Abstract
Understanding of the present-day genetic diversity, population structure, and evolutionary history of tree species can inform resource management and conservation activities, including response to pressures presented by a changing climate. Cercis canadensis (Eastern Redbud) is an economically valuable understory tree species native to the United States (U.S.) that is also important for forest ecosystem and wildlife health. Here, we document and explain the population genetics and evolutionary history of this deciduous tree species across its distributed range. In this study, we used twelve microsatellite markers to investigate 691 wild-type trees sampled at 74 collection sites from 23 Eastern U.S. states. High genetic diversity and limited gene flow were revealed in wild, natural stands of C. canadensis with populations that are explained by two major genetic clusters. These findings indicate that an ancient population bottleneck occurred coinciding with the last glacial maximum (LGM) in North America. The structure in current populations likely originated from an ancient population in the eastern U.S. that survived LGM and then later diverged into two contemporary clusters. Data suggests that populations have expanded since the last glaciation event from one into several post-glacial refugia that now occupy this species’ current geographic range. Our enhanced understanding benchmarks the genetic variation preserved within this species and can direct future efforts in conservation, and resource utilization of adaptively resilient populations that present the greatest genetic and structural diversity.
Similar content being viewed by others
Introduction
The genetic structure and demographics of many North American plant species have been greatly influenced by climate fluctuations that occurred during the Pleistocene epoch1,2. During the last glacial maximum (LGM), which occurred approximately 18,000–21,000 years ago3,4,5, the Laurentide Ice Sheet extended from the northernmost portions of North America to 39°N3. These events reduced the range of many temperate tree species, forcing them into glacial refugia, which included unglaciated southern regions and suitable micro-environments that were present in northern glaciated regions1,6. Many historically contiguous or closely occurring refugia have been identified in the eastern United States (U.S.), but location delineations and number of refugia continue to be debated6,7. These refugia are poorly represented within the fossil record, yet the spatial genetic structure and evolutionary histories of many species have been used as evidence of historical refugia6,8,9,10.
According to the “range shift following last glacial maximum” hypothesis, many temperate species recolonized and spread into their current distributions after the LGM11,12. The result of this recolonization process can be inferred from the genetic structure within extant plant populations, manifested as reduced genetic diversity along colonization routes, and distinct spatial genetic clusters across the newly expanded range of a species2,13,14. Patterns of reduced genetic diversity across a range in expansion are common among European temperate plant species15. This trend is not as evident among North American tree species, such as Carya cordiformis [Wagenh.] K. Koch (bitternut hickory) and C. ovata [Mill.] K.Koch (shagbark hickory), which present relatively uniform genetic variation in their distributions15. Genetic homogeneity in North American plant species can be explained by slow post-glacial expansion into new areas, presence of many refugia occurring in close proximity, and high gene flow across time15,16. Nevertheless, phylogeographic studies provide evidence that range dynamics of the post-glacial species populations have contributed more to the current patterns of genetic diversity in temperate tree species than any other ecological force (e.g., central- periphery theory17, particularly for the populations in the northern edge of a species distribution11.
The distribution of tree species, their genetic diversity, and population structure are shaped by many factors including climate oscillations, demographic incidents, ecological and environmental variables, and by their distinct biology11,18,19. Yet, the role that glaciation has played in the distribution, range, genetic variation, and spatial genetic structure of outcrossing tree species that span a geographically wide range are not well understood, especially across the eastern U.S. To better understand the role of LGM in structuring current species distributions and population structure of temperate tree species in the eastern U.S., we evaluated the spatial population structure of widely distributed forest understory tree Cercis canadensis L. (C. canadensis var. canadensis L.; Fabaceae; eastern redbud). Cercis canadensis is a great system to study the role of LGM in eastern U.S. tree species, due to its wide and continuous geographic distribution without any major geographical barriers across eastern U.S.
Cercis canadensis is a self-incompatible20, deciduous tree native to the midwestern and eastern U.S., as well as northeastern Mexico21,22. This species grows well in partial shade, is well adapted to a wide-range of climate conditions and elevations, and can be found in the USDA hardiness zones 4 through 923,24. This relatively small ornamental tree is characterized by its wide, colorful, umbrella-shaped canopy25, and is a popular landscape tree due to its heart-shaped foliage, compact form, and early spring flowers23.
When a fine-scale, smaller, and fragmented population of C. canadensis was examined with microsatellite loci26, wild trees maintained high genetic diversity, gene flow, and moderate to high genetic differentiation27. Although C. canadensis is ecologically important, there is limited knowledge of the contemporary genetic diversity, spatio-temporal genetic structure, gene flow, and past evolutionary history of this species across its native range in the U.S. To address this knowledge gap, we used microsatellite loci to accomplish the following: (1) characterize the genetic diversity of the wild populations of C. canadensis within its native range in the U.S.; (2) infer patterns in the spatial genetic structure of C. canadensis; and (3) reveal the evolutionary demographics in its native range. We hypothesized that C. canadensis wild populations will be genetically diverse and genotypes will be spatially clustered across its native range. We also hypothesized that the genetic structure would be consistent with range expansion that occurred from one of several southern glacial refugia. More specifically, we aimed to explore the following questions: (1) Do genetic diversity and population structures patterns reflect evidence for northern glacial refugium? We expected to detect high genetic diversity with distinct population structure in northern range limit, otherwise a trend of high to low genetic diversity from south to north would indicate recolonization of C. canadensis in north from one or more southern glacial refugia; (2) Is there any evidence for micro-refugia in current C. canadensis species distribution range? If there were multiple micro-refugia present in close proximity, especially in southern region, we expected to detect genetic homogeneity in the studied populations.
Materials and methods
Sample collection
Leaf samples of C. canadensis were collected by authors, collaborators, and citizen scientists (see acknowledgements), who sampled specimens across the native range of this species in 23 states in the midwestern and eastern U.S. (Table 1). The use of trees in the present study complied with international, national and/or institutional guidelines. Plants were identified based on collection guide provided to our collectors and confirmed by co-authors (voucher specimen deposited at the University of Tennessee Vascular Herbarium, catalog # TENN-V-0246136). For each collection site, at least 10 non-cultivated C. canadensis trees occurring within a one-mile radius were selected and their geographical coordinates were recorded. From each tree, five to seven young and disease-free leaves were collected, held between several pieces of absorbent paper and stored at ambient room temperature in a paper envelope until processing. Leaf samples from 1193 individual trees were collected at 117 collection sites. To avoid the over-representation of trees within a geographical area, we randomly selected a subset of collection sites from geographical areas with more than one collection site sampled. This study used total of 790 trees representing 79 collection sites that span much of the current native geographic range of C. canadensis.
DNA extraction
From each tree, 60 to 100 mg of dry leaf tissue was used to isolate DNA. Samples were homogenized four times for 30 s each at 6 m/s using a Beadmill 24 homogenizer (Fisher Scientific, Pittsburgh, Pennsylvania, U.S.) and were kept in liquid nitrogen for 5 min between each pulverization step. The Qiagen DNeasy Plant Mini Kit (Qiagen, Valencia, California, U.S.) was used to isolate genomic DNA (gDNA) from the pulverized samples with the following minor modifications in the manufacturer’s provided protocol. Specifically, 2% w/v polyvinylpyrrolidone (PVP) was mixed into the lysis buffer (AP1). Then 8 µl of RNase was added into each sample tube and incubated at 65 °C in a water bath for 45 min. Every two mins each sample tube was inverted gently to mix the sample well. Lastly, samples were incubated at − 20 °C for at least one hour. Ethanol was used to wash the spin columns if there was any visible remaining debris and elution buffer added. Elution buffer was heated to 65 °C before 50 µl was added to the spin columns twice. Concentrations of gDNA were quantified using ND1000 Ultraviolet-Vis Spectrophotometer (NanoDrop Technologies, Wilmington, Delaware, U.S.) and the gDNA was stored at − 20 °C until further use.
Microsatellite primers and genotyping conditions
Initially, gDNA was isolated from five C. canadensis individuals from the University of Tennessee Gardens (Knoxville, Tennessee, U.S.) and used to evaluate 68 candidate microsatellite loci26. Primers for twelve polymorphic microsatellite loci (Table 2) were selected for this study based on the successful amplification and PCR product size agreement with the published data. Microsatellite loci were amplified with polymerase chain reaction (PCR) in a 10 µl reaction mixture containing the following: 1 µl gDNA, 1 µl of 10 µM of each forward and reverse primer, 0.5 µl of dimethyl sulfoxide, 4 µl of GoTaq G2 Hot Start Master Mix (Promega Corp, Madison, Wisconsin U.S.), and 2.5 µl sterile molecular grade water. To assure validity of the data, both a negative control (reaction mixture with water instead of any DNA sample) and a positive control (a DNA sample from the initial screens that amplified across all microsatellite primers) were incorporated for every primer-pair tested. Amplification of DNA with 12 microsatellite loci across all samples was completed in 96 well plates using an Eppendorf Thermocycler (Eppendorf AG, Hamburg, Germany) with the following thermal profile: initial denaturation at 94 °C for 3 min, followed by 35 cycles of denaturation at 94 °C for 30 s, annealing at 55 °C for 30 s, and an extension at 72 °C for 30 s, with a final extension of 72 °C for 4 min.
Amplified PCR products were visualized with QIAxcel Capillary Electrophoresis System (Qiagen) and analyzed with a 15/600 bp internal alignment marker and a 25 to 500 bp DNA ladder. All C. canadensis gDNA samples were amplified and visualized against each of the 12 microsatellite loci using the procedure described above. Reactions not producing any amplified products were rerun once before they were considered missing data. Samples with ≥ 40% missing data were discarded. Also, collection sites with more than four samples having ≥ 40% missing data were excluded from the dataset.
Genetic diversity
Using the Excel macro FLEXBIN version28, raw allele sizes were converted into allelic classes. In this program, alleles were binned into base-pair (bp) size categories by statistical similarities. This binned genetic dataset was used for all of the following statistical analyses, which were completed using R version 3.5.329. Clone-correction of the data was implemented to identify presence of clonal individuals at the collection site level using the R package POPPR version 2.8.230,31. For each collection site, only multi-locus genotypes (MLG) were used to obtain unbiased estimates of allelic frequency from the dataset32.
R package POPPR was used to calculate the total number of alleles per locus, observed heterozygosity (HO; number of the heterozygotes present at a locus which is divided by sample size), expected heterozygosity (HE; calculated as expected heterozygosity per locus33), and linkage disequilibrium (rbard); non-random association of alleles between loci). Additionally, the Shannon-Weiner diversity index (H) was calculated for each collection site using POPPR. H considers both allele richness and evenness of the allelic distribution34. The number of unique private alleles in collection sites and different loci was estimated in POPPR package. Allelic richness (Ar), a measure of rarefied allelic counts per locus, was estimated using package HIERFSTAT version 0.04–2235. Allelic richness is used as an estimate of the long-term evolutionary potential to adapt and persist in a given population36,37. The genetic fixation index (FST), inbreeding coefficient (FIS), and allelic differentiation (F’ST)38,39 were calculated using HIERFSTAT package. Gene flow (Nm) was estimated using GenAlEx 6.5 software (Peakall & Smouse, 2006; Peakall & Smouse, 2012). In the program, Nm was estimated as the effective number of the migrants per locus based on F-statistics.
Population structure
Population structure within the native range of wild C. canadensis trees was analyzed using the program STRUCTURE version 2.3.440 to which an admixture model was applied. This Bayesian clustering method with Monte Carlo Markov Chain (MCMC) approach was used with the following parameters: 500,000 burn-in period with 1,000,000 MCMC repetitions for 30 independent chains for K values from 1 to 18. The resulting output was visualized with STRUCTURE HARVESTER web version 6.9441. The optimum K value, indicator of population clusters present in the dataset, was calculated utilizing the Evanno method42. The estimation of ΔK criterion obtained from STRUCTURE HARVESTER were visualized using POPHELPER 2.2.643 that merged the 30 independent chains. R packages MAPS version 3.3.044 and PLOTRIX version 3.8–145 were used to generate pie charts of admixture proportions at K = 2.
Several model-free methods were utilized to investigate the population structure of C. canadensis samples. A Neighbor-Joining (NJ) dendrogram was constructed using Nei’s genetic distance in POPPR46,47. Discriminant Analysis of Principal Components (DAPC) was implemented using package ADEGENET version 2.1.148 to visualize the underlying genetic structure of this species in its wide geographical range. This is a two-step multivariate analysis that investigates the genetic variations within populations among the sampled collection sites49. At first, a principal component analysis (PCA) was conducted, and then the number of PCA vectors (to explain majority of variance with minimizing over-fit of the DAPC) was selected. Then, a selected number of PCAs were used to reveal differences between groups while minimizing within group variations, as well as ordination of collection sites into distinct groups using discriminant analysis49,50. Moreover, missing values were calculated as mean allele frequency and cross-validation analysis was performed to select appropriate PC numbers.
Isolation by distance (IBD) was estimated using the Mantel test51,52 with 10,000 permutations in package VEGAN version 2.5–653 using Euclidean distance. IBD checks for a correlation between genetic distance and geographical distance among the individuals in a dataset. The Mantel test was implemented across the 74 collection sites while considering the whole dataset as one population.
Analysis of Molecular Variance (AMOVA)54 was carried out using POPPR with 10,000 permutations by sorting the individuals into hierarchical groups to assess the degree of molecular variance partitioned within, between, and among the collection sites. The levels of population hierarchy included: (1) 74 collection sites as one hierarchical group; (2) two groups on the basis of the STRUCTURE analysis; and (3) four major groups on the basis of five major eco-region divisions namely hot continental division (mountain provinces), hot continental division, warm continental division, subtropical division, and prairie division (see Supporting Information Fig. S2) in the midwestern and eastern U.S. (Bailey, 1994). C. canadensis collection sites in warm continental division were grouped with hot continental division collection sites. As C. canadensis is found in wide range of climate and elevations, we tested if there was any influence of the regional climate patterns in the population structures of C. canadensis in AMOVA analysis.
Demographic histories
To investigate and interpret the evolutionary history of C. canadensis, we used DIYABC program version 2.155,56 that utilized Approximate Bayesian Computation (ABC) statistical methods. For this analysis, collected individuals were pooled into two major groups, based on the STRUCTURE results. To elucidate the evolutionary history of C. canadensis, we analyzed competing scenarios in two ABC steps. In the first step, we tested five demographic scenarios using 200,000 simulated pseudo-observed datasets (PODs) wherein: (1) the first two scenarios suggested stepwise divergence of the current two major groups from an ancient population, (2) a third scenario suggested a single, two-way split of contemporary groups from an ancient unsampled population, and (3) the last two scenarios were based on the hypothesis of divergence of current groups from two separate ancient un-sampled populations. Once the analysis of these scenarios was completed, the two scenarios from the first step that yielded higher logistic regression support were selected as the basis for assessing the second step of ABC. In the second step, seven scenarios were constructed that addressed the possibility of a bottleneck occurrence within the evolutionary history of the species. Over 1,000,000 pseudo-observed datasets (PODs) were simulated under the assumed prior parameter ranges for each scenario. Posterior probabilities of the compared scenarios were estimated to select the best supported scenario55.
Results
Microsatellite genetic diversity and hierarchical fixation indices
Twelve microsatellite loci were amplified from 790 C. canadensis trees sampled in this study. Due to presence of missing data (missing ≥ 40% SSRs), five of 79 collection sites were excluded and 49 individuals from the remaining 74 sites were discarded, resulting in 691 individuals from 74 collections. Additionally, after deleting two clonal individuals, 689 unique multilocus genotypes from 74 collection sites remained for further data analyses (Table 1). The average null alleles or missing data across the dataset were overall 2.89% (see Supporting Information Fig. S1). Nei’s genetic diversity index (HE) value in the studied 74 collection sites was 0.67, ranging from 0.32 (Anderson county, TN10) to 0.68 (Sarpy county, NE1) (Table 1). Moreover, weak (rbard = 0.05, P value = 0.01) but significant linkage disequilibrium value detected in the dataset. Nineteen private alleles were detected within the 74 collection sites (Table 1). In addition, 9 of the 12 microsatellite loci yielded private alleles, and the highest number of private alleles was recovered from locus 168a (Pa = 4, Table 2). The number of alleles per locus ranged from 6 to 13 with a mean of 10 alleles per locus (Table 2). Overall allelic richness (Ar) ranged from 1.23 for locus 995a to 1.79 for locus 780b, with a mean of 1.55, implying a presence of high allelic richness in wild C. canadensis individuals. Observed heterozygosity (HO) across all loci was 0.32, ranging from 0.01 (locus 995a and 680a) to 0.99 (locus 127spa). The overall expected heterozygosity (HE) across all 12 microsatellite loci was high (HE = 0.69), ranging from 0.52 (locus 995a) to 0.86 (locus 780b).
The overall Shannon-Weiner diversity index (H) for the 12 loci was 1.42 and ranged from 0.84 (locus 995a) to 2.19 (locus 780b; Table 2). Additionally, high population fixation (FST = 0.19; ranging from 0.05 to 0.53; Table 2) and population differentiation (F’ST = 0.19; ranging from 0.05 to 0.54) were identified among C. canadensis populations. We estimated an inbreeding coefficient (FIS) of 0.43 across all loci, indicating excess homozygotes (Table 2) among the studied C. canadensis populations. The average estimated gene flow was 0.75, which indicates that a limited amount of gene flow has occurred among the studied populations (Table 2).
Population structure
Using Nei’s genetic distance, we estimated pairwise FST values among the 74 collection sites and the values ranged from 0.02 to 0.33 (see Supporting Information Table S1). STRUCTURE results revealed an optimum ∆K = 2, implying that across its wide native range, C. canadensis collection sites are divided into two major clusters. Collection sites in the northern-most collection region of the U.S. (Ohio to Nebraska) and mid-south to mid-north (from Texas to Nebraska) were part of the first cluster (designated as north genetic cluster) (Fig. 1). The rest of the collection sites from the northeast (New York) to mid-south (Mississippi) along the Atlantic Ocean coastline belonged to the second cluster (designated as south genetic cluster). Whereas, a constructed NJ dendrogram revealed the presence of two major groups (except KS and TX collection sites that did not group with any major group), which supported the STRUCTURE findings of presence of two genetic clusters (Fig. 2). In addition, the collection site distribution in these two major groups (NJ dendrogram) is similar to the distribution of collection sites in the two STRUCTURE-based clusters. The DAPC biplot further confirmed the presence of genetic structures, primarily along the x-axis with two overlapping clusters (Fig. 3). Therefore, based on additional analyses used in this study, the grouping of C. canadensis individuals is best explained with two genetic clusters (Fig. 1–3). These analyses also showed that the majority of the collection sites (except two collection sites from Georgia) grouped in clusters based on their geographical origin.
In the Analysis of Molecular Variance (AMOVA) analysis, the first data arrangement showed that most of the genetic variation was present within 74 collection sites (74.2%, P < 0.001) (Table 3). A significant amount of variation was also partitioned among collection sites (25.8%, P < 0.001). When the dataset was divided into two genetic clusters based on STRUCTURE results, 69.2% (P < 0.001) of the genetic variability was attributed to the location of collection sites (Table 3). There was also a significant amount of variability between the two different genetic clusters (13.9%, P < 0.001) and between collection sites within the clusters (17%, P < 0.001) (Table 3). When the dataset was partitioned by the major eco-regions groups that are represented across the distribution of C. canadensis, only 7.9% (P < 0.001) of the variability could be attributed among eco-region groups, versus 18.6% (P < 0.001) variability that was attributed among the collection sites within groups (Table 3). The majority of genetic variation was explained among individuals within a collection site, rather than among populations or group levels for all three tested scenarios (Table 3). Nevertheless, the extent of variation that was observed within collection sites and between clusters revealed the presence of genetic structure. AMOVA results were, therefore, congruent with the hierarchical fixation indices and indicated the presence of population structure. However, the lowest amount of variation was found within the groups when the data were divided according to major eco-regions. This finding suggests that partitioning trees within eco-regions cannot be expected to explain the genetic differentiation and population structure observed in C. canadensis wild populations. Results from the isolation-by-distance analysis indicated that among C. canadensis populations, the geographical distance effect was weak, but linearly correlated (r = 0.08, P < 0.001) with genetic distance (see Supporting Information Fig. S2).
Demographic histories
The DIYABC program with the ABC approach, however, supported the presence of population structure and found evidence for an ancient bottleneck event occurring in C. canadensis wild populations. From the first step of the analysis, two probable scenarios were chosen according to their posterior relative support (Scenario 2, posterior probability (P) = 0.39 and Scenario 3, posterior probability (P) = 0.37; Fig. 4A). In these analyses, Scenario 2 provided evidence that the contemporary C. canadensis population originated in the southeastern U.S. region from an ancient population and then later, a north population group (first genetic cluster) diverged from the south population group (second genetic cluster). Alternatively, Scenario 3 suggested that both current C. canadensis groups (north and south) have split from an ancient, as yet unsampled group (Fig. 4A). In the second step of the ABC analysis (Fig. 4B), the principal component analysis and relative posterior probability tests revealed that Scenario 2a (posterior probability (P) = 0.74, Fig. 4B) was the most supported and therefore had the greatest likelihood of accurately describing the evolutionary processes that are evident within native stands of C. canadensis. Thus, we infer from Scenario 2a that from an ancient population, a group of C. canadensis in the south first endured a bottleneck, and then later, a group of northern C. canadensis diverged from the southern group (Scenario 2a, Fig. 4B).
Estimated posterior parameters of Scenario 2a suggested that a population bottleneck occurred approximately 4,950 generations ago (ranging from 722 to 9,650 generations in simulated datasets), which is approximately 25,000 years ago, given the average time for a C. canadensis tree to reach reproductive maturity is six to seven years57 (Fig. 4B). Therefore, the bottleneck event most probably occurred during the last glacial period, which ended about 21,000 years ago3,5. Later, the northern population diverged from the southern population about 493 generations ago (ranging from 102 to 1,490 generations in simulated datasets) (Fig. 4B). Post-hoc analyses provided goodness-of-fit for this scenario, with the original dataset well embedded in the prior PODs population and nested in the posterior PODs population (see Supporting Information Table S2 for details of this analysis).
Discussion
Wild populations of C. canadensis that were sampled across its native range in the U.S. revealed high levels of genetic diversity and population differentiation, the presence of population structure, limited gene flow, and an ancient bottleneck that temporally coincides with the last glacial period in North America. We detected the presence of geographical clusters, longitudinally in the southern region (along U.S. coastal plains), and northern region. Evolutionary history analyses revealed an ancient bottleneck event occurring in the C. canadensis population in the south followed by divergence of the northern population from the southern population of C. canadensis.
When populations were compared across the ecoregion divisions from which they were collected, ecoregion designations were not associated with population structure and genetic diversity of wild populations of C. canadensis. Low genetic variation in C. canadensis across ecoregions is not surprising, given that this tree species is well-adapted to a wide range of soil types, environmental conditions and habitats, has not been constrained by any major geographical barriers, and is widespread among the eastern U.S. However, we found evidence of relatively higher genetic diversity among the northernmost collection sites (IA, IN, MI, NE, and OH, and in mid-latitude North America) that are located at the periphery of the contemporary northern range of this species. A plausible explanation for this discrepancy between northernmost samples when compared to the southern collection sites is the probability of one major refugium or admixture between small, but genetically rich, populations in refugial contact zones58,59,60. This effect is most evident in species that experience reproduction via long distance gene flow and local adaptation among sensitive individuals in the distributed population margins. However, unlike European temperate species, eastern north American species maintained high genetic diversity in northern populations61,62. This high level of genetic diversity in the C. canadensis northern population could be maintained by long-distance seed dispersal events during range expansion from post-glacial northern refugia62,63.
The ability of C. canadensis to maintain high genetic diversity can be influenced by several factors, including wide and continuous geographic distribution, an outcrossing reproductive system, and large effective population size59,64,65,66. Many other temperate tree species sustain high genetic diversity and allelic richness across a wide geographical range even in the presence of environmental stressors11,14,58, pressure from insect and plant pathogens67,68, and human disturbances67,69,70. A study using microsatellite loci revealed high genetic variation among five Asian Cercis spp., averaging 5.7 alleles per locus71. A recent study focused on smaller and fragmented C. canadensis populations also determined that trees in this species maintain high genetic diversity and allelic richness across the native range27, congruent with Asian Cercis species and several other hardwood tree species66,72,73,74,75,76.
In addition to high genetic variability, C. canadensis populations also display a wide range of morphological variation across diverse environmental conditions20,26,77,78,79. For example, Cercis leaf shape, size, surface pubescence, and other structural features were found to be strongly related to environmental factors, such as temperature and moisture content79,80,81. In Cercis spp., these characteristics likely originated through local adaptation to varying climatic pressure26,82,83. Morphological variation in C. canadensis has led efforts to differentiate the species into the following three varieties: C. canadensis var. canadensis L., distributed across mesophytic habitats of the eastern U.S., C. canadensis var. mexicana (Rose) M. Hopkins (Mexican redbud) and C. canadensis var. texensis (S. Watson) M. Hopkins (Texas redbud), which are both commonly found in semi-arid regions of central Mexico and southwestern Texas26,84,85,86. However, the validity of such sub-specific classification has been questioned because of the highly continuous pattern of morphological variation in C. canadensis populations across its range. Moreover, current phylogenetic studies were unable to provide sufficient support to validate these divisions84,87. Because C. canadensis var. mexicana and C. canadensis var. texensis were not represented in our study, our data will not assist in resolving this question.
Widely distributed tree species that grow in large populations usually have low genetic differentiation and have limited population structure across their geographic range15,60,65,66,68,88. Populations of Viburnum rufidulum Raf.89 and Cornus florida L.68 are temperate tree species that are widely distributed in the southeastern U.S. and have low genetic differentiation with weak population structures. Across populations of V. rufidulum and C. florida, high levels of gene flow via pollen and seed dispersal may have reduced genetic variability75,89. Contrary to these studies, high genetic differentiation observed among widely distributed populations of C. canadensis may be attributed to a limited gene flow, as well as the demographic history of the species.
Similar to many other self-incompatible forest tree species20, gene flow in C. canadensis is dependent upon various pollen and seed dispersal mechanisms. Flight distance of insect pollinators varies from one to several miles90,91, which would limit long distance gene flow by pollen dispersal among trees. Seedpods and seeds of C. canadensis are relatively heavy and typically fall in close proximity to the parental tree. Progeny that survive grow as non-reproductive seedlings during the next several years57,92 and yield half-sib “neighborhoods” within a localized spatial scale89,93,94,95,96. Several seed-feeding mammals, such as eastern woodrats (Neotoma floridana Ord)97, and birds including quail98 contribute to the dispersal of C. canadensis seed to some extent. Small rodents and deer may repeatedly eat from the same tree, thus carrying the closely related, half-sibling propagules (if eaten when seeds have matured) for distances restricted to the retention time of fecal scats57,97,99,100,101. To fully understand gene flow patterns and predict changes to C. canadensis distribution patterns, it may be helpful to unravel the seed and pollen dispersal methods and efficacy of seed transport by animals that have been associated with this tree species. However, seed transport efficacy will likely be limited to the relatively short distances traveled by these animals during foraging. Fruit consumption rate by animals is also restricted by reliance upon C. canadensis fruits as emergency food in late fall or winter, and this behavior would lower efficiency of functional seed dispersal57,97,99,100,101. These events likely limit the gene flow to short distances, create spatial genetic structures, and increase the likelihood of inbreeding at a local level, as revealed in fine scale level assessments of C. canadensis27,69,102. We also collected C. canadensis samples from New York (U.S.) that represent individuals occurring farther north than the reported geographic range of the species. These individuals also could result from open-pollinated escapes subsequent to introduction of C. canadensis into managed landscapes.
STRUCTURE analysis of the C. canadensis dataset revealed the presence of two geographically distinct clusters, designated as northern and southern clusters, that are divided longitudinally northwest by southeast along a Kentucky-Tennessee-Mississippi transition zone. From this evidence, the southern Appalachian Mountains have not posed a barrier, as populations belonging to the northern STRUCTURE clusters were found on both sides of the Appalachian Mountains. The presence of only two genetic clusters is congruent with the simple postglacial lineage theory presented for eastern North American tree species59,103. The most recent glacial event ended approximately 21,000 years ago. By its conclusion, boreal and temperate tree species had shifted closer to mid-latitudes within the eastern U.S., where many species survived within bottlenecked refugial populations104. According to our DIYABC supported scenario, a southeastern refugium was also likely to be the major postglacial refugium for C. canadensis. This scenario is further supported by several phylogenetic studies that have indicated that southeastern U.S. populations served as one prominent large-scale, post-glacial refugium for many temperate species14,103,104. Modern day temperate species including Fagus grandifolia Ehrh. (American beech), Acer rubrum L. (Red maple), and C. florida (Flowering dogwood), for example, likely originated from this southeastern refugium4,14,105. Cercis canadensis also shares the same geographic distribution as these temperate tree species, and modern-day wild populations of C. canadensis are ubiquitous throughout this region.
Our analyses also revealed support for several possible micro-refugia across the eastern U.S., evident in genetic differences among populations and presence of substructures that lack distinct centers. Several studies on different tree species indicated presence of refugia in the eastern U.S. such as southern Appalachian Mountains, southeastern coastal plains, and lower Mississippi River Valley (McLachlan et al., 2005; Potter et al. 2011). Post-glacial C. canadensis populations from this geographical range may have spread northward to establish the current species distribution. Post-glacial populations of other tree species from this range are adapted to semi-arid to xeric environments9,14,103 and present adaptive characteristics that are similarly evident in mid-western C. canadensis populations. Moreover, the presence of a number of refugia or fragmented refugia is also can be supported by the high genetic diversity and allelic richness of the modern-day C. canadensis populations60.
Phylogeographical studies of other tree species and animals indicate that they survived as northern cryptic micro-refugia104,106,107,108. Due to insufficient fossil data from the Late Pleistocene in the northern region, it is difficult to conclude the definite presence of northern refugia of the C. canadensis populations5,6. The best supported DIYABC evolutionary scenario suggests that at the time of the last glacial period C. canadensis populations persisted within a southern population group. Therefore, we also find little evidence for the possibility of a northern cryptic refugium for this species. Instead, pre-glacial C. canadensis was distributed in midwestern and southeastern U.S. populations, which later survived in one or more post-glacial midwestern and southeastern U.S. refugia. As a consequence of long-term population isolation within the refugial areas, a post-glacial refugial population in the midwestern U.S. may have diverged from the southeastern large refugium population, giving rise to a genetically differentiated northern spatial cluster70,95. Also, this post-glacial northern population may have later migrated from the midwestern U.S. to establish the current range distribution.
Ancestors of North American Cercis species are thought to have originated under mesic conditions and may have dispersed into North America across the North Atlantic Land Bridge81,84,109. According to several studies, ancestral Cercis population adapted to the drier environment and then spread into the Northern hemisphere during the mid-Miocene period84,87. It is possible, then, that this ancestral, un-sampled, mid-Miocene Cercis population gave rise to the southern C. canadensis population as suggested by DIYABC Scenario 2a.
This economically and ecologically significant deciduous shade tree species has a number of desirable morphological variations and ornamental characteristics including foliar color and texture, flower color variation, drought tolerance, pathogen resistance, as well as a wide variety of architectural forms20,26,57. Fruits and seeds of C. canadensis are consumed by several bird species and small mammals57,97,99,101, and many pollinators depend on this tree for an early season food source110. More than three dozen cultivars are available commercially and nursery stock sales of the species contribute to more than $27 M annually in the U.S.71,111. When paired with recent introductions of novel horticultural cultivars with highly desirable characteristics, the value of adaptive traits that are likely to exist across different geographic localities supports the importance in conserving local level diversity of C. canadensis. These populations are genetic reservoirs of potential variability that can provide breeding programs with the resources needed to improve selected traits (e.g., limiting seed pod productivity in landscape specimens) and provide additional opportunities for developing high-value cultivars for commercial trade. Future work should also focus on identifying important adaptive traits in wild populations that can be used to help ensure that C. canadensis populations will persist and will continue to adapt to a changing climate that is occurring across portions of the current species distribution.
Data Archiving Statement
After the manuscript is accepted, data will be publicly available and deposited to Dryad Depository.
References
Hewitt, G. The genetic legacy of the Quaternary ice ages. Nature 405, 907–913. https://doi.org/10.1038/35016000 (2000).
Hewitt, G. Genetic consequences of climatic oscillations in the Quaternary. Philos. Trans. R. Soc. Lond. 359, 183–195. https://doi.org/10.1098/rstb.2003.1388 (2004).
Ehlers, J. & Gibbard, P. Quaternary Glaciations-Extent and Chronology: Part I: Europe Vol. 2 (Elsevier, New York, 2004).
Call, A. et al. Genetic structure and post-glacial expansion of Cornus florida L. (Cornaceae): Integrative evidence from phylogeography, population demographic history, and species distribution modeling. J. Syst. Evol. 54, 136–151. https://doi.org/10.1111/jse.12171 (2016).
Jackson, S. et al. Vegetation and environment in eastern North America during the Last Glacial Maximum. Quatern. Sci. Rev. 19, 489–508. https://doi.org/10.1016/S0277-3791(99)00093-1 (2000).
Nadeau, S. et al. Contrasting patterns of genetic diversity across the ranges of Pinus monticola and P. strobus: A comparison between eastern and western North American postglacial colonization histories. Am. J. Bot. 102, 1342–1355. https://doi.org/10.3732/ajb.1500160 (2015).
Beaulieu, J. & Simon, J. Genetic structure and variability in Pinus strobus in Quebec. Can. J. For. Res. 24, 1726–1733. https://doi.org/10.1139/x94-223 (1994).
Provan, J. & Bennett, K. Phylogeographic insights into cryptic glacial refugia. Trends Ecol. Evol. 23, 564–571. https://doi.org/10.1016/j.tree.2008.06.010 (2008).
Soltis, D., Morris, A., McLachlan, J., Manos, P. & Soltis, P. Comparative phylogeography of unglaciated eastern North America. Mol. Ecol. 15, 4261–4293. https://doi.org/10.1111/j.1365-294X.2006.03061.x (2006).
Mee, J. & Moore, J. The ecological and evolutionary implications of microrefugia. J. Biogeogr. 41, 837–841. https://doi.org/10.1111/jbi.12254 (2014).
Hoban, S. et al. Range-wide distribution of genetic diversity in the North American tree Juglans cinerea: A product of range shifts, not ecological marginality or recent population decline. Mol. Ecol. 19, 4876–4891. https://doi.org/10.1111/j.1365-294X.2010.04834.x (2010).
Hampe, A. & Petit, R. Conserving biodiversity under climate change: The rear edge matters. Ecol. Lett. 8, 461–467. https://doi.org/10.1111/j.1461-0248.2005.00739.x (2005).
Excoffier, L., Foll, M. & Petit, R. Genetic consequences of range expansions. Annu. Rev. Ecol. Evol. Syst. 40, 481–501. https://doi.org/10.1146/annurev.ecolsys.39.110707.173414 (2009).
McLachlan, J., Clark, J. & Manos, P. Molecular indicators of tree migration capacity under rapid climate change. Ecology 86, 2088–2098. https://doi.org/10.1890/04-1036 (2005).
Bemmels, J. & Dick, C. Genomic evidence of a widespread southern distribution during the Last Glacial Maximum for two eastern North American hickory species. J. Biogeogr. 45, 1739–1750. https://doi.org/10.1111/jbi.13358 (2018).
Jaramillo-Correa, J., Beaulieu, J., Khasa, D. & Bousquet, J. Inferring the past from the present phylogeographic structure of North American forest trees: Seeing the forest for the genes. Can. J. For. Res. 39, 286–307. https://doi.org/10.1139/X08-181 (2009).
Eckert, C., Samis, K. & Lougheed, S. Genetic variation across species’ geographical ranges: The central–marginal hypothesis and beyond. Mol. Ecol. 17, 1170–1188. https://doi.org/10.1111/j.1365-294X.2007.03659.x (2008).
Foll, M. & Gaggiotti, O. Identifying the environmental factors that determine the genetic structure of populations. Genetics 174, 875–891. https://doi.org/10.1534/genetics.106.059451 (2006).
Loveless, M. & Hamrick, J. Ecological determinants of genetic structure in plant populations. Ann. Rev. Ecol. Syst. 15, 65–95. https://doi.org/10.1146/annurev.es.15.110184.000433 (1984).
Roberts, D., Werner, D., Wadl, P. & Trigiano, R. Inheritance and allelism of morphological traits in eastern redbud (Cercis canadensis L.). Hortic. Res. 2, 1–11 (2015).
Couvillon, G. Cercis canadensis L. seed size influences germination rate, seedling dry matter, and seedling leaf area. HortScience 37, 206–207 (2002).
Li, S. et al. Methods for breaking the dormancy of eastern redbud (Cercis canadensis) seeds. Seed Sci. Technol. 41, 27–35 (2013).
Cheong, E. & Pooler, M. Micropropagation of Chinese redbud (Cercis yunnanensis) through axillary bud breaking and induction of adventitious shoots from leaf pieces. In Vitro Cell. Dev. Biol. Plant 39, 455–458 (2003).
Pooler, M., Jacobs, K. & Kramer, M. Differential resistance to Botryosphaeria ribis among Cercis taxa. Plant Dis. 86, 880–882. https://doi.org/10.1094/PDIS.2002.86.8.880 (2002).
Trigiano, R., Beaty, R. & Graham, E. Somatic embryogenesis from immature embryos of redbud (Cercis canadensis). Plant Cell Rep. 7, 148–150. https://doi.org/10.1007/BF00270127 (1988).
Wadl, P., Trigiano, R., Werner, D., Pooler, M. & Rinehart, T. Simple sequence repeat markers from Cercis canadensis show wide cross-species transfer and use in genetic studies. J. Am. Soc. Hortic. Sci. 137, 189–201. https://doi.org/10.21273/JASHS.137.3.189 (2012).
Ony, M. et al. Habitat fragmentation influences genetic diversity and differentiation: Fine-scale population structure of Cercis canadensis (eastern redbud). Ecol. Evol. 10, 3655–3670. https://doi.org/10.1002/ece3.6141 (2020).
Amos, W. et al. Automated binning of microsatellite alleles: Problems and solutions. Mol. Ecol. Resour. 7, 10–14. https://doi.org/10.1111/j.1471-8286.2006.01560.x (2007).
R: A language and environment for statistical computing (R Foundation for Statistical Computing, Vienna, Austria, 2019).
Kamvar, Z., Tabima, J. & Grünwald, N. Poppr: An R package for genetic analysis of populations with clonal, partially clonal, and/or sexual reproduction. PeerJ 2, e281. https://doi.org/10.7717/peerj.281 (2014).
Kamvar, Z., Brooks, J. & Grünwald, N. Novel R tools for analysis of genome-wide population genetic data with emphasis on clonality. Front. Genet. 6, 208. https://doi.org/10.3389/fgene.2015.00208 (2015).
Tsui, C. et al. Population structure and migration pattern of a conifer pathogen, Grosmannia clavigera, as influenced by its symbiont, the mountain pine beetle. Mol. Ecol. 21, 71–86. https://doi.org/10.1111/j.1365-294X.2011.05366.x (2012).
Nei, M. Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics 89, 583–590 (1978).
Shannon, C. E. A mathematical theory of communication. Bell System Tech. J. 27, 379–423 (1948).
Goudet, J. Hierfstat, a package for R to compute and test hierarchical F-statistics. Mol. Ecol. Notes 5, 184–186. https://doi.org/10.1111/j.1471-8286.2004.00828.x (2005).
Hurlbert, S. The nonconcept of species diversity: A critique and alternative parameters. Ecology 52, 577–586. https://doi.org/10.2307/1934145 (1971).
El Mousadik, A. & Petit, R. High level of genetic differentiation for allelic richness among populations of the Argan tree [Argania spinosa (L.) Skeels] endemic to Morocco. Theor. Appl. Genet. 92, 832–839. https://doi.org/10.1007/BF00221895 (1996).
Bird, C., Karl, S., Smouse, P. & Toonen, R. In Phylogeography and Population Genetics in Crustacea Vol. 19 (eds Held Christoph, Koenemann Stefan, & Schubart Christoph) pp. 31–55 (Boca Raton, FL: CRC Press, 2011).
Meirmans, P. & Hedrick, P. Assessing population structure: FST and related measures. Mol. Ecol. Resour. 11, 5–18. https://doi.org/10.1111/j.1755-0998.2010.02927.x (2011).
Pritchard, J., Stephens, M. & Donnelly, P. Inference of population structure using multilocus genotype data. Genetics 155, 945–959 (2000).
Earl, D. & Bridgett, V. STRUCTURE HARVESTER: A website and program for visualizing STRUCTURE output and implementing the Evanno method. Conserv. Genet. Resour. 4, 359–361. https://doi.org/10.1007/s12686-011-9548-7 (2012).
Evanno, G., Regnaut, S. & Goudet, J. Detecting the number of clusters of individuals using the software STRUCTURE: A simulation study. Mol. Ecol. 14, 2611–2620. https://doi.org/10.1111/j.1365-294X.2005.02553.x (2005).
Francis, R. Pophelper: An R package and web app to analyse and visualize population structure. Mol. Ecol. Resour. 17, 27–32. https://doi.org/10.1111/1755-0998.12509 (2017).
Becker, R. & Wilks, A. MAPS: An R Package to Drae Geographical Maps (Version package 3.3.0, 2018).
Lemon, J. Plotrix: An R Package for Various Plotting Functions (Version R package 3.8–1, 2006).
Bruvo, R., Michiels, N., D’souza, T. & Schulenburg, H. A simple method for the calculation of microsatellite genotype distances irrespective of ploidy level. Mol. Ecol. 13, 2101–2106. https://doi.org/10.1111/j.1365-294X.2004.02209.x (2004).
Grünwald, N., Everhart, S., Knaus, B. & Kamvar, Z. Best practices for population genetic analyses. Phytopathology 107, 1000–1010. https://doi.org/10.1094/PHYTO-12-16-0425-RVW (2017).
Jombart, T. & Ahmed, I. adegenet 1.3–1: New tools for the analysis of genome-wide SNP data. Bioinformatics 27, 3070–3072. https://doi.org/10.1093/bioinformatics/btr521 (2011).
Jombart, T., Devillard, S. & Balloux, F. Discriminant analysis of principal components: A new method for the analysis of genetically structured populations. BMC Genet. 11, 9. https://doi.org/10.1186/1471-2156-11-94 (2010).
Cullingham, C., Cooke, J. & Coltman, D. Effects of introgression on the genetic population structure of two ecologically and economically important conifer species: Lodgepole pine (Pinus contorta var. latifolia) and jack pine (Pinus banksiana). Genome 56, 577–585. https://doi.org/10.1139/gen-2013-0071 (2013).
Diniz-Filho, J. et al. Mantel test in population genetics. Genet. Mol. Biol. 36, 475–485. https://doi.org/10.1590/S1415-47572013000400002 (2013).
Mantel, N. The detection of disease clustering and a generalized regression approach. Can. Res. 27, 209–220 (1967).
Vegan: Community ecology package v. R package version 2.5–3 (R package version 2.5–3). (2018).
Excoffier, L., Smouse, P. & Quattro, J. Analysis of molecular variance inferred from metric distances among DNA haplotypes: Application to human mitochondrial DNA restriction data. Genetics 131, 479–491 (1992).
Cornuet, J., Ravigné, V. & Estoup, A. Inference on population history and model checking using DNA sequence and microsatellite data with the software DIYABC (v1.0). BMC Bioinform. 11, 401–411. https://doi.org/10.1186/1471-2105-11-401 (2010).
Cornuet, J. et al. DIYABC v2.0: A software to make approximate Bayesian computation inferences about population history using single nucleotide polymorphism, DNA sequence and microsatellite data. Bioinformatics 30, 1187–1189. https://doi.org/10.1093/bioinformatics/btt763 (2014).
Dickson, J. In Silvics of North America Vol. 2 (eds Burns, R. & Honkala, B.) 266–269 (United States Department of Agriculture-Forest Service, 1990).
Thomson, A., Dick, C. & Dayanandan, S. A similar phylogeographical structure among sympatric North American birches (Betula) is better explained by introgression than by shared biogeographical history. J. Biogeogr. 42, 339–350. https://doi.org/10.1111/jbi.12394 (2015).
Petit, R. et al. Glacial refugia: Hotspots but not melting pots of genetic diversity. Science 300, 1563–1565 (2003).
David, R. & Hamann, A. Glacial refugia and modern genetic diversity of 22 western North American tree species. Proc. R. Soc. B Biol. Sci. 282, 20142903. https://doi.org/10.1098/rspb.2014.2903 (2015).
Lumibao, C., Hoban, S. & McLachlan, J. Ice ages leave genetic diversity ‘hotspots’ in Europe but not in Eastern North America. Ecol. Lett. 20, 1459–1468. https://doi.org/10.1111/ele.12853 (2017).
Bialozyt, R., Ziegenhagen, B. & Petit, R. Contrasting effects of long distance seed dispersal on genetic diversity during range expansion. J. Evol. Biol. 19, 12–20. https://doi.org/10.1111/j.1420-9101.2005.00995.x (2006).
Petit, R. Early insights into the genetic consequences of range expansions. Heredity 106, 203–204. https://doi.org/10.1038/hdy.2010.60 (2011).
Dubreuil, M. et al. Genetic effects of chronic habitat fragmentation revisited: Strong genetic structure in a temperate tree, Taxus baccata (Taxaceae), with great dispersal capability. Am. J. Bot. 97, 303–310. https://doi.org/10.3732/ajb.0900148 (2010).
Hamrick, J., Godt, M. & Sherman-Broyles, S. In Population Genetics of Forest Trees Vol. 42 (eds Adams, W., Strauss, S., Copes, D. & Griffin, A) 95–124 (Springer, Dordrecht, 1992).
Hamrick, J. & Godt, M. Effects of life history traits on genetic diversity in plant species. Philos. Trans. R. Soc. Lond. Ser. B Biol. Sci. 351, 1291–1298 (1996).
Spaulding, H. & Rieske, L. The aftermath of an invasion: Structure and composition of central appalachian hemlock forests following establishment of the hemlock woolly adelgid, Aelges tsugae. Biol. Invasions 12, 3135–3143. https://doi.org/10.1007/s10530-010-9704-0 (2010).
Hadziabdic, D. et al. Analysis of genetic diversity in flowering dogwood natural stands using microsatellites: The effects of dogwood anthracnose. Genetica 138, 1047–1057. https://doi.org/10.1007/s10709-010-9490-8 (2010).
Marquardt, P., Echt, C., Epperson, B. & Pubanz, D. Genetic structure, diversity, and inbreeding of eastern white pine under different management conditions. Can. J. For. Res. 37, 2652–2662 (2007).
Potter, K. et al. Widespread inbreeding and unexpected geographic patterns of genetic variation in eastern hemlock (Tsuga canadensis), an imperiled North American conifer. Conserv. Genet. 13, 475–498. https://doi.org/10.1007/s10592-011-0301-2 (2012).
Thammina, C., Kidwell-Slak, D., Lura, S. & Pooler, M. SSR markers reveal the genetic diversity of asian Cercis taxa at the US National Arboretum. HortScience 52, 498–502. https://doi.org/10.21273/hortsci11441-16 (2017).
Chang, C., Bongarten, B. & Hamrick, J. Genetic structure of natural populations of black locust (Robinia pseudoacacia L.) at Coweeta, North Carolina. J. Plant Res. 111, 17–24. https://doi.org/10.1007/BF02507146.pdf (1998).
Marquardt, P. & Epperson, B. Spatial and population genetic structure of microsatellites in white pine. Mol. Ecol. 13, 3305–3315. https://doi.org/10.1111/j.1365-294X.2004.02341.x (2004).
Victory, E., Glaubitz, J., Rhodes-Jr, O. & Woeste, K. Genetic homogeneity in Juglans nigra (Juglandaceae) at nuclear microsatellites. Am. J. Bot. 93, 118–126. https://doi.org/10.3732/ajb.93.1.118 (2006).
Hadziabdic, D. et al. Genetic diversity of flowering dogwood in the Great Smoky Mountains National Park. Tree Genet. Genomes 8, 855–871. https://doi.org/10.1007/s11295-012-0471-1 (2012).
Nybom, H. Comparison of different nuclear DNA markers for estimating intraspecific genetic diversity in plants. Mol. Ecol. 13, 1143–1155. https://doi.org/10.1111/j.1365-294X.2004.02141.x (2004).
Donselman, H. Variation in native populations of eastern redbud (Cercis canadensis L.) as influenced by geographic location [USA]. In Proceedings, of the Florida State Horticulture Society Vol. 89. 370–373 (1976).
Dirr, M. Manual of Woody Landscape Plants: Their Identification, Ornamental Characteristics, Culture, Propagation and Uses (Stipes Publishing Co, Champaign, 1990).
Fritsch, P., Schiller, A. & Larson, K. Taxonomic implications of morphological variation in Cercis canadensis (Fabaceae) from Mexico and adjacent parts of Texas. Syst. Bot. 34, 510–520. https://doi.org/10.1600/036364409789271254 (2009).
Nevo, E. et al. Drought and light anatomical adaptive leaf strategies in three woody species caused by microclimatic selection at evolution canyon, Israel. Israel J. Plant Sci. 48, 33–46 (2000).
Fritsch, P. et al. Leaf adaptations and species boundaries in North American Cercis: Implications for the evolution of dry floras. Am. J. Bot. 105, 1577–1594. https://doi.org/10.1002/ajb2.1155 (2018).
Raulston, J. Redbud. Am. Nurseryman 171, 39–51 (1990).
Robertson, K. Cercis: The redbuds. Arnoldia 36, 37–49 (1976).
Davis, C., Fritsch, P., Li, J. & Donoghue, M. Phylogeny and biogeography of Cercis (Fabaceae): Evidence from nuclear ribosomal ITS and chloroplast ndhF sequence data. Syst. Bot. 27, 289–302. https://doi.org/10.1043/0363-6445-27.2.289 (2002).
Hopkins, M. In Rhodora Vol. 44 (eds M Fernald, C Eatherby, L Griscom, & S Marris) 193–211 (New England Botanical Club, Inc., 1942).
Griffin, J., Ranney, T. & Pharr, D. Heat and drought influence photosynthesis, water relations, and soluble carbohydrates of two ecotypes of redbud (Cercis canadensis). J. Am. Soc. Hortic. Sci. 129, 497–502. https://doi.org/10.21273/JASHS.129.4.0497 (2004).
Fritsch, P. & Cruz, B. Phylogeny of Cercis based on DNA sequences of nuclear ITS and four plastid regions: Implications for transatlantic historical biogeography. Mol. Phylogenet. Evol. 62, 816–825. https://doi.org/10.1016/j.ympev.2011.11.016 (2012).
Chung, M., Chung, M., Oh, G. & Epperson, B. Spatial genetic structure in a Neolitsea sericea population (Lauraceae). Heredity 85, 490–497. https://doi.org/10.1046/j.1365-2540.2000.00781.x (2000).
Dean, D. et al. Analysis of genetic diversity and population structure for the native tree Viburnum rufidulum occurring in Kentucky and Tennessee. J. Am. Soc. Hortic. Sci. 140, 523–531. https://doi.org/10.21273/JASHS.140.6.523 (2015).
Hagler, J., Mueller, S., Teuber, L., Machtley, S. & Van-Deynze, A. Foraging range of honey bees, Apis mellifera, in alfalfa seed production fields. J. Insect Sci. 11, 144. https://doi.org/10.1673/031.011.14401 (2011).
Pasquet, R. et al. Long-distance pollesn flow assessment through evaluation of pollinator foraging range suggests transgene escape distances. Proc. Natl. Acad. Sci. 105, 13456–13461 (2008).
Hayden, W. Redbud seedpods hold surprises. Bull. Virginia Native Plant Soc. 32, 1–6 (2013).
Schnabel, A., Laushman, R. & Hamrick, J. Comparative genetic structure of two co-occurring tree species, Maclura pomifera (Moraceae) and Gleditsia triacanthos (Leguminosae). Heredity 67, 357–364. https://doi.org/10.1038/hdy.1991.99 (1991).
Nakanishi, A., Tomaru, N., Yoshimaru, H., Manabe, T. & Yamamoto, S. Effects of seed- and pollen-mediated gene dispersal on genetic structure among Quercus salicina saplings. Heredity 102, 182–189. https://doi.org/10.1038/hdy.2008.101 (2008).
Vekemans, X. & Hardy, O. New insights from fine-scale spatial genetic structure analyses in plant populations. Mol. Ecol. 13, 921–935. https://doi.org/10.1046/j.1365-294X.2004.02076.x (2004).
Gonzales, E., Hamrick, J., Smouse, P., Trapnell, D. & Peakall, R. The impact of landscape disturbance on spatial genetic structure in the Guanacaste tree, Enterolobium cyclocarpum (Fabaceae). J. Hered. 101, 133–143. https://doi.org/10.1093/jhered/esp101 (2009).
Post, D. Change in nutrient content of foods stored by eastern woodrats (Neotoma floridana). J. Mammal. 73, 835–839 (1992).
Surrency, D. & Owsley, C. (ed. Natural Resources Conservation Service United States Department of Agriculture) 146 (United States Department of Agriculture, Natural Resources Conservation Service, 2001).
Wakeland, B. & Swihart, R. Ratings of white-tailed deer preferences for woody browse in Indiana. Proceedings of the Indiana Academy of Science 118, 96–101 (2009).
Wright, V., Fleming, E. & Post, D. Survival of Rhyzopertha dominica (Coleoptera, Bostrichidae) on fruits and seeds collected from woodrat nests in Kansas. J. Kansas Entomol. Soc. 63, 344–347 (1990).
Sullivan, J. (ed. Forest Service U.S. Department of Agriculture, Rocky Mountain Research Station) (U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station. Fire Sciences Laboratory, 1994).
Weir, B. & Ott, J. Genetic data analysis II. Trends Genet. 13, 379 (1997).
Magni, C., Ducousso, A., Caron, H., Petit, R. & Kremer, A. Chloroplast DNA variation of Quercus rubra L. in North America and comparison with other Fagaceae. Mol. Ecol. 14, 513–524. https://doi.org/10.1111/j.1365-294X.2005.02400.x (2005).
Peterson, B. & Graves, W. Chloroplast phylogeography of Dirca palustris L. indicates populations near the glacial boundary at the Last Glacial Maximum in eastern North America. Journal of Biogeography 43, 314–327, doi:https://doi.org/10.1111/jbi.12621 (2016).
Shaw, J. & Small, R. Chloroplast DNA phylogeny and phylogeography of the North American plums (Prunus subgenus Prunus section Prunocerasus, Rosaceae). Am. J. Bot. 92, 2011–2030. https://doi.org/10.3732/ajb.92.12.2011 (2005).
Rowe, K., Heske, E., Brown, P. & Paige, K. Surviving the ice: Northern refugia and postglacial colonization. Proc. Natl. Acad. Sci. 101, 10355–10359 (2004).
Graignic, N., Tremblay, F. & Bergeron, Y. Influence of northern limit range on genetic diversity and structure in a widespread North American tree, sugar maple (Acer saccharum Marshall). Ecol. Evol. 8, 2766–2780. https://doi.org/10.1002/ece3.3906 (2018).
Bemmels, J., Knowles, L. & Dick, C. Genomic evidence of survival near ice sheet margins for some, but not all, North American trees. Proc. Natl. Acad. Sci. 116, 8431–8436. https://doi.org/10.7302/Z2JS9NNG (2019).
Jia, H. & Steven, R. Fossil leaves and fruits of Cercis L. (Leguminosae) from the Eocene of western North America. International Journal of Plant Sciences 175, 601–612, doi:https://doi.org/10.1086/675693 (2014).
Kraemer, M. & Favi, F. Emergence phenology of Osmia lignaria subsp lignaria (Hymenoptera: Megachilidae), its parasitoid Chrysura kyrae (Hymenoptera: Chrysididae), and bloom of Cercis canadensis. Environ. Entomol. 39, 351–358. https://doi.org/10.1603/en09242 (2010).
USDA. Census of horticultural specialties. Volume 3 AC-12-SS-3, Washington, DC (2014).
Acknowledgements
This work was supported, in part, by the United States Department of Agriculture (USDA; Grant 58‐6062‐6), the USDA National Institute of Food and Agriculture (NIFA; Hatch project 1009630: TEN00495), and the University of Tennessee’s Open Publishing Support Fund. We are extremely grateful for the enthusiastic sampling help of more than 52 individual contributors, friends, and family who assisted us by gathering leaf tissue samples and collection coordinates for wild growing redbud trees across the U.S. Without each of your support, this effort would not have been possible. Also, we are sincerely thankful to Adrienne Gorny (Cornell University), David Held (Auburn University), Caterina Villari and Megan Buland (University of Georgia), Chris Wyman (University of Tennessee), Christine Nalepa and John Banask (North Carolina Department of Agriculture & Consumer Services), Cory Tanner (Clemson University), Donn Johnson and Lizabeth Herrera (University of Arkansas), Erfan Vafaie (Texas A&M University), Eric Day (Virginia Tech), Eric Rebek (Oklahoma State University), Erin Pfar (Rutgers University), Frank Hale (University of Tennessee), Gary Bachman (Mississippi State University), Grace Pietsch (University of Tennessee), Jackie Lee (University of Arkansas), Jason Griffin (Kansas State University), Juang-Horng Chong (Clemson University), John Olive (Auburn University), Katherine Kilbourne (Tennessee Department of Agriculture), Matt Ginzel and Geoffrey Williams (Purdue University), Michelle Clayson (Michigan, US), Natalie Diesel and Robbie Doerhoff (Missouri Department of Conservation), Nathan Schiff (United States Department of Agriculture, Mississippi), Philip Marshal (Vallonia State Nursery, Indiana), Raymond Moore (Tennessee Valley Authority, Alabama), Rob Pivar (University of Tennessee), Ron Winston (Florida), Sandra Wilson (University of Florida), Sarah White (Clemson University), Scott Goldman (Tennessee Department of Agriculture), Scott Ludwig (Amvac Chemical Corp., Texas), Shimat Joseph (University of Georgia), Stephen Clarke, Mr. Horne, Ms. Standard, and Mr. Gras (United States Department of Agriculture- Forest Service, Texas), Steve Meyers (Mississippi State University), Sydney Everhart and Eldon Everhart (University of Nebraska-Lincoln), and Will Hudson (University of Georgia).
Author information
Authors and Affiliations
Contributions
D.H., W.K., and R.T. conceived and designed the experiments including the major conceptual ideas and proof outline. All authors assisted in sample collection and preparation. M.O. carried out the experiments and S.B. contributed to the processes. M.O., M.N., S.B., and D.H. troubleshot technical details of the experiments. M.O., M.N., J.Z., and D.H. contributed to data analyses. SE provided technical advice on data preparation and analyses. All authors contributed to the interpretation of the results, manuscript writing, and editing. All authors provided critical feedback to shape the experiments, analyses, and finally produce the manuscript.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Additional information
Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Ony, M., Klingeman, W.E., Zobel, J. et al. Genetic diversity in North American Cercis Canadensis reveals an ancient population bottleneck that originated after the last glacial maximum. Sci Rep 11, 21803 (2021). https://doi.org/10.1038/s41598-021-01020-z
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-021-01020-z
This article is cited by
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.