Molecular Detection, Phylogenetic Analysis, and Identification of Transcription Motifs in Feline Leukemia Virus from Naturally Infected Cats in Malaysia

A nested PCR assay was used to determine the viral RNA and proviral DNA status of naturally infected cats. Selected samples that were FeLV-positive by PCR were subjected to sequencing, phylogenetic analysis, and motifs search. Of the 39 samples that were positive for FeLV p27 antigen, 87.2% (34/39) were confirmed positive with nested PCR. FeLV proviral DNA was detected in 38 (97.3%) of p27-antigen negative samples. Malaysian FeLV isolates are found to be highly similar with a homology of 91% to 100%. Phylogenetic analysis revealed that Malaysian FeLV isolates divided into two clusters, with a majority (86.2%) sharing similarity with FeLV-K01803 and fewer isolates (13.8%) with FeLV-GM1 strain. Different enhancer motifs including NF-GMa, Krox-20/WT1I-del2, BAF1, AP-2, TBP, TFIIF-beta, TRF, and TFIID are found to occur either in single, duplicate, triplicate, or sets of 5 in different positions within the U3-LTR-gag region. The present result confirms the occurrence of FeLV viral RNA and provirus DNA in naturally infected cats. Malaysian FeLV isolates are highly similar, and a majority of them are closely related to a UK isolate. This study provides the first molecular based information on FeLV in Malaysia. Additionally, different enhancer motifs likely associated with FeLV related pathogenesis have been identified.


Introduction
Feline leukaemia virus (FeLV) is a gammaretrovirus associated with anaemia, immunodeficiency, leukaemia, and lymphoma in cats [1]. FeLV has been studied extensively as a model for human immunodeficiency virus (HIV) and human T-cell lymphoma virus (HTLV) infections [2]. FeLV is distributed worldwide; however, prevalence varies greatly with geography and with risk factors that include age, health status, and population density [3,4]. A recent study reported FeLV seroprevalence of 5.1% and 18.9% in healthy and sick Malaysian cats, respectively [4]. On the other hand, studies carried out in other Asian regions reported 0% FeLV prevalence in Vietnam [5]; 14.7% among cats in Singapore [6]; 2.9% in Japan [7]; and 6% each from Taiwan and Thailand [8,9]. In non-Asian countries, FeLV prevalence has been reported to be 4.8% on Prince Edward Island, Canada [10]; 5.3% and 3.7% in Raleigh and Gainesville, USA, respectively [11]; 3.4% in all Canada [12]; and 4.6% in Egypt [13]. These discrepancies in prevalence of FeLV may arise due to differences in cat's lifestyle and FeLV vaccination practices in different countries [4].
Diagnosis of FeLV is usually performed by detection of p27 antigen [14]. However, demonstrating p27 antigen is difficult during early viraemia and with latent infections. Studies have shown that FeLV viral RNA and provirus DNA are better predictors of progressive and latent infections, respectively [15,16].

Veterinary Medicine International
Apart from the envelope gene of FeLV, the long terminal repeats (LTRs) play important role in determining disease outcome and in differentiating exogenous from endogenous FeLV [15,17]. Vaccination against FeLV is not carried out in Malaysia and, to date, FeLV clinical status of Malaysian cats has not been investigated using molecular assays. Additionally, unlike the ubiquitous feline infectious peritonitis (FIP) [18] sequence and phylogenetic characteristics of the Malaysian FeLV isolates have not been elucidated. The objectives of this study are to evaluate the presence of FeLV viral RNA and provirus DNA in selected antigenaemic and nonantigenaemic cats, respectively. Sequence characteristics, enhancer motifs as well as phylogenetic relationships of the Malaysian FeLV also were determined.

Animals and Sampling.
Heparinized blood samples were collected from cats presented at University Veterinary Teaching Hospital, Universiti Putra Malaysia (UVH-UPM). The samples were tested for the presence of FeLV p27 viral antigen using a commercially available test kit [4]. These cats were divided into p27 antigen positive and p27 antigen-negative groups. From each group, 39 cats were selected by convenience sampling method and the samples were subjected to PCR analysis. All cats had no history of vaccination against FeLV as vaccination against FeLV is not practiced in Malaysia. All samples were collected by the attending veterinary clinicians, as part of routine practices. In addition, consent for evaluation was obtained from the cat owners, prior to sampling.

Nucleic
Acid and PCR Amplification. Viral RNA was extracted from the plasma of p27-positive cats, using high pure viral RNA purification kit (Roche, Germany). On the other hand, genomic DNA was isolated from whole blood of p27-negative cats, using QIAGEN DNA extraction kits (QIAGEN, Germany). All nucleic acid extraction procedures were carried out according to manufacturers' instructions. RNA was reverse transcribed and subjected to nested PCR, using a one-step access RT PCR (Promega, USA). Genomic DNA was amplified by nested PCR assay.
Two sets of primers (outer and inner primers) were synthesised (1st BASE, Malaysia) and used to amplify a 601 bp segment of FeLV-U3LTR and partial gag regions. This segment recognises exogenous but not endogenous FeLV segments presence in cat genome; thus the primers used in this study are specific for exogenous FeLV detection. Outer PCR reaction was carried out using U3-F(1) (5 -ACA GCA GAA GTT TCA AGG CC -3 ) and G-R(1) (5 -GAC CAG TGA TCA AGG GTG AG-3 ) primers. The inner PCR reaction was carried out with U3-F(2) (5 -GCT CCC CAG TTG ACC AGA GT-3 ) and G-R(2) (5 -GCT TCG GTA CCA AAC CGA AA-3 ) primers [15].
The PCR mixture was prepared in 25 L reaction volume containing 10 mM each of dNTPs mix, 0.2 mM Tfl DNA polymerase (5 U/ L), 0.1 U AMV (5 U/ L), 0.1 U recombinant RNasin ribonuclease inhibitor (400 U/ L), 0.8 U MgSO 4 (25 mM), 20 pmol of each of the forward and reverse primer, 5.0 L of 1 times buffer, and 1 L RNA or DNA template. Nuclease-free water was used to bring the mixture to its final volume of 25 L. AMV reverse transcriptase enzymes and RNasin ribonuclease inhibitor were included only when RNA was a starting template for the PCR assay. In the nested PCR step, 1 L of outer PCR product was used as template.

Sequence and Phylogenetic Analyses.
In order to gain insight on the characteristics of Malaysian FeLV sequences, 29 nested PCR-positive samples (RNA = 14; DNA provirus = 15) were selected and purified using an Accuprep purification kit (Bioneer, Daejeon, Korea). Sequencing was carried out based on the amplified U3LTR-gag segment using a standard ABI Big Dye terminator version 3.1 sequence kit (Applied Biosystem). The obtained sequences were analysed for homology using the NCBI Basic Local Alignment Search Tool (BLAST: http://www.ncbi.nlm.nih.gov). In addition, multiple sequence alignment was carried out using ClustalW and the percentage nucleotide identity was determined using DNA identity matrix [19,20]. On the other hand, single nucleotide polymorphism (SNP), DNA distance matrix, and transcription binding proteins prediction analyses were carried out using geneious software version R7 [20]. A neighbour-joining (NJ) phylogenetic tree was constructed based on the U3LTR-gag sequences using MEGA5 software. The tree reliability was assessed using 100 bootstrap replicates [21]. All nucleotide sequences were deposited with the NCBI GenBank (Table 1).

Results and Discussion
FeLV infection is of concern to cat owners due to its ability to induce tumours and immunodeficiency, thus predisposing cats to other secondary diseases. In this study, a U3-LTR and gag regions of exogenous without endogenous FeLV sequences were amplified by nested PCR methods. Post-PCR analysis using electrophoresis revealed an expected amplicon size of 770 bp in the outer PCR and 601 bp in the nested inner PCR assay ( Figure 1). Overall, it was found that 97.4% (38/39) of p27 antigen-negative cats were positive for FeLV provirus DNA suggesting that this category of cats likely goes undetected when only p27 detection is used to judge their FeLV clinical status. Similar studies reported high prevalence of FeLV provirus DNA in Brazilian cats [22]. However, Hofmann-Lehmann et al. [23] reported a lower provirus DNA rate in cats in Switzerland. The observed differences  in prevalence among different countries could be associated with cat lifestyle, as well as variations in factors known to favour FeLV transmission [3,4]. Provirus DNA detection rate observed in this study could be associated with regressive or latent FeLV infection, which is characterized by integration of DNA provirus into the host cell genome and absence of viral antigen in circulation [1,15]. The consequence of latent FeLV infection is that provirus DNA could reactivate to an infectious state, especially following stress and/or immunosuppression. Thus, cats that are p27 antigen-negative, but provirus DNA positive, could serve as sources of infection of FeLV-naïve cats [24]. A previous study has established an association between feline lymphoma and provirus DNA positivity in p27 antigen-negative cats, though this has not been evaluated in the present study [25]. Moreover, transmission of FeLV has been shown to occur in cats following blood transfusion from cats with provirus DNA, thus highlighting the importance of screening blood donor cats for provirus DNA [26].
Viral RNA was detected in 87.2% (34/39) of p27 antigenpositive cats whereas 13% (5/39) tested negative using RT-PCR assay. Since plasma viral RNA is an indicator of FeLV viraemia, cats that are positive for FeLV p27 antigen and viral RNA are likely to harbour replicating virus [27]. Cats in this category may progress to a persistent viraemic stage, succumbing to FeLV-associated illness [28].
Failure to detect FeLV viral RNA in about 13% p27 antigen-positive cats (p27-positive/viral RNA-negative) could result from atypical infection, wherein the virus is sequestered and replicates locally in tissues such as salivary gland, mammary gland, and urinary epithelium, causing intermittent or low-grade antigenaemia, although there is no Note: FeLV01-FeLV14 sequences were amplified from plasma viral RNA while the remaining local sequence (FeLVUPM13-FeLVUPM29) were amplified from proviral DNA.
detectable viraemia [28,29]. Our findings are consistent with the results of an earlier study that failed to isolate FeLV from about 10% of p27 antigen-positive cats, irrespective of the antigen detection methods used. Such cats were considered as 4 Veterinary Medicine International "discordant, " suggesting that p27 antigen-positive status may not always correlate with viraemia [30]. Another potential explanation for p27-positive-RNA-negative status might be false positive antigen or false negative RNA tests that arise occasionally because of low positive predictive value of p27 antigen tests in regions with low FeLV prevalence [31]. Clinical relevance of atypical FeLV infection is not well-understood, and it has been recommended to monitor the status of discordant cats over time [27,28]. No additional follow-up was carried out in the present study, because most owners were not willing to subject cats to repeated venepunctures [4]. Based on the U3LTR and partial gag regions, nucleotide sequence analyses revealed homology of 91-100% among Malaysian FeLV isolates. However, homology decreased to 84.6% when local isolates were compared with reference isolates. Previous studies reported strong sequence conservation (>97%) among FeLV isolates of different geographic and temporal clusters [32,33]. In agreement with Jackson et al. [17], we do observe point mutations and nucleotide deletion in Malaysian FeLV isolates (see Supplementary Material available online at http://dx.doi.org/10.1155/2014/760961).While U3LTR is conserved in FeLV, field isolates have been reported to exhibit sequence variation within the terminally repeated LTRs regions [17,33]. Mutational changes in the LTR regions have been implicated with enhanced transcriptional and/or insertion activities of FeLV, thus supporting T-cell lymphomagenesis [34,35].
In this study, several transcription binding motifs were predicted within the amplified U3LTR-gag region ( Table 2). Of these, NF-GMa, Krox-20/WT1I-del2, BAF1, AP-2, TBP, TFIIF-beta, TRF, and TFIID motifs were found to be conserved between local FeLV isolates and the two characterized FeLV-Rickard subgroup A and FeLV-FAIDS reference Veterinary Medicine International 5  isolates. On the other hand, E1A-F, ELP, Sp1, C\EBPbeta, BAF1, GCF, HNF-3, and PEA3 motifs are found in some local isolates but were absent in reference sequences. These motifs may have implication for viral oncogenicity or probably favours viral replication. For example, an Sp1 enhancer, a member of Sp/Kruppel-like factor, was reported to activate gene transcription and contribute to abnormal metabolism of cancer cells [36,37] whereas C\EBPbeta regulates the growth and differentiation of myeloid as well as lymphoid cells [38]. Studies have shown that, the U3-LTR sequence contains multiple transcription binding sites that aid viral replication and pathogenesis. Interactions of different transcription binding factors, via the U3-LTRs, may contribute to cellular gene transactivation and viral leukemogenesis [39,40]. Enhancer motifs observed in this study appeared in multiple locations such as in the case of E1A-F, BAF1, and TFIID, each occurring in duplicate; GCF appeared in triplicate while AP-2 is repeated 5 times at different positions. An enhancer duplication and triplication has been reported in naturally occurring cases of FeLV-induced T-cell lymphomas [41,42]. The clinical relevance of multiple enhancers in cats used in the present study is not determined, although some FeLV positive cats had evidence of different tumour forms at post-mortem (result not shown). Previous studies reported that E1A-F, a member ets-oncogene family transcription factor, upregulates the multiple matrix metalloproteinase (MMP) genes thus contributing to the malignant phenotypic activity by increasing the invasion and metastatic activities of cancerous cells [43]. TFIID, a potential protooncogene with TATA-box protein and a TBP-associated factor also plays role in transcription initiation and genome expression [44]. On the other hand, AP2 and SP1 are known to activate epidermal growth factor receptor (EGFR) gene. In addition, overexpression of these gene has been reported to cause cellular transformation [45,46]. Surprisingly we also identified a triplicate of GCF binding factor that has suppressor effect on EGR gene; these discrepancies, however, need further elucidation with quantitative real-time PCR [47]. Absence of length mutation (nucleotide position 473-481) in Malaysian FeLV isolates, as observed in FeLV isolates Veterinary Medicine International 9 from Taiwan (FeLV-TW-25 and FeLV-TW-30) and a European isolate (FeLV-GM1), might suggest limited influence of geography in evolutionary patterns of FeLV, unlike its lentiviral counterpart, feline immunodeficiency virus [33,48].
Phylogenetic analysis based on the U3LTR-gag sequence revealed that Malaysian FeLV isolates are closely related (Tables 3(a), 3(b), and 3(c)) but when compared with reference isolates, separated into two distinct clusters, with the majority (86.2%) being closely related to FeLV-K01803 isolate from UK. The remaining local FeLV isolates (13.8%) clustered with FeLV-GM1 ( Figure 2). The reason for the observed similarity between local FeLV isolates and European isolates, but not with Taiwanese isolates, may suggest the lack of geographical influence, this should be explored further. It is possible also that FeLV might have been introduced into Malaysia as a result of translocation of domestic pets from Europe. Due to a somewhat conserved nature of the U3LTR region, conclusion about the FeLV subgroup requires further investigations of FeLV envelope protein gene.

Conclusion
This study revealed the occurrence of FeLV viral RNA and provirus DNA among naturally infected Malaysian cats. Based on the U3LTR-gag sequence, Malaysian FeLV isolates are highly conserved and more closely related to K01803 isolate from UK compared to Taiwanese and other reference isolates. Presence of multiple enhancers some of which have been linked with FeLV induced tumours may contribute to the development of poor prognostic outcome in naturally infected Malaysian cats although this needs further investigation. Overall, this is the first molecular study for evidence of FeLV in Malaysia. We also identified several motifs that have potential implications in FeLV-induced leukemogenesis. Future studies need to explore association between FeLV positive status and occurrence of feline tumour in Malaysian cats. The present findings is useful in designing molecular diagnostics for clinical applications and for improved understanding of FeLV infection outcome and epidemiology.