Enzyme-Linked Immunosorbent Assay Based on Recombinant Human Group C Rotavirus Inner Capsid Protein (VP6) To Detect Human Group C Rotaviruses in Fecal Samples

Abstract

A recent study showed that 43% of a population in the United Kingdom were seropositive for group C rotavirus. The higher than expected incidence may be due to limited diagnosis of acute human group C rotavirus infections because no routine test is available. Human group C rotavirus infections are routinely diagnosed by electron microscopy (EM) and a negative group A rotavirus enzyme-linked immunosorbent assay (ELISA) result. An antigen-detection ELISA was developed with hyperimmune antibodies raised to human group C rotavirus recombinant VP6 (Bristol strain) expressed in insect cells. The assay was used to screen fecal samples to determine the prevalence of group C rotavirus infection. Samples positive by ELISA were confirmed by EM, polyacrylamide gel electrophoresis of double-stranded RNA, or detection of the VP6 gene by reverse transcription-PCR. Retrospective analysis indicated a 1 to 2% detection rate of positivity among samples from patients with acute diarrhea.

Rotaviruses, one of nine genera belonging to the family Reoviridae, were first identified as an important cause of gastroenteritis in 1973 (2). Intact virions are 70 nm in diameter and possess a double-layered protein capsid surrounding a core containing 11 segments of double-stranded RNA. Each double-stranded RNA segment codes for a single viral protein ranging in size from 20 to 125 kDa (5).

Rotaviruses are subdivided into seven serogroups (serogroups A to G) on the basis of their antigenic and genetic properties (14). The members of each serogroup share a common group antigen located on major inner capsid protein VP6. Only rotaviruses in groups A to C have been associated with disease in humans, and group A rotaviruses are the major cause of gastroenteritis in children worldwide (3). A recent serologic study has shown that antibodies to human group C rotavirus are present in 43% of a geographically limited population in the United Kingdom (9).

Group A rotaviruses are routinely detected by enzyme-linked immunosorbent assay (ELISA), which identifies the group-specific antigen. The lack of a standardized diagnostic assay for the detection of group C rotaviruses has meant that no prevalence figures are available in the United Kingdom. A recent study in Japan by the reverse passive hemagglutination assay (RPHA) showed incidences of 0 to 13% in 10 regions (15). Diagnosis in other countries is dependent on electron microscopy (EM), which is available in only a limited number of laboratories. A prototype ELISA for the detection of group C rotaviruses in fecal samples has been described (12, 21); however, in that ELISA hyperimmune sera against the porcine (Cowden) strain was used, and recent studies have shown that the Cowden strain is distinct from the human group C rotaviruses. An ELISA with monoclonal antibodies raised to group C rotaviruses from a human fecal sample has also been described (6). The monoclonal antibodies reacted with the outer capsid protein. The genes corresponding to outer capsid protein VP7 of human strains that have been sequenced are highly conserved (97.8 to 99.8 and 97.1 to 100% nucleotide sequence identities in two studies [7, 11], respectively). Some minor differences have been demonstrated between the electrophoretic profiles of human group C rotaviruses (6, 16); however, these strains retain 95.7% nucleotide sequence identity (16), indicating that they belong to the same genotype and are likely to belong to the same serotype. Evidence of serologic and genetic diversity within group C rotaviruses infecting animals has been described (13, 22), and thus, diagnosis relying on the outer capsid proteins in which this diversity has been demonstrated may limit detection to a single serotype.

Recently, human group C rotaviruses have been grown in cell culture (9, 20). Culture required 4 μg of trypsin per ml, and virus yields were poor, so this was not a useful means of producing antigen. Our objective was to develop an antigen-capture ELISA for the specific detection of human group C rotaviruses in fecal samples with recombinant VP6 derived from human group C rotavirus (9). By using recombinant VP6 as an immunogen, all group C rotaviruses, regardless of serotype, should be detected.

MATERIALS AND METHODS

Viruses.

Twenty-three fecal samples containing various enteric viruses (identified by EM, polyacrylamide gel electrophoresis [PAGE] of double-stranded RNA [dsRNA], or ELISA) were supplied by the Public Health Laboratories in Bristol and Southampton, United Kingdom. Five hundred sixty samples for screening by the human group C rotavirus ELISA were collected by the Bristol and Southampton Public Health Laboratories between March 1994 and April 1995 from patients with gastroenteritis negative for group A rotaviruses, salmonella, shigella, and campylobacter. All samples were stored at 4°C.

Antibody production.

Recombinant VP6 was produced in insect cells (9) and was used to immunize laying hens and pathogen-free rabbits prescreened by immunoblotting for evidence of prior infection with group C rotavirus. Immunization was with 100 μg of VP6 in complete Freund’s adjuvant, followed by three boosts with VP6 in incomplete Freund’s adjuvant at approximately 10-day intervals. Eggs were collected daily, and immunoglobulin Y (IgY) was extracted by a commercial method (Eggstract; Promega, Southampton, United Kingdom). Rabbits were bled prior to immunization and after the third and fourth immunizations.

Immunoblot analysis.

Recombinant VP6 was resolved by sodium dodecyl sulfate (SDS)-PAGE and was electroblotted onto nitrocellulose (0.8 mA/cm² for 1 h with a Trans-blot semidry blotting apparatus [Bio-Rad, Hemel Hempstead, United Kingdom]) with transfer buffer containing SDS (48 mM Tris, 39 mM glycine, 20% methanol, 1.3 mM SDS). The membrane was cut into strips for immunostaining, blocked with 5% nonfat dried milk in Tris-buffered saline (20 mM Tris, 500 mM NaCl [pH 7.5]), and reacted with IgY or rabbit serum at a range of dilutions. Bound antibodies were detected with anti-species alkaline phosphatase-conjugated antibodies and Bio-Rad alkaline phosphatase stain (Bio-Rad).

Antigen-detection ELISA.

Polyclonal rabbit antibodies at an optimum dilution (1:10,000) in coupling buffer (15 mM Na₂CO₃, 35 mM NaHCO₃ [pH 9.6]), pre- or postimmunization, were immobilized on alternate columns of wells of polyvinyl microtiter plates (ICN Flow, Irvine, United Kingdom) by overnight incubation at 37°C. Coating antibodies were aspirated, and the plates were blocked by incubation at 37°C for 1 h with phosphate-buffered saline (PBS) containing 5% nonfat milk. Unbound material was removed by washing (four times) with PBS containing 0.05% Tween 20 (PBST). Suspensions of fecal samples (1 to 10% [vol/vol] in PBS) were centrifuged (10,000 × g for 2 min) to remove particulate material. The supernatants were incubated with antibody-coated wells for 2 h at 37°C. After washing with PBST, the detector IgY antibodies were bound to immobilized antigens by incubation for 1 h at 37°C. Following a further wash with PBST, bound IgY antibodies were detected with goat anti-chicken horseradish peroxidase enzyme conjugate (incubation at 37°C for 1 h). Substrate, 3,3′,5,5′-tetramethylbenzidine (0.1 M sodium acetate, 1% 3,3′,5,5′-tetramethylbenzidine [3.6 mg/ml {wt/vol} in dimethyl sulfoxide, 0.01% H₂O₂), was added after washing with PBST, and after 5 min the reaction was stopped with 2 M H₂SO₄. The colorimetric reaction was measured by recording the A₄₅₀ with an automated spectrophotometer (Anthos II; Anthos Labtech Instruments, Salzburg, Austria).

cDNA synthesis.

Viral RNA was extracted as described previously (18) with RNAzolB and purification of RNA in the aqueous phase by binding to silica particles in the presence of NaI (Geneclean II). Purified dsRNA was denatured by heating it to 90°C for 5 min in 7% (vol/vol) dimethyl sulfoxide in the presence of 500 ng of random primers prior to cDNA synthesis in a 50-μl reaction mixture containing 50 mM Tris-HCl (pH 8.3), 75 mM KCl, 3 mM MgCl₂, 10 mM dithiothreitol, 100 μg of bovine serum albumin per ml, 40 U of RNAsin (Promega, Southampton, United Kingdom), and 500 U of Superscript (Life Technologies, Paisley, United Kingdom). The reaction mixture was incubated at 42°C for 1 h, and the reaction was stopped by the addition of 0.01 M EDTA. Genomic RNA was hydrolyzed by the addition of 0.1 M NaOH and heating to 65°C for 60 min, and the reaction was neutralized by the addition of 0.1 M HCl and 0.1 M Tris-HCl (pH 7.5). The cDNA was purified with spin columns (Promega) containing a Sephacryl S-400 gel matrix (Sigma, Poole, United Kingdom) and was stored at −20°C.

PCR.

A nested PCR was designed with four oligonucleotide primers based on the human group C rotavirus gene 5 (VP6) (4). Outer primer pairs 5′-¹⁷⁵GGAACACAGCCTCAGAAAG¹⁹⁴-3′ and 5′-¹⁰¹⁷GGACTCTGTGGTAGCATC¹⁰⁰⁰-3′ formed an 843-bp product, and inner primer pair 5′-⁴⁴⁰ATGCTCAATCAAGACGTGAG⁴⁶⁰-3′ and 5′-⁹¹³CAGCTGGTCTAATCATGTC⁸⁹⁵-3′ formed a 474-bp product. Oligonucleotides were synthesized on an Expedite automated synthesizer (Millipore, Watford, United Kingdom) by β-cyanoethyl phosphoramidite chemistry. cDNA samples were amplified in a GeneAmp 9600 (Perkin-Elmer Cetus) thermal cycler. Amplification conditions comprised 30 cycles of denaturation at 94°C for 20 s, primer annealing at 50°C for 20 s, and DNA polymerization at 72°C for 1 min with the external primers and 30 s with the internal primers. A typical 25-μl PCR mixture comprised 50 mM KCl, 10 mM Tris-HCl (pH 8.8), 1.5 mM MgCl₂, and 0.1% (vol/vol) Triton X-100 (all supplied as a 10× buffer), each deoxynucleotide triphosphate at a concentration of 0.2 mM, 125 ng of each primer, 2.5 μl of template DNA, and 0.625 U of Taq polymerase (Promega). Aliquots (2.5 μl) of the first-round product were amplified with the internal primer pair. The PCR products were resolved on a 2% agarose gel and were stained with ethidium bromide.

RESULTS

ELISA design, specificity, and sensitivity.

The rabbits and hens were shown by immunoblotting to have no preexisting antibodies to recombinant group C rotavirus VP6. Anti-group C rotavirus VP6 antibodies were detected by immunoblotting postimmunization rabbit serum and IgY at a dilution of 1:100,000 (data not shown). The rabbit sera, IgY, and recombinant VP6 were used to develop an antigen-trap ELISA to detect human group C rotaviruses. Twenty-three fecal samples containing enteric viruses, as identified previously by EM, from patients with acute gastroenteritis were tested to determine the criteria for positivity. The panel consisted of 6 samples containing small round-structured viruses (SRSVs), 1 sample containing enteric adenoviruses, 11 samples containing human group A rotaviruses (confirmed by IDEIA Rotavirus; Dako, Ely, United Kingdom), and 5 samples containing human group C rotavirus (as confirmed by PAGE).

In the ELISA, the five group C rotavirus-positive samples gave A₄₅₀ values greater than 0.38 and postimmune capture antibody to preimmune capture antibody ratios (P:N) of greater than 3.1. The non-group C rotavirus samples all gave A₄₅₀ values of 0.18 or less (mean, 0.09; standard deviation [SD], 0.03], and P:N ratios of less than 1.25 (mean, 1.0; SD, 0.14). Samples were considered positive for group C rotavirus if the A₄₅₀ was greater than 0.18 (mean for the non-group C rotavirus samples + 3 SDs), and the P:N ratio was greater than 1.4 (mean + 3 SDs) (Fig. 1).

FIG. 1.

ELISA of fecal samples containing identified viruses. Plot of the A₄₅₀ with hyperimmune rabbit antisera against the P:N ratios for fecal samples containing various viruses identified by EM. The cutoff was derived from the mean values for the non-group C rotavirus-containing fecal samples +3 SDs, as indicated. ⧫, non-group C rotavirus-containing fecal samples; ■, samples confirmed to contain human group C rotavirus.

To determine the sensitivity of the ELISA the number of group C rotavirus particles in a fecal sample (detected by EM and PAGE of dsRNA) was estimated by EM by comparison to the number of polystyrene latex beads present at a known concentration. A dilution series of the fecal sample was made and was tested by ELISA, and the endpoint was determined to be 1:2,000. The limit of detection was estimated to be 3.3 × 10⁵ particles per ml.

Screening of the sample population.

The ELISA was used to screen 560 fecal samples that were negative for group A rotaviruses and bacterial pathogens and that had been collected over a 14-month period. Positive and borderline samples (fulfilling only one of the criteria for positivity) were retested by ELISA and were examined by EM and PAGE. Rotavirus particles were identified by EM in three samples. RNA profiles characteristic of group C rotaviruses were obtained from two of these samples (A₄₅₀s, 0.39 and 0.58, respectively; P:N ratios, 2.34 and 2.29, respectively). These profiles had genome pattern I (6, 15), similar to that of the human Bristol strain. No RNA profile could be identified by PAGE in the third sample (group A rotavirus ELISA negative). SRSVs were identified by EM in a further sample.

RT-PCR of fecal samples.

ELISA results were confirmed by reverse transcription-PCR (RT-PCR) developed with primers designed specifically from the published human VP6 sequences (4). The samples tested were chosen for further investigation according to the ELISA results and a selection of group C rotavirus ELISA-negative samples containing other enteric viruses identified by EM. All the ELISA-negative samples and the samples already shown to contain other viruses were RT-PCR negative. Nine samples were positive by RT-PCR, including one sample with a high A₄₅₀ (0.489) and a P:N ratio of 1.32 (0.08 below the cutoff for positivity). The mean A₄₅₀ for all group C rotavirus ELISA-negative samples was 0.099 (SD, 0.053), and the mean P:N ratio was 1.057 (SD, 0.185).

Clinical details.

Of the 560 samples examined, 9 (1.6%) were positive by the human group C rotavirus ELISA and were confirmed to be positive by RT-PCR. Only three were positive by EM, and two of these were positive by PAGE. The patients, both adults and children, had diarrhea ranging in duration from 1 day to 3 weeks. Other symptoms were noted in the five children ages 1 and 2 years and included vomiting, fever, convulsions, and respiratory symptoms (Table 1).

TABLE 1.

Details about patients ELISA positive for group C rotavirus

Age (yr)	Date of sample retrieval (mo/yr)	Result by the following:
		EM	PAGE	ELISA	RT-PCR
1	4/1994	−	−	+	+
1	5/1994	−	−	+	+
1	12/1994	−	−	+	+
1	2/1995	+	−	+	+
2	7/1994	+	+	+	+
4	5/1995	+	+	+	+
25	2/1995	−	−	+	+
27	11/1994	−	−	+	+
90	4/1994	−	−	±	+

DISCUSSION

To minimize nonspecificity in an ELISA, antibodies from two different animal species are generally used: one for virus capture and the second to act as a detector antibody. Rabbits were chosen as the source of capture antibodies because they have traditionally been used for the production of polyclonal antibodies, whereas hens were chosen to produce the detector antibodies. Immunization of laying hens produces high yields of polyclonal antibody in the yolk of the egg. M. Haak-Frendscho (8) described an extraction method, using proprietary polyethylene glycol solutions, that resulted in similar quantities of immunoglobulin from one egg to that from a single bleed (10 ml) from a rabbit. No cross-reaction occurs between chicken antibodies and mammalian IgG (1), and IgY does not bind to bacterial or mammalian Fc receptors (10) or interact with rheumatoid factor. Hens were therefore chosen because the use of IgY would reduce the risk of false-positive reactions caused by these factors.

This is the first description of an ELISA for the detection of the group-specific antigen of group C rotaviruses based on reagents derived from human group C rotavirus. The assay was shown to be at least 10-fold more sensitive than EM. A previous ELISA used to detect group C rotaviruses in human fecal samples relied on reagents derived from the Cowden porcine group C rotavirus (12). However, the Cowden-based ELISA failed to detect 81% of EM-positive samples, indicating a lower sensitivity of assays based on reagents made from viruses infecting a different animal species. In the present study of nine samples positive by ELISA (confirmed by RT-PCR), only two were confirmed to be positive by EM and electropherotyping. The limited detection by EM suggests that group C rotaviruses can be present at low concentrations in feces. However, the storage of the samples in this study might have resulted in the loss of intact virions, reducing the possibility of detection by EM.

A recent study in Japan (15) found that 6.8% of diarrheal samples collected over a 6-month period were positive by RPHA for group C rotaviruses. The incidence was higher than that found in the study presented here and could represent a large-scale outbreak in the winter of 1992 and 1993 in Japan. In their original paper (17), it was shown that the RPHA detects 1.5 × 10⁷ particles/ml. In our study, which could detect group C rotavirus particles at 3.3 × 10⁵ virions/ml, only one-third of the positive samples were positive by EM or PAGE; however, all were confirmed to be positive by RT-PCR. Kuzuya et al. (17) were unable to confirm the positivity of 60% their samples, suggesting that their relatively high positivity rate may be a result of nonspecific reactions.

In the initial experiment for determination of the criteria for positivity, a high A₄₅₀ value was observed with one of the group A rotavirus-positive samples (Fig. 1). This suggests some cross-reactivity of the hyperimmune sera with group A rotaviruses. One fecal sample containing SRSVs was also positive by the group C rotavirus ELISA but negative by the RT-PCR. An internal standard was not included in the PCR, so the possibility that RT-PCR inhibitors were present in the sample cannot be ruled out. Assuming that the RT-PCR is the “gold standard” for detection of group C rotaviruses, the specificity of the ELISA could be improved by redefining the cutoff criteria, although this may result in a loss of sensitivity. The combination of the ELISA as a rapid screening assay with confirmation by RT-PCR proved to be effective. The specificity of the ELISA could be improved in future by using monoclonal antibodies to the inner capsid protein.

The rate of detection of human group C rotaviruses in fecal samples remains low compared to that achieved in our serological study (9). In a recent serological study in Western New York, one-third of young adults had evidence of previous infection with group C rotaviruses (19). This level is remarkably similar to the level (35%) detected in the group ranging in age from 16 to 35 years in our local serologic study in the United Kingdom. Because human group C rotaviruses appear to belong to a single genotype, it is likely that repeated infections, as seen with group A rotaviruses, do not occur. Although it should be noted that five of the positive samples were from hospitalized patients, another explanation for the low detection rate is that infection causes mild symptoms that are not routinely investigated or that virus shedding is transient and samples have been collected too late after the acute onset of infection.

The ELISA results indicate that human group C rotavirus infections are more commonly detected by ELISA than by EM and PAGE of dsRNA, which until now have been the only methods routinely used for their detection in the United Kingdom.