryo-EM reconstruction membrane before they become part of lipid droplets in the
generated a 3.5 Å map (Supplementary Figs. 11,12 and Supple- cytosol. Unexpectedly, though SOAT1 and SOAT2 shares
mentary Table 1). This map was compared with the CI-976 extensive sequence homology, different groups reported that
bound map and was found to be similar overall, with a real space SOAT2 may contain only 2 or 5 TMDs, as reviewed in41. The
correlation of 0.9. We anticipated that BiSAS might mimic both uncertainty in the transmembrane topology of ACAT2 can be
substrates of hSOAT1 and might occupy the substrate-binding settled by further structural studies.
pocket while cholesterol might only bind at the activating site. In
contrast to our prediction, in the BiSAS map, the strong
continuous density of CI-976 found in the central cavity of the
CI-976 bound map was replaced by weak residual densities that
were not continuous (Supplementary Fig. 8b), indicating the
absence or low occupancy of BiSAS molecule, probably due to the
low affinity or incompatibility of BiSAS in the nanodisc sample
preparation conditions. In addition, we did not observe any extra
cholesterol-like density that would suggest the presence of
activating cholesterol either, probably due to the low affinity of
activating cholesterol on hSOAT1 (ref.23). Instead, the BiSAS map
most likely represents the apo resting state of hSOAT1.
Interestingly, all three additional non-protein densities (density
A−C) present in the CI-976 map were also observed in the BiSAS
map (Supplementary Fig. 8d, f, h, j). Therefore, BiSAS map serves
as a reference to validate the identity of CI-976 previously
observed in the catalytic center and also suggests the tight
associations of non-protein ligands A−C with the hSOAT1
protein and likely their functional importance as well.
Methods
Cell culture. Sf9 insect cells (Thermo Fisher Scientific #12659017) were cultured
in Sf-900 III serum-free medium (SFM; Thermo Fisher Scientific) or in SIM SF
(Sino Biological) at 27 °C. HEK293F (Thermo Fisher Scientific #R79007) cells
were cultured at 37 °C with 6% CO2 and 70% humidity in Free Style 293
medium (Thermo Fisher Scientific) supplemented with 1% fetal bovine serum
(FBS). HEK293S cell line was kindly generated and provided by Professor Gobind
Khorana at MIT42. The CHO mutant cell line AC29, deficient in ACAT1/SOAT1
activity, was isolated in Chang Lab43
.
Constructs. We cloned human SOAT1 (Hs_SOAT1), human SOAT2
(Hs_SOAT2), Xenopus laevis SOAT1 (Xl_SOAT1), chicken SOAT1 (Gg_SOAT1)
and zebrafish SOAT1 (Dr_SOAT1) into NGFP tagged BacMaM vector for
screening by FSEC method44 (Please see Supplementary Table 2 for primer list).
The screening procedures identified human SOAT1 as a putative target with rea-
sonable expression level and elution profile. cDNAs of human SOAT1 full-length
(1–550), SOAT1-dimer (66–550) were cloned into a modified BacMaM vector,
with N-terminal His7-strep-GFP tags45
.
SOAT1 activity assays using 3H-oleyl-CoA in vitro. The assay was conducted as
described in22, by using solubilized cell extracts as the enzyme source and pre-
formed taurocholate/cholesterol/phosphatidylcholine mixture and tritium-labeled
oleoyl-CoA as the substrates. This assay provides large amounts of exogenous
phospholipids and cholesterol at optimized concentrations, such that the mem-
brane phospholipids and cholesterol endogenously associated with SOAT1 become
significantly diluted and can no longer affect the enzyme activity in vitro. Cells
were harvested by spinning down at 800 × g for 5 min at room temperature. After
two washes with PBS, cell pellets were solubilized in 2.5% CHAPS, 1 M KCl in
50 mM KH2PO4 at pH 7.4 with protease inhibitors. 5 µl of cell homogenates at
protein concentration of 2–5 mg/ml were diluted into 50 µl of preformed 9.3 mM
taurocholate/1.6 mM cholesterol/11.2 mM phosphatidylcholine mixture and incu-
bated on ice for 10 min. The reactions were initiated by adding 10 nmole of
Discussion
SOAT catalyzes the esterification reaction between acyl-CoA and
cholesterol. The surface representation of the hSOAT1 monomer
shows an intra-membrane tunnel from outside of the molecule
into the reaction chamber. The tunnel is located between M4 and
M5, and is mainly hydrophobic. This lateral tunnel within the
transmembrane domain of hSOAT1 might be the substrate or
product transfer pathway for hydrophobic molecules. The other
substrate, fatty acyl-CoA, is amphipathic with a hydrophobic tail
and a highly hydrophilic CoA group. The cytosolic acyl-CoA can
access the central reaction chamber only from the cytosolic side
of hSOAT1. However, the surface representation of the hSOAT1
dimer shows that the reaction chamber is completely shielded
from the cytosolic side by the two short αD-αE helices and
associated intracellular loops (Fig. 5h–j). This suggests that the
current structure represents a substrate-unbound resting state
with relatively low catalytic activity, in which the putative cata-
lytic residue H460 is less accessible to the acyl-CoA substrate
3H-oleoyl-CoA. The reactions were continued at 37 °C for 10 min (ref.23).
Thermostability assay. The hSOAT1 dimer mutants were generated by Quick
Change methods (Please see Supplementary Table 2 for primer list). hSOAT1
dimer constructs were transfected into HEK293F cells (grown in FreeStyle 293
medium + 1% FBS, 37 °C) and the cells were harvested 48 h post transfection. The
cells were washed with TBS buffer (20 mM Tris pH 8.0 at 4 °C, 150 mM NaCl) and
incubated with 0.9% DMSO (vehicle group) or 0.9% DMSO together with 100 μM
CI-976 (CI-976 group). Cells were solubilized by TBS with 1% GDN and protein
inhibitors (2 µg ml−1 leupeptin, 2 µg ml−1 pepstatin A, 2 µg ml−1 aprotinin and
1 mM PMSF) at 4 °C for 30 min. Cell debris was removed by centrifugation at
13,000 × g for 10 min and centrifuged at 60,000 × g for 30 min. Each supernatant
was split into two samples. One sample was kept at 4 °C as unheated sample and
(Supplementary Fig. 13). Therefore, acyl-CoA binding would the other one was heated at 45 °C for 10 min using a thermo cycler. After heating,
8
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
|
(
2
0
2
0
)
1
1
:
2
4
7
8
|
h
t
t
p
s
:
/
/
d
o
i
.
o
r
g
/
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
2
0
-
1
6
2
8
8
-
4
|
w
w
w
.
n
a
t
u
r
e
.
c
o
m
/
n
a
t
u
r
e
c
o
m
m
u
n
i
c
a
t
i
o
n
s

N
A
T
UR
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
|
h
t
t
p
s
:
/
/
d
o
i
.
o
r
g
/
10
.
1
0
3
8
/
s
4
1
4
6
7
-
0
2
0
-
1
6
2
8
8
-
4
A R T I C L E
each sample was centrifuged at 60,000 × g for additional 30 min to remove pre-
β-cyclodextrin (Sigma, HZB1102) was added and the mixture was incubated at 37 °
C for 2 min. To start the enzymatic reaction, 1 μl 2.5 mM oleoyl-CoA (Sigma,
O1012) was added and the reaction mixture was incubated at 37 °C for 15 min. The
reaction was terminated by adding 2:1 chloroform/methanol, the extract was
separated on an HPLC column at 0.2 ml/min (Agilent, Poroshell HPH-C18,
2.7 μm) running in 100% ethanol and detected via fluorescence detector on an
HPLC (SHIMADZU). NBD-cholesterol eluted at 1.3 min and its ester eluted at
1.9 min. The peak areas of the NBD-cholesteryl ester products and remaining
NBD-cholesterol were integrated separately to obtain the relative ratio of NBD-
cholesterol that was converted into NBD-cholesteryl esters.
cipitates and the supernatants were analyzed on a Superose 6 increase 5/150 (GE
Healthcare) column running in 20 mM Tris pH 8.0 at 4 °C, 150 mM NaCl and
0.5 mM DDM. The fluorescence peak height at the hSOAT1 dimer elution position
was divided by the peak height at the monomer elution position to generate the
ratio which was used to quantify the thermostability. The enhancement effect of
CI-976 on hSOAT1 dimer thermostability was quantified by dividing the ratio in
the presence of CI-976 by the ratio in the presence of DMSO alone.
Protein expression and purification. The BacMam expression system was used
for large-scale expression of human SOAT1. The BacMaM viruses were added into
suspended HEK293F cells (grown in FreeStyle 293 medium + 1% FBS, 37 °C).
Sodium butyrate (10 mM) was added to the culture 12 h post-infection to promote
protein expression and the temperature was lowered to 30 °C. Cells were harvested
60 h post-infection and washed with TBS buffer. Cell pellets were frozen at −80 °C
for later use.
Nanodisc preparation. MSP2X was constructed by linking two MSP1E3D1 by
PCR overlap extension, with Gly-Thr as the linker. The MSP2X gene was con-
structed into a pET vector, with N-terminal His6 tag and HRV 3C site. The MSP2X
and MSP2N2 proteins were purified by Talon resin48. The eluted SOAT1 protein in
TBS buffer with 0.1% digitonin and 10 mM desthiobiotin from strep-tactin column
was concentrated by a 100-kDa cut-off ultrafiltration device (Millipore) and
exchanged into buffer without desthiobiotin. The SOAT1 protein was mixed with
soybean polar lipids extract (SPLE, Avanti) and purified MSP (MSP2N2 for SOAT1
tetramer, MSP2X for SOAT1 dimer) at a molar ratio of SOAT1: MSP: SPLE =
1:7:100. For the hSOAT1 dimer in nanodisc that contained cholesterol, SPLE and
cholesterol (at 4:1 ratio) were supplemented in GDN detergent micelles. The
SOAT1 protein was mixed with micelles and purified MSP2X at a molar ratio of
SOAT1: MSP2X: SPLE = 1:4:100. After incubating at 4 °C for 30 min, Bio-beads
SM2 (Bio-Rad) were added and rotated at 4 °C for 1 h to initiate the reconstitution.
Another batch of fresh bio-beads was added and rotated at 4 °C overnight. The next
day, the Bio-beads were removed and the mixture was loaded into a streptactin
column to remove the empty nanodisc. The eluted hSOAT1 in the nanodisc was
concentrated and cleaved by Prescission protease to remove GFP tags. The
nanodisc was centrifuged at 60,000 × g for 30 min, and then loaded onto a
superose-6 increase column running in TBS buffer containing 0.5 mM TCEP. The
collected fractions were detected by SDS−PAGE and peak fractions were con-
centrated to A280 = 1.2. Because the hSOAT1 tended to aggregate after nanodisc
reconstitution, fractions of each peak were combined and concentrated for cryo-
EM grids preparations and screening and only fractions containing high ratio of
SOAT1 dimer were used for data collection.
The cell pellets were resuspended in TBS buffer supplemented with protease
inhibitors (2 µg ml−1 leupeptin, 2 µg ml−1 pepstatin A, 2 µg ml−1 aprotinin and
1 mM PMSF). Unless stated otherwise, all buffers used for purification were
supplemented with inhibitor either 1 μM inhibitors CI-976 or BiSAS. The cells
were broken by sonication and centrifuged at 5000 × g for 10 min with JA25.5 rotor
(Beckman) to remove cell debris. The supernatant was centrifuged at 60,000 × g for
1 h in Ti45 rotor (Beckman) to harvest cell membrane in pellets. The membrane
pellets were homogenized in TBS, solubilized by 1% digitonin for 2 h at 4 °C, and
centrifuged at 60,000 × g for 1 h. The supernatant was loaded onto a 1 ml
prepacked strep-tactin superflow high-capacity column (IBA) and washed using
TBS buffer with 0.1% digitonin. The binding protein was eluted using TBS buffer
with 0.1% digitonin and 10 mM desthiobiotin. The eluted proteins were digested
with PreScission protease to remove tags and further purified by superose-6
increase column (GE Healthcare) in TBS buffer with 0.1% digitonin or 40 μM
GDN. Buffers were not supplemented with inhibitors during the purification of
proteins used for the activity assay. Purified protein were either directly used for
experiments or flash frozen in liquid nitrogen for storage at −80 °C. Stored protein
was thawed on ice and centrifuged to remove precipitates before further
experiments. The original gels for protein purification were provided as a
Source Data file.
Cryo-EM data collection. The nanodisc samples were loaded onto glow-
discharged GiG R1/1 holey carbon gold grids (Lantuo) and plunged into liquid
ethane by Vitrobot Mark IV (Thermo Fisher Scientific). Cryo-grids were screened
by Talos Arctica electron microscope (Thermo Fisher Scientific) operated at the
voltage of 200 kV using a Ceta 16 M camera (Thermo Fisher Scientific). Optimal
grids were transferred to Titan Krios electron microscope (Thermo Fisher Scien-
tific) operated at the voltage of 300 kV, with an energy filter set to a slit width of
20 eV. Super-resolution movies (50 frames per movie) were collected with a dose
SOAT1 activity assays using 3H-oleate in intact cells. This assay is designed to
measure the rate of 3H-cholesteryl oleate biosynthesis in intact cells (i.e., without
having to break up the cells). Mutant CHO AC29 cells that lack endogenous SOAT
activity were cultured in 6-well plates at 37 °C and transiently transfected with
DNAs from various single amino acid substitutions of hSOAT1 as indicated.
Unless stated otherwise, each construct contained the 6x His tag at the N-terminus.
We used hSOAT1 that contained the C92A substitution as the wild type hSOAT1.
Control experiments showed that the SOAT activity of the C92A mutant remained
the same as the wild type27 but the C92A substitution greatly diminished the
SOAT1 protein aggregation that occurred in vitro during the SDS-PAGE process.
At the third day after transfections, the cells were split into three equal parts by
trypsinization. One part was used to monitor the hSOAT protein expression by
western blot analysis, using the SOAT1 specific antibodies DM10 as the probe, as
described46, with the intensity of the WT hSOAT protein set as 1.0. The other two
parts of the transfected cells were used to monitor the SOAT enzyme activity, by
incubating the intact cells to 20 μl of 3H-oleate/fatty acid free BSA (7.5 × 106 dpm
μl−1) for 20 min. The amount of 3H-cholesteryl oleate produced was determined
by the procedure previously described27,36. Briefly, after 3H-oleate/BSA pulse, cells
were rinsed with PBS, harvested in 1 mL per well of 0.1 M NaOH, incubated at RT
for 30 min. The cell homogenates were transferred to13 × 100 mm size glass tubes;
then neutralized by adding 67 μl of 3 M HCl, and buffered with 50 μl of 1 M
K2HPO4. 80 μg per tube of non-radiolabeled cholesteryl oleate was added as carrier
for identification purposes. Cellular lipids in each tube were extracted with 3 mL of
CHCl3:MeOH at 2:1 followed by adding 1 mL of H2O. The bottom chloroform
phase that contained the lipid samples were dried under N2, then redissolved in
80 μl of ethyl acetate and spotted onto TLC plates (Anatech), with petroleum ether:
ether:acetic acid (90:10:1) as the solvent system. After TLC, the plates were air-
dried, the cholesteryl oleate bands (at Rf 0.9) were identified by briefly staining the
TLC plates with iodine vapor. The cholesteryl oleate bands were scraped into
scintillation vials and counted in a scintillation counter after addition of 3 mL per
vial of Ecoscint O. The enzyme activity of each mutant hSOAT1 was estimated
relative to that of WT hSOAT1, with the protein content of each mutant hSOAT1
normalized with that of the WT hSOAT1.
rate of 5.4 e−pixel−1 −1
using K2 Summit direct electron camera (Thermo Fisher
s
Scientific) at a nominal magnification of 130,000 ×, equivalently to a calibrated
super-resolution pixel size of 0.5225 Å, and with defocus ranging from −1.3 μm to
−2.3 μm. All data acquisition was performed automatically using SerialEM49
.
Cryo-EM image processing. For the CI-976 complex, super-resolution movie
stacks were motion-corrected, dose-weighted and 2-fold binned by MotionCor2
1.1.0 using 9 × 9 patches50. Micrographs with ice or ethane contamination were
manually removed. Contrast transfer function (CTF) parameters were estimated
using Gctf v1.06 (ref.51). Particles were picked by Gautomatch (developed by Kai
Zhang) and subjected to reference-free 2D classification. Unless otherwise stated,
all classification and reconstruction were performed with Relion 2.0 (ref.52). Initial
model was generated by cryoSPARC53 using the selected particles from 2D clas-
sification. The selected particles were further subjected to 3D classification using
C1 symmetry. The particles selected from good 3D classes were re-centered and
their local CTF parameters were determined using Gctf v1.06. These particles were
further refined by cisTEM54 using C2 symmetry imposed. The resolution estima-
tion was based on the Part.FSC curve in cisTEM at FSC = 0.143 cut-off. Local
resolution estimation for hSOAT1 dimer and CI-976 complex was calculated using
blocres55. For the apo state, images were processed in the same way, except the
finial refinement were done using Relion 3.0 (ref.56) with a soft mask that excluded
the MSP and lipids. The resolution estimations of the apo state were based on the
gold standard FSC of 0.143 cut-off after correction of the masking effect57. Local
resolution estimation for hSOAT1 dimer in the apo state was calculated using
Resmap58
.
SOAT1 activity assay using NBD-cholesterol. 2.8 mM cholesterol/11.2 mM PC/
18.6 mM taurocholate were mixed to generate the cholesterol donor47. The tetra-
meric or dimeric hSOAT1 enzyme was prepared in GDN detergent. First, 10 μl 2 M
KCl, 5 μl 5% BSA, 1 μl SOAT1 protein (A280 = 0.5), 5 μl micelles, 40 μl TBS buffer
with 40 μM GDN were mixed with TBS buffer with 0.5% CHAPS to reach the
volume of 100 μl and incubated at 37 °C for 2 min. In order to measure the IC50 of
inhibitors, different concentrations of given inhibitors were added, as indicated.
Then, 1.0 μl 0.2 mg ml−1 NBD-cholesterol (Sigma, N2161) solubilized in 35%
Model building and refinement. The sharpened map in the presence of CI-976
from cisTEM was converted into mtz file by Phenix59. The model was manually
built de novo in Coot60. The assignments of transmembrane domain helices were
based on their connectivity aided by the less-sharpened map. The register
assignment and modeling building were based on the features of large aromatic
side chains and partly aided by the further sharpened map. The manually built
model was refined by Phenix59. The sterol-like ligand in non-protein density A was
built as cholesterol for visualization. The model in the presence of CI-976 were
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
ON
S
|
(
2
0
2
0
)
1
1
:
2
4
7
8
|
h
t
t
p
s
:
/
/
d
o
i
.
o
r
g
/
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
2
0
-
1
6
2
8
8
-
4
|
w
w
w
.
n
a
t
u
r
e
.
c
o
m
/
n
a
t
u
r
e
c
o
m
m
u
n
i
c
a
t
i
o
n
s
9

A
R
T
I
C
L
E
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
|
h
t
t
p
s
:
/
/
d
o
i
.
o
r
g
/
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
2
0
-
1
6
2
8
8
-
4
fitted into the map in the presence of BiSAS by Chimera61 and further refined by
Phenix.
5. Shibuya, Y., Chang, C. C. & Chang, T. Y. ACAT1/SOAT1 as a therapeutic
target for Alzheimer’s disease. Future Med. Chem. 7, 2451–2467 (2015).
6. Rong, J. X. et al. ACAT inhibition reduces the progression of preexisting,
advanced atherosclerotic mouse lesions without plaque or systemic toxicity.
Arterioscler. Thromb. Vasc. Biol. 33, 4–12 (2013).
7. Geng, F. et al. Inhibition of SOAT1 suppresses glioblastoma growth via
blocking SREBP-1-mediated lipogenesis. Clin. Cancer Res. 22, 5337–5348
(2016).
8. Yang, W. et al. Potentiating the antitumour response of CD8(+) T cells by
modulating cholesterol metabolism. Nature 531, 651–655 (2016).
9. Yue, S. et al. Cholesteryl ester accumulation induced by PTEN loss and PI3K/
AKT activation underlies human prostate cancer aggressiveness. Cell Metab.
19, 393–406 (2014).
ConSurf calculation. The SOAT1 sequences (TF105767) were download from
Treefam62, aligned by Clustal Omega and uploaded into the Consuf server63
together with the CI-976 pdb files. The figures were prepared by UCSF Chimera61
.
Chemical synthesis of BiSAS. To the solution of α-bromo stearic acid (727 mg,
2 mmol, 1.0 equiv), pregnenolone (632 mg, 2 mmol, 1.0 equiv) and dicyclohex-
ylcarbodiimide (DCC, 495 mg, 2.4 mmol, 1.2 equiv) in dichloromethane (DCM,
30 mL) was added 4-dimethylaminopyridine (DMAP, 293 mg, 2.4 mmol, 1.2
equiv). The solution was stirred at room temperature for 24 h. The crude product
was purified by column chromatography (Hexanes: Ethyl acetate = 10:1) to obtain
the α-bromo ester (747 mg, 57 % yield) as a white solid. 1H NMR (400 MHz,
CDCl3) δ 5.39 (d, J = 5.1 Hz, 1H), 4.67 (qd, J = 11.3, 9.5, 4.2 Hz, 1H), 4.17 (t, J =
7.4 Hz, 1H), 2.54 (t, J = 8.9 Hz, 1H), 2.40–2.32 (m, 2H), 2.12 (s, 4H), 2.08–1.84
(m, 6H), 1.75–1.58 (m, 4H), 1.56–1.39 (m, H), 1.37–1.11 (m, 32H), 1.03 (s, 3H),
0.88 (t, J = 6.7 Hz, 3H), 0.63 (s, 3H). 13C NMR (101 MHz, CDCl3) δ 209.69, 169.49,
139.39, 122.83, 77.48, 77.16, 76.84, 75.51, 63.81, 56.96, 49.99, 46.72, 44.12, 38.92,
37.69, 37.04, 36.74, 35.04, 32.08, 31.94, 31.91, 31.71, 29.84, 29.83, 29.81, 29.73,
29.61, 29.52, 29.45, 28.97, 27.61, 27.40, 24.62, 22.97, 22.84, 21.18, 19.46, 14.28,
10. Jiang, Y. et al. Proteomics identifies new therapeutic targets of early-stage
hepatocellular carcinoma. Nature 567, 257–261 (2019).
11. Hofmann, K. A superfamily of membrane-bound O-acyltransferases with
implications for wnt signaling. Trends Biochem. Sci. 25, 111–112 (2000).
12. Chang, C. C. et al. Immunological quantitation and localization of ACAT-1
and ACAT-2 in human liver and small intestine. J. Biol. Chem. 275,
28083–28092 (2000).
13. Anderson, R. A. et al. Identification of a form of acyl-CoA:cholesterol
acyltransferase specific to liver and intestine in nonhuman primates. J. Biol.
Chem. 273, 26747–26754 (1998).
+
13.37. HRMS(ESI): m/z calcd for C39H66BrO3 [M + H]+: 661.418985, found
661.420600
14. Bocan, T. M., Mueller, S. B., Uhlendorf, P. D., Newton, R. S. & Krause, B. R.
Comparison of CI-976, an ACAT inhibitor, and selected lipid-lowering agents
for antiatherosclerotic activity in iliac-femoral and thoracic aortic lesions. A
biochemical, morphological, and morphometric evaluation. Arterioscler.
Thromb. 11, 1830–1843 (1991).
15. Field, F. J., Albright, E. & Mathur, S. Inhibition of acylcoenzyme A: cholesterol
acyltransferase activity by PD128O42: effect on cholesterol metabolism and
secretion in CaCo-2 cells. Lipids 26, 1–8 (1991).
16. Roth, B. D. et al. Inhibitors of acyl-CoA:cholesterol acyltransferase. 1.
Identification and structure-activity relationships of a novel series of fatty acid
anilide hypocholesterolemic agents. J. Med. Chem. 35, 1609–1617 (1992).
17. Chang, C. C. Y., Sun, J. & Chang, T.-Y. Membrane-bound O-acyltransferases
(MBOATs). Front. Biol. 6, 177 (2011).
To the solution of α-bromo ester obtained above (27 mg, 40 μmol, 1.0 equiv) and
CoA-SH (62 mg, 80 μmol, 2.0 equiv) in N,N-dimethyllformamide (DMF, 1 mL) was
added triethylamine (TEA, 56 μL, 0.4 mmol, 10 equiv). The solution was stirred
under nitrogen atmosphere at 35 °C overnight. The crude product was purified by
reverse phase HPLC (Water: Acetonitrile = 50: 50 to 5: 95) to obtain BiSAS in bis
(triethylammonium) salt form (23.8 mg) as a colorless solid. 1H NMR (500 MHz,
D2O) δ 8.44 (s, 1H), 8.06 (s, 1H), 6.04 (s, 1H), 5.27 (s, 1H), 4.89–3.72 (m, 9H),
3.65–3.20 (m, 6H), 3.10 (q, J = 7.2 Hz, 8H), 2.89–1.73 (m, 19H), 1.42–1.00 (m,
48H), 0.88 (s, 3H), 0.77 (s, 6H), 0.72–0.60 (m, 3H), 0.50 (s, 3H). HRMS(ESI): m/z
calcd for C60H99N7O19P3S− [M - H]−: 1346.593480, found 1346.591330.
Quantification and statistical analysis. Global resolution estimations of cryo-EM
density maps are based on the 0.143 Fourier Shell Correlation criterion64. Fluor-
escence values were plotted versus the log of the concentration of inhibitor, and
GraphPad Prism 6 was used to generate a curve fit with dose-response inhibition
equation: Y = 100/1 + 10[Log(IC50−X) *HillSlope]. IC50 values were calculated from
the curve fit using GraphPad Prism software. The number of biological replicates
(N) and the relevant statistical parameters for each experiment (such as mean or
standard error) are described in the figure legends. No statistical methods were
used to pre-determine sample sizes.
18. Ma, D. et al. Crystal structure of a membrane-bound O-acyltransferase.
Nature 562, 286–290 (2018).
19. She, J. et al. Structural insights into the voltage and phospholipid activation of
the mammalian TPC1 channel. Nature 556, 130–134 (2018).
20. Yu, C. et al. Human acyl-CoA:cholesterol acyltransferase-1 is a
homotetrameric enzyme in intact cells and in vitro. J. Biol. Chem. 274,
36139–36145 (1999).
21. Yu, C. et al. Role of the N-terminal hydrophilic domain of acyl-coenzyme A:
cholesterol acyltransferase 1 on the enzyme’s quaternary structure and
catalytic efficiency. Biochemistry 41, 3762–3769 (2002).
22. Chang, C. C. et al. Recombinant acyl-CoA:cholesterol acyltransferase-1
(ACAT-1) purified to essential homogeneity utilizes cholesterol in mixed
micelles or in vesicles in a highly cooperative manner. J. Biol. Chem. 273,
35132–35141 (1998).
23. Chang, C. C. et al. Purification of recombinant acyl-coenzyme A:cholesterol
acyltransferase 1 (ACAT1) from H293 cells and binding studies between the
enzyme and substrates using difference intrinsic fluorescence spectroscopy.
Biochemistry 49, 9957–9963 (2010).
Reporting summary. Further information on research design is available in
the Nature Research Reporting Summary linked to this article.
Data availability
Data supporting the findings of this manuscript are available from the corresponding
author upon reasonable request. A reporting summary for this Article is available as a
Supplementary Information file.
The source data underlying Figs. 1b, c, e, 4g, 5e, g, and Supplementary 1b, f, h, 3b, d, f,
9e, 10b are provided as a Source Data file.
The cryo-EM maps of hSOAT1 tetramer in oval shape, hSOAT1 tetramer in rhombic
shape, hSOAT1 dimer bound with CI-976 and in the apo resting state have been
deposited in the EMDB under ID codes EMD-0829, EMD-0830, EMD-0831 and EMD-
0832.
The atomic coordinates of hSOAT1 dimer bound with CI-976 and in the apo resting
state have been deposited in the PDB under ID codes PDB 6L47 [https://doi.org/10.2210/
pdb6L47/pdb] and PDB 6L48 [https://doi.org/10.2210/pdb6L48/pdb].
24. Cases, S. et al. ACAT-2, a second mammalian acyl-CoA:cholesterol
acyltransferase. Its cloning, expression, and characterization. J. Biol. Chem.
273, 26755–26764 (1998).
25. Lada, A. T. et al. Identification of ACAT1- and ACAT2-specific inhibitors
using a novel, cell-based fluorescence assay: individual ACAT uniqueness. J.
Lipid Res. 45, 378–386 (2004).
26. Hattori, M., Hibbs, R. E. & Gouaux, E. A fluorescence-detection size-exclusion
chromatography-based thermostability assay for membrane protein
precrystallization screening. Structure 20, 1293–1299 (2012).
27. Guo, Z. Y., Lin, S., Heinen, J. A., Chang, C. C. & Chang, T. Y. The active site
His-460 of human acyl-coenzyme A:cholesterol acyltransferase 1 resides in a
hitherto undisclosed transmembrane domain. J. Biol. Chem. 280,
37814–37826 (2005).
Received: 4 February 2020; Accepted: 24 April 2020;
28. Huang, L. H. et al. Acyl-coenzyme A:cholesterol acyltransferase 1 -
significance of single-nucleotide polymorphism at residue 526 and the role
of Pro347 near the fifth transmembrane domain. FEBS J. 281, 1773–1783
(2014).
29. Joyce, C. W. et al. ACAT1 and ACAT2 membrane topology segregates a serine
residue essential for activity to opposite sides of the endoplasmic reticulum
membrane. Mol. Biol. Cell 11, 3675–3687 (2000).
References
1. Gimpl, G. Interaction of G protein coupled receptors and cholesterol. Chem.
Phys. Lipids 199, 61–73 (2016).
2. Levitan, I., Singh, D. K. & Rosenhouse-Dantsker, A. Cholesterol binding to ion
channels. Front. Physiol. 5, 65 (2014).
3. Chang, T. Y., Chang, C. C., Ohgami, N. & Yamauchi, Y. Cholesterol sensing,
trafficking, and esterification. Annu. Rev. Cell Dev. Biol. 22, 129–157 (2006).
4. Rogers, M. A. et al. Acyl-CoA:cholesterol acyltransferases (ACATs/SOATs):
Enzymes with multiple sterols as substrates and as activators. J. Steroid
Biochem. Mol. Biol. 151, 102–107 (2015).
30. Neumann, B., Chang, C. C. Y. & Chang, T. Y. Triton X-100 or octyl glucoside
inactivates acyl-CoA:cholesterol acyltransferase 1 by dissociating it from a
two-fold dimer to a two-fold monomer. Arch. Biochem. Biophys. 671, 103–110
(2019).
10
N
A
T
U
R
E
C
O
M
M
U
NI
C
A
T
I
O
N
S
|
( 2 0 2 0 ) 1 1 : 2 47 8 | h t t p s : / / d o i . o r g / 1 0 . 1 0 3 8 / s 41 4 6 7 - 0 2 0 - 1 6 2 8 8 - 4 | w ww . n a t u r e . c o m / n a t u re c o m m u n i c a t i o n s

N
A
T
U
R
E
C
O
M
M
U
N
I
CA
T
I
O
NS
|
h
t
t
p
s
:
/
/
d
o
i
.
o
r
g
/
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
2
0
-
1
6
2
8
8
-
4
A R T I C L E
31. Jogl, G. & Tong, L. Crystal structure of carnitine acetyltransferase and
60. Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. Features and development
of Coot. Acta Crystallogr. D Biol. Crystallogr. 66, 486–501 (2010).
61. Pettersen, E. F. et al. UCSF Chimera–a visualization system for exploratory
research and analysis. J. Comput. Chem. 25, 1605–1612 (2004).
62. Ruan, J. et al. TreeFam: 2008 Update. Nucleic Acids Res. 36, D735–D740
(2008).
63. Ashkenazy, H., Erez, E., Martz, E., Pupko, T. & Ben-Tal, N. ConSurf 2010:
calculating evolutionary conservation in sequence and structure of proteins
and nucleic acids. Nucleic Acids Res. 38, W529–W533 (2010).
64. Rosenthal, P. B. & Henderson, R. Optimal determination of particle
orientation, absolute hand, and contrast loss in single-particle electron
cryomicroscopy. J. Mol. Biol. 333, 721–745 (2003).
implications for the catalytic mechanism and fatty acid transport. Cell 112,
113–122 (2003).
32. Das, A., Davis, M. A. & Rudel, L. L. Identification of putative active site
residues of ACAT enzymes. J. Lipid Res. 49, 1770–1781 (2008).
33. Guo, Z. Y., Chang, C. C. & Chang, T. Y. Functionality of the seventh and
eighth transmembrane domains of acyl-coenzyme A:cholesterol
acyltransferase 1. Biochemistry 46, 10063–10071 (2007).
34. Lu, X., Lin, S., Chang, C. C. & Chang, T. Y. Mutant acyl-coenzyme A:
cholesterol acyltransferase 1 devoid of cysteine residues remains catalytically
active. J. Biol. Chem. 277, 711–718 (2002).
35. Guo, Z., Cromley, D., Billheimer, J. T. & Sturley, S. L. Identification of
potential substrate-binding sites in yeast and human acyl-CoA sterol
acyltransferases by mutagenesis of conserved sequences. J. Lipid Res. 42,
1282–1291 (2001).
Acknowledgements
The cDNAs of human SOAT1 and SOAT2 were kindly provided by Jiahuai Han. Cryo-
EM data collection was supported by Electron microscopy laboratory and Cryo-EM
platform of Peking University with the assistance of Xuemei Li, Daqi Yu, Xia Pei, Bo
Shao, Guopeng Wang, and Zhenxi Guo. Part of structural computation was also per-
formed on the Computing Platform of the Center for Life Science and High-performance
Computing Platform of Peking University. We thank Li Quan and Kunlun Cheng for
their assistance on HPLC-MS experiments. This work is supported by grants from the
Ministry of Science and Technology of China (National Key R&D Program of China,
2016YFA0502004 to L.C.), National Natural Science Foundation of China (91957201,
31622021, 31870833 and 31821091 to L.C., 31521004, 21672011 and 21822101 to
T.L., 31900859 to J.-X.W.), Beijing Natural Science Foundation (5192009 to L.C.), Young
Thousand Talents Program of China to L.C., an NIH grant in U.S.A. (R01 AG063544 to
T.Y.C. and C.C.Y.C.), and the China Postdoctoral Science Foundation (2016M600856,
2017T100014, 2019M650324, and 2019T120014 to J.-X.W.). J.-X.W. is supported by the
Boya Postdoctoral Fellowship and Peking-Tsinghua Center for Life Sciences Postdoctoral
Fellowship of Peking University. C.G. is partly supported by a postdoctoral fellowship
endorsed by Boehringer-Ingelheim.
36. Chang, C. C., Huh, H. Y., Cadigan, K. M. & Chang, T. Y. Molecular cloning
and functional expression of human acyl-coenzyme A:cholesterol
acyltransferase cDNA in mutant Chinese hamster ovary cells. J. Biol. Chem.
268, 20747–20755 (1993).
37. Das, A., Davis, M. A., Tomoda, H., Omura, S. & Rudel, L. L. Identification of
the interaction site within acyl-CoA:cholesterol acyltransferase 2 for the
isoform-specific inhibitor pyripyropene A. J. Biol. Chem. 283, 10453–10460
(2008).
38. Barnett, B. P. et al. Glucose and weight control in mice with a designed ghrelin
O-acyltransferase inhibitor. Science 330, 1689–1692 (2010).
39. Rogers, M. A. et al. Cellular pregnenolone esterification by acyl-CoA:
cholesterol acyltransferase. J. Biol. Chem. 287, 17483–17492 (2012).
40. Qian, H. et al. Structural basis for catalysis and substrate specificity of human
ACAT1. Preprint at bioRxiv https://doi.org/10.1101/2020.01.06.896597 (2020).
41. Luo, J., Yang, H. & Song, B. L. Mechanisms and regulation of cholesterol
homeostasis. Nat. Rev. Mol. Cell Biol. 21, 225–245 (2020).
42. Bruel, C., Cha, K., Reeves, P. J., Getmanova, E. & Khorana, H. G. Rhodopsin
kinase: expression in mammalian cells and a two-step purification. Proc. Natl
Acad. Sci. USA 97, 3004–3009 (2000).
43. Cadigan, K. M., Heider, J. G. & Chang, T. Y. Isolation and characterization of
Chinese hamster ovary cell mutants deficient in acyl-coenzyme A:cholesterol
acyltransferase activity. J. Biol. Chem. 263, 274–282 (1988).
44. Kawate, T. & Gouaux, E. Fluorescence-detection size-exclusion
chromatography for precrystallization screening of integral membrane
proteins. Structure 14, 673–681 (2006).
45. Goehring, A. et al. Screening and large-scale expression of membrane proteins
in mammalian cells for structural studies. Nat. Protoc. 9, 2574–2585 (2014).
46. Chang, C. C. et al. Regulation and immunolocalization of acyl-coenzyme A:
cholesterol acyltransferase in mammalian cells as studied with specific
antibodies. J. Biol. Chem. 270, 29532–29540 (1995).
Author contributions
L.C. initiated the project. C.G. developed the fluorescence-based activity assay. C.G.
purified proteins and prepared the cryo-EM samples with the help of Y.N. C.G. collected
the cryo-EM data with the help of J.-X.W. and Y.K. C.G. processed the cryo-EM data
with the help of L.C. L.C. built and refined the atomic model. S.-C.C. synthesized the
BiSAS inhibitor under the guidance of T.L. K.N., C.C.Y.C. and T.-Y.C. did the enzymatic
activity assays based on radioactive substrates. All authors contributed to the manuscript
preparation.
Competing interests
The authors declare no competing interests.
47. Zhang, Y. et al. Cholesterol is superior to 7-ketocholesterol or 7 alpha-
hydroxycholesterol as an allosteric activator for acyl-coenzyme A:cholesterol
acyltransferase 1. J. Biol. Chem. 278, 11642–11647 (2003).
48. Ritchie, T. K. et al. Chapter 11 - Reconstitution of membrane proteins in
phospholipid bilayer nanodiscs. Methods Enzymol. 464, 211–231 (2009).
49. Mastronarde, D. N. Automated electron microscope tomography using robust
prediction of specimen movements. J. Struct. Biol. 152, 36–51 (2005).
50. Zheng, S. Q. et al. MotionCor2: anisotropic correction of beam-induced
motion for improved cryo-electron microscopy. Nat. Methods 14, 331–332
(2017).
51. Zhang, K. Gctf: Real-time CTF determination and correction. J. Struct. Biol.
193, 1–12 (2016).
52. Kimanius, D., Forsberg, B. O., Scheres, S. H. W. & Lindahl, E. Accelerated
cryo-EM structure determination with parallelisation using GPUs in RELION-
2. eLife 5, e18722 (2016).
Additional information
Supplementary information is available for this paper at https://doi.org/10.1038/s41467-
020-16288-4.
Correspondence and requests for materials should be addressed to L.C.
Peer review information Nature Communications thanks David Drew, Savvas Savvides
and the other, anonymous, reviewer(s) for their contribution to the peer review of this
work. Peer reviewer reports are available.
Reprints and permission information is available at http://www.nature.com/reprints
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
53. Punjani, A., Rubinstein, J. L., Fleet, D. J. & Brubaker, M. A. cryoSPARC:
algorithms for rapid unsupervised cryo-EM structure determination. Nat.
Methods 14, 290–296 (2017).
Open Access This article is licensed under
a Creative Commons
54. Grant, T., Rohou, A. & Grigorieff, N. cisTEM, user friendly software for
single-particle image processing. eLife 7, e35383 (2018).
55. Heymann, J. B. Guidelines for using Bsoft for high resolution reconstruction
and validation of biomolecular structures from electron micrographs. Protein
Sci. 27, 159–171 (2018).
56. Zivanov, J. et al. New tools for automated high-resolution cryo-EM structure
determination in RELION-3. eLife 7, e42166 (2018).
57. Chen, S. X. et al. High-resolution noise substitution to measure overfitting and
validate resolution in 3D structure determination by single particle electron
cryomicroscopy. Ultramicroscopy 135, 24–35 (2013).
Attribution 4.0 International License, which permits use, sharing,
adaptation, distribution and reproduction in any medium or format, as long as you give
appropriate credit to the original author(s) and the source, provide a link to the Creative
Commons license, and indicate if changes were made. The images or other third party
material in this article are included in the article’s Creative Commons license, unless
indicated otherwise in a credit line to the material. If material is not included in the
article’s Creative Commons license and your intended use is not permitted by statutory
regulation or exceeds the permitted use, you will need to obtain permission directly from
the copyright holder. To view a copy of this license, visit http://creativecommons.org/
licenses/by/4.0/.
58. Kucukelbir, A., Sigworth, F. J. & Tagare, H. D. Quantifying the local resolution
of cryo-EM density maps. Nat. Methods 11, 63–65 (2014).
59. Adams, P. D. et al. PHENIX: a comprehensive Python-based system for
macromolecular structure solution. Acta Crystallogr. D Biol. Crystallogr. 66,
213–221 (2010).
© The Author(s) 2020
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
|
(
2
0
2
0
)
1
1
:
2
4
7
8
|
h
t
t
p
s
:
/
/
d
o
i
.
o
r
g
/
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
2
0
-
1
6
2
8
8
-
4
|
w
w
w
.
n
a
t
u
r
e
.
c
o
m
/
n
a
t
u
r
e
c
o
m
m
u
n
i
c
a
t
i
o
n
s
11