In Situ Tagged nsp15 Reveals Interactions with Coronavirus Replication/Transcription Complex-Associated Proteins

ABSTRACT Coronavirus (CoV) replication and transcription are carried out in close proximity to restructured endoplasmic reticulum (ER) membranes in replication/transcription complexes (RTC). Many of the CoV nonstructural proteins (nsps) are required for RTC function; however, not all of their functions are known. nsp15 contains an endoribonuclease domain that is conserved in the CoV family. While the enzymatic activity and crystal structure of nsp15 are well defined, its role in replication remains elusive. nsp15 localizes to sites of RNA replication, but whether it acts independently or requires additional interactions for its function remains unknown. To begin to address these questions, we created an in situ tagged form of nsp15 using the prototypic CoV, mouse hepatitis virus (MHV). In MHV, nsp15 contains the genomic RNA packaging signal (P/S), a 95-bp RNA stem-loop structure that is not required for viral replication or nsp15 function. Utilizing this knowledge, we constructed an internal hemagglutinin (HA) tag that replaced the P/S. We found that nsp15-HA was localized to discrete perinuclear puncta and strongly colocalized with nsp8 and nsp12, both well-defined members of the RTC, but not the membrane (M) protein, involved in virus assembly. Finally, we found that nsp15 interacted with RTC-associated proteins nsp8 and nsp12 during infection, and this interaction was RNA independent. From this, we conclude that nsp15 localizes and interacts with CoV proteins in the RTC, suggesting it plays a direct or indirect role in virus replication. Furthermore, the use of in situ epitope tags could be used to determine novel nsp-nsp interactions in coronaviruses.

IMPORTANCE Despite structural and biochemical data demonstrating that the coronavirus nsp15 protein contains an endoribonuclease domain, its precise function during coronavirus infection remains unknown. In this work, we created a novel in situ tagged form of nsp15 to study interactions and localization during infection. This in situ tag was tolerated by MHV and did not affect viral replication. Utilizing this tag, we established that nsp15 localized to sites of replication but not sites of assembly throughout infection. Furthermore, we found that nsp15 interacted with the putative viral primase nsp8 and polymerase nsp12 during CoV infection. The strong association of nsp15 with replication complexes and interactions with replicative CoV enzymes suggest nsp15 is involved in CoV replication. These data and tools developed in this study help elucidate the function of nsp15 during infection and may be used to uncover other novel viral protein interactions.
C oronaviridae, members of the Nidovirales order, are a family of positive-sense RNA (ϩssRNA) viruses that infect a wide range of host species. Generally, human coronavirus (CoV) infections cause mild disease with upper respiratory tract and gastrointestinal symptoms. In contrast, two human CoVs, severe acute respiratory syndrome (SARS)-CoV and Middle East respiratory syndrome (MERS)-CoV, recently emerged from zoonotic sources into the human population and caused severe respiratory disease with high morbidity and mortality rates (1)(2)(3). After the emergence of SARS-CoV in 2002 to 2003, efforts were made to better understand CoV replication and to develop therapies and vaccines to reduce CoV-mediated morbidity and mortality. These efforts expanded our understanding of the structure and function of several CoV proteins and of CoV replication; however, there are many aspects of the replication cycle that require further investigation (4).
Following binding and internalization of the virion, the CoV genome is deposited into the cytoplasm and translated into two large polyproteins, which account for two-thirds of the genome. These polyproteins are then cleaved by viral proteases into the nonstructural proteins nsp1 to -16. The nsps then establish a replication/transcription complex (RTC) on endoplasmic reticulum (ER) membranes, which have been restructured by viral transmembrane proteins (5,6). To date, all studied nsps have been demonstrated to localize to replication compartments (6)(7)(8)(9)(10)(11)(12), except nsp14 and nsp16, which have not been studied. However, the precise configuration of the RTC, the binding partners of specific nsps, and the role of each nsp in replication of genomic RNA (gRNA) and transcription of subgenomic RNA (sgRNA) are not well understood.
Our current understanding of most nsp interactions comes from two-hybrid screens (13)(14)(15), cell-free in vitro assays (16), structural assays (17)(18)(19), or overexpression studies (11). To date, two CoV complexes containing nsp12, the RNA-dependent RNA polymerase (RdRp), have been described: (i) a complex of nsp7, nsp8, nsp12, and nsp14 demonstrated processive RNA synthesis in vitro (16), and (ii) a complex of nsp5, nsp8, nsp9, and nsp12 was immunoprecipitated from mouse hepatitis virus (MHV)-infected cells (9), but its function was not demonstrated. Since the majority of nsps localize to RTCs, it is likely additional interactions drive virus RNA replication and subgenomic transcription. However, due to relatively low levels of nsps produced during infection, it has been difficult to identify these interactions during a natural infection.
nsp15 contains a conserved uridine-specific endoribonuclease domain with an unknown function in CoV infection (20,21). The endoribonuclease activity of nsp15 is conserved in CoVs and arteriviruses, but is not conserved among other nidoviruses (roniviruses and mesoniviruses) (22,23). This lack of conservation raises the possibility that nsp15 does not function only in virus replication, but rather is also involved in immune evasion or another host-specific function. nsp15 forms a homohexamer, which is required for RNA binding and cell-free cleavage assays (24)(25)(26)(27)(28). CoVs and arteriviruses with endoribonuclease catalytic mutations have reduced levels of replication, as assessed by levels of infectious virus, gRNA, and subgenomic RNA (sgRNA); these effects are more pronounced in arteriviruses than in CoVs (29,30). nsp15 was shown to colocalize with replicating RNA, but its precise localization throughout infection, interactions with other viral proteins, and physiological role are poorly understood (10). nsp15 also contains the only known RNA packaging signal (termed "P/S" herein) in lineage A ␤-CoVs; this signal has been partially characterized and contains a stem-loop structure. We took advantage of the known structure of the P/S to introduce an epitope tag into nsp15 that would be useful for subsequent studies. Due to their cleavage from a larger polyprotein, N-and C-terminal nsp epitope tags are not always feasible. It was previously reported that an epitope tag inserted into the native location of nsp4 was lethal to the virus; however, the virus was viable if the epitopetagged-nsp4 was expressed as an sgRNA (11). Notable exceptions include green fluorescent protein (GFP)-tagged nsp2 (31,32), which is not required for replication (33), and GFP-nsp3 (32), which results in significant virus attenuation. To circumvent these problems, we inserted an influenza A virus hemagglutinin (HA) epitope tag into the P/S of MHV strain A59 (rA59 Nsp15-HA ). This HA tag was predicted to be useful for identification of protein-protein interactions and localization during infection.
The P/S is conserved among lineage A ␤-coronaviruses but is not present in nsp15 of other closely related ␤-coronaviruses and is essential for selective packaging of gRNA (34)(35)(36)(37). Previous studies have shown that this region may be removed with no effect on viral titers, suggesting that specific amino acids in this region are not important for nsp15 function (36). rA59 Nsp15-HA replicated equivalently to the wild-type strain, rA59 WT , demonstrating that this tag did not substantially affect virus replication or, by inference, significantly affect nsp15 function. We found that nsp15-HA colocalizes and interacts with CoV RTC-associated proteins nsp8 and nsp12 during infection and that this interaction was independent of RNA intermediates. Together, these data indicate that nsp15 is a component of the CoV RTC. Our data also highlight the potential utility of using internal tags to monitor the expression, localization, and interactions of CoV proteins.

RESULTS
Constructing recombinant MHV with HA-tagged nsp15. To study the role of nsp15 during infection and address whether nsp15 interacts with other viral proteins, we constructed an in situ tagged form of nsp15 in MHV. Current antibodies used for the study of MHV nsps are limited to rabbit polyclonal sera. In many cases, these antibodies have significant background binding to host proteins, limiting their downstream applications. To circumvent this barrier, we constructed an HA-tagged form of nsp15 in MHV. In MHV, nsp15 contains the P/S, which is an ideal site for in situ tag insertion. The P/S, which forms a conserved stem-loop structure (Fig. 1A, left), is conserved among lineage A ␤-coronaviruses but not other closely related ␤-coronaviruses (e.g., SARS-CoV) (Fig. 1C) (38). Studies have demonstrated that the P/S can be deleted from MHV without causing significant growth defects (36). Finally, the amino acids encoded by the P/S are predicted to form a flexible loop, which is surface exposed on the monomer and hexamer of nsp15 ( Fig. 1B) (24). With these characteristics in mind, we decided to design our in situ epitope tags around the MHV P/S. Initially, we replaced 66 bp of the P/S with that of a 3ϫFLAG sequence, rA59 Nsp15-FLAG (Fig. 1A, right). We were able to rescue recombinant virus but were unable to detect any FLAG-specific signal by immunoblotting or immunofluorescence. Sequence analyses of the P/S revealed a near complete deletion of the 3ϫFLAG sequence from these viruses during initial virus replication after 5 passages (data not shown). The 3ϫFLAG tag was likely unsuccessful due to large differences predicted in the secondary structure and Gibbs free energy (ΔG) of the rA59 WT and the 3ϫFLAG-containing P/S (Fig. 1A, right). Considering these data, we set out to create an internal tag that maintained a stem-loop structure in the P/S more closely mimicking the wild-type stem-loop. To achieve this, we inserted the HA tag sequence into the ascending stem of the P/S and its complement into the descending stem of the P/S, creating rA59 Nsp15-HA (Fig. 1A, middle, C). A recombinant virus containing this HA tag was created using an in vitro ligation system as previously described (39).
nsp15-HA is expressed and is stable during serial passaging. Following rescue of rA59 Nsp15-HA , we first examined the expression of nsp15-HA during MHV infection. 17Cl-1 cells were infected with rA59 Nsp15-HA or rA59 WT and analyzed by confocal microscopy and immunoblotting for HA signal. We found that anti-HA antibody could detect nsp15-HA in rA59 Nsp15-HA -infected cells with high specificity and low background ( Fig. 2A and B). Furthermore, nsp15 was localized to tight perinuclear puncta in both rA59 Nsp15-HA -and rA59 WT -infected cells, indicating the localization of nsp15 was not altered in rA59 Nsp15-HA . We also detected nsp15-HA in rA59 Nsp15-HA cell lysates and found nsp15 levels were equivalent in rA59 Nsp15-HA -and rA59 WT -infected cells (Fig. 2C). These data indicated our HA tag was detectable, specific, and did not alter nsp15 expression or localization during MHV infection.
Next to examine the stability of the in situ HA sequence, rA59 Nsp15-HA was serially passaged on 17Cl-1 cells. Progeny viruses were collected at each passage and used to infect the next set of 17Cl-1 cells using a multiplicity of infection (MOI) of approximately 0.1 PFU/cell. Subsequently, each passage was analyzed for the presence of the in situ HA epitope tag by immunoblotting. nsp15-HA protein levels did not diminish over 5 passages (Fig. 2F), suggesting the in situ HA tag was stable over multiple passages, unlike the in situ 3ϫFLAG tag. rA59 Nsp15-HA has similar replication kinetics to rA59 WT . We next compared the kinetics of rA59 Nsp15-HA and rA59 WT replication, by infecting 17Cl-1 cells and measuring virus production under multistep growth conditions. rA59 Nsp15-HA replicated similarly compared to rA59 WT (Fig. 2D). To further support these data, 17Cl-1 cells were infected Input virus was collected following the adsorption step. Virus titers were determined by plaque assay on HeLa-MVR cells (E). 17Cl-1 cells were infected with rA59 WT or rA59 Nsp15-HA at an MOI of 5. Cells were collected at the indicated time points. Levels of gRNA and sgRNA were determined by RT-qPCR and normalized (Continued on next page) nsp15 Interacts with CoV RTC-Associated Proteins ® with rA59 WT or rA59 Nsp15-HA , and viral gRNA and sgRNA (sgRNA7) were measured in single-step growth curves. Levels of sgRNA and gRNA in rA59 Nsp15-HA -and rA59 WTinfected cells differed by no more than 2-fold from 4 to 12 h postinfection (hpi) (Fig. 2E). Furthermore, amounts of both nsp8 and N proteins were equivalent in rA59 Nsp15-HAand rA59 WT -infected 17Cl-1 cells (Fig. 2C). From these data, we conclude that insertion of an HA tag into the P/S within nsp15 did not substantially affect virus replication in vitro.
rA59 Nsp15-HA has altered selective gRNA packaging. During our investigation of viral growth kinetics, we noted that rA59 Nsp15-HA had significantly increased levels of sgRNA at 2 hpi. We hypothesized that this increase in sgRNA levels was too early to be due to sgRNA transcription and attributed it to increased packaging of sgRNA into virions by rA59 Nsp15-HA . This would be in agreement with previous reports, which demonstrated that P/S mutants altered selective gRNA packaging, so that sgRNA was now incorporated into virions, but did not affect viral replication (36). To determine the efficiency of selective packaging in rA59 Nsp15-HA , we measured the ratio of sgRNA (sgRNAs 4, 5, 6, and 7) to gRNA in virus isolated from infected cell supernatants by ultracentrifugation. Our results showed that viral RNA obtained from rA59 Nsp15-HAinfected cell supernatants had at least a 25-fold increase in the sgRNA/gRNA ratio (Fig. 3A). Previous work demonstrated that P/S mutants and wild-type virus have identical growth kinetics, but P/S mutants are readily outcompeted by rA59 WT when cells are dually infected (36). In order to test whether the rA59 Nsp15-HA virus would be outcompeted by rA59 WT virus, cells were infected with 1:1 or 3:1 mixtures of rA59 Nsp15-HA and rA59 WT at a total MOI of 0.1. This low MOI was used to ensure that only a small number of cells would be infected by both viruses. At 16 hpi, progeny viruses were collected and subsequently used for both immunoblotting and further passaging. The rA59 Nsp15-HA was rapidly outcompeted by rA59 WT , as nsp15-HA signal was greatly diminished by passages 2 and 3 for 1:1 and 3:1 infection ratios, respectively (Fig. 3B). Taken together, these data indicated that while rA59 Nsp15-HA replicates normally, the HA insertion significantly hampers its capacity to selectively package gRNA over sgRNA, diminishing its competitive advantage. This suggests that while the in situ tag maintained a stem-loop within this region of nsp15, it slightly attenuated the virus by likely altering the primary sequence of the P/S or the internal secondary structure of the RNA in this region or, possibly, a function of nsp15.
nsp15 colocalizes with RTC members nsp8 and nsp12. Previously, it was shown that nsp15 localizes with replicating viral RNA (10), potentially to replication compartments and with RTC members. To confirm that nsp15 was in fact localized to RTCs and to expand upon these results, we investigated the colocalization of nsp15 with nsp8 and nsp12 throughout infection. nsp8 and nsp12, the proposed viral primase and RdRp, respectively, are both established members of the replication complex and often used as markers for RTC localization (8,9). Following infection with either rA59 Nsp15-HA or rA59 WT , 17Cl-1 cells were fixed at indicated time points throughout infection and then costained with anti-HA and anti-nsp8 or anti-nsp12 antibody. nsp15 strongly colocalized with both nsp8 and nsp12 in rA59 Nsp15-HA -infected cells throughout infection (Fig. 4). The average Pearson's correlation coefficients (PCCs) ranged from 0.84 to 0.71 for nsp8 and 0.78 to 0.53 for nsp12. There was a small decrease in PCCs for both nsp8 and nsp12 at later time points; however, the PCCs remained above 0.5 throughout infection, which is considered strong colocalization. This colocalization was present in both syncytia and individual infected cells (Fig. 4A and B). These data demonstrated that nsp15 strongly colocalized with two proteins associated with the CoV RTC by confocal microscopy. nsp15 does not localize to sites of assembly. These results demonstrated strong colocalization of nsp15 with RTCs but did not address whether nsp15 also localized to sites of virus assembly. To this end, we costained rA59 WT -infected cells with anti-nsp15 and anti-M protein antibodies. M protein is localized to the endoplasmic reticulum-Golgi intermediate compartment (ERGIC) and is a marker for sites of assembly (40). To more formally confirm a lack of colocalization, we used rabbit anti-nsp15 to enable the use of a mouse monoclonal antibody (MAb) to M protein. M protein displayed tight punctate staining, most likely in the ERGIC compartment, at 6 and 8 hpi (Fig. 5B, top and middle). At 12 hpi, most cells were syncytial, and M puncta were scattered throughout the cytoplasm and potentially the cell membrane, most likely due to Golgi fragmentation (Fig. 5B, bottom) (41). At all stages of infection, nsp15 did not colocalize  with M protein (Fig. 5B), resulting in low PCCs of Ͻ0.3 throughout infection (Fig. 5C). Similar results were obtained when cells were infected with rA59 Nsp15-HA in lieu of rA59 WT (data not shown). These results indicated that nsp15 was not localized to sites of assembly during infection.
nsp15 interacts with CoV RTC-associated proteins. Due to the strong colocalization of nsp15 with nsp8 and nsp12, we next investigated whether nsp15 physically interacted with these proteins. 17Cl-1 cells were infected with rA59 Nsp15-HA or rA59 WT , and lysates were incubated with HA or nsp8 antibodies to identify interacting partners. Using the anti-HA antibody, no proteins were immunoprecipitated in rA59 WT -infected cells, while nsp8, nsp12, and nsp15-HA coprecipitated in rA59 Nsp15-HA -infected cells (Fig. 6A). To confirm these interactions, we next examined whether anti-nsp8 antibody would coprecipitate nsp15-HA. We found that nsp12 coprecipitated with nsp8 in both rA59 Nsp15-HA -and rA59 WT -infected cells, in agreement with previous results (Fig. 6B) (9). Furthermore, we found that nsp15-HA protein was immunoprecipitated with nsp8 nsp15 Interacts with CoV RTC-Associated Proteins ® antibody only in rA59 Nsp15-HA -infected cells (Fig. 6B). In these experiments, no precipitation of the irrelevant control protein GAPDH (glyceraldehyde-3-phosphate dehydrogenase) was observed, and no viral proteins were immunoprecipitated with normal rabbit serum (Fig. 6B). Taken together, these data demonstrated that nsp15 interacted with RTC proteins during MHV infection.
Since nsp8, nsp12, and nsp15 are all RNA binding or processing enzymes, we next investigated the dependence of these interactions on RNA binding. To accomplish this, we added a general RNA/DNA nuclease (Pierce Universal Nuclease), which we showed was active in our immunoprecipitation (IP) buffer (Fig. 6C), to cleave all unprotected RNA. We found no difference in the interactions between nsp8 and nsp15-HA when nuclease was present (Fig. 6D), demonstrating that these interactions were not dependent on RNA intermediates.

DISCUSSION
While nsp15 is completely conserved among CoVs and is known to have endoribonuclease activity in vitro, the role of this protein during infection is still largely unknown. nsp15 is likely to have a role in enhancing virus replication since mutation of catalytic amino acids resulted in reduced accumulation of viral RNA (29). Further, it was previously shown that nsp15 localized with replicating RNA (10), but its localization and interactions with other viral proteins were previously unknown. In this work, utilizing our in situ HA tag, we found strong colocalization of nsp15 with nsp8 and nsp12 throughout infection (Fig. 4). Further, we found only low levels of staining outside the RTC and no colocalization with M protein (Fig. 5). These data indicate that nsp15 localized to replication compartments with little or no detectable localization to other cellular or viral compartments. We found that nsp15 interacted with two markers of the RTC, nsp8 and nsp12, both with and without nuclease treatment, indicating that these interactions occur without the involvement of an RNA intermediate (Fig. 6). Our methods cannot distinguish direct from indirect interactions, so it remains possible that a viral or host protein mediates these binding events. While these interactions were detected in MHV-infected cells, a previous yeast two-hybrid (Y2H) screen along with a glutathione S-transferase (GST) pulldown identified potential binding partners of SARS-CoV nsp15, including nsp8 and nsp12 (15), suggesting these interactions are conserved among CoVs. These data collectively suggest that nsp15 is required for optimal viral production. However, further work will be required to determine if this occurs by directly or indirectly enhancing CoV RNA synthesis, processing RNA intermediates, or countering host cell defenses.
Incubation with anti-HA or anti-nsp8 antibody only coprecipitated a small fraction of the respective binding partner (Fig. 6). We hypothesize that this could be due to one or a combination of the following possibilities: (i) only a small percentage of viral nsps are engaged in RNA replication at any point in time as reported for hepatitis C virus (42), (ii) nsp15 has multiple functions, only one of which requires interaction with other nsps, or (iii) nsp15 only transiently interacts with the RTC before completing its function and subsequently dissociates from the complex. Previous data have demonstrated a link between nsp15's oligomeric state and RNA binding and cleavage (24)(25)(26). Oligomer formation was required for binding and cleavage in vitro, and cycling of nsp15 between a hexamer and a monomer was hypothesized to be required for RNA cleavage (25). Thus, nsp15 could be available for binding to RTC components intermittently.
In the present study, both the sequence and structure of the P/S stem were altered by our in situ tag. While viral replication was unaffected, viral fitness was modestly decreased compared to wild-type virus (Fig. 3B). This decrease in viral fitness was most likely due to a significant decrease in selective gRNA packaging in rA59 Nsp15-HA , as a similar phenotype was described with a P/S deletion virus (36). In rA59 Nsp15-HA , we found a significant Ͼ25-fold (Fig. 3A) increase in sgRNA packaging. This decrease in selective packaging may reflect requirements for the primary sequence of the P/S stem, proper secondary structure, or a combination of both, but at this point, we cannot distinguish between these possibilities. It has been hypothesized that internal conserved secondary structures, e.g., 2-bp bulges, play a functional role in selective packaging (38). From our data, we hypothesize that the pentaloop and the remaining 2-bp bulge are not sufficient for efficient selective packaging. However, it remains possible that they are necessary for some degree of selective packaging, perhaps mediating interactions with M protein (34). This may explain why the HA tag sequence was stable, whereas the FLAG tag sequence was quickly lost following passaging of the virus ( Fig. 3B; data not shown). Further work is required to determine what factors within the P/S are important for selective packaging of genomic RNA.
In conclusion, we demonstrated that the use of internal tags is tolerated by the MHV polyprotein and suggest that this approach may be generally useful for defining CoV RTC components and for identifying other interactions within the RTC. To create the in situ tag, we first identified a site with favorable characteristics for the insertion of a tag. These characteristics include (i) lack of conservation in other ␤-CoVs, suggesting that the site would tolerate mutation, (ii) being nonessential for replication, and (iii) surface exposure. The MHV P/S located within nsp15 met all of these criteria. We found that our rA59 Nsp15-HA in situ tag had little to no effect on viral RNA transcription/replication, viral protein expression, or virus production (Fig. 2). Using criteria similar to those presented here, we believe other CoV proteins could be internally tagged and studied during infection from native expression. The creation of internal tags and use of their corresponding monoclonal antibody (MAb) open new applications for the study of CoV replication, most notably the identification of novel in vivo RTC interactions.
Generation of recombinant strain rA59 Nsp15-HA . An epitope derived from the influenza A virus hemagglutinin (HA), YPYDVPDYA, was introduced into the F plasmid of the A59 in vitro ligation system (39). The F plasmid was linearized, and homologous arms, with the HA substitution, were added by PCR using the following primers: 5= GAGCCCACAAGGTAATCCGGGTGGTTGCGTAATCAGGAACGTCGTAAGGGT AGAAGCTCTAGCGCGTGGCAC 3= and 5= GATTACCTTGTGGGCTCCGGTAATGTGCGTAATCAGGAACGTCGTA nsp15 Interacts with CoV RTC-Associated Proteins ® AGGGTAGAAGATAACATCGTCACCGT 3=. The plasmid was recombined using In-Fusion (Clontech) according to the manufacturer's protocol and screened for the HA substitution. MHV-A59 fragments were digested and ligated as previously reported (46). Transcripts were generated using mMessage Machine T7. RNAs were then electroporated (Bio-Rad GenePulser Xcell) into 6 ϫ 10 7 BHK-R cells seeded into a T75 flask. Flasks were monitored for cytopathic effect (CPE [24 to 48 h]), and viruses were collected by freeze-thawing at peak CPE. After thawing, cell debris was cleared by centrifugation at 1,500 ϫ g, and supernatants were then aliquoted and termed passage 0 (P0). To create working stocks, rA59 Nsp15-HA and rA59 WT P0 viruses were passaged on 17Cl-1 cells and collected at maximum CPE.
Virus infection. Recombinant versions of the A59 strain of MHV (termed rA59 herein) were used in all experiments. The cDNA clone used to develop recombinant rA59 Nsp15-HA and rA59 WT has been described previously (GenBank accession no. AY910861) (39). Unless otherwise stated, 17Cl-1 cells were infected with the indicated viruses at an MOI of 0.1 for 30 min at 37°C. Viruses from supernatants and cells were combined prior to determination of viral titers. Viral titers. Virus titers were determined on HeLa-MVR cells as previously described with some modifications (47). HeLa-MVR cells under agarose overlays were fixed with 3.7% formaldehyde at 16 to 24 hpi and stained with 0.1% crystal violet.
Passaging and competition assay. The rA59 Nsp15-HA P1 viruses were passaged on 17Cl-1 cells at an initial MOI of 0.1. Each passage was collected at maximum CPE. For each subsequent passage, an estimated MOI of 0.1 was used by infecting 17Cl-1 cells with 5 l of infected supernatants. Following the fifth passage, stocks of each virus were used to infect 17Cl-1 cells and collected for immunoblotting as described below.
RNA analysis. Subconfluent monolayers of 17Cl-1 cells were infected at an MOI of 5 and collected by Trizol (Thermo Fisher Scientific) at the indicated times. RNA was isolated according to the manufacturer's instructions and transcribed into cDNA using Moloney murine leukemia virus reverse transcriptase (MMLV RT) (Thermo Fischer Scientific). Viral gRNA, sgRNA, and cellular hypoxanthine phosphoribosyltransferase (HPRT) RNA levels were measured by quantitative PCR (qPCR) with the indicated primers (see Table S1 in the supplemental material) using an Applied Biosystems 7300 real-time PCR system (Applied Biosystems). The levels of gRNA and sgRNA were normalized to HPRT by the following threshold cycle (C T ) equation: ΔC T ϭ C T of gene of interest Ϫ C T of HPRT. All results are shown as a ratio to HPRT calculated as 2 ϪΔCT .
Selective viral RNA packaging. Cell supernatants were collected at 16 hpi from 1.5 ϫ 10 7 17Cl-1 cells infected with rA59 WT or rA59 Nsp15-HA at an MOI of 0.1. Supernatants were centrifuged at 1,000 rpm to remove cellular debris and then passed through a 45-m-pore filter. Viruses were pelleted through a 30% sucrose cushion for 4 h at 27,000 rpm using a Beckman ultracentrifuge and an SW32 Ti rotor (Beckman, Indianapolis, IN). Viral pellets were resuspended in Dulbecco's modified Eagle's medium (DMEM), and viral RNA was isolated using Trizol. The ratio of sgRNA to gRNA was measured using reverse transcription-quantitative PCR (RT-qPCR).
Confocal microscopy. Subconfluent monolayers of 17Cl-1 cells on glass coverslips were infected at an MOI of 1 with either rA59 WT or rA59 Nsp15-HA virus. At the indicated times, monolayers were fixed with 4% paraformaldehyde. Cells were then permeabilized with 0.1% Triton X-100 in phosphate-buffered saline (PBS) before blocking with 1% goat serum. Following blocking, cells were stained with indicated primary antibodies for 3 h at room temperature using the following concentrations: anti-HA, 1:500 (clone 16B12 [BioLegend]); anti-nsp8, 1:1,000 (8); anti-nsp12, 1:500 (9); anti-M (5B11.5), 1:10,000 (48); and anti-nsp15 (D23 [a generous gift from Susan Baker, Loyola University, Chicago, IL]), 1:500 (10,49). Following washing with blocking buffer, fluorophore-conjugated secondary antibody was applied. Goat anti-rabbit 488 and goat anti-mouse 568 (Thermo Fisher Scientific) were used 1:500 in blocking buffer for 1 h at room temperature. Cells were washed before mounting with Vectashield antifade reagent (Vector Laboratories). In most cases, 4 to 5 images, with an average of~10 cells per image for each condition and for each experiment, were taken using a Leica STED Sp8 scanning confocal microscope. Images were analyzed using LAS X software. Pearson's correlation coefficients (PCCs) were obtained by analyzing individual cells from at least four images from two independent experiments using the CoLoc2 plugin for FIJI (50). Background was subtracted for data analysis as required.
Immunoprecipitation. 17Cl-1 cells were plated on 10-cm dishes 2 days prior to infection. For each condition, 4.5 ϫ 10 7 17Cl-1 cells were infected with either rA59 WT or rA59 Nsp15-HA at an MOI of 0.1. When plates reached maximum CPE (16 to 20 hpi), cells were collected and pelleted by low-speed centrifugation prior to lysis. Cell pellets were lysed with IP buffer (0.5% NP-40, 150 mM NaCl, 5% glycerol, and 50 mM Tris [pH 8.0], with or without Pierce Universal Nuclease [Thermo Fischer Scientific]), as indicated for 1 h at 4°C. Next, nuclei were pelleted by centrifugation (16,000 ϫ g for 15 min at 4°C) and applied to protein G Dynabeads (Thermo Fischer Scientific) conjugated to the indicated antibodies according to the manufacturer's protocol. Cell lysate and antibody-bound beads were mixed overnight at 4°C. Beads were washed with PBS-Tween before elution with 2ϫ SDS sample buffer at 95°C for 5 min. Eluates and lysates were then subjected to immunoblotting as described above.
Statistics. Student's unpaired t test or Mann-Whitney U test was used to analyze differences in mean values between groups. All results are expressed as means Ϯ range or standard error of the mean. P values of Յ0.05 were considered statistically significant.