HSV-1–inducedIfnb1transcription was atte-
nuated inHnrnpa2b1-KO PMs (Fig. 6C). Vaccinia
virus (VACV), another DNA virus, replicates in
the cytoplasm and is sensed by cytosolic DNA
sensors ( 29 ). VACV infection inHnrnpa2b1-KO
PMs inducedIfnb1production to a certain level
after 4 hours but showed no subsequent in-
creases inIfnb1as it did in wild-type cells (Fig. 6C).
Wild-type macrophages showed higher and more
sustainedIfnb1expression thanHnrnpa2b1-KO
macrophages in response to both viruses (Fig. 6C).
Similar results were obtained in BMDMs (fig. S11).
Thus, hnRNPA2B1 appears to be required for
cGAS-, IFI16-, and STING-mediated pathways
tofullyactivatetypeIinterferonproduction
against DNA viruses.
hnRNPA2B1 promotes
nucleocytoplasmic trafficking of CGAS,
IFI16, and STING mRNAs to amplify
cytoplasmic innate sensor signaling
We next investigated why hnRNPA2B1 is re-
quired for the efficient induction of IFN-I by
cGAS, IFI16, and STING in response to HSV-1
infection. Up to 6 hours after HSV-1 infection,
the endogenous amounts of cGAS, p204 (the
functional mouse ortholog of human IFI16),
and STING protein were similar in both wild-
type andHnrnpa2b1-KO RAW264.7 cells. cGAS
expression began to increase in wild-type
RAW264.7 cells 6 hours after infection, whereas
p204 began to increase 12 hours after infection.
However, these proteins failed to increase in
Hnrnpa2b1-KO RAW264.7 cells following HSV-1
infection (Fig. 6D). STING abundance decreased
more rapidly inHnrnpa2b1-KO RAW264.7 cells
than in wild-type RAW264.7 cells 6 hours after
infection (Fig. 6D). Thus, hnRNPA2B1 appears to
be required for the efficient induction of cGAS,
IFI16, and STING after DNA virus infection,
which subsequently generates an antiviral IFN-I
response.
We examined the effects of hnRNPA2B1 on
Cgas,p204, andStingmRNA expression. The
transcriptional levels of these genes were sim-
ilar in wild-type andHnrnpa2b1-KO RAW264.7cells, indicating that the splicing of these mRNAs
was unaffected (fig. S12A). Similarly, the stability
of these mRNAs did not significantly differ
between these cell lines (fig. S12B). However,
depletion of hnRNPA2B1 led to the nuclear re-
tention ofCgas,p204,andStingmRNAs (Fig. 7A).
Analysis of mRNAs associated with hnRNPA2B1-
immunoprecipitated complexes revealed that
hnRNPA2B1 was able to bindCgas,p204,and
StingmRNAs in macrophages after HSV-1 infec-
tion (Fig. 7B). Thus, hnRNPA2B1 appears to play
a role in mediating the nucleocytoplasmic trans-
location of these mRNAs.
N6-methyladenosine (m^6 A) is the predominant
methylated base in mammalian mRNAs and
has been recently revealed to promote mRNA
translocation from the nucleus to the cytoplasm
( 30 , 31 ). A greater number of methylated mRNAs
were precipitated after HSV-1 infection, although
the amounts of methylatedCgas,p204,andSting
mRNA were much lower inHnrnpa2b1-KO
RAW264.7 cells than in controls (Fig. 7C). Thus,
specific classes of mRNAs involved in antiviral
response such asCgas, p204,andSting,undergo
m^6 AmodificationafterDNAvirusinfectioninan
hnRNPA2B1-dependent manner.
Two RNA demethylases, alkylated DNA repair
protein alkB homolog 5 (ALKBH5) and fat mass
and obesity-associated protein (FTO), have been
identified to date ( 31 , 32 ). Our MS data suggested
an interaction between hnRNPA2B1 and FTO.
hnRNPA2B1 was constitutively associated with
FTO,andhnRNPA2B1disassociatedwithFTO
after HSV-1 infection in mouse PMs and human
THP1 cells (Fig. 7D and fig. S12C). Knockdown of
FTO led to increased HSV-1–inducedIfnb1ex-
pression in macrophages (Fig. 7E and fig. S12D).
The METTL3–METTL14 complex mediates
mRNA m^6 A methylation. To explore whether
METTL3 was involved in the innate immune re-
sponse, we studied myeloid cell–specificMettl3-
KO mice ( 33 ).Ifnb1expression was impaired in
Mettl3-KO PMs and BMDMS after HSV-1 infec-
tion (fig. S12E). METTL3 deficiency did not af-
fect hnRNPA2B1 binding withCgas,p204,or
StingmRNAs (fig. S12F).Cgas,p204,andSting
m^6 A levels were lower inMettl3-KO macro-
phages than in wild-type cells (fig. S12G). Thus,
METTL3 contributes to the m^6 Amodification
of hnRNPA2B1-bound mRNAs in macrophages,
which promotes IFN-bexpression in response
to DNA virus infection.
The binding ofCGAS, IFI16,andSTINGmRNAs
by demethylated hnRNPA2B1 was severely im-
paired compared to that by hnRNPA2B1-FL
(Fig. 7F and fig. S12H). The m^6 Alevelsofthese
mRNAs in JMJD6 inhibitor-treated cells were
similar to control cells (Fig. 7G). Thus, RNA
binding by hnRNPA2B1 requires Arg^226 methyl-
ation, and demethylated hnRNPA2B1 cannot bind
these mRNAs or affect their m^6 Amodification.
Analysis of RNA from nuclear and cytoplasmic
fractions in both wild-type andHnrnpa2b1-KO
macrophages before and after HSV-1 infection
revealed that hnRNPA2B1 deficiency decreased
the nuclear export of mRNAs involved in sev-
eral biological processes, including pheromoneWanget al.,Science 365 , eaav0758 (2019) 16 August 2019 5of11
A
Ifnb1activation (fold)0510152025100125150MockFLS17AY28AY29AS58AS59AS73AS90AS137AT147AY162AT164AS177AS189AS200AS213AS219AS224AY232AY235AY243AS245AY248AY254AS259AY261AS266AS272AY279AY284AS289AS292AY295AY301AR188AR191AR201AR216AR226AR273A**
*B
Ifnb1 mRNA (fold)A2B1-FL
A2B1-R226AUI HSV-1010203040(^50)
C
IFN-
β (pg/ml)
A2B1-FL
A2B1-R226A
UI HSV-1
0
200
400
(^600)
D
IP:anti-A2B1
HSV-1 (h)
MMA
A2B1
++IgG
0 22
E
A2B1-Flag:
IP: anti-FlagMMA
Flag
R188AR191AR201AR216AR226AFL
F
R226-Deme-A2B1
A2B1
HSV-1 (h) 0 2 4 0
Hnrnpa2b1+/+ +/+ +/+ –/–
Fig. 4. Arg^226 demethylation is essential for hnRNPA2B1-mediated type I IFN induction.
(A) HEK293T cells were transiently transfected with hnRNPA2B1 or its mutant expression vectors
with anIfnb1reporter vector as indicated. The activation of theIfnb1reporter was examined by dual
luciferase reporter assays. (BandC)Hnrnpa2b1-KO RAW264.7 cells transfected with hnRNPA2B1,
and mutant expression vectors were infected with HSV-1 for 7 hours (B) or 18 hours (C).Ifnb1mRNA
was measured by qPCR (B), and IFN-bconcentrations were measured by ELISA (C). (D) Mouse
PMs were infected with HSV-1 (MOI, 10) for 2 hours. Cell lysates were immunoprecipitated with
anti-hnRNPA2B1 and then examined for the level of Arg methylation by immunoblot. (E) HEK293T
cells were transfected with the indicated vectors. Cell lysates were immunoprecipitated with
anti-Flag and then examined for the level of Arg methylation by immunoblot. (F) The demethylation
of hnRNPA2B1 was detected by using a specific antibody against R226-demethylated hnRNPA2B1
in RAW264.7 cells in response to HSV-1 infection (MOI, 10). Similar results were obtained for
three independent experiments. One representative experiment is shown [(D) to (F)]. Data are
displayed as means ± SEM of three [(A) to (C)] independent experiments performed in triplicate.
***P< 0.001, two-tailed, unpaired Student’sttest [(A) to (C)]. See also fig. S9.
RESEARCH | RESEARCH ARTICLE