Genetic analysis of the central untranslated genome region and the proximal coding part of the F gene of wild-type and vaccine canine distemper morbilliviruses

Located between the open reading frames encoding the matrix (M) and the fusion (F) protein the morbillivirus genome contains an unusually large non-coding intercistronic region (M-F UTR) of up to 5.6% of the full length genome. Any function(s) of this region have largely remained obscure. Here, we analyze the M-F UTR and the proximal coding part of the downstream F gene of several recent canine distemper morbillivirus (CDV) wild-type (wt) isolates and vaccine strains. While the F gene coding part appeared to be highly conserved (about 93% homology), a considerable degree of strain-specific variation of up to 21.4% was evident when comparing the M-F UTR. Phylogenetic analysis revealed a co-circulation of several contemporary CDV genotypes within a close geographic range (central Europe). A remarkably distinct CDV wt lineage, so far detected only in mustelids, is displayed. A rather non-scattered pattern of mutations within the M-F UTR suggested superimposition of RNA sequence and/or secondary structure constraints. Extensive folding in the long (460 nt) and moderately GC-rich 5'-UTR of the F mRNA was evident, particularly around the putative F protein translation initiation codon (AUG461 of the Onderstepoort vaccine strain). The region immediately preceding the putative F initiation site also harbored the only mutation unique to both vaccine strains within the F-5'UTR (position 455: Awt vs. Cvac). The putative F protein start codon, AUG461, was found to be mutated to AUA or GUA in all wt isolates analyzed and in another vaccine strain (Rockborn). Possible consequences for F protein translation initiation in wt CDV are discussed.