Occurrence and molecular characterization of Turkish isolates of Turnip mosaic virus

Authors


*E-mail: skorkmaz@comu.edu.tr

Abstract

A total of 142 samples of plants showing symptoms of Turnip mosaic virus (TuMV) were collected from fields planted to Brassicaceae and non-Brassicaceae crops in the southwest Marmora region of Turkey, during the 2004−06 growing seasons. Using enzyme-linked immunosorbent assay (ELISA) TuMV was detected in the main brassica-crop fields of Turkey, with an overall incidence of 13·4%. TuMV was detected in samples from Brussels sprouts, cabbage, wild mustard, radish and wild radish, but not cauliflower or broccoli. The full-length sequences of the genomic RNAs of two biologically distinct isolates, TUR1 and TUR9, were determined. Recombination analyses showed that TUR1 was an intralineage recombinant, whereas TUR9 was a non-recombinant. Phylogenetic analyses of the Turkish isolates with those from the rest of the world showed that the TUR1 and TUR9 isolates belonged to world-Brassica and Asian-Brassica/Raphanus groups, respectively. This study showed that TuMV is widely distributed in the Asia Minor region of Turkey.

Introduction

Turnip mosaic virus (TuMV) belongs to the genus Potyvirus within the Potyviridae family of plant viruses (Fauquet et al., 2005). Potyviruses have flexuous filamentous particles 700−750 nm long, each of which contains a single copy of the genome, a single-stranded positive sense RNA molecule about 10 000 nt long. TuMV infects a wide range of plant species, mostly from the family Brassicaceae. It is a widely occurring and economically important virus, infecting both crop and ornamental species of this family. It occurs in many parts of the world, including the temperate and tropical regions of Africa, Asia, Oceania and North/South America (Provvidenti, 1996). TuMV is transmitted by a wide range of aphid species in a non-persistent manner. Climatic conditions that favour the build-up of aphid populations increase the spread of the virus.

Previous studies showed that the different TuMV subpopulations have probably emerged from the more ancient Eurasian subpopulations, such as those found in the Mediterranean region, including southeast Europe, Asia Minor and mid-Eurasia (Ohshima et al., 2002; Tomimura et al., 2004). In these regions, Brassica crops are an important component of local agriculture; in Europe, the crops are mostly Brassica species; and in Asia Minor, both Brassica and Raphanus species are important. Although TuMV was recently reported in Turkey (Korkmaz et al., 2007), little data on the incidence of TuMV and its biological and molecular characteristics are available. Recent studies of the genetic structures of TuMV populations in East Asia and Japan, including recombination, phylogeny, selection, neutrality tests and mismatch distribution, suggested that recent Chinese and Japanese TuMV isolates came from discrete lineages of the same population (Tomitaka & Ohshima, 2006). Furthermore, it was shown that the Asian TuMV population was composed of three phylogenetic lineages: Asian-Brassica/Raphanus (Asian-BR), world-Brassica (world-B) and basal-Brassica/Raphanus (basal-BR). Here, the occurrence of TuMV in the Asia Minor region of Turkey is reported from different hosts, mostly Brassicaceae. Furthermore, two full genomes of these isolates were sequenced and their phylogenetic relationships with worldwide isolates are discussed.

Materials and methods

Disease survey and sample collection

The Brassicaceae crop-producing areas of the southwest Marmora region of Turkey, including Canakkale, Balikesir and Bursa provinces, were surveyed during the growing seasons of 2004–06. Such crops (including cabbage, cauliflower, broccoli, radish and Brussels sprouts) and some non-Brassicaceae plants were visually examined and samples taken from plants showing TuMV-like symptoms. Some non-Brassicaceae crops were also collected because of previous reports of TuMV infection in these crops (Ohshima et al., 2002). All commercial fields and home gardens were randomly selected and samples collected from plants showing symptoms such as mosaic, mottling, necrotic spots, malformation and chlorosis. Several leaves were collected from each plant, placed in plastic bags and transported in an ice chest to the laboratory. All collected samples were tested by direct double antibody sandwich enzyme-linked immunosorbent assay (DAS-ELISA).

Enzyme-linked immunosorbent assay

DAS-ELISA, as previously described by Clark & Adams (1977), was used to detect TuMV. Briefly, 0·5 g leaf tissue was ground to a powder with a mortar and pestle in 10 mL phosphate-buffered saline, pH 7·4, containing 0·05% Tween 20, 2·0% polyvinylpyrrolidone (MW 40 000) and 0·2% bovine serum albumin. In the meantime, microtitre plates were coated with TuMV-specific broad-spectrum polyclonal antibody (Loewe) according to the manufacturer's instructions. The presence of TuMV in the samples was detected in 200 µL homogenate in duplicate wells by TuMV-specific antibody conjugated to alkaline phosphatase using p-nitrophenol phosphate substrate (Loewe) according to the manufacturer's instructions. Absorbance values at 405 nm were measured on a Stat Fax-2100 microtitre plate reader (Awareness Technology Inc.). Absorbance values that were at least twice the standard deviations of the negative controls were considered positive.

Virus isolates and host tests

Nineteen ELISA-positive TuMV isolates collected from the southwest Marmora region of Turkey (Table 1) were used to inoculate Chenopodium quinoa and serially cloned through single lesions at least three times. They were subsequently propagated in Brassica rapa cv. Hakatasuwari or Nicotiana benthamiana. Leaves from plants infected systemically with each of the TuMV isolates were homogenized in 0·01 m potassium phosphate buffer (pH 7·0), and the extract used to mechanically inoculate young indicator plants. Initially, all of the 19 TuMV isolates were inoculated onto Raphanus sativus cvs Akimasari and Taibyo-sobutori, B. pekinensis cvs Nozaki-1 go and Kyoto-3 go, B. napus cv. Norin-32 go and B. rapa cv. Hakatasuwari and the host-range type of each isolate was roughly evaluated. Two selected isolates, TUR1 and TUR9, were then inoculated onto R. sativus cvs Karagulle, Kartopu, Akimasari, and Taibyo-sobutori (the former two are Turkish cultivars and the latter two are Japanese), B. oleracea var. capitata cvs Yalova-1, Zencibasi, Mohrenkopt, Ryosan and Shinsei (the former three Turkish and the latter two Japanese), B. pekinensis cv. Nozaki-1 go and Kyoto-3 go, B. rapa cvs Ada-202 and Hakatasuwari (the former Turkish and the latter Japanese) and some additional Brassicaceae and non-Brassicaceae plants. Inoculated plants were kept for at least 4 weeks in a glasshouse at 25°C and symptoms recorded.

Table 1.  Double antibody sandwich enzyme-linked immunosorbent assay (DAS-ELISA) for the detection of Turnip mosaic virus (TuMV) in the southwest Marmora region of Turkey during 2004–06
ProvincePlantNumber of collected samplesNumber of ELISA-positive samplesIncidence of virus infection (%)Isolate
CanakkaleBrassica oleracea var. italica (broccoli)  3 011·4 
B. oleracea var. botrytis (cauliflower) 22 0  
B. oleracea var. capitata (cabbage) 45 6 TUR1, TUR2, TUR3, TUR5, TUR6, TUR7
B. oleracea var. gemmifera (Brussels sprouts)  1 0  
Raphanus sativus (radish)  2 0  
R. raphanistrum (wild radish)  5 3 TUR16, TUR17, TUR18
Sinapis arvensis (wild mustard)  1 1 TUR19
Lactuca sativa (lettuce)  3 0  
Spinacia oleracea (spinach)  6 0  
BalikesirB. oleracea var. botrytis  20 011·9 
B. oleracea var. capitata 10 1 TUR4
B. oleracea var. gemnifera  8 1 TUR8
R. sativus  4 3 TUR9, TUR10, TUR11
BursaB. oleracea var. capitata   1 033·3 
B. oleracea var. gemmifera  1 0  
R. sativus  5 4 TUR12, TUR13, TUR14, TUR15
R. raphanistrum   1 0  
Spinacia oleracea  3 0  
Allium ampeloprasum (leek)  1 0  
Total 1421913·4 

Viral RNA and sequencing

Two Turkish isolates, TUR1 and TUR9, collected from B. oleracea and R. sativus plants were selected for sequencing. Viral RNAs were extracted from TuMV-infected B. rapa and N. benthamiana leaves using Isogen (Nippon Gene). The RNAs were reverse-transcribed and amplified using high-fidelity PlatinumTM Pfx DNA polymerase (Invitrogen). The reverse transcription–polymerase chain reaction (RT-PCR) products were separated by electrophoresis in agarose gels and purified using a QIAquick Gel Extraction kit (Qiagen). Sequences from each isolate were determined using three or four overlapping independent RT-PCR products to cover the complete genome. The sequences of the RT-PCR products of adjacent regions of the genome overlapped by at least 100 bp to ensure that they were from the same genome and were not from different components of a genome mixture. Each RT-PCR product was sequenced by primer walking in both directions using a BigDye Terminator v3·1 Cycle Sequencing Ready Reaction kit (Applied Biosystems) and an Applied Biosystems Genetic Analyzer DNA model 310. Sequence data were assembled using bioedit version 5·0·9 (Hall, 1999). The genomic sequences of Turkish isolates determined in this study are lodged in the DDBJ/EMBL/GenBank databases (Accession Nos. AB362512 and AB362513).

Recombination analyses

The genomic sequences of the 97 isolates available from the international sequence databases and two Turkish isolates analysed in this study were used for evolutionary analyses. Two sequences of Japanese yam mosaic virus (JYMV) (AB027007 and AB016500), one of Scallion mosaic virus (ScMV) (AJ316084) and one of Narcissus yellow stripe virus (NYSV) (AM158908) were used as outgroups as blast searches identified them as the sequences in the international sequence databases most closely and consistently related to those of TuMV. TuMV protein 1 (P1) genes were more closely related to those of JYMV than ScMV, whereas for some other genomic regions between the helper-component proteinase protein (HC-Pro) and nuclear inclusion b protein (NIb) sequences the opposite was true; the TuMV coat protein (CP) gene was most closely related to that of NYSV. All 97 TuMV P1 sequences were aligned with those of the two JYMV isolates as outgroups, the CP sequences with that of NYSV, and the remaining sequences with those of JYMV and ScMV using clustal x (Jeanmougin et al., 1998). However, this procedure resulted in some gaps that were not in multiples of three nucleotides. Therefore, the amino acid sequences corresponding to individual regions were aligned with the appropriate outgroups indicated above using clustal x with transalign (kindly supplied by Georg Weiller, Australian National University, Canberra, Australia) to maintain the degapped alignment of the encoded amino acids, and the aligned subsequences were then reassembled to form complete sequences of 9321 nt. The aligned sequences were first checked for incongruent relationships that might have resulted from recombination, using rdp (Martin & Rybicki, 2000), geneconv (Sawyer, 1999), bootscan (Salminen et al., 1995), maxchi (Maynard Smith, 1992), chimaera (Posada & Crandall, 2001) and siscan (Gibbs et al., 2000) programs in rdp3 (Martin et al., 2005). These analyses were performed using default settings for the different detection programs and a Bonferroni corrected P-value cutoff of 0·05 or 0·01. Next, all sequences that had been identified as likely recombinants, together with all those used in this study, were checked again using the original phylpro (Weiller, 1998) and siscan version 2, with all nucleotide sites, as well as with synonymous and non-synonymous sites separately. Both 100 and 50 nt slices of all sequences were checked for evidence of recombination using these programs. These analyses also assessed which non-recombinant sequences had regions that were closest to regions of the recombinant sequences and hence indicated the likely lineages that provided those regions of the recombinant genomes. For simplicity, these are referred to as the ‘parental isolates’ of recombinants, although in reality they were just the most closely related sequences in the set analysed.

Phylogenetic analyses

The phylogenetic relationships of the sequences were determined by two methods: the maximum-likelihood (ML) algorithm of treepuzzle version 5·0 (Strimmer & von Haeseler, 1996; Strimmer et al., 1997) and the neighbour-joining (NJ) algorithm of phylip version 3·5 (Felsenstein, 1993). For ML analyses, 1000 puzzling steps were calculated using the Hasegawa-Kishino-Yano model of substitution (Hasegawa et al., 1985). For NJ analyses, distance matrices were calculated by dnadist with the Kimura two-parameter option (Kimura, 1980) and trees were constructed from these matrices by the NJ method (Saitou & Nei, 1987). A bootstrap value for each internal node of the NJ trees was calculated using 1000 random resamplings with seqboot (Felsenstein, 1985). Trees were displayed using treeview (Page, 1996). One ScMV and two JYMV sequences were used as outgroups to construct a phylogenetic tree of the polyprotein regions, as the CP sequence of only one NYSV isolate was available. The JYMV and ScMV sequences corresponding to individual gene regions within the genomes were aligned as the encoded amino acids using clustal x with transalign to maintain the degapped alignment of the nucleotides and then reassembled to form sequences of 8988 nt. The nucleotide and amino acid similarities were estimated using the Kimura two-parameter model (Kimura, 1980) and the Dayhoff PAM matrix (Dayhoff et al., 1983), respectively.

Results

Field surveys and detection of TuMV by DAS-ELISA

A total of 142 samples collected during the 2004–06 growing seasons were tested by DAS-ELISA (Table 1). Of those tested, 19 plants (13·4%) including B. oleracea vars capitata (cabbage; seven plants) and gemmifera (Brussels sprouts; one plant), R. sativus (radish; seven plants), R. raphanistrum (wild radish; three plants) and Sinapis arvensis (wild mustard; one plant) were found to be infected with TuMV. No infection was detected in B. oleracea vars botrytis and italica (cauliflower and broccoli). Viruses were found in commercial fields as well as in home gardens. Bursa Province had the highest rate of infection (33·3%), followed by Balikesir (11·9%) and Canakkale provinces (11·4%).

Host range

The 19 Turkish TuMV isolates were examined to determine their specific host ranges; 10 came from Canakkale, five were from Balikesir and four were from Bursa (Table 1). Two of 19 isolates collected from R. raphanistrum and S. arvensis infected Brassica plants systemically, but only occasionally and did not infect R. sativus; thus, these belonged to the occasional Brassica [(B)]-infecting host-range type. Three isolates from B. oleracea plants infected most Brassica plants systemically, but did not infect R. sativus; these belonged to the Brassica [B]-infecting host-range type. In contrast, five isolates from B. oleracea plants infected most Brassica plants systemically, but occasionally infected R. sativus; thus, these belonged to the Brassica/(Raphanus) [B(R)]-infecting host-range type. The remaining seven isolates from R. sativus infected not only most Brassica plants systemically, but also Raphanus plants, and were therefore regarded as belonging to the Brassica/Raphanus [BR] host-range type.

Two isolates TUR1 and TUR9, representative of the B- and BR-infecting host-range types, respectively (Ohshima et al., 2002), were selected for a more detailed study of host range using more variable test plants (Table 2). The greatest difference in host specificity of the two isolates was that TUR1 infected B. oleracea but not R. sativus, whereas the opposite was true for TUR9, regardless of whether cultivars were of Japanese or Turkish origin. TUR1 and TUR9 showed similar pathogenicities to B. pekinensis, B. rapa and Eruca sativa. Interestingly, there was a difference in infectivity of TUR1 and TUR9 isolates on Japanese and Turkish cultivars of B. pekinensis; both isolates infected the two Japanese cultivars of B. pekinensis, Nozaki-1 go and Kyoto-3 go, but did not infect the Turkish cultivar. TUR1 infected N. glutinosa systemically, whereas TUR9 infected only the inoculated leaves.

Table 2.  Host reaction of two Turkish Turnip mosaic virus (TuMV) isolates
PlantCommon nameSeed originTuMV Isolate
TUR1TUR9
  • M, mosaic; SM, severe mosaic; NR, necrotic ringspot; CS, chlorotic spot; +, symptomless infection; –, no infection; ( ), occasionally.

  • All the leaves were tested for TuMV infection by DAS-ELISA.

  • a

    Inoculated leaves/upper leaves.

Brassicaceae
 Brassica chinensis cv. ChoyoQing geng caiDenmark+/Ma+/M
 B. juncea cv. HakarashinaMustardJapan+/SM+/SM
 B. napus cv. Norin-32 goOilseed rapeJapan–/––/–
 B. narinosa cv. TatsuaiRosette pakchoiAustralia+/M+/M
 B. oleracea var. capitata cv. Ryozan-2 goGreen cabbageJapan+/M–/–
 B. oleracea var. capitata cv. ShinseiGreen cabbageJapan+/M–/–
 B. oleracea var. capitata cv. Yalova-1White cabbageTurkey+/M–/–
 B. oleracea var. capitata cv. ZencibasiRed cabbageTurkey(+/M)–/–
 B. oleracea var. capitata cv. MohrenkoptRed cabbageTurkey+/M(+/+)
 B. pekinensis cv. Nozaki-1 goChinese cabbageJapan+/M(+/+)
 B. pekinensis cv. Kyoto-3 goChinese cabbageJapan+/SM+/M
 B. pekinensis cv. unknownChinese cabbageTurkey–/––/–
 B. rapa cv. HakatasuwariTurnipJapan+/M+/M
 B. rapa cv. Ada–202TurnipTurkey(+/M)(+/SM)
 Cheiranthus cheiri cv. Vega yellowWallflowerJapan–/––/–
 Eruca sativa cv. IzmirRocketTurkey+/SM+/SM
 E. sativa cv. OdysseyRocketItaly(+/SM)+/SM
 Nasturtium officinale cv. unknownWatercressDenmark–/––/–
 Raphanus sativus cv. AkimasariJapanese radishJapan–/–+/M
 R. sativus cv. Taibyo-sobutoriJapanese radishJapan–/–+/M
 R. sativus cv. KaragulleBlack radishTurkey–/–(+/M)
 R. sativus cv. KartopuWhite radishTurkey–/–(+/M)
Solanaceae
 Nicotiana glutinosaTobaccoJapanNR/NRNR/–
 N. benthamiana Japan+/M+/M
Chenopodiaceae
 Chenopodium quinoa JapanCS/–CS/CS

Recombination sites

The polyprotein sequences were analysed for evidence of recombination. The 5′ and 3′ non-coding regions were omitted because the aim was to examine polyprotein sequences using not only all of the nucleotide sites, but also the synonymous and non-synonymous sites separately. After all gaps and nucleotides homologous to them had been removed from the aligned sequences, the likely recombination sites were assessed using RDP3. Each of the identified sites was examined individually and a phylogenetic approach was used to verify the parent/donor assignments made by RDP3. Having examined all sites with an associated P value of < 1× 10−6 (i.e. the most obvious events), the intralineage recombinants (parents from the same major lineage) were retained and the interlineage recombinants (parents from different major lineages) were removed by treating the identified recombination sites as missing data in subsequent analyses. Moreover, the ‘phylogenetic profiles’ of the polyprotein sequences, which were examined by phylpro and siscan version 2 programs, were then used to check for evidence of recombination, not only in the total nucleotide sites, but also in synonymous and non-synonymous sites separately. This complex approach was adopted to find all of the recombination sites, as well as to decrease the possibility of obtaining false evidence of recombination. Figure 1 shows the likely recombinant of TUR1 and it recombination site, as indicated by the original software of siscan version 2 program, together with the sequences most closely related to their parents (‘parental sequences’). TUR1 showed significant affinities (Z values > 3·0) with sequence PV376Br in the region corresponding to the N terminus of the P1 gene to the C-terminal region of the HC-Pro gene (nt 1–2300 in degapped sequence). However, from the C-terminal region of the HC-Pro gene to the N-terminal region of the VPg gene (nt 2300–5700 in the degapped sequence), TUR1 had greatest affinity to KEN1. This indicated that TUR1 is at least a single recombinant and an intralineage recombinant of world-B parents. A ‘clear’ intralineage recombinant (P values smaller than 1 × 10−6) was also confirmed using the rdp (9·2 × 10−59), geneconv (5·0 × 10−57), bootscan (1·2 × 10−56), maxchi (6·4 × 10−34), chimaera (7·1 × 10−19) and siscan (2·2 × 10−2) programs in rdp3 software. On the other hand, no recombination site in the genome of the R. sativus isolate TUR9 was detected by any programs of the rdp3 software, the original programs of siscan version 2 or phylpro.

Figure 1.

Graph showing siscan analysis of the polyprotein sequence of Turnip mosaic virus isolate TUR1 with that of isolates PV376Br (black thick line) and KEN1 (grey thick line). The sequences of PV376Br and KEN1 represent the likely parental sequences of TUR1. Window comparison involved subsequences of 100 nucleotides and a step between window positions of 50 nucleotides. Note the strong support (i.e. Z value > 3·0) for TUR1 being more closely related to PV376Br than KEN1 in the degapped 5′ 2300 nt and the reverse between the degapped nt 2300 and 5700. Arrow indicates where recombination occurred.

Phylogenetic relationships

Trees were initially calculated from the genome sequences of the 97 isolates, including all of the recombinants identified in this study. However, there were inconsistencies in and poor bootstrap support for some lineages in the resulting trees, as found previously (Ohshima et al., 2002, 2007). Therefore, the trees were recalculated from the genomes of only 63 isolates, omitting the interlineage recombinants. The relationships of these isolates were investigated by ML and NJ methods and the ML tree is shown in Fig. 2. All of the trees partitioned most of the sequences into the same four consistent groups: basal-B, basal-BR, Asian-BR and world-B, as reported previously (Ohshima et al., 2007). The basal-B group was a sister group to all others in the ML and NJ phylogenies and was supported by high bootstrap values. TUR1, the intralineage recombinant isolate from B. oleracea var. capitata, fell into the world-B group, the largest group of TuMV isolates, consisting of many of the B-infecting host-range type isolates from the world. On the other hand, TUR9, the non-recombinant isolate from R. sativus fell into the Asian-BR group, confined to BR-infecting host-range type isolates from Asia. As TUR9 was found to be non-recombinant, the nucleotide and amino acid similarities between it and the other two isolates in the Asian-BR group were calculated (Table 3). The similarities showed that the TUR9 isolate was distinct from the two Chinese isolates in the Asian-BR group.

Figure 2.

Maximum likelihood (ML) tree calculated from the polyprotein sequences of 63 Turnip mosaic virus isolates. The ML tree was calculated from the sequences of all isolates, excluding the interlineage recombinants identified in this study and in an earlier study (Ohshima et al., 2007). Numbers at each node indicate the percentage of supporting puzzling steps (only values > 50 are shown). Horizontal branch lengths are drawn to scale with the bar indicating 0·1 nt replacements per site. The homologous sequences of two isolates (mild and j1) of Japanese yam mosaic virus (JYMV), an isolate of Scallion mosaic virus (ScMV) and an isolate of Narcissus yellow stripe virus (NYSV) were used as the outgroups. Isolates in boxes show Turkish isolates obtained in this study. For details of the phylogenetic groups, basal-B, basal-BR, Asian-BR and world-B, see Ohshima et al. (2002). The name of each isolate, its country of origin, original host plant, year of isolation, host-range type and accession code in the international gene sequence database are listed.

Table 3.  Nucleotide and amino acid similarities between Turnip mosaic virus (TuMV) isolates in the Asian-BR group
 HRDCH6TUR9
  1. Nucleotide (bottom) and amino acid (top) similarities were calculated using Kimura 2-parameter (Kimura, 1980) and Dayhoff PAM matrix (Dayhoff et al., 1983).

HRD 0·006790·03495
CH60·0162 0·03562
TUR90·13820·1424 

Discussion

Tomimura et al. (2004) previously assessed the genetic structure of populations of TuMV in Europe and Asia (West and East Eurasia), using the concatenated gene sequences of 142 isolates. The simplest interpretation of the phylogenetic analyses was that the original TuMV population is the basal-B group and the most diverse, consists of isolates only from Eurasia and is a sister group to all the others. Moreover, based on analyses of phylogenetic relationships between worldwide isolates they suggested that the original TuMV population would be from Southeast European, Asia Minor or mid-Eurasian regions.

In the present study, the natural hosts and the distribution of TuMV were surveyed in the southwest Marmora region of Turkey and TuMV was found to be widespread in most of the vegetable-growing fields. However, the incidence of TuMV infections differed among plant species surveyed. This variation could be caused by many ecological factors such as the presence or abundance of virus reservoirs. In Turkey, brassica crops are usually sown in spring and autumn and they mature in summer and winter. Thus, the spring-sown crops may act as an inoculum source for the second crop in the same region. This study shows that B. oleracea var. capitata, B. oleracea var. gemmifera, R. sativus and R. raphanistrum are natural hosts of TuMV in Turkey. Moreover, the natural occurrence of TuMV on S. arvensis (wild mustard) has not previously been reported.

Host-range studies divided Turkish isolates into (B)-, B-, B(R)- and BR-infecting host-range types and showed that TUR1 and TUR9 isolates were distinct in biological characterization (Table 2). However, the host reaction of the two isolates showed that infectivity to the host plants depended not only on plant species, but also on cultivar. For instance, TUR9 infected all four R. sativus cultivars from both Turkey and Japan, but TUR1 did not infect any, indicating that R. sativus showed a clear-cut host reaction. On the other hand, isolate TUR1 infected cultivars of cabbage from both Japan and Turkey, while isolate TUR9 did not infect these plants. However, both isolates infected cv. Mohrenkopt (Turkish cultivar), indicating it was difficult to group them on the basis of infectivity in cabbage, although Turkish isolates are likely to be more variable biologically compared with east Asian TuMV (Tomimura et al., 2004; Tomitaka & Ohshima, 2006).

Recombination plays a major role in plant RNA and DNA virus variability and evolution. In the case of TuMV, a potyvirus, a total of 44 recombination sites were found in the genomes using 92 full sequences (Ohshima et al., 2007). In this study, 46 out of 97 sequences were found to be recombinants, and one Turkish isolate, TUR1, was shown to be an intralineage recombinant of world-B parents with one recombination site in the HC-Pro gene (nt 2300 in the degapped sequence). This site appeared to be different from the site found in isolate DNK2 collected in Denmark (nt 2250 in the degapped sequence), as reported in an earlier study (Ohshima et al., 2007). Thus, the TUR1 isolate may be a new type of recombinant, although this needs to be confirmed using representative isolates of both populations.

In contrast, isolate TUR9, which fell into the Asian-BR group, seemed not to be a recombinant. In east Asia, many recombinant isolates having Asian-BR sequences have been identified (Tan et al., 2004), although non-recombinant isolates of the Asian-BR group were only recently found, and only two isolates, HRD and CH6, have been observed in the Chinese population so far (Ohshima et al., 2007). There is therefore only scant information on the ancestral sequence of this lineage group. Nucleotide and amino acid sequence similarities between TUR9 and the two Chinese isolates HRD and CH6 showed that the TUR9 isolate was distinct from the two Chinese isolates. Therefore, Turkish and Chinese Asian-BR isolates seemed to be genetically distinct, as shown in the ML tree (Fig. 2). This indicates that different subpopulations of the Asian-BR group may be distributed in the Asia Minor region. The finding of an Asian-BR non-recombinant isolate in Turkey is the first evidence that Asian-BR isolates are not only distributed in east Asia, but also in the Asia Minor region, the place considered to be one of the origins of TuMV populations.

Although Turkish isolates belonging to Asian-BR and world-B groups were found for the first time in the Asia Minor population, it is not yet known whether these are dominant isolates in this region. Evolutionary comparisons of a large number of isolates from southeast Europe, Asia Minor and mid-Eurasia with representative worldwide isolates are necessary to determine this. However, the present study shows, apparently for the first time, the wide distribution and evolutionary relationships of TuMV in the Asia Minor region of Turkey.

Acknowledgements

We thank Ayako Mori (Saga University, Japan) for her technical assistance. This work was supported by grant no. 106O675 from the Scientific and Technological Research Council of Turkey and Grant-in-Aid for Scientific Research (B) no. 18405022 from the Japan Society for the Promotion of Science.

Ancillary