Changes in mumps virus neurovirulence phenotype associated with quasispecies heterogeneity

Abstract

Mumps virus is a highly neurotropic virus with evidence of central nervous system invasion (CNS) in approximately half of all cases of infection. In countries where live attenuated mumps virus vaccines were introduced, the number of mumps cases declined dramatically; however, recently, the safety of some vaccine strains has been questioned. For example, one of the most widely used vaccines, the Urabe AM9 strain, was causally associated with meningitis, leading to the withdrawal of this product from the market in several countries. This highlights the need for a better understanding of the attenuation process and the identification of markers of attenuation. To this end, we further attenuated the Urabe AM9 strain by serial passage in cell culture and compared the complete nucleotide sequences of the parental and passaged viruses. Interestingly, despite a dramatic decrease in virus virulence (as assayed in rats), the only genomic changes were in the form of changes in the level of genetic heterogeneity at specific genome sites, i.e., either selection of one nucleotide variant at positions where the starting material exhibited nucleotide heterogeneity or the evolution of an additional nucleotide to create a heterogenic site. This finding suggests that changes in the level of genetic heterogeneity at specific genome sites can have profound neurovirulence phenotypic consequences and, therefore, caution should be exercised when evaluating genetic markers of virulence or attenuation based only on a consensus sequence.

Identification of genetic mutations associated with attenuation and changes in tropism of Urabe mumps virus

“…All of the sites identified by us in L, except the two characterizing Gw7, have been noted to vary or be mutated in either Urabe vaccines or in at least one other Urabe virus [Amexis et al, 2001;Sauder et al, 2006]. We believe that the mutations in the polymerase of Gw7 have not been generated by passage in culture since purification, as we have confirmed these mutations in a second ''G'' virus isolated independently from the same vaccine lot (unpublished results).…”

“…These mutations were verified by independent RT/PCR, so were not generated during the cloning process. At 8 sites, 1004-10/2 was identical to the published sequence for 87-1005, but unlike any other Urabe sequences (shaded in Table I , has been detected in the SKB vaccine [Sauder et al, 2006]. Both viruses were identical to the short stretches of sequence available for wild-type Urabe P [Yamada et al, 1989] and the wild type SH gene was identical to that of Gw7 SH [Takeuchi et al, 1991].…”

“…One of these was the non-coding change in NP at nt 1,387, and the remaining differences predicted aa changes at F . Sequences generated from vaccines directly showed variation at two of these sites, HN 335 and F 120 [Amexis et al, 2001;Sauder et al, 2006]. At nt15328, Gw7 had the same mutation as 87-1004, but this change has not been described for any other Urabe virus.…”

“…If mutations were not confirmed, an additional sequence was generated from another independent RT/PCR reaction. The sequences of the trailer and leader of each virus were confirmed by direct sequencing of PCR products spanning these regions, while genome end sequences were resolved using 3 0 and 5 0 RACE (Invitrogen) as previously described [Sauder et al, 2006].…”

“…The full genomes include representative sequences from SKB, Chiron and Biken Urabe vaccines [Amexis et al, 2001;Sauder et al, 2006], three sequences from a vaccine in Japan have been banked, AM9 parental, AM9-A, AM9-B (Genbank AB000388, AB000387, AB000386) and sequences from isolates associated with post-vaccination meningitis in Canada, 87-1004 and 87-1005 (Genbank AF314560, AF314562). All sequences are assumed to represent partially attenuated Urabe viruses, although we think 87-1004 and 87-1005 are among the most virulent viruses associated with the vaccine, and hence closer in sequence to wild-type Urabe virus.…”