- Research article
- Open Access
Genetic and morphological evidence for introgression between three species of willows
BMC Evolutionary Biology volume 15, Article number: 193 (2015)
Hybridization and introgression are said to occur relatively frequently in plants, and in particular among different species of willows. However, data on the actual frequency of natural hybridization and introgression is rare. Here, we report the first fine-scale genetic analysis of a contact zone shared between the three basket willow species, Salix dasyclados, S. schwerinii and S. viminalis in the vicinity of the Lake Baikal in Southern Siberia. Individuals were sampled in fourteen populations and classified as pure species or hybrids based on a set of morphological characters. They were then genotyped at 384 nuclear SNP and four chloroplast SSR loci. The STRUCTURE and NewHybrids softwares were used to estimate the frequency and direction of hybridization using genotypic data at the nuclear SNP loci.
As many as 19 % of the genotyped individuals were classified as introgressed individuals and these were mainly encountered in the centre of the contact zone. All introgressed individuals were backcrosses to S. viminalis or S. schwerinii and no F1 or F2 hybrids were found. The rest of the genotyped individuals were classified as pure species and formed two clusters, one with S. schwerinii individuals and the other with S. viminalis and S. dasyclados individuals. The two clusters were significantly genetically differentiated, with F ST = 0.333 (0.282–0.382, p < 0.001). In contrast, for the chloroplast haplotypes, no genetic differentiation was observed as they were completely shared between the species. Based on morphological classification only 5 % of the individuals were classified as introgressed individuals, which was much less than what was detected using genotypic data.
We have discovered a new willow hybrid zone with relatively high frequency of introgressed individuals. The low frequency of F1 hybrids indicates that ongoing hybridization is limited, which could be because of the presence of reproductive barriers or simply because the conditions are not favorable for hybridization. We further conclude that in order to get a complete picture of the species composition of a hybrid zone it is necessary to use a combination of morphological characters and genetic data from both nuclear and chloroplast markers.
Hybridization is a phenomenon that happens when mating between individuals from different species generates viable offspring and is a process that is estimated to occur in approximately 25 % of all plant species [1, 2]. Hybridization can sometimes lead to introgression if F1 hybrids backcross to one or both parental species and foreign genetic material is integrated into the genomes of either parent . If hybrids or backcrossed individuals have increased fitness and develop reproductive isolation, hybridization and introgression can lead to diversification and speciation [3–5], which has been demonstrated in for example wild sunflowers and Louisiana irises . In regions where previously geographically isolated species or populations meet, hybrid zones can form  if hybridization and introgression happens. Hybrid zones thus provide the opportunity to study evolutionary processes associated with interactions between species or differentiated populations. The genetic composition of hybrid zones depends to a large extent on the intensity of reproductive isolation and gene flow among interacting populations, which determines the frequency of F1 relative to introgressed F2 individuals [8–11].
In the past, hybrids were characterized solely based on morphology, however the use of morphological traits alone may not be sufficient to identify all hybrids or introgressed individuals , which was demonstrated by Hardig et al.  that failed to detect several hybrids between Salix eriocephala and S. sericea with only morphological characters. Therefore, in addition to morphological characteristics, molecular markers provide a good tool for assessing the degree of gene flow and genetic structure of hybrid zones. Due to the contrasting modes of inheritance of chloroplast and nuclear DNA they are expected to generate different patterns of genetic diversity within and among populations . Thus a combination of nuclear and chloroplast markers can be used to disentangle the complex mode of gene flow in hybrid zones. For example, maternally inherited chloroplast markers allow the assessment of the maternal parent of the hybrids and the direction of hybrid matings. In addition, chloroplast DNA should due to its smaller effective population size, be more sensitive to genetic drift and therefore display higher levels of population differentiation than nuclear DNA .
Willows (Salix; Salicaceae) is a plant genus with tree and shrub species that are common in riparian habitats worldwide. The genus contains about 400 species , however due to an unusually complex taxonomy this number has varied greatly over time . Willow species are notoriously difficult to classify as they are often morphologically similar and display large intraspecific diversity. In addition, hybridization might further complicate classification as hybrids can display unusual morphologies. The majority of willow species are diploid with a basis chromosome number of n = 19. However, ploidy levels can vary extensively both within and among species and range from diploids up to dodecaploids [17, 18] and possibly some of these polyploids are allopolyploids and the result of hybridization. The degree of hybridization has for centuries been a matter of great debate, and still is. In the late 1800’s and early 1900’s the general opinion was that hybridization was common and it was proposed that the great intraspecific morphological diversity was the result of hybridization . Still, some taxonomists believed that the role of hybridization was greatly exaggerated and that many so called hybrids were not hybrids, but mere variants within species. In order to fully understand the role of hybridization for the evolution of willows, more studies are indeed needed. Nevertheless, there are several well-known cases of hybridization, for example between S. fragilis and S. alba in Europe , S. sericea and S. eriocephala in the U.S.  and between species in the subg. Vetrix e.g. S. caprea, S. cinerea, S. atrocinerea and S. aurita [20, 21].
Salix viminalis L., S. dascyclados Wimm. and S. schwerinii E. Wolf are closely related basket willows in the subg. Vetrix and are tall shrubs or trees found along streams and rivers and in other wet habitats . Salix viminalis has a huge natural distribution , ranging from UK in the west to Siberia in the east (Fig. 1) and is the most well studied of the three species. It was in the past heavily used by humans for basketry and weaving but also for management of riverbanks and waterways . Trading of cutting materials has contributed to the dispersal of the species, which is the most likely reason why it is present in Sweden, despite the fact that it is not naturally occurring in Scandinavia . Salix viminalis and S. dasyclados are morphologically very similar and their distributions almost completely overlap (Fig. 1). According to Skvortsov , the two species have taxonomically been constantly confused and since hybrids between S. viminalis and other species of subg. Vetrix resemble S. dasyclados, S. dasyclados was also often regarded as a hybrid. However, Skvortsov  argued that no evidence exist that support a hybrid nature of S. dasyclados. S. schwerinii is found in East Asia and parallels S. dasyclados and S. viminalis in a north-south direction over a long distance across Siberia (Fig. 1). According to Skvortsov , S. viminalis and S. schwerinii meet in many geographical regions, however, they are rarely found growing together, and S. schwerinii is regarded as “vicarious” to S. viminalis. In a previous study, large genetic differentiation was found between the species as F ST was estimated to 0.56 and coalescent simulations suggested a divergence time of around 600,000 years, however sampling was done outside of the contact zone .
Interspecific crosses between S. viminalis × S. dasyclados and S. viminalis × S. schwerinii have been produced experimentally and pure species as well as hybrids have been extensively used in pre-breeding programs aiming at producing high-yielding varieties for biomass production in Europe. For example, an experimental cross between S. viminalis × (S. viminalis × S. schwerinii) [25, 26] has been used in several QTL studies of different traits [27–33]. Interestingly in a crossing experiment between S. viminalis × S. dasyclados it was found that the S. dasyclados parent was hexaploid and the offspring tetraploid , indicating that some S. dasyclados individuals could be allopolyploids possessing genomes from several species. However, observations indicate that these three species are predominantly diploid.
In the present study we performed fine-scale genetic analysis of a contact zone with the three species Salix dasyclados, S. schwerinii and S. viminalis located in the vicinity of the Lake Baikal in Southern Siberia. The main objectives were to morphologically and genetically characterize individuals in the contact zone and to estimate the frequency and direction of hybridization using 384 nuclear SNP and four chloroplast SSR markers together with a set of morphological characters. Samples were collected in fourteen populations. We specifically asked the following questions: i) Is the contact zone a hybrid zone, that is, do we find hybrids or introgressed inidviduals? ii) If yes, what is the population structure of the hybrid zone? Do we primarily find intermediate F1 hybrids or introgressed individuals? iii) Is introgression uni- or bidirectional? iv) Can morphology be used to predict the structure of the hybrid zone? v) What are the characteristics of the reproductive barriers between the species?
A total of 375 individuals were collected from fourteen populations and based on morphological characterization nineteen individuals (5 %) were classified as hybrids.
Eleven were classified as S. viminalis × S. dasyclados hybrids, seven as S. viminalis × S. schwerinii hybrids and one as either a S. viminalis × S. dasyclados hybrid or a S. schwerinii × S. dasyclados hybrid (Additional file 1). Twenty-two individuals were classified as pure S. dasyclados, 123 as pure S. schwerinii and 209 as pure S. viminalis (for two individuals info was missing) (Additional file 1). Hybrids were encountered in eight different populations, two populations were composed of only S. schwerinii individuals and four populations were composed of only S. viminalis individuals (Additional file 1). The populations with hybrids were located in the central parts of the sampled area, pure S. schwerinii populations were located in the eastern parts and populations with only S. viminalis were located in the western parts (Fig. 2).
Nuclear SNP variation and population structure
Seventy-nine individuals were successfully genotyped at 323 of the 384 nuclear SNP loci, of which 206 were polymorphic in at least one of the individuals. Both pure species and hybrids were present among these 79 individuals (two were classified as S. dasyclados, 45 as S. viminalis, 27 as S. schwerinii, two as S. viminalis × S. dasyclados hybrids and one as a S. schwerinii × S. dasyclados hybrid (for two individuals info was missing), thus they form a representative subset of all collected samples. Two SNPs in S. schwerinii and three in S. viminalis deviated significantly from Hardy-Weinberg equilibrium (p ≤ 0.01) (Additional file 2). One hundred and forty-six SNPs were polymorphic in both species, 40 were polymorphic only in S. viminalis and 20 were polymorphic only in S. schwerinii. No fixed SNPs were found. F ST was estimated to 0.057 (AMOVA, 95 % CI: 0.033–0.075, p < 0.001) in S. viminalis, 0.101 (AMOVA, 95 % CI: 0.079–0.123, p < 0.001) in S. schwerinii and 0.333 between the species (0.282–0.382, p < 0.001). When all samples were included in clustering analysis with the STRUCTURE software [35–38], the number of clusters (K) was estimated to 2, both when the original method from Pritchard et al.  was used and with the ΔK statistics given in Evanno et al.  (Additional file 3). Sixty-four individuals had Q-values ≥ 0.9 and 44 of these were assigned to cluster 1 and twenty were assigned to cluster 2 (Fig. 3a). Based on the morphological classification, all but two individuals that originated from cluster 1 were S. viminalis. The two other individuals were S. dasyclados, meaning that genetically S. dasyclados and S. viminalis are indistinguishable. In cluster 2, nineteen out of the twenty individuals were morphologically S. schwerinii and one was a S. viminalis × S. schwerinii hybrid. Individuals of S. viminalis origin were found in populations in the western parts of the sampled area and individuals with S. schwerinii origin were found in populations in the eastern parts (Fig. 2).
Fifteen of the genotyped individuals (19 %) had Q-values between 0.1 and 0.9 and were identified as hybrids (Table 1; Fig. 3a; Additional file 4). These individuals derive from six populations located in the centre of the sampling area (Fig. 2). Morphologically, only two were classified as hybrids (S. viminalis × S. dasyclados), while eight were classified as S. schwerinii and three as S. viminalis (for two others, morphological classification was missing) (Table 1). As no Q-value was close to 0.5 (Table 1), this was interpreted as no F1 hybrid was present among the individuals. Instead, their Q-values suggest that they were backcrosses both to S. viminalis and S. schwerinii. Thirteen of the fifteen introgressed individuals had posterior probabilities > 0.9 (Table 1; Fig. 3b), of which seven were classified as S. schwerinii backcrosses, four as S. viminalis backcrosses and two as pure S. viminalis, while none was classified as F1 or F2 hybrids (cross between two F1s).
Chloroplast SSR variation
Seventy individuals originating from Europe and the Baikal region were successfully genotyped with the four chloroplast SSR markers. As expected, one allele per locus and individual was found for each of the markers. A total of seventeen alleles were detected and ccmp2 was the most variable with six alleles and ccmp4 the least variable with two alleles, ccmp5 had four and ccmp6 had five alleles. When combined, these alleles formed 24 different haplotypes. When only including the individuals from the Baikal region (N = 50) the PhiPT (an analog to F ST , i.e genetic differentiation among populations) was estimated to -0.023 (p = 0.647) in S. viminalis, to 0.107 (p = 0.100) in S. schwerinii and to -0.006 (p = 0. 371) between the species. Thus, the chloroplast haplotypes did not display any significant genetic differentiation, either within or between the species (Fig. 2; Additional file 5). No sequence variation was found in the two sequenced chloroplast regions (trnL and rbcL).
The three closely related basket willow species Salix dasyclados, Salix schwerinii and Salix viminalis are phenotypically similar and several morphological characteristics are needed to distinguish them from each other. Given that the three species recently diverged, it is possible that they would hybridize where they come into contact in nature, unless some strong reproductive barriers have evolved. To study this, individuals from all three species were sampled in fourteen populations in a region in Southern Siberia where the distributions of the species meet and overlap. 81 % of the genotyped individuals were assigned to one of two clusters; one composed of S. schwerinii and one of S. viminalis and S. dasyclados individuals. Relatively high level of genetic differentiation was observed between the clusters (F ST of 0.333), suggesting that they are largely reproductively isolated. However, 19 % of the genotyped individuals were identified as hybrids, which shows that the species boundaries are partly permeable. Both the STRUCTURE and the NewHybrid analyses showed that none of these individuals were intermediate F1 hybrids but backcrosses to either S. schwerinii or S. viminalis. This observation demonstrates that introgression is bidirectional in this willow hybrid zone. Furthermore, the absence of F1 hybrids is striking and is best explained by the presence of pre- and/or postmating reproductive barriers that form reproductive isolation between the species . Currently, we can only speculate about the nature of these barriers in this system and the relative contribution of different barriers. Putative premating barriers are differences in flower phenologies or pollinator specificity  and as willows generally are both wind and insect pollinated [42, 43], both are likely to be involved. We do not know whether or not S. schwerinii and S. viminalis display asynchronous flowering in the region of sampling. However, at our field site outside of Uppsala in Sweden (59.80° N, 17.67° E, 25 m AOD), S. schwerinii plants flower considerably earlier than S. viminalis plants, bearing in mind that those S. viminalis plants originate from Central Europe and not from the region near Lake Baikal. Possibly, postmating barriers are operating together with premating barriers to produce strong reproductive isolation, as premating barriers alone are often permeable and leaky . However as artificial crosses between the species are routinely produced in our laboratory and F1 hybrids are fertile and also used to backcross to either parents, strong postmating barriers are unlikely to be present. The age of the hybrid zone may also affect the occurrence of F1 hybrids as in a very young zone, not enough time may have elapsed for intermediate hybrids to reach larger numbers. However as we found a substantial number of backcrossed individuals, the hybrid zone could not be very young. Perhaps instead, the hybrid zone is relatively old and the intermediate hybrids have been lost as they are no longer being produced or produced at very low frequencies. It could also be, that environmental conditions and the relative proportion of species  were conducive to hybridization during the formation of the hybrid zone but are no longer so.
An interesting observation was that chloroplast haplotypes were shared between the species and no genetic differentiation was found, even when including samples from Europe. This could indeed be the result of homoplasy, however as extensive haplotype sharing has been reported in other willow [13, 45, 46] and tree species, e.g. Betula spp.  and Quercus spp.  using entirely different markers, it is unlikely that homoplasy alone could create the observed pattern. Instead, haplotype sharing is likely the result of past hybridization and/or introgression. The observation that none of the SNPs were fixed between the species furthermore supports the scenario that introgression happens where the two species meet. While this may well be the case, our data also suggest that to some degree the shared haplotypes and polymorphisms are ancestral polymorphism. Shared ancestral polymorphism between species is expected from incomplete lineage sorting, which is particularly likely in species such as the willows that have large effective population sizes as it under those circumstances take very long time for alleles to go to fixation . In a previous study, using sequences from S. viminalis and S. schwerinii individuals from Europe and Siberia, respectively, i.e. long way from the hybrid zone, divergence was estimated at 600,000 ya . The effective population size in our previous study was estimated to 40,000 in both species and the generation time assumed to be 10 years. Thus, the divergence time would be roughly 1.5 Ne and would be indeed insufficient to observe reciprocal monophyly. This shows that even if gene flow is not currently taking place, we expect to observe shared polymorphism due to historical gene flow and/or incomplete lineage sorting.
The chloroplast and nuclear DNA gave contrasting results as there was no significant genetic differentiation with the chloroplast haplotypes, while with the nuclear genotypic data there was relatively high level of genetic differentiation. This pattern is a likely consequence of the differences in intraspecific gene flow that the nuclear and the chloroplast DNA experience . Since nuclear DNA is transmitted by both pollen and seeds and as chloroplast DNA is transmitted only by seeds, intraspecific gene flow will be higher for nuclear DNA compared to chloroplast DNA. Introgressed nuclear alleles will therefore be diluted and maintained at low frequencies, however introgressed chloroplast haplotypes will not be diluted to the same degree and can therefore persist in the population. This effect should be particularly significant in species where pollen can travel long distances by wind and where seeds are less prone to long distance dispersal. Willow seeds are also primarily wind dispersed, , however compared to pollen, seeds are expected to travel less far.
Compared to the 19 % of the genotyped individuals that were assessed to be of hybrid origin, only 5 % were classified as hybrids using traditional morphological characteristics. Most noteworthy was that eight individuals that were classified as pure S. schwerinii were in fact introgressed individuals. This demonstrates the difficulties and uncertainties in detecting hybrids solely based on morphology, which could be particularly difficult at low levels of introgression. Another surprising finding was that individuals classified as S. dasyclados clustered with individuals classified as S. viminalis. As we were unable to distinguish the species from each other with this large set of markers, one conclusion could be that the sampled individuals should be regarded as one and the same species. Still, they display enough morphological differences to be classified as different species by trained botanists. In order to solve this, more samples need to be analysed, both genetically and morphologically.
We have discovered a new hybrid zone between three willow species with rather high frequency of introgressed individuals, which were identified both with morphological characters and genotypic data. We found extensive sharing of chloroplast haplotypes between the species and a large number of shared nuclear polymorphisms, a pattern that is expected from introgression but also from ancestral polymorphisms that remain in the populations due to incomplete lineage sorting. We also demonstrated rather high levels of reproductive isolation between S. viminalis and S. schwerinii. This was further supported by the lack of intermediate F1 hybrids, which could be a consequence of some of reproductive barrier(s). In addition we found that introgression was bidirectional so that backcrosses happen with both S. schwerinii and S. viminalis. Another surprising observation was the genetic similarities between S. dasyclados and S. viminalis, which leads us to doubt the taxonomic status of S. dasyclados.
Sampling and morphological classification
Shoots and leaves from S. dasyclados, S. schwerinii and S. viminalis and putative hybrids were collected from plants in the field in July 2012. Samples were collected in fourteen different populations (Table 2 and Fig. 2) in the vicinity of the Lake Baikal in Southern Siberia, Russia. In each population, between 25 and 30 individuals were collected (Additional file 1). If all species were present in one population, the number of individuals collected per species approximately corresponded to the relative abundance of that species to the other. In the laboratory, a herbarium collection was established and each individual was carefully examined and classified as pure species or hybrids. The species and hybrids were morphologically discriminated using a set of characters that are listed in Additional file 6. Leaf tissue samples for each individual was prepared, dried in silica gel and shipped to Uppsala, Sweden for genetic analyses.
DNA extractions and genotyping
Genomic DNA from dried leaf tissue samples was isolated using the DNeasy Mini Plant Kit (Qiagen, Crawley, UK). Genotypes were determined at 384 previously developed single nucleotide polymorphism (SNP) loci  with the Illumina GoldenGate Assay at the SNP&SEQ Technology Platform, Science for Life Laboratory at Uppsala University, Sweden. The SNP markers were identified in S. viminalis and S. schwerinii individuals and have random genomic locations . Alleles were scored at four chloroplast SSR loci; ccmp2, ccmp4, ccmp5 and ccmp6 . The chloroplast SSR markers were amplified by PCR, multiplexed and separated on an ABI3730XL instrument at the Uppsala Genome Center, Science for Life Laboratory at Uppsala University, Sweden. The PCRs were performed in 10 μl reactions with 10 ng of DNA, 1xPCR buffer (Qiagen, Stockholm, Sweden), 0.75 mM MgCl2 (Qiagen, Crawley, UK), 0.9 mM of a mix of dNTPs (Applied Biosystems, Stockholm, Sweden), 0.1 μM of each primer (Life Technologies, Stockholm, Sweden) and 0.25 U Taq HotStar Plus (Qiagen, Stockholm, Sweden). The PCR program included 95 °C for 5 min, then 30 cycles at 95 °C for 30 s, Tm for 30 s, 72 °C for 30 s, then 8 °C on hold. Tm was 58 °C for ccmp2, ccmp4 and ccmp6 and 56 °C for ccmp5. The forward primers were fluorescently labeled at the 5´end; ccmp2 with 6-FAM, ccmp4 with NED, ccmp5 with VIC and ccmp6 with PET. Allele sizes were determined with the PeakScanner TM Software v1.0 (Life Technologies, Stockholm, Sweden). In addition to individuals sampled in the fourteen populations in Southern Siberia another twenty samples, morphologically classified as S. viminalis were scored with the chloroplast SSRs; four originated from western Russia and sixteen originated from Europe. Two chloroplast regions were sequenced and analysed for sequence variation; an intron in the tRNA-Leu gene (trnL) amplified with primers c and d  and a fragment in the rbcL gene (Rubisco) amplified with primers carbcLF and carbcLR . The PCRs were performed in 30 μl reactions with approximately 10 ng of DNA, 1xPCR buffer (Qiagen Crawley, UK), 2 mM MgCl2 (Qiagen Crawley, UK), 0.8 mM of a mix of dNTPs (Applied Biosystems Thermo Fisher Scientific, Stockholm, Sweden), 0.2 μM of each primer (Life Technologies Thermo Fisher Scientific, Stockholm, Sweden) and 1.5 U Amplitaq GOLD (Applied Biosystems, Stockholm, Sweden). The PCR program included 94 °C for 5 min, five cycles of 94 °C for 30 s, 65-60 °C for 30 s, (- 1 °C every cycle), then 25 cycles at 94 °C for 30 s, 60 °C for 30 s and 72 °C for 90 s, then 72 °C for 10 min and finally 10 °C on hold. Amplification was determined by agarose gel electrophoresis. PCR products were cleaned and sequenced at Macrogen Inc. (Macrogen, Seoul, South Korea) on both strands using the forward and reverse PCR primers as sequencing primers. The chromatograms were visually inspected and edited using Lasergene SeqMan vs 9.1.0 (DNASTAR).
The R package  adegenet v1.4-2 [55, 56] was used to calculate observed and expected heterozygosity and departure from Hardy-Weinberg equilibrium for each of the nuclear SNP markers. Genetic divergence was estimated as F ST both within and between S. schwerinii and S. viminalis populations using Arlequin v. 3.5 . Shared and fixed number of SNP loci between and within S. schwerinii and S. viminalis were estimated. For these analyses, individuals were grouped based on morphological characteristics and the hybrids and S. dasyclados were not included.
Genetic differentiation was investigated using Bayesian cluster analyses with the STRUCTURE software v 184.108.40.206 [35–38]. An admixture model was employed where correlated allele frequencies were assumed, and the K-value (i.e. the number of clusters) was set from one to eight, a value within this range being expected to be the most likely result. The length of the burn-in period was set to 10,000 and the Monte Carlo Markov Chain (MCMC) model after burn-in was run for an additional 100,000 iterations. For each K, 40 replicates were run and runs with an outlier value of lnPD were removed (χ 2-test, α = 0.05, as implemented in R-package outliers). The optimal value of K was determined by examination of the L(K) and Evanno’s ∆K statistics  using the R package ‘CorrSieve’ .
For each of the four chloroplast SSR markers, allele numbers and size ranges were assessed. Alleles were combined in haplotypes for every individual. PhiPT within and between S. schwerinii and S. viminalis was assessed using GenAIEx v 6.5 [59, 60]. Haplotypes in each population were plotted in Fig. 2. All maps were created using ArcGIS® software by Esri. ArcGIS® and ArcMap™ are the intellectual property of Esri and are used herein under license. Copyright © Esri. All rights reserved. For more information about Esri® software, please visit www.esri.com.
Identification and classification of hybrids
Hybrids and backcrosses were identified using the nuclear genotypic data and the softwares STRUCTURE [35–38] and NewHybrids [61, 62]. The Q-value from STRUCTURE, which shows the proportion of an individual’s genome that originates from each of the K clusters, was used as hybrid index (similar to what was done in Devitt et al. ). Thus, for every K, individuals with Q-values between 0.1 and 0.9 were considered to be hybrids or backcrosses, while individuals with Q-values ≥ 0.9 were considered to be pure species. NewHybrids was used to classify individuals as pure species, F1s, F2s or backcrosses. NewHybrids computes the posterior probabilities that an individual belongs to these different classes. For NewHybrids, we used the default genotype categories for first- and second-generations of crossing and ran 100,000 sweeps of five chains started from overdispersed starting values after a burn-in period of 50,000 sweeps following the software author’s recommendation. Jeffrey-type priors were used for the mixing proportions and allele frequencies. To check for convergence, we visually inspected P(z) values from the different runs which were then averaged across the 5 runs.
Availability of supporting data
The data set supporting the results of this article is archived in LabArchives (doi: 10.6070/H48P5XJ2).
Ellstrand NC, Whitkus R, Rieseberg LH. Distribution of spontaneous plant hybrids. Proc Natl Acad Sci U S A. 1996;93:5090–3.
Mallet J. Hybridization as an invasion of the genome. Trends Ecol Evol. 2005;20:229–37.
Baack EJ, Rieseberg LH. A genomic view of introgression and hybrid speciation. Curr Opin Genet Dev. 2007;17:513–8.
Hegarty MJ, Hiscock SJ. Hybrid speciation in plants: new insights from molecular studies. New Phytol. 2005;165:411–23.
Abbott R, Albach D, Ansell S, Arntzen JW, Baird SJ, Bierne N, et al. Hybridization and speciation. J Evol Biol. 2013;26:229–46.
Arnold ML, Buckner CM, Robinson JJ. Pollen-mediated introgression and hybrid speciation in Louisiana irises. Proc Natl Acad Sci U S A. 1991;88:1398–402.
Harrison RG. Hybrid zones: windows on evolutionary process. Oxf Surv Evol Biol. 1990;7:69–128.
Broyles SB. Hybrid bridges to gene flow: a case study in milkweeds (Asclepias). Evolution. 2002;56:1943–53.
Ma YP, Milne RI, Zhang CQ, Yang JB. Unusual patterns of hybridization involving a narrow endemic Rhododendron Species (Ericaceae) in Yunnan, China. Am J Bot. 2010;97:1749–57.
Nagano Y, Hirao AS, Itino T: Genetic structure of a hybrid zone between two violets, Viola rossii Hemsl. and V. bissetii Maxim.: dominance of F1 individuals in a narrow contact range. Plant Species Biol 2014.
Milne RI, Terzioglu S, Abbott RJ. A hybrid zone dominated by fertile F1s: maintenance of species barriers in Rhododendron. Mol Ecol. 2003;12:2719–29.
Rieseberg LH, Ellstrand NC. What can molecular and morphological markers tell us about plant hybridization. Crit Rev Plant Sci. 1993;12:213–41.
Hardig TM, Brunsfeld SJ, Fritz RS, Morgan M, Orians CM. Morphological and molecular evidence for hybridization and introgression in a willow (Salix) hybrid zone. Mol Ecol. 2000;9:9–24.
Petit RJ, Duminil J, Fineschi S, Hampe A, Salvini D, Vendramin GG. Comparative organization of chloroplast, mitochondrial and nuclear diversity in plant populations. Mol Ecol. 2005;14:689–701.
Renner SS. The relative and absolute frequencies of angiosperm sexual systems: Dioecy, monoecy, gynodioecy, and an updated online database. Am J Bot. 2014;101:1588–96.
Skvortsov AK. Willows of Russia and adjacent countries. Taxonomical and geographical review (translated by Kadis, IN from: Skvortsov AK (1968) Willows of the USSR. Taxonomic and Geographic Revision. Nauka, Moscow). In: University of Joensuu Faculty of Mathematics and Natural Sciences Report Series. Edited by Erkama T, Simola H, vol. 39. Joensuu, Finland; 1999.
Suda Y, Argus GW. Chromosome numbers of some North American Salix. Brittonia. 1968;20:191–7.
Zsuffa L, Mosseler A, Raj Y. Prospects for interspecific hybridization in willow for biomass production. In: Perttu KL, editor. Ecology and management of forest biomass production systems. Uppsala: Swedish University of Agricultural Sciences; 1984. p. 261–81. Report 15.
Triest L, De Greef B, Vermeersch S, Van Slycken J, Coart E. Genetic variation and putative hybridization in Salix alba and S. fragilis (Salicaceae): evidence from allozyme data. Plant Syst Evol. 1999;215:169–87.
Reichinger KH. Salix taxonomy in Europe - problems, interpretations and observations. Proc R Soc Edinb. 1992;98B:1–12.
Meikle RD. Britsh willows; some hybrids and some problems. Proc R Soc Edinb. 1992;98B:13–20.
Kuzovkina Y, Weih M, Abalos Romero M, Belyaeva I, Charles J, Hurst S, et al. Salix: Botany and global horticulture. Hortic Revs. 2008;34:447–89.
Lascoux M, Thorsen J, Gullberg U. Population structure of a riparian willow species, Salix viminalis L. Genet Res. 1996;68:45–54.
Berlin S, Fogelqvist J, Lascoux M, Lagercrantz U, Rönnberg-Wästljung AC. Polymorphism and divergence in two willow species, Salix viminalis L. and Salix schwerinii E. Wolf. G3: Genes Genom Genet. 2011;1:387–400.
Berlin S, Lagercrantz U, von Arnold S, Öst T, Rönnberg-Wästljung AC. High-density linkage mapping and evolution of paralogs and orthologs in Salix and Populus. BMC Genomics. 2010;11:129.
Tsarouhas V, Gullberg U, Lagercrantz U. An AFLP and RFLP linkage map and quantitative trait locus (QTL) analysis of growth traits in Salix. Theor Appl Genet. 2002;105:277–88.
Tsarouhas V, Gullberg U, Lagercrantz U. Mapping of quantitative trait loci controlling timing of bud flush in Salix. Hereditas. 2003;138:172–8.
Tsarouhas V, Gullberg U, Lagercrantz U. Mapping of quantitative trait loci (QTLs) affecting autumn freezing resistance and phenology in Salix. Theor Appl Genet. 2004;108:1335–42.
Rönnberg-Wästljung AC, Glynn C, Weih M. QTL analyses of drought tolerance and growth for a Salix dasyclados x Salix viminalis hybrid in contrasting water regimes. Theor Appl Genet. 2005;110:537–49.
Weih M, Rönnberg-Wästljung AC, Glynn C. Genetic basis of phenotypic correlations among growth traits in hybrid willow (Salix dasyclados x S. viminalis) grown under two water regimes. New Phytol. 2006;170:467–77.
Ghelardini L, Berlin S, Weih M, Lagercrantz U, Gyllenstrand N, Rönnberg-Wästljung AC. Genetic architecture of spring and autumn phenology in Salix. BMC Plant Biol. 2014;14:31.
Berlin S, Ghelardini L, Bonosi L, Weih M, Rönnberg-Wästljung AC. QTL mapping of biomass and nitrogen economy traits in willows (Salix spp.) grown under contrasting water and nutrient conditions. Mol Breeding. 2014;34:1987–2003.
Samils B, Rönnberg-Wästljung AC, Stenlid J. QTL mapping of resistance to leaf rust in Salix. Tree Genet Genomes. 2011;7:1219–35.
Rönnberg-Wästljung AC, Tsarouhas V, Semerikov V, Lagercrantz U. A genetic linkage map of a tetraploid Salix viminalis x S. dasyclados hybrid based on AFLP markers. For Genet. 2003;10:185–94.
Pritchard JK, Stephens M, Donnelly P. Inference of population structure using multilocus genotype data. Genetics. 2000;155:945–59.
Falush D, Stephens M, Pritchard JK. Inference of population structure using multilocus genotype data: linked loci and correlated allele frequencies. Genetics. 2003;164:1567–87.
Falush D, Stephens M, Pritchard JK. Inference of population structure using multilocus genotype data: dominant markers and null alleles. Mol Ecol Notes. 2007;7:574–8.
Hubisz MJ, Falush D, Stephens M, Pritchard JK. Inferring weak population structure with the assistance of sample group information. Mol Ecol Resour. 2009;9:1322–32.
Evanno G, Regnaut S, Goudet J. Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study. Mol Ecol. 2005;14:2611–20.
Coyne JA, Orr HA. Speciation. Sunderland: Sinauer Associates Inc.; 2004.
Widmer A, Lexer C, Cozzolino S. Evolution of reproductive isolation in plants. Heredity. 2009;102:31–8.
Vroege PW, Stelleman P. Insect and wind pollination in Salix repens L. and Salix caprea L. Israel J Bot. 1990;39:125–32.
Argus GW. An experimental study of hybridization and pollination in Salix (willow). Can J Bot. 1974;52:1613–9.
Currat M, Ruedi M, Petit RJ, Excoffier L. The hidden side of invasions: massive introgression by local genes. Evolution. 2008;62:1908–20.
Palmé AE, Semerikov V, Lascoux M. Absence of geographical structure of chloroplast DNA variation in sallow, Salix caprea L. Heredity. 2003;91:465–74.
Percy DM, Argus GW, Cronk QC, Fazekas AJ, Kesanakurti PR, Burgess KS, et al. Understanding the spectacular failure of DNA barcoding in willows (Salix): does this result from a trans-specific selective sweep? Mol Ecol. 2014;23:4737–56.
Palmé AE, Su Q, Palsson S, Lascoux M. Extensive sharing of chloroplast haplotypes among European birches indicates hybridization among Betula pendula, B. pubescens and B. nana. Mol Ecol. 2004;13:167–78.
Petit RJ, Csaikl UM, Bordacs S, Burg K, Coart E, Cottrell J, et al. Chloroplast DNA variation in European white oaks - Phylogeography and patterns of diversity based on data from over 2600 populations. Forest Ecol Manag. 2002;156:5–26.
Hudson RR, Coyne JA. Mathematical consequences of the genealogical species concept. Evolution. 2002;56:1557–65.
Steyn EMA, Smith GF, Van Wyk AE. Functional and taxonomic significance of seed structure in Salix mucronata (Salicaceae). Bothalia. 2004;34:53–9.
Weising K, Gardner RC. A set of conserved PCR primers for the analysis of simple sequence repeat polymorphisms in chloroplast genomes of dicotyledonous angiosperms. Genome. 1999;42:9–19.
Taberlet P, Gielly L, Pautou G, Bouvet J. Universal primers for amplification of three non-coding regions of chloroplast DNA. Plant Mol Biol. 1991;17:1105–9.
Hamza-Babiker N, Heinze B, Glössl J, Arnold C. Chloroplast DNA identification of eight closely related European Salix species. Austrian J For Sci. 2009;126:175–93.
R Core Team. R: A language and environment for statistical computing. Vienna: R Foundation for Statistical Computing; 2014.
Jombart T. adegenet: a R package for the multivariate analysis of genetic markers. Bioinformatics. 2008;24:1403–5.
Jombart T, Ahmed I. adegenet 1.3-1: new tools for the analysis of genome-wide SNP data. Bioinformatics. 2011;27:3070–1.
Excoffier L, Lischer HE. Arlequin suite ver 3.5: a new series of programs to perform population genetics analyses under Linux and Windows. Mol Ecol Resour. 2010;10:564–7.
Campana MG, Hunt HV, Jones H, White J. CorrSieve: software for summarizing and evaluating Structure output. Mol Ecol Resour. 2011;11:349–52.
Peakall R, Smouse PE. GenAlEx 6.5: Genetic analysis in Excel. Population genetic software for teaching and research - an update. Bioinformatics. 2012;28:2537–9.
Peakall R, Smouse PE. GENALEX 6: genetic analysis in Excel. Population genetic software for teaching and research. Mol Ecol Notes. 2006;6:288–95.
Anderson EC, Thompson EA. A model-based method for identifying species hybrids using multilocus genetic data. Genetics. 2002;160:1217–29.
Anderson EC. Bayesian inference of species hybrids using multilocus dominant genetic markers. Phil Trans R Soc B. 2008;363:2841–50.
Devitt TJ, Baird SJE, Moritz C. Asymmetric reproductive isolation between terminal forms of the salamander ring species Ensatina eschscholtzii revealed by fine-scale genetic analysis of a hybrid zone. BMC Evol Biol. 2011;11:245.
SNP genotyping was performed by the SNP&SEQ Technology Platform, Science for Life Laboratory at Uppsala University, a national infrastructure supported by the Swedish Research Council (VR-RFI) and the Knut and Alice Wallenberg Foundation. This research was financed by a Project grant (2011-3544) for junior researchers to SBK from the Swedish Research Council. We thank the three anonymous reviewers for insightful comments that greatly improved the manuscript.
The authors declare that they have no competing interests.
JF carried out the genetic and statistical analyses and helped to draft the manuscript. AVV and AIK collected and morphologically characterized the plant material. PP participated in the in house lab work. CD and ACRW helped to draft the manuscript. ML participated in the design of the study and helped to draft the manuscript. SB acquired funding, designed the study, carried out in house lab work and drafted the manuscript. All authors read and approved the final manuscript.
Collected samples and classification. (XLSX 14 kb)
H O and H E and test of Hardy-Weinberg equilibrium for every marker for S. schwerinii and S. viminalis separately. (XLSX 52 kb)
lnPD and deltaK from the Bayesian cluster analyses with the STRUCTURE software. (PDF 672 kb)
Q -values with K = 2 (STRUCTURE) and posterior probabilities (NewHybrids). (XLSX 15 kb)
Chloroplast DNA haplotype distribution across Europe and Russia. (PNG 236 kb)
Morphological characters used to classify species as S. dasyclados, S. schwerinii or S. viminalis. (XLS 25 kb)
About this article
Cite this article
Fogelqvist, J., Verkhozina, A.V., Katyshev, A.I. et al. Genetic and morphological evidence for introgression between three species of willows. BMC Evol Biol 15, 193 (2015). https://doi.org/10.1186/s12862-015-0461-7
- Reproductive Isolation
- Hybrid Zone
- Single Nucleotide Polymorphism Marker
- Reproductive Barrier
- Incomplete Lineage Sorting