Mutations in Rotavirus Nonstructural Glycoprotein ... - Journal of Virology

JOURNAL OF VIROLOGY, May 1998, p. 3666–3672 0022-538X/98/$04.0010 Copyright © 1998, American Society for Microbiology

Vol. 72, No. 5

Mutations in Rotavirus Nonstructural Glycoprotein NSP4 Are Associated with Altered Virus Virulence MINGDONG ZHANG,1 CARL Q.-Y. ZENG,1 YANJIE DONG,1 JUDITH M. BALL,1 LINDA J. SAIF,2 ANDREW P. MORRIS,3 AND MARY K. ESTES1* Division of Molecular Virology, Baylor College of Medicine,1 and Department of Pharmacology, Physiology, and Integrative Biology, University of Texas Health Science Center,3 Houston, Texas 77030, and Food Animal Health Research Program, Ohio Agricultural Research and Development Center, Wooster, Ohio 446912 Received 12 May 1997/Accepted 20 January 1998

Rotaviruses are major pathogens causing life-threatening dehydrating gastroenteritis in children and animals. One of the nonstructural proteins, NSP4 (encoded by gene 10), is a transmembrane, endoplasmic reticulum-specific glycoprotein. Recently, our laboratory has shown that NSP4 causes diarrhea in 6- to 10-day-old mice by functioning as an enterotoxin. To confirm the role of NSP4 in rotavirus pathogenesis, we sequenced gene 10 from two pairs of virulent and attenuated porcine rotaviruses, the OSU and Gottfried strains. Comparisons of the NSP4 sequences from these two pairs of rotaviruses suggested that structural changes between amino acids (aa) 131 and 140 are important in pathogenesis. We next expressed the cloned gene 10 from the OSU virulent (OSU-v) and OSU attenuated (OSU-a) viruses by using the baculovirus expression system and compared the biological activities of the purified proteins. NSP4 from OSU-v virus increased intracellular calcium levels over 10-fold in intestinal cells when added exogenously and 6-fold in insect cells when expressed endogenously, whereas NSP4 from OSU-a virus had little effect. NSP4 from OSU-v caused diarrhea in 13 of 23 neonatal mice, while NSP4 from OSU-a caused disease in only 4 of 25 mice (P < 0.01). These results suggest that avirulence is associated with mutations in NSP4. Results from site-directed mutational analyses showed that mutated OSU-v NSP4 with deletion or substitutions in the region of aa 131 to 140 lost its ability to increase intracellular calcium levels and to induce diarrhea in neonatal mice, confirming the importance of amino acid changes from OSU-v NSP4 to OSU-a NSP4 in the alteration of virus virulence.

rotavirus pathogenesis cannot be confirmed by traditional mutation and gene knockout studies. Therefore, we have used an alternative approach to further evaluate and confirm the function of NSP4 in rotavirus pathogenesis. We sequenced NSP4 from two pairs of virulent and tissue culture-attenuated porcine rotaviruses, the OSU and Gottfried strains, and compared the biological properties of NSP4 from the virulent (OSU-v) and tissue culture-attenuated (OSU-a) OSU viruses. We made site-directed mutants in NSP4 from OSU-v virus and compared the biological properties of these mutants with those of wild-type NSP4 from OSU-v virus. The results from this study suggest that mutations in NSP4 are associated with the altered virus virulence of the attenuated rotavirus strain.

Rotaviruses are major pathogens causing life-threatening dehydrating gastroenteritis in children and animals. Rotaviruses are classified as a separate genus within the family Reoviridae. They are large (100 nm in diameter), nonenveloped particles with icosahedral symmetry. Mature viral particles have a triple-layered protein capsid which surrounds the genome of 11 segments of double-stranded RNA (dsRNA). The genome codes for six structural proteins and five nonstructural proteins. One of the nonstructural proteins, NSP4, is a transmembrane, endoplasmic reticulum-specific glycoprotein with pleiotropic functions in viral replication and pathogenesis (15). NSP4 serves as an intracellular receptor for double-layered particles and interacts with viral capsid proteins during virus morphogenesis (1). Recently, NSP4 has been shown to be an enterotoxin, causing diarrhea in mouse pups (4). Increasing evidence indicates that this enterotoxin functions to activate a signal transduction pathway that increases intracellular calcium levels in cells by mobilizing calcium from the endoplasmic reticulum, resulting in chloride secretion (4, 14, 37, 38). Despite extensive studies in different animal models, rotavirus pathogenesis is still not well understood. Malabsorption secondary to the destruction of the enterocytes (20) and alterations in transepithelial fluid balance (10) are among the proposed pathophysiologic mechanisms by which rotaviruses induce diarrhea after virus replication. Because of the lack of a reverse genetics system for rotaviruses, the role of NSP4 in

MATERIALS AND METHODS Cells and viruses. Virulent and tissue culture-attenuated porcine rotaviruses (both OSU and Gottfried strains) were characterized previously (6, 36), and provided by Linda Saif, Ohio State University. Virulent viruses were maintained by serial passage of infected intestinal contents in gnotobiotic pigs and provided as intestinal contents. The OSU-a virus was derived from the OSU-v virus by passaging it 5 times in gnotobiotic pigs, 13 times in primary pig kidney cells, 11 times in MDBK cells, and 39 times in monkey kidney MA104 cells, with multiple plaque isolations during cell culture passage. The Gottfried attenuated virus was derived from the Gottfried virulent virus by passaging it 6 times in PK-15 porcine kidney cells and 39 times in MA104 cells, with multiple plaque isolations during cell culture passage. The 50% diarrhea dose (DD50) was #0.1 focus-forming unit (FFU) for virulent OSU and Gottfried viruses and $106 FFU for the attenuated viruses in 3- to 42-day-old gnotobiotic piglets (19, 32, 34a). The titers of OSU-v and OSU-a viruses were determined in MA104 cells by a focus-forming assay as previously described (9). Spodoptera frugiperda Sf9 cells were grown and maintained in TNM-FH medium with 10% fetal calf serum as previously described (16). The HT-29 clone 19A cells (2) were routinely cultured in Dulbecco’s modified Eagle medium with 4.5 g of glucose per liter, supplemented with 4 mM L-glutamine, penicillin-streptomycin (100 U/ml), gentamicin (5 mg/ml), and 10% fetal calf serum. The HT-29 cells were used at passages 25 to 40.

* Corresponding author. Mailing address: Division of Molecular Virology, Baylor College of Medicine, One Baylor Plaza, Houston, TX 77030. Phone: (713) 798-3585. Fax: (713) 798-3586. E-mail: mestes @bcm.tmc.edu. 3666

VOL. 72, 1998

ASSOCIATION OF NSP4 MUTATIONS WITH ROTAVIRUS VIRULENCE

RT-PCR. Genomic dsRNAs were extracted from piglet intestinal contents (containing virulent OSU or Gottfried virus) and from tissue culture lysates containing the attenuated OSU or Gottfried virus. Briefly, 2 ml of attenuated virus stock or 1 ml of intestinal homogenate was extracted with 1,1,2-trichloro1,2,2-trifluoroethane (Dupont, Wilmington, Del.), and virus particles were pelleted by centrifugation for 1 h at 35,000 rpm at 4°C in a TLS-55 rotor. The virus pellet was suspended in 500 ml of Tris-EDTA buffer (pH 8.0) and digested with proteinase K (1 mg/ml) and 5 mM EDTA (pH 8.0) for 30 min at 37°C. dsRNAs were extracted with phenol-chloroform, ethanol precipitated, and suspended in 10 ml of H2O (MilliQ sterile) with 10% RNasin (Promega, Madison, Wis.). One microliter of dsRNA was used as the template in a reverse transcriptase (RT)mediated PCR (RT-PCR) mixture that included 2 ml of 103 PCR buffer, 4 ml of 5 mM deoxynucleoside triphosphates, 0.5 ml of RNasin, 7 ml of dimethyl sulfoxide, 1 ml of H2O, and 2 ml each of the forward (59-GGCTTTTAAAAGTTCTG TTCCGAG-39) and reverse (59-GGTCACACTAAGACCATTCC-39) primers made with the sequence from the SA11 gene 10. The reaction mixture was heated for 5 min at 95°C, then quenched on dry ice, and thawed. Avian myleloblastosis virus RT (0.5 ml; Life Technologies, Baltimore, Md.) was added to the mixture, and reverse transcription was carried out for 1 h at 42°C. After reverse transcription, 8 ml of 103 PCR buffer, 71 ml of distilled H2O, and 1 ml of Taq polymerase (Perkin-Elmer, Norwalk, Conn.) were added, and gene 10 cDNA was amplified for 40 cycles of denaturation for 1 min at 94°C, annealing for 1.5 min at 42°C, and extension for 2 min at 72°C. Cloning and sequencing of gene 10. The TA vector (Invitrogen, San Diego, Calif.) was used to clone RT-PCR-amplified gene 10 cDNA as suggested by the manufacturer, except that MAXEfficiency DH5a competent cells (Life Technologies) were used for transformation. Gene 10 cDNA was sequenced by dideoxy sequencing with M13 universal and reverse primers and SA11 internal primers. Gene 10 sequences were confirmed by sequencing two additional clones from independent RT-PCR products. Expression and purification of NSP4. Gene 10 cloned in the TA vector was subcloned into the baculovirus transfer vector pFastBac1 (Life Technologies). The sequence of each gene 10 subcloned into pFastBac1 was confirmed by dideoxy sequencing. Recombinant baculoviruses expressing NSP4 were generated as described by the manufacturer, and recombinant virus stocks were plaque purified. NSP4 was purified from Sf9 cells infected with the recombinant baculovirus. Infected cells were harvested and lysed with lysis buffer {10 mM Tris-HCl [pH 8.1], 0.1 mM EDTA, 1% 3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate (CHAPS)}. NSP4 was first semipurified by fast-performance liquid chromatography (FPLC) using a quaternary methylamine anion-exchange column (Waters Chromatography Division, Milford, Mass.) preequilibrated with equilibration buffer (20 mM Tris-HCl [pH 8.1]). The NSP4-rich fractions were pooled for further purification by using an agarose immunoaffinity column onto which rabbit immunoglobulin G (IgG) against SA11 NSP4 (aa 114 to 135) had been immobilized (4, 27). NSP4 was eluted from the column with 0.1 M Tris-HCl buffer at pH 2.8. The eluate was then dialyzed against 50 mM NH4HCO3 and lyophilized. NSP4 was dissolved immediately before use in endotoxin-free (Limulus amebocyte lysate assay; Associates of Cape Cod, Inc., Falmouth, Ma.) phosphate-buffered saline (PBS), and the resulting solutions were tested and found to be endotoxin free. Construction of OSU-v NSP4 mutants. Mutants were made by using overlapping extension PCR as described by Cormack (11). Mutagenesis was performed on OSU-v gene 10 cloned in the pFastBac1 vector, a baculovirus transfer vector. The two primers flanking the region to be mutated were BsaHIF (59-CAAAG AAATGAGGCGTCAACTGG-39) and NdeIR (59-GTCACTTCTGATGGTTC ATATGG-39). These primers contained unique restriction sites. The two primers used to make the deletion mutant were D131-140F (59-TAAACGCATAGCTA TAGATATGTCGAAAG-39) and D131-140R (59-TATCTATAGCTATGCGT TTAAGCAACTCAAC-39). The two primers used to make the substitution mutant were VVP3F (59-GCTGCTAGATCAGTTGACGCTATAG-39) and VVP3R (59-GATCTAGCAGCTAACTTATCATGTATG-39). The two primers used to make the point mutant were P138SF (59-TTAGATCAGTTGACGCTA TAGATATG-39) and P138SR (59-AACTGATCTAACAACTAACTTATC-39). All mutations were confirmed by dideoxy sequencing. Measurement of [Ca21]i. The intracellular calcium concentration ([Ca21]i) in HT-29 cells was measured by calcium imaging using the fluorescent Ca21 indicator fura-2/AM as previously described (14). Calibration of the fura-2 dye fluorescence was carried out by using the ionophore ionomycin as well as EGTA under Ca21-free and Ca21-saturating conditions as described previously (29), and the [Ca21]i was calculated according to the Grynkiewicz equation (21). Six to ten cells from each camera field were chosen for time-dependent analysis of [Ca21]i. The averaged ratio signal obtained from each cell was digitally saved as a log file. The collected values from cells imaged within a single experiment were then averaged to give an experimental observation of 1 (n 5 1). For each experimental condition, three to six experimental observations from different dye loadings were routinely collected (14). In Sf9 cells, [Ca21]i was measured essentially as for HT-29 cells except that fura-2/AM-loaded cells were superfused continuously with Na-HEPES (containing 1 mM Ca21) at room temperature, instead of 37°C, to remove extracellular dye. Sf9 cells grown on coverslips were infected with a recombinant baculovirus expressing NSP4 at a multiplicity of infection (MOI) of 20 for 36 h prior to fura-2/AM loading.

3667

TABLE 1. DD50 of OSU virulent and attenuated viruses in neonatal mice inoculated orally Virus

OSU-v

DD50 OSU-a

DD50

Dose (FFU)

3,300 330 110 37 12 160 500,000 50,000 16,667 5,556 1,852 26,000

No. of mice with diarrhea/total inoculated (%)

10/10 (100) 7/10 (70) 4/10 (40) 1/10 (10) 1/10 (10) 8/8 (100) 8/10 (80) 3/10 (30) 1/10 (10) 0/10 (0)

Diarrhea induction in neonatal mice. Six- to seven-day-old CD1 mice (Charles River Laboratories, Wilmington, Mass.) were given OSU-v or OSU-a virus orally, and the DD50 was determined. Purified NSP4 from either virus was inoculated intraperitoneally into mice. The severity of diarrhea was scored on a scale of 1.0 to 4.0 as previously described (4). All animal studies were done with coded samples. SDS-PAGE and silver staining. Protein expression in virus-infected cell lysates and purified NSP4 was analyzed by polyacrylamide gel electrophoresis (PAGE) on reducing sodium dodecyl sulfate (SDS)–12% polyacrylamide gels as previously described (8). Gels were stained with a silver staining kit (Sigma, St. Louis, Mo.) as described by the manufacturer. ELISAs. A monoclonal antibody (MAb) capture enzyme-linked immunosorbent assay (ELISA) (using MAb 60-F2D4 as the capture antibody) was used to detect VP7 in virus inocula and intestinal homogenates from virus-infected mice as previously described (26). A similar ELISA using MAb 631-7-54 as the capture antibody was used to detect VP6 in intestinal homogenates from virus-infected mice as previously described (26). A sandwich ELISA was developed to detect NSP4 in intestinal homogenates from virus-infected mice, using anti-NSP4 IgG purified from rabbit anti-NSP4 sera as the coating antibody and guinea pig anti-NSP4 as the detecting antibody. Goat anti-guinea pig Ig conjugated with horseradish peroxidase (Sigma) was the conjugate, and the substrate was the TMB Microwell ELISA substrate (Kirkegaard & Perry Laboratories, Gaithersburg, Md.) as previously described (31).

RESULTS Porcine OSU-v and OSU-a rotaviruses differ in pathogenicity in neonatal mice. OSU-v rotavirus induces severe diarrhea in all piglets when inoculated orally (6, 19). After OSU-v rotavirus was serially passaged in tissue culture, the passaged virus became attenuated and induced only slight diarrhea in piglets (6). We first tested if the pathogenicities of OSU-v and OSU-a viruses were different in the neonatal mouse model of rotavirus diarrhea. Six- to seven-day-old CD1 mice were given either 330 or 3,300 FFU of virus orally, and the occurrence of diarrhea was scored. Significantly more mice (8 of 10) given either dose of virulent virus had diarrhea compared with diarrhea in mice (1 of 10) given attenuated virus (P , 0.01). In a separate experiment, we determined the DD50 for OSU-v and OSU-a viruses by inoculating groups of mice with five different doses of virulent or attenuated virus (Table 1). A greater than 160-fold difference in DD50 was observed between the two viruses: 160 FFU for the virulent virus and 26,000 FFU for the attenuated virus. To test if the difference in pathogenicity between OSU-v and OSU-a viruses was due to the difference in replication efficiencies in mice, we inoculated 6- to 7-day-old CD1 mice with 3,300 FFU of each virus, made intestinal homogenates from infected mice at 12, 24, 48, and 72 h postinoculation, and compared virus replication by measuring the amount of VP7 in the homogenates by ELISA. Similar amounts of VP7 were detected

3668

ZHANG ET AL.

FIG. 1. Amino acid sequences of NSP4 from OSU and Gottfried virulent and attenuated porcine rotaviruses. Amino acid residues are shown in a single-letter format. Dashes represent the same amino acid residue as in the top line. Residues in bold indicate differences between NSP4 from attenuated virus and NSP4 from virulent virus in either the OSU or Gottfried strain. OSU-v, NSP4 from OSU virulent virus; OSU-a, NSP4 from OSU attenuated virus; Gott-v, NSP4 from Gottfried virulent virus; Gott-a, NSP4 from Gottfried attenuated virus.

in the intestinal homogenates from either OSU-v or OSU-a virus-infected mice at each time point (data not shown), indicating similar replication efficiencies of the two viruses in mice. Because virulent virus may be less efficient than tissue culture-attenuated virus in growing (forming foci) in vitro, a second approach, in addition to the focus-forming assay, was used to determine the titer of infectious virulent and attenuated virus in each inoculum. The amount of virus in each inoculum was quantitated by measuring the amount of VP7 by ELISA. We then inoculated 6- to 7-day-old CD1 mice with the same amount of OSU-v and OSU-a viruses based on this VP7 ELISA titer. These studies showed that eight of nine mice given virulent virus, but none of the nine mice given attenuated virus, had diarrhea, a statistically significant difference (P , 0.01). These results independently confirm that the differences in pathogenicity between OSU-v and OSU-a viruses seen in mice were not due to inoculation of the mice with an excess of virulent virus because virulent virus may form foci in vitro with low efficiency. We then compared the replication efficiencies of the same amount of OSU-v and OSU-a viruses (based on VP7 titer) in mice as described above. Again, these viruses replicated to the same efficiencies in mice, confirming that the difference in pathogenicity between the two was not due to a difference in their replication efficiencies in mice. Sequencing of NSP4 from OSU and Gottfried virulent and attenuated rotaviruses. Having shown that the porcine OSU-v and OSU-a rotaviruses exhibit differences in pathogenicity in mice, we cloned and sequenced gene 10 from each rotavirus. The amino acid sequences of NSP4 from OSU rotaviruses showed that of 175 amino acid residues, 6 differences, at residues 59, 72, 103, 135, 136, and 138, were found in the OSU-a NSP4 (Fig. 1). Three of the six mutations were clustered between aa 131 and 140, suggesting that this region may be important in virulence. To further test our hypothesis that changes in NSP4 are associated with rotavirus virulence, the 10th genes from another pair of virulent and tissue culture-attenuated porcine (Gottfried strain) rotaviruses were sequenced. As shown in Fig. 1, four amino acid changes, at residues 12, 94, 135, and 138, were found in NSP4 from the Gottfried attenuated virus compared to the sequence of NSP4 from Gottfried virulent

J. VIROL.

FIG. 2. [Ca21]i in Sf9 cells expressing NSP4 and NSP4 mutants. Fura-2/AMloaded cells were superfused continuously with Na-HEPES buffer, and intracellular calcium was measured by ratio imaging. The averaged ratio signal obtained from each cell was digitally saved as a log file. The collected values from cells imaged within a single experiment (6 to 10 cells) were then averaged to give an experimental observation of 1 (n 5 1). WT, cells infected with wild-type baculovirus (n 5 3); OSU-a, cells infected with recombinant baculovirus expressing OSU-a NSP4 (n 5 5); OSU-v, cells infected with recombinant baculovirus expressing OSU-v NSP4 (n 5 3); D131-140, cells infected with recombinant baculovirus expressing the deletion mutant; VVP3, cells infected with recombinant baculovirus expressing the substitution mutant; P138S, cells infected with recombinant baculovirus expressing the point mutant. Standard errors of the means are shown by the bars.

virus. Of interest, the changes at aa 135 (from valine to alanine) and 138 (from proline to serine) were exactly the same as those seen in the OSU NSP4 pairs (Fig. 1). This finding provided additional evidence to support the hypothesis that changes between aa 131 and 140 on NSP4 are important in rotavirus virulence. NSP4 from OSU-v virus increases [Ca21]i in Sf9 cells when expressed endogenously and in HT-29 cells when added exogenously, while NSP4 from OSU-a virus and OSU-v NSP4 mutants lack the ability to mobilize calcium. SA11 NSP4 can increase [Ca21]i in recombinant baculovirus-infected Sf9 cells when NSP4 is expressed endogenously (38). To determine if there was a difference in intracellular calcium mobilization between OSU-v and OSU-a NSP4 expressed endogenously, Sf9 cells were infected with the same MOI of recombinant baculovirus expressing either OSU-v NSP4 or OSU-a NSP4, and [Ca21]i was measured by calcium imaging fluorescence microscopy at 36 h postinfection (Fig. 2). [Ca21]i levels were as follows: for wild-type baculovirus-infected cells, 114.9 6 7.0 nM (n 5 3); for cells expressing OSU-a NSP4, 140.3 6 34.2 nM (n 5 5); and for cells expressing OSU-v NSP4, 628.7 6 151.8 nM (n 5 3). [Ca21]i was slightly higher in cells expressing OSU-a NSP4 than in cells infected with wild-type baculovirus, while expression of OSU-v NSP4 in Sf9 cells increased [Ca21]i approximately sixfold. When expressed endogenously, NSP4 from OSU-v virus increased [Ca21]i significantly more than did NSP4 from OSU-a virus (P , 0.01, Student’s t test). Western blot analysis of the same volume of infected cell lysates made from the same number of cells infected with the same MOI of recombinant baculoviruses showed that the levels of expression of OSU-v and OSU-a NSP4 were similar (data not shown). Therefore, observed differences in calcium mobilization induced by the two NSP4 proteins correlate with differences in the virulence of these viruses. To test the hypothesis that changes between aa 131 and 140 in NSP4 are important in rotavirus virulence, we constructed three mutants of OSU-v NSP4: (i) D131-140, a deletion mutant with aa 131 to 140 deleted; (ii) VVP3, a mutant converting the

VOL. 72, 1998


FIG. 3. Silver staining of purified NSP4 from OSU virulent and attenuated viruses. Five microliters of infected Sf9 cell lysate (6 3 107 cells/ml; lanes L) or 1 mg of purified protein (lanes P) was analyzed by SDS-PAGE (12% gel). M, molecular weight markers. Arrows indicate that two different glycosylated forms of NSP4 are present in the purified material.

three changes in this region of OSU-v NSP4 to the corresponding amino acid sequence of OSU-a NSP4 (VVP to AAS at 135, 136, and 138, respectively); and (iii) P138S, a point mutant changing P to S at position 138. Recombinant baculoviruses expressing these mutants were generated and plaque purified. All three mutants expressed the NSP4 protein to similar levels, as shown by Western blot analysis of the same volume of infected cell lysates made from the same number of cells infected with the same MOI of recombinant baculoviruses (data not shown). We tested the biological activities of the mutants by examining whether they could increase [Ca21]i in insect cells when expressed endogenously. [Ca21]i levels were as follows: for cells expressing deletion mutant D131-140, 101.8 6 20.1 nM (n 5 5); for cells expressing substitution mutant VVP3, 80.1 6 11.3 nM (n 5 5); and for cells expressing point mutant P138S, 78.8 6 8.4 nM (n 5 5). Compared to the wild-type OSU-v NSP4, all three mutants lost the ability to increase intracellular calcium concentrations (Fig. 2). Recently, our laboratory showed that SA11 NSP4 also can increase [Ca21]i in HT-29 cells, a human intestinal epithelial cell line (14). We next sought to determine if OSU-v and OSU-a NSP4 would mobilize intracellular calcium in these

3669

human cells. The NSP4 from each virus was cloned into a baculovirus recombinant, expressed in insect cells, and purified from infected cell lysates using FPLC and immunoaffinity chromatography. Silver-stained gel analysis (Fig. 3) showed the NSP4 from both OSU-v and OSU-a viruses was purified to homogeneity. Two bands representing the two different glycosylated forms of NSP4 were present in the purified material. Purified OSU-v or OSU-a NSP4 was added exogenously to HT-29 cells, and [Ca21]i was measured by calcium imaging fluorescence microscopy (Fig. 4). The basal level of intracellular calcium in HT-29 cells, about 100 nM, was elevated slightly following the addition of 100 nM of NSP4 from OSU-a virus to the cells (Fig. 4A); higher levels (500 nM) of OSU-a NSP4 had a similar effect (data not shown). In contrast, NSP4 (100 nM) from OSU-v virus increased [Ca21]i more than 10-fold above the basal level. The calcium mobilization was transient, lasting for approximately 1 to 2 min (Fig. 4A). Next we purified the point mutant P138S from infected Sf9 cell lysates and tested if purified P138S could increase [Ca21]i in HT-29 cells when added exogenously. Mutant P138S lost the ability to increase [Ca21]i in HT-29 cells when added at 100 nM (Fig. 4B) or 500 nM (data not shown). NSP4 from OSU-v virus was significantly more pathogenic in inducing diarrhea in neonatal mice than NSP4 from OSU-a virus and NSP4 mutant P138S. The pathogenicities of the purified OSU-v and OSU-a NSP4 proteins were compared by inoculating 6- to 7-day-old CD1 mice intraperitoneally in a total volume of 50 ml of endotoxin-free PBS. The occurrence of diarrhea was scored as described previously (4). Thirteen of 23 mice given OSU-v NSP4, but only 4 of 25 mice given OSU-a NSP4, had diarrhea (Table 2), a statistically significant difference (P , 0.01). One of the 25 mice given PBS had minor diarrhea (2.01, the lowest score counted as diarrhea) at one time point. To confirm the difference in pathogenicity between OSU-v and OSU-a NSP4, the DD50s for these proteins were determined. The DD50 for OSU-v NSP4 was 3.2 mg, while the DD50 for OSU-a NSP4 could not be determined because 50% of the mice, even those in the group given the highest dose (10 mg) of OSU-a NSP4, did not develop diarrhea; higher doses of NSP4

FIG. 4. Effects of NSP4 and NSP4 mutant on [Ca21]i in HT-29 cells. Fura-2/AM-loaded cells were superfused continuously with Na-HEPES buffer, and intracellular calcium was measured by ratio imaging. The averaged ratio signal obtained from each cell was digitally saved as a log file. The collected values from cells imaged within a single experiment (6 to 10 cells) were then averaged to give an experimental observation of 1 (n 5 1). Seven experimental observations from different dye loadings were averaged and presented. (A) NSP4 from OSU-a virus, (NSP4-a) or OSU-v virus (NSP4-v) was added as indicated. (B) NSP4 point mutant (P138S) or NSP4 from OSU virulent virus (NSP4-v) was added as indicated.

3670

ZHANG ET AL.

J. VIROL.

TABLE 2. Comparison of pathogenicities of OSU virulent and attenuated NSP4 and mutant P138S NSP4 given intraperitoneally to 6- to 7-day-old CD1 mice Expta

Inoculum

Dose

No. of mice with diarrhea/no. inoculated

Avg scored

DD50 (mg)

1

OSU-v NSP4 OSU-a NSP4 PBS

5 mg 5 mg 50 ml

13/23b 4/25 1/25c

2.8 2.5 2.0

3.2 .10e

2

OSU-v NSP4 P138S P138S PBS

5 mg 5 mg 10 mg 50 ml

7/12b 0/12 0/12 0/12

2.8

a

All experiments were performed with coded samples. P , 0.01, chi-square test. One mouse exhibited minimal diarrhea at one time point. d Average score of disease severity for animals with diarrhea (4). e Unable to be determined because of limitations in NSP4 solubility. b c

could not be given due to limitations in the solubility of NSP4 in PBS. To confirm the importance of amino acid changes in the region from aa 131 to 140 in the observed difference in pathogenicity between OSU-v and OSU-a NSP4, we compared the pathogenicity of purified NSP4 from mutant P138S with that of wild-type OSU-v NSP4. While 7 of 12 mice given 5 mg of OSU-v NSP4 had diarrhea, none of the mice given either 5 mg or 10 mg of mutant P138S had diarrhea (Table 2). This difference in pathogenicity was statistically significant (P , 0.01). This finding indicated that the mutation at proline 138 is associated with changed biological functions of OSU-v NSP4. DISCUSSION Genes encoding several rotavirus proteins, such as VP4 in mice (30) and humans (18) and NSP1 and NSP2 in mice (7), have been associated with pathogenicity in different hosts. In piglets, genes coding for VP3, VP4, VP7, and NSP4 were shown to be associated with diarrhea induction by studying single-gene reassortants between a porcine rotavirus strain which causes diarrhea in piglets (SB-1A) and a human strain which does not cause disease in piglets (DS-1) (23). However, the relationship of these proteins to the diarrheal response remained unknown. The association of NSP4 with rotavirus pathogenesis was first suggested when SA11 NSP4, expressed and purified from insect cells, was shown to induce a dosedependent diarrhea in neonatal mice and rats when NSP4 was administered intraperitoneally or intraileally (4, 3). NSP4 has also been shown to function as an enterotoxin by stimulating chloride secretion through a calcium-dependent signaling pathway (4, 3). By comparing the sequences of NSP4 from two pairs of virulent and tissue culture-attenuated porcine rotaviruses (the OSU and Gottfried strains) with the biological activities of NSP4 from the virulent and attenuated OSU viruses, and by performing site-directed mutational analyses, we now have shown that mutations in NSP4 are associated with altered rotavirus virulence. In our study, we showed that NSP4 from OSU-a virus and mutant P138S NSP4 significantly or completely lost their biological activities to mobilize intracellular calcium in cells and induce diarrhea in neonatal mice, compared to NSP4 from OSU-v virus. The loss of biological activity of NSP4 from attenuated virus or mutant P138S NSP4 cannot be explained by inactivation of these proteins during purification, because

we used the same procedures to purify all three forms of NSP4 (NSP4 from virulent virus, NSP4 from attenuated virus, and NSP4 point mutant P138S). In addition, different lots of purified proteins were tested, and the results were repeatable. The differences in pathogenicity in mice between NSP4 from OSU-v virus and NSP4 from OSU-a virus or point mutant P138S NSP4 also cannot be explained by differences between litters, because the results from multiple experiments, in which multiple litters were inoculated with each of the NSP4s or the control, were repeatable. To consider the biological relevance of our study, it is of interest to compare how much NSP4 is produced during virus infection in vivo and how the amount of NSP4 produced in vivo correlates with the DD50 of NSP4 from OSU-v virus, which was 3.2 mg. However, the validity of comparisons of the amount of NSP4 produced in vivo with the 3.2 mg of NSP4 given intraperitoneally is limited because this comparison assumes that (i) all of the NSP4 produced in the intestine of a mouse from the initiation of infection will be present at the time of sampling (no degradation of NSP4) and (ii) the pharmacological effects of 3.2 mg of an agent given by the intraperitoneal route will mimic the effects of the same amount of agent synthesized intracellularly. To clarify this issue with respect to biological relevance, we performed experiments to directly detect NSP4 in intestinal homogenates. Using an ELISA capable of detecting 39 ng of NSP4, we were unable to detect NSP4 in intestinal homogenates of either SA11 or EDIM virus-infected mice at 12, 24, 48, or 72 h postinoculation. In contrast, an ELISA that could detect approximately 250 ng of VP6, the most abundant viral structural protein, was able to detect VP6 in EDIM, but not in SA11, virus-infected mouse intestinal homogenates, suggesting that active viral replication was required to detect viral proteins by this method. It is possible that we could not detect NSP4 in infected mouse intestinal homogenates because NSP4 is degraded in vivo. Our inability to detect NSP4 in vivo in intestinal homogenates does not necessarily minimize the biological relevance of our results, since the pharmacology of 3.2 mg of NSP4 given by the intraperitoneal route is not known; therefore, the amount of biologically effective NSP4 in the mouse intestine when 3.2 mg of NSP4 is given intraperitoneally is not known. Previously, NSP4 was detected at levels able to cause disease in stool samples from mice infected with SA11 rotavirus (4). When OSU-a virus was given to neonatal mice, diarrhea development was more attenuated compared to disease in mice given virulent OSU virus. This phenomenon was also observed when gnotobiotic pigs were inoculated with virulent and tissue culture-attenuated human rotavirus Wa strains, as lesion development and disease were limited to virulent Wainoculated pigs (40). The lack of disease induction following inoculation of attenuated viruses may be explained by several potential mechanisms. For example, the number of enterocytes infected by attenuated virus may be low and, therefore, only minimal mucosal damage occurs, or the site of replication may vary. Attenuation through tissue culture passage may also select for a virus which is more efficient at replication in the cell culture enzymatic environment than in the intestinal microenvironment (35). This is not the mechanism of attenuation of the porcine rotaviruses, as each attenuated virus replicates in animals (this study and reference 34a). In addition, we inoculated animals with the same amount of virus based on measuring focus forming units or VP7. In both cases, differences in pathogenicity between OSU virulent and attenuated viruses were seen in mice. The results of this study provide a new mechanism to explain the loss of pathogenicity of virus by passage in tissue culture.

VOL. 72, 1998


As viruses are passaged in tissue culture, mutations may be introduced in gene 10. These mutations then lead to the loss of biological properties of NSP4. Given that NSP4 by itself causes disease, a mutated NSP4 could account for the attenuation in virus virulence. However, this does not imply that a mutated NSP4 is the only mechanism for attenuation of a rotavirus; changes in other viral genes may also contribute to virus attenuation. For example, Ward et al. recently reported that attenuation in one human rotavirus vaccine candidate, 89-12, did not correlate with mutations in NSP4 (41). They obtained a consensus sequence of NSP4 from unpassaged and tissue culture-attenuated 89-12 and identified one change, from threonine to alanine at amino acid residue 45, a substitution found in many symptomatic and asymptomatic human rotaviruses. Although analysis of RT-PCR consensus sequences from non-plaque-purified viruses may have missed the finding of mutations in NSP4, attenuation of this particular human rotavirus strain 89-12 likely results from changes in other viral genes. Changes in virus virulence resulting from mutations in viral gene products have been observed in many other viral systems (39). Mutations in the glycoprotein E1 and E2 of Sindbis virus produce a highly attenuated strain (33). A mutation in the acidic polymerase (PA) protein in a cold-adapted influenza virus has been associated with attenuated virulence (13). The mutated PA protein and wild-type PA are known to differ by six amino acid substitutions, and substitution of a hydrophobic residue (valine for methionine) at position 431 is found to be involved in altered virulence. Even a single amino acid mutation can drastically change virus virulence. A single amino acid substitution at position 340 or 419 of the sigma-1 protein in reovirus type 3 (Dearing) can markedly attenuate its neurovirulence (5). Cytotoxicity secondary to B19 parvovirus infection is abolished by a single amino acid mutation in the nucleoside triphosphate-binding domain of B19 nonstructural protein (28). A single amino acid change in the E2 spike protein of a virulent strain of Semliki Forest virus attenuates virus pathogenicity (17). In our study, we found six amino acid substitutions in NSP4 from a pair of OSU viruses and four amino acid changes in NSP4 from a pair of Gottfried viruses. Two changes, at residues 135 (from valine to alanine) and 138 (from proline to serine), were exactly the same for NSP4 from both the OSU and Gottfried pairs of virus. Amino acid residue 135 was also changed from valine in NSP4 from isolates of human rotaviruses from symptomatic children to isoleucine in NSP4 from viruses from asymptomatic children (24), leading the authors to predict an association of this substitution with rotavirus virulence in humans. We propose that the two conserved mutations at 135 and 138 are important for changes in NSP4 function and virus virulence of OSU and Gottfried strains. It is theoretically of interest to look for a potential correlation between the presence of proline 138 in any rotavirus strain and diarrhea induction in mice, but this idea is restricted by the lack of information available on the pathogenicity of the specific virus strains in mice. However, of the available NSP4 sequences for a variety of human and animal rotaviruses, six strains, OSU-v, Gottfried-v, H1, RRV, YM, and Wa, have a proline at 138 in NSP4. OSU-v and RRV can induce diarrhea in mice, and Wa does not induce diarrhea in mice (34), the pathogenicity of the other three strains in mice is not yet clear. Thus, this analysis did not yield a clear answer. Our study of parent-derived virus pairs is a more powerful method of analysis, and our results clearly correlate amino acid changes with altered virus virulence. It is possible that the conformation in this region of NSP4, rather than the presence of a specific

3671

amino acid residue, is the key factor relevant to the function of NSP4 in virus pathogenesis. Mutations in this region may change the overall structure of NSP4 and disrupt interactions of NSP4 with host cell proteins. The importance of these changes was confirmed by our site-directed mutagenesis studies, with the caveat that we did not study a mutant which does not change the biological activities of NSP4. Additional neutral mutants such as P138A, which replaces proline with alanine at amino acid residue 138 in NSP4 from OSU virulent virus, should be useful to confirm the conclusions of our present work and to show whether structural changes are the key to altered biologic properties of NSP4. Analysis of NSP4 sequences in this study with available sequences from other strains indicate that the N-terminal 130 aa of the protein are conserved among all strains, while the Cterminal 45 aa are more variable (12, 22, 25, 42). We propose that the NSP4 sequence in the C-terminal region may coevolve with host cell proteins with which NSP4 interacts. This proposed interaction between NSP4 and host cell proteins may be crucial for the biological activities of NSP4. Elucidation of the mechanisms by which NSP4 functions and how mutations in NSP4 change its biological activities may help us to develop new drugs against rotavirus infection and improve our current strategies to develop an effective rotavirus vaccine. ACKNOWLEDGMENTS This work was supported by Public Health Service grant DK 30144 awarded to M.K.E. and Texas Advanced Technology Program grant 004949-062 to M.K.E. and A.P.M. We thank Sue Crawford for technical assistance. REFERENCES 1. Au, K.-S., W.-K. Chan, J. W. Burns, and M. K. Estes. 1989. Receptor activity of rotavirus nonstructural glycoprotein NS28. J. Virol. 63:4553–4562. 2. Augeron, C., and C. L. Laboisse. 1984. Emergence of permanently differentiated cell clones in a human colonic cancer cell line in culture after treatment with sodium butyrate. Cancer Res. 44:3961–3969. 3. Ball, J. M., and M. K. Estes. Unpublished data. 4. Ball, J. M., T. Peng, C. Q.-Y. Zeng, A. P. Morris, and M. K. Estes. 1996. Age-dependent diarrhea induced by a rotaviral nonstructural glycoprotein. Science 272:101–104. 5. Bassel-Duby, R., A. Jayasuriya, D. Chatterjee, N. Sonenberg, J. V. Maizel, and B. N. Fields. 1985. Sequence of reovirus haemagglutinin predicts a coiled-coil structure. Nature 315:421–423. 6. Bohl, E. H., K. W. Theil, and L. J. Saif. 1984. Isolation and serotyping of porcine rotaviruses and antigenic comparison with other rotaviruses. J. Clin. Microbiol. 19:105–111. 7. Broome, R. L., P. T. Vo, R. L. Ward, H. F. Clark, and H. B. Greenberg. 1993. Murine rotavirus genes encoding outer capsid proteins VP4 and VP7 are not major determinants of host range restriction and virulence. J. Virol. 67:2448– 2455. 8. Burns, J. W., H. B. Greenberg, R. D. Shaw, and M. K. Estes. 1988. Functional and topographical analysis of epitopes on the hemagglutinin (VP4) of the simian rotavirus SA11. J. Virol. 62:2164–2172. 9. Ciarlet, M., M. Hidalgo, M. Gorziglia, and F. Liprandi. 1994. Characterization of neutralization epitopes on the VP7 surface protein of serotype G11 porcine rotaviruses. J. Gen. Virol. 75:1867–1873. 10. Collins, J., W. G. Starkey, T. S. Wallis, G. J. Clarke, K. J. Worton, A. J. Spencer, S. J. Haddon, M. P. Osborne, D. C. Candy, and J. Stephen. 1988. Intestinal enzyme profiles in normal and rotavirus-infected mice. J. Pediatr. Gastroenterol. Nutr. 7:264–272. 11. Cormack, B. 1991. Mutagenesis of cloned DNA, p. 8.5.7. In F. M. Ausubel et al. (ed.), Current protocols in molecular biology. John Wiley & Sons, Inc., New York, N.Y. 12. Cunliffe, N. A., P. A. Woods, J. P. Leite, B. K. Das, M. Ramachandran, M. K. Bhan, C. A. Hart, R. I. Glass, and J. R. Genstch. 1997. Sequence analysis of NSP4 gene of human rotavirus allows classification into two main genetic groups. J. Med. Virol. 53:41–50. 13. Donabedian, A. M., D. C. DeBorde, S. Cook, C. W. Smitka, and H. F. Maassab. 1988. A mutation in the PA protein gene of cold-adapted B/Ann Arbor/1/66 influenza virus associated with reversion of temperature sensitivity and attenuated virulence. Virology 163:444–451. 14. Dong, Y.-J., C. Q.-Y. Zeng, J. M. Ball, M. K. Estes, and A. P. Morris. 1997.

3672

15. 16.

17. 18.

19. 20. 21. 22. 23.

24.

25. 26. 27.

ZHANG ET AL.

The rotavirus enterotoxin NSP4 mobilizes intracellular calcium in human intestinal cells by stimulating phospholipase C mediated inositol 1,4,5triphosphate production. Proc. Natl. Acad. Sci. USA. 94:3960–3965. Estes, M. K. 1996. Rotavirus and their replication, p. 1625–1655. In B. N. Fields, D. M. Knipe, P. M. Howley, et al. (ed.), Fields virology. LippincottRaven Publishers, Philadelphia, Pa. Estes, M. K., S. E. Crawford, M. E. Penaranda, B. L. Petrie, J. W. Burns, W.-K. Chan, B. Ericson, G. E. Smith, and M. D. Summers. 1987. Synthesis and immunogenicity of the rotavirus major capsid antigen using a baculovirus expression system. J. Virol. 61:1488–1494. Glasgow, G. M., H. M. Killen, P. Liljestrom, B. J. Sheahan, and G. J. Atkins. 1994. A single amino acid change in the E2 spike protein of a virulent strain of Semliki Forest virus attenuates pathogenicity. J. Gen. Virol. 75:663–668. Gorziglia, M., K. Green, K. Nishkawa, K. Taniguchi, R. Jones, A. Z. Kapikian, and R. M. Chanock. 1988. Sequence of the fourth gene of human rotaviruses recovered from asymptomatic or symptomatic infections. J. Virol. 64:2978–2984. Graham, D. Y., G. R. Dufour, and M. K. Estes. 1987. Minimal infective dose of rotavirus. Arch. Virol. 92:261–271. Graham, D. Y., J. W. Sackman, and M. K. Estes. 1984. Pathogenesis of rotavirus-induced diarrhea: preliminary studies in miniature swine piglet. Digest. Dis. Sci. 29:1028–1035. Grynkiewicz, G., M. Poenie, and R. Y. Tsien. 1985. A new generation of Ca21 indicators with greatly improved fluorescence properties. J. Biol. Chem. 260:3440–3450. Horie, Y., O. Masamune, and O. Nakagomi. 1997. Three major alleles of rotavirus NSP4 proteins identified by sequence analysis. J. Gen. Virol. 78: 2341–2346. Hoshino, Y., L. J. Saif, S. Y. Kang, M. M. Sereno, W. K. Chen, and A. Z. Kapikian. 1995. Identification of group A rotavirus genes associated with virulence of a porcine rotavirus and host range restriction of a human rotavirus in the gnotobiotic piglet model. Virology 209:274–280. Kirkwood, C. D., B. S. Coulson, and R. F. Bishop. 1996. G3P2 rotaviruses causing diarrheal disease in neonates differ in VP4, VP7 and NSP4 sequence from G3P2 strains causing asymptomatic neonatal infection. Arch. Virol. 141:1661–1676. Kirkwood, C. D., and E. A. Palombo. 1997. Genetic characterization of the rotavirus nonstructural protein, NSP4. Virology 236:258–265. Matson, D. O., M. K. Estes, J. W. Burns, H. B. Greenberg, K. Taniguchi, and S. Urasawa. 1990. Serotype variation of human group A rotaviruses in two regions of the USA. J. Infect. Dis. 162:605–614. Mckinney, M. M., and A. Parkinson. 1987. A simple, non-chromatographic procedure to purify immunoglobulins from serum and ascites fluid. J. Immunol. Methods 96:271–278.

J. VIROL. 28. Momoeda, M., S. Wong, M. Kawase, N. S. Young, and S. Kajigaya. 1994. A putative nucleoside triphosphate-binding domain in the nonstructural protein of B19 parvovirus is required for cytotoxicity. J. Virol. 68:8443–8446. 29. Morris, A. P., K. L. Kirk, and R. A. Frizzell. 1990. Simultaneous analysis of cell Ca21 and Ca21-stimulated chloride conductance in colonic epithelial cells (HT-29). Cell Regul. 1:953–963. 30. Offit, P. A., G. Blavat, H. B. Greenberg, and H. F. Clark. 1986. Molecular basis for rotavirus virulence: role of rotavirus gene segment 4. J. Virol. 57:46–49. 31. O’Neal, C. M., S. E. Crawford, M. K. Estes, and M. E. Conner. 1997. Rotavirus virus-like particles administered mucosally induce protective immunity. J. Virol. 71:8707–8717. 32. Payment, P., and E. Morin. 1990. Minimal infective dose of the OSU strain of porcine rotavirus. Arch. Virol. 112:277–282. 33. Polo, J. M., and R. E. Johnson. 1990. Attenuating mutations in glycoproteins E1 and E2 of Sindbis virus produce a highly attenuated strain when combined in vitro. J. Virol. 64:4438–4444. 34. Ramig, R. F. 1988. The effects of host age, virus dose, and virus strain on heterologous rotavirus infection of suckling mice. Microb. Pathog. 4:189– 202. 34a.Saif, L. J. Personal communication. 35. Steel, R. B., and A. Torres-Medina. 1984. Effects of environmental and dietary factors on human rotavirus infection in gnotobiotic piglets. Infect. Immun. 43:906–911. 36. Theil, K. W., E. H. Bohl, and A. G. Agnes. 1977. Cell culture propagation of porcine rotavirus (reovirus-like agent). Am. J. Vet. Res. 38:1765–1768. 37. Tian, P., M. K. Estes, Y. Hu, J. M. Ball, C. Q.-Y. Zeng, and W. P. Schilling. 1995. The rotaviral nonstructural glycoprotein NSP4 mobilizes Ca21 from the endoplasmic reticulum. J. Virol. 69:5763–5772. 38. Tian, P., Y. Hu, W. P. Schilling, D. A. Lindsay, J. Eiden, and M. K. Estes. 1994. The nonstructural glycoprotein of rotavirus affects intracellular calcium levels. J. Virol. 68:251–257. 39. Tyler, K. L., and B. N. Fields. 1990. Pathogenesis of viral infection, p. 191–239. In B. N. Fields and D. M. Knipe (ed.), Fields virology, 2nd ed. Raven Press, New York, N.Y. 40. Ward, L. A., B. I. Rosen, L. Yuan, and L. J. Saif. 1996. Pathogenesis of an attenuated and a virulent strain of group A human rotavirus in neonatal gnotobiotic pigs. J. Gen. Virol. 77:1431–1441. 41. Ward, R. L., B. B. Mason, D. I. Bernstein, D. S. Sander, V. E. Smith, G. A. Zandle, and R. S. Rappaport. 1997. Attenuation of a human rotavirus vaccine candidate did not correlate with mutations in the NSP4 protein gene. J. Virol. 71:6267–6270. 42. Zhang, M.-D., and M. K. Estes. Unpublished data.