How anthropogenic changes may affect soil-borne parasite diversity? Plant-parasitic nematode communities associated with olive trees in Morocco as a case study

Background Plant-parasitic nematodes (PPN) are major crop pests. On olive (Olea europaea), they significantly contribute to economic losses in the top-ten olive producing countries in the world especially in nurseries and under cropping intensification. The diversity and the structure of PPN communities respond to environmental and anthropogenic forces. The olive tree is a good host plant model to understand the impact of such forces on PPN diversity since it grows according to different modalities (wild, feral and cultivated olives). A wide soil survey was conducted in several olive-growing regions in Morocco. The taxonomical and the functional diversity as well as the structures of PPN communities were described and then compared between non-cultivated (wild and feral forms) and cultivated (traditional and high-density olive cultivation) olives. Results A high diversity of PPN with the detection of 117 species and 47 genera was revealed. Some taxa were recorded for the first time on olive trees worldwide and new species were also identified. Anthropogenic factors (wild vs cultivated conditions) strongly impacted the PPN diversity and the functional composition of communities because the species richness, the local diversity and the evenness of communities significantly decreased and the abundance of nematodes significantly increased in high-density conditions. Furthermore, these conditions exhibited many more obligate and colonizer PPN and less persister PPN compared to non-cultivated conditions. Taxonomical structures of communities were also impacted: genera such as Xiphinema spp. and Heterodera spp. were dominant in wild olive, whereas harmful taxa such as Meloidogyne spp. were especially enhanced in high-density orchards. Conclusions Olive anthropogenic practices reduce the PPN diversity in communities and lead to changes of the community structures with the development of some damaging nematodes. The study underlined the PPN diversity as a relevant indicator to assess community pathogenicity. That could be taken into account in order to design control strategies based on community rearrangements and interactions between species instead of reducing the most pathogenic species.


Background
A biological community refers to an assemblage of populations from different organisms living together in a habitat. This biological assemblage within a community could be described by several traits such as the number of species (richness), their relative abundance (evenness), the present species (taxonomical structure), the interactions among them as well as their temporal and spatial variation [1]. Species diversity is important for the stability of the community and consequently that of the ecosystems [2]. For instance, functional consequences on ecosystem processes are related to species richness and to speciesspecific traits. Moreover, species diversity can play a crucial role in ecosystems resilience and/or resistance to human disturbances and to environmental changes [1].
Soil communities have been described as the "poor man's tropical rainforest", because of the relatively high level of biodiversity and the large proportion of undescribed species, as well as the limited information available about their community structure and dynamics [3]. Human interventions in ecosystems such as land-use changes, invasive species and over-exploitation, lead to biodiversity loss and/or species extinction [4]. For example, in agrosystems, crop intensification greatly disturbs the soils, affecting composition and functions of their biota [5,6].
Among soil biota, nematodes are ubiquitous soil inhabitants and among the most abundant and diversified biota [7]. They reflect several feeding behaviors that make it possible to allocate them to different trophic groups: bacterivores, fungivores, carnivores and plant feeders [8]. Due to the various life strategies of nematodes (r and K for colonizer and persister nematodes, respectively), their diversity and their co-existence in communities are closely related to short response time, to environmental changes and to disturbances in their habitats [9].
Plant-parasitic nematodes (PPN) are known to attack a wide range of crop plants (cereals, vegetables, tubers, fruits, flowers, etc.), causing annual crop losses estimated at billions of dollars in worldwide [10,11]. On the olive tree (Olea europaea L.), PPN are able to reduce tree growth [12] and may be responsible for 5-10% yield losses [13]. Their impact is especially strengthened in nurseries and in intensive cultivation systems where irrigation conditions favor the development of roots and, as a result, nematode multiplication [14]. A high diversity of PPN on olive trees was reviewed worldwide [14,15].
In Morocco, olive tree is a good example of ecological, botanical and genetic diversity. Spontaneous trees are distinguished under three different forms: (i) autochthonous wild trees, usually referred to as oleasters (O. europaea subsp. europaea var. sylvestris (Mill.) Lehr.), are common in coastal and mountainous regions [16]; (ii) the Moroccan hexaploid olive subspecies O. europaea subsp. maroccana is endemic in the High Atlas Mountains [17]; (iii) feral forms are wild-looking olive trees that correspond either to abandoned cultivated olive trees or to olive trees grown from cultivated olive seeds spread by birds. Additionally, cultivated forms (O. europaea subsp. europaea var. europaea) are also widespread. Different olive cropping systems can be distinguished according to tree density [18]: traditional orchards (ca. 80-400 trees/ ha) vs high-density orchards (up to 1800 trees/ha). However, these new intensive techniques, accompanied by the replacement of traditional low-intensive production with highly intensified and mechanized cultivation, including the use of herbicides to remove weeds, are expected to induce a possible degradation of the plant communities and their associated fauna [19]. As for olive propagation, it is generally performed from root cuttings that could be accompanied by soil transport and, consequently, by the spread of soil-borne parasites. Thus, PPN could be spread by soil transport or by unsanitized plant material (e.g. from uncertified nurseries). The local PPN populations in olive-growing areas could therefore have originated from historical mixtures made up of native (before olive introduction) and invasive (with root stocks from oleasters) communities. In this context, we hypothesize that PPN communities may have adapted to olive propagation processes and to cultivation practices. These anthropogenic forces could exist in Morocco where high-density cultivated areas have been extended and where ancestral or traditional cultivars have often been discarded in favor of a few highly productive varieties [20]. These new conditions of cultivation might have to face a resurgence of several pests, including PPN. To address these hypotheses, this study was undertaken in order to: (i) describe the species diversity of PPN communities associated with wild, feral and cultivated olives in Morocco where their diversity is completely unknown, and (ii) assess how anthropogenic forces (propagation and intensification practices) could impact the diversity and the structure of PPN communities by comparing them between different olive growing modalities.

Site description
Sampling of soil and olive leaves took place in Morocco from March to April 2012. Wild olive locations were as far as possible from current orchards. In contrast, feral olive locations were sampled within the proximity of cultivated olive stands or near main roads. The survey was conducted at 94 sites in several geographic regions all along a northeast-southwest 900-km long transect ( Fig. 1; Table 1 cultivation), tillage and other human activities are frequent, which could lead to the homogenization of the PPN communities in an orchard. Each orchard was therefore considered as a repetition per modality. The sampling was carried out in each orchard along transects under four trees located at a distance of approximately 10 m. Five sub-samples were collected from each tree. These 20 sub-samples were thoroughly mixed to obtain a single representative sample per orchard. Contrary to cultivated orchards, heterogeneous PPN communities were expected in wild and feral olive trees because human interventions are scarce or absent. Each tree was thus taken as a repetition. Five sub-samples were also collected from each tree and then combined to form one 1-dm 3 reference sample per tree.

Genetic characterization of the olive tree
In order to confirm the determination of olive-growing modalities, three olive branches corresponding to soil samples were collected to determine the chloroplast haplotype of each tree (according to [22]

Nematode extraction, identification and quantification
All of the nematode analyses were performed in the nematode quarantine area (French Government Agreement No 80622) of the Research Unit, "Centre de Biologie pour la Gestion des Populations" (Montpellier, France). A 250-cm 3 wet aliquot was taken from each soil sample for nematode extraction using the elutriation procedure [23]. PPN belonging to the Aphelenchida, Dorylaimida, Triplonchida and Tylenchida orders were enumerated in 5-cm 3 counting chambers [24] and identified at the genus level based on dichotomous keys [25] and at the species level with genus-specific keys. The population levels were expressed per dm 3 of fresh soil. Concerning specific identification, the nematode suspensions were preserved in mixture of formalin and glycerine [26], and then adult specimens were processed according to Seinhorst method [27] and mounted onto slides [28] for microscopic observation. Root-knot nematodes (Meloidogyne spp.) were identified at the species level by biochemical (esterase patterns) and molecular (SCAR markers and 28S rDNA D2-D3 expansion segments) approaches [29].

Analyses of nematode diversity
Several ecological indices were used: a. Taxonomical diversity: (i) the total number of PPN in a community (N); (ii) the species richness (S); (iii) the Shannon-Wiener diversity index H' (H′ = −∑p i lnp i , where p i is the proportion of individuals in each species (iii) that quantifies the local diversity or the heterogeneity of diversity (H′ ranges from 0 to ln(S)); and (iv) the evenness (E = H′/ln S) that quantifies the regularity of species distribution within the community (E varies between 0 and 1). b. Functional diversity: PPN species detected in communities were distributed into life-strategy groups according to the colonizer/persister value (cp-value) of the family to which they belong [30]. The diversity of the community was described by calculating: (i) the plant-parasitic index (PPI = ∑cp i n i /N), which quantifies the plant-feeding diversity of the communities; (ii) the relative mean abundance (%) of each cp-value class in a community calculated as follows: Rcp i = cp i n i /N; (iii) the genus richness included in each cp-value class. PPN species were also assigned to the trophic groups according to their feeding habits [31,32]: obligate plant feeders (OPF), facultative plant feeders (FPF) that alternatively feed on fungi, and fungal feeders (FF) that alternatively feed on plants. These trophic groups were also described according to (i) the relative mean abundance (%) of individuals within each of them, and (ii) the genus richness included in each [33]. c. The structure of PPN communities was designed at the genus level. The dominance of each nematode genus in the samples was first estimated by modeling the abundance (A) and the frequency (F) of each genus in the whole samples [34]. Afterwards, PPN community structures were described according to multivariate statistical analyses.

Data analyses
These diversity indices were calculated using the Vegan library [35]. In order to evaluate the impact of anthropogenic changes on biodiversity and community structures, different olive variables were defined according to olivegrowing modalities: wild (WO), feral (FO), traditional or low-density cultivation (TR) and modern or high-density cultivation (HD), and according to olive irrigation conditions: irrigated or rainfed. The mean values of the different nematode diversity indices were compared according to olive propagation (wild vs cultivated) and to intensification practices (traditional vs high-density, irrigated vs rainfed). Principal Component Analysis (PCA) was carried out on nematode genera data in order to describe PPN community structures. To assess the impact of olive anthropogenic changes on taxonomical structures, a co-Inertia Analysis (CIA) was applied between olive-growing modality data (WO-FO-TR-HD) and PPN genera. The scarcest genera (with total abundance less than 1%) were then excluded from the dataset prior to running the analysis. These different multivariate analyses and graphs were performed using ade4 library [36,37]. All analyses were done using R version 3.3.2 [38]. The Wilcox (nonparametric) test was used for all pair-wise multiple comparisons. Differences obtained at levels of P < 0.05 were considered to be significant.

PPN diversity associated with olive trees in Morocco
The PPN communities associated with olive trees in Morocco were highly diversified. A total of 117 species and 47 genera were identified. They belong to two families of Aphelenchida, to a family of Dorylaimida, to a family of Triplonchida and to 14 families of Tylenchida (Table 2). At the family level, the Tylenchidae and Telotylenchidae were dispersed in all the regions sampled; they were the most diversified families, including 11, 9 genera in each, respectively. However, each genus was often represented by one or two species only (e.g. Amplimerlinus, Bitylenchus, Tylenchus). Most of these species were very rare as they were detected in one or two sites only (e.g. Aglenchus agricola, Coslenchus gracilis and Paratrophurus loofi in the Rif region). In contrast, the Hoplolaimidae family was represented by two genera only (Helicotylenchus and Rotylenchus), but the number of species identified in each genus was high (11 and 4 species, respectively), and they were distributed in all the regions, except in eastern Morocco (the Kandar and Jel regions). Longidoridae and Trichodoridae nematodes were detected mostly in the Rif region. Root-lesion nematodes (e.g. Pratylenchus) and Pin nematodes Paratylenchidae (e.g. Paratylenchus) were dispersed at all the sites surveyed. Four root-knot nematodes species were identified: Meloidogyne arenaria and M. hapla were detected in the Rif region, M. javanica was generally detected in southern Morocco (in the Souss and Haouz regions) and in the Guerouane and Tadla regions. M. spartelensis is a new species identified in the Rif region; another new species seems to occur in the Souss region (identification is in progress). Other families such as Criconematidae and Psilenchidae were detected in a few sites.
Among the 47 identified genera, Filenchus, Helicotylenchus, Merlinius, Paratylenchus, Pratylenchus, Rotylenchus, Tylenchorhynchus and Xiphinema were the most widespread in olive soils. Considering the species level, 11 Helicotylenchus species (Hoplolaimidae) were frequently collected in olive samples. Among them, H. crassatus was clearly the most dominant species (occurring in 58% of the samples). It was present in all regions except in the Jel and Kandar regions. H. dihystera and H. varicaudatus also occurred in 43 and 32% of the samples, respectively. In contrast, H. exallus and H. minzi, detected in the Guerouane region, and H. pseudorobustus, detected in the Haouz region, were scarcer. In addition, Merlinius brevidens (Telotylenchidae) and Filenchus filiformis (Tylenchidae) were also frequently recovered (51 and 40% of the samples, respectively).

Diversity of PPN communities according to anthropogenic changes
Diversity indices mean values were compared between to the four olive-growing modalities and between rainfed and irrigated olive samples.

(a) Taxonomical diversity
The total number of PPN (N) was up to two times higher on cultivated (HD & TR) than on non-cultivated olive (WO & FO). Similarly on irrigated olive, the total number of PPN was higher (Table 3). In contrast, the PPN communities were significantly richer in species (S), more diversified (H′) and more homogenously distributed (E) in communities on WO and FO and on rainfed olive than on TR and HD and on irrigated olive.

(b) Functional diversity
The PPN identified were allocated in all the parasitic cpvalues (cp-2 to cp-5 groups, Table 2). The WO and HD modalities revealed nematode communities with significantly higher plant-parasitic indices (PPI) than those in FO and in TR orchards (Table 4). This means that WO and HD olive areas had significantly more plant-feeding nematodes with higher cp values than other olive systems. The most opportunist/colonizer PPN (cp-2 and cp-3) dominated in all the communities (44 and 48%, respectively; Table 2). The overall abundance and occurrence of the persister nematodes (cp-4 and cp-5) was very low (4% for each cp class). Any effect was recorded on the cp-4 class. Cp-2 and cp-3 nematodes were more abundant in TR and HD, while cp-5 nematodes occurred more often in WO areas and were completely absent in HD orchards.
Concerning the trophic groups within communities, the OPF nematodes were the most dominant (62%), while the FPF and the FF nematodes were the least frequent (26 and 12%, respectively). FF nematodes were significantly more numerous in WO areas (Table 4). FPF and OPF nematodes were more abundant in TR and        HD orchards,w respectively. The ratio between FPF and OPF nematodes was unbalanced in favor of OPF in HD orchards, and in favor of FPF in TR orchards and in FO areas. The rainfed-irrigation modalities did not have any effect on the trophic groups.
The cp-2, cp-3, FPF and OPF functional groups were represented by the highest number of genera (44,48,26 and 62%, respectively). Comparing this richness in each group between olive-growing modalities only, the PPN communities detected in WO and FO demonstrated higher richness and diversity compared to those detected in TR and HD (Table 5).

(c) Community patterns
Community structure was described at the genus level. Modeling the dominance of each genus in the samples (Fig. 2a), 83% of the genera were classified as less frequent (F < 30%) according to the model and 35% as occasional (F < 5%). A total of 62.5% of the nematode genera were classified as highly abundant according to the abundance threshold defined by the model (A = 200 nematodes/ dm 3 of soil). Eight genera were classified as dominant (F ≥ 30% and A ≥ 10,000 nematodes/dm 3 of soil): Filenchus and Helicotylenchus (F > 80%); and Rotylenchus, Merlinius, Paratylenchus, Xiphinema, Pratylenchus and Tylenchorhynchus (40 < F < 70%). Six other highly abundant genera were less frequent, including root-knot nematodes (Meloidogyne spp., F = 12.2%) and cyst nematodes (Heterodera spp., F = 10%). No genus was found to be frequent and in low abundance.
As shown by the PCA loading plot of the nematode taxa (Fig. 2b)

Correspondences between PPN community patterns and olive-growing modalities
Considering olive-growing modalities, the loading plot of the Co-Inertia Analysis (CIA) analysis between nematode and olive data (Fig. 3) indicated an important contribution of the anthropogenic gradient (WO-FO-TR-HD) to the CIA1 axis. The CIA2 axis was essentially correlated with the feral growing modality (FO, positive values) and with the wild olive (WO, negative values). Regarding the projection of the nematode genera in the loading plot (Fig. 3), the analysis indicated that the genera Merlinius, Xiphinema, Heterodera, Nothotylenchus, Rotylenchulus and Boleodorus were correlated with WO. In contrast,

Table 3 Taxonomical diversity indices in PPN communities associated with olive (mean values) according to olivegrowing modalities and water supply
The letters (a-c) indicate significant differences among the variables measured according to ANOVA and Wilcoxon tests.     Table 6. Dotted lines indicate delineation between low and high abundances and frequencies as described in [34]. b Plant-parasitic nematode community patterns (PCA loading plot for the nematode genera) confirmed that Meloidogyne and Tylenchorhynchus nematodes were significantly more abundant in HD orchards compared to TR orchards or to WO + FO. Some significant differences were also detected between traditional and non-cultivated olive orchards. However, other nematodes such as Merlinius, Xiphinema and Heterodera were found to be significantly more abundant in WO + FO compared to cultivated olive conditions (HD, TR).

Discussion
Biodiversity is an essential ecological phenomenon because it represents a complex set of interacting ecological, evolutionary, biogeographical and physical processes [39]. Native biodiversity is being lost at a rapid rate owing to anthropogenic causes, including habitat destruction, pollution and the spread of non-native species [1,40]. In this context, the main focus of this study was to understand how human activities (e.g. agricultural practices) in ecosystems could impact the diversity of PPN communities. The Mediterranean olive tree is particularly suitable for this study because it concerns ancient ecosystems with post-glacial refugia [16], many spots of Oleaster and many cases of feral olive. It also offers a large range of varieties, cultivated traditionally or at high-density, as present in Morocco.

PPN diversity associated with olive trees in Morocco
The PPN fauna and their distribution was totally unknown in Morocco before this study, except for a few reports on some nematodes such as root-knot nematodes Meloidogyne morocciensis [41] and cereal cyst nematodes [42]. This study clearly highlights a high taxonomical diversity of PPN communities where 117 species belonging to 47 genera were recorded. In addition, the study adds taxa (seven genera and 60 species) that were recorded for the first time in association with olive trees worldwide. The dominance pattern was also revealed by PCA analyses that demonstrated that the nematode dataset was mainly structured by the most frequent and abundant genera, and by less frequent but abundant nematodes to a lesser extent. The communities observed were mainly dominated by Filenchus and Helicotylenchus genera, and other nematodes such as Rotylenchus, Merlinius, Paratylenchus, Xiphinema, Pratylenchus and Tylenchorhynchus. Some of them have been previously reported as widespread on olive trees worldwide [15]. High population levels of some nematode genera such as root-knot nematodes (Meloidogyne spp.) and cyst nematodes (Heterodera spp.), considered as very dangerous soil-borne plant pests were also recorded [43]. The taxonomical diversity of PPN analyzed in Morocco is the greatest when compared to other surveys on olive trees that documented 223 species worldwide (reported in [14,15,[44][45][46][47]). This high diversity and the detection of new taxa could be essentially explained by: (i) a large sampling effort (213 soil samples corresponding to 363 trees sampled), conducted along a long transect (about 900 km) covering a wide range of olive-growing regions in Morocco; and (ii) a large proportion of samples collected in wild and feral olive areas (163 samples). These olive habitats could be considered as reservoirs of high diversity where a part remains unknown [48]. As evidence, a new root-knot nematode species, Meloidogyne spartelensis, was detected on wild olive in Northern Morocco [49]. However, other species could not be detected because they may occur only under unidentifiable life stages (e.g. juveniles), or their development may be linked to other periods of the year or to specific microhabitats [50]. As an example, no Rotylenchulus could be identified at the species level because all individuals were in the juvenile stage.

Impact of anthropogenic changes on the PPN communities associated with olive trees in Morocco
Taxonomical diversity indices were revealed impacted by olive propagation practices (from wild to cultivated olive): a high PPN richness was found in non-cultivated olive areas (wild and feral), with an equal distribution of species within communities (high evenness), contrary to what was observed in cultivated orchards (traditional and high-density). Nematode abundance was also significantly higher in orchards. A main conclusion also arose in this study that showed that PPN are abundant in cultivated conditions while richness, local diversity and evenness are low, and vice versa in non-cultivated conditions. In other words, a high PPN species diversity within a community may prevent the multiplication of the species as a potential effect of trade-off interactions between nematode species and/or between them and other soil microorganisms [51,52]. The study also highlighted the impact of anthropogenic practices on the functional diversity in communities: persisters and fungal-feeders were more diverse and numerous in wild olive conditions, whereas colonizers were frequently present under high-density conditions. Colonizer nematodes were represented by fewer genera, confirming imbalance between the high relative abundance and the low-genus richness and vice versa. Moreover, cp-5 nematodes were particularly related to wild olive, and totally absent under high-density olive cultivation conditions. This is consistent with other studies that demonstrated that cultivation intensification usually does not reduce the number of nematode trophic groups, but may change the composition of these groups [53]. The taxonomical structures of the communities were also distinguished between wild and cultivated olive: genera such as Xiphinema and Heterodera were detected in relation to natural ecosystems (wild olive), while others (e.g. Meloidogyne and Tylenchorhynchus) were favoured in cultivated areas. Dominant taxa such as Helicotylenchus, Rotylenchus and Filenchus did not appear to be impacted, which could explain their high dominance in the samples.