SELECTION AND DISPERSAL IN A MULTISPECIES OAK HYBRID ZONE Author(s): Richard S. Dodd and Zara Afzal-Rafii Source: Evolution, 58(2):261-269. 2004. Published By: The Society for the Study of Evolution DOI: 10.1554/03-288 URL: http://www.bioone.org/doi/full/10.1554/03-288

BioOne (www.bioone.org) is an electronic aggregator of bioscience research content, and the online home to over 160 journals and books published by not-for-profit societies, associations, museums, institutions, and presses. Your use of this PDF, the BioOne Web site, and all posted and associated content indicates your acceptance of BioOne’s Terms of Use, available at www.bioone.org/page/terms_of_use. Usage of BioOne content is strictly limited to personal, educational, and non-commercial use. Commercial inquiries or rights and permissions requests should be directed to the individual publisher as copyright holder.

BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers, academic institutions, research libraries, and research funders in the common goal of maximizing access to critical research.

Evolution, 58(2), 2004, pp. 261–269

SELECTION AND DISPERSAL IN A MULTISPECIES OAK HYBRID ZONE RICHARD S. DODD1,2

AND

ZARA AFZAL-RAFII3,4

1 Department

of Environmental Science Policy and Management, University of California, 145 Mulford Hall, Berkeley, California 94720 2 E-mail: [email protected] 3 IMEP, Faculte ´ des Sciences St. Je´roˆme, Universite´ d’Aix-Marseille III, 13397 Marseille, France 4 E-mail: [email protected]

Abstract. The four western North American red oak species (Quercus wislizeni, Q. parvula, Q. agrifolia, and Q. kelloggii) are known to produce hybrid products in all interspecific combinations. However, it is unknown whether hybrids are transitory resulting from interspecific gene flow or whether they are maintained through extrinsic selection. Here, we examine cryptic hybrid structure in Q. wislizeni through a broad region including contact and isolation from three other western North American red oaks using amplified fragment length polymorphism molecular markers. All four species were simultaneously detected in the genetic background of individuals morphologically assigned to Q. wislizeni, although the contribution of Q. kelloggii was minor. In some cases, introgression was detected well outside the region of sympatry with one of the parental species. Molecular structure at the individual level indicated this was due to long-distance pollen dispersal and not to local extinction of parental species. Species admixture proportions were correlated with climatic variables and greater proportions of Q. agrifolia and Q. parvula were present in the genetic background of Q. wislizeni in sites with cooler and more humid summers, corresponding with habitat preferences of the parental species. Partial Mantel tests indicated that climate was more important than distance from pollen source in this association. Despite high levels of introgression, species integrity was maintained in some populations in close proximity to the other species, providing further support to environmental selection in determining population genetic structure. Thus, the contribution of species mixtures to population genetic structure varies across the landscape according to availability of pollen, but more importantly to varying environmental selection pressures that produce a complex pattern of hybrid and pure gene pools. Key words.

Climate, hybridization, introgression, pollen dispersal, Quercus wislizeni, red oaks. Received May 15, 2003.

Accepted September 9, 2003.

The extent of hybrid zones depends on the balance between dispersal that leads to recombinant genotypes and selection that sieves out parental and hybrid genotypes according to their relative fitnesses. Interpretation of the importance of dispersal and selection and the likelihood of hybrids being fitter than parental types has fueled the debate over the importance of hybrid zones in evolution. At one extreme, reinforcing selection is seen as maintaining species isolation so that hybrid zones are narrow and transient (Mayr 1963; Wagner 1969, 1970; Hardin 1975). At the other extreme hybrid zones are viewed as a very potent source of genetic recombination and diversity (Anderson 1949; Stebbins 1959; Arnold 1997). Because parental lineages are the only gene pools that have experienced long-term selection, it is commonly assumed that, unless environments change, hybrids will be less fit and will be selected against. Under this model a tension zone is formed that commonly generates a steep cline from one parental form through a transient hybrid zone to the second parental form. Success of hybrids would then depend on a spatial or temporal environmental change, which encompasses the Andersonian view of the hybridized habitat. However, several studies have revealed a mosaic hybrid structure (Harrison 1986; Howard 1986; Harrison and Rand 1989; Howard et al. 1997) that could be determined by extrinsic selection associated with patchy environmental conditions. Interspecific transfer of genes through introgressive hybridization is expected to increase levels of genetic diversity, to provide new gene combinations on which selection can act, and may lead to speciation if reproductive isolation is established. However, introgression is difficult to identify. In

polymorphic populations, introgression may have gone unnoticed because appropriate methods for its detection were not used, or introgression may have been invoked without any experimental evidence to support it. Of 165 reports of introgression in plants, only 65 cases were found that justified this claim (Rieseberg and Wendell 1993). However, the frequency of introgression is probably underestimated as phylogenetic closeness may mean that there are few diagnostic features to separate parental taxa. Also, differences between introgressed individuals and the parent species follow a decay function with time since first introgression occurred (Rieseberg and Wendell 1993). Quercus (oaks) has had a long history as a difficult taxon that defies the biological species concept (Burger 1975; Van Valen 1976), because of the seemingly unlimited potential for hybrid combinations within a taxonomic section (Hardin 1975; Nixon 2002). Although hybridization is widespread among oak taxa, it is still unclear how important it is as a source of novel genotypes that contribute to evolution of the genus. Most hybrid combinations of the white oaks of eastern North America occur in nature (Hardin 1975), but only rarely are generations beyond the F1 detected. The few molecular studies have found little evidence for extensive nuclear gene flow in oaks. Nuclear recombinants were found over short distances within a mosaic structured hybrid zone between the white oaks Q. grisea and Q. gambelii (Howard et al. 1997). However, contrasting patterns from nuclear and chloroplast genes were reported in several white oaks from eastern North America (Whittemore and Schaal 1991). Chloroplast genes were exchanged readily among geographically close, but not conspecific, oaks, whereas a nuclear ribosomal gene followed

261 q 2004 The Society for the Study of Evolution. All rights reserved.

262

R. S. DODD AND Z. AFZAL-RAFII

taxonomic boundaries. Other reports from nuclear genes suggest little evidence for hybrid gene exchange (Guttman and Weigt 1989; Ducousso et al. 1993), but this may be more lack of detection than lack of gene exchange as differentiation of parental species was very low. The red oaks of western North America include three evergreen and one deciduous species. Fine-scale habitat differentiation among the species results in some differences in geographic ranges, but areas of sympatry are extensive. In most cases, the four species are morphologically distinct, but in some areas a range of intermediate phenotypes may be encountered. Hybrids have been reported for all species combinations based on morphological or biochemical traits (Brophy and Parnell 1974; Tucker 1980; Vasey 1980; Nason et al. 1992; Dodd et al. 1993), but no molecular studies have been reported that address hybrid status. Diagnostic biochemical traits have indicated that populations of Q. wislizeni from Northern California are complex, suggesting introgression from Q. agrifolia and Q. parvula var. shrevei (Dodd et al. 2002). This could be a result of transient gene flow where reproductive barriers among species are weak, or it could reflect long-term mixing of gene pools accompanied by differential selection of gene complexes. Here we report the dynamics of interspecific gene flow among these four red oaks of the western United States using amplified fragment length polymorphism (AFLP) molecular markers. This study was aimed at determining if the genomes of Q. kelloggii, Q. agrifolia, and Q. parvula var. shrevei could be detected in populations of Q. wislizeni from coastal California; investigating associations between habitat characteristics and population hybrid structure to determine the relative importance of transient interspecific gene flow as opposed to long-term incorporation of alien genes through extrinsic selection pressure; and examining concordance between AFLP markers and diagnostic biochemical traits. MATERIALS

AND

METHODS

Populations Sampled Interior live oak (Q. wislizeni A. DC.) from western North America was investigated for evidence of introgression from three other partially sympatric red oak species (Q. agrifolia Nee, Q. parvula var. shrevei [Muller] Nixon & Muller, and Q. kelloggii Newb.). Reference populations (number sampled) of Q. agrifolia (2), Q. kelloggii (3), Q. parvula var. shreveii (4), and Q. wislizeni (7) were selected to represent parental types based on morphological identification and biochemical typing. In addition, we selected 10 populations of Q. wislizeni from coastal California as putative introgressed populations and one population as pure Q. wislizeni based on our earlier biochemical evidence (Dodd et al. 2002). An average of 15 trees per population were sampled, spaced as widely apart as possible and not less than 50 m apart. Only individuals having Q. wislizeni morphological characters were sampled in the putative introgressed populations, even though some individuals were present that exhibited intermediate morphology.

TABLE 1. Primer combinations and numbers of amplified fragment length polymorphism band classes identified among species of California red oaks (Quercus). Restriction enzyme

Selective primer combinations

EcoRI MseI

AC TC GTA TAC CAA GTG GTG TC

Species

Q. Q. Q. Q.

agrifolia kelloggi parvula wislizeni

Total

Number of bands

70 60 66 73

56 65 67 72

63 67 71 75

67 63 69 74

256 255 273 294

DNA Analysis Genomic DNA was extracted following a simplified Doyle and Doyle method (Cullings 1992). The AFLP method (Vos et al. 1995) was performed following the protocol of Life Technologies (Rockville, MD). DNA was restricted with EcoRI and MseI (1.25 U/ml of each) in a reaction buffer (10 mM Tris-HCl [pH 7.5], 10 mM magnesium acetate and 50 mM potassium acetate) in a total volume of 12.5 ml. The restriction reaction was carried out at 378C for 2 h followed by denaturation of restriction enzymes at 658C for 10 min. Adapter ligation was conducted in the ligation buffer using T4 DNA ligase for 2 h at 208C. Subsequently, the ligation product was diluted 1:10 for primary amplification. Primary amplifications were performed in a standard polymerase chain reaction (PCR) cocktail containing 1.5 mM MgCl2 and 0.5 mM of each of the primary amplification primers. For the primers, the EcoRI primer sequence was identical to the adapter sequence, whereas the MseI primer had an extra cytosine (C) as a selective nucleotide. The PCR reaction was performed in a Techne (Cambridge, U.K.) Genius thermocycler for 20 cycles using the following cycling parameters: 30 sec at 948C, 60 sec at 568C, and 60 sec at 728C. The primary amplification product was then diluted 1: 50 for selective amplification. Selective amplification was performed in a standard PCR reaction cocktail containing 2.5 mM MgCl2 and 0.5 mM of the selective amplification primers. The PCR program comprised the following cycles: 13 cycles of 30 sec at 948C, 30 sec at 658C (annealing temperature was lowered 0.78C at each cycle), and 60 sec at 728C, followed by 23 cycles of 30 sec at 948C, 30 sec at 568C, and 60 sec at 728C. Four sets of selective primers were adopted from a wide range of sets of primers that were tested in pilot runs (Table 1). Selective amplification products were resolved on 6% polyacrylamide native gels. Gels were stained with Gelstar (FMS Bioproducts, Rockland, ME) following the manufacturer’s protocols and documented with a Kodak (Rochester, NY) DC 120 digital camera. Bands were scored as present/absent using Gelcompar II software (Applied Maths, Kortrijk, Belgium). Data Analysis AFLP fragments were treated as single amplified products representing a locus with two allelic states; one in which amplification occurred and the null allele in which base sub-

263

INTROGRESSION IN OAKS

stitutions resulted in lack of enzyme cutting or amplification at the selected site. Mendelian segregation of the alleles was assumed and allele frequencies were calculated assuming a random mating model (FIS 5 0). Frequencies of the recessive allele were calculated by the Taylor expansion (Lynch and Milligan 1994). Nei’s expected heterozygosities were obtained using Popgene ver. 1.31 (Yeh et al. 1999). Linkage disequilibrium was tested for all pairwise combinations of AFLP fragment classes for each of the four primer pair sets independently using the software RAPDLD (Apostol et al. 1996). This software also provides estimates of the decomposition of linkage disequilibrium into a genetic drift component and a component due to epistasis among loci. To test for admixture in the putative introgressed populations, we took two approaches. First, hybrid indices were computed for pairwise comparisons among the four species using a maximum-likelihood (ML) approach to compare band frequencies for each individual with average band frequencies in the parental taxa (Rieseberg et al. 1998, 1999; Rogers et al. 2001). The ML hybrid index (HI) is the sum, over each individual’s alleles, of the probability that the allele was derived from one parent population rather than another. Second, admixture proportions were estimated using a Bayesian approach described by Pritchard et al. (2000). This model-based approach uses Markov chain Monte Carlo (MCMC) methods to assign sample genotypes to K groups in a way that minimizes within-group linkage disequilibrium and assumes within-group Hardy-Weinberg equilibrium. Sample genotypes may be assigned to a single group or jointly to two or more groups probabilistically if their genotypes indicate admixture. A burn-in length of 30,000 runs was used based on test runs for parameter convergence. To infer the number of groups, we used several runs of the Gibbs sampler for different values of K, from K 5 1 to 6, on samples from the reference populations and selected the number of groups that yielded the maximum posterior probability of K. Because our AFLP banding patterns are interpreted as dominant data, we first used the no-admixture model on our reference populations to determine how well species classifications matched the assigned groupings. This model treats each class of genotype as a haploid allele. As the agreement was good, we then used the using-prior-population model to assign admixture proportions of individuals from putative hybrid populations using several different values for migration rate (v). Because the different values for v had little effect on the results, we chose a value of 0.05 for analyses presented here. Climate data was obtained from the Daymet database developed by the University of Montana, Numerical Terradynamic Simulation Group and available at http:// www.daymet.org/. Daymet generates daily surfaces of temperature, precipitation, humidity, and radiation over complex terrain at a 1-km level of resolution, taking into account elevation. The model uses an 18-year daily dataset from 1980 to 1997. Monthly averages of temperatures (8C), precipitation (cm), vapor pressure deficits (Pa), and solar radiation (Wm22) were obtained for each of the collection sites. Canonical correlation analysis was carried out between a matrix of admixture proportions for each individual and a matrix of monthly climate data per population. To test for an association between distance to an alien pollen source and admix-

ture proportion, we carried out simple correlations between distance from the nearest locality of the alien species and its proportion of admixture in the sampled populations. This was carried out separately for Q. agrifolia and Q. parvula as alien species. We then tested the relative importance of environmental conditions and availability of alien pollen on the admixture structure of each population of Q. wislizeni using simple and partial Mantel tests. For this, the multivariate AFLP and climate data were transformed into average taxonomic distance matrices. Average taxonomic distance is a Euclidean metric in character space that calculates the average squared differences among character values and is suitable for quantitative data. Simple Mantel tests were carried out between the taxonomic distance matrix of admixture proportions of all species and both the taxonomic distance matrix of climate data and a Euclidean data matrix of distance from nearest alien pollen source. Partial Mantel tests were then computed for the climate effect with distance from pollen source removed and the distance from pollen source effect with climate removed. RESULTS Hybrid Indices Hybrid indices estimated from band frequencies in parental populations indicated introgression from both Q. parvula and Q. agrifolia into the putative admixed populations of Q. wislizeni. The HI scores were estimated separately for Q. wislizeni–Q. parvula and Q. wislizeni–Q. agrifolia combinations. The joint ML estimate of HI over all putative hybrids between Q. wislizeni and Q. parvula was 0.66 (two units of support: 0.58–0.75), indicating an important level of mixture of these two gene pools (pure Q. wislizeni would have an HI of one and pure Q. parvula an HI of zero). Hybrid contribution of Q. agrifolia to Q. wislizeni was lower, with a joint ML estimate of HI of 0.88 (two units of support: 0.84–0.94). Introgression from Q. kelloggii was undetectable. In most populations HI scores were intermediate between parental scores for Q. wislizeni and Q. parvula (see Fig. 1). In the putative pure population at Clayton, a high proportion of individuals were assignable to Q. wislizeni. HI scores contrasting Q. wislizeni and Q. agrifolia were skewed mostly in the direction of Q. wislizeni, but in all populations some evidence of introgression of Q. agrifolia was detectable. Admixture Proportions All 10 putative introgressed populations showed some degree of admixture varying from significant admixture in all individuals in a population to a mix of pure and admixed individuals in other populations (see Figs. 2, 3). The presumed pure coastal population from Clayton showed minimal levels of admixture. Again, Q. kelloggii appeared only at background levels in some of the individuals. Admixture proportions from Q. parvula and Q. agrifolia were variable, reaching high proportions in some populations. Indeed, four populations showed higher proportions of Q. parvula than of Q. wislizeni in the individual’s genetic background. In northernmost populations, the contribution of Q. agrifolia to the gene pool reached very low levels.

264

R. S. DODD AND Z. AFZAL-RAFII

FIG. 2. Sampled populations of red oaks of California. Contribution from different species to mean admixture in a population denoted by pie charts. Population codes beginning with W, Quercus wislizeni; P, Q. parvula; A, Q. agrifolia; K, Q. kelloggii; U, admixed populations of Q. wislizeni. Admixture was estimated from amplified fragment length polymorphism molecular banding patterns using a Bayesian approach.

Over all populations, the drift component of LD was consistently greater than the epistasis among loci component, the latter accounting on average for only 2% of total linkage disequilibrium. FIG. 1. Maximum-likelihood estimates of hybrid indices between Quercus wislizeni and Q. agrifolia and Q. wislizeni and Q. parvula for populations morphologically identified as Q. wislizeni, based on amplified fragment length polymorphism molecular markers. Pure Q. wislizeni would have a hybrid index (HI) score of one and the other two species would have (HI) scores of zero. Numbers of individuals (N) assigned HI scores.

Population Genetics Genetic diversity in putative hybrid populations was no greater than in pure populations of Q. wislizeni, Q. parvula, or Q. agrifolia (t 5 0.20, P 5 0.84; t 5 0.54, P 5 0.60; t 5 21.5, P 5 0.18, respectively), but was greater than populations of Q. kelloggii (t 5 2.92, P 5 0.01; see Table 2). However, the highest levels of genetic diversity detected in any populations were in two of the putative hybrid populations (Mt. Tamalpais and Clearlake). Over all species and populations, the proportion of pairwise loci showing significant linkage disequilibrium (LD) averaged 6.8%, a little greater than a Type 1 error rate of a 5 0.05. This suggests that LD was not important among the AFLP loci detected here. Higher proportions of LD were detected in three pure populations W21 (Q. wislizeni), P5 (Q. parvula), and K1 (Q. kelloggii), but not in any of the putative hybrid populations.

Climate and Pollen Source Canonical correlation analysis yielded a highly significant association (Wilk’s l 5 0.309, approximate F 5 7.44, P , 0.0001) between the species genetic admixture proportions of all 167 trees and climatic data for each population. The first pair of canonical variables had a canonical correlation of 0.74 and explained 74% of the multivariate association. Of this pair, the canonical variable representing species admixture reflected increasing proportions of Q. wislizeni and decreasing proportions of Q. agrifolia and Q. parvula in the genetic background of individuals (Table 3). Monthly mean temperatures and monthly mean vapor pressure deficits dominated the first canonical variable representing climatic data and contrasted summer (April to October) climate with winter climate. This pair of canonical variables ordinated individuals from those with higher levels of admixture of all three species under conditions of lower summer temperatures and vapor pressure deficits (left in Fig. 4) to those with less admixture and corresponding more closely to pure Q. wislizeni under conditions of higher summer temperatures and higher vapor pressure deficits (right in Fig. 4). To test whether admixture proportions were more closely associated with climate than with distance from alien pollen

265

INTROGRESSION IN OAKS

(distance matrix). Mantel correlations were only significant for climate with the genetic matrix and the climate effect remained significant in partial correlations with the genetic matrix (see Table 4). Canonical variates analysis showed that the chemotypic background of parental contributions that we observed earlier (Dodd et al. 2002) was faithfully represented by the molecular background detected here for the putative hybrid individuals (see Fig. 5). Only the first pair of canonical variates was significant (P , 0.0001), and these explained most of the variation in the multivariate data (multiple correlation coefficient equal to 0.77). DISCUSSION

FIG. 3. Radius plots showing admixture proportions of three red oak species in the genetic background of individuals morphologically identified to Q. wislizeni. Radii are individuals with proportional admixture of Q. wislizeni, Q. parvula, and Q. agrifolia. Origin of radius is zero proportion and the perimeter of radius is 100% proportion. Admixture was estimated from amplified fragment length polymorphism molecular banding patterns using a Bayesian approach.

sources, we carried out simple and multivariate correlations among the putative causal variables. Simple correlations of the proportions of admixture with distance from nearest known alien pollen source were negative and significant (Q. agrifolia: r 5 20.18, P 5 0.02; Q. parvula: r 5 20.27, P 5 0.0005), indicating that proximity to current pollen source reflects historic interspecific gene flow in these oaks. Correlations were then carried out between three multivariate sets of data; proportions of each species contribution to an individual’s genetic background (genetic matrix), climatic variables (climate matrix), and distance from pollen source

Hybrid zone studies commonly focus on the contribution made by two parental species to mixed gene pools. Contribution from more than two parental species necessarily means that exchange of genes has had more than a transient effect on at least one species’ gene pool and should be considered evidence of an introgressive event (Rieseberg and Wendell 1993). We have shown that admixture in the red oaks of California has incorporated genes of at least two species into the genetic background over a broad geographic region of a third species. From earlier evidence, we suspected that coastal populations of Q. wislizeni had complex gene pools including admixture from other oaks of the same section in the genus Quercus. Although this was evident from mixed morphological characters in many cases, we felt that we had detected cryptic introgression also (Dodd et al. 2002). Therefore, the present study is confined to individuals that showed no morphological evidence of mixed ancestry and so is not providing an estimate of the total frequency of introgression in a population. Overall estimates of mixed gene pools among the four red oaks suggest very little contribution of Q. kelloggii in a cryptic hybrid structure. However, both Q. parvula and Q. agrifolia contributed significantly to the overall genetic background of populations assigned to Q. wislizeni. High levels of introgression in the absence of morphologically identifiable hybrids have been reported for other oak species (Whittemore and Schaal 1991; Dumolin-Lape´gue et al. 1999) and in Aesculus (DePamphilis and Wyatt 1990), reflecting both historic and actual interspecific gene flow. Several studies of animal hybrid zones have suggested that the geographic space over which hybridization occurs is commonly limited to small patches or to long, but narrow clines (Barton and Hewitt 1985; Hewitt 1990; Bridle et al. 2001). However, pollen dispersal in plants may provide fundamentally different dynamics to hybrid structure. First, it permits allopatric hybridization, and this may be reinforced year after year if the vectors of pollen dispersal repeat the same geographic trajectories. Second, in wind-pollinated taxa such as oaks, pollen is an invasive agent of gene flow that is relatively unrestricted by habitat variations. Our sampling in the coastal range of Q. wislizeni is not complete, but the molecular patterns that have emerged suggest that introgression is widespread and probably not confined to long, narrow clines, but more closely resembles a mosaic hybrid structure (Harrison 1986; Howard 1986; Harrison and Rand 1989; Howard et al. 1997).

266

R. S. DODD AND Z. AFZAL-RAFII

TABLE 2. Proportions of amplified fragment length polymorphism fragments in linkage disequilibrium and genetic diversity of fragments in populations of red oaks of California. Variance components of linkage disequilibrium partitioned into drift D2ST and epistasis D2IS . Standard deviations are shown in parentheses. Populations in the left three columns are reference populations of parental species, populations in the right three columns are putative hybrid populations assigned in the field to Quercus wislizeni. Population

% of Dij . 0 (P 5 0.05)

He

Q. wislizeni 6.2 7.2 7.7 8.6 7.2 6.0 6.2

(0.4) (0.5) (0.6) (0.7) (0.7) (0.8) (0.6)

0.20 0.13 0.12 0.18 0.16 0.14 0.16

(0.19) (0.18) (0.16) (0.18) (0.19) (0.18) (0.17)

7.1 7.0 8.1 5.9

(0.5) (0.6) (0.8) (0.4)

0.15 0.15 0.18 0.12

(0.19) (0.19) (0.19) (0.18)

Q. agrifolia A1 A2

5.7 5.5

(0.4) (0.4)

0.13 (0.18) 0.12 (0.19)

Q. kelloggii K1 K2 K3 Average D 2IT Drift D 2ST Epistasis D 2IS

9.9 6.3 5.7 0.126 0.124 0.002

(0.8) (0.4) (0.6) (0.049) (0.047) (0.002)

0.09 (0.17) 0.09 (0.16) 0.09 (0.17)

W18 W19 W20 W21 W22 W23 W24 Q. parvula P3 P4 P5 P6

% of Dij . 0 (P 5 0.05)

Population

We detected hybrid structure within the present-day region of sympatry, but also, in allopatric populations separated by at least 300 km from one of the putative hybrid parents. This could be a result of historic hybrid events that have been followed by local extinction of one of the parental species, or it could result from long-distance pollen dispersal. Our estimates of species admixture to the genetic background of an individual were able to shed light on these two possibilities. Assuming regional extinctions would have occurred many generations in the past as a result of habitat modifications through climate change, the admixture from an allopatric species at the individual level should always be low

Admixed populations Cachuma Farley Mt. Tamalpais Ornbaum Clearlake Whiskeytown Cloverdale Ukiah Covelo Douglas City Clayton

6.4 6.8 5.8 6.1 7.3 6.6 6.8 7.3 6.9 6.6 6.9

(0.4) (0.5) (0.4) (0.4) (0.6) (0.6) (0.6) (0.6) (0.5) (0.4) (0.5)

He

0.18 0.20 0.21 0.18 0.21 0.15 0.16 0.13 0.14 0.12 0.15

(0.18) (0.18) (0.18) (0.17) (0.17) (0.17) (0.17) (0.18) (0.18) (0.17) (0.16)

and might be relatively uniform among individuals in a population as a result of dilution of the introgressed genome. However, recent long-distance pollen dispersal should lead to rare individuals in a population exhibiting high proportions of the immigrant genome. In the most northeasterly population of Q. wislizeni (Whiskeytown), which is about 300 km north of the present-day range of Q. agrifolia, only one individual showed a two-parent admixed genetic structure including 88% Q. wislizeni and 12% Q. agrifolia, suggesting a second generation backcross of Q. agrifolia into Q. wislizeni. In the allopatric Douglas City population, again, one individual accounted for most of the contribution of Q. agri-

TABLE 3. Canonical correlation between the matrix of admixture proportions from amplified fragment length polymorphism molecular markers and the matrix of climate data. Loadings for the first canonical variable representing admixture (CV1adm) and the first canonical variable representing climate (CV1clim) are correlations between the respective canonical variables and their original variables. Climate variables CV1 clim Admixture proportion CV1 adm Species

Quercus wislizeni Q. parvula Q. agrifolia

0.976 20.712 20.488

Month

Temperature

Precipitation

Vapor pressure deficit

January February March April May June July August September October November December

20.650 20.541 20.405 0.045 0.628 0.655 0.659 0.674 0.562 0.055 20.463 20.591

20.431 20.224 20.200 20.251 0.083 0.379 0.349 20.358 0.480 20.228 20.226 20.256

20.484 20.317 20.187 0.103 0.640 0.599 0.642 0.674 0.546 0.110 20.267 20.387

267

INTROGRESSION IN OAKS

TABLE 4. Simple and partial Mantel correlations between admixture proportion from amplified fragment length polymorphism molecular markers, climatic variables, and distance from alien pollen source in putative hybrid populations of Quercus wislizeni. Partial Mantel (RM)

Simple Mantel (RM)

Climate Distance

0.135** 20.018

distance removed climate removed

0.148** 20.052

** Mantel correlation significant at P , 0.01.

FIG. 4. Canonical correlation analysis for a set of admixture proportions (based on amplified fragment length polymorphism molecular markers) of three red oak species in the genetic background of individuals morphologically identified to Q. wislizeni and a set of climatic data. Vertical bars are 95% confidence limits about the mean canonical score per population for the admixture dataset. Numbers adjacent to the points are codes of the admixed populations (U) in Figure 2.

folia to the population gene pool. Its admixed proportions (51% Q. wislizeni, 18.6% Q. parvula, and 29% Q. agrifolia) could be explained by a first-generation backcross of Q. agrifolia into an individual already backcrossed from Q. parvula. These two examples indicate that long-distance pollen dispersal is contributing to introgressed genotypes. Other individuals in these allopatric populations having very low proportions of admixed genes can be best interpreted as repeated backcrosses of ancient hybrids that have escaped more recent interspecific gene flow. The presence of hybrids in the absence of one of the parental species is not uncommon in plants (Stebbins 1950; Grant 1981; Wheeler and Guries 1987; DePamphilis and Wyatt 1990) and serves to emphasize the differences between animal and plant hybrid zones. The evolutionary significance of hybrid genotypes depends on the relative importance of dispersal and selection in maintaining the hybrid zone and has been strongly debated (Barton and Hewitt 1985; Arnold and Hodges 1995). Introgression has been defined as ‘‘the permanent incorporation of alien alleles into a new, reproductively integrated population system’’ (Rieseberg and Wendell 1993). It is difficult to determine whether alleles have been permanently incorporated into a genome, but, if selection for hybrids is strong, this should be expected. The association between levels of admixture and climatic variables provides indirect support for environmental selection acting on these mixed genotypes of red oaks. Although pollen dispersal may be important in maintaining low-frequency hybrid genotypes in some populations, this appeared to take secondary importance to environmental gradients in determining levels of introgression. Canonical correlations showed that the frequencies of Q. agrifolia and Q. parvula AFLP markers were higher in regions where summer temperature and summer vapor pressure deficits were lower. This corresponds well with the less xeric habitat preferences of Q. agrifolia and Q. parvula and their

restriction to coastal California. In allopatric populations, Q. wislizeni occupies more xeric sites and higher elevations than either Q. agrifolia or Q. parvula and is distributed both along the coast and in the Sierra Nevada foothills. Despite high levels of introgression, species integrity can be maintained in some sympatric populations, suggesting that selection is operating to maintain a range of genotypes. The Clayton population of Q. wislizeni and two reference populations of Q. agrifolia were all in close proximity to sources of alien pollen, yet we detected minimal levels of alien genes in these populations. This landscape pattern of patches of pure and mixed population structure is consistent with reports for other hybridizing oaks (Howard et al. 1997) and is further evidence that environmental gradients are more important than pollen sources in determining mixed species genetic structure in these oaks. The high degree of introgression of genomic AFLP markers is supported by our detection of introgression of biochemical traits that are presumably the expression of nuclear genes. This contrasts with other studies that have indicated extensive exchange of organellar DNA, but marked differentiation of nuclear DNA in a number of organisms (Ferris et al. 1983; Powell 1983; Marchant at al. 1988) including oaks (Guttman and Weigt 1989; Whittemore and Schaal 1991). This may suggest that genes that are under selection in these hybrid

FIG. 5. Canonical correlation analysis for a set of admixture proportions based on amplified fragment length polymorphism molecular markers of three red oak species in the genetic background of individuals morphologically identified to Quercus wislizeni and admixture proportions based on biochemical markers. Characterization of individuals from biochemical markers was based on earlier reported data (Dodd et al. 2002). Arrows represent CV vectors with canonical loadings for admixture proportions indicated; W, Q. wislizeni; P, Q. parvula; A, Q. agrifolia.

268

R. S. DODD AND Z. AFZAL-RAFII

populations are frequent and dispersed in the genome and are therefore detectable by the AFLP technique. Lack of differences in levels of linkage disequilibrium between hybrid and parental populations suggests that if coadapted gene complexes are selected in parental species, they may occur unrecombined in successful hybrid genotypes. Genetic linkage mapping offers the opportunity to investigate effects of recombination and the genetic architecture of barriers to hybridization in natural hybrid populations (Rieseberg et al. 1999; Carney et al. 2000), and AFLPs are one of the marker systems of choice for developing QTL maps (Mueller and Wolfenbarger 1999). Genetic markers can rarely be considered as truly diagnostic because of lack of knowledge of genetic variation in the total species gene pool. The choice of reference populations can be problematic, as sympatric populations may be less divergent than allopatric populations. Although sympatric populations are preferable, lack of knowledge of their possible hybrid ancestry may confound estimates of introgression. Reference populations for Q. wislizeni were allopatric, whereas reference populations of Q. agrifolia, Q. parvula, and Q. kelloggii were sympatric. This could have exaggerated the differences between species and resulted in higher levels of admixture in the coastal gene pools. Also, estimates of admixture were based on differences in allele frequencies among parental species as there were very few AFLP bands that were unique to a species. This also introduces some error into the estimation process. Nevertheless, our estimates of introgression measured as a hybrid index and as admixture proportions using a Bayesian approach were both consistent with our earlier findings based on chemotype proportions (Dodd et al. 2002), providing strong support for our conclusions. ACKNOWLEDGMENTS We thank N. Kashani and M. Cigna for technical assistance, P. Palsbøll for helpful discussions, and the USDA Forest Service, Pacific Southwest Research Station (through research agreement number 01-JV-11272135-173) for partial funding for this project. LITERATURE CITED Anderson, E. 1949. Introgressive hybridization. Wiley, New York. Apostol, B. L., W. C. Black, P. Reiter, and B. R. Miller. 1996. Population genetics with RAPD-PCR markers: the breeding structure of Aedes aegypti in Puerto Rico. Heredity 76:325–334. Arnold, M. L. 1997. Natural hybridization and evolution. Oxford Univ. Press, Oxford, U.K. Arnold, M. L., and S. A. Hodges. 1995. Are natural hybrids fit or unfit relative to their parents? Trends Ecol. Evol. 10:67–71. Barton, N. H., and G. M. Hewitt. 1985. Analysis of hybrid zones. Annu. Rev. Ecol. Syst. 16:113–148. Bridle, J. R., S. J. E. Baird, and R. K. Butlin. 2001. Spatial structure and habitat variation in a grasshopper hybrid zone. Evolution 55:1832–1843. Brophy, W. P., and D. R. Parnell. 1974. Hybridization between Quercus agrifolia and Q. wislizenii (Fagaceae). Madron˜o 22: 290–302. Burger, W. C. 1975. The species concept in Quercus. Taxon 24: 45–50. Carney, S. E., K. A. Gardner, and L. H. Rieseberg. 2000. Evolutionary changes over the fifty-year history of a hybrid population of sunflowers (Helianthus). Evolution 54:462–474.

Cullings, K. W. 1992. Design and testing of a plant-specific PCR primer for ecological and evolutionary studies. Mol. Ecol. 1: 233–240. DePamphilis, C. W., and R. Wyatt. 1990. Electrophoretic confirmation of interspecific hybridization in Aesculus (Hippocastanaceae) and the genetic structure of a broad hybrid zone. Evolution 44:1295–1317. Dodd, R. S., Z. A.-Rafii, and S. Bojovic. 1993. Chemosystematic study of hybridization in Californian live oak: Acorn steroids. Biochem. Syst. Ecol. 21:467–473. Dodd, R. S., N. Kashani, and Z. A.-Rafii. 2002. Population diversity and evidence of introgression among the black oaks of California. Pp. 775–785 in R. B. Standiford, D. McCreary, and K. L. Purcell, eds. Oak woodlands: oaks in California’s changing landscapes. U.S. Dept. of Agriculture, Albany, CA. Ducousso, A., H. Michaud, and R. Lumaret. 1993. Reproduction and gene flow in the genus Quercus L. Ann. Sci. For. 50: 91S–106S. Dumolin-Lape´gue, S., A. Kremer, and R. J. Petit. 1999. Are chloroplast and mitochondrial DNA variation species independent in oaks? Evolution 53:1406–1413. Ferris, S. D., R. D. Sage, C.-M. Huang, J. T. Nielsen, U. Ritte, and A. C. Wilson. 1983. Flow of mitochondrial DNA across a species boundary. Proc. Natl. Acad. Sci. USA 80:2290–2294. Grant, V. 1981. Plant speciation. Columbia Univ. Press, New York. Guttman, S. I., and L. A. Weigt. 1989. Electrophoretic evidence of relationships among Quercus (oaks) of eastern North America. Can. J. Bot. 67:339–351. Hardin, J. W. 1975. Hybridization and introgression in Quercus alba. J. Arnold Arbor. 56:336–363. Harrison, R. G. 1986. Pattern and process in a narrow hybrid zone. Heredity 56:337–349. Harrison, R. G., and D. M. Rand. 1989. Mosaic hybrid zones and the nature of species boundaries. Pp. 111–133 in D. Otte and J. Endler eds. Speciation and its consequences. Sinauer, Sunderland, MA. Hewitt, G. M. 1990. Divergence and speciation as viewed from an insect hybrid zone. Can. J. Zool. 68:1701–1715. Howard, D. J. 1986. A zone of overlap and hybridization between two ground cricket species. Evolution 40:34–43. Howard, D. J., R. W. Preszler, J. Williams, S. Fenchel, and W. J. Boecklen. 1997. How discrete are oak species? Insights from a hybrid zone between Quercus grisea and Quercus gambelii. Evolution 51:747–755. Lynch, M., and B. G. Milligan. 1994. Analysis of population genetic structure with RAPD markers. Mol. Ecol. 3:91–99. Marchant, A. D., M. L. Arnold, and P. Wilkinson. 1988. Gene flow across a chromosomal tension zone. I. Relicts of ancient hybridization. Heredity 61:321–328. Mayr, E. 1963. Animal species and evolution. Cambridge Univ. Press, Cambridge, U.K. Mueller, U., and L. Wolfenbarger. 1999. AFLP genotyping and fingerprinting. Trends Ecol. Evol. 14:389–394. Nason, J. D., N. C. Ellstrand, and M. L. Arnold. 1992. Patterns of hybridization and introgression in populations of oaks, manzanitas, and irises. Am. J. Bot. 79:101–111. Nixon, K. C. 2002. The oak (Quercus) biodiversity of California and adjacent regions. Pp. 3–20 in R. B. Standiford, D. McCreary and K. L. Purcell, eds. Oak woodlands: oaks in California’s changing landscapes. U.S. Dept. of Agriculture, Albany, CA. Powell, J. R. 1983. Interspecific cytoplasmic gene flow in the absence of nuclear gene flow: evidence from Drosophila. Proc. Natl. Acad. Sci. USA 80:492–495. Pritchard, J. K., M. Stephens, and P. Donnelly. 2000. Inference of population structure using multilocus genotype data. Genetics 155:945–959. Rieseberg, L. H., and J. F. Wendell. 1993. Introgression and its consequences in plants. Pp. 70–109 in R. G. Harrison, ed. Hybrid zones and the evolutionary process. Oxford Univ. Press, Oxford, U.K. Rieseberg, L. H., S. J. E. Baird, and A. M. Desrochers. 1998. Patterns of mating in wild sunflower hybrid zones. Evolution 52: 713–726.

INTROGRESSION IN OAKS

Rieseberg, L. H., J. Whitton, and K. A. Gardner. 1999. Hybrid zones and the genetic architecture of a barrier to gene flow between two sunflower species. Genetics 152:713–727. Rogers, S. M., D. Campbell, S. J. E. Baird, R. G. Danzmann, and L. Bernatchez. 2001. Combining the analyses of introgressive hybridization and linkage mapping to investigate the genetic architecture of population divergence in the lake whitefish (Coregonus clupeaformis, Mitchill). Genetica 111:25–41. Stebbins, G. L. 1950. Variation and evolution in plants. Columbia Univ. Press, New York. ———. 1959. The role of hybridization in evolution. Proc. Am. Philos. Soc. 103:231–251. Tucker, J. M. 1980. Taxonomy of California oaks. Pp. 19–29 in T. R. Plumb, ed. Ecology, management and utilization of oaks. U.S. Dept. of Agriculture, Albany, CA. Van Valen, L. 1976. Ecological species, multispecies and oaks. Taxon 25:233–239. Vasey, M. 1980. Natural hybridization between two evergreen black oaks in the north central Coast Ranges of California. Pp. 30–35

269

in T. R. Plumb, ed. Ecology, management and utilization of oaks. U.S. Dept. of Agriculture, Albany, CA. Vos, P., R. Hogers, M. Bleeker, M. Reijans, T. van de Lee, M. Hornes, A. Frijters, J. Pot, J. Peleman, M. Kuiper, and M. Zabeau. 1995. A new technique for DNA fingerprinting. Nucleic Acids Res. 23:4407–4414. Wagner, W. H. 1969. The role and treatment of hybrids. Bioscience 19:785–789. ———. 1970. Biosystematics and evolutionary noise. Taxon 19: 146–151. Wheeler, N. C., and R. P. Guries. 1987. A quantitative measure of introgression between lodgepole and jack pines. Can. J. Bot. 65: 1876–1885. Whittemore, A. T., and B. A. Schaal. 1991. Interspecific gene flow in sympatric oaks. Proc. Natl. Acad. Sci. USA 88:2540–2544. Yeh, F., R. C. Yang, and T. Boyle. 1999. POPGENE. Ver. 1.31. Molecular Biology and Biotechnology Centre, University of Alberta, Edmonton, Canada. Corresponding Editor: L. Galloway

selection and dispersal in a multispecies oak hybrid ...

Berkeley, California 94720 .... from Northern California are complex, suggesting introgres- .... veloped by the University of Montana, Numerical Terradyn-.

194KB Sizes 2 Downloads 206 Views

Recommend Documents

Dispersal evolution and resource matching in a spatially and ...
tion ability, move between patches at no cost, and have perfect ... develop an analytically tractable asexual model of dispersal .... APP dП ч╪ FPN 1юgюσ=2.

Seed dispersal and changing seed characteristics in a ... - Springer Link
from the crown seed-bank, whereas seeds stored in the soil are .... account for the larger seed densities observed in S2. .... C.A.B. International, Wallingford, UK.

Reactive dynamics in a multispecies lattice-gas ...
phenomena occur on long distance and time scales. It is well known that in such .... chemical rate laws and reaction-diffusion equations. T'he links established ...

A new hybrid method for gene selection - Springer Link
Jul 15, 2010 - Abstract Gene selection is a significant preprocessing of the discriminant analysis of microarray data. The classical gene selection methods can ...

Dispersal evolution and resource matching in a ...
Sage software (Sage Foundation, 2009), we found that the singular ... 4. Discussion. We developed an analytical model to investigate the distribu-.

Fine-scale genetic structure and gene dispersal in Centaurea ... - ULB
+32 2 650 9169; fax: +32 2 650 9170; e-mail: ..... the model (for the best fitting a,b parameters) using a ..... The pollen dispersal kernel best fitting the data had a.

Natural Selection and Cultural Selection in the ...
... mechanisms exist for training neural networks to learn input–output map- ... produces the signal closest to sr, according to the con- fidence measure, is chosen as ...... biases can be observed in the auto-associator networks of Hutchins and ..

Natural Selection and Cultural Selection in the ...
generation involves at least some cultural trans- ..... evolution of communication—neural networks of .... the next generation of agents, where 0 < b ≤ p. 30.

Signatures of seaway closures and founder dispersal in the phylogeny ...
Aug 15, 2007 - uals may arrive simultaneously at a new habitat to estab- ..... of the transisthmian seahorse lineages took place when a land bridge formed in Central ...... Crossley Foundation ex gratia bursary awarded to PRT, a grant from the.

Signatures of seaway closures and founder dispersal in the phylogeny ...
Aug 15, 2007 - A phylogeny was reconstructed using sequence data from mitochondrial ...... Wallace AR: On the zoological geography of the Malay Archi-.

A Stand-Development Approach to Oak Afforestation in ...
mixtures are being planted with little knowledge of subsequent stand development, leading to an inability to predict future stand composition for management purposes. .... More recently, managers have established species mixtures with ...... HAYNES,

Fine-scale genetic structure and gene dispersal in Centaurea ... - ULB
Our model is Centaurea corymbosa Pourret (Asteraceae), ... within a 3 km2 area along the French Mediterranean ..... defined previously (using the best fitting a, b parameters ..... Programme Diversitas, Fragmented Populations network,.

Herbivory, seed dispersal, and the distribution of a ...
studied by combining seedling transplants, herbivore exclosures, and data on seed ... capacity of recovery after disturbance, which include ..... Press, London.

A Hybrid Approach to Error Detection in a Treebank - language
of Ambati et al. (2011). The figure shows the pseudo code of the algo- rithm. Using this algorithm Ambati et al. (2011) could not only detect. 3False positives occur when a node that is not an error is detected as an error. 4http://sourceforge.net/pr

A Hybrid Approach to Error Detection in a Treebank - language
recall error identification tool for Hindi treebank validation. In The 7th. International Conference on Language Resources and Evaluation (LREC). Valleta, Malta.

Kin Selection, Multi-Level Selection, and Model Selection
In particular, it can appear to vindicate the kinds of fallacious inferences ..... comparison between GKST and WKST can be seen as a statistical inference problem ...

Self-Selection and Screening in Law Firms
Dec 3, 2007 - Workers with high effort costs do not find the investment worthwhile, ...... significant lower profit level when firms switch from one track to two ...

CROWD-IN-THE-LOOP: A Hybrid Approach for ...
regardless of the design of the task, SRL is simply .... We required workers to complete a short tutorial2, followed .... to the sentence- and task-level features of ai.

Hybrid Service-Oriented Architectures: A Case-Study in ...
execute the functionality locally or remotely, respectively. The optional attributes, set ..... QoS issues, like performance, trustworthiness, and security. Our future plans include a more ... JXTA SOAP Project home page, 2004. soap.jxta.org/.

Variable selection in PCA in sensory descriptive and consumer data
Keywords: PCA; Descriptive sensory data; Consumer data; Variable selection; Validation. 1. Introduction. In multivariate analysis where data-tables with sen-.

Molecular evidence of male-biased dispersal in ...
(Musick and Limpus, 1997). As adults, they migrate between the same or different seasonal foraging areas of juveniles, and also between foraging and reproductive areas. (Musick and Limpus, 1997). So far, the latter was the only kind of movement thoug