Vanilla

CymMV is the most prevalent and economically damaging virus infecting orchids (Zettler et al., 1990). Since its first description by Jensen in California (Jensen, 1950; Jensen and Gold, 1955), CymMV has been reported in most cultivated orchid species in many countries and is now considered to be present worldwide. CymMV was first detected in vanilla during a survey in FP (Wisler et al., 1987). It was subsequently found in vanilla plots of many countries: Cook Islands, Fiji, Niue, and Tonga (Pearson et al., 1993), Madagascar (Grisoni et al., 1997), Reunion Island (Pearson, 1997), Mauritius (Rassaby, 2003), India (Bhat et al., 2006) as well as in material conserved in botanical gardens (Grisoni et al., 2007).

Virus Structure and Genetic Diversity

As a member of the genera Potexvirus, family Flexiviridae (Adams et al., 2004), CymMV has flexuous particles (~13 nm × 480 nm) containing a single-stranded RNA (+) genome of about 6.3 kb with five open reading frames (ORFs) flanked by 5′ and 3′ noncoding regions plus a 3′ polyA tail (Francki, 1970; Wong et al., 1997).

The GenBank database contained (in October 2009) the complete genome sequence for nine isolates of CymMV and over 100 sequences for the coat protein (CP) gene, most of which are of Asian origin. The overall diversity between isolates is low at the amino acid level, with less than 14% divergence for CP (Ajjikuttira et al., 2002; Bhat et al., 2006; Gourdel and Leclercq-Le Quillec, 2001; Moles et al., 2007; Sherpa et al., 2006) and less than 3% divergence for RNA-dependent RNA polymerase (Moles et al., 2007). Nucleotide sequence analyses revealed, however, that the CymMV population splits into two diverging haplogroups, which may reflect a dual origin for the isolates found worldwide. However, no biological difference has been observed between the two clusters of strains, which can coexist and recombine within the same host (Sherpa et al., 2007; Vaughan et al., 2008). Strains belonging to both CymMV subgroups have been found in vanilla (Moles et al., 2007).

Symptoms and Diagnosis

In ornamental orchids, CymMV causes chlorotic or necrotic spots and streaks on leaves and flowers (Albouy and Devergne, 1998; Gibbs et al., 2000; Yamane et al., 2008) and reduces plant growth (Izaguirre-Mayoral et al., 1993; Pearson and Cole, 1991; Wannakrairoj, 2008). In vanilla, CymMV infection is generally symptomless but has occasionally been associated with flecking on leaves (on Vanilla planifolia and Vanilla tahitensis) and necrotic spots on the stem and leaves (on V. planifolia) (Grisoni et al., 1997, 2004; Leclercq Le Quillec et al., 2001) (Figure 7.1). It has been suggested from field observations that CymMV weakens the vine and increases decline due to stresses, resulting from overcropping or infection with another pathogen (Benezet et al., 2000), but this remains to be demonstrated. Even in the absence of symptoms, CymMV infection can be responsible for up to 40% reduction of stem growth (Bartet, 2005). The impact of CymMV on aromatic content of the pods has not been investigated so far.

FIGURE 7.1 (See color insert following page 136.) Chlorotic (left) and necrotic (right) flecks induced by CymMV on vanilla leaves.

Observations in field surveys indicated that CymMV symptoms were more severe on V. planifolia (in Indian Ocean area) than on V. tahitensis (in the Pacific). The evaluation of disease severity in comparable greenhouse conditions using a local subgroup A strain in Reunion Island did not show any difference in virus disease between these two Vanilla species. Surprisingly however, a Vanilla pompona accession included in the trial exhibited partial resistance to this virus. This resistance consisted of two components: (1) resistance to inoculation (lower rate of infected plants in mechanical inoculation tests compared to V. tahitensis and V. planifolia accessions, and (2) reduction of virus titer in the infected leaves, resulting several months after inoculation with apparent elimination of the virus (Table 7.1). This resistance evokes transient inhibition of posttranscriptional gene silencing by CymMV in V. pompona rather than a hypersensitive response (Baures et al., 2008), but the mechanism underlying this resistance is unclear and needs further investigation.

Nd, Not determined.

CymMV cannot be reliably detected in vanilla on the basis of symptoms and therefore diagnosis has to be based on detection of the virus. Numerous and innovative molecular techniques have been described for CymMV detection, such as: quartz crystal microbalance-based DNA biosensors (Eun et al., 2002), immuno-capillary zone electrophoresis (Eun and Wong, 1999), DIG-labeled cRNA probes (Hu and Wong, 1998), antisera produced to recombinant capsid proteins (Lee and Chang, 2008), super paramagnetic beads (Ooi et al., 2006), TD/RT-PCR (Seoh et al., 1998), rapid immunofilter paper assay (Tanaka et al., 1997), and wash-free antibody-assisted magnetoreduction assays (Yang et al., 2008).

Since CymMV particles are highly immunogenic (Francki, 1970) and abundantly and evenly distributed in host tissues (Lawson and Hearon, 1974; Leclercq-Le Quillec and Servé, 2001), serological techniques, such as ELISA or DIBA, which are cheap and robust, are particularly suitable for routine diagnosis of the virus. As would be expected from the low amino acid divergence of the CP, monoclonal as well as polyclonal antibodies have detected a wide range of isolates (Lee and Chang, 2008; Vejaratpimol et al., 1998; Wisler et al., 1982). However, Read et al. (2007) reported a strain (Thailand 306) that did not react with a commercial monoclonal antibody. This strain was characterized by aspartic acid instead of asparagine in position 123 of the CP, unique to this isolate.

More specific and sensitive detection of the virus can be achieved using nucleic acids-based methods, and a number of primer pairs and various formats have been designed for RT-PCR detection of CymMV (Barry et al., 1996; Bhat et al., 2006; Lim et al., 1993b; Martos, 2005; Ryu et al., 1995; Seoh et al., 1998; Yamane et al., 2008).

Epidemiology

CymMV incidence in vanilla plots was assessed in several surveys conducted in various agronomic conditions and revealed very disparate situations, not explicable by sampling bias alone. In Comores, for instance the entire planting material seems to be healthy since a recent and extended survey did not find a single virus infected vine (Grisoni and Abdoul-Karime, 2007). In Madagascar, the situation was similar at the beginning of 2000s with the large majority of traditional farmer plantations being CymMV free. However, most of the cuttings from new cultivars propagated in the germplasm repository at Ambohitsara (Antalaha district) were infected (Grisoni et al., 1997; Leclercq-Le Quillec and Nany, 1999, 2000). In Reunion and Society Islands relatively high CymMV incidence was recorded after intensive cultivation programs were initiated in the 1990s (Benezet et al., 2000; Grisoni et al., 2004; Leclercq Le Quillec et al., 2001) and epidemiological analysis could trace back the propagation of the virus in vanilla plots.

Owing to the high stability of the virion and its elevated concentration in host tissues, CymMV is readily transmitted by mechanical means (Jensen and Gold, 1955). It can therefore spread efficiently in the field or in green houses in the absence of specific vectors, solely as a result of virus from infected sap entering micro injuries in epidermal cells of the host. This has been specifically demonstrated for vanilla in shade house programs in Reunion Island (Leclercq Le Quillec et al., 2001) and FP (Grisoni et al., 2004). CymMV monitoring in these shade houses showed a rapid spread in large clusters primarily along the rows. These results supported the assumption that CymMV is introduced into shade houses via infected cuttings collected in the field, and then spread from plant to plant mainly by cultural practices (artificial pollination, looping) and possibly by root anastomosis between adjacent vines.

It is most important for vanilla growers to appreciate that the absence of symptoms increases the probability of both primary and secondary dissemination of this virus, and therefore the importance of planting material that has been certified free of the virus.

Control

In the absence of curative means or resistant commercial vanilla varieties, prophylaxis is the only approach to avoid CymMV losses in vanilla crops. Since humans are the principal vector of the virus (by planting infected cutting or as a mechanical transmission agent) a simple strategy combining plant phytosanitary certification and good management practices is sufficient to prevent the disease. This was successfully implemented in Reunion Island and FP in new plantation programs at the beginning of the 2000s (Benezet et al., 2000; Richard et al., 2009).

A basic requirement of phytosanitary certification is the need for virus-free material, which should be multiplied according to a specific protocol. A healthy starting material can generally be found within the cultivated stock but where a particular variety is required for which only infected germplasm is available, plants need to be cured of the virus. Vanilla is a clonal crop with a high degree of heterozygosity (see Chapter 2), and although seedlings give rise to virus-free vanilla (Jensen and Gold, 1955; Yuen et al., 1979), they are not true to type. Virus elimination has been achieved for several orchid species by chemotherapy in vitro (Albouy et al., 1988; Lim et al., 1993a) and similar strategies should be transferable to vanilla.

Genetic resistance or tolerance is particularly desirable when prophylaxis fails to provide sufficient control of viral disease. Apart from the resistance described in V. pompona, no natural resistance is available in V. planifolia. Pathogen-derived acquired resistance to CymMV has been bio-engendered in some orchid species (Chang et al., 2005; Liao et al., 2004; Lim et al., 1999), and although this strategy has not yet been developed for vanilla, it is theoretically feasible, since regeneration of vanilla protoplasts has been recently achieved (Minoo et al., 2008).

The genus Potyvirus represents the second largest group of plant pathogenic viruses (about 25% of all known plant viruses species), making the potyviruses a worldwide agricultural concern (Fauquet et al., 2005). A potyvirus causing leaf distortion and mosaic in V. tahitensis vines was identified in 1986 in FP (Wisler et al., 1987). This virus, named Vanilla mosaic virus (VanMV), was shown to be serologically, then genetically, related to Dasheen mosaic virus (DsMV) (Farreyrol et al., 2006; Wang and Pearson, 1992; Wisler et al., 1987). Contemporaneously another potyvirus causing necrosis in V. planifolia was identified in 1986 in Tonga and tentatively named Vanilla necrosis virus (VNV) before being characterized as a strain of Watermelon mosaic virus (WMV-Tonga) (Pearson and Pone, 1988; Pearson et al., 1990). The biological properties of WMV-Tonga have been well characterized and its full CP gene sequence determined (Pearson et al., 1990; Wang et al., 1993). Wang and Pearson (1992) demonstrated that WMV-Tonga and DsMV-Vanilla were distinct viruses. DsMV-Vanilla was subsequently detected serologically in V. tahitensis in the Cook Islands, and in V. planifolia in Fiji and Vanuatu (Pearson et al., 1993), and WMV was detected in V. planifolia in FP (Grisoni et al., 2004). Several potyvirus isolates that did not react with either WMV or DsMV antisera were also reported (Grisoni et al., 2004; Pearson, 1997; Pearson et al., 1993), suggesting that other poty-viruses were present in vanilla crops.

The Viruses

The Potyvirus genus (family Potyviridae) encompasses 129 recognized species plus about 15 tentative species (Carstens and Ball, 2009; Fauquet et al., 2005). Potyviruses are found worldwide and infect more than 30 plant families, although individual viruses often have a restricted host range. They have flexuous, nonenveloped, rod-shaped particles (680–900 nm long, 12–15 nm in diameter) that contain a positive-sense single-stranded RNA molecule (Hull, 2002). The monopartite genome is about 10 kb in size and contains a single ORF that encodes a polyprotein of 3000–3300 amino acids. The polyprotein is cleaved co- and post-translationally by three virus-encoded proteinases, into 10 mature proteins, most of which are multifunctional (Adams et al., 2005a; Revers et al., 1999; Urcuqui-Inchima et al., 2001). An extensive study of gene diversity within the family Potyviridae (Adams et al., 2005b) concluded that the cytoplasmic inclusion (CI) protein gene was superior for potyvirus identification when using only a subportion of the genome. Although less informative than the CI, the CP gene can also efficiently separate the species, the demarcation criterion being defined as 76–77% nt identity by these authors.

The CP gene analysis from 36 vanilla samples collected in the Indian Ocean and Pacific regions between 1997 and 2005 added five potyvirus species to the two previously described infecting vanilla, namely: Bean common mosaic virus (BCMV), Cowpea aphid-borne mosaic virus (CABMV), Wisteria vein mosaic virus (WVMV), Bean yellow mosaic virus (BYMV), and Ornithogalum mosaic virus (OrMV) (Grisoni et al., 2006).

Remarkably, among the seven species that infect vanilla, five (BCMV, CABMV, DsMV, WMV, and WVMV) belong to the BCMV subgroup, which preferentially infect leguminous crops and weeds. They were detected in Reunion Island (BCMV and CABMV), in Madagascar (BCMV), in Mauritius (CABMV), in FP (BCMV, DsMV, and WMV), in Tonga (WMV), and in Samoa (WVMV). The two other species, BYMV and OrMV, were only detected in Reunion Island. The mosaic inducing potyvirus reported in India (Bhai et al., 2003; Bhat et al., 2004; Thomas et al., 2002) were not identified but may involve BYMV as suggested by the two GenBank accessions (AY845011 and AY845012) originating from vanilla material collected in Karnataka (India). Potyvirus-like symptoms have also been reported in Papua New Guinea (Kokoa, 2000) and in Puerto Rico (Childers and Cibes, 1948).

The DsMV strains infecting vanilla were at first tentatively referred to as VanMV because they showed only distant serological relationship to dasheen-infecting strains of DsMV and because of the failure of cross-inoculation between the strains originating from the two hosts. Subsequently, the nucleotide analysis of the 3′ end of the genome of two VanMV isolates (from Cook Islands [CI] and FP [FP]) confirmed that they should be classified as DsMV strains, since they shared more than 78% identities, notably in the core CP (Farreyrol et al., 2006). Interestingly, the DsMV-Vanilla [CI] and DsMV-Vanilla [FP] were as divergent from each other as from published DsMV-dasheen (Colocasia esculenta) strains (Farreyrol et al., 2006). However, recently obtained sequences for dasheen isolates from Cook Islands and FP were very similar to vanilla isolates from the same country (M. Pearson et al., unpubl. data). Since dasheen has been grown on these islands for much longer than vanilla, this suggests that “virus jump” from dasheen to vanilla may have occurred independently in the two Pacific Islands. In addition, the DsMV-Vanilla [FP] showed unusual features which makes it a particular virus among potyviruses (Farreyrol et al., 2006): a DVG aphid transmission motif (instead of the more common DAG motif) upstream of an unusual stretch of amino acid repeats (GTN) typical of natively unfolded proteins and an uncommon NIb/CP proteolytic cleavage site (Q//V).

Symptoms and Diagnosis

In vanilla, potyviruses cause a more or less pronounced mosaic and deformation of leaf blades sometimes associated with necrotic lesions (Figure 7.2). The mosaic and necrosis can also be visible on the stem. These symptoms are particularly spectacular on the young leaves and are much attenuated on older leaves. In Reunion Island, potyvirus symptoms were most obvious during the cool months with a remission of symptoms observed during the hot season (Benezet et al., 2000; Leclercq Le Quillec et al., 2001). The impact of the viruses on flowering and fruiting of the vanilla vines has not been established, although some reduction in growth can be inferred from the severe mosaic and deformation, affecting the shoots. Field observations revealed that the virus particles were unevenly distributed in the plant and symptomatic vines may test negative in ELISA (Leclercq Le Quillec et al., 2001). As a consequence, visual observation of mosaic on leaves and ELISA tests are complementary to assess potyvirus in vanilla plots, none of the technique used alone being sufficiently reliable for diagnosis.

FIGURE 7.2 (See color insert following page 136.) Potyvirus symptoms on leaves of V. planifolia (a = BYMV-Réunion, b = BCMV-Madagascar, c = necrotic strain of WMV-Tonga) and V. tahitensis (d = WMV-FP, e = DsMV-FP) and on stem (f = WMV-FP).

Several molecular techniques are available to diagnose potyvirus infection of vanilla, at the generic or specific level. The generic anti-potyvirus monoclonal antibodies (Jordan and Hammond, 1991), commercialized by Agdia (USA), proved to be efficient in detecting all the potyvirus species identified in vanilla by Indirect Simple Sandwich ELISA. Likewise, several motifs conserved in the genome of potyviruses have been used to design a number of degenerate primers for RT-PCR amplification of the corresponding sequence in the potyviruses genome (Chen et al., 2001; Colinet et al., 1998; Gibbs and Mackenzie, 1997; Ha et al., 2008; Langeveld et al., 1991; Marie-Jeanne et al., 2000; Pappu et al., 1993). Both techniques can be used to assess the potyvirus status of material but do not specifically identify the potyvirus species detected.

For potyviruses, specific identification by serology is unreliable because of the numerous cross reactions between virus species (Shukla et al., 1994). To overcome this problem a simple one tube, one-step, RT-PCR assay using degenerate primers followed by direct sequencing of a short fragment can be used (Grisoni et al., 2006). As RT-PCR and sequencing are getting easier and cheaper technologies, this method can be used in epidemiological surveys requiring high throughput. Microarray technologies, which are a promising way to get broad spectrum, specific, and sensitive detection tool for potyviruses, are under development (Boonham et al., 2007; Wei et al., 2009).

Epidemiology

Potyviruses are responsible for major losses of many economically important crops because they spread readily in the fields. Several potyviruses are seed transmitted, most are mechanically transmitted (albeit inefficiently), but all are efficiently transmitted by aphids in a nonpersistent manner (Brunt, 1992).

VanMV was readily transmitted from V. tahitensis to V. pompona using the peach aphid Myzus persicae (Wisler et al., 1987). Experimental transmission tests of WMV-Tonga from infected to healthy Nicotiana clevelandii using Aphis gossypii was successful but transmission tests involving V. planifolia as a donor plant with the same aphid had failed (Pearson and Pone, 1988; Pearson et al., 1990; Wang and Pearson, 1992; Wang et al., 1993, 1997). However, several field data on potyvirus infections support the role of aphids in the spread of potyviruses in vanilla plots (Richard et al., 2009). Successful aphid transmission trials were consistent with a nonpersistent mode of transmission. In nonpersistent transmission, a few seconds or minutes are sufficient for virus acquisition or virus inoculation by the vector in feeding probes. It is therefore not surprising to observe heavy potyvirus infection in the absence of established aphid had colonies on vanilla (excepting Cerataphis orchidarum, for which only wingless forms have been described and which is not considered an efficient virus vector).

A few potyviruses are transmitted by seeds, notably CABMV in several leguminous species (Gillaspie, 2001), BCMV in some Phaseolus and Vigna cultivars, and BYMV in several leguminous species (Aftab and Freeman, 2006). In addition, BCMV is transmitted by bean pollen (Card et al., 2007). DsMV, WMMV, and WVMV are not known to be seed transmissible, but it has been suggested that seed transmission might be more common than currently recognized particularly within the BCMV subgroup (Gibbs et al., 2008a).

Seed transmission of viruses has epidemiological implications through intercrop-ping and weed management. In the field, primary inoculum of potyvirus may come from infected vanilla cuttings or from surrounding plants, either perennial plants or annual weeds, in which the virus can overwinter when it is transmitted by seeds. From this primary inoculum, the potyvirus is then disseminated in the plot by aphids. Molecular similarity between isolates infecting vanilla and weeds provides support to the passage of viruses from one species to another. For instances, in Madagascar, BCMV isolates were identified in vanilla and in the bordering weed Senna sp. (Leguminosae), which were 100% identical in CP gene sequences (Grisoni et al., 2006). In Reunion Island, a rapid outbreak of CABMV in a recently planted vanilla plot using virus-free cutting was correlated to the intercropping with Vigna unguicu-lata in which CABMV is seed transmitted. On the contrary, the use of Gliricidia sepium, a legume tree frequently used as support in vanilla plots and which is, because of frequent pruning, heavily infested by Aphis craccivora aphids (an effi-cient vector of BCMV subgroup of viruses) does not seem to increase potyvirus incidence in vanilla plots compared to when aphid-free support trees are used, such as Leucena glauca (Grisoni et al., 2004). However, in the presence of other virus sources the high aphid populations could become an issue.

Control

Epidemiological data showed that severe potyvirus outbreaks occur primarily in intensive vanilla farming system and may originate from a diversity of virus species involving a diversity of host species, virus reservoirs, and probably aphid species.

Thus, apart from the universal recommendation of planting virus-free material, the control strategy needs to be adapted to each local circumstance. As a consequence, it is advisable to assess potyvirus risk prior to any development program in a new area or when changing the cultivation system. In addition, molecular tools that enable rapid identification of potyviruses help provide insight into the appropriate control strategy to use in case of unexpected potyvirus outbreaks.

CMV is a widespread virus infecting a very broad range of plants. It was identified in vanilla in the early 2000s, associated with severe distortions of vines in FP (Farreyrol et al., 2001). The virus was later found in vanilla in Reunion Island and in India (Madhubala et al., 2005).

Virus and Genetic Diversity

CMV is the type member of the Cucumovirus genus in the family Bromoviridae. It has a tripartite positive-sense RNA genome encoding five proteins (two polymerase components on RNA1 and RNA2, the 2b protein involved in the breaking of plant defense and virus movement in the plant, on RNA2, and a 30 K protein and the CP on RNA3). Each RNA is encapsulated in a distinct polyhedral particle (28 nm diameter) hence the infection of a plant requires the inoculation of the three types of virions. Some isolates also encapsulate a subgenomic RNA or a satellite (sat) RNA. The biology and ecology of this virus have been extensively reviewed by Palukaitis et al. (1992) and Gallitelli (2000).

The CMV strains are divided into two serologically distinct subgroups diverging in about 25% of their nucleotides and also differing in biological traits. The thermo-tolerant strains of subgroup 1 are further divided into two clades (1A and 1B) on the basis of genomic sequences of their RNAs (Roossinck et al., 1999). The subgroup 1A emerged from the subgroup 1B and (until recently) the 1A strains were only found in Asia. The thermosensitive strains of subgroup 2 are restricted to cool climate areas. Three mechanisms have been associated with the evolution of CMV: RNA reassortments between the RNA components, recombination in the noncoding regions and mutations in the coding regions. Several amino acid positions on the CP were shown to be positively selected in relation to virus transmission (Moury, 2004).

To our knowledge, 65 CMV–CP sequences have been obtained for vanilla isolates, originating from FP (58), Reunion Island (6), and India (1) (Farreyrol, 2005; Madhubala et al., 2005; Mongredien, 2002), 53 of which are deposited in Genbank. The majority of the sequences were most closely related to subgroup 1B isolates and clustered in a few clades (Farreyrol et al., 2009). Trees generated from 3′ UTR of RNA3 (Farreyrol, 2005) and ORF2b of RNA2 (Mongredien, 2002) sequences were congruent with CP analysis. Interestingly, an isolate belonging to group 2 (NZ100, AY861389) was mechanically inoculated in vanilla plants, which induced leaf mosaic and deformation (Farreyrol et al., 2009), suggesting that a wide range of CMV isolates can infect vanilla. Attempts to detect CMV sat-RNA associated to vanilla isolates of CMV from FP by RT-PCR using degenerate primers (Escriu et al., 2000; Varveri and Boutsika, 1999) have failed (M. Grisoni, unpubl. data).

Phylogenetic analysis of CMV isolates collected from vanilla plus several associated crops or weeds revealed a clustering according to the geographic origin rather than to the host plant. For instance in the Leeward Islands (FP), the CMV isolates from the islands of Raiatea and Huahine were in distinct clusters, while high similarities were observed between isolates from vanilla and from Commelina diffusa (Farreyrol et al., 2009). These findings supported the hypothesis of transmission of the virus from one plant to another, particularly from C. diffusa to V. tahitensis. Similarly, the CP gene of CMV from vanilla sequenced in Kerala (India) showed the highest identities (99%) with that of another Indian isolate infecting black pepper in Karnataka (Madhubala et al., 2005).

Symptoms and Diagnosis

A variety of symptoms has been associated with CMV on its numerous hosts; most common are mosaics and stunting, but symptoms can be as severe as complete systemic necrosis. Some strains are symptomless on certain hosts or at elevated temperature (subgroup two strains) and symptom expression may also vary over time due to temporary remission of the virus infection or be attenuated or aggravated by the coinfection with a sat-RNA.

In vanilla, the CMV isolates described in FP induced severe leaf and shoot deformation and the infected plants are often sterile, either because the flowers do not fully develop or because they are abnormally formed and contain no viable pollen. On V. tahitensis vines mechanically inoculated with a field isolate of CMV the first foliar symptoms appeared two months after inoculation with an embossing of the young leaves, which usually had a slender and asymmetrical (comma-like) shape (Figure 7.3a). The vines subsequently developed an abnormal phyllotaxy and irregular growth of the apices and flowers (Figure 7.3b and c), typical of the disease. Similar leaf symptoms were described in India and the Reunion Island on V. planifolia. In a few instances, shoot deformations on CMV infected vines have been observed (Farreyrol, 2005) resembling the abnormal tubular leaves described by Jacob de Cordemoy (1899) (Figure 7.4), which are possibly historic evidence of remote infections by CMV in the Reunion Island.

FIGURE 7.3 Symptoms caused by CMV infection in V. tahitensis: (a) young leaves embossed and deformed (healthy leaf on the right); (b) proliferation on infected young shoot; and (c) abnormal flower (right) on CMV-infected vine.

FIGURE 7.4 (Left) Abnormal tubular leaf a, observed on V. planifolia in 1899 by M.H. Jacob de Cordemoy (Jacob de Cordemoy, 1899); f, normal leaf; p, peduncle. (Right) Tubular leaf symptom observed during a survey in FP in 2005.

CMV symptoms are spectacular on vanilla and rather specific and can be good indicators of virus infection. However, these symptoms can be confused, in the early stages, with potyvirus infection, and in the later stages, with physiological disorders. In addition, temporary symptom remission may occur. Therefore, tests based on serology (Devergne et al., 1981; Hu et al., 1995; Maeda and Inouye, 1991; Yu et al., 2005) or nucleic acid amplification (Hu et al., 1995; Wylie et al., 1993; Yu et al., 2005) are recommended to diagnose the disease, or to select virus-free cuttings.

Epidemiology

CMV is a plant virus having the largest known host range of more than 1000 species belonging to more than 30 families (Douine et al., 1979). The virus is transmitted by more than 80 aphid species in a nonpersistent manner (Gallitelli, 2000). In several hosts, CMV is also transmitted at high frequency through the seeds (Palukaitis et al., 1992). Several studies have pointed out the key role of the co-occurrence of reservoir plants (that overwinter the primary inoculum) and aphids (that ensure secondary spread of the virus) in the epidemics of CMV in different crops and countries (Hobbs et al., 2000; Kiranmai et al., 1998; Lavina et al., 1996; McKirdy, 1994; Rist and Lorbeer, 1991; Skoric et al., 2000).

In vanilla plots of the Society Islands (FP), high prevalence of CMV was recorded in the early 2000s, with one out of three plots infected, and an incidence frequently exceeding 25% of vines (Farreyrol et al., 2009). Much lower virus prevalence was observed in the Reunion Island with only 2 infected plots out of 25 with sporadic occurrence of the virus (Farreyrol, 2005) as well as in India where mosaic viruses were observed in <5% of vines (Bhat et al., 2004; Madhubala et al., 2005). In Marquisas and Australes (FP), Comoros archipelagos, and Madagascar no CMV was found in vanilla despite intensive surveys (Grisoni, 2003; Grisoni, 2009; Grisoni and Abdoul-Karime, 2007; Leclercq-Lequillec and Nany, 2000).

The spread of CMV was monitored in several shade houses planted with virus-free cuttings and conducted in various locations in the Leeward Islands (Richard et al., 2009). The results showed that in the absence of insect-proofing and in favorable locations CMV can spread very quickly, with 50% of vines infected only two years after planting. In contrast, with insect proofing or in another location, a very low incidence of CMV was recorded. These results demonstrated the role of aphids in spreading the disease and also the importance of the environment possibly through the presence or absence of CMV reservoir among weeds.

C. diffusa (common name climbing or spreading dayflower, ma’apape in Tahitian or herbe d’eau in French, Figure 7.5) was shown to play a predominant role in CMV epidemics in FP (Richard et al., 2009): (1) C. diffusa was the only species that exhibited simultaneously high frequencies of CMV infection and aphid (A. gossypii) infestation, (2) the occurrence of CMV in vanilla vines was highly correlated with the proximity of C. diffusa, (3) the incidence of CMV in the plots surveyed was significantly correlated with the occurrence of infected C. diffusa within the plot but not with the other CMV-infected species, and (4) the ability of A. gossypii and A. craccivora to transmit CMV from C. diffusa to young plants of V. tahitensis was demonstrated in laboratory tests. All these data, along with the high sequence similarities between vanilla and C. diffusa CMV isolates (Farreyrol et al., 2009) strongly support the view that C. diffusa is a major active reservoir of CMV and contributes greatly to the spread of this virus in vanilla plots. This epidemiological scenario resembles that described for CMV in banana plantations (Eiras et al., 2004; Lockard, 2000; Magnaye and Valmayor, 1995).

FIGURE 7.5 C. diffusa Burm. f., the main aphid and virus reservoir involved in CMV epidemics in FP: (a) virus-infected plants; (b) flower; and (c) aphid-infested leaf.

Control

In view of the very severe symptoms induced on leaves and flowers and its potentially high incidence in the field, CMV is a major threat to vanilla plantations and should be a priority for risk assessment and control when implementing intensive cultivation systems.

In FP, an efficient control strategy was set up in 2003 relying on (1) a supply of virus-free cuttings, (2) roofing the shade houses with insect-proof netting, (3) roguing the diseased plants as soon as they are detected, (4) decontaminating tools, and (5) avoiding weeds (particularly C. diffusa) inside and at the proximity of the plantation. A postimplementation survey in 2007 showed that the viral impact (CMV and other viruses) was drastically reduced in the new plantations that followed the recommendations (Richard et al., 2009).

Epidemiological data indicate that CMV prevalence may vary greatly from one area to another, notably in relation to the presence of virus reservoirs. Therefore, control strategies have to be adapted according to each specific cultivation context.

Besides the major viruses already detailed other viruses with minor economic importance such as Odontoglossum ringspot virus (ORSV) or barely characterized such as a Rhabdolike-virus and a Clostero-like viruses (Bhat et al., 2004; Pearson et al., 1993) have also been reported in vanilla.

Odontoglossum Ringspot Virus (Orsv)

ORSV is a member of the Tobamovirus genus of which the type member is Tobacco mosaic virus (Jensen and Gold, 1951). The rod-shaped particles of ORSV (300 × 18 nm) encapsulate a single-stranded positive RNA molecule of approximately 6.6 kb coding for five proteins (Ryu et al., 1995). ORSV’s natural host range is limited to orchids in which it may cause severe symptoms such as mosaic, ringspots, necrosis on leaves and flowers (Albouy and Devergne, 1998; Gibbs et al., 2000). As a tobamovirus, ORSV has a high stability in sap (temperature inactivation of 90°C and a dilution endpoint of 10⁻⁵–10⁻⁶) and is readily transmitted by mechanical means. It is frequently found in cultivated ornamental species worldwide (Freitas et al., 1999; Khentry et al., 2006; Zettler et al., 1990).

On vanilla, however, the virus has a very limited incidence and prevalence. It has been serologically detected in few vanilla vines in Tonga, Fiji, Cook Islands, Niue, FP, and the Reunion Island (Farreyrol et al., 2001; Pearson and Pone, 1988; Pearson et al., 1993; Wisler et al., 1987), but no symptoms were consistently associated with the virus. However, in plant quarantine in the Reunion Island, we recently observed a V. pompona vine originating from a botanical garden that exhibited mosaic on young leaves (Figure 7.6). This plant tested positive for ORSV and negative for CymMV, potyviruses, and CMV, by ELISA and RT-PCR.

FIGURE 7.6 Mosaic on leaf of V. pompona infected by ORSV.

Since it is relatively infrequent in vanilla, ORSV has so far been of little concern in production plots. However, due to its possible pathogenicity and absence of a natural vector it is worth testing the planting material for this virus in order to avoid its propagation. As with CymMV, a number of innovative methods have been developed to detect ORSV.

Unidentified Rhabdo-Like Virus

During a survey conducted in Pacific Islands (Pearson et al., 1993), unusual symptoms on leaves, consisting of necrotic spots were observed in vanilla plots (Figure 7.7a) in Vanuatu and Fiji. Leaf dip preparations from symptomatic leaves revealed the presence of enveloped bacilliform particles suggesting a possible Rhabdovirus.

FIGURE 7.7 (See color insert following page 136.) (a) Enations and necrosis on V. planifo-lia leaves infected with (b) enveloped virus-like particles visible under electron microscope.

The Rhabdoviridae family (order Mononegavirales) contains viruses infecting animals and plants, which are transmitted by arthropods and may multiply in the vector (Fauquet et al., 2005). The plant infecting rhabdoviruses are assigned to two genera (Cytorhabdovirus and Nucleorhabdovirus) but a number is still unassigned, including the putative rhabdovirus Orchid fleck virus (OFV), which has become more prevalent in orchids over the last decade (Kitajima et al., 2001; Kondo et al., 2006). OFV has been tentatively classified in a new Dichorhabdovirus genus because its genome, although showing sequence similarities with nucleorhabdoviruses, is bipartite and the particles are not enveloped (Kondo et al., 2009). This last feature contrasts with the putative rhabdovirus particles seen in vanilla which are clearly enveloped (Figure 7.7b). In addition, repeated attempts to amplify OFV sequences from vanilla symptomatic leaves using specific primers (Blanchefield et al., 2001) have failed (Mongredien, 2002). Since the statement from 1993, no further damage has been reported in vanilla that is reminiscent of this still uncharacterized putative rhabdovirus, and this virus should be considered as anecdotal in vanilla plots.