Genetic diversity, clonal propagation, and insularity
Ruta microcarpa, with its small, isolated populations, and phenotypic evidence of clonality, provides a distinctive model to study the effects of clonal reproduction on the genetic structure of rare island species. The population genetic results reported here show that clonality represents a common reproductive strategy for all analyzed populations and that it appears to counteract some of the effects of small population size and isolation by increasing heterozygosity, polymorphism, and allele richness in R. microcarpa populations.
Although the amount of genetic variability we found in R. microcarpa is low, it is higher than expected if considering the geographic restriction to a single island, the small population sizes, and the low total number of individuals in the species. According to population genetic theory, in fact, rare insular species should be characterized by overall low levels of gene diversity, a low number of alleles per locus, low polymorphism (i.e., several fixed loci), and a high rate of linkage disequilibrium among loci (Hamrick and Godt 1996; Frankham 1998; Frankham et al. 2002; Ouborg et al. 2006). The high number of heterozygotes detected in R. microcarpa (Ho = 0.651; Table 3) together with the relatively high levels of gene diversity (He = 0.410; Table 3) and the detection of just one monomorphic locus are unexpected results for rare insular species and may represent the genetic effects of the high allelic divergence driven by clonality (Halkett et al. 2005). Support for this interpretation comes from the observation that, contrary to our results in R. microcarpa, low values of genetic diversity were found for sexually reproducing Canarian endangered species (He = 0.2 for Anagyris latifolia, González-Pérez et al. 2009; Ho = 0.113, He = 0.306 for Lotus kunkelii, Oliva-Tejera et al. 2006; Ho = 0.100, He = 0.112 for Cistus chinamadensis ssp. gomerae, Batista et al. 2001), while values of genetic diversity were similar to those found in this study for other endangered clonal species such as the Canarian endemic Sambucus palmensis (Ho = 0.550, He = 0.499; Sosa et al. 2010) and the Southern Appalachian endemic Spiraea virginiana (Ho = 0.503, He = 0.391; Brzyski and Culley 2011).
As commonly detected in other plant species (Eckert et al. 2003; Travis et al. 2004; Tsyusko et al. 2005), we found that clonality does not equally affect the different populations of R. microcarpa. According to our data, RC1, RC2, and MUL1 are strictly clonal, while in populations ALO, MUL2, and CED sexual and asexual recruitment strategies seem to contribute equally to reproduction (Table 2). Two hypotheses may explain the pattern of strict clonality that we found in populations RC1, RC2, and MUL1: i) there is no sexual reproduction in these populations, for even few events of sexual reproduction per generation should be sufficient to prevent an extreme monoclonal genotypic pattern (Bengtssom 2003) and ii) no seedling recruitment occurred over a relatively long period of time. However, a few seedlings were observed during field sampling in RC1 and RC2 (À. Marrero, pers. comm.), suggesting that occasional events of sexual reproduction take place in these populations thus supporting the hypothesis of no seedling recruitment. Moreover, allelopathy has been observed for some R. microcarpa populations (R. M. Coelo, pers. comm.), further suggesting that some allelochemicals might inhibit seedlings growth in RC1, RC2 and MUL1.
Many plant species combine sexual and vegetative reproduction (Richards 1986). The balance between sex and clonal growth varies between and within species (Honnay and Bossuyt 2005) and is mainly driven by environmental fluctuations (including both episodic and continuous changes), making the two modes of reproduction successful under different circumstances (Honnay and Bossuyt 2005; Silvertown 2008). Vegetative propagation has ecological costs mainly related to the increased size of clonal plants, resulting in higher resource uptake, increased space occupied, higher probability to interact with other conspecific or heterospecific plants, reduced pollen dispersal, and increased geitonogamous self-pollination, all leading to fitness costs associated with inbreeding depression and pollen discounting (Bushakra et al., 1999; Honnay and Jacquemyn 2008; Vallejo-Marín et al. 2010). Despite the mentioned costs, species that can reproduce clonally have several potential ecological and evolutionary advantages: they can persist in habitats that may not be consistently favorable for sexual reproduction, can better uptake resources in heterogeneous environments, spread the risk of death among ramets, and can increase the attraction of pollinators by increasing floral display size (Honnay and Jacquemyn 2008; Vallejo-Marín et al. 2010).
In the case of R. microcarpa, clonality could provide advantages on two fronts: (1) in small, isolated populations clonal reproduction may provide a form of reproductive assurance by guaranteeing the survival of the species in case of limited pollinator service or absence of mates (Lhuillier et al. 2006; Silvertown 2008); (2) in harsh environments, including steep and windswept ridges or areas with rocky soil affected by frequent landslides, germination of seeds is unlikely, whereas new individuals can be easily generated through clonal propagation (Lhuillier et al. 2006). The combination of population size and type of habitat characterizing each population of R. microcarpa can explain the different levels of clonal propagation we found in different populations. A higher rate of asexual reproduction, in fact, is found in MUL1 (DG = 0.200; Table 2) than in MUL2 (DG = 0.514; Table 2), the former consisting of only a few individuals located on a cliff below a road, the latter comprising more individuals and located on a gentle slope in an open area. In population ALO (composed of many large individuals, located in an open area in the NW part of the island, and with no obvious human impact detected) we found the highest genotypic diversity. Lhuillier et al. (2006) found a similar pattern in Santalum insulare, where populations more subjected to overexploitation, environmental stress, and human impact showed higher levels of clonality. A higher incidence of clonal reproduction in populations threatened by human activities was also found in non insular species (Kenningtom and James 1997; Warburton et al. 2000; Smith et al. 2003).
The low values of genotypic diversity (G/N = 0.215, DG = 0.263; Table 2) discovered in the analyzed populations of R. microcarpa confirm the high overall degree of clonality of this species, especially when compared with other species characterized by small, naturally isolated populations that occur on continents. Lower levels of clonality, for example, were inferred in the endangered species Cypripedium calceolus (DG = 0.97; Brzosko et al. 2002), in the rare species Allium triccocum (DG = 0.87; Vasseur 2001) and in threatened populations of Eucalyptus curtisii (G/N = 0.53, DG = 0.72; Smith et al. 2003). Levels of clonality similar to those of R. microcarpa were retrieved in the endangered insular Pacific tree Santalum insulare (G/N = 0.35, DG = 0.43; Lhuillier et al. 2006). The observation in R. microcarpa of levels of clonal reproduction similar to those of a few other island species for which such information is available, while lower levels of clonality have been reported for endangered, mainland species, implies that clonality might play a more important ecological and evolutionary role in rare insular than mainland species. Even though there is a shortage of studies on the extent of clonal reproduction specifically on islands, it is reasonable to propose that clonal growth may offer an advantage especially in small and isolated populations, where clones may have a greater ability to persist than sexually reproducing individuals (Silvertown 2008). High rates of clonal propagation were actually found in mainland populations that, similar to those of island endemics, were small and marginal (i.e., rare or endangered species, populations of alien plants, or at the edges of species' geographic range; Silvertown 2008).
The occurrence of genetically identical individuals in all R. microcarpa populations results in a reduction in the already small population size of these populations. This further complicates the conservation status of this species, especially if considering that the high number of clonal individuals detected in R. microcarpa populations (74% of the sampled plants shared the same MLG with other samples) may represent an underestimation of the real incidence of clonality in this species, for adjacent plants (which might represent ramets of the same genet) were avoided during sampling. Notably, our results also showed that spatial distances among R. microcarpa individuals do not necessarily reflect the degree of genetic relatedness among individuals, highlighting the importance of molecular techniques in assessing the genetic characteristics and spatial distribution of individuals in populations thought to be affected by clonal propagation.
Genetic differentiation among populations
The results show R. microcarpa to be genetically structured with high differentiation among populations (FST = 0.446). This finding is expected for island species with highly fragmented distribution (Frankham 1997; Carrol and Fox 2008) and is congruent with results of genetic analyses in other Canarian endemics. Francisco-Ortega et al. (2000) reviewed the genetic diversity of 69 species endemic to the Canary Islands and concluded that most of the genetic variation was explained by differences between populations.
The presence of private alleles in all populations, the high values of FST (Table 4) and the low migration rate indicate that populations of R. microcarpa are genetically isolated. Since isolation by distance was not detected, other factors affecting gene flow are more likely to explain genetic isolation than geographic distance. The two most differentiated populations (MUL1 and MUL2, FST = 0.920; Table 4), in fact, are spatially very close to each other, with only a road separating them. This suggests a lack on dispersal ability for R. microcarpa and highlights the susceptibility of this species to habitat fragmentation. Several factors could explain the low dispersal ability of R. microcarpa. For example, its seeds do not show any characteristics typical of a high ability to disperse (i.e., they are not fleshy and have no wings). Lizards, which are thought to be responsible for seed dispersal, are short-range vectors. Furthermore, since allelopathy is suspected to occur in some R. microcarpa populations (R. M. Coelo, pers. comm.), individuals that disperse to a different population might not necessarily be able to establish. Therefore, the presence and intensity of allelopathy could further reduce the already low migration rate among populations.
This study provided important insights into the genetic structure of R. microcarpa and demonstrated the high susceptibility of this species to extinction. The very small effective population size, low genetic diversity, and low levels of gene flow put at severe risk the persistence R. microcarpa and highlight the immediate necessity of measures for conservation. In situ conservation is essential and should aim to preserve as many individuals as possible, including the ones belonging to very small populations, since they can harbor unique genotypes. Concentrating conservation efforts only on the few, large populations or only on part of the populations, in fact, would result in the likely loss of genetic and genotypic variability for the species.
The main threats to R. microcarpa are habitat fragmentation, grazing, and competition with introduced exotic plants (i.e., Opuntia maxima; Bañares et al. 2004; Moreno 2008). Accordingly, in situ conservation should include agricultural and grazing control, in addition to measures to reduce introgression of alien plants. Ex situ conservation in seed orchards is also advisable, for the eventual reintroduction of seedlings belonging to the same population should restore genetic diversity and sustain fitness (Wilkinson 2001). However, this measure would only be successful if seedling establishment is not prohibited by allelopathy (R. M. Coelo, pers. comm.). Further research on the reproductive biology, dispersal ability, the presence of allelopathy, and its influence on seedling establishment is fundamental for planning more specific, potentially successful long-term conservation programs.