Two cluster II Frankia genomes contain close homologs of rhizobial nodABC genes

There has been only one report of nodulation of a compatible host from soil that was devoid of native host plants , although only after 18 months of bait plant incubation. Recently, a cluster II Frankia strain was isolated in pure culture from a nodule of Coriaria japonica collected in Japan . These two examples indicate that not all cluster II Frankia strains are obligate symbionts. That being said, it seems clear that the cluster II Frankia strains form a strongly restrictive association with their host plants. Our study adds to the complex nature of cluster II Frankia strain association with their hosts. Two degrees of ecological association between a microsymbiont and its host may be considered: persistence and enrichment. Members of cluster I and III Frankia spp. have been known to persist in soil devoid of host plants for many years and to readily nodulate their respective hosts . Moreover, the number of infective units of cluster I Frankia spp. has been known to increase under Alnus host plants up to 30-fold, which makes the presence of the host plants a major factor in ampliftying Frankia populations . interestingly, the rhizosphere of Betula sp., a non-host closely related to the genus Alnus, showed enrichment of cluster I Frankia spp. compared to the rhizosphere of Alnus spp. at a host-plant-present site and abundance comparable to Alnus spp. at a host-plant-absent site . In our study, as evidenced in the microbiome analysis, an enrichment effect of cluster II Frankia was present but was much subtler. Cluster II Frankia spp. were detected in the host-plant-present sites and were also detected in rhizosphere soil of a non-host, H. arbutifolia, in the host-plant-present site, but no cluster II Frankia spp. were detected in the host-plant-absent site. In contrast, plastic flower bucket strains of the other clusters of Frankia were detected in both host-plant-present and -absent sites.The relatively greater abundance of cluster I Frankia spp. among the typical symbiotic and nitrogen-fixing subgroups that we observed is in congruence with results of previous studies . On the other hand, we found that cluster III was the least abundant: 6 reads total and only two samples with any reads detected.

This contrasts with previous findings where cluster III Frankia spp. was the dominant or codominant subgroup in a noncompatible host rhizosphere . This difference may be due to the fact that the soil conditions among these studies are at the opposite end of the ecological spectrum. In reference 35 the soil was moist and in a university campus arboretum, whereas our MiSeq samples were collected in dry nutrient-poor serpentine soil in summer. It is possible that there was a specific effect of serpentine soil conditions on cluster III Frankia spp. For example, Oline showed that while microbiomes from serpentine soils were similar to those of nearby non-serpentine soils at the phylum level, distinct subgroups were adapted to the specialized environment of serpentine soil. Not very much is known about the population distribution of the atypical cluster IV Frankia strains in soil. However, since they comprise the most abundant subgroup inour samples, it is clear that at least some strains of cluster IV Frankia are adapted to serpentine soil. The observed trend of decreasing clusters I and IV Frankia spp. in host-plant-present sites compared to that in host-plant-absent sites suggests that the host plant may be influencing the population sizes of strains of clusters I and IV as well as of cluster II Frankia. This may be related to an inhibitory factor in the host soil ecosystem, as discussed below. Nearly 34% of the reads that mapped to the genus Frankia mapped to OThus of indeterminate identity. Thus, depending on the true identity of these OThus, our results are subject to change. However, all of the reads that mapped to indeterminate OThus were  1% different from any of the cluster II OThus; with the mapping threshold set to 99%, no reads that mapped to these OThus would map to cluster II OThus, and all indeterminate OThus are more closely related to known cluster I or III strains than to cluster II strains.While the relative abundance of cluster II Frankia spp. in the microbiome was significantly higher in the host-plant-present site, it is still clear that this group of OThus is quite rare in the microbiome, representing only 0.05‰ in any sample.

This suggests that the host factor is not a source of energy to sustain a large population; i.e., it is not promoting the prolifteration of cluster II Frankia outside the host, in contrast to some strains within the cluster I Frankia that can utilize a host-derived soil carbon source . Further, permutational MANOVA and factorial ANOVA both showed that the presence of the host plant did not significantly affect the relative abundances of all OThus in the microbiome overall nor the -diversity; in fact PCoA with an outgroup showed that the microbiomes of the samples from the host-plant-present and host-plant absent sites were very similar to each other. This minimal influence of host/non-host on the overall microbiome together with the trend observed for clusters I and IV Frankia OThus suggests that a host-derived factor has an effect specifically on cluster II Frankia OThus. Cluster II Frankia OThus were found in the rhizosphere soil of the non-host H. arbutifolia in host-plant-present sites, but not in the rhizosphere of H. arbutifolia in host-plant-absent sites. Thus, the host-plant effect was not limited to the host-plant rhizosphere but rather extended to the level of the site where the host plant was present. Extracts from host-plant roots and shoots have also been found to enhance growth of Frankia strains in culture particularly strongly if the strain is compatible . Taking this into account, we propose that possible mechanisms for the dispersal of this host-dependent factor might be diffusion of a compound or compounds originating as root exudates, leaf litter decomposition and leaching into the soil profile, or chemical signals dispersed via mycorrhizal networks . A plant factor with specific targets that might have a strong impact on Frankia strain presence in the rhizosphere might be a plant-derived secondary compound related to nodulation signaling, such as a flavonoid. Flavonoids are known signaling molecules in rhizobial symbioses . Additionally, in Myrica sp. symbioses, a suite of hydroxy-chalcones, termed myrigalones, has been shown to promote infective Frankia and inhibit noninfective Frankia strains . Alternatively, this factor might be a terpenoid with effects similar to strigolactone, known to stimulate spore germination and limited hyphal growth of fungal symbionts in the rhizosphere in arbuscular mycorrhizal endosymbioses . The host signaling pathway that initiates AM symbiosis is ancestral to signal exchange in RNS; and the microsymbiont signal molecules are also similar between these two types of endosymbiosis . These nod gene homologs have been shown to be expressed in developing root nodules of D. glomerata . In summary, our data demonstrate the existence of a host site enrichment of cluster II Frankia strains that is independent of host plant species, vegetation type, geographical location, climate, soil type, and soil pH. The dependency of Frankia spp. on theirhosts for survival and persistence has been postulated, based on in vitro experiments . Host and non-host rhizospheres were recently compared with respect to cluster I Frankia strains in one location , showing no significant host rhizosphere effect. Cluster III Frankia strains were shown to have a distribution independent of the host rhizophere, implying a broad physiological adaptation in soil . The current report is the first comprehensive study showing a significant host plant influence on cluster II Frankia strains in natural soil environments, an effect that is distinct from that of other ecological parameters. Identification and testing of potential specific host-plant compounds and transmission pathways remain to be carried out in more controlled environment experiments.All soil, leaf, flower buckets wholesale and root nodule samples were collected from three locations in northern Califtornia: in and near Anderson Lake County Park, Santa Clara County, Califtornia , McLaughlin Natural Reserve, Lower Lake, Califtornia , or Sagehen Experimental Forest, Truckee, Califtornia , as shown in Fig. 6 . Maps of these locations were generated with R with the ggmap package .

These locations varied considerably in altitude, monthly precipitation, monthly temperature, and soil type. The altitudes were approximately 200 m, 650 m, and 1,950 m above sea level in ALCP, MNR, and SEF, respectively. Multiple sampling sites were selected within each location: ALCP, n 4 sampling sites; MNR, n 5 sampling sites; and SEF, n 2 sampling sites. The main factors that were tested across these sites were host plant presence/absence , host plant species , soil type , soil pH, vegetation type, and climate type. Factors 1 to 5 are described in Table 2. The soil type in each site was determined by methods described below, and the vegetation types were determined to the macro group level according to the U.S. National Vegetation Classification using A Manual of Califtornia Vegetation . The annual patterns of temperatures and precipitation at each location were determined according to the WorldClim database . Sites within each location were paired according to host-plant presence/absence for comparison. In ALCP, AR, WA, and K were compared with CC on serpentine and non-serpentine soil . In MNR, SE02 was compared with SE05 , both on serpentine soil, and NS02 was compared with NS01 , both on non-serpentine soil. In SEF, SH06 was compared with SH01 , both on non-serpentine soil. Vegetation composition was similar for pairs in ALCP and in MNR . In the Sagehen Creek site, the pairing was between sites with dissimilar but adjacent vegetation types . In ALCP, the host plant Ceanothus ferrisiae was present in AR, WA, K but absent in CC. In MNR, the host plant Ceanothus jepsonii was present in SE02, and the host plant Cercocarpus betuloides was present in NS02. In SEF, Ceanothus velutinus was present in SH06. Site SE06 in MNR was used to collect leaf samples of Avena fatua for 15N comparison .Strawberry production in Califtornia accounts for more than 80% of total U.S. production, with an annual farm gate value of $1.10 billion , which is four times greater than all other states combined . In addition, Califtornia produces nearly one billion strawberry transplants each year in nurseries, and these transplants must meet strict phytosanitary standards for local production and export. Such a profitable industry in Califtornia has been made possible by the fumigation technology developed in the 1950s with methyl bromide and chloropicrin . Since then, preplant fumigation with methyl bromide and chloropicrin has become an integral part of the Califtornia strawberry production industry , and nearly all conventional strawberry production occurs in fumigated soils . Annual soil fumigation has contributed to the control of soilborne pathogens, nematodes, and weeds while also boosting the yields of strawberry plants. Historically, this also allowed breeding programs to focus on improving horticultural characteristics of strawberry cultivars in lieu of emphasizing disease resistance. Because of the negative effects of methyl bromide on stratospheric ozone, the fumigant was designated as a class I stratospheric ozone depleting substance by the Montreal Protocol and as a significant risk to human health . The continued availability of this efficient fumigant for agricultural soil fumigation beyond the 2005 phase-out date will be through critical-use exemptions. It has been estimated that annual losses in short term net farm income in Califtornia will be more than $162 million, with strawberry accounting for more than 60% of these losses . Over the past 10 years, research has focused on identifying alternative fumigants with efficacy comparable with methyl bromide . Alternative fumigants such as chloropicrin and Telone C35 have been identified, and improved application techniques have been developed to reduce emissions . Although chloropicrin is as efficacious as methyl bromide + chloropicrin at high rates, these are not feasible for the growers due to regulatory limits placed on application rates. Regardless, chemical alternatives to methyl bromide will be subjected to increasing review and regulation and they may not be readily available over the longer term. It has been estimated that soilborne diseases caused by Pythium, Phytophthora, Cylindrocarpon, Macrophomina, Rhizoctonia, and Verticillium spp. result in 20 to 30% strawberry yield losses in the absence of fumigation .