SEARCH

SEARCH BY CITATION

Keywords:

  • co-evolution;
  • compatibility;
  • molecular phylogeny;
  • mutualism;
  • pathogen;
  • specialization;
  • symbiosis

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Natural history of the model system
  5. Materials and methods
  6. Steinernema/Xenorhabdus stock
  7. Aposymbiotic nematodes
  8. Culture and count of bacteria
  9. Combinations of S. carpocapsae with non-native bacteria
  10. Estimating nematode fitness
  11. Transmission of non-native symbionts by S. carpocapsae
  12. Statistical analyses
  13. Phylogenetic analyses
  14. Does the effect of a bacterial strain depend on its position in the phylogenetic tree?
  15. Results
  16. Parasitic success of S. carpocapsae with non-native symbionts
  17. Reproductive rate of S. carpocapsae with non-native symbionts
  18. Transmission of non-native symbionts by S. carpocapsae
  19. Discussion
  20. Acknowledgments
  21. References

In this paper, we investigate the level of specialization of the symbiotic association between an entomopathogenic nematode (Steinernema carpocapsae) and its mutualistic native bacterium (Xenorhabdus nematophila). We made experimental combinations on an insect host where nematodes were associated with non-native symbionts belonging to the same species as the native symbiont, to the same genus or even to a different genus of bacteria. All non-native strains are mutualistically associated with congeneric entomopathogenic nematode species in nature. We show that some of the non-native bacterial strains are pathogenic for S. carpocapsae. When the phylogenetic relationships between the bacterial strains was evaluated, we found a clear negative correlation between the effect a bacterium has on nematode fitness and its phylogenetic distance to the native bacteria of this nematode. Moreover, only symbionts that were phylogenetically closely related to the native bacterial strain were transmitted. These results suggest that co-evolution between the partners has led to a high level of specialization in this mutualism, which effectively prevents horizontal transmission. The pathogenicity of some non-native bacterial strains against S. carpocapsae could result from the incapacity of the nematode to resist specific virulence factors produced by these bacteria.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Natural history of the model system
  5. Materials and methods
  6. Steinernema/Xenorhabdus stock
  7. Aposymbiotic nematodes
  8. Culture and count of bacteria
  9. Combinations of S. carpocapsae with non-native bacteria
  10. Estimating nematode fitness
  11. Transmission of non-native symbionts by S. carpocapsae
  12. Statistical analyses
  13. Phylogenetic analyses
  14. Does the effect of a bacterial strain depend on its position in the phylogenetic tree?
  15. Results
  16. Parasitic success of S. carpocapsae with non-native symbionts
  17. Reproductive rate of S. carpocapsae with non-native symbionts
  18. Transmission of non-native symbionts by S. carpocapsae
  19. Discussion
  20. Acknowledgments
  21. References

In species that engage in mutualistic interaction, traits involved in specialization are potentially under strong selective pressure: each species should increase its compatibility with its partner to gain more benefits from the interaction (Thompson, 1982). Under this scenario, congeneric populations of both partners can experience rapid parallel differentiation, so that two genotypes are more compatible when taken from the same population than from distinct ones (Wilkinson et al., 1996). When transmission is horizontal, the ultimate limit to specialization is that extremely specialized mutualists risk not finding suitable partners. Some authors have therefore proposed that mutualistic species should remain generalists (Law & Koptur, 1986; Herre et al., 1999; Leigh, 1999; Bergstrom & Lachmann, 2003).

The work of Frank (1996) sheds a different light on this issue. He proposed that host and symbiont are in conflict over the mixing of symbiont lineages. Within their host, symbionts compete with close relatives, and should therefore experience strong selection to disperse and infect new hosts. However, the host has no interest in its benefits from the symbionts being reduced by such competition and should oppose incoming horizontal transfers, as they may increase competition among symbiont lineages and therefore promote intra-host evolution towards higher virulence. This conflict can potentially determine how transmission modes evolve in mutualisms. It may also have consequences for the evolution of specialization in the host, if specialization is seen as a trait that limits symbiont mixing. Vertical transmission and specialization are thus complementary ways to reduce mixing of symbiont lineages.

Much information can be gained on these issues from the study of experimentally tractable systems such as mutualistic symbioses involving a micro-organism partner. In these systems, the interaction between partners that do not co-occur naturally (non-native combinations) can be compared with that of native combinations (interaction between partners that do co-occur naturally). Such experiments have been performed on two well-known symbiotic associations: the SepiolidVibrio system (Nishiguchi et al., 1998) and the legume–Bradyrhizobium system (Wilkinson et al., 1996). In both systems, combinations between non-native partners are potentially frequent in nature because symbionts are horizontally transmitted (Spoerke et al., 1996; Nyholm et al., 2000). In either case, host and non-native mutualistic symbiont are able to associate but the interaction that ‘functions best’ is the one between native partners. These experimental results strongly suggest co-evolution between native partners, but this mutual specialization does apparently not lead to incompatibility between non-native partners. The similar type of study has been performed in the attine ants, which are obligatorily dependent on fungal symbionts that are vertically transmitted. Experimental confrontation of ant hosts with symbionts from other nest of the same population demonstrated some incompatibilities between them (Bot et al., 2001). The authors proposed that these incompatibilities were explained in part by genetic differences between cultivars, but also by a hostile behaviour of the ants towards ‘foreign’ cultivars, so that mixing of symbiotic lineages was effectively avoided (Bot et al., 2001). As stated above, transmission mode and specialization appear tightly coupled: incompatibilities are found in systems in which symbionts are vertically transmitted; specialization exists but to a lesser degree when transmission is horizontal.

Here, we have experimentally combined non-native mutualistic partners to examine the degree of specialization between the nematode Steinernema carpocapsae and its native symbiotic bacteria Xenorhabdus nematophila during the parasitism of an insect host, Galleria mellonella. We tested the compatibility between S. carpocapsae and different foreign bacteria which are each associated with a different entomopathogenic nematode in nature (Fischer-Le Saux et al., 1997). The tested bacteria were two conspecific strains of X. nematophila, six congeneric strains and two further strains belonging to Photorhabdus, the sister genus of Xenorhabdus. The two Photorhabdus species are associated with entomopathogenic nematodes of the genus Heterorhabditis, which are very different from the genus Steinernema (Poinar, 1993; Blaxter et al., 1998), but share a similar entomoparasitic biological life cycle. We partially sequenced the 16S rDNA of the symbionts to test whether phylogenetic distance between native and introduced symbionts affected the compatibility between partners and fitness of the nematode.

Natural history of the model system

  1. Top of page
  2. Abstract
  3. Introduction
  4. Natural history of the model system
  5. Materials and methods
  6. Steinernema/Xenorhabdus stock
  7. Aposymbiotic nematodes
  8. Culture and count of bacteria
  9. Combinations of S. carpocapsae with non-native bacteria
  10. Estimating nematode fitness
  11. Transmission of non-native symbionts by S. carpocapsae
  12. Statistical analyses
  13. Phylogenetic analyses
  14. Does the effect of a bacterial strain depend on its position in the phylogenetic tree?
  15. Results
  16. Parasitic success of S. carpocapsae with non-native symbionts
  17. Reproductive rate of S. carpocapsae with non-native symbionts
  18. Transmission of non-native symbionts by S. carpocapsae
  19. Discussion
  20. Acknowledgments
  21. References

The mutualistic symbiosis between the nematode S. carpocapsae and the bacterium X. nematophila (Poinar & Thomas, 1966; Poinar & Himsworth, 1967) is a good model for investigating specialization in mutualism (Sicard et al., 2003). The infective juveniles (IJs) of S. carpocapsae harbour their extra-cellular clonal symbionts in a special intestinal organ called vesicle (Bird & Akhurst, 1983). These nematode–bacteria associations are pathogenic for a wide range of insect hosts such as Lepidoptera, Diptera, Coleoptera, Orthoptera and Hymenoptera (Laumond et al., 1979; Poinar, 1979). The life cycle of Steinernema–Xenorhabdus insect interactions is presented Fig. 1. After entering the insect, all the penetrated nematodes release their Xenorhabdus symbionts, which kill the insect, after which the gonochoric (separate sexes) adults reproduce for several generations leading to the production of offspring (IJs) (Poinar & Thomas, 1966). The IJs recover the symbionts from the insect cadaver and form a dispersing generation of symbiotic IJs. A recent study has revealed that S. carpocapsae naturally associated with native bacteria produce more juveniles during the parasitism of insects than aposymbiotic (i.e. nonsymbiotic) nematodes (Sicard et al., 2003). However, we do not know whether this strongly positive effect on fitness is due to specialization of the interaction between S. carpocapsae and its native symbiont X. nematophila.

image

Figure 1. The life cycle of the Steinernema–Xenorhabdus insect interactions. After entering the insect, the nematodes release their Xenorhabdus symbionts which multiply, causing a septicaemia that finally overwhelms the insect. After the death of the insect, the males and females of the nematodes reproduce sexually for a few generations leading to the production of eggs that develop into L1, L2, L3 and L4 larvae. The special L3 take up symbionts in their intestinal vesicle before leaving the cadaver are called Infective Juveniles (IJs).

Download figure to PowerPoint

The pattern of transmission of Xenorhabdus throughout generations of Steinernema is determined by (i) the ability of a symbiont genotype to outcompete its competitors in the insect and (ii) the level of specificity of the re-association between the symbionts and new generation of nematode larvae (IJs). In case of a single infestation by nematodes of the same population, these hosts release their bacterial symbionts into the insect's hemocoele where they multiply alone. They can thus be regarded as being transmitted ‘pseudo-vertically’ (Wilkinson, 2001). However, multiple-infestations of the same insect host with individuals from different Steinernema–Xenorhabdus populations (or species) may induce host-switching if the different bacterial clones are all able to multiply and if the associations are not too specific. As predicted by Frank (1996), we can suppose that symbionts have an interest in maintaining their options for horizontal transmission, but that the nematodes have been selected to counteract these tendencies. Previous studies in vitro to investigate the compatibility between Steinernema spp. and Xenorhabdus spp. have shown that most of the cultures of Steinernema spp. can be supported by non-native symbionts and that re-associations between some non-native partners were possible (Akhurst, 1983; Aguillera & Smart, 1993; Grewal et al., 1997). This suggests that this mutualism is not highly specialized and that host-switching between different populations or species of Steinernema can occur. The present study shows that strong fitness effects occur in vivo in spite of these in vitro compatibilities between hosts and symbionts.

Steinernema/Xenorhabdus stock

  1. Top of page
  2. Abstract
  3. Introduction
  4. Natural history of the model system
  5. Materials and methods
  6. Steinernema/Xenorhabdus stock
  7. Aposymbiotic nematodes
  8. Culture and count of bacteria
  9. Combinations of S. carpocapsae with non-native bacteria
  10. Estimating nematode fitness
  11. Transmission of non-native symbionts by S. carpocapsae
  12. Statistical analyses
  13. Phylogenetic analyses
  14. Does the effect of a bacterial strain depend on its position in the phylogenetic tree?
  15. Results
  16. Parasitic success of S. carpocapsae with non-native symbionts
  17. Reproductive rate of S. carpocapsae with non-native symbionts
  18. Transmission of non-native symbionts by S. carpocapsae
  19. Discussion
  20. Acknowledgments
  21. References

All Steinernema species and their associated bacteria (naturally symbiotic nematodes) were established in laboratory culture as soon as they were sampled by successive experimental infestations of the last instar of the wax moth Galleria mellonella (Lepidoptera: Pyralidae). Insect hosts were reared in the dark in aired plastic boxes at 28 °C, 65% RH, on a diet of pollen and wax. The IJs of S. carpocapsae can be preserved alive during a couple of month in sterile water at 8 °C.

Aposymbiotic nematodes

  1. Top of page
  2. Abstract
  3. Introduction
  4. Natural history of the model system
  5. Materials and methods
  6. Steinernema/Xenorhabdus stock
  7. Aposymbiotic nematodes
  8. Culture and count of bacteria
  9. Combinations of S. carpocapsae with non-native bacteria
  10. Estimating nematode fitness
  11. Transmission of non-native symbionts by S. carpocapsae
  12. Statistical analyses
  13. Phylogenetic analyses
  14. Does the effect of a bacterial strain depend on its position in the phylogenetic tree?
  15. Results
  16. Parasitic success of S. carpocapsae with non-native symbionts
  17. Reproductive rate of S. carpocapsae with non-native symbionts
  18. Transmission of non-native symbionts by S. carpocapsae
  19. Discussion
  20. Acknowledgments
  21. References

In order to produce aposymbiotic IJs of the nematodes from the same in vivo conditions as the symbiotic ones, we disinfected the surface of the eggs of the nematodes by crushing 40 mature females in sterile Ringer solution (NaCl 0.9%, w/v) with sodium hypochloride (10%, w/v) during 18 min. The disinfected eggs were then rinsed twice with sterile Ringer and transferred to ‘liver-agar’ plates (Sicard et al., 2003) for incubation at 24 °C. Three weeks later, axenic (i.e. grown without any germs) IJs were obtained from these plates, which were then available for further experimentations. Sterility of the IJs was assessed by inoculating nutrient broth tubes with samples from the liver plates. Aposymbiotic nematodes were produced by infesting insects with axenic IJs. The IJs emerging from these infestations were not axenic as they came from insect cadavers containing their own microflora. However, they were aposymbiotic because they grew within the insect in the absence of Xenorhabdus and did not obtain any saprophytic bacteria from the insect in their intestines (Sicard et al., 2003).

Culture and count of bacteria

  1. Top of page
  2. Abstract
  3. Introduction
  4. Natural history of the model system
  5. Materials and methods
  6. Steinernema/Xenorhabdus stock
  7. Aposymbiotic nematodes
  8. Culture and count of bacteria
  9. Combinations of S. carpocapsae with non-native bacteria
  10. Estimating nematode fitness
  11. Transmission of non-native symbionts by S. carpocapsae
  12. Statistical analyses
  13. Phylogenetic analyses
  14. Does the effect of a bacterial strain depend on its position in the phylogenetic tree?
  15. Results
  16. Parasitic success of S. carpocapsae with non-native symbionts
  17. Reproductive rate of S. carpocapsae with non-native symbionts
  18. Transmission of non-native symbionts by S. carpocapsae
  19. Discussion
  20. Acknowledgments
  21. References

The bacterial strains used in this study and their native nematode species are listed in Table 1. After extraction from the nematodes, they were maintained in liquid nitrogen in a 17% glycerol solution. Before use, they were streaked onto nutrient agar supplemented with 0.004% (w/v) triphenyltetrazolium chloride and 0.0025% (w/v) bromothymol blue (NBTA). The Petri dishes were incubated at 28 °C for 48 h. Subcultures were obtained by transferring a single colony of bacteria to 5 mL of Luria–Bertani broth for liquid culture incubated at 28 °C during 15 h. A quantity of 100 μL of this liquid subculture was used to perform a culture to reach an optical density of 0.7. Before inoculation of the insects, the number of Xenorhabdus or Photorhabdus cells in the culture was counted on Thoma cells in order to inoculate a diluted suspension of an approximately constant concentration of 2000 cells/20 μL. A control of the actual number of bacteria inoculated was carried out by colony-forming units by plating diluted suspension out on nutrient agar plates.

Table 1.  List of bacterial strains, native nematode species with their geographical origin, accession numbers of bacterial 16S rDNA partial sequences and number of combination experiments for each bacterium tested.
Bacterial species and strainsNative nematode speciesAccession no. of the 16S rDNA sequenceGeographical originNo. of combination experiments
Xenorhabdus nematophila F1S. carpocapsaeAY521241France80
X. nematophila PL31S. carpocapsaeAY521242Poland80
X. nematophila AN6S. carpocapsaeD78009USA160
X. poinarii SK72S. glaseriAY521239USA100
X. poinarii G6S. glaseriD78010USA120
X. beddingii Q58Steinernema sp.D78006Australia80
X. bovienii FR10S. feltiaeAY521240France80
Xenorhabdus sp. USTX62S. riobraveAY521244USA80
Xenorhabdus sp. UY61S. scapterisciAY521243Uruguay80
Photorhabdus luminescens TT01H. bacteriophoraAJ007404Trinidad80
P. temperata XlNachH. megidisAJ007405Russia80

Combinations of S. carpocapsae with non-native bacteria

  1. Top of page
  2. Abstract
  3. Introduction
  4. Natural history of the model system
  5. Materials and methods
  6. Steinernema/Xenorhabdus stock
  7. Aposymbiotic nematodes
  8. Culture and count of bacteria
  9. Combinations of S. carpocapsae with non-native bacteria
  10. Estimating nematode fitness
  11. Transmission of non-native symbionts by S. carpocapsae
  12. Statistical analyses
  13. Phylogenetic analyses
  14. Does the effect of a bacterial strain depend on its position in the phylogenetic tree?
  15. Results
  16. Parasitic success of S. carpocapsae with non-native symbionts
  17. Reproductive rate of S. carpocapsae with non-native symbionts
  18. Transmission of non-native symbionts by S. carpocapsae
  19. Discussion
  20. Acknowledgments
  21. References

Twenty aposymbiotic IJs were counted and deposited in 1.5-mL Eppendorf tubes containing a piece of filter paper. A last instar larva of G. mellonella was then introduced into each Eppendorf and the aposymbiotic nematodes and the insect host were incubated together at 24 °C during 24 h. For the inoculation with each bacterial strain, 80–160 insects were infested with aposymbiotic IJs. These infested insects were then inoculated with 20 μL (about 2000 cells) of each bacterial culture. Thus, nematode and bacteria were independently introduced into the same insect host. Insects were then re-incubated at 24 °C and their death occurred within the next 24 h. At that time, the insect cadavers were placed on a ‘White trap’ (White, 1927). The nematode offspring (IJs) escaping from the insect cadaver migrated to the sterile water of the ‘White trap’. Two months after infestation, all the IJs produced from single cadavers were collected in sterile water.

Estimating nematode fitness

  1. Top of page
  2. Abstract
  3. Introduction
  4. Natural history of the model system
  5. Materials and methods
  6. Steinernema/Xenorhabdus stock
  7. Aposymbiotic nematodes
  8. Culture and count of bacteria
  9. Combinations of S. carpocapsae with non-native bacteria
  10. Estimating nematode fitness
  11. Transmission of non-native symbionts by S. carpocapsae
  12. Statistical analyses
  13. Phylogenetic analyses
  14. Does the effect of a bacterial strain depend on its position in the phylogenetic tree?
  15. Results
  16. Parasitic success of S. carpocapsae with non-native symbionts
  17. Reproductive rate of S. carpocapsae with non-native symbionts
  18. Transmission of non-native symbionts by S. carpocapsae
  19. Discussion
  20. Acknowledgments
  21. References

We measured two fitness components of S. carpocapsae: (i) its parasitic success and (ii) its reproductive rate. The parasitic success is the percentage of infestations leading to the emergence of offspring. It represents the ability of the nematode to complete its life cycle within the insect with a given bacterium and was estimated as the ratio of the number of ‘White traps’ containing IJs over the total number of ‘White traps’. The reproductive rate is the total number of IJs produced per successful infestation. The entire offspring of each infestation was harvested separately in 50-mL Falcon flasks and preserved at 8 °C. The total number of IJs produced was then counted under a binocular microscope, using 1 mL of the suspension taken from the Falcon flask on a grid drawn on a 6-cm Petri dish.

Transmission of non-native symbionts by S. carpocapsae

  1. Top of page
  2. Abstract
  3. Introduction
  4. Natural history of the model system
  5. Materials and methods
  6. Steinernema/Xenorhabdus stock
  7. Aposymbiotic nematodes
  8. Culture and count of bacteria
  9. Combinations of S. carpocapsae with non-native bacteria
  10. Estimating nematode fitness
  11. Transmission of non-native symbionts by S. carpocapsae
  12. Statistical analyses
  13. Phylogenetic analyses
  14. Does the effect of a bacterial strain depend on its position in the phylogenetic tree?
  15. Results
  16. Parasitic success of S. carpocapsae with non-native symbionts
  17. Reproductive rate of S. carpocapsae with non-native symbionts
  18. Transmission of non-native symbionts by S. carpocapsae
  19. Discussion
  20. Acknowledgments
  21. References

The bacteriological analysis of the intestinal contents of the emerging IJs of S. carpocapsae allowed us to assess the transmission of non-native Xenorhabdus spp. and Photorhabdus spp. by S. carpocapsae. A total of 500 IJs of each set resulting from non-native combinations were counted and immersed separately for 10 min in a sterile Ringer solution with 10% (w/v) sodium hypochloride, to induce the exsheathing of the L2 cuticle kept by the L3 (IJs) during starvation, and to avoid any external contamination from the L3 cuticle. Successful completion of this shedding was checked under a stereomicroscope. The set of 500 IJs were then rinsed twice in sterile Ringer and crushed in 1 mL of sterile phosphate-buffered saline without Mg2+ and Ca2+ salts (8 g NaCl, 0.2 g KCl, 1.15 g Na2HPO4, 0.2 g KH2PO4, 1 L H2O sterile). A quantity of 100 μL of the pure and the 1/10 diluted suspensions were streaked onto NBTA medium. The Petri dishes were incubated at 28 °C for 48 h. After incubation, the presence/absence of symbiont colonies was determined for each set of crushed IJs. Phenotypic tests were conducted at 28 °C to verify whether the bacteria belonged to the Xenorhabdus and Photorhabdus genera (Akhurst & Boemare, 1988).

Statistical analyses

  1. Top of page
  2. Abstract
  3. Introduction
  4. Natural history of the model system
  5. Materials and methods
  6. Steinernema/Xenorhabdus stock
  7. Aposymbiotic nematodes
  8. Culture and count of bacteria
  9. Combinations of S. carpocapsae with non-native bacteria
  10. Estimating nematode fitness
  11. Transmission of non-native symbionts by S. carpocapsae
  12. Statistical analyses
  13. Phylogenetic analyses
  14. Does the effect of a bacterial strain depend on its position in the phylogenetic tree?
  15. Results
  16. Parasitic success of S. carpocapsae with non-native symbionts
  17. Reproductive rate of S. carpocapsae with non-native symbionts
  18. Transmission of non-native symbionts by S. carpocapsae
  19. Discussion
  20. Acknowledgments
  21. References

Parasitic successes (percentages) were compared across groups using general linear model with binomial error (logistic analysis in JUMP JMP, 1989, ver. 4.04; SAS Institute, Cary, NC, USA). This test was used to define groups of non-native bacterial strains leading to a similar parasitic success of S. carpocapsae. The reproductive rate data were found to deviate strongly from normality (Bartlett's test: inline image, P < 0.001), so we performed the nonparametric Kruskal–Wallis test (KW) to check for significant differences between the reproductive rates of S. carpocapsae with native or non-native symbionts. When the KW tests showed significant heterogeneity between samples, the Multiple Comparisons Test of Noether (Scherrer, 1984) was applied to determine which pairs of bacterial strains had a significantly different effect on the reproduction of S. carpocapsae.

Phylogenetic analyses

  1. Top of page
  2. Abstract
  3. Introduction
  4. Natural history of the model system
  5. Materials and methods
  6. Steinernema/Xenorhabdus stock
  7. Aposymbiotic nematodes
  8. Culture and count of bacteria
  9. Combinations of S. carpocapsae with non-native bacteria
  10. Estimating nematode fitness
  11. Transmission of non-native symbionts by S. carpocapsae
  12. Statistical analyses
  13. Phylogenetic analyses
  14. Does the effect of a bacterial strain depend on its position in the phylogenetic tree?
  15. Results
  16. Parasitic success of S. carpocapsae with non-native symbionts
  17. Reproductive rate of S. carpocapsae with non-native symbionts
  18. Transmission of non-native symbionts by S. carpocapsae
  19. Discussion
  20. Acknowledgments
  21. References

The 16S rDNA of X. nematophila (strain F1, strain PL31), X. poinarii (strain SK72), X. bovienii (strain FR10), Xenorhabdus sp. (strain UY61) and Xenorhabdus sp. (strain USTX62) were amplified by PCR under conditions previously described (Brunel et al., 1997). The prokaryote-specific primers used were the following: 5′-GAAGAGTTTGATCATGGCTC-3′ and 5′-AAGGAGGTGATCCAGCCGCA-3′. The amplified products were purified from agarose gels with a QIAamp DNA mini kit (QIAGEN). The sequencing was performed by Act GeneEuro Sequence Genes Services (Paris, France) on an ABI377 sequencer using an ABI PRISM Dye Terminator Cycle Sequencing Ready Reaction kit with Ampli Taq DNA Polymerase. The phylogenetic analyses were performed by Phylo_win (Galtier et al., 1996). The trees were reconstructed with bio-Neighbour-joining (bioNj), Maximum Parsimony (MP) and Maximum Likelihood (ML) methods. Bootstrap values were calculated for bioNj, MP, ML with 500 replicates. The genetic distances between a non-native and native symbiont were calculated by adding all branch lengths obtained from the ML analysis that separated the two taxa.

Does the effect of a bacterial strain depend on its position in the phylogenetic tree?

  1. Top of page
  2. Abstract
  3. Introduction
  4. Natural history of the model system
  5. Materials and methods
  6. Steinernema/Xenorhabdus stock
  7. Aposymbiotic nematodes
  8. Culture and count of bacteria
  9. Combinations of S. carpocapsae with non-native bacteria
  10. Estimating nematode fitness
  11. Transmission of non-native symbionts by S. carpocapsae
  12. Statistical analyses
  13. Phylogenetic analyses
  14. Does the effect of a bacterial strain depend on its position in the phylogenetic tree?
  15. Results
  16. Parasitic success of S. carpocapsae with non-native symbionts
  17. Reproductive rate of S. carpocapsae with non-native symbionts
  18. Transmission of non-native symbionts by S. carpocapsae
  19. Discussion
  20. Acknowledgments
  21. References

In order to test the potential effect of phylogenetic distance, we correlated the effect of each bacterial strain on the nematode's reproductive rate, with the genetic distances between this strain and the native one. The effect of a specific bacterial strain on nematode reproduction was estimated as the reproductive rate (number of IJs) that it induced for the nematodes, divided by the median number of IJs produced by all aposymbiotic nematodes (i.e. 15 565, see Sicard et al., 2003). These correlation analyses raised several statistical issues. First, both the relative effects and the distances were estimated on pairs of bacterial strains. These pairs were not independent from each other as they all had the native strain in common. Secondly, for each strain, we performed several independent measures of the nematode's reproductive rate, but we had only a single estimate of the distances. To solve both issues, we first estimated the slope of the regression line between the distances and the relative reproductive success of the nematode. We then performed 100 000 permutations where we randomly assigned a phylogenetic distance to each strain and re-estimated the slope of the regression line. From these permutations, we were able to estimate the distribution of this slope under the null hypothesis that the effect of a strain on the nematode does not depend on its position in the phylogenetic tree.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Natural history of the model system
  5. Materials and methods
  6. Steinernema/Xenorhabdus stock
  7. Aposymbiotic nematodes
  8. Culture and count of bacteria
  9. Combinations of S. carpocapsae with non-native bacteria
  10. Estimating nematode fitness
  11. Transmission of non-native symbionts by S. carpocapsae
  12. Statistical analyses
  13. Phylogenetic analyses
  14. Does the effect of a bacterial strain depend on its position in the phylogenetic tree?
  15. Results
  16. Parasitic success of S. carpocapsae with non-native symbionts
  17. Reproductive rate of S. carpocapsae with non-native symbionts
  18. Transmission of non-native symbionts by S. carpocapsae
  19. Discussion
  20. Acknowledgments
  21. References

The sequences performed for this study were compared with the other 16S rDNA sequences of Xenorhabdus and Photorhabdus strains already available in the genbank database (Suzuki et al., 1996; Fischer-Le Saux et al., 1997; Szallas et al., 1997). The best outgroups for an intrageneric phylogeny of Xenorhabdus were found in the genus Photorhabdus (Suzuki et al., 1996). The topology of the tree was the same for the three different methods used and was supported by high bootstrap values in all cases (Fig. 2).

image

Figure 2. Phylogeny of the symbiotic bacterial strains obtained with maximum likelihood analysis. Boostrap values are indicated on the branches of the tree. The assessment of S. carpocapsae fitness is indicated for each symbiont combination by parasitic success (PS), the median of the distribution of the reproductive rate (RR), and the ability of the nematode to transmit the bacterial strain (T).

Download figure to PowerPoint

Parasitic success of S. carpocapsae with non-native symbionts

  1. Top of page
  2. Abstract
  3. Introduction
  4. Natural history of the model system
  5. Materials and methods
  6. Steinernema/Xenorhabdus stock
  7. Aposymbiotic nematodes
  8. Culture and count of bacteria
  9. Combinations of S. carpocapsae with non-native bacteria
  10. Estimating nematode fitness
  11. Transmission of non-native symbionts by S. carpocapsae
  12. Statistical analyses
  13. Phylogenetic analyses
  14. Does the effect of a bacterial strain depend on its position in the phylogenetic tree?
  15. Results
  16. Parasitic success of S. carpocapsae with non-native symbionts
  17. Reproductive rate of S. carpocapsae with non-native symbionts
  18. Transmission of non-native symbionts by S. carpocapsae
  19. Discussion
  20. Acknowledgments
  21. References

The infestations with aposymbiotic nematodes followed by injections of the native symbiont (X. nematophila, strain F1) resulted in a parasitic success (i.e. 79%) similar to that of infestations using the naturally symbiotic nematodes (i.e. 80% see Sicard et al., 2003) (inline image = 0.04, n.s.). The parasitic success of S. carpocapsae was dependent on the non-native symbionts injected in the insect as the global comparison of all combinations performed showed significant heterogeneity (inline image = 123.61, P < 0.001). The injections of symbionts conspecific to the native one (AN6 and PL31) led to significantly lower emergences of juvenile nematode (parasitic success) than the infestations with naturally symbiotic nematodes (inline image = 5.5, P < 0.05). The injections of congeneric symbionts resulted in important differences in the parasitic success of S. carpocapsae. The two X. poinarii strains (G6 and SK72) did not have the same effect. Injections with strain G6 led to 82.5% parasitic success, a percentage similar to the one obtained after injections of the native strain and to infestations with natural symbiotic nematodes (inline image = 0.37, n.s.). Conversely, injections of strain SK72 led to only 40% parasitic success which was similar to the results obtained with aposymbiotic nematodes (i.e. 46% see Sicard et al., 2003) (inline image = 0.66, n.s.). The injections of X. beddingii significantly reduced the parasitic success of the nematode compared with that of aposymbiotic nematodes (inline image = 24.10, P < 0.001). Finally, the injections of X. bovienii (strain FR10), Xenorhabdus sp. (strain USTX62 and UY61), Photorhabdus luminescens (strain TT01) and P. temperata (strain Xlnach) completely prevented reproduction of the nematode within the insect.

Reproductive rate of S. carpocapsae with non-native symbionts

  1. Top of page
  2. Abstract
  3. Introduction
  4. Natural history of the model system
  5. Materials and methods
  6. Steinernema/Xenorhabdus stock
  7. Aposymbiotic nematodes
  8. Culture and count of bacteria
  9. Combinations of S. carpocapsae with non-native bacteria
  10. Estimating nematode fitness
  11. Transmission of non-native symbionts by S. carpocapsae
  12. Statistical analyses
  13. Phylogenetic analyses
  14. Does the effect of a bacterial strain depend on its position in the phylogenetic tree?
  15. Results
  16. Parasitic success of S. carpocapsae with non-native symbionts
  17. Reproductive rate of S. carpocapsae with non-native symbionts
  18. Transmission of non-native symbionts by S. carpocapsae
  19. Discussion
  20. Acknowledgments
  21. References

The reproductive rate of S. carpocapsae was strongly affected by the identity of the non-native bacteria injected into the insect (Kruskal–Wallis Hcorr = 141.942, P < 0.001). The results of the Noether Multiple Comparisons test for each combination between S. carpocapsae and symbiont are given in Table 2. The reproductive rate of S. carpocapsae when its native symbiont was injected into the insect did not significantly differ from the one of naturally symbiotic nematodes (i.e. median = 109 037 IJs, see Sicard et al., 2003). Thus, this result also allows us to reject that any bias was induced by the experimental combination. The injections of the conspecific strain AN6 also resulted in a similar reproductive rate as the one of the naturally symbiotic nematodes. On the contrary, the injections with the conspecific strain PL31 resulted in a lower reproduction rate than infestations with the naturally symbiotic nematodes, but a significantly higher one than simple infestations with aposymbiotic nematodes (i.e. median = 15 565 IJs, see Sicard et al., 2003). The injections of X. poinarii (strain G6, strain SK72) yielded reproductive rates similar to that of aposymbiotic nematodes, and thereby led to a lower reproductive rate for S. carpocapsae compared with the one of naturally symbiotic nematodes. The nonsignificant results of the experiments involving X. beddingii are most likely due to the small number of data available which may be related to the low parasitic success of the nematode when this bacterium was injected into the insect. Note that there was no nematode offspring when Xenorhabdus sp. (strain USTX62 and UY61), X. bovienii (FR10), Photorhabdus luminescens (strain TT01) and P. temperata (strain Xlnach) were injected into the insect.

Table 2.  Comparison of the distribution of reproductive rates of S. carpocapsae in experimental combinations with that of aposymbiotic (first column) and naturally symbiotic worms of the same species (second column). The distribution of reproductive rates of aposymbiotic and naturally symbiotic S. carpocapsae that we use as references here were estimated in a previous study (Sicard et al., 2003). For each case, the median number of the reproductive rate distribution is given between brackets.
Strains of bacteriaAposymbiotic nematodes (15 565)Naturally symbiotic nematodes (109 037)
ZcalcPZcalcP
F1 (94 050)5.9<0.0012.2n.s.
PL31 (38 000)4.05<0.015.2<0.001
AN6 (72 300)4.8<0.0012.3n.s.
SK72 (17 200)0.7n.s.6.8<0.001
G6 (12 300)0.7n.s.9<0.001
Q58 (33 450)0.8n.s.3n.s

The effect of a bacterial strain on the reproductive rate of the nematode sharply decreased when its genetic distance to the native strain increased (b = −75, P < 0.001) (Fig. 3). The negative correlation was still significant when the analyses were performed without the Photorhabdus strains (b = −71.6, P < 0.001).

image

Figure 3. Nematode reproductive rate for each combination of bacterial strain, as a function of the genetic distance between non-native and native bacteria. The reference number of the bacterial strain used in each combination is noted below the corresponding distribution (see Table 1). The regressions indicate a negative correlation between the effect of each non-native strain on nematode fitness and its genetic distance to the native symbiont. The reproductive rates of S. carpocapsae with its own and non-native symbionts were divided by the median of the reproductive rate of aposymbiotic nematodes (i.e. 15 565 IJs) to evaluate the gain or the loss of fitness of the nematode due to a given bacterium. The broken line corresponds to a ratio equal to one. A ratio higher than 1 indicates a positive effect of the bacteria on nematode fitness, a ratio lower than 1, a negative effect. The slope of the regression line is −75. The permutation test indicates that this slope is significantly lower than zero (P < 0.001).

Download figure to PowerPoint

Transmission of non-native symbionts by S. carpocapsae

  1. Top of page
  2. Abstract
  3. Introduction
  4. Natural history of the model system
  5. Materials and methods
  6. Steinernema/Xenorhabdus stock
  7. Aposymbiotic nematodes
  8. Culture and count of bacteria
  9. Combinations of S. carpocapsae with non-native bacteria
  10. Estimating nematode fitness
  11. Transmission of non-native symbionts by S. carpocapsae
  12. Statistical analyses
  13. Phylogenetic analyses
  14. Does the effect of a bacterial strain depend on its position in the phylogenetic tree?
  15. Results
  16. Parasitic success of S. carpocapsae with non-native symbionts
  17. Reproductive rate of S. carpocapsae with non-native symbionts
  18. Transmission of non-native symbionts by S. carpocapsae
  19. Discussion
  20. Acknowledgments
  21. References

Only the symbionts belonging to the X. nematophila species were transmitted by S. carpocapsae (Fig. 2). This was true for the native strain (F1) and for the non-native conspecific strains (AN6 and PL31). The other bacterial species (X. poinarii, X. beddingii) were not transmitted although they permitted nematode reproduction (Fig. 2). The bacterial strains which were lethal for the nematode during the parasitism of the insect were obviously not transmitted as no nematode offspring was produced.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Natural history of the model system
  5. Materials and methods
  6. Steinernema/Xenorhabdus stock
  7. Aposymbiotic nematodes
  8. Culture and count of bacteria
  9. Combinations of S. carpocapsae with non-native bacteria
  10. Estimating nematode fitness
  11. Transmission of non-native symbionts by S. carpocapsae
  12. Statistical analyses
  13. Phylogenetic analyses
  14. Does the effect of a bacterial strain depend on its position in the phylogenetic tree?
  15. Results
  16. Parasitic success of S. carpocapsae with non-native symbionts
  17. Reproductive rate of S. carpocapsae with non-native symbionts
  18. Transmission of non-native symbionts by S. carpocapsae
  19. Discussion
  20. Acknowledgments
  21. References

The level of specialization between partners of mutualistic symbioses provides information on the common evolutionary history between interacting species. Several empirical studies have indicated that partners of mutualistic symbioses were less specialized than parasitic ones (Thompson, 1982; Law & Koptur, 1986; Hoeksema & Bruna, 2000). However, our study performed on the mutualistic associations between a bacterial symbiont and a host, showed that the interaction ‘functions best’ when the host is associated to its native symbiont. More precisely, in our model system, we classified three main classes of bacteria according to their effect on the fitness of the studied nematode host, S. carpocapsae (Fig. 2): (i) mutualistic bacteria that increased both the infestation success and offspring production of the nematode, (ii) ‘neutral’ bacteria that have a slightly beneficial or no impact on the fitness of the nematode, and (iii) pathogenic or even lethal bacteria that reduce or even suppress offspring production of the nematode.

The classification of a bacterial strain into one of these three classes seems to depend mostly on its evolutionary history, as suggested by the phylogenetic tree in Fig. 2. A striking illustration of this is that the bacterial strains in the first class are conspecific of the native symbiont. However, even within this group, the native symbiont is more beneficial than the other strains. These results strongly suggest a high degree of specialization between the native bacterial strain and its nematode host. Neutral bacteria (the second class) belong to a different bacterium species (X. poinarii). Finally, pathogenic and lethal bacteria (the third class) mostly belong to very different Xenorhabdus species or to an even less related sister genus Photorhabdus. This phylogenetic trend is strongly supported by the significant correlation between the relative fitness effect of each bacterial strain and its genetic distance to the native symbiont (see Fig. 3).

The specificity of re-association and transmission in this system is highly congruent with the classification presented above: the only bacteria that are transmitted by the nematode are the mutualistic symbionts. These results also show that the specificity of transmission is not absolute, as intraspecific exchanges of genetically different symbionts can occur when two individuals of S. carpocapsae carrying different strains of X. nematophila infect the same insect. Under such circumstances, the bacterial strains would compete for dispersal by the same nematode (Frank, 1996). The symbionts that are able to eliminate the other strains would thus benefit from all nematode individuals as vectors of dispersion from the insect host. Any trait allowing a bacterium to exclude other strains, such as the production of antibiotic against competing strains, should therefore be strongly favoured by natural selection (Thaler et al., 1995, 1997). Competition can also occur between S. carpocapsae and congeneric nematode species. In this case, our results indicate that interspecific exchanges of symbionts seem unlikely. The nematode cannot re-associate with bacteria which are not closely related to its native symbiont.

Frank (1996, 2003) suggested that hosts should avoid mixing of symbiont lineages. The experimental approach of Bot et al. (2001) supports this view. We may also interpret the high degree of specialization in bacterial transport, through finely tuned recognition mechanisms, as being part of a similar strategy. This strategy effectively yields a close to vertical pattern of transmission. Each bacterial strain should therefore strive to give a competitive advantage to its nematode vector in order to increase its own dispersal. It could for instance produce toxic compounds that specifically eliminate competitors of its nematode. It is well known that Xenorhabdus produces virulence factors to kill insects (Bowen et al., 1998; Dowds & Peters, 2002), and it can be hypothesized that some of these factors may have an indirect pathogenic effect on nematodes. The factors of pathogenicity developed against the insect would be also efficient against non-native nematodes. Nevertheless, some toxins could also be specifically produced to kill ‘foreign’ nematodes. Under this hypothesis, each nematode strain would be resistant to the molecules produced by its native symbiont and by other closely related bacterial strains. A nematode, which has co-evolved with a bacterial strain, should be selected to be resistant to virulence factors produced by its native symbiont, but not to those developed by phylogenetically distant ones. This hypothesis could explain why phylogenetically distant bacterial strains have pathogenic effects when they co-infect an insect with S. carpocapsae.

Whatever the mechanism involved, the fundamental result of this study is that bacterial strains that are mutualistic for their associated nematode species (Sicard et al., 2003) can be pathogenic or even lethal for other species of nematodes within the same genus. As we suggest above, the evolution of this system will be strongly affected by the frequency of competition between the Steinernema–Xenorhabdus associations within the same insect host. In other words, in this system as well as in other mutualisms in which symbionts are not strictly vertically transmitted, the maintenance of mutualism is determined by interactions between symbionts and hosts at multiple levels. Thus, the association between Steinernema and Xenorhabdus is a promising experimental system to investigate the evolution of mutualistic interactions at the level of entire communities of interactants of varying relatedness.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Natural history of the model system
  5. Materials and methods
  6. Steinernema/Xenorhabdus stock
  7. Aposymbiotic nematodes
  8. Culture and count of bacteria
  9. Combinations of S. carpocapsae with non-native bacteria
  10. Estimating nematode fitness
  11. Transmission of non-native symbionts by S. carpocapsae
  12. Statistical analyses
  13. Phylogenetic analyses
  14. Does the effect of a bacterial strain depend on its position in the phylogenetic tree?
  15. Results
  16. Parasitic success of S. carpocapsae with non-native symbionts
  17. Reproductive rate of S. carpocapsae with non-native symbionts
  18. Transmission of non-native symbionts by S. carpocapsae
  19. Discussion
  20. Acknowledgments
  21. References
  • Aguillera, M.M. & Smart, G. 1993. Development, reproduction, and pathogenicity of Steinernema scapterisci in monoxenic culture with different species of bacteria. J. Invertebr. Pathol. 62: 289294.
  • Akhurst, R. 1983. Neoplectana species: specificity of association with bacteria of the genus Xenorhabdus. Exp. Parasitol. 55: 258263.
  • Akhurst, R. & Boemare, N. 1988. A numerical taxonomic study of the genus Xenorhabdus (Enterobacteriaceae) and proposed elevation of the subspecies of X. nematophilus to species. J. Gen. Microbiol. 134: 18351845.
  • Bergstrom, C.T. & Lachmann, M. 2003. The red king effect: when the slowest runner wins the coevolutionary race. Proc. Natl. Acad. Sci. U.S.A. 100: 593598.
  • Bird, A.F. & Akhurst, R. 1983. The nature of the intestinal vesicle in nematodes of the family Steinernematidae. Int. J. Parasitol. 13: 599606.
  • Blaxter, M., De Ley, P., Garey, J., Liu, L., Scheldeman, P., Vierstraete, A., Vanfleteren, J., Mackey, L., Dorris, M., Frisse, L., Vida, J. & Thomas, W. 1998. A molecular evolutionary framework for the phylum Nematoda. Nature 392: 7175.
  • Bot, A.N.M., Rehner, S.A. & Boomsma, J.J. 2001. Partial incompatibility between ants and symbiotic fungi in two sympatric species of Acromyrmex leaf-cutting ants. Evolution 55: 19801991.
  • Bowen, D., Rocheleau, T.A., Blackburn, M., Andreev, O., Golubeva, E., Bhartia, R. & Ffrench-Constant, R.H. 1998. Insecticidal toxins from the bacterium Photorhabdus luminescens. Science 280: 21292132.
  • Brunel, B., Givaudan, A., Lanois, A., Akhurst, R. & Boemare, N. 1997. Fast and accurate identification of Xenorhabdus and Photorhabdus species by restriction analysis of PCR-amplified 16S rRNA genes. Appl. Environ. Microbiol. 63: 574580.
  • Dowds, B.C.A. & Peters, A. 2002. Virulence mechanisms. In: Entomopathogenic Nematology (R.Gaugler, ed.), pp. 7998. CABI Publishing, Wallingford, UK.
  • Fischer-Le Saux, M., Mauléon, H., Constant, P., Brunel, B. & Boemare, N. 1997. PCR-ribotyping of Xenorhabdus and Photorhabdus isolates from the Caribbean region to the taxonomy and geographic distribution of their nematode hosts. Appl. Environ. Microbiol. 64: 42464254.
  • Frank, S. 1996. Host-symbiont conflict over mixing of symbiotic lineages. Proc. R. Soc. Lond. B 263: 339344.
  • Frank, S. 2003. Perspective: repression of competition and the evolution of cooperation. Evolution 57: 693705.
  • Galtier, N., Gouy, M. & Gautier, C. 1996. Seaview and phylowin: two graphic tools for sequence alignment and molecular phylogeny. Comp. Appl. Biosci. 12: 543548.
  • Grewal, P.S., Matsuura, M. & Converse, V. 1997. Mechanisms of specificity of association between the nematode Steinernema scapterisci and its symbiotic bacterium. Parasitology 114: 483488.
  • Herre, E.A., Knowlton, N., Meuller, U.G. & Rehner, S.A. 1999. The evolution of mutualisms: exploring the paths between conflict and cooperation. Trends Ecol. Evol. 14: 4952.
  • Hoeksema, J. & Bruna, E. 2000. Pursuing the big questions about interspecific mutualism: a review of theoretical approaches. Oecologia 125: 321330.
  • Laumond, C., Mauléon, H. & Kermarrec, A. 1979. Données nouvelles sur le spectre d'hôtes et le parasitisme du nématode entomophage Neoplectana carpocapsae. Entomophaga 24: 1327.
  • Law, R. & Koptur, S. 1986. On the evolution of non specific mutualism. Biol. J. Linn. Soc. 27: 251267.
  • Leigh, E.G. 1999. Levels of selection, potential conflicts, and their resolution – the role of the ‘common good’. In: Levels of Selection in Evolution (L.Keller, ed.), pp. 1530. Princeton University Press, Princeton, NJ.
  • Nishiguchi, M.K., Ruby, E. & McFall-Ngai, M. 1998. Competitive dominance among strains of luminous bacteria provides an unusual form of evidence of parallel evolution in sepiolid squid-Vibrio symbioses. Appl. Environ. Microbiol. 64: 32093213.
  • Nyholm, S.V., Stabb, E.V., Ruby, E.G. & McFall-Ngai, M. 2000. Establishment of an animal-bacterial association: recruiting symbiotic vibrios from the environment. Proc. Natl. Acad. Sci. USA. 97: 1023110235.
  • Poinar, G.O. 1979. Nematodes for Biological Control of Insects. CRC Press, Boca Raton, FL.
  • Poinar, G.O. 1993. Origins and phylogenetics relationship of the entomophilic rhabditids, Heterorhabditis and Steinernema. Fundam. Appl. Nematol. 16: 333338.
  • Poinar, G.O. & Himsworth, P.T. 1967. Neoplectana Parasitism of larvae of the greater wax moth Galleria mellonella. J. Invertebr. Pathol. 9: 241246.
  • Poinar, G.O. & Thomas, G.M. 1966. Significance of Achromobacter nematophilus Poinar and Thomas (Achromobacteriaceae: Eubacteriales) in the development of the nematode, DD-136 (Neoplectana sp. Steinernematidae). Parasitology 56: 385390.
  • Scherrer, B. 1984. Biostatistique. Gaetan Morin, Paris.
  • Sicard, M., Le Brun, N., Pagès, S., Godelle, B., Boemare, N. & Moulia, C. 2003. Effect of native Xenorhabdus on the fitness of their Steinernema hosts: contrasting types of interaction. Parasitol. Res. 91: 520524.
  • Spoerke, J.M., Wilkinson, H.H. & Parker, M.A. 1996. Nonrandom genotypic associations in Legume-Bradyrhizobium mutualism. Evolution 50: 146154.
  • Suzuki, T., Yabusaki, H. & Nishiguchi, M.K. 1996. Phylogenetic relationship of entomopathogenic nematophilic bacteria: Xenorhabdus spp. and Photorhabdus spp. J. Basic Microbiol. 5: 351354.
  • Szallas, E., Koch, C., Fodor, A., Burghardt, J., Buss, O., Szentirmai, A., Nealson, K.H. & Stackebrandt, E. 1997. Phylogenetic evidence for the taxonomic heterogeneity of Photorhabdus luminescens. Int. J. Syst. Bacteriol. 47: 402407.
  • Thaler, J.-O., Baghdiguian, S. & Boemare, N. 1995. Purification and characterization of Xenorhabdicin a phage tail-like bacteriocin, from the lysogenic strain F1 of Xenorhabdus nematophilus. Appl. Environ. Microbiol. 61: 20492052.
  • Thaler, J.-O., Boyer-Giglio, M.-H. & Boemare, N. 1997. New antimicrobial barriers produced by Xenorhabdus spp. and Photorhabdus spp. to secure the monoxenic development of entomopathogenic nematodes. Symbiosis 22: 205215.
  • Thompson, J.N. 1982. Interaction and Co-evolution. Wiley-Interscience, New York.
  • White, G. 1927. A method for obtaining infective nematode larvae from culture. Science 66: 302303.
  • Wilkinson, D.M. 2001. Horizontally acquired mutualisms, an unsolved problem in ecology? Oïkos 92: 377384.
  • Wilkinson, H.H., Spoerke, J.M. & Parker, M.A. 1996. Divergence in symbiotic compatibility in a legume-Bradyrhizobium mutualism. Evolution 50: 14701477.