Biological nitrogen fixation as a bacterial process in symbiosis with plants accounts for a major fraction of nitrogen input in terrestrial ecosystems . The spectrum ranges from obligate endosymbioses, such as arbuscular mycorrhizal fungi , to generally facultative nitrogen-fixing symbioses; the latter, including the rhizobia-legume nodules , Frankia-actinorhizal nodules , and cyanobacteriumGunnera associations . Nitrogen-fixing root nodule symbiosis began with one evolutionary event that occurred in the common ancestor of a clade of angiosperms known as the nitrogen- fixing clade approximately 100 million years ago . The Frankia-actinorhizal symbioses are of particular interest in the interpretation of the evolutionary underpinnings of RNS because these symbioses developed, probably independently, within several distinct lineages across the NFC . Genetically distinct subgroups within the genus Frankia form compatible associations with different lineages of plants within the NFC , while cluster IV includes atypical non-symbiotic and/or non-nitrogen-fixing strains . Of these subgroups, cluster II Frankia strains have the broadest host range among nitrogen-fixing plants: species in four different families in two orders form RNS with cluster II Frankia strains. Representatives of these families are particularly abundant in western North America, where the current study has been carried out. Comparisons of representative sequenced genomes from these subgroups suggested a general correlation of Frankia genomic features with the corresponding host bio-geographic distribution . That being said, since RNS is first established between the host-plant root and the bacterial symbiont by the two partners coming into physical contact with each other in the rhizosphere, procona valencia it is to be expected that all of the symbiotic Frankia subgroups must have at least some capability to persist in a free living state in soil, in addition to living in symbiosis.
Clusters I, III, and IV are known to have a free-living phase in soil: infective units of clusters I and III are ubiquitously distributed in soils . Moreover, infective units of cluster I Frankia can be as abundant, or more abundant, in nearby non-host rhizosphere soil than in hosts . Similarly, strains of cluster III Frankia are well known to occur in soil independent of host-plant presence . Cluster II Frankia strains, however, have been characterized as highly host dependent, with only one report of nodule formation from soil devoid of a host, 18 months after inoculation . Previous DNA-based surveys of cluster II Frankia strain distribution have been conducted using field-collected nodules . Thus, there is only limited knowledge about the soil ecology of cluster II Frankia strains in relation to their host plants, particularly concerning the degree of association with the host presence. Since the cluster II Frankia strains have been shown to be phylogenetically early divergent within the genus , a better understanding of host association in this subgroup may shed light on the process of evolutionary acquisition of RNS. The paucity of evidence for cluster II Frankia strains living independently of host plant influences compared with the other subgroups led us to hypothesize that cluster II Frankia strains may be unusually dependent on the unique environment provided by the host plant for their long-term persistence . To test this, we conducted a comparative survey to determine the soil distribution pattern of free-living cluster II Frankia strains in three geographical locations in northern Califtornia; in different soil types, including serpentine soils, a depauperate soil type that harbors several actinorhizal hosts; and under different climatic conditions. The distribution of cluster II Frankia strains was assessed through detection of DNA sequences unique to cluster II Frankia strains in host and non-host rhizosphere soils, sampled in sites with species of host and non-host plants. The microbiome of a subset of these soil samples was also analyzed via high-throughput sequencing.
Because of the seasonally arid climate and soil environment, which has been shown to limit nitrogen fixation activity in Ceanothus , nitrogen-stable isotope analysis was conducted to confirm an active state of nitrogen-fixing symbiosis in each host plant sampled. In addition, stable isotope analysis enabled us to assess the degree of site-level influence of nitrogen-fixing host plants on cluster II Frankia strain distribution.Our study has demonstrated, based both on presence/absence of glnA sequences and relative abundance of OThus in rhizosphere soil, a significant correlation between the presence of cluster II Frankia strains and the presence of host plants in a particular site, that is consistent across all three experimental locations. There was complete correlation between cluster II strains and host plant presence in two of the three locations . Even though a high proportion of host-plant-absent site samples in MNR were positive for cluster II Frankiastrains , this was nevertheless statistically significantly different from the host-plant-present site samples . This confirms and expands upon previous observations of increased abundance of cluster II Frankia strains in soil of one host-plant-present site relative to the soil of an adjacent host-plant-absent site, using a bait plant bio-assay . While the effect of host-plant presence was consistent, the following ecological factors did not correlate with cluster II Frankia strain presence/ absence across the three locations: host species present , soil type , elevation, geographic location, summer or annual precipitation, average monthly temperature, or soil pH. The consistency of the host influence is evidence that a host-plant-derived factor present in the soil of host-plant-present sites is the major determinant of cluster II Frankia distribution in these ecosystems. We found that pH was not a significant ecological factor determining presence/ absence of cluster II Frankia strains. The pH differed significantly among the three locations, but within each location, the pattern was consistent between sites, regardless of host presence or absence. Earlier studies found pH to have an influence on Frankia nodulation units in cluster I strains , but, on the other hand, both cluster I Frankia and cluster II Frankia strains have been reported to nodulate successfully in a wide range of pH between 3.5 to 8.0 and 5.0 to 9.0 , respectively. The range of soil pH in our study was between 5.0 and 7.2. Although no studies to date have examined the effect of pH directly on cluster II Frankia strain populations in soil, in culture, the optimal growing condition seems to be neutral to basic.
The cultured cluster II Frankia strain BMG5.1 was shown to grow optimally at pH between 9.0 and 10.0 , and in another study, Coriaria myrtifolia had optimal growth and nodulation at pH between 7.0 and 9.0 . It seems, therefore, that in general Frankia spp. can tolerate a range of pH while having a more specific optimum. It is possible that at a more basic pH, where cluster II Frankia spp. are known to grow optimally, there will be a difference in cluster II Frankia abundance; however, Frankia presence/absence is not affected.ANOVA of the rhizosphere microbiomes showed that while host-plant presence/absence at a given site significantly correlated with the relative abundance of cluster II Frankia spp., this correlation did not differ significantly between the rhizospheres of C. jepsonii and H. arbutifolia , when H. arbutifolia was growing in the host-plant-present site. This suggests that the effect of the host plant is not limited to its rhizosphere but can extend to the site level. A site-level influence was also observed for nitrogen fixation, determined by the ANOVA of 15N. While host plants all showed significantly more negative 15N values than the co-occurring non-host plants, flower bucket indicating active engagement in nitrogen fixation via RNS , the non-host species in host-plant-present sites also showed significantly more negative 15N values in host-plant-present sites than in the same species in host-plant-absent sites. This pattern of 14N stable isotope enrichment at the site level is similar to the facilitative effect of nitrogen-fixing plants on primary succession, reported in other nitrogen-fixing ecosystems . It was shown previously that Thissue extracts from both roots and shoots of a host plant Casuarina cunninghamiana enhanced growth of a range of Frankia spp. grown in culture but particularly enhanced the in vitro growth of the Casuarina-compatible strains . This suggests that some factor derived from the host plant, either as leaf litter or root exudate, can explain the site-level effect observed in our experiments.Our findings suggest that the abundance of cluster II Frankia spp. in a given site is associated with the strength of the influence that a host plant exerts. In the case of MNR, where the host-plant-absent site had a relatively close proximity to the nearest host plant growing in a similar vegetation type , cluster II Frankia spp. were detectable; however, no cluster II Frankia spp. were detectable in the CC site in ALCP where no host plant was proximate , despite similarities in vegetation types. At the SEF site, cluster II Frankia spp. were not detected in the wetland site despite the high proximity to the nearest host plants ; however, the wetland site had dramatically different soil conditions and vegetation type than the adjacent montane chaparral. Cluster II Frankia itself can survive in moist wetland soil: DaThisca glomerata, another host for cluster II Frankia spp. found in Califtornia, Nevada, and Baja Califtornia, grows and nodulates in seasonally flooded streambeds or washes throughout its range . Rather, it is likely that cluster II Frankia spp. were absent in the wetland site at SEF because the more xeric-adapted host plants could not grow in wetland soil, thus eliminating a host effect. The strength of the influence that the host plants may have on a given site might be affected by factors other than proximity or vegetation type. Since the cluster II Frankia enhancement factor may be able to persist in the soil over time, a legacy effect from host plants that were previously present might play a role. For example, cluster II Frankia spp. have been found in sites devoid of host plants for more than 100 years ; Elaeagnus-compatible Frankia spp. were present at a site that had not had Elaeagnus for many years , and Comptonia peregrina seed was viable after being in a pine-dominated forest seed bank for more than 70 years . Alternatively, since Ceanothus and Cercocarpus support mycorrhizal symbiosis , the mycorrhizal mycelia may be a means of conveying a chemical signal.There was low glnA sequence variability throughout this study with all samples within 0.01 substitution/site from each other; in fact, of the 74 sequences generated in this study, 56 had at least one other sequence that was identical .
Low genetic diversity of cluster II Frankia spp. based on DNA from field nodule collections, rather than from rhizosphere soil DNA, has been well documented in Oregon with repetitive sequence PCR or PCR-restriction fragment length polymorphism and in Califtornia and Nevada using single-sequence analysis . interestingly, a glnA sequence identical to one found in the current study has been detected in DNA extracted from a nodule of Dryas drummondii collected in Alaska . On the other hand, DNA from nodules collected in the southeastern United States showed a substantial divergence, based on RFLP, from that of nodules from western North America . A study of nodules collected in a circumscribed area in southern coastal Califtornia found diversity of cluster II Frankia at the species level, using rep-PCR . The sampling area in the southern coastal Califtornia study corresponds to a zone of biodiversity for the genus Ceanothus as documented previously . Taken together, these findings suggest that a suite of closely related strains of cluster II Frankia is dominant in major portions of western North America, with some localized diversity. Various factors have been suggested to account for the observed strain diversity within cluster II, however low it may be. Previously, a weak geographic pattern of strain diversity by Ceanothus collection sites was found . In reference 15, diversity was attributed to two ecological groupings of the host plants , but the sequence groupings were separated by 1%. Both of these findings reflect the generally low cluster II Frankia strain sequence diversity in western North America. Whole-genome sequencing has been carried out on strains of cluster II Frankia from Pakistan and Japan ; additional whole-genome sequencing from a greater number of sources will be valuable for assessing the degree of global geographic and genetic diversity within this clade.It has been postulated that cluster II Frankia is an obligate symbiont, based on the difficulty of isolating any strain in this group and the difficulty of detecting it in soils devoid of host plants .