Diversity of root-knot nematodes in Moroccan olive nurseries and orchards: does Meloidogyne javanica disperse according to invasion processes?

Background Root-knot nematodes (RKN) are major pest of olive tree (Olea europaea ssp. europaea), especially in nurseries and high-density orchards. Soil samples were collected from main olive growing areas of Morocco, to characterize Meloidogyne species and to discuss the contribution of biotic and abiotic factors in their spatial distribution. Results RKN were found in 159 soil samples out of 305 from nurseries (52.1% occurrence) and in 11 out of 49 soil samples from orchards (23.2% occurrence). Biochemical and molecular characterisation (PAGE esterase and SCAR) revealed the dominance of M. javanica both in nurseries and orchards with minor presence of M. incognita only in nurseries, and M. arenaria in only one nursery. RKN were distributed on aggregated basis. Frequent presence of M. javanica in orchards might have come from nurseries. In contrast, the detection of M. incognita in nurseries alone suggests that this species could not reproduce in orchards because of either the competition with other plant-parasitic nematodes or unfit local habitats. The impact of environmental variables (climate, habitat origin and physicochemical characteristics of the substrates) on the distribution of Meloidogyne species is also discussed. Conclusion Olive nurseries in Morocco are not able to guarantee the safety of rooted plants. As a result, olive production systems are exposed to strong RKN invasion risks. Consequently, the use of healthy substrates in nurseries may prevent plant-parasitic nematode induction in orchards.


Background
Sustainable management of key taxa depends upon understanding their distribution and behaviour towards biotic and abiotic factors. Studying the factors of RKN population dispersal can facilitate to understand their spatial structure [1]. Ecologists and conservation managers depend on spatial models to assess environmental effects on distribution of a species. These models facilitate to develop reserve selection and survey design to manage various species.
Distribution models are categorized into two groups: (1) some simulate interactive processes between environment and organisms, (2) others use pattern analysis to reveal correlation among target taxa and environmental variables. Biological (capacity of an organism to disperse and reproduce), physical (mountains or oceans) and environmental (soil texture, moisture conditions) factors hinder species dispersal [2]. These models require detailed information about organism and environment over a period of time to predict spatial and temporal patterns of an organism [3]. Aït Hamza et al. BMC Ecol (2017) 17:41 The production of commercial olive plantlets in the Mediterranean basin, especially in Morocco, provides a favourable environment for the development of plant pests [4]. The geographic location of Morocco, compared to other Mediterranean countries, offers specific orography and bio-climates, with endemic vegetation [5]. High mountain ranges (exceeding 4000 m in altitude) create a complex and highly compartmentalized structure, with extensive plateaux and plains. Climate of the country is deeply influenced by the Atlantic Ocean with annual rainfall of 30-2000 mm. In Morocco, nursery substrates are often prepared with soil from cropped fields or natural environments that could be potentially infested with soil-borne pathogens such as Verticillium dahliae Kleb. (i.e., Verticillium wilt) [5] and plant-parasitic nematodes (PPN). The use of pathogen-free planting materials and non-infested soils is necessary during the early years of olive cultivation. The threat of these pathogens to olive production has also been recognised by the European Union [6].
PPN are microscopic, round and filiform worms living in soil and/or inside plant root tissues. They generally parasite the underground parts (roots, tubers, rhizomes) of the plants. They cause significant agricultural damage in the world (about 14% yield loss), that reaches over $100 billion per year [7]. Abiotic factors such as pH, soil type, organic matter content, moisture [8], and local climatic conditions affect PPN development [9]. They move only short distances and, thus their dissemination is via water [10] and wind [11]. Human activities such as the introduction of infected planting material or diffusion of infested soil with nursery practices also contribute to spreading [12]. Meloidogyne spp. are major PPN, causing a worldwide loss of about 50 billion Euros [13,14]. Meloidogyne species like M. javanica, M. inconita, M. arenaria, M. hapla and M. lusitanica are known to infect olive trees [15,16]. Recently M. baetica and M. spartelensis were also identified on wild olive trees in southern Spain and northern Morocco [17]. Nevertheless, little information is available about PPN host-parasite relationship between RKN and olive plantlets. RKN are known as major pest of olive trees, especially in nurseries having favourable irrigation conditions [18]. Experiments have demonstrated effect of RKN on olive plant growth and of susceptible olive cultivars [19].
RKN control is a challenging task, because: (i) their occurrence is worldwide especially under hot climate; (ii) they are highly diversified; and (iii) they exhibit various reproduction methods (mitotic and meiotic parthenogenicity and amphimixis) [20]. Therefore, in order to manage their infestation, identification of Meloidogyne species is basic requirement to understand its ecology, physiology and reproduction [21]. Conventional methods for RKN identification based on morphological traits require a great deal of skill and are often inconclusive. Polyacrylamide gel electrophoresis (PAGE) isozyme analysis is a relatively fast way to identify Meloidogyne species [22]. However, isozyme analysis can only be done with mature females embedded in roots, and not with secondstage juveniles or eggs in the soil. SCAR (Sequence Characterized Amplified Region) based molecular biomarkers are used to confirm Meloidogyne species [23].
Previous surveys revealed scarce populations of M. arenaria, M. hapla and M. spartelensis in wild olive whereas M. javanica dominated in cultivated areas of Morocco [16]. We hypothesized that its widespread distribution might be due to inductions from nurseries. Therefore, objective of this study is to test the hypothesis by: (i) characterizing the Meloidogyne species in nurseries and orchards of main olive-producing areas of Morocco; (ii) analysing their distribution in nurseries, impact of the climate, habitat and physicochemical characteristics of the substrates; and (iii) discussing their introduction from nurseries to orchards, especially in relevance to M. javanica.

Soil sampling
In nurseries, olive plantlets are grown in 2-3-l plastic bags containing solid substrates from different origins (alluvial sandy soils, forest soils, loamy open-field soils) with different proportions of sand, peat fertilizer and animal manure. Five olive plantlets were sampled from each nursery for each variety. Information of variety, origin and substrates was recorded for each sample. In total, 305 olive plantlets were taken to the laboratory and maintained under greenhouse conditions. In orchards, only soil samples were collected, as PPN spend a part of their life cycle in it. Samples were collected from upper rhizosphere of soil under the foliage. In each orchard, four trees at 10 m distance along transects, were chosen and from each tree five sub-samples were taken. Considering that the cultivation activities could homogenize the distribution of nematodes, twenty  sub-samples were pooled into one (1-dm 3 ) reference sample per orchard.

Root-knot nematode extraction and culture
From both, nursery olive plantlets and orchard soil samples, nematodes were extracted from a 250-cm 3 soil aliquot according to elutriation procedure [26]. RKN were identified according to genus, counted and expressed per dm 3 of fresh soil. Susceptible tomato variety (cv. Roma) was grown in 500-cm 3 soil of each sample under greenhouse (12 h light at 25 °C, 12 h dark at 20 °C) to multiply the populations. Presence of RKN galls and egg masses was observed after 60 days of tomato transplantation.

Identification of Meloidogyne species Isozyme phenotype analysis
Tomato roots were lightly washed and adult females were collected using forceps and transfer needles. 25 females and their eggs were collected per sample. Females were individually crushed in 250-µL micro-tubes containing 5 µL of Trugdill buffer with 20% sucrose (pH 8.0) [27], and stored at − 20 °C. Females of pure M. javanica population were prepared as above and used as the reference population. Micro-tubes were centrifuged (9500 rpm for 10 min) and 0.01% bromophenol-blue was added. Supernatants were transferred to 70 × 80 × 0.5 mm separating (7% bis-acrylamide, pH 8.4) and stacking (3.5% bis-acrylamide, pH 6.7) gels [22] whereas PAGE was processed in a Mini Protean II electrophoresis unit (BioRad ® ) at 7 °C. Each gel included two reference M. javanica females. Gels were incubated with α-naphthyl acetate and Fast Blue (37 °C for 1 h) to reveal Esterase (Est) phenotype bands. The band stain was fixed by placing gels in 10% acetic acid for several hours and sandwiched between cellophane sheets to dry for 48 h [28]. Est phenotype patterns were identified and labelled by bands (Rm) in reference to M. javanica.

Molecular identification
Female egg masses of Est analyses were individually incubated in distilled water for hatching.

Data analyses
Tukey's range test was used to compare the frequency of Meloidogyne species between the regions (P value < 0.05). In order to assess the distribution of RKN in olive nurseries according to substrates, physicochemical characteristics and climate (Table 2), k + 1 multivariate method (MultiBlock Partial Least Squares (MBPLS) was followed and anlyzed in readxl, ade4 and R [31][32][33]. MBPLS regression is widely used for exploring and modelling relationships between several datasets to be predicted from several other datasets, and reveals contribution of each set in predicting response variables.

Results
Meloidogyne spp. specimens were detected in 52.1% of the nursery plants and in 23.2% of the orchard soil samples with a population range of 20-4000 individuals per dm 3 of soil. All the RKN isolates reproduced on susceptible tomato plants, except one isolate detected in a traditional orchard from the Haouz region (isolate no 255).

Species characterisation
M. javanica reference population was confirmed as a J3 Est phenotype [22] with three bands (Rm of 46, 54.5 and 58.9%). Six phenotypes were detected among tested females ( Fig. 3a and Table 3), labelling six different Est bands. J3 phenotype was detected widespread in both nurseries and orchards. Two phenotypes of M. javanica J2a (Rm of 46 and 58.9%) and J2b (Rm of 46 and 54.5%) were detected as being mixed with J3 in orchard isolates 260 and 252, respectively. One I1 Est phenotype, specific to M. incognita (Rm of 46%), was detected only as being mixed with J3 in low proportions of Souss and Haouz nurseries. Two a Est phenotypes specific to M. arenaria such as A2 (Rm of 53.75 and 56.25%) and A3 (Rm of 51, 53.75 and 56.25%) occurred as mixed populations in the Jabla nursery. These Est patterns were confirmed with species-specific SCAR patterns ( Fig. 3b and Table 3).

Impact of environmental factors on Meloidogyne species distribution in nurseries
Analysis of variance (Fig. 4) showed that RKN populations were more abundant in the southern regions (Souss    Table 3). M. incognita was found to be associated with the nurseries of southern regions (Souss and Haouz), characterised by a higher MACM. Habitat origin of the substrates contributed less to the MBPLS analysis as compared to climate (Fig. 5c, second axis). However, it was clearly established that M. javanica, widespread in nearly all nurseries (96%, Table 3), was primarily associated with substrates prepared with large amounts of riverbank soils in the Haouz region and of crop soils in the Souss region. Physio-chemical soil factors did not greatly contribute to the structuring of Meloidogyne diversity.

Distribution of the Meloidogyne species in nurseries and orchards
Biochemical and molecular diagnoses confirmed the occurrence of Meloidogyne populations on cultivated olives in Morocco. M. javanica phenotype J3 was detected in 96% of the nurseries and in 92% of the orchards (Table 3), either traditionally or high-density cultivated, and was widespread throughout the main olive producing areas (Fig. 6). Two phenotypes of M. javanica J2a and J2b were detected only in traditional orchards as being mixed with J3 in the Haouz and the Souss regions, respectively (Fig. 6b). M. incognita phenotype I1 was detected only in nurseries and as mixed in low proportions with J3 (Fig. 6a, Table 3). Two M. arenaria phenotypes A2 and A3 occurred as mixed populations in the nursery of Jbala region (Fig. 6a, Table 3). M. javanica populations were more frequently abundant in the southern nurseries (up to 10 3 juveniles/dm 3 of soil in the Souss and Haouz regions) as compared to northern nurseries (less than 10 2 juveniles in the Guerouane region). They were, however, less common in the Souss orchards than elsewhere (Fig. 6b). M. incognita populations flourished more in the Souss nurseries than in Haouz (with 16 and 4.5% of the total RKN populations, respectively), and the population of M. arenaria in Jbala nursery was low (< 10 2 juveniles/dm 3 of soil).

Discussion
Main objective of this study was to understand the dispersal process of RKN from nurseries to orchards. RKN were detected in 1/2 of the nurseries and in less than 1/4 of the surveyed orchards. This corroborates with other reports which reveal only scarce detection of RKN in olive-producing areas [16]. Potential damage of these species to olive has never been properly investigated. However, Meloidogyne spp. are major pest of olive trees as high occurrence is usually noticed [34]. They induce considerable damage in nurseries and reduce olive  growth in orchards [18,19]. RKN population thresholds to olive plantlets are unknown, yet the population densities noticed during this study (up to 10 3 juveniles/dm 3 of soil) may present a potential risk to olive plantlets, in nurseries and fields. M. incognita and M. javanica significantly reduce shoot growth of olive cultivars in nurseries, implying a potential impact of RKN [18]. RKN damage in the irrigated sandy soils of the Souss and Haouz regions (high temperature and moisture) can be even higher as many new olive plantations have been established in context of the Moroccan Green Plan [35,36]. Meloidogyne species were characterised through biochemical (PAGE esterase phenotypes) and molecular RAPD primer (SCAR) methods [16,21]. Est phenotype is the most instructive biochemical identification technique for Meloidogyne species [37,38] because of its species-specificity within the Meloidogyne genus [39][40][41]. All the phenotypes have previously been reported on other crops [22,42] including olive [15,16]. Nevertheless, some variability within M. javanica populations in orchards and within M. arenaria populations in northern nurseries was spotted. The phenotype J2a was already reported [43] whereas other phenotype J2b was previously diagnosed on peanut [44]. Phenotypes A2 and A3 [39,45] were apparently clustered by geographic origin. Biochemical diagnostics were further confirmed by the molecular approach with SCAR markers that demonstrated their specificity to M. incognita, M. arenaria and M. javanica [46].
M. arenaria, M. incognita and M. javanica have previously been reported on olive trees in the Mediterranean basin, Asia and South America [47]. It is suggested that these species have the same geographic distribution on various hosts [23]. M. javanica and M. incognita have been reported as the most common species in the olive nurseries of Spain [18], but M. incognita was not detected in orchards. Widespread distribution of M. javanica was noticed in the southern regions (Souss and Haouz), known for RKN-susceptible vegetable production. M. javanic-M. incognita distribution in Moroccan olive nurseries is in line with their distribution in Iran [48] and Spain [18] where more than 20% of olive plantlets were infested by M. javanica alone, and 10% were co-infected by M. javanica and M. incognita.
Distribution of these RKN species in nurseries could be related to human activities and favourable environmental factors. As in orchards, intensive nursery monoculture is very susceptible to build-up of nematode populations and ultimately damaging the tree seedlings. High temperature is a favourable factor for the reproduction and multiplication of M. javanica and M. incognita populations. In fact, temperature is the key feature for their survival and fecundity [49]. High temperatures also favour hatching, mobility and root invasion of M. javanica. There are reports from southern Spain and northern Iran revealing alarming Meloidogyne population densities (28.6 and 22.3% yield loss, respectively) in nurseries [18,48]. Besides, M. arenaria, usually found in tropical, subtropical, temperate mild regions and in glasshouses under cooler climates [50], was found in the coldest regions of northern Morocco with high annual rainfall.
In short, M. javanica and M. incognita were most probably introduced into nurseries through soil substrates from agricultural fields and riverbanks. Their widespread distribution in nurseries with highly infested substrates confirmed high fitness properties. It clearly confirms that their adaptation and reproductive success is mainly due to their mitotic parthenogenetic reproduction [20] and ability to infect various plant species [51]. M. arenaria was only found in the northern nursery (Jbala region), grown on forest soil substrates. It can be linked to previous reports about phenotype A2 on wild olives [16].
Introduction of RKN in orchards after the transplantation of rooted plantlets, seems obvious, as in case of endoparasitic nematodes. Soil nematodes move very short distances so their long-distance dissemination is only possible through human activities [52]. Consequently, their widespread distribution in cultivated olivegrowing areas might have been inducted from nurseries [53]. Therefore, in case of olive production systems, the invasive species status can be attributed to M. javanica [54], which out-competes native species especially in high-density orchards [55]. However, selection processes might have occurred as no M. incognita was detected in orchards despite their occurrence in nurseries. This extinction could be explained either by non-adequate life conditions (no fitted niches) or competition during longterm cohabitation with M. javanica or with native PPN species. Species selection after invasion might influence the capacity to disperse [56] along with the physiological tolerance to the new environment [57]. Presence of M. arenaria only in one nursery supports the hypothesis of refuge conditions of the area for this species [58], especially in wild olive [59], and that could explain why M. arenaria did not disperse in a context of the low human activity.

Conclusion
To conclude, introductions of pest species through cropping practices are usually irreversible and frequently cause undesirable impacts. Therefore, we can assume