key: cord-0860907-kcpunfxa
authors: Wang, Li; Muneer, Adil; Xie, Ling; Zhang, Feng; Wu, Bing; Mei, Liu; Lenarcic, Erik M; Feng, Emerald Hillary; Song, Juan; Xiong, Yan; Yu, Xufen; Wang, Charles; Gheorghe, Ciprian; Torralba, Karina; Cook, Jeanette Gowen; Wan, Yisong Y.; Moorman, Nathaniel John; Song, Hongjun; Jin, Jian; Chen, Xian
title: Novel gene-specific translation mechanism of dysregulated, chronic inflammation reveals promising, multifaceted COVID-19 therapeutics
date: 2020-11-16
journal: bioRxiv
DOI: 10.1101/2020.11.14.382416
sha: 8aa772d904d97fa7f7d36e9dc7874990a608f7dd
doc_id: 860907
cord_uid: kcpunfxa

Hyperinflammation and lymphopenia provoked by SARS-CoV-2-activated macrophages contribute to the high mortality of Coronavirus Disease 2019 (COVID-19) patients. Thus, defining host pathways aberrantly activated in patient macrophages is critical for developing effective therapeutics. We discovered that G9a, a histone methyltransferase that is overexpressed in COVID-19 patients with high viral load, activates translation of specific genes that induce hyperinflammation and impairment of T cell function or lymphopenia. This noncanonical, pro-translation activity of G9a contrasts with its canonical epigenetic function. In endotoxin-tolerant (ET) macrophages that mimic conditions which render patients with pre-existing chronic inflammatory diseases vulnerable to severe symptoms, our chemoproteomic approach with a biotinylated inhibitor of G9a identified multiple G9a-associated translation regulatory pathways that were upregulated by SARS-CoV-2 infection. Further, quantitative translatome analysis of ET macrophages treated progressively with the G9a inhibitor profiled G9a-translated proteins that unite the networks associated with viral replication and the SARS-CoV-2-induced host response in severe patients. Accordingly, inhibition of G9a-associated pathways produced multifaceted, systematic effects, namely, restoration of T cell function, mitigation of hyperinflammation, and suppression of viral replication. Importantly, as a host-directed mechanism, this G9a-targeted, combined therapeutics is refractory to emerging antiviral-resistant mutants of SARS-CoV-2, or any virus, that hijacks host responses.

The Coronavirus Disease pandemic, caused by severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2), is an unprecedented global public health crisis. As of October 16, 2020 , the virus has claimed over 1.1-million lives. High mortality is observed for SARS-CoV-2-infected patients who have preexisting chronic conditions such as recovery from sepsis, chronic pulmonary diseases, metabolic diseases (e.g., diabetes), asthma, cardiovascular diseases, thrombosis, chronic liver disease and cirrhosis, and cancer 1, 2 . A dysregulated immune system or hyperinflammatory response to SARS-CoV-2 infection is the leading cause of severe illness and mortality. 3, 4 Correspondingly, excessive release of inflammatory factors in a 'cytokine storm' aggravates acute respiratory distress syndrome (ARDS) or widespread tissue damage, which results in respiratory or multi-organ failure and death. The immunopathologic characteristics of COVID-19 include reduction and functional exhaustion of T cells (lymphopenia) and increased levels of serum cytokines (hyperinflammation) 5 . Ordinarily, macrophages regulate the innate immune response to viral threats by producing inflammatory molecules that activate T cells for viral containment and clearance. However, macrophage function/activation is dysregulated in severe COVID-19 patients 6, 7 , and proinflammatory monocyte-derived macrophages are abundant in their bronchoalveolar lavage fluids. 8 These observations indicated the crucial effects of dysregulated macrophages in SARS-CoV-2 immunopathogenesis and associated cytokine storms. Therapeutic options have been severely restrained by a lack of understanding of the pathways and mechanisms that trigger the SARS-CoV-2-induced hyperinflammatory response.

identification of METTL3 cofactors as TL-specific G9a interactors indicated that G9a may regulate translation via interaction with METTL3. We also identified most of the other translation regulatory proteins in Flag-tagged METTL3 pull-down from TL macrophages (Extended Data Table 2 ). For example, ribosomal proteins that bind to actively translated mRNA were identified, including forty-six 39S ribosomal proteins (Mrpl1-57), 28S (Mrps), 40S ribosomal proteins (Rps), 60S ribosomal proteins (RPl), 60S ribosome subunit biogenesis protein NIP7 homolog, ribosome-binding protein 1 (Rrbp1), ribosomal RNA processing proteins (Rrp1, Rrp36, Rrp7a, Rrp8) as well as multiple RNA-binding proteins (Fig. 1b) . These results from unbiased interactome screening 31 indicated that G9a and METTL3 function in the same translation regulatory pathways. Immunoblotting showed TL-specific association with endogenous G9a with METTL3 and major translation regulators (Fig. 1c) , in agreement with the ChaC result. In macrophages with G9a knock-down, UNC0965 pulled down only proteins that interacted with the residual G9a. These results confirmed the new function of G9a in METTL3-mediated translation in ET macrophages.

We then determined by polysome analysis 32 the effect of G9a and METTL3 on protein synthesis. First, we used CRISPR/Cas9 to knock out (ko) G9a (EHMT2) or METTL3 in macrophages subjected to LPS stimulation to generate cells with different inflammatory conditions (NL, T, or TL). G9a depletion caused reduced protein expression of METTL3 in TL/ET. Also, as indicated by reduced levels of the proinflammatory markers such as phosphor-p65 and phospho-IκB in the macrophages under prolonged LPS stimulation (Extended Data Fig.   1c ), depletion of either METTL3 or G9a similarly caused reduced ET and resensitized the ET macrophages to LPS stimulation. These results suggested that ET is coregulated by G9a and METTL3. Next, we measured global protein synthesis from polysome abundance in wild type versus G9a ko or METTL3 ko macrophages. 32 Fig. 1d shows that the polysome abundance decreased in G9a ko or G9a inhibitor(UNC0642) treated or METTL3 ko cells compared with abundance in wild type. Thus, depletion of either G9a or METTL3 suppressed protein synthesis that was dependent on G9a and METTL3 in the endotoxin-tolerant macrophages.

G9a and METTL3 co-upregulate translation of m 6 A-marked mRNA subsets. Because global protein synthesis in ET macrophages required G9a and METTL3, we next sought to identify specific mRNAs whose translation was dependent on G9a and METTL3. Thus, we performed RNA-seq and m 6 A RNA immunoprecipitation-sequencing (MeRIP-Seq) on wild type, G9a ko, and METTL3 ko THP1 macrophages in N, NL, TL. First, in the G9a-depleted macrophages under TL, 440 immune or inflammatory response genes exhibited increased mRNA abundance (Extended Data Fig. 2a) . This finding of the G9a-suppressed genes aligned with our previous report 14 that, via interactions with transcriptional repressors such as cMyc, constitutively active G9a suppresses the transcription of proinflammatory genes. Conversely, we identified, in both G9a ko and METTL3 ko cells under TL, 136 genes with decreased mRNA expression compared to wild type cells; these genes are associated mostly with regulation of cell cycle progression, cell proliferation, and antiviral or anti-inflammatory responses. Like the co-existence of G9a and METTL3 in TL-specific translation regulatory complexes, this new finding of G9a/METTL3-co-upregulated genes indicated that G9a and METTL3 cooperate to posttranscriptionally activate expression of pro-survival and anti-inflammatory genes in TL.

Further, 62 of 136 G9a/METTL3-co-upregulated genes were tagged with m 6 A (Fig. 2a) . The Integrative

Genomics Viewer plots showed that the m 6 A level of most mRNAs that encode these G9a/METTL3-coupregulated genes decreased in either G9a ko or METTL3 ko macrophages with TL (Extended Data Fig. 2b) .

Notably, the 'stabilized' or 'actively translated' mRNA (total input) that was proportional to the amount of m 6 Atagged transcripts also showed a dependence on G9a and METTL3. We validated by quantitative PCR the G9a-and METTL3-dependence of the abundance of m 6 A mRNA coding these genes (Extended Data Fig. 2c) .

To determine the predominant mechanism responsible for the G9a-mediated translation of specific genes, we examined whether G9a ko affected the m 6 A level of total RNA under N, NL, TL. We observed a relatively reduced m 6 A level in G9a ko and METTL3 ko cells (Extended Data Fig. 2d ). In addition, we performed LFQ proteomics 26 to identify proteins that exhibited G9a-or METTL3-dependent expression changes in the same THP macrophage set (e.g., wild type versus G9a ko versus METTL3). Principle component analysis (Extended Data Fig. 2e) showed that, in TL, G9a ko or METTL3 ko produced clusters of protein expression profiles that were well separated from the clusters of wild type ET macrophages. This result indicated that depletion of G9a and METTL3 led to characteristic protein expression patterns that represented similar inflammatory phenotypes. Importantly, the G9a/METTL3-co-upregulated m 6 A mRNA or proteins that we identified by nonbiased MeRIP-Seq or LFQ proteomics functionally overlapped with pathways related to antiviral immune response and cell fate determination. Immunoblotting confirmed that the protein expression of select genes in these functional clusters had a dual dependence on G9a and METTL3 (Fig. 2b) , which indicated that G9a and METTL3 coactivate translation of the same sets of m 6 A-tagged mRNA. In agreement with our ChaC-MS finding of enhanced interactions of G9a with METTL3 and associated translational regulatory proteins (Fig. 1b) , this epitranscriptomic-to-proteomic correlation for specific genes indicated that G9a is directly involved in gene-specific translation regulatory processes in ET macrophages.

G9a and METTL3 promote proliferation of ET macrophages that produce organ-damaging inflammatory factors. One cluster of the proteins co-upregulated by G9a and METTL3, BIRC5, 33 CDC20, 34, 35 NUSAP1, 36 CDCA7, 37 TOP2A, and ALCAM, is functionally associated with cell division, mitotic cell cycle, and cell proliferation 36, [38] [39] [40] [41] (Fig. 2a) . This observation not only aligned with the function of METTL3 in promoting cell cycle progression and survival 42 , but, more importantly, the observation implicated constitutively active G9a in promoting METTL3-mediated translation of prosurvival proteins, e.g., METTL3-mediated m 6 A promotes translation of cMyc mRNA and is responsible for cMyc stability in AML cells 43 .

We therefore used flow cytometry to determine the effect of G9a and METTL3 on the cell cycle in different inflammatory conditions. As shown in Fig. 2c , conditions that mimicked chronic inflammation (TL/ET) caused a dramatic accumulation of macrophages in G1 compared with nonstimulated (N) or acutely inflamed (NL) conditions (72% vs 46%). This G1 accumulation was accompanied by fewer S phase cells (23% vs 43%).

Neither G9a nor METTL3 knockout delayed the G1 phase under the TL condition. Prolonged LPS stimulation (TL) caused these G9a-or METTL3-knockout lines to spend more time in S phase relative to wild type cells and relative to the same cells under acute inflammation conditions. The clonogenic assay (Fig. 2d) showed both knockout lines survived less well than controls during TL treatment, and the G1 delay could serve a protective function. Akin to our proteomic finding that G9a/METTL3 coactivate expression of multiple prosurvival proteins, these cell cycle analyses indicated that constitutively active G9a and METTL3 cooperate to restrain cell cycle progression and promote increased survival of the slow-processing macrophages during chronic inflammation, possibly by targeting prosurvival mRNAs for translation specifically in the growth phase (G1).

G9a and METTL3 co-upregulated ET overexpression of PD-L1, and depletion of G9a or METTL3 restored T cell function. Another cluster of G9a/METTL3-co-upregulated proteins, PD-L1, CX3CR1, and IRF8, are functionally associated with immune checkpoint regulation and antimicrobial response. PD-L1 is overexpressed in sepsis patients with impaired T cell function, 27 and, likewise, our MeRIP-Seq data showed an increased level of PD-L1 (CD274) m 6 A mRNA under ET (Extended Data Fig. 2b) . Although PD-L1 m 6 A mRNA exhibited little dependence on G9a or METTL3 similar to certain METTL3-regulated genes, 20 results from LFQ proteomics and immunoblotting consistently showed that the ET overexpression of PD-L1 was dependent on G9a and METTL3 ( Fig. 2b and Extended Data Fig.2e) . Thus, we hypothesized that G9a and METTL3 impair T cell function via promoting translation or overexpression of PD-L1 in ET macrophages.

To test the hypothesis, we performed a similar T cell proliferation assay 44 to determine the effect of either G9a ko or METTL3 ko on T cell function under the TL/ET condition. We compared T-cell activation and proliferation by incubation of wild type, G9a ko, or METTL3 ko Raw cells collected in N, NL, and TL with T cells from a P14 transgenic mouse. By monitoring multiple markers of T-cell activation, including CD25, CD44, and CD69, we observed that the co-existing TL/ET wild type cells suppressed activation of CD8 T cells, whereas incubation with G9a ko or METTL3 ko cells produced efficiently activated T cells (Fig. 2e, upper panel) .

Similarly, we found a greater number of proliferated P14 CD8 T cells after six days of co-incubation of T cells with either G9a ko or METTL3 ko, whereas wild type cells lost the ability to promote proliferation of T cells in ET (Fig. 2e, lower panel) . In line with downregulation of proteins that were highly enriched in T cell activation in severe COVID-19 patients, these results showed that G9a and METTL3 impair T cell function by promoting overexpression of immune checkpoint proteins such as PD-L1.

gene-specific translation. We next investigated the function of the G9a-METTL3 interaction in ET macrophages. Briefly, we co-transfected equal amounts of HA-tagged G9a and Flag-tagged METTL3 into the LPS-responsive, TLR4/CD14/MD2 293 cells 45 under N, NL, or TL. Notably, both METTL3 and G9a (input) showed significantly increased co-expression specifically in T or TL (ET). Immunoprecipitation with anti-HA antibody confirmed TL-specific interaction of the two proteins. Importantly, the METTL3-G9a interactions also increased with prolonged LPS stimulation that mimicked ET (Fig. 3a) . Notably, the composition of the UNC0965-captured G9a interactome from ET/TL macrophages showed significant overlap with the composition of the G9a lysine methylation (Kme) proteome 46, 47 ; eight nonhistone G9a substrates were identified as ET-specific G9a interactors. These results indicated that certain G9a interactors are ET-specific G9a substrates.

Considering that some G9a interactors are ET-specific G9a substrates, we examined whether the ETspecific interaction between G9a and METTL3 implied that G9a methylates METTL3 in ET. With domaintruncated versions, we found that the catalytic domain of G9a interacts with the C-terminus (aa 201-580) of METTL3 (Fig. 3b) . We then immunoprecipitated Flag-METTL3 from the cotransfected TLR4/CD14/MD2 293 cells in TL; immunoblotting with an anti-mono-or di-methylysine antibody showed that METTL3 was methylated (Extended Data Fig. 3) . We then used LC-MS/MS to sequence the tryptic digests of Flag-METTL3 and unambiguously identified four mono-or di-methyl-lysine (Kme) sites in the C-terminus of METTL3 that directly interact with the G9a catalytic domain (Fig. 3c) . In addition, using LFQ MS to measure the differences in these Kme sites in different inflammatory conditions, we identified one Kme site (K215) that showed increased abundance in ET (Fig. 3d) , thereby confirming that these METTL3 lysines are targets of constitutively active G9a. Du et al. reported that SUMOylation of METTL3 at lysine 215 modulates the RNA N 6adenosine-methyltransferase activity of METTL3. 48 Also, E3 SUMO-protein ligases PIAS1, RanBP2, and CBX4, and a SUMO1-specific protease SENP1, were identified as ET-specific interactors of METTL3. SENP1 reduced METTL3 SUMOylation at multiple lysine sites including K215. 48 These results confirmed the propensity of METTL3 to be post-translationally modified and the functional relevance of distinct modification (e.g., lysine methylation) types in ET macrophages. Notably, the transient nature of enzyme-substrate   interaction probably explained why ChaC-MS did not unambiguously identify the endogenous G9a-METTL3 interaction.

Initiation is usually the rate-limiting step in translation, and METTL3 enhances translation of target mRNAs by recruiting eIF3 to the initiation complex. 20 We made 49 single and double, nonmethylatable mutants of METTL3, i.e., K215, K281, and K327-to-R (KXR), and used immunoblotting to compare the strength of binding of indicated proteins to the Flag-METTL3 versus Flag-K215R or K281R METTL3 mutants in the ET TLR4/CD14/MD2 293 cells. We observed that Kme absence (confirmed with Kme antibodies) weakened METTL3 interaction with eIF3 by at least 30% (Fig. 3e) , a deficiency that was expected to impair METTL3mediated translation. 20 Notably, the METTL3 interaction with the 5' cap-bound eIF4e was unaffected by the K215R or K281R mutation, a finding that agreed with the fact that METTL3 is critical for 5'UTR m 6 A mRNA to promote cap-independent translation. 50 A mechanism underlying G9a methylation-activated, METTL3mediated translation (Fig. 3f) is postulated based on results from the abovementioned combined approaches.

identifications of additional translation regulators as G9a interactors indicated that the G9a-METTL3 axis is only one part of a translation regulatory network involving G9a (Fig. 1a) . Thus, we employed our improved of AACT 51 or SILAC pulse-labeling translatome 52 strategy ( Fig. 4a and Supplementary Methods) to profile, in a nonbiased proteome-wide manner, all genes that showed G9a-dependent translation, i.e., the 'G9a-translated proteins'. Accordingly, we determined individual protein rates of synthesis, degradation, and overall turnover in snapshot samples collected at different times from cultures of wild type, G9a ko, and UNC0642 53,54 -treated macrophages. Here, genetic (G9a ko) and pharmacologic inhibition (UNC0642 treatment) of G9a decreased synthesis or turnover of G9a-translated proteins.

Briefly, cells grown with Lys0-Arg0 (K0/R0, 'light', L) were pulse-labeled with Lys4-Arg6 (K4/R6, 'medium', M) supplemented media. At 2h, 4h, 8h, 24h, 48h, and 72h after the media switch, proteins extracted from harvested cells were subjected to tryptic digestion, fractionation, and LC-MS/MS. This experimental design yields (i) increasing signals from the K4R6-labeled protein molecules due to nascent protein synthesis, and (ii) decreasing signals from K0R0-labeled proteins due to degradation or secretion of pre-existing protein molecules. Accordingly, the inhibitor-induced rate changes in nascent protein synthesis or protein degradation were quantified by the intensities of L or M labels at different time points in wild type, G9a ko, and UNC0642treated macrophages, respectively, under non-stimulated (N) or prolonged endotoxin stimulation (T) conditions (Fig. 4a) . The model assumed steady-state equilibrium conditions, in which the rate of increase was counterbalanced by the rate of decrease, leading to stable intracellular protein levels. Effects of amino acid recycling and differences in cell division rate between different conditions were also considered. Fit qualities were estimated using least-squared regression ( ܴ ଶ ), root-mean-squared error (RMSE), with additional thresholds on fitted parameters to ensure good/meaningful estimates of protein turnover (Extended Data We obtained synthesis and/or degradation half-lives for 6,243 protein groups in the combined dataset (i.e., wild-type/Ctrl, G9a-KO, UNC0642-treated cells under N and T conditions). For 78-94% of fitted proteins, information was available for both AACT-label increase and decrease, which provided an internal duplicate measurement of protein turnover time for each sample in a steady-state system ( Fig. 4b and Extended Data Table 3b ). Half-life determination was reliable across labeling pairs (R = 0.99) and cell culture replicates (R = 0.29-0.99) as evidenced by high Pearson's correlation coefficients and covariance (Extended Data Fig. 4a &   4b) . Under N and T conditions, whereas the principle component analysis hinted towards distinct proteostasis landscapes in macrophages, G9a knock-out or inhibition produced large effects on protein turnover compared with wild type macrophages (Extended Data Fig. 4c) . Consequently, we observed significant pairwise differences in global protein turnover time upon G9a inhibition as well as ET (Extended Data Fig. 4d ).

Estimated protein turnover times spanned four orders of magnitude (with some outliers), from minutes to thousands of hours. Similar to SARS-CoV-2-upregulated global translation in multiple organs of severe patients, ET also increased the rates of global translation or protein turnover as evidenced by shorter median half-lives in T (24.8-32.9 h) compared with N (34.7-38.4 h). The ET-accelerated translation rates were reduced in G9a ko or UNC0642-treated macrophages specifically under T (32.9, 26.9h) (Fig. 4c) . In agreement with the polysome results that showed G9a/METTL3-dependent protein synthesis (Fig. 1c) , this nonbiased translatome profiling indicated that constitutively active G9a accelerated global protein synthesis and degradation (i.e., global proteostasis) in ET. Table 3b ). G9a mediated methylation of a nonhistone substrate FOXO1 has already been shown to induce proteasomal degradation 57 . Therefore, our results not only support our identification of METTL3 as both G9a interactor and a nonhistone substrate of G9a ( Fig. 3c) but also indicate that G9a may upregulate gene-specific turnover by interacting with or methylating select translation regulators.

Notably, 472 G9a-translated proteins were encoded by G9a and/or METTL3-regulated m 6 A mRNAs ( Fig. 4d and Extended Data Table 3b), of which 282 proteins (~59.7%) showed more than two-fold difference in turnover time in a G9a-dependent manner (Fig. 4e) . These m 6 A mRNA-encoded, G9a-translated proteins are associated with cell cycle, cytokine/chemokine/interleukin signaling, myeloid/leukocyte activation, blood coagulation and wound healing, proteostasis including ubiquitination, localization (i.e. endocytosis, vesicle transport) signaling, organo-nitrogen metabolism, i.e., lipid, nucleotide, amino-acid synthesis, and ion transport (Fig. 4e) . Thus, combined results from ChaC-MS, m 6 ARIP-Seq, and translatome analysis validated the regulatory function of the G9a-METTL3-m 6 A axis in ET-phenotypic translation activation, which accounted for a subset of G9a-translated genes (472 out of 3994). Clearly, via interactions with distinct translation regulators other than METTL3 (Fig. 1b) , G9a coordinates additional, as yet unknown, mechanisms to facilitate genespecific translation (Fig. 4d) .

Strikingly, almost all of the G9a-translated pathways that we identified by translatome profiling (Figs 4d and 4e) have been implicated in SARS-CoV-2 life cycle and COVID-19 pathogenesis 4, 8, 10, 15, 19, 58, 59 . Indeed, we Fig 5b) . Several of these G9a-translated proteins were identified as ET-specific G9a interactors, nonhistone G9a substrates or G9a/METTL3-dependent m 6 A targets (Extended Data Fig. 5b ; also see Extended Data Table 3b ), which supported the translation regulatory function of G9a in COVID-19 pathogenesis. More importantly, genetic perturbation of several of these G9a translated host interactors adversely affected SARS-CoV-2 replication and infection 13, 30 (Fig. 4f) . Similarly, several host factors critical for SARS-CoV-2 infection identified in siRNA/CRISPR based screens are closely related to G9a complex 13, 30 (Extended Data Fig. 7 ). In line with COVID-19 hallmarks of a systemic cytokine storm, excessive infiltration of monocytes, dysregulated macrophages, and impaired T cells 4 , we observed faster turnover of proteins that belong to immune response pathways involving B-cell receptor, T-cell receptor, NK-cell, chemokine, interferon, interleukin, Jak/Stat, NF-kB and CXCR4 signaling. Further, pharmacologic inhibition of G9a downregulated these pathways by reducing the translation/turnover rates of major pathway components (Extended Data Fig.   5c ). Similarly, proteins involved in splicing, unfolded protein response, and translation initiation/elongation were upregulated following SARS-CoV-2 infection 19 . Consistent with these findings, we observed increased turnover for >150 proteins related to translation/proteostasis including EIF2/4, unfolded protein response, SUMOylation and ubiquitination signaling and RNA biogenesis such as spliceosomal cycle and RNA degradation pathways in ET macrophages, whereas G9a inhibition reversed these effects by reducing their turnover times (Extended Data Fig. 5e & 5g) . In short, constitutively active G9a regulates specific genes at the translational or posttranslational level to drive ET-related, SARS-Cov-2-induced pathogenesis, and inhibition of G9a and its associated proteins hinders coronavirus replication and infection. Thus, G9a and its associated proteins are potential drug targets to treat COVID-19 and other coronavirus-related ailments, and these targets merit further molecular and clinical study.

proteins. Interestingly, we also identified as ET-specific interactors of G9a the histone methyltransferase Ezh2 (Enhancer of zeste homolog 2) and two Polycomb Repressive Complex 2 (PRC2) components, SUZ12 and EED (Fig. 1b) . Further, we were intrigued by the G9a-dependent translation of the PRC2 complex components, which includes Ezh2, in ET (Extended Data Fig. 6c inhibitor (UNC1999) 60 with a G9a inhibitor. We used various quantitative proteomic approaches [24] [25] [26] , i.e., a multiplex TMT quantitative proteomic method to analyze the intracellular proteins collected from the pellets of N, NL, and T Raw macrophages, with or without inhibitor treatment. In parallel, LFQ proteomic approach was used to comparatively identify the proteins whose secretion showed dependence on the treatment by either G9a or Ezh2 inhibitor or both, respectively. Based on their overexpression in ET macrophages, we identified 43 proteins (Fig. 5a ) whose abundances were suppressed by either or both inhibitors. Strikingly, >80% of these proteins that showed G9a-or Ezh2-dependent overexpression in ET macrophages were associated with ARDS-related cytokine storm or clinical ARDS, with some proteins found in the sera of severe COVID patients 6 ( Fig. 5b and Extended Data Table 4 ).

Discovery of nonepigenetic function of G9a in driving SARS-CoV-2 immunopathogenesis. In 10-20% of patients, SARS-CoV-2 infection progresses to ARDS or/and severe pneumonia or multi-organ damage/failure; most of these patients are vulnerable because of pre-existing chronic conditions. The hyperinflammatory response mediated by SARS-CoV-2-dysregulated macrophage activation results in a cytokine storm, which is the major cause of disease severity and death. 6 Also, the immunologic characteristics of COVID-19, especially in severe cases that require ICU care, include reduced counts of CD3+, CD4+ and CD8+ T lymphocytes or lymphopenia, 61 and significantly increased levels of certain serum cytokines or hyperinflammation. In addition, as indicated by significantly higher levels of a T cell exhaustion marker PD-1, the surviving T cells in severe patients appeared to be functionally exhausted. We discovered a new gene-specific translation mechanism of hyperinflammation, lymphopenia and viral replication in which the constitutively active G9a/GLP interactome coordinates the networked, SARS-CoV-2-dysregulated pathways that determine COVID-19 severity.

Endotoxin-tolerant macrophages have molecular characteristics similar to chronic inflammationassociated complications (e.g., ARDS and sepsis), including downregulation of inflammatory mediators and upregulation of other antimicrobial factors. These complications systemically contribute to impaired adaptive immunity (e.g., T-cell function impairment or a poor switch to the adaptive response) and susceptibility to secondary infection with an organ-damaging cytokine storm. 11 The current model for inflammation control in ET macrophages was derived from mRNA expression studies. 62 Foster et al. reported that the chromatin modification landscape was differentially programmed in a gene-specific (pro-inflammatory versus antimicrobial genes) manner. That is, with prolonged LPS stimulation, the promoter chromatin of pro-inflammatory genes transitioned from transcriptionally activate to transcriptionally silent and LPS-induced transcripts regulated increased expression of other antimicrobial genes in ET. 62 In parallel with G9a's canonical epigenetic function for transcriptional silencing of pro-inflammatory genes, particular G9a interactors associated with chromatin regulation such as BRD2/4 were found to affect SARS-CoV-2 replication 30 . Additional G9a interactors including the components of chromatin remodeling complexes such as SMARCA2/4, SMARCC2, CREBBP, were characterized as proviral host factors essential for SARS-CoV-2 infection 13 (Fig. 1c) .

Conversely, upon restimulation, antimicrobial proteins are unlikely to be derived from time-consuming transcription processes as the turnover of mRNAs is an energy-cost-effective process to respond to an infection, which suggests that overexpression of antimicrobial genes in ET is regulated by under-characterized translation mechanisms. Coincident with its constitutive activity in ET 14 (Fig. 3f) . Fig. 1a ) which 'reshapes' the translation regulatory pathways and, in turn, activates the translation of a range of genes for SARS-CoV-2 immunopathogenesis. Briefly, Bojkova et al. 19 reported that SARS-CoV-2 infection led to increased expression of proteins associated with translation initiation and elongation, alternative splicing, mRNA processing, and nucleic acid metabolism. Likewise, in the ET macrophages that showed similar immunologic features to severe COVID-19, we found that most of these SARS-CoV-2-upregulated translation components in infected cells were ET-phenotypic interactors and nonhistone substrates of constitutively active G9a (Fig. 4) . Further, by nonbiased screening of the ET macrophage translatome with time-dependent inhibition of G9a, we identified genes that are translated at higher rates in a G9a-dependent manner as 'G9atranslated proteins'. Clustered by functional pathway enrichment, ~4,000 G9a-translated proteins were merged into two major networks associated, respectively, with the host response to SARS-CoV-2 infection (Fig. 4e ) or host-virus interactions (Fig. 4f) . For example, COVID-19 severity was linked to multiple SARS-CoV-2-induced, G9a-translated host response pathways, primarily associated with immune complement and coagulation dysregulation 15 , IFNs-and IL-6-dependent inflammatory responses 63 , and ER-associated degradation (ERAD) 64 (Fig. 4e) . The G9a-translated components of complement and coagulation pathways including C5aR1, SERPINE1, CR1L were upregulated in severe patients 65 . Overexpression of the C5a-C5aR1 axis was associated with ARDS in COVID-19 patients 66 . Specifically, increased PD-L1 levels in monocytes and dendritic cells 67 and elevated levels of C5aR1 in blood and pulmonary myeloid cells 59 contribute to COVID-19characteristic hyperinflammation and ARDS. G9a inhibition reduced the turnover rates of both proteins in ET macrophages (Extended Data Fig. 4h) . These COVID-19-associated networks composed of G9a-translated proteins provide a compelling rationale to suspect that G9a inhibition will adversely affect SARS-CoV-2upregulated pathways associated with not only the impaired host response but also with viral infection and replication.

Inhibitor mechanism of action for severe COVID-19 therapeutics. Once immunopathologic complications such as ARDS or pneumonia occur, particularly in patients with pre-existing chronic inflammatory diseases, antiviral treatment alone is less effective and should be combined with appropriate antiinflammatory treatment. However, current immunomodulatory measures failed to improve prognosis of patients vulnerable to a cytokine storm. 68 19 , as ET-phenotypic G9a interactors, G9a-targeted therapy can be combined with Rps14-or SF3B1-inhibition antiviral therapy to improve the efficacy of single target therapy. Also, because an Ezh2 inhibitor showed inhibitory effects similar to the effects of a G9a inhibitor, the Ezh2 inhibitors with proven safety for cancer therapy can be repurposed for COVID-19 therapy.

The ability to develop targeted therapies to minimize mortality of severe patients depends on a detailed understanding of SARS-CoV-2-dysregulated chronic inflammatory pathways that determine COVID-19 severity.

Our results from chemoproteomics, MeRIP-seq, translatome proteomics, and molecular/cell biology revealed a novel mechanism of inhibitor action on COVID-19. Specifically, via ET-phenotypic G9a-interacting translation machinery, SARS-CoV-2 may evolve a G9a-associated mechanism of gene-specific translation activation to modulate host response, evade the host immune system, and promote viral replication and infection. In endotoxin-tolerant macrophages that share similar immunopathologic characteristics with SARS-CoV-2 dysregulated macrophages, we dissected the G9a-associated pathways that cause individuals with preexisting chronic inflammatory diseases to be highly susceptible to secondary infection by SARS-CoV-2.

Specifically, we discovered a widespread translational function of constitutively active G9a in the impairment/depletion of T cell function and the production of organ-damaging cytokine storm. Importantly, combined results from COVID-19 patients with pre-existing conditions, genome-wide CRISPR screening, and CoV-host interactome mapping validated our mechanistic findings in ET and identified numerous G9a interactors or G9a-translated proteins in the interconnected networks associated, respectively, with the host response to SARS-CoV-2 infection or host-virus interactions. Compared with current trials focused on either antiviral or anti-inflammatory therapy, this single-target inhibition of G9a-associated pathways was responsible for multifaceted, systematic effects, namely, restoration of T cell function to overcome lymphopenia, mitigation of hyperinflammation, and suppression of viral replication. Our studies have paved new roads that combine immunomodulatory and antiviral therapies with more effective COVID-19 therapies with minimal side effects.

Importantly, because both G9a and its complex components were overexpressed in infected host cells by other CoV strains (e.g., SARS-CoV-1 and MERS-CoV 30 ), our G9a-targeted therapy is refractory to complications induced by emerging antiviral-resistant mutants of SARS-CoV-2, or any virus, that hijacks host responses. W., C. G., and K. T. assisted with clinical analysis. Y.X. and J.J. provided UNC0965, UNC0642, and UNC1999. X.C. conceived and designed the project and experiments, analyzed and interpreted data, and wrote the manuscript. 

Cell culture media, other components, and fetal bovine serum were obtained from Gibco. Trypsin was purchased from Promega. LPS (Escherichia coli 0111:B4, ultrapure) was purchased from InvivoGen (USA, Invivogen, cat# tlrl-pelps). HEK 293 stable TLR4-MD2-CD14 cell line was also purchased from InvivoGen. All chemicals were HPLC-grade unless specifically indicated. Raw264.7 and THP1 cells were purchased from ATCC (Manassas,VA). UNC0642 (G9a inhibitor) and UNC1999 (EZH2 inhibitor) 60 

Plasmid pHA-G9a was described 12 . pFlag-CMV2-METTL3, Plasmid pFlag-ms2-METTL3 and truncated expression plasmids were kindly provided by Dr. Richard I. Gregory 17 . Full-length pFlag-G9a-SBP and truncated plasmids (D1-D4) were kindly provided by Dr. Xiaochun Yu 69 .

Raw264.7 cells were cultured in DMEM medium (Gibco). The human monocytic cell line THP-1 was maintained in RPMI 1640 medium (Gibco). All media were supplemented with 10% fetal bovine serum (Gibco), 100 U/ml penicillin and streptomycin. Cells were grown at 37 °C in humidified air with 5% carbon dioxide. For ChaC pull-down experiments, Raw264.7 cells were either unstimulated ('N') or subjected to a single LPS stimulation with 1 µg/ml ('NL') or first primed with 100 ng/ml LPS to induce endotoxin tolerance for 24h ('T'), followed by a second LPS challenge at 1 mg/ml ('TL'). For the condition requiring G9a inhibitor treatment, 1 µM UNC0642 was added at the time of cell plating. For global protein profiling, Raw264.7 cells were pre-treated with 0.1 μg/ml LPS for 24 hours, then inhibitor UNC1999 (1 µM) or UNC0642 (1 µM) was added. The cells were collected after inhibitor treatment for 8 h, 24 h and 48 h. For MeRIP-Seq, THP-1 cells were first incubated in the presence of 60 nM PMA overnight to differentiate into macrophages followed by 48 h resting in PMA-free medium. Cells were either left unstimulated ('N') or subjected to a single LPS stimulation at 1 µg/ml ('NL'), or first primed with 100 ng/ml LPS to induce endotoxin tolerance for 24 h ('T'), followed by the second LPS challenge at 1 µg/ml ('TL').

Similar to our recent report, 16 1 mg nuclear protein extracted from Raw 264.7 macrophage cells was incubated overnight at 4 o C with 2 nmole UNC0965 pre-coupled to 50 μ l neutravidin-agarose (Thermo Fisher), and washed three times with 1 ml lysis buffer to remove non-specific bound proteins. For on-beads sampling and processing, five additional washes with 50 mM Tris-HCl pH 8.0, 150 mM NaCl were used to remove residual detergents. On-beads tryptic digestion was performed with 125 μ l buffer containing 2 M urea, 50 mM Tris-HCl pH 8.0, 1 mM DTT, 500 ng trypsin (Promega) for 30 min at room temperature on a mixer (Eppendorf). The tryptic digests were eluted twice with a 100 μ l elution buffer containing 2 M urea, 50 mM Tris-HCl pH 8.0, 5 mM iodoacetamide. Combined eluates were acidified with trifluoroacetic acid at final concentration of 1% (TFA, mass spec grade, Thermo Fisher) and desalted with a C18 stage tip.

Cell pellets were resuspended in 8 M Urea, 50 mM Tris-HCl pH 8.0, reduced with dithiothreitol (5 mM final) for 30 min at room temperature, and alkylated with iodoacetamide (15 mM final) for 45 min in the dark at ambient temperature. Samples were diluted 4-fold with 25 mM Tris-HCl pH 8.0, 1 mM CaCl 2 and digested with trypsin at 1:100 (w/w, trypsin : protein) ratio overnight at room temperature. Peptides were desalted on homemade C18 stage tips. Each peptide sample (100 μg) was labeled with 100 μg of TMT reagent following the optimal protocol 70 . The mixture of labeled peptides was desalted and fractionated into 12 fractions in 10 mM TMAB containing 5-40% acetonitrile.

METTL3 KO, G9A KO, and Control Raw 264.7 cells were cultured in 10 cm plates until 80% confluent at the time of harvest. Cells were treated with 100 µg/ml cycloheximide (CHX; Sigma) for 10 min at 37 0 C. Media were removed and the cells were washed twice with 10 ml PBS containing 0.1 mg/ml CHX, scraped, pelleted by spinning for 10 min at 2200rpm at 4 0 C. After removing the supernatant, the cells were resuspended with 1ml lysis buffer (20 mM Tris-HCl, pH 7.4, 140 mM KCl, 5 mM MgCl 2 , 1% Triton X-100, 10 mM DTT) containing 0.1 mg/ml CHX and swelled on ice for 10 min followed by passing through a 27 gauge needle 5 times to break the cell membrane. Spin down the lysate at max speed for 10 min, the supernatant were carefully layered onto 10-50% sucrose gradients and centrifuged at 32,000rpm (no brake) in a Beckman SW-40 rotor for 2 h at 4 0 C. Gradients were fractionated and monitored by absorbance 254 nm.

293TLR4-MD2-CD14 cells were seeded in 6-well plates for 24 h before transfection, and the constructs were transfected or co-transfected into MCF7 cells using reagent jetPRIME (Polyplus). After 24 h, the cells were lysed directly in the plates by adding SDS-PAGE sample buffer, heating at 95°C for 5 min, and sonicating for 5 seconds to clear the lysate for immunoblotting. For the CRISPR/Cas9 constructs, oligonucleotides for the sgRNA of human and mouse G9a and METTL3 as described below, were annealed and cloned in BsmBI-digested lentiCRISPRv2 (Addgene plasmid #52961). The empty vector was used as a negative control. Viral production was performed with a standard protocol. In brief, a total of 10 μ g of plasmid, including target plasmid, pMD2.G (Addgene plasmid #12259) and psPAX2 (Addgene plasmid #12260) with a ratio of 10:5:9, was co-transfected into 293T cells with jetPRIME™. Viruscontaining media were collected 48 hours after transfection. MDA-MB-231 cells at 60-80% confluency were incubated with the virus containing media for 24-48 hours, and then subjected to 1.0 μ g/mL puromycin selection. After 4-7 days puromycin selection, the stably transfected cells were collected for further analysis.

Total RNA was isolated using RNeasy kit (Qiagen). First-strand cDNA was synthesized by M-MLV reverse transcriptase (Promega, car# M170B) and diluted by a factor of 10 for quantitative PCR. Real-time PCR was performed using SYBR Green Master Mix (Thermo Fisher, cat# 0221). All measurements were normalized to GAPDH and represented as relative ratios. PCR primers for real-time PCR are summarized in Supplementary Table. The m6A level of isolated RNA was measured by using the m6A RNA methylation quantification kit.

For flow cytometry analyses of cell cycle, cells were incubated with 10 uM EdU (Santa Cruz, sc-284628) for 30 minutes before harvesting by trypsinization. Cells were washed with PBS, and fixed with 4% paraformaldehyde (Electron Microscopy Sciences, 1571s) in PBS for 15 min at room temperature. Then 1% BSA-PBS was added and the cells were centrifuged. Fixed cells were permeabilized with 0.5% Triton X-100 in 1% BSA-PBS at room temperature for 15 min, then centrifuged. Cells were processed for EdU conjugation with 1 µM AF647-azide (Life Technologies, A10277) in 100 mM ascorbic Acid, 1 mM CuSO4, and PBS for 30 minutes at room temperature in the dark. Lastly, cells were washed and incubated in1 µg/mL DAPI (Life Technologies, D1306) overnight at 4°C. Samples were run on an Attune NxT (Beckman Coulter) and analyzed with FCS Express 7 (De Novo Software).

P14 CD8+ T cell proliferation and activation assay was performed as described previously 44 . Briefly, CD8+ T cells from a P14 transgenic mouse were first isolated with CD8a microbeads according to manufacturer's instruction. Isolated CD8+ T cells were resuspended in 1 mL 1640 medium and either labeled with 1 mL of the 10 μ M Carboxyfluorescein diacetate succinimidyl ester (CFSE) for 8 min at RT for proliferation assay or kept unlabeled for activation assay. METTL3 KO, G9a KO, and control Raw 264.7 cells were seeded in 6-well plates with indicated treatments followed by 50 μ g/mL Mitomycin C treatment for 30 min at 37 °C. Cells were washed with 1 mL PBS twice and then labeled with 0.4 mL of the 30 μ g/mL GP33-41 peptide in PBS for 30 min at 37 °C. One hundred thousand P14 T cells were co-cultured with 2 × 10 5 control, METTL3 KO, and G9a KO cells in 96-well plates in RPMI medium containing 50 U/mL mIL-2. The proliferation and activation of CD8+ T cells were assessed by flow-cytometry at day six.

For ChaC pull-down samples, desalted peptide mixtures were dissolved in 30 μ l 0.1% formic acid (Thermo Fisher). Peptide concentration was measured with Pierce™ Quantitative Colorimetric Peptide Assay (Thermo Fisher). In the Easy nanoLC-Q Exactive HFX setup, peptides were loaded on to a 15 cm C18 RP column (15 cm × 75 μ m ID, C18, 2 μ m, Acclaim Pepmap RSLC, Thermo Fisher) and eluted with a gradient of 2-30% buffer B at a constant flow rate of 300 nl/min for 30 min followed by 30% to 45% B in 5 min and 100% B for 10 min. The Q-Exactive HFX was also operated in the positive-ion mode but with a data-dependent top 20 method. Survey scans were acquired at a resolution of 60,000 at m/z 200. Up to the top 20 most abundant isotope patterns with charge ≥ 2 from the survey scan were selected with an isolation window of 1.5 m/z and fragmented by HCD with normalized collision energies of 27. The maximum ion injection time for the survey scan and the MS/MS scans was 100 ms, and the ion target values were set to 3e6 and 1e5, respectively. Selected sequenced ions were dynamically excluded for 30 seconds. For global protein profiling, 0.5 μg of each fraction was analyzed on a Q-Exactive HF-X coupled with an Easy nanoLC 1200 (Thermo Fisher Scientific, San Jose, CA). Peptides were loaded on to a nanoEase MZ HSS T3 Column (100Å, 1.8 µm, 75 µm x 150 mm, Waters). Analytical separation of all peptides was achieved with 100min gradient. A linear gradient of 5 to 10% buffer B over 5 min, 10% to 31% buffer B over 70 min and 31% to 75% buffer B over 15 minutes was executed at a 300 nl/min flow rate followed a ramp to 100%B in 1 min and 9-min wash with 100%B, where buffer A was aqueous 0.1% formic acid, and buffer B was 80% acetonitrile and 0.1% formic acid.

Peptides were separated with 45-min gradient, a linear gradient of 5 to 30%B over 29 min, 30 to 45%B over 6 min followed a ramp to 100%B in 1 min and 9-min wash with 100%B. LC-MS experiments were also performed in a data-dependent mode with full MS (externally calibrated to a mass accuracy of <5 ppm and a resolution of 120,000 for TMT-labeled samples or 60,000 for secretome samples at m/z 200) followed by high energy collision-activated dissociation-MS/MS with a resolution of 45,000 for TMT-labeled global samples and 15,000 for secretome samples at m/z 200. High energy collision-activated dissociation-MS/MS was used to dissociate peptides at a normalized collision energy of 32 eV (for TMT-labeled sample) or 27 eV in the presence of nitrogen bath gas atoms. Dynamic exclusion was 45 or 20 seconds. Each fraction was subjected to three technical replicate LC-MS analyses. There were two biological replicates of samples and two technical replicates were executed for each sample.

Mass spectra were processed, and peptide identification was performed using the MaxQuant software version 1.6.10.43 (Max Planck Institute, Germany). All protein database searches were performed against the UniProt human protein sequence database (UP000005640). A false discovery rate (FDR) for both peptide-spectrum match (PSM) and protein assignment was set at 1%. Search parameters included up to two missed cleavages at Lys/Arg on the sequence, oxidation of methionine, and protein N-terminal acetylation as a dynamic modification. Carbamidomethylation of cysteine residues was considered as a static modification. Peptide identifications are reported by filtering of reverse and contaminant entries and assigning to their leading razor protein. The TMT reporter intensity found in MaxQuant was for quantitation. Data processing and statistical analysis were performed on Perseus (Version 1.6.0.7). Protein quantitation was performed using TMT reporter intensity found in MaxQuant and a one-sample t-test statistics on three technical replicates was used with a pvalue of 5% to report statistically significant protein abundance fold-changes. Label-free quantification (LFQ) was for ChaC interactome and secretome data analysis.

The biological processes and molecular functions of the G9a-interacting proteins were categorized by IPA (http://www.ingenuity.com/), DAVID (http://david.abcc.ncifcrf.gov/), and STRING (http://string-db.org/).

m 6 A-RIP-Seq was performed as described previously with slight modifications 71 . Messenger RNA from 10 ug total RNA extracted from Ctrl, METTL3 KO and G9a KO cell samples was purified with Dynabeads Oligo (dT) 25 (Thermo Fisher; 61006). Ten percent of 150 ng mRNA was used as Input mRNA, and the remainder was incubated with 3 ug anti-m 6 A polyclonal antibody (Synaptic Systems; 202003) which was preconjugated to Dynabeads Protein A (Thermo Fisher; 10001D) in 500uL IP buffer (50 mM Tris, pH 7.4, 150 mM NaCl, 0.1% Igepal CA-630) for 2 hours at 4°C. After washing twice with IP-buffer and twice with High-Saltwash buffer (50 mM Tris pH 7.4, 500 mM NaCl, 0.1% Igepal CA-630) for 5 minutes each, the m6A mRNA was eluted with 100 uL IP-buffer containing 6.7 mM N 6 -Methyladenosine (Sigma-Aldrich; M2780) and 40 U RNase Inhibitor (NEB, M0314S) and then recovered with RNA Clean and Concentrator-5 spin columns (Zymo; R1015).

The Input mRNA and m 6 A-IPed mRNA were subjected to library generation using the SMART-seq protocol as described , and only uniquely mapping reads at the exon level for each gene were quantified and summarized to gene counts, which were further analyzed in R v.3.6.2. After normalization, sorting, and indexing with Samtools-1.1 software 72 , the corresponding BAM files for each sample were loaded to IGV software to generate the peaks plots.

G9a KO and control Raw 264.7 cells were cultured in AACT/SILAC DMEM medium supplemented with regular lysine and arginine (K0R0) and 10% dialyzed fetal bovine serum (Thermo Fisher), 1% penicillin, and streptomycin for five passages. G9a KO, control and (1 µM) UNC0642 treated control cells were either untreated (N) or treated with low dose (100 ng/ml) LPS to induce endotoxin tolerance. The cells were washed with PBS twice to remove light medium (K0R0), and then switched to heavy AACT DMEM medium containing stable isotope-enriched D 4 -lysine and 13 C 6 -arginine to label newly synthesized proteins. The cells were harvested at 2h, 4h, 8h, 24h, 48h, and 72h, and lysed in 8 M urea containing 50 mM Tris-HCl pH 8.0. One hundred micrograms protein from each condition was digested with trypsin, desalted, and fractionated with C18 material (High pH) into eight fractions followed by LC-MS/MS analysis.

Supplementary Methods of AACT Pulse-labeling Translatome Analysis: Detailed description of kinetic model used for curve fitting and applied methods for data normalization

Mathematical equations used to model synthesis and degradation of proteins in cell-lines/conditions with different doubling times were inspired from mathematical models described previously [73] [74] [75] [76] and adapted to yield interpretable analytical solutions for our system. Briefly, we assumed that there is a source of amino acids and a pool of degraded waste products, both of which contain medium and heavy amino acids. Assuming we switched from medium to heavy label (i.e.,

՜ ‫ܪ‬ medium switch at ‫ݐ‬ ൌ 0 ), the heavy source pool was assumed to be the inexhaustible medium in which the cells were grown ‫ݐ(‬ 0 ). Additionally, there is a contamination of the heavy amino acid source pool via recycling so that fraction

The following items are other assumptions of the model include:

Proteins are synthesized at a constant rate (ܵ) using medium and heavy amino acids from the source pool. Therefore, synthesis is a zero-order process with respect to protein concentration b)

Proteins are degraded at constant rate (݇) into the degradation pool. Probability of a protein being degraded is the same for old and newly synthesized protein molecules and remains constant during the lifetime of these proteins c)

Rate of cell division ( ݇ ௗ ௩ ሻ is constant (for each cell-line/condition) and contributes to overall degradation and synthesis rates (i.e., a protein with degradation rate of 0 will still show a loss of pre-existing M label because of its distribution equally to daughter cells. Similar thing occurs for newly synthesized proteins as intensity of H-label will increase owing to cell doubling even if protein's per cell concentration is constant) d)

Cells are at steady state. The concentration per cell of any given protein remains constant during the experiment. Therefore, synthesis and degradation rates of a protein are equal (i.e.,

) and first order labelling kinetics were adopted for curve fitting and intensities were fitted to exponential equations.

All model equations/diagrams are for a protein ݅ with the index dropped for clarity. Master equations describing the system are as follows:

where ‫ܯ‬ ൌ ‫ܯ‬ ሺ ‫ݐ‬ ሻ and ‫ܪ‬ ൌ ‫ܪ‬ ሺ ‫ݐ‬ ሻ are dimensionless abundances of medium and heavy proteins, respectively.

In steady state we expect the contamination fraction to stabilize at a certain level,

. In such case the equations (1) have a trivial solution: 

We require a stationary solution where the total amount of protein is constant, . Therefore, using equations (2) and (3), we can find ‫ܥ‬ ଵ and ‫ܥ‬ ଶ and plugging their value back into equations in (2), we get: ሻ . In such case the first equation of (1) assumes the form

ሻ ‫ܯ‬ (11) which can be solved analytically

ሻ (12) and in a case of balanced synthesis and degradation (ܵ

This is a modified exponential decay with a constant asymptotic offset, ߛ ஶ . The shape of this curve resembles a simple exponential decay with an offset. The offset, ߛ ஶ , is interpreted again as the asymptotic contamination fraction of the source pool.

Cell lysates were collected at 2h, 4h, 8h, 24h, 48h and 72h from wildtype (Ctrl), G9a knockout (G9a-KO) and 1 ߤ M UNC0642-treated Raw 264.7 cells under N and T conditions, digested with trypsin, fractionated by HPLC and finally subjected to LC-MS/MS analysis. To correct for errors introduced during sample-preparation/dataacquisition, protein intensities were normalized based on the premise that the sum of K0/R0 (M) and the K4/R6 (H) intensities should be constant across different time-points. After normalization, curves for estimation of turnover rates, determined from the kinetics of AACT label incorporation or loss, were fitted based on the assumption of exponential protein degradation (or synthesis) and constant offset (γ). Proteins identified in at-least 4 out of 6 time points were selected for curve fitting using the model described above. The turnover rates ( Points where synthesis and degradation curves intersect depict apparent half-life which is corrected for differences in cell division rates (between the 6 conditions) to yield actual protein synthesis and degradation rates for the protein (shown in the table ) were performed to identify 3994 proteins that were translated in G9a dependent manner in N and T condition. G9a/ET translated proteins (>2FC increase (red)/decrease (blue) in half-life) are highlighted. f. Results of canonical pathway analysis for genes highlighted in (e) are summarized. Sizes of the circles depicts number of proteins in each pathway, and color depicts log transformed p-value. G9a primarily controls turnover of proteins involved in immunity, cell cycle, translation/proteostasis, cellular energetics, RNA biogenesis and coronavirus related pathways. G9a inhibition appears to reverse the effect of ET on global proteostasis landscape. g. G9a may control global protein turnover by multiple pathways because several of the differentially translated proteins are G9a interactors, substrates or m 6 A regulated targets. h. 

Clinical Characteristics of Hospitalized Covid-19 Patients

SARS-CoV-2 receptor networks in diabetic kidney disease, BK-Virus nephropathy and COVID-19 associated acute kidney injury

COVID-19: consider cytokine storm syndromes and immunosuppression

The trinity of COVID-19: immunity, inflammation and intervention

Reduction and Functional Exhaustion of T Cells in Patients With Coronavirus Disease 2019 (COVID-19)

Proteomic and Metabolomic Characterization of COVID-19 Patient Sera

Alveolar macrophage dysfunction and cytokine storm in the pathogenesis of two severe COVID-19 patients

Single-cell landscape of bronchoalveolar immune cells in patients with COVID-19

Pathological inflammation in patients with COVID-19: a key role for monocytes and macrophages

Viral sepsis is a complication in patients with Novel Corona Virus Disease (COVID-19)

Recent advances in endotoxin tolerance

Pathophysiology of endotoxin tolerance: mechanisms and clinical consequences

Genome-wide CRISPR screens reveal host factors critical for SARS-CoV-2 infection

A chromatin activity-based chemoproteomic approach reveals a transcriptional repressome for gene-specific silencing

Immune complement and coagulation dysfunction in adverse outcomes of SARS-CoV-2 infection

Multi-omic Dissection of Oncogenically Active Epiproteomes Identifies Drivers of Proliferative and Invasive Breast Tumors

Analysis of the human endogenous coregulator complexome

The BioPlex Network: A Systematic Exploration of the Human Interactome

Proteomics of SARS-CoV-2-infected host cells reveals therapy targets

The m(6)A Methyltransferase METTL3 Promotes Translation in Human Cancer Cells

Mettl3 Deficiency Sustains Long-Chain Fatty Acid Absorption through Suppressing Traf6-Dependent Inflammation Response

Curcumin Attenuates Lipopolysaccharide-Induced Hepatic Lipid Metabolism Disorder by Modification of m(6) A RNA Methylation in Piglets

METTL3 regulates alternative splicing of MyD88 upon the lipopolysaccharide-induced inflammatory response in human dental pulp cells

Relative and Absolute Quantitation in Mass Spectrometry-Based Proteomics

Novel RNA-Affinity Proteogenomics Dissects Tumor Heterogeneity for Revealing Personalized Markers in Precision Prognosis of Cancer

Label-free Quantitative Proteomics Reveals a Role for the Mycobacterium tuberculosis SecA2 Pathway in Exporting Solute Binding Proteins and Mce Transporters to the Cell Wall

PD-L1 Overexpression During Endotoxin Tolerance Impairs the Adaptive Immune Response in Septic Patients via HIF1alpha

Decreased expression of the fractalkine receptor CX3CR1 on circulating monocytes as new feature of sepsis-induced immunosuppression

Virus-Host Interactome and Proteomic Survey Reveal Potential Virulence Factors Influencing SARS-CoV-2 Pathogenesis

Comparative host-coronavirus protein interaction networks reveal pan-viral disease mechanisms

Reversing SKI-SMAD4-mediated suppression is essential for TH17 cell differentiation

The 5' Untranslated Region of the Major Immediate Early mRNA Is Necessary for Efficient Human Cytomegalovirus Replication

BIRC5 (survivin): a pejorative prognostic marker in stage II/III breast cancer with no response to neoadjuvant chemotherapy

CDC20, a potential cancer therapeutic target, is negatively regulated by p53

Cdc20: a potential novel therapeutic target for cancer treatment

NUSAP1 promotes invasion and metastasis of prostate cancer

The MYC-associated protein CDCA7 is phosphorylated by AKT to regulate MYC-dependent apoptosis and transformation

SKA3 promotes cell proliferation and migration in cervical cancer by activating the PI3K/Akt signaling pathway

The prognostic significance of Cdc6 and Cdt1 in breast cancer

Overexpression of denticleless E3 ubiquitin protein ligase homolog (DTL) is related to poor outcome in gastric carcinoma

B-Myb): a central regulator of cell proliferation, cell survival and differentiation involved in tumorigenesis

METTL3 plays multiple functions in biological processes

The N(6)-methyladenosine (m(6)A)-forming enzyme METTL3 controls myeloid differentiation of normal hematopoietic and leukemia cells

Proteomic dissection of LPS-inducible, PHF8-dependent secretome reveals novel roles of PHF8 in TLR4-induced acute inflammation and T cell proliferation

Flightless I homolog negatively modulates the TLR pathway

A general molecular affinity strategy for global detection and proteomic analysis of lysine methylation

Defining efficient enzyme-cofactor pairs for bioorthogonal profiling of protein methylation

SUMOylation of the m6A-RNA methyltransferase METTL3 modulates its function

Non-canonical Bromodomain within DNA-PKcs Promotes DNA Damage Response and Radioresistance through Recognizing an IR-Induced Acetyl-Lysine on H2AX

5' UTR m(6)A Promotes Cap-Independent Translation

Amino acid residue specific stable isotope labeling for quantitative proteomics

Global analysis of cellular protein translation by pulsed SILAC

Inhibition of EHMT1/2 rescues synaptic and cognitive functions for Alzheimer's disease

Targeting the histone methyltransferase G9a activates imprinted genes and improves survival of a mouse model of Prader-Willi syndrome

Bioorthogonal profiling of protein methylation using azido derivative of S-adenosyl-Lmethionine

Defining efficient enzyme-cofactor pairs for bioorthogonal profiling of protein methylation

FOXO1 degradation via G9a-mediated methylation promotes cell proliferation in colon cancer

Systems biological assessment of immunity to mild versus severe COVID-19 infection in humans

Association of COVID-19 inflammation with activation of the C5a-C5aR1 axis

An orally bioavailable chemical probe of the Lysine Methyltransferases EZH2 and EZH1

Deep immune profiling of COVID-19 patients reveals distinct immunotypes with therapeutic implications

Gene-specific control of inflammation by TLR-induced chromatin modifications

The Global Phosphorylation Landscape of SARS-CoV-2 Infection

Coronavirus interactions with the cellular autophagy machinery

Multi-organ Proteomic Landscape of COVID-19 Autopsies. medRxiv

Association of COVID-19 inflammation with activation of the C5a-C5aR1 axis

Disharmonic Inflammatory Signatures in COVID-19: Augmented Neutrophils' but Impaired Monocytes' and Dendritic Cells' Responsiveness

Transmission, Diagnosis, and Treatment of Coronavirus Disease 2019 (COVID-19): A Review

Correction: The zinc finger proteins ZNF644 and WIZ regulate the G9a/GLP complex for gene repression

TMT labeling for the masses: a robust and cost-efficient, in-solution labeling approach

Epitranscriptomic m6A regulation of axon regeneration in the adult mammalian nervous system

The sequence alignment/map format and SAMtools

A quantitative spatial proteomics analysis of proteome turnover in human cells

Time-resolved analysis of proteome dynamics by tandem mass tags and stable isotope labeling in cell culture (TMT-SILAC) hyperplexing

Dynamic profiling of the protein life cycle in response to pathogens

Proteome-wide analysis of protein abundance and turnover remodelling during oncogenic transformation of human breast epithelial cells

NL: 1.71E5 20180311_Met1_2#16298 RT: 20.93 AV: 1 T: FTMS + c NSI d Full ms2 589.5616@hcd27.00 [161.6667-2425.0000] NL: 7.52E5 20180311_Met1_2#16286 RT: 20.91 AV: 1 T: FTMS + c NSI d Full ms2 586

. Hence, we can rewrite equations in (4):These are trivial equations of pure degradation and synthesis with exponential solutions of the form (at ‫ݐ‬ ൌ 0 ) as shown:From this we can find an approximate relation between half-life and the degradation/synthesis coefficient given by: Substituting these into the original equations (1) yields