Background and Objectives
Interactions between plants and beneficial soil organisms (e.g. rhizobial bacteria, mycorrhizal fungi) are models for investigating the ecological impacts of such associations in plant communities, and the evolution and maintenance of variation in mutualisms (e.g. host specificity and the level of benefits provided). With relatively few exceptions, variation in symbiotic effectiveness across wild host species is largely unexplored.
We evaluated these associations using representatives of several legume genera which commonly co-occur in natural ecosystems in south-eastern Australia and an extensive set of rhizobial strains isolated from these hosts. These strains had been previously assigned to specific phylotypes on the basis of molecular analyses. In the first of two inoculation experiments, the growth responses of each host species was evaluated with rhizobial strains isolated from that species. The second experiment assessed performance across genera and the extent of host specificity using a subset of these strains.
While host growth responses to their own (sympatric) isolates varied considerably, rhizobial phylotype was a significant predictor of symbiotic performance, indicating that bacterial species designations on the basis of molecular markers have ecological importance. Hosts responded in qualitatively different ways to sympatric and allopatric strains of rhizobia, ranging from species with a clear preference for their own strains, to those that were broad generalists, through to species that grew significantly better with allopatric strains.
Theory has focused on trade-offs between the provision of benefits and symbiont competitive ability that might explain the persistence of less beneficial strains. However, differences in performance among co-occurring host species could also drive such patterns. Our results thus highlight the likely importance of plant community structure in maintaining variation in symbiotic effectiveness.
Citation: Thrall PH, Laine A-L, Broadhurst LM, Bagnall DJ, Brockwell J (2011) Symbiotic Effectiveness of Rhizobial Mutualists Varies in Interactions with Native Australian Legume Genera. PLoS ONE 6(8): e23545. https://doi.org/10.1371/journal.pone.0023545
Editor: Matthias Rillig, Freie Universität Berlin, Germany
Received: June 10, 2011; Accepted: July 19, 2011; Published: August 26, 2011
Copyright: © 2011 Thrall et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: The work described in this study was funded by the National Action Plan for Salinity and Water Quality (project #202749). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Interactions between plants and symbiotic soil microbes are major determinants of ecosystem productivity and diversity. Plants can receive substantial benefits from root associated symbionts, such as rhizobial bacteria and mycorrhizal fungi, particularly where increased nutrient availability can provide hosts with significant fitness advantages. The presence of effective mutualists enhances the growth and competitive ability of host plants, and in turn can influence successional dynamics , plant productivity , and community restoration –. Plant-associated microbes, including pathogens, are also important regulators of plant community dynamics and structure , . The general importance of plant-soil microbe interactions for community assembly and coevolutionary processes, and the value of utilising these associations in the management and restoration of functioning native ecosystems , ,  is widely recognised. However, considerable empirical and theoretical gaps remain in our ecological and evolutionary understanding of plant-soil symbiont interactions in diverse natural host communities.
Characterisation of the degree to which symbiotic microbes vary in the provision of mutualistic benefits in relation to environmental quality, host species and plant community structure is critical to developing an understanding of their role as agents of productivity and selection in natural populations . It is becoming increasingly clear that plants and microbes interact within a diverse community of potential partners and competitors , within which interactions vary widely in both specificity (i.e. host range) and position along the mutualism-parasitism continuum. Furthermore, variation in host and microbe genetic identity can strongly influence the strength, net fitness effect and even direction of symbiotic interactions , as has been shown for Acacia spp. and associated rhizobia –. Much less information is available regarding interactions across host genera, although some previous work indicates that patterns of host specificity and symbiotic effectiveness are likely to be complex , .
The persistence of ineffective symbionts (‘cheaters’) in mutualistic associations has been a focus of evolutionary theory as well as empirical studies, with a particular emphasis on legume-rhizobial interactions –. For some mutualists such as mycorrhizal fungi, there is evidence for trade-offs between growth promotion and inter-strain competitiveness (i.e. a cost of mutualism) which could mediate ‘main-effect’ differences among fungal strains , . Such trade-offs are also likely to play a role in maintaining variation in legume-rhizobial interactions . However, given the potential for considerable host specificity in such interactions, whether particular rhizobial strains are characterised as ineffective is likely to be at least partly context dependent (i.e. strains perceived as essentially parasitic on one host may well be beneficial on another host), as has been shown for mycorrhizal fungi . Thus, likely determinants of the extent of variation in mutualistic benefits within and among host species include spatial structure in host distribution and community structure, and environmental quality , . Overall, characterisation of these associations in nature is still insufficient to determine the extent to which outcomes are determined by variation among rhizobial strains, host species which may differ in their ability to discriminate among strains, or by their interaction, although clearly genotype×genotype interactions are an important determinant of the level of mutualistic benefits conferred .
Development of a more quantitative understanding of the genetic, ecological and environmental factors that drive the evolution of host range and mutualistic benefit also demands characterisation of the correspondence between phylotypes identified using molecular approaches and their ecological performance. For example, in soil microbial ecology, molecular studies have revealed a myriad of previously cryptic bacterial species – which could represent functionally meaningful diversity should genetically distinct taxa also have correspondingly different physiologies and ecologies. However, examples confirming distinct ecologies of phylogenetically distinct taxa (phylotypes) within a lineage are limited –. Thus, the generality of the correspondence between genetic delineation of microbial species and ecological function is uncertain –, although a recent study  provides strong support for such a correspondence in native legume-rhizobial associations. This is problematic because the link between genes and ecology is a basic assumption underlying the rapidly growing fields of molecular ecology and environmental microbiology .
In the past decade, several studies have examined the phenotypic and genetic diversity, and phylogenetic relationships of rhizobia in native Australian associations –. Bradyrhizobium is the most common genus of root-nodule bacteria reported as nodulating Australian native legumes , but many other nodule-forming bacterial genera are recorded on leguminous genera represented by Australian natives, including Rhizobium, Mesorhizobium, Sinorhizobium (Ensifer), Burkholderia, Devosia, Phyllobacterium , , ,  and Ochrobactrum . To what extent these genera are involved in cross nodulation and meaningful symbiotic relationships in the field with Australian native legumes is unknown. However, recent work suggests that at least some Acacia spp. form significant associations with many of these bacterial genera , , and that these generic associations may vary both in relation to environmental factors such as soil salinity ,  and geographic distance , .
Here we use native representatives of several host genera in the Fabaceae, and an extensive collection of rhizobial strains from these hosts to: a) characterise variation within host genera in symbiotic effectiveness; b) evaluate how these patterns change across co-occurring host genera, both in terms of average effectiveness and host specificity; and c) evaluate the extent to which rhizobial phylotypes (i.e. designated on the basis of molecular markers) have ecological significance with regard to predicting plant growth responses. As noted above, in addition to the potential of such studies to contribute to basic understanding of host-symbiont associations, there is considerable applied value in understanding these relationships better with regard to improving the potential to re-establish functional and diverse native plant communities , , .
Materials and Methods
In this study, we examined the growth performance of eight native legume species from different genera (all in the Fabaceae) in replicated glasshouse inoculation trials using a broad range of rhizobial strains. The host species were: Bossiaea foliosa A. Cunn., Daviesia ulicifolia C.R.P. Andrews, Dillwynia retorta (Wendl.) Druce, Goodia lotifolia Salisb., Hardenbergia violacea (Schneev.) Stearn, Indigofera australis Willd., Oxylobium ellipticum (Vent.) R. Br. and Podolobium ilicifolium (Andrews) Crisp and P.H. Weston. All eight species are common inhabitants of infertile soils in south-eastern Australia.
Symbiotic relationships were evaluated in two experiments in which we measured the extent of nodulation and plant growth responses. In the first experiment, we inoculated each plant host species with rhizobial strains isolated from nodules of that species growing naturally in the field. In the second experiment, we cross-inoculated a subset of these host species with rhizobial strains from the other hosts, as well as with its own strains.
In addition to the field isolates obtained from the target host genera, an additional strain (2836) was used in the first glasshouse experiment. This strain was from the collection of Lafay & Burdon  and it has previously been shown experimentally to fix nitrogen effectively with a diverse range of Australian native species of Acacia. Strain 2836 was originally isolated from Acacia melanoxylon R. Br. and is a component of “Wattlegrow” (a commercial inoculant for native Australian legumes; Becker Underwood, Somersby, New South Wales). In the results that follow, control strain 2836 has been designated as “WG”.
Isolation of rhizobial strains from native hosts.
Nine general localities in south-eastern Australia were chosen (Table 1), with one or more of the eight legume hosts used in this study), occurring at each site . Vigorous adult plants were dug up in order to isolate rhizobial bacteria from their root nodules. Isolates of root-nodule bacteria (Table 1) were extracted from nodules using standard techniques . Pure cultures of 130 isolates were suspended in yeast mannitol broth  and stored under glycerol at −80°C.
Classification of rhizobial phylotypes and generic affiliation.
The 130 field isolates from the nine sites (Table 1) and the Bradyrhizobium control strain (WG) were genetically characterised and assigned at the generic level by Lafay & Burdon  as part of a larger research effort investigating rhizobial diversity on native legumes in southeastern Australia. In that study, field-collected strains were assigned unique phylotype profiles on the basis of RFLP banding patterns from multiple enzymes. Generic affiliations of representative isolates from each phylotype were determined following phylogenetic analyses of SSU rDNA sequences. The majority of these strains were Bradyrhizobium spp (representing 12 different phylotypes) with ten Rhizobium strains (representing a single phylotype) isolated from B. foliosa, Daviesia ulicifolia, H. violacea and I. australis. Summary information on host species, collection sites and generic affiliation of the rhizobial strains used in the present study are given in Table 1; further details of the distribution of phylotypes by host species and geographic location are given in Lafay & Burdon .
Glasshouse Inoculation Studies
For the glasshouse inoculation studies of symbiotic effectiveness and host specificity, seed of each legume species was obtained from the Australian Seed Company, Hazelbrook, New South Wales, or from the CSIRO Australian Tree Seed Centre, Canberra, ACT. In the first glasshouse experiment evaluating within-host performance (Experiment I), seedlings of all 8 host genera were separately inoculated with rhizobial strains that had been isolated from nodules of that host. For each host species, a total of 17 strains were used as individual inocula (Table 1), with the exception of P. ilicifolium (15 strains) and I. australis (13 strains). As noted above, each host species was also separately inoculated with acacia strain WG. Two uninoculated control treatments were also included where plants either received a full nutrient solution including nitrogen or a nutrient solution lacking nitrogen (designated as N+ and N− and respectively). All inoculated plants received the N− nutrient solution.
The second glasshouse experiment was designed to evaluate symbiotic benefits across host genera (Experiment II). The three most effective strains from each host species (except for B. foliosa and O. ellipticum which showed a general lack of responsiveness to rhizobial inoculation in the first trial) were selected on the basis of their N2-fixing performance in Experiment I. These strains were used as inocula in Experiment II (Table 1). In this trial, each host was inoculated with its own three strains (referred to as ‘sympatric’) as well as those from each of the other genera (i.e. ‘allopatric’). As in Experiment I, uninoculated control treatments (N−, N+) were also included.
For both experiments, seeds were pretreated with boiling water for 1 minute, allowed to cool and imbibe for 24 hrs, surface sterilized with ethanol (98%) for 30 seconds then with sodium hypochlorite (5%) for 10 minutes, rinsed 10 times with sterile distilled water, sown into a shallow dish of sterile, moist horticultural vermiculite, and incubated at 25°C until emergence. Newly-emerged seedlings were transplanted (1 per pot) into cylindrical (8 cm×15 cm) polyethylene pots containing a mixed substrate (1∶1 by volume) of steam-sterilised vermiculite and washed river sand. Seedlings were inoculated with a heavy suspension (approx. 1×109 cells per plant) of monocultures of each rhizobial strain, or were left uninoculated as noted above. There were 10 replicates of each host×strain (or control) treatment.
Pots of each species were arranged in randomised blocks in a temperature-controlled glasshouse under standard day/night conditions (16 hrs at 25°C; 8 hrs at 18°C). Plants were watered daily with UV-sterilised tap water or as needed, and weekly with N-free McKnight's  solution. Plants in the N+ control group were given an additional 10 ml of H20 containing 0.05% KNO3 once a week. Plants were harvested approximately 90 days after inoculation. At harvest, the roots were cut off and scored for occurrence and extent of nodule formation : a) nodule number (0, <10, 10–50, >50), b) nodule functionality based on nodule colour and size [ranging from 1 (small non-N2-fixing nodules with white centres) to 5 (large N2-fixing nodules with pink/red centres), and c) nodule distribution [low scores (<2) represented plants with nodules distributed mostly in the root crown and higher scores (3–5) represented plants with nodules more broadly distributed throughout the root system]. All scoring was done by a single observer. Shoots were oven-dried (70°C for 48 hours) and weighed. Shoot dry weight was used as an index of rhizobial strain effectiveness at N2 fixation, given that this was the only source of N available for plant growth in the inoculation treatments.
There was a low level of nodulation of some uninoculated controls by rhizobial contaminants. In almost all instances where it occurred, contaminant nodulation did not lead to any appreciable N2 fixation as evidenced by the small plant dry weights in these treatments; therefore, it was ignored in data treatment. Plants in the N+ uninoculated control groups for B. foliosa, D. ulicifolia and D. retorta performed poorly in Experiment I, thus N− controls were used as a more reliable benchmark of plant-to-plant variation in both glasshouse trials.
We measured the interaction between rhizobial strains and their host plants by calculating a response variable ‘symbiotic response’. For this purpose the dry weight of the host plants inoculated with different rhizobial strains was divided by the average dry weight of the uninoculated N− control plants for that species. Hence, a value of one means that the host did not gain or lose anything from the rhizobial interaction relative to the N− control treatment. Symbiotic response values <1 indicate that the rhizobial strain has a negative effect on its hosts compared to the null situation, and values >1 indicate a positive association between the host plant and a given rhizobial strain. For all host plant×rhizobia interactions we also examined nodule formation. This response variable (nodule presence/absence) has a binomial probability distribution and a logit link function. For all analyses, symbiotic response was confirmed to be normally distributed from the normal quantile plots following a log transformation. All analyses were done using SAS 9.1 . Only significant interactions were included in the final models.
Experiment I: within-host variation.
We first analyzed interactions between the different rhizobial isolates and their host plants using the entire dataset. We used a generalized linear model  for analysing whether isolates formed nodules with their hosts or not, and an ANOVA to analyze the symbiotic response (as described above). Host plant and isolate, nested under host plant, were the explanatory variables in the models. We then compared symbiotic response and nodule formation of strains originating from a given host to the generally effective strain WG on that same host as a ‘standardised’ measure of host response across the species used in our study. In this model WG vs. other strains was the fixed categorical explanatory variable nested under host plant. Strains, nested under host plant, were defined as random variables in the models.
We then analyzed whether phylotypes that were identified on the basis of earlier RFLP analyses  also represented biologically different functional units in their interaction with their host plant. We first asked whether rhizobial phylotypes colonizing the same host species differed in their interaction with that host. For this analysis we only included host genera (Bossiaea, Daviesia, Goodia, Hardenbergia, Podolobium) from which several rhizobial phylotypes had been identified. We used generalized linear mixed models (GLMMs) to analyze nodule formation and symbiotic response with host genus as a fixed factor in the model.
Phylotype, also a fixed factor, was nested within host genus, because different phylotypes were identified from different hosts. To control for variation among rhizobial strains, for both models we defined strains as random variables, nested under their phylotype and host genus, respectively. To test whether phylotypes performed differently across host species, we then analyzed a dataset of three host species (B. foliosa, D. ulicifolia and G. lotifolia) and four phylotypes that were isolated from each of them (A, F, I and P; Table 2). The models for analyzing patterns of nodule formation and symbiotic response were identical to the first phylotype analyses, except that here the effect of phylotype could be estimated across different host species and, hence, it was not nested in the analyses.
Experiment II: among-genera interactions.
The majority of the subset of rhizobial isolates chosen for the across-genera inoculation experiment were grouped within a single phylotype (Table 1). Analyses therefore focused on host origin of rhizobial strains as a predictor of plant growth. In particular, we tested whether host plant response to inoculation with their own sympatric rhizobial strains differed from their performance with strains isolated from other host species, measured as nodule formation and symbiotic response. In the first analysis the explanatory fixed variables in the GLMM models were host plant, the origin of the rhizobial strain, and their interaction. In the second model host plant and origin of the rhizobial strain were again designated as fixed explanatory variables, and we included a third explanatory variable for whether the interaction was sympatric or allopatric (as defined above). Rhizobial strain, nested within host origin, was defined as a random variable in all models.
Experiment I: within-host species inoculations
There was considerable variation in host species responses, both in terms of nodulation as well as in growth responses (P<0.0001 for both; Table 2 and Fig. 1). For example, all but one plant (out of 130) of I. australis had nodules, while 70% of O. ellipticum plants were un-nodulated. For all the other genera, there were at least some host×strain inoculation combinations that did not result in nodules being formed (overall average nodulation across hosts = 75%; P<0.0001; Table 2). The broadly effective acacia strain (WG) nodulated 100% of individual plants for most genera; a notable exception was O. ellipticum where nodulation with this strain was only successful for 55% of the plants inoculated. This was consistent with the generally poor nodulation observed across the set of strains isolated from that host.
The solid line at 1 on the y-axis indicates the growth response level where the host did not gain or lose anything from the rhizobial interaction relative to the N− control treatment. Symbiotic response values <1 indicate a negative response, and values >1 indicate a positive effect of inoculation. The dashed line is the average symbiotic response to all rhizobial strains. Error bars are based on standard errors of means (if not visible, they are smaller than the symbols). BOS = Bossiaea foliosa, DAV = Daviesia ulicifolia, DIL = Dillwynia retorta, GOO = Goodia lotifolia, HAR = Hardenbergia violacea, IND = Indigofera australis, OXY = Oxylobium ellipticum, and POD = Podolobium ilicifolium. Note the different y-axis scales between figure panels.
When the outcome of the interaction was measured as host dry weight, host species differed significantly in their overall level of symbiotic response (P<0.0001; Table 2). Some host species clearly gained little additional benefit from inoculation (e.g. B. foliosa; Fig. 1) while others demonstrated substantial increases in dry weight with rhizobial partners relative to the uninoculated controls (e.g. I. australis; Fig. 1). In addition to the overall differences between host species in their response to inoculation per se, there was also significant variation among individual rhizobial strains with regard to the benefits conferred on their host plants (P<0.0001; Table 2 and Fig. 1), although the degree of variation in performance among strains differed substantially for different host species. With regard to control strain WG, nodule formation did not differ between this strain and the others (P = 0.9985). However, plant growth performance was consistently as good as or better with this strain than the average performance of sympatric strains (Fig. 1). This difference was not statistically significant across all host species (P = 0.2025), but contrast tests for individual hosts revealed that for P. ulicifolium, although WG only nodulated 56% of the plants, it was significantly more effective than the sympatric strains (P = 0.0165).
When rhizobial strains were classified into phylotypes based on RFLP banding patterns, the groups did not differ in whether they nodulated their host or not (average nodulation across phylotypes = 85%; Table 2). However, there were clear differences among phylotypes with regard to the growth responses elicited in their host plants (P = 0.0051; Table 2 and Fig. 2). Several rhizobial phylotypes were represented among the isolates originating from three of the host species (B. foliosa, D. ulicifolia and G. lotifolia). For this subset of the data, we were able to evaluate the extent to which phylotype performance varied across hosts. The results showed that nodule formation was only affected by the host species (P = 0.0207; Table 2). However, symbiotic response (host growth) varied both among the three host species and among the four phylotypes, and there were also strong phylotype×host species interactions (P = 0.0078; Table 2 and Fig. 3). For example, rhizobial phylotype F had a negative effect on the growth response of D. ulicifolia (relative to the uninoculated N− control), while the highest positive growth response was measured in the interactions between phylotype F and G. lotifolia (Fig. 3). This host could be characterised as a generalist as it responded well to inoculation with most rhizobial strains, both its own (Fig. 1) and those of other host species (see below and Fig. 4).
The solid line at 1 on the y-axis indicates the level where the host did not gain or lose anything from the rhizobial interaction relative to the N− control treatment. The dashed line is the average symbiotic response to all rhizobial phylotypes. Error bars are based on standard errors of means (if not visible, they are smaller than the symbols). Host species abbreviations are as in Fig. 1. Note the different y-axis scales between figure panels.
The solid line at 1 on the y-axis indicates the level where the host did not gain or lose anything from the rhizobial interaction relative to the N− control treatment. Error bars are based on standard errors of means (if not visible, they are smaller than the symbols). Host species abbreviations are as in Fig. 1.
(a) The symbiotic response of six host plants to rhizobial strains originally collected from a given host species and tested with all hosts. The three strains collected from the same host are depicted with the same colour and the colours correspond to those in Fig. 4b. (b) The symbiotic response of hosts to the same rhizobial strains grouped as sympatric (originally collected from the same host species), and allopatric (strains originally collected from other hosts species). Error bars are based on standard errors of means. In both (a) and (b) the solid line at 1 on the y-axis indicates the level where the host did not gain or lose anything from the rhizobial interaction relative to the N− control treatment. Host species abbreviations are as in Fig. 1.
Experiment II: among-host species inoculations
As expected, there was a consistently high level of nodulation in the second experiment (average percentage of plants nodulated in Experiment I by the subset of symbiotically effective strains used for Experiment II was >95%). Host species differed significantly in their symbiotic response to these strains, while strains did not consistently elicit high or low growth responses in the hosts (P<0.0001 and P = 0.5373, respectively; Fig. 4a). However, the interaction between host species and strain origin was highly significant (P<0.0001; Fig. 4a). In the second analysis where strains were classified as sympatric or allopatric, host species did not have consistently higher or lower growth responses when they were inoculated with sympatric strains (i.e. ones originally sampled from that species) as compared to allopatric strains (those originating from other host species).
Instead, host plants differed significantly in how they responded to strains of sympatric and allopatric origin (host species×sympatry/allopatry interaction: P<0.0001; Table 3 and Fig. 4b). Thus, the symbiotic response of D. ulicifolia and I. australis did not differ with sympatric strains compared to strains originating from other host species, although clearly D. ulicifolia was overall far less responsive to inoculation per se. In contrast, two of the host species (G. lotifolia, H. violacea) demonstrated significantly higher growth with their sympatric strains while two host species (D. retorta, P. ilicifolium) achieved the highest growth with strains originating from other host species (Fig. 4b). Nodulation was overall 5% higher for allopatric interactions compared to sympatric host-strain interactions, and this difference was statistically significant (88% vs. 93%, respectively; P = 0.0297; Table 3).
Native legumes and their associated soil symbionts (rhizobial bacteria) are of ecological importance in many plant communities, and knowledge of the evolutionary history and distribution of these associations has advanced considerably in recent decades , –. Much of the more detailed work on symbiotic effectiveness and rhizobial diversity has focused on variation within host taxa , , , or among host taxa within a genus , . However, little is known regarding how associations might vary at higher taxonomic levels. Here we evaluate differences in the provision of mutualistic benefits across naturally co-occurring host genera, using comprehensive inoculation experiments. A key finding was that strain origin with respect to these hosts (sympatry-allopatry) was not a consistent predictor of symbiotic response. The direction and magnitude in the response to sympatric and allopatric strains varied significantly among host species indicating that host community structure is likely to play an important role in the maintenance of symbiont variation. These results thus have relevance for theory focused on the persistence of cheaters in mutualistic associations , , .
Variation in symbiotic effectiveness within host species
As in previous studies of associations between native legumes and rhizobia , , we found considerable variation in symbiotic effectiveness (i.e. host growth promotion) among strains associated with a particular host species. Not only were there differences in effectiveness among rhizobial strains, there was significant variation among hosts in their growth responses to inoculation with their own strains. For example, both Goodia lotifolia and Indigofera australis nodulated effectively, and grew well, with all sympatric strains (Fig. 1). These two species generally also nodulated and grew well with allopatric strains derived from the other host genera, indicating that these hosts can be considered as generalists. Some species, such as Daviesia ulicifolia and Hardenbergia violacea, exhibited considerable heterogeneity in their growth responses to their own strains; some strain combinations were highly effective, while a significant proportion were clearly worse than the N− control (Fig. 1). Finally, for hosts such as Oxylobium ellipticum, few if any strain combinations were effective (Fig. 1), despite these strains having been originally isolated from that host.
Of particular interest was the finding of significant variation among rhizobial phylotypes in symbiotic performance within a given host (over and above variation among individual strains). Not only do these results indicate that rhizobial phylotypes have relevance in an ecological context (i.e. predicting host growth responses), but they contribute to the ongoing debate about the functional relevance of bacterial species in the wider microbial literature –. Our study provides additional data that further support results from a recent comprehensive study . In that study, molecular analyses of rhizobial community structure from 60 sites across southeastern Australia, together with extensive inoculation trials using two Acacia spp. present at those sites, found that site-level differences in rhizobial genetic diversity could explain a significant proportion of the variance in growth performance observed in the glasshouse. The mechanisms underpinning such differences among phylotypes are unclear. However, our results suggest that contrary to expectations based on the extensive horizontal transfer of genes associated with symbiosis , –, evolutionary history and genetic background are likely to be important determinants of ecological performance in symbiotic bacteria.
When we cross-inoculated a subset of host species with their own and each others most effective strains, we found that host species responded in qualitatively different ways to sympatric and allopatric rhizobial strains (Fig. 4). This ranged from species that clearly preferred their own strains (G. lotifolia, H. violacea), to species that responded well to both sympatric and allopatric strains (Daviesia ulicifolia, I. australis), and those that, somewhat surprisingly, performed significantly better with allopatric strains (P. ilicifolium, Dillwynia retorta). This diversity of responses was observed despite the fact that the rhizobial strains selected for the cross-species inoculation experiment were the most effective symbionts on their own hosts (as determined from the nodulation and growth data from Experiment I). Moreover, the majority of the strains used in Experiment II were grouped within a single phylotype (although clearly there can be genotypic variation within a phylotype). Interestingly, strains from both indiscriminate hosts (D. ulicifolia, I. australis) as well as those preferring sympatric strains (G. lotifolia, H. violacea) were primarily sampled from the same sites where these hosts co-occurred (Table 1). Thus, not only is there considerable variability in species responses, but clearly the outcome of these interactions can be difficult to predict from knowledge of within-species patterns of symbiotic effectiveness.
Clearly, as shown by the highly significant main effect of host (Table 3) some hosts responded much more to inoculation overall than others (e.g. compare G. lotifolia, D. ulicifolia; Fig. 4b). Not only did hosts differ in their overall level of responsiveness, but among-strain differences in symbiotic effectiveness varied considerably between hosts. Both G. lotifolia and D. ulicifolia showed little among-strain variation, while the responses of P. ilicifolium spanned the entire observed range (Fig. 4b). Our analyses did not find an overall strain effect, indicating that symbiotic performance of strains also varied across hosts. We note that the highly significant host×strain interaction was observed despite the fact that Experiment II was conducted with a deliberately selected set of symbiotically effective rhizobial strains, representing a small number of phylotypes. Nevertheless, there were a few strains that, across different hosts, either consistently provided clear symbiotic benefits or were ineffective. The maintenance of different rhizobial strategies (mutualism vs. parasitism) is likely to be at least partly a consequence of evolutionary trade-offs in life-history components associated with symbiosis, among-strain competition and reproductive success , as has been shown for mycorrhizal fungi .
Considerable effort has focused on advancing a conceptual framework for understanding the factors that determine the magnitude and direction of ecological and coevolutionary trajectories in host-symbiont interactions that fall along the parasitism-mutualism continuum . A recent review predicted that symbiotic associations should become less beneficial with increasing environmental quality and that the association of productivity with symbiont specificity depends on tradeoffs between host range and other life-history parameters . At the same time, biotic complexity is expected to favour generalist pathogens but more specific mutualists. Our results demonstrate significant within and among-species variation in symbiotic effectiveness, ranging from essentially parasitic to highly beneficial associations, but also provide empirical support for the role of host community structure in shaping these interactions. Similarly, negative feedbacks in plant performance caused by specificity in mycorrhizal associations have been implicated as an important determinant of coexistence .
Theoretical and empirical studies on the maintenance of variation in host-symbiont associations have largely focused on main-effect differences (i.e. cheaters vs. beneficial mutualists) as might be mediated by trade-offs between the provision of mutualistic benefits and competitive ability among symbionts, or host sanctions , . However, we suggest that a broader perspective requires evaluating such effects in concert with a consideration of other factors that are also likely to influence variability within host×symbiont interactions (e.g. local host diversity, physical environment). For example, the generic composition and diversity of rhizobial communities will partly depend on factors such as soil pH , salinity  or nitrogen levels . Continuing efforts to elucidate the systematics of legumes known to form rhizobial associations  will provide tools for exploring community phylogenetic patterns in host-symbiont associations and how these relate to environmental hetergeneity. Combining such approaches will not only enhance basic understanding of symbiotic interactions, but will ultimately result in a greater ability to predict how manipulation of soil biota will contribute to desired ecological outcomes , .
We thank Jim Bever, Andrew Young and Luke Barrett for helpful discussions, and Jacqui McKinnon and Jo Slattery for technical assistance associated with strain preparation and maintenance of glasshouse inoculation experiments.
Conceived and designed the experiments: PHT DB. Performed the experiments: PHT DB LMB. Analyzed the data: A-LL PHT JB. Contributed reagents/materials/analysis tools: DB LMB. Wrote the paper: PHT A-LL JB DB LMB.
- 1. Vitousek PM, Walker LR (1989) Biological invasion by Myrica fava in Hawaii: plant demography, nitrogen fixation, ecosystem effects. Ecol Monogr 59: 247–265.
- 2. Spehn EM, Scherer-Lorenzen M, Schmid B, Hector A, Caldeira MC, et al. (2002) The role of legumes as a component of biodiversity in a cross-European study of grassland biomass nitrogen. Oikos 98: 205–218.
- 3. Requena N, Perez-Solis E, Azcón-Aguilar C, Jeffries P, Barea1 J-M (2001) Management of indigenous plant-microbe symbioses aids restoration of desertified ecosystems. Appl Environ Microb 67: 495–498.
- 4. Rodriguez-Echeverria S, Perez-Fernandez MA (2005) Potential use of Iberian shrubby legumes and rhizobia inoculation in revegetation projects under acidic soil conditions. Appl Soil Ecol 29: 203–208.
- 5. Thrall PH, Millsom D, Jeavons AC, Waayers M, Harvey GR, et al. (2005) Seed inoculation with effective root-nodule bacteria enhances revegetation success. J Appl Ecol 42: 740–751.
- 6. Reynolds HL, Packer A, Bever JD, Clay K (2003) Grassroots ecology: plant-microbe-soil interactions as drivers of plant community structure and dynamics. Ecology 84: 2281–2291.
- 7. Van der Heijden MGA, Bardgett RD, van Straalen NM (2008) The unseen majority: soil microbes as drivers of plant diversity and productivity in terrestrial ecosystems. Ecol Lett 11: 296–310.
- 8. Herrera MA, Salamanca CP, Barea JM (1993) Inoculation of woody legumes with selected arbuscular mycorrhizal fungi and rhizobia to recover desertified mediterranean ecosystems. Appl Environ Microb 59: 129–133.
- 9. Thrall PH, Hochberg ME, Burdon JJ, Bever JD (2007a) Coevolution of symbiotic mutualists and parasites in a community context. Trends Ecol Evol 22: 120–126.
- 10. Rangin C, Brunel B, Cleyet-Marel JC, Perrineau MM, Béna G (2008) Effects of Medicago truncatula genetic diversity, rhizobial competition, and strain effectiveness on the diversity of a natural Sinorhizobium species community. Appl Environ Microb 74: 5653–5661.
- 11. Heath KD (2009) Intergenomic epistasis and coevolutionary constraint in plants and rhizobia. Evolution 64: 1446–1458.
- 12. Burdon JJ, Gibson AH, Searle SD, Woods MJ, Brockwell J (1999) Variation in the effectiveness of symbiotic associations between native rhizobia and temperate Australian Acacia: within-species interactions. J Appl Ecol 36: 398–408.
- 13. Thrall PH, Bever JD, Slattery JF (2008) Rhizobial mediation of Acacia adaptation to soil salinity: evidence of underlying trade-offs and tests of expected patterns. J Ecol 96: 746–755.
- 14. Thrall PH, Broadhurst LM, Hoque MS, Bagnall DJ (2009) Diversity and salt tolerance of native Acacia rhizobia isolated from saline and non-saline soils. Austral Ecol 34: 950–963.
- 15. Brockwell J, Evans CM, Bowman AM, McInnes A (2010) Distribution, frequency, and symbiotic properties of the Australian native legume Trigonella suavissima Lindl. and its associated root-nodule bacteria. Rangeland J 32: 395–406.
- 16. Thrall PH, Burdon JJ, Woods MJ (2000) Variation in the effectiveness of symbiotic associations between native rhizobia and temperate and Australian legumes: interactions within and between genera. J Appl Ecol 37: 52–65.
- 17. Yates RJ, Howieson JG, Nandasana KG, O'Hara GW (2004) Root-nodule bacteria from indigenous legumes in the north-west of Western Australia and their interaction with exotic legumes. Soil Biol Biochem 36: 1319–1329.
- 18. Simms EL, Taylor DL (2002) Partner choice in nitrogen-fixation mutualisms of legumes and rhizobia. Integr Comp Biol 42: 369–380.
- 19. West SA, Kiers ET, Simms EL, Denison RF (2002) Sanctions and mutualism stability: why do rhizobia fix nitrogen? Proc Roy Soc Lond B 269: 685–694.
- 20. Kiers ET, Rousseau RA, West SA, Denison RF (2003) Host sanctions and the legume-rhizobium mutualism. Nature 425: 78–81.
- 21. Foster KR, Kokko H (2006) Cheating can stabilize cooperation in mutualisms. Proc Roy Soc Lond B 273: 2233–2239.
- 22. Bennett AE, Bever JD (2009) Trade-offs between arbuscular mycorrhizal fungal competitive ability and host growth promotion in Plantago lanceolata. Oecologia 160: 807–816.
- 23. Bever JD, Richardson SC, Lawrence BM, Holmes J, Watson M (2009) Preferential allocation to beneficial symbiont with spatial structure maintains mycorrhizal mutualism. Ecol Lett 12: 13–21.
- 24. Bever JD (2002) Negative feedback within a mutualism: host-specific growth of mycorrhizal fungi reduces plant benefit. Proc Roy Soc Lond B 269: 2595–2601.
- 25. Johnson NC, Wilson GWT, Bowker MA, Wilson JA, Miller RM (2010) Resource limitation is a driver of local adaptation in mycorrhizal symbioses. P Natl Acad Sci USA 107: 2093–2098.
- 26. Torsvik V, Goksøyr J, Daae FL (1990) High diversity in DNA of soil bacteria. Appl Environ Microb 56: 782–787.
- 27. Torsvik V, Øvreås L (2002) Microbial diversity and function in soil: from genes to ecosystems. Curr Opin Microbiol 5: 240–245.
- 28. Torsvik V, Daae FL, Sandaa R-A, Øvreås L (1998) Novel techniques for analysing microbial diversity in natural and perturbed environments. J Biotechnol 64: 53–62.
- 29. Palys T, Berger E, Mitrica I, Nakamura LK, Cohan FM (2000) Protein-coding genes as molecular markers for ecologically distinct populations: the case of two Bacillus species. Int J Syst Evol Microbiol 50: 1021–1028.
- 30. Palys T, Nakamura LK, Cohan FM (1997) Discovery and classification of ecological diversity in the bacterial world: the role of DNA sequence data. Int J Syst Bacteriol 47: 1145–1156.
- 31. Portier P, Fischer-Le Saux M, Mougel C, Lerondelle C, Chapulliot D, et al. (2006) Identification of genomic species in Agrobacterium biovar 1 by AFLP genomic markers. Appl Environ Microb 72: 7123–7131.
- 32. Cohan FM (2006) Toward a conceptual and operational union of bacterial systematics, ecology and evolution. Philos Trans Roy Soc B 361: 1985–1996.
- 33. Doolittle WF, Papke RT (2006) Genomics and the bacterial species problem. Genome Biol 7: 1161–1167.
- 34. Konstantinidis KT, Ramette A, Tiedje JM (2006) The bacterial species definition in the genomic era. Philos Trans Roy Soc B 361: 1929–1940.
- 35. Bever JD, Broadhurst LM, Thrall PH (2011) Microbial phylotype composition and diversity predicts ecological variation within plant-soil community interactions. P Natl Acad Sci USA (in revision).
- 36. von Mering C, Hugenholtz P, Raes J, Tringe SG, Doerks T, et al. (2007) Quantitative phylogenetic assessment of microbial communities in diverse environments. Science 315: 1126–1130.
- 37. Lafay B, Burdon JJ (1998) Molecular diversity of rhizobia occurring on native shrubby legumes in southeastern Australia. Appl Environ Microb 64: 3989–3997.
- 38. Lafay B, Burdon JJ (2001) Small-subunit rRNA genotyping of rhizobia nodulating Australian Acacia spp. Appl Environ Microb 67: 396–402.
- 39. Marsudi NDS, Glenn AR, Dilworth MJ (1999) Identification and characterization of fast- and slow-growing root nodule bacteria from South-Western Australian soils able to nodulate Acacia saligna. Soil Biol Biochem 31: 1229–1238.
- 40. Parker MA, Lafay B, Burdon JJ, van Berkum P (2002) Conflicting phylogeographic patterns in rRNA and nifD indicate regionally restricted gene transfer in Bradyrhizobium. Microbiology 148: 2557–2565.
- 41. Hoque MS, Broadhurst LM, Thrall PH (2010) Genetic characterization of bacteria associated with root nodules of Acacia salicina and A. stenophylla across southeastern Australia. Int J Syst Evol Microbiol 61: 299–309.
- 42. Ngom A, Nakagawa Y, Sawada H, Tsukahara J, Wakabayashi S, et al. (2004) A novel symbiotic nitrogen-fixing member of the Ochrobactrum clade isolated from root nodules of Acacia mangium. J Gen Appl Microbiol 50: 17–27.
- 43. Thrall PH, Slattery JF, Broadhurst LM, Bickford S (2007b) Geographic patterns of symbiont abundance and adaptation in native Australian Acacia-rhizobial interactions. J Ecol 95: 1110–1122.
- 44. Bissett A, Richardson AE, Baker G, Wakelin S, Thrall PH (2010) Life history determines biogeographical patterns of soil bacterial communities over multiple spatial scales. Mol Ecol 19: 4315–4327.
- 45. Harris J (2009) Soil microbial communities and restoration ecology: facilitators or followers? Science 325: 573–574.
- 46. Somasegaran P, Hoben HJ (1994) The handbook for rhizobia: Methods in legume-rhizobia technology. New York: Springer-Verlag. 450 p.
- 47. Vincent JM (1970) A manual for the practical study of root-nodule bacteria. IBP Handbook No. 15. Oxford: Blackwell. 176 p.
- 48. McKnight T (1949) Efficiency of isolates of Rhizobium in the cowpea group, with proposed additions to this group. Queensland Journal of Agricultural Science 6: 61–76.
- 49. SAS Institute (1999) SAS/STAT Software User's Guide. Release 8.00. Cary, NC: SAS Institute Inc. 1464 p.
- 50. Littell RC, Milliken GA, Stroup WW, Wolfinger RD (1996) SAS (R) system for mixed models. Cary, NC: SAS Institute Inc. 656 p.
- 51. Sprent JI (2001) Nodulation in legumes. Kew: Royal Botanical Gardens. 146 p.
- 52. Sprent JI (2009) Legume nodulation. Wiley Blackwell. 183 p.
- 53. Parker MA, Doyle JL, Doyle JJ (2004) Comparative phylogeography of Amphicarpaea legumes and their rootnodule symbionts in Japan and North America. J Biogeogr 31: 425–434.
- 54. Lawrie AC (1983) Relationships among rhizobia from native Australian legumes. Appl Environ Microb 45: 1822–1828.
- 55. Parker MA (1999) Relationships of bradyrhizobia from the legumes Apios americana and Desmodium glutinosum. Appl Environ Microb 65: 4914–4920.
- 56. Murray BR, Thrall PH, Woods MJ (2001) Acacia species and rhizobial interactions: implications for restoration of native vegetation. Ecological Management and Restoration 2: 213–219.
- 57. Kiers ET, Denison RF (2008) Sanctions, cooperation, and the stability of plant-rhizosphere mutualisms. Annu Rev Ecol Evol S 39: 215–236.
- 58. Marco DE, Carbajal JP, Cannas S, Pérez-Arnedo R, Hidalgo-Perea A, et al. (2009) An experimental and modelling exploration of the host-sanction hypothesis in legume–rhizobia mutualism. J Theor Biol 259: 423–433.
- 59. Laguerre G, Nour SM, Macheret V, Sanjuan J, Drouin P, et al. (2001) Classification of rhizobia based on nodC and nifH gene analysis reveals a close phylogenetic relationship among Phaseolus vulgaris symbionts. Microbiology 147: 981–993.
- 60. Finan TM (2002) Evolving insights: symbiosis islands and horizontal gene transfer. J Bacteriol 184: 2855–2856.
- 61. MacLean AM, Finan TM, Sadowsky MJ (2007) Genomes of the symbiotic nitrogen-fixing bacteria of legumes. Plant Physiol 144: 615–622.
- 62. Denison RF, Kiers ET (2004) Lifestyle alternatives for rhizobia: mutualism, parasitism, and forgoing symbiosis. FEMS Microbiol Lett 237: 187–193.
- 63. Thompson JN, Nuismer SL, Gomulkiewicz R (2002) Coevolution and maladaptation. Integr Comp Biol 42: 381–387.
- 64. Han TX, Wang ET, Han LL, Chen WF, Sui XH, et al. (2008) Molecular diversity and phylogeny of rhizobia associated with wild legumes native to Xinjiang, China. Syst Appl Microbiol 31: 287–301.
- 65. Elliott GN, Chou J-H, Chen W-M, Bloemberg GV, Bontemps C, et al. (2009) Burkholderia spp. are the most competitive symbionts of Mimosa, particularly under N-limited conditions. Environ Microbiol 11: 762–778.
- 66. Wall DH, Bardgett RD, Kelly EF (2010) Biodiversity in the dark. Nat Geosci 3: 297–298.