Introduction

Root-knot nematodes (RKN), Meloidogyne spp., belong to the order Tylenchida. These small round worms (typically from 300 μm to 2 mm for vermiform juveniles and pyriform females, respectively; Figure 1) live in soils and are obligate and sedentary endoparasites of plant roots. They harbour at their anterior end a hollow, protrusible stylet, which they use to both inject secretions into and withdraw nutrients from the infected root cells. They have evolved very sophisticated interactions with their host (Abad et al, 2003). As a result of the typical gall symptoms induced on the root system of the infested plants (Figure 1), their subsequent stunted growth and wilting, RKN are considered as significant agricultural pests. Their life cycle and reproduction mode vary considerably between species, ranging from classical amphimixis to obligatory mitotic parthenogenesis (Triantaphyllou, 1985). The occurrence of parthenogenesis is broadly correlated with increasing importance as crop pathogens.

Figure 1
figure 1

The root-knot nematode, Meloidogyne incognita. (a) Infective second-stage juveniles. Bar=100 μm. (b) Females dissected from the root tissues. Bar=500 μm, h=head. (c) Typical gall symptoms on tomato roots.

Full size image

Pluricellular eukaryotic organisms that reproduce exclusively by parthenogenesis are extremely rare, and often distributed in marginal, disturbed or ressource-poor environments, a pattern known as ‘geographic parthenogenesis’ (Peck et al, 1998; Kearney, 2003). Moreover, due to the benefits of sex in the long-term, that is the higher level of genetic diversity released by sexual reproduction, asexual lineages are generally considered as evolutionary dead ends (Maynard Smith, 1978; Kondrashov, 1993), although the debate is still running (Victoir and Dujardin, 2002; Lushai et al, 2003). From all these points of view, the mitotic parthenogenetic species of RKN appear as an outstanding exception. First, they are ubiquitous on the Earth's surface. For example, Meloidogyne incognita, the most widespread RKN species, is found from temperate to tropical regions, in any place where the lowest temperature is above 3°C (Sasser et al, 1983). Second, these biotrophic parasites have an extremely wide host range encompassing most of the flowering plants (Trudgill, 1997). In that respect, their actual distribution goes far beyond the restricted and highly specialized ecological niches supposedly inhabited by asexual organisms under the geographic parthenogenesis hypothesis. And third, it recognized that the parthenogenetic RKN species exhibit a high capacity to respond to environmental selection (eg their ability to overcome plant resistance genes; Castagnone-Sereno, 2002). Taken together, all these biological characteristics account for the extreme success of these clonal nematodes as parasites, and explain why M. incognita has been considered as ‘the world's most damaging plant pathogen’ (Trudgill and Blok, 2001).

The purpose of this review is to synthesize what is currently known on the genetics of Meloidogyne spp. in relation with their mode of reproduction, and the possible route(s) that lead them from sex to parthenogenesis, along with a discussion on the mechanisms that could generate genetic variability in these clonal organisms. We emphasize the potential interest in RKN as a model system for evolutionary studies into the origin and relative advantages of parthenogenetic lineages vs their sexual relatives.

Modes of reproduction and chromosomes

The 80 RKN species currently described (Karssen and Van Hoenselaar, 1998) have diversified to an extent that is, arguably, unparalleled in the animal kingdom. First, RKN display very different modes of reproduction. A few Meloidogyne species are amphimictic, and produce crossfertilized eggs after copulation. In that case, male and female genetic material is fused, and recombination creates genetic variation (ie new combination of genes) at each generation. However, these amphimictic species (including M. carolinensis, M. megatyla, M. microtyla, M. pini) are considered as minor RKN species because of their very restricted distribution, host range and economic impact (Jepson, 1987). In fact, most RKN reproduce by parthenogenesis, with many variations (Figure 2). Some species (eg M. chitwoodi, M. exigua, M. fallax, most M. hapla populations, etc.) can reproduce by both crossfertilization and meiotic (automictic) parthenogenesis. When males are present, mating occurs freely, and reproduction is by crossfertilization. When males are not present (thelytokous populations), the egg undergoes reduction of the chromosome number through meiosis, and the somatic chromosome number is re-established after fusion of the second polar nucleus with the egg pronucleus (Triantaphyllou, 1966; Dalmasso and Bergé, 1975; Van der Beek et al, 1998). The third mode of reproduction is obligatory mitotic (apomictic) parthenogenesis, and it is encountered in the most important species in terms of geographic distribution and agronomic impact, that is M. arenaria, M. incognita and M. javanica. In this case, there is neither reduction nor fusion of nuclei, and the egg directly develops into an embryo. When males are present, they can inseminate females, but the sperm nucleus degenerates and does not participate in fertilization (Triantaphyllou, 1962, 1963, 1981). In parthenogenetic RKN species, sex determinism is epigenetic, under strong influence of environmental factors (eg population size, host quality, etc.). When conditions are favourable, juveniles develop as females, but under poor resource conditions, they develop as males. In addition, some infrequent events of sex reversal and intersexuality have also been reported, for example female juveniles that develop into males (Triantaphyllou, 1973; Papadopoulou and Triantaphyllou, 1982). In parthenogenetic RKN, males play no role in reproduction, but they have an important biological function in terms of population density and dynamics, and thus ecological adaptation. Indeed, since only the apomictic females can produce offspring, abundance of males at generation n will significantly reduce the population size at generation n+1 (Triantaphyllou, 1973).

Figure 2
figure 2

Schematic representation of parthenogenetic pathways in root-knot nematodes.

Full size image

RKN are also highly variable with respect to their chromosomal complement. It is generally admitted that the haploid number of the genus is n=18, but most populations have somatic chromosome numbers ranging from 30 to 50, and thus are thought to be either diploids or triploids (Triantaphyllou, 1985). In fact, somatic chromosome numbers that are perfect multiple of 18 are not frequently observed, implying that there has been extensive aneuploidy or polysomy and structural rearrangements such as deletions, duplications, and translocations. These events may have been frequent, in part because, like most nematodes, RKN have a diffuse centromere lacking localized kinetochore activity (Triantaphyllou, 1983). Amphimictic RKN species are exclusively diploid, while diploid, triploid and rare tetraploid forms are encountered within parthenogenetic species (Table 1). As an example, most populations of M. incognita, the most prevalent apomictic RKN species, are considered to be (hypo)triploid, with a set of 3n=40–48 chromosomes, although diploid populations with chromosome numbers ranging from 2n=30 to 39 are not so infrequent (eg Janati et al, 1982; Marais and Kruger, 1991).

Table 1 Typical examples of cytogenetic variation in root-knot nematodes, Meloidogyne spp.
Full size table

Hypothesis on the origin of parthenogenesis in RKN

As they appear to have reproduced exclusively asexually for a long stretch of evolutionary time, the apomictic RKN species have been considered as one of the putative ‘ancient asexual scandals’ (Judson and Normark, 1996). Indeed, although the calibration of dates used indirect evidence, the divergence of the parthenogenetic RKN species from the amphimictic meiotic ones has been estimated to have occurred about 43 Myr ago (Esbenshade and Triantaphyllou, 1987) and might be far older (Hugall et al, 1997). However, no definitive evidence of asexuality has been provided for these nematodes (such as the Meselson effect described in Judson and Normark, 1996). Molecular studies have nevertheless confirmed that the apomictic RKN species share a common lineage, and that they diverged early from meiotic species (Castagnone-Sereno et al, 1993b; Baum et al, 1994). There are several ways in which parthenogenetic lineages could arise (Simon et al, 2003). In the case of RKN, no fossil records are available, and the ancestors of the genus are unknown. However, based on cytogenetic (Triantaphyllou, 1985) and isoenzyme data (Dalmasso and Bergé, 1983; Esbenshade and Triantaphyllou, 1987), the following assumptions are currently widely accepted: (1) the ancestral RKN were amphimictic animals, and the rare amphimictic species encountered today (eg M. carolinensis, M. megatyla) are considered as their closest relatives; (2) parthenogenetic species evolved from amphimictic species; (3) obligatory parthenogenetic (mitotic) species evolved from facultatively parthenogenetic (meiotic) species, following suppression of meiosis during oocyte maturation (Figure 3).

Figure 3
figure 3

Schematic representation of putative evolutionary relationships in root-knot nematodes (adapted from Dalmasso and Bergé, 1983). Lengths of arrows are not proportional to evolutionary time. Hypothetic datation (see Esbenshade and Triantaphyllou, 1987).

Full size image

Hybridization is a major route to parthenogenesis in animals, and may be implicated in the RKN. Polyploidization probably occurred by either intra- or interspecific hybridization (ie fertilization of an unreduced diploid oocyte by an haploid spermatozoon). Since functional, parthenogenetically produced males may be present in populations under poor environmental conditions, they could be involved in such exceptional fertilization events. For example, M. javanica is suspected to be a triploid interspecific hybrid species (Dalmasso and Bergé, 1983). In the same line, a possible reticulate hybrid origin of apomictic RKN has been hypothesized as the result of combinations of closely related females with more diverse parental lineages (Hugall et al, 1999). So far, no evidence has been provided for alternative hypotheses about the origin of parthenogenesis in RKN, such as spontaneous origin by mutation in genes involved in the production of sexual forms, or due to infection by microorganisms such as Wolbachia. However, we cannot definitely exclude the possibility that there have been several different routes to apomixis in RKN species, as examplified by the insects Otiorhynchus scaber (Stenberg et al, 2003) and Rhopalosiphum padi (Delmotte et al, 2003).

The species definition and clonal diversity in parthenogenetic RKN, from morphology to molecular data

The utility of species concepts has recently been challenged (see De Meeûs et al, 2003), but such debate is beyond the scope of this review. The widely used biological species concept defines species on the basis of reproductive isolation, and therefore applies only to organisms exclusively reproducing through bi-parental sexual reproduction. It is not, therefore, appropriate for most RKN species. In the case of (parthenogenetic) RKN, species definitions have therefore been based primarily on morphological traits from males, females and second-stage juveniles used in combination (eg stylet length and shape, tail and spicule shape of males and perineal patterns of females; Jepson, 1987). Genetic data, when available, generally support these groupings. Extensive allozyme electrophoresis studies have shown that the major RKN can be differentiated, in particular by esterase phenotypes, which correlate with the species previously recognized on the basis of morphology, thus supporting the concept of parthenogenetic species (Dalmasso and Bergé, 1978, 1983; Esbenshade and Triantaphyllou, 1985, 1990). The methodology works reliably on single female nematodes, and is still currenly used on a routine basis for diagnotic purposes. For the last 20 years, studies using DNA-based techniques have provided an additional perspective, and largely confirmed that the previously described species do constitute distinct genetic units. An exhaustive review of RKN molecular diagnostics would not be appropriate for this review, but relevant information about the main genetic regions commonly used for species characterization, among which ribosomal DNA, mitochondrial DNA, and anonymous loci (such as RFLP, RAPD and AFLP), can be found elsewhere (for reviews, see Hyman and Powers, 1991; Powers, 2004). Repetitive genomic sequences known as satellite DNA are thought to be involved in the stability of genome structure via chromosome folding and location during mitotic/meiotic events (Charlesworth et al, 1994; Csink and Henikoff, 1998), thus participating to the reproductive isolation of species. In meiotic parthenogenetic RKN, the distribution of such sequences is often restricted to populations belonging to the same species, thus confirming their specific status. In particular, this was experimentally shown for M. hapla (Piotte et al, 1995) and M. exigua (Randig et al, 2002).

Intraspecific RKN variability was initially detected during host-range studies, and led to the concept of ‘host race’, that is a set of populations sharing distinctive physiological characters that result in their ability to reproduce or not on selected differential hosts (Sasser, 1979). The further development of biochemical and molecular markers led to the production of an abundant literature and confirmed that a high level of clonal diversity exists among populations within RKN species, in particular within the apomictic species. However, cluster analyses did not reveal any correlation between genomic similarity and geographic origin of the populations (Blok et al, 1997; Semblat et al, 1998). Similarly, RKN populations from the same host-race do not have correlated molecular fingerprints (Cenis, 1993; Baum et al, 1994), suggesting that Meloidogyne races are not monophyletic groups, but rather evolved in a convergent manner.

High genetic diversity is not expected in obligatory apomictic taxa, yet is observed in RKN species. Its origin and maintenance demands explanation. There can be little doubt that the transition from sexual to asexual reproduction has happened many times during evolution of RKN (see above), and the current observed diversity probably results from a combination of (i) genetic variation accumulating during hybridization and persisting through subsequent polyploidization; (ii) multiple origins from a common sexual ancestor; and (iii) accumulation of independant mutations in different clonal lineages. So far, no correlation between RKN clonal diversity and biological parameters has been found (eg virulence interactions with resistant host genotypes; Semblat et al, 2000)). Taking into account their extremly wide host range and geographic distribution, apomictic Meloidogyne clones within each spesies may comprise a single (or a few) well-adapted genotype that has accumulated neutral variation, in accordance with the ‘general-purpose genotype’ model (Lynch, 1984). The resolving power of molecular tools has been increasing continuously, and allows us to distinguish larger and larger numbers of clonal genotypes. However, this extra information might actually mask the differences between ancient distinct clones and the more recent related lineages (ie those that have evolved from them only through accumulation of neutral or poorly selected mutations).

Response to plant resistance as an example of rapid adaptation

Based upon worldwide surveys, a general relationship between host-specificity and reproductive mode can be proposed for RKN, although some exceptions may occur (Jepson, 1987). Most amphimictic species are host-specific, either on woody or perennial herbaceous hosts (eg M. megatyla and M. pini, which are both restricted to Pinus spp.). Apart from M. hapla, the meiotic parthenogenetic species tend to have a narrow host range, while the mitotic species, including the extremely polyphagous major species M. arenaria, M. incognita and M. javanica, have a potential host range encompassing the majority of the flowering plants (Trudgill and Blok, 2001). However, because of the lack of cytogenetic data and complete host records for a significant number of Meloidogyne spp., it is currently not possible to try to relate RKN evolution with the evolution of their hosts. In RKN collected from natural populations, adaptation is well-documented in the form of virulence, that is overcoming of plant resistance genes. Such genes normally control avirulent populations of the parasite, and resistance is in most cases expressed as a localized necrosis of plant cells at the infection site (the so-called hypersensitive reaction (HR)), which prevents nematode installation and further development (Williamson and Hussey, 1996). However, virulent RKN populations, able to infest and develop on resistant plants without eliciting the HR, have been reported in field conditions. In particular, virulent genotypes belonging to the mitotic species M. arenaria, M. incognita and M. javanica are found in most of the World's tomato-growing areas. Their wide distribution and not infrequent occurrence is probably a consequence of the selection resulting from the monoculture of cultivars bearing the Mi resistance gene (Castagnone-Sereno, 2002). In the laboratory, artificial selection experiments demonstrated that the shift from avirulence to virulence can occur in a progressive manner in 5–10 generations of repeated inoculation of M. incognita on a Mi-resistant tomato (Jarquin-Barberena et al, 1991; Castagnone-Sereno et al, 1994). In the meiotic parthenogenetic species M. chitwoodi, a complete phenotypic change was observed after one single generation of selection on resistant Solanum fendleri carrying the Rmc2 gene (Janssen et al, 1998). One might speculate that the parthenogenetic RKN are better able to exploit new plant genotypes than their amphimictic relatives, which would explain their very wide host-range. However, such changes have been noted to have fitness costs, notably in the reproductive potential of the nematodes, which could make these genotypes less successful over the course of time (Trudgill and Blok, 2001; Castagnone-Sereno et al, submitted).

The molecular mechanisms by which plant-parasitic nematodes are able to overcome resistance genes are not yet elucidated. However, the comparative analysis of M. incognita isogenic lines, selected for their avirulence or virulence against the tomato Mi resistance gene for 25 generations only, allowed the identification of a candidate gene, map-1, coding for an amphid-secreted protein, present in the avirulent lines and lacking in the corresponding virulent lines (Semblat et al, 2001). Its detailed analysis suggested that it could be located in an unstable region of the nematode genome, which can be deleted without dramatically affecting the viability of the nematode (Semblat et al, 2001). More recently, a transcriptome analysis showed differences in gene expression profiles between M. incognita isogenic lines, that have been selected for their avirulence or virulence against the tomato Mi resistance gene for as few as 10 generations (Neveu et al, 2003). Interestingly, a previous study had shown that virulent nematodes could not revert back to an avirulent phenotype after 18 generations of continuous propagation on a susceptible host, that is without the selective pressure of the resistance gene (Castagnone-Sereno et al, 1993a). Overall, these data suggest that selection induces rapid changes in the genome of mitotic parthenogenetic RKN that could be maintained in subsequent generations, and thus could contribute to the emergence of new genotypes.

Putative mechanisms of genome evolution in parthenogenetic RKN

As mitotic parthenogenesis should theoretically produce clonal progenies, the adaptation of RKN to their environment (eg resistant hosts) raises questions about genome plasticity leading to genetic variation and adaptive evolution in apomictic animals. Besides nematodes, a wide range of other asexual taxa have been shown to undergo rapid genetic changes, including, among others, crustaceans (Schön et al, 2003) and insects (Wilson et al, 2003).

Among the mechanisms of genetic change that are independent from sexual recombination, those involving transposable elements are well documented in eukaryotes. Indeed, such mobile DNA sequences can provide host genomes with the ability to enhance their own evolution and can be a major source of genetic diversity, allowing response to environmental changes (Kidwell and Lisch, 2000). However, the dynamics and fate of transposable elements are strongly influenced by the mode of reproduction, asexual genomes having substantially reduced numbers of active transposons (Wright and Finnegan, 2001). In parthenogenetic organisms, theoretical predictions suggest that transposable elements could be beneficial in the short-term, but deleterious if present and still active in the long-term (Nuzhdin and Petrov, 2003). Indeed, it was shown that obligately parthenogenetic populations of Daphnia pulex contain only fixed (ie inactive) transposons (Sullender and Crease, 2001) and that ancient asexual bdelloid rotifers have lost all of the retrotransposables elements (Arkhipova and Meselson, 2000). Similarly, a molecular survey revealed that the genome of M. incognita totally lack sequences homologous to the Tc(s) transposable elements of Caenorhabditis elegans (Abad et al, 1991). However, mariner-like elements were found in the facultative meiotic parthenogenetic species M. chitwoodi and M. hapla, and in the obligate mitotic parthenogenetic species M. incognita and M. javanica (Leroy et al, 2000). Further cloning of the full-length Mcmar-1 mariner element in M. chitwoodi showed the presence of an uninterrupted ORF that encodes a putatively active transposase (Leroy et al, 2003), and detection of transcripts of the transposase in juveniles and females (unpublished data) suggests that this element may be mobile in the nematode. A very recent survey of the nematode EST database NEMBASE2, an upgraded version of NEMBASE (Parkinson et al, 2004), identified a number of transposase transcripts in RKN, but showed no drastic differences in their abundance in mitotic and meiotic species (unpublished). This preliminary result may nevertheless suggest that the occurrence of transposable elements may be comparable in the amphimictic and apomictic RKN. Whether such sequences are still active remains an open question, and the possibility that transposon-based mutation could generate genetic variability in these asexual organisms is currently under active investigation.

Qualitative and quantitative alteration of the genetic information can also occur at the level of the whole chromosome, and includes fissions, fusions, inversions and duplications. RKN exhibit exceptional levels of karyotypic variation even within species, and this is particularly true for species that have completely abandoned sexual reproduction. The map-1 gene is a good example of such variation in M. incognita, since the gene was shown to be present or absent in isogenic progenies selected for their avirulence or virulence to the tomato Mi resistance gene, respectively (Semblat et al, 2001). A more detailed analysis of this gene showed in its internal sequence the occurrence of highly conserved repetitive motives (Semblat et al, 2001), and cloning experiments revealed that different copies of the gene harbour a variable number of such motives (unpublished data). This result suggests that the gene may be present in an unstable region of a chromosome where amplification/deletion events might occur, resulting in the observed variability. Polyploidy can also contribute to genome evolution, and the emergence of new phenotypic and molecular variation shortly after polyploid formation has been documented (Soltis and Soltis, 1999; Osborn et al, 2003). The potential mechanisms that could generate such changes include homologous recombination and other types of DNA rearragements, point mutations and gene-conversion-like events (Song et al, 1995). Since RKN species exhibit variable levels of ploidy, these processes could be important in their genomes.

Conclusion

RKN are soil-living organisms, with limited capacity for dissemination and gene flow. Parthenogenesis gives advantages in terms of colonization (males and females do not need to meet for reproduction) and protection of the advantageous genotypes. More unexpectedly, although parthenogenesis is generally thought to reduce diversity, these species have retained sufficient genetic variation to allow rapid adaptation to unfavourable environments, for example resistant hosts. Lastly, the combination of parthenogenesis and rapid adaptive responses are reflected in an extremely wide host range and geographic distribution, and are probably the main reasons for the outstandingly successful establishment of asexual Meloidogyne spp. as plant parasites (Trudgill and Blok, 2001).

Understanding of the potential of these asexual organisms for adaptation will unquestionably provide knowledge of major fundamental and applied significance. From that point of view, RKN definitely constitute a very attractive model system. Indeed, the recently funded genome sequencing projects for the meiotic parthenogenetic species M. hapla (in the USA) and of the mitotic parthenogenetic species M. incognita (in France), and the comparative analysis of the resulting data, will represent a significant step forward in the quest to understand the molecular and genetic basis of genome structure and evolution in asexual organisms.