Small molecule inhibition of cGAS reduces interferon expression in primary macrophages from autoimmune mice

Article abstract of DOI:10.1038/s41467-017-00833-9

Cyclic GMP-AMP synthase is essential for innate immunity against infection and cellular damage, serving as a sensor of DNA from pathogens or mislocalized self-DNA. Upon binding double-stranded DNA, cyclic GMP-AMP synthase synthesizes a cyclic dinucleotide that initiates an inflammatory cellular response. Mouse studies that recapitulate causative mutations in the autoimmune disease Aicardi-Goutières syndrome demonstrate that ablating the cyclic GMP-AMP synthase gene abolishes the deleterious phenotype. Here, we report the discovery of a class of cyclic GMP-AMP synthase inhibitors identified by a high-throughput screen. These compounds possess defined structure-activity relationships and we present crystal structures of cyclic GMP-AMP synthase, double-stranded DNA, and inhibitors within the enzymatic active site. We find that a chemically improved member, RU.521, is active and selective in cellular assays of cyclic GMP-AMP synthase-mediated signaling and reduces constitutive expression of interferon in macrophages from a mouse model of Aicardi-Goutières syndrome. RU.521 will be useful toward understanding the biological roles of cyclic GMP-AMP synthase and can serve as a molecular scaffold for development of future autoimmune therapies.

Full text of DOI:10.1038/s41467-017-00833-9

A
R
T
I
C
L
E
DOI: 10.1038/s41467-017-00833-9
OPEN
S
m
a
r
l
f
l
e
m
o
n
l
e
e
c
x
u
p
l
e
i
s
n
s
h
i
b
n
i
t
i
i
n
o
n
p
o
f
m
c
a
G
A
S
r
a
e
c
d
u
p
c
h
e
s
i
n
t
e
r
o
r
e
i
o
r
i
r
y
m
r
o
a
g
e
s
f
r
o
m
a
u
t
o
i
m
m
u
n
e
G
m
i
c
e
1
2
3
,
4
2
,
7
2
,
5
6
6
1
J
e
s
s
i
c
a
V
i
n
c
e
n
t
,
C
a
r
o
l
i
n
a
A
d
u
r
a
,
P
u
a
o
,
A
n
t
o
n
i
o
L
u
z
,
L
o
d
o
e
L
a
a
,
Y
a
s
u
t
o
m
i
A
s
a
n
2
o
,
R
e
i
O
k
a
m
o
t
o
,
,
6
6
1
,
5
T
o
s
h
i
h
i
r
o
I
m
A
a
e
d
a
,
J
u
m
p
e
i
A
i
d
a
,
K
a
t
h
e
r
i
n
e
R
o
t
h
a
m
e
l
3
,
T
a
s
o
s
G
o
g
a
k
o
s
,
J
o
s
h
u
a
S
t
e
i
n
b
e
r
g
,
S
e
t
h
R
e
a
s
o
n
e
r
6
2
,
5
2
1
K
a
z
u
y
o
s
h
i
s
o
,
T
h
o
m
a
s
T
u
s
c
h
l
,
D
i
n
s
h
a
w
J
.
P
a
t
e
l
,
J
.
F
r
a
s
e
r
G
l
i
c
k
m
&
M
a
n
u
e
l
A
s
c
a
n
o
C
y
c
l
i
c
G
M
P
-
A
M
g
P
s
y
n
t
h
a
s
e
i
s
e
s
s
e
n
t
i
a
l
f
o
r
i
n
n
a
t
e
i
m
m
u
n
i
t
y
a
g
a
i
n
s
t
i
n
f
e
c
t
i
o
n
a
n
d
c
e
l
l
u
l
a
r
d
d
a
m
a
g
e
,
s
e
r
v
i
n
a
s
a
s
e
n
s
o
r
o
f
D
N
A
f
r
o
m
p
a
t
h
o
g
e
n
s
o
r
m
i
s
l
o
c
a
l
i
z
e
d
s
e
l
f
-
D
N
A
.
U
p
o
n
b
i
n
d
i
n
g
o
u
b
l
e
-
s
t
r
a
n
d
e
d
D
N
A
,
c
y
c
l
i
c
G
M
P
-
A
M
P
s
y
n
t
h
a
s
e
s
y
n
t
h
e
s
i
z
e
s
a
c
y
c
l
i
c
d
i
n
u
c
l
e
o
t
i
d
e
t
h
a
t
e
g
e
i
n
i
t
i
a
t
e
s
a
n
i
n
fl
a
m
m
a
t
o
r
y
c
e
l
l
u
l
a
r
r
e
s
p
o
n
s
e
.
M
o
u
s
e
s
t
u
d
i
e
s
t
h
a
t
r
e
c
a
p
i
t
u
l
a
t
e
c
t
a
u
s
a
t
i
v
n
h
m
u
t
a
t
i
o
n
s
i
n
t
h
e
a
u
t
o
i
m
m
u
n
e
d
i
s
e
a
s
e
a
A
i
c
a
r
d
e
i
-
G
o
u
t
i
è
r
e
s
s
y
n
d
r
o
m
e
d
e
m
o
n
s
t
r
a
t
e
t
h
e
a
a
b
l
a
t
i
t
h
e
c
y
c
l
i
c
G
o
M
P
-
A
M
P
s
y
n
t
h
a
s
e
g
e
n
e
b
o
l
i
s
h
s
t
h
e
d
e
l
e
t
e
r
i
o
u
s
p
h
e
n
o
t
y
p
e
.
H
e
r
e
,
w
r
e
p
o
r
t
t
d
i
s
c
o
v
e
r
y
f
a
c
l
a
s
s
o
f
c
y
c
l
i
c
G
M
P
-
A
M
P
s
y
n
t
h
a
s
e
i
n
h
i
b
i
t
o
r
s
i
d
e
n
t
i
fi
e
d
b
y
a
h
i
g
h
-
t
h
r
o
u
g
h
p
u
t
s
c
r
e
e
n
.
T
h
e
s
e
c
o
m
p
o
u
n
d
s
p
o
s
s
e
s
s
d
e
fi
n
n
e
d
s
t
r
u
c
t
u
r
e
-
a
c
t
i
v
i
t
y
r
e
l
a
t
i
o
n
s
h
i
p
s
a
n
d
w
e
p
r
e
s
e
n
t
c
r
y
s
t
a
l
s
t
r
u
c
t
u
r
e
s
o
f
c
y
c
l
i
c
G
M
P
-
A
M
P
s
y
n
t
h
a
s
e
,
d
o
u
b
l
e
-
s
t
r
a
n
d
e
d
D
N
A
,
a
n
d
i
n
h
i
b
i
t
o
,
r
s
s
w
i
t
h
i
n
t
h
e
e
n
z
y
m
a
t
i
c
a
c
t
i
v
e
s
i
t
e
.
W
e
fi
d
t
h
a
t
a
c
h
e
m
i
c
a
l
l
y
i
m
p
r
o
v
e
d
m
e
m
b
e
r
,
R
U
.
5
2
1
i
a
a
c
t
i
v
e
a
n
d
s
e
l
e
c
t
i
v
e
i
n
c
e
l
l
u
l
a
r
a
s
s
a
y
s
o
f
c
y
c
l
i
c
G
M
P
-
A
M
P
s
y
n
t
h
a
s
e
-
m
e
d
i
a
t
e
d
s
i
g
n
a
l
i
n
g
n
d
r
e
d
u
c
e
s
c
o
n
s
t
i
t
u
t
i
v
e
e
x
p
.
r
e
s
s
i
o
n
o
f
i
n
t
e
r
f
e
r
o
n
i
n
m
a
c
r
o
p
h
a
g
e
s
f
r
o
m
a
m
o
u
s
e
m
o
d
e
l
o
f
A
i
c
a
r
d
i
-
G
o
u
t
i
è
r
e
s
s
y
n
d
r
o
m
e
R
U
.
5
2
1
w
i
l
l
b
e
e
u
s
e
f
u
l
t
o
w
a
r
d
u
n
d
e
r
s
t
a
n
d
i
n
g
e
t
h
e
b
i
o
l
o
g
i
c
a
f
l
r
o
l
e
s
o
f
c
y
c
l
i
c
G
M
P
-
A
M
P
s
y
n
t
h
a
s
e
a
n
d
c
a
n
s
e
r
v
a
s
a
m
o
l
e
c
u
l
a
r
s
c
a
f
f
o
l
d
f
o
r
d
v
e
l
o
p
m
e
n
t
o
f
u
t
u
r
e
a
u
t
o
i
m
m
u
n
e
t
h
e
r
a
p
i
e
s
.
1 Vanderbilt University School of Medicine, Nashville, TN 37027, USA. 2 The Rockefeller University, New York, NY 10065, USA. 3 Structural Biology Program,
Memorial Sloan-Kettering Cancer Center, New York, NY 10065, USA. 4 Key Laboratory of Infection and Immunity, CAS Center for Excellence in
Biomacromolecules, Institute of Biophysics, Chinese Academy of Sciences, Beijing 100101, China. 5 Howard Hughes Medical Institute Laboratory for RNA
Molecular Biology, New York, NY 10065, USA. 6 Tri-Institutional Therapeutics Discovery Institute, New York, NY 10021, USA. 7Present address: Regeneron
Pharmaceuticals Incorporated, Tarrytown, NY 10591, USA. Jessica Vincent, Carolina Adura, and Pu Gao contributed equally to this work. Correspondence
and requests for materials should be addressed to D.J.P. (email: pateld@mskcc.org) or to J.F.G. (email: fglickman@mail.rockefeller.edu) or to
M.A. (email: manuel.ascano@vanderbilt.edu)
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
8
:
7
5
0
D
O
I
:
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
1
7
-
0
0
8
3
3
-
9
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
|
D
O
I
:
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
1
7
-
0
0
8
3
3
-
9
he innate immune system contains protein sensors to initiating a signal transduction cascade which leads to the
detect aberrantly modified and/or mislocalized nucleic activation of the transcription factor interferon regulatory factor
acids, perceiving such molecules as foreign or markers of 3 (IRF3) and the upregulation of cytokines including type I
T
cellular stress1–5. Sensing of aberrant nucleic acids leads to the interferon beta 1 (IFNB1)11, 17–21. Many studies have now
activation of downstream signal transduction pathways and implicated the importance of cGAS in the innate immune
inflammatory responses through the upregulation of type I response to intracellular and prokaryotic pathogens such as
interferon genes. One such pattern recognition receptor is cyclic L. monocytogenes22, C. trachomatis23, L. pneumophila24, and
GMP-AMP synthase (cGAS, official gene name MB21D1)6, 7
,
M. tuberculosis24–26. Moreover, the cGAS pathway can detect or
which detects cytoplasmic double-stranded DNA (dsDNA), be suppressed by viruses during infection, including members of
indicative of an infection by a virus or bacterial pathogen the herpes family27, 28, the oncogenic viruses hAD5 and
or mislocalization of nuclear or mitochondrial DNA8–10
.
HPV1829, as well as retroviruses including HIV30, highlighting its
Upon binding to dsDNA, cGAS utilizes ATP and GTP to role as a key sentinel against pathogenic infection.
synthesize the only known metazoan cyclic-dinucleotide, cyclic
While the cellular immune response to dsDNA plays an
GMP-AMP (cGAMP or c[G(2′,5′)pA(3′,5′)p])11–16, a molecule indispensable role in pathogen defense, an abnormal response to
in which guanosine and adenosine are linked with a 2′,5′- and a dsDNA has been shown to be an important factor in the etiology
3′,5′-phosphodiester bond following sequential ligation at the of hyper-inflammatory or autoimmune disorders. A hallmark of
same active site. cGAMP acts as a second messenger, diffusing autoimmune disorders such as systemic lupus erythematosus
and binding to the endoplasmic reticulum membrane-bound (SLE) is the development of patient serum immunoreactivity
adapter protein, Stimulator of interferon genes (STING), thus against dsDNA31. Inactivating mutations in the enzyme
a
b
c
100
80
60
40
20
0
cGAS + dsDNA
DMSO
No dsDNA
Compounds
ATP + GTP
cGAMP
cGAMP
Putative
antagonist
80
60
40
20
–20
0
80
ATP
GTP
–20
0
20
40
60
80
100
60
40
20
Day 2 NPI (%)
d
60
50
40
30
20
10
0
0
80
60
40
20
0
A01
N
C01
E01
G01
I01
K01 M01
O01
Well location
dsDNA (μM)
e
H
H
N
O
N
N
O
S
O
N
HN
N
O
O
N
N
H
O
N
HO
O
N
OH
Cl
S
Cl
N
H
RU166365
1.89
RU114757
RU191752
0.07
RU100840
0.03
IC50 (μM)
1.4
Fig. 1
D
e
s
g
v
o
e
F
e
f
n
o
l
o
c
r
r
p
y
a
t
m
e
n
t
G
o
i
o
M
f
f
a
h
A
A
i
g
h
-
P
P
t
h
r
o
u
g
h
p
)
u
. b
g
e
t
s
c
T
n
r
h
e
e
R
0
e
n
e
f
o
y
S
r
t
m
T
s
h
a
h
D
e
t
e
N
i
c
d
a
e
c
s
n
t
a
w
t
i
i
fi
c
a
t
o
s
i
o
f
i
d
n
c
n
o
f
c
S
a
G
A
d
d
u
S
i
n
e
r
h
r
1
i
m
2
m
b
i
i
t
n
o
e
m
u
r
d
s
.
a
S
m
c
h
o
g
S
e
n
6
a
m
a
r
n
t
i
i
n
M
c
g
o
t
f
h
G
i
e
m
c
m
o
.
G
u
n
N
A
n
e
s
m
m
P
t
p
a
i
m
u
n
e
d
l
a
o
t
o
f
x
t
r
A
y
d
s
D
(blue
N
A
-
)
n
d
a
e
n
c
a
p
d
o
s
e
n
G
n
m
d
T
t
e
u
e
P
v
n
t
c
G
A
S
)
s
y
n
t
h
e
s
i
c
l
i
c
P
-
M
M
( cGAMP
( red
n
z
i
v
i
t
y
G
A
i
s
e
t
b
y
i
t
o
s
u
t
i
o
T
P
(green
l
,
a
n
d
t
h
e
e
t
h
i
o
e
n
c
G
)
s
)
u
c
a
s
r
t
i
n
a
F
-
M
.
d
a
s
y
s
w
u
.
a
s
5 μ
c
u
b
t
e
f
l
o
a
m
0
i
n
u
c
s
i
n
0
c
C
c
;
A
S
o
M
a
r
l
i
z
d
e
c
i
t
r
a
c
t
e
d
i
o
u
a
u
e
s
a
r
e
p
l
o
t
t
e
d
.
c
h
o
g
h
-
t
h
r
o
u
g
h
p
u
t
e
a fi
n
,
3
0
l
n
n
M
c
A
n
a
e
t
o
s
t
i
m
t
e
o
s
i
e
c
G
A
t
i
v
i
t
y
%
y
c
f
p
c
r
o
u
g
f
o
r
m
a
t
i
o
n
e
w
o
a
s
u
r
e
d
a
g
a
i
n
s
t
1
2
6
6
c
o
m
p
u
n
d
s
(
b
l
u
e
d
o
t
s
n
a
c
o
e
n
t
r
a
t
i
o
o
f
1
2
M
u
s
i
n
g
S
i
g
a
-
A
l
d
r
h
L
O
P
A
(
l
i
b
r
a
r
o
p
h
r
m
a
o
l
o
c
a
l
l
y
a
c
t
i
v
c
m
p
o
n
d
)
l
i
b
r
a
r
y
p
l
a
n
r
m
t
t
r
e
r
e
s
o
.
l
T
D
h
M
n
r
e
r
S
w
0
e
O
a
0
s
u
i
l
s
0
0
t
s
s
o
f
t
w
w
n
T
a
o
i
h
i
n
d
e
p
e
n
a
t
l
d
n
s
e
e
d
u
c
n
t
t
i
m
d
a
y
p
s
o
,
a
s
t
v
P
e
e
c
c
A
e
h
o
C
n
n
i
c
t
a
l
r
l
e
(
y
p
n
r
g
l
o
i
c
d
s
f
a
s
u
o
t
e
D
s
N
e
,
a
A
r
)
e
p
l
o
s
e
t
h
a
e
t
o
e
w
s
c
d
a
n
y
g
i
h
a
i
n
s
t
e
a
c
a
h
o
w
p
a
t
h
e
e
r
r
e
e
a
a
a
f
r
s
n
a
a
n
a
o
r
z
r
r
m
e
o
a
d
f
t
l
i
u
z
e
d
n
6
.
p
g
. e
e
V
r
c
o
o
t
e
n
e
t
x
o
s
f
i
o
g
r
n
h
w
i
b
a
e
a
i
r
t
e
h
i
o
a
n
n
h
o
(
d
-
n
N
t
P
h
r
e
I
)
e
u
t
.
T
h
e
p
g
e
f fi
u
e
n
c
t
p
e
i
s
r
g
e
c
o
a
n
r
c
t
t
e
e
i
v
o
n
s
e
f
c
c
o
o
h
o
n green
e
m
h
e
s
i
t
i
r
o
i
s
n red
a
o
.
D
t
a
l
y
s
i
r
t
f
t
c
o
l
a
t
i
o
s
,
.
8
6
s
.
d
r
o
e
s
u
l
u
s
o
n
g
t
h
e
n
L
d
O
s
l
o
i
b
r
a
r
e
l
t
d
w
i
n
t
h
s
s
a
v
i
n
g
d
a
Z
r
l
i
m
c
t
o
e
0
i
.
7
n
F
l
e
l
o
w
i
n
t
h
i
g
t
h
o
g
h
e
f
r
o
o
v
e
1
0
0
m
l
l
-
l
e
c
p
o
u
,
t
h
f
l
l
o
w
i
n
u
r
e
r
e
e
l
t
e
d
f
r
a
d
i
t
i
o
n
c
h
c
t
i
z
a
o
S
e
e
x
t
f
o
d
e
t
i
l
s
t
h
r
i
a
s
2
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
8
:
7
5
0
D
O
I
:
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
1
7
-
0
0
8
3
3
-
9
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
|
D
OI
:
1
0
.
1
0
3
8
/
s
4
1
46
7
-
01
7
-
0
0
8
3
3
-
9
A
R
T
I
C
L
E
a
b
RU.365
Active pocket
c
F234
R364
R364
Y421
A233
F367
C419
Y421
90°
C419
F473
I470
L475
H422
H467
d
e
f
g
Arg364
Arg364
Arg364
Arg364
RU.365
ATP
5’-pppGpG
2’,3’-cGAMP
Tyr421
Tyr421
Tyr421
Tyr421
Fig. 2
T
e
r
n
a
r
y
c
r
y
s
t
a
l
s
t
r
u
c
t
u
r
e
o
f
c
G
A
S
w
i
t
h
d
s
D
N
A
a
n
d
R
U
.
3
6
5
.
a
O
v
e
r
a
l
l
s
t
r
u
c
t
u
r
e
o
f
t
h
e
c
G
A
S
-
d
D
N
A
-
R
U
.
3
6
5
c
o
m
p
l
e
x
,
w
i
t
h
c
G
A
S
i
n
v
i
o
l
e
t
,
d
s
D
N
A
i
n
l
i
g
h
t
orange
,
a
n
d
l
i
g
a
n
d
i
n
y
e
l
l
o
w
.
R
U
.
3
6
5
i
s
s
h
o
w
n
i
n
s
t
i
c
k
r
e
p
r
e
s
e
n
t
a
t
i
o
n
.
b
C
l
o
s
e
-
u
p
o
f
t
h
e
c
G
A
S
-
b
i
n
d
i
n
g
p
o
c
k
e
t
(
s
h
o
w
n
i
n
s
u
r
f
a
c
e
r
e
p
r
e
s
e
n
t
a
t
i
o
n
a
n
d
v
i
o
l
e
t
)
a
n
d
b
dots
o
u
n
a
b
d
n
o
R
U
.
3
6
,
U
5
i
.
(
s
h
o
w
i
)
)
n
s
,
s
u
A
p
a
o
P
c
u
e
n
( e
-
fi
l
n
l
i
n
g
g
t
-
r
h
pp
e
e
p
p
b
G
r
e
o
p
s
u
G
e
n
n
d
(f
t
a
R
)
t
U
,
i
o
.
a
n
3
n
a
n
d
y
e
l
l
o
w
)
i
.
T
h
e black dashed box
n stick n yellow . d–g
(g
i
T
g
n
h
a
d
e
n
i
c
c
s
a
o
t
e
n
n
s
s
t
e
d
h
e
e
h
e
d
n
s
e
t
t
i
r
a
t
e
c
w
b
i
n
n
g
d
i
i
n
n
g
p
o
c
s
c
t
a
k
e
t
n
o
m
f
c
e
G
di
n
A
S
e
. c
d
T
h
A
e
a
m
i
n
4
o
a
c
d
i
d
T
s
(
s
h
o
w
n
i
n
d
s
t
i
c
k
n
v
i
o
l
e
t
r
r
d
i
6
5
(
s
h
o
w
n
,
i
) .
)
r
v
k
i
t
e
r
a
i
o
a
t
b
y
r
g
3
6
a
n
y
r
4
2
1
w
i
t
h
t
h
e
u
n
d
R
3
6
5
(
d
T
)
,
5
′
d
2
′
,
3
′
-
c
G
A
M
P
T
h
e
b
o
u
n
d
l
i
d
a
t
e
s
o
r
e
s
i
d
u
e
r
e
s
h
o
w
i
n
s
t
i
c
k
r
e
p
r
e
s
e
n
t
a
t
i
o
n
responsible for degrading cytoplasmic DNA, the human DNA
Here, we report the discovery and characterization of a class of
exonuclease TREX1, cause Aicardi-Goutières syndrome32 (AGS) compounds that binds to the catalytic pocket of cGAS and
or Chilblain lupus33, both autoimmune disorders manifesting inhibits its dsDNA-stimulated activity. We define structure-
overproduction of interferon and whose clinical symptoms activity relationships and present the crystal structure of an
overlap with SLE. Knockout studies using bone marrow-derived inhibitor–enzyme–dsDNA. We further demonstrate the potency
macrophages (BMDMs) of Trex1 null mice have shown elevated and selectivity of a chemically improved inhibitor, RU.521, in
levels of interferon, which are normalized upon knockout of cellular assays showing that while it inhibits cGAS-mediated
cGAS34, 35, Sting36, 37, or IRF336, substantiating the contribution interferon upregulation, it has reduced to no effect on inflam-
of cGAS function in these diseases. Moreover, recent work has matory pathways independent of cGAS. Furthermore,
implicated mitochondria as a cell-intrinsic source of aberrantly RU.521 suppresses the chronically elevated levels of type I
localized DNA, indicating that mitochondrial stress can promote interferon observed in primary macrophages from Trex1
an inflammatory response by triggering cGAS activity38, 39. Since null mice, a model of AGS. We expect this chemical scaffold
damaged mitochondria can be the consequence of many kinds of to facilitate studies of the physiological roles played by cGAS,
cellular insults, the number of inflammatory responses for which which will foster a greater understanding of innate immune
cGAS is a major driver may still be underestimated40.
mechanisms.
Given its central importance in innate immunity, small-mole-
cule inhibitors of the enzymatic activity of cGAS could be used to
enhance research efforts to dissect cGAS mechanism and regula-
tion of innate immunity. Additionally, small molecules that target
cGAS can be an important step toward the development of drugs
in the treatment of human disorders relevant to a dysregulated
cGAS pathway. Recently, an in silico screen has identified existing
anti-malarial small molecules with some efficacy toward inhibiting
cGAS by interfering with its association with dsDNA40. However,
anti-malarial drugs can intercalate dsDNA, and their mechanism
of inhibition appears to rely on non-specific interactions with
nucleic acid rather than direct association and interference of
cGAS and its enzymatic activity. Intercalating compounds have a
higher propensity of having side effects in the cell.
Results
Discovery of a small-molecule inhibitor of cGAS. We sought
to develop a high-throughput screen using a small-molecule
library to identify compounds that could modulate enzymatic
production of cGAMP. In this report, we describe our discovery
and characterization of cGAS inhibitors. The enzymatic synthesis
of the cyclic dinucleotide cGAMP is readily performed in cell-free
conditions by combining purified recombinant mouse cGAS
enzyme with pre-annealed 45-bp dsDNA in a reaction buffer, and
adding ATP and GTP as substrates (Fig. 1a). We used an Agilent
RapidFire mass spectrometry system (RF-MS) to simultaneously
quantify the abundance of substrates and product in reaction
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
8
:
7
5
0
D
O
I
:
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
1
7
-
0
0
8
3
3
-
9
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
|
D
O
I
:
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
1
7
-
0
0
8
3
3
-
9
Table 1 X-ray statistics for cGAS-DNA-inhibitor complexes
cGAS-dsDNA-RU.365
cGAS-dsDNA-RU.332
cGAS-dsDNA-RU.521
Data collection
S
p
a
ce
g
r
o
u
p
I2
2
2
I2
2
2
I2
2
2
C
e
l
l
d
i
m
e
n
s
i
o
n
s
a
α
, b, c
, β, γ
(
(
n
Å
)
8
9
5
0
.
.
5
0
2
5
,
,
9
9
8
0
.
8
,
1
9
3
0
0
.
2
.
0
2
8
9
5
0
.
.
3
0
2
7
,
,
9
9
8
0
.
6
,
1
9
2
3
2
0
4–
)
9
.
.
2
2
0
.
8
9
5
5
0
0 –
.
.
6
0
1
4
,
,
.
4
9
9
8
0
.
.
(
.
4
0
,
,
8
2
1
3
0
.
4
°
)
.
0
,
.
0
,
.
9 0 . 0
7–
0
a
a
a
a
R
Rs
I
C
e
s
o
l
u
t
i
o
(
e r g e
Å
)
5
0 –
.
1
3
(
(
.
2
5
.
4
1
9
–
.
1
3
)
)
5
0
–
.
1
8
(
.
2
3
1
8
8
)
)
8
3
(
1
.
1
.
8
3
3
)
)
o
r
R
0
.
0
3
(
1
)
9
)
4
)
0
.
0
3
(
(
1
6
0
.
0
1
)
5
)
y
m
m
/
σ
I
1
3
.
3
1
.
1
1
2
.
9
6
8
.
1
.
3
)
1
7
.
8
7
(
1
.
1
o
m
p
l
e
t
e
n
e
y
s
s
(
%
)
9
8
2
.
8
(
(
9
.
7
)
9
0
(
(
9
5
)
.
1
)
9
4
.
4
(
9
4
.
1
)
R
e
d
u
n
d
a
n
c
3
.
3
.
3
3
3
.
3
.
4
.3 ( 4 .3 )
Refinement
a
a
R
N
e
s
o
l
u
t
i
o
n
(
Å
t
)
5
3
0 –
0
2
3
.
1
(
2
.
1
9 –
2
.
1
3
5
2
0 –
8
2
3
.
1
8
(
2
.
2
4
–
2
.
1
5
0
–
1
.
8
3
(
1
.
8
7
–
1
.
8
o
.
o
f
r
e
fl
e
c
i
o
n
s
,
0
3
,
7
3
4
5
,
2
3
9
Rw / Rf r
0
.
2
1
3
/
0
.
2
4
8
0
.
2
4
2
/
0
.
2
8
1
0
.
1
9
4
/
0
.
2
3
2
o
r
k
e
e
N
o
.
o
f
a
t
o
m
s
P
r
o
t
e
i
n
3
5
2
9
3
5
2
9
3
5
2
9
L
W
i
g
a
t
o
g
a
a
t
o
t
a
t
n
e
r
e
n
e
d
r
s
/
i
o
n
n
2
6
2
8
2
6
2
1
8
0
B-
f
a
c
r
i
P
L
W
i
n
7
2
.
4
4
0
6
6
8
5
7
6
.
.
9
6
7
7
3
2
5
8
4
8
.
2
6
d
r
/
i
o
1
5
6
.
5
0
.
7
3
5
6
.
4
3
.
9
.9
0
R
.
m
.
s
.
d
ev i a t i o n s
B
B
o
o
n
n
d
d
l
a
e
n
n
g
g
t
h
s
s
(
(
°
Å
)
)
0
.
0
2
0
8
9
0
.
0
5
0
3
9
0 .0 0 9
1 . 1 5 3
l
e
1
.
2
1
.
2
a
O
n
e
c
r
y
s
t
a
l
f
o
r
e
a
c
h
s
t
r
u
c
t
u
r
e
;
v
a
l
u
e
s
i
n
p
a
r
e
n
t
h
e
s
e
s
a
r
e
f
o
r
h
i
g
h
e
s
t
-
r
e
s
o
l
u
t
i
o
n
s
h
e
l
l
mixtures containing recombinant cGAS and dsDNA. Figure 1b We obtained a linear regression coefficient of 0.86 (Fig. 1c) and
shows a typical RF-MS chromatogram where ATP and GTP were a Z′ value of 0.76 (Fig. 1d). Subsequently, a total of 123,306
present at equimolar concentration (0.3 mM), while the dsDNA compounds were screened (Supplementary Table 1) at a
concentration was varied between 0 to 3 µM. After 120 min, concentration of 12.5 µM. A cutoff of 60% inhibition normalized
reactions were stopped and the abundance of ATP, GTP, and by the intra-plate controls and a Z′ criterion of > 0.5 was applied
cGAMP were measured. The EC50 for dsDNA is 0.017 µM yielding 229 (0.19%) compounds. These were re-tested in
(Supplementary Fig. 1a). At 0.3 µM dsDNA, the rate of product triplicate concentration response experiments. One-hundred
formation was linear over 120 min at room temperature yielding and four compounds showed an IC50 ≤ 10 µM. Of these, 17
approximately 50% of cGAMP (Supplementary Fig. 1b). We then contained structural motifs that are recurrent in PAINS42 and,
measured under steady-state conditions the effect of substrate therefore, were removed from further consideration, yielding
concentration on catalytic rate. We found that the saturation 87 compounds. Further high-performance liquid chromatogra-
curves consistently and significantly fit better to a sigmoidal curve phy (HPLC) mass spectrometry analysis determined that
rather than a hyperbolic function. This suggests that the substrate 49 compounds showed the correct molecular mass and had a
binding works cooperatively, and is in agreement with the purity ≥ 85% in the first chromatography gradient run in positive
structural findings that show cGAS can exist as a 2:2 dimer ion mode (Supplementary Table 2). These 49 compounds
with two enzymes binding to two DNA molecules41. Thus, it is were evaluated for potential DNA intercalation by measuring
possible that initial binding of ATP and GTP to the active site of a the displacement of acridine orange bound to dsDNA by means
cGAS molecule, could cause a higher affinity binding of substrates of fluorescence polarization (FP). Fourteen compounds showed
to occur with the second cGAS molecule. We used an allosteric > 50% of acridine orange displacement from dsDNA compared
sigmoidal fitting and found that ATP has a Km of 86 6.7 and a to 30 µM mitoxantrone—a known DNA intercalator43. From
kcat 23 (min−1) The Km and kcat for GTP are 96.4 5.1 µM and 23 35 remaining compounds, 4 compounds (Fig. 1e) had 50%
(min−1) respectively (Supplementary Fig. 1c, d). Taken together, inhibitory concentrations (IC50’s) ranging from 0.03 µM to
these results show the utility of RF-MS for measuring in vitro 1.9 µM (confirmed with independent lots of compound) and
cGAS activity and that reaction conditions were amenable to were then prioritized based on stability, potency, and structural
high-throughput screening.
diversity for further testing in crystallization trials.
We also tested whether human cGAS could be used in the
high-throughput screen by performing an enzyme progress curve
(Supplementary Fig. 2). The signal for human cGAS is
undetectable under the same conditions used with mouse cGAS.
Due to the low signal, human cGAS was not suitable for accurate
kinetic characterization using the technique presented in this
paper and the screen and subsequent validation assays were
performed with the murine version.
RU.365 binds to the active site of cGAS. We were able to solve
the crystal structure of RU166365 (hereafter as RU.365) in
complex with cGAS and dsDNA (Fig. 2a and Supplementary
Fig. 4), allowing us to better examine the structural basis for the
affinity of this inhibitor. The ternary complex was obtained by
using the co-crystallization method and the structure was solved
at 2.13 Å resolution by molecular replacement based on PDB:
4K96 (X-ray statistics in Table 1). The final cGAS-dsDNA-
RU.365 ternary complex shows good refinement statistics
A pilot study was conducted in two different days using
1268 compounds from the Sigma Aldrich LOPAC compound
collection in order to test the statistical robustness of the assay.
4
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
8
:
7
5
0
D
O
I
:
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
1
7
-
0
0
8
3
3
-
9
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
|
D
O
I
:
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
1
7
-
0
0
8
3
3
-
9
A
R
T
I
C
L
E
a
b
RU.365
RU.521
RU.365
RU.521
c
d
e
G290
Y421
Water
K350
RU.521
R364
f
g
h
GTP
2’3’-cGAMP
5’-pppGpG
RU.365
GTP
RU.365
RU.365
ATP
i
j
k
2’3’-cGAMP
5’-pppGpG
ATP
RU.521
RU.521
RU.521
Fig. 3
A
s
o
6
R
t
s
5
U
r
i
u
c
o
t
n
u
r
o
a
f
l
c
o
r
r
m
p
t
h
a
f
a
u
o
n
r
r
i
e
w
s
s
o
n
o
i
3
t
o
f
R
U
n
,
y
.
d
3
6
R
i
2
5
U
w
i
t
5
h
i
t
s
d
e
r
i
)
v
a
a
s
i
t
n
s
i
v
d
h
e
R
o
d
R
U
w
i
U
.
5
2
2
1
1
. a
( cyan
n dots
O
m
i
t
F
T
S
w
o
h
( violet
-
e
F
c
d
e
l
h
e
l
i
c
o
s
2
t
r
r
o
s
1
o
n
s
o
d
e
s
n
n
s
t
i
u
t
y
(
g
r
e
e
n
m
e
s
h
)
e
T
o
f
R
U
.
t
3
h
c
3
6
e
5
c
n
0
a
G
g
n
A
d
S
t
R
R
p
U
o
.
c
5
k
2
e
o
1
t
n
c
a
c
.
b
o
R
e
m
n
U
t
p
t
.
w
o
5
e
u
2
e
x
r
1
e
d
a
t
3 σ
)
.
b
S
u
p
e
r
p
t
i
s
t
u
e
c
s
f
b
o
u
.
3
6
( yellow
c
a
U
.
5
)
. c
i
c
u
b
t
i
t
i
o
n
t
a
r
ge
t
d
e
p
e
r
i
n
( yellow
n
a
n
d
R
U
.
3
(
s
i
l
v
e
r
)
a
n
R
c
n
s
t
i
c
k
w
4
b
t
h
t
w
o
h
e
3
l
o
r
i
m
n
e
d
n
i
.
c
G
A
)
U
o
h
w
i
n
s
u
r
f
a
c
e
v
i
e
w
.
d
h
e
)
s
t
a
k
i
i
n
e
r
a
c
t
i
R
U
.
o
5
w
n
2
1
n
n
(yellow
)
/
.
3
6
r
5
o
( silver
i
)
o
d
t
6
4
/
T
r
f
1
u
.
e
W
t
.
r-
e
a
t
e
n
t
e
r
a
2
c
t
i
o
n
b
e
t
h
e
e
n
s
R
e
.
5
(yellow (violet
b
)
,
a
n
d
c
G
A
S
.
L
y
s
d
5
c
i
n
U
.3
6
5
g
o
l
e
i
s
s
h
i
n
s
i
l
v
e
r
.
S
u
p
e
p
s
t
i
o
n
s
r
u
u
r
e
s
o
o
n
d
R
6
5
f
–
h
a
n
d
R
U
.
5
1
i
–
k
w
i
t
t
h
o
f
s
u
s
t
r
a
t
e
s
i
n
t
e
r
m
ed
i
a
t
e
,
a
n
d
p
r
o
u
t
.
A
l
l
t
h
e
l
i
n
d
s
a
r
e
s
h
o
w
i
stick
r
e
p
r
e
s
e
n
t
a
t
i
o
n
.
I
n
h
i
b
i
t
o
r
s
a
r
e
s
h
o
wn
i
n
y
e
l
l
o
w
.
S
u
b
s
t
r
a
t
e
s
/
I
n
t
e
r
m
e
d
i
a
t
e
/
P
r
o
d
u
c
t
a
r
e
s
h
o
w
n
i
n
c
y
a
n
(Rwork: 21.3%; Rfree: 24.8%) and a ligand geometry well-defined by Tyr421, His422, His467, and Phe473, the hydrophobic amino
electron density (Fig. 3a). cGAS adopts an active conformation in acids are Ala233, Ile470, and Leu475), and the charged or polar
this ternary complex characterized by a DNA-induced “open amino acids are Arg364 and Cys419 (Fig. 2c). The benzimidazole
pocket” (Fig. 2a, b), similar to our previously reported structure of ring and a part of the pyrazole ring of bound RU.365 partially
the complex of cGAS with bound dsDNA and cGAMP14. RU.365 stack between the guanidinium group of Arg364 and Tyr421
occupies only one side of the cGAS pocket, as does cGAMP in (Fig. 2c, d), which represent the key intermolecular contacts
our previously reported structure14, leaving the remainder of between RU.365 and cGAS. Of note, point mutations of Arg364
the pocket filled with solvent (Fig. 2b). Of the amino acids and Tyr421 render it incapable of responding to dsDNA, and
surrounding RU.365, the aromatic ones are Phe234, Phe367, no interferon upregulation is observed14. Stacking interactions
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
8
:
7
5
0
D
O
I
:
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
1
7
-
0
0
8
3
3
-
9
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
|
D
O
I
:
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
1
7
-
0
0
8
3
3
-
9
a
100
80
60
40
20
0
RU.365
RU.521
–2
–1
0
1
2
Log compound (μM)
b
3.4
3.3
3.2
3.1
3.0
2.9
0
–5
–10
–15
0
1000
2000
Time (s)
3000
0.0
0.5
1.0
1.5
2.0
2.5
Molar ratio
Fig. 4
c
G
A
d
e
R
S
i
n
n
-
t
h
i
b
u
i
t
s
n
t
i
o
e
S
n
,
u
a
b
y
.
l
d
R
U
1
e
.
,
3
i
n
6
n
5
t
r
a
n
d
p
a
d
i
t
s
s
e
D
c
h
c
.
A
e
e
V
m
o
a
i
c
f
u
m
a
d
e
l
a
c
N
r
x
t
i
v
A
a
e
,
v
a
A
e
n
T
a
P
g
l
o
a
e
g
n
s
s
d
o
. a
G
I
T
n
v
a
i
t
s
a
r
o
m
e
c
e
d
o
a
e
n
s
c
u
e
r
n
e
t
d
i
r
a
b
a
t
y
t
i
o
n
R
n
r
e
M
w
s
p
S
o
.
h
n
T
S
s
h
D
e
e
c
I
u
C
d
r
5
v
e
0
a
s
f
o
l
. b
r
u
R
U
f
.
o
o
3
6
r
h
5
R
e
a
U
n
d
6
l
t
5
t
h
,
e
R
m
o
s
2
t
1
c
p
a
o
n
o
t
d
r
e
n
t
a
n
p
y
a
o
-
l
o
t
i
g
e
n
s
y
n
t
h
e
s
i
z
e
i
h
o
d
R
U
5
2
h
e
r
e
e
n
s
D
P
l
F
-
v
a
e
s
.
3
U
.
5
l
e
s
s
n
d
t
i
a
c
n
a
v
l
o
e
g
s
s
f
a
r
l
U
i
s
e
i
p
p
e
m
t
a
y
T
/
b
s
l
4
N
l
s
p
a
e
r
a
f
t
r
i
p
i
c
t
t
e
r
m
n
i
o
s
i
t
i
n
i
c
t
e
d
I
s
t
r
m
a
i
t
r
a
t
i
o
n
a
l
i
m
e
t
r
b
n
g
u
r
o
r
.
5
2
1
i
t
r
t
e
i
n
t
o
c
G
A
S
c
o
l
e
were also found in substrate- (ATP, Fig. 2e), intermediate- of cGAS and subsequently push the benzimidazole ring away
(5′-pppGpG, Fig. 2f), and product-bound (cGAMP, Fig. 2g) from the protein (Fig. 3b, c), which in turn increases the stacking
structures14, indicating a conserved ligand-binding strategy of surface of RU.521 with Arg364 and Tyr421 (Fig. 3d). In addition,
cGAS. The angle between the pyrazole ring and phthalide ring of we observed a water-mediated interaction between the aldehyde
bound RU.365 is approximately 120°, which is mainly caused by group of the phthalide ring and the main-chain of Gly290 and the
the interaction between Cys419 and the hinge region of these two side-chain of Lys350 (Fig. 3e), which is absent in the
rings. The remaining parts of the phthalide ring face the solvent RU.365 structure. The above structural features of RU.521 make
and form no interactions with cGAS. Surprisingly, there is no it a more potent inhibitor than RU.365.
hydrogen bonding interaction between the inhibitor and the
To our surprise, by superposing the structures of inhibitor
protein, despite the presence of four nitrogen and three oxygen complexes with the structure of cGAS in free state, we found that
atoms in RU.365.
RU.365/RU.332/RU.521 could also fit the “smaller active site” of
apo-cGAS in the dsDNA-free state without notable steric effects
(Supplementary Fig. 5). To confirm this we used isothermal
titration calorimetry (ITC) to measure the affinity of RU.365 in
the absence or presence of dsDNA. There was not a large
difference between the affinity of RU.365 with and without
dsDNA, both being between 64.1 nM and 98.5 nM, respectively
(Supplementary Fig. 5).
RU.365 analogs bind to the active site of cGAS. We also solved
the crystal structure of RU281332 (hereafter as RU.332), an
analog of RU.365, in complex with cGAS and dsDNA at 2.18 Å
resolution (Supplementary Figs. 3 and 4; X-ray statistics in
Table 1). The only difference between RU.365 and RU.332 is the
substitution of the benzimidazole ring with a benzothiazole ring;
the two bound inhibitors can be superimposed very well with
each other (Supplementary Fig. 3a, b). We can distinguish S from A structural comparison of cGAS ternary complexes. To
N of the benzothiazole ring by checking the Fo-Fc map during investigate how RU.365/RU.332/RU.521 might inhibit cGAS
refinement. The sulfane group is facing the aromatic/hydrophobic activity, we compared the inhibitor complexes to cGAS-DNA
surface of the binding pocket, while the remaining methanamine in complex with substrate, intermediate, and product.
group is pointing into the solvent region. The interactions When comparing the bound inhibitors with the substrates
between RU.332 and cGAS are very similar to that of RU.365.
(ATP/GTP)14, 16, we observed heavy steric clashes between
To improve the binding affinity and inhibitory activity of inhibitors with both ATP and GTP. The benzimidazole-pyrazole
the lead compound, we set-up a structure-guided chemical and phthalide moieties of RU.365 overlap with the adenine ring
synthesis process. Among many synthesized RU.365 derivatives, of ATP and the α/β-phosphates of GTP, respectively (Fig. 3f).
the compound RU320521 (hereafter RU.521) with a dichloro The inhibitors also show overlapped occupancy with the
substitution (Fig. 3a) showed the best inhibition activity bound covalent intermediate analog 5′-pppG(2,5)pG14, with the
(see below). We determined the crystal structure of benzimidazole-pyrazole moiety clashing against the guanine ring
cGAS-dsDNA-RU.521 ternary complex at 1.83 Å resolution by of pppG and the phthalide ring against the sugar ring of pppG, as
using a co-crystallization approach (X-ray statistics in Table 1). well as the linker phosphate (Fig. 3g). Compared to the substrates
The two modified chlorines in RU.521 target deeper in the pocket and intermediate, the bound cGAMP product14 shows better
6
N
A
TU
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
8
:
7
5
0
D
O
I
:
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
1
7
-
0
0
8
3
3
-
9
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
|
D
O
I
:
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
1
7
-
0
0
8
3
3
-
9
A
R
T
I
C
L
E
superimposition with RU.365 (Fig. 3h). The entire RU.365 modifications of the isobenzofuran-3-one group showed only
molecule overlaps well with the AMP moiety of cGAMP, with minor variations in potency (RU320469 and RU320468,
the benzimidazole ring of RU.365 inserting deeper into the Supplementary Table 3).
pocket than the adenine ring of cGAMP. RU.332 (Supplementary
Importantly, and consistent with our structural model,
Fig. 3c–e) and RU.521 (Fig. 3i–k) show similar superposition modifications that increased the hydrophobic character of
results to that of RU.365. The above comparison suggests the benzimidazole improved potency, suggesting an interaction
that RU.365/RU.332/RU.521 could potentially inhibit every of this moiety with a hydrophobic environment in the target
catalytic step (substrate binding, intermediate and product protein. Substantiating this hypothesis, we found that compounds
formation) of cGAS, or could displace substrates from the in which aromatic hydrogens of the benzimidazole were
catalytic pocket.
replaced by halogens or methyl groups showed increased
To further explore this, we performed ITC-binding experi- inhibitory potency (RU.320519, RU320461, RU320462,
ments titrating RU.365 into the complex cGAS/dsDNA in the RU320520, RU320467, RU320582, and RU.521, Supplementary
presence or absence of ATP or GTP. We found that the affinity of Table 3). A 17-fold improvement in potency was observed
RU.365 decreased in presence of GTP or ATP (64 nM in absence in analogs with 6,7- or 5-7-dichloro substitution in compounds
of GTP vs. 298.2 nM, in the presence of GTP, 104 nM Kd in the RU.521 and RU320582. Additional analogs and their potencies
presence of ATP, Supplementary Fig. 6). Taken together, RU-365 are shown in Supplementary Table 4. Concentration response
appears to impede the binding of both GTP and ATP to the curves of RU.365 and its most potent analog RU.521 are shown
enzyme active site. To determine whether RU.365 showed a in Fig. 4a. ITC experiments were used to confirm the binding
competitive mechanism of inhibition enzyme kinetics, experi- of RU.521 to the cGAS/dsDNA complex (Kd = 36.2 nM, Fig. 4b
ments were performed to examine the effect of inhibitor and Table 2).
app
concentration on Vmax and Km. We found that Km
of both
substrates did not increase in the presence of increasing
concentrations of RU.365. Under the same conditions the Vmax
varied significantly (Supplementary Fig. 7).
Inhibition of dsDNA-induced signaling in macrophage cells.
We next evaluated the activity of the in vitro validated
compounds in cellular assays to determine their efficacy in
suppressing dsDNA-dependent activation of cGAS and
subsequent upregulation of type I interferon genes. Introduction
of dsDNA into RAW macrophage cells leads to marked
upregulation of type-I interferons through activation of IRF3 in a
cGAS and Sting dependent manner6, 11–14, 21. We used RAW cells
that stably express an interferon-responsive element coupled to a
luciferase gene (IFNB1-Luc), enabling us to utilize luciferase
signal as the readout of cGAMP production. We performed a
series of cellular concentration-response analyses to determine
the IC50 values of the small molecules in dsDNA-activated RAW
cells. We observed that the compounds exhibited IC50 values
ranging from 700 nM to 13.60 µM. The cellular IC50 value for
RU.365 (4.98 µM, Fig. 5a) was approximately twofold lower than
RU.332 (13.60 µM, Fig. 5b). The derivatized analog, RU.521, had
an IC50 of 700 nM (Fig. 5c) and was the best at suppressing
dsDNA-activated reporter activity.
Improving RU.365 potency through structure-activity studies.
To examine the structure-activity relationship of cGAS inhibitors
and seek molecules with improved potency, we tested in-house
synthesized and commercially available analogs of RU.365
harboring substitutions of the atoms in the benzimidazole, the
isobenzofuran-3-one, or the central pyrazol-5-ol ring systems. We
performed concentration-response experiments in the RF-MS
assay and derived 50% inhibitory concentration (IC50) values for
each compound. We tested analogs with targeted substitutions in
groups that can function as H-bond donors or acceptors since we
previously observed that replacing the 1-N benzimidazole’s
nitrogen by sulfur (RU.332, Supplementary Table 3) did not
change the potency of cGAS inhibition, whereas ring conversion
from isobenzofuranone to isoindolinone (RU320465) abolished
inhibitory activity (Supplementary Table 4). Furthermore,
Selective inhibition of cGAS-mediated signaling by RU.521. To
evaluate whether RU.521 and its analogs might affect other innate
immune signaling pathways beyond dsDNA activation, we
stimulated RAW macrophage cells with a selection of other
immunogenic ligands. Specifically, we exposed cells to ligands
for RIG-I (5′ppp-HP20 RNA44), Tlr2/1 (Pam3CSK4), Tlr3
(poly(I:C)), Tlr4 (lipopolysaccharide, LPS), and JAK/STAT
signaling (recombinant Ifnb) in the presence or absence of
Table 2 Dissociation constant and thermodynamic
parameters of RU.521
rM
Kd (nM)
ΔH
(cal/mol)
T*ΔS
(cal/mol)
ΔG
(cal/mol)
R
U
.
5
2
1
0
.
80
71
3
6
.
1
7
−1
6
2
9
0
−6
1
3
4
−1 0 1 5 0
a
b
c
1.5
1.0
0.5
0.0
1.5
1.5
1.0
0.5
0.0
1.0
0.5
IC50 = 4.98 μM
–2 –1
IC50 = 13.60 μM
–1
IC50 = 0.70 μM
0.0
0
1
2
3
0
1
2
3
–4 –3 –2 –1
0
1
2
3
log RU.365 (μM)
log RU.332 (μM)
log RU.521 (μM)
Fig. 5
S
m
a
t
o
l
i
l
v
n
-
e
m
n
o
e
n
l
s
e
s
s
c
u
l
e
i
e
i
n
p
n
h
r
h
i
b
i
i
t
n
o
i
o
g
n
c
r
o
G
c
f
A
o
c
S
u
G
a
A
c
S
-
v
t
d
e
y
a
p
i
e
n
n
d
c
d
e
D
e
n
t
N
a
i
A
s
n
t
-
e
s
g
r
t
e
f
e
r
u
e
o
l
.
n
a
T
i
n
d
e
d
u
R
o
c
A
n
t
i
o
n
i
m
t
n
a
c
c
e
l
l
p
n
u
h
l
a
a
e
r
g
s
a
e
p
s
s
o
s
.
n
a
T
s
y
h
e
s
e
c
.
R
U
h
v
.
i
e
3
b
s
6
5
,
R
n
R
U
.
3
t
3
y
6
2
,
a
I
( a
n
I
)
d
R
U
.
s
.
5
p
3
2
o
1
w
e
r
v
,
e
i
a
t
a
n
e
d
d
s
s
t
D
R
e
d
N
.
i
A
5
n
c
w
1
e
a
(c
l
s
l
u
m
)
l
a
o
n
r
a
s
s
a
y
d
c
s
f
v
o
r
t
h
e
i
r
e
f
f
e
c
i
n
i
b
i
t
t
i
i
t
s
f
i
m
t
e
W
c
r
o
i
n
i
t
i
o
o
f
p
e
F
N
r
e
n
s
e
n
i
t
o
r
e
i
a
a
n
i
n
t
e
r
f
e
r
s
e
i
t
i
v
o
m
t
e
p
l
e
d
o
l
u
i
r
e
n
h
c
e
n
r
a
t
i
o
r
u
r
f
o
r
U
.
3
5
,
R
U
3
2
(
b
)
U
2
i
R
A
W
e
l
l
s
a
r
e
s
h
o
w
n
.
E
r
r
o
r
b
a
r
s
r
e
p
r
e
s
e
n
t
S
E
M
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
8
:
7
5
0
D
O
I
:
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
1
7
-
0
0
8
3
3
-
9
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
IO
N
S
|
D
O
I
:
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
1
7
-0
0
8
3
3
-
9
a
b
c
d
e
3.0
2.0
1.0
0.0
1.5
1.0
0.5
0.0
3.0
2.0
1.0
0.0
1.5
1.0
0.5
0.0
3.0
2.0
1.0
0.0
unt
* P ≤ 0.05
DMSO
RU.365
RU.332
RU.521
*
*
*
*
*
*
*
*
*
*
dsDNA
5’ppp-HP20
Pam3CSK4
poly(I:C)
LPS
g
f
h
i
150
100
50
1.5
1.0
0.5
0.0
1.5
1.0
0.5
0.0
1.5
KO-cGAS
RAW 264.7
BMDM
1.0
0.5
0.0
*
*
*
RU.365: 45.17 mM
RU.332: 47.80 mM
RU.521: 61.82 mM
*
*
0
–2
–1
0
1
2
3
WT
DMSO
365332 521
log inhibitor (mM)
Trex1–/–
dsDNA
cGAMP
Ifnb
Fig. 6
P
o
e
t
e
r
n
t
a
n
t
d
s
c
e
e
l
e
c
s
t
i
w
v
e
e
i
n
h
e
i
x
b
p
i
t
o
i
o
s
n
e
o
d
f
t
c
o
G
A
D
S
N
a
A
c
t
i
(a
v
i
)
t
y
,
i
5 ′
n
p
R
A
-
W
m
a
c
R
r
o
N
p
A
h
a
(b
g
e
)
a
P
n
a
d
m
B
3
M
D
K
M
4
c
(c
e
)
l
l
,
s
p
f
o
r
o
m
a
C
n
)
A
(d
i
c
a
r
d
p
i
o
-
G
o
l
u
y
t
s
i
è
a
r
c
e
c
s
h
S
y
n
e
d
r
(
o
L
m
e
)
m
( e
o
)
u
,
s
o
e
r
m
o
c
d
o
e
m
l
.
R
A
a
W
l
u
c
i
f
e
r
a
s
e
p
o
r
e
r
l
l
r
e
d
s
p
p
H
P
2
0
,
C
S
l
y
(
I
:
)
,
l
i
p
o
a
r
i
d
P
S
r
e
b
i
n
n
t
m
u
r
u
i
n
e
a
i
n
n
e
q
t
o
R
e
t
e
u
T
f
y
t
t
r
s
f
e
r
o
t
R
s
s
e
B
n
-
β
(
I
f
n
b
)
f
t
i
O
o
n
-
p
r
o
m
o
t
s
e
d
a
c
p
e
m
p
e
a
i
t
l
t
h
l
e
m
r
a
t
c
y
u
w
t
A
o
p
l
e
e
e
d
T
I
(
i
o
n
r
t
v
e
e
p
A
e
n
r
f
h
e
r
o
n
)
t
a
(
a
)
y
s
,
p
(
b
)
,
i
(
f
)
r
r
a
a
f
,
o
r
r
A
o
N
F
r
-
κ
B
p
n
(
c
)
s
s
o
,
e
i
( d
w
)
a
,
(
e
)
r
e
s
p
o
n
s
e
u
n
d
e
r
d
i
f
f
e
r
e
n
t
i
m
m
u
n
e
s
t
i
m
u
l
i
a
n
d
s
i
m
l
t
l
y
r
e
a
t
e
d
w
i
t
h
d
i
c
a
t
e
S
d
r
o
l
e
i
c
l
e
.
T
e
I
n
t
e
f
e
o
n
e
s
o
n
s
r
e
a
o
d
u
n
v
i
a
t
h
n
l
r
e
t
u
e
c
a
i
t
f
e
n
b
e
d
g
i
r
a
w
e
t
s
e
r
h
s
f
e
e
p
a
d
e
o
c
r
h
i
t
e
o
r
f
t
,
w
h
e
d
i
l
s
e
m
s
M
N
a
e
F
- κ
m
B
o
o
s
r
e
c
s
o
s
u
e
p
l
o
e
n
s
s
. h
e
w
a
e
s
r
e
a
d
v
o
i
a
-
e
w
e
P
c
C
t
a
t
. g
o
m
d
R
A
W
K
c
G
A
p
e
t
o
r
e
r
c
e
l
l
s
r
e
e
x
o
W
s
e
d
m
e
e
o
d
D
N
A
g
a
o
s
c
G
c
M
P
a
d
m
u
l
t
a
n
e
s
g
l
y
t
i
t
t
h
l
l
l
e
T
h
.
c
y
t
o
t
x
x
i
c
f
f
t
h
e
l
e
a
d
c
o
m
o
u
n
u
s
w
e
t
o
n
r
e
t
e
s
f
f
M
e
i
P
4
n
(
h
R
C
c
r
o
p
h
a
e
t
n
c
e
n
t
r
a
t
i
n
s
s
p
a
n
n
i
r
a
t
y
e
o
t
e
n
h
e
o
D
r
e
s
p
n
c
u
r
v
e
s
−
/
−
C
y
t
o
e
x
t
o
i
e
i
c
i
e
a
s
u
r
e
d
b
y
d
t
h
q
n
o
r
t
i
t
a
i
o
n
n
o
d
l
l
T
i
t
r
-
G
l
o
a
a
s
s
y
)
t
7
2
h
a
n
d
C
s
h
o
n
w
n
a
s
.
p
S
e
r
c
e
v
i
a
l
i
c
l
l
s
.
i
B
M
c
e
l
l
f
r
m
T
r
e
x
1
m
i
c
w
r
e
r
y
a
e
w
i
t
h
t
h
e
i
n
i
c
a
e
d
m
p
u
t
r
2
a
d
t
h
n
h
r
v
e
s
t
e
d
o
r
q
R
T
-
P
R
a
a
l
y
s
i
s
h
o
w
n
a
r
e
t
h
e
r
e
s
u
l
t
s
,
r
e
l
a
t
i
v
e
t
o
I
F
N
B
1
e
x
p
r
e
s
s
i
o
n
i
n
T
r
e
1
w
l
d
-
p
M
D
M
s
.
E
r
r
o
r
b
a
r
s
r
e
e
s
e
S
E
each small molecule, and compared their ability to suppress RU.521 is active in murine bone marrow-derived macrophages.
these immunogenic stimuli (Fig. 6a–f). As compared to its The phenotype and elevated cytokine expression levels observed
ability to inhibit dsDNA-dependent reporter activation, RU.521 in Trex1−/− knockout mice reflect the clinical presentation of
was unable to potently suppress activation of cells by essentially some human autoimmune disorders—notably AGS and Chilblain
all the immunogenic stimuli tested (Fig. 6a–f). These lupus where mutations within human TREX1 render the primary
results differed from what we observed with RU.365 and RU.332. cellular DNA 3′-to-5′ exonuclease non-functional32, 33. Previous
We found that RU.365 led to increased Il-6 messenger groups have shown that Trex null mice lacking cGAS, Sting,
RNA (mRNA) expression in cells activated by Pam3CSK4, or IRF3 expression develop normally and bear no phenotype
poly(I:C), or LPS (Fig. 6c–e); none of the compounds associated with interferonopathy35–37. Thus, the chronically
caused reporter activation on their own (Fig. 6a). RU.332 inhi- elevated levels of cytokines observed in Trex1 null mice are a
bited the activation of cells by most of the immunogenic ligands consequence of constitutively activated cGAS, due to the inability
tested.
to eliminate aberrantly localized self-DNA. We harvested
To further assess selectivity, we used RAW cells that do not BMDMs from 6–8-week old Trex1−/− mice, treated them with
express cGAS (KO-cGAS) and stimulated them with dsDNA or each compound, and measured expression levels of IFNB1 by
cGAMP. Although the introduction of dsDNA into these cells do quantitative reverse transcription PCR (qRT-PCR) (Fig. 6i).
not stimulate reporter activity given the lack of cGAS, treatment Treatment of primary BMDMs with RU.521 or its analogs
of these cells with cGAMP leads to activation via Sting reduced IFNB1 expression, indicating their effectiveness in
stimulation. Under conditions in which the dsDNA pattern suppressing intrinsic DNA-dependent, constitutively-activated
recognition pathway is stimulated by using cGAMP, we observed type I interferon expression in cells deficient of a cytoplasmic
that RU.521 did not show any significant inhibition of reporter DNA exonuclease.
activity (Fig. 6g)—which is distinct from its activities in cells
exposed to dsDNA when cGAS is present.
The inhibitory effects of RU.521 and RU.365 on cGAS
signaling does not appear to be a consequence of cytotoxicity
since we only observed ≥ 50% cellular viability loss at concentra-
tion levels significantly over their cellular IC50 values (Fig. 6h).
The least toxic compound was RU.521, where we measured
an LD50 value of 62 µM (88-fold), followed by RU.365 (45 µM,
9-fold), then RU.332 (48 µM, 3.5-fold).
Discussion
Using an in vitro mass-spectrometry based high-throughput
compound screen, we initially identified RU.365 and its
benzothiazole analog RU.332 as inhibitors of the dsDNA-induced
enzymatic activity of cGAS. The generation of cGAMP by cGAS
requires that the enzyme bind to dsDNA and use its two
substrates, ATP and GTP, to generate the cyclic dinucleotide.
8
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
8
:
7
5
0
D
O
I
:
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
1
7
-
0
0
8
3
3
-
9
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
|
D
O
I
:
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
1
7
-
0
0
8
3
3
-
9
A
R
T
I
C
L
E
Thus, an inhibitor could potentially be found to disrupt dsDNA selectivity, and nominal cytotoxicity. As it was a compound that
binding, block ATP and/or GTP from entering the active site, or was specifically designed to be a more potent inhibitor of cGAS,
somehow inhibit the generation of either phosphodiester linkage our results underscore the value of a structurally-guided and
to prevent cyclization of cGAMP. Our ITC experiments interdisciplinary approach to rational drug design. Nonetheless, it
demonstrated that the presence of RU.365 reduced the binding should be noted that while we did test a reasonable number of
affinity of cGAS for either GTP or ATP, whereas the kinetic inflammatory pathways to assess the selectivity of these small
experiments indicated that RU.365 only affected the enzyme’s molecules, it is naturally a limited set. Further testing in whole
Vmax. These findings suggest that RU.365 is a non-competitive animals and subsequent pharmacokinetic optimization will likely
inhibitor against ATP or GTP. Noncompetitive inhibition is reveal the full extent of potential off-target effects and toxicities of
common in multi-reactant systems and can occur with active site RU.521. These in turn would point to where improvements can
inhibitors of multi-substrate enzymes for a variety of reasons, be made—and what the therapeutic index might be, which would
including: (1) the presence of exosites, (2) rate-limiting step not be an unusual course toward the eventual development of a
isomerization of the catalytic site, (3) two-step mechanisms, and clinically relevant derivative.
(4) bisubstrate/byproduct enzymes that must follow an ordered
cGAS has emerged to be an essential protein for the innate
substrate binding or product release45, 46. In the case of cGAS, it immune response to cytosolic DNA given its activation of Sting
is unlikely that an exosite exists because the substrates are by the production of cGAMP. Like Sting, cells that lack cGAS fail
not high molecular weight polymers such as DNA or protein. to substantially upregulate IFNB1 expression in response to
However, the other three mechanisms, or a combination thereof, infection by bacterial, viral, and eukaryotic pathogens27, 47, 48, or
remain a possibility.
to the accumulation of self-DNA as in the case of knockout
Our previous structural studies have shown that upon dsDNA mouse models of AGS49–52. Nevertheless, there are other DNA
binding, cGAS is activated through conformational transitions sensing proteins5, 48, and significant questions remain as to why
resulting in the formation of a catalytically competent and multiple sensors exist. Although cell-type specific expression of
accessible nucleotide-binding pocket for generation of cGAMP14. sensors might provide some clues, cGAS is found co-expressed in
An analysis of the ternary complexes between cGAS, dsDNA and cells with other DNA sensors such as Interferon gamma inducible
either compound indicated that the inhibitors occupy only one protein 16 (IFI16, IFI204 in M. musculus), which partly
side of the active pocket, similar to bound cGAMP. Arg364 and signals through STING53. While IFI16 is partially found in the
Tyr421 are two highly conserved amino-acid residues found in cytoplasm, it is predominantly in the nucleus where it can act as a
mouse and human cGAS, and the stacking interaction mediated transcription co-regulator48. Interestingly, there are an increasing
by these residues appears critical for inhibitor binding and also number of reports that indicate that cGAS can interact with
represents a conserved strategy used by different cGAS ligands IFI16 under various cellular and pathogenic contexts54–56. The
(substrates/intermediate/product). Given the position of the underlying mechanisms that dictate the manner in which cGAS
inhibitors within the larger active pocket and its interaction with cooperates with other DNA sensors like IFI16 remains an actively
Arg364 and Tyr421, we reasoned that designing a compound that researched area—one that can benefit from the use of selective
could fill the rest of the cGAS pocket and further stabilize the cGAS inhibitors.
interaction with the two amino acids might lead to a more potent
In summary, we have discovered and characterized a small
inhibitor. From a series of chemically synthesized derivatives, we molecule that shows potent and selective inhibition of cGAS
subsequently identified RU.521 as exhibiting nanomolar cGAS dsDNA-dependent enzymatic activity in vitro and in cells,
inhibitor activity in vitro. The ternary complex with cGAS, including primary macrophages taken from a Trex1 null mouse
dsDNA, and RU.521 showed that the compound had an model of AGS autoimmunity. Based on our structural and kinetic
increased stacking interaction with Arg364 and Tyr421 due to the studies, RU.521 can occupy the catalytic site of cGAS and reduce
orientation of the benzimidazole ring as well as a water-mediated its affinity for ATP and GTP. Moreover, it is likely that RU.521
interaction between RU.521, Gly290, and Lys350—which was not does not interfere with dsDNA directly since an acridine orange
observed with RU.365.
competition screen did not show any indication that it could act
Using mouse cell lines and primary cells from Trex1−/− mice to in a manner that could intercalate or interact with DNA. RU.521
evaluate the lead compounds, we determined that RU.521 was the will be a useful experimental chemical probe toward deciphering
most potent inhibitor of dsDNA-induced immune activation. cGAS biology and with further derivatization, may prove to be an
Using
a panel of pattern recognition receptor ligands, instrumental structural scaffold toward the development of an
RU.521 suppressed dsDNA-dependent activation and showed immunomodulatory therapeutic agent in treating cGAS-related
little to no inhibitory effects for the other pathways. Importantly, human disorders.
we demonstrated that the inhibition of the dsDNA pathway by
RU.521 requires the presence of cGAS, as the small molecule did
Methods
not have inhibitory activity in KO-cGAS cells under conditions in
which the dsDNA sensing pathway is stimulated using cGAMP to
directly activate Sting. Moreover, RU.521 did not inhibit cells
directly stimulated with recombinant Ifnb, indicating that it does
not appreciably target IFNB1 protein, interferon receptors, and/or
downstream signaling components of the JAK/STAT pathway.
Curiously, we observed that treatment of cells with RU.365,
particularly in conjunction with Tlr stimulation, led to the
activation of Il-6 expression. While we cannot explain the
molecular basis of these RU.365 results, presumably off-target
activity, none of these effects were observed with the improved
compound RU.521.
Chemical libraries. A total of 123,306 compounds were screened in the primary
screen. Compounds were selected from vendor files using fingerprint-based
(FCFP6) clustering algorithms with Biovia Pipeline Pilot software (http://accelrys.
com/products/collaborative-science/biovia-pipeline-pilot). The average cluster size
was set to 30 and cluster centers were identified and chosen for purchase. For most
of the collections, PAINS were removed. However, the non-drug libraries and
natural product collections contain a large number of PAINS that were permitted.
Selected compounds were purchased from the following vendors: ChemDiv (7022),
San Diego, CA; Spectrum (1360), New Brunswick, NJ; Prestwick (471), Illkirch,
France; Enamine (53,108), Monmouth Junction, NJ; Pharmakon-900 (905),
(MicroSource, Gaylordsville, CT); LifeChem (11,218), Life Chemicals, Niagara-on-
the-Lake, Canada; Specs (4051), Specs, Zoetermeer, The Netherlands; Chiral Center
Diversity Library (3289), NIH Clinical Collection (727); Tocris (480), Bristol, UK;
Biofocus (4416), Charles River, Wilmington, MA; HTSRC Clinical Collection
(294), NIH Small Molecule Repository; Cerep (640), Cerep, Poitiers, France;
Chembridge (34,134), ChemBridge, San Diego, CA; AMRI (2833), AMRI, Albany,
NY; Analyticon (352), AnalytiCon discovery, Potsdam, Germany; ChemX (983).
Taken together, we find that RU.521 exhibited the most
optimal characteristics among this class of compounds, having
potent nanomolar activity in vitro and in cells, improved One-thousand sixty-eight compounds were screened in the pilot study using Lopac
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
8
:
7
5
0
D
O
I
:
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
1
7
-
0
0
8
3
3
-
9
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
|
D
O
I
:
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
1
7
-
0
0
8
3
3
-
9
follows: Z′ = 1−[3∗(standard deviation positive control + standard deviation
negative control)/(average positive control—average negative control)].
library, Sigma, Carlsbad, CA. All compounds were dissolved in DMSO at 5 mM
and stored at −28 °C.
Cheminformatics and data handling. Data from all screening studies performed
at The Rockefeller University are acquired and analyzed using the CDD Vault from
collaborative Drug discovery (Burlingame, CA. www.collaborativedrug.com).
MarvinSketch (ChemAxon, Budapest, version 14.9.8.0) was used for drawing and
naming the structures. Pipeline pilot was used to search for analogs in various
commercial databases.
Preparation of dsDNA for activity assays of mouse cGAS. Non-phosphorylated
oligodeoxynucleotides were purchased from IDT and full-length content was
verified by separation of an aliquot on urea-containing 10% polyacrylamide gels
followed by ultraviolet shadowing. Fully complementary strands were annealed at
200 μM strand concentration in buffer containing 20 mM Tris-HCl, pH 7.5, and
150 mM NaCl by first incubating at 95 °C for 90 s followed by slow cooling to 25 °C
at a rate of 0.1 °C per second in a Peltier thermocycler (BioRad). Complete
annealing was verified by electrophoresis using 2.5% low melting temperature
agarose gel (Life Technologies, 15517-014) containing ethidium bromide and
comparing the mobility with the single-stranded oligodeoxynucleotides. The
sequences used for annealing can be found in Supplementary Table 5.
DNA intercalators. Validated compounds of the primary screen were tested in a
high-throughput FPassay for their capacity to intercalate DNA following the
protocol developed at the Broad Institute (National Center for Biotechnology
Information. PubChem BioAssay Database; AID = 504727, https://pubchem.ncbi.
nlm.nih.gov/bioassay/504727 (accessed 27 April, 2016)). The assay was performed
in a final volume of 30 µl in 384-solid bottom opaque plates. Ten microliters of
HEN buffer (10 mM HEPES pH 7.5, 1 mM EDTA pH 7.5, 100 mM NaCl) were
dispensed per well using a Thermo Multidrop Combi dispenser (Thermo
Scientific). The liquid was collected at the well bottom using centrifugation for 30 s
at 180×g. Compounds were dissolved in DMSO and 0.18 µl at 5 mM were
dispensed with a Janus 384 MDT NanoHead (Perkin Elmer). The final
concentration of compounds in the assay was 30 µM. Ten microliters of a solution
of 150 nM acridine orange in HEN buffer was dispensed per well using a Thermo
Multidrop Combi dispenser (Thermo Scientific). The liquid was collected at the
well bottom using centrifugation for 30 s at 180×g. Subsequently, 10 µl of a solution
of 45-bp dsDNA at 37.5 µg ml−1 was dispensed per well using a Thermo Multidrop
Combi dispenser (Thermo Scientific). The liquid was collected at the well bottom
using centrifugation for 30 s at 180 × g and the plates were incubated for 30 min.
Mitoxantrone, a known DNA intercalator was used at 50 µM as a positive control,
while DMSO alone was used as negative control. FP was measured using a Biotek
Synergy Neo plate reader. Samples were excited by a flash Xenon lamp filtered by a
485 nm excitation filter, collected 10 mm from the top of the well, and fluorescence
emission was detected through a 530 nm filter, taking ten measurements per data
point. All FP values are expressed in millipolarization (mP) units. The mP values
were calculated using the equation mP = 1000 × [(IS−ISB)−(IP−IPB)]/[IS—ISB) +
(IP−IPB)], where IS and ISB refer to the parallel and perpendicular emission
intensity, respectively, and ISB and ISP corresponds to parallel emission intensity
perpendicular emission intensity of the buffer, respectively58. The experimental G-
factor, defined as G = IS/IP, was set to 1.0 in all experiments.
High-throughput screening. Screening reactions were carried out in 20 µl volumes
in 384 small volume deep well polypropylene plates. The final concentration of
cGAS enzyme, dsDNA, ATP, and GTP were 60 nM, 300 nM, 300 µM, and 300 µM,
respectively. Five microliters of reaction buffer composed of 20 mM Tris-HCl
pH 7.4, 150 mM NaCl, 5 mM MgCl2, 1 mM dithiothreitol (DTT), and 0.01%
Tween-20 were dispensed per well using a Thermo Multidrop Combi dispenser
(Thermo Scientific). The liquid was collected at the well bottom using cen-
trifugation for 30 s at 180×g. Compounds were dissolved in DMSO and 0.05 µl of 5
mM were dispensed with a Janus 384 MDT NanoHead (PerkinElmer). Final
concentration of the compounds in the assay was 12.5 µM. A concentration of 0.5%
DMSO did not interfere with cGAMP production from recombinant cGAS. Next,
10 µl of a master mix containing reaction buffer supplemented with 0.6 mM ATP,
0.6 mM GTP, and 0.6 µM dsDNA was added to wells in columns 1–23 using a
Thermo Multidrop Combi dispenser, while 10 µl of the master mix devoid of
dsDNA (control for no enzymatic activity) was added to wells in column 24. Plates
were centrifuged for 30 s at 180×g to collect all liquid at the bottom of the wells.
The reaction was started by adding 5 µl of 0.24 µM recombinant full-length mouse
cGAS in reaction buffer to each well of the plate followed by centrifugation for 30 s
at 180×g and incubation for 120 min at room temperature. The reaction was
stopped by addition of 65 µl of 0.5% (v/v) formic acid per well. The plates were
centrifuged for 30 s at 180×g and sealed with a Velocity11 PlateLoc thermal plate
sealer. Compounds that inhibited cGAS activity by ≥ 60% were retested in con-
centration response experiments to determine half maximal inhibitory con-
centration (IC50). Compounds were serially diluted by half for a total of ten
dilutions where the highest final concentration in the assay was 25 µM. The
compounds selected for follow-up studies were reordered from the vendors, dis-
solved in DMSO to a concentration of 10 mM and re-tested in concentration
response experiments. The IC50 values for the enzymatic assay were calculated
using GraphPad Prism (7.01), from three replicate experiments; the error bars
represent SD.
Kinetics parameter of mcGAS in presence of RU.365. Kinetics parameters
of cGAMP formation were determined in the reaction mixtures carried out in
20 μl volumes in 384 small volume deep well polypropylene plates. The final
concentration of m-cGAS enzyme and dsDNA were 60 nM and 300 nM,
respectively. To determine if RU.365 was competitive against ATP and/or GTP,
different concentrations of RU.365 ranging from 0.8 µM to 6.3 µM were tested,
app
keeping one substrate fixed at 300 µM while varying the other one. Apparent Km
and apparent Vmax
app
values were also calculated in the presence and absence of
RapidFire 365 mass spectrometry high-throughput assay. All samples were
analyzed using a RapidFire 365 high-throughput solid phase extraction system
integrated with a 6230 TOF mass spectrometer (Agilent Technologies, Wakefield,
MA, USA). The solvent used for loading/washing process was an aqueous solution
of 5 mM ammonium acetate, pH 10. For elution of the analytes, the organic solvent
used was 50% water, 25% acetone, and 25 % acetonitrile containing 5 mM
ammonium acetate, pH 10. Each sample of approximately 35 µl were aspirated
from a 384-well plate and separated using a Graphitic carbon Type D
cartridge. After each sample was loaded onto the cartridge it was washed for 4 s at
1.5 ml min−1 using the aqueous solvent. ATP, GTP and cGAMP were eluted for 5 s
using the organic solvent at a flow rate of 1.5 ml min−1. The cartridge was then
re-equilibrated with the aqueous solvent for 5 s at a flow rate of 1.5 ml min−1. All
samples were analyzed using a negative ionization mode in the mass spectrometer,
with a gas temperature of 350 °C, nebulizer pressure of 35 psig, gas flow rate of
15 l min−1. The acquisition range of all chromatograms was between 300 and 800
m/z and the molecular masses of the peaks detected were the following: ATP:
505.9835, GTP: 521.9854 and cGAMP: 673.0906. The Agilent RapidFire Integrator
software was used to calculate the area under the curve (AUC) of the extracted ion
counts for each analyte. Data points represent values of product formation.
Catalytic rates expressed as percent product formation were calculated using the
AUC of cGAMP and using the AUC’s of ATP and GTP to normalize for variations
in the capacity of the solid phase extraction column during a screening run, as
shown in the following formula: product formation (%) = [(AUCcGAMP × 100)/
(AUCcGAMP + ½ AUCATP + ½ AUCGTP)].
the small molecule. RU.365 was diluted by half in DMSO for a total of four
dilutions where the highest final concentration in the assay was 6.3 μM. Ten
microliters of reaction buffer composed of 20 mM Tris-HCl pH 7.4, 150 mM NaCl,
5 mM MgCl2, and 0.01% Tween-20 were dispensed per well using a Thermo
Multidrop Combi dispenser (Thermo Scientific). The liquid was collected at the
well bottom by centrifugation for 30 s at 180×g. In all, 0.1 μl of RU.365 serially
diluted in DMSO were dispensed with a Janus 384 MDT NanoHead (PerkinElmer).
A concentration of 0.5% DMSO did not interfere with cGAMP production from
recombinant cGAS.
Each reaction containing a variable substrate concentration was performed in a
different plate. A separate column containing different amounts of cGAMP was
also included in each plate in order to construct the calibration curve and calculate
the µmoles of product formed. Twenty different master mix solutions were
prepared in reaction buffer supplemented with 2 mM DTT and 1.2 µM dsDNA. Of
these, Ten master mix solutions were prepared with different concentrations of
GTP ranging from 2.36 µM to 1.2 mM while keeping the concentration of ATP
fixed at 1.2 mM and vice versa for the other 10 master mix solutions. Five
microliters of these solutions with variable substrate concentration were added for
each reaction. Finally, the reaction was started by adding 5 μl of 0.24 μM
recombinant full-length mouse cGAS in reaction buffer supplemented with 2 mM
DTT to each well of the plate followed by centrifugation for 30 s at 180×g and
incubation for 120 min at room temperature. The reaction was stopped by addition
of 60 μl of 0.5% (v/v) formic acid per well. The plates were centrifuged for 30 s at
180×g and sealed with a Velocity11 PlateLoc thermal plate sealer. To calculate IC50
log (inhibitor) vs. response, a variable slope non-linear fit from GraphPad Prism
(7.01) was used, using three replicate experiments; error bars represent SD. The
kinetics parameters were obtained fitting the curves to an allosteric sigmoidal
nonlinear fit using the same software.
The EC50 of dsDNA (six replicates) was calculated using GraphPad Prism
(7.01); error bars represent SD. The time course of mouse and human cGAS assays
were plotted using GraphPad Prism (7.01), with three replicates each; error bars
represent SD. To determine the percent inhibition of mouse cGAS the data was
normalized against the positive control (column 24) and negative control
(column 23). The % inhibition is calculated as follows: % inhibition = 100 ×
(sample—average negative control)/(average positive control—average negative
control)]. The quality of the screen was assessed by Z′ factor57 calculated as
Isothermal titration calorimetry. ITC measurements were performed at 25 °C
using a MicroCal auto-iTC200 calorimeter (MicroCal, LLC). Purified truncated
10
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
8
:
7
5
0
D
O
I
:
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
1
7
-
0
0
8
3
3
-
9
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
|
D
OI
:
1
0.
1
0
3
8
/
s
4
1
4
6
7
-
0
1
7
-
0
0
8
3
3
-
9
A
R
T
I
C
L
E
mouse cGAS was dialyzed against 40 mM HEPES buffer (pH 7.5) and 300 mM
NaCl for 24 h at 4 °C. Dialysis led to some precipitation, which were removed
through centrifugation followed by collection of the supernatant. The concentra-
tion of the protein in supernatant was measured using Pierce® BCA protein assay.
The protein was flash frozen in liquid N2 and stored at −80 °C. For the ITC assay,
the dialyzed protein was prediluted to 20 µM in the dialysis buffer and diluted to a
final concentration of 10 µM in a final buffer containing 40 mM HEPES, 1%
DMSO, 150 mM NaCl, 0.01% Tween-20 and + /− dsDNA 10 µM and + /− ATP or
GTP 1 mM. Then, 2 μl RU.365 100 µM, dissolved in the buffer aforementioned,
were injected into 0.2 ml of protein in the chamber every 150 s. Data for raw ITC
and thermodynamic curves, each from one experiment, were downloaded after
analysis using Affinimeter software and plotted using GraphPad Prism (7.01). The
software calculates error bars as an indication of the contribution of the raw data
noise that is determined from the integral of the standard deviation of the baseline
during the injection.
(ethyl 3-oxo-2-(3-oxo-1H-isobenzofuran-1-yl) butanoate (126.86 mg, 483.74 umol,
1.05 eq) in one portion at 25 °C under N2. The mixture was stirred at 25 °C for
12 h, then heated to 50 °C stirred for another 3 h. The reaction mixture was diluted
with water (20 ml) and extracted with EtOAc (3 × 50 ml). The combined
organic layers were washed with brine (50 ml), dried over Na2SO4, filtered and
concentrated under reduced pressure to give a residue. The residue was purified by
prep-HPLC (FA) to give 3-[1-(6,7-dichloro-1H-benzimidazol-2-yl)-5-hydroxy-3-
methyl–pyra zol-4-yl]-3H-isobenzofuran-1-one (37.00 mg, 19.27% yield, 99.617%
purity) as a white solid.
LCMS: (M + H+) : 415.1 @ 2.887 min (WUXIAB10, 4.5 min).
1H NMR: (DMSO-d6, 400 MHz) δppm 7.83 (d, J = 7.7 Hz, 1 H), 7.75–7.67 (m, 1
H), 7.60–7.49 (m, 2 H), 7.42–7.35 (m, 1 H), 7.33–7.27 (m, 1 H), 6.58 (s, 1 H), 2.19
(br s, 3 H).
Protein for high-throughput screening and crystallization. The gene
encoding mouse cGAS was purchased from Open Biosystems Inc. The sequences
corresponding to full-length (for high-throughput screening) and residues 147–507
(for structural study) of cGAS were inserted into a modified pRSFDuet-1 vector
(Novagen), in which cGAS was separated from the preceding His6-SUMO tag
by an ubiquitin-like protease (ULP1) cleavage site. The gene sequences were
subsequently confirmed by sequencing. The fusion proteins were expressed in BL21
(DE3) RIL bacteria (Agilent Technologies). The cells were grown at 37 °C until
OD600 reached approximately 0.6. The temperature was then shifted to 18 °C and
the cells were induced by addition of isopropyl β-D-1-thiogalactopyranoside to the
culture medium at a final concentration of 0.3 mM. After induction, the cells were
grown overnight. The fusion protein was purified over a Ni-NTA affinity column.
The His6-SUMO tag was removed by ULP1 cleavage during dialysis against buffer
containing 40 mM Tris-HCl pH 7.5, 0.3 M NaCl, 1 mM DTT. After dialysis, the
protein sample was further fractionated over a Heparin column, followed by gel
filtration on a 16/60 G200 Superdex column. The final sample of cGAS (full-length)
and cGAS (147–507) contain about 30 mg ml−1 protein, 20 mMTris-HCl pH 7.5,
300 mMNaCl, 1 mM DTT. The yield of final cGAS protein is ~ 0.7 mg (full-length)
and ~ 1.2 mg (aa 147–507) from 1 l bacteria medium.
Synthesis of RU320521. Please see Supplementary Fig. 8 for an illustration of the
reaction scheme that was used to synthesize RU320521 (RU.521).
2, 3-dichloro-6-nitro-aniline (2). A mixture of 1, 2, 3-trichloro-4-nitro-benzene
(5.00 g, 22.08 mmol, 1.00 eq) and MeOH-NH3 (50 ml) was heated in a steel
reaction vessel at 120 °C and stirred for 24 h. The mixture was filtered. The filter
cake was washed with a mixture of Petroleum ether and EtOAc (15:1, 300 ml). The
solid was concentrated in vacuum, and the residue was washed with a mixture of
PE: EA (15:1, 300 ml) to give 2, 3-dichloro-6-nitro-aniline (4.1 g, crude product) as
yellow solid.
1H Nuclear magnetic resonance (NMR): (METHANOL-d4, 400 MHz) δppm 8.08
(d, J = 9.2 Hz, 1 H), 6.86 (d, J = 9.2 Hz, 1 H).
3, 4-dichlorobenzene-1, 2-diamine (3). To a solution of 2, 3-dichloro-6-nitro-
aniline (2.20 g, 10.63 mmol, 1.00 eq) in MeOH (44.00 ml) was added a solution of
NH4Cl (568.47 mg, 10.63 mmol, 371.55 ul, 1.00 eq) in H2O (2 ml). Then Zn (3.48 g,
53.15 mmol, 5.00 eq) was added slowly at 25 °C under N2. The mixture was stirred
at 25 °C for 2 h. The mixture was filtered through celite, and the residue was
washed with EtOAc (3 × 150 ml). The combined organic layers were concentrated
in vacuum to get the crude product. The crude product was purified by column
chromatography (Petroleum ether: Ethyl acetate = 50:1 to 2:1) to give 3, 4-
dichlorobenzene-1, 2-diamine (1.90 g, crude) as blank brown solid.
Crystallization. The co-crystallization approach employed for murine cGAS
(147–507), in complex with dsDNA and small-molecule inhibitors (RU.365,
RU.332, and RU.521), was performed largely as previously described14. Briefly,
samples were prepared by directly mixing protein with a 16-bp DNA (containing
1-nt 5′-overhang at either end) in a 1:1.2 molar ratio, and with inhibitors in a final
concentration of 1 mM. The crystals of cGAS-dsDNA-RU.365 and cGAS-dsDNA-
RU.332 were generated by hanging drop vapor diffusion method at 20 °C, from
drops mixed from 1 μl of cGAS-dsDNA-inhibitor solution and 1 μl of reservoir
solution (0.1 M MES, pH 6.3, 26% PEG400, 0.1 M MgCl2). The crystals of cGAS-
dsDNA-RU.521 were generated by sitting drop vapor diffusion method at 20 °C,
from drops mixed from 0.2 μl of cGAS-dsDNA-inhibitor solution and 0.2 μl of
reservoir solution (0.1 M MES, pH 6.9, 22.5% PEG400, 0.08 M MgCl2).
1H NMR: (CHLOROFORM-d, 400 MHz) δppm 6.78 (d, J = 8.4 Hz, 1 H), 6.56 (d, J
= 8.6 Hz, 1 H), 3.91 (br s, 2 H), 3.40 (br s, 2 H).
4, 5-dichloro-1, 3-dihydrobenzimidazol-2-one (4). To a solution of 3, 4-
dichlorobenzene-1, 2-diamine (1.90 g, 10.73 mmol, 1.00 eq) in THF (50.00 ml) was
added pyridine (1.70 g, 21.46 mmol, 1.73 ml, 2.00 eq). CDI (3.48 g, 21.46 mmol,
2.00 eq) in DCM (50.00 ml) was dropwise. The mixture was stirred at 25 °C for
12 h. The mixture was concentrated in vacuum to get crude solid. The solid
was washed with Petroleum ether: Ethyl acetate = 15:1 several times to afford
4, 5-dichloro-1, 3–dihydrobenzimidazol– 2-one (1.70 g, 78.03% yield) as
white solid.
Structure determination. The diffraction data sets for cGAS-dsDNA-inhibitor
ternary complexes were collected at the Advanced Photo Source at the Argonne
National Laboratory. The diffraction data were indexed, integrated and scaled
using the NECAT RAPD online server, and the structures of cGAS-dsDNA-inhi-
bitor ternary complexes were solved using molecular replacement method in
PHASER_ENREF_259 using our previous cGAS-dsDNA binary complex
structure14 (RCSB code: 4K96) as the search model. Model building and structural
refinement was carried out using COOT60 and PHENIX61 software, respectively.
The statistics of the data collection and refinement are shown in Supplementary
Table 3.
1H NMR: (DMSO-d6, 400 MHz) δppm 11.34 (s, 1 H), 11.00 (s, 1 H), 7.12
(d, J = 8.2 Hz, 1 H), 6.87 (d, J = 8.4 Hz, 1 H).
2, 6, 7-trichloro-1H-benzimidazole (5). A flask was charged with 4, 5-dichloro-1,
3-dihydrobenzimidazol-2-one (500.00 mg, 2.46 mmol, 1.00 eq) and POCl3 (24.75 g,
161.43 mmol, 15.00 ml, 65.62 eq) in one portion and was stirred at 110 °C for 12 h.
Liquid chromatography-mass spectrometry (LCMS) showed one major peak with
desired MS was detected. The reaction mixture was concentrated at 50 °C to afford
a residue. The residue was washed with NaHCO3 aqueous (50 ml) and extracted
with CH2Cl2 (3 × 80 ml). The combined organic layers were dried with Na2SO4 and
concentrated in vacuum to give 2, 6, 7-trichloro-1H-benzimidazole (540.00 mg,
2.44 mmol, 99.19% yield) as white solid.
Cell lines. Mouse RAW macrophages (RAW-Lucia ISG, RAW-Lucia ISG-KO-
TREX1, RAW-Lucia ISG-KO-cGAS, and RAW-Blue Invivogen) were cultured in
DMEM with high glucose, L-glutamine, phenol red, and sodium pyruvate (Life
Technologies) supplemented with 10% FBS, 100 U ml–1 penicillin-streptomycin,
Normocin (100 µg ml–1, Invivogen), and Zeocin (200 µg ml–1, Invitrogen) and an
additional 20 mM L-glutamine and 1 mM sodium pyruvate. RAW-Lucia cells stably
expressed an interferon sensitive response element from the mouse Isg54 minimal
promoter and five interferon-stimulated response elements (Isre-Isg54) coupled to
a synthetic coelenterazine-utilizing luciferase. RAW-Blue cells have a chromosomal
integration of a secreted embryonic alkaline phosphatase (SEAP) reporter construct
inducible by nuclear factor-κB (NF-κB) and AP-1. Wild-type and Trex1−/− null
derived BMDMs were maintained in RPMI 1640 with L-glutamine and phenol red
(Life Technologies) supplemented with 10% FBS, 10 mM HEPES, 1× non-essential
amino acids, 1 mM sodium pyruvate, 0.055% 2-mercaptoethanol, 100 U ml−1
penicillin-streptomycin, 10 ng ml−1 of recombinant murine M-CSF (PeproTech),
and an additional 2 mM L-glutamine. All cell lines were routinely checked for
mycoplasma contamination using MycoAlert (Lonza).
LCMS: (M + H+) : 223.0 @ 0.792 min (5–95% ACN in H2O, 1.5 min).
1H NMR: (DMSO-d6, 400 MHz) δppm 7.59–7.53 (m, 1 H), 7.53–7.48 (m, 1 H).
(6,7-dichloro-1H-benzimidazol-2-yl)hydrazine (6). A flask was charged with 2,
6, 7-trichloro-1H-benzimidazole (300.00 mg, 1.35 mmol, 1.00 eq) and NH2NH2.
H2O (30.90 g, 617.27 mmol, 30.00 ml, 455.69 eq). The mixture was stirred at 100 °C
for 12 h. After being cooled to ambient temperature (25 °C), water (4 ml) was
added to the reaction mixture under ice cooling. The resulting precipitates were
collected by filtration and washed with water (3 × 3 ml), then dried in vacuum to
give (6,7- dichloro-1H- benzimidazol-2-yl)hydrazine (230.00 mg, 1.06 mmol,
78.49% yield) was obtained as a white solid.
1H NMR: (DMSO-d6, 400 MHz) δppm 8.29 (br s, 1 H), 7.08–7.03 (m, 1 H),
7.01–6.96 (m, 1 H), 4.59 (br s, 2 H).
Mice and isolation of BMDMs. C57BL/6 wild type and Trex1−/− offspring from
Trex1+/− crosses (a kind gift from Dr Dan Stetson, University of Washington)
were harvested at 6–8 weeks of age. Femurs and tibias were removed from age- and
3-[1-(6,7-dichloro-1H-benzimidazol-2-yl)-5-hydroxy-3-methyl-pyrazol-4-yl]-
3H-isobenzofuran-1-one (8). To a suspension of (6,7-dichloro-1H-benzimidazol-
2-yl)hydrazine (100.00 mg, 460.70 umol, 1.00 eq) in acetic acid (5.00 ml) was added
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
8
:
7
5
0
D
O
I
:
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
1
7
-
0
0
8
3
3
-
9
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
|
|
|

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
|
D
O
I
:
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
1
7
-
0
0
8
3
3
-
9
sex-matched mice. The ends were clipped and marrow pushed from bones
via syringe. Clumps were broken apart via pipetting, rinsed, and pelleted in
PBS, passed through a cell strainer, and plated in seven 100 mm petri dishes
per mouse. Cells were matured in media described under the methods for
cell lines with an additional 10 ng ml−1 M-CSF (for a total of 20 ng ml−1 M-CSF
during the maturation period). On Day 7, cells were counted and re-plated at
5 × 105 cells ml−1 in 100 mm petri dishes. BMDMs were maintained by adding an
equal volume of media to each dish every 2 days after harvesting and performing a
complete media change every 4 days after harvesting. IACUC protocols for
humane termination of mice for Vanderbilt University were strictly followed.
Data availability. The pdb files representing the ternary complexes of cGAS +
dsDNA + RU.365/RU.332/RU.521 have been deposited in the RCSB Protein Data
Bank and can be accessed with the following codes, respectively: 5XZB, 5XZE,
5XZG. Further information regarding the high-throughput screen can be found in
the Supplementary Information file. All other relevant data are available upon
reasonable request.
Received: 20 June 2016 Accepted: 31 July 2017
Cellular luciferase assays. A concentration range of dsDNA from 0.004 to 40 µg
was complexed with Lipofectamine LTX (Life Technologies) per the manu-
facturer’s suggested protocol. The sequence and annealing conditions of the
dsDNA used is described above. Typically, 0.4 µg or 0.6 µg of dsDNA was used to
stimulate RAW-Lucia ISG-KO-TREX1/ISG-KO-cGAS or RAW-Lucia ISG cells,
respectively, which equated to 90% of maximal luciferase activity observed. dsDNA
stimulation of RAW macrophages and subsequent luciferase measurements
were performed with or without the addition of small molecules at indicated
concentrations in each figure, or with DMSO as vehicle control. For the dose
curve experiments, serial dilutions of inhibitors were applied to 6.7 × 105 cells,
plated in 24-well dishes, and co-stimulated with 0.4 µg of dsDNA. In subsequent
experiments using inhibitors in RAW cells, the small molecules were used at
concentrations representing the derived EC75 values, while co-stimulated with one
of the following: 0.4 µg dsDNA, 50 nM 5′ppp-HP20 hairpin RNA (a kind gift from
Dr Anna Pyle, Yale University; sequence located in Supplementary Table 5), 10 µM
cGAMP, or 100 U ml−1 murine recombinant Ifnb (R&D Systems). dsDNA was
transfected using Lipofectamine LTX. 5′ppp-HP20 RNA was transfected using
Lipofectamine 2000. Cells were harvested 24 h post-transfection, washed with PBS,
and lysed in 1× Luciferase Cell Lysis Buffer (Pierce). Luciferase luminescence was
recorded with a microplate reader (BioTek Synergy HTX) using QUANTI-Luc
Luciferase reagent (Invivogen) and its suggested protocol (20 µl lysate/well of a
96-well plate; microplate reader set with the following parameters: 50 µl injection,
end-point measurement with 4 s start time and 0.1 s reading time).
References
1. Wu, J. & Chen, Z. J. Innate immune sensing and signaling of cytosolic nucleic
acids. Annu. Rev. Immunol. 32, 461–488 (2014).
2. Barbalat, R., Ewald, S. E., Mouchess, M. L. & Barton, G. M. Nucleic acid
recognition by the innate immune system. Annu. Rev. Immunol. 29, 185–214
(2011).
3. Barrat, F. J., Elkon, K. B. & Fitzgerald, K. A. Importance of Nucleic Acid
Recognition in Inflammation and Autoimmunity. Annu. Rev. Med. 67,
052814–023338 (2015).
4. Christensen, M. H.. & Paludan, S. R. Viral evasion of DNA-stimulated innate
immune responses. Cell Mol. Immunol. doi:10.1038/cmi.2016.06 (2016).
5. Dempsey, A. & Bowie, A. G. Innate immune recognition of DNA: A recent
history. Virology 479–480, 146–152 (2015).
6. Sun, L., Wu, J., Du, F., Chen, X. & Chen, Z. J. Cyclic GMP-AMP synthase is
a cytosolic DNA sensor that activates the type I interferon pathway. Science
(New York, NY) 339, 786–791 (2013).
7. Schoggins, J. W. et al. A diverse range of gene products are effectors of the type
I interferon antiviral response. Nature 472, 481–485 (2011).
8. Cai, X., Chiu, Y.-H. & Chen, Z. J. The cGAS-cGAMP-STING pathway of
cytosolic DNA sensing and signaling. Mol. Cell 54, 289–296 (2014).
9. Hornung, V., Hartmann, R., Ablasser, A. & Hopfner, K.-P. OAS proteins and
cGAS: unifyingconcepts in sensing and respondingto cytosolic nucleic acids.
Nat. Rev. Immunol. 14, 521–528 (2014).
10. Bhat, N. & Fitzgerald, K. A. Recognition of cytosolic DNA by cGAS and other
STING-dependent sensors. Eur. J. Immunol. 44, 634–640 (2014).
11. Wu, J. et al. Cyclic GMP-AMP is an endogenous second messenger in innate
immune signaling by cytosolic DNA. Science (New York, NY) 339, 826–830
(2013).
The minimum numbers of biological and technical replicates for the luciferase
assays were two and three, respectively. Analysis of luminescence values were
evaluated for outliers (one standard deviation above and below the mean) for each
biological replicate, and the resulting means were used to generate graphs in
GraphPad Prism (v. 7.0c); where applicable, an unpaired t-test with Welch’s
correction was used to compare the results of each individual compound to dsDNA
or other stimulant. The experimental variances observed within control samples
(untreated or vehicle) and treated samples were similar.
12. Diner, E. J. et al. The innate immune DNA sensor cGAS produces a
noncanonical cyclic dinucleotide that activates human STING. Cell Rep. 3,
1355–1361 (2013).
13. Ablasser, A. et al. cGAS produces a 2‘-5’-linked cyclic dinucleotide second
messenger that activates STING. Nature 498, 380–384 (2013).
14. Gao, P. et al. Cyclic [G(2’,5’)pA(3‘,5’)p] is the metazoan second messenger
produced by DNA-activated cyclic GMP-AMP synthase. Cell 153, 1094–1107
(2013).
15. Zhang, X. et al. Cyclic GMP-AMP containing mixed phosphodiester
linkages is an endogenous high-affinity ligand for STING. Mol. Cell 51,
226–235 (2013).
16. Civril, F. et al. Structural mechanism of cytosolic DNA sensing by cGAS.
Nature 498, 332–337 (2013).
17. Ishikawa, H. & Barber, G. N. STING is an endoplasmic reticulum adaptor that
facilitates innate immune signalling. Nature 455, 674–678 (2008).
18. Ishikawa, H., Ma, Z. & Barber, G. N. STING regulates intracellular DNA-
mediated, type I interferon-dependent innate immunity. Nature 461, 788–792
(2009).
19. Zhong, B. et al. The adaptor protein MITA links virus-sensing receptors to
IRF3 transcription factor activation. Immunity 29, 538–550 (2008).
20. Sun, W. et al. ERIS, an endoplasmic reticulum IFN stimulator, activates innate
immune signaling through dimerization. Proc. Natl Acad. Sci. 106, 8653–8658
(2009).
21. Gao, P. et al. Structure-function analysis of STING activation by c[G(2’,5’)pA
(3‘,5’)p] and targeting by antiviral DMXAA. Cell 154, 748–762 (2013).
22. Hansen, K. et al. Listeria monocytogenes induces IFNβ expression through
an IFI16-, cGAS- and STING-dependent pathway. EMBO. J 33, 1654–1666
(2014).
23. Zhang, Y. et al. The DNA sensor, cyclic GMP-AMP synthase, is essential for
induction of IFN-β during Chlamydia trachomatis infection. J. Immunol. 193,
2394–2404 (2014).
Cellular cytokine analysis of RAW macrophages. The small molecules were used
at concentrations representing the EC75 values and were co-stimulated with either
200 ng Pam3CSK4 (Invivogen), 25 µg poly(I:C) or 100 ng lipopolysaccharide
(Sigma-Aldrich) at the cell counts and conditions previously described for the
cellular luciferase assays. Cells were harvested 24 h post-transfection, washed with
PBS, and RNA was extracted using Trizol (Ambion) and chloroform (Sigma-
Aldrich). In all, 1–2 µg of total RNA was reverse-transcribed using random hex
primers and SuperScript III (Life Technologies) into cDNA. Real-time PCR was
carried out with 1× FastSYBR Green Plus Master Mix (Applied Biosystems) and
run on an Applied Biosystems StepOne Plus PCR machine. The expression of
mRNAs repressing murine Actb1 and Il-6 were measured (see Supplementary
Table 5 for primer sequences used). Target CT values were normalized to Actb1 CT
values and used to calculate ΔCT. mRNA expression of target genes were then
calculated using the ΔΔCT method (2ΔΔCT). Expression values were analyzed as
described in the method for cellular luciferase assays.
Cytotoxicity assays. Small-molecule compounds were serially diluted to
concentrations spanning the range tested in the response curves were added
to 6.7 × 105 RAW-Blue macrophages plated 16 h prior in 96-well dishes, then
harvested 72 h after compound addition. ATP was measured using CellTiter Glo
Viability Assay (Promega) using 50 µM Tamoxifen (Sigma) as a positive control for
cytotoxicity. Viability values were generated using vehicle (DMSO) or the first dose
(RU.521) as 100% and Tamoxifen as 0%. Outliers were removed as described
previously.
Cytokine expression analysis of RAW and BMDM cells. Wild type and
Trex1−/− BMDM cells were plated at 1 × 105 cells/well (24-well dishes) 16 h prior
to addition of inhibitors at their respective IC50 values. BMDMs were incubated in
the presence of DMSO or the various inhibitors, then harvested 24 h later. RNA
was extracted using Trizol and chloroform. Genomic DNA was removed
using TURBO DNase (Ambion) followed by a phenol-chloroform: chloroform
extraction. 480–1000 ng of total RNA was reverse-transcribed using random hex
primers and SuperScript III into cDNA. Real-time PCR against murine Actb1 and
IFNB1 (see Supplementary Table 5 for primer sequences used) was carried out and
the data was analyzed as previously described in the method section for cytokine
analysis in RAW macrophages.
24. Watson, R. O. et al. The Cytosolic Sensor cGAS Detects Mycobacterium
tuberculosis DNA to Induce Type I Interferons and Activate Autophagy. Cell
Host Microbe 17, 811–819 (2015).
25. Collins, A. C. et al. Cyclic GMP-AMP synthase is an innate immune DNA
sensor for Mycobacterium tuberculosis. Cell Host Microbe 17, 820–828 (2015).
12
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
8
:
7
5
0
D
O
I
:
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
1
7
-
0
0
8
3
3
-
9
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
CA
T
I
O
N
S
|
D
O
I
:
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
1
7
-
0
0
8
3
3
-
9
A
R
T
I
C
L
E
26. Wassermann, R. et al. Mycobacterium tuberculosis differentially activates cGAS-
and inflammasome-dependent intracellular immune responses through ESX-1.
Cell Host Microbe 17, 799–810 (2015).
57. Zhang, J. H. A simple statistical parameter for use in evaluation and
validation of high throughput screening assays. J. Biomol. Screen. 4, 67–73
(1999).
27. Li, X.-D. et al. Pivotal roles of cGAS-cGAMP signaling in antiviral defense and
immune adjuvant effects. Science (New York, NY) 341, 1390–1394 (2013).
28. Ma, Z. et al. Modulation of the cGAS-STING DNA sensing pathway by
gammaherpesviruses. Proc. Natl Acad. Sci. 112, E4306–E4315 (2015).
29. Lau, L., Gray, E. E., Brunette, R. L. & Stetson, D. B. DNA tumor virus
oncogenes antagonize the cGAS-STING DNA-sensing pathway. Science
(New York, NY) 350, 568–571 (2015).
30. Gao, D. et al. Cyclic GMP-AMP synthase is an innate immune sensor of HIV
and other retroviruses. Science (New York, NY) 341, 903–906 (2013).
31. Pisetsky, D. S. Anti-DNA antibodies-quintessential biomarkers of SLE.
Nat. Rev. Rheumatol. 12, 102–110 (2016).
58. Du, Y. et al. High-throughput screening fluorescence polarization assay for
tumor-specific Hsp90. J. Biomol. Screen 12, 915–924 (2007).
59. McCoy, A. J. et al. Phaser crystallographic software. J. Appl. Crystallogr. 40,
658–674 (2007).
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. Adams, P. D. et al. PHENIX: a comprehensive Python-based system for
macromolecular structure solution. Acta Crystallogr. D Biol. Crystallogr. 66,
213–221 (2010).
32. Crow, Y. J. et al. Mutations in the gene encoding the 3‘-5’ DNA exonuclease
TREX1 cause Aicardi-Goutières syndrome at the AGS1 locus. Nat. Genet. 38,
917–920 (2006).
33. Rice, G. et al. Heterozygous mutations in TREX1 cause familial chilblain lupus
and dominant Aicardi-Goutieres syndrome. Am. J. Hum. Genet. 80, 811–815
(2007).
34. Gray, E. E., Treuting, P. M., Woodward, J. J. & Stetson, D. B. Cutting edge:
cGAS is required for lethal autoimmune disease in the trex1-deficient mouse
model of Aicardi-Goutieres syndrome. J. Immunol. 195, 1939–1943 (2015).
35. Gao, D. et al. Activation of cyclic GMP-AMP synthase by self-DNA causes
autoimmune diseases. Proc. Natl Acad. Sci. 201516465 doi:10.1073/
pnas.1516465112 (2015).
36. Stetson, D. B., Ko, J. S., Heidmann, T. & Medzhitov, R. Trex1 prevents cell-
intrinsic initiation of autoimmunity. Cell 134, 587–598 (2008).
37. Ahn, J., Ruiz, P. & Barber, G. N. Intrinsic self-DNA triggers inflammatory
disease dependent on STING. J. Immunol. 193, 4634–4642 (2014).
38. White, M. J. et al. Apoptotic caspases suppress mtDNA-induced STING-
mediated type I IFN production. Cell 159, 1549–1562 (2014).
39. West, A. P. et al. Mitochondrial DNA stress primes the antiviral innate immune
response. Nature 520, 553–557 (2015).
Acknowledgements
We would like to thank Dr Dan Stetson and his laboratory (University of Washington)
for providing the Trex1 null mice (originally generated by Dr Tomas Lindahl). We would
like to thank the synchrotron beam line staff at the Argonne National Laboratory for
their assistance. We would like to thank Dr Lauren Frick at Agilent Technologies
for technical support in running the RF-MS. We thank the High Throughput and
Spectroscopy Resource Center at the The Rockefeller University for providing equipment
infrastructure. We also would like to thank Dr Taku Kamei (TDI) for reviewing the data
on the chemical synthesis of derivatives of RU.365. Finally, we would like to thank
members of the Ascano laboratory for their support, collegiality, and critical review of the
manuscript. This work was supported, in part, by the following agencies:
1R35GM119569-01 (M.A.), Vanderbilt University Dept. Biochemistry start-up funds
(M.A.), Rockefeller University Robertson Therapeutic Development Funds (J.F.G., T.T.,
and M.A.), Cancer Research Institute Irvington Postdoctoral Fellowship and
NSFC31670903 (P.G.), NIGMS training grant support (4T32GM065086-14) (K.R.), and
GM104962 and MSKCC core grant (P30 CA008748) (D.J.P). The crystallographic
research was conducted at the Northeastern Collaborative Access Team beamlines, which
are funded by NIGMS (P41 GM103403) and DOE (DE-AC02-06CH11357). The RFF-
MSS instrument was purchased with funds from The Leona M. and Harry B. Helmsley
Charitable Trust.
40. An, J., Woodward, J. J., Sasaki, T., Minie, M. & Elkon, K. B. Cutting edge:
Antimalarial drugs inhibit IFN-β production through blockade of cyclic GMP-
AMP synthase-DNA interaction. J. Immunol. 194, 4089–4093 (2015).
41. Zhang, X. et al. The cytosolic DNA sensor cGAS forms an oligomeric complex
with DNA and undergoes switch-like conformational changes in the activation
loop. Cell Rep. 6, 421–430 (2014).
42. Baell, J. B. & Holloway, G. A. New substructure filters for removal of pan assay
interference compounds (PAINS) from screening libraries and for their
exclusion in bioassays. J. Med. Chem. 53, 2719–2740 (2010).
43. Mazerski, J., Martelli, S. & Borowski, E. The geometry of intercalation complex
of antitumor mitoxantrone and ametantrone with DNA: molecular dynamics
simulations. Acta Biochim. Pol. 45, 1–11 (1998).
Author contributions
J.V. contributed to the design, execution, and interpretation of all the cellular assays. C.A.
contributed to the design, execution, and interpretation of the screening, biochemical,
and biophysical assays, with experimental assistance from L.L., A.L., T.G., and J.S. P.G.
conducted the structural biology work and the protein preparation for high-throughput
screening. Y.A. was the lead chemist with R.O., T.I., and J.A. responsible for the design
and preparation of the compounds based on the RU.365 scaffold. K.A. supervised the
medicinal chemistry team. K.R. and S.R. ran qRT-PCR experiments and analyses. M.A.,
J.F.G., D.J.P., and T.T. supervised and contributed to the design and interpretation of the
data. J.F.G. designed, curated, and annotated the compound library, and designed the
enzymology and medicinal chemistry experiments. M.A. wrote the manuscript with
editing and writing assistance from all authors.
44. Kohlway, A., Luo, D., Rawling, D. C., Ding, S. C. & Pyle, A. M. Defining the
functional determinants for RNA surveillance by RIG-I. EMBO Rep. 14,
772–779 (2013).
45. Segel, I. H. Enzyme Kinetics. 125 (Wiley-Interscience, 1993).
46. Blat, Y. Non-competitive inhibition by active site binders. Chem. Biol. Drug
Des. 75, 535–540 (2010).
Additional information
Supplementary Information accompanies this paper at doi:10.1038/s41467-017-00833-9.
47. Burdette, D. L. & Vance, R. E. STING and the innate immune response to
nucleic acids in the cytosol. Nat. Immunol. 14, 19–26 (2013).
48. Paludan, S. R. & Bowie, A. G. Immune sensing of DNA. Immunity 38, 870–880
(2013).
Competing interests: The authors declare no competing financial interests.
Reprints and permission information is available online at http://npg.nature.com/
49. Reijns, M. A. M. et al. Enzymatic removal of ribonucleotides from DNA is
essential for mammalian genome integrity and development. Cell 149,
1008–1022 (2012).
50. Rice, G. I. et al. Mutations in ADAR1 cause Aicardi-Goutières syndrome
associated with a type I interferon signature. Nat. Genet. 44, 1243–1248 (2012).
51. Hiller, B. et al. Mammalian RNase H2 removes ribonucleotides from DNA to
maintain genome integrity. J. Exp. Med. 209, 1419–1426 (2012).
52. Morita, M. et al. Gene-targeted mice lacking the trex1 (DNase III) 3′->5′ DNA
exonuclease develop inflammatory myocarditis. Mol. Cell Biol. 24, 6719–6727
(2004).
53. Unterholzner, L. et al. IFI16 is an innate immune sensor for intracellular DNA.
Nat. Immunol. 11, 997–1004 (2010).
54. Orzalli, M. H. et al. cGAS-mediated stabilization of IFI16 promotes innate
signaling during herpes simplex virus infection. Proc. Natl Acad. Sci 112,
E1773–E1781 (2015).
reprintsandpermissions/
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
Open Access This article is licensed under
a Creative Commons
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/.
55. Diner, B. A., Lum, K. K., Toettcher, J. E. & Cristea, I. M. Viral DNA sensors
IFI16 and cyclic GMP-AMP synthase possess distinct functions in regulating
viral gene expression, immune defenses, and apoptotic responses during
herpesvirus infection. mBio 7, e01553–16–15. (2016).
56. Almine, J. F. et al. IFI16 and cGAS cooperate in the activation of STING
during DNA sensing in human keratinocytes. Nat. Commun. 8, 14392 (2017).
© The Author(s) 2017
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
8
:
7
5
0
D
O
I
:
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
1
7
-
0
0
8
3
3
-
9
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
13
|
|
|