Phylogenetic and structural diversity in the feline leukemia virus env gene

Abstract

Feline leukemia virus (FeLV) belongs to the genus Gammaretrovirus, and causes a variety of neoplastic and non-neoplastic diseases in cats. Alteration of viral env sequences is thought to be associated with disease specificity, but the way in which genetic diversity of FeLV contributes to the generation of such variants in nature is poorly understood. We isolated FeLV env genes from naturally infected cats in Japan and analyzed the evolutionary dynamics of these genes. Phylogenetic reconstructions separated our FeLV samples into three distinct genetic clusters, termed Genotypes I, II, and III. Genotype I is a major genetic cluster and can be further classified into Clades 1-7 in Japan. Genotypes were correlated with geographical distribution; Genotypes I and II were distributed within Japan, whilst FeLV samples from outside Japan belonged to Genotype III. These results may be due to geographical isolation of FeLVs in Japan. The observed structural diversity of the FeLV env gene appears to be caused primarily by mutation, deletion, insertion and recombination, and these variants may be generated de novo in individual cats. FeLV interference assay revealed that FeLV genotypes did not correlate with known FeLV receptor subgroups. We have identified the genotypes which we consider to be reliable for evaluating phylogenetic relationships of FeLV, which embrace the high structural diversity observed in our sample. Overall, these findings extend our understanding of Gammaretrovirus evolutionary patterns in the field, and may provide a useful basis for assessing the emergence of novel strains and understanding the molecular mechanisms of FeLV transmission in cats.

Figure 1. Incidence of FeLV in blood samples collected from private veterinary hospitals located in each prefecture of Japan.

The incidence of samples testing positive for the FeLV antigen was divided into six color-coded groups in increments of 5%. A two-letter code was assigned to each prefecture as described in Table S2.

Figure 2. Detection of FeLV env genes by PCR, and strategies for analysis of these genes.

(A) Strategy used for generating PCR products. Schematics of coding sequences for the FeLV env gene are shown. FeLV proviral env sequences were amplified by PCR with the primer pairs Fe-8S/Fe-4S and Fe-3R, PRB-1 and Fe-3R, Fe-9S and Fe-7R. The lengths of the expected products from amplifications using each primer pair are shown. The abbreviations s.p., SU and TM represent, respectively, signal peptide, surface glycoprotein, and transmembrane subunit. (B) The DNA templates for PCR amplifications were as follows: neg. (genomic DNA isolated from FeLV-negative AH927 cells), GA5 (DNA from FeLV-A Glasgow-1-infected AH927 cells), GB (DNA from Gardner-Arnstein FeLV-B-infected AH927 cells), SC (DNA from FeLV-C Sarma-infected AH927 cells), DNA from FT-1 cell line, DNA from sample SN5, DNA from sample MZ29, DNA from a FeLV-positive cat (SN22), and DNA from a FeLV-negative cat (FS23). c-myc was amplified as a positive control . PCR products were electrophoresed and stained with ethidium bromide. Asterisk indicates atypical bands of env gene. (C) Each detected PCR fragment was cloned into a cloning plasmid, and full-length env gene libraries were constructed. Several unique env genes were isolated from these full-length env gene libraries using information on fragment size, or by screening using FeLV-B specific PCR. In addition to non-recombinant env genes, FeLV-B-type and other recombinants were isolated and analyzed. Furthermore, env genes smaller than full length were also analyzed. N indicates negative and P positive for FeLV-B detection or recombination detection. (D) For the most part, only non-recombinant sequences were used for phylogenetic analysis, but where recombinant sequences were included, their endogenous-derived regions were removed from the alignment. However, partial sequences from some recombinants as well as representatives of the FeLV-C and FeLV-B subgroups were included in the analyses after excluding their recombinant regions, to determine the FeLV genotypes from cats from which non-recombinants were not identified.

Figure 3. The best maximum-likelihood (ML) tree from phylogenetic analysis of near-full-length env nucleotide sequences generated in this study and obtained from the NCBI database.

FeLV sequence information obtained from the NCBI database is listed in Table S1.

Figure 4. Geographic distribution of the major FeLV genotypic groups (I, II, III) and the seven clades of Genotype I.

Each color-coded dot represents one infected cat (small dot) or 5 infected cats (large dot). Colored areas indicate provinces of Japan such as Hokkaidō, Tōhoku, Kantō, Chūbu, Kinki, Chūgoku, Shikoku and Kyūshū. A two-letter code was assigned to each prefecture as described in Table S2. A detailed summary of the geographic distribution of all groups is provided in Table S3.

Figure 5. Analyses of FeLV env gene recombination.

(A) Plots of similarity between a set of indicated sequences. Each curve is a comparison between the title sequence and the color-coded reference FeLV env sequences. The horizontal axis indicates physical position along the env sequences, and the vertical axis indicates % of the permuted tree. Non-recombinant FeLV was derived from a similar case that had each indicated recombinant except for WY24-1RC. Non-recombinant WY22-3 was used as a reference for WY24-1RC. (B) Schematic representation of the various recombination structures identified using similarity plot analysis. The motifs are abbreviated s.p. (signal peptide), SPHQ (SPHQ motif), VRA (variable region A), VRC (variable region C), VRB (variable region B), PRR (proline-rich region) and C-dom. (C-terminal domain). (C) Positions of recombination (breakpoints) in env genes from 80 recombinant clones (76 belonging to the FeLV-B subgroup, and 4 non-typical recombinants). Breakpoints of 5′ exogenous and 3′ endogenous env sequences, and 5′ endogenous and 3′ exogenous env sequences are indicated on the FeLV-A 61E sequence.

Figure 6. FeLV env gene mutants.

(A) The start (5′ breakpoint) and end (3′ breakpoint) positions of segments deleted from env genes are indicated on the FeLV-A 61E sequence. (B) The various env deletion mutants isolated from sample MZ29. PCR amplifications were performed with primers Fe-8S and Fe-3R, and the product was electrophoresed and stained with ethidium bromide. Five deletion mutants (MZ29-9, MZ29-7, MZ29-4, MZ29-3, MZ29-1) and one prototype sequence (MZ29-6) are represented schematically. s.p.: signal peptide. ins.: insertion. del.: deletion. Abbreviations for specific motifs are as for Fig. 5. Asterisk indicates stop codon.

Figure 7. Analysis of structural diversity in small regions of the env genes.

Characteristic indels in different versions of the env gene are indicated by shaded boxes and labeled with lowercase letters, shown relative to the FeLV-A 61E sequence. (A) Three insertions identified in the Genotype I group: ‘a’ – insertion of AGT or AAT; ‘b’ – insertion of AATACAAGCAGT; ‘c’ – insertion of CCCCAC. (B) Three additional indels identified in the Genotype I group: ‘d’ – insertion of ACTACT; ‘e’ – insertion of CAGGGC; ‘f’ – deletion of three nucleotides at position 535–537 of the FeLV 61E sequence. Boxes with darker shading at ‘d’ indicate atypical deletions. (C) A single insertion identified in the Genotype II group: ‘g’ – insertion of CTT. (D) – (F) show these indels (‘a’ to ‘g’) plotted on various representations of the best phylogenetic tree shown in Fig. 3.