Phytoparasitica DOI 10.1007/s12600-016-0548-8

Molecular analysis of three new Cherry mottle leaf virus isolates reveals intra-species genetic diversity and inter-species gene transfer Li Su & Basdeo Bhagwat & Mike Bernardy & Paul A. Wiersma & Zhihui Cheng & Yu Xiang

Received: 30 June 2016 / Accepted: 12 October 2016 # Her Majesty the Queen in Right of Canada 2016

Abstract Cherry mottle leaf virus (CMLV) is a pathogen of sweet cherry and several other plant species in the genus Prunus, but only two isolates of the virus were reported previously. Here we determined the complete genome sequences of three CMLV isolates that were collected during 1980s. The coat protein of the CMLV was estimated to be about 22 kDa by protein western blot assay, and agree with the size expected based on their genome sequences. The three isolates share 78.4 to 82.7 % identities at the nucleotide sequence level between each other and 78.7 to 84.3 % identities to the two previously published CMLV isolates SA1162-21 and 95CI215. The high genome sequence divergences among the CMLV isolates were supported by phylogeny analysis. Possible inter-species gene transfer between CMLV and the other species in the genus Trichovirus was revealed by the RDP recombination analysis algorithms.

Electronic supplementary material The online version of this article (doi:10.1007/s12600-016-0548-8) contains supplementary material, which is available to authorized users. L. Su : Z. Cheng (*) College of Horticulture, Northwest A&F University, Yangling, Shaanxi 712100, China e-mail: [email protected] L. Su : B. Bhagwat : M. Bernardy : P. A. Wiersma : Y. Xiang (*) Agriculture and Agri-Food Canada, Summerland Research and Development Centre, Summerland, BC V0H 1Z0, Canada e-mail: [email protected]

Keywords Cherry mottle leaf virus . Genetic divergence . Recombination . Sweet cherry . Trichovirus

Cherry mottle leaf virus (CMLV), which has long been reported in North America and recently in China, is a pathogen causing irregular chlorotic leaf mottle, distortion and puckering in some varieties of sweet cherry (Prunus avium) (Ma et al. 2014; Zeller 1934). CMLV also infects ornamental flowering cherry, peach and apricot but is mostly symptomless in these species (James and Mukerji 1993; Mekuria et al. 2013). The virus is transmissible by grafting and by Eriophyes inaequalis experimentally (James and Mukerji 1993). CMLV is a member of the genus Trichovirus in the family Betaflexiviridae. It forms filamentous particles approximately 790 × 11 nm, and has a single-stranded positive-sense RNA genome of about 8 kb, excluding a poly(A) tail at the 3’ end. Only two CMLV genome sequences have been determined to date, including the cherry isolate SA1162-21 (James et al. 2000) and the peach isolate 95CI215 (Mekuria et al. 2013). The virus genome contains four putative open reading frames (ORFs), coding for the virus replicase polyprotein, movement protein, coat protein and a putative nucleic acid-binding protein of unknown function, respectively (James et al. 2000; Mekuria et al. 2013). A previous report implied that CMLV may be a product of a recombination event between members within the family Betaflexiviridae but no direct evidence was provided (James et al. 2006). Progress in understanding the origins, divergence, evolution and pathogenicity of the

Phytoparasitica

virus has been slow at least partially due to the lack of new virus isolates for comparative study. In the present report, we determined and analyzed complete genome sequence of three CMLV isolates from the collections conserved in the Canadian Plant Virus Collections (CPVC) at Agriculture and Agri-Food Canada. Analysis of the CMLV genome sequence data suggests high genetic divergence among different CMLV isolates and possible recombination events between CMLV and the other virus species in the genus Trichovirus. Samples designated as S85-58HP, S85-79, and LC3 were originally collected in 1980s and conserved in the CPVC (Online Resource 1). Viral double-stranded RNA (dsRNA) was extracted from the three samples as described by Morris and Dodds (1979) and used for preparation of cDNA libraries (Online Resource 1). Pyrosequencing of the cDNA libraries yielded 1,012, 2,853 and 31,060 quality-filtered nucleotide sequence reads for samples S85-58HP, S85-79 and LC3, respectively. The average length of the reads was above 340 bp. De-novo assembly of the sequence reads resulted in 34, 94 and 185 contigs for S85-58HP, S85-79 and LC3, respectively. The contig datasets were used as queries to search against a plant virus database downloaded from the Descriptions of Plant Viruses website (DPVweb, http://www.dpvweb.net) (Adams and Antoniw 2006) for virus discovery based on homology with existing virus sequences (Online Resource 1). Only CMLV-like sequences were detected in S85-58HP in six contigs, ranging from 422 to 1226 bp with 76 to 85 % nucleotide (nt) identity with CMLV (Genbank accession: KC207480 and AF170028). Virus-like sequences, including CMLV, Cherry rusty mottle associated virus (CRMaV) and Prune dwarf virus (PDV) were identified in sample S85-79. Sample LC3 was obtained from a tree conserved as part of a collection of Little cherry virus 2 (LChV2) isolates. Surprisingly, it contained genome sequences of several virus populations of Cherry virus A (CVA), Cherry green ring mottle virus (CGRMV), CMLV, Little cherry virus 1 (LChV1), LChV2, PDV and Prunus necrotic ringspot virus (PNRSV). However, only CMLV but no other viruses in the three samples was validated by RTPCR and Sanger re-sequencing. Poor quality of cDNA libraries, mostly from contamination of adapter dimers was probably one of the main reasons causing the very low number of quality-filtered pyrosequencing reads with samples S85-58HP and S85-79 (data not shown). In general, much stricter library evaluation and

quantification steps are required to reduce such artifacts from library preparation in our future works (Head et al. 2014). The pyrosequencing data detected CMLV in the three samples but the assembled CMLV contigs did not cover complete CMLV genome sequences. Therefore, we decided to re-sequence the CMLV genome in the tree samples by Sanger-sequencing. Using the primers designed from the pyrosequencing results (Online Resource 2), single PCR products of 7912 bp (nucleotide position 15 to 7926) from sample S85-58HP, 7829 bp (nt 60 to 7888) from sample S85-79, and 7918 bp (nt 21–7938) from sample LC3 were amplified by reverse transcription PCR (RT-PCR) from the three samples, respectively. Furthermore, two smaller overlapped fragments (4–4.5 kb in size) were amplified from each main PCR product (7912 bp, 7829 bp or 7918 bp, as described above) and served as the templates for Sanger-sequencing. The 5’- and 3’-end of the CMLV genome were determined by sequencing the clones of RACE-PCR products (Online Resource 1). Three complete or near complete mono-partite CMLV RNA genome sequences were assembled from sequences of the specific PCR fragments. To keep consistent with the historical records in the CPVC, the CMLV isolates from the three samples were designated as isolate S85-58HP, S85-79 and LC3 respectively. The CMLV genome was determined to be 7989 nucleotides (nts) in length for isolate S85-58HP (GenBank accession: KX443697), 7988 nts for isolate S85-79 (KX443698) and 7986 nts for isolate LC3 (KX443699), all excluding a poly(A) tail at the 3’ end. The three isolates share 78.4 to 82.7 % identities at the nucleotide sequence level between each other and 78.7 to 84.3 % identities to the two previously published CMLV isolates SA1162-21 and 95CI215 (James et al. 2000; Mekuria et al. 2013), suggesting relatively high nucleotide divergence among the CMLV genome sequences. The genome organization of all three CMLV isolates is similar to that of CMLV-SA1162-21 and CMLV95CI215, consisting of four presumed open reading frames (ORF1 to ORF4) (James et al. 2000; Mekuria et al. 2013). The ORF1 encodes the virus replicase polypeptide P1 (1187 aa, approximately 215 kDa). The deduced P1 proteins from the different isolates share 87.7–93.2 % identity and contain conserved peptide domains for viral methyltransferase (amino acid residue position 43–336, pfam01660), oxygenase (aa position 734–823, pfam13532), peptidase (aa position

Phytoparasitica

844–935, pfam05413), RNA helicase (aa position 1060–1299, pfam01443) and RNA-dependent RNA polymerase (aa position 1477–1773, pfam00978). A further analysis of the alignment of the P1 protein sequences revealed a variable region downstream of the methyltransferase domain from aa position 563 to 666, showing only 32–69 % identity between/among the isolates and no protein domain was detected for the region. The ORF2 encodes a single viral movement protein (MP, 422 aa, ~47 kDa), belonging to the ‘p30like’ MP superfamily (Martelli et al. 2007). The protein sequence similarity ranges from 81.0 to 88.7 % among the different isolates including CMLV-SA1162-21 and CMLV-95CI215. The virus coat protein (CP), encoded by ORF3, has a predicted length of 193 aa for all four CMLV cherry isolates, with an estimated molecular mass of 22 kDa. In contrast, the CP of the peach isolate CMLV-95CI215 is composed of 260 aa residues with a molecular mass of 29.5 kDa (Mekuria et al. 2013). It is because the initiating start codon of CMLV-95CI215 ORF3 was predicted 201 bp further upstream that of the cherry isolates, which results in coding for an additional 67 aa at the N-terminus. The C-terminal 193 aa of CMLV-95CI215 shares 93 to 96 % sequence identity with the CPs of the four cherry isolates. The predicted size of 22 kDa for the CPs of S85-58HP, S85-79 and LC3 was confirmed by immunoblot analysis (data not shown). Alignments of the CMLV protein sequences indicated that CMLV CP shares about 93.3–97.9 % identity at the amino acid level among the different CMLV isolates. The ORF4 encodes a putative P4 protein (135 aa, ~16 kDa) of unknown function but containing a putative nucleic acid-binding motif (James et al. 2000). The putative P4 proteins from the different isolates share 80.7–88.9 % aa identity. The CMLV genome contains three untranslated regions (UTR), including the 5’UTR, the 3’UTR and the intergenic region (IR) between the ORF3 and ORF4. Noticeably, the homology of the 3’UTR is very high, showing 96.1 to 99.2 % identities among the five CMLV isolates. Taken together, the data of the genome structure and divergence in the complete genome sequences and the putative protein sequences suggests that CMLV isolates S8558HP, S85-79, LC3 are likely new strains of the virus. Using MEGA6 software (Tamura et al. 2013), a phylogenetic tree was constructed based on alignment of the complete genome sequences of the three CMLV isolates presented in this report along with CMLVSA1162-21 and CMLV-95CI215, and the other virus

species in the genus Trichovirus, including: Apple chlorotic leaf spot virus (ACLSV), Apricot pseudo-chlorotic leaf spot virus (APCLSV), Grapevine berry inner necrosis virus (GINV), Grapevine Pinot gris virus (GPGV), and Peach mosaic virus (PcMV) (Fig. 1). The long genetic distances represented by the wide branch lengths and the high bootstrap percentages displayed in the topology tree support high genetic divergence among the five CMLV isolates. The CMLV genomes cluster into two sub-lineage groups, with CMLV-SA1162-21, CMLV-S85-79 and CMLV-LC3 as one group, and CMLV-95CI215 and CMLV-S85-58HP as the other one. The two CMLV groups cluster together and separate from the other trichoviruses, suggesting that the CMLV isolates originated from a common ancestor. Among the different species in the genus Trichovirus, CMLV is phylogenetically more related to PcMV than to APCLSV, ACLSV, GINV and GPGV. Recombination analysis, using the methods packed in the RDP4 programs (Martin et al. 2015), of the complete genomes of the five CMLV isolates and the other five species, ACLSV, APCLSV, GINV, GPGVand PcMV in the genus Trichovirus, reveals possible interspecies recombination events among the trichoviruses (Online Resource 3). A site in the ORF1 region of about 300 bp with breakpoints between nt positions 1300/1320 and 1630/1633 is predicted to be recombined among CMLV-LC3, CMLV-SA1162-21 and PcMV (GenBank accession: NC011552) and the recombination events are supported with moderate to high probability depending on performance of the individual

Fig. 1 A phylogeny tree on the basis of the alignment of the full genome nucleotide sequences of the CMLV isolates and the representative virus species in the genus Trichovirus. The distance tree was constructed in MEGA6 using the Neighbor-Joining method based on Maximum Composite Likelihood model. Values on branches indicate percentage of support out of 1000 bootstrap replications. The scale bar indicates numbers of substitutions per base. The GenBank accessions of the viruses are included within brackets

Phytoparasitica

programs. Other recombination signals in ORF1 suggest that CMLV-LC3-like genome could be a donor (minor parent) for nt 4718–4794 of GINV (KU234317), and GINV-like virus might contribute sequence to CMLVLC3 between nt 4861–4919 or CMLV-95CI215 between nt 4877–4923 as well. The programs also predict unknown viruses, possibly ACLSV-derived, might be the minor parents for some genome sequences within the regions of ORF2, ORF3, IR and ORF4 of PcMV or the five CMLV isolates. A recent report also identified possible recombination between GINV and GPGV (Glasa et al. 2014). These results suggest inter-species gene transfer through recombination is probably a common phenomenon for the viruses in the genus Trichovirus. Intra-species recombination was previously reported in ACLSV (Dhir et al. 2013). Possible recombination among the different CMLV isolates was also detected by the RDP programs. The putative CMLV intra-species recombination sites were suggested in the regions of the ORF3 and ORF4, but the events were detected by only four methods in the RDP programs and supported with low probability signals, which cannot rule out possibility of local divergence of the genome sequences (data not shown). Apparently, genome sequences of more CMLV isolates are required for this type of analysis. Acknowledgments LS was supported by the Ministry of Education of China and Agriculture and Agri-Food Canada (MOEAAFC) joint PhD student research program.

References Adams, M. J., & Antoniw, J. F. (2006). DPVweb: a comprehensive database of plant and fungal virus genes and genomes. Nucleic Acids Research, 34, 382–385.

Dhir, S., Zaidi, A. A., & Hallan, V. (2013). Molecular characterization and recombination analysis of the complete genome of Apple chlorotic leaf spot virus. Journal of Phytopathology, 161, 704–712. Glasa, M., Predajňa, L., Komínek, P., Nagyová, A., Candresse, T., & Olmos, A. (2014). Molecular characterization of divergent Grapevine Pinot gris virus isolates and their detection in Slovak and Czech grapevines. Archives of Virology, 159, 2103–2107. Head, S. R., Komori, H. K., Lamere, S. A., Whisenant, T., Van, N. F., Salomon, D. R., & Ordoukhanian, P. (2014). Library construction for next-generation sequencing: overviews and challenges. Biotechniques, 56, 61–77. James, D., & Mukerji, S. (1993). Mechanical transmission, identification, and characterization of a virus associated with mottle leaf in cherry. Plant Disease, 77, 271–275. James, D., Jelkmann, W., & Upton, C. (2000). Nucleotide sequence and genome organisation of Cherry mottle leaf virus and its relationship to members of the Trichovirus genus. Archives of Virology, 145, 995–1007. James, D., Varga, A., Croft, H., Rast, H., Thompson, D., & Hayes, S. (2006). Molecular characterization, phylogenetic relationships, and specific detection of Peach mosaic virus. Phytopathology, 96, 137–144. Ma, Y. X., Li, J. J., Li, G. F., & Zhu, S. F. (2014). First report of Cherry mottle leaf virus infecting cherry in China. Plant Disease, 98, 1161. Martelli, G. P., Adams, M. J., Kreuze, J. F., & Dolja, V. V. (2007). Family Flexiviridae: a case study in virion and genome plasticity. Annual Review of Phytopathology, 45, 73–100. Martin, D. P., Murrell, B., Golden, M., Khoosal, A., & Muhire, B. (2015). RDP4: detection and analysis of recombination patterns in virus genomes. Virus Evolution, 1, vev003. Mekuria, T. A., Druffel, K. L., Susaimuthu, J., & Eastwell, K. C. (2013). Complete nucleotide sequence of a strain of Cherry mottle leaf virus associated with peach wart disease in peach. Archives of Virology, 158, 2201–2203. Morris, T. J., & Dodds, J. A. (1979). Isolation and analysis of double-stranded RNA from virus-infected plant and fungal tissue. Phytopathology, 69, 854–858. Tamura, K., Stecher, G., Peterson, D., Filipski, A., & Kumar, S. (2013). MEGA6: molecular evolutionary genetics analysis version 6.0. Molecular Biology and Evolution, 30, 2725–2729. Zeller, S. M. (1934). Cherry mottle leaf. Oregon State Horticultural Society Annual Report, 26, 92–95.