USP44 positively regulates innate immune response to DNA viruses through deubiquitinating MITA

Mediator of IRF3 activation (MITA, also known as stimulator of interferon genes, STING) senses the second messenger cyclic GMP-AMP (cGAMP) which is synthesized upon DNA virus infection and activates innate antiviral immune response. It has been demonstrated that the activity of MITA is delicately regulated by various post-translational modifications including polyubiquitination. In this study, we identified the deubiquitinating enzyme USP44 as a positive regulator of MITA. USP44 is recruited to MITA following DNA virus infection and removes K48-linked polyubiquitin moieties from MITA at K236, therefore prevents MITA from proteasome mediated degradation. USP44-deficiency results in acceleration of HSV-1-induced degradation of MITA and reduced induction of type I interferons (IFNs) and proinflammatory cytokines. Consistently, Usp44-/- mice are more susceptible to HSV-1 infection as indicated by higher tissue viral titers, greater tissue damage and lower survival rate. These findings suggest that USP44 plays a specific and critical role in the regulation of innate immune response against DNA viruses.


Author summary
Cyclic GMP-AMP synthase (cGAS) senses cytosolic dsDNA and initiates signal transductions, leading to activation of innate immune response. MITA is the key adaptor protein downstream of cGAS and plays a critical role in cGAS-mediated signaling. The activity of MITA is tightly regulated by various post-translational modifications including polyubiquitination and deubiquitination. Here we found that the deubiquitinating enzymes USP44 associates with MITA and removes the K48-linked polyubiquitin chains from MITA, therefore maintains the stability of MITA after DNA virus infection. Deficiency of USP44 results in accelerated degradation of MITA, impaired induction of type I IFNs and proinflammatory cytokines, and increased viral replication. These findings suggest that USP44 is a positive regulator of MITA and plays an important role in the regulation of innate immune response against DNA viruses.
As a key adaptor protein in innate immune response against DNA viruses, the activity of MITA is delicately regulated. Several post-translational modifications, such as phosphorylation, sumoylation and polyubiquitination, have been reported to play important roles in regulation of MITA [28]. For example, phosphorylation of MITA at Ser358 and Ser366 is crucial for its activation and recruitment of IRF3 [18,26,27]; sumoylation of MITA in the early phase of viral infection by TRIM38 promotes its stability and activation whereas desumoylation of MITA in the late phase by SENP2 leads to its degradation thus avoiding sustained activation of MITA [29]. In addition, polyubiquitination of MITA have also been reported to distinctly regulate its activity. The E3 ubiquitin ligases TRIM56 and TRIM32 mediated K63-linked polyubiquitination of MITA and AMFR mediated K27-linked polyubiquitination of MITA have been shown to enhance the antiviral innate immune response [30][31][32]. RNF5 catalyzes K48-linked polyubiquitination of MITA that results in its degradation [33], and this could be inhibited by RNF26 mediated K11-linked polyubiquitination of MITA at the same site [34].
Deubiquitination is a reverse biochemical process of polyubiquitination, in which polyubiquitin chains previously added to target proteins are removed by deubiquitinating enzymes (DUBs). Recently, several DUBs have been reported to regulate the antiviral innate immune response. For example, USP14 removes K48-linked polyubiquitination of cGAS therefore impairs autophagic degradation of cGAS [35]. It has been reported that USP13 removes K27-linked polyubiquitin chains from MITA and inhibits interaction between MITA and TBK1, resulting in impaired antiviral innate immune response [36]. USP49 has been shown to deconjugate K63-linked polyubiquitin chains from MITA and inhibits the aggregation and activation of MITA [37].
In this study, we screened for DUBs that involve in DNA virus-triggered induction of type I IFNs and identified ubiquitin-specific protease 44 (USP44) as a positive regulator. USP44 consists of 712 amino acids, which contains a Zn-finger domain at the N-terminus and a USP domain at the C-terminus [38]. USP44 has been shown to be involved in many cellular processes, including embryonic stem cell (ESC) differentiation, cell proliferation, DNA repair and tumor progression [39][40][41][42][43]. However, it is dispensable for the differentiation of hematopoietic stem cells (HSC) and barely affects the development of immune cells [44]. So far, whether USP44 is involved in immune response is unclear. We found that overexpression of USP44 potentiated DNA-but not RNA virus-triggered production of type I IFNs and proinflammatory cytokines. Conversely, USP44 deficiency suppressed cytosolic DNA-and DNA virus-triggered innate immune response. Further study revealed that USP44 was recruited to MITA and selectively removed K48-linked polyubiquitin chains from MITA at K236, therefore inhibited proteasome-mediated degradation of MITA and promoted antiviral response against DNA viruses. Our study shed new light on the function of USP44 and further demonstrated the importance of deubiquitination in the regulation of antiviral innate immune response.

USP44 positively regulates DNA virus-induced innate immune response
To identify potential molecules involved in DNA-triggered signaling. We screened~40 independent human DUB cDNA expression plasmids for their abilities to regulate the IFNB promoter in reporter assays. These efforts led to identification of USP44 as a candidate protein that could potentiate activation of the IFNB promoter triggered by the DNA virus herpes simplex virus 1 (HSV-1) (S1 Fig). Overexpression of USP44 potentiated HSV-1-, but not SeVinduced activation of the IFNB promoter, interferon-stimulated response element (ISRE) and NF-κB in a dose-dependent manner in HEK293 cells ( Fig 1A). Consistently, overexpression of USP44 enhanced cGAS-MITA-induced activation of the IFNB promoter in a dose-dependent manner ( Fig 1B). We next established a THP-1 cell line stably expressing USP44. qPCR experiments indicated that USP44 enhanced transcription of IFNB1, IFIT1, IL6 and CXCL10 genes in response to HSV-1 ( Fig 1C). In addition, transcription of downstream genes induced by transfected HSV120 (120-mer dsDNA representing the genome of HSV-1) was also enhanced in USP44-expressing THP-1 cells (Fig 1D). In contrast, the transcription of downstream genes induced by SeV was not markedly affected in USP44-expressing THP-1 cells (Fig 1E). Moreover, USP44 enhanced HSV-1-induced phosphorylation of TBK1, IRF3, IKKβ and p65 ( Fig  1F). These results suggest that USP44 specifically regulates in DNA virus-triggered innate immune response.

Knockdown of USP44 inhibits DNA virus-triggered signaling
To confirm the role of endogenous USP44 in DNA virus-triggered signaling, we designed two pSuper-RNAi constructs which could efficiently knockdown the expression of USP44 (Fig  2A). qPCR experiments indicated that transcription of IFNB1, IL6 and CXCL10 induced by HSV-1 infection and transfected dsDNA, such as HSV120 and interferon stimulating DNA (ISD), was markedly reduced in USP44-RNAi expressing THP-1 cells compared with control cells (Fig 2B and S2 Fig). In contrast, knockdown of USP44 had little effects on SeV-induced the transcription of downstream genes in THP-1 cells (Fig 2C). Knockdown of USP44 also reduced HSV-1-induced phosphorylation of TBK1, IRF3, IKKβ and p65 in THP-1 cells ( Fig  2D).
To determine whether USP44 regulates in a cell-and virus-specific manner, we used independent siRNA construct to knockdown USP44 in human foreskin fibroblasts (HFFs), which are permissive to human cytomegalovirus (HCMV) infection. We found that knockdown of USP44 inhibited HCMV-and HSV120-induced transcription of IFNB1, IFIT1, IL6 and CXCL10 genes (Fig 2E), but had no marked effects on SeV-or dsRNA mimic poly(I:C)induced transcription of antiviral genes in HFFs (Fig 2F). These results suggest that USP44 plays a general role in modulating innate immune response to DNA virus in various cell types.
To investigate whether USP44 is important for host antiviral response in vivo, wild-type and Usp44 knockout mice were intra-peritoneally infected with HSV-1 at lethal or non-lethal dose. While the concentrations of serum cytokines including IFN-β, IL-6 and CXCL10 were obviously lower in Usp44 -/than that in the wild type mice (Fig 4A and 4B), the lung and brain levels of HSV-1 ICP22 and ICP27 mRNA as well as the brain viral titre were significantly increased in Usp44 -/mice ( Fig 4C and 4D). Similar results were obtained when mice are intranasally infected ( Fig 4E). Furthermore, hematoxylin and eosin staining indicated that the lung tissues of Usp44 -/mice exhibited greater infiltration of immune cells and tissue damage after HSV-1 infection compared with that of Usp44 +/+ mice ( Fig 4F). Consistently, Usp44 -/mice were more susceptible to HSV-1-, but not VSV-induced death compared to their wild-type littermates (Fig 4G and 4H). Collectively, these data suggest that USP44 is essential for host defense against DNA virus infection.

USP44 acts at the level of MITA
To investigate the mechanisms on how USP44 modulates innate immune responses to DNA viruses, we examined the effects of USP44 on activation of the IFNB promoter mediated by components of the DNA-triggered signaling pathways. Reporter assays indicated that cotransfection of USP44 enhanced activation of the IFNB promoter mediated by cGAS and MITA but had no marked effects on activation of the IFNB promoter mediated by their USP44 expression plasmids. Twenty-four hours after transfection, cells were left uninfected, infected with HSV-1 (top) or SeV (bottom) (MOI = 1) for 12 hours before luciferase assays were performed. (B) HEK293T cells (1 x 10 5 ) were cotransfected with empty vector or cGAS and MITA, IFN-β reporter (0.05 μg), and increased amounts of USP44 for 24 hours before luciferase assays were performed. (C and D) THP-1 cells (4 x 10 5 ) stably expressing USP44 were left untreated, infected with HSV-1 (MOI = 1) (C) or transfected with HSV120 (2 μg/ml) (D) for 12 hours before qPCR analysis. (E) THP-1 cells (4 x 10 5 ) stably expressing USP44 were left uninfected or infected with SeV (MOI = 1) for 12 hours before qPCR analysis. (F) THP-1 cells (4 x 10 5 ) stably expressing USP44 were left uninfected or infected with HSV-1 (MOI = 1) for the indicated times followed by immunoblot analysis. Graphs show mean ± S.D. n = 3. � P < 0.05,  downstream components TBK1 and IRF3-5D (a constitutively active mutant of IRF3) ( Fig  5A). Interestingly, knockdown of USP44 in THP-1 cells or knockout of Usp44 in mouse BMDMs inhibited cGAMP-triggered transcription of downstream genes ( Fig 5B). These results indicate that USP44 acts downstream of cGAMP and upstream of TBK1-IRF3. Consistently, transient transfection and coimmunoprecipitation experiments showed that USP44 interacted with MITA but not cGAS, TBK1, IKKε or IRF3 ( Fig 5C). Unfortunately, after extensive efforts, our study, as well as the published studies on USP44 [39][40][41][42]44] have not been able to identify a USP44 antibody that can detect endogenous USP44 protein. To overcome this obstacle, we reconstituted immortalized Usp44 -/-MLF cells (iMLFs) with Flag-tagged USP44 (USP44-Flag) by retrovirus-mediated transduction. Using this cell line, we found that USP44 was associated with MITA after HSV-1 infection (Fig 5D). USP44 have been reported to mainly localize in the nucleus [38,43]. However, in the subcellular fractionation experiments, USP44 was detected in both the nuclear and cytoplasmic fractions (S4 Fig). Since MITA is a transmembrane protein [19,21], we further examined whether USP44 was recruited to membranes during HSV-1 infection. As shown in Fig 5E (top panel), while USP44 was detected in both the membraneous and cytosolic fractions in uninfected cells, it migrated from cytosol to membranes after HSV-1 infection (Fig 5E, top panel). Importantly, co-immunoprecipitation experiments with the membraneous fraction demonstrated that USP44 associated with MITA on membranes (Fig 5E, bottom panel). Taken together, these results suggest that USP44 is recruited to membrane-localized MITA after HSV-1 infection.

USP44 deubiquitinates and stabilizes MITA
Since USP44 is a DUB, we determined whether USP44 functions by deubiquitinating MITA. Reporter assays indicated that USP44(C282A), a deubiquitinase inactive mutant of USP44, lost its ability to enhance cGAS-MITA-mediated activation of the IFNB promoter ( Fig 6A). qPCR experiments showed that USP44 but not USP44(C282A) promoted transcription of downstream effector genes induced by HSV-1 infection or transfected HSV120 in THP-1 cells ( Fig  6B). Consistently, USP44(C282A) also lost its ability to potentiate phosphorylation of TBK1, IRF3, IKKβ and p65 triggered by HSV-1 infection (Fig 6C). To further confirm whether the deubiquitinase activity of USP44 was required for its function, we reconstituted Usp44 -/-iMLFs with wild-type USP44 or USP44(C282A) respectively. qPCR results showed that HSV-1-induced transcription of Ifnb1, Il6 and Cxcl10 genes was rescued by reconstitution of USP44 but not USP44(C282A) (Fig 6D). These results suggest that the enzymatic activity of USP44 is essential for its regulation of innate immune response to DNA viruses.
https://doi.org/10.1371/journal.ppat.1008178.g002 To further examine which type of polyubiquitin moieties USP44 removes from MITA, we performed deubiquitination assays with wild-type ubiquitin and the K-only mutants of ubiquitin including K11O, K27O, K29O, K33O, K48O and K63O, in which only the indicated lysine residue was retaining. As shown in Fig 7D, USP44 selectively removed K48-but not other linkage-mediated polyubiquitin moieties from MITA (Fig 7D). In addition, deubiquitination assays with ubiquitin mutants K6R, K11R, K27R, K29R, K33R, K48R and K63R, in which only the indicated lysine residue was mutated to arginine, showed that USP44 impaired MITA ubiquitination of all tested linkages but failed to deubiquitinate MITA when the K48 residue of ubiquitin was mutated (Fig 7E), indicating that USP44 specifically removes K48-linked polyubiquitin chains from MITA. To further confirm this conclusion, we examined the level of K48and K63-linked ubiquitination of endogenous MITA induced by HSV-1 infection in WT and Usp44 -/cells. The results showed that knockout of USP44 potentiated K48-but not K63-linked ubiquitination of endogenous MITA (Fig 7F).
Next we investigated which residue(s) of MITA was targeted by USP44. We performed deubiquitination assays with MITA and a series of its KR mutants where the lysine residues have been individually mutated to arginine. The results showed that among all the KR mutants of MITA, the K236R mutant was resistant to USP44-mediated deubiquitination while the K150R and K224R mutants also exhibited minimal resistance, suggesting that USP44 mainly deubiquitinates MITA at K236 (Fig 7G). Taken together, these findings suggest that USP44 mainly deconjugates K48-linked polyubiquitin chains from MITA at K236.
It has been shown that viral infection-triggered K48-polyubiquitination of MITA promotes its degradation by proteosomes [33,[45][46][47]. As shown in Fig 7H & 7I, HSV-1 induced degradation of MITA was aggravated in Usp44 -/-MLFs compared with the wild-type cells. Importantly, such aggravation was rescued by reconstitution of Usp44 -/-MLFs with USP44 but not USP44(C282A), or by proteasomal but not autophagic or lysosomal inhibitors (Fig 7H and 7I). Taken together, these findings indicate that USP44 prevents MITA from undergoing proteasome-mediated degradation by removing K48-linked polyubiquitination chains at K236 of MITA after DNA virus infection.

Discussion
The cGAS-MITA pathway plays an important role in innate immune response to DNA viruses [15,16,18,19,24]. This pathway is tightly regulated so that the host can efficiently initiate innate antiviral response and timely terminate it to avoid immune dysfunctions. Various posttranslationally modifications, including different types of polyubiquitination, play critical roles in regulation of cGAS-MITA-mediated innate immune response. In this study, we identified USP44 as a positive regulator of innate immune response to DNA viruses by deubiquitinating and stabilizing MITA after viral infection.
Overexpression of USP44 increased cytosolic DNA-and DNA virus-triggered activation of downstream effector genes whereas knockdown of USP44 inhibited HSV-1-induced transcription of downstream effector genes in various cell types. Consistently, USP44-deficiency impaired HSV-1-and cytosolic DNA-induced transcription of downstream genes in murine immune cells. Furthermore, Usp44 -/mice showed lower serum cytokines levels, higher viral loads in lungs and brains after HSV-1 infection, and were more susceptible to HSV-1-induced death. In these experiments, USP44 did not affect innate immune response to the RNA viruses SeV, VSV or cytosolic dsRNA, suggesting that USP44 plays a specific role in modulating innate immune response to DNA viruses.
USP44 interacted with MITA following HSV-1 infection. Although previous studies have reported that USP44 is mainly localized in nucleus [38,43], subcellular fractionation experiments showed that USP44 is distributed in both nucleus and cytoplasm. Interestingly, a portion of cytoplasmic USP44 migrated to membranes after viral infection. Given that MITA is a transmembrane protein and that the MITA-USP44 interaction was detected in the membraneous fraction after HSV-1 infection, a simple explanation is that USP44 interacts with MITA and is recruited to membranes following viral infection.
https://doi.org/10.1371/journal.ppat.1008178.g004 from MITA at K150 [49]. However, CYLD-mediated deubiquitination lacks target specificity in the regulation of innate immune response as it has also been shown to deubiquitinate TRAF2, RIG-I and TBK1 [53][54][55][56]. Importantly, reconstitution experiments demonstrated that USP44mediated regulation of MITA signaling is independent of USP20 and CYLD. In light of these and our studies, it is possible that MITA is distinctly regulated by different USPs in a spatial and temporal manner, so that it could be properly activated and inactivated at the onset and termination of innate immune response to DNA viruses.
Previous studies have shown that USP44 plays roles in tumorigenesis, cell cycle and DNA damage response [39,40,42,43,57]. It has also been shown that the DNA damage factors, meiotic recombination 11 homolog A (MRE11) and Ku70/80 complexes play important roles in initiation of dsDNA-induced type I IFN production [58,59]. It would be interesting for the future studies to investigate whether USP44 is involved in the cross-talk between DNA damage and innate immune response.

Stable cell lines
The HEK293T cells were transfected with two packaging plasmids (pGAG-Pol and pVSV-G) together with empty vector, or the indicated plasmids respectively by calcium phosphate USP44 deubiquitinates MITA to prevent its degradation precipitation. Twenty-four hours later, medium was replaced. Fourty-eight hours later, the recombinant virus-containing medium was then filtered with 0.45 μm filter and added to THP-1 or iMLF cells in the presence of polybrene (8 μg/mL). Twenty-four hours post infection, cells were selected with puromycin (1 μg/mL) for 7 days before experiments.

qPCR
Total RNA was isolated for qPCR analysis to measure mRNA levels of the indicated genes as previously described [63,64,[67][68][69]. Data shown are the relative abundance of the indicated mRNA derived from human or mouse cells normalized to that of GAPDH. Primer sequences for qPCR assays are:

Preparations of BMDMs and MLFs
Bone marrow cells were isolated from tibia and femur. For preparation of bone marrowderived macrophages (BMDMs), bone marrow cells were cultured in 10% M-CSF-containing conditional medium for 5 days. For murine lungs fibroblasts (MLFs), lungs were minced and digested in calcium and magnesium free HBSS containing 10 μg/ml type II collagenase and 20 μg/ml DNase I for 1 h at 37˚C with shaking. Cell suspension was centrifuged at 500 g for 5 min. The cells were then plated in culture medium (1:1 [v/v] DMEM/Ham's F-12 containing 10% fetal bovine serum (FBS), 50 U/ml penicillin, 50 μg/ml streptomycin, 15 mM HEPES, 2 mM L -glutamine). For generation of immortalized MLFs, MLFs were infected with SV40 in the presence of polybrene (8 μg/mL), 24 h later cells were cultured with fresh medium and immortalized cells were selected.

HSV-1 infection of mice
Eight-week-old Usp44 +/+ and Usp44 -/mice were intra-peritoneally (i.p.) or intra-nasally (i.n.) infected with HSV-1. The serum was collected at 6 h post infection for measurement of IFN-β, IL-6 and CXCL10 by ELISA. The viability of the infected mice was monitored for 9 days.

Plaque assay
The brain from HSV-1 infected mice were collected 5 days post infection. The brains were weighed and homogenized with PBS, followed by centrifugation at 1620 g for 30 min, and the supernatants were collected for plaque assays as previously described [60,62,67].

Subcellular fractionation
HEK293T cells were transfected with USP44-Flag for 24 h, nuclear and cytoplasmic fractions were extracted using NE-PER Nuclear and Cytoplasmic Extraction Reagents Kit (Thermo 78835).
For isolation of membraneous and cytosolic fractions, cells were washed with PBS and lysed by bouncing for 50 times in a homogenization buffer (10 mM Tris-HCl [PH 7.4], 10 mM potassium chloride, 2 mM magnesium chloride and 250 mM saccharose). The homogenate was centrifuged at 500 g for 10 min and the supernatant was centrifuged at 5000 g for 10 min. The supernatant from this step was further centrifuged at 20000 g for 30 min. The pellet was membraneous fraction and the supernatant was cytosol.

Deubiquitination assays
Cells were lysed with the lysis buffer (100 μl) and the supernatants were denatured at 95˚C for 5 min in the presence of 1% SDS. The denatured lysates were diluted with NP-40 lysis buffer until the concentration of SDS was reduced < 0.1% followed by immunoprecipitation with the indicated antibodies.

In vitro deubiquitination assays
Denature-IP was performed to obtain ubiquitin-modified MITA from HEK293T cells cotransfected with Flag-tagged MITA and HA-Ubiquitin. The immunoprecipitates were eluted by Flag peptide (0.5 mg/ml, 60 μl). USP44 or USP44(C282A) protein was obtained utilizing an in vitro transcription and translation kit (Promega). The ubiquitin-modified Flag-MITA was incubated with USP44 or USP44(C282A) at 37˚C for 2 h followed with overnight incubation at 16˚C in the presence of ATP (1 μM). The mixtures were analyzed by immunoblots with the indicated antibodies.

Ethics statement
All mice were housed in the specific pathogen-free facility and viral infection experiments were carried out in an ABSL-2 facility at Wuhan Institute of Virology. The experimental protocol was adhered to the International Guiding Principles for Biomedical Involving Animals. The protocol for animal experiments were approved by the Institutional Animal Care and Use Committee of Wuhan Institute of Virology (approval number WIVA31201601).

Statistical analysis
Student's t test was used for statistical analysis with Microsoft Excel and GraphPad Prism Software, P<0.05 was considered significant. For the mouse survival study, Kaplan-Meier survival curves were generated and analyzed by Log-Rank test.