The Role of the VP4 Attachment Protein in Rotavirus Host Range Restriction in an In Vivo Suckling Mouse Model

ABSTRACT

The basis for rotavirus (RV) host range restriction (HRR) is not fully understood but is likely multigenic. RV genes encoding VP3, VP4, NSP1, NSP2, NSP3, and NSP4 have been associated with HRR in various studies. With the exception of NSP1, little is known about the relative contribution of the other RV genes to HRR. VP4 has been linked to HRR because it functions as the RV cell attachment protein, but its actual role in HRR has not been fully assessed. We generated a collection of recombinant RVs (rRVs) in an isogenic murine-like RV genetic background, harboring either heterologous or homologous VP4 genes from simian, bovine, porcine, human, and murine RV strains, and characterized these rRVs in vitro and in vivo. We found that a murine-like rRV encoding a simian VP4 was shed, spread to uninoculated littermates, and induced diarrhea comparably to rRV harboring a murine VP4. However, rRVs carrying VP4s from both bovine and porcine RVs had reduced diarrhea, but no change in fecal shedding was observed. Both diarrhea and shedding were reduced when VP4 originated from a human RV strain. rRVs harboring VP4s from human or bovine RVs did not transmit to uninoculated littermates. We also generated two rRVs harboring reciprocal chimeric murine or bovine VP4. Both chimeras replicated and caused disease as efficiently as the parental strain with a fully murine VP4. These data suggest that the genetic origin of VP4 partially modulates HRR in the suckling mouse and that both the VP8* and VP5* domains independently contribute to pathogenesis and transmission.

IMPORTANCE Human group A rotaviruses (RVs) remain the most important cause of severe acute gastroenteritis among infants and young children worldwide despite the introduction of several safe and effective live attenuated vaccines. The lack of knowledge regarding fundamental aspects of RV biology, such as the genetic basis of host range restriction (HRR), has made it difficult to predictively and efficiently design improved, next-generation live attenuated rotavirus vaccines. Here, we engineered a collection of VP4 monoreassortant RVs to systematically explore the role of VP4 in replication, pathogenicity, and spread, as measures of HRR, in a suckling mouse model. The genetic and mechanistic bases of HRR have substantial clinical relevance given that this restriction forms the basis of attenuation for several replication-competent human RV vaccines. In addition, a better understanding of RV pathogenesis and the determinants of RV spread is likely to enhance our ability to improve antiviral drug and therapy development.

INTRODUCTION

Rotaviruses (RVs) belong to the Reoviridae family and are comprised of a variety of icosahedral, nonenveloped multisegmented double-stranded RNA (dsRNA) viruses. RVs have three concentric layers of proteins that surround the 11 dsRNA segments encoding six structural (VP1 to VP4, VP6, VP7) and six nonstructural proteins (NSP1 to NSP6) (1). The outermost capsid layer is formed by trimers of VP7 that constitute the smooth surface of the virion and by trimers of the spike protein VP4, which serves as the RV cell attachment protein (1, 2).

Group A RVs are small bowel pathogens that infect the young of many avian and mammalian species and remain the most common cause of acute, severe gastroenteritis among infants and young children worldwide, responsible for between 128,000 and 215,000 deaths each year, primarily in developing countries. This morbidity and mortality continue despite the availability of several safe and effective RV vaccines (3, 4).

Besides their broad host range, RVs exhibit marked species specificity. In nature, RVs from one species (homologous host) are relatively infrequently isolated from another species (heterologous host), and these infrequent isolations rarely persist and spread efficiently over time in the heterologous host species. The limited capacity of certain viruses, such as RVs, to grow and transmit efficiently in an animal species that is distinct from the species they naturally infect is called host range restriction (HRR) (5–8). Under experimental conditions, cross-species infection with RVs tends to have diminished virulence and replication capacity and appears not to spread efficiently in the new species. The viral factors underlying RV HRR are not fully understood but are likely to be multigenic. Early studies with reassortant RVs in various animal model systems implicated genes encoding VP3, VP4, NSP1, NSP2, NSP3, and NSP4 as being associated with HRR (6–13). Of note, RV HRR has been defined differently in different studies: some studies have focused specifically on host-restricted replication (7, 9, 11), others on host-restricted virulence (12, 13), while at least one other on the host-restricted capacity to spread (14).

NSP1 is the product of RV gene 5 and has been studied in some detail in vitro and in vivo. NSP1 antagonizes the host innate immune response, especially the type I interferon (IFN-I) response by degrading several key components of IFN signaling, including IRF3, IRF5, IRF7, and β-TrCP. The abilities of NSP1 to effectively antagonize host IFN signaling (15, 16) and potentially IFN inhibition-independent activities are thought to contribute to HRR (7, 8, 10, 17–19).

Gene 4 encodes the RV attachment protein VP4, which interacts with a cell receptor(s) during cell entry and is identified as a significant determinant of RV cell and host tropism (1, 3, 20). Once assembled on the virus particles, VP4 requires cleavage by trypsin-like proteases into the N-terminal fragment VP8* (28 kDa) and C-terminal fragment VP5* (60 kDa) to become fully infectious (1, 2, 20). Although virus-cell contact is essential to initiate infection, the precise role and relevance of the species origin of VP4 in defining RV HRR is not clear. In a previous study (8), we used genetic reassortants generated from a mixed infection with the non-cell culture-adapted murine RV strain EW and a cell culture-adapted simian RV RRV strain to study the association of specific RV genes with replication and diarrhea in the suckling mouse model of infection. We found RRV VP4 only moderately reduced murine RV infectivity; however, VP4 from a bovine RV UK strain severely restricted intestinal replication in suckling mice (8). Due to the labor-intensive nature of isolating monoreassortants and low likelihood of incorporating of heterologous VP4s in a murine RV genetic background when using traditional coinfection approaches, the role of VP4 in HRR has not been interrogated in detail until now, with the assistance of the RV reverse genetics (RG) system.

RESULTS

Monoreassortant rRVs in an isogenic murine-like RV, harboring homologous or heterologous VP4s, can be rescued efficiently using an optimized RV RG system.

The role of VP4 as a viral determinant of HRR has been suggested in several studies (8, 9, 11, 13, 21). However, the lack of a highly tractable RG system to engineer the required recombinatnt RV (rRV) monoreassortants made it difficult to directly test the specific role of VP4 in HRR in a tractable small animal model.

We previously reported the rescue of a murine-like reassortant rRV (rD6/2-2g) via RG (22). In that murine-like recombinant RV, RV genes 2, 3, 5, 6, 7, 8, 9, and 11 were derived from the wild-type (WT) non-culture-adapted murine EW RV strain (EW_wt), gene 4 was derived from the cell culture-adapted simian RRV strain, and genes 1 and 10 came from the cell culture-adapted simian SA11 strain. This murine-like RV was genetically stable, efficiently infected suckling mice, reproducibly induced diarrhea, and was capable of spreading to uninfected littermates, as did the original D6/2 reassortant RV (8, 22). We reasoned that if the prototype rD6/2-like simian gene 4 were to be interchanged with homologous (murine) or heterologous (nonmurine) gene 4s in this isogenic murine-like RV background, we could directly evaluate the role of VP4 as a determinant of HRR in a suckling mouse model, with a focus on RV replication, pathogenicity, and the ability to spread to uninoculated littermates as measures of HRR.

We selected gene 4s from the following RV strains for this study: (i) simian SA11 (GenBank LC178567.1), (ii) bovine UK (M22306.1), (iii) porcine OSU (KJ450845.1), (iv) human CDC-9 p44/45 (23–25), and (v) the cell culture-adapted murine ETD strain (GQ479950.1). The cDNA sequences of all studied VP4s, including murine EW VP4 and RRV VP4, were used to generate a phylogenetic tree (Fig. 1A). Based on the phylogenetic tree, the simian and porcine RV VP4s are more closely related to murine RV VP4s than either bovine or human RV VP4s.

FIG 1.

VP4 monoreassortants based on an isogenic murine-like RV background can be successfully rescued via an optimized RG system. (A) A neighbor-joining phylogenetic tree was generated using gene 4 sequences of all VP4s in this study. Tree branches are scaled by the number of nucleotide substitutions per site, and values are shown on each branch. SH-like support values are indicated (percentages), and the GenBank accession numbers are shown in brackets. For CDC-9 gene 4 [¤], see reference 25. (B) The dsRNA genomic profile from VP4 monoreassortants based on murine-like RV were examined by RNA-PAGE. The positions of segment 4 are marked with black arrowheads. The segment numbers are indicated.

Using an optimized RG protocol (22), we generated a collection of rRVs based on the murine-like RV backbone and harboring heterologous or homologous gene 4s from the cell culture-adapted simian, bovine, porcine, human, and murine RV strains. Because all rRVs carrying different VP4s are based on the same genetic background (murine-like RV), we refer to them as VP4 monoreassortants (SA11-VP4, UK-VP4, OSU-VP4, CDC-9-VP4, and ETD-VP4) in the rest of this report.

To validate the identity of rescued VP4s monoreassortants, the viral RNA was extracted from polyethylene glycol (PEG)-concentrated precipitates or sucrose cushion-concentrated RV-infected lysates and analyzed by RNA polyacrylamide gel electrophoresis (PAGE) (Fig. 1B). The dsRNA migration patterns among murine-like RV and VP4 monoreassortants were identical for all segments except for segment 4, which comigrated with the corresponding segment 4 from the parental RV strain (Fig. 1B, inset). The identities of gene 4 of each monoreassortant were also verified by Sanger sequencing (data not shown).

Genetic origin of VP4 impacts cell culture replication and plaque sizes of VP4 monoreassortants.

The growth of these VP4 monoreassortants was first characterized and compared to that of the prototype murine-like RV in MA104 cells. All VP4 monoreassortants were serially passaged three times in MA104 cells, and the subsequent progeny titer was determined in a focus-forming assay. SA11-VP4, UK-VP4, and OSU-VP4 monoreassortants reached titers similar to the murine-like RV (RRV-VP4) (Fig. 2A). Monoreassortant RVs harboring murine (ETD) and human (CDC-9) VP4s had significantly lower titers than the murine-like RV (Fig. 2A). We also performed multistep growth curves of these rRVs. The replication curves of SA11-VP4, UK-VP4, and OSU-VP4 monoreassortants were comparable to that of murine-like RV (RRV-VP4). In contrast, replication curves for ETD-VP4 (murine VP4) and CDC-9-VP4 (human VP4) monoreassortants were significantly lower (Fig. 2B). Finally, we compared plaque sizes of the VP4 monoreassortants. All VP4 monoreassortants formed plaques, though plaque sizes varied significantly (Fig. 2C and D). Recombinant RVs harboring murine or human VP4s had smaller plaques than the prototype murine-like RV (RRV-VP4), while the porcine VP4 (OSU-VP4) monoreassortant formed significantly larger plaques. No significant differences in plaque size were noticed among simian SA11-VP4, bovine UK-VP4, and murine-like RV (RRV-VP4) monoreassortants (Fig. 2D). Taken together, these findings indicate that cell culture replication and plaque size of the VP4 monoreassortant in MA104 cells are at least partially determined by the genetic origin of VP4.

FIG 2.

Cell culture characterization of VP4 monoreassortants. (A) The VP4 monoreassortants were serially passaged 3 times in MA104 cells. Cells were harvested at day 7 postinfection or when complete cytopathic effects were observed; the virus titer was determined by an immunoperoxidase focus-forming assay. (B) Growth kinetics of VP4 monoreassortants. MA104 cell monolayers were infected with the indicated rRVs at a multiplicity of infection of 0.01 in the presence of trypsin (0.5 μg/mL) and harvested at the indicated times by freeze-thawing. The viral titers were determined in an immunoperoxidase focus-forming assay. (C) Comparison of plaques size. Plaques were generated on MA104 monolayers and detected at 7 days postinfection by crystal violet staining. Representative photographs of indicated recombinant RVs are shown. (D) The diameters of at least 50 randomly selected plaques were measured using a bright-field microscope. Mean values and the standard deviations (SD) from three (Fig. 2A) and two (Fig. 2B to D) independent experiments are shown. Statistical significance was evaluated by the ordinary one-way analysis of variance (ANOVA) and Dunnett's multiple comparisons post hoc test. The evaluation was performed using the parental virus (RRV-VP4) as reference. The asterisks indicate significant difference: *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001, ****, P ≤ 0.0001.

Genetic origins of VP4 partially determine RV replication and shedding in vivo.

The host range of viruses is determined by species-specific interactions between virus and host cell factors. These interactions include the ability to bind and enter cells, replicate their genome, evade host innate and acquired immune restriction factors, and transmit between individuals (26). To assess the contribution of VP4 to HRR in the suckling mouse model, we evaluated the effect of homologous or heterologous VP4s on in vivo replication, diarrhea, and spread to uninoculated littermates using our prototype murine-like RV as the standard. Four- to five-day-old suckling 129sv mice were orally gavaged with 1 × 10⁴ focus-forming units (FFUs) of the individual VP4 monoreassortant RVs. The heterologous simian RRV strain and the highly pathogenic non-cell culture-adapted WT murine RV (EW_wt) served as additional controls. For most experiments, eight pups were inoculated with the indicated RVs, and two additional pups in the litter were inoculated with phosphate-buffered saline (PBS) to determine if the specific recombinant RVs being tested would spread to the mock-inoculated littermates under these conditions.

The time course of diarrhea among infected littermates (Fig. 3, in blue) and the level of intestinal replication, measured by the shedding of RV in the feces by enzyme-linked immunosorbent assay (ELISA) (Fig. 3, in red), were monitored for 12 days. As reported, the heterologous simian RRV did not replicate well in nonimmunodeficient suckling mice (8, 22, 27, 28); at the inoculum dose used in this study, ELISA analysis failed to detect any RRV antigen in feces collected during the infection time course (Fig. 3A). However, a brief diarrhea peak was noted for 1 to 3 days after RRV inoculation (Fig. 3A), consistent with previous studies (8, 22, 27). In contrast, suckling mice infected with the homologous murine EW_wt RV developed severe diarrhea, which started as early as 1 day postinoculation and lasted until day 11 postinfection. The EW_wt fecal shedding curve demonstrated the efficient and self-limiting replication of fully homologous non-cell culture-adapted EW_wt RV in the intestine of suckling mice (Fig. 3B).

FIG 3.

Fecal shedding (red), daily diarrhea patterns (blue), and presence of transmission (↔) of indicated RVs in wild-type 129sv mice. Five-day-old 129sv mice were orally gavaged with the indicated RV. RV fecal shedding was monitored by ELISA. Virus shedding within the same group on each day is shown as the mean ± SD (in red). Diarrheal development (in blue) was scored from days 0 to 12 postinfection (see Table 1 for statistical analysis). To evaluate the presence of transmission in the litters of each VP4 monoreassortant, 2 mock-infected pups were cohoused with the inoculated littermates in the same cage. The ↔ symbol indicates that transmission to uninoculated pups occurred.

The time course of diarrhea and fecal shedding displayed distinct patterns for each of the VP4 monoreassortants (Fig. 3C to H). Based on these shedding curves, each VP4 monoreassortant replicated in the mouse small intestines but to somewhat different levels. To quantify the differences in replication, the areas under the fecal RV shedding curves during the experimental period were analyzed and compared with the murine-like RV (RRV-VP4) and with EW_wt by a nonparametric Mann-Whitney U test; these comparisons are summarized in Table 1.

TABLE 1.

Comparison of fecal shedding (by ELISA) and diarrhea development of VP4 monoreassortants on a murine-like RV background in wild-type 129sv suckling mice^a

RV strain origin of VP4 in prototype murine-like RV		Phenotypic differences
		Fecal shedding comparison		Diarrhea comparison
		Prototype murine-like RV (RRV-VP4)	EWwt	Prototype murine-like (RRV-VP4)	EWwt
Simian	RRV	NA	P < 0.05	NA	NS
	SA11	NS	P < 0.05	NS	NS
Bovine	UK	NS	P < 0.05	P < 0.05	P < 0.05
Porcine	OSU	NS	P < 0.05	P < 0.05	P < 0.05
Human	CDC-9	P < 0.01	P < 0.01	P < 0.01	P < 0.01
Murine	ETD	NS	P < 0.05	NS	NS

^aThe areas under the curve for fecal virus shedding were analyzed and compared with those for parental murine-like RV (RRV-VP4) or EW_wt by the nonparametric Mann-Whitney U test. P values are shown. NS, P > 0.5 (not significant); NA, not applicable.

Taking the prototype murine-like RV (RRV-VP4) as the reference, RV replication in the suckling mice was not significantly affected by the presence of VP4s from simian SA11, bovine UK, or porcine OSU RV strains. There was, however, a significant decrease (P < 0.01) in replication when the human CDC-9 VP4 replaced the simian RRV VP4 (Fig. 3G; Table 1). Of note, the homologous (derived from a cell culture-adapted strain of EW_wt) murine ETD VP4 did not enhance virus replication, as no significant difference was observed between the prototype murine-like RVs harboring an RRV VP4 and murine-like RVs harboring a murine ETD VP4 (P > 0.05) (Fig. 3H; Table 1).

In contrast, murine EW_wt was shed at significantly greater levels than any murine-like rRVs studied (Fig. 3B). When WT murine RV was used as the basal comparator for the effect of heterologous or homologous VP4 genes on recombinant murine-like RVs, we noticed that the four cell culture-adapted animal origin VP4s (simian, bovine, murine, and porcine) had a similar shedding pattern: a significant reduction in fecal shedding (P < 0.05) was seen compared to murine EW_wt, suggesting that either EW VP4 is more infectious than the cell culture-adapted ETD VP4 or that the other two heterologous gene segments (VP1 and NSP4 derived from SA11) play a role in restricting RV replication in suckling mice. In order to test the latter hypothesis, we generated an all-murine recombinant RV (rD6/2 ETD-VP4), in which SA11 genes 1 and 10 were also replaced with the corresponding genes from the D6/2 reassortant RV and the RRV gene 4 was replaced with the ETD VP4. We found that our prototype murine-like RV was shed in an identical fashion to rD6/2 ETD-VP4 (see Fig. S1 in the supplemental material), suggesting that VP1 and NSP4 from SA11 did not contribute to growth restriction in the pups. Regardless, human CDC-9 VP4 was clearly associated with a dramatically reduced RV fecal shedding (P < 0.01) compared to the other cell culture-adapted strains (Table 1).

To identify potentially significant but more subtle differences in RV shedding that an ELISA might not readily detect, we also quantified the amount of RV in the feces by a sensitive reverse transcription-quantitative PCR (RT-qPCR) assay (10, 29). RT-qPCR shedding results are presented as a conversion to PFU equivalents per microliter of stool over the 12-day postinoculation period (Fig. 4). The areas under the fecal RV shedding curves during the experimental period were analyzed and compared with those for the murine-like RV (RRV-VP4) and with EW_wt by a nonparametric Mann-Whitney U test; these comparisons are summarized in Table 2.

FIG 4.

Fecal shedding as measured by qRT-PCR of the indicated VP4 monoreassortants on a murine-like RV background in wild-type 129sv suckling mice. Five-day-old 129sv mice were orally gavaged with the indicated RV. Fecal samples were collected daily. Samples over 2 consecutive days were pooled and subjected to RNA extraction and RT-qPCR. Virus shedding within the same group on a 2-day period is shown as the mean ± SD. See Table 2 for statistics analysis.

TABLE 2.

Comparison of fecal shedding measured by qRT-PCR of VP4 monoreassortants on a murine-like RV background in wild-type 129sv suckling mice^a

RV strain origin of VP4 in prototype murine-like RV		Fecal shedding differences
		Prototype murine-like (RRV-VP4)	EWwt
Simian	RRV	NA	P < 0.01
	SA11	NS	P < 0.01
Bovine	UK	NS	P < 0.01
Porcine	OSU	NS	P < 0.01
Human	CDC-9	P < 0.01	P < 0.01
Murine	ETD	NS	P < 0.01

^aThe areas under the curve for fecal virus shedding were analyzed and compared with parental murine-like RV (RRV-VP4) or EW_wt fecal shedding areas by the nonparametric Mann-Whitney U test. P values are shown. NS, not significant (P > 0.5); NA, not applicable.

This analysis confirmed the ELISA shedding findings (Fig. 3): (i) murine EW_wt was shed at significantly higher levels in the suckling mice than any of the tested cell culture-adapted VP4 monoreassortants. (ii) Among the murine-like recombinants with heterologous VP4s, the simian, bovine, and porcine VP4 RV genes had no significant differences in their effect on total RV shedding. (iii) The RV VP4 gene from the human CDC-9 RV strain was associated with significantly reduced fecal shedding in the suckling mouse model (Fig. 4; Table 2).

Animal and human RV VP4s reduce the pathogenesis of murine-like RVs in vivo.

To determine whether the replacement of RRV-VP4 by either a homologous murine VP4 from a cell culture-adapted murine strain or several heterologous VP4s from selected human or other animal RV strains affects murine-like RVs capability to induce diarrhea, the areas under the diarrhea curves over the entire experimental time course were analyzed and compared with our prototype murine-like RV or with EW_wt using a nonparametric Mann-Whitney U test (Fig. 3 and Table 1). We found that replacing VP4s in the murine-like RV background with the simian SA11 VP4 or cell culture-adapted murine ETD VP4 did not affect the efficiency of the prototype murine-like RVs’ ability to produce diarrhea in suckling mice (Fig. 3D and H). However, the substitution of the RRV VP4 with either a bovine (UK) or a porcine (OSU) VP4 significantly decreased the amount of diarrhea (P < 0.05) (Fig. 3E and F). Remarkably, the presence of human CDC-9 VP4 completely abolished the capability of the murine-like RV prototype to induce diarrhea (P < 0.01) (Fig. 3G), consistent with its attenuated in vitro replication phenotype (Fig. 2).

VP4 is involved in RV transmission efficiency in suckling mice.

Along with examining the ability to induce diarrhea and replicate in the intestinal tract, we measured the ability of various murine-like RVs with several distinct VP4 substitutions to spread among mock-inoculated littermates. Mock-inoculated pups were housed in the same cage as RV-infected littermates and monitored daily for diarrhea and stool collection for RV testing. Fecal specimens from mock-infected pups were then assayed for RV by ELISA and RT-qPCR to document horizontal transmission. Transmission occurrence is indicated by double-headed arrows in Fig. 3.

Using RT-qPCR to examine selected, pooled fecal samples, we confirmed that excreted virus from mock-infected mice was the same virus as that present in their virus-inoculated littermates. In suckling mice inoculated with the prototype murine-like RVs (RRV-VP4), i.e., SA11-VP4, OSU-VP4, or ETD-VP4, the monoreassortants transmitted to uninoculated littermates that tested positive by RT-qPCR using specific VP4 primers. No virus was detected, either by ELISA or qPCR, in stools from the mock-infected pups kept in the same cage with pups inoculated with CDC-9-VP4 or UK-VP4 monoreassortants. These findings suggest that horizontal transmission among suckling mice can occur efficiently with RVs harboring some heterologous VP4s but not with the selected human or bovine origin VP4s. Of note, the origin of RV VP4 in this study did not correlate perfectly with viral shedding, as the bovine UK VP4 was associated with lack of transmission but did not show diminished fecal shedding.

VP8* and VP5* domains of VP4 are independently involved in RV pathogenesis and spread.

The in vitro and in vivo trypsin and trypsin-like proteases cleave VP4 to produce two fragments: the N-terminal VP8* and the larger C-terminal VP5*. This cleavage is required to promote efficient RV infection (30). VP8* is responsible for initial attachment of the virus to the surface of target cell, while VP5* is involved in cell membrane penetration (2). We next tried to examine the contribution of the individual N-terminal VP8* and C-terminal VP5* domains of VP4 as determinants of replication, diarrhea, and spread using this murine RV model. We first constructed two chimera VP4 genes between RRV VP4 and human CDC-9 VP4 and attempted to rescue rRVs carrying these two VP4 chimera proteins. However, we could not generate these rRVs after multiple attempts (n > 8).

Since the bovine UK VP4 was also associated with reduced diarrhea and lack of transmission and was more distantly related to murine ETD VP4 (Fig. 1A), we alternatively generated chimera VP4 genes between murine ETD and bovine UK VP4s (Fig. 5A) and rescued two murine-like rRVs harboring chimera VP4s consisting of ETD VP8*-UK VP5* or UK VP8*-ETD VP5* (Fig. 5A). Both viruses were rescuable and replication competent. We then examined virus titers (Fig. 5B), replication kinetics (Fig. 5C), and plaque sizes (Fig. 5D) of the new rRVs and compared them with those of both ETD-VP4 and UK-VP4 monoreassortants in MA104 cells. As shown in Fig. 5B to D, these phenotypes were virtually identical for the RV strains with the same VP8* origin, suggesting that the VP8* domain of VP4 most directly modulates replication and plaque size in vitro.

FIG 5.

Generation and in vitro characterization of recombinant murine-like RVs carrying chimera VP4 proteins. (A) Schematics of the gene 4 RNAs from ETD, UK parental strain ETD*UK-VP4 and UK*ETD-VP4 chimeras. The RNA schematic includes coding positions for VP4 and untranslated regions (UTRs); nucleotide positions are labeled. Nucleotides for amino acids 231 to 248 are common between ETD and UK VP4s sequences and were used as recombination sites to generate VP4 chimeras. (B) The VP4 chimera rRVs were serially passaged 3 times in MA104 cells. Cells were harvested at day 7 postinfection or when complete cytopathic effects were observed; the virus titer was determined by an immunoperoxidase focus-forming assay. (C) Growth kinetics of VP4 chimera rRVs. The viral titers were determined by immunoperoxidase focus-forming assay. (D) Comparison of plaque sizes. Plaques were generated on MA104 monolayers and detected at 7 days postinfection by crystal violet staining. The diameter of at least 50 randomly selected plaques were measured using a bright-field microscope. Mean values and the standard deviations from three (B and C) and two (D) independent experiments performed in duplicate are shown. Statistical significance was evaluated by the ordinary one-way ANOVA and Tukey’s post hoc test. For panel C, statistical analysis of th area under the curve were performed. The asterisks indicate significant differences: *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001.

Next, we examined the replication, diarrhea, and spread phenotypes of both chimeras in suckling mice (Fig. 6). Fecal shedding among VP4 UK-ETD chimeras and UK-VP4 and ETD-VP4 monoreassortants was indistinguishable, with no significant differences identified (Fig. 6A; Table 3). Diarrhea results for VP4 chimeras are shown in Fig. 6B. Taking the UK-VP4 monoreassortant as the reference, the replacement of VP8* in bovine UK-VP4 with ETD VP8* significantly increased diarrhea (Fig. 6B; Table 3). Interestingly, the same result was observed when instead of the UK VP8* domain the UK VP5* domain was replaced with the murine ETD VP5* in UK-VP4 (Fig. 6B; Table 3). We did not observe any significant differences in replication or diarrhea among chimera VP4 rRVs and the ETD-VP4 monoreassortant (Fig. 6; Table 3). The efficiency of horizontal transmission of chimera VP4 rRVs was further evaluated by inoculating two pups with the indicated rRVs and a minimum of six pups with PBS. ETD and UK-VP4 monoreassortants were included as controls. The results, presented as the percentage of efficiency of spread, are shown in Table 4. The ETD-VP4 monoreassortant spread efficiently among the mock-infected littermates, with 100% of noninfected pups developing diarrhea and testing positive for RVs antigen in feces over the 10-day postinoculation period. In contrast, UK-VP4 appeared to completely lack the ability to spread among littermates. Consistent with the diarrhea data, both VP4-chimeric rRVs spread among littermates as efficiently as ETD-VP4 did (Table 4).

FIG 6.

Fecal shedding and diarrhea patterns of recombinant murine-like RVs carrying chimera VP4 proteins. Five-day-old 129sv mice were orally gavaged with the indicated rRV. (A) RV fecal shedding was monitored by ELISA. Virus shedding within the same group on each day is shown as the mean ± SD. (B) Diarrheal development was scored from days 1 to 12 postinfection.

TABLE 3.

Comparison of fecal shedding (by ELISA) and diarrhea development of murine-like RV harboring chimeric VP4s in wild-type 129sv suckling mice^a

RV strain origin of VP4 in prototype murine-like RV	Phenotypic differences
	Fecal shedding		Diarrhea
	ETD-VP4	UK-VP4	ETD-VP4	UK-VP4
ETD VP8*-UK VP5*	NS	NS	NS	P < 0.05
UK VP8*-ETD VP5*	NS	NS	NS	P < 0.05

^aThe areas under the curve for fecal virus shedding and diarrhea were analyzed and compared with those for VP4 parental murine-like RVs ETD-VP4 or UK-VP4 by the nonparametric Mann-Whitney U test. P values are shown. NS, not significant (P > 0.5).

TABLE 4.

Transmission efficiency of chimeric VP4 murine-like RVs^a

RV strain origin of VP4 in the prototype murine-like RV	% Transmission efficiencyb
Bovine UK	0 (0/10)
Murine ETD	100 (10/10)
Murine-bovine ETD VP8*-UK VP5*	100 (6/6)
Murine-bovine UK VP8*-ETD VP5	100 (6/6)

^aTo evaluate the transmission ability of rRVs, 2 5-day-old 129sv mice were orally inoculated with the indicated rRV and cohoused with at least 6 non inoculated litters mattes in the same cage. RV Fecal shedding was monitored by ELISA from days 1 to 12 postinfection. Transmission efficiency is reported as % and was determined based on the presence of RV antigen in fecal samples.

^bFor the values in parentheses, the first number represents the number of noninoculated pups that tested positive for RV antigen, and the second number is the total number of noninoculated littermates.

DISCUSSION

The VP4 protein plays a key role in RV attachment and entry, but its role is less well defined regarding cell and tissue tropism and in terms of viral transmission efficiency (1, 20, 31). Although VP4 has also been identified as a determinant of HRR in several animal model studies (7–9, 13, 14), its actual role and importance as an HRR determinant is still ambiguous. Previously, in the suckling mouse model, we reported that a simian VP4 reduced murine RV infectivity only moderately; however, a reassortant expressing VP4 from a bovine RV strain (UK) severely restricted intestinal replication in the suckling mice (8). Here, using a plasmid-only-based RG system, we were able to engineer a collection of recombinant rRVs in an isogenic murine-like RV genetic background which harbored a collection of either heterologous or homologous VP4s from cell culture-adapted simian, bovine, porcine, human, and murine RVs (Fig. 1). We were then able to systematically study the role and relevance of individual VP4s in defining RV HRR in an in vivo murine model of enteric infection. As a first step, we characterized the selected recombinant VP4 monoreassortants in vitro (Fig. 2). We confirmed results from prior studies showing that the genetic origin of VP4 plays an important role in cell culture replication and plaque size regulation (32, 33).

To directly evaluate the contribution of VP4 to HRR, we focused on in vivo replication, pathogenicity as measured by diarrhea and the ability to spread to uninoculated littermates. To this end, all the VP4 monoressortants were studied in suckling mice with a focus on fecal shedding, diarrhea, and virus transmission within litters (Fig. 3). We found that in a murine-like recombinant RV harboring heterologous VP4s from other species the ability to replicate in the suckling mouse small intestine did not appear to be strictly dependent on the species origin of VP4. However, at least one human origin VP4 was clearly not well adapted to replication in suckling mice. The human CDC-9-VP4 monoreassortant was the only one, among all tested, that showed a significantly lower level of replication in mouse intestine, consistent with its attenuated diarrhea phenotype in vivo. We expected that replacement of VP4 by a homologous culture-adapted murine VP4 would increase virus replication and diarrhea; however, no significant differences were noticed between either fecal shedding or diarrhea in mice infected with prototype murine-like RV with a simian VP4 and ETD-VP4 monoreassortant expressing a VP4 from a cell culture-adapted EW_wt murine RV. In an earlier comparison study (34), using infant and adult mouse models and four murine RV strains (EW included) and their cell culture-adapted counterparts, we showed that wt murine RVs appeared to be equally infectious in suckling mice with regard to the level and duration of fecal shedding and spread to uninfected littermates. However, all the cell culture-adapted murine RVs described appeared to be attenuated when administered to naive pups, and they did not spread as efficiently as the wt counterparts in these early studies (34). Of note, the cell culture-adapted strain of the murine RV EW used in these initial studies was a separate and distinct cell culture derivative from the ETD cell culture-derived version of EDIM used in the present report. Hence, it is difficult to directly compare transmission results from the current study with results from our earlier studies of the EDIM-derived cell culture-adapted EW strain (17). In addition, the earlier studies of EDIM were undertaken using BALB/c mice, while these studies were carried out with the 129sv strain, also making direct comparisons difficult.

The current results reveal that VP4 is involved in regulating RV growth in vitro and in vivo, but they also indicate that VP4 does not exclusively function in a host range-restricting manner. Our findings provide evidence supporting the conclusion that RV HRR is not primarily regulated at the attachment or entry steps of the viral life cycle, although we have not ruled out a host-related regulatory role for VP7 in RV entry. These findings are consistent with and support prior work indicating that the RV NSP1 plays the major role in regulating RV HRR (8, 18, 19).

We also explored the role of VP4 in modulating diarrheal disease. The replacement of a simian RRV VP4 with either a bovine UK-VP4 or a porcine OSU-VP4 was associated with less diarrhea (measured as the area under the curve), and diarrhea was completely abolished when the human CDC-9-VP4 replaced RRV-VP4. As might be expected, a correlation between diarrhea and transmission was seen. If diarrhea was diminished, transmission was either compromised or abolished. For example, the replacement of RRV VP4 by UK VP4 was associated with decreased diarrhea and the lack of transmission to uninfected littermates. The OSU-VP4 monoreassortant also exhibited diminished diarrhea (compared with prototype murine-like RV). In an independent experiment to further characterize the transmission of the OSU-VP4 monoreassortant, we found that three out of eight PBS-inoculated pups tested negative when they were cohoused with two infected pups. In contrast, 10 out of 10 PBS-inoculated pups tested positive in similar conditions when cohoused with the prototype murine RV (data not shown). Of note, the CDC-9-VP4 monoreassortant, which did not induce diarrhea but was shed to some degree, was not able to spread to uninfected littermates in our study (Fig. 3). Consistent with the phylogenetic tree data (Fig. 1A), human (CDC-9) and bovine (UK) VP4 are not closely related to murine VP4, and they both showed a poor ability to induce diarrhea as well as exhibiting reduced ability to spread to uninfected littermates.

In order to further study the relative contributions of the N-terminal VP8* and C-terminal VP5* domains of VP4 to the diarrhea and transmission phenotypes, we generated two additional chimera rRVs harboring either murine VP8*-bovine VP5* or bovine VP8*-murine VP5* on the murine-like RV backbone. Both RV chimeras had comparable in vitro growth kinetics and plaque sizes as the parental RV strain with which they shared the same VP8*. This association suggests that the VP8* domain of VP4 directly regulates these in vitro growth characteristics. Interestingly, both chimeras induced diarrhea and spread to uninfected littermates as efficiently as the murine ETD-VP4 parental strain. This finding suggests that in the suckling mouse model, the genetic origin of VP4 partially dictates in vivo replication, pathogenesis, and spread of infection and that both tryptic fragments VP5* and VP8* can independently modify the in vivo diarrheal and spread phenotypes but that these phenotypes are not strictly linked to the host origin of VP4.

MATERIALS AND METHODS

Cell culture and viruses.

MA104 cells (ATCC CRL-2378) were grown in Medium 199 (Sigma-Aldrich) supplemented with 10% heat-inactivated fetal bovine serum (FBS), 100 IU penicillin/mL, 100 μg/mL streptomycin, and 0.292 mg/mL l-glutamine (complete medium). Vero cells were cultivated in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% FBS, 100 IU penicillin/mL, 100 μg/mL streptomycin, 0.292 mg/mL of l-glutamine. The described BHK-T7 cell line (35) was kindly provided by Ursula Buchholz (Laboratory of Infectious Diseases, NIAID, NIH, USA) and cultured in completed DMEM supplemented with 0.2 μg/mL of G418 (Promega). MA104 N*V cells (36) were cultured in complete medium 199 in the presence of 3 μg/mL of puromycin (InvivoGen, San Diego, CA) and 3 μg/mL of blasticidin (InvivoGen).

The RV strains used in this study include murine EW (G16P[17]) (37, 38), simian RRV (G3P[3]) (39), simian SA11(G3P[2]) (39), human CDC-9 P50 (G1P[8]) (23), bovine UK (G6P[5]) (39), porcine OSU (G5P[7]) (39), and the murine reassortant D6/2. They were propagated in MA104 cells, and prior to infection all RV inocula were activated with 5 μg/mL of trypsin (Gibco Life Technologies, Carlsbad, CA) for 30 min at 37°C.

Plasmids.

The murine D6/2 plasmids pT7-VP2D6/2, pT7-VP3D6/2, pT7-VP4D6/2, pT7-VP6 D6/2, pT7-VP7D6/2, pT7-NSP1D6/2, pT7-NSP2D6/2, pT7-NSP3D6/2, and pT7-NSP5D6/2, as well pT7-VP4UK and pT7-VP4ETD were commercially synthesized (GenScript USA Inc.) (22); pT7-VP1SA11, pT7VP4SA11, and pT7-NSP4SA11 were originally made by Takeshi Kobayashi (Research Institute for Microbial Diseases, Osaka University, Japan) and obtained from Addgene (40). The pT7-VP4CDC-9 plasmid was provided by Baoming Jiang (CDC, Atlanta, USA). The pT7-VP4OSU and pT7-VP4 chimera plasmids were generated in our lab. The C3P3-G1 plasmid (41) was provided by Philippe H Jaïs. The purification of all the plasmids was performed using a Qiagen Plasmid Midiprep or Maxiprep kit per the manufacturer’s instructions.

Plasmid construction.

The gene 4 nucleotide sequence of bovine OSU RV was determined using a self-priming oligonucleotide DNA linker, as described elsewhere (42). Briefly, the C9 anchor primer (5′-p-GACCTCTGAGGATTCTAAAC/iSp9/TCCAGTTTAGAATCC-3′) was attached to the 3′ ends of viral dsRNAs using T4 RNA ligase 2 (NEB), and adaptor-ligated RV dsRNAs were purified by 1% agarose gel electrophoresis. Purified viral cDNAs were synthesized using SuperScript III reverse transcriptase (Invitrogen). Full-length viral cDNAs were amplified with a primer (5′-GAGTTAATTAAGCGGCCGCAGTTTAGAATCCTCAGAGGTC-3′) complementary to the C9 anchor primer using PrimeSTAR HS DNA polymerase (TaKaRa). Amplified viral cDNAs were subcloned into pBluescript KS(+) and sequenced. Finally, full cDNA of OSU gene 4 was amplified using specific VP4 primers and cloned into a pT7 rescue plasmid (40). The complete sequence of this gene corresponded to that of KJ450845.1.

The chimera VP4 plasmids were constructed via NEBuilder HiFi DNA assembly (NEB). For the pT7-VP8*ETD-VP5*UK plasmid, the nucleotide sequence from amino acid (aa) 219 to 775 of ETD-VP4 was interchanged for PCR amplicons with the nucleotide sequence aa 221 to 777 of UK-VP4. The pT7-VP8*UK-VP5*ETD was generated with aa 221 to 777 of UK-VP4 replaced by the nucleotide sequence of aa 219 to 775 of ETD-VP4.

Phylogenetic analysis of RV VP4 and VP7 genes.

Phylogenetic trees of VP4 genes were generated by the neighbor-joining method using MEGA 11 software (43).

Generation of recombinant rotaviruses.

Murine-like RV was generated using the following pT7 plasmids: pT7-VP1SA11, pT7-VP2D6/2, pT7-VP3D6/2, pT7-VP4D6/2, pT7-VP6D6/2, pT7-VP7D6/2, pT7-NSP1D6/2, pT7 NSP2D6/2, pT7-NSP3D6/2, pT7-NSP4SA11, and pT7-NSP5D6/2, according to the optimized, entirely plasmid-based RG system (22). The pT7-VP4D6/2 plasmid was replaced by plasmid carrying the simian SA11 VP4 (pT7-VP4SA11), bovine UK VP4 (pT7-VP4UK), porcine OSU VP4 (pT7-VP4OSU), murine ETD VP4 (pT7-VP4ETD), or human CDC-9 VP4 (pT7-VP4CDC-9) or chimera VP4 plasmids to generate the reassortant viruses. The rescued rRVs were propagated for two passages in MA104 cells in a 6-well plate and then propagated in a T75 flask to produce the virus stock. Murine-like RV harboring CDC-9 VP4 was propagated in MA104 cells in a 12-well plate and then expanded in Vero cells in a 12-well plate for the mouse experiments or dsRNA extraction.

Purification of RV particles by sucrose gradient centrifugation.

Murine-like RV harboring CDC-9 VP4 was concentrated by pelleting through a sucrose cushion as described previously (44). Briefly, Vero cells grown in a 12-well plate were infected and harvested 72 h postinfection, the viral lysates were freeze-thawed three times, and viral particles were concentrated by ultracentrifugation for 1 h at 30,000 rpm at 4°C. Viral pellets were resuspended in TNC buffer (10 mM Tris-HCl [pH 7.5], 140 mM NaCl, 10 mM CaCl₂) and extracted by Genetron, and the aqueous phase was pelleted through a 40% sucrose cushion by centrifugation for 1 h at 30,000 rpm at 4°C. The pelleted RV was resuspended with 500 μL of PBS with Ca²⁺ and Mg²⁺, and this suspension was used to perform mouse infections or to obtain genomic dsRNA.

Polyethylene glycol concentration for rRVs.

rRVs were concentrated by using PEG as reported elsewhere (45). In brief, 5 to 10 mL of clarified rRV supernatant was mixed with PEG 1500 (Sigma-Aldrich) to a final concentration of 10% (wt/vol), stirred for 2 h at room temperature, and then centrifuged at 10,000 × g for 20 min. The pellet was resuspended with 1 mL of TRIzol for viral dsRNA analysis.

Plaque and focus-forming assays.

Cell culture supernatants or virus samples were serially diluted 2-fold or 10-fold and added to monolayers of MA104 cells for 1 h at 37°C. Samples were removed and replaced with 0.1% agarose (SeaKem ME agarose; Lonza) in 199 serum-free medium supplemented with 0.5 μg/mL of trypsin. Cultures for plaque assays were incubated for 7 days at 37°C, then fixed with 10% formaldehyde and stained with 1% crystal violet (Sigma-Aldrich). To measure the sizes of the plaques, we captured pictures of more than 50 plaques with a bright-field microscope in two different experiments for each RV. Then, plaques were automatically identified using a Fiji macro (46), and finally the diameters of the plaques were measured with the annotation tool of the microscope (Advanced Analysis Software BZ-X).

Cultures for focus-forming assays were incubated for 16 to 18h at 37°C, then fixed with 10% paraformaldehyde, permeabilized with 1% Tween 20, and stained with rabbit hyperimmune serum to rotavirus (anti-DLPs) produced in our lab and previously described (47) and anti-rabbit horseradish peroxidase antibody. Viral foci were stained with 3-amino-9-ethylcarbazole (AEC substrate kit; Vector Laboratories) per the manufacturer’s instructions and enumerated visually.

RNA gels.

Viral dsRNA was extracted with TRIzol (Invitrogen) according to the manufacturer’s protocol and then mixed with gel loading dye Purple (6×), no SDS (NEB). Samples were subjected to PAGE (10%) for 18 h at 25 mA and silver stained (48).

Mice, RV infection, and phenotypic characterization.

Wild-type 129sv mice were originally purchased from the Jackson Laboratory and maintained in in-house breeding colonies. Four- to 5-day-old pups were orally inoculated with the indicated rRV (10⁴ PFU), or wt RRV (10⁴ PFU), or EW RV (10⁴ DD₅₀; 1 DD₅₀ is the dose that causes diarrhea in 50% of animals). Litters, consisting of 8 to 13 pups, were divided into two groups as indicated. One group was inoculated with the test virus, and the second was left uninoculated within the same cage to determine whether diarrheal disease spread within the litter. Pup abdomens were gently palpated to check for diarrhea as described elsewhere (14) and scored positive if evidence of unformed or liquid stool was observed. Spread was deemed to have occurred when any uninoculated pup developed diarrhea that was determined to be RV positive by ELISA or RT-qPCR.

RVs from fecal specimens collected from infected pups were amplified on MA104 to analyze dsRNA profiles or tested by qRT-PCR (VP4-specific primers) to evaluate possible cross-contamination among infected litters.

All mice were housed at the Veterinary Medical Unit of the Palo Alto VA Health Care System. All animal studies were approved by the Stanford Institutional Animal Care Committee.

Quantitation of virus shedding.

Fecal specimens were collected on the indicated days postinfection. The amount of virus shedding in the stools was determined by a double-antibody capture ELISA to detect viral antigen (14) and also subjected to a RT-qPCR-based assay measuring RV gene NSP5 levels with standard curves to determine infectious virus particle equivalents per gram of stool sample (10, 29).

Statistical analysis.

The abilities of recombinant or wt RVs to cause diarrhea in inoculated mouse pups (virulence) and to cause disease in inoculates’ littermates (a measure of host range restriction) were scored. All statistical tests were performed as described in the relevant figure legend using GraphPad Prism 9.0 or the IBM SPSS Statistics Grad Pack 26. All experiments, unless otherwise noted, were repeated at least three times.