Verticillium wilt of olive: a case study to implement an integrated strategy to control a soil-borne pathogen

Verticillium dahliae shows a wide morphological and physiological diversity: broad differences in shape and color of colonies develop in different culture media and variable types of conidiophores and MS can be found (Pegg and Brady 2002 and references therein). However, morphological criteria cannot solely be used to establish a suitable distinction among subspecific groups which may display, for instance, differential pathogenicity or virulence. Thus, diverse studies have indicated that morphology of MS produced by V. dahliae on different culture media might be correlated with the virulence of the isolate. For instance, López-Escudero and Blanco-López (2005c) reported significant differences in the average length/width ratio among MS produced by different isolates on modified sodium polypectate agar (Butterfield and DeVay 1977), a common culture medium used to assess pathogen inoculum density in soil (see below). Indeed, MS produced by highly-virulent isolates were elongated and showed an average length/width ratio of 4.06, whereas those produced by mildly-virulent isolates were more rounded and showed an average ratio of 2.92. Verticillium dahliae isolates causing a defoliating syndrome in cotton and olive can detoxify the benzylisoquinoline alkaloid sanguinarine, and have a growth optimum temperature of 24–27°C, whereas isolates that do not induce defoliation in these hosts are not able to detoxify this compound and their optimum temperature ranged between 21 and 24°C (Bejarano-Alcázar et al. 1996).

Severity of VWO attacks depends on the virulence of the V. dahliae pathotype that infects the tree. Thus, V. dahliae isolates infecting olive are traditionally classified into defoliating (D) and non-defoliating (ND) pathotypes, according to their ability to cause the complete fall of green leaves (Rodríguez-Jurado et al. 1993). Interestingly, differential virulence displayed by isolates infecting cotton (Gossypium hirsutum L.) was also described in olive (Schnathorst and Sibbett 1971a). Thus, an isolate recovered from cotton or olive displays cross virulence in both crops. The D pathotype is highly virulent and causes severe symptoms including wilt, chlorosis, defoliation, drastic weight and height reduction and, eventually, death of the tree (Blanco-López et al. 1984; Rodríguez-Jurado et al. 1993; López-Escudero and Blanco-López 2001; Birem et al. 2009). The ND pathotype causes nearly the same symptoms in olive as D-pathotype infections, although at milder (defoliation of green leaves) or moderate (die-back, chlorosis or necrosis) levels. Similarly, it induces reduction in weight and, to a lesser extent, in height under favorable environmental conditions (López-Escudero and Blanco-López 2001; López-Escudero et al. 2004, 2007a; Martos-Moreno et al. 2006). Differences in virulence seem to be consistent among olive cultivars (López-Escudero et al. 2004, 2007a; Martos-Moreno et al. 2006; Dervis et al. 2010).

The broad host range and apparent low host specificity suggests that V. dahliae populations contain low pathogenic and genetic diversity. However, V. dahliae virulence variability is apparent by the aforementioned presence of the D and ND pathotypes infecting both cotton and olive. Isolates of the D pathotype are more virulent than ND isolates (Schnathorst and Mathre 1966a; Bejarano-Alcázar et al. 1996, 1997; López-Escudero and Blanco-López 2001, 2007a). Consequently, olive and cotton cultivars tolerant to the less virulent pathotype (ND) are severely affected by the more virulent one (D), and epidemics caused by the latter start earlier and develop faster and more severely.

Therefore, understanding the genetic and pathogenic diversity of V. dahliae populations is of relevance for the effective implementation of an integrated disease management strategy. For instance, the successful use of resistant cultivars in a given geographical area can be compromised by the variability of the pathogen population structure (i.e., presence of different races or pathotypes) found in that region. Thus, the optimal use of resistant genotypes and/or the efficient application of additional disease control measures need of an appropriate knowledge on the structure, history, and evolutionary potential of the pathogen populations (McDonald and Linde 2002).

The possibility to find a correlation between V. dahliae VCGs and differential virulence in relevant crops would be valuable for disease prediction and breeding for resistance (Katan 2000b). So far, however, contradictory results have been reported and no definitive conclusions can be drawn. Indeed, some studies point to a worldwide, prevalent distribution of VCG1, VCG2, and VCG4, suggesting a limited VCG diversity. Other investigations suggested that some VCGs may prevail on a crop in a given geographic area (Joaquim and Rowe 1991; Strausbaugh 1993; Katan 2000b; Subbarao et al. 1995; Dobinson et al. 1998; Elena 2000; Korolev et al. 2000; Bhat et al. 2003; Hiemstra and Rataj-Guranowska 2003). On the other hand, considering VCGs as fully genetically isolated groups may be a strict statement. In fact, the existence of weak complementation and formation of heterokaryons between isolates belonging to different VCGs (or certain subgroups), or even between different Verticillium species, has been described (Daayf et al. 1995; Hiemstra and Rataj-Guranowska 2003; Joaquim and Rowe 1990; Strausbaugh et al. 1992). This may suggest the existence of a continuum of genetic variation surpassing the genetic barrier of the VCG. This can be further supported by recent findings suggesting that parasexuality within the genus Verticillium can be an important mechanism to enhance genetic variability. Thus, the appearance of interspecific hybrids may contribute to diversity and evolution, and more importantly, could be of outstanding significance for the emergence of novel traits such as new (or modified) host adaptation and/or enhanced virulence to host genotypes (Collado-Romero et al. 2010).

Examing Verticillium wilts from a disease control perspective, the relationship between pathogenicity and the presence of a particular V. dahliae genetic group (VCG) linked in a geographical context or host plant can be crucial. However, a clear-cut association between isolates of a particular VCG and their pathogenicity or virulence in a given host plant is not always easy to ascertain. Thus, some studies support the correlation between VCG assignment of isolates and their differential virulence displayed in a given plant (Joaquim and Rowe 1990; Bhat and Subbarao 1999; Tsror et al. 2001; Korolev et al. 2001), whereas others conclude that different isolates from the same VCG show different degrees of virulence in a same crop (Elena 1999; Korolev et al. 2000, 2001). Interesting examples pointing to an association between genetic groups and host specificity are: i) the predominance of isolates assigned to VCG4 in potato (Solanum tuberosum L.) infections in the United States and Israel (Joaquim and Rowe 1991; Strausbaugh et al. 1992; Korolev et al. 2000); ii) the aforementioned association of the highly-virulent, D-pathotype isolates of cotton (Gossypium hirsutum L.) with their assignment to VCG1A described in various areas such as Central Asia, China, Israel, Peru, Spain, Turkey and USA (Daayf et al. 1995; Korolev et al. 2001; Zhengjun et al. 1998; Korolev et al. 2008; Dervis et al. 2008), in what may constitute an example of dispersal of isolates from an original focus of infected cotton genotypes; or iii) the prevalence of VCG2A and 2B isolates infecting artichoke (Cynara scolymus L.) in a defined artichoke-cultivating region of eastern-central Spain, with VCG2B being the most virulent in this host (Jiménez-Díaz et al. 2006). Studies regarding to olive-infecting V. dahliae isolates, are indicating the following scenario: 1) highly-virulent isolates, inducing severe syndromes in olive, generally belong to the D pathotype; 2) all olive D-pathotype isolates have so far been assigned to VCG1A in all countries where this group/pathotype has been reported (Pérez-Artés et al. 2000; Collado-Romero et al. 2006; Dervis et al. 2010); and 3) olive isolates assigned to other VCGs (namely, VCG2 and VCG4) have been characterized as ND pathotype either by pathogenicity testing or by PCR-based molecular pathotyping (Fig. 4). However, a correlation has not always been reported between VCG and pathogenicity in olive isolates (Tantaoui et al. 2002), or a continuum of virulence has been found within the VCG1A/D pathotype group, revealing that severity of symptoms caused by some D isolates was not significantly different to that induced by ND isolates in artificial inoculation experiments (Dervis et al. 2010). Thus, despite VCG assignment provides useful information on the diversity of V. dahliae populations, it could be insufficient to correlate genetic diversity with traits such as level of virulence or host specificity. Therefore, more refined tools such as molecular methodologies are necessary. They provide higher resolution than VCG analysis to unravel the complex structure of V. dahliae populations.

The diversity and the structure of V. dahliae populations have also been studied using different neutral molecular markers that are best suited to make inferences on migration and drift genotype. For strictly asexual-reproducing fungi, these markers also assist in estimating the variability in ecologically-important characters (i.e., pathogenicity, virulence). Thus, characterization and genetic differentiation of V. dahliae (and/or other Verticillium spp.) isolates has been carried out by restriction fragment length polymorphism (RFLP) (Carder and Barbara 1991; Typas et al. 1992; Carder et al. 1994; Okoli et al. 1994; Dobinson et al. 2000), random amplified polymorphic DNA (RAPD) (Messner et al. 1996; Koike et al. 1996; Zeise and von Tiedemann 2002; Bhat and Subbarao 1999), and amplified fragment length polymorphism (AFLP) (Collins et al. 2003; Fahleson et al. 2003; Radišek et al. 2003; Collado-Romero et al. 2006). Analysis and comparison of DNA sequences such as the intergenic spacer (IGS) and internal transcribed spacer (ITS) regions of genes encoding ribosomal RNA, as well as repeated DNA sequences, have also been used to study variation within V. dahliae (Morton et al. 1995a, b; Subbarao et al. 1995; Dobinson et al. 1998; Pramateftaki et al. 2000).

Olive-infecting V. dahliae isolates have been studied using RAPD (Cherrab et al. 2000; Pérez-Artés et al. 2000; Lachquer and Sedra 2002; Bellahcene et al. 2005a; Nigro et al. 2005), AFLP (Collado-Romero et al. 2006), and PCR-specific markers (Pérez-Artés et al. 2000; Bellahcene et al. 2005a; Mercado-Blanco et al. 2003b; Collins et al. 2005; Nigro et al. 2005; Collado-Romero et al. 2006; Dervis et al. 2010). RAPD fingerprints have provided variable results, for instance, this approach was successful in correlating the geographical origin of a collection of olive-infecting Moroccan isolates, but not their colony morphology (Cherrab et al. 2000). On the other hand, RAPD fingerprinting showed that olive-infecting V. dahliae isolates in Algeria showed limited diversity, which correlated with the presence of only one VCG, and solely the ND pathotype (Bellahcene et al. 2005a). Previously, no correlation was found between RAPD markers, the geographic origin or the host plants of the isolates examined, although four RAPD groups were described (Lachquer and Sedra 2002). Finally, RAPD fingerprinting indicated a low level of genetic diversity in V. dahliae olive-infecting population in the Apulia region (Italy). No correlation was found among pathogen isolates, olive cultivars and geographical location, and PCR analysis showed that all isolates (with the exception of one inconclusive result) belonged to the ND pathotype (Nigro et al. 2005).

Early studies on molecular variability of V. dahliae populations infecting cotton and olive in southern Spain unraveled the presence of different D- and ND-specific RAPD patterns from which SCARs (“sequence characterized amplified regions”) markers were subsequently identified (Pérez-Artés et al. 2000). These studies already revealed significant molecular diversity in V. dahliae cotton/olive-infecting populations. Indeed, specific RAPD fingerprints were found in all the examined D-pathotype isolates which were assigned to VCG1A, therefore distinguishing this group as genetically homogeneous. However, the rest of isolates evaluated (characterized as ND pathotype) showed greater diversity than the previous group and no specific RAPD pattern could be associated to the different VCGs (VCG2A, 2B or 4B) to which these isolates were assigned (Pérez-Artés et al. 2000). More recently, the molecular variability amongst V. dahliae VCGs infecting artichoke, cotton and olive from diverse regions of the Mediterranean Basin was assessed by AFLP, which offers a much higher resolution than RAPD fingerprinting, and by PCR markers analyses (Collado-Romero et al. 2006) (Fig. 4). The aim of this study was to explore associations between molecular patterns (AFLP and PCR markers) and phenotypic features (virulence), vegetative compatibility diversity, and geographical origin of host plants. The results demonstrated that V. dahliae isolates within a VCG subgroup were similar at the molecular level, and that clustering of isolates correlated with VCG subgroups regardless of host source and geographic origin. An important conclusion to be drawn from this study was that VCGs are clonal groups, each one genetically differentiated and/or isolated from the others. As for olive-infecting isolates concerns, a group exclusively found in artichoke (namely VCG2B³³⁴/2β³³⁴ cluster) was molecularly similar to isolates of VCG1A group (which harbors cotton and olive-infecting D pathotype; Fig. 4). Interestingly, VCG1A isolates from artichoke and cotton induced a severe (defoliating) syndrome in cotton but not in artichoke, with this VCG being the most virulent group to cotton over any other group. In contrast, isolates of VCG1A were the least virulent to artichoke, whereas VCG2B and VCG4B isolates were the most virulent ones in this host (Jiménez-Díaz et al. 2006). Indeed, molecular (AFLP and PCR-based pathotyping) and VCG diversity of isolates showed a variable pattern of correlation with regards to virulence in different hosts.

RAPD analyses, as stated above, showed low molecular variability among VCG1A/D-pathotype isolates (cotton and olive) originating from southern Spain. However, the use of AFLP fingerprints and PCR-based pathotyping in a wider collection of isolates revealed molecular differences between VCG1A/D pathotype isolates from Spain and isolates of the same group from Greece and Turkey. This would support the idea that new strains of V. dahliae VCG1A might develop in different geographic areas or, alternatively, that they may have spread from a common focus. Moreover, the lack of a 462-bp sequence in VCG1A isolates from Greece and Turkey, which is a consistent PCR marker of Spanish cotton and olive D isolates (Mercado-Blanco et al. 2003b) would support each of these two possibilities (Collado-Romero et al. 2006). Interestingly, the absence of the 462-bp PCR marker has been further confirmed in three VCG1 cotton-infecting isolates in Israel, however no further information on the subgroups was reported (Korolev et al. 2008). In addition, this maker was absent in all VCG1A/D pathotype olive-infecting isolates in Turkey so far examined (Dervis et al. 2010). Despite this, the VCG1A/D-pathotype isolates from Turkey, infecting olive are molecularly different from their Spanish relatives, yet pathogenicity tests did not revealed any significant differences in virulence among them (Dervis et al. 2010).

Molecular tools have also served to study the evolutionary relationships among V. dahliae VCG (sub)groups. Using parsimony analysis of AFLP fingerprints and sequences of six DNA regions (actin, β-tubulin, calmodulin, and histone-3 genes, the ITS 1 and 2 regions of the rDNA, and a V. dahliae-specific polymorphic DNA sequence) from a broad collection of isolates has provided interesting conclusions (Collado-Romero et al. 2008, 2010). A low number of polymorphisms in gene sequences was found among isolates of different VCGs, and individual gene genealogies provided very little resolution at the VCG level. However, evolutionary relationships among V. dahliae VCGs (that is, at the intraspecific level) were resolved by phylogenetic analyses of AFLP fingerprints, multiple gene genealogies, and the combined data of AFLP fingerprinting and multiple gene genealogies. Thus, two main lineages were identified. One including VCG1A (which harbors olive D pathotype isolates), VCG1B, and VCG2B³³⁴, and another comprising two closely related subgroups of VCGs (where olive ND isolates are allocated): VCG2A and VCG4B (subclade 1) and VCG2B⁸²⁴, VCG4A, and VCG6 (subclade 2) (Fig. 4). This study demonstrated that V. dahliae VCG subgroups were monophyletic except for VCG2B that appeared to be polyphyletic. An additional interesting conclusion from this study was that limiting the parsimony analysis either to AFLP fingerprints or DNA sequences would have obscured intra-VCG differentiation, and that the dual approach represented by the independent and combined analyses of AFLP fingerprints and DNA sequences was a highly valuable approach (Collado-Romero et al. 2008).

In summary, all these studies have provided fundamental knowledge on the structure of V. dahliae populations, including those isolates infecting olive. Moreover, relevant association between VCG and molecular diversity has been found, although correlation with host specificity or pathogen virulence is sometimes difficult to establish and/or interpret. In addition, these results offer a valuable practical perspective as well: the availability of molecular tools for in planta and in soil pathogen detection (see below). Indeed, from the variety of molecular markers identified, proper selection can be made from those ones clearly associated with specific groups or linked to a particular phenotypic trait (i.e., pathotype). Thus, specific DNA sequences can be identified to design VCG-specific or pathotype-specific primers to be used in PCR-based detection and diagnosis procedures of V. dahliae isolates infecting olive or any other host (Pérez-Artés et al. 2000; Mercado-Blanco et al. 2003b; Collado-Romero et al. 2006, 2009) (Fig. 4).

The possibility that V. dahliae persists in symptomless plants alerts about the precautions that must be taken to ensure that healthy propagation and/or planting material is used to establish new olive orchards in pathogen-free soils. Therefore, the early, rapid and reliable in planta detection of the pathogen (particularly of the D pathotype because of its alarming spread) would be of importance for the management of VWO. Thus, the production of pathogen-free certified olive plants in nurseries is a key control measure to avoid pathogen dispersal to new areas. In addition, this should be bound to the use of healthy (pathogen-free or de-infested) soils (see below).

Pathogen-free qualification of olive plants by means of traditional isolation procedures suffers from the inconsistency of recovering V. dahliae from affected woody tissues (Blanco-López et al. 1984). Moreover, isolation of the pathogen from plant tissues is time-consuming and, in the case of VWO, does not provide information about the pathotype (D and/or ND). Alternative diagnostic methods have therefore been developed to overcome these limitations. Thus, molecular detection and differentiation of Verticillium spp. is generally performed by PCR-based procedures. For instance, V. dahliae-specific PCR primers derived from the mitochondrial small rDNA gene region were designed to differentiate V. dahliae isolates from diverse Ascomycetes genera and also from its close relative V. albo-atrum (Li et al. 1994). Other PCR primers based upon ITS rDNA sequences or RAPDs markers have been designed for the specific differentiation, and in some cases detection and quantification, of Verticillium wilt pathogens in diverse herbaceous host plants and soil samples (Nazar et al. 1991; Robb et al. 1993, 1994; Moukhamedov et al. 1994; Morton et al. 1995a; Heinz et al. 1998; Li et al. 1999; Heinz and Platt 2000). However, the differentiation of V. dahliae D and ND pathotypes was never addressed in these studies. The availability of pathotype-specific PCR primers (Pérez-Artés et al. 2000) raised the possibility to develop detection procedures for the early, rapid and specific molecular diagnosis of V. dahliae isolates, especially aimed at nursery-produced olive plants. Nested-PCR protocols, necessary to accurately detect the presence of very low quantities of pathogen DNA in samples of total genomic olive DNA, are already available for the in planta differential detection of the ND (Mercado-Blanco et al. 2001) and D (Mercado-Blanco et al. 2002) pathotypes. The tuning up of these molecular detection procedures was performed in artificially-inoculated plants. Detection of each V. dahliae pathotype DNA was effective from the very earliest moments after inoculation; that is, positive pathogen DNA detection was achieved in symptomless plants, much earlier than first VWO symptoms developed. Interestingly, olive plants inoculated with the ND pathotype which developed VWO but became symptomless 7 months after inoculation (i.e., they showed the recovery phenomenon, see below), were positive to the presence of pathogen DNA when examined by the nested-PCR assay (Mercado-Blanco et al. 2001). These results confirmed that asymptomatic plants may constitute a reservoir of V. dahliae propagules, as previously discussed in this review. On the other hand, the availability of a specific molecular detection procedure for the D pathotype is of much help to avoid spread of the most damaging pathotype to new areas, especially if they are free of V. dahliae. This is of particular relevance now, since the crop is expanding into Australia and South American countries, which currently import large amounts of rooted, nursery-produced olive plants (S. Ruano, A. Trapero-Casas, personal communications).

Subsequent refinements of nested-PCR protocols have been accomplished. Thus, an improved duplex, nested-PCR procedure was developed eliminating occasional ambiguities observed when detecting D isolates, and reducing time and cost for an effective detection of D and ND V. dahliae in olive tissues (Mercado-Blanco et al. 2003b). The duplex, nested-PCR procedure is based on the simultaneous amplification of both the previously described ND PCR markers (Mercado-Blanco et al. 2001) and a D-associated PCR marker designed from a V. dahliae species-specific genomic sequence (Carder et al. 1994), which has been revealed as highly polymorphic in subsequent studies (Collins et al. 2005; Collado-Romero et al. 2006; Collado-Romero et al. 2010). The procedure was effective in the rapid, unequivocal and simultaneous detection of the two infecting V. dahliae pathotypes in both artificially-inoculated, self-rooted olive plants and naturally-infected adult trees of different cultivar, age and growing conditions (Mercado-Blanco et al. 2003b). Therefore, the improved molecular detection protocol enabled to discover that both pathotypes can infect and colonize the same olive tree under natural conditions. This finding poses important epidemiological consequences and may be of interest for the implementation of additional control measures. Finally, V. dahliae (no pathotype differentiation) has also been successfully detected in olive seeds by nested-PCR assays using PCR primers designed from the ITS region of nuclear rDNA genes (Karajeh 2006).

Besides being informative on the pathotype present in infected olive tissues, all PCR-based procedures so far developed for V. dahliae detection always yielded more consistent and sensitive results compared to traditional microbiological isolation methods. This was confirmed by a comparative study between the nested-PCR strategy, in this case using PCR primer pairs which did not differentiate between D and ND pathotypes (Li et al. 1994), and the traditional pathogen isolation method in culturing media from olive tissues (Morera et al. 2005). It is worth mentioning here the recent development of a multiplex-nested-PCR procedure which can differentiate among V. dahliae VCGs simultaneously infecting artichoke plants (Collado-Romero et al. 2009). This procedure is applicable to olive and many other plants which can be infected by any V. dahliae subspecific group(s). Obviously, despite these molecular procedures were originally aimed to qualify nursery-produced olive plants, they also have a huge value in VWO diagnosis of adult olive trees under field conditions.

All previous molecular detection methods are qualitative and provided essential information to assess the sanitary status of planting material, or for specific and accurate disease diagnostics in established, commercial orchards. However, they do not provide information on differences of pathogen biomass present in plant tissues or on the relation between susceptibility of host genotypes and/or virulence of isolates. Thus, for Verticillium wilt diseases, quantitative measurement of the pathogen in infected tissues is necessary to characterize the disease reaction of host genotypes to specific strains of the pathogen. Quantification of Verticillium spp. DNA in several infected herbaceous hosts has been previously performed by PCR-based quantitative procedures (Hu et al. 1993; Heinz et al. 1998; Dan et al. 2001). However, real-time quantitative PCR (RT-qPCR) technology enables the quantification of nucleic acids in unknown samples by a direct comparison to standards amplified in parallel reactions (Morrison et al. 1999; Deepak et al. 2007). It is more accurate and less time consuming than conventional, end-point quantitative PCR because it monitors PCR products before reaction components become limiting. Currently, RT-qPCR technology is commonly used in plant pathology for the precise detection and quantification of pathogens in infected plants and infested soils (reviewed by Schena et al. 2004b; Ma and Michailides 2007), and it has been used to quantify V. dahliae DNA as well (Schena et al. 2004a; Atallah et al. 2007; Gayoso et al. 2007). RT-qPCR has also been successfully implemented to quantify the biomass of V. dahliae D and ND pathotypes in infected olive plants (Mercado-Blanco et al. 2003a; Markakis et al. 2009). These studies not only provided accurate quantification of V. dahliae DNA in infected olive tissues, but also served to test whether severity of VWO and virulence of pathotypes are correlated with the amount of pathogen DNA present in plant tissues. This was achieved by performing time-course infection bioassays of susceptible and resistant olives genotypes, and by monitoring pathogen colonization in root and stem tissues by means of RT-qPCR.

Using this approach Mercado-Blanco et al. (2003a) reached four interesting conclusions. 1) The amount of pathogen DNA quantified in an olive genotype correlated with its susceptibility to Verticillium wilt. Thus, pathogen biomass (either D or ND pathotype) in roots and in stems was higher in the most severely diseased genotype (‘Picual’, highly susceptible) and lesser in ‘Arbequina’ (susceptible) and ‘Acebuche-L’ (resistant) plants. These two latter genotypes also produced fewer diseased plants, although all roots were infected as demonstrated by nested-PCR assays. 2) Overall, differences in the amount of pathogen DNA measured in planta were influenced more by the olive genotype than by the virulence of the infecting pathotype. This suggested that the extent of pathogen colonization does not clearly determine the virulence phenotype. Thus, with regard to V. dahliae persistence in the infected olive tissues, the more susceptible was the olive genotype and the more virulent was the V. dahliae pathotype in the interaction, the longer the time period was for which significant amounts of pathogen DNA were measured in plant tissues. 3) Maximum amounts of pathogen DNA in roots of plants of all olive genotypes and stems of ‘Picual’ occurred before maximum disease expression. Thereafter, pathogen DNA in root decreased sharply (‘Arbequina’ and ‘Acebuche-L’ plants) or varied slightly along several weeks after inoculation (‘Picual’ plants). This has been recently confirmed by the increase of pathotype D biomass observed by CLSM in roots of artificially-inoculated ‘Arbequina’ plants (Prieto et al. 2009) (see above). Similarly, the gradual decrease in pathogen DNA corroborated earlier work suggesting that pathogen biomass may decrease once the disease is fully developed (Rodríguez-Jurado 1993). Therefore, evidences of the cyclical colonization pattern observed in herbaceous hosts upon V. albo-atrum infection (Hu et al. 1993; Heinz et al. 1998) were not observed in olive. It might well be that putative cycles of colonization in olive may require a time period longer than the extent (100 days) of bioassays performed in this study. 4) Finally, and from a practical perspective, RT-qPCR has revealed as an outstanding tool to assess the reaction of olive genotypes to V. dahliae pathotypes. An earlier recommendation on the necessity of accurate in planta quantification of V. dahliae biomass during breeding programs for resistance to Verticillium wilt in potato (Dan et al. 2001) is of application in VWO as well. The RT-qPCR approach can definitively assist in pursuing such an aim.

Recently, a similar approach was carried out to correlate the relative amount of DNA of V. dahliae D and ND pathotypes to VWO susceptibility in the Greek cultivars Amfissis (susceptible) and Kalamon and Koroneiki (resistant) (Markakis et al. 2009). Overall, results confirmed main conclusions reached in the RT-qPCR study with Spanish olive genotypes. Thus, for example, it was demonstrated that the D and ND pathotypes were present at a significantly higher level in susceptible ‘Amfissis’ plants than in resistant ‘Kalamon’ and ‘Koroneiki’ genotypes. However, some differences were also found. Interestingly, the amount of pathogen (both DNA and percentage of positive isolations) in roots was lower than in stems and shoots, and declined in plant tissues over time regardless the olive cultivar examined. This was in contrast to the lower pathogen DNA content found in stem tissues compared to that in root tissues reported by Mercado-Blanco et al. (2003a). This discrepancy was attributed to the different infection methods used (root system dipping in conidial suspension of Spanish cultivars vs. transplant to MS-infested soil for Greek ones) and to the distinct duration of bioassays and sampling time-points strategies (about 100 days, although numerous sampling time-points for the Spanish cultivars vs. over a year but only three sampling times for the Greek bioassay). Markakis et al. (2009) also observed a steady decline in the amount of DNA over time in all cultivars examined, as overall reported by Mercado-Blanco et al. (2003a). This observation was attributed to defense mechanisms operating in olive (for instance, occlusion of infected vessels, inactivation of the pathogen in the xylem, increase of certain phenolic compounds, etc) within the framework of the recovery phenomenon (Markakis et al. 2009).

Traditional microbiological methods used to quantify V. dahliae MS in soil has been thoroughly reviewed by Harris et al. (1993). These techniques involve the plating of a defined amount of soil, previously processed by diverse methods (i.e. suspension, sieving, centrifugation etc), onto plates of semi-selective culturing media (Harris et al. 1993). Each viable pathogen propagule, either as a single MS or as a cluster of them, free or attached to soil or plant debris particles, will germinate and produce a recognizable V. dahliae colony. Indeed, MS produced from mycelium form a typical star-shaped distribution upon incubation (López-Escudero et al. 2003). Colonies are then counted and the result is expressed as propagules per gram of soil.

None of the available media seems to be completely selective for V. dahliae, and growth of other soil-borne microorganisms is generally observed (Butterfield and DeVay 1977; Harris et al. 1993; Termorshuizen et al. 1998; López-Escudero et al. 2003). Moreover, it should be mentioned that precision in detection and quantification of V. dahliae in soil is particularly aimed to: 1) disease risk assessment, predicting the onset and development of future VWO epidemics, and 2) evaluate the effectiveness of control measures for pathogen eradication in soil. Therefore, several studies have evaluated advantages and disadvantages of available selective media and propagule isolation procedures (Termorshuizen et al. 1998; Goud and Termorshuizen 2003; Goud et al. 2003; López-Escudero et al. 2003). Remarkably, these studies conclude that results obtained upon the implementation of the available detection/quantification methods are greatly influenced by: 1) diverse soil features, such as texture, (i.e. average size of pathogen propagules when the process includes sieving); 2) the amount of soil distributed on a plate which could influence competition and/or antagonism events taking place in vitro among the culturable soil microorganisms present in the soil sample; and/or 3) the variability and abundance of the culturable soil microbiota (Harris et al. 1993). One of the most common techniques for V. dahliae quantification in soil is based on the wet sieving method (Huisman and Ashworth 1974). This procedure separates the 35–150 μm soil fractions in water. Subsequently, aliquots of this suspension are plated onto several media, such as the previously-mentioned modified sodium polypectate agar medium (Butterfield and DeVay 1977). Other techniques are based on dry-plating procedures which seem to be less variable at higher inoculum levels than the wet sieving. The latter is less sensitive but more effective than the former when the number of MS in the soil sample is low (Harris et al. 1993; Goud and Termorshuizen 2003).

As mentioned above, dry and wet plating methods have traditionally been used for the detection and quantification of V. dahliae in soil. Nevertheless, plating methods are time-consuming, inconsistent, soil-dependent and usually not informative about V. dahliae pathotypes present in the soil sample (Harris et al. 1993; Goud and Termorshuizen 2003; López-Escudero and Blanco-López 2005c). Thus, PCR-based techniques have been effectively used for the molecular detection (Mahuku et al. 1999; Volossiouk et al. 1995) and quantification (Krishnamurty et al. 2001; Mahuku and Platt 2002; Debode et al. 2009; Peters et al. 2009) of Verticillium spp. in soil. Success in the detection procedure always depends upon efficiency in extracting PCR-quality DNA from soil samples. One of the main barriers for the efficient and consistent amplification of target DNA extracted from soil samples is the presence of humic acids and related compounds originated from degradation of organic matter (García-Pedrajas et al. 1999). These compounds are known to be powerful inhibitors of Taq polymerases (Tsai and Olson 1992). Therefore, diverse methods have been designed to prevent co-purification of such inhibitors from soil samples, or to avoid their inhibitory effects on PCR (Volossiouk et al. 1995; García-Pedrajas et al. 1999; Heinz and Platt 2000).

Besides procedures for isolating and quantifying V. dahliae propagules in soil, which provide useful information for disease risk assessment, soils chosen to establish new olive orchards could be previously sanitized by different approaches. Before-planting strategies for soil disinfection are based on the application of physical, cultural, biological or chemical practices to reduce pathogen inoculum in soil. It is important to remark that complete eradication of the pathogen is very difficult, and soil re-infestation can take place from pre-existing, non-eliminated propagules after disinfection treatment. An additional aspect to be considered is the high cost of some of these control measures that makes them advisable only at low scale (i.e. for well-defined focus where the pathogen has been particularly detected, Fig. 9a), or for crop areas where the cost of a selected eradication practice is acceptable (i.e. super-intensive hedgerow olive orchards).

Open image in new window

Soil disinfestation prior to orchard establishment can be accompanied by cultural and biological-based practices. For instance, a delay in planting the olive orchard combined with crop rotation and/or cultural interventions such as cropping of non-susceptible species, or their sowing and subsequent burying of green manures such as Sudan grass (Sorghum halepense var. sudanense) have been widely advised. These practices increase the antagonistic microbiota (see below) and have shown some effectiveness in controlling V. dahliae (Devay and Pullman 1984).

Soil solarization is one of the most effective physical control methods to reduce inoculum density of soil-borne pathogens such as V. dahliae (Katan 1987; Katan 2000a; Goicoechea 2009). Soil covering with transparent polyethylene films, combined with abundant watering, makes it possible to heat soil by sun irradiation to temperatures lethal for many soil-borne pathogens and weed seeds. Temperatures reached by soil solarization normally range from 35 to 60°C (depending on soil type and depth) and these are enough to kill V. dahliae MS (Katan 1987). It has been effectively implemented to control Verticillium wilts of herbaceous and woody hosts before sowing or planting (Pegg and Brady 2002, and references therein). However, it is interesting to emphasize that temperatures above 70°C can produce detrimental effects on soil microbiota and may result in eradication of beneficial microorganisms (i.e. rhizobia, mycorrhizal fungi, and antagonists [Katan 2000a; Goicoechea 2009]). Another limitation of this physical method is that its efficacy depends on particular environmental and soil conditions, such as physical soil features and irrigation. Moreover, the survival of V. dahliae MS at deep layers in the soil profile can constitute a source of infective propagules years after soil solarization treatment(s) (Katan 1981; Tjamos et al. 1991; López-Escudero and Blanco-López 2001).

Soil solarization is a common pre-planting practice in modern agricultural systems, predominantly for high-value herbaceous hosts such as artichoke, cotton, strawberry, etc. In some of these crops (i.e., cotton) it has been successfully combined with tolerant cultivars (Melero-Vara et al. 1995). In others (i.e., artichoke), soil solarization was combined with reduced dosage of some of the most common chemical biocides used in intensively-managed crops (methyl bromide, chloropicrin, etc) (Tjamos and Paplomatas 1988). The high levels of solar irradiation in most of the olive growing regions would make soil solarization an adequate control action to reduce MS viability. Indeed, it has been used to diminish VWO symptoms in established orchards (see below). Nevertheless, to the best of our knowledge, soil solarization as extensive before-planting practice in commercial olive orchards has not been used, probably due to its high cost. However, it has been successfully applied in combination with low doses of metam-sodium to disinfest new plots for planting at olive germoplasm collections sites (Caballero and Del Rio 2008b). These chemical treatments combined with soil solarization may be also very effective and economical-cost justified for soil disinfestations at specific spots of olive orchards affected by VWO (see below).

Soil-borne pathogens disinfestation by chemical treatments has been a common before-planting practice in intensive agriculture of numerous herbaceous crops. The efficacy of soil fumigants or combinations of them (Metam-sodium, 1,3-dichloropropene, chloropicrin, or methyl bromide, which is being internationally phased out under the Montreal Protocol) against V. dahliae has been recently reviewed by Goicoechea (2009). This and other classical reviews (i.e. Katan 2000b) on the subject strongly emphasize the need of searching for alternatives to the use of these noxious chemicals, within a sustainable agriculture framework (Hamblin 1995). The final goal would be the implementation of a control strategy based on chemical treatments, either alone or in combination with some of the aforementioned physical and cultural methods, but always employed at reduced dosages. For instance, the combination of diverse soil amendments with low dosages of biocides has been minutely documented, and its use particularly encouraged by Goicoechea (2009). Thus, the individual application of low dosages of methan-sodium or dazomet (dichloropropene/methyl isothiocyanate mixtures), or in conjunction with soil amendments and/or soil solarization, would be an advisable action in olive orchards. This action is particularly interesting in treatments of well-defined focus in order to avoid problems after replanting (i.e., replacing dead trees as a consequence of the disease), as well as to limit the spreading of pathogen infective propagules throughout the plantation.

The development of breeding programs aimed to produce resistant cultivars is likely the most economically-effective and environmentally-friendly control measure to be implemented in an integrated disease management strategy. Therefore, research efforts are aimed to find effective sources of VWO resistance in different countries where olive is a relevant commodity. Obviously, new disease-resistant genotypes must also display agronomic traits attractive from a commercial perspective. Once an agronomically-interesting and VWO-resistant genotype is identified it may be directly used to either replace dead trees in affected olive orchards, or as rootstocks to graft scions of existing cultivars, or to be incorporated as genitor in breeding programs. An important advantage in olive breeding for resistance to V. dahliae is the ample source of genetic variability available (Angiolillo et al. 1999; Rallo et al. 2005). Indeed, putative resistant cultivars could be identified either in genotypes preserved in different olive germoplasm banks, or in wild olives, or in progenies generated from diverse breeding programs (Rallo et al. 2005). However, significant efforts are still needed and only limited results are available (Trapero et al. 2009b). We would like to stress that for the olive/V. dahliae pathosystem we use the term resistance instead of tolerance (see Robb 2007), on the basis of histopathological observations showing that infected olive plants indeed prevent, delay or restrict pathogen penetration and further proliferation in host tissues (Rodríguez-Jurado 1993; López-Escudero and Blanco-López 2005a). That is, plant-mediated mechanisms seem to operate for hindering pathogen colonization.

Studies on VWO resistance under field conditions have been carried out either in experimental or in commercial olive orchards established in infested soils. It is worth mentioning that distribution and inoculum density of V. dahliae were unknown (or not reported) for some of these studies, whereas the presence of D and/or ND pathotypes was reported in others (Tables 2 and 3). These studies usually consisted of long-term observations (Wilhelm and Taylor 1965; López-Escudero and Blanco-López 2001; Martos-Moreno 2003) and, in some cases, natural infections were complemented with artificial pathogen inoculation such as root spraying before planting (Wilhelm and Taylor 1965). Overall, information so far gathered shows that most of the olive cultivars examined to date are susceptible to VWO. Moreover, these cultivars are the economically (and historically) relevant ones in their areas of cultivation: ‘Hojiblanca’ or ‘Picual’ in Spain, ‘Konservolia’ or ‘Kalamon’ in Greece, or the worldwide cultivated cultivar ‘Manzanillo’ (Tables 2 and 3). For instance, a long-term study aimed at investigating VWO resistance in a plantation established in a soil with high inoculum density since 1998, has provided interesting information about some cultivars (Table 2) (Martos-Moreno 2003). In this study, while ‘Cornicabra’, ‘Picual’ and ‘Picudo’ trees were extremely susceptible, ‘Arbequina’ showed to be moderately resistant. However, this cultivar has been qualified as highly susceptible to V. dahliae D pathotype in artificial inoculations (see below) (López-Escudero et al. 2004) (Table 3). On the contrary, low DI’s scored for ‘Changlot Real’, ‘Empeltre’, ‘Frantoio’ and ‘Oblonga’ trees in the same field comfirmed resistance levels earlier found for these cultivars in artificially-inoculated plants under controlled conditions (López-Escudero et al. 2004; Martos-Moreno et al. 2006).

One of the first known studies on the assessment of olive resistance to V. dahliae using artificial inoculation of V. dahliae is that by Hartman et al. (1971). Since then, numerous studies have employed artificial inoculation approaches to screen for resistance to VWO, predominantly assessing olive cultivars of local importance in a given olive cultivation area/country, or genotypes preserved in germoplasm collections (Table 3). These evaluations have been conducted in controlled conditions (greenhouse or growth chambers), using pathogen isolates displaying virulence representative of the study zone (i.e. D or ND), and by root-dipping or stem-injection inoculation methods. For instance, disease reaction displayed by olive cultivars upong infection with the two V. dahliae pathotypes is being continuously evaluated since 1994 in the Olive World Germplasm Bank (López-Escudero et al. 2004, 2007; Raya-Ortega 2005; Martos-Moreno et al. 2006; Birem et al. 2009; Abo Shkeers 2010). This is a worldwide collection of olive varieties, established in 1970 in Córdoba (southern Spain), comprising more than 400 varieties (250 originating from Spain) (Caballero and Del Rio 2008b). The first important conclusion from this research program is that all cultivars so far assessed (>100) were always more susceptible to D than to ND pathotype. However, the most remarkable and alarming result is that most of the agronomically- and economically-relevant olive cultivars are susceptible or extremely susceptible to the D pathotype. This is the case of Greek cultivars Kalamon, Amfissis or Konservolia, the Italian ‘Ascolana’ or ‘Leccino’, or Spanish ‘Picual’, ‘Hojiblanca’, ‘Cornicabra’, ‘Manzanilla de Sevilla’ or ‘Arbequina’. On the contrary, high resistance level has been found in cultivars Empeltre, Frantoio, Oblonga and Changlot Real. These genotypes have shown a delayed disease onset, disease recovery and low percentage of dead plants, making them attractive to farmers to replace dead trees of susceptible cultivars. Moreover, the moderate susceptibility of cultivars such as Cipresino, Sevillenca and, particularly, Koroneiki, makes them an interesting alternative in situations of low inoculum potential in soil (Paplomatas and Elena 2001; Martos-Moreno et al. 2006; López-Escudero et al. 2004; 2007a; Antoniou et al. 2008).

Validation and standardization of V. dahliae inoculation methods to provide accurately-assessed resistant plant genotypes has been an additional goal of the above-mentioned studies (Blanco-López et al. 1998; López-Escudero et al. 2004; Martos-Moreno et al. 2006; Cirulli et al. 2008; Trapero et al. 2009b). Indeed, a rapid and effective inoculation procedure is needed to effectively differentiate resistant genotypes from susceptible ones. Thus, despite root-dip inoculation (using V. dahliae conidia suspension) has been the most widely-used method in evaluating VWO resistance, stem injection of conidia suspensions has proved to be a valid approach. Certainly, good correlation in susceptibility/resistance responses has been found between the two inoculation methods (Blanco-López et al. 1998; López-Escudero et al. 2007a; Paplomatas and Elena 2001) (Table 3).

Evaluation of resistance from olive genotypes crossings was early carried out by Wilhelm and Taylor (1965). They reported interesting results on the VWO resistance reaction in a number of olive seedlings grown from seeds produced under open-pollination conditions, as well as from seeds collected from wild olives planted in infested fields (Table 2). Particularly, one descendant from ‘Arbequina’, called ‘Allegra’, showed high VWO resistance levels (Wilhelm 1981). Later, Tjamos et al. (1985) informed about the high level of VWO resistance of one line obtained by self-pollination of ‘Oblonga’ plants. More recently, the evaluation of wild olives populations as well as progenies obtained from crosses between commercial cultivars (under both controlled and field conditions) is being pursued (Cirulli et al. 2008; Colella et al. 2008; Trapero et al. 2009b). Overall, several wild olive genotypes showing a complete VWO resistance level at the seedling stage have been identified in these studies. Hopefully, they will constitute the first step to generate new olive varieties resistant to V. dahliae. In addition, ongoing studies involving pathogen inoculation of a large number of individuals from progenies produced out of parentals with different VWO resistant levels, will serve to provide relevant information about VWO resistance inheritance (Colella et al. 2008; Trapero et al. 2009b). The improvement of V. dahliae inoculation methods is also a remarkable achievement from these works. Thus, by using the stem-inoculation method, the disease incubation period has been considerably reduced, a consistent expression of disease symptoms was achieved, and the possibility to work with large numbers of plants, thus reducing time and space, is facilitated. On the other hand, the inoculation of very young (3-to-5 week old) olive seedlings by root dipping was successful as well, since it allowed rapid but consistent VWO symptoms expression enabling the evaluation of a large number of plants at early stages of the breeding program. This provides that pre-selected plant genotypes can pass to the next evaluation step in the program in a shorter period of time, quickly discarding genotypes displaying a susceptible response (Trapero et al. 2009b).

The use of resistant olive rootstocks is another alternative to genetically control VWO, since pathogen proliferation can be suppressed within them, hindering tissue colonization of the grafted variety. Related to this, Rodríguez-Jurado (1993) showed that VWO expression in susceptible olive cultivars may be related to a pathogen biomass threshold present in the aerial part of the plant, while the roots could serve as a pathogen reservoir. Thus, olive rootstocks with high level of resistance may serve to hamper pathogen multiplication and therefore to protect the variety grafted. For example, after artificial inoculation with a D isolate, Porras-Soriano et al. (2003) reported a reduction of VWO symptoms developed in the susceptible cultivar Cornicabra after grafting it onto a rootstock of the moderately resistant cultivar Oblonga. Under field conditions, Hartman et al. (1971) observed that scions of cv. Sevillano grafted on 'Oblonga' rootstocks did not develop VWO symptoms during an evaluation period of 16 years, suggesting that the pathogen was not transmitted through the graft. However, other combinations tested, including graftings on several Olea species and diverse O. europaea commercial cultivars, yielded reactions ranging from moderate to high susceptibility. Tjamos et al. (1985) also found that ‘Conservolia’ plants grafted onto an ‘Oblonga’ clone became resistant to V. dahliae infection. This result was based on the absence of symptoms and the failure to isolate the pathogen from plant tissue. Our long-term field observations conducted in Spain since 1998 in an experimental olive orchard established in a heavily infested (D pathotype) soil, aimed to assess the influence of different combinations of rootstocks and grafted cultivars on VWO resistance are providing preliminary results. For instance, susceptible (‘Hojiblanca’ and ‘Lechín de Granada’) or highly susceptible cultivars (‘Cornicabra’ and ‘Picudo’) grafted on ‘Oblonga’, ‘Frantoio’, ‘Empeltre’ and ‘Changlot Real’ rootstocks showed low DI (8–20%) 12 years after plantation, whereas the same non-grafted cultivars have reached DI ranging from 73 to 84% (López-Escudero and Blanco-López, unpublished data).

Finally, it is worth mentioning that the extrapolation of results on disease reaction of olive genotypes obtained from artificial inoculation bioassays to observations under field conditions should be made with the necessary caution. A resistant or moderately-resistant reaction for a given olive genotype obtained upon artificial inoculation of the pathogen may be overcome by diverse environmental conditions, by the inoculum potential of the pathogen in soil and its virulence, or by different crop management practices (Blanco-López et al. 1998). Therefore, it is strongly recommended to corroborate results obtained under controlled conditions by approaches simulating field conditions and that consistently reproduce the onset and development of the disease. This would also enable to control sources of experimental variation and to use reliable information about genotypes resistance to predict risk levels. This strategy has been successfully implemented in assays performed in above-ground microplots (1 m width, 1 m length, 0.5 depth, nine olives per microplot) artificially-infested with MS of the D pathotype (López-Escudero and Blanco-López 2007). Indeed, this study enabled to determine the relationship between V. dahliae inoculum densities present in soil and VWO onset and progress in ‘Picual’ plants. After nearly three years after planting, no differences on final DI were found among microplots infested with inoculum densities ranging from 0.4 to 1 MS g⁻¹ soil, reaching a mean DI of 11% (Fig. 10). Nevertheless, in microplots infested with 3 to 10 MS g⁻¹ soil, percentage of wilted trees was significantly higher (mean DI of 55.5%), suggesting that the defensive reaction in ‘Picual’ plants is overcome when the amount of MS in soil is above a threshold level of 1 MS per gram of soil (López-Escudero and Blanco-López 2007).

Open image in new window

Resistance under field conditions can be modified by the natural recovery phenomenon. This can be observed in trees which have developed symptoms during a growing season and that may eventually recuperate. This phenomenon is commonly associated to mechanisms which allow trees to overcome injury and decay (Shigo 1984; Hiemstra 1998), but it can also be activated after infections caused by vascular pathogens such as V. dahliae (Tippett and Shigo 1981). Indeed, Verticillium wilt recovery has been reported in diverse woody hosts (for example, Taylor and Flentje 1968; Emechebe et al. 1974; Latorre and Allende 1983; Hiemstra 1995; Cirulli et al. 1998; Goud and Hiemstra 1998), including VWO-affected trees and nursery-produced young olives under field and controlled-growing conditions (Wilhelm and Taylor 1965; Vigouroux 1975; Tjamos et al. 1991; Rodríguez-Jurado et al. 1993; Mercado-Blanco et al. 2001; Levin et al. 2003b; López-Escudero and Blanco-López 2001, 2005a; Markakis et al. 2009).

Recovery is evidenced by a gradual reduction in DI and by the emission of new sprouts at the base of the trunk or under affected branches. This phenomenon has been explained as an active response by the tree. Indeed, the highly compartmentalized structure of the xylem in woody plants in combination with cambial activity producing new layers of tissue that replace old or diseased xylem facilitate plant recovery (Shigo 1984). Early steps of recovery are probably a consequence of tree responses such as occlusion of infected vessels, inactivation of the fungus in the xylem, hindrance to new infections and/or enhanced levels of certain phenolic compounds (Wilhelm and Taylor 1965; Talboys 1968; Sinclair et al. 1981; Rodríguez-Jurado 1993; Hiemstra 1995; López-Escudero and Blanco-López 2005a; Baídez et al. 2007; Markakis et al. 2010). Since trees affected by VWO are subjected to seasonal and transient infections by the pathogen (Wilhelm and Taylor 1965; Tjamos et al. 1991), new infections seem to be necessary for a new disease cycle after recovery. This situation is particularly interesting from a disease management perspective since it may favor the implementation of effective control measures. For instance, soil solarization may be used to reduce V. dahliae inoculum density in soil around the tree, thus diminishing the probability of new infection events (Tjamos et al. 1991; López-Escudero and Blanco-López 2001; Saremi et al. 2006). This beneficial effect has been observed especially in young commercial olive orchards, in which soil-solarized trees frequently overcome seasonal infections by V. dahliae (Rodríguez-Jurado et al. 1993; López-Escudero and Blanco-López 2001; Martos-Moreno et al. 2001).

Reduction of symptoms may be a consequence of inactivation and/or elimination of the pathogen, since it is correlated with the difficulty or impossibility of a successful isolation of V. dahliae from affected branches that recovered (Wilhelm and Taylor 1965; Rodríguez-Jurado 1993; Levin et al. 2003b; López-Escudero and Blanco-López 2005a), or to detect pathogen DNA in aerial tissues of infected but recovered plants (Mercado-Blanco et al. 2001). These results again support that new infections are likely needed for the development of new symptoms, as first suggested by Wilhelm and Taylor (1965). Alternatively, the persistence of V. dahliae propagules within infected parts of the tree, i.e. MS produced within tracheal elements (Rodríguez-Jurado 1993), may eventually lead to new symptoms when optimal conditions for disease development are again present. Indeed, V. dahliae DNA could be detected in root samples of artificially-inoculated ‘Picual’ plants more than 6 months after they recovered from the disease (Mercado-Blanco et al. 2001), or the pathogen could be isolated at least 1 year later (Rodríguez-Jurado 1993).

It is worth mentioning, however that the phenomenon of VWO recovery is strongly influenced by diverse critical factors such as the olive genotype susceptibility, environmental conditions, pathogen virulence, etc (López-Escudero and Blanco-López 2005a). For example, recovery strongly depends upon the virulence of the pathotype present in soil and on cultivar resistance, being more frequent when ND isolates and resistant cultivars were tested under field (López-Escudero et al. 2004; Martos-Moreno et al. 2001) or in artificial inoculation conditions (López-Escudero and Blanco-López 2005a). Moreover, it also depended upon the resistance level of the olive cultivar examined (López-Escudero and Blanco-López 2005a), as previously suggested by Resende et al. (1995) in cocoa (Theobroma cacao L.). In these cases, recovery could be favored by the irregular distribution, or discontinuous colonization, of the pathogen conferred by the natural resistance of the cultivar. This is particularly important since integrated disease management and risk assessment of VWO rely on, among other control measures, the use of resistant cultivars. However, this is strongly conditioned by the presence and spreading of V. dahliae D isolates to new olive orchards (López-Escudero and Blanco-López 2001; López-Escudero et al. 2004). Therefore, the complete understanding of the mechanisms underlying the natural recovery phenomenon still needs in-depth insight to be used in combination with control measures.

The use of biological control agents (BCAs) can be either considered as a before-planting (preventive) measure or as a post-planting (palliative) action in established orchards. It is a sustainable and environmentally-friendly approach that could be used in combination with other control tools within the framework of an integrated disease management strategy. Thus, application of antagonistic microorganisms in pathogen-free certified olive plants during the propagation process in nurseries, to protect them from eventual V. dahliae infections after planting, has been proposed (Tjamos 1993; Mercado-Blanco et al. 2004). It is worth mentioning that a successful biological control strategy depends on a complex and subtle balance of interactions among the pathogen, the host plant, the BCA and the environment (Fravel 1988; Weller 1988; Thomashow and Weller 1996). Therefore, the efficient use of BCAs requires of an in-depth knowledge of the physiology and ecology of the targeted systems in situ (Edwards et al. 1994). In addition, any beneficial effect that a BCA may deploy against a given phytopathogen requires of a previous and successful colonization of the target niche (rhizosphere, inner root tissues, etc.). Thus, an adequate bacterial population level is required, which is variable and highly dependent on the BCA, the plant species, and/or the specific environment (O’Sullivan and O’Gara 1992; Haas and Défago 2005). In the particular situation of root colonization by BCAs, numerous factors are dynamically interacting: water content and availability, temperature, pH, soil type, composition of root exudates, presence of other microorganisms, host plant genotype, BCA motility, etc. (Berg et al. 2006; Costa et al. 2006). All these parameters determine how a soil-resident bacterial population can be established in the rhizosphere and/or within the plant root. The contribution of each of these factors, the interactions among them and the set-up of the so-called phenomenon “rhizospheric competence” should be studied for each particular interaction (reviewed by O’Sullivan and O’Gara 1992; Lugtenberg et al. 2001; Mercado-Blanco and Bakker 2007; Raaijmakers et al. 2009).

Biological control of soil-borne pathogens of woody plants has been investigated to a lesser extent compared to herbaceous species, although examples are available in the literature (Scheffer 1990; Paul et al. 1997; Reddy et al. 1997; Utkhede et al. 2001; Ran et al. 2005). The anatomy, longevity, and/or inherent particularities of crop management of woody hosts are factors contributing to the difficulty to carry out biocontrol research of their diseases, and olive is not an exception. On the other hand, while studies on biological control of Verticillium wilts are available, they are again focused in herbaceous hosts and in just a small number of soil-borne fungi (i.e., Talaromyces, Trichoderma) or bacteria (i.e. Bacillus, Paenibacillus, Pseudomonas, Serratia, Streptomyces) species which have so far been demonstrated to be effective BCA against V. dahliae (Jordan and Tarr 1978; Leben et al. 1987; Berg et al. 1994; Madi et al. 1997; Alström 2001; Nagtzaam et al. 1998; Narisawa et al. 2002; Tjamos et al. 2004, 2005; Antonopoulos et al. 2008; Uppal et al. 2008; Erdogan and Benlioglu 2010; Meschke and Schrempf 2010). Finally, arbuscular mycorrhizal fungi (i.e. Glomus spp.) have been also investigated for their Verticillium wilt suppressive effects in different herbaceous hosts (Karagiannidis et al. 2002; Garmendia et al. 2004, 2005; Kobra et al. 2009).

Regarding biological control of VWO, indigenous Pseudomonas fluorescens and Pseudomonas putida strains originated from olive roots have been demonstrated as effective BCAs in artificial-inoculation experiments of young nursery-produced olive plants. Some of these strains showed: 1) in vitro antagonism against the ND and D pathotypes of V. dahliae in different culturing media; 2) ability to colonize and persist in the rhizosphere of young olive plants upon artificial inoculation; and 3) effective control of VWO caused by infections of the D pathotype (Mercado-Blanco et al. 2004). These strains were also characterized on the ability to in vitro produce the siderophore pyoverdine, salicylic acid (SA), or hydrogen cyanide (HCN), metabolites which have been largely studied as involved in biocontrol activity displayed by beneficial Pseudomonas spp. (reviewed by Mercado-Blanco and Bakker 2007). The biocontrol capability of these strains was variable in different types of nursery-produced plants (i.e. either micropropagated or self-rooted) of the highly susceptible cv. Picual. Moreover, effective biocontrol was achieved under different culturing environments: growth-chamber (controlled) and greenhouse (semi-controlled) conditions. One of the best-performing strains was P. fluorescens PICF7, which in some situations reduced the disease intensity index and the final DI by 95% and 70%, respectively, with regards to control (non-bacterized) plants (Mercado-Blanco et al. 2004). On the other hand, some strains displaying strong in vitro antagonism against V. dahliae were not so efficient in controlling the disease, a lack of correlation which has been frequently reported (Fravel 1988; Paulitz et al. 1992). Interestingly enough, a good VWO biocontrol strain was not always that one to reach highest population levels at the target niche, a situation exemplified by strain PICF7. The mechanism(s) involved in biocontrol of VWO by these Pseudomonas spp. strains have not yet been elucidated. However, some interesting findings have been recently gathered by using biotechnological and microscopy tools. Indeed, autoflorescent protein-tagging of bacterial cells, vibratome plant tissues sectioning and CLSM have been employed to demonstrate that P. fluorescens PICF7 endophytically colonize the intercellular spaces of the olive roots cortex (Prieto and Mercado-Blanco 2008). This poses additional advantages to be exploited in biocontrol of VWO since endophytic BCAs, such as P. fluorescens PICF7, are already adapted to colonize and persist in the ecological niche (i.e. olive roots) where their beneficial effects can be displayed. Our recent studies suggest that for an effective biocontrol of VWO, strain PICF7 should colonize intact olive roots at both superficial and endophytic levels. Moreover, bacterial root colonization should take place before V. dahliae gains entrance to the root interior by, for example, severe root damage (see above). Although further studies are needed to elucidate the mechanism(s) operating in the PICF7-mediated biocontrol of this disease, CLSM studies have revealed that root hairs seem to play a pivotal role in the internal colonization of the roots by PICF7, and that antagonism may be important as direct contact between PICF7 cells and V. dahliae hyphae has been observed (Prieto et al. 2009). Regarding to other Pseudomonas spp. as BCA for VWO, Hameed et al. (2005) informed of in vitro antagonism against V. dahliae of selected indigenous Pseudomonas strains from olive. However, no information on their possible biocontrol activity has been further reported.

Serratia plymuthica HRO-C48 has been also examined on its ability to suppress VWO in V. dahliae artificially-inoculated olive plants cv. Arbequina (Müller et al. 2007). This bacterium was isolated from oilseed rape rhizosphere, and showed in vitro antagonism against different fungal pathogens, including V. dahliae (Kalbe et al. 1996; Berg 2000; Erdogan and Benlioglu 2010). Diverse mechanisms have been proposed to be involved in biocontrol by HRO-C48: production of lytic enzymes such a chitinases and glucanases, antibiotics such as prodigyosine and pyrrolnitrin, siderophores and phytohormones (Kalbe et al. 1996; Frankowski et al. 2001; Kamensky et al. 2003). Strain HRO-C48 was shown to colonize the rhizosphere of olive plants for a prolonged period of time. However, its biocontrol performance depended on the method used for pathogen inoculation. Indeed, suppression of VWO was found when V. dahliae (D pathotype) was applied by soil inoculation rather than by root dipping (Müller et al. 2007).

The ubiquitous fungal genus Trichoderma has been widely used as BCAs against many soil-borne plant phytopathogenic fungi. Its biocontrol activity relies on multiple mechanisms such as competition for nutrients and space, antibiosis and mycoparasitism. Trichoderma spp. are also able to promote plant growth and activation of plant defense responses (reviewed by Howell 2003; Harman et al. 2004; Benítez et al. 2004). A commercial bioformulation of Trichoderma asperellum and T. gamsii has been tested against VWO produced by the D pathotype under conditions conducive for the disease (Jiménez-Díaz et al. 2009). Interestingly, biocontrol efficiency also depended on the method of pathogen inoculation in artificial inoculation experiments and controlled conditions. Thus, effectiveness and consistency of biocontrol was higher in root-dipped inoculated plants than in the case of plants transplanted to V. dahliae-infested soil. Suppression of VWO to a degree was also achieved by this Trichoderma spp. mixture under field conditions. However, no information was provided on the possible mechanism involved.

Endophytic fungi, which can provide diverse beneficial effects to their host plants, including enhanced resistance to diverse (a)biotic stresses (Arnold 2007; Mejía et al. 2008; Scheffer et al. 2008; Rodríguez et al. 2009b) may also represent an interesting approach for Verticillium wilts biocontrol. However, only a scarce number of studies using root endophytic fungi as BCA against V. dahliae have been performed (Matta and Garibaldi 1977; Narisawa et al. 2002, 2004). Regarding VWO, Martos-Moreno (2003) reported that artificial root-dipping inoculation of olive plants with a ND isolate of V. dahliae can protect the plant from subsequent infections by a D isolate. This may suggest that, among others mechanisms, an enhanced resistance could be acquired upon application of a ND isolate.

Finally, the possibility to control VWO and other soil-borne pathogens by mycorrhizal fungi has also been explored (Porras-Soriano et al. 2006). For instance, it has been shown that olive seedlings inoculated with Glomus intraradices had lower severity of infections by Phytophthora spp. than non-inoculated seedlings under controlled conditions (Guerrero 1999). However, this effect was not observed in the case of plants inoculated with V. dahliae (Jiménez-Díaz et al. 2009; Porras-Soriano et al. 2006).

All these studies are encouraging the use of BCAs, either at the olive propagation stage or during the first years after planting, as an effective action complementary to other disease control measures already explained: i.e., use of certified planting material, low pathogen inoculum density in soil (particularly of the D pathotype), use of tolerant/resistant olive cultivars, or the advantageous situation offered by the recovery phenomenon. An interesting perspective in VWO biocontrol is the possibility to develop combinations of different microbial agents (bioformulations), especially those involving microorganisms that have different mechanisms of action and/or can be targeted to different root niches (Whipps 2001). For instance, bioformulations which might incorporate a beneficial endophyte together with another BCA which exerts its biocontrol action on the root surface provide an interesting research field to be explored. Moreover, different biocontrol mechanisms (for instance, induction of systemic resistance triggered by an endophyte and antibiosis by a root-surface colonizer) could be operating synergistically. However, the development of commercial products that combine different BCAs still needs of in-depth studies to avoid negative interactions that may occur between the BCAs present in the bioformulation, or between them and the indigenous beneficial microflora, or between them and the target plant (Whipps 2001).