High-Resolution Functional Profiling of the ... - Journal of Virology

High-Resolution Functional Profiling of the Norovirus Genome Lucy Thorne, Dalan Bailey,* and Ian Goodfellow Calicivirus Research Group, Section of Virology, Department of Medicine, Imperial College London, London, United Kingdom

Human noroviruses (HuNoV) are a major cause of nonbacterial gastroenteritis worldwide, yet details of the life cycle and replication of HuNoV are relatively unknown due to the lack of an efficient cell culture system. Studies with murine norovirus (MNV), which can be propagated in permissive cells, have begun to probe different aspects of the norovirus life cycle; however, our understanding of the specific functions of the viral proteins lags far behind that of other RNA viruses. Genome-wide functional profiling by insertional mutagenesis can reveal protein domains essential for replication and can lead to generation of tagged viruses, which has not yet been achieved for noroviruses. Here, transposon-mediated insertional mutagenesis was used to create 5 libraries of mutagenized MNV infectious clones, each containing a 15-nucleotide sequence randomly inserted within a defined region of the genome. Infectious virus was recovered from each library and was subsequently passaged in cell culture to determine the effect of each insertion by insertion-specific fluorescent PCR profiling. Genome-wide profiling of over 2,000 insertions revealed essential protein domains and confirmed known functional motifs. As validation, several insertion sites were introduced into a wild-type clone, successfully allowing the recovery of infectious virus. Screening of a number of reporter proteins and epitope tags led to the generation of the first infectious epitope-tagged noroviruses carrying the FLAG epitope tag in either NS4 or VP2. Subsequent work confirmed that epitope-tagged fully infectious noroviruses may be of use in the dissection of the molecular interactions that occur within the viral replication complex.

H

uman norovirus (HuNoV) is the leading cause of nonbacterial gastroenteritis in developed countries, with initial reports also indicating high incidence and mortality in developing countries (47). Due to the low infectious dose, multiple transmission routes, and high stability of HuNoV in the environment (41, 72), outbreaks typically occur in places with close living conditions, such as hospitals. Infections in hospitals alone cost the United Kingdom National Health Service an estimated £115 million and result in 47,000 lost bed days per year, compromising hospital services (39, 53). The majority of HuNoV infections are acute and self-resolving; however, links with conditions such as inflammatory bowel disease (36), infantile seizures (18), and neonatal necrotizing enterocolitis (65, 74) have been established. There are also increasing reports of persistent infections in the elderly and the immunocompromised, such as transplant recipients and some cancer patients (2, 13, 20, 40, 55, 56). Details of the life cycle and replication of HuNoV are relatively unknown due to the lack of an efficient cell culture system. The discovery of murine norovirus (MNV) in 2003 has since facilitated studies on norovirus replication. Unlike HuNoV, MNV can be propagated in culture, displaying a tropism for macrophage and dendritic cells (75), and can be manipulated by reverse genetics (17, 76). The first strain of MNV identified (MNV-1) was found to be the cause of a lethal infection in immunocompromised mice but can be cleared from immunocompetent hosts following a subclinical infection (33). Further studies indicate that the majority of other MNV strains cause persistent infections in immunocompetent hosts (28, 73). MNV has also been isolated from wild mice, confirming its widespread distribution (59). Studies with MNV have begun to probe various aspects of the norovirus replication cycle, including entry mechanisms and attachment factors (24, 48, 49, 67), viral protein function (8, 19, 30, 43, 64, 66), determinants of virulence (7), and host immune responses (14, 15, 42); however, our understanding of the molecular mechanisms of norovi-

November 2012 Volume 86 Number 21

rus genome translation and replication lags far behind that for other RNA viruses. MNV is a positive-sense single-stranded RNA virus with a genome of approximately 7.4 kb covalently attached to a viral genome-linked protein (VPg) at the 5= end and polyadenylated at the 3= end (33). The 5= and 3= untranslated regions (UTRs) are extremely short, 5 and 78 nucleotides (nt), respectively, for MNV-1 (33). The 5= and 3= extremities of the MNV genome, including the coding regions, contain evolutionarily conserved RNA structures important for viral replication and pathogenesis (6, 58). The MNV genome is organized into four open reading frames (ORFs). ORF1 is translated as a large polyprotein, which is co- and posttranslationally cleaved by the virus-encoded protease (NS6) to release six mature nonstructural proteins, including NS6 itself (63). The other nonstructural proteins include the viral RNA-dependent RNA polymerase (RdRp) (NS7), VPg (NS5), the putative nucleoside triphosphatase (NTPase)/RNA helicase (NS3) and NS1-2 and NS4, which have both been implicated in replication complex formation (30). ORF2 and ORF3 are translated from a subgenomic RNA and encode the major and minor capsid proteins, VP1 and VP2, respectively. ORF4 is translated from an alternative reading frame within ORF2 and encodes the recently

Received 20 February 2012 Accepted 12 August 2012 Published ahead of print 22 August 2012 Address correspondence to Ian Goodfellow, [email protected], or Lucy Thorne, [email protected]. * Present address: Dalan Bailey, Institute for Animal Health, Surrey, United Kingdom. Supplemental material for this article may be found at http://jvi.asm.org/. Copyright © 2012, American Society for Microbiology. All Rights Reserved. doi:10.1128/JVI.00439-12 The authors have paid a fee to allow immediate free access to this article.

Journal of Virology

p. 11441–11456

jvi.asm.org

11441

Thorne et al.

identified virulence factor 1 (VF1) protein, a novel innate immune regulator (43). The rational development of antivirals for noroviruses is dependent on the elucidation of the specific viral protein functions as well as their viral and cellular interaction partners. Transposon (Tn)-mediated insertional mutagenesis has been widely used to identify the essential functional domains of many viral proteins (23, 35, 37, 54, 69) and cis-acting elements (3, 4) and more recently for comprehensive functional analysis of entire viral genomes, such as the hepatitis C virus genome (3, 9, 44). In comparison to targeted mutagenesis, Tn mutagenesis allows the structure and function of viral proteins to be randomly probed. When combined with insertion profiling techniques, this approach can be used to simultaneously determine the effect of a large number of insertions on viral replication (3, 4, 9, 44). Identification of tolerated insertion sites by Tn mutagenesis also provides the opportunity to engineer exogenous nucleotide sequences into a viral genome. These have included sequences encoding reporter proteins or small peptide tags to facilitate further characterization of the viral proteins and their interaction partners, as well as to enable imaging of viral replication (1, 5, 45, 70, 71). This strategy has been particularly successful for labeling small positive-strand RNA viruses (69, 70), for which traditional approaches, such as making insertions in the viral UTRs (51) or targeted viral protein-reporter fusions (11), are challenging. Such approaches are further complicated for MNV due to the compact nature of the genome, the short length of the UTRs, the presence of cis-acting RNA within the coding region, and the little structural information available for the viral proteins. Consequently, to date there are no reports of studies using infectious reporter- or epitope-tagged noroviruses. In this study, to determine the viral protein domains that are required for norovirus replication, Tn-mediated insertional mutagenesis was coupled with insertion-specific fluorescent PCR profiling to probe the entire MNV-1 genome for sites that tolerate 15-nt insertions. Insertion mapping and functional analysis of each viral protein revealed new functional domains and allowed the importance of known functional motifs to be probed. The introduction of a selection of the identified insertion sites into the MNV-1 wild-type (WT) cDNA clone, followed by recovery by reverse genetics, validated the profiling approach. Screening of a number of reporter proteins and epitope tags led to the generation of the first infectious epitope-tagged noroviruses carrying the FLAG epitope tag in either NS4 or VP2. MATERIALS AND METHODS Cells and plasmid constructs. Murine macrophage cells (RAW264.7) were maintained in Dulbecco’s modified Eagle’s medium (DMEM; Gibco) containing 10% fetal calf serum (FCS), penicillin (100 SI units/ ml), streptomycin (100 ␮g/ml), and 10 mM HEPES (pH 7.6) at 37°C with 10% CO2. The murine microglial BV-2 cell line (10) was provided by Jennifer Pocock (University College London). BV-2 cells were maintained in DMEM supplemented with 10% FCS (Biosera), 2 mM L-glutamine and 0.075% sodium bicarbonate (both from Gibco), and the antibiotics penicillin and streptomycin, as described above. BHK cells engineered to express T7 RNA polymerase (BSR-T7 cells, obtained from Karl-Klaus Conzelmann, Ludwig Maximilian University, Munich, Germany) were maintained in DMEM containing 10% FCS, penicillin (100 SI units/ml), streptomycin (100 ␮g/ml), and 0.5 mg/ml G418. The cDNA clone pT7: MNV 3= RZ, containing the WT MNV-1 genome under the control of a truncated T7 polymerase promoter, and a modified version that contains

11442

jvi.asm.org

a frameshift in the NS7 region of ORF1 (pT7:MNV 3= RZ F/S) were previously constructed (48). Reverse genetics and virus titration. Virus was rescued from cDNA clones (as indicated in the text) by using the reverse genetics system based on recombinant fowlpox virus expressing T7 RNA polymerase, as previously described (17). Briefly, 1 ␮g of each cDNA clone was transfected, using Lipofectamine 2000 transfection reagent (Invitrogen), into BSR-T7 cells previously infected with recombinant fowlpox virus expressing T7 RNA polymerase at a multiplicity of infection (MOI) of approximately 0.5 PFU per cell. Where indicated, the RNA-mediated reverse genetics system was used as recently described (76). In short, RNA was produced by in vitro transcription in a reaction mixture containing 200 mM HEPES, pH 7.5, 32 mM magnesium acetate, 40 mM dithiothreitol, 2 mM spermidine, 7.5 mM each nucleoside triphosphate (NTP), 60 ng of linearized DNA template, and 50 ␮g/ml of T7 RNA polymerase. Reaction mixtures were incubated at 37°C for 5 to 7 h, and then RNA was precipitated with lithium chloride and resuspended in RNA storage solution (Ambion). Posttranscriptional enzymatic capping was performed using a ScriptCap system from Epicentre, according to the manufacturer’s instructions. After the reaction, RNA was again precipitated in lithium chloride and stored in RNA storage solution (Ambion). One microgram of RNA was transfected into BSR-T7 cells using Lipofectamine 2000 (Invitrogen). At 48 h posttransfection of cDNA or RNA, cells were freeze-thawed at ⫺80°C to release virus particles, and 50% tissue culture infectious dose (TCID50) titrations were performed on RAW264.7 cells using 10-fold serial dilutions. The viral TCID50/ml was determined by scoring for signs of cytopathic effect (CPE) at 5 days postinfection by visual inspection. Transposon mutagenesis. The HyperMuA transposase enzyme (Epicentre Biotechnologies) was used as detailed in the manufacturer’s instructions for transposition of a kanamycin-resistant Tn (Mutation Generation System kit; Finnzymes) into the MNV-1 infectious clone pT7: MNV 3= RZ. The transposition reaction was performed using the recommended ratio of transposon to pT7:MNV 3= RZ. The ratio was selected to promote a single transposition event per copy of the infectious clone, but the possibility of a double insertion cannot be entirely excluded. Mutagenized clones were transformed into XL10 Gold ultracompetent cells (Stratagene) and selected through resistance to kanamycin and carbenicillin, to create the original Tn library. The resulting library contained an estimated 225,000 clones, each containing a single randomly inserted kanamycin-resistant Tn. Ten micrograms of this library was subsequently digested with one of five different pairs of restriction enzymes (New England BioLabs) to release a defined region of the genome (Fig. 1). Nonmutagenized pT7:MNV 3= RZ was digested with the same enzyme pairs and ligated to each mutagenized insert. Each library was screened by digestion with the same pair of enzymes to confirm that all clones had a Tn in the appropriate fragment. To create the 15-nt insertion libraries, the five Tn-mutagenized libraries were extensively digested with NotI and the Tn was removed, before self-ligation. The DNA libraries are available upon request. RNA transfections were performed for each 15-nt insertion library (see above for details) in triplicate, and the titer of the recovered virus was determined from the TCID50. Insertion profiling. For tissue culture selection of tolerated insertions, virus recovered from each 15-nt insertion library was passaged 3 times in RAW264.7 cells at an MOI of 0.01 TCID50 per cell for the first passage and 0.1 TCID50 per cell thereafter for 16 h with two independent biological replicates. The titer of the virus was determined between each sequential passage, and total RNA was extracted using a GenElute system (Sigma). Reverse transcription (RT) of the extracted RNAs, the in vitro-transcribed library input RNA, and wild-type RNA was performed using SuperScript III reverse transcriptase (Invitrogen). The cDNAs were then used as the templates for PCR amplification of overlapping fragments spanning the mutagenized region of each library, using KOD Hot Start DNA polymerase (Novagen). In the case of the VP1/VP2/3= UTR library, the original DNA library was also PCR amplified and profiled to control for a loss of library diversity during in vitro transcription. All fragments were

Journal of Virology

Murine Norovirus Functional Profiling

FIG 1 (A) Strategy for Tn mutagenesis of the MNV-1 genome. A HyperMu transposase enzyme was used to insert one Tn per pT7:MNV 3= RZ infectious clone to generate a Tn-mutagenized library. This was subcloned to generate five libraries containing Tns only in a well-defined region of the genome (gray shading).*, SpeI cleaves at a position within the plasmid vector; other numbers indicate genome positions. Removal of the Tns from each library leaves residual 15-nt insertions (black shading). Dashed lines, vector of the pT7:MNV 3= RZ infectious clone. (B) Composition of the 15-nt insertion. Numbers in italics represent the 5-nt target site for the original transposition, 10 nt that originated from the transposon are shown and contain a NotI restriction site highlighted in bold, and the 5 underlined numbers represent target nucleotides duplicated during transposition. (C) Recovery of infectious virus from 15-nt insertion libraries. BSR-T7 cells were transfected with in vitro-transcribed, enzymatically capped RNA produced from each cDNA 15-nt insertion library. Virus was harvested at 48 h posttransfection, and the yield of recovered infectious virus was determined in RAW264.7 cells. The limit of detection is 50 TCID50/ml. Transfections were performed in triplicate; error bars represent the standard error of the mean. Significance was tested using one-way analysis of variance and Dunnett’s multiple-comparison posttest to compare each insertion virus to the WT. FS, frameshift clone; ***, P ⬍ 0.001; *, P ⬍ 0.5.

approximately 400 bp in size and overlapped the following fragment by about 100 bp. The pT7:MNV 3=RZ clone was also used as a template to generate corresponding PCR products to later control for nonspecific products in the second round of PCR and detection of the insertions. Each RT-PCR or PCR product was gel purified, and 50 ng was used as a template in a second round of PCR using the same virus-specific forward primer (from the first round of PCR) and a NotI miniprimer (5=TGCGG CCGCA3=), labeled at the 5= end with the fluorescent dye 6-carboxyfluorescein. One microliter of each fluorescently labeled product was then mixed with 8.6 ␮l HiDi formamide (Applied Biosystems) and 0.4 ␮l car-

November 2012 Volume 86 Number 21

boxy-X-rhodamine 400HD size markers (Applied Biosystems) and analyzed in three technical replicates using an ABI 3730xl DNA analyzer at the Genomic Core Facility (Hammersmith Campus, Imperial College London). Data analysis and interpretation. The data generated by the DNA analyzer were processed using Peak Scanner software (Applied Biosystems) and modified amplified fragment length polymorphism analysis settings. The minimum fluorescence intensity of detection was set according to the background peaks observed for WT DNA across the same mutagenized region of each library, which presumably occurred due to GC-

jvi.asm.org 11443

Thorne et al.

rich regions in the WT nonmutagenized genome. These were typically 600 arbitrary units for all libraries except the VP1/VP2/3= UTR insertion library, for which the background was set at 200 arbitrary units, and the NS1-2 library, for which it was set at 400 arbitrary units. Any peaks that were greater than this limit in the WT were later excluded from the library data sets to exclude the possibility of false positives. The data were visualized as an electropherogram, with each peak representing a 15-nt insertion. Data regarding the fragment lengths (in base pairs) and corresponding peak heights in each fragment were exported into the Microsoft Excel program for further analysis. Fragments smaller than 65 bp in size typically corresponded to primer noise and were excluded from the data set, as this region would be covered by the overlapping nature of the PCR amplicons. The data from triplicate analyses of each fragment were compiled, and the peak heights were averaged for insertions appearing in all three. Insertion profiles were also compiled for the duplicate passages. To calculate the position of each insertion within the genome, the size of the peak was added to the genome position of the virus-specific forward primer, and 10 nt were subtracted to allow for the size of the 3= NotI miniprimer. The final whole-genome insertion map was compiled from each passage of each of the libraries in GraphPad Prism software. The Chimera modeling software (UCSF) was used to map the positions of insertions onto the MNV NS7 structure (Protein Data Bank [PDB] accession number 3NAH) and the insertions in VP1 onto the structure of the MNV protruding (P) domain (PDB accession number 3lq6B). Sequence alignments were performed with Align X software, and protein sequence structure predictions were performed using the TMPred, PHDht, and PSIPRED programs, all available online. Isolation and cloning of insertion mutants. To isolate individual insertion mutants, limiting dilutions of the virus recovered from each library were performed on RAW264.7 cells seeded in a 96-well plate. RNA was extracted from wells with cells showing a cytopathic effect using TRIzol. The mutagenized region was amplified by RT-PCR using a one-step Access Quick RT-PCR system (Promega) and primers to span the relevant mutagenized region (details are available upon request). PCR products were then digested with NotI to confirm the presence of the 15-nt insertion, and the sequences of the insertion sites were verified to determine the genomic position. The sequences of selected insertion sites from the profiling were predicted using Vector NTI software (Invitrogen). Insertion sites were introduced into the MNV-1 cDNA clone pT7:MNV 3=RZ using overlap PCR mutagenesis (primer sequences are available upon request) by KOD Hot Start polymerase (Novagen), and virus was recovered using the aforementioned DNA reverse genetics system. Western blotting. Samples were prepared for Western blotting by harvesting the cells directly into radioimmunoprecipitation assay buffer (50 mM Tris-HCl [pH 8.0], 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.1% SDS). The cell lysates were resolved by SDS-PAGE and transferred to a nitrocellulose membrane (polyvinylidene difluoride; Immobilon). The membranes were blocked (5% milk with 0.1% Tween 20 in phosphate-buffered saline [PBS]) and probed for the FLAG tag, NS7, and cellular GADPH (glyceraldehyde-3-phosphate dehydrogenase) using anti-FLAG– horseradish peroxidase (HRPO; Sigma), rabbit anti-MNV NS7 serum (noncommercial antibody), and mouse anti-GAPDH (Applied Biosystems) antibodies, respectively. HRPO-conjugated primary and secondary goat anti-rabbit or anti-mouse antibodies were detected by enhanced chemiluminescence (ECL; Amersham, GE Healthcare). NS1-2 was detected using mouse immune sera provided by Herbert Virgin and Larissa Thackray (Washington University, St. Louis, MO). Confocal microscopy. BV-2 cells were seeded onto glass coverslips at a density of 5 ⫻ 105 and infected on the same day at an MOI of 1 TCID50 per cell. At 12 h postinfection (hpi), cells were fixed using 4% paraformaldehyde (PFA) in PBS, permeabilized with 0.2% Triton X-100 in 4% PFA in PBS, and quenched with 200 mM glycine in PBS. PBS washes were performed between each step. Cells were blocked in 10% FCS in PBS and stained with monoclonal anti-FLAG M2 (Sigma) and rabbit anti-NS7 serum, followed by donkey anti-mouse and anti-rabbit antibodies (Jack-

11444

jvi.asm.org

son Laboratories) conjugated with Alexa Fluor 546 and Alexa Fluor 488, respectively, using a Molecular Probes conjugation kit (provided by Micheal Hollinshead, Imperial College London). Cells were then mounted onto a glass slide with Mowiol medium containing DAPI (4=,6-diamidino-2-phenylindole) stain. Images were captured using a Zeiss 510 Meta laser confocal microscope. Immunoprecipitations. RAW264.7 cells were infected with passage 1 sequence-verified stocks of NS4-FLAG, VP2-FLAG, and WT MNV-1 at an MOI of 0.01 TCID50/ml. At 36 hpi, cell lysates were harvested (lysis buffer 50 mM Tris-HCl, pH 7.4, with 150 mM NaCl, 1 mM EDTA, and 1% Triton X-100), and a proportion was kept aside as the total cell input. Immunoprecipitations were then performed using anti-FLAG M2 affinity gel (Sigma-Aldrich), as per the manufacturer’s instructions. Elution into SDS-PAGE sample buffer was performed, and input lysates and eluates were analyze by Western blotting (see above).

RESULTS

Transposon-based insertional mutagenesis of the murine norovirus genome. In order to probe the MNV-1 genome for sites that can tolerate the insertion of an exogenous sequence, a strategy based on Tn mutagenesis was designed to create five libraries of mutagenized MNV-1 cDNA clones, between them covering the entire genome (Fig. 1A). To create the original mutagenized library, from which the five mutagenized libraries were subcloned, the full-length MNV-1 infectious clone pT7:MNV 3=RZ (17) was subjected to an in vitro transposition reaction catalyzed by HyperMuA transposase (Fig. 1A). To exclude clones containing a Tn in the backbone of the vector, which would later produce WT virus likely to dominate in a population of insertion mutants, five adjacent regions of the original library were subcloned into the nonmutagenized pT7:MNV 3=RZ infectious clone. For this purpose, five different pairs of restriction enzymes were used to divide the genome into regions spanning NS1-2, NS3/4, NS5/6/7, NS7/VP1/ VF1, and VP1/VP2/3= UTR, without overlaps (Fig. 1A). Restriction digestion using the original enzyme pairs confirmed that Tns were present only in the appropriate region for each library (data not shown). The Tns were then removed from each library by NotI digestion, to leave behind a 15-nt insertion, generating five 15-nt insertion libraries spanning the entire MNV-1 genome. The 15-nt insertion left behind after Tn removal was composed of 10 nt derived from the Tn containing a GC-rich NotI site (not otherwise found in the MNV-1 genome) and 5 nt that originated from target DNA adjacent to the insertion site, which is duplicated during the original transposition event (Fig. 1B). Importantly for insertions in ORF1, all insertions result in an in-frame addition of 5 amino acids, with the amino acid sequence of each insertion varying on the basis of the reading frame in which it is made and the sequence of the duplicated target DNA (35). At least 10 clones from each 15-nt insertion library were screened by restriction digest for the 15-nt insertion NotI site, which revealed seemingly unbiased differential digestion patterns, suggesting that insertion sites were spread throughout each library region (data not shown). To recover virus from the 15-nt cDNA insertion libraries, in vitro-transcribed capped RNA was generated for direct transfection into BSR-T7 cells. The RNA reverse genetics system, recently developed in our lab (76), was employed over the DNA-based reverse genetics system due to its higher efficiency, which is preferable for recovering a complex library. Infectious virus was recovered from each of the 15-nt insertion libraries, although at significantly lower titers in comparison to those obtained with

Journal of Virology

Murine Norovirus Functional Profiling

FIG 2 Schematic diagram showing selection and detection of insertions in the 15-nt insertion library. RNA was produced from each 15-nt cDNA insertion library by in vitro transcription and enzymatic capping of the DNA library, and virus was recovered using the RNA-mediated reverse genetics system, in duplicate. For selection in tissue culture, the recovered virus was sequentially passed (in duplicate) three times in RAW264.7 cells at a low MOI. RNA was extracted after each passage and alongside the input RNA was used as a template for RT-PCR amplification of overlapping fragments covering the respective mutagenized region of each library. Each fragment was used as a template in a second round of PCR with a virus-specific primer and the insertion-specific NotI miniprimer, labeled with a fluorescent tag (black circles). Fluorescently labeled products were detected and analyzed by capillary electrophoresis.

RNA derived from the nonmutagenized WT cDNA clone, with the exception of the NS1-2 library (Fig. 1C). Passage of insertion libraries. To simultaneously identify and determine the effect of each insertion on viral replication, an insertion profiling method using insertion-specific fluorescent PCR was established for MNV (Fig. 2). The virus recovered from each 15-nt insertion library was first subjected to tissue culture selection by 3 serial passages in RAW264.7 cells, performed as two independent biological replicates. Each passage was performed at a low MOI to reduce the possibility of transencapsidation and complementation of defective RNAs. Virus and RNA were harvested from each passage at 16 hpi and used to identify the insertion sites present within the population. Multiple RT-PCRs were performed on the extracted RNA, alongside the unselected input RNA for each library, to generate overlapping fragments spanning the appropriate mutagenized region. The purified products were then used as a template for a second round of PCR using a fluorescently labeled insertion-specific primer (Fig. 2) and the same virus-specific forward primer. The resulting fluorescently labeled products were analyzed by capillary electrophoresis to determine the size of each peak corresponding to an insertion, from which the genome position could be calculated (see Materials and Methods for details of data handling). To ensure that library complexity was maintained during the in vitro transcription of the cDNA libraries (prior to recovery), the VP1/VP2/3= UTR cDNA 15-nt insertion library was also profiled and was found to be identical to that of the corresponding in vitro-transcribed RNA. Example elec-

November 2012 Volume 86 Number 21

tropherograms from a region of the VP1/VP2/3= UTR library are shown for the cDNA library, input RNA, and products from each of the three passages (Fig. 3). A list of insertions detected prior to selection and at each stage thereafter was compiled for each 15-nt insertion library to generate an insertion map covering the entire MNV-1 genome (with the exception of the first 55 nt) (Fig. 4; see Data Set S1 in the supplemental material). In the unselected input population, 2,049 insertions were detected, corresponding to approximately 1 insertion every 3 to 4 nt, although the distribution is not entirely even. After the first passage, the number of insertions detected was reduced to 918, and then 468 and 324 insertions were detected after the second and third passages, respectively, indicating the selection of stable and viable insertion mutants. The immediate loss of insertions during the first round of selection reveals which regions of the genome encode protein domains that must be intact for viral replication. Conversely, insertions that are tolerated for all three passages must not significantly compromise protein function or replication, and as such, this rationale can be used to narrow down regions that encode important functional protein domains across the entire genome. Mapping of insertion sites. (i) NS1-2. The NS1-2 protein is the only norovirus protein that lacks sequence homology with any other viral protein, therefore providing no clue as to its function (26). In a recent study, NS1-2 was found to be associated with the replication complex in MNV-1-infected cells, and it has been implicated in complex formation by recruiting cellular membranes,

jvi.asm.org 11445

Thorne et al.

FIG 3 Example electropherograms showing the detection and positioning of 15-nt insertions in the fragment covering nt 6546 to 6931 (ORF2) during selection in tissue culture. Peak size is depicted on the x axis, and fluorescence intensity is indicated on the y axis. The similar insertion profiles of the cDNA library and the RNA input suggest that diversity was not lost during in vitro transcription. Comparison of sequential passages reveals selection occurring within the insertion mutant populations. Insertions within the solid box were maintained, whereas the dashed box highlights a group of insertions that were rapidly lost. Peaks under 65 nt correspond to primer noise and were excluded from the analysis.

although a definitive function has not yet been established (30). Of all the nonstructural proteins, NS1-2 was found to have the highest percentage of tolerated insertions remaining after the three passages (22% of those detected in the input). This is in line with the observation that it is the least conserved of the norovirus nonstructural proteins. The insertions remaining after passage 3 were almost entirely located in the N-terminal half of the protein (residues 1 to 144), recently reported to account for the high sequence divergence of NS1-2 between norovirus genogroups, that at least in isolation behaves as an intrinsically disordered region (8). In particular, the insertions clustered between nucleotide positions 57 to 90, 154 to 185, and 350 to 404, corresponding to amino acid residues 18 to 29, 50 to 60, and 111 to 133, respectively. These tolerated clusters could also be probed for the potential insertion of other exogenous sequences. Nearly all insertions between the second and third clusters were lost at the first passage, suggesting

11446

jvi.asm.org

that this region encodes an important functional domain. It is also worth noting that the first 250 nt of the MNV genome contains predicted RNA secondary structures (58) that may interact with cellular proteins, which suggests that the tolerated insertions within the first two clusters do not compromise the RNA structures or any RNA-protein interactions essential for replication in cell culture. Two caspase 3 cleavage sites, 118DXXD121 and 128DXXD131, previously described for MNV, overlap the second cluster of tolerated insertions and are thought to be cleaved late in infection (63). Three insertions were tolerated for all three passages in the first site, two insertions were tolerated in the second site, and all of these disrupted the cleavage sites by introducing 5 amino acids between the aspartic acid residues. This indicates that cleavage of these sites is not essential for replication in permissive cells and, as previously suggested, may simply be a result of the induction of

Journal of Virology

Murine Norovirus Functional Profiling

FIG 4 (A) Genome-wide insertion profiling of the murine norovirus genome. Insertion sites detected in each passage (passage 1 [P1], passage 2 [P2], and passage 3 [P3]) for each library were compiled to generate a map of insertion sites across the entire genome, with the exception of the first 55 nt. A total of 2,089 insertions were detected in the input profile, with only 324 remaining after the three passages, indicating selection of viable mutants. Insertion sites found in the input and passage 3 populations at the end of NS4 (B) and the end of VP2 (C) are shown in greater detail. The insertions highlighted in red correspond to the position where a FLAG tag was later inserted in the NS4- and VP2-coding regions.

apoptosis at late stages of infection (63). However, it is interesting to speculate that as many MNV strains appear to have maintained these sites (63), perhaps the cleavage sites and the subsequent cleavage products play a role in vivo that is not apparent in cell culture. A third protease site in NS1-2 (100DXXD103) has been reported and is thought to be cleaved by an unknown cellular proteinase (63). Interestingly, all insertions within this site that disrupted the positioning of the aspartic acid residues relative to one another were lost at the first round of selection, suggesting that cleavage of this site is important for replication and that the cleavage products may have a specific function. The C-terminal half of NS1-2 exhibits a greater degree of sequence conservation between norovirus genogroups (8), and accordingly, only several insertions were detected in this region after the third passage, indicating that disruption of its structure or function must compromise viral fitness or replication. The largely

November 2012 Volume 86 Number 21

hydrophobic C terminus contains a predicted transmembrane (TM) domain spanning residues 266 to 318 (8). All except three insertions in this region were lost after the first two passages, suggesting that disruption of the putative transmembrane domain compromises viral fitness. The C-terminal end of the HuNoV NS1-2 equivalent protein (p48) (Table 1) is also largely hydrophobic and contains a predicted TM domain between residues 361 and 379 (21, 22). This TM domain is thought to potentially act as a membrane anchor and be involved in intracellular membrane rearrangements and Golgi apparatus disassembly (22). In contrast, the MNV NS1-2 localizes to the endoplasmic reticulum (ER) with no apparent Golgi association and promotes ER reorganization into distinct vesicular structures when transfected in isolation (30). It is therefore possible that the TM regions of NS1-2 and p48 mediate similar functions but do so in different organelles. Finally, as regions that did not tolerate any insertions were located within

jvi.asm.org 11447

Thorne et al.

TABLE 1 The murine norovirus proteins and the names of their HuNoV and FCV orthologs Ortholog MNV protein

HuNoV

FCV

NS1-2 NS3 NS4 NS5 (VPg) NS6 NS7 VP1 VP2 VF1

p48 p41 (NTPase) p22 VPg Pro Pol VP1 VP2 No equivalent

p32 p39 (NTPase) p30 VPg Pro-Pola Pro-Pol VP1 VP2 No equivalent

a

The FCV protease and polymerase exist as enzymatically active precursors.

both the C- and N-terminal halves, MNV NS1-2 may contain two separate functional domains. This is consistent with the recent report that NS1-2 appears to be comprised of two distinct halves with differential properties, which may afford the protein multiple functions (8). (ii) NS3. NS3 is the putative helicase, due to the presence of three canonical superfamily 3 helicase motifs (motifs A, B, and C); however, its role in replication has not yet been clearly demonstrated. In vitro characterization of the HuNoV NS3 (p41) (Table 1) demonstrated that the protein had NTP binding and hydrolysis activity, as conferred by canonical superfamily 3 helicase motifs A and B (50). All insertions that disrupted motif A [GXXGXGK(S /T)] were lost at the first passage, although an insertion that altered the identity of the last 2 residues (K/N and S/A, respectively) was tolerated for the first passage but subsequently lost, most likely due to a reduction in viral fitness. It is also possible, however, that the virus altered this insertion to restore the identity of those residues, meaning that it could no longer be detected. Only 1 insertion was detected in motif B (DD/E), which is involved in Mg2⫹ binding, and it appeared to be tolerated until passage 3. However, prediction of the insertion sequence revealed that it would not disrupt or alter the identity of the motif due to duplication of the target-site nucleotides. Together these data provide the first evidence that the NTPase activity is essential for MNV replication. The proposed helicase activity of NS3 is thought to be conferred by motif C, although HuNoV p41 failed to demonstrate unwinding during in vitro characterization, but it is possible that an extra cofactor is required. Motif C (S/TTN) is thought to confer the helicase activity and contains an invariant asparagine residue located downstream of motif B. While insertions in the preceding residues were found to tolerate insertions for the first two passages, an insertion that changed the asparagine to methionine was immediately lost from the input profile upon passage in cell culture. This indicates that this motif may also play an important role in replication; however, further studies are required to confirm this observation and determine the exact function of motif C. By passage 3, although there were not well-defined clusters of tolerated insertions (as observed for NS1-2), the regions between nt 1928 to 1933 and 2012 to 2020 appeared to be the most likely positions that could be probed for insertions of other exogenous sequences. (iii) NS4. Like NS1-2, NS4 has been associated with the replication complex in MNV-1-infected cells and is also thought to recruit cellular membranes to establish the complex (30); how-

11448

jvi.asm.org

ever, its precise role in the MNV life cycle has not been determined. Both the N- and C-terminal regions of NS4 tolerated insertions throughout all three passages (Fig. 4B) and represent the most likely regions which could be probed for the potential insertion of other exogenous sequences. The central domain, however, did not tolerate any insertions through even the first passage, and amino acids 64 to 129 appeared to be essential for replication. Accordingly, sequence comparison of four different MNV strains (strains 1 to 4) revealed 95% amino acid identity over these residues (data not shown). The central domain of the HuNoV NS4 (p22) (Table 1) contains a predicted transmembrane helix involving residues 112 to 127, which is thought to mediate membrane association (57). Sequence analysis using three different protein structure prediction tools (TMPred, PSIPRED, and PHDhtm) predicted a transmembrane helix in the central domain of MNV-1 NS4, with the most agreement for its position involving residues 110 to 121 (data not shown), which suggests that the central domain of NS4 may also mediate membrane association. The TM domain of HuNoV p22 is thought to be responsible for localization of the protein to the Golgi apparatus, resulting in its dispersion and disruption of protein trafficking in transfected cells (57). Similarly, the MNV NS4 protein has also been shown to localize to the Golgi apparatus; however, only minor dispersion of the transGolgi apparatus has been observed in MNV-infected cells without gross changes in morphology (30, 31). Furthermore, NS4 did not disrupt protein secretion in transfected cells (30, 31), which suggests that during MNV replication the predicted membrane-associated domain of NS4 may exert an alternative function to its HuNoV counterpart. (iv) NS5/VPg. VPg is found to be covalently bound to the 5= end of the genomic and subgenomic RNAs and plays a role in translation initiation and also in RNA synthesis by acting as a protein primer for NS7 (19, 25). For this role, VPg is guanylated on Y26 by NS7, and mutation of this residue has been shown to abolish viral replication (66). In accordance with these data, an insertion compromising Y26 was immediately lost from the profile at the first passage, further validating the profiling results. VPg mediates translation initiation by interacting with the cellular translation initiation factors eIF3, eIF4G1, and eIF4E, in addition to the S6 ribosomal protein, although it is worth noting that a direct interaction was demonstrated only for VPg and eIF3 and eIF4E (16, 19). Insertions tolerated through passage 3 were largely found in the C-terminal end of VPg, which could now also be examined for insertion of other small exogenous sequences. Many sites in the N-terminal half of the protein were still detectable at passage 2; however, the majority were not maintained through the third passage, suggesting that they in some way compromise viral fitness and are altered or lost from the population. Together this analysis indicates that the binding sites for the cellular translation initiation proteins may map to the N terminus of VPg, or otherwise, the selected C-terminal insertions must not disrupt any of these interactions. If the structure of VPg becomes available, then mapping these insertions should facilitate identification of the translation initiation factor binding sites on VPg. (v) NS6. The viral protease NS6 tolerated the fewest number of insertions, with only 9% of identified input insertions still detectable after three passages in cell culture. The structure and function of the HuNoV protease (3CLpro) have been extensively characterized, and the active site is thought to be a catalytic triad of His30, Glu54, and Cys139, although Glu54 is largely thought to deter-

Journal of Virology

Murine Norovirus Functional Profiling

FIG 5 Structural mapping of the 15-nt insertion sites. The insertions found in the input and passage 3 profiles were mapped onto the structures of NS7 (PDB accession number 3NAH) (A and B, respectively) and the P domain of VP1 (PDB accession number 3LQ6) (C and D, respectively). Blue residues highlight positions after which an insertion was detected. Where two insertions were made in a single codon with different outcomes, the most tolerated one was marked. In the P domain of VP1, the majority of insertions tolerated after three passages lie within flexible loop regions. The catalytic center in NS7 did not tolerate any insertions, further validating the profiling. The structures were mapped using Chimera software (UCSF).

mine substrate specificity rather than to be essential for catalysis (60, 61, 77). From sequence alignments, His30 and Cys139 are conserved between the Norwalk virus 3CLpro and MNV-1 NS6, whereas position 54 is an aspartic acid in MNV-1 NS6, although this implies that it has a similar function. An insertion was detected in the input profile only in Cys139, which, as predicted, was lost at the first passage. Taken together, this indicates that the catalytic triad and mechanism of catalysis may be conserved between the HuNoV 3CLpro and MNV NS6, which has not yet been established. The low degree of tolerated insertions would imply that the majority of the NS6 structure is important for maintaining protease activity without compromising viral replication. This also makes it very unlikely that other larger insertions will be tolerated in NS6, although any of the tolerated sites could be tested. The cleavage sites of MNV NS6 have been well mapped (63). Within the input profile, at least one insertion was identified in every cleavage junction within ORF1. Interestingly, none of these insertions would result in a change in the identity of the immediate cleavage-site residues, due to duplication of the target DNA

November 2012 Volume 86 Number 21

within the insertion. However, the position of the cleavage site is shifted 5 amino acids downstream and the identity of the preceding residues is altered. None of the insertions in the cleavagejunction residues were tolerated beyond passage 1 or 2, suggesting that altering the positioning or the surrounding residues may affect cleavage efficiency and viral fitness. (vi) NS7. NS7 is the viral RNA-dependent RNA polymerase (RdRp), which synthesizes the genomic, subgenomic, and antigenomic viral RNAs during replication. The structure of the MNV NS7 has been determined (27), which allowed the insertions present in the input and passage 3 profiles to be mapped (Fig. 5A and B). The majority of insertions were lost by passage 3, particularly those surrounding the active-site motifs near the center of the protein and loops involved in RNA binding and adjacent to the NTP binding site (residues 433 to 440 and 374 to 381, respectively), which, as expected, would likely compromise viral replication. The majority of the insertions remaining after passage 3, which could now be probed for other insertions, can be found in unstructured loops and in the bundle of eight ␣ helices on the

jvi.asm.org 11449

Thorne et al.

left-hand side of the structure in Fig. 5A and B. This bundle of ␣ helices comprises the fingers region of NS7, which, like other RdRps, is described as having a right-hand conformation (27) (Fig. 5B). (vii) VP1 and VF1. VP1 is the major capsid protein and is divided into an N-terminal arm, a shell (S) domain, and a protruding (P) domain. The S domain forms a smooth shell around the viral genome, while the P domain extends from the surface of the capsid and dimerizes into arch-like structures (52). The majority of insertions across VP1 were largely lost after a single passage in tissue culture (Fig. 4). Those that were still detected in the S domain after three passages clustered around residues 133 to 137, 176 to 177, and 195 to 199, suggesting that these regions may not be involved in tight intermolecular interactions within the capsid. These clusters also represent the regions which could be probed for other insertions, although any larger insertions could be more disruptive to the capsid. The region of ORF2 which encodes the S domain also overlaps the alternative reading frame ORF4 and has been shown to be highly conserved across MNV strains with suppression of synonymous-site variability (43). Thus, in addition to the role of the S domain in capsid formation, it is unsurprising that the majority of insertions in this region were lost, as any insertions may affect the function of both proteins. The insertion clusters tolerated for three passages in the S domain corresponded to amino acid residues 130 to 133, 172 to 173, and 192 to 196 in VF1. It is worth noting, however, that VF1 is not required for replication in cell culture, and as such, functional analysis of this kind is of limited value. The P domain is further divided into the P1 domain, which forms the base of the arch and is heavily involved in dimerization interactions, and the P2 domain, which forms the top of the arch, is the least conserved domain, and contains the receptor binding sites and sites for antigenicity (34, 38, 68). The structure of the P domain of MNV VP1 has recently been determined to a high resolution (68), which allows structural mapping of the insertion sites present in the input and passage 3 profiles (Fig. 5C and D, respectively). After passage 3, the remaining insertions map largely to flexible loops or unstructured regions that may be able to better accommodate small additional sequences. Interestingly, a small cluster that is stable throughout the passages maps to the C=-D= loop of the P2 domain, which is involved in dimerization interactions (68). This loop may therefore be sufficiently flexible to accommodate insertions while maintaining the interactions, or modification of the interactions does not significantly compromise the overall capsid structure. The remaining insertions appear to lie on the side of the P1 domain that is not thought to be at the dimer interface. Receptor binding sites and defined antigen sites have been mapped to the E=-F= loop of the P2 domain (68). Four insertions in this loop were detected in the input RNA; however, only one at the base of the loop was present at passage 3 (after amino acid 380), despite the reported high degree of diversity observed at the antigenic sites (46). As this domain is also thought to mediate cell binding, it suggests that there may be some structural constraints to maintain and that insertions here may reduce the viral fitness. Likewise, an insertion at amino acid 300, in the A=-B= loop in the P2 domain, which is also recognized by neutralizing antibodies, was not tolerated past passage 1, nor was an insertion in R396, one of the residues thought to be involved in an ionic lock between P2 dimers in the capsid that may mediate conformational changes required for infection (68).

11450

jvi.asm.org

(viii) VP2 and the 3= UTR. VP2 is a minor capsid protein, with reports from feline calicivirus (FCV) suggesting that two copies of it are incorporated into the capsid (62), although it is also possible that VP2 plays other roles in the viral life cycle. VP2 tolerated the highest percentage of insertions after the three passages, with 30% of the insertions present in the input remaining after three passages in cell culture. Clusters of stable insertion sites were found in particular at the start of VP2, near the VP1-VP2 boundary, and at the 3= end of the VP2-coding region (Fig. 4C). Any of these clusters could be tested in a trial for insertion of other small exogenous sequences. It is likely that the region between these clusters contains a functional domain or structure important in capsid interactions. The same region has also been demonstrated to be essential for replication of related FCVs (62). The basic nature of VP2 has led to the proposal that it may be involved in binding the viral RNA and also acidic regions of VP1, suggesting a structural basis of genome encapsidation, although a direct interaction between VP2 and RNA has not yet been demonstrated. Interestingly, the cluster of stable mutations in the 3= end of the VP2-coding region corresponds to the position of a putative stemloop structure (6). While the function of this structure is unknown, the nucleotides involved have been found to be largely conserved between different strains of MNV (6). However, insertions on both sides of the stem-loop appear to be possible, suggesting that this loop has a high degree of flexibility. Two stem-loop structures have been mapped to the 3= UTR and were also able to tolerate insertions at specific locations. The integrity of these stem-loops is essential for replication in tissue culture (6), which suggests that the stable insertions do not completely disrupt the structure of these stem-loops. The second of these stem-loops (otherwise called stem-loop 3 in previous literature [6]) contains a single-stranded polypyrimidine tract that is the binding site for the cellular proteins polypyrimidine tract binding protein (PTB) and poly(rC) binding protein isoform2 (PCBP2). An insertion in this region was lost after a single passage, which again validates this approach as disrupting this loop, and its interactions have previously been shown to incur a fitness cost in RAW264.7 cells (6). It is likely that these RNA structures also bind other cellular proteins during MNV replication, and so the insertional mutagenesis data should aid future studies to map potential protein binding sites. Validation of insertion sites. Several approaches were taken in order to validate the results of the profiling and confirm that the insertions identified during tissue culture selection represent replication-competent insertion viruses. First, a small selection of individual insertion viruses was isolated from each of the 15-nt insertion libraries by limiting dilution, and the sequences of the insertion sites were verified. All except one of the sites identified in the individually isolated, sequence-verified, and replication-competent viruses were found to within 1 nt in the tissue cultureselected profiles, which is similar to the accuracy previously published for this technique (3) (Table 2). The failure to identify a single insertion site in the profiling from 12 individually isolated viruses serves to highlight the fact that profiling using capillary electrophoresis is unlikely to identify all minority species within a population. A selection of insertion sites identified in the profiling was next validated by introducing them into the pT7:MNV 3= RZ WT clone. All of those predicted to be stable tolerated insertions (on the basis of the fact that they could be detected in the passage 3

Journal of Virology

Murine Norovirus Functional Profiling

TABLE 2 Validation of insertion profiles by identifying insertions in viruses isolated from 15-nt insertion libraries by limiting dilution Present in profilingb

No. of viruses isolated

Protein

Nucleotide position

Strain

Amino acid sequencea

NS1-2

218

WT Mutant

LAGLPV LAGCGRTGLPV

No

2

NS4

2523

WT Mutant

RPSFYW RPLRPQPSFYW

Yes

2

NS4

2604

WT Mutant

GWYHSE GWLRPQWYHSE

Yes

1

VPg

2840

WT Mutant

ATRRSR ATRCGRTRRSR

Yes

2

VP1

5073

WT Mutant

SDGAAP SDGCGRNGAAP

Yes

1

VP1

5585

WT Mutant

QEESMR QEDAAAEESMR

Yes

1

VP2

6688

WT Mutant

MAGALF MAGAAAAGALF

Yes

1

VP2

7229

WT Mutant

PRDHTP PRDCGRSDHTP

Yes

1

VP2

7239

WT Mutant

TPATQG TPVRPQPATQG

Yes

1

a Underlined residues correspond to the inserted residues; residues in italics indicate a change in residue identity due to insertion within a codon. b Insertion detected within selected profiles to within 1 nt.

profile; Table 3) successfully yielded infectious virus by reverse genetics recovery (Fig. 6) and were stable for three passages in RAW264.7 cells (data not shown). Importantly, a site identified to be not tolerated in the profiling (nucleotide position 6722), i.e., present in the input but lost upon a single passage in cell culture (Table 3), was also introduced into the WT clone and, as expected, failed to yield infectious virus (Fig. 6). While the use of low-multiplicity infections during passage of the recovered libraries in cell culture would minimize coinfection, we could not formally exclude the possibility that transencapsidation or complementation was occurring. To address this specifically, we examined the ability of a replication-incompetent insertion virus (nucleotide position 6722; Fig. 6) to be detected when cotransfected and passaged with a replication-competent virus (position 6686; Fig. 6). In vitro-transcribed capped RNA was generated for the validated cloned viruses with insertions at nt 6686 (viable) and 6722 (replication incompetent). Cotransfection of the nt 6722 insertion RNA in combination with either the nt 6686 insertion RNA or WT RNA was performed, the titer of the resulting virus was determined, and the virus was passaged at an MOI of 0.01, which had been used for the first passage of each library. While the nt 6722 insertion could be detected in the input RNA for both corecoveries, it was not detected after a single passage in the presence of either the nt 6686 insertion virus or WT virus (data not shown), thereby confirming that detection of replication-incompetent insertion mutants is unlikely using this experimental setup.

November 2012 Volume 86 Number 21

TABLE 3 Sequence and detection of insertions cloned into the WT to validate the profiling

Protein

Nucleotide position

Strain

Amino acid sequencea

NS1-2

84

WT Mutant

SSKASV SSKVRPHKASV

NS1-2

218c

WT Mutant

LAGLPV LAGCGRTGLPV

NS1-2

382

WT Mutant

NS4

2600

VP2

Profiling detectionb Input Passage 3 4358

31032

SHAEDA SHAAAAAHADEA

407

23352

WT Mutant

DDGWYH DDGCGRNGWYH

6348

31617

6686

WT Mutant

MAGALF MAGAAAAGALF

9181

17900

VP2

6688d

WT Mutant

MAGALF MAGAAAAGALF

VP2

6722

WT Mutant

GLMGII GLMCGRMMGII

4736

VP2

7229

WT Mutant

PRDHTP PRDCGRSDHTP

308

3399

VP2

7239

WT Mutant

TPATQG TPVRPQPATQG

283

4262

618

a Underlined residues correspond to the inserted residues; residues in italics indicate a change in residue identity due to insertion within a codon. b Numbers reflect detection of the insertion to within 1 nt and refer to fluorescence intensity of the corresponding peak. c Insertions were not detected in the profiling but were isolated by limiting dilution. d Insertion 6688 was not identified in the input profile.

Generation of tagged MNV. The identification of tolerated insertion sites provides the opportunity for insertion of other exogenous sequences into the genome. While attempts were made to produce infectious virus bearing green fluorescent protein (GFP), luciferase, or other reporters at a number of the validated sites, none resulted in the production of infection virus (data not shown). Therefore, the FLAG tag (DYKDDDDK) was next selected for insertion due to its small size and success with other small positive-strand RNA viruses (71). The FLAG tag was inserted into a site in both NS4 (nt 2600) and VP2 (nt 6686), both of which were stable validated sites in proteins of largely unknown function (Fig. 7A). Although the NS4 site had originally resulted in recovery of virus at reduced levels compared to WT (Fig. 6), the 2600 position translates to the highly charged C terminus of NS4, which seemed to be a suitable location for the equally highly charged FLAG tag. Using a DNA-based reverse genetics system, infectious FLAG-tagged viruses were recovered for both NS4FLAG and VP2-FLAG, although at slightly lower titers for the latter (Fig. 7B). Western blot analysis of NS4-FLAG- and VP2FLAG-infected cell lysates with an anti-FLAG antibody detected proteins of the expected molecular masses for NS4 (18 kDa) and VP2 (29 kDa), respectively (Fig. 7C). The NS4-FLAG insertion was stable for at least three passages in RAW264.7 cells, as confirmed by Western blotting and sequencing of the insertion site

jvi.asm.org 11451

Thorne et al.

FIG 6 Validation of insertion profiling data. Tolerated insertion sites maintained for three passages were introduced into the pT7:MNV 3= RZ (WT) infectious clone and yielded infectious virus by use of the DNA-based reverse genetics system. Numbers refer to the genome position after which the insertion was made. Insertions colored white lie within NS1-2, those in light gray are insertions in NS4, and those in dark gray are in VP2. Importantly, introduction of an insertion at position 6722, which was deemed replication incompetent in the profiling, failed to yield infectious virus by reverse genetics. The pT7:MNV 3= RZ NS7 frameshift clone (FS) was used as a negative control for virus recovery. Significance was tested using one-way analysis of variance and Dunnett’s multiple-comparison posttest to compare each insertion virus to the WT control, n ⫽ 3. **, P ⬍ 0.01.

(data not shown). In contrast, the VP2-FLAG virus was stable for only two passages, after which VP2-FLAG could no longer be detected by Western blotting with anti-FLAG antibodies, and sequence verification revealed that the insertion had been entirely lost (data not shown). To determine the subcellular localization of NS4 and VP2 during replication, the MNV permissive microglial BV-2 cells were infected with sequence-verified passage 1 stocks of NS4-FLAG, VP2-FLAG, and WT MNV-1. At 12 hpi, cells were fixed and stained with antibodies against the FLAG tag and NS7 and analyzed by confocal immunofluorescence microscopy. As previously reported (30), NS4-FLAG colocalized with NS7 in dense perinuclear foci (Fig. 8), in a position most likely consistent with the microtubule-organizing center, which has recently been reported to be the location for the replication complex (29). NS7 also appeared to show some diffuse staining throughout the cytoplasm of all infected cells, as has previously been observed (30). VP2-FLAG was also found to colocalize with NS7 in perinuclear foci (Fig. 8). Finally, to determine whether the FLAG tag could be used to immunoprecipitate NS4 and VP2, RAW264.7 cells were infected with passage 1 stocks of either NS4-FLAG, VP2-FLAG, or WT MNV-1, and the tagged proteins were enriched using anti-FLAG agarose. Immunoprecipitation and subsequent Western blot analysis of the input lysates and eluted proteins with an anti-FLAG

FIG 7 Engineering infectious FLAG-tagged murine noroviruses. (A) The FLAG epitope tag was inserted into sites in NS4 and VP2. (B) Infectious virus was recovered from NS4-FLAG and VP2-FLAG using the DNA-based reverse genetics system. Significance was tested using one-way analysis of variance and Dunnett’s multiple-comparison posttest to compare each insertion virus to the WT control, n ⫽ 3. **, P ⬍ 0.01. (C) Detection of the FLAG-tagged proteins in infected cell lysates by Western blotting. Detection with an anti-FLAG antibody gave proteins at the expected molecular mass for NS4 and VP2. *, a nonspecific protein with the same molecular mass as NS7 is detected by the anti-NS7 antibody. (D) Immunoprecipitation of NS4 and VP2 via the FLAG tag. RAW264.7 cells were infected with either NS4-FLAG and VP2-FLAG at a low MOI, at 36 hpi cell lysates were prepared, and anti-FLAG agarose was used to immunoprecipitate NS4 and VP2, respectively, as confirmed by Western blot analysis. NS1-2 was coimmunoprecipitated with NS4 only. Lanes M, molecular mass markers.

11452

jvi.asm.org

Journal of Virology

Murine Norovirus Functional Profiling

FIG 8 NS4 and VP2 localize with NS7 in infected cells. Permissive microglial cell line BV-2 was infected with NS4-FLAG, VP2-FLAG, or WT virus at an MOI of 1 or mock infected. At 12 hpi, cells were fixed and dual stained for the FLAG tag and NS7. Both NS4 and VP2 colocalized with NS7 in perinuclear foci.

antibody confirmed that NS4 and VP2 could be detected in the appropriate fractions (Fig. 7D). In addition, we also detected the previously reported interaction between the NS4 and the NS1-2 proteins observed in FCV (Fig. 7D). This work confirms that epitope-tagged replication-competent noroviruses may be of use in the dissection of the molecular interactions that occur within the viral replication complex. DISCUSSION

Despite significant recent advances in the study of norovirus replication, the role of many of the viral proteins has not yet been fully elucidated. In this study, we applied comprehensive Tn-mediated insertional mutagenesis to the MNV-1 genome, coupled with insertion-specific fluorescent PCR profiling, to identify the essential functional domains of each viral protein. Five 15-nt insertion libraries of MNV-1 infectious clones were generated by Tn mu-

November 2012 Volume 86 Number 21

tagenesis and were then subjected to selection of tolerated insertion sites in a permissive cell line. Functional analysis of over 2,000 insertions across the genome suggested previously unknown functional domains and allowed known functional motifs to be probed, shedding light on the role of several of the viral proteins. Cloning insertion sites into the WT MNV-1 genome served to validate the results of the profiling approach and facilitated insertion of a FLAG tag into sites within NS4 and VP2. Infectious FLAG-tagged viruses were recovered, making this the first report of viable MNV insertion viruses, which can be use to characterize the tagged viral proteins. Passage of the 15-nt insertion libraries in permissive cells resulted in selection of viable and stable insertion sites, while over 84% of insertions were lost during the three passages. Loss of the insertion sites during tissue culture selection could occur for several reasons. First, insertions that disrupt the function of a viral

jvi.asm.org 11453

Thorne et al.

protein domain or an RNA structure essential for replication are likely to be lethal or seriously debilitating to the virus and will therefore not be present within the replicating population. If an insertion results in a fitness cost, weaker viruses may eventually be outcompeted to undetectable levels. Furthermore, due to the high error rate of the RdRp (12), any changes in an insertion sequence that increase viral fitness may be favored and such insertions will no longer be detected by the insertion-specific primer. While these interpretations provide insights into which sites are required for viral replication, it is also possible that nonhomologous genomic recombination occurs within the complex population and contributes to the loss of insertions or generation of wild-type virus. It is worth noting, however, that low-MOI infections were performed, and viruses were harvested relatively early in the multicycle growth curve (16 hpi) in order to minimize this occurrence. The accuracy of detection of the insertions was to within 1/2 nt, which is similar to that previously reported for use of insertionspecific PCR profiling and capillary electrophoresis in the functional analysis of the hepatitis C virus genome (3). It was also evident that these techniques did not give resolution sufficient to detect all the minority insertions within such complex library populations, suggesting that this is not an exhaustive list of possible insertions. While the precise reason for the lack of sensitivity was not determined, the need to use detection thresholds based on the average peak intensity obtained from a WT nonmutagenized amplicon covering the same region undoubtedly contributed. Profiling the large number of insertions across the genome allowed the functional domains of the viral proteins to be probed and identified. For the profiling analysis, the focus was placed on the insertions that are immediately lost during selection and those which are stabilized for three passages, as a strategy to narrow down the functionally important protein regions. While insertions that were present for only the first or second passage most likely compromise viral replication, analysis of these insertions is complicated by the numerous reasons for their loss within a large library population, as described above. However, studying these insertions in isolation can also yield information on protein function and amino acid requirements, which could contribute to future, more detailed studies of the viral proteins, particularly NS4 and VP2. Likewise, site-directed mutagenesis provides a more direct, less disruptive mutagenesis strategy than 15-nt insertional mutagenesis. While site-directed mutagenesis is not a feasible approach for genome-wide study, it can now be used for more targeted analysis of the regions identified here to be functionally important. Despite the identification of numerous tolerated 15-nt insertion sites throughout the genome, generation of a reporter-tagged virus was not achieved, despite the use of different engineering strategies, multiple reporter proteins (GFP, Gaussia luciferase, improved light-, oxygen-, and voltage-sensing domain [iLOV], and the FlAsH tag), and several targeted insertion sites. This suggests that such large insertions were too disruptive to the viral protein structure and functions and as such abolished viral replication. It is unlikely that the insertions exceeded a genome packaging limit, as similarly sized FCV has been tagged with GFP and Discosoma sp. red fluorescent protein (DsRed) (1). The reporter-tagged FCVs were also generated using Tn mutagenesis to identify tolerated insertion sites. Due to differences in the Tn mutagenesis strategy, genome-wide functional analysis was not performed, but several tolerated insertion sites were identified in VP2 and the leader

11454

jvi.asm.org

capsid protein, which is the only FCV protein for which MNV does not have an equivalent. The site in the leader capsid protein was found to be the only one that would subsequently tolerate insertion of GFP and DsRed (1). It has been suggested that generating a reporter-tagged virus in this way can be a largely empirical exercise of matching an appropriate reporter gene with an appropriate site and protein context. For example, DsRed was stably inserted into a site in the poliovirus genome, whereas a GFP insertion was quickly expelled from the same site (70). Further studies on the generation of infectious reporter-tagged MNV-1 are under way. An alternative insertion to reporter proteins is small peptide tags, which can still be used to characterize viral proteins, but not for live in vivo imaging, but they have the advantage that the size of the insertion is most likely much less disruptive to viral proteins. The FLAG tag was successfully inserted into sites in both NS4 and VP2 and yielded infectious tagged virus that could be characterized through Western blotting, immunofluorescence analysis, and immunoprecipitation. Via the FLAG tag, VP2 was shown for the first time to colocalize with NS7 in infected cells, suggesting that it also associates with the replication complex. Both NS4 and VP2 could be immunoprecipitated by means of the FLAG tag, and NS1-2 was coimmunoprecipitated with NS4, an interaction previously demonstrated for the FCV equivalent proteins p32 (NS2) and p30 (NS4) (Table 1) (32). This suggests that immunoprecipitation of FLAG-tagged viral proteins, generated using the insertion sites identified here, can be used to enable the characterization of the viral and cellular interaction partners of NS4 and VP2. For VP2, it will be interesting to determine if viral RNA can be coimmunoprecipitated to probe its potential role in genome encapsidation. Although the insertion sites eventually used for FLAG insertion were carefully selected, it is also possible that many of the other stable insertion sites identified in the profiling could accommodate a FLAG tag or a similarly sized tag, such as the His or c-myc tags. This would allow other viral proteins and their essential domains to be characterized in a similar way. It is also possible that these insertions could be engineered into the HuNoV proteins at similar sites; for example, the C terminus of HuNoV p22 is also highly charged, suggesting that it may similarly accommodate a FLAG tag to generate replication-competent RNA for transfection studies. In summary, transposon-mediated insertional mutagenesis has allowed us to perform a wide-scale functional analysis of the entire MNV genome, which led to the identification of protein domains required for replication in viral proteins of largely unknown function. Future characterization of the FLAG-tagged viruses and generation of other peptide-tagged viruses will shed further light on the specific functions of each viral protein. ACKNOWLEDGMENTS This work was supported by funding from the Wellcome Trust to L.T. in the form of a Ph.D. studentship and to I.G. as a senior fellowship in basic biomedical science. I.G. is a Wellcome senior fellow.

REFERENCES 1. Abente EJ, Sosnovtsev SV, Bok K, Green KY. 2010. Visualization of feline calicivirus replication in real-time with recombinant viruses engineered to express fluorescent reporter proteins. Virology 400:18 –31. 2. Alkhouri N, Danziger-Isakov L. 2011. Norovirus and severe chronic

Journal of Virology

Murine Norovirus Functional Profiling

3. 4. 5. 6. 7. 8. 9.

10. 11. 12. 13.

14.

15. 16. 17. 18. 19. 20. 21.

22. 23.

24.

25.

gastroenteritis in pediatric stem cell transplantation: the plot thickens. Pediatr. Transplant. 15:671– 672. Arumugaswami V, et al. 2008. High-resolution functional profiling of hepatitis C virus genome. PLoS Pathog. 4:e1000182. doi:10.1371/ journal.ppat.1000182. Arumugaswami V, et al. 2009. High-resolution functional profiling of a gammaherpesvirus RTA locus in the context of the viral genome. J. Virol. 83:1811–1822. Atasheva S, Gorchakov R, English R, Frolov I, Frolova E. 2007. Development of Sindbis viruses encoding nsP2/GFP chimeric proteins and their application for studying nsP2 functioning. J. Virol. 81:5046 –5057. Bailey D, et al. 2010. Functional analysis of RNA structures present at the 3= extremity of the murine norovirus genome: the variable polypyrimidine tract plays a role in viral virulence. J. Virol. 84:2859 –2870. Bailey D, Thackray LB, Goodfellow IG. 2008. A single amino acid substitution in the murine norovirus capsid protein is sufficient for attenuation in vivo. J. Virol. 82:7725–7728. Baker ES, Krause LSKL, Lambden PR, Clarke IN, Ward VK. 2012. Inherent structural disorder and dimerisation of murine norovirus NS1-2 protein. PLoS One 7:e30534. doi:10.1371/journal.pone.0030534. Beitzel BF, Bakken RR, Smith JM, Schmaljohn CS. 2010. Highresolution functional mapping of the Venezuelan equine encephalitis virus genome by insertional mutagenesis and massively parallel sequencing. PLoS Pathog. 6:e1001146. doi:10.1371/journal.ppat.1001146. Blasi E, Barluzzi R, Bocchini V, Mazzolla R, Bistoni F. 1990. Immortalization of murine microglial cells by a v-raf/v-myc carrying retrovirus. J. Neuroimmunol. 27:229 –237. Brown DD, et al. 2005. Rational attenuation of a morbillivirus by modulating the activity of the RNA-dependent RNA polymerase. J. Virol. 79: 14330 –14338. Bull R, et al. 2011. Comparison of the replication properties of murine and human calicivirus RNA-dependent RNA polymerases. Virus Genes 42:16 –27. Capizzi T, Makari-Judson G, Steingart R, Mertens W. 2011. Chronic diarrhea associated with persistent norovirus excretion in patients with chronic lymphocytic leukemia: report of two cases. BMC Infect. Dis. 11: 131. Chachu KA, LoBue AD, Strong DW, Baric RS, Virgin HW. 2008. Immune mechanisms responsible for vaccination against and clearance of mucosal and lymphatic norovirus infection. PLoS Pathog. 4:e1000236. doi:10.1371/journal.ppat.1000236. Chachu KA, et al. 2008. Antibody is critical for the clearance of murine norovirus infection. J. Virol. 82:6610 – 6617. Chaudhry Y, et al. 2006. Caliciviruses differ in their functional requirements for eIF4F components. J. Biol. Chem. 281:25315–25325. Chaudhry Y, Skinner MA, Goodfellow IG. 2007. Recovery of genetically defined murine norovirus in tissue culture by using a fowlpox virus expressing T7 RNA polymerase. J. Gen. Virol. 88:2091–2100. Chen S-Y, et al. 2009. Norovirus infection as a cause of diarrheaassociated benign infantile seizures. Clin. Infect. Dis. 48:849 – 855. Daughenbaugh K, Wobus C, Hardy M. 2006. VPg of murine norovirus binds translation initiation factors in infected cells. Virol. J. 3:33–39. Ebdrup L, Böttiger B, Mølgaard H, Laursen AL. 2011. Devastating diarrhoea in a heart-transplanted patient. J. Clin. Virol. 50:263–265. Ettayebi K, Hardy ME. 2003. Norwalk virus nonstructural protein p48 forms a complex with the SNARE regulator VAP-A and prevents cell surface expression of vesicular stomatitis virus G protein. J. Virol. 77:11790 – 11797. Fernandez-Vega V, et al. 2004. Norwalk virus N-terminal nonstructural protein is associated with disassembly of the Golgi complex in transfected cells. J. Virol. 78:4827– 4837. Frolov I, Garmashova N, Atasheva S, Frolova EI. 2009. Random insertion mutagenesis of Sindbis virus nonstructural protein 2 and selection of variants incapable of downregulating cellular transcription. J. Virol. 83: 9031–9044. Gerondopoulos A, Jackson T, Monaghan P, Doyle N, Roberts LO. 2010. Murine norovirus-1 cell entry is mediated through a non-clathrin-, noncaveolae-, dynamin- and cholesterol-dependent pathway. J. Gen. Virol. 91:1428 –1438. Han KR, et al. 2010. Murine norovirus-1 3Dpol exhibits RNA-dependent RNA polymerase activity and nucleotidylylates on Tyr of the VPg. J. Gen. Virol. 91:1713–1722.

November 2012 Volume 86 Number 21

26. Hardy M. 2005. Norovirus protein structure and function. FEMS Microbiol. Lett. 253:1– 8. 27. Högbom M, Jäger K, Robel I, Unge T, Rohayem J. 2009. The active form of the norovirus RNA-dependent RNA polymerase is a homodimer with cooperative activity. J. Gen. Virol. 90:281–291. 28. Hsu CC, Riley LK, Wills HM, Livingston RS. 2006. Persistent infection with and serologic cross-reactivity of three novel murine noroviruses. Comp. Med. 56:247–251. 29. Hyde JL, Gillespie LK, Mackenzie JM. 2012. Mouse norovirus-1 utilises the cytoskeleton network to establish localization of the replication complex proximal to the microtubule organizing centre. J. Virol. 86:4110 – 4122. 30. Hyde JL, Mackenzie JM. 2010. Subcellular localization of the MNV-1 ORF1 proteins and their potential roles in the formation of the MNV-1 replication complex. Virology 406:138 –148. 31. Hyde JL, et al. 2009. Mouse norovirus replication is associated with virus-induced vesicle clusters originating from membranes derived from the secretory pathway. J. Virol. 83:9709 –9719. 32. Kaiser WJ, Chaudhry Y, Sosnovtsev SV, Goodfellow IG. 2006. Analysis of protein-protein interactions in the feline calicivirus replication complex. J. Gen. Virol. 87:363–368. 33. Karst S, Wobus C, Lay M, Davidson J, Virgin H. 2003. STAT1dependent innate immunity to a Norwalk-like virus. Science 299:1575– 1578. 34. Katpally U, et al. 2010. High-resolution cryo-electron microscopy structures of murine norovirus 1 and rabbit hemorrhagic disease virus reveal marked flexibility in the receptor binding domains. J. Virol. 84:5836 – 5841. 35. Kekarainen T, Savilahti H, Valkonen JPT. 2002. Functional genomics on potato virus A: virus genome-wide map of sites essential for virus propagation. Genome Res. 12:584 –594. 36. Khan RR, et al. 2009. Gastrointestinal norovirus infection associated with exacerbation of inflammatory bowel disease. J. Pediatr. Gastroenterol. Nutr. 48:328 –333. 37. Laurent LC, Olsen MN, Crowley RA, Savilahti H, Brown PO. 2000. Functional characterization of the human immunodeficiency virus type 1 genome by genetic footprinting. J. Virol. 74:2760 –2769. 38. Lochridge VP, Hardy ME. 2007. A single amino acid substitution in the P2 domain of VP1 of murine norovirus is sufficient for escape from antibody neutralization. J. Virol. 81:12316 –12322. 39. Lopman BA, et al. 2004. Epidemiology and cost of nosocomial gastroenteritis, Avon, England, 2002-2003. Emerg. Infect. Dis. 10:1827–1834. 40. Ludwig A, Adams O, Laws HJ, Schroten H, Tenenbaum T. 2008. Quantitative detection of norovirus excretion in pediatric patients with cancer and prolonged gastroenteritis and shedding of norovirus. J. Med. Virol. 80:1461–1467. 41. Marks PJ, et al. 2000. Evidence for airborne transmission of Norwalk-like virus (NLV) in a hotel restaurant. Epidemiol. Infect. 124:481– 487. 42. McCartney SA, et al. 2008. MDA-5 recognition of a murine norovirus. PLoS Pathog. 4:e1000108. doi:10.1371/journal.ppat.1000108. 43. McFadden N, et al. 2011. Norovirus regulation of the innate immune response and apoptosis occurs via the product of the alternative open reading frame 4. PLoS Pathog. 7:e1002413. doi:10.1371/journal. ppat.1002413. 44. Moorman NJ, Lin CY, Speck SH. 2004. Identification of candidate gammaherpesvirus 68 genes required for virus replication by signature-tagged transposon mutagenesis. J. Virol. 78:10282–10290. 45. Moradpour D, et al. 2004. Insertion of green fluorescent protein into nonstructural protein 5A allows direct visualization of functional hepatitis C virus replication complexes. J. Virol. 78:7400 –7409. 46. Müller B, Klemm U, Mas Marques A, Schreier E. 2007. Genetic diversity and recombination of murine noroviruses in immunocompromised mice. Arch. Virol. 152:1709 –1719. 47. Patel MM, et al. 2008. Systematic literature review of role of noroviruses in sporadic gastroenteritis. Emerg. Infect. Dis. 14:1224 –1231. 48. Perry JW, Taube S, Wobus CE. 2009. Murine norovirus-1 entry into permissive macrophages and dendritic cells is pH-independent. Virus Res. 143:125–129. 49. Perry JW, Wobus CE. 2010. Endocytosis of murine norovirus 1 into murine macrophages is dependent on dynamin II and cholesterol. J. Virol. 84:6163– 6176. 50. Pfister T, Wimmer E. 2001. Polypeptide p41 of a Norwalk-like virus is a

jvi.asm.org 11455

Thorne et al.

51. 52. 53. 54. 55. 56. 57.

58. 59.

60. 61. 62. 63.

nucleic acid-independent nucleoside triphosphatase. J. Virol. 75:1611– 1619. Pierson TC, et al. 2005. An infectious West Nile virus that expresses a GFP reporter gene. Virology 334:28 – 40. Prasad BV, et al. 1999. X-ray crystallographic structure of the Norwalk virus capsid. Science 286:287–290. Rhinehart E, Walker S, Murphy D, O’Reilly K, Leeman P. 2012. Frequency of outbreak investigations in US hospitals: results of a national survey of infection preventionists. Am. J. Infect. Control 40:2– 8. Rothenberg SM, Olsen MN, Laurent LC, Crowley RA, Brown PO. 2001. Comprehensive mutational analysis of the Moloney murine leukemia virus envelope protein. J. Virol. 75:11851–11862. Schorn R, et al. 2010. Chronic norovirus infection after kidney transplantation: molecular evidence for immune-driven viral evolution. Clin. Infect. Dis. 51:307–314. Schwartz S, et al. 2011. Norovirus gastroenteritis causes severe and lethal complications after chemotherapy and hematopoietic stem cell transplantation. Blood 117:5850 –5856. Sharp TM, Guix S, Katayama K, Crawford SE, Estes MK. 2010. Inhibition of cellular protein secretion by Norwalk virus nonstructural protein p22 requires a mimic of an endoplasmic reticulum export signal. PLoS One 5:e13130. doi:10.1371/journal.pone.0013130. Simmonds P, et al. 2008. Bioinformatic and functional analysis of RNA secondary structure elements among different genera of human and animal caliciviruses. Nucleic Acids Res. 36:2530 –2546. Smith DB, McFadden N, Blundell RJ, Meredith A, Simmonds P. 2012. Diversity of murine norovirus in wild-rodent populations: speciesspecific associations suggest an ancient divergence. J. Gen. Virol. 93:259 – 266. Someya Y, Takeda N. 2009. Insights into the enzyme-substrate interaction in the norovirus 3C-like protease. J. Biochem. 146:509 –521. Someya Y, Takeda N, Wakita T. 2008. Saturation mutagenesis reveals that GLU54 of norovirus 3C-like protease is not essential for the proteolytic activity. J. Biochem. 144:771–780. Sosnovtsev SV, Belliot G, Chang K-O, Onwudiwe O, Green KY. 2005. Feline calicivirus VP2 is essential for the production of infectious virions. J. Virol. 79:4012– 4024. Sosnovtsev SV, et al. 2006. Cleavage map and proteolytic processing of the murine norovirus nonstructural polyprotein in infected cells. J. Virol. 80:7816 –7831.

11456

jvi.asm.org

64. Strong DW, Thackray LB, Smith TJ, Virgin HW. 2012. Protruding domain of capsid protein is necessary and sufficient to determine murine norovirus replication and pathogenesis in vivo. J. Virol. 86:2950 –2958. 65. Stuart RL, et al. 2010. An outbreak of necrotizing enterocolitis associated with norovirus genotype GII. 3. Pediatr. Infect. Dis. J. 29:644 – 647. 66. Subba-Reddy CV, Goodfellow I, Kao CC. 2011. VPg-primed RNA synthesis of norovirus RNA-dependent RNA polymerases by using a novel cell-based assay. J. Virol. 85:13027–13037. 67. Taube S, et al. 2009. Ganglioside-linked terminal sialic acid moieties on murine macrophages function as attachment receptors for murine noroviruses. J. Virol. 83:4092– 4101. 68. Taube S, et al. 2010. High-resolution X-ray structure and functional analysis of the murine norovirus 1 capsid protein protruding domain. J. Virol. 84:5695–5705. 69. Teterina NL, et al. 2011. Identification of tolerated insertion sites in poliovirus non-structural proteins. Virology 409:1–11. 70. Teterina NL, Levenson EA, Ehrenfeld E. 2010. Viable polioviruses that encode 2A proteins with fluorescent protein tags. J. Virol. 84:1477–1488. 71. Teterina NL, Pinto Y, Weaver JD, Jensen KS, Ehrenfeld E. 2011. Analysis of poliovirus protein 3A interactions with viral and cellular proteins in infected cells. J. Virol. 85:4284 – 4296. 72. Teunis PFM, et al. 2008. Norwalk virus: how infectious is it? J. Med. Virol. 80:1468 –1476. 73. Thackray LB, et al. 2007. Murine noroviruses comprising a single genogroup exhibit biological diversity despite limited sequence divergence. J. Virol. 81:10460 –10473. 74. Turcios-Ruiz RM, et al. 2008. Outbreak of necrotizing enterocolitis caused by norovirus in a neonatal intensive care unit. J. Pediatr. 153:339 – 344. 75. Wobus CE, et al. 2004. Replication of norovirus in cell culture reveals a tropism for dendritic cells and macrophages. PLoS Biol. 2:e432. doi: 10.1371/journal.pbio.0020432. 76. Yunus MA, Chung LMW, Chaudhry Y, Bailey D, Goodfellow I. 2010. Development of an optimized RNA-based murine norovirus reverse genetics system. J. Virol. Methods 169:112–118. 77. Zeitler CE, Estes MK, Venkataram Prasad BV. 2006. X-ray crystallographic structure of the Norwalk virus protease at 1.5-Å resolution. J. Virol. 80:5050 –5058.

Journal of Virology