Current status of cucurbit viruses in Venezuela and characterization of Venezuelan isolates of Zucchini yellow mosaic virus



In Venezuela, cucurbit viruses have been associated with important yield losses. Therefore, an extensive survey was conducted to determine the major cucurbit viruses in this country. Leaf samples from 284 cucurbit plants exhibiting virus-like symptoms were collected mainly in 2009–2010 from several states of Venezuela. They were assessed for viral infection by polymerase chain reaction (PCR) for Melon chlorotic mosaic virus (MeCMV) and reverse transcriptase (RT)-PCR for Papaya ringspot virus (PRSV), Zucchini yellow mosaic virus (ZYMV), Watermelon mosaic virus (WMV), Cucumber mosaic virus (CMV), Squash mosaic virus (SqMV) and Cucurbit aphid-borne yellows virus (CABYV). The most common virus in cucurbit fields, MeCMV, was present in 65·8% of samples. Its associated alphasatellite was found in 78% of samples positive for MeCMV. PRSV, ZYMV and WMV were found with different prevalence: 34·2, 32·4 and 1·1% respectively. CMV was also detected (6·7%) but SqMV and CABYV were not found. Single infections were more frequent than mixed infections (56·4 and 38·6%, respectively). For ZYMV, comparison and phylogenetic analyses of either polymerase and coat protein (NIb-CP) partial sequences or CP complete sequences revealed a low genetic diversity within Venezuelan isolates. Thirty-four ZYMV isolates were used for serological and biological analysis. Thirteen monoclonal antibodies showed a major group of isolates spread in several states and two groups located in Zulia only. Venezuelan ZYMV isolates showed biological variability on cucurbit cultivars susceptible, tolerant or resistant to ZYMV. Resistance to ZYMV in cucumber appears potentially durable, whereas resistance or tolerance in zucchini and melon may be easily overcome.


Cucurbit crops are widespread in tropical, subtropical and temperate areas where they are grown mostly for human consumption, and their economic importance is second only to solanaceous crops. In Venezuela, the latest official reports show that approximately 30 000 ha are dedicated to cucurbit crops. The main cultivated cucurbits are watermelon (Citrullus lanatus; 14 229 ha), squash (Cucurbita moschata and Cucurbita pepo; 7404 ha), melon (Cucumis melo; 5334 ha) and cucumber (Cucumis sativus; 842 ha) (Anonymous, 2011).

Plant diseases, including viral diseases, are one of the limiting factors to cucurbit production worldwide. More than 59 well-characterized plant viruses have been found naturally infecting cultivated cucurbits (Lecoq et al., 1998; Lecoq & Desbiez, 2012). In Venezuela, six cucurbit viruses have been described: Cucumber mosaic virus (CMV) (genus Cucumovirus), Squash mosaic virus (SqMV) (genus Comovirus), Papaya ringspot virus (PRSV), Watermelon mosaic virus (WMV) and Zucchini yellow mosaic virus (ZYMV) (genus Potyvirus) and Melon chlorotic mosaic virus (MeCMV) (genus Begomovirus) (Lastra, 1968; Hernández et al., 1989; Ramírez et al., 2004). More recently, MeCMV was found in association with an atypical satellite molecule infecting watermelon in a western state of Venezuela (Romay et al., 2010). PRSV, ZYMV, WMV and CMV are transmitted by aphids in a nonpersistent manner (Lecoq & Desbiez, 2012). SqMV can be transmitted by several species of beetles, although seed transmission is probably the major way for its long distance dissemination (Lecoq & Desbiez, 2012). Begomoviruses, including MeCMV, are transmitted by the whitefly Bemisia tabaci in a circulative manner (Navas-Castillo et al., 2011). All these viruses cause more or less severe mosaics and deformations of leaves in cucurbits and they can reduce crop yields significantly (Lecoq & Desbiez, 2012). PRSV has a limited host range including cucurbits and papaya (Carica papaya). ZYMV has an experimental host range that includes species of 11 families of dicotyledons, although it infects mostly cucurbits in natural conditions. WMV infects over 170 species in 26 botanical families, which is a broad host range for a potyvirus. CMV is able to infect over 1000 species in more than 85 botanical families, whereas SqMV infects only a few species outside cucurbits (Lecoq & Desbiez, 2012). MeCMV has been found infecting melon and watermelon (Ramírez et al., 2004; Romay et al., 2010), but its host range has not yet been fully characterized.

Viral diseases are among the major constraints for cucurbit production in Venezuela (Chirinos & Geraud-Pouey, 2011). PRSV and CMV have been associated with remarkable economic losses, up to 90%, in melon fields of the central region of Venezuela (Acosta et al., 1973). PRSV was found to be the most prevalent virus in a survey carried out 44 years ago (Lastra, 1968). Since then, there is no data on whether the status of PRSV, or of other cucurbit viruses, has changed. In the late 1980s, the potyvirus ZYMV was reported infecting several cucurbit species in Venezuela with a prevalence of 90% (Hernández et al., 1989). However, in that survey a set of only 25 cucurbit samples was analysed. More recently, symptoms associated with begomovirus infections have been reported in melon and watermelon fields of Venezuela (Chirinos & Geraud-Pouey, 2011). However, the frequency, distribution and impact of these viruses are currently unknown.

ZYMV was first isolated in Italy in 1973 and 10 years later was found in all continents, becoming one of the most economically important cucurbit viruses (Desbiez & Lecoq, 1997). In the New World, ZYMV has been reported from Canada to Argentina and Chile (Desbiez & Lecoq, 1997; García, 2000; Prieto et al., 2001). Sequences of the ZYMV coat protein gene available from public molecular databases revealed three major groups of this virus worldwide (Coutts et al., 2011). Serological studies using a set of monoclonal antibodies have shown at least 16 serotypes for ZYMV (Desbiez & Lecoq, 1997; Yakoubi et al., 2008) and biological variability includes a large spectrum of strains from severe to mild (Lecoq & Purcifull, 1992). In South America, ZYMV has been reported in Argentina, Brazil, Chile and Venezuela (Hernández et al., 1989; García, 2000; Moura et al., 2001; Prieto et al., 2001). However, little is known about the molecular, serological and biological diversity of ZYMV in Latin America. Sources of resistance to ZYMV have been identified; however, in some cases the resistance can be overcome by aggressive variants of the virus (Desbiez et al., 2003). Hence, characterization of ZYMV isolates in Latin America is necessary to estimate the durability of ZYMV resistance in local programmes of genetic improvement of cucurbits.

In order to assess the relative importance of viruses infecting cucurbits in Venezuela and the possibility of genetic control for ZYMV, this study focuses on: (i) detection, distribution and prevalence of viruses infecting either cultivated or wild cucurbits in Venezuela, and (ii) the molecular, serological and biological characterization of ZYMV strains.

Materials and methods

Sample collection

A sample collection was carried out in 2009 and 2010 (224 samples) and some additional samples collected from 2001 to 2009 (60 samples) were included in the study. The samples were collected in 11 out of 24 Venezuelan states (Fig. 1). A total of 284 cucurbit plants, both cultivated and wild, and most of them exhibiting virus-like symptoms, were sampled in 51 cucurbit fields. Sample collections were mainly undertaken in Falcón, Zulia, Guárico and Lara because of their importance for national cucurbit production. Thus, 12, 12, 8 and 7 fields were sampled in Falcon, Zulia, Guárico and Lara state, respectively (Table S1). Cucurbit species collected in this survey were C. melo (130 plants), C. lanatus (54 plants), C. moschata (54 plants), C. pepo (seven plants), C. sativus (nine plants), Cucumis anguria (19 plants), Cucumis dipsaceus (six plants) and C. melo var. agrestis (five plants).

Figure 1.

Map of Venezuela with location of states included in this survey (highlighted in grey). Falcón (Fa); Lara (La); Zulia (Zu); Trujillo (Tr); Barinas (Ba); Aragua (Ar); Yaracuy (Ya); Nueva Esparta (NE); Sucre (Su); Bolívar (Bo), Guárico (Gu).

Three to seven plants with symptoms per field were sampled. Immediately after collection, each sample was placed over a filter paper in a receptacle containing Silicagel® for at least 3 days. Then, dried samples were wrapped with tissue paper and stored at −20°C in sealed 2-mL vials containing Silicagel® until use.

Virus detection

Total RNA was extracted from 10 mg of dried leaf tissue from each sample using Tri-Reagent (Molecular Research Center, Inc.) according to the manufacturer's instructions. RT-PCR was performed using specific primers (Table 1) to detect ZYMV, PRSV, WMV, CMV, SqMV and Cucurbit aphid-borne yellows virus (CABYV, Polerovirus). RT-PCR was performed according to a standard protocol used in the laboratory (Yakoubi et al., 2008). For begomovirus detection, total DNA was extracted from each dried leaf sample following the protocol described by Gilbertson et al. (1991). PCR assays were performed using universal primers for begomoviruses (Rojas et al., 1993; Wyatt & Brown, 1996), as well as specific primers for MeCMV and its associated satellite Melon chlorotic mosaic alphasatellite (MeCMA) (Table 1). In order to test the putative presence of other begomoviruses, PCR fragments amplified with begomovirus universal primers (Wyatt & Brown, 1996) were digested with restriction enzymes AclI and MseI and the digestion products were analysed on a 2% agarose gel. Isolates that were negative for all viruses listed above were also tested by RT-PCR with specific primers for Cucurbit yellow stunting disorder virus (CYSDV, genus Crinivirus) and with universal primers for potyviruses and some flexiviruses (Table 1).

Table 1. List of primers used for virus detection in cucurbit samples from Venezuela
PrimerSequence (5′–3′)VirusFragment size expected (bp)Reference
  1. a

    M. Jacquemond, INRA Pathologie Végétale, Montfavet, France.

ZYMV-CP-5′GGTTCATGTCCCACCAAGCZYMVc. 1300Yakoubi et al., 2008
ZYMV-1500-5′CAAGACGAATTGGACTTAGCZYMV1600Desbiez et al., 2003
AV494GCCYATRTAYAGRAAGCCMAGBegomovirus DNA-Ac. 550Wyatt & Brown, 1996
PBLv2040CTCTCTGCAGCARTGRTCKATCTTCATACABegomovirus DNA-Bc. 1000Rojas et al., 1993
WMV5GGCTTCTGAGCAAAGATGWMVc. 400Lecoq & Desbiez, 2012
Poty-5CCACGGATCCGGBAAYAAYAGYGGDCARCCPotyviridaec. 1600Gibbs & Mackenzie, 1997
PDO-F1iTITTYATKAARWSICARYWITGIACFlexiviridaec. 360Foissac et al., 2005

In addition, a DAS-ELISA test was performed for 99 samples collected in 2010, which were previously tested by RT-PCR as mentioned above (the amount of dry material available for the older samples was not sufficient to test them this way), using antisera against ZYMV (Desbiez & Lecoq, 1997), PRSV, WMV, CMV, CABYV, SqMV, Melon necrotic spot virus (MNSV, genus Carmovirus) and Cucumber green mottle mosaic virus (CGMMV, genus Tobamovirus) produced in the laboratory at INRA Montfavet.

Molecular variability of ZYMV isolates

The molecular diversity of ZYMV in Venezuela was studied using sequence comparison of a c. 600 bp fragment that includes the C-terminal part of the polymerase coding region (NIb) and the N-terminal part of the coat protein (CP) coding region (C-terNIb-N-terCP). This fragment is known to be highly variable, and is frequently used to study the intraspecific variability and genetic structure of potyviruses (Yakoubi et al., 2008; Lecoq & Desbiez, 2012). RT-PCR products from 89 samples positive for ZYMV, obtained in the previous section, were directly sequenced. In addition, the complete sequence of the viral CP coding region (837 bp) was obtained for 16 representative isolates from different cucurbit hosts and different locations, and compared with 66 CP sequences from worldwide isolates available in public databases (Fig. 2). To obtain complete CP sequences, an RT-PCR assay was performed using primers ZYMV-CP-5 and ZYMVdebNC-3 (Table 1) following the protocol mentioned above. To obtain partial sequences in the P3 coding region (1600 bp), RT-PCR reactions were also performed on some isolates using primers ZYMV-1500-5′ and ZYMV-3100-3′ (Table 1).

Figure 2.

Distance trees showing the relationships of 16 ZYMV isolates from Venezuela (in bold) with 66 ZYMV sequences from worldwide isolates available in GenBank using full-length nucleotide sequences of the coat protein coding region. Neighbour-joining trees were built with the program mega. Bootstrap values (500 bootstraps) above 50% are indicated for each node.

Sequence alignment was performed using dambe program and clustalW (Xia, 2000). A phylogenetic tree was constructed from aligned sequences using the neighbour-joining method (500 bootstrap replicates) and Kimura two-parameter correction for multiple substitutions in nucleotide sequences with the program mega v. 3.1 (Kumar et al., 2004). The C-terNIb-N-terCP and CP fragments were analysed for recombination events using gard (genetic algorithms for recombination detection; Kosakovsky Pond et al., 2006). Search for codons under positive selection was performed by estimating the ratio of nonsynonymous to synonymous substitution rates (ω) for each codon using three methods: single likelihood ancestor counting (SLAC), random effects likelihood (REL) and fixed effects likelihood (FEL) (Kosakovsky Pond et al., 2005) available in the HyPhy package (

Biological recovery of ZYMV isolates

Dried leaf tissues from ZYMV-positive samples from 2009 and 2010 (the amount of dried material available was not sufficient for older samples) were used for mechanical inoculations of 4–6 young seedlings (10-day old) of zucchini squash cv. Diamant to recover isolates of the virus. Mechanical inoculations were performed according to Lecoq & Pitrat (1984). Inoculated plants were kept in an ‘S3’ quarantine greenhouse in case mechanically transmissible quarantine viruses were present in the original samples. Four weeks after inoculation, the presence of ZYMV, PRSV, CMV and WMV was tested by DAS-ELISA. Plants that were found to be infected by ZYMV were used as sources for transmission by Myzus persicae, in order to eliminate quarantine begomoviruses, nontransmissible by aphids, which could be present in the samples. Nonviruliferous aphids were starved for at least 1 h before an acquisition access period (AAP) of 1 min on a detached infected leaf. Groups of 10 aphids per plant were then deposited on cotyledons of 10-day old zucchini squash cv. Diamant seedlings and allowed an inoculation access period (IAP) of 2 h before insecticide treatment.

For samples co-infected with ZYMV and another virus, aphid transmissions were performed using a similar protocol to 10 zucchini plantlets, but using only one M. persicae per plantlet. Because very few viral particles are transmitted by a single aphid, this is an efficient way to separate the components of mixed infections. Inoculated plants were tested by DAS-ELISA for the presence of ZYMV, PRSV, WMV and CMV. Leaf tissues from plants infected by ZYMV only were stored as dry material on calcium chloride until use.

Serological variability of ZYMV isolates

Thirty-four ZYMV isolates were recovered, as mentioned above, and mechanically inoculated to 10-day old Diamant seedlings. Four weeks after inoculation, plants exhibiting symptoms were tested by triple antibody sandwich (TAS)-ELISA using a set of 13 monoclonal antibodies (MAbs) developed against ZYMV as previously described (Desbiez et al., 2002; Yakoubi et al., 2008). In addition, all plants were also tested by DAS-ELISA using the polyclonal antiserum against ZYMV obtained at INRA Montfavet as previously described. The serotyping assay was repeated three times independently.

Biological variability of ZYMV isolates

Biological performance of the 34 ZYMV isolates was analysed by using zucchini squash cvs Diamant (susceptible) and Tigress (tolerant), cucumber cvs Beit Alpha (susceptible) and Taichung Mou Gua (TMG) (resistant) and melon cvs Vedrantais (susceptible), Doublon (possessing the Fn gene which produces a rapid lethal wilting reaction with ZYMV isolates belonging to pathotype F) and PI 414723 (possessing the Zym resistance gene) (Lecoq & Pitrat, 1984; Lecoq et al., 1998). Leaf tissue from infected zucchini plants was used to mechanically inoculate plantlets of cucumber, zucchini and melon cultivars as described by Lecoq & Pitrat (1984). Viral symptoms were recorded weekly for 4 weeks (Table S2) and a DAS-ELISA test was carried out 1 month after inoculations. Two replicates were analysed using three plants per test for each cucumber, melon and zucchini cultivar. In addition, a principal component analysis (PCA) was performed to estimate relationships among variables based on biological responses to ZYMV isolates of each cucurbit cultivar using InfoStat v. 2011 (InfoStat group, National University of Córdoba, Argentina).


Virus detection and cucurbit virus frequency

From all samples showing virus-like symptoms, 95% (270/284) were positive for at least one virus. There was a clear association between symptoms observed in the fields and the viral disease presence. Negative samples were also analysed by RT-PCR using additional primers for CYSDV, as well as universal primers for potyviruses and some flexiviruses. No evidence for the presence of viruses from these groups was found (data not shown). Cucurbit samples were positive for five out of seven viruses tested by RT-PCR/PCR (Table 2). The most prevalent virus was MeCMV (65·8%) mainly infecting melon and watermelon. Furthermore, the satellite MeCMA associated with MeCMV (Romay et al., 2010) was detected in 146 out of 187 samples that were positive for this virus. Based on enzyme restriction profiles, no other begomovirus was found in the survey. All samples positive for begomovirus using universal primers were also positive for MeCMV when using specific primers for both DNA components of this virus (data not shown). The potyviruses PRSV and ZYMV were present in similar proportion (34·2 and 32·4%, respectively), whilst WMV, the other potyvirus tested, was detected in a very low percentage (1·1%). CMV was detected at a rather low frequency (6·7%), as expected in tropical conditions. Neither CABYV nor SqMV infection was observed in the cucurbit samples collected. ELISA tests against MNSV and CGMMV infection were negative for the 99 samples tested, whilst ELISA tests for ZYMV, PRSV, CMV and WMV were in accordance with the results obtained by RT-PCR (data not shown).

Table 2. Occurrence of viruses infecting cucurbits from Venezuela
Virus detectedCucumis melo n = 130bCitrullus lanatus n = 54Cucurbita moschata n = 54Cucurbita pepo n = 7Cucumis sativus n = 9Wild cucurbitsa = 30Total = 284
  1. a

    Cucumis anguria, Cucumis dipsaceus, Cucumis melo var. agrestis.

  2. b

    Number of plants sampled.

  3. c

    Percentage of positive plants.

No virus detected4·63·73·728·611·16·75

Single infections were more frequent than mixed infections (56·4 and 38·6%, respectively), reaching 75·4% in melon samples (Table 3). MeCMV was found infecting 187 cucurbits; 63 and 37% of MeCMV infections were single and mixed infections, respectively. In contrast, PRSV and ZYMV, the other most prevalent viruses, showed single infections in only 18·5 and 13% of samples positive for each virus, respectively. Infections of PRSV together with ZYMV were the most frequent mixed infection for all samples (14·1%), reaching up to 50% in squash (Table 3). Twenty-six of the 270 positive samples were wild cucurbits belonging to three species: C. anguria, C. dipsaceus, and C. melo var. agrestis, which were found either at the edge or within cucurbit fields. PRSV was the most common virus detected in the three species of wild cucurbits collected. Furthermore, CMV, an infrequent virus in cultivated cucurbits in this survey, was relatively common in wild cucurbits (Table 2). Infections of either four or five viruses in conjunction were rare and only found in wild cucurbits (Table 3).

Table 3. Occurrence of mixed and single infections of viruses infecting cucurbit species sampled in Venezuela
Virus detectedCucumis melo n = 130bCitrullus lanatus n = 54Cucurbita moschata n = 54Cucurbita pepo n = 7Cucumis sativus n = 9Wild cucurbitsa = 30Total = 284
  1. a

    Cucumis anguria, Cucumis dipsaceus, Cucumis . melo var. agrestis.

  2. b

    Number of plants sampled.

  3. c

    Zucchini yellow mosaic virus.

  4. d

    Papaya ringspot virus.

  5. e

    Watermelon mosaic virus.

  6. f

    Cucumber mosaic virus.

  7. g

    Melon chlorotic mosaic virus.

  8. h

    Percentage of positive plants.

Single infection75·451·942·542·811·123·356·4
No virus detected4·63·73·728·611·16·75

Although sample collection was distributed across 11 states of Venezuela, Zulia, Guárico, Lara and Falcón were the principal states involved in this survey because of their importance for the national cucurbit production. Interestingly, ZYMV and PRSV were rarely found in the plants tested from Falcón state (7 and 4·6%, respectively), as shown in Table S1, regardless of the fact that traditional intensive melon production is developed in this state for the national and international market. Meanwhile, in the same state, MeCMV was present in almost all samples positive for viral infection (in 36 out of 37 positive samples). Overall, the results showed a predominance of MeCMV in melon and watermelon (>80% in both cases), while in squash ZYMV and PRSV infections were more frequent (c. 70%; Table 2).

Molecular variability of ZYMV isolates

Eighty-nine ZYMV isolates were partially sequenced and the sequences were deposited in the GenBank database (accession numbers JX310103 to JX310118 and JX414042 to JX414114). Sequence analysis, using C-terNIb-N-terCP coding region, revealed nucleotide identity of 97–100%, suggesting a low genetic diversity among isolates of ZYMV found in Venezuela. Although the survey spanned major regions of cucurbit production, including cultivated and wild cucurbits, no evidence of a distribution pattern was observed. Nucleotide sequence analysis of the full-length ZYMV CP coding region of 16 Venezuelan isolates (accession numbers JX310103 to JX310118) confirmed the results obtained from sequencing of the C-terNIb-N-terCP region. ZYMV isolates found in this survey clustered into subcluster I of the cluster A (Fig. 2) which is apparently the most widespread throughout the world (Desbiez et al., 2002; Coutts et al., 2011; Lecoq & Desbiez, 2012). No evidence for recombination events was found among ZYMV Venezuelan isolates either in the C-terNIb-N-terCP sequence or in the CP complete sequence. When searching for positive selection, similar results were obtained with REL, FEL and SLAC methods. With SLAC, 13 negatively selected sites and no positive selection was detected at a significance level of 0·1 in the C-terNIb-N-terCP region for 89 isolates, with an average ω value of 0·19. When analysing the complete CP nucleotide sequence of 16 isolates only four negatively selected sites and no positively selected sites were observed (at 0·1 significance level) and the average ω value was 0·24.

Serological variability of ZYMV isolates

Using 13 MAbs against ZYMV, three different serotypes could be distinguished among the 34 ZYMV isolates tested. Interestingly, these serotypes did not correspond to the 16 serotypes previously described worldwide (Desbiez & Lecoq, 1997; Yakoubi et al., 2008). Thus, three new ZYMV serotypes are proposed, namely serotypes 17, 18 and 19 (Table 4). The frequency of serotypes 17, 18 and 19 was 79, 18 and 3%, respectively (Table 4). Isolates belonging to serotype 17, the most abundant, were present in all states included for serological analysis (Aragua, Bolívar, Guárico, Lara and Zulia). Meanwhile, isolates belonging to either serotype 18 or serotype 19 were only present in Zulia state.

Table 4. Serological characterization of ZYMV isolates from Venezuela using a set of 13 monoclonal antibodies
SerotypeAB6CC11CE11CH10DD2DE6AE11CC1BC2AF4DD3ED3BG1Frequency in Venezuela (%)Geographical origin References
  1. a

    Serotype 1 described in Desbiez & Lecoq (1997).

  2. b

    Absorbance at 405 nm (A): 0, A<0·1; +, 0·1>A>0·5; ++, A>0·5.

1a++b+++++++++++++++++++++++0France, UK, Italy, Greece, Spain, Pakistan, Sudan, Algeria, Syria, Israel, Australia, USA, TunisiaDesbiez & Lecoq, 1997; Yakoubi et al., 2008
7++0++++++++00++++++++++79VenezuelaThis work
18+0++++0++00++++++++++18VenezuelaThis work
19+++++0++0++0+++++++++3VenezuelaThis work

Biological variability of ZYMV isolates

Thirty-four isolates of ZYMV, one from 2009 (VE09-091) and 33 from 2010, were recovered after mechanical inoculation of dried leaf tissue followed by aphid transmission (Table S2). Most of the 34 ZYMV isolates induced mosaic and leaf deformation on Diamant, Beit Alpha and Vedrantais cultivars of squash, cucumber and melon, respectively (Table S2). In contrast, no isolate was able to produce symptoms on TMG, a ZYMV-resistant cucumber. ELISA tests were also negative when using the apex of inoculated TMG plants. Twenty-six out of 34 isolates produced few or very scarce systemic chlorotic spots on plants of zucchini squash cv. Tigress, a ZYMV-tolerant cultivar. Eight isolates induced mosaic symptoms in this zucchini squash cultivar. In plants of melon cv. Doublon, seven isolates generated a lethal wilting reaction in this cultivar, which possesses the Fn gene. Plants of melon accession PI 414723, possessing the Zym resistance gene developed systemic chloro-necrotic spots with 33 of 34 isolates indicating that most ZYMV isolates studied in this work belong to pathotype 1. Only one isolate produced symptoms of mosaic in melon PI 414723, and it was therefore classified as pathotype 2. PCA based on biological responses of cucurbit cultivars to each ZYMV isolate did not reveal evidence of clustering among isolates from different regions or different hosts (data not shown).


Most samples in this survey were infected with at least one virus (95%). However, 5% of samples remained negative, even when using additional sets of primers designed to detect other cucurbit viruses. Symptoms observed in negative samples could be related to other causes such as nutritional deficiency, ageing of plants or phytotoxicity rather than viral infections, or possibly to other viruses not tested for.

MeCMV was the most common virus detected in melon and watermelon plants sampled throughout the country (>80% of positive samples), whereas PRSV and ZYMV were the most common viruses detected in squash, zucchini and cucumber (65·7% of positive samples for both cases). In the late 1960s, PRSV was the most prevalent virus not only in squash, but also in melon and watermelon fields of Venezuela (Lastra, 1968). The present study reveals a shift in virus prevalence with the emergence of MeCMV in melon and watermelon and also of ZYMV in squash and cucumber. Interestingly, PRSV still remains the most common virus detected in wild cucurbits, possibly indicative of a long co-adaptation to the local flora, although no data are available on this point. The low prevalence of MeCMV as compared to potyviruses in squash, zucchini and wild cucurbits could be related to a poor adaptation of MeCMV to these hosts, and/or to the fact that squash and zucchini are usually grown in Venezuela in areas or seasons with environmental conditions more favourable to aphids than to whiteflies. A similar shift in virus populations in melon and watermelon has been observed in Florida (USA) where aphid-transmitted potyviruses were the most frequent cucurbit viruses in the early 2000s (Webb et al., 2003), whilst nowadays whitefly-transmitted viruses are more important (Adkins et al., 2011). During the 1980–90 decade, high populations of Bemisia tabaci were observed in Venezuela (Geraud-Pouey et al., 1997) as well as in Florida (Oliveira et al., 2001). Furthermore, a study of whitefly and aphid populations conducted in melon fields of Venezuela showed that whitefly populations were six-fold greater than aphid populations in insecticide-free conditions and up to 70-fold under insecticide-treated conditions (Geraud-Pouey et al., 1997). It has also been shown that the B biotype of B. tabaci is the most widespread and dominant biotype in the main agricultural regions of Venezuela (Romay et al., 2011). Seventy-eight percent of the samples positive for MeCMV were also positive for the satellite MeCMA, confirming their frequent association (Romay et al., 2010). The satellite was not found in MeCMV-negative samples. Forty-one samples were negative for the alphasatellite but positive for the virus, indicating that infection is possible in the absence of the alphasatellite. However, satellite molecules may also be highly variable and additional sets of primers may be needed to detect them all. Development of infectious clones for both MeCMV and its satellites is required to elucidate their relationship and their impact in the infection process.

CMV was found at a low frequency, as previously observed in Venezuela and other countries in subtropical and tropical regions of America (Lastra, 1968; Yuki et al., 2000; Webb et al., 2003; Félix-Gastélum et al., 2007). WMV was infrequently detected in this study (c. 1%) in contrast to PRSV and ZYMV, as expected in a tropical climate. In Florida (USA) WMV was more prevalent than PRSV in the northern region (subtropical climate), whilst PRSV was more prevalent in the southern region (tropical climate; Webb et al., 2003). Double infections with PRSV and ZYMV were quite prevalent, probably related to the fact that they are transmitted by the same aphid vectors and so may be spread simultaneously. SqMV, whose presence in cucurbit fields has been previously established in Venezuela (Lastra, 1968), was not detected in this work although the survey involved samples from 2001 to 2010, as well as regions where SqMV had been reported (Lastra, 1968; Hernández et al., 1989). In South America, SqMV has been reported in Central valleys of Chile causing significant yield reduction in melon fields (Prieto et al., 2001) but it was rarely found in northeastern Brazil (Moura et al., 2001) where tropical conditions are similar to those of Venezuela. Interestingly, CABYV was not detected in this survey. This virus is probably one of the most common cucurbit viruses worldwide (Lecoq, 1999, 2003) and it can infect a variety of weeds that may be reservoirs of the virus (Lecoq et al., 1998). CABYV has been found in California (USA) where it is widespread (Lemaire et al., 1993), and detected in samples from Brazil and Honduras (H. Lecoq, C. Wipf-Scheibel & C. Desbiez, INRA Pathologie Végétale, Monfavet, France, unpublished data). In contrast, despite extensive surveys, CABYV has not been detected in the Caribbean islands of Martinique and Guadeloupe (H. Lecoq, C. Wipf-Scheibel, P. Millot & C. Desbiez, INRA Pathologie Végétale, Monfavet, France, unpublished data).

In the second part of this work, the molecular, serological and biological variability of 34 Venezuelan isolates of ZYMV was studied. Molecular analysis of the C-terNIb-N-terCP-coding region revealed a low genetic diversity among Venezuelan isolates. This suggests that a unique putative introduction event of ZYMV occurred in Venezuela, as has been suggested for Martinique (Desbiez et al., 2002). Low molecular divergence was also observed when analysing complete CP nucleotide sequences of 16 representative isolates.

Despite their low genetic variability, ZYMV isolates of Venezuela showed a relatively high serological variability when analysed with a set of 13 MAbs: they clustered in three serotypes (serotypes 17, 18 and 19), not observed previously among worldwide isolates. For MAbs, whose corresponding epitopes are known (Desbiez & Lecoq, 1997), antigenic reactivity of Venezuelan isolates was correlated with changes in the amino acid sequence within the epitopes (Fig. 3). Nevertheless, the reaction of MAb DE6 with isolates belonging to serotype 17 was not abolished, even if some isolates had a mutation located in this epitope (Fig. 3). Similar results have been observed with a serotype from Tunisia (Yakoubi et al., 2008). Serotype 19 gave a negative reaction with MAb CH10 (for which the epitope is not known) unlike serotype 1, the most abundant serotype throughout the world, as well as the other serotypes previously described (Desbiez & Lecoq, 1997; Yakoubi et al., 2008). The relatively high number of nonsynonymous mutations in the N-terminal part of the CP of Venezuelan ZYMV isolates, most of them being located within epitopes recognized by MAbs, could suggest that positive selection applies to that part of the genome; however, bioinformatic analysis did not reveal any evidence for such selection.

Figure 3.

Amino acid sequence alignment of the N-terminal coat protein region from ZYMV isolates from serotype 1 (reference) and the three serotypes found in Venezuela. The Q/S is the NIb/CP cleavage site. Epitopes corresponding to MAbs CC11/AE11, AB6/DD2 and DE6 are indicated in boxes.

As in previous studies on ZYMV (Desbiez & Lecoq, 1997; Desbiez et al., 2002; Yakoubi et al., 2008), a high biological variability was observed among Venezuelan ZYMV isolates despite their low genetic diversity in the studied genomic regions. There was no correlation between the severities of Venezuelan isolates on different hosts. Thus, some isolates causing mosaic and leaf deformation in susceptible zucchini squash cv. Diamant were not very severe in susceptible melon or cucumber cvs (Védrantais and Beit Alpha, respectively), whereas isolates causing milder mosaic in zucchini squash cv. Diamant caused severe mosaic and leaf deformation in melon cv. Védrantais and cucumber cv. Beit Alpha. In ZYMV-tolerant zucchini squash cv. Tigress, eight isolates induced mosaic symptoms, indicating that they overcame the tolerance. Sequence analysis of the P3 protein from five of these isolates revealed the R to W mutation in the ‘MREK’ motif (data not shown), associated with tolerance breaking (Desbiez et al., 2003). Sequence analysis of the original samples corresponding to these isolates indicated that they originally did not have the R to W mutation. Therefore, the mutation emerged in these isolates within the 4 weeks of incubation in zucchini squash cv. Tigress. This suggests that the tolerance will not be very durable if used in field conditions in Venezuela, unless there is a high fitness cost associated with tolerance-breaking, as observed in France (Desbiez et al., 2003).

In melon PI 414723, one isolate out of 34 overcame ZYMV resistance conferred by the Zym gene (Table S2), and thus it belongs to pathotype 2. A low proportion of pathotype 2 isolates has also been observed in previous studies of ZYMV biological variability (Lecoq & Pitrat, 1984; Yakoubi et al., 2008). Only in Sudan was the frequency of pathotype 2 found to be high in natural ZYMV populations. This was related to the frequent occurrence of wild melons (C. melo var. agrestis) possessing the Zym gene (Mahgoub et al., 1998). However, the finding of at least one isolate overcoming the resistance conferred by Zym, which has not been used in commercial cultivars in Venezuela so far, suggests a potentially low durability of this resistance in the fields. In contrast, cucumber TMG was not infected by any isolate analysed in this study. Hence, cucumber TMG represents a potentially reliable source of ZYMV resistance for future management programmes in Venezuela.

Overall, the present work outlined the current situation of cucurbit viruses in Venezuela, where the use of cucurbit cultivars resistant to ZYMV and/or PRSV may be included in integrated disease management programmes. New studies are required to find sources of resistance to MeCMV for genetic improvement programmes in melon and watermelon.


This work was partially supported by the cooperation agreement between the Venezuelan Foundation Fundayacucho and the Embassy of France in Venezuela with a scholarship to GR.