key: cord-0266530-9ajnms4f
authors: Park, Chorong; Peng, Chen; Rahman, M. Julhasur; Haller, Sherry L.; Tazi, Loubna; Brennan, Greg; Rothenburg, Stefan
title: Orthopoxvirus K3 orthologs show virus- and host-specific inhibition of the antiviral protein kinase PKR
date: 2020-06-24
journal: bioRxiv
DOI: 10.1101/2020.02.20.958645
sha: 68d383b10b8ccc59db3a6acd7a0f6ac6688aac1c
doc_id: 266530
cord_uid: 9ajnms4f

The antiviral protein kinase R (PKR) is an important host restriction factor, which poxviruses must overcome to productively infect host cells. To inhibit PKR, many poxviruses encode a pseudosubstrate mimic of the alpha subunit of eukaryotic translation initiation factor 2 (eIF2), designated K3 in vaccinia virus. Although the interaction between PKR and eIF2α is highly conserved, some K3 orthologs were previously shown to inhibit PKR in a species-specific manner. To better define this host range function, we compared the sensitivity of PKR from 17 mammals to inhibition by K3 orthologs from closely related orthopoxviruses. The K3 orthologs showed species-specific inhibition of PKR and exhibited three distinct inhibition profiles. In some cases, PKR from closely related species showed dramatic differences in their sensitivity to K3 orthologs. Vaccinia virus expressing the camelpox virus K3 ortholog replicated more than three orders of magnitude better in human and sheep cells than a virus expressing vaccinia virus K3, but both viruses replicated comparably well in cow cells. Strikingly, in site-directed mutagenesis experiments between the variola virus and camelpox virus K3 orthologs, we found that different amino acid combinations were necessary to mediate improved or diminished inhibition of PKR derived from different host species. Because there is likely a limited number of possible variations in PKR that affect K3-interactions but still maintain PKR/eIF2α interactions, by chance PKR from some potential new hosts may be susceptible to K3-mediated inhibition from a virus it has never previously encountered. We conclude that neither the sensitivity of host proteins to virus inhibition nor the effectiveness of viral immune antagonists can be inferred from their phylogenetic relatedness but must be experimentally determined. Authors summary Most virus families are composed of large numbers of virus species. However, in general, only a few prototypic viruses are experimentally studied in-depth, and it is often assumed that the obtained results are representative of other viruses in the same family. In order to test this assumption, we compared the sensitivity of the antiviral protein kinase PKR from various mammals to inhibition by multiple orthologs of K3, a PKR inhibitor expressed by several closely related orthopoxviruses. We found strong differences in PKR inhibition by the K3 orthologs, demonstrating that sensitivity to a specific inhibitor was not indicative of broad sensitivity to orthologs of these inhibitors from closely related viruses. We also show that PKR from even closely related species displayed markedly different sensitivities to these poxvirus inhibitors. Furthermore, we identified amino acid residues in these K3 orthologs that are critical for enhanced or decreased PKR inhibition and found that distinct amino acid combinations affected PKRs from various species differently. Our study shows that even closely related inhibitors of an antiviral protein can vary dramatically in their inhibitory potential, and cautions that results from host-virus interaction studies of a prototypic virus genus member cannot necessarily be extrapolated to other viruses in the same genus without experimental verification.

weakly inhibiting cow and mouse PKR [17] . Similarly, M156, the K3 ortholog from the 111 rabbit-and hare-specific myxoma virus (MYXV), inhibited European rabbit PKR but not 112 PKR derived from seven other mammals. Moreover, this work demonstrated that changes 113 in viral antagonists can be rapidly fixed in the population, as demonstrated by the 114 discovery that M156 evolved a loss of function mutation that might contribute to the 115 attenuation of the virus in European rabbits [18] . 116 Most research on orthopoxviruses has focused on the prototypic poxvirus VACV 117 as a model for poxvirus virology and host-virus interactions. However, very few studies 118 have systematically investigated the activity of VACV orthologs in other orthopoxviruses. 119 It is therefore unclear whether the activity of VACV proteins is representative of their 120 orthologs in other orthopoxvirus. Orthopoxviruses such as CPXV and VARV represent 121 substantial public health or bioterror threats, and human infections with orthopoxviruses 122 are predicted to become more common as immunity due to the cessation of the smallpox 123 vaccination decreases in the population [19] . Therefore, it is critically important to identify 124 orthopoxvirus determinants of host range and to identify viruses most likely to infect new 125 species.

In this study, we compared inhibition of PKR from a panel of mammalian species 127 by K3 orthologs from VACV, VARV, CMLV and CPXV-rat09 to determine whether the 128 ability of VACV K3 to inhibit PKR from a given host was predictive of the ability for other 129 orthopoxvirus K3 orthologs to inhibit this host antiviral protein. The results presented here 130 show that the tested K3 orthologs exhibited distinct inhibition profiles and also that the 131 phylogenetic relatedness of PKR was not a good predictor for how susceptible they were 132 to inhibition by K3 orthologs. Mapping K3 residues that governed the species-specificity 133 of these interactions revealed that multiple amino acid substitutions were necessary to 134 convert a weak K3 ortholog into a better inhibitor of most tested PKRs. Furthermore, 135 different amino acid combinations were needed to most effectively inhibit PKR from 136 different species.

141 Previous yeast-based and transfection assays indicated differential inhibition of 142 PKR from several vertebrate species by VACV K3 [8, 9] . In order to get a more complete 143 picture about the breadth of species-specific PKR inhibition, we analyzed the sensitivity 144 of PKRs from 17 mammals to inhibition by VACV K3 in an established luciferase-based 145 reporter (LBR) assay (Fig. 1A) [9]. In agreement with previous findings, human, and 146 sheep PKRs were largely resistant, whereas mouse and cow PKRs were sensitive to K3 147 inhibition in this assay [17] . Among the primate PKRs tested, tamarin PKR was most 148 strongly inhibited, consistent with previous yeast assays [8] . Because vaccinia virus is 149 thought to have originated from a rodent virus, it was surprising that only Syrian hamster 150 and mouse PKRs were sensitive, whereas PKRs derived from other rodents, rats and 151 guinea pigs, were resistant. Among PKRs from ungulates, horse, cow and camel PKR 152 were sensitive, whereas sheep PKR was only weakly sensitive and pig PKR was To compare the inhibition profiles of VACV K3 orthologs from other 159 orthopoxviruses, we cloned VARV C3L, CMLV (CMS strain) 032, and CPXV 043 from the 160 CPXV-rat09 strain, the latter of which is identical to multiple CPXV isolates from the D 161 clade [20, 21] , into the mammalian expression vector pSG5. CPXV-rat09 043 is also 162 identical to protein 037 from taterapox virus (TATV) and will henceforth be referred to as 163 CPXV-D 043/TATV 037. Relative to VACV K3, CPXV-D 043/TATV 037 has 11 amino acid 164 (aa) differences (87.5 % identity), CMLV 032 has 12 aa differences (86.4 % identity), and 165 VARV C3 has 17 aa differences (80.7 % identity) ( Fig. 2A To visualize if the phylogenetic relatedness of a given PKR may be predictive of 178 susceptibility to these K3 orthologs, we generated a phylogenetic tree using PKR 179 sequences derived from 40 mammalian species. We included PKRs from several species 180 that were not tested in the LBR assays to achieve better statistical support for the 181 branches. We then mapped the relative sensitivities of the tested PKRs to each K3 182 ortholog obtained from multiple LBR assays using a susceptibility scale ranging from 183 weakly sensitive to highly sensitive on a scale from 1 to 6, respectively (Fig. 4) VACV-K3L replicated to slightly higher titers than input (~10-fold increase) in HeLa and 207 Tert-GF cells indicating that the weak inhibition of PKR allowed limited virus replication.

Identification of amino acid residues that confer differential PKR inhibition between 210 VARV C3L and CMLV 032 211 VARV C3L and CMLV 032 differ by seven aa residues (Fig. 2) . In order to identify 212 which residues are responsible for the virus-specific differences in PKR inhibition, we first 213 exchanged single aa residues in VARV C3L with the ones found in CMLV 032 and 214 performed the LBR assay with human, tamarin, mouse, dog, cow and sheep PKR. None 215 of these mutants inhibited human PKR better, whereas some mutants showed better 216 inhibition of some other PKRs. Only one VARV C3L mutant (V25A) reached comparable 217 PKR inhibition as CMLV 032, and only for dog PKR. VARV C3L-V25A showed more than 218 2-fold increased PKR inhibition of tamarin and sheep PKR and more than 50% higher 219 inhibition of cow and mouse PKR. VARV C3L-I70V showed more than 2-fold increased 220 PKR inhibition of tamarin PKR and more than 50% higher inhibition of mouse and dog 221 PKR. VARV C3L-V3A displayed more than 50% higher inhibition of dog and sheep PKR, 222 whereas VARV C3L-T44V only showed higher inhibition of mouse PKR (more than 2-223 fold).

These experiments confirmed that a single aa residue was not responsible for the 225 differential PKR inhibition found for most of the tested PKRs. Therefore, we generated Most research on poxviruses has focused on VACV, which is often called the 261 prototypic poxvirus. VACV K3 has been shown to inhibit PKR from several mammals in 262 a species-specific manner, as defined in both LBR assays and yeast-based studies [8, 9, 263 17]. The data presented here support and extend this species-specific sensitivity of PKR 264 to VACV K3L and add to a growing body of evidence that virus-specific differences play This also demonstrates that one efficient K3 ortholog can substitute for both VACV E3 287 and K3, in at least some cell lines.

The strong activity of CMLV K3 against PKR from its natural host ( context, these chance susceptibilities of host antiviral factors to viral inhibitors that a host 331 has otherwise never been exposed to, might play an important role in cross-species 332 transmission of viruses [24] .

Although VARV is a human-specific pathogen, this study and previous yeast-334 based assays demonstrated that the VARV K3L ortholog does not inhibit human PKR 335 efficiently [8, 9] . However, mutation of some of the positively selected residues in human 336 PKR to residues found in other species differentially affected the sensitivity to VACV K3 337 and VARV C3, further illustrating that these differences are mediated by only a few 338 residues in either the host or viral protein [9] . The close relatedness between VARV C3 339 and CMLV 032 allowed the identification of aa residues that are responsible for the 340 differential inhibition of PKR. In 5 of the 6 tested PKR mutants, single aa substitutions in 341 VARV C3 led to little or no change in activity. Distinct combinations of mutations were 342 necessary to make VARV C3 as efficient as CMLV 032 depending on the particular PKR 343 ortholog being tested. For example, only V25A was necessary to make VARV C3 inhibit 344 dog PKR as efficiently as CMLV 032; however, four aa changes were necessary to 345 phenocopy CMLV 032 inhibition of human PKR by VARV C3, and only three aa changes 346 were necessary for this effect against sheep PKR. It is noteworthy that the loss of PKR 347 inhibition by CMLV 032 was more easily achieved by single aa substitutions than an 348 increase in PKR inhibition. The differences in VARV C3L were all acquired after the split 349 from a CMLV/TATV ancestor, and all of these C3L mutations are non-synonymous, which 

After 48 hours, cells and supernatants were collected and subjected to three rounds of 465 freezing at -80 ˚C and thawing at 37 ˚C. Lysates were sonicated for 15s, 50% amplitude 466 (Qsonica Q500). Viruses were titered by 10-fold serial dilutions on RK13+E3L+K3L cells.

Infections and viral titer were performed in duplicate. high inhibition (7 to 10-fold); 6 = very high inhibition (≥ 10-fold). 

Poxviruses and the evolution of host 511 range and virulence

Zoonotic cases of camelpox infection in India

Human and Dromedary Camel Infection with Camelpox Virus 517 in Eastern Sudan. Vector Borne Zoonotic Dis

Protein kinase PKR and RNA adenosine 520 deaminase ADAR1: new roles for old players as modulators of the interferon response

The role of the PKR-inhibitory genes, E3L and K3L, in 558 determining vaccinia virus host range

A survey of host range genes in poxvirus 561 genomes

Species-specific inhibition of antiviral protein 564 kinase R by capripoxviruses and vaccinia virus

Myxoma virus M156 is a 567 specific inhibitor of rabbit PKR but contains a loss-of-function mutation in Australian virus isolates

PubMed 569 PMID: 26903626; PubMed Central PMCID: PMCPMC4833222. 570 19. Shchelkunov SN. An increasing danger of zoonotic orthopoxvirus infections

PubMed 572 PMID: 24339772; PubMed Central PMCID: PMCPMC3855571

Cowpox virus: 574 What's in a Name?

Cowpox Viruses into Several Distinct Clades and Identification of a Novel Lineage

Seamless Generation of 581

Recombinant Poxviruses using Host Range and Visual Selection

How host genetics dictates successful viral zoonosis

Species-Specific Host-Virus Interactions: Implications for Viral 587

Genome-wide comparison of cowpox 593 viruses reveals a new clade related to Variola virus

Use of 597 Next Generation Sequencing to study two cowpox virus outbreaks

Isolation of poxvirus from an African Rodent

Monkeypox virus 604 induces the synthesis of less dsRNA than vaccinia virus, and is more resistant to the anti-poxvirus 605 drug, IBT, than vaccinia virus

Ectromelia 609 virus accumulates less double-stranded RNA compared to vaccinia virus in BS-C-1 cells

Poxvirus encoded eIF2alpha homolog, K3 family 613 proteins, is a key determinant of poxvirus host species specificity

Comparative analysis of poxvirus orthologues of the vaccinia virus E3 protein: modulation of 617 protein kinase R activity, cytokine responses, and virus pathogenicity

Variola virus immune evasion design: expression 621 of a highly efficient inhibitor of human complement

Structure and 625 regulatory profile of the monkeypox inhibitor of complement: comparison to homologs in vaccinia 626 and variola and evidence for dimer formation

Identification of hot spots in the variola virus 629 complement inhibitor (SPICE) for human complement regulation

Species 637 Specificity of Vaccinia Virus Complement Control Protein for the Bovine Classical Pathway Is 638 Governed Primarily by Direct Interaction of Its Acidic Residues with Factor I

Identification from diverse mammalian poxviruses of host-range 642 regulatory genes functioning equivalently to vaccinia virus C7L

The poxvirus C7L host range factor superfamily

Epub 2012/10/30

Structural basis for antagonizing a host 649 restriction factor by C7 family of poxvirus host-range proteins

M062 is a host range factor essential for myxoma 653 virus pathogenesis and functions as an antagonist of host SAMD9 in human cells

SAMD9 is an innate antiviral host factor with stress response 657 properties that can be antagonized by poxviruses

A paralogous pair of 661 mammalian host restriction factors form a critical host barrier against poxvirus infection

PubMed 663 PMID: 29447249; PubMed Central PMCID: PMCPMC5831749

Identification of CP77 as 665 the Third Orthopoxvirus SAMD9 and SAMD9L Inhibitor with Unique Specificity for a Rodent 666 SAMD9L

Myxoma virus protein M029 673 is a dual function immunomodulator that inhibits PKR and also conscripts RHA/DHX9 to promote 674 expanded host tropism and viral replication

MUSCLE: multiple sequence alignment with high accuracy and high 678 throughput

Improvement of phylogenies after removing divergent and 681 ambiguously aligned blocks from protein sequence alignments

New algorithms 684 and methods to estimate maximum-likelihood phylogenies: assessing the performance of PhyML 685 3.0

SMS: Smart Model Selection in PhyML

Epub 2017/05/05

Higher-order substrate recognition of eIF2alpha by the RNA-691 dependent protein kinase PKR

Residues that correspond to amino acids of eIF2a that contact 777 human PKR in a co-crystal structure [51] are shown in purple (residue A41) or red (other 778 residues). Residues that differ between VARV C3 and CMLV 032 are shown in purple 779 (residue A41) or blue (other residues). Residue H47

Cell lysates from transfected cells were separated on 12% SDS-PAGE gels and 784 analyzed by immunoblot analysis with anti-K3 and anti-b-actin antibodies. (D) HeLa-785 PKR kd cells were transfected with expression vectors encoding firefly luciferase (0.05 μg), 786 with VARV C3L, the indicated mutants of VARV C3L, or CMLV 032 (0.4 μg), and PKR 787 (0.2 μg) from the indicated species. Luciferase activities were measured 48 hours after 788 transfection and normalized to PKR-only transfected cells to obtain relative luciferase 789 activities. Error bars represent the standard deviations from three independent 790 transfections