American Journal of Botany 97(2): 303–310. 2010.

GENETIC EFFECTS OF CHRONIC HABITAT FRAGMENTATION REVISITED: STRONG GENETIC STRUCTURE IN A TEMPERATE TREE, TAXUS BACCATA (TAXACEAE), WITH GREAT DISPERSAL CAPABILITY1

Marta Dubreuil2, Miquel Riba2, Santiago C. González-Martínez3, Giovanni G. Vendramin4, Federico Sebastiani5, and Maria Mayol2,6 2CREAF

(Center for Ecological Research and Forestry Applications), Autonomous University of Barcelona, E-08193 Bellaterra, Spain; 3CIFOR-INIA (Center of Forest Research), Carretera de La Coruña km 7.5, 28040 Madrid, Spain; 4Plant Genetics Institute, National Research Council, Via Madonna del Piano, 10 50019 Sesto Fiorentino, Firenze, Italy; and 5Department of Agricultural Biotechnology, GenExpress, University of Florence, Via della Lastruccia 14/16, 50019 Sesto Fiorentino, Firenze, Italy Tree species are thought to be relatively resistant to habitat fragmentation because of their longevity and their aptitude for extensive gene flow, although recent empirical studies have reported negative genetic consequences, in particular after long-term habitat fragmentation in European temperate regions. Yet the response of each species to habitat loss may differ greatly depending on their biological attributes, in particular seed dispersal ability. In this study, we used demographic and molecular data to investigate the genetic consequences of chronic habitat fragmentation in remnant populations of Taxus baccata in the Montseny Mountains, northeast Spain. The age structure of populations revealed demographic bottlenecks and recruitment events associated with exploitation and management practices. We found a strong genetic structure, both at the landscape and within-population levels. We also detected high levels of inbreeding for a strictly outcrossing species. Chronic forest fragmentation resulting from long-term exploitation in the Montseny Mountains seems the most plausible explanation for the strong genetic structure observed. Our results support the view that, contrary to some predictions, tree species are not buffered from the adverse effects of habitat fragmentation, even in the case of species with a high dispersal potential. Key words:

forest fragmentation; gene flow; microsatellites; spatial genetic structure; Taxaceae; Taxus baccata.

Habitat fragmentation can have major genetic and demographic consequences (e.g., Young et al., 1996; Oostermeijer et al., 2003; Honnay et al., 2005). Theoretical and empirical studies predict that reduction in effective population size and the increased isolation of fragmented populations can lead to genetic erosion through enhanced random genetic drift (Ellstrand and Elam, 1993), greater inbreeding (Keller and Waller, 2002), restricted gene flow, and reduced immigration rates (Couvet, 2002). Comparative studies based on a range of molecular markers have shown that trees tend to have less among-population differentiation and more within-population genetic diversity than do herbaceous species (e.g., Hamrick and Godt, 1996; Nybom, 2004) as a result of their outcrossed mating system, their aptitude for extensive gene flow, and their large population sizes (Petit and Hampe, 2006). These findings led to the common belief that trees, especially wind-pollinated species, are less

1

susceptible to the effects of habitat fragmentation (but for exceptions, see Jump and Peñuelas, 2006; Kettle et al., 2007). During the last few years, a substantial amount of empirical evidence has accumulated indicating that both angiosperms and conifer species are sensitive to human disturbances and fragmentation (see a recent review in Pautasso, 2009). However, the available evidence for the negative impact of fragmentation on tree species has also been questioned, particularly in the case of temperate forests (Kramer et al., 2008), because in many cases fragmentation may have been relatively recent and fragments may contain large remnant populations. Detecting the effects of fragmentation in trees might indeed be difficult if the process of habitat fragmentation has occurred over a short time (i.e., <200 yr) relative to the generation time (typically very long in tree species), making it unlikely that the effects of genetic drift would be measurable (Young et al., 1993; Victory et al., 2006; Kramer et al., 2008). Recently, Jump and Peñuelas (2006) reported the negative effects of >600 yr of habitat fragmentation on the genetic diversity and structure of Fagus sylvatica in the Montseny Mountains, northeast Spain. Comparing fragmented and continuous forest areas, they showed that fragmentation of the beech forest resulted in genetic bottlenecks and disruption of the species’ breeding system, leading to significantly elevated levels of inbreeding, population divergence, and reduced within-population genetic diversity. However, tree species may differ greatly in their response to habitat fragmentation depending on life-history traits, in particular seed dispersal ability (Montoya et al., 2008). Although comparative studies indicate that tree species rely predominantly on pollen for gene exchange, in some species seed dispersal might also contribute substantially to gene flow (Petit

Manuscript received 9 April 2009; revision accepted 13 November 2009.

We are grateful to J. M. Espelta and A. Rodrigo for kindly providing part of the data for the TM population. We also thank A. Briz, A. Juarez, and O. Verdeny for field assistance. This research was supported by funds from the Spanish Ministry of Education and Science (Projects CGL2007-63107/ BOS and Consolider-Ingenio Montes CSD2008-00040), the National Park Autonomous Organism (Spanish Ministry of Environment, Project 26/2007), and Generalitat de Catalunya (Emergent Research Group 2005SGR00381), and by a grant (2004FI01255) from DURSI (Departament d’Universitats, Recerca i Societat de la Informació, Generalitat de Catalunya) to M.D. 6 Author for correspondence (e-mail: [email protected]) doi:10.3732/ajb.0900148

303

American Journal of Botany

304

et al., 2005). Therefore, comparisons across forest tree species with contrasting life-history traits within the same landscape may provide useful tools to test the long-term genetic consequences of habitat fragmentation in relation to biological traits. Human activities in the Mediterranean Basin during the last 10 000 years have had a significant impact on the landscape (Thompson, 2005). In Catalonia (northeastern Spain), a number of palynological records indicate that human pressure intensified during the Iberian times (6th–5th centuries BC) and heavily expanded by the 7th century AD through rangewide deforestation associated with grazing and farming (Riera, 2005). Furthermore, the Montseny Mountains are known to have been largely deforested since the 15th century (Monreal, 1989; Oliver, 2003) up to the mid 20th century, when exploitation for firewood and charcoal production was progressively abandoned. Thus, this area provides a pertinent system to further investigate the genetic effects of chronic habitat fragmentation in tree species with different biological attributes. With this purpose in mind, we selected the English yew (Taxus baccata L., Taxaceae) to assess the extent to which the long history of severe habitat degradation in this area has affected its patterns of genetic variation. Taxus baccata coexists with F. sylvatica in the Montseny Mountains. Both species are wind-pollinated, but they differ in other biological attributes, such as dispersal ability: while European beech has a relatively limited dispersal by birds and mammals (median seed dispersal distance of 6.49 m, Sagnard et al., 2007), yew has fleshy fruits that are expected to be dispersed by birds to very long distances from the source (Thomas and Polwart, 2003; see also Jordano et al., 2007). In this study, we used seven nuclear microsatellite markers to investigate the genetic consequences of habitat fragmentation in remnant populations of T. baccata in the Montseny Mountains. Our specific objectives were to quantify the levels of genetic diversity of T. baccata populations and to determine the patterns of genetic structure, both at the landscape and withinpopulation levels. MATERIALS AND METHODS Study species—Yew is a slow-growing evergreen tree, which is very longlived (in some cases up to 500–1000 years), dioecious (strictly outcrossing) and wind-pollinated. Reproduction begins when individuals are about 20 yr old when growing under open canopy conditions (M. Riba et al., unpublished data), or c. 70 yr under closed canopies (Thomas and Polwart, 2003). Seeds are mainly dispersed by frugivorous birds of the genus Turdus (García et al., 2000; Thomas and Polwart, 2003). Yew is widely distributed in Europe, but currently forms small, scattered populations, in particular in the Mediterranean Basin. It is found in a wide range of habitat and community types, from grasslands to closed forests (a complete description of habitat types is given in Thomas and Polwart, 2003). Its overall distribution in the Iberian Peninsula is highly irregular, much more frequent and forming larger populations in the northwestern regions than in the south and along the Mediterranean coast (Vaquero and Iglesias, 2007). Study site, plant sampling, and population age structure—In the Montseny Mountains, T. baccata is currently restricted to a small number of isolated patches, most of them comprising one or a few individuals (Fig. 1). The patches occupy a complex and highly heterogeneous matrix of vegetation types, such as beech and holm oak forests, scrublands (dominated by Erica scoparia, Buxus sempervirens, Juniperus communis, and Calluna vulgaris) and grasslands. Only the four largest populations were included in this study (FN, LBS, TH, TM; Table 1). Trees within these populations are also found growing on a variety of microhabitats, including rocky outcrops, closed canopy forest, and open scrubland. Leaves were collected from 125 individuals (15–43 per population; Table 1), dried in silica gel, and stored at −20°C prior to DNA extraction. Sampling at LBS was limited because only 15 individuals were accessible.

[Vol. 97

In each population, we counted all individuals and recorded the diameter at breast height (DBH, 1.3 m) of all accessible trees with DBH ≥1 cm (Table 1). For large multistemmed trees, we measured the size of the biggest and dominant stem. Tree age at 1.3 m was estimated from DBH using the average annual growth obtained from cores of 12–20 individuals per population. For population LBS, where access to trees was very limited, we used an averaged value of the three other populations. Because most T. baccata saplings take some 20–30 yr to reach 1.3 m in height (M. Riba et al., unpublished data), we added 25 yr to the initial estimates to provide an approximate age at ground level. It should be noted that in some cases, true individual ages were probably underestimated because the largest trees usually consisted of multiple stems/trunks arising near the ground surface. Sex ratio was impossible to determine; most trees were not reproducing at the time of sampling. DNA extraction and microsatellite analysis—Total DNA was isolated using the DNeasy Plant Mini Kit (Qiagen, Hilden, Germany) from 50–100 mg of dry leaf material. Seven primer pairs developed and optimized for T. baccata (Tax23, Tax26, Tax31, Tax36, Tax60, Tax86, and Tax92) were used for the genetic analysis (Dubreuil et al., 2008). Microsatellite amplifications, electrophoresis conditions, and allele-size scoring were carried out as described by Dubreuil et al. (2008). Data analysis—To estimate overall levels of genetic diversity, we calculated the following parameters for all populations: observed heterozygosity (HO), Nei’s unbiased expected heterozygosity (HE, Nei, 1978) and Wright’s inbreeding coefficient (FIS) according to Weir and Cockerham (1984). Departures from Hardy–Weinberg equilibrium at each locus were assessed by a permutation test using 10 000 replicates. All the estimates were computed using the program GENETIX version 4.04 (Belkhir et al., 2001). The presence of null alleles was estimated using the program GENEPOP version 4.0.7 (Rousset, 2008). Maximum likelihood estimates of null allele frequencies were obtained using the expectation–maximization (EM) algorithm (Dempster et al., 1977), and the adjusted allele frequencies were then used to recompute expected heterozygosity values (HE null). Allelic richness was calculated with the program FSTAT version 2.9.3.2 (Goudet, 2001), using the rarefaction approach proposed by El Mousadik and Petit (1996) to correct for differences in sample size. Standard sample size consisted of 15 diploid individuals, which corresponded to the smallest population sample (LBS). We used the program BOTTLENECK version 1.2.02 (Cornuet and Luikart, 1996) to test for recent reductions in population size, as opposed to being historically small populations. Nonparametric Wilcoxon sign-rank tests were computed to determine whether populations showed a significant number of loci with genetic diversity excess, under both the infinite-allele (IAM) and the stepwise-mutation models (SMM). Ten thousand simulations were performed for each sample. Levels of genetic differentiation among populations were assessed by computing the FST estimator (Weir and Cockerham, 1984) with GENETIX version 4.04 (Belkhir et al., 2001). The statistical significance of FST values was tested using 10 000 permutations. Because we detected a significant presence of null alleles at most loci (see Results), we used the program FREENA to estimate FST values from corrected genotype frequencies (available at http://www.montpellier. inra.fr/URLB/; Chapuis and Estoup, 2007). We performed a Mantel test using GENETIX version 4.04 (Belkhir et al., 2001) to investigate for a pattern of isolation by distance, by testing the correlation between the matrix of pairwise [FST / (1 − FST)] and the matrix of geographic distances (logarithmic scale). In addition, Bayesian clustering methods were applied to infer the number of genetic units and their spatial delimitation. The program STRUCTURE version 2.2 (Pritchard et al., 2000) was used to identify clusters (K) of genetically similar individuals. Five independent runs were performed for each K, from K = 2 to 10. The maximum number of clusters used was greater than the number of populations to detect possible substructuring within the samples. All runs were performed with the admixture model without prior population information, assuming correlated allele frequencies among populations, with burn-in and run lengths of 10 000 and 100 000 iterations, respectively. The optimal number of clusters was determined following the guidelines from the authors (Pritchard and Wen, 2004) and the recommendations of Evanno et al. (2005).

RESULTS Population age structure— We found significant differences in annual-ring growth among populations (ANOVA: P < 0.05). Mean annual growth in diameter was very similar for populations

February 2010]

Dubreuil et al.—Strong genetic structure in TAXUS BACCATA

305

Fig. 1. Geographical distribution of Taxus baccata within the Montseny Mountains, northeast Iberian Peninsula. Sampled populations are indicated by black circles (for population codes, see Table 1). White circles indicate nonsampled populations (size < 5 individuals).

TH and TM (3.095 and 3.737 mm/yr, respectively), but was significantly higher for population FN (5.003 mm/yr). The age frequency distribution for each population is shown in Fig. 2. Most single-stemmed trees, or trunks in the case of multistemmed trees, were aged less than 70 yr, dating back to the end of the intensive forest exploitation in this region. About 10% of the individuals from TH were old-aged trees, with ages comprising between 145 and 245 yr. In the smallest population (LBS) about 20% of the individuals were aged between 150 and 300 yr, and another 20% were recruited 50 yr ago. Because in all populations large trees older than 70 yr were usually composed of multiple trunks arising from the ground, their true age might have been underestimated. Genetic diversity and population differentiation— All seven microsatellite loci were polymorphic and the number of alleles per locus ranged from seven (Tax23, Tax31, Tax60) to 16 (Tax92). The four T. baccata populations analyzed had rather similar levels of genetic variability (AR, HE), with overlapping 95% confidence intervals (Table 1). The observed heterozygosity (HO) was in general much lower than the expected heterozygosity (HE), producing significant deviations from Hardy–Weinberg expectations in three of the populations analyzed (Table 1). The frequency of null alleles varied widely among loci and populations (ranging from 6.2% for Tax31 to 26.8% for Tax23). Thus, values of expected heterozygosity slightly increased when the data set was adjusted according to the presence of null alleles (Table 1), although major changes in population ranking were not detected. The BOTTLENECK software identified recent demographic bottlenecks in populations TM, FN, and LBS (P < 0.05), but only under the IAM model. On the other hand, we found a high level of population differentiation even though the maximum distance separating

populations was only ~10 km and in one case <2 km (FN-TM). The overall FST value was 0.227 (P < 0.001), and all pairwise FST values differed significantly from zero (P < 0.001, values ranged from 0.128 to 0.332; Table 2). Correction for null alleles only marginally decreased absolute values, indicating that null alleles were not strongly biasing the differentiation indices (Table 2). The Mantel test revealed a significant correlation between genetic and geographic distances (r = 0.694, P < 0.05), indicating a pattern of isolation by distance among populations. Bayesian clustering of individuals— The five independent runs produced very similar outcomes. The largest increase in the posterior probability occurred at K = 4, suggesting that this was the best model. At this point, almost all individuals from each population were assigned to four independent clusters, with a high proportion of corresponding membership (>81%, Fig. 3). However, five individuals from TH population were always assigned to the same genetic pool as LBS, and some individuals from FN, TH, and TM had mixed ancestry (Fig. 3). For K higher than four, the model inferred two additional genetic units within populations LBS and TM (Fig. 3, K = 5 and K = 6, respectively). In these populations, the distribution of individuals within each cluster was associated with a clear geographical pattern (Fig. 4). For K higher than six, clustering did not increase the number of genetic units within the populations considered. DISCUSSION Strong genetic structure in an animal-dispersed tree with great dispersal capability— The patterns of genetic diversity

American Journal of Botany

306

[Vol. 97

Table 1.

Location, description, and genetic diversity for Taxus baccata populations in the Montseny Mountains (northeast Iberian Peninsula). n = number of measured individuals to build age structure distributions; N = number of genotyped individuals; AR, allelic richness, computed for a standardized sample of 15 diploid individuals; HO, observed heterozygosity; HE, Nei’s unbiased expected heterozygosity (Nei, 1978); HE null, Nei’s unbiased expected heterozygosity corrected for null allele frequencies; FIS, inbreeding coefficient (calculated following Weir and Cockerham, 1984). P values: *** P < 0.001; ns = not significant. Standard errors are in parentheses.

Population

Code

Latitude

Longitude

Population size (n)

N

AR

HO

HE

HE null

FIS

Font Negra La Besa Turó de l’Home Torrent de la Mina

FN LBS TH TM

41.78 41.80 41.77 41.77

2.33 2.38 2.45 2.34

56 (56) 30 (15) 90 (80) 160 (140)

31 15 43 36

3.249 (0.506) 4.857 (0.404) 4.784 (0.663) 4.641 (0.577)

0.529 (0.112) 0.505 (0.046) 0.353 (0.061) 0.561 (0.089)

0.509 (0.042) 0.681 (0.045) 0.636 (0.079) 0.661 (0.042)

0.571 (0.056) 0.743 (0.046) 0.743 (0.069) 0.713 (0.034)

−0.038 ns 0.266*** 0.448*** 0.153***

Fig. 2. Age frequency distributions of Taxus baccata populations in the Montseny Mountains.

and structure obtained for T. baccata in this study contrast with common expectations for temperate tree species, namely, high genetic diversity, low level of inbreeding within populations, and low genetic differentiation among populations (Hamrick et al., 1992). The significant population differentiation detected at the landscape level, combined with significant inbreeding coefficients in three of the four populations analyzed, suggest a

strong limitation to gene flow, both within and among populations. The results obtained using classical population genetic statistics were confirmed by Bayesian model-based clustering methods: STRUCTURE analysis identified as the best model the one that considered each population as a separate gene pool (K = 4). Moreover, Bayesian methods also detected significant substructuring associated with a clear geographical pattern within LBS and TM populations (Figs. 3 and 4), suggesting that mating within these populations has not been at random and occurred within clusters of genetically related individuals. The high values of inbreeding reported here are somewhat surprising for a long-lived tree in which selfing is prevented by dioecy. High values of FIS can be related to the presence of null alleles, which might potentially bias population genetic analysis (Shaw et al., 1999). However, in populations experiencing true inbreeding, the algorithms developed to estimate null alleles (Dempster et al., 1977; Brookfield, 1996) may substantially overestimate their frequency because these methods are based on the a priori assumption of random mating (Van Oosterhout et al., 2006). Despite the moderate-high frequencies of null alleles found for these markers in T. baccata, several lines of evidence suggest that heterozygote deficiencies are not produced (at least only) by null alleles. First, the presence of null alleles must be suspected when some loci show significant excess of homozygotes and others do not deviate from Hardy– Weinberg proportions. In the present study, most populations and loci showed significant heterozygote deficiencies (Appendix S1, see Supplemental Data with the online version of this article), suggesting that populations might not be panmictic. Second, age structure distributions (Fig. 2) show the presence of a reduced number of old and large reproducing trees in each population, strongly suggesting that reproduction occurred between few individuals in the past, increasing the number of consanguineous matings in consecutive generations. This interpretation is in accordance with the observed patterns of genetic structure in some populations, where clear discontinuities separating clusters probably reflect sporadic but successful recruitment events from a reduced number of old-aged, reproducing females. Third, variation in the frequency of null alleles might also increase FST values between populations. However, genetic structuring similar to T. baccata has been found at a wider scale in other Taxus species, such as T. brevifolia (allozymes: El Kassaby and Yanchuk, 1994; Wheeler et al., 1995), T. canadensis (allozymes: Senneville et al., 2001), T. fuana (RAPD: Shah et al., 2008), and T. wallichiana (cpDNA: Gao et al., 2007). Moreover, previous studies on T. baccata conducted at a very local scale (2–8 km) using RAPDs have also detected significant among-population genetic differentiation (Hilfiker et al., 2004a, b). All this evidence suggests that the FIS and FST values reported in our study might be the result of true biological processes rather than artifacts of the markers used. These

Dubreuil et al.—Strong genetic structure in TAXUS BACCATA

February 2010]

307

Table 2.

Pairwise FST comparisons among populations of Taxus baccata considered in this study (below the diagonal). Null-allele corrected values using FREENA (Chapuis and Estoup, 2007) are also given (above the diagonal). For population codes, see Table 1. All values were significantly different from zero (P < 0.001) as shown by a permutation test with 10 000 replicates.

Population

FN LBS TH TM

FN

LBS

TH

TM

— 0.274 0.332 0.128

0.255 — 0.153 0.146

0.323 0.143 — 0.245

0.107 0.129 0.216 —

findings are paradoxical for a species with a high potential for dispersal (i.e., wind-pollinated and animal-dispersed with fleshy fruits) and suggest that gene flow through both pollen and seeds is not enough to counteract the effects of other factors leading to the strong genetic structure observed, as discussed next. Effects of forest fragmentation on spatial genetic structure— Jump and Peñuelas (2006) have recently demonstrated that historic fragmentation of beech forests in the Montseny Mountains has resulted in significant negative genetic effects in this species. Compared to a continuous population, fragmented beech forest patches showed significant inbreeding, isolation by distance patterns of genetic differentiation, and genetic bottlenecks. We obtained similar results when analyzing the largest and isolated yew populations presently found in the same area, which suggests that chronic forest fragmentation in the Montseny Mountains has also negatively influenced its genetic structure. Furthermore, the patterns of genetic structuring obtained for T. baccata in this study were much stronger than those reported by Jump and Peñuelas (2006) for fragmented beech populations within the same landscape (FIS = 0.226 vs. 0.127, FST = 0.227 vs. 0.029 for yew vs. beech, respectively). Alternative explanations for the genetic patterns of differentiation observed, in fragmented populations of both yew and beech, could be the establishment after planting or long-distance colonization. Yew management through planting has not been documented in the study area and seems highly unlikely because the oldest trees occupy rocky, inaccessible sites. On the other hand, and based on the current scattered distribution and the absence of genetic isolation by distance among populations of T. baccata in Switzerland, its patterns of population differentiation on a wider scale have been interpreted as a result of long-distance colonization and metapopulation-like dynamics (Hilfiker et al., 2004b). Certainly, according to this evidence and the species’ traits potentially favoring long-distance dispersal, yew seems to be naturally adapted to persistence through extinction and recolonization dynamics. Therefore, metapopulation dynamics may interact in complex ways with habitat fragmentation influencing population differentiation. However, the genetic patterns found in our study, the demographic structure of populations, and the known history of forest fragmentation in the study area, provide strong evidence to support a major role of habitat degradation in Montseny Mountains’ yew populations. First, long-distance colonization events from different genetic sources and, more generally, extensive gene exchange through both pollen and seeds, is difficult to reconcile with the restricted gene flow and strong inbreeding observed in our study. In fact, the few studies available dealing with temperate tree species that naturally occur in scattered and lowdensity populations do not show evidence of genetic bottlenecks

Fig. 3. Genetic structure of Taxus baccata populations in the Montseny Mountains inferred by the Bayesian approach of Pritchard et al. (2000). Each vertical bar represents an individual, and each color a distinct gene pool. The y-axis indicates the estimated membership percentage in each of the K-inferred clusters. Site codes are defined in Table 1. For further explanation, see Results.

or isolation (e.g., Craft and Ashley, 2007). Second, despite highly restricted gene flow, we found a significant pattern of isolation by distance, suggesting historical gene flow among contiguous populations and therefore a much larger or continuous distribution across the landscape in the past. In addition, ample evidence indicates that European populations of T. baccata have declined in many areas of its natural range during the last 4000 yr (Paule et al., 1993; Thomas and Polwart, 2003). This decline has been attributed to a combination of factors, among them suboptimal climate conditions (particularly in southern Europe), heavy grazing, and competition for light against some deciduous trees such as beech (Hulme, 1996; Svenning and Magård, 1999; Thomas and Polwart, 2003). However, widespread deforestation and excessive felling have undoubtedly contributed to its decline (Tittensor, 1980; García et al., 2000), yew being one of the European tree species that has been most negatively affected by this process (Svenning and Magård, 1999). A compelling evidence of the negative effects of forest exploitation on T. baccata at the Montseny Mountains is the recent expansion of populations: many single-stemmed trees have been recruited during the last 70 yr, by the end of intensive forest exploitation. Past intensive forestry practices have reduced the density of adult trees in remnant populations, reducing the number of both pollen donors and seed-producing females, increasing the mating between related individuals, and consequently, leading to high levels of biparental inbreeding. This process could be exacerbated if deviations from equal sex ratio exist in

308

American Journal of Botany

[Vol. 97

Fig. 4. Geographical distribution of individuals within the distinct genetic units detected using Bayesian methods in the LBS and TM populations in the Montseny Mountains. Different colors indicate different genetic pools. There is no correspondence between the colors used to define pools from one population to another.

the studied populations. A female-biased sex ratio has been reported for small populations of T. baccata in Switzerland (Hilfiker et al., 2004a) and in southern Spain (García et al., 2000). The reduction of population sizes would also have increased the isolation of the remaining fragments, limiting the connectivity between them, and thereby reducing gene flow among populations. The presence of old yew trees scattered in the surroundings of the analyzed populations suggests that more continuous or well-connected populations might have existed in the past. Forest fragmentation might have a negative impact on dispersal and seedling recruitment, because the ever-smaller patches of woodland can become less attractive to frugivorous birds and mammals. For example, some studies have shown a significant loss of dispersal agents for trees dispersed by animals in small fragments of African forests, with a subsequent reduction in tree recruitment (Cordeiro and Howe, 2001). Similarly, studies on fragmented populations of Juniperus thurifera support that dispersal and recruitment in this species are seriously reduced by the loss of the main dispersers, Turdus spp. (Santos and Tellería, 1994; Santos et al., 1999), which are also the main seed dispersers for English yew (Hulme, 1997; García et al., 2000; Thomas and Polwart, 2003). Taken together, reductions in effective population sizes, increased isolation, and reduction of animal-mediated dispersal may explain the strong genetic differentiation found in a tree with high pollen and seed dispersal potential. Conclusions—Our study shows that a strictly outcrossing species with a high dispersal potential is severely affected by fragmentation and habitat loss. Our results are not consistent with recent studies suggesting that animal-dispersed trees in European temperate forests are less vulnerable than wind-dispersed species

to the effects of forest loss (Montoya et al., 2008). A possible explanation, as already pointed out by Montoya et al. (2008), is that animal-dispersed species might be especially sensitive to forest fragmentation when this process also involves the removal of dispersal agents. On the other hand, our study has also stressed the impact of interactions among ecological and anthropogenic factors in the distribution of genetic diversity in T. baccata. Some authors have recently called into question the negative genetic consequences of habitat fragmentation, arguing that empirical support for those predictions are still scarce (Kramer et al., 2008). The lack of conclusive evidence about the genetic effects of forest fragmentation might be attributed, to a large extent, to the long lifespan of many tree species, and perhaps that a larger number of generations are needed for genetic effects to become evident. In this context, temperate relict trees, like T. baccata, provide a good model system to ascertain the long-term effects of habitat fragmentation. LITERATURE CITED Belkhir, K., P. Borsa, L. Chikhi, N. Raufaste, and F. Bonhomme. 2001. GENETIX vs. 4.04, logiciel sous Windows™ pour la génétique des populations. Laboratoire Génome, Populations, Interactions, CNRS UMR 5000, Université de Montpellier II, Montpellier, France. Brookfield, J. F. Y. 1996. A simple new method for estimating null allele frequency from heterozygote deficiency. Molecular Ecology 5: 453–455. Chapuis, M. P., and A. Estoup. 2007. Microsatellite null alleles and estimation of population differentiation. Molecular Biology and Evolution 24: 621–631. Cordeiro, N. J., and H. F. Howe. 2001. Low recruitment of trees dispersed by animals in African forest fragments. Conservation Biology 15: 1733–1741.

February 2010]

Dubreuil et al.—Strong genetic structure in TAXUS BACCATA

Cornuet, J. M., and G. Luikart. 1996. Description and power analysis of two tests for detecting recent population bottlenecks from allele frequency data. Genetics 144: 2001–2014. Couvet, D. 2002. Deleterious effects of restricted gene flow in fragmented populations. Conservation Biology 16: 369–376. Craft, K. J., and M. V. Ashley. 2007. Landscape genetic structure of bur oak (Quercus macrocarpa) savannas in Illinois. Forest Ecology and Management 239: 13–20. Dempster, A. P., N. M. Laird, and D. B. Rubin. 1977. Maximum likelihood from incomplete data via the EM algorithm (with discussion). Journal of the Royal Statistical Society, B, Methodological 39: 1–38. Dubreuil, M., F. Sebastiani, M. Mayol, S. C. González-Martínez, M. Riba, and G. G. Vendramin. 2008. Isolation and characterization of polymorphic nuclear microsatellite loci in Taxus baccata L. Conservation Genetics 9: 1665–1668. Ellstrand, N. C., and D. R. Elam. 1993. Population genetic consequences of small population size: Implications for plant conservation. Annual Review of Ecology and Systematics 24: 217–242. El Kassaby, Y. A., and A. D. Yanchuk. 1994. Genetic diversity, differentiation, and inbreeding in Pacific yew from British-Columbia. Journal of Heredity 85: 112–117. El Mousadik, A., and R. J. Petit. 1996. High level of genetic differentiation for allelic richness among populations of the argan tree [Argania spinosa (L.) Skeels] endemic to Morocco. Theoretical and Applied Genetics 92: 832–839. Evanno, G., S. Regnaut, and J. Goudet. 2005. Detecting the number of clusters of individuals using the software STRUCTURE: A simulation study. Molecular Ecology 14: 2611–2620. Gao, L. M., M. Möller, X. M. Zhang, M. L. Hollingsworth, J. Liu, R. R. Mill, M. Gibby, and D. Z. Li. 2007. High variation and strong phylogeographic pattern among cpDNA haplotypes in Taxus wallichiana (Taxaceae) in China and North Vietnam. Molecular Ecology 16: 4684–4698. García, D., R. Zamora, J. A. Hodar, J. M. Gomez, and J. Castro. 2000. Yew (Taxus baccata L.) regeneration is facilitated by fleshy-fruited shrubs in Mediterranean environments. Biological Conservation 95: 31–38. Goudet, J. 2001. FSTAT, a program to estimate and test gene diversities and fixation indices, version 2.9.3.2. Website http://www2.unil. ch/popgen/softwares/fstat.htm. Hamrick, J. L., and M. J. W. Godt. 1996. Effects of life history traits on genetic diversity in plant species. Philosophical Transactions of the Royal Society of London, B, Biological Sciences 351: 1291–1298. Hamrick, J. L., M. J. W. Godt, and S. L. Sherman-Broyles. 1992. Factors influencing levels of genetic diversity in woody plant species. New Forests 6: 95–124. Hilfiker, K., F. Gugerli, J. P. Schütz, P. Rotach, and R. Holderegger. 2004a. Low RAPD variation and female-biased sex ratio indicate genetic drift in small populations of the dioecious conifer Taxus baccata in Switzerland. Conservation Genetics 5: 357–365. Hilfiker, K., R. Holderegger, P. Rotach, and F. Gugerli. 2004b. Dynamics of genetic variation in Taxus baccata: Local versus regional perspectives. Canadian Journal of Botany 82: 219–227. Honnay, O., H. Jacquemyn, B. Bossuyt, and M. Hermy. 2005. Forest fragmentation effects on patch occupancy and population viability of herbaceous plant species. New Phytologist 166: 723–736. Hulme, P. E. 1996. Natural regeneration of yew (Taxus baccata L.): Microsite, seed or herbivore limitation? Journal of Ecology 84: 853–861. Hulme, P. E. 1997. Post-dispersal seed predation and the establishment of vertebrate dispersed plants in Mediterranean scrublands. Oecologia 111: 91–98. Jordano, P., C. García, J. A. Godoy, and J. L. García-Castaño. 2007. Differential contribution of frugivores to complex seed dispersal patterns. Proceedings of the National Academy of Sciences, USA 104: 3278–3282. Jump, A. S., and J. Peñuelas. 2006. Genetic effects of chronic habitat fragmentation in a wind-pollinated tree. Proceedings of the National Academy of Sciences, USA 103: 8096–8100.

309

Keller, L. F., and D. M. Waller. 2002. Inbreeding effects in wild populations. Trends in Ecology & Evolution 17: 230–241. Kettle, C. J., P. M. Hollingsworth, T. Jaffré, B. Moran, and R. A. Ennos. 2007. Identifying the early genetic consequences of habitat degradation in a highly threatened tropical conifer, Araucaria nemorosa Laubenfels. Molecular Ecology 16: 3581–3591. Kramer, A. T., J. L. Ison, M. V. Ashley, and H. F. Howe. 2008. The paradox of forest fragmentation genetics. Conservation Biology 22: 878–885. Monreal, L. 1989. Montclús, poblat ibèric i castell medieval. In Associació d’Amics del Montseny [eds.], Monografies del Montseny, vol. 2, 147–158. Diputació de Barcelona, Barcelona, Spain. Montoya, D., M. A. Zavala, M. A. Rodríguez, and D. W. Purves. 2008. Animal versus wind dispersal and the robustness of tree species to deforestation. Science 320: 1502–1504. Nei, M. 1978. Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics 89: 583–590. Nybom, H. 2004. Comparison of different nuclear DNA markers for estimating intraspecific genetic diversity in plants. Molecular Ecology 13: 1143–1155. Oliver, M. J. 2003. Tagamanent a l’Edat Mitjana: Senyors, Pagesos i Paisatge al Montseny Occidental (segles X–XV). Ajuntament de Tagamanent, Tagamanent, Barcelona, Spain. Oostermeijer, J. G. B., S. H. Luijten, and J. C. M. den Nijs. 2003. Integrating demographic and genetic approaches in plant conservation. Biological Conservation 113: 389–398. Paule, L., D. Gömöry, and R. Longauer. 1993. Present distribution and ecological conditions of the English yew (Taxus baccata L.) in Europe. In C. R. Temple [ed.], Extended abstracts of International yew resources conference: Yew (Taxus) conservation biology and interaction, Berkeley, California, 1993, 189–196. University of California, Berkeley, California, USA. Pautasso, M. 2009. Geographical genetics and the conservation of forest trees. Perspectives in Plant Ecology, Evolution and Systematics 11: 157–189. Petit, R. J., J. Dumunil, S. Fineschi, A. Hampe, D. Salvini, and G. Vendramin. 2005. Comparative organization of chloroplast, mitochondrial and nuclear diversity in plant populations. Molecular Ecology 14: 689–701. Petit, R. J., and A. Hampe. 2006. Some evolutionary consequences of being a tree. Annual Review of Ecology Evolution and Systematics 37: 187–214. Pritchard, J. K., M. Stephens, and P. Donnelly. 2000. Inference of population structure using multilocus genotype data. Genetics 155: 945–959. Pritchard, J. K., and W. Wen. 2004. Documentation for STRUCTURE software version 2. Department of Human Genetics, University of Chicago, Chicago, Illinois, USA. Website: http://pritch.bsd.uchicago. edu. Riera, S. 2005. Canvis ambientals i modelació antròpica del territori entre l’època ibèrica i l’altmedieval a Catalunya: Aportacions de la palinologia. Cota Zero 20: 99–107. Rousset, F. 2008. GENEPOP’007: A complete re-implementation of the GENEPOP software for Windows and Linux. Molecular Ecology Resources 8: 103–106. Sagnard, F., C. Pichot, P. Dreyfus, P. Jordano, and B. Fady. 2007. Modelling seed dispersal to predict seedling recruitment: Recolonization dynamics in a plantation forest. Ecological Modelling 203: 464–474. Santos, T., and J. L. Tellería. 1994. Influence of forest fragmentation on seed consumption and dispersal of Spanish juniper, Juniperus thurifera. Biological Conservation 70: 129–134. Santos, T., J. L. Tellería, and E. Virgós. 1999. Dispersal of Spanish juniper, Juniperus thurifera, by birds and mammals in a fragmented landscape. Ecography 22: 193–204. Senneville, S., J. Beaulieu, G. Daoust, M. Deslauriers, and J. Bousquet. 2001. Evidence for low genetic diversity and metapopulation structure in Canada yew (Taxus canadensis): Considerations for conservation. Canadian Journal of Forest Research 31: 110–116.

310

American Journal of Botany

Shah, A., D. H. Li, L. M. Gao, H. T. Li, and M. Möller. 2008. Genetic diversity within and among populations of the endangered species Taxus fuana (Taxaceae) from Pakistan and implications for its conservation. Biochemical Systematics and Ecology 36: 183–193. Shaw, P. W., G. J. Pierce, and P. R. Boyle. 1999. Subtle population structuring within a highly vagile marine invertebrate, the veined squid Loligo forbesi, demonstrated with microsatellite DNA markers. Molecular Ecology 8: 407–417. Svenning, J. C., and E. Magård. 1999. Population ecology and conservation status of the last natural population of English yew Taxus baccata in Denmark. Biological Conservation 88: 173–182. Thomas, P. A., and A. Polwart. 2003. Biological flora of the British Isles. Journal of Ecology 91: 489–524. Thompson, J. D. 2005. Plant evolution in the Mediterranean. Oxford University Press, New York, New York, USA. Tittensor, R. M. 1980. Ecological history of yew Taxus baccata L. in southern England. Biological Conservation 17: 243–265. van Oosterhout, C., D. Weetman, and W. F. Hutchinson. 2006. Estimation and adjustment of microsatellite null alleles in nonequilibrium populations. Molecular Ecology Notes 6: 255–256.

Vaquero, J., and S. Iglesias. 2007. Conservación del tejo (Taxus baccata L.) en España. In L. Serra [ed.], El tejo en el Mediterráneo occidental. Jornadas internacionales sobre el tejo y las tejeras en el Mediterráneo occidental, 13–23. Generalitat Valenciana, Conselleria de Territori i Habitatge, Valencia, Spain. Victory, E. R., J. C. Glaubitz, O. E. Rhodes Jr., and K. E. Woeste. 2006. Genetic homogeneity in Juglans nigra (Juglandaceae) at nuclear microsatellites. American Journal of Botany 93: 118–126. Weir, B. S., and C. C. Cockerham. 1984. Estimating F-statistics for the analysis of population structure. Evolution 38: 1358–1370. Wheeler, N. C., K. S. Jech, S. A. Masters, C. J. O’Brien, R. W. Stonecypher, D. W. Timmons, and A. Lupkes. 1995. Genetic variation and parameter estimates in Taxus brevifolia (Pacific yew). Canadian Journal of Forest Research 25: 1913–1927. Young, A., T. Boyle, and T. Brown. 1996. The population genetic consequences of habitat fragmentation for plants. Trends in Ecology & Evolution 11: 413–418. Young, A. G., H. G. Merriam, and S. I. Warwick. 1993. The effects of forest fragmentation on genetic variation in Acer saccharum Marsh. (sugar maple) populations. Heredity 71: 277–289.