A Sweep of Earth’s Virome Reveals Host-Guided Viral Protein Structural Mimicry and Points to Determinants of Human Disease

•Structural mimicry of host proteins is a pervasive strategy across earth’s virome•Majority the >6,000,000 cataloged mimics cannot be discerned through sequence•Protein structure space utilized by viruses dictated proteomes•Knowledge viral provides clues about pathophysiology COVID-19 Viruses deploy genetically encoded strategies to coopt machinery and support replicative cycles. Here, we use protein similarity scan for molecular mimicry, manifested structural between endogenous proteins, thousands hosts spanning broad ecological niches taxonomic range, including bacteria, plants fungi, invertebrates, vertebrates. This survey identified over 6,000,000 instances mimicry; more than 70% sequence alone. We demonstrate that manner degree which exploit varies genome size nucleic acid type identify 158 human are mimicked coronaviruses, providing cellular processes driving pathogenesis. Our observations point as employed indicate used given virus proteome.A record this paper’s transparent peer review process included in Supplemental Information. proteome. A an array Among strategies, protein-protein interactions (PPIs), mediated promiscuous, multifunctional widely documented. Targeted discovery tools focused largely on public-health importance have experimentally mapped virus-host complexes, structure-informed prediction algorithms allowed such all fully sequenced human-infecting (Lasso et al., 2019Lasso G. Mayer S.V. Winkelmann E.R. Chu T. Elliot O. Patino-Galindo J.A. Park K. Rabadan R. Honig B. Shapira S.D. atlas human-virus interactions.Cell. 2019; 178: 1526-1541.e16Abstract Full Text PDF PubMed Scopus (77) Google Scholar). Molecular allows harness or disrupt functions metabolism modulation immune responses. Yet, while examples latter pepper literature (Wimmer Schreiner, 2015Wimmer P. Schreiner S. Viral usurp ubiquitin SUMO pathways.Viruses. 2015; 7: 4854-4872Crossref (44) Scholar; Alcami, 2003Alcami A. cytokines, chemokines their receptors.Nat. Rev. Immunol. 2003; 3: 36-50Crossref (452) Guven-Maiorov 2016Guven-Maiorov E. Tsai C.J. Nussinov Pathogen interfaces modulates immunity.Semin. Cell Dev. Biol. 2016; 58: 136-145Crossref (33) Scholar), most (Elde Malik, 2009Elde N.C. Malik H.S. The evolutionary conundrum pathogen mimicry.Nat. Microbiol. 2009; 787-797Crossref (144) Felix Savvides, 2017Felix J. Savvides S.N. Mechanisms immunomodulation mammalian decoy receptors: insights from structures.Nat. 2017; 17: 112-129Crossref (45) systematic analysis pathogen-encoded has not been performed. vast genomic landscape occupied hampers relationships hosts. As well known, however, since three-dimensional (3D) much better conserved sequence, information can interrogate Bamford 2005Bamford D.H. Grimes J.M. Stuart D.I. What does tell us evolution?.Curr. Opin. Struct. 2005; 15: 655-663Crossref (305) Scholar) uncover virus-encoded detected (see STAR Methods). virally proteomes. Briefly, first employ sequence-based methods similar structures queried then alignment find “structural neighbors” (Figure 1A). refer corresponding host-encoded neighbors. applied approach set 337,493 representing 7,486 majority (>70%) alone below). results dictated, at least part, further observe type. For example, arthropod-infecting occupy proteome, arboviruses, transmitted humans insect vectors, encode promiscuous mimic both proteins. In addition, that, relative proteome size, single-stranded RNA (ssRNA) viruses, coronaviruses (CoVs), circumvented limitations small genomes mimicking greater extent large dsDNA counterparts like Pox Herpes viruses. Interrogation points diversification pathways targeted via identifies biological may underly autoimmune disorders, reveals leveraged engineer synthetic metabolic circuits serve targets therapeutics. Finally, 150 (including members complement activation pathway critical regulators innate adaptive immunity) CoV, underlying pathogenesis ongoing pandemic. order relationships, implemented summarized Figure 1A. Sequence-based searches Methods) (269,077 unique sequences) against PDB templates 92,868 (34.5%), with disparities coverage: were 64.7% whereas coverage non-vertebrate-infecting ranged 31% 38% S1A; Table S2). shown S1B, vertebrate came primarily other while, bacterial provided bacteria-infecting Most invertebrate-infecting likely reflecting limited non-vertebrate invertebrate PDB. measure Ska, extensively validated tool inference structure-based functional even absence detectable (Garzón 2016Garzón J.I. Deng L. Murray D. Petrey computational interactome annotation proteome.eLife. 5: e18715Crossref (46) Hwang 2017Hwang H. Dey F. Structure-based ligand-protein genome-wide scale.Proc. Natl. Acad. Sci. USA. 114: 13685-13690Crossref (36) Lasso Zhang 2012Zhang Q.C. Qiang Shi Y. Thu C.A. Bisikirska Lefebvre C. Accili Hunter al.Structure-based scale.Nature. 2012; 490: 556-560Crossref (525) 2003Petrey Xiang Z. Tang C.L. Xie Gimpelev M. Mitros Soto C.S. Goldsmith-Fischman Kernytsky Schlessinger al.Using multiple alignments, fast model building, energetic fold recognition homology modeling.Proteins. 53: 430-435Crossref (274) Scholar, 2009Petrey Fischer Structural among different global topologies implications function strategies.Proc. 106: 17377-17382Crossref (70) Yang Honig, 2000Yang A.S. An integrated modeling sequences structures. III. comparative study conservation families using alignments.J. Mol. 2000; 301: 691-711Crossref (79) addition conservative criteria (structural score, SAS < 2.5 Å) (Budowski-Tal 2010Budowski-Tal I. Nov Kolodny FragBag, accurate representation structure, retrieves neighbors entire quickly accurately.Proc. 2010; 107: 3481-3486Crossref (88) 2005Kolodny Koehl Levitt Comprehensive evaluation methods: scoring geometric measures.J. 346: 1173-1188Crossref (237) Subbiah 1993Subbiah Laurents D.V. DNA-binding domains bacteriophage repressors globin core.Curr. 1993; 141-148Abstract (161) infer minimize biases imposed local similarities S2, enables rediscovery known mimics. 6,083,167 involving 88,715 26,542 present (data available download https://github.com/gorkaLasso/VirusHost_mimicry). these (72.2%) could retrieved based requirement no e value obtained any one three was 1 × 10?6 Taxonomic enrichment (Figures 1B 1C) clear correlation identity host. Since proteomes represented PDB, coli-infecting phages, natural coli, budding yeast Saccharomyces cerevisiae, human. Significant only observed phages 1B). 1C, exception constitute special case below), broader 38,363 4,045 putative demonstrates every group exhibits significant division exhibit divisions close distance reduced distant (e.g., enriched human, vertebrates, but bacteria plant 1C). Such reflected divisions. instance, accumulating evidence describes some compartments phytosphere environmental spots horizontal gene transfer (HGT) facilitate genetic exchange (Pontiroli 2009Pontiroli Rizzi Simonet Daffonchio Vogel T.M. Monier Visual transplastomic tobacco.Appl. Environ. 75: 3314-3322Crossref (65) underscoring relationship fungi-infecting Moreover, agreement categories, found non-human 1C)—pointing possible path zoonotic infections jump animal reservoirs cause disease. Unlike do hosts—perhaps owing fact invertebrates account lowest (containing just 2,687 proteins). To bridge knowledge gap, modeled mosquito Ae. Aegypti compared each 1D, mosquito-infecting significantly larger infect mosquito-borne mosquitoes engage vectors (p 2.3 10?3; 1D). Together, data highlight pressures utilize defined host’s range. While recursive virome, belonging same family astroviruses, filoviruses, orthomyxoviruses, paramyxoviruses, rhabdoviruses, picornaviruses; https://github.com/gorkaLasso/VirusHost_mimicry) tend share fraction (where >75% within mimics). 2A displays number per families. families, Hepadnaviridae, Bunyaviridae, Papillomaviridae, Parvoviridae, Retroviridae, Arenaviridae, Polyomaviridae, far less others. On hand, poxviruses herpesviruses, constrained 2B) retain horizontally transferred genes (Alcami, Elde many 2A). sizes vary greatly 2B), computed effort assess utilizations 2C). display enhanced promiscuity, residue. So, restricted ability expand HGT, they limitation encoding These underscore role evolving expanded functionality genomes. nature PPIs engaging fulfill life cycles led convergence evolutionarily unrelated around key DNA, retro-transcribing converge, sets PPIs, pathways) contrast, 2D; S3, structurally revealed little convergence. when observed, it often involved 2 3 Baltimore classification. propensity acquire Herpesviridae, Adenoviridae, Poxviridae, largest pathways, 113, 38, 35, respectively S3). harbor mimics, components DNA metabolism—likely polymerases (as Poxviridae) and/or integrate into therefore bystander participants replication. cytokine-related poxvirus- herpesvirus-mimicked chemokine-related herpesviruses (Table S3; 2D). Indeed, evolved repertoire evasion acquiring immunomodulatory recombination (often cytokines cytokine receptors). Undetectable complicated past attempts identification (Odom 2009Odom M.R. Hendrickson R.C. Lefkowitz E.J. Poxvirus evolution: wide assessment events.Virus Res. 144: 233-249Crossref (48) However, herpesvirus chemokine receptors list include 61 herpesviral discernible complete 2D, US27D Cercopithecine betaherpesvirus 5, cytomegalovirus, CXCR4 (sequence identity: 10.4%), receptor regulates T cell inflammatory programs co-receptor HIV (Busillo Benovic, 2007Busillo Benovic J.L. Regulation signaling.Biochim. Biophys. Acta. 2007; 1768: 952-963Crossref (450) It unclear if result divergence products acquired HGT convergent evolution. factors mimicry. highlighted above, accordance phenotypes (like fibrinolysis hemorrhagic fever) associated Dengue (DENV; Flaviridae) infection (Chuang 2014Chuang Y.C. Lin Y.S. Liu Yeh dengue coagulation induces antibodies inhibit thrombin activity enhance fibrinolysis.J. Virol. 2014; 88: 13759-13768Crossref (27) 2011Lin C.F. Wan S.W. Chuang Hsu T.K. C.C. Anderson Lei H.Y. its disease pathogenesis.Exp. Med. (Maywood). 2011; 236: 515-523Crossref (96) related blood pathway—an observation mirrored (we previously PPI networks 51 56 flaviviruses “complement cascades”) DENV C, prM, E, NS1 (Lin four positive-sense ssRNA ((+)ssRNA) (168, 166, 1,376, 1,293 astroviridae, caliciviruses, picornaviruses, respectively; <20%, immunization exposure elicits cross-react factor-associated pathological states our suggest also imbalances SARS-Coronaviridae, hepatitis A-Picornaviridae, C-Flaviridae) disorders thrombocytopenia, thrombosis, hemorrhage (Goeijenbier 2012Goeijenbier Van Wissen De Weg Jong Gerdes V.E. Meijers J.C. Brandjes D.P. Gorp E.C. Review: mechanisms thrombosis bleeding.J. 84: 1680-1696Crossref (183) Knowledge precise regulatory control cycle mediate pathology provide valuable determinants. closer look underscores level promiscuity 2C, (+)ssRNA CoVs so (?96%) shared seasonal pathogenic CoV MERS, SARS-1, SARS-CoV-2). amino identities sites ultimately shape intervened end, experimental interrogation CoV-encoded will necessary determine residues. target anti-viral programs. First, context infection, intracellular sensing DDX58 (also RIG-I; retinoic acid-inducible I), receptor, triggers signaling cascade production secretion I interferons (IFN?/?) (Abe Shapira, 2019Abe Negative regulation cytosolic DNA.Int. 344: 91-115Crossref (16) Jensen Thomsen, 2012Jensen Thomsen A.R. Sensing viruses: recognizing invasion.J. 86: 2900-2910Crossref (419) IFN?/? turn binds surface (IFNRs) nearby cells and, following signal-transduction cascade, leads STAT1-mediated (signal transducer activator transcription 1) upregulation hundreds IFN-stimulated (ISGs) govern responses render refractory infection. indispensable replication initiation long-term 3, NSP13 DDX3X RIG-I helicase cooperate direct indirect initiate ligands ssRNAs contain 5?-triphosphate moiety previous biochemical characterizations attributed enzymatic activities NSP13, hydrolysis NTPs dNTPs 5?-triphosphatase (Ivanov 2004Ivanov K.A. Thiel V. Dobbe Der Meer Snijder Ziebuhr Multiple severe acute re