Functional Analysis of the Murine Coronavirus Genomic RNA 2 Packaging Signal 3 4 5

25 26 Coronaviruses selectively package genomic RNA into assembled virions, despite the 27 great molar excess of subgenomic RNA species that is present in infected cells. The genomic 28 packaging signal (PS) for the coronavirus mouse hepatitis virus (MHV) was originally identified 29 as an element that conferred packaging capability to defective interfering RNAs. The MHV PS is 30 an RNA structure that maps to the region of the replicase gene encoding the nonstructural protein 31 15 subunit of the viral replicase-transcriptase complex. To begin to understand the role and 32 mechanism of action of the MHV PS in its native genomic locus, we constructed viral mutants in 33 which this cis-acting element was altered, deleted, or transposed. Our results demonstrated that 34 the PS is pivotal in the selection of viral genomic RNA for incorporation into virions. Mutants in 35 which PS RNA secondary structure was disrupted or entirely ablated packaged large quantities 36 of subgenomic RNAs, in addition to genomic RNA. Moreover, the PS retained its function when 37 displaced to an ectopic site in the genome. Surprisingly, the PS was not essential for MHV 38 viability, nor did its elimination have a severe effect on viral growth. However, the PS was found 39 to provide a distinct selective advantage to MHV. Viruses containing the PS readily outcompeted 40 their otherwise isogenic counterparts lacking the PS. 41 42 43 44 on July 6, 2017 by gest http/jvi.asm .rg/ D ow nladed fom JVI00100-13 Version 2 02.10.13 page 3 of 28 INTRODUCTION 45 46 Virion assembly culminates in the packaging of viral genetic material in a protected form 47 that can be transmitted to a new host. Many viruses exhibit a marked preference for packaging 48 their own genomes, to the exclusion of other available nucleic acids, but the molecular basis for 49 this preference is understood in only a handful of cases. Coronaviruses are very large RNA 50 viruses with an elaborate mechanism of gene expression that entails the synthesis of a 3'-nested 51 set of subgenomic (sg) mRNAs (1, 2). Each sgRNA contains a leader segment, which is identical 52 to the 5' end of the genome, joined at a downstream transcription-regulating sequence (TRS) to a 53 segment identical to the 3' end of the genome. Despite the multiplicity of full-length and sgRNA 54 species of both polarities that come into play during the course of intracellular replication, it is 55 generally agreed that coronaviruses selectively package positive-strand genomic RNA (gRNA) 56 into virions. Purified virions of mouse hepatitis virus (MHV) and transmissible gastroenteritis 57 virus (TGEV) have been shown to almost exclusively contain gRNA (3-6), but it is not well 58 established whether this high degree of selectivity is shared by all members of the coronavirus 59 family (7, 8). 60 In MHV, the genomic packaging signal (PS) was localized through comparative analysis 61 of packaged and unpackaged defective interfering (DI) RNAs (3, 4). DI RNAs are variants of 62 gRNA that contain multiple, extensive deletions and therefore can only propagate by parasitizing 63 the replicative machinery of a helper virus. The exact functional boundaries of the MHV PS are 64 not precisely defined, but different studies have converged on a 220-nucleotide region that is 65 centered some 20.3 kb from the 5' end of the genome (5, 9, 10). This situates the PS in rep 1b, 66 the downstream portion of the gene for the viral replicase-transcriptase complex (Fig. 1). Such a 67 location would ensure that the PS appears only in gRNA and not in any of the sgRNAs. The 68 on July 6, 2017 by gest http/jvi.asm .rg/ D ow nladed fom JVI00100-13 Version 2 02.10.13 page 4 of 28 product of the replicase gene is a huge polyprotein that is autoproteolytically processed into 16 69 nonstructural proteins (nsps). The PS is embedded in the coding sequence for nsp15, a hexameric 70 uridylate-specific endoribonuclease (11-13). 71 Although early work suggested a 69-nucleotide RNA secondary structure as the minimal 72 functional MHV PS (5, 14), larger forms of the element were shown to be required for optimal 73 efficiency (10, 15, 16). A folding for one larger version of the PS was predicted but was not 74 experimentally verified (15). Recently, a novel structural model for the MHV PS was proposed 75 by the Olsthoorn laboratory (17). This new structure, a 95-nt bulged stem-loop, was strongly 76 supported by chemical and enzymatic probing experiments. It was also found to be highly 77 conserved among lineage A betacoronaviruses, a subgenus that includes human coronavirus 78 HKU1 (HCoV-HKU1), as well as bovine coronavirus (BCoV), human coronavirus OC43, and 79 related viruses within the betacoronavirus 1 species (18). This phylogenetic conservation is 80 consistent with the earlier demonstration that the BCoV PS is functionally interchangeable with 81 its MHV counterpart (15). However, the structural and functional homology of the PS does not 82 extend to other lineages within the betacoronaviruses, nor to other genera. It is clear, for 83 example, that there is not a counterpart of the MHV PS in the nsp15 open reading frame (ORF) 84 of the lineage B betacoronavirus severe acute respiratory syndrome coronavirus (SARS-CoV), 85 the alphacoronavirus TGEV, or the gammacoronavirus infectious bronchitis virus (IBV) (19, 20). 86 Moreover, studies characterizing packaged DI RNAs of TGEV and IBV suggest that the 87 packaging signals for these viruses map to distant sites (6, 21). 88 We have begun to genetically manipulate the MHV PS with the goal of identifying its 89 interacting partners and defining the mechanism of packaging. Although the PS was discovered 90 by use of DI RNA systems, the role of this element in its native locus in gRNA has never been 91 tested. Our study found that the PS does govern the selective incorporation of gRNA into virions, 92 on July 6, 2017 by gest http/jvi.asm .rg/ D ow nladed fom JVI00100-13 Version 2 02.10.13 page 5 of 28 and it remains functional when transposed to an ectopic genomic site. Remarkably, however, the 93 PS is not essential for MHV viability, but it confers a selective advantage to genomes that harbor 94 it. 95 96 on July 6, 2017 by gest http/jvi.asm .rg/ D ow nladed fom JVI00100-13 Version 2 02.10.13 page 6 of 28 MATERIALS AND METHODS 97 98 Cells and viruses. All MHV-A59 wild type and mutants were grown in mouse 17 clone 99 1 (17Cl1) cells. Plaque titrations and plaque purifications were performed with mouse L2 cells. 100 The interspecies chimeric virus designated fMHV.v2 (22), which was used for mutant 101 construction, was grown in feline FCWF cells. 102 MHV mutant construction. All mutants in this study were constructed by targeted RNA 103 recombination, as described in detail previously (22, 23). In brief, feline cells were infected with 104 fMHV.v2, which bears the ectodomain of the feline coronavirus spike (S) protein, and were then 105 transfected with synthetic donor RNA containing the mutations to be incorporated. Progeny virus 106 that reacquired the MHV S gene were selected, based on restoration of the ability to grow in 107 murine cells, and mutants were subsequently identified through screening by reverse 108 transcription (RT)-PCR. 109 The wild-type parent transcription vector for synthesis of donor RNA was pPM9 (Fig. 2), 110 which was constructed, in multiple steps, from the previously described pMH54 (23). Vector 111 pPM9 contains 5' elements of the MHV-A59 genome linked to the 3' end of the genome, the 112 latter starting from codon 236 of nsp14. Two alterations were made in pPM9 with respect to the 113 authentic wild-type sequence. First, two unique coding-silent restriction sites were created to 114 flank the region of the PS. An XbaI site (underlined) was generated by changing codons 124-126 115 of nsp15 from GCTCTTGAA to GCTCTAGAA, and a BspEI site (underlined) was produced by 116 changing codons 313-314 of nsp15 from AGTGGT to TCCGGA. Second, a 2,079-bp deletion 117 was created downstream of the nsp16 stop codon, removing transcription-regulating sequence 2 118 (TRS2), gene 2a, and all except 99 bp of the hemagglutinin-esterase (HE) gene. A 31-bp linker 119 containing unique SalI and AscI sites was inserted in place of the deleted region (24). 120 on July 6, 2017 by gest http/jvi.asm .rg/ D ow nladed fom JVI00100-13 Version 2 02.10.13 page 7 of 28 Downstream of the start of the S gene, other engineered sequence differences carrying over from 121 pMH54 to pPM9 have been noted previously (23). 122 Transcription vectors for the silPS mutant (pPM17) and the ΔPS mutant (pPM22) were 123 constructed from pPM9 by replacement of the XbaI-BspEI segment with a fragment synthesized 124 either from overlapping oligonucleotides by PCR or else by two-step PCR. Transcription vectors 125 for the silPS-PS2 mutant (pPM20) and the ΔPS-PS2 mutant (pPM19) were constructed by 126 insertion of a PCR product containing bp 514-681 of nsp15 in place of the SalI-AscI fragment of 127 pPM17 or pPM22, respectively. 128 Virus purification. Purifications of wild-type and mutant viruses by ultracentrifugation 129 in glycerol-tartrate gradients (25) or in sucrose gradients (4, 5) were carried out as described 130 previously. For immunopurifications, we employed a modification of the procedure of Escors et 131 al. (6). Virus was grown in 12 T150 flasks of 17Cl1 cell monolayers in Eagle's medium 132 containing 10% fetal bovine serum. Following polyethylene glycol precipitation from growth 133 medium, virions were resuspended in magnesiumand calcium-free phosphate-buffered saline, 134 pH 7.4 (PBS), and were pelleted onto cushions of 60% sucrose in PBS by centrifugation for 2 h 135 at 151,000 x g in a Beckman SW41 rotor at 4C. An aliquot equivalent to one-sixteenth of the 136 collected virions was incubated with 1 ml of anti-M monoclonal antibody J.1.3 for 3 h at 4C and 137 then for an additional 3 h with 1 ml of a 75% slurry of nProtein A Sepharose (GE Healthcare) in 138 PBS. Sepharose beads were collected by centrifugation at 200 x g for 10 min at 4C, washed 139 three times with 10 ml PBS, and used directly for RNA purification or SDS-PAGE sample 140 preparation. 141 Analysis of viral RNA. RNA was extracted from purified virions or from infected cell 142 monolayers with Ultraspec reagent (Biotecx) per the manufacturer's instructions or by using 143 Direct-zol RNA MiniPrep spin columns (Zymo Research). For verification of constructed 144 on July 6, 2017 by gest http/jvi.asm .rg/ D ow nladed fom JVI00100-13 Version 2 02.10.13 page 8 of 28 mutants and for analysis of competition experiments, reverse transcription of isolated RNA was 145 conducted with a random hexanucleotide primer and avian myeloblastosis virus reverse 146 transcriptase (Life Sciences). PCR amplification of cDNA was performed with the Expand High 147 Fidelity PCR System (Roche). RT-PCR products were analyzed by agarose gel electrophoresis 148 or were purified with QIAquick spin columns (Qiagen) prior to DNA sequencing. Northern 149 blotting analysis of purified virion RNA was carried out as detailed previously (24). RNA was 150 probed with a PCR product corresponding to the 3'-most 539 nucleotides of the N ORF and the 151 entire 3' untranslated region of the MHV genome, which was labeled with an AlkPhos Direct kit 152 and was visualized using CDP-Star detection reagent (GE Healthcare). 153 Protein analysis. Western blotting of purified virions or of NP40 lysates prepared from 154 infected 17Cl1 cell monolayers was carried out exactly as described previously (26). Prestained 155 protein standards (SeeBlue Plus2; Invitrogen) were included in adjacent lanes in SDS-PAGE. 156 Proteins were detected with anti-N monoclonal antibody J.3.3 and anti-M monoclonal antibody 157 J.1.3 (both generously provided by John Fleming, University of Wisconsin, Madison). Bound 158 antibodies were visualized by enhanced chemiluminescence detection (Pierce), which was 159 quantitated with a BioRad ChemiDoc XRS+ instrument. 160 Virus growth and competition assays. The growth kinetics of mutant and wild-type 161 viruses were measured in infections begun at multiplicities of 0.01 or 5.0 PFU per cell exactly as 162 described previously (27). To assay the relative fitness of the silPS mutant and its wild-type 163 counterpart, 10-cm wells of 17Cl1 cells were infected with a mixture of both viruses at a 164 constant multiplicity of 1 PFU per cell of the silPS virus and multiplicities of either 1, 0.1, or 165 0.01 PFU per cell of the wild type. At 12 h postinfection, when monolayers exhibited maximal 166 syncytia formation but little detachment, released virus was harvested, and RNA was purified 167 from the monolayer. Harvested virus (300 μl of 4 ml total) was used to infect a fresh set of 168 on July 6, 2017 by gest http/jvi.asm .rg/ D ow nladed fom JVI00100-13 Version 2 02.10.13 page 9 of 28 monolayers, and the process was repeated for a total of five passages. The silPS and silPS-PS2 169 mutant pair were assayed in the same manner. The composition of purified infected cell RNA 170 was monitored by RT-PCR. The primers for analysis of the native PS locus were PM166 (nsp15 171 nucleotides 352-369) and PM406 (complementary to nsp15 nucleotides 920-952); the primers 172 for analysis of the PS2 marker were CM6 (nsp16 nucleotides 634-651) and CM31 173 (complementary to nucleotides 65-82 upstream of the S gene). 174 175 on July 6, 2017 by gest http/jvi.asm .rg/ D ow nladed fom JVI00100-13 Version 2 02.10.13 page 10 of 28 RESULTS 176 177 Construction of an MHV mutant with a disrupted PS. The structure proposed by the 178 Olsthoorn laboratory (17) for the PS of lineage A betacoronaviruses is a 95-nt bulged stem-loop, 179 which, for MHV, encompasses bases 20,273 20,367 of the genome (Fig. 3A). This element 180 contains four copies of an AGC/GUAAU motif repeated at regular intervals, each copy of which 181 displays an AA or GA bulge on its 3' side. An internal loop, which varies among different 182 viruses, divides the PS into two quasi-symmetric halves. 183 To test the functional significance of the MHV PS we constructed a mutant, designated 184 silPS, in which 20 coding-silent mutations were created in the interval of the nsp15 ORF 185 containing this element (Fig. 3B). Changes were chosen to generate only codons of equivalent or 186 greater codon usage frequency for the mouse (ref. 28; http://www.kazusa.or.jp/codon/). 187 Collectively, these mutations widely altered the primary sequence and completely abolished the 188 RNA secondary structure of the PS. None of the three structures for silPS predicted by mfold 189 (29), one of which is shown in Fig. 3B, contained any apparent vestige of the features of the PS. 190 It should be noted that the 20 mutations also disrupted alternative proposed structures for the PS 191 (5, 15). Additionally, in both the silPS mutant and its corresponding wild-type counterpart, the 192 nonessential genes 2a and HE were deleted, and their associated TRS (TRS2) was knocked out 193 (Fig. 3D). The net effect of these latter alterations was the creation of a region of 128 nucleotides 194 downstream of the replicase gene that is unique to gRNA and is not transcribed into any of the 195 sgRNAs. The deletion of the 2a and HE genes of MHV has been shown to have no effect on 196 growth of the virus in tissue culture (30-32). 197 The silPS mutant and its isogenic wild-type control virus were constructed by targeted 198 recombination of synthetic donor RNAs with the interspecies chimeric coronavirus fMHV (22, 199 on July 6, 2017 by gest http/jvi.asm .rg/ D ow nladed fom JVI00100-13 Version 2 02.10.13 page 11 of 28 23). This method of coronavirus reverse genetics relies on selection for the reacquisition of the 200 MHV S gene. However, because all of the mutations of interest in this case fell upstream of the S 201 gene, silPS and wild-type recombinants were subsequently identified through screening. Two 202 independent silPS mutants, designated Alb650 and Alb651, were isolated. Since both behaved 203 identically in preliminary experiments, one of them (Alb651) was chosen for further analysis in 204 comparisons with an isogenic wild-type recombinant, Alb649. 205 Phenotype of the silPS mutant. Our initial expectation had been that disruption of the 206 PS would be severely debilitating, if not lethal, for the virus. We were therefore surprised to find 207 that the silPS mutant formed plaques that were indistinguishable in size from those of the wild 208 type at either 33, 37, or 39C (Fig. 4A). Moreover, there was little detectable difference between 209 the growth kinetics of the silPS mutant and the wild type, either in multicycle, low-multiplicity 210 infections (Fig. 4B) or in single-step, high-multiplicity infections (Fig. 4C). A second, 211 independent growth comparison gave results identical to those in Fig. 4B and C. One 212 reproducibly observed dissimilarity was that in low-multiplicity infections, after reaching a peak 213 at 30 h postinfection, titers of the silP mutant dropped some 0.5 to 1 log10 relative to those of the 214 wild type by 48 h postinfection. Also, in these experiments and in infections for preparative 215 purposes it was consistently noted that the silPS mutant, despite producing widespread syncytia, 216 never caused as extensive a cytopathic effect and detachment of the cell monolayer as did the 217 wild type. (The same was true for the subsequently isolated ΔPS mutant.) We do not yet know 218 the basis for these observations. 219 To determine the effect of the silPS mutations on packaging, we analyzed the RNA and 220 protein content of extensively purified virions. In preliminary experiments, we prepared virions 221 by two rounds of centrifugation in either glycerol-tartrate gradients (25) or sucrose gradients (4, 222 5); alternatively, virions were isolated by immunopurification (6). In agreement with the findings 223 on July 6, 2017 by gest http/jvi.asm .rg/ D ow nladed fom JVI00100-13 Version 2 02.10.13 page 12 of 28 of Escors and coworkers with TGEV (6), we observed the highest stringency of packaging, i.e., 224 the least incorporation of sgRNAs, in wild-type virions that were immunopurified (data not 225 shown). Consequently, this procedure was adopted for the remainder of the current study. The 226 silPS mutant and wild-type virus were grown, titrated by plaque assay, and immunopurified with 227 a monoclonal antibody specific for the ectodomain of the M protein, as outlined in Fig. 5A and 228 described in detail in Materials and Methods. Following quantitation of virions by Western 229 blotting for N protein, the protein compositions of equal amounts of silPS and wild-type virions 230 were examined by Western blots probed with monoclonal antibodies specific for the N and M 231 proteins (Fig. 5B) and by Coomassie blue staining of virions separated by SDS-polyacrylamide 232 gel electrophoresis (Fig. 5C). This showed that there were no differences between the mutant and 233 the wild type in either the profiles or the ratios of virion structural proteins that each contained, 234 nor did either exhibit obvious incorporation of nonviral proteins. Thus, disruption of the PS did 235 not appear to cause any gross aberrations in viral protein assembly. 236 Northern blot analysis of RNA from the cells in which virus had been grown 237 demonstrated the presence of the expected set of gRNA and sgRNAs for both the silPS mutant 238 and its wild-type counterpart (Fig. 5D). Owing to the 3'-nested set structure of coronavirus RNAs 239 (Fig. 1), all species were detected with a probe specific for the 3' end of the genome, and, 240 because of the knockout of TRS2, sgRNA2 was not synthesized. In contrast to the pattern of 241 intracellular viral RNA, RNA isolated from purified wild-type virions was almost entirely devoid 242 of sgRNA (Fig. 5E), consistent with previous work showing a high degree of selective packaging 243 of gRNA by MHV (3-5). Virions of the silPS mutant, however, were seen to have packaged 244 substantial amounts of sgRNAs, and these were present in direct proportion to their relative 245 intracellular abundance. These results established that the PS, in its native genomic locus, does 246 indeed govern the selectivity of viral gRNA packaging. 247 on July 6, 2017 by gest http/jvi.asm .rg/ D ow nladed fom JVI00100-13 Version 2 02.10.13 page 13 of 28 Construction and characterization of mutants in which the PS was deleted or 248 transposed to a different genomic site. It still remained possible that the silPS mutations failed 249 to abolish all essential features of the PS. To learn if the silPS mutant represented the null 250 phenotype for packaging we next sought to entirely delete the PS from the MHV genome. 251 Although the MHV PS falls within the ORF for nsp15, a critical component of the viral 252 replicase-transcriptase complex, the PS encodes a surface loop joining the aminoand carboxy253 terminal domains of the mature nsp15 molecule (Fig. 6A). This localization was first noted in a 254 comparison of the MHV and SARS-CoV nsp15 structures (19). A portion of the MHV nsp15 255 loop is presumed to be flexible, as it is missing from the crystal structure (33); moreover, the 256 loop falls on the outward-facing surface of each unit in the nsp15 hexamer. Accordingly, we 257 designed a packaging element deletion (ΔPS) based on an alignment of nsp15 sequences of 258 lineage A betacoronaviruses with those of other coronaviruses that lack this PS (Fig. 6B). In the 259 ΔPS mutant, nsp15 amino acids F184 through E214 were removed and replaced with a 260 heterologous pentapeptide sequence, NGNGN. At the RNA level, the corresponding short 261 substitution (AACGGCAATGGCAAC) is predicted by mfold to be completely unstructured 262 (Fig. 6C). The ΔPS mutant was also engineered to contain the same abbreviated intergenic 263 region downstream of the replicase gene as was made in the silPS mutant (Fig. 6D). As an 264 adjunct, we constructed a related mutant, ΔPS-PS2, in which the PS was deleted from its original 265 position and inserted ectopically in the intergenic region, at a position that was 15 nucleotides 266 downstream of the replicase gene stop codon and 105 nucleotides upstream of TRS3. This move 267 placed the center of the PS 1,525 nucleotides from its original position. The PS2 insertion 268 comprised the structure shown in Fig. 6C plus 35 and 38 nucleotides, respectively, of its 5' and 3' 269 flanking regions from the nsp15 ORF. 270 For each mutant, two independent isolates were obtained by targeted RNA recombination 271 on July 6, 2017 by gest http/jvi.asm .rg/ D ow nladed fom JVI00100-13 Version 2 02.10.13 page 14 of 28 and screening (Alb743 and Alb746 for ΔPS, Alb758 and Alb760 for ΔPS-PS2). Data is 272 presented for one set (Alb746 and Alb760), but both independent isolates of each mutant 273 behaved identically. As we had seen with the silPS mutant, the ΔPS mutant formed plaques at 274 33, 37, or 39C that were of equal size to those of its wild-type or ΔPS-PS2 counterparts (Fig. 275 7A). The ΔPS virus also grew to high titers comparable to those of the wild type. Thus, the virus 276 was not significantly impaired by loss of either the PS or the nsp15 peptide linker that it encodes. 277 Northern blot analysis of immunopurified virions showed that, to the same extent as the silPS 278 mutant, the ΔPS mutant packaged large amounts of sgRNAs (Fig. 7B). Side-by-side Northern 279 blots of the silPS and ΔPS mutants were indistinguishable (data not shown), which indicated that 280 the ΔPS mutant was not more extensively defective in RNA packaging selectivity than the silPS 281 mutant. By contrast, the ΔPS-PS2 mutant exhibited the same degree of gRNA packaging 282 selectivity as did the wild type (Fig. 7B). This result established that the PS, with a minimal 283 amount of flanking sequence, is sufficient to impart packaging selectivity, and it can operate at a 284 genomic site other than its native locus in the replicase gene. Additionally, the phenotype of the 285 ΔPS-PS2 mutant showed that the functionality of the PS is not coupled to translation of the 286 region in which it resides. 287 Fitness advantage provided by the PS. Although our results clearly demonstrated a 288 discriminating role for the PS in virion particle assembly, we had not observed a measurable 289 impact of that role on virus growth. To further probe whether the MHV PS provides any 290 detectable benefit to viruses that harbor it, we carried out competition assays. In the initial set of 291 experiments, monolayers of cells were infected with mixtures of the ΔPS mutant and the wild 292 type, and progeny viruses were serially propagated for a total of five passages. The input virus 293 for the first passage consisted of equal numbers of PFU of each virus, or else was weighted 10:1 294 or 100:1 in favor of the ΔPS mutant. Following each passage, infected cell RNA was isolated 295 on July 6, 2017 by gest http/jvi.asm .rg/ D ow nladed fom JVI00100-13 Version 2 02.10.13 page 15 of 28 and analyzed by RT-PCR using primers in the nsp15 ORF that would monitor the presence or 296 absence of the PS. It should be noted that the same primer pair was used to detect RNA from 297 both viruses. Thus, if there had been any bias in the PCR, it would have been expected to favor 298 the smaller amplicon (that of the ΔPS mutant). In all cases, a strong advantage for the wild-type 299 virus was obvious (Fig. 8A). At PFU input ratios of 1:1 or 10:1, the wild type was seen to 300 predominate as early as the first or second passage and the ΔPS mutant was no longer detectable 301 by the third or fourth passage. Even at a PFU input ratio of 100:1, the wild type decisively 302 overtook the ΔPS mutant by the fourth passage. An independent duplicate of the experiment 303 shown in Fig. 8A yielded the same results. This outcome implied that, in growth in tissue culture, 304 a virus containing the PS was rapidly selected over one that lacked the PS. 305 We could not completely rule out, however, that the deletion in the ΔPS mutant produced 306 some deleterious effect on nsp15. It has been shown previously that particular mutations 307 constructed in the carboxy-terminal domain of MHV nsp15 can alter viral phenotype (13, 34). 308 That this was not the case for the ΔPS mutant was suggested by experiments competing the ΔPS 309 and the silPS mutants, which did not reveal a dominance of one virus over the other. To resolve 310 this uncertainty, we performed a competition experiment between the silPS virus and a silPS-PS2 311 mutant (Alb767), the latter of which had been constructed in a manner entirely analogous to the 312 ΔPS-PS2 mutant (Fig. 8B). Thus, the two competing viruses contained the same coding-silent 313 disruption of the PS in the nsp15 ORF and differed only in the presence or absence of a copy of 314 the PS in the intergenic region downstream of the replicase gene. As with the initial competition 315 assay, the input inoculum for the first passage contained equal PFU of each virus, or else was 316 weighted 10:1 or 100:1 in favor of the silPS mutant. In this case, RT-PCR analysis was carried 317 out with a primer pair chosen to amplify the intergenic region between the replicase and S genes. 318 For a PFU input ratio of 1:1, we observed that the silPS-PS2 virus overtook the silPS virus by the 319 on July 6, 2017 by gest http/jvi.asm .rg/ D ow nladed fom JVI00100-13 Version 2 02.10.13 page 16 of 28 second or third passage. (An RT-PCR product of intermediate mobility [marked with an asterisk 320 in Fig. 8B] was shown by both sequencing and further PCR analysis to be a heteroduplex 321 composed of opposite strands of the silPS and silPS-PS2 products.) The predominance of the 322 silPS-PS2 virus was only delayed until the third or fourth passage when the PFU input ratio was 323 increased to 10:1. Even at the highly skewed input ratio of 100:1, where the silPS-PS2 mutant 324 was barely detectable at the first passage, it was seen to be increasing disproportionately by the 325 fourth and fifth passages, indicating that it was outcompeting the silPS mutant. The same 326 outcome was obtained in an independent repeat of the experiment shown in Fig. 8B. These 327 results, although not as dramatic as those with the ΔPS mutant, confirmed that the presence of 328 the PS affords a definite fitness advantage to MHV. 329 330 on July 6, 2017 by gest http/jvi.asm .rg/ D ow nladed fom JVI00100-13 Version 2 02.10.13 page 17 of 28 DISCUSSION 331 332 The MHV PS was originally defined as an element that enabled DI RNAs to be packaged 333 in the presence of co-replicating helper virus (3, 4), but the in situ role of the PS in the MHV 334 genome has until now remained unexplored. We approached this question by isolating and 335 characterizing viruses in which the PS was either disrupted (the silPS mutant) or totally ablated 336 (the ΔPS mutant). Three main conclusions can be drawn from the current study. First, the PS 337 ensures the selective inclusion of gRNA over sgRNA in virions, but it is not an absolute 338 prerequisite for gRNA packaging or for particle assembly. The silPS and ΔPS viruses each 339 exhibited a complete loss of packaging selectivity. Purified virions from both mutants contained 340 large quantities of sgRNAs, in contrast to wild-type virions, which overwhelmingly packaged 341 gRNA (Fig. 5E and 7B). However, contrary to our original assumptions, the PS mutants were 342 fully viable and exhibited little or no differences with respect to the wild type in plaque size, 343 plaque morphology, virus titers, or growth kinetics. Moreover, side-by-side preparation and 344 analysis of silPS and wild-type virions revealed that they had essentially the same particle-to345 PFU ratio (Fig. 5A). 346 A second conclusion from our work is that the PS is functional if moved to a different 347 genomic locus. The loss of packaging selectivity in the ΔPS mutant was rescued by placement of 348 a copy of the wild-type PS (PS2) downstream of the replicase gene, showing that the PS is able 349 to act at an ectopic site. The size of the PS2 insertion establishes that a segment corresponding to 350 nucleotides 20,238-20,405 of MHV gRNA is sufficient to confer packaging functionality. 351 Although this further delimits the boundaries set by other studies (9, 10, 15), the precise extent of 352 the PS remains to be defined. Finally, we were able to demonstrate that, over the course of 353 multiple passages, the PS provides a fitness advantage to the virus (Fig. 8). Even at initial PFU 354 on July 6, 2017 by gest http/jvi.asm .rg/ D ow nladed fom JVI00100-13 Version 2 02.10.13 page 18 of 28 input ratios that were biased 100:1 to their disadvantage, viruses containing the PS outgrew their 355 otherwise identical counterparts that lacked the PS. This suggests that there must have existed 356 strong evolutionary pressure to retain the PS, once it had been acquired by some common 357 ancestor of the lineage A betacoronaviruses. 358 For some RNA viruses, such as poliovirus (35) and brome mosaic virus (36), packaging 359 is thought to be primarily driven by a close association between capsid proteins and nascent 360 genomes. Such an association can result from co-compartmentalization or, more directly, from a 361 mechanistic coupling between packaging and RNA replication. For other RNA viruses, such as 362 alphaviruses (37) and retroviruses (38), distinct genomic RNA sequences or structures have been 363 found to play determining roles in the packaging process. Our current results indicate that, for 364 MHV, both gRNA and sgRNAs have a propensity to be incorporated into assembling virions. 365 This may be a consequence of their mutual localization in the double-membrane vesicle 366 compartments that are set up by the viral replicase-transcriptase complex (39). Alternatively, 367 there may exist elements other than the PS, common to all viral RNAs, that contribute to the 368 recognition process. The role of the PS appears to be to tilt the balance in the competition 369 between gRNA and the stoichiometrically much more numerous sgRNAs. 370 Perhaps the closest parallel with our findings can be found in the recently discovered PS 371 of Venezuelan equine encephalitis virus (VEEV) and related alphaviruses (37). VEEV packaging 372 discriminates between gRNA and a single, abundantly transcribed sgRNA. This is accomplished 373 through recognition of a PS that has been mapped to the nsP1 gene, which is unique to gRNA. 374 Like the MHV PS, the VEEV PS contains a single-stranded purine repeat, in this case a GGG 375 motif displayed in the loops of four to six adjacent hairpin structures. Also, similar to the MHV 376 PS, the VEEV PS retained its function when transposed to a different genomic site. Coding-silent 377 mutational disruption of the VEEV PS caused a high level of packaging of sgRNA. However, 378 on July 6, 2017 by gest http/jvi.asm .rg/ D ow nladed fom JVI00100-13 Version 2 02.10.13 page 19 of 28 unlike the MHV PS mutants, VEEV PS mutants exhibited more severe defects in viral growth 379 kinetics and 30to 100-fold drops in peak infectious titers, compared to the wild type. This more 380 drastic phenotype may result from a limitation on the amount of RNA that can be packaged into 381 the icosahedral alphavirus capsid. By contrast, and highly unusually for positive-strand RNA 382 viruses, coronaviruses have helically symmetric nucleocapsids and pleomorphic envelopes that 383 are likely more able to accommodate additional RNA. 384 Our demonstration of the principal, and possibly exclusive, role of the MHV PS in the 385 selection of gRNA for virion assembly raises a number of further issues. One of the most 386 immediate questions is the identity of the interacting partner that carries out the selection of 387 gRNA. The obvious candidate for this activity is the nucleocapsid (N) protein, which is the only 388 known protein constituent of the viral nucleocapsid. Indeed, specific binding of N protein to the 389 PS has been detected in vitro by gel-shift and UV-crosslinking assays (9). On the other hand, a 390 study with MHV virus-like particles found that the principle component of the virion envelope, 391 the membrane (M) protein, was responsible for the selective packaging of nonviral RNA 392 containing the PS, and this packaging occurred in the absence of N protein (40). We hope to use 393 genetic approaches to clarify these apparently contradictory results. Another significant question 394 arising from the role of the MHV PS is how packaging selectivity is determined in coronaviruses 395 that fall outside of betacoronavirus lineage A. It can readily be seen that the region of the nsp15 396 ORF that contains the MHV PS is absent from most other coronaviruses (Fig. 6B). Dissection of 397 packaged DI RNAs has shown that the PS of the alphacoronavirus TGEV is confined to the 5'398 most 649 nucleotides of the genome of that virus (6). Similarly, the composition of a packaged 399 DI RNA of the gammacoronavirus IBV rules out almost all of the nsp15 ORF as a potential 400 locus for the PS of that virus (21). Thus, the PSs of different groups of coronaviruses have 401 evolved at different genomic locations, in a manner analogous to the disperse positioning of cre 402 on July 6, 2017 by gest http/jvi.asm .rg/ D ow nladed fom JVI00100-13 Version 2 02.10.13 page 20 of 28 elements in picornavirus genomes (41, 42). Our MHV PS mutants, if combined with the 403 appropriate selective factor(s), may provide an opportunity to identify heterologous PS elements. 404 Specifically, we would like to be able to test the prediction that some PSs encompass particular 405 repeated loop motifs found in the 5' UTRs of various alphaand betacoronavirus genomes (20). 406 Such future work will potentially reveal unifying principles governing coronavirus gRNA 407 packaging. 408 409 410 411 412 413 414 415 on July 6, 2017 by gest http/jvi.asm .rg/ D ow nladed fom JVI00100-13 Version 2 02.10.13 page 21 of 28 ACKNOWLEDGMENTS 416 417 We are grateful to Joachim Jaeger for providing the PyMol figure of MHV nsp15. We 418 thank the Applied Genomics Technology Core Facility of the Wadsworth Center for DNA 419 sequencing. 420 This work was supported by National Institutes of Health (National Institute of Allergy 421 and Infectious Diseases) grants R01 AI064603 and R56 AI064603. 422 423 REFERENCES 424 425 1. Masters PS. 2006. The molecular biology of coronaviruses. Adv. Virus Res. 66:193-292. 426 427 2. Perlman S, Netland J. 2009. Coronaviruses post-SARS: update on replication and 428 pathogenesis. Nat. Rev. Microbiol. 7:439-450. 429 430 3. Makino S, Yokomori K, Lai MMC. 1990. Analysis of efficiently packaged defective 431 interfering RNAs of murine coronavirus: localization of a possible RNA-packaging 432 signal. J. Virol. 64:6045-6053. 433 434 4. van der Most RG, Bredenbeek PJ, Spaan WJM. 1991. A domain at the 3' end of the 435 polymerase gene is essential for encapsidation of coronavirus defective interfering RNAs. 436 J. Virol. 65:3219-3226. 437 438 5. Fosmire JA, Hwang K, Makino S. 1992. Identification and characterization of a 439 coronavirus packaging signal. J. Virol. 66:3522-3530. 440 441 6. Escors D, Izeta A, Capiscol C, Enjuanes L. 2003. Transmissible gastroenteritis 442 coronavirus packaging signal is located at the 5' end of the virus genome. J. Virol. 443 77:7890-7902. 444 445 7. Hofmann MA, Sethna PB, Brian DA. 1990. Bovine coronavirus mRNA replication 446 continues throughout persistent infection in cell culture. J. Virol. 64:4108-4114. 447 448 8. Zhao X, Shaw K, Cavanagh D. 1993. Presence of subgenomic mRNAs in virions of 449 coronavirus IBV. Virology 196:172-178. 450 451 9. Molenkamp R, Spaan WJM. 1997. Identification of a specific interaction between the 452 coronavirus mouse hepatitis virus A59 nucleocapsid protein and packaging signal. 453 Virology 239:78-86. 454 455 on July 6, 2017 by gest http/jvi.asm .rg/ D ow nladed fom JVI00100-13 Version 2 02.10.13 page 22 of 28 10. Narayanan K, Makino S. 2001. Cooperation of an RNA packaging signal and a viral 456 envelope protein in coronavirus RNA packaging. J. Virol. 75:9059-9067. 457 458 11. Ivanov KA, Hertzig T, Rozanov M, Bayer S, Thiel V, Gorbalenya AE, Ziebuhr J. 459 2004. Major genetic marker of nidoviruses encodes a replicative endoribonuclease. Proc. 460 Natl. Acad. Sci. USA 101:12694-12699. 461 462 12. Ricagno S, Egloff MP, Ulferts R, Coutard B, Nurizzo D, Campanacci V, Cambillau 463 C, Ziebuhr J, Canard B. 2006. Crystal structure and mechanistic determinants of SARS 464 coronavirus nonstructural protein 15 define an endoribonuclease family. Proc. Natl. 465 Acad. Sci. USA 103:11892-11897. 466 467 13. Kang H, Bhardwaj K, Li Y, Palaninathan S, Sacchettini J, Guarino L, Leibowitz 468 JL, Kao CC. 2007. Biochemical and genetic analyses of murine hepatitis virus nsp15 469 endoribonuclease. J. Virol. 81:13587-13597. 470 471 14. Woo K, Joo M, Narayanan K, Kim KH, Makino S. 1997. Murine coronavirus 472 packaging signal confers packaging to nonviral RNA. J. Virol. 71:824-827. 473 474 15. Cologna R, Hogue BG. 2000. Identification of a bovine coronavirus packaging signal. J. 475 Virol. 74:580-583. 476 477 16. Bos ECW, Dobbe JC, Luytjes W, Spaan WJM. 1997. A subgenomic mRNA transcript 478 of the coronavirus mouse hepatitis virus strain A59 defective interfering (DI) RNA is 479 packaged when it contains the DI packaging signal. J. Virol. 71:5684-5687. 480 481 17. Chen SC, van den Born E, van den Worm SH, Pleij CW, Snijder EJ, Olsthoorn RC. 482 2007. New structure model for the packaging signal in the genome of group IIa 483 coronaviruses. J. Virol. 81:6771-6774. 484 485 18. Carstens EB. 2009. Ratification vote on taxonomic proposals to the International 486 Committee on Taxonomy of Viruses. Arch. Virol. 155:133-146. 487 488 19. Joseph JS, Saikatendu KS, Subramanian V, Neuman BW, Buchmeier MJ, Stevens 489 RC, Kuhn P. 2007. Crystal structure of a monomeric form of severe acute respiratory 490 syndrome coronavirus endonuclease nsp15 suggests a role for hexamerization as an 491 allosteric switch. J. Virol. 81:6700-6708. 492 493 20. Chen SC, Olsthoorn RC. 2010. Group-specific structural features of the 5'-proximal 494 sequences of coronavirus genomic RNAs. Virology 401:29-41. 495 496 21. Penzes Z, Tibbles K, Shaw K, Britton P, Brown TDK, Cavanagh D. 1994. 497 Characterization of a replicating and packaged defective RNA of avian coronavirus 498 infectious bronchitis virus. Virology 203:286-293. 499 500 22. Goebel SJ, Hsue B, Dombrowski TF, Masters PS. 2004. Characterization of the RNA 501 components of a putative molecular switch in the 3' untranslated region of the murine 502 coronavirus genome. J. Virol. 78:669-682. 503 on July 6, 2017 by gest http/jvi.asm .rg/ D ow nladed fom JVI00100-13 Version 2 02.10.13 page 23 of 28 23. Kuo L, Godeke GJ, Raamsman MJ, Masters PS, Rottier PJ. 2000. Retargeting of 504 coronavirus by substitution of the spike glycoprotein ectodomain: crossing the host cell 505 species barrier. J. Virol. 74:1393-1406. 506 507 24. Kuo L, Masters PS. 2010. Evolved variants of the membrane protein can partially 508 replace the envelope protein in murine coronavirus assembly. J. Virol. 84:12872-12885. 509 510 25. Ye R, Montalto-Morrison C, Masters PS. 2004. Genetic analysis of determinants for 511 spike glycoprotein assembly into murine coronavirus virions: distinct roles for charge512 rich and cysteine-rich regions of the endodomain. J. Virol. 78:9904-9917. 513 514 26. Hurst KR, Koetzner CA, Masters PS. 2009. Identification of in vivo-interacting 515 domains of the murine coronavirus nucleocapsid protein. J. Virol. 83:7221-7234. 516 517 27. Züst R, Miller TB, Goebel SJ, Thiel V, Masters PS. 2008. Genetic interactions 518 between an essential 3' cis-acting RNA pseudoknot, replicase gene products, and the 519 extreme 3' end of the mouse coronavirus genome. J. Virol. 82:1214-1228. 520 521 28. Nakamura Y, Gojobori T, Ikemura T. 2000. Codon usage tabulated from international 522 DNA sequence databases: status for the year 2000. Nucleic Acids Res. 28:292. 523 524 29. Zuker M. 2003. Mfold web server for nucleic acid folding and hybridization prediction. 525 Nucleic Acids Res. 31:3406-3415. 526 527 30. de Haan CA, Masters PS, Shen X, Weiss S, Rottier PJ. 2002. The group-specific 528 murine coronavirus genes are not essential, but their deletion, by reverse genetics, is 529 attenuating in the natural host. Virology 296:177-189. 530 531 31. Koetzner CA, Kuo L, Goebel SJ, Dean AB, Parker MM, Masters PS. 2010. 532 Accessory protein 5a is a major antagonist of the antiviral action of interferon against 533 murine coronavirus. J. Virol. 84:8262-8274. 534 535 32. Zhao L, Jha BK, Wu A, Elliott R, Ziebuhr J, Gorbalenya AE, Silverman RH, Weiss 536 SR. 2012. Antagonism of the interferon-induced OAS-RNase L pathway by murine 537 coronavirus ns2 protein is required for virus replication and liver pathology. Cell Host 538 Microbe 11:607-616. 539 540 33. Xu X, Zhai Y, Sun F, Lou Z, Su D, Xu Y, Zhang R, Joachimiak A, Zhang XC, 541 Bartlam M, Rao Z. 2006. New antiviral target revealed by the hexameric structure of 542 mouse hepatitis virus nonstructural protein nsp15. J. Virol. 80:7909-7917. 543 544 34. Bhardwaj K, Liu P, Leibowitz JL, Kao CC. 2012. The coronavirus endoribonuclease 545 nsp15 interacts with retinoblastoma tumor suppressor protein. J. Virol. 86:4294-4304. 546 547 35. Nugent CI, Johnson KL, Sarnow P, Kirkegaard K. 1999. Functional coupling between 548 replication and packaging of poliovirus replicon RNA. J. Virol. 73:427-435. 549 550 36. Annamalai P, Rao ALN. 2006. Packaging of brome mosaic virus subgenomic RNA is 551 on July 6, 2017 by gest http/jvi.asm .rg/ D ow nladed fom JVI00100-13 Version 2 02.10.13 page 24 of 28 functionally coupled to replication-dependent transcription and translation of coat 552 protein. J. Virol. 80:10096-10108. 553 554 37. Kim DY, Firth AE, Atasheva S, Frolova EI, Frolov I. 2011. Conservation of a 555 packaging signal and the viral genome RNA packaging mechanism in alphavirus 556 evolution. J. Virol. 85:8022-8036. 557 558 38. D'Souza V, Summers MF. 2005. How retroviruses select their genomes. Nat Rev 559 Microbiol. 3:643-655. 560 561 39. Knoops K, Kikkert M, Worm SH, Zevenhoven-Dobbe JC, van der Meer Y, Koster 562 AJ, Mommaas AM, Snijder EJ. 2008. SARS-coronavirus replication is supported by a 563 reticulovesicular network of modified endoplasmic reticulum. PLoS Biol. 6:e226. 564 565 40. Narayanan K, Chen CJ, Maeda J, Makino S. 2003. Nucleocapsid-independent specific 566 viral RNA packaging via viral envelope protein and viral RNA signal. J. Virol. 77:2922567 2927. 568 569 41. Cordey S, Gerlach D, Junier T, Zdobnov EM, Kaiser L, Tapparel C. 2008. The cis570 acting replication elements define human enterovirus and rhinovirus species. RNA 571 14:1568-1578. 572 573 42. Steil BP, Barton DJ. 2009. Cis-active RNA elements (CREs) and picornavirus RNA 574 replication. Virus Res. 139:240-252. 575 576 on July 6, 2017 by gest http/jvi.asm .rg/ D ow nladed fom JVI00100-13 Version 2 02.10.13 page 25 of 28 FIGURE LEGENDS 577 578 FIG. 1. MHV RNA species. The 31.3-kb MHV genome (gRNA) is shown with an expanded 579 segment detailing the 3' terminus of the replicase gene (rep 1a and 1b), including the region 580 encoding nsp14 nsp 16. The position of the packaging signal (PS) within the coding region for 581 nsp15 is indicated. Downstream of the replicase gene are the genes for structural proteins (spike 582 [S], envelope [E], membrane [M], and nucleocapsid [N]) and for accessory proteins (2a, 583 hemagglutinin-esterase [HE], 4, 5a, and internal [I]). Beneath the gRNA is the 3'-nested set of 584 transcribed subgenomic RNA (sgRNA) species that is a defining characteristic of coronaviruses 585 and other members of the order Nidovirales. 586 587 FIG. 2. Donor RNA transcription vector for construction of MHV PS mutants by targeted RNA 588 recombination (22, 23). All mutants in the current work originated with the parent vector pPM9, 589 which was derived from the previously described pMH54 (23) by the addition of genomic cDNA 590 upstream of the S gene. The locus containing the PS is indicated by a bar above nsp15. Shown 591 are coding-silent unique XbaI and BspEI sites flanking the PS and unique SalI and AscI sites in 592 the truncated nonessential intergenic region between nsp16 and the S gene. 593 594 FIG. 3. Construction of the silPS mutant. (A) Model for the MHV PS proposed by Chen et al. 595 (17). The four repeating units with AA (or GA) bulges are boxed. (B) Positions and identities of 596 20 mutations (circled nucleotides) made to disrupt the structure of the PS without altering the 597 encoded amino-acid sequence of nsp15. The resulting lowest free energy structure predicted by 598 mfold (29) for the mutated sequence is shown on the right. (C) Schematics of the relevant 599 genomic regions of the constructed silPS mutant and its isogenic wild-type counterpart. (D) 600 on July 6, 2017 by gest http/jvi.asm .rg/ D ow nladed fom JVI00100-13 Version 2 02.10.13 page 26 of 28 Detail of the region downstream of the replicase gene, in which TRS2 was knocked out, all of 601 gene 2a was deleted, and all except 99 nucleotides of the HE gene was deleted. The positions of 602 SalI and AscI sites created in parent transcription vector pPM9 are shown. Filled circles above 603 the sequence indicate nucleotides that were mutated to knock out TRS2. 604 605 FIG. 4. Growth characteristics of the silPS mutant. (A) Plaques of the silPS mutant at 33, 37 or 606 39C compared with those of the isogenic wild-type virus. Plaque titrations were carried out on 607 L2 cells; monolayers were stained with neutral red at 72 h postinfection and were photographed 608 18 h later. (B and C) Growth kinetics of the silPS mutant relative to the wild type. Confluent 609 monolayers of 17Cl1 cells were infected at a multiplicity of 0.01 PFU per cell (B) or 5.0 PFU per 610 cell (C). At the indicated times postinfection, aliquots of medium were removed, and infectious 611 titers were determined by plaque assay on L2 cells. 612 613 FIG. 5. RNA packaging phenotype of the silPS mutant. (A) Outline of the procedures for 614 purification, normalization, and analysis of silPS and wild-type virions, as detailed in Materials 615 and Methods. Infectious titers were determined by plaque assay on L2 cells at 37C; 616 chemiluminescence values are given in arbitrary volume units, as measured with a BioRad 617 ChemiDoc XRS+ instrument. (B) Western blot of normalized amounts of immunopurified wild618 type and silPS virions probed with monoclonal anti-N antibody J.3.3 and monoclonal anti-M 619 antibody J.1.3. (C) Coomassie blue-stained SDS-polyacrylamide gel of normalized amounts of 620 immunopurified wild-type and silPS virions; samples on the right are a five-fold dilution of those 621 on the left. In (B) and (C) molecular mass standards (kDa) are indicated on the left of each panel. 622 (D) Northern blot of total RNA isolated from infected 17Cl1 cells, from which wild-type or silPS 623 virions were purified, or from mock-infected cells. (E) Northern blot of RNA isolated from 624 on July 6, 2017 by gest http/jvi.asm .rg/ D ow nladed fom JVI00100-13 Version 2 02.10.13 page 27 of 28 normalized amounts of immunopurified wild-type and silPS virions. In (D) and (E) MHV RNA 625 was detected with a probe specific for the 3' end of the genome. 626 627 FIG. 6. Construction of the ΔPS and ΔPS-PS2 mutants. (A) Ribbon diagram and molecular 628 surface rendering of the MHV nsp15 monomer (PDB accession code 2GTH; ref. 33) generated 629 with PyMol (http://pymol.org). The nsp15 amino and carboxy termini are labeled N and C, 630 respectively. The dotted line represents part of the surface loop connecting the aminoand 631 carboxy-terminal domains of the protein. (B) Alignment of the central region of nsp15 of MHV 632 with that of representative coronaviruses: lineage A betacoronaviruses BCoV and HCoV-HKU1, 633 lineage B betacoronavirus SARS-CoV, alphacoronavirus TGEV, and gammacoronavirus IBV. 634 GenBank accession numbers for the sequences shown are: MHV, AY700211; BCoV, U00735; 635 HCoV-HKU1, AY597011; SARS-CoV, AY278741; TGEV, AJ271965; and IBV, AJ311317. 636 Boxed residues in the MHV sequence are those that are encoded by the PS and that were deleted 637 in the ΔPS mutant. The bar above the alignment indicates flexible loop residues (Y195 through 638 L216) that are missing from the crystal structure of MHV nsp15 (33). Dotted lines connect 639 deleted and missing residues to their corresponding positions in the structure in (A). At the 640 bottom of the alignment, the five heterologous amino acids encoded by the substituted sequence 641 in the ΔPS mutant are underlined. (C) Representation of the unstructured 15-nt sequence 642 substituted for the PS in the ΔPS mutant, compared to the wild-type PS. (D) Schematics of the 643 relevant genomic regions of the constructed ΔPS mutant, the ΔPS-PS2 mutant, and their isogenic 644 wild-type counterpart. 645 646 FIG. 7. Growth and RNA packaging by the ΔPS and ΔPS-PS2 mutants. (A) Plaques of the ΔPS 647 and ΔPS-PS2 mutants at 33, 37, or 39C compared with those of the isogenic wild-type virus. 648 on July 6, 2017 by gest http/jvi.asm .rg/ D ow nladed fom JVI00100-13 Version 2 02.10.13 page 28 of 28 Plaque titrations were carried out on L2 cells; monolayers were stained with neutral red at 72 h 649 postinfection and were photographed 18 h later. (B) Northern blot of RNA isolated from 650 normalized amounts of immunopurified wild-type, ΔPS, and ΔPS-PS2 virions detected with a 651 probe specific for the 3' end of the genome. (C) Western blot of normalized amounts of 652 immunopurified wild-type, ΔPS, and ΔPS-PS2 virions probed with monoclonal anti-N antibody 653 J.3.3 and monoclonal anti-M antibody J.1.3. 654 655 FIG. 8. Relative fitness of PS mutants. (A) Monolayers of 17Cl1 cells were co-infected with ΔPS 656 and wild-type viruses at an initial input PFU ratio of 1:1, 10:1 or 100:1, as detailed in Materials 657 and Methods. Harvested released virus was serially propagated for a total of five passages. At 658 each passage, RNA was isolated from infected cells and analyzed by RT-PCR, using a pair of 659 primers flanking the central region of the nsp15 ORF to assay the presence or absence of the PS. 660 PCR products were analyzed by agarose gel electrophoresis; the sizes (bp) of DNA markers are 661 indicated at the left of each gel. Control lanes show RT-PCR products obtained from infections 662 with wild-type virus alone or with ΔPS mutant alone or from uninfected cells (mock). (B) 663 Competition between silPS and silPS-PS2 viruses was evaluated by the same procedure as in 664 (A), except that RT-PCR was carried out with a pair of primers flanking the intergenic region 665 between the replicase and the S ORFs to assay the presence or absence of the transposed PS 666 element (PS2). The asterisk to the right of each agarose gel marks the position of an artefactual 667 heteroduplex band formed by opposite strands of the 501-bp silPS-PS2 product and the 341-bp 668 silPS product. In the schematics in (A) and (B), the positions and sizes of PCR primers are not 669 drawn to scale. 670 671 on July 6, 2017 by gest http/jvi.asm .rg/ D ow nladed fom