Provenance of rhizobial symbionts is similar for invasive and noninvasive acacias introduced to California

Abstract

Plant–soil interactions can be important drivers of biological invasions. In particular, the symbiotic relationship between legumes and nitrogen-fixing soil bacteria (i.e. rhizobia) may be influential in invasion success. Legumes, including Australian acacias, have been introduced into novel ranges around the world. Our goal was to examine the acacia–rhizobia symbiosis to determine whether cointroduction of non-native mutualists plays a role in invasiveness of introduced legumes. To determine whether acacias were introduced abroad concurrently with native symbionts, we selected four species introduced to California (two invasive and two noninvasive in the region) and identified rhizobial strains associating with each species in their native and novel ranges. We amplified three genes to examine phylogenetic placement (16S rRNA) and provenance (nifD and nodC) of rhizobia associating with acacias in California and Australia. We found that all Acacia species, regardless of invasive status, are associating with rhizobia of Australian origin in their introduced ranges, indicating that concurrent acacia–rhizobia introductions have occurred for all species tested. Our results suggest that cointroduction of rhizobial symbionts may be involved in the establishment of non-native acacias in their introduced ranges, but do not contribute to the differential invasiveness of Acacia species introduced abroad.

invasive, mutualism: nitrogen-fixation, nifD, nodC, rhizobia

Introduction

Species that are introduced to novel regions likely often become isolated from their native mutualists. Thus, species that are introduced concurrently with their native mutualistic partners may have a competitive advantage over other non-native or native organisms in their novel ranges (Richardson et al. 2000). However, few studies have taken a broad geographical approach to examine the cointroduction of non-native mutualistic organisms to new regions, and those that have frequently focus on individual species or locations (Wandrag et al. 2020). One category of cointroduced mutualistic organisms is plants and their microbial mutualistic symbionts (Rodríguez‐Echeverría et al. 2011). Microbial mutualists may be transported along with plant materials in accompanying soils, on seeds, or as intentional inoculants to improve establishment of species abroad (Pérez-Ramírez et al. 1998). The cointroduction of plant species with their microbial mutualists may influence their effective establishment, colonization, and subsequent invasion in novel ranges.

Legumes may benefit from cointroduction to a novel range with their sympatric, coadapted bacterial mutualists (Parker 2001). Legumes form beneficial symbiotic relationships with nitrogen-fixing bacteria (i.e. rhizobia) that influence their establishment success, particularly in low-nitrogen habitats (Sprent 2001). Recent research indicates that invasive legumes, in particular Australian acacias, have frequently been cointroduced with their native rhizobial mutualists (Rodríguez‐Echeverría 2010, Rodríguez‐Echeverría et al. 2011, Crisóstomo et al. 2013, Ndlovu et al. 2013).

Australian acacias have been introduced widely to regions outside their native ranges. They are a model system for examining mechanisms by which suites of introduced species become invasive (Richardson et al. 2011). More than 1000 Acacia species are native to Australia (Miller et al. 2003), at least 386 of which have been introduced around the world for a variety of purposes, including forestry, fuel wood, erosion control, and ornamental use (Richardson et al. 2011). In areas where they are introduced, acacias range from highly invasive (23 species), naturalized (48 species), to noninvasive (315 species), with some species causing drastic changes to natural areas, and others having only minor impact (Richardson et al. 2011, Rejmánek and Richardson 2013). Understanding the driving forces behind why certain Acacia species are successful invaders and why others are not is important for developing methods to control their spread and can elucidate the mechanisms behind legume invasion in general.

One region where multiple Acacia species have been introduced abroad is the state of California (USA). Acacias were introduced to California beginning in the mid-1800s through the ornamental trade (Butterfield 1938). There are currently 16 Acacia species that occur and vary in invasiveness in this region (two invasive, five naturalized, and nine noninvasive; Cal-IPC, 2006; Jepson eFlora, 2015; CalFlora, 2015). While Acacia species are differentially invasive within California, all species introduced to this region except for one have become invasive in at least one part of the globe (Richardson et al. 2011). Acacias are markedly less successful in California as invasive species than in most other parts of the world. The variation in invasiveness of Acacia species within the localized region of California provides an opportunity to test mechanisms of invasion by comparing multiple closely related species and examining what drives their differential establishment and colonization success abroad.

Mutualisms are important in the life history of Acacia species and may have a key role in their ability to invade novel regions. Acacias are recognized as having mutualistic relationships with a range of species, including ants that protect them from herbivores (Holmes 1990) and birds that disperse their seeds (Glyphis et al. 2010), but it is their interaction with nitrogen-fixing soil bacteria (i.e. rhizobia), that may be particularly important to their establishment and colonization success in novel ranges. Rhizobia are Gram-positive bacteria that occur in the soil and infect roots of many legumes, creating nodules on the roots in which the rhizobia are housed (Beringer et al. 1979). Rhizobia fix atmospheric nitrogen and convert it to amino acids and proteins, which are available to the plant; in turn, the plant provides carbon substrates to the bacteria (Sprent and Sprent 1990, Sprent 2001). For acacias with effective symbiotic rhizobia, the interaction provides an available source of nitrogen for the plant. Acacias that encounter mutualistic bacterial partners in their novel ranges may have competitive advantages when introduced abroad.

Parker (2001) developed models to predict the likely success of legume invasion in novel ranges under different levels of rhizobial availability; without suitable rhizobial partners, legumes may fail in their ability to colonize novel ranges. In a comparison of over 3500 legume species, Simonsen et al. (2017) found that nonsymbiotic legumes have invaded more novel regions than symbiotic legumes, suggesting that symbiotic legumes are constrained in establishment abroad due to their need to find adequate symbiotic rhizobial partners. Cointroduction with coevolved native mutualistic symbionts would likely provide immediate ecological benefits for acacias and remove the requirement to form partnerships with beneficial novel rhizobia (should these exist).

The acacia–rhizobia symbiosis is especially useful for testing a contemporary hypothesis that focuses on the role of mutualisms in invasions, the Enhanced Mutualism (EM) Hypothesis (Reinhart and Callaway 2006, Rodríguez‐Echeverría 2010, Crisóstomo et al. 2013, Ndlovu et al. 2013). The EM Hypothesis suggests that exotic species benefit from mutualisms that originate in their introduced range (Reinhart and Callaway 2006). If Acacia species are not introduced abroad with their native rhizobial mutualists, those that become invasive may represent symbiotic generalists that are more capable of developing novel associations with rhizobia native to their introduced range. By studying groups of closely related introduced species that vary in their invasion success, it may be possible to pinpoint specific mechanisms that enable certain species to become invasive.

The ability of some acacias to associate with a wider diversity of rhizobial strains, hereafter called “host promiscuity,” may also influence the invasiveness of introduced Acacia species. Those species that can associate with more rhizobial strains may more easily find compatible rhizobial mutualists when introduced abroad (Thrall et al. 2000). Higher host promiscuity has been shown in acacias that have become invasive in multiple regions of the world (Klock et al. 2015), as well as acacias that are more widely distributed in their native continent (Thrall et al. 2000). The cointroduction of non-native acacia species with their native rhizobial symbionts, paired with the ability to associate with a wider diversity of rhizobial strains should their native symbionts not be available, may greatly promote the ability of certain Acacia species to become invasive when introduced abroad.

The aim of this study was to examine the role of rhizobial provenance and diversity of rhizobial symbionts in the establishment success of Acacia species introduced abroad. We took a culture-based approach in conjunction with a multilocus phylogenetic analysis (studying three different loci) to assess whether invasive acacias were introduced abroad with their native rhizobial mutualists. We selected four Australian Acacia species that differ in invasiveness in California (two invasive and two noninvasive) and identified rhizobial strains associating with each species in their native and novel ranges using three genes, 16S rRNA, nifD, and nodC. The 16S rRNA gene has been commonly used for identification of bacterial genera associated with legumes, whereas the nifD and nodC genes have been used to determine biogeographic placement of rhizobial strains (Parker et al. 2002, Qian et al. 2003, Moulin et al. 2004, Stępkowski et al. 2005, Andam and Parker 2008, Barrett et al. 2016). Barrett et al. (2016) have shown that Australian rhizobia likely have very different nod and nif genes, and that 16S rRNA and functional genes can give incongruent results. Strains that have identical 16S rRNA may have very divergent nod and nif genes, and because of this it is necessary to rely on genes other than 16S rNRA to determine rhizobial provenance.

A previous review of nine studies examining the role of acacia–rhizobia symbiosis in invasion showed that acacias, whether invasive or not, are not limited in finding suitable rhizobial symbionts abroad (Wandrag et al. 2020). However, the provenance of these symbiotic rhizobia in California has not previously been determined, and thus it is unknown whether acacias have been cointroduced to this range with their native rhizobia, and if so, whether cointroduction provides a beneficial invasion response in this region. Thus, we hypothesized that provenance and diversity of rhizobia associating with acacias in California would influence invasiveness, making the following predictions: (1) invasive acacias harbor different rhizobial communities than noninvasive acacias; (2) invasive acacias host a greater diversity of rhizobia than noninvasive acacias; and (3) invasive acacias are more promiscuous hosts, associating with a greater diversity of rhizobial strains where they are native and where they are introduced abroad.

Materials and methods

Study species

We examined bacterial communities from the root nodules of four Acacia species native to Australia and introduced to California: two Acacia species that are considered to be invasive (A. dealbata and A. melanoxylon; CalFlora, 2015; Cal-IPC, 2006), and two species that have not become invasive in this region (A. longifolia and A. verticillata; Jepson eFlora, 2015). In California, A. dealbata and A. melanoxylon occur along the entire coastal region and have also expanded inland. Acacia longifolia and A. verticillata occur along the California coast from the San Francisco Bay Area south to San Diego (Jepson, eFlora, 2015). Within their native continent, all four of these Acacia species are widely distributed [i.e. > 1000 herbarium records as reported by the Australian Virtual Herbarium (AVH, 2015)]. They are found primarily in southeastern Australia, with A. melanoxylon extending further north (AVH, 2015; Fig. 1).

Figure 1.

Distribution maps and photos of Acacia species used in this study in their native continent of Australia [based on herbarium records from the Australian National Herbarium, Canberra, Australia (AVH, 2015)].

Bacterial isolation

Nodule collection in northern California was performed in August 2012 and 2013. Nodules were collected from nine A. dealbata populations, 18 A. melanoxylon populations, five A. longifolia populations, and two A. verticillata populations (Fig. 2). Variation in the number of collection sites among species was due to relative differences in presence of these species in this region. Nodules were obtained from one to five plants per population by carefully digging ∼30 cm around the base of seedlings or adult plants. When seedlings were present in a population, the entire plant was excavated, attached soil was gently removed, and nodules were clipped from the roots, leaving 0.5 cm of root tissue on either side of the nodule (Somasegaran and Hoben 1994). For adult trees, nodules were extracted from roots attached to the base of the plant. Nodules were stored in collection vials containing silica gel and sterile cotton wool and maintained at room temperature for 1 month until use, as recommended by Somasegaran and Hoben (1994).

Figure 2.

Nodule collection sites in (A) California and (B) Australia.

To culture bacteria, desiccated nodules were first rehydrated by soaking in sterile water overnight. A total of one to three nodules per plant were individually surface sterilized by immersion in 95% ethanol for less than 10 s, followed by soaking in 3% NaCl for 2–4 min, and then rinsed at least five times with sterile water (Somasegaran and Hoben 1994). Nodules were individually crushed using sterile plastic pestles. A volume of 100 ml of the suspension was transferred to plates containing yeast mannitol agar (YMA; 5.0 g mannitol, 0.5 g K₂HPO₄, 0.2 g MgSO₄.7H₂O, 0.1 g NaCl, 0.5 g yeast extract, 1 l distilled H₂O, and 20 g agar; Somasegaran and Hoben 1994). Plates were incubated at 25°C for up to 21 days. Single well-isolated colonies were selected from original plates and successively subcultured two to three times to obtain pure cultures. Pure cultures were stored on YMA plates at 4°C and at −80°C in in YMA broth supplemented with 20% glycerol for long-term storage.

Root nodules were collected from A. longifolia and A. verticillata in Australia in April 2013. Nodules were collected from five populations of A. longifolia and seven populations of A. verticillata (Fig. 2), and from four to five plants per population, stored in plastic bags and kept on ice, then transferred to 4°C for no longer than 2 days before processing. Three nodules per plant were processed and bacterial isolates were cultured as described above. Bacterial isolates from A. dealbata and A. melanoxylon root nodules were obtained from cultures stored in the Commonwealth Scientific and Industrial Research Organization (CSIRO) rhizobial culture collection. These isolates were originally obtained during a collection expedition across the east coast of mainland Australia and Tasmania from 1993 to 1995, as described by Burdon et al. (1999). We cultured 19 isolates from 10 A. dealbata populations, and 16 isolates from nine A. melanoxylon populations.

To extract DNA from pure cultures, single isolates were suspended in 100 ml TE buffer and boiled at 95°C for 5 min. Extractions were examined for concentration and integrity of DNA using gel electrophoresis in 1% agarose in 1X TBE buffer. Lysed cells were stored at 4°C until use but not longer than 4 h.

PCR and sequencing

To determine phylogenetic and biogeographic placement of rhizobial strains associating with different Acacia species, we sequenced three genes from DNA extractions including 16S rRNA, nifD, and nodC. The 16S rRNA gene was used for bacterial identification and phylogenetic placement. The 16S rRNA gene (∼1200 bp) was amplified using the primers forward 27F (5′-AGAGTTTGATCMTGGCTCAG-3′) and reverse 1429R (5′-TACGGYTACCTTGTTACGACTT-3′) (Lane et al. 1985) for cultures obtained in California, and forward GM3 (5′-AGAGTTTGATCMTGGC-3′) and reverse GM4 (5′-TACCTTGTTACGACTT-3′) (Muyzer et al. 1995) for cultures obtained in Australia. The PCR program was as follows: initial predenaturation cycle at 94°C for 5 min, followed by 35 cycles of denaturation at 94°C for 1 min, annealing at 50°C for 1 min, extension at 72°C for 2 min, and a final extension at 72°C for 10 min. Reactions were carried out in a volume of 50 µl containing: 50 ng template DNA, 0.2 µM of each primer, 200 µM of each dNTP (Astral Scientific, Taren Point, Australia), 4.8 mM MgCl₂, and 0.3 U of Taq DNA polymerase (New England Biolabs, Ipswich, MA, USA) in Taq DNA polymerase reaction buffer, using a Hybaid PCR Express thermocycler (Integrated Sciences, Willoughby, Australia).

We also sequenced the nifD (∼620 bp) and nodC (∼620 bp) genes, which play a role in nitrogen fixation and root nodulation, respectively (Sprent 2001). We selected a subset of samples identified by the 16S rRNA analysis at 98% identity as Bradyrhizobium and Rhizobium to use for the biogeographic analysis. These genera were identified in our analysis as the ones predominantly associating with acacias in both Australia and California, and primers, as well as existing sequences are readily available to identify variation in genetic and biogeographic structure of these rhizobial genera. The gene nifD was amplified using the primers forward nifD2F (5′-CATCGGIGACTACAAYATYGGYGG-3′) and reverse nifD1R (5′-CCCAIGARTGCATYT GICGGAA-3′) (Fedorov et al. 2008). PCR touchdown program was as follows: initial predenaturation cycle at 96°C for 2 min, followed by seven cycles of denaturation at 96°C for 30 s, annealing at 55°C* for 45 s (*reduced by 1°C for each cycle), extension at 45°C for 1 min, followed by 25 cycles of denaturation at 96°C for 30 s, annealing at 48°C for 45 s, extension at 72°C for 90 s, and a final extension at 72°C for 5 min. The gene nodC was amplified using the primers forward nodCF540 (5′-TGATYGAYATGGARTAYTGGCT-3′) and reverse nodCR1160 (5′-CGYGACARCCARTCGCTRTTG-3′) (Sarita et al. 2005). PCR touchdown program was as follows: initial predenaturation cycle at 96°C for 2 min, followed by two cycles of denaturation at 96°C for 20 s, annealing at 63°C for 30 s, extension at 72°C for 30 s; two cycles of denaturation at 96°C for 20 s, annealing at 62°C for 30 s, extension at 72°C for 35 s; three cycles of denaturation at 96°C for 20 s, annealing at 59°C for 30 s, extension at 72°C for 40 s; four cycles of denaturation at 96°C for 20 s, annealing at 56°C for 30 s, extension at 72°C for 30 s; five cycles of denaturation at 96°C for 20 s, annealing at 53°C for 30 s, extension at 72°C for 50 s; 25 cycles of denaturation at 96°C for 20 s, annealing at 50°C for 30 s, extension at 72°C for 60 s, and a final extension at 72°C for 5 min.

Aliquots of 5 µl of the PCR product were examined for successful amplification and concentration using gel electrophoresis in 1% agarose in 1X TBE buffer. PCR product for bacterial isolates obtained in California were sequenced by Beckman Coulter Genomics (Danvers, MA, USA), and by the John Curtain School of Medical Research for isolates obtained in Australia (Canberra, Australia). DNA sequences generated through this study were submitted to the GenBank online repository [https://www.ncbi.nlm.nih.gov/genbank/, accession numbers MG588396–MG588695 (16S rRNA); MG588696–MG588938 (nifD); and MG588155–MG588395 (nodC)].

Phylogenetic analysis

Nucleotide sequences were assembled using Geneious v. 7.1.7 and aligned with ClustalW (Larkin et al. 2007). We conducted a BLAST search of the 16S rRNA, nifD, and nodC gene sequences to verify identity as rhizobia and to compare our sequences with those available in Genbank (http://blast.ncbi.nlm.nih.gov). Cultures that were identified by the 16S rRNA gene as rhizobia were selected for further biogeographic analysis using nifD and nodC sequences. Sequences with the highest similarity values for each gene, respectively, were used in phylogenetic analyses.

We clustered 16S rRNA sequences into Operational Taxonomic Units (OTUs) using the software program Uclust (Edgar 2010) at 98% similarity. We clustered sequences into OTUs because we did not have a way to test the changes in amino acids; in other words, our goal was not to show that one protein was more efficient than another. We then blasted OTU representative sequences as well as the complete alignment against a curated alignment of 16S rRNA gene sequences from the Greengenes database (http://greengenes.lbl.gov/cgi-bin/nph-index.cgi). We also clustered nifD and nodC sequences into OTUs as above. NifD and nodC contig sequences were translated to proteins using the software Geneious v. 7.1.7 in which alignments were performed. We conducted a maximum likelihood (ML) phylogenetic analysis for each gene using RAxML software (Stamatakis 2014) with the nucleotide substitution general time reversible (GTR) model for 16S rRNA; for nifD and nodC analyses we used GTRCAT with two partitions (partition 1, first and second codon; partition 2, third codon). Support for ML analysis was estimated using 1000 bootstrap replicates; for nifD and nodC phylogenetic trees were obtained under the PROTGAMMAWAG model of amino acid evolution with Azorhizobium doebereinerae as the designated outgroup taxon. We used FigTree software v. 1.4.2 to edit the phylogenetic tree.

Statistical analysis

We analyzed data obtained from the 16S rRNA, nifD, and nodC genes to determine whether there was a difference in rhizobial strains associating with different Acacia host species between native and novel continents and invasiveness categories using principal coordinate analysis (PCoA) with an unweighted UniFrac distance matrix. We used the R statistical package “phyloseq” (McMurdie and Holmes 2013) to conduct ordinations. We tested for differences in the rhizobial species complex between continents and among invasiveness categories using PerManova. Ordination analyses were conducted using the R statistical programming language version 3.2.0 (R Core Team 2015), and PerManova analyses were conducted using QIIME (Caporaso et al. 2010).

We examined the provenance of rhizobia associating with acacias in Australia and California using analysis of molecular variance (AMOVA) (Excoffieret al. 1992). These analyses were conducted using the software Arlequin ver. 3.5 (Excoffier and Lischer 2010). Pairwise FST values were calculated using 10000 permutations with a Bonferroni adjustment for multiple comparisons.

Results

Culture and molecular identification of nodulating bacteria

We obtained a total of 304 isolates from the four Acacia species examined in this study that were identified as nitrogen-fixing root nodule bacteria by the 16S rRNA gene analysis. Additional reference sequences (26) were obtained from rhizobia nodulating with or in soil directly adjacent to other Acacia species native to Australia. Isolates obtained corresponded to nine OTUs based on the 16S rRNA gene; two OTUs were from reference rhizobial strains previously collected in Australia, while the remaining seven were from strains isolated from Acacia species that were the focus of this study. Phylogenetic analysis of the 16S rRNA gene showed that most isolates clustered with rhizobia in the genus Bradyrhizobium (OTU1, OTU6, and OTU9, 285 isolates; Fig. 3). We also found isolates clustering with the genera Mesorhizobium (OTU4, three isolates), Rhizobium (OTU2, OTU5 and OTU8, 29 isolates), Sinorhizobium (OTU3, 10 isolates), and Phyllobacterium (OTU7, three isolates; Fig. 3).

Figure 3.

Phylogenetic tree based on the 16S rRNA gene. Isolates are clustered into OTUs (sequences with 98% similarity or higher were considered the same OTU). Species highlighted in gray indicate rhizobia collected from nodules in this study. Final optimized ML = −10820.69.

nifD and nodC sequence diversity

Amplification of the nifD gene was successful in a subset of isolates identified by the 16S rRNA gene as rhizobia (242 isolates total: 133 from Australia and 109 from California). The nifD sequences were clustered into 19 OTUs. The majority of isolates were identified as a nifD gene amplified from Bradyrhizobium sp.1 (∼93% total; ∼94% CA; ∼92% AU), a slow-growing rhizobium commonly found associating with acacias in Australia (Lafay and Burdon 1998, 2001). The remainder of isolates were identified as Rhizobium (∼7% total; ∼6% CA; ∼8% AU).

Amplification of the nodC gene was successful in a subset of rhizobial isolates (240 isolates total: 129 from Australia and 111 from California). Isolates clustered into 26 OTUs. As with nifD, the majority of isolates were identified as a nodC gene amplified from Bradyrhizobium sp. 1 (∼98% total; ∼97% CA; ∼98% AU), with the reminder identified as Rhizobium spp. (∼2% total; 3% CA; 2% AU).

Host biogeographical pattern based on 16S, nifD, and nodC loci

We obtained 136 bacterial isolates from root nodules of acacias in Australia and 166 isolates from root nodules of acacias in California. For A. dealbata we obtained 20 (AU) and 44 (CA) isolates, respectively, for A. melanoxylon 18 (AU) and 51 (CA) isolates, for A. longifolia 45 (AU) and 41 (CA) isolates, and for A. verticillata 53 (AU) and 30 (CA) isolates. This total (302) differs from the 304 total isolates from acacias in this study because two isolates from Australia were cultured from soil near the base of A. melanoxylon plants rather than from nodules.

The most common bacterial genus isolated from both California and Australia acacia nodules was Bradyrhizobium sp.1 (OTU1; Fig. 4). This was true across all Acacia species from which we collected nodules, in both their native and introduced continents. Other rhizobial genera isolated from acacia nodules included Mesorhizobium, Phyllobacterium, and Rhizobium, although these genera were found in lower abundance for all Acacia species on both continents. Sinorhizobium was also isolated, but only from soils previously collected near acacias in Australia. We did not detect a significant geographical difference between rhizobial communities in Australia and California based on the 16S rRNA phylogeny.

Figure 4.

Rhizobial genera associated with different Acacia species from different locations based on the 16S rRNA gene. OTUs 3 and 8 not included as they were collected from reference species. Genera were determined by megaBlast comparison of sequences with GenBank online database (http://blast.ncbi.nlm.nih.gov).

Phylogenetic analysis based on the nifD gene showed a similar pattern as 16S rRNA. Isolates collected from Australia and California had the same OTUs represented on both continents (Fig. 5). Our nifD gene analysis also showed that there was few differences in OTUs between isolates collected from invasive and noninvasive acacias in California. Phylogenetic analysis based on the nodC gene showed similar patterns as 16S rRNA and nifD analyses. Thus, for the nodC gene we also did not detect a difference among continents or invasiveness categories in rhizobial OTUs (Fig. 6). This indicates that there is little geographic difference in rhizobial isolates collected from different Acacia species in Australia and California.

Figure 5.

Phylogenetic tree based on the nifD gene. Isolates are clustered into OTUs based on phylogenetic analysis (sequences with 98% similarity or higher were considered the same). Icon shape indicates the continent from which rhizobial isolates were collected (circle = Australia and square = California) and color indicates invasiveness category (gray = invasive, white = noninvasive). Codes indicate species from which isolates were collected (de = A. dealbata, lo = A. longifolia, me = A. melanoxylon, and ve = A. verticillata). C_nif indicates OTUs belonging to different clades. Final optimized ML = −4130.15.

Figure 6.

Phylogenetic tree based on the nodC gene. Isolates are clustered into OTUs based on phylogenetic analysis (sequences with 98% similarity or higher were considered the same OTU). Icon shape indicates the continent from which rhizobial isolates were collected (circle = Australia and square = California) and color indicates invasiveness category (gray = invasive, white = noninvasive). Codes indicate species from which isolates were collected (de = A. dealbata, lo = A. longifolia, me = A. melanoxylon, and ve = A. verticillata). C_nod indicates OTUs belonging to different clades. Final optimized ML = 4716.35.

Rhizobial diversity in relation to host invasiveness

While the phylogenetic analysis found no evidence of structure in relation to either geography or invasion category, A. dealbata and A. melanoxylon, both of which are invasive in California, as shown from the 16S rRNA phylogeny, did associate with a wider diversity of strains than A. longifolia or A. verticillata in their introduced ranges (Fig. 4). In Australia, all host species associated with the same number of bacterial genera except for A. longifolia, which was found associating with a greater diversity of bacterial genera in its native continent.

Bradyrhizobium was the most prominent genus associating with all Acacia species. Most of the rhizobial isolates could not be identified to species, however, in Australia, A. dealbata and A. longifolia were found associating with B. diazoefficiens. Acacia longifolia was additionally found associating with Rhizobium leguminosarum. In California, all Acacia species were found associating with B. canariense and B. japonicum. Acacia dealbata and A. melanoxylon were both found associating with P. brassicacearum. Acacia melanoxylon was also found associating with R. leguminosarum and R. miluonense.

We did not detect a difference in the biogeographic structure of rhizobial strains collected from Acacia species between California and Australia, based on either nifD (PerManova, F = 1.279, P = .272; Fig. 7A) or nodC (PerManova, F = 1.022, P = .394; Fig. 7B) PCoA analyses. In addition, we did not detect a difference in the genetic structure of rhizobial strains collected from different invasiveness categories (i.e. invasive or noninvasive in California, and native in Australia) for the nifD (PerManova, F = 1.262, P = .249) or nodC (PerManova, F = 1.864, P = .090) genes.

Figure 7.

Ordination of rhizobial strain identity and abundance for the (A)nifD and (B)nodC genes for rhizobial strains collected from Acacia species in Australia (circles) and California (triangles) derived from PCoA base on a UniFrac matrix. Rhizobial strains more similar in identity are closer together in ordination space.

Provenance of rhizobia among acacia hosts and geographic locations

We performed AMOVA to examine whether there was a difference in nucleotide sequences of the 16S rRNA, nifD, and nodC genes in rhizobia collected from acacias in Australia and California that vary in invasiveness in their introduced region. For all host species examined (regardless of invasive status), provenance did not influence rhizobial genetic variation (F_ST values ranged from −0.007 to 0.183; Table 1).

Table 1.

AMOVA examining genetic differences in rhizobial symbionts of Acacia species between Australia and California. F_ST and significance levels are shown for three genes, 16S rRNA, nifD, and nodC. *P < .05 and^**P < .001.

Category	Host	16 s rRNA	nifD	nodC
Invasive	A. dealbata	−0.007	0.026	0.032
Invasive	A. melanoxylon	0.039	0.161**	0.064
Noninvasive	A. longifolia	0.183**	0.073	0.049
Noninvasive	A. verticillata	0.007	0.139**	0.154**

Discussion

The goal of this study was to examine whether invasive and noninvasive Acacia species were cointroduced to California with their native Australian rhizobial symbionts, and the implication for host establishment and their ability to become invasive. We predicted that invasive and noninvasive acacias would harbor different rhizobial communities. In fact, our results suggest that all the acacias examined here were cointroduced with their Australian rhizobial symbionts, as evidenced by the lack of geographic structure with respect to variation in nifD and nodC genes found among isolates collected across both ranges, and regardless of acacia invasiveness category. This suggests that cointroduction of rhizobial symbionts does not provide an advantage for invasive compared to noninvasive Acacia species introduced to California.

Diversity of nodulating bacteria

Based on the 16S rRNA gene, acacias in both California and Australia associate with a diversity of rhizobia from several genera. In both ranges, however, Acacia species tested were found primarily associating with rhizobia from the genus Bradyrhizobium. These results align with previous studies showing that, especially in Australia, acacias are primarily nodulated by Bradyrhizobium spp. (Lafay and Burdon 1998, 2001). Similar results have been found for acacias introduced to multiple other regions of the globe, such as Portugal (Rodríguez‐Echeverría 2010) and South Africa (Ndlovu et al. 2013). We also found Acacia species associating with Rhizobium spp., a faster-growing rhizobial genus nodulating acacias in their native continent (Barnet and C. 1991).

Cointroduction and host promiscuity with rhizobia in native and novel ranges

This study provides evidence for the cointroduction of acacias with their native rhizobial mutualists to California, regardless of invasive status. Previous research has shown similar patterns. Keet et al. (2017) found no difference in rhizobial diversity or community composition between localized and widespread introduced acacias in South Africa. Rodríguez‐Echeverría (2010) found evidence that A. longifolia, which was introduced to Portugal and has become invasive in dune ecosystems of this region, was cointroduced with its native Australian rhizobial symbionts. In addition, she found two locally native legumes co-occurring with A. longifolia and being nodulated by beneficial rhizobial symbionts containing nifD and nodA genes of Australian origin, suggesting that native legumes in this region may also benefit from rhizobia cointroduced with A. longifolia (Rodríguez‐Echeverría 2010). Crisóstomo et al. (2013) found further evidence for the cointroduction of Australian acacias with their native rhizobial mutualists, showing that A. saligna was likely cointroduced to Portugal with rhizobia of Australian origin, particularly those in the genus Bradyrhizobium (Crisóstomo et al. 2013). Cointroductions of acacias with their native rhizobia appear to be geographically widespread. Ndlovu et al. (2013) found that A. pycnantha, which has become invasive in South Africa, associated with native rhizobial strains as well as those of Australian origin.

Our results indicate that acacias introduced to California, whether invasive or noninvasive, were cointroduced with their native rhizobial mutualists. These results agree with a recent review by Wandrag et. al (2020), which indicated that acacias introduced to multiple parts of the globe can form effective rhizobial associations, regardless of invasive status. It may also be possible that individual Acacia species were introduced to California with their native symbionts, and that other Acacia species have been able to subsequently associate with strains cointroduced with those species. A large-scale review of more than 3500 legume species showed evidence that symbiotic legumes are less likely to establish in novel regions than nonsymbiotic legumes. This suggests that the ability to find suitable symbionts is key for legume establishment abroad, although it may not determine which species subsequently become invasive (Simonsen et al. 2017). The association of acacias with their native rhizobial symbionts, while likely promoting establishment and colonization of acacias, does not appear to be the sole mechanism determining which plant species become invasive when introduced abroad. Since both invasive and noninvasive acacias have the benefit of their native symbionts in California there are likely additional, or perhaps different mechanisms influencing the invasive success of certain Acacia species in this region.

Host promiscuity with respect to rhizobial symbionts may be a mechanism that differentiates invasive status of acacias introduced abroad. More promiscuous acacia hosts can effectively associate with a wider range of beneficial rhizobial symbionts. Within Australia, Thrall et al. (2000) showed that Acacia species with wider distributions are more promiscuous hosts, suggesting that their ability to associate with larger numbers of rhizobial symbionts may influence their ability to colonize a wider native range. Klock et al. (2015) found that acacias invasive in multiple regions of the globe are more promiscuous hosts of rhizobial symbionts in that more positive growth responses across a greater number of rhizobial strains were observed for invasive Acacia species than for naturalized or noninvasive acacias. However, contrary results have also been found, with acacias varying in invasiveness within California displaying similar levels of host-promiscuity (Klock et al. 2016). Based on the phylogenetic analysis based on the 16S rRNA gene that we conducted, there was a difference in host promiscuity regarding the number of rhizobial strains with which acacias introduced to California associate. Acacia dealbata and A. melanoxylon, both invasive in California, associated with a greater number of rhizobial strains in their introduced range than A. longifolia and A. verticillata, which are not invasive in California. In Australia, all Acacia species examined associated with similar or lower numbers of strains than in California, with the exception of A. longifolia, which was found associating with three more rhizobial species in Australia than in California. The capacity to associate with a greater diversity of strains when introduced abroad may influence invasion success when introduced to novel ranges. Our results also show that A. longifolia is one of the most promiscuous species in its home range, associating with the widest diversity of rhizobial strains out of all Acacia species examined in Australia. Although not currently invasive in California, A. longifolia has become invasive in eight regions of the globe, as designated by Richardson and Rejmánek (2011) [later updated by Rejmánek and Richardson (2013)]. The ability of A. longifolia to associate with multiple rhizobial strains in its home range, coupled with evidence of cointroduction with its native symbionts, as well as records of invasiveness in multiple other regions of the world, make this species a prime contender for future invasive status, and one that should be managed before populations expand beyond their current distribution in California.

We collected nodules from acacia populations within their native and introduced ranges. Despite limitations in collection breadth, we found that two key genes involved in symbiosis, nifD and nodC have been transferred from Australia to California. These genes are important in nitrogen-fixation (nifD) and host specificity (nodC; Barrett et al. 2016). Thus, the cointroduction of these rhizobial genes with their host and their ability to establish in novel regions likely provides a benefit to acacias that form a symbiosis with rhizobia possessing their native genetic signature. Future studies may benefit from expanding the number of populations as well as the number of Acacia species collected from, to determine whether all species introduced to California are found in association with rhizobia of Australian origin.

Acacias are important species economically, environmentally, and culturally in their native range. Many Acacia species are used to restore degraded habitats, and a great deal of effort has gone into determining the best rhizobial strains to use as seed inoculants, thereby reducing the need for nitrogenous fertilizers that may be harmful to the environment (Thrall et al. 2005). As important as these species are in their native range, many have become ecologically detrimental in the novel ranges where they have been introduced. For example, Australian acacias have changed the community structure of native habitats (van Wilgen et al. 2011), altered soil chemistry that has led to the invasion of weedy grasses (Yelenik et al. 2004), and outcompeted native vegetation in sensitive dune habitats (Rodríguez-Echeverría et al. 2012). In California, where 16 Australian Acacia species have been introduced, multiple species appear to be expanding their ranges, two of which are recognized as invasive in this region. From a global perspective, it is, therefore, key to determine the mechanisms that enable certain Acacia species to become invasive, so that we can use this information to better prevent their further introduction, manage current invasive populations, and make more nuanced choices when species are introduced abroad for plantation forestry or other purposes.

To better understand mechanisms driving Acacia invasions it is essential to study aspects of the biology of these species in their native and introduced ranges (Hierro et al. 2005). Here, we evaluated an important mutualistic relationship, the acacia–rhizobia symbiosis, in areas where these species are native and introduced. By studying these species in both their native and introduced ranges we were able to develop a more complete understanding of the role of the acacia–rhizobia symbiosis in acacia invasiveness. In addition, comparing closely related species that differ in their invasive capacity allowed us to assess whether or not the cointroduction of native rhizobial symbionts was limited to species that have become invasive. Although we did not find that association with native rhizobial symbionts differs between our focal invasive and noninvasive acacia species, it may still be an important mechanism contributing to the current establishment and colonization of other invasive species, and the potential future invasion of species that are currently naturalized or noninvasive. Continued investigation of the mechanisms that do and do not influence invasiveness of acacias is important for the control and management of these species abroad.

Acknowledgements

We thank Mohammad S. Hoque and Kristy Lam for assistance and help in the laboratory. We thank Meredith Blackwell, James T. Cronin, Hallie Dozier, Bret Elderd, and Richard Stevens for helpful feedback and support throughout the course of this project. We thank one anonymous reviewer for helpful comments that improved the manuscript.

Conflicts of interest statement

None declared.

Funding

This work was supported by funding from the National Science Foundation Division of Environmental Biology (DEB) (grant number DEB-1311290).