SARS-CoV-2 protein encoded by ORF8 contains a histone mimic that disrupts chromatin regulation

SARS-CoV-2 emerged in China at the end of 2019 and caused the global pandemic of COVID- 19, a disease with high morbidity and mortality. While our understanding of this novel virus is 24 rapidly increasing, gaps remain in our understanding of how SARS-CoV-2 can effectively suppress host cell antiviral responses. Recent work on other viruses has demonstrated a novel mechanism through which viral proteins can mimic critical regions of human histone proteins. Histone proteins are responsible for governing genome accessibility and their precise regulation 28 is critical for a cell’s ability to control transcription and respond to viral threats. Here, we show that 29 the protein encoded by ORF8 (Orf8) in SARS-CoV-2 functions as a histone mimic of two critical 30 histone 3 sites containing an ARKS motif. Orf8 expression in cells disrupts multiple critical histone 31 post-translational modifications (PTMs) while Orf8 lacking this histone mimic motif does not. Orf8 32 binds to numerous histone-associated proteins and to DNA, and is itself acetylated within the 33 histone mimic site. Importantly, SARS-CoV-2 infection of multiple susceptible cell types causes 34 the same global changes of histone post-translational modifications that are disrupted by Orf8 35 expression; these include induced pluripotent stem cell-derived alveolar type 2 cells (iAT2) and 36 cardiomyocytes (iCM) and postmortem patient lung tissue. These findings demonstrate a novel 37 function for the poorly understood SARS-CoV-2 ORF8 encoded protein and a mechanism through 38 which SARS-CoV-2 disrupts host cell epigenetic regulation. Notably, this work provides a potential 39 mechanism for emerging findings from human patients indicating that ORF8 deletion results in 40 less severe illness and describes a potentially druggable pathway that may contribute to the 41 virulence of SARS-CoV-2. SARS-CoV-2 MOI=5. ChIP-RX

2 suppress host cell antiviral responses. Recent work on other viruses has demonstrated a novel 26 mechanism through which viral proteins can mimic critical regions of human histone proteins.

27
Histone proteins are responsible for governing genome accessibility and their precise regulation 28 is critical for a cell's ability to control transcription and respond to viral threats. Here, we show that

56
While SARS-CoV-2 likely employs numerous mechanisms to dampen this response, we 57 examined whether SARS-CoV-2 employs histone mimicry to disrupt histone regulation, to better 58 understand how it evades host cell antiviral responses.

60
To investigate whether histone mimicry is utilized by the SARS-CoV-2 virus, we first performed a 61 bioinformatic comparison of all SARS-CoV-2 viral proteins 17 with all human histone proteins ( Fig.   62 S1a). Most SARS-CoV-2 proteins are highly similar to those in the coronavirus strain that caused 63 the previous major SARS-CoV outbreak with the notable exception of proteins encoded by ORF3b 64 and ORF8 18 . Remarkably, we detected an identical match between a region of the protein 65 encoded by ORF8 (henceforth called Orf8) and critical regions within the histone H3 amino 66 terminal tail ( Fig. 1a-b, S1a-b). Furthermore, Orf8 aligns to a longer stretch of amino acids (6 67 identical, sequential amino acids) than any previously described and validated case of histone 68 mimicry 4,6,7,19,20 (Fig. S1c). Based on a crystal structure of Orf8, this region of the protein falls in 69 a disordered region that is potentially exposed to the cell in an Orf8 monomer 21 . Most compelling 70 is that the motif we detected contains the 'ARKS' sequence, which is found at two distinct sites in 71 the histone H3 tail and is well-established as one of the most critical regulatory regions within H3.

72
Both of these ARKS sites are modified with multiple crucial PTMs, including methylation and 73 acetylation at H3 Lysine 9 (H3K9me3 and H3K9ac) and at H3 Lysine 27 (H3K27me3 and 74 H3K27ac). Strikingly, this amino acid stretch is absent from the previous SARS-CoV virus Orf8-75 encoded protein (both before and after a deletion generated two distinct peptides, Orf8a and 4 Orf8b 22 ) (Fig. S1d). Our proposed histone mimicry motif is also a considerably closer match than 77 a previously proposed histone mimic in protein E of SARS-CoV-2 ( Fig. S1e) 23 . These findings 78 indicate that Orf8 may act as a histone mimic to disrupt regulation of ARKS sites on histone H3, 79 providing a novel mechanism through which this relatively poorly understood and highly divergent 80 protein 24-26 functions during infection.

82
To determine whether Orf8 may act as a histone mimic, we examined whether Orf8 expression 83 disrupts histone PTM regulation using an unbiased mass spectrometry approach. HEK cells were 84 transfected with Orf8 containing a Strep tag or with a control GFP-expressing plasmid and 85 transfected cells were isolated using fluorescence-activated cell sorting (FACS). Histones were 86 purified using an acid-extraction method, and bottom-up unbiased mass spectrometry was 87 performed to quantify all detected histone PTMs. Fitting with its potential role as a histone mimic, 88 we found that numerous histone modifications were disrupted in response to Orf8 expression 89 (Table S1). We focused on significantly disrupted histone PTMs with well-defined links to gene 90 expression that contributed to at least 1% of the total peptide detected. Remarkably, we found 91 numerous histone modifications associated with active gene expression were depleted in cells 92 expressing Orf8 while histone modifications associated with chromatin compaction or 93 transcriptional repression were increased in cells expressing Orf8 (Fig. 1c)

99
These data support a role for Orf8 as a putative histone mimic and demonstrate that it is capable 100 of disrupting histone PTM regulation at numerous critical sites within histones.  H3K9ac staining compared to control plasmid transfected cells ( Fig. 1e-j). To determine whether 118 these effects are due to the proposed histone mimic site within Orf8, we generated a deletion 119 construct lacking the ARKSAP histone mimic site (Orf8-del). While Orf8-del was expressed at 120 similar levels to Orf8 (Fig. S2b), it did not increase H3K9me3 or H3K27me3, and showed a trend 121 toward decreasing the effect on H3K9ac (Fig. 1e-j). Thus, the ability of Orf8 to disrupt histone 122 PTMs largely relies on the presence of the ARKSAP motif. Next, we examined another dominant 123 form of Orf8 containing an acquired mutation (S84L) commonly found in strains SARS-CoV-2.

124
This site lies outside the histone mimic region and thus is not expected to affect its ability to 125 regulate histone PTMs. We found that Orf8-S84L also increased H3K9me3 and H3K27me3, while 126 decreasing H3K9ac (Fig. S2c-e), indicating that, as predicated, this common mutation does not 127 alter the potential histone mimicry. We did not detect significant global changes in H3K27ac using 6 these methods (Fig. S2f), potentially due to low H3K27ac basal levels and fitting with mass 129 spectrometry results.

131
To understand the mechanism through which Orf8 disrupts host cell chromatin, we began by 132 examining its intracellular localization. Notably, while Orf8 does not have a well-defined NLS, it is 133 15kD in size and thus is small enough to diffuse into the nucleus. We first transfected HEK cells 134 with Strep-tagged Orf8 and using a cellular fractionation assay, detected Orf8 in both the 135 cytoplasm and the nucleus (Fig. 2a). We performed immunofluorescence to confirm these 136 findings through an independent approach. We found that Orf8 was present in the cytoplasm and 137 was located at the periphery of the nucleus as well as in nuclear puncta (Fig. 2b). This expression 138 pattern matches those described in a previous report 27 , although this study focused on a 139 cytoplasmic role of Orf8. Given the observed expression pattern of Orf8, we next asked whether 140 Orf8 is associated with Lamin. We found that Orf8 colocalized with LaminB1 and LaminA/C ( Fig.   141   2c, Fig. S3a). Furthermore, Orf8 bound LaminB1, histone H3, and HP1a, a protein associated 142 with both Lamin and histones (Fig. 2d). Similarly, reciprocal co-immunoprecipitation for LaminB1 143 and H3 confirmed Orf8 binding (Fig. 2d). These findings show nuclear localization of Orf8 and 144 indicate association with chromatin.

146
We further examined the strength of the Orf8 association with chromatin, utilizing increasing salt 147 concentrations to examine chromatin binding. We found that Orf8 dissociates from the chromatin 148 fraction at salt concentrations between those at which Lamin dissociates and the peak at which 149 histones dissociate while Orf8-del dissociates at lower salt concentrations (Fig. 2e). We next used 150 ChIP-sequencing of Orf8 itself to determine whether and where Orf8 associates with genomic 151 DNA. We discovered that Orf8 was enriched at transcription start sites and in genic regions within 152 the human genome relative to input DNA (or compared to a control ChIP performed with cells that 153 do not express Orf8) (Fig. S3b-c), although Orf8 binding does not show clearly defined peaks at 7 specific genes as would be expected for an endogenous histone PTM (Fig. S3d). To confirm Orf8 155 association with open chromatin regions, we used ChIP-qPCR and observed greater Orf8 156 association with euchromatic compared heterchomatin genomic regions (Fig. S3e).

158
We next used targeted mass spectrometry to determine whether the proposed Orf8 histone mimic 159 site is modified similarly to histones. Using a bottom-up approach, Orf8 was purified from cells,  (Fig. 2f, S4a). This demonstrates that Orf8 is  Orf8, while regions of enriched H3K9me3 typically appear at or in close proximity to Orf8 puncta 178 (Fig. S4b). This localization pattern is consistent with a role for Orf8 as a histone mimic that 179 disrupts host cell chromatin regulation causing both local and global changes in histone PTMs.
8 Finally, we used mass spectrometry to identify additional binding partners beyond Lamin-181 associated complexes (Table S2). Top hits included the HAT complex protein MORF4L, several 182 zinc finger proteins, and the transcription factor SP2 which we confirmed by co-183 immunoprecipitation (Fig. S4c). Together, these results support a model in which Orf8 associates  with LaminA/C (Fig. 3a, S5b), similar to the patterns observed in cells transfected with Orf8 ( Fig.   197 2a-c). The requirement for viral inactivation through methods such as fixation prevented 198 subsequent biochemical analysis of virally expressed Orf8. However, we confirmed that in 199 A549 ACE cells exogenously expressing Orf8, sequential salt extractions showed similar Orf8 200 chromatin association as in HEK cells (Fig S5c) and similar Orf8 localization (Fig. S5d)  H3K27me3 and decreased H3K9ac (Fig. 3b-g), replicating the effects of Orf8 expression.

9
Together, these data indicate that both Orf8 expression and SARS-CoV-2 infection result in global 207 changes in histone regulation and chromatin accessibility, providing a novel mechanism through 208 which SARS-CoV-2 can disrupt host cell function.

210
New data from COVID-19 human patients, recently published in The Lancet, found that a 382-211 nucleotide deletion variation in SARS-CoV-2 that blocks expression of the ORF8 gene (Fig. 4a) 212 is associated with a milder infection in COVID-19 patients 28 . Furthermore, Orf8 expression has 213 been shown to block type 1 interferon and NF-kB responsive promoters and to inhibit induction of infection in A549 ACE cells. We found that, despite widespread differential gene expression (Fig.   218 S6a-b), interferon viral response genes were only mildly induced by infection as measured by 219 gene ontology analysis, overlap with a defined set of A549 interferon response genes 30 , or 220 examination of specific response genes (Fig. S6b-e). These data support recent findings  sequencing with ChIP-RX normalization (Fig. S7a) to allow for detection for global changes in 231 histone PTMs. Strikingly, we found that infected iAT2s showed globally increased H3K9me3 and H3K27me3 and decreased H3K9ac (Fig. 4b), again matching the effects of Orf8 expression. In 233 addition to global changes, increased H3K9me3 and H3K27me3 and decreased H3K9ac were 234 found at interferon-stimulated genes such as IFITM2, ADAR and FOSL2 (Fig. 4c) (Fig. S7b).

242
Lastly, we obtained postmortem lung tissue samples from three COVID-19 patients and matched 243 controls. We stained tissue for SARS-CoV-2 Nucleocapsid protein to identify infected cells and 244 for H3K9me3 to measure histone PTM changes. We found that in all patient samples, infected 245 cells showed increased H3K9me3 staining compared to neighboring cells within the same tissue 246 as well as compared to control tissue (Fig. 4d-e, S7c). While limited sample availability limits the 247 conclusions that can be drawn from this assay, this finding indicates that histone PTMs are also

256
The work described here identifies a novel case of histone mimicry in the SARS-CoV-2 virus and 257 defines a mechanism through which SARS-CoV-2 acts to disrupt host cell chromatin regulation.

11
We found that the protein encoded by the SARS-CoV-2 ORF8 gene contains an ARKS motif and 259 that Orf8 expression disrupts histone PTM regulation. Orf8 is associated with chromatin-