LETTER

Top Letter References

Mumps immunization has been effective at controlling epidemic mumps infection, but a high rate of vaccine-associated disease for Urabe AM9 has been documented (6). Since the mumps virus (MuV) hemmagglutinin-neuraminidase (HN) antigen is the major surface protein and an important candidate with respect to vaccine failure, we compared parts of the nucleotide sequences of the HN genes of clinical isolates associated with the disease to determine whether sequence differences could explain the manifestation of increased virulence. Since studies of neutralization escape mutants of MuV have shown the putative region for the epitope of selected neutralizing monoclonal antibodies to be in the region encompassing amino acids 350 to 500 of HN antigen (7, 9), we have examined the HN genes (nucleotides [nt] 900 to 1827) of the wild strains of MuV detected in 10 samples drawn from five patients with aseptic meningitis and five patients with parotitis. The HN region of the genome was amplified directly from the sample (cerebrospinal fluid or pharyngeal swab). Virus RNA was purified from samples by using a previously described method (5), and it was subjected to reverse transcription-PCR. PCR products were sequenced by dideoxy chain termination with a Sequenase kit (United States Biochemical Corp., Cleveland, Ohio), as described by the manufacturer. On comparison of the sequences of the 10 samples, seven different strains of MuV were recognized (Fig. 1). There are no peculiar base mutations which allow for differentiation between the virus strain that causes parotitis and the one that provokes meningitis. Our study showed that all the MuV strains had an A residue at HN nt 1081. All samples, with the exception of three derived from patients with parotitis, presented a C residue at position 1470 and a T residue at position 1476. Of these three parotitis samples, one presented an A residue at position 1470 and two presented a C residue at position 1476. While the base change at position 1476 corresponds to a silent mutation, the change at position 1470 determines an amino acid substitution (C/AsnA/Lys). Our results agree with those of Brown et al. (3), who suggested that the A form was pathogenic at least in infections which cause disease, but we could not find a correlation with the other substitutions at positions 1470 and 1476, which are supposedly necessary for discriminating between the attenuated form and the virulent form (2). We did not detect any change at position 1570 in any sample. On the contrary, we found some other noncorrelated point mutations in the HN sequence considered in this study. We found a mutation in six samples at position 1074, where a C was substituted by a T. Another six samples had a G instead of an A at position 1134. None of these mutations, however, determined an amino acid change. Only one sample, derived from a patient with meningitis, presented a mutation at position 1420 which determined an amino acid change (AsnAsp). All the data regarding the sequence were generated directly from the samples, avoiding cell culture, since the genetic heterogeneity may be an artifact of the cell culture system. We are in agreement with others (2, 3) that the A residue at position 1081 is an important genetic marker, but we are skeptical in considering it the only change responsible for the pathogenicity of the virus. It has been reported that Urabe AM9 is a mixture of two viruses (as is the Jeryl Lynn vaccine) (1) and that the virus containing the A1081 residue is like the wild type, but we would define it as a variant that has been selected during passages in chicken embryo amniotic cavities and in quail embryo fibroblasts (8). The finding that an A1470 associated with the A1081 has been found in the "pathogenic" form of the Urabe strain, and that it is rarely found (1 in 10 strains) in virulent strains, confirms the fact that the A1470 mutation is not involved in virus pathogenicity. Likewise, the TC mutation at position 1476 (2 in 10 strains) is not always associated with G1081.

View larger version (22K): [in this window] [in a new window]   FIG. 1.   Comparison of two representative fragments of the MuV HN gene (Urabe AM9 strain) with the sequences derived from strains from patients with meningitis (1-3) and parotitis (4-7). Letters in bold type correspond to mutations at positions 1081, 1470, and 1476.

Thus, the study of the virulence markers of the mumps virus is very complex, and further studies which could correlate the nucleotide sequences of the HN and fusion genes with the recognition of the most important epitopes are necessary. However, our data demonstrate that all the virulent strains always contain an A at position 1081, which appears to be the most relevant marker of virulence.