7-9 October 2009, Freiburg Workshop Opportunities, challenges and limitations of genomic-based technologies in forest tree breeding and forest genetics Widespread but genetically depauperate forest tree species: Pinus pinea, a case study G.G. Vendramin Istituto di Genetica Vegetale CNR, Firenze, Italy Genetic variation is generally considered a prerequisite for adaptation to new environmental conditions. Thus the discovery of genetically depauperate but geographically widespread species is unexpected. Even more significant, the finding of species that are largely devoid of genetic variation questions the central dogma of conservation genetics : the belief that genetic diversity is essential and worthy of conservation (Lehman 1998).
In most cases, however, the absence or near absence of genetic diversity does not present a challenge to existing theories. For instance, lack of diversity is not unexpected (although not universal) in a species restricted to a single locale with very small population size and that has experienced intensive genetic drift. Similarly, genetic homogeneity is not surprising in asexual or self-fertilizing species, at least at the population level. On the contrary, the finding of widely distributed, obligatory sexually reproducing species that are genetically impoverished is a priori more surprising. One of the most famous cases is that of the African cheetah (O Brien et al. 1985). This species may have experienced a prolonged bottleneck resulting in extreme inbreeding during the Pleistocene when large mammals such as the giant ground sloth or cave bears became extinct (O Brien 1994). Although the cheetah has regained a large distribution range, it has poor sperm quality and poor reproductive success in captivity, stimulating the interest in the role of genetics in conservation (Lehman 1998; Amos and Balmford 2001).
However, other species that are also nearly completely devoid of genetic variation, such as the northern elephant seal, do not seem to suffer from fertility problems (Bonnell and Selander 1974; Hoelzel et al. 1993). Hence, some authors have concluded that continued demographic recovery of a species may owe more to immediate non-genetic factors, such as local availability of food resources, than to the level of genetic variability (Weber et al. 2004). Trees provide an excellent opportunity to test such hypotheses. Given their larger effective population sizes, higher reproductive capacity and greater longevity, trees should be less prone than large animal species to prolonged bottlenecks (Petit and Hampe 2006), as testified by their typically high levels of genetic diversity (Hamrick et al. 1992). If they do experience bottlenecks and massive diversity loss, can they recover and thrive, as observed in some large mammals? Such cases should be particularly useful for testing the conservationists dogma.
Pinus pinea This tree has been used since ancient times for timber, landscape architecture and for its large edible seeds. Consequently, it was extensively planted around the Mediterranean by Etruscans, Greeks, Romans and Arabs Due to its intense cultivation since well before the pre-roman age, its natural range is now difficult to circumscribe Palaeobotanical records pre-dating its cultivation are very scarce, a possible indication that its large present-day distribution is of recent and artificial nature; however, macrofossils and charcoals indicate that P. pinea was present in Spain before the last glacial maximum and in Lebanon long before the coming of man We used chloroplast microsatellite markers. Paternal in conifers Maternal in angiosperms Maternal in conifers and angiosperms
Pinus halepensis Pinus pinaster Pinus pinea 55 populations analysed 1400 individuals analysed 13 cpssrs used (diplaying high variation in other pine species) Chloroplast SSR variation in Pinus pinea Vendramin et al. (2008), Evolution
Variation at chloroplast microsatellites in Pinus pinea compared to that in other conifers Database of genetically depauperate plants Pinus pinea Other conifers 1 Number of species 1 8 Number of loci per species (>6 repeats) 13 4.9 Proportion of polymorphic loci 2 23% 71% H (haplotypic diversity) 0.019 0.37 Â (mean number of alleles/locus) 1.23 4.62 Â[100] 3 1.06 3.90 Â[100;10] 4 1.08 2.58 Mean number of repeats per locus 10.0 13.0 Maximum number of repeats 14.0 15.1 1 Data for eight species combined: Abies alba, Picea abies, Pinus brutia, P. cembra, P. halepensis, P. lambertiana, P. mugo and P. pinaster; see ref. 33. 2 Proportion of polymorphic loci computed without any threshold. 3 Mean number of alleles per locus in equal-sized samples of 100 individuals. 4 Mean number of alleles per locus in equal-sized samples of 100 individuals after standardization to a mean number of repeats per locus of 10. The finding of reduced levels of marker diversity in P. pinea can best be appreciated in a comparative framework. We therefore conducted a survey of the literature for studies reporting genetically depauperate wild plant species based on allozyme data. We selected a threshold of H (overall species nuclear diversity) < 0.05; this corresponds to the heterozygosity of a locus whose most frequent allele exceeds 97%. Attributes such as plant size, woody/herbaceous habit, geographic range size, native or introduced status, existence of asexual reproduction (through apomixis or vegetative growth), and mating system, were noted. Sex repartition Form Vendramin et al. (2008), Evolution Range
Species lacking allozyme diversity were estimated to represent about 4% of the cases, a small but significant number. In agreement with this finding, our literature survey of allozyme data identified as many as 96 species with H (species level diversity) below 0.05. However, further analysis indicates that genetically depauperate plants with attributes similar to those of P. pinea are truly exceptional. First, 26 of the 96 species identified can reproduce clonally. In these species, the lack of diversity can stem from a single clone spreading across a large area. Second, 34 of the 96 species are predominantly selffertilizing. This mode of reproduction allows a single pure line to persist or expand, potentially accounting for the finding of genetically depauperate selfing plants. Third, the list includes 59 endemics, some growing in a single population, often in isolated places such as in islands. Finally, 15 of the cases involve introduced and/or weedy species sampled only in their introduced range.
List of genetically depauperated but widespread plant species, excluding clonal and self-fertilising plants, ranked by order of increasing diversity Species 1 Family Stature 2 Habit 3 Pop 4 Loci 5 Hes 6 Reference Pinus resinosa Pinaceae 32 W 2 27 0.001 (27) Berchemia berchemiaefolia Rhamnaceae 17 W 4 14 0.001 (64) Schwalbea americana Scrophulariaceae 0.6 H 13 15 0.006 (65) Pinus pinea Pinaceae 30 W 17 32 0.015 (16) Lespedeza capitata Fabaceae 1.6 H 12 34 0.020 (66) Juglans cinerea Juglandaceae 30 W 9 12 0.029 (67) Heuchera americana Saxifragaceae 0.6 H 12 14 0.039 (68) Desmodium nudiflorum Fabaceae 0.29 H 5 13 0.043 (69) Tsuga canadensis Pinaceae 35 W 17 10 0.043 (70) 1 Trees in heavy font 2 Stature in meters 3 Habit (W: woody, H: herbaceous) 4 Pop: number of populations sampled 5 Loci: number of loci scored 6 Hes: expected heterozygosity at the species level Vendramin et al. (2008), Evolution Can the near absence of diversity in Pinus pinea be due to technical problems? 1) Absence of the SSRs stretches and/or limited sample sizes? All cp-ssrs sequenced and presence of NO microsatellite stretches confirmed Near absence of diversity confirmed!!! Increased sample size in terms of number of cpssrs (from 9 to 13) and populations (from 12 to 55)
Genomic resources Pinus pinea cdna library construction A non-normalized cdna library was constructed from mrna of 1 year-old needles of stone pine seedlings germinated from a population of open-pollinated seeds Double stranded cdna was prepared and cloned into λ phage vector using the ZAP Express cdna Synthesis Kit (Stratagene) Sequence analysis and assembly A total of 6144 sequences from the 3 end of each clone sequences were obtained. Sequences were subjected to base calling through Phred, pre-processed by means of SeqClean and assembled with Cap3. A total of 3134 unique sequences were identified 81.2% had sequence similarity with GenBank entries, using the BlastX algorithm. Microsatellites Pinus pinea EST-SSRs marker development (collaboration with INIA and INRA) Pinus pinea ESTs were searched for SSR motifs Of the sequences available, 49 (3%) contained SSRs. Primer pairs flanking SSR sequences for 12 loci were designed 16 produced high quality SSRs markers 9 were polymorphic in 35 populations analysed so far, among which populations sampled in Spain where some cpssr variation was detected. Two alleles per locus were detected
New SSRs in Pinus pinea Variation was detected using nuclear SSRs, but diversity is still low (H e =0.154).higher in the Iberian Peninsula p24 pest4c pest5a pest8 pest1489 pest1871 pest2081 pest2669 p114 Sequencing Pinus pinea CANDIDATE GENES: Pinus halepensis 10 Individuals In 10 Populations Mega-gametopyhte DNA isolation 10 populations, 10 individuals each
Sequencing Resequencing in Conifers (CRIEC initiative) Out of 47 amplicons, other 240 already sequenced (in collaboration with UC Davis (David Neale) ACE-SAP project) N # of polymorphic amplicons (%) # of well-placed SNPs Abies pinsapo Juniperus communis 3 3 6 (13) 2 (4) 8 2 Taxus baccata 3 2 (4) 2 Pinus leucodermis 7 19 (40) 31 Pinus brutia 12 28 (60) 56 Pinus halepensis 48 17 (36) 31 Pinus nigra 12 36 (77) 46 Pinus canariensis 12 35 (74) 80 Pinus pinea 12 9 (19) 12 Austrocedrus chilensis 3 4 (8) 4 Cedrus atlantica 4 2 (4) 2 Cedrus libani 4 13 (28) 16 Cedrus brevifolia 4 20 (43) 29 Preliminary estimates to be validated by visual checking of chromatograms
This extreme situation is probably the result of a combination of factors that have acted in the same direction. First, Pinus pinea, like other circum-mediterranean tree species, has a fragmented distribution range, due in part to the heterogeneous landscape and to its limited climatic tolerance. Climatic fluctuations, together with a narrow ecological niche, were suggested to have depleted genetic variation in Pinus torreyana, a Californian coastal pine with similar ecological requirements as P. pinea and no cpssr variation. Environmental constraints and global climate change also resulted in the extinction from the European continent of Pinus canariensis, a closely-related species to P. pinea. Second, the loss of diversity might have been accelerated by the particular mating system of the species. The relatively inbred mating system, along with other reproductive traits of the species (e.g., high energetic investment in seed production; abundant seed predation; the three-year cone maturation cycle; late age of first flowering) might have promoted low historical effective population sizes. P. pinea is one of the very few tree species that are largely devoid of inbreeding depression: even two successive generations of selfpollination do not appreciably reduce height growth in this species (Ammannati 1988).
Third, contrary to other pines, its ability to disperse seeds depends on the presence of mutualistic birds, not on wind. The scarcity of seed dispersers during critical periods of the species history may have compromised its ability to colonize new territories, making it more susceptible to range contractions (and ultimately affecting its genetic diversity). Interestingly, some of the most extreme cases of genetic bottlenecks described so far in pines involve species with birddispersed seeds: the Californian-endemic P. torreyana (Ledig and Conkle 1983) and the Mexican-endemic P. maximartinezii (Ledig et al. 1999). Finally, and most conclusively, there was only limited differentiation between Pinus pinea at the two extremities of its Mediterranean range (Lebanon and Iberia). Since the species was present in these areas much before the advent of ship sailing, as indicated by the fossil record, it is more likely that the bottleneck experienced by Pinus pinea does not date from the Holocene nor from the Last Glacial Maximum but much earlier during the Quaternary, and hence has nothing to do with recent human impacts and cultivation. Such a conclusion is in line with recent findings showing that phylogeographic surveys of trees have often underestimated the antiquity of biogeographic events responsible for the current genetic structure (Petit et al. Taxon 2005)
Demographic expansion of P. pinea The most unusual finding of this study is not the bottleneck and subsequent loss of diversity per se but the fact that this genetically depauperate species has retained its low level of diversity while spreading across a diverse and fragmented region. The maintenance of a low level of genetic diversity implies that the expansion of P. pinea was particularly recent or that mutation rates are low enough. The expansion of P. pinea across the Mediterranean is not dated; however, it must have started at least 3000 years ago, when humans began to cultivate the species (Badal 1998) Thus, P. pinea has been unable to regain much genetic diversity several hundreds or more probably thousands of generations after its decline. Hence, low evolutionary rates seem necessary to account for the existence of such widespread yet genetically depauperate species. Substitution rates are known to be particularly low in trees (Petit and Hampe 2006), including pines (Willyard et al. 2006), a likely consequence of their long generation time (Kay et al. 2006; Petit and Hampe 2006), potentially accounting for the relatively high proportion of trees among widespread yet still genetically depauperate plant species.
While a reduced evolutionary rate should help preserve a low level of diversity after the expansion, it does not account for the success of the expansion itself. First, the expansion may owe more to the idiosyncratic presence of a suitable disperser than to the existence of a large store of genetic variation. In this respect, humans have played a significant role by cultivating the species since at least three millennia. Second, the striking rebound of P. pinea may be in part attributed to the loss of specific parasites and diseases during the bottleneck phase (Amos and Balmford 2001). This release from enemies hypothesis is similar to that proposed for the success of invasive species in their introduced range (Keane and Crawley 2002). As a matter of fact, P. pinea is known to have comparatively few parasites and diseases (Fady et al. 2004).
Third, successful adaptation to new environmental conditions encountered during the expansion depends on the presence of variation at phenotypic traits, not on marker diversity. The non-negligible amount of heritable variation found at adaptive traits in P. pinea (Mutke et al. 2005) could have played a role in the successful expansion of the species. Fourth, epigenetic variation can also accumulate quickly after a bottleneck (Rapp and Wendel 2005); it might therefore be responsible for some of the differences among trees identified in clonal tests Finally, P. pinea has a high level of phenotypic plasticity (Chambel 2006), which could have helped it colonize new environments despite reduced genetic variation. Conclusions While there is little doubt that genetic variation is the raw material for adaptation, the connections between neutral diversity, quantitative trait variation and adaptability are not straightforward. As a consequence, a causal relationship between genetic variability, as measured by marker loci, and the evolutionary precariousness of a species, cannot be taken for granted (Lehman 1998). This study illustrates the great potential of genetically depauperate species for critically evaluating some of the classical assumptions of conservation genetics.
Conclusions However, we should keep in mind that even if some species can survive and thrive following a severe population bottleneck (such as stone pine, red pine or the northern elephant seal), these represent the few lucky ones rather than the norm (Godt and Hamrick 1997). Moreover, the time needed to re-establish variation is considerable, especially in long-lived species, and it is unclear how such species will react to novel conditions, not experienced during the bottleneck or recovery phase. Federico Jerome Remy Bruno David Santiago Anna...and you for your attention
SCIENCE (2003), 300: 1563-1565 Molecular Ecology (2005), Invited Review The American Naturalist (2007), in press Level and distribution of the diversity can be strongly influenced by human activities..i.e. Castanea sativa and Pinus pinea
Phylogeographic analysis Genetic structure 2 haplogroups Geographic structure 2 phylogroups Population 1 Population 2 N ST >> G ST N ST = G ST N ST << G ST
Fagus sylvatica Chloroplast markers Postglacial migration from two main refugia: - Italian Peninsula - Balkan Peninsula Fagus sylvatica Nuclear markers More complex pattern: - Italian refugia confirmed - Refugia in the Pyrenees - Refugia in the South of France
Estimate of odna genetic subdivision based on similarities between haplotypes (N ST ) as a function of G ST. 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 G ST 0 NST A majority of values are found above the diagonal, including all but one of the significant differences between N ST and G ST (shown in black). Note the trend for N ST based on maternallyinherited markers to increasingly depart from the diagonal when G ST increases. 0 maternal paternal significant moving average cpssr variation in some Pinus species pop N n e H P. pinea 51 4 1.10 0.18 P. halepensis 13 25 3.18 0.60 P. pinaster 42 179 3.75 0.74 P. sylvestris 23 178 3.96 0.81 P. resinosa 7 23 3.02 0.57
Fourth, while adult tree are tolerant to fires, thanks to their lack of low branches and thick bark, repeated fires might have reduced population sizes. Instead, other fireadapted pines, thanks to their serotinous cones, can quickly regenerate after fire events and hence rely on rapid generation turnover. This could mitigate genetic drift and decrease generation time, thereby increasing the recovery of genetic diversity following bottlenecks, in stark contrast with the situation found in Pinus pinea. If diversity had been lost independently in different parts of the range, then different alleles would have been fixed at some loci, as observed in the two surviving populations of P. torreyana in California (Ledig and Conkle 1983). The three rare chloroplast DNA haplotypes from Lebanon and central Spain, all differing by a single mutation from the ubiquitous haplotype, do not point to independent survival of haplotypes existing prior to the bottleneck but rather to post-bottleneck mutations in areas of expansion (Slatkin and Hudson 1991).