Southeastern Naturalist
147
H.E. Dorman and L.E. Wallace
22001199 SOUTHEASTERN NATURALIST 1V8o(1l.) :1184,7 N–1o6. 41
Diversity of Nitrogen-Fixing Symbionts of Chamaecrista
fasciculata (Partridge Pea) Across Variable Soils
Hanna E. Dorman1 and Lisa E. Wallace2,*
Abstract - We evaluated whether geographic distance and soil characteristics influence
genetic structure of nitrogen-fixing bacterial symbionts associated with the host plant
Chamaecrista fasciculata (Partridge Pea). We tested phylogeographic clustering and associations
between genetic distance, geographic distance, and soil variables using sequences
of 2 bacterial genes and soil chemistry across 23 sites in Mississippi. We identified rhizobia
isolated from Partridge Pea as Bradyrhizobium. We detected significant genetic structure at
a regional level, and determined that rhizobia within each region were more phylogenetically
related than expected. Significant correlation between genetic distance and distances
based on soil chemistry suggests environmental influences on rhizobia diversity. High levels
of diversity among rhizobia over small spatial scales suggest that symbionts respond to local
factors. Understanding geographic diversity in natural assemblages of rhizobia aids in
predicting how hosts and symbionts respond to environmental per turbations.
Introduction
Approximately 88% of legume species are known to form symbiotic relationships
with nitrogen-fixing bacteria known as rhizobia (Graham and Vance 2003).
The great diversity of legume taxa, estimated at 20,000 species (Cronk et al. 2006),
coupled with their presence in many different habitats around the world (Dolye
and Luckow 2003, Yahara et al. 2013), may indicate that symbioses with rhizobia
have contributed to promoting diversification across this family (Martínez-Romero
and Caballero-Mellado 1996, Sprent 2009). Béna et al. (2005) showed that some
Medicago (medick) species capable of forming symbioses with multiple rhizobia
strains have larger ranges, yet other medicks evolved toward highly specialized
relationships with few rhizobia as a result of lower fitness associated with hosting
numerous symbionts. Characterization of diversity of rhizobia within and across
host species in natural ecosystems is key to understanding the influence of rhizobia
on legume diversification and the impact of such symbioses on ecosystem functioning
historically and in the face of climate and land-use change s.
Numerous studies of microbial communities have found significant geographic
structure in diverse environments and across diverse bacterial taxa (O’Malley
2008, Papke et al. 2003, Rout and Callaway 2012, Staley and Gosink 1999,
Whitaker et al. 2003). The composition of free-living soil bacterial communities
can be due to abiotic factors (Pasternak et al. 2013, Xiong et al. 2012) as well as
1 Department of Biology, University of Massachusetts, 611 North Pleasant Street, Morrill
Science Center, RM 427, Amherst, MA 01007. 2Department of Biological Sciences, Old
Dominion University, Mills Godwin Building 110, Norfolk, VA 23529. *Corresponding
author - lewallac@odu.edu.
Manuscript Editor: Richard Baird
Southeastern Naturalist
H.E. Dorman and L.E. Wallace
2019 Vol. 18, No. 1
148
impacts from other soil organisms (Djigal et al. 2004, Toljander et al. 2007) and
plants, particularly through root exudates (Marschner et al. 2004). Among abiotic
factors, climate, precipitation, organic matter, and soil texture have been found to
influence soil microbial biogeographic patterns, but the most prominent variable
influencing soil microbial diversity may be soil pH (Chong et al. 2012, Fierer and
Jackson 2006, Griffiths et al. 2011, Lauber et al. 2008).
Rhizobia have intimate relationships with their plant hosts; thus, biogeographic
patterns of symbiotic rhizobia may reflect selection by host species at the centers of
origin and associated diversification with hosts across landscapes (Martínez-Romero
and Caballero-Mellado 1996). As new legume species evolve and species expand
their distributions, plants may maintain original relationships with symbionts if rhizobia
are widespread, or they may evolve the ability to recognize and associate with
new rhizobia types. While rhizobia and legume hosts may not be strictly co-evolving,
plant hosts could influence biogeographic patterns of rhizobia by selecting for
certain genotypes. For example, Sachs et al. (2009) found that plant-host identity
significantly explained nodulating rhizobia diversity, as host plants were infected
by a small subset of rhizobia available in the soil. Thus, geographic structure in
symbiotic rhizobia may also be expected, but the relative strength of environment
versus host species in determining biogeographic patterns of the symbionts remains
understudied across diverse host species and ecosystems.
In this study, we characterized genetic diversity of rhizobia symbionts of a common
and widespread legume, Chamaecrista fasciculata (Michx.) Greene (Partridge
Pea), which is an important species in many natural ecosystems because it provides
cover, nectar, and pollen for animals. It has also been of interest in agricultural systems,
for example in crop rotation to enhance soil nitrogen (Reeves 1994) and to
manage root-knot nematodes (Rodríguez-Kábana et al. 1995). Given its annual habitat,
herbaceous growth form, and phylogenetic position as one of the only nodulating
Caesalpinioids, there is growing interest in developing Partridge Pea as a model for
studies of legume evolution (Singer et al. 2009). Partridge Pea has a geographic range
that extends from Minnesota to the Gulf of Mexico and from the east coast of the US
to New Mexico (USDA–NRCS 2015). Plants grow in open habitats such as prairies,
bluffs, riverbanks, and upland woods, and can grow in a variety of soils (Pullen 1963,
USDA–NRCS 2015). Phenotypic differences in Partridge Pea have been noted in
plant morphology at local scales (e.g., Pullen 1963; L. Wallace, pers. observ.; Weakley
2012) and among widely separated populations (e.g., Galloway and Fenster 2000,
Henson et al. 2013), suggesting great potential for locally adapted populations. The
morphological and ecological variation exhibited by this species has been recognized
by some taxonomists to represent distinct taxa (e.g., Weakley 2012), but these
have not yet been shown to be genetically differentiated or reproductively isolated.
Whether geographic structure extends to rhizobia symbionts is unclear. Thus, in this
study, we tested 2 hypotheses: (1) nodulating rhizobia of Partridge Pea are geographically
structured, and (2) geographic structure is associated with variation in soil
composition. We expected phylogenetic clustering of rhizobia strains by their physiographic
region, which are partially defined by soil type. Furthermore, we expected
Southeastern Naturalist
149
H.E. Dorman and L.E. Wallace
2019 Vol. 18, No. 1
that soil pH would be associated with rhizobia diversity, as pH often shapes the
structure of free-living microbial communities in soils (Fierer and Jackson 2006). We
characterized phylogenetic diversity using a housekeeping gene and a gene on the
symbiosis island to evaluate whether horizontal gene exchange (HGE) may influence
local community structure of the rhizobia, as reported for other symbiotic Bradyrhizobium
(Parker 2012, Parker and Rousteau 2014).
Field-site Description
We selected sites across the physiographic regions of Mississippi (Fig. 1) to
represent variation in soil habitat. The sampling design was intended to capture
variation in nodulating rhizobia at a regional geographic scale reflecting variation
in plant communities and soils, rather than to characterize complete rhizobia
diversity at individual sites or across the entire range of the host species. We
sampled rhizobia in nodules of Partridge Pea from June to July 2013 from a total
of 23 locations in the Blackbelt Prairie (n = 4), Tombigbee Hills (n = 3), North-
Central Hills (n = 3), Loess Hills (n = 3), Delta (n = 3), Jackson Prairie (n = 3), and
South-Central Hills (n = 4) (Table 1; Fig. 1; see Table S1 in Supplemental File 1,
available online at http://www.eaglehill.us/SENAonline/suppl-files/s18-1-S2494-
Wallace-s1, and for BioOne subscribers, at https://dx.doi.org/10.1656/S2494.s1).
Figure 1. Sample locations of
Partridge Pea within the designated
physiographic regions of
Mississippi.
Southeastern Naturalist
H.E. Dorman and L.E. Wallace
2019 Vol. 18, No. 1
150
We excavated roots from individual flowering plants randomly selected at each site
across the area of growth of the species. We did not quantify or consider plant size
or herbivory when selecting plants to sample. Sampling continued until we had
collected roots with at least 1 nodule from a minimum of 12 plants per site; we did
not sample plants growing immediately adjacent to a sampled plant. We placed each
sampled root with its nodule(s) in a distinct bag to reduce contamination by bacteria
from other plants. At each sampling location, we collected a voucher-plant specimen,
soil sample from the top 25 cm within an area of extant host plants, and GPS
coordinates (see Table S1 in Supplemental File 1, available online at http://www.
eaglehill.us/SENAonline/suppl-files/s18-1-S2494-Wallace-s1, and for BioOne
subscribers, at https://dx.doi.org/10.1656/S2494.s1). We deposited plant vouchers
in the Mississippi State University (MISSA) herbarium (Mississippi State, MS).
Upon return to the laboratory, we stored roots with nodules at -80 °C until processing
for DNA extraction.
Methods
Genetic analysis
We were unable to obtain high-quality DNA or sequence data for some sampled
plants. Thus, the final data-set included sequences collected from 117 distinct
plants and 183 nodules (see Table 1 for sample sizes in each physiographic region).
Rhizobia DNA from 1 or more nodules for 6 plants per collection site were
extracted and diluted in 200 μl buffer using the Qiagen DNeasy plant Mini Kit
(Qiagen, Valencia, CA). Prior to grinding, we surface-sterilized nodule samples in
a 1% bleach solution for 5 min and then washed them in sterile water for 5 min. We
placed nodules in 70% ethanol for 5 min, followed by a final 5-min sterile-water
wash. We characterized rhizobia diversity across the sample sites based on partial
Table 1. Genetic diversity of Bradyrhizobium species across physiographic regions of Mississippi.
n = number of clean, readable sequences generated; S = segregating sites; H = number of unique
sequences; Hd = haplotype diversity; π = nucleotide diversity; BBP = Blackbelt Prairie; TH = Tombigbee
Hills; NCH = North-Central Hills; LH = Loess Hills; D = Delta; JP = Jackson Prairie; and SCH
= South-Central Hills.
BBP TH NCH LH D JP SCH
truA
n 36 38 29 17 18 13 32
S 123 173 87 116 96 73 89
H 19 25 12 9 12 9 19
Hd 0.954 0.950 0.894 0.831 0.948 0.936 0.944
π 0.1235 0.1103 0.0543 0.0817 0.1193 0.1127 0.0846
nifH
n 35 28 27 16 18 10 27
S 113 101 86 70 46 67 53
H 19 16 14 8 8 7 12
Hd 0.923 0.947 0.875 0.808 0.752 0.867 0.872
π 0.0757 0.0775 0.0270 0.0493 0.0513 0.0484 0.0576
Southeastern Naturalist
151
H.E. Dorman and L.E. Wallace
2019 Vol. 18, No. 1
sequences of a symbiosis-island gene, nifH, and a housekeeping gene, truA. The
nifH gene is involved in nitrogen fixation (Laguerre et al. 2001), and the truA gene
is involved in translation and ribosomal biogenesis (Ahn et al. 2004). Both of these
markers have been used by others to characterize rhizobia diversity (Vinuesa et al.
2005, Zhang et al. 2012). We amplified and sequenced nifH using primers outlined
in Vinuesa et al. (2005), whereas truA was amplified and sequenced using primers
from Zhang et al. (2012). We employed PCR to amplify the regions in 12.5-μl volumes
containing 1.5 μl DNA, 1x LongAmp buffer (New England Biolabs, Ipswich,
MA), 0.8% DMSO, 1.5 U LongAmp Taq (New England Biolabs), 0.32 mM dNTPs,
0.4 μM forward primer, and 0.4 μM reverse primer. For both genes, we heated the
reaction tubes to 95 °C prior to the addition of DNA. The nifH program consisted
of denaturation at 95 ºC for 3.5 min, 30 cycles of 93.5 ºC for 1 min, 58 ºC for 1
min, 72 ºC for 1 min, and an elongation step of 72 ºC for 5 min. truA required a
touchdown thermal-cycler program as follows: denaturation at 95 ºC for 5 min, 11
cycles of 94 ºC for 45 sec, 60 ºC for 1 min decreased by 1.0 ºC per cycle, 72 ºC
for 1 min, 26 cycles of 94 ºC for 45 sec, 50 ºC for 1 min, 72 ºC for 1 min, and an
elongation step of 72 ºC for 10 min. We determined amplification of PCR products
by agarose gel electrophoresis and ethidium bromide staining. We included a
negative control with each set of reactions to check for contamination. We cleaned
PCR products using 0.2x Antarctic Phosphatase buffer (New England Biolabs), 5
units Exonuclease I (New England Biolabs), and 1.25 units Antarctic Phosphatase
followed by cycle sequencing in 10-μl reactions using forward and reverse primers
and Big Dye version 3.1 (Life Technologies, Carlsbad, CA). We dried and sent
sequenced samples to Arizona State University DNA Lab for capillary electrophoresis.
Forward and reverse sequences were edited and assembled into a consensus
sequence using Sequencher version 4.7 (Gene Codes Corporation, Ann Arbor, MI).
We employed the AUTO option in MAFFT (Katoh et al. 2005) provided by the
Computational Biology Research Center (http://mafft.cbrc.jp/alignment/server/) to
align sequences. We deposited sequences in GenBank (NCBI 1988) as accessions
KR186321–KR186443.
We determined sequence variation for each gene in each of the 7 geographic
regions using DNAsp v. 5 (Librado and Rozas 2009) by calculating number of variable
sites (S), number of haplotypes (H), haplotype diversity (Hd), and nucleotide
diversity (π). We conducted analyses of molecular variance (AMOVA; Excoffier
et al. 1992) to quantify the distribution of variation within and across geographic
regions for each gene; analyses were performed in Arlequin v. 3.5 (Excoffier and
Lischer 2010) using pairwise distance of DNA sequences and designating sequences
by physiographic region of origin. We generated phylogenetic trees for each of the
genes to assess evolutionary relatedness of the strains. We downloaded reference
sequences of taxonomically valid Bradyrhizobium taxa from GenBank (NCBI
1988) for use in the analyses to place the newly collected sequences in the broader
context of known Bradyrhizobium diversity. For truA, the reference sequences
included B. canariense (BTA1), B. elkanii (USDA 76), B. japonicum (USDA 6),
B. liaoningense (USDA 3622), and B. yuanmingense (CCBAU 10071). For nifH,
Southeastern Naturalist
H.E. Dorman and L.E. Wallace
2019 Vol. 18, No. 1
152
reference sequences included B. canariense (BES1, BC-C2), B. elkanii (USDA 46,
USDA 94), B. japonicum (DSMZ 30131, USDA 122, X3-1, X6-9, Nep1, Blup-
MR1, FN13), B. liaoningense (LMG 18230, Spr 3-7), B. yuanmingense (CCBAU
10071), and unidentified Bradyrhizobium isolated from Partridge Pea (cf1, cf4,
cfhr1a, cfrr1) or C. nictitans (L.) Moench (Sensitive Patridge Pea) (cnw15) by
Parker (2012). We subjected each aligned data set to an assessment by jModeltest2
(Darriba et al. 2012, Guindon and Gascuel 2003) to determine the best-fitting
model of molecular evolution under the BIC. We conducted these analyses using
the Cipres Science Gateway (Miller et al. 2010). We selected TrN+I+G (Tamura and
Nei 1993) as the best model for the nifH data set, and HKY+G (Hasegawa et al.
1985) as the best model for the truA sequences. We implemented GTR+I+G (Tavaré
1986), the closest model to TrN+I+G that can be used in MrBayes, for nifH and
HKY+G for truA in independent Bayesian phylogenetic analyses using MrBayes
v. 3.2.3 (Ronquist et al. 2012) in the Cipres Science Gateway (Miller et al. 2010).
For each analysis, we conducted Markov Chain Monte Carlo (MCMC) for 5 million
generations with a sampling frequency of 1000, after which the split standard
deviation was less than 0.006. We set a burn-in of 1250 trees prior to determining
the posterior probability of the trees with the highest likelihood. Consensus trees
are reported with posterior probability indicated as support for clades. We used
Phylocom 4.0.1 (Webb et al. 2008) to evaluate if rhizobia sequences were phylogenetically
clustered by physiographic region. For each region, we calculated mean
phylogenetic distance (MPDsample ; Webb et al. 2008) across all pairs of bacterial sequences,
and determined significance by comparing MPDsample to a null distribution
inferred from 1000 random permutations of sequences across the tips of the tree.
We calculated a standardized measure of clustering across unequal sample sizes, the
net relatedness index (NRI), as the difference between MPDsample and the MPDnull,
divided by the standard deviation of MPDnull (Webb et al. 2008). We conducted
these analyses independently on the data sets.
Soil analysis
Soil samples were allowed to air dry for 20 d before we ground them using
a mortar and pestle. We removed any visible organic fragments >5 mm prior to
grinding. We sent the soil samples to the University of Arkansas for analysis of
pH, nitrate (NO3), ammonium (NH4), phosphorus, potassium, magnesium, sulfur,
sodium, iron, manganese, zinc, copper, and boron. We used these variables to assess
possible soil factors affecting rhizobia assemblages. We employed Spearman’s
non-parametric correlation (Spearman 1907) to identify redundant soil variables in
the dataset based on a correlation coefficient of 0.6 or higher between any 2 variables.
We conducted correlation analyses in SPSS v. 21 (IBM Corporation 2012).
Boron and magnesium exhibited significant correlation coefficients greater than 0.6
with several other variables; thus, we eliminated them from further analysis. Boron
levels were correlated with levels of nitrate, calcium, zinc, and copper. Magnesium
and potassium levels were correlated. We tested spatial autocorrelation of soils using
a Mantel test (Mantel 1967) between a composite-soil distance and geographic
Southeastern Naturalist
153
H.E. Dorman and L.E. Wallace
2019 Vol. 18, No. 1
distance in PASSaGE 2 (Rosenberg and Anderson 2011). We generated the geographic-
distance matrix using the linear distance between each sampling site and
the haversine formula (Sinnott 1984). We used a squared Euclidian distance with
a Z-score variance correlation in SPSS v. 21 (IBM Corporation 2012) to calculate
a composite pairwise distance-matrix of sites based on all soil variables except
boron and magnesium. We assessed significance of the correlation in the Mantel
test with 999 permutations in a 1-tailed test, and defined significant correlations
as P < 0.05. We also examined associations of soil properties and geography with
genetic variability of rhizobia using partial Mantel tests (Smouse et al. 1986) in
comparisons of: (1) genetic distance vs. a composite distance of soil variables while
controlling for geographic distance, (2) genetic distance vs. distance based on soil
pH while controlling for geographic distance, (3) genetic distance vs. geographic
distance while controlling for soil characteristics. We tested soil pH separately
from soil-mineral variables because previous studies suggested pH as the factor
most significantly affecting soil bacterial assemblages (e.g., Fierer and Jackson
2006). We also calculated the matrix of soil pH distances in SPSS using squared
Euclidean distance and Z-score variance. We determined pairwise genetic distance,
calculated as p-distance, between sample sites using MEGA v. 6 (Tamura et al.
2013). We conducted Mantel tests for each genetic data set and the geographic- and
environmental-distance matrixes independently using PASSaGE 2 (Rosenberg and
Anderson 2011) as described above. We excluded site R63 from the Mantel tests
due to the small number of sequences collected from this site.
Results
Genetic diversity and phylogenetic patterns
In total, we generated 161 nifH sequences and 183 truA sequences from rhizobia
in nodules that were successfully sequenced. Sequences generated in this project
most closely matched Bradyrhizobium sequences in BLAST-n searches with
GenBank (NCBI 1988). The aligned lengths were 706 nucleotides for nifH (~69%
coverage of the gene in Bradyrhizobium) and 497 nucleotides for truA (~67% coverage
of the gene in Bradyrhizobium). We found high levels of genetic diversity
in each of the data sets. Nucleotide diversity was higher in truA compared to nifH
for all geographic regions. Haplotype diversity varied from 0.831 to 0.954 for truA
and from 0.752 to 0.947 for nifH. Regions for which we obtained more samples
exhibited a greater number of haplotypes and increased haplotype diversity in nifH,
but differences in sampling intensity were not apparent in the number of segregating
sites or nucleotide diversity for either gene or for haplotype diversity in truA.
Diversity measures for each of the regions are reported in Table 1.
The diversity of Bradyrhizobium symbionts found in Mississippi encompasses
multiple recognized species in this genus, as evidenced by the phylogenetic
clustering of newly collected samples with reference sequences of B. elkanii, B. japonicum,
B. liaoningense, and B. yuanmingense (Figs. 2, 3). The topologies of
the phylogenies both depict 2 large, well-supported clades (labeled as A and B in
Figs. 2, 3; posterior probability > 0.99), with several smaller clades and singleton
Southeastern Naturalist
H.E. Dorman and L.E. Wallace
2019 Vol. 18, No. 1
154
strains clustering external to these clades. The identities of clades A and B are
relatively conserved between the data sets, but some samples are phylogenetically
incongruent between the trees (see Figure S1 in Supplemental File 1, available online
at http://www.eaglehill.us/SENAonline/suppl-files/s18-1-S2494-Wallace-s1,
and for BioOne subscribers, at https://dx.doi.org/10.1656/S2494.s1).
Figure 2. Phylogram
of sampled
Bradyrhizobium
strains and reference
sequences
based on variation
in truA.
Strains have been
grouped into
clades that exhibit
strong support.
Samples of outgroups
and related
taxa downloaded
from GenBank
are indicated on
the tree or beside
the clade in
which they clustered.
Pie charts
for clades A and B
indicate proportional
make-up
from each of the
designated physiographic
regions
as follows: BBP
= Blackbelt Prairie;
TH = Tombigbee
Hills; NCH
= North-Central
Hills; LH = Loess
Hills; D = Delta;
JP = Jackson Prairie;
SCH = South-
C e n t r a l H i l l s .
Support values
are indicated on
branches by asterisks
(** >95%
posterior probability).
Southeastern Naturalist
155
H.E. Dorman and L.E. Wallace
2019 Vol. 18, No. 1
The majority of hosts with multiple nodules exhibited rhizobia sequences that
were less than 10% divergent, although some plants did contain divergent strains.
Considering greater than 6% difference in truA to signal interspecific differences
(based on the assessment by Zhang et al. [2012] that included truA sequences), we
identified 23 plants in our data set that contained highly divergent rhizobia reflecting
species-level differences among Bradyrhizobium. Considering greater than 5%
difference as a signal of species-level differences in nifH (Gaby and Buckley 2012),
we identified 11 plants containing different species of symbionts. Eight plants
Figure 3. Phylogram of sampled Bradyrhizobium strains and reference sequences based on
variation in nifH. Strains have been grouped into clades that exhibit strong support. Samples
of outgroups and related taxa downloaded from GenBank are indicated on the tree or beside
the clade that they clustered in. Pie charts for clades A and B indicate proportional make-up
from each of the designated physiographic regions as follows: BBP = Blackbelt Prairie, TH
= Tombigbee Hills, NCH = North-Central Hills, LH = Loess Hills, D = Delta, JP = Jackson
Prairie, SCH = South-Central Hills. Support values are indicated on branches by asterisks
(** >95% posterior probability; * > 90% posterior probability).
Southeastern Naturalist
H.E. Dorman and L.E. Wallace
2019 Vol. 18, No. 1
156
exhibited concordance between truA and nifH data sets in having highly divergent
symbionts between nodules. These were distributed across 7 sample sites.
Analyses of sequence and phylogenetic diversity detected geographic structure
in the symbionts of Partridge Pea. AMOVA for both genes indicated significant
molecular variation among the defined physiographic regions (truA F ST = 0.134,
P < 0.001; nifH FST = 0.123, P < 0.001; Table 2). Strains also exhibited significant
phylogenetic clustering for 4 of the 7 physiographic regions in the truA dataset
and for 5 of the regions in the nifH data set (Table 3). When we observed significant
patterns, rhizobia contained gene variants that were more similar, on average,
compared to the null distribution of permuted sequences across the trees for all
cases except nifH in the Blackbelt Prairie. For this region, nifH gene variants were
significantly over-dispersed across the phylogeny relative to the null distributi on.
Comparisons of genetic and soil variation
Soils were highly variable across sampling sites, and the correlation between
geographic distance and the composite soil-distance was weak (r = -0.018, t =
-0.205, P > 0.05; Table 4), although sites are distinguishable in canonical discriminant
analysis (see Figure S2 in Supplemental File 1, available online at http://www.
eaglehill.us/SENAonline/suppl-files/s18-1-S2494-Wallace-s1, and for BioOne
Table 3. Phylogenetic clustering of Bradyrhizobium symbionts in relation to physiographic region.
n = sample size, NRI = net relatedness index, and P = level of significance. Asterisks (*) indicate
significant correlations (P < 0.05).
Region truA n NRI P nifH n NRI P
Blackbelt Prairie 36 -1.1276 0.112 35 -2.243 0.009*
Tombigbee Hills 38 1.7639 0.048* 28 -0.401 0.349
North-Central Hills 29 7.1819 less than 0.001* 27 5.614 less than 0.001*
Loess Hills 17 1.7079 0.063 16 1.974 0.012*
Delta 18 1.9522 0.042* 18 2.729 0.005*
Jackson Prairie 13 0.5532 0.233 10 1.177 0.100
South-Central Hills 32 4.9874 less than 0.001* 27 2.757 0.004*
All 183 3.7790 less than 0.001* 161 3.174 0.003*
Table 2. Analysis of molecular variance (AMOVA) of Bradyrhizobium genotypes determined by truA
and nifH sequences across the physiographic regions of Mississippi (see Fig. 1). DF = degrees of
freedom.
Sum of Variance Percent
DF squares component variation Fixation index
truA
Among regions 6 556.801 2.896 13.42 FST = 0.134; P < 0.001
Within regions 176 3289.270 18.689 86.58
Total 182 3846.071 21.585 100.00
nifH
Among regions 6 348.832 1.960 12.32 FST = 0.123; P < 0.001
Within regions 154 2148.367 13.950 87.68
Total 160 2497.199 15.911 100.00
Southeastern Naturalist
157
H.E. Dorman and L.E. Wallace
2019 Vol. 18, No. 1
subscribers, at https://dx.doi.org/10.1656/S2494.s1). Many sites within the Blackbelt
Prairie contained high levels of calcium, which is characteristic of the white
soils common to this area. Many sites in the Delta, which is highly agricultural,
contained higher levels of phosphorus and potassium. In contrast to a lack of autocorrelation
in soil variables, we identified significant associat ions between genetic
distance and distance based on soil properties. When controlling for geographic
distance, genetic distance was correlated with soil distance for both genes (truA:
r = 0.413, t = 2.465, P = 0.01; nifH: r = 0.303, t = 2.047, P = 0.049) as well as
distance based solely on pH (truA: r = 0.202, t = 1.783, P = 0.04; nifH: r = 0.366,
t = 3.550, P = 0.001). Genetic distance was not correlated with geographic distance
when controlling for variation in soil characteristics for either gene (truA r =
-0.082, t = -0.971, P = 0.20; nifH r = -0.048, t = -0.591, P = 0.320).
Discussion
Genetic diversity in relation to geography and soils
In this study, we found that rhizobia symbionts of the host plant Partridge Pea
exhibited high levels of genetic diversity and geographically relevant structure.
Our results substantially extend understanding of the diversity of native rhizobia
symbionts associated with Partridge Pea and in ecosystems that have not been
examined previously (Koppell and Parker 2012, Parker 2015, Parker and Kennedy
2006). All sequences generated in this study are most similar to other Bradyrhizobium
strains, a finding that is consistent with the exclusive use of Bradyrhizobium
by Chamaecrista (sensitive pea) host plants (Andrews and Andrews 2017). Many
Bradyrhizobium species are considered to be generalists because they nodulate
wild legume species from different genera (Ehinger et al. 2014, Koppell and Parker
2012) and some agriculturally important species that are widely planted (Appunu et
al. 2008). Our results mirror other studies that have reported diverse symbiotic partners
in widespread legume hosts, including Medicago sativa L. (Alfalfa) (Paffetti
et al. 1996), Vicia faba L. (Broad Bean) (Tian et al. 2007), and Acacia pycnantha
Benth. (Golden Wattle) (Ndlovu et al. 2013). For many of these species, it has been
suggested that enhanced diversity of symbionts may underlie the success of host
plants in non-native habitats (e.g., Ndlovu et al. 2013) and influence the evolution
Table 4. Mantel and partial Mantel tests of matrix correlations between distances based on genetic,
geographic, soil, and soil pH variables. r = correlation; t = t-statistic; P = significance level. Asterisks
(*) indicate significant correlations.
Comparison r t P
Soil vs. geographic -0.018 -0.205 0.47
truA nifH
Comparison r t P r t P
Genetic vs. geographic, soil constant -0.082 -0.971 0.20 -0.048 -0.591 0.320
Genetic vs. soil, geography constant 0.413 2.465 0.01* 0.303 2.047 0.049*
Genetic vs. soil pH, geography constant 0.202 1.783 0.04* 0.366 3.550 0.001*
Southeastern Naturalist
H.E. Dorman and L.E. Wallace
2019 Vol. 18, No. 1
158
of legume species (Zahran 2001). For Partridge Pea, we also found that symbionts
can be highly diverse among nodules on an individual plant. As many as 23
plants in this study (~20%) contained species-level differences of Bradyrhizobium
symbionts in their nodules. Intra-plant variation in rhizobia has been reported
for other widespread legume hosts, including Trifolium pratense L. (Red Clover)
(Hagen and Hamrick 1996). Given that legume success is highly dependent on the
functionality of specific rhizobia genotypes through genotype x genotype interactions
(Heath 2010), the ability to utilize diverse strains may underlie this host
plant’s ability to maintain such a wide geographic range across the e astern US.
It is now well known that free-living and symbiotic soil microbes in diverse
geographic areas frequently exhibit genetic structure (O’Malley 2008, Papke et al.
2003, Rout and Callaway 2012, Staley and Gosink 1999, Whitaker et al. 2003), but
the causes underlying this structure are not well understood. We also hypothesized
that rhizobia symbionts of Partridge Pea would be geographically structured and
that soil variables may underlie these patterns of distribution. In our study, analyses
of molecular variance by region indicated that a significant proportion of the
observed variation is distributed among the regions (Table 2). Nevertheless, we did
not find a significant correlation between geographic and genetic distance in the
Mantel test, indicating that the symbionts do not follow a strict isolation by distance
pattern. This finding is consistent with other studies demonstrating that genetic
structure of soil bacteria is largely independent of geographic distance (Fierer and
Jackson 2006). When examined by phylogenetic relatedness, our Phylocom analyses
indicated significant clustering of strains for both genes in the North-Central
Hills, Delta, and South-Central Hills, which supports the resul ts of AMOVA. Only
Jackson Prairie showed non-significant values for both genes in these analyses.
This result may be due to the small sample size for this region (10–13 sequences)
compared to the other areas. Also noteworthy, the Blackbelt Prairie strains were
more dispersed in both phylogenies than any other area, resulting in negative NRI
values for both genes (Table 3). The Blackbelt Prairie contains a mixture of habitats,
some with chalky, calcareous soils and others with rich dark soils. For this
study, we did not characterize the type of soil where individual plants were growing,
but note that we did sample in both soil types. Such a dif ference may underlie
the reason for the negative NRI values and reflect a finer scale of structure among
Bradyrhizobium symbionts if they (or their host plants) are locally adapted to very
different soil types of this region.
Relative to other genera of rhizobia, Bradyrhizobium is considered to be a generalist
group that is symbiotic with diverse host plants (Parker 2015). However,
within this diverse genus there is evidence of host specificity and geographic localization
of strains. Koppell and Parker (2012) identified superclades of B. elkanii and
B. japonicum sampled from 41 legume genera from Alaska to Panama; individuals
in these clades spanned all sampled regions, indicating little evidence of regional
endemism at this deep phylogenetic level. At finer scales, Bradyrhizobium strains
within the 2 superclades did exhibit geographical structure because distinct bacterial
strains were restricted to geographically disparate areas (Koppell and Parker
Southeastern Naturalist
159
H.E. Dorman and L.E. Wallace
2019 Vol. 18, No. 1
2012). Our results provide support for geographic localization of strains at fine
taxonomic and geographic scales. These results are also consistent with the largerscale
study by Koppell and Parker (2012) across multiple legume hosts and much
of North America.
Given that a strict isolation-by-distance pattern was not satisfactory in explaining
the phylogenetic structure of rhizobia isolated from Partridge Pea, other factors
must be considered, including traits of the environment and plant host. Diversity
and composition of soil microbial communities is often dictated by soil properties.
Among these, pH has been found to have a strong effect on soil microbial communities
(Chong et al. 2012, Fierer and Jackson 2006) because many bacteria are
limited in their ability to survive in basic or acidic soils. For certain rhizobia, soil
properties have also been found to influence their presence and diversity. For example,
highly acidic soils show less rhizobia diversity than soils where the pH has
been artificially increased with the addition of lime (Andrade et al. 2002). Multiple
studies have identified differences in soil preferences between Bradyrhizobium
and Sinorhizobium isolates and biogeographic patterns that are associated with
soil pH (Li et al. 2011, Zhang et al. 2011) as well as available N, P, and K (Zhang
et al. 2011). Soil pH and site elevation were found to correlate with diversity of
Mesorhizobium symbionts (Lemaire et al. 2014). We identified a positive association
between genetic distance and soil variation for Partridge Pea. The Mantel tests
revealed significant positive correlations between the composite soil distance and
rhizobia genetic distance for both genes, as well as distance based only on pH
and genetic distance. Given that we found wide variation in soil minerals of sites
considered to be in the same physiographic region, our sampling approach may
not have captured the continuum of environmental variation to pinpoint specific
variables of the soil and their relative influence on rhizobia diversity. Additionally,
plant host could also influence symbiont structure because host plants can select
their rhizobia partners (Hirsch et al. 2001) via signaling prior to the establishment
of nodules (Yang et al. 2010). Sachs et al. (2009) found that the rhizobia housed in
nodules were a subset of those in the surrounding soil, indicating a strong role for
plant host to choose particular rhizobia genotypes. We did not characterize rhizobia
in soil samples or host genotypes, thus these alternative factors cannot be fully
evaluated in this study. Nevertheless, given the signal we identified between genetic
distance and environmental variation, we expect that a combination of biotic
and abiotic factors is likely to dictate biogeographic patterns of Bradyrhizobium
symbionts. Additionally, given that Partridge Pea is an annual species, the diversity
of symbiotic rhizobia may vary from year to year .
Horizontal gene exchange and symbiont diversity
In previous studies on diversity of Bradyrhizobium symbionts, genes located on
the symbiosis island exhibited incongruent phylogenies when compared to those
based on housekeeping genes (Koppell and Parker 2012, Parker 2012, Parker and
Rousteau 2014, Parker et al. 2002,), suggesting strong potential for horizontal
transfer of symbiosis genes among soil microbes. We did not find overwhelming
evidence of incongruence between the phylogenies based on nifH and truA because
Southeastern Naturalist
H.E. Dorman and L.E. Wallace
2019 Vol. 18, No. 1
160
most of the samples clustered in 1 of 2 large and well-supported clades (Figs. 2, 3).
For those samples that contained mismatched placements between the trees, acquisition
of 1 of these genes may have occurred horizontally via exchange with other
Bradyrhizobium. Nevertheless, both data sets show evidence of similar geographic
structure. AMOVA indicated 12–13% of the variation is distributed among regions.
There was also evidence of phylogenetic clustering in both genes for 3 of the physiographic
regions, and Mantel tests indicated significant and positive correlations
between matrixes of genetic distance and soil variables for both genes. Based on
the similarities in phylogenetic signal of the 2 genes, we suggest that horizontal
gene exchange is not likely a significant driver of the genetic structure observed
in this study. The mismatched individuals occur across multiple sample sites and
physiographic regions, which suggests the absence of local selection for particular
variants. Although Parker and colleagues (Koppell and Parker 2012, Parker 2012,
Parker and Rousteau 2014, Parker et al. 2002) found evidence consistent with horizontal
gene-exchange among Bradyrhizobium at a broad geographic scale, they also
found that most strains (as deduced from symbiosis genes) were associated with
few host plants. Thus, within any given host plant–rhizobia symbiosis, there may be
selection for maintenance of a more limited pool of compatible symbionts. Further
studies of horizontal gene exchange in natural systems would help in understanding
its importance in generating novel symbionts and their interactions with host plants.
Partridge Pea has a wide distribution and occupies a diversity of habitats; thus, it
is an ideal system in which to investigate the breadth of the effects of geography
and environmental variables on the establishment of legume–rhizobia symbioses.
Understanding natural assemblages of rhizobia associated with native legumes in
the context of environmental heterogeneity will aid in predicting how hosts and
symbionts are likely to respond to environmental perturbations.
Acknowledgments
The authors thank Mark Welch, Ronn Altig, Gary Ervin, and Matthew Brown for
assistance in experimental design, data analyses, and editing previous versions of the manuscript.
We are also grateful to reviewers who provided comments to improve the manuscript.
This research was supported by grants from Mississippi State University and a Graduate
Student Research Award from the Botanical Society of America to HED.
Literature Cited
Ahn, K.S., U. Ha, J. Jia, D. Wu, and S. Jin. 2004. The truA gene of Pseudomonas aeruginosa
is required for the expression of type III secretory genes. Mic robiology 150:539–547.
Andrade, D., P.J. Murphy, and K.E. Giller. 2002. The diversity of Phaseolus-nodulating
rhizobial populations is altered by liming of acid soils planted with Phaseolus vulgaris
L. in Brazil. Applied and Environmental Microbiology 68:4025–4034.
Andrews, M., and M.E. Andrews. 2017. Specificity in legume–rhizobia symbioses. International
Journal of Molecular Sciences 18:705. DOI:10.3390/ijms18 040705.
Appunu, C., A. N’Zoue, and G. Laguerre. 2008. Genetic diversity of native Bradyrhizobium
isolated from soybeans (Glycine Max L.) in different agricultural–ecological–climatic
regions of India. Applied and Environmental Microbiology 74:5991–5996.
Southeastern Naturalist
161
H.E. Dorman and L.E. Wallace
2019 Vol. 18, No. 1
Béna, G., A. Lyet, T. Huguet, and I. Olivieri. 2005. Medicago–Sinorhizobium symbiotic
specificity evolution and the geographic expansion of Medicago. Journal of Evolutionary
Biology 18:1547–1558.
Chong, C., D. Pearce, P. Convey, W. Yew, and I. Tan. 2012. Patterns in the distribution of
soil bacterial 16S rRNA gene sequences from different regions of Antarctica. Geoderma
181–182:45–55.
Cronk, Q., I. Ojeda, and R.T. Pennington. 2006. Legume comparative genomics: Progress
in phylogenetics and phylogenomics. Current Opinion in Plant Bi ology 9:99–103.
Darriba, D., G.L. Taboada, R. Doallo, and D. Posada. 2012. jModelTest 2: More models,
new heuristics, and parallel computing. Nature Methods 9:772.
Djigal, D., A. Brauman, T.A. Diop, J.L. Chotte, and C. Villenave. 2004. Influence of bacterial-
feeding nematodes (Cephalobidae) on soil microbial communities during Maize
growth. Soil Biology and Biochemistry 36:323–331.
Doyle, J.J., and M.A. Luckow. 2003. The rest of the iceberg: Legume diversity and evolution
in a phylogenetic context. Plant Physiology 131:900–910.
Ehinger, M., T.J. Mohr, J.B. Starcevich, J.L. Sachs, S.S. Porter, and E.L. Simms. 2014. Specialization–
generalization trade-off in a Bradyrhizobium symbiosis with wild legume
hosts. BMC Ecology 14:8. DOI:10.1186/1472-6785-14-8.
Excoffier, L., and H.E.L. Lischer. 2010. Arlequin suite ver 3.5: A new series of programs
to perform population genetics analyses under Linux and Windows. Molecular Ecology
Resources 10:564–567.
Excoffier, L., P.E. Smouse, and J.M. Quattro. 1992. Analysis of molecular variance inferred
from metric distances among DNA haplotypes: Application to human mitochondrial
DNA restriction data. Genetics 131:479–491.
Fierer, N., and R.B. Jackson. 2006. The diversity and biogeography of soil bacterial communities.
Proceedings of the National Academy of Sciences, USA. 103:626–631.
Gaby, J.C., and D.H. Buckley. 2012. A comprehensive evaluation of PCR primers to
amplify the nifH gene of nitrogenase. PLoS One 7(7):e42149. DOI:10.1371/journal.
pone.0042149.
Galloway, L.F., and C.G. Fenster. 2000. Population differentiation in an annual legume:
Local adaptation. Evolution 54:1173–1181.
Graham, P.H., and C.P. Vance 2003. Legumes: Importance and constraints to greater use.
Plant Physiology 131:872–877.
Griffiths, R.I., B.C. Thomson, P. James, T. Bell, M. Bailey, and A.S. Whiteley. 2011. The
bacterial biogeography of British soils. Environmental Microbio logy 13:1642–1654.
Guindon, S., and O. Gascuel. 2003. A simple, fast, and accurate method to estimate large
phylogenies by maximum likelihood. Systematic Biology 52:696–70 4.
Hagen, M.J., and J.L. Hamrick. 1996. A hierarchical analysis of population genetic structure
in Rhizobium leguminosarum vs. trifolii. Molecular Ecology 5:177–186.
Hasegawa, M., K. Kishino, and T. Yano. 1985. Dating the human–ape splitting by a molecular
clock of mitochondrial DNA. Journal of Molecular Evolution 22:160–174.
Heath, K.D. 2010. Intergenomic epistasis and co-evolutionary constraint in plants and rhizobia.
Evolution 64:1446–1458.
Henson, T.M., W. Cory, and M.T. Rutter. 2013. Extensive variation in cadmium tolerance
and accumulation among populations of Chamaecrista fasciculata. PLoS One
8(5):e63200. DOI:10.1371/journal.pone.0063200.
Hirsch, A.M., M.R. Lum, and J.A. Downie. 2001. What makes the rhizobia–legume symbiosis
so special? Plant Physiology 127:1484–1492.
Southeastern Naturalist
H.E. Dorman and L.E. Wallace
2019 Vol. 18, No. 1
162
IBM Corporation. 2012. IBM SPSS Statistics for Windows, Version 21.0. IBM Corporation,
Armonk, NY.
Katoh, K., K. Kuma, H. Toh, and T. Miyata. 2005. MAFFT version 5: Improvement in accuracy
of multiple-sequence alignment. Nucleic Acids Research 33:511–518.
Koppell, J.H., and M.A. Parker. 2012. Phylogenetic clustering of Bradyrhizobium symbionts
on legumes indigenous to North America. Microbiology 158:2050–2059.
Laguerre, G., S.M. Nour, V. Macheret, J. Sanjuan, P. Drouin, and N. Amarger. 2001. Classification
of rhizobia based on nodC and nifH gene analysis reveals a close phylogenetic
relationship among Phaseolus vulgaris symbionts. Microbiology 147:981–993.
Lauber, C.L., M.S. Strickland, M.A. Bradford, and N. Fierer. 2008. The influence of soil
properties on the structure of bacterial and fungal communities across land-use types.
Soil Biology and Biochemistry 40:2407–2415.
Lemaire, B., O. Dlodlo, S. Chimphango, C. Stirton, B. Schrire, J.S. Boatwright, O. Honnay,
E. Smets, J. Sprent, E.K. James, and A.M. Muasya. 2014. Symbiotic diversity, specificity,
and distribution of rhizobia in native legumes of the core Cape subregion (South
Africa). FEMS Microbiology Ecology 91:1–17.
Li, Q.Q., E.T. Wang, Y.Z. Zhang, Y.M. Zhang, C.F. Tian, X.H. Sui, and W.X. Chen. 2011.
Diversity and biogeography of rhizobia isolated from root nodules of Glycine max
grown in Hebei Province, China. Microbial Ecology 61:917–931.
Librado, P., and J. Rozas. 2009. DnaSP v5: A software for comprehensive analysis of DNA
polymorphism data. Bioinformatics 25:1451–1452.
Mantel, N. 1967. The detection of disease clustering and generalized regression approach.
Cancer Research 27:209–220.
Marschner, P., D. Crowley, and C.H. Yang. 2004. Development of specific rhizosphere bacterial
communities in relation to plant species, nutrition, and soil type. Plant and Soil
261:199–208.
Martínez-Romero, E., and J. Caballero-Mellado. 1996. Rhizobium phylogenies and bacterial
genetic diversity. Critical Reviews in Plant Sciences 15:113–140.
Miller, M.A., W. Pfeiffer, and T. Schwartz. 2010. Creating the CIPRES Science Gateway
for inference of large phylogenetic trees. Pp. 1–8, In Institute of Electrical and Electronics
Engineers (Eds.). Proceedings of the Gateway Computing Environments Workshop
(GCE), New Orleans, LA. 115 pp.
National Center for Biotechnology Information (NCBI). 1988. National Library of Medicine
(US), National Center for Biotechnology Information; Bethesda, MD. Available
online at https://www.ncbi.nlm.nih.gov/. Accessed 14 June 2018.
Ndlovu, J., D.M. Richardson, J.R.U. Wilson, and J.J. Le Roux. 2013. Co-invasion of South
African ecosystems by an Australian legume and its rhizobial symbionts. Journal of
Biogeography 40:1240–1251.
O’Malley, M.A. 2008. “Everything is everywhere: But the environment selects”: Ubiquitous
distribution and ecological determinism in microbial biogeography. Studies in History
and Philosophy of Biological and Biomedical Sciences 39:31 4–325.
Paffetti, D., C. Scotti, S. Gnocchi, S. Fancelli, and M. Bazzicalupo. 1996. Genetic diversity
of an Italian Rhizobium meliloti population from different Medicago sativa varieties.
Applied and Environmental Microbiology 62:2279–2285.
Papke, R.T., N.B. Ramsing, M.M. Bateson, and D.M. Ward. 2003. Geographical isolation
in hot spring cyanobacteria. Environmental Microbiology 5:650–6 59.
Parker, M.A. 2012. Legumes select symbiosis-island–sequence variants in Bradyrhizobium.
Molecular Ecology 21:1769–17778.
Southeastern Naturalist
163
H.E. Dorman and L.E. Wallace
2019 Vol. 18, No. 1
Parker, M.A. 2015. The spread of Bradyrhizobium lineages across host legume clades:
From Abarema to Zygia. Microbial Ecology 69:630–640.
Parker, M.A., and D.A. Kennedy. 2006. Diversity and relationships of Bradyrhizobium
from legumes native to eastern North America. Canadian Journal of Microbiology
52:1148–1157.
Parker, M.A., and A. Rousteau. 2014. Mosaic origins of Bradyrhizobium legume symbionts
on the Caribbean island of Guadeloupe. Molecular Phylogenetics and Evolution
77:110–115.
Parker, M.A., B. Lafay, J.J. Burdon, and P. van Berkum. 2002. Conflicting phylogeographic
patterns in rRNA and nifD indicate regionally restricted gene transfer in Bradyrhizobium.
Microbiology 148:2557–2565.
Pasternak, Z., A. Al-Ashhab, J. Gatica, R. Gafny, S. Avraham, D. Minz, O. Gillor, and E.
Jurkevitch. 2013. Spatial and temporal biogeography of soil microbial communities in
arid and semiarid regions. PLoS One 8(7):e69705. DOI:10.1371/journal.pone.0069705.
Pullen, T.M. 1963. The Cassia fasciculata complex (Leguminosae) in the United States.
Ph.D. Dissertation. University of Georgia, Athens, GA. 194 pp.
Reeves, D.W. 1994. Cover crops and rotations. Pp. 125–172, In J.L. Hatfield and B.A.
Stewart (Eds.). Advances in Soil Science: Crops Residue Management. Lewis Publishers,
CRC Press, Boca Raton, FL. 240 pp.
Rodríguez-Kábana, R., N. Kokalis-Burelle, D.G. Robertson, C.F. Weaver, and L. Wells.
1995. Effects of Partridge Pea–Peanut rotations on populations of Meloidogyne arenaria,
incidence of Sclerotium rolfsii, and yield of peanut. Nematropica 25:27–34.
Ronquist, F., M. Teslenko, P. van der Mark, D.L. Ayres, A. Darling, S. Höhna, B. Larget,
L. Liu, M.A. Suchard, and J.P. Huelsenbeck. 2012. MrBayes 3.2: Efficient Bayesian
phylogenetic inference and model choice across a large model space. Systematic Biology
61:539–542.
Rosenberg, M.S., and C.D. Anderson. 2011. PASSaGE: Pattern analysis, spatial statistics,
and geographic exegesis, Version 2. Methods in Ecology and Evolution 2:229–232.
Rout, M.E., and R.M. Callaway. 2012. Interactions between exotic invasive plants and
soil microbes in the rhizosphere suggest that “everything is not everywhere”. Annals of
Botany 110:213–222.
Sachs, J.L., S.W. Kembel, A.H. Lau, and E.L. Simms. 2009. In situ phylogenetic structure
and diversity of wild Bradyrhizobium communities. Applied and Environmental Microbiology
75:4727–4735.
Singer, S.R., S.L. Maki, A.D. Farmer, D. Ilut, G.D. May, S.B. Cannon, and J.J. Doyle.
2009. Venturing beyond beans and peas: What can we learn from Chamaecrista? Plant
Physiology 151:1041–1047.
Sinnott, R.W. 1984. Virtues of the haversine. Sky and Telescope 68:158.
Smouse, P.E., J.C. Long, and R.R. Sokal. 1986. Multiple regression and correlation extensions
of the Mantel test of matrix correspondence. Systematic Z oology 35:627–632.
Spearman, C. 1907. Demonstration of formulae for true measurement of correlation. The
American Journal of Psychology 18:161–169.
Sprent, J.I. 2009. Legume Nodulation: A Global Perspective. Wiley–Blackwell, Oxford,
UK. 200 pp.
Staley, J.T., and J.J. Gosink. 1999. Poles apart: Biodiversity and biogeography of sea-ice
bacteria. Annual Review of Microbiology 53:189–215.
Tamura, K., and M. Nei. 1983. Estimation of the number of nucleotide substitutions in the
control region of mitochondrial DNA in human chimpanzees. Molecular Biology and
Evolution 10:512–526.
Southeastern Naturalist
H.E. Dorman and L.E. Wallace
2019 Vol. 18, No. 1
164
Tamura, K., G. Stecher, D. Peterson, A. Filipski, and S. Kumar. 2013. MEGA6: Molecular
evolutionary-genetics analysis version 6.0. Molecular Biology and Evolution
30:2725–2729.
Tavaré, S. 1986. Some probabilistic and statistical problems in the analysis of DNA sequences.
Lectures on Mathematics in the Life Sciences 17:57–86.
Tian, C.F., E.T. Wang, T.X. Han, X.H. Sui, and W.X. Chen. 2007. Genetic diversity of
rhizobia associated with Vicia faba in three ecological regions of China. Archives of
Microbiology 188:273–282.
Toljander, J.F., B.D. Lindahl, L.R. Paul, M. Elfstrand, and R.D. Finlay. 2007. Influence of
arbuscular mycorrhizal mycelia exudates on soil bacterial growth and community structure.
FEMS Microbiology Ecology 61:295–304.
US Department of Agriculture-Natural Resource Conservation Service (USDA–NRCS).
2015. The PLANTS Database. National Plant Data Team, Greensboro, NC. Available
online at http://plants.usda.gov. Accessed 15 June 2018.
Vinuesa, P., C. Silva, D. Werner, and E. Martínez-Romero. 2005. Population genetics and
phylogenetic inference in bacterial molecular systematics: The roles of migration and
recombination in Bradyrhizobium species cohesion and delineation. Molecular Phylogenetics
and Evolution 34:29–54.
Weakley, A.S. 2012. Flora of the Southern and Mid-Atlantic States. University of North
Carolina Herbarium, Chapel Hill, NC. Available online at http://www.herbarium.unc.
edu/flora.htm. Accessed 15 June 2018.
Webb, C.O., D.D. Ackerly, and S.W. Kembel. 2008. Phylocom: Software for the analysis
of phylogenetic community structure and trait evolution. Bioinformatics 24:2098–2100.
Whitaker, R.J., D.W. Grogan, and J.W. Taylor. 2003. Geographic barriers isolate endemic
populations of hyperthermophilic archaea. Science 301:976–978.
Xiong, J., Y. Liu, X. Li, H. Zhang, J. Zeng, J. Hou, Y. Yang, T. Yao, R. Knight, and H. Chu.
2012. Geographic distance and pH drive bacterial distribution in alkaline lake sediments
across the Tibetan Plateau. Environmental Microbiology 14:2457–2466.
Yahara, T., F. Javadi, Y. Onoda, L. Paganucci de Queiroz, D.P. Faith, D.E. Prado, M. Akasaka,
T. Kadoya, F, Ishihama, S. Davies, J.W.F. Slik, J.W. Ferry, Y. Ting-Shuang, M.
Ke-Ping, C. Bin, D. Dedy, P.R. Toby, T. Midori, S. Masakazu, I. Motomi, E.N. Ashley, B.
Sven, R. Niels, K. Tadashi, V. Mohammad, M. Makiko, T. Hidenori, I. Yoh, S.F. Gideon,
V.E. Janine, and N. Tandiwe. 2013. Global legume diversity assessment: Concepts, key
indicators, and strategies. Taxon 62:249–266.
Yang, S., F. Tang, M. Gao, H.B. Krishnan, and H. Zhu. 2010. R-gene–controlled host
specificity in the legume–rhizobia symbiosis. Proceedings of the National Academy of
Sciences USA. 107:18735–18740.
Zahran, H.H. 2001. Rhizobia from wild legumes: Diversity, taxonomy, ecology, nitrogen
fixation, and biotechnology. Journal of Biotechnology 91:143–153.
Zhang, Y.M., Y. Li Jr., W.F. Chen, E.T. Wang, C.F. Tian, Q.Q. Li, Y.Z. Zhang, and W.X.
Chen. 2011. Biodiversity and biogeography of rhizobia associated with soybean
plants grown in the North China Plain. Applied and Environmental Microbiology
77:6331–6342.
Zhang, Y.M., C.T. Tian, X.H. Sui, W.F. Chen, and W.X. Chen. 2012. Robust markers reflecting
phylogeny and taxonomy of rhizobia. PLoS One 7(9):e44936. DOI:10.1371/journal.
pone.0044936.