AoB Plants
Oxford University Press
Comparative genetic and epigenetic diversity in pairs of sympatric, closely related plants with contrasting distribution ranges in south-eastern Iberian mountains
Volume: 12, Issue: 3
DOI 10.1093/aobpla/plaa013
•
•
•
• Altmetric

### Notes

Abstract

Genetic diversity is generally considered the chief determinant of evolutionary change, but epigenetic diversity is now recognized as another layer of heritable variation with potential adaptive consequences. Epigenetic diversity could sometimes (fragmented populations, stressing habitats) alleviate the loss of genetic diversity and provide an ‘evolutionary backup’ mechanism for wild plants. This study compares genetic and epigenetic diversity in seven congeneric species pairs with restricted and widespread distributions in south-eastern Spain. Results suggest that higher epigenetic diversity could alleviate the loss of genetic diversity in some populations of endemic plants, but also other plant features can be essential to understand the relationship between genetic and epigenetic diversities.

Keywords
Medrano, Alonso, Bazaga, López, Herrera, and Abdelaziz: Comparative genetic and epigenetic diversity in pairs of sympatric, closely related plants with contrasting distribution ranges in south-eastern Iberian mountains

## Introduction

The genetic diversity of species and populations has multiple implications for their ecology, evolution and survival. For example, reduced genetic diversity arising from inbreeding, fragmentation, bottlenecks or founder effects have been long known to pose threats on long-term survival of species and populations (Ellstrand and Elam 1993; Frankham 2005; Allendorf and Luikart 2007; Allendorf 2017). An increasing number of recent studies are showing, however, that genetic diversity (i.e. depending on variation in DNA nucleotide sequence) is not the only heritable genomic information that could influence the ecology, evolution or survival of populations. Epigenetic variations that depend on DNA methylation or chromatin states can influence phenotypic traits and are often inherited over generations in plant populations (Richards 2006; Verhoeven et al. 2010; Quadrana and Colot 2016; Herrera et al. 2018). One defining feature that sets epigenetic variation apart from genetic variation is the capacity to exhibit modifications in response to environmental factors (see, e.g., Dowen et al. 2012; Alonso et al. 2016).

There is now a growing consensus that natural epigenetic diversity could endow wild plant populations with an extra layer of heritable phenotypic variation that could complement genetically based variation and contribute to local adaptation and survival (Grativol et al. 2012; Medrano et al. 2014; Schulz et al. 2014; Herrera et al. 2017; Gáspár et al. 2019; but see Herden et al. 2019). Compelling evidence supporting a role for epigenetic variation as an additional component of genomic variation has been provided by studies showing, for instance, that in wild plant populations epigenetic diversity is often greater than genetic variation (Herrera et al. 2016, 2017; and references therein) and can compensate for the complete or nearly complete loss of genetic variation in small colonizing populations, apomictic species or genetically homogeneous clones (Richards et al. 2012; Spens and Douhovnikoff 2016; Wilschut et al. 2016; Guarino et al. 2019; Jueterbock et al. 2019; Shi et al. 2019; Wang et al. 2020). Interestingly, in experimental populations of Arabidopsis thaliana it has been recently shown that epigenetic variation is under selection and contributes to rapid phenotypic adaptive responses in absence of consistent genetic changes (Schmid et al. 2018). More studies with non-model plants growing in a real-world context are essential to understand the potential role of epigenetic variation in plant adaptation and in ecological and evolutionary processes.

Narrowly endemic plants provide a hitherto unexplored study system particularly well suited for assessing the hypothesis that epigenetic diversity could mitigate the loss of genetic diversity in wild plants. On one side, narrow endemics with restricted distributions often have small, discrete, isolated populations with lower levels of genetic diversity than close relatives with broader geographical distributions (Kruckeberg and Rabinowitz 1985; Cole 2003). And on the other hand, at least in the Mediterranean, endemic plants often live in stressful environments, ecologically marginal habitats and/or highly specific habitat disturbance regimes (Lavergne et al. 2004; Thompson et al. 2005; Totté et al. 2015). Since biotic and abiotic stresses (e.g. herbivory, aridity, extreme temperatures) are able to induce heritable epigenetic modifications in plant genomes (Alonso et al. 2016; Quadrana and Colot 2016), the association of endemic plants with stressful environments could in itself promote epigenetic diversity via environmental induction of epigenetic variants. Recent models suggest that, when populations are small, epigenetic variation would particularly promote adaptation by rapidly restoring rare adaptive states that would otherwise be lost by genetic drift, or in divergent, peripheral environments, especially when epimutations are adaptively biased (Smithson et al. 2019).

This study examines comparative patterns of genetic and epigenetic diversity in wild populations of seven congeneric species pairs occurring sympatrically in the Sierra de Cazorla mountain range, one important glacial refuge and plant biodiversity hotspot in south-eastern Spain (Médail and Quézel 1999; Médail and Diadema 2009), associated to various habitat types. Specifically, we test the expectation that endemic plants with restricted distributions should have lower genetic diversity but higher epigenetic diversity than widespread ones. The paired-species approach used here has been often applied in previous comparisons of ecological, biological or genetic features of endemic and widespread species, as it allows to control for possible phylogenetic effects on interspecific differences (e.g. Karron 1987; Lavergne et al. 2004). By analysing pairs in seven different genera within the same region we also controlled for geographic effects. One novelty of our sampling design was the concurrent estimation of genetic and epigenetic diversity in replicated population samples of each species in each congeneric pair. In this way, robust statistical tests of hypothesized relationships between genetic and epigenetic diversity could be undertaken by applying linear mixed models to the data and treating plant species and populations as random effects. Linear mixed models allow drawing conclusions with reference to a broad inference space, which makes them particularly well suited to answer questions whose scope transcends the limits of the particular samples studied (Bolker et al. 2009; Bolker 2015). Specifically, we expected that (i) epigenetic diversity should be greater than genetic diversity at population level; (ii) after controlling by species relatedness and geographic location, endemic species with restricted distribution should have lower genetic diversity and higher epigenetic diversity than their widespread congeners. If so, then (iii) genetic and epigenetic diversities should have an inverse relationship particularly significant across populations of endemic plants.

## Methods

### Site and study species

Field sampling for this study was carried out in Sierras de Cazorla-Segura-Las Villas Natural Park (Jaén Province, Spain), one of several mountain units comprising the Baetic Ranges complex in south-eastern Iberian Peninsula (Fig. 1A). This Natural Park is the largest protected space in Spain with a total area of 209 920 ha and altitudes ranging from 590 to 2107 m above sea level. The region is characterized by rugged topography, large expanses of well-preserved habitats, a large proportion of endemic species and outstanding floristic diversity (Médail and Quézel 1999; Mota et al. 2002; Melendo et al. 2003).

Figure 1.
Geographical location of our study site, the Natural Park of Sierras de Cazorla Segura and Las Villas (red rectangle), in south-eastern Iberian Peninsula (A). Maps showing the approximate location of the 21 populations from the seven restricted endemic species (left panel), and the 21 populations from the seven widespread congeners (right panel) included in this study, showing that they were similarly distributed across the study area (B).

For the present study we selected seven congeneric species pairs. Each pair consisted of one narrow endemic species with a restricted geographic distribution and that is a specialist of stressing Mediterranean microhabitats or that occurs only in highly specific habitats, and one species with a widespread geographic distribution which utilizes a broad range of habitat types (see Table 1 and Fig. 2 for more details). The seven pairs of restricted and widespread selected species, mentioned in this order, were: Anthyllis ramburii and Anthyllis vulneraria (Fabaceae, Fig. 2A and B); Aquilegia pyrenaica subsp. cazorlensis and Aquilegia vulgaris subsp. vulgaris (Ranunculaceae, Fig. 2C and D); Convolvulus boissieri and Convolvulus arvensis (Convolvulaceae, Fig. 2E and F); Daphne oleoides and Daphne laureola (Thymelaeaceae, Fig. 2G and H); Erodium cazorlanum and Erodium cicutarium (Geraniaceae, Fig. 2I and J); Teucrium rotundifolium and Teucrium similatum (Lamiaceae, Fig. 2K and L); and Viola cazorlensis and Viola odorata (Violaceae, Fig. 2M and N). All except E. cicutarium are perennial plants. In addition, within the genera Anthyllis, Convolvulus, Erodium and Viola the restricted species has woody stocks, whereas the widespread congener is herbaceous. Data on species distribution and habitat requirements were obtained from Flora Iberica (Castroviejo et al. 1986–2012), Flora Vascular de Andalucía Oriental (Blanca et al. 2011) and Proyecto Anthos (http://www.anthos.es/; Aedo and Castroviejo 2005).

Table 1.
Description of the habitat type and geographic distribution for the 14 species included in this study, and details of the location (longitude and latitude), altitude and sample sizes (N) for each of the three populations sampled per species. Longitudinal and latitudinal coordinates are given as decimal coordinates (WGS84). Rrestricted distribution; Wwidespread distribution.
TaxaHabitat typeDistributionPopulationLongitudeLatitudeAltitude (m asl)N
Anthyllis ramburiiRHigh mountain dwarf procumbent scrubs, sandy dolomitic soilsBaetic Rangesanra1−2.8416437.91791146025
anra2−2.8683237.92836163625
anra3−2.9324437.86575147325
Anthyllis vulnerariaWDry grasslands and rocky environments with calcareous soils, broad altitudinal rangeEuropean, Mediterranean basin east to the Caucasusanvu1−2.9755937.83554140225
anvu2−2.8685037.92843164525
anvu3−2.8323537.94364133225
Aquilegia p. cazorlensisRRifts of limestone outcrops, sandy soils in shady, damp sites at cliff basesBaetic Rangesaqca1−2.9602237.78849149525
aqca2−2.9557337.80753190125
aqca3−2.9320237.86538147525
Aquilegia v. vulgarisWStream margins, poorly drained open meadows around springs, broad altitudinal rangeEuropean temperate element and Mediterraneanaqvu1−2.8323137.94395135625
aqvu2−2.8535137.96322118625
aqvu3−2.7866537.92610171325
Convolvulus boissieriRHigh mountain dwarf procumbent scrubs, sandy dolomitic soilsBaetic Rangescboi1−2.9557237.80722191925
cboi2−2.7448037.99704171128
cboi3−2.9723137.90047153525
Convolvulus arvensisWCultivated areas, wasteland, roadsides, grassy slopes, broad altitudinal rangeTemperate and tropical regions, except Australiacarv1−2.8413437.92845151125
carv2−2.8666437.92967163025
carv3−2.9213437.9513281825
Daphne oleoidesRCalcareous xerophytic woodlands and scrublands, rocky slopes North Africa, southern Europe, Asia Minordole1−2.7461237.99791172525
dole2−2.9530337.81240193125
dole3−2.7896537.92117170325
Daphne laureolaWSclerophyllous and semi-deciduous forests, understory of montane forests, mostly in basic soilsPalaearcticdlau1−2.7840737.92632165325
dlau2−2.8683737.92602162925
dlau3−2.9630337.81203177425
Erodium cazorlanumRHigh mountain dry grasslands and rocky environments, sandy dolomitic soilsBaetic RangesecazF−2.8412637.92870152632
ecazL−2.7546937.92780175433
ecazT−2.9723637.89824159840
Erodium cicutariumWMeadows, flood plains, gravel areas, roadsides and disturbed areas, broad altitudinal rangeEurosiberian Southern-temperate element, widely naturalized now CircumpolarecicC−2.8657437.92942162228
ecicF−2.9694437.90388154130
ecicT−2.8421937.92835152023
Teucrium rotundifoliumRRocky outcropsIberian Peninsula–Moroccotrot1−2.8420937.91830146625
trot2−2.9660237.83489147525
trot3−2.8658337.93001163725
Teucrium similatumWSclerophyllous and semi-deciduous forests, calcareous scrublands and grasslands, rocky mountain slopes, broad altitudinal rangeIberian Peninsulatsim1−2.9757137.83560141025
tsim2−2.8416337.91844144125
tsim3−2.8659037.93026162225
Viola cazorlensisRRocky outcrops, cliffs, sandy dolomitic soils Baetic Rangesvcaz1−2.8407537.91833143725
vcaz2−2.9714537.89952153725
vcaz3−2.8715837.92985160025
Viola odorataWOpen woodlands, hedge banks and scrublands, edges of forests and clearings, broad altitudinal rangeEurope and Asiavodo1−2.8398837.93238148324
vodo2−2.8670537.92842163125
vodo3−2.9758837.83988135025
Figure 2.
Photographs showing the general aspect of flowering individuals from each of the 14 study species (A–N), grouped by genera: the restricted species of each genus always on the left and the widespread on the right. Maps showing the exact location of the three populations sampled from each pair of study species (O–U). Anthyllis ramburii (A), Anthyllis vulneraria (B), Aquilegia p. cazorlensis (C), Aquilegia v. vulgaris (D), Convolvulus boissieri (E), Convolvulus arvensis (F), Daphne oleoides (G), Daphne laureola (H), Erodium cazorlanum (I), Erodium cicutarium (J), Teucrium rotundifolium (K), Teucrium similatum (L), Viola cazorlensis (M), Viola odorata (N). Population names refer to Table 1. In each map purple points represent populations of restricted species while blue crosses populations of its widespread congener.

### Field sampling

In order to assess intra- and interspecific patterns of genetic and epigenetic diversity we sampled three populations per species and in each population 23–40 widely spaced flowering individuals (Table 1, Fig. 1B, Fig. 2O and U). A total of 1088 individuals were sampled for the 14 species included in this study. The sampling design aimed to fulfil the appropriate number of individuals analysed to obtain right estimates of population diversities (Nybom 2004) and a balanced sample for the comparisons across species within a specific geographic location. Young leaves were collected from each plant, placed in paper envelopes and dried at ambient temperature in containers with silica gel. Field collections were carried out during the flowering season (April to June) of each studied taxa.

### Laboratory methods

Total genomic DNA was extracted from dry leaf material of all plants sampled using ISOLATE II Plant DNA Kit (Bioline, London, UK) and the manufacturer’s protocol. Genetic and epigenetic analyses were conducted on the same DNA extracts.

Genetic fingerprints were obtained for each plant using amplified fragment length polymorphism (AFLP) markers (Vos et al. 1995; Weising et al. 2005; Meudt and Clarke 2007). The AFLP analyses were performed using standard protocols involving the use of fluorescent dye-labelled selective primers (Weising et al. 2005). After testing different combinations of selective primer pairs, four different combinations of MseI + 3/PstI + 2 primer pairs that resolved more reproducible and easier to score bands were selected on each genus and AFLP analyses were conducted using them (see Supporting Information—Table S1 for a complete overview of all primers used for each genus with the AFLP protocol. Note that the comparisons between species within genera were done using the same four primer pairs).

Plants were also characterized epigenetically using the methylation-sensitive amplified polymorphism (MSAP) technique (Reyna-López et al. 1997; Schulz et al. 2013; Fulneček and Kovařik 2014; Guevara et al. 2017). Methylation-sensitive amplified polymorphism is a modification of the standard AFLP technique that uses the methylation-sensitive restriction enzymes HpaII and MspI in parallel runs in combination with another restriction enzyme (here the frequent cutter MseI is used). HpaII and MspI are two isoschizomers that recognize and cleave the same tetranucleotide sequence 5′-CCGG, but differ in their sensitivity to the methylation state of cytosine. HpaII can cut at sites that are either non-methylated or contain one methylated external cytosine, whereas Msp I cuts non-methylated sites and those with one or two methylated internal cytosine (see Schulz et al. 2013; Fulneček and Kovařic 2014 for further details). Namely, both enzymes cut the DNA if the restriction site is not methylated, but they cut in a different way in the presence of cytosine methylation. After testing different combinations of selective primer pairs three or four MseI + 3/HpaII–MspI + 2 primer combinations that resolved more reproducible and easier to score bands were selected on each genus and MSAP assays were conducted using them (see Supporting Information—Table S1 for a complete overview of the primers used in each genus for the MSAP protocol. Again, the same primer pairs were used for the two species within each genus).

Amplified products from both AFLP and MSAP protocols were analysed on an ABI PRISM 3130xl DNA sequencer, and fingerprint profiles were scored manually by visualizing electrophoregrams with GeneMapper 5.0 software. All primer combinations used with each species were merged in a single binary data table to generate finally one genetic (AFLP) and two epigenetic (MSAP) raw data matrices (see below). We assessed the repeatability of banding patterns for each species by repeating the entire AFLP/MSAP protocol in a number of randomly selected samples. For each species an average of 11.2 % (AFLP) and 25.8 % (MSAP) of the samples were used as replicates (for further details, see Supporting Information—Table S2). After elimination of inconsistent bands, overall error rates ranged from 1.05 to 3.48 % and from 2.79 to 4.85 %, respectively, for AFLP and MSAP raw data sets [see Supporting Information—Table S2]. In the whole group of 1088 plants a total of 2918 AFLP markers were scored, with a mean of 208.4 markers scored per plant (range = 86–418).

Some of the greatest advantages of AFLP and MSAP techniques are their wide genome sampling, high reproducibility and ability to generate many polymorphic bands per reaction without prior knowledge of genomic sequence of the organisms being assayed (Meudt and Clarke 2007; Paun and Schönswetter 2012). On the contrary, one of the most important disadvantages of these two techniques is that bands obtained are considered dominant, implying that polymorphism is scored only in terms of presence or absence, and thus it is not viable in any of them to distinguish between individuals being heterozygous or homozygous for the dominant allele (Paun and Schönswetter 2012). Although different attempts of codominant scoring using band intensities have been proposed (see, for instance, Gort and van Eeuwijk 2010; Foll et al. 2010), various drawbacks, like high unreliable genotype assignments or important loss of information, advised against the regular use of those approaches to obtain codominant markers. For typical population diversity studies this problem can be at least partially mitigated by the high number of polymorphisms that are generated per reaction (Paun and Schönswetter 2012) and by sampling a large number of individuals per population (Lynch and Milligan 1994), as we have done here. Another important limitation of these two techniques is that the sequence content of each AFLP or MSAP marker remains unknown throughout the whole process, i.e. they are anonymous markers, restricting their usefulness to the description of population patterns but obstructing it for further analyses. Although new developments in next-generation sequencing technologies are offering affordable ways to overcome marker anonymity, adopting this approach in a study like the present, with such a high number of individuals and populations from quite a few non-model species, is nowadays still unfeasible in terms of money and time. A specific drawback of the MSAP technique is that it only detects differences in methylation that occur at the restriction sites of the cutter endonucleases, underestimating the overall level of methylation (Schrey et al. 2013; Fulneček and Kovařik 2014; Alonso et al. 2016). In spite of this limitation, MSAP markers have been validated as an alternative to whole-genome bisulfite sequencing (WGBS) (Lauria et al. 2017) and proven useful to investigate variation in genome-wide patterns of cytosine methylation in non-model plants that lack sequenced genomes (Medrano et al. 2014; Foust et al. 2016; Herrera et al. 2016; Wilschut et al. 2016; Thiebaut et al. 2019). In particular, for comparative interspecific studies of non-model organisms with large sample sizes like this, the application of these two cost-effective techniques are probably the only affordable methods that currently can provide reliable, robust and relatively simple genome-wide information simultaneously on both DNA polymorphisms as well as putative changes in DNA cytosine methylation.

### Data analysis

#### DNA methylation analyses

The MSAP profiles were analysed with the R script ‘MSAP_calc’ (Schulz et al. 2013) using the ‘Extract_MSAP_epigenotypes’ function applied independently to each species with the following parameters: Epicode = ‘Mix1’, delete.monomorphic.loci = TRUE and MinPoly = 2. This software analyses the MSAP binary matrix based on four types of methylation pattern according to the presence or absence of one or both fragments of MseI/HpaII and MseI/MspI: (1) fragments present in both profiles (1/1), indicating an unmethylated state; (2) fragments present only in MseI/HpaII profiles (1/0), indicating hemi-methylated CHG sites; (3) fragments present only for MseI/MspI (0/1), indicating hemi- or fully methylated CG sites; and (4) absence of fragments in both profiles (0/0), representing an uninformative state caused either by different types of methylation, or due to restriction site polymorphism. Under the ‘Mix1’ scoring scheme the MSAP profiles are transformed into two binary data sets: one data set of methylated epiloci (hereafter M-MSAP) which scores conditions (2) and (3) as 1 and all other conditions as 0; and a data set of unmethylated epiloci (U-MSAP) which scores condition (1) as 1 and conditions (2) and (3) as 0. All plants sampled were characterized epigenetically by presence–absence scores for U- and M-type MSAP markers. In total 1450 U-MSAP and 2197 M-MSAP markers were scored in the whole group of 1088 plants sampled, with a mean per individual plant of 103.6 (range = 54–181) and 156.9 (range = 95–213), respectively, for U-MSAP and M-MSAP markers.

### Diversity indices

Binary AFLP and MSAP raw data sets were analysed following the same framework using a band-based strategy which did not require calculating allele frequencies (Bonin et al. 2007). Genetic (AFLP) and epigenetic (MSAP) diversity within populations were quantified using four different indices: (i) proportion of polymorphic fragments (PPOL); (ii) Shannon’s diversity index (SI); (iii) proportion of private fragments (PPRIV), i.e. bands unique of each single population; and (iv) the frequency-down-weighted marker value, in the following termed ‘rarity index’ (RI). Shannon’s diversity index (SI) was calculated for each locus within each population using the formula:

$\begin{array}{l}\mathrm{S}\mathrm{I}=-\sum ^{}{P}_{i}\cdot {\mathrm{log}}_{e}\left({P}_{i}\right)\end{array}$

where Pi is the frequency of the presence or absence of the band. The mean SI per population is given by an average of the index values over individual loci (Pérez-Figueroa 2013). The rarity index for individual x (RIx ) was calculated according to Schönswetter and Tribsch (2005) using the formula:

$\begin{array}{l}\mathrm{R}{\mathrm{I}}_{x}=\underset{i=1}{\sum ^{n}}\frac{{s}_{ix}}{\sum _{j=1}^{k}{s}_{ij}}\end{array}$

where n is the number of markers, six is the binary state of the ith marker in individual x (either 1 or 0) and k is the total number of individuals in the population data set (i.e. within the population sample). In the denominator the number of occurrences of the ith marker in the total population data set is calculated. Population rarity index (RI) was estimated as the average of individual values. Calculations were carried out using ‘AFLPdat’ (function ‘Rarityrarity 1 ’, Version 20.10.2010; Ehrich 2006). Altogether we finally obtained 42 population data for each diversity index and marker type.

### Statistical analysis

All statistical analyses were carried out using the R environment (R Development Core Team 2019). Differences in genetic and epigenetic diversity between restricted and widespread species were analysed using linear and generalized mixed effect models with each of the four population diversity indices (PPOL, PPRIV, SI and RI) as response variables, ‘Markers’ (with three levels: AFLP, U-MSAP and M-MSAP) and ‘Distribution’ (with two levels: restricted and widespread) and their interaction as fixed effects, and ‘Populations’ nested within ‘species’ as random effect to specify paired comparisons at population level. The ‘lmer’ and ‘glmer ’ functions from the ‘lme4’ package were used to fit linear mixed and generalized linear mixed models, respectively (Bates et al. 2015). The number of loci was included as a weighing factor to account for variance across species and type of markers [see Supporting Information—Table S3]. In generalized linear models, proportions (PPOL, PPRIV) were modelled as binomial processes. Residuals were graphically inspected for linearity and homoscedasticity, and normal distribution of residuals was not rejected (P > 0.05) in any of the fitted models presented. Statistical significance of the two fixed effects and their interaction on response variables was determined in all cases with ordinary likelihood ratio tests using the ‘anova ’ function from the R ‘stats’ library (Zuur et al. 2009). In each analysis, estimated marginal means (sensu Searle et al. 1980) and associated confidence intervals for the response variable at each factor level were obtained with the ‘emmeans ’ function of the ‘emmeans’ library (Lenth 2018). Post hoc analysis was done by conducting multiple pairwise comparisons of the estimated marginal means with Tukey adjustment. Marginal means from generalized linear models involving proportions were back-transformed to the original scale of measurement.

The relationship between epigenetic and genetic diversity of study populations was separately explored for restricted and widespread species using correlation analyses.

## Results

### Variation in genetic and epigenetic diversity in populations of restricted and widespread plants

Table 2.
Summary of the genetic (AFLP) and epigenetic (U-MSAP and M-MSAP) diversity estimates obtained per species. Means and standard errors (in parenthesis) are shown for proportion of polymorphic fragments (PPOL), Shannon’s diversity index (SI), proportion of private fragments (PPRIV) and rarity index (RI). Rrestricted distribution; Wwidespread distribution.
PPOLSIPPRIVRI
SpeciesAFLPU-MSAPM-MSAPAFLPU-MSAPM-MSAPAFLPU-MSAPM-MSAPAFLPU-MSAPM-MSAP
Anthyllis ramburiiR0.707 (0.023)0.593 (0.020)0.769 (0.010)0.364 (0.013)0.340 (0.019)0.382 (0.006)0.104 (0.020)0.147 (0.019)0.069 (0.024)1.96 (0.01)1.15 (0.10)1.61 (0.10)
Anthyllis vulnerariaW0.647 (0.043)0.604 (0.000)0.855 (0.006)0.295 (0.012)0.312 (0.011)0.389 (0.007)0.118 (0.025)0.168 (0.006)0.023 (0.004)3.12 (0.26)1.35 (0.02)2.72 (0.07)
Aquilegia p. cazorlensisR0.595 (0.068)0.556 (0.076)0.759 (0.056)0.377 (0.019)0.333 (0.012)0.369 (0.019)0.127 (0.053)0.176 (0.067)0.038 (0.018)2.17 (0.34)1.41 (0.31)2.55 (0.33)
Aquilegia v. vulgarisW0.614 (0.035)0.550 (0.008)0.680 (0.019)0.342 (0.009)0.416 (0.016)0.439 (0.013)0.093 (0.041)0.169 (0.032)0.068 (0.023)2.01 (0.25)1.11 (0.09)2.03 (0.16)
Convolvulus boissieriR0.656 (0.090)0.678 (0.073)0.762 (0.039)0.394 (0.034)0.382 (0.033)0.369 (0.018)0.099 (0.032)0.085 (0.025)0.064 (0.010)2.97 (0.38)1.61 (0.23)2.09 (0.17)
Convolvulus arvensisW0.616 (0.066)0.739 (0.027)0.770 (0.064)0.367 (0.012)0.403 (0.014)0.372 (0.011)0.129 (0.027)0.068 (0.010)0.046 (0.020)2.38 (0.31)1.90 (0.06)2.48 (0.35)
Daphne oleoidesR0.0529 (0.099)0.627 (0.059)0.784 (0.024)0.366 (0.029)0.309 (0.015)0.373 (0.015)0.132 (0.037)0.136 (0.034)0.023 (0.005)1.52 (0.22)0.79 (0.14)1.93 (0.08)
Daphne laureolaW0.554 (0.101)0.747 (0.022)0.772 (0.035)0.314 (0.028)0.303 (0.009)0.306 (0.007)0.159 (0.125)0.068 (0.012)0.011 (0.006)1.15 (0.41)0.72 (0.06)1.27 (0.06)
Erodium cazorlanumR0.0757 (0.039)0.724 (0.026)0.864 (0.010)0.410 (0.011)0.360 (0.007)0.394 (0.007)0.064 (0.027)0.103 (0.009)0.027 (0.004)1.56 (0.31)0.99 (0.05)1.40 (0.07)
Erodium cicutariumW0.629 (0.048)0.0639 (0.021)0.707 (0.034)0.429 (0.019)0.303 (0.005)0.309 (0.003)0.091 (0.024)0.111 (0.024)0.087 (0.012)1.60 (0.11)1.04 (0.06)1.23 (0.05)
Teucrium rotundifoliumR0.675 (0.025)0.541 (0.074)0.762 (0.002)0.327 (0.005)0.241 (0.002)0.306 (0.002)0.111 (0.014)0.187 (0.053)0.056 (0.003)5.57 (0.36)1.73 (0.34)2.84 (0.02)
Teucrium similatumW0.636 (0.006)0.520 (0.015)0.794 (0.034)0.317 (0.001)0.288 (0.006)0.351 (0.014)0.132 (0.014)0.217 (0.007)0.047 (0.017)4.25 (0.16)1.33 (0.06)2.48 (0.35)
Viola cazorlensisR0.596 (0.021)0.764 (0.083)0.678 (0.089)0.292 (0.010)0.357 (0.016)0.293 (0.007)0.112 (0.006)0.048 (0.024)0.110 (0.059)3.48 (0.03)2.41 (0.20)2.47 (0.61)
Viola odorataW0.657 (0.064)0.627 (0.097)0.776 (0.062)0.378 (0.006)0.341 (0.017)0.355 (0.011)0.126 (0.059)0.115 (0.044)0.065 (0.034)4.26 (0.84)1.25 (0.18)1.44 (0.23)
Figure 3.
Variation in genetic (AFLP) and epigenetic (U-MSAP and M-MSAP) diversity estimates between populations of restricted and widespread species included in this study. Values of the four diversity indices: proportion of polymorphic fragments (PPOL, A–B), Shannon’s diversity index (SI, C–D), proportion of private fragments (PPRIV, E–F) and rarity index (RI, G–H) are depicted. In each figure each dot denotes a population, the lower and upper boundaries of the boxplot indicate the 25th and 75th percentiles, the horizontal line within the box marks the median and the whiskers indicate data range.

Table 3.
Results of the linear and generalized mixed models examining the effect of the fixed predictors: type of Markers (AFLP, U-MSAP and M-MSAP), Distribution (restricted and widespread) and their interaction, on each of the four response variables, the diversity indices: proportion of polymorphic fragments, PPOL; Shannon’s diversity index, SI; proportion of private fragments, PPRIV; and rarity index, RI. Signif. codes: ***P < 0.001; **P < 0.01; *P < 0.05; P < 0.1.
Response variablePredictorsχ 2P
PPOLMarkers325.41<2.2e-16***
Distribution1.030.30958
Markers * Distribution15.790.00077***
SIMarkers 5.530.06284
Distribution0.200.65303
Markers * Distribution4.990.08268
PPRIVMarkers207.66<2.0e-16***
Distribution0.330.56618
Markers * Distribution7.610.02222*
RIMarkers 93.68<2.2e-16***
Distribution0.360.54758
Markers * Distribution9.570.00834**
Figure 4.
Comparisons of model-estimated marginal means (EMMs) in restricted and widespread species between genetic (AFLP) and epigenetic (U-MSAP and M-MSAP) markers obtained for the four diversity indices included in this study: proportion of polymorphic fragments (PPOL), Shannon’s diversity index (SI), proportion of private fragments (PPRIV) and rarity index (RI). Estimated marginal means are represented by horizontal lines. The vertical bars are confidence intervals for the EMMs, and the red arrows are for the comparisons among them. Statistically significant differences (P < 0.05) are indicated by non-overlapping arrows.

### Relationship between genetic and epigenetic diversity within populations

In restricted species, populations that had more genetic diversity tended to have also more epigenetic diversity, as revealed by the statistically significant (or marginally significant) correlation between most of the diversity indices measured for AFLP markers, on one side, and on the other for the group of methylated (M-MSAP) markers (r = 0.374, 0.653 and 0.555; and P = 0.095, 0.0013 and 0.0089, for PPOL, SI and RI, respectively), or for the group of unmethylated (U-MSAP) markers (r = 0.372 and 0.540; P = 0.096 and 0.0113, for SI and RI, respectively; Fig. 5). However, in widespread species genetic and epigenetic diversity of study populations were largely independent of each other, as shown by statistically non-significant correlations between three of the four diversity indices measured for AFLP markers (PPOL, PPRIV and SI), and M-MSAP (r = 0.104, 0.118 and −0.267; and P = 0.65, 0.61 and 0.24); or U-MSAP type of markers (r = 0.044, 0.084 and 0.181; and P = 0.848, 0.716 and 0.431, respectively, for PPOL, PPRIV and SI; Fig. 5). Just RI in widespread species had significant correlation between genetic and epigenetic markers (r = 0.398 and P = 0.073 for the correlation between AFLP and M-MSAP markers; r = 0.563 and P = 0.007 for the correlation between AFLP and U-MSAP markers), concurring with the pattern found in restricted species (Fig. 5C).

Figure 5.
Pairwise relationships within populations between genetic (AFLP) and epigenetic (U-MSAP and M-MSAP) diversity indices for restricted and widespread species. Note that the relationship is illustrated only for three of the four indices that we have studied, in which at least one statistical significant correlation was found: (A) proportion of polymorphic fragments (PPOL); (B) Shannon’s diversity index (SI); and (C) rarity index (RI). Only statistically significant linear regressions (solid lines) and their 95 % confidence intervals (coloured area) were depicted. In all the figures the identity line (x = y) is also shown as a reference (dashed lines), and to emphasize that all points above these lines represent a greater epigenetic value compared to the same genetic value, whereas points below dashed lines indicate the reverse pattern.

## Discussion

Genetic diversity of natural populations is a fundamental trait in their evolutionary trajectory that summarizes past events and defines the potential for future adaptation to environmental changes (Jump et al. 2009; Allendorf 2017). Epigenetic variation, in turn, results from genetic and environmental factors, as well as from stochastic epimutations (Richards 2006; Johannes and Schmitz 2019), and it can also shape the evolutionary trajectories of populations (e.g. Kronholm and Collins 2016; Smithson et al. 2019). Epigenetic diversity can enhance plant population response under environmental stress (e.g. Latzel et al. 2013) but its magnitude and association with genetic diversity in natural populations is still poorly understood (Herrera et al. 2016; Moler et al. 2018). The concurrent analysis of the genetic and epigenetic diversity of 14 species presented here illustrates the wide range of genetic and epigenetic diversities that plant populations may harbour within a relatively small area of well-preserved montane habitats. In the next sections we will discuss the patterns obtained and how the absence of global consistent relationships between estimates of genetic and epigenetic diversity might be interpreted as the variable outcome of multiple determinants of genetic diversity (Leimu et al. 2006; Stuessy et al. 2014; Ellegren and Galtier 2016) and possibly other factors that were not accounted for in our study design.

### Genetic diversity

Endemic species account for the most original, phylogenetically distinctive component of local floras and a good portion of the rich biodiversity associated to hotspots (Médail and Quézel 1999; Melendo et al. 2003). Consequently, the magnitude of genetic diversity of endemic and threatened plant species has been investigated for decades with the dual interest of understanding their evolution and contributing to their conservation (Allendorf 2017). At a global geographic scale, pioneer analyses based on allozyme markers found lower percentages of polymorphic loci in endemic plants compared to widespread congeners (Karron 1987; Gitzendanner and Soltis 2000; Cole 2003). Likewise, López-Pujol et al. (2009) found a similar trend within a Mediterranean region, although without specifically controlling for relatedness due to scarcity of available congeneric comparisons. Later studies based on large numbers of anonymous DNA markers (e.g. AFLP, RAPD) found only partial support for this trend and confirmed the need to control for phylogenetic relatedness (Reisch and Bernhardt-Römermann 2014; Ellegren and Galtier 2016; and references therein). In the Mediterranean region, many narrow endemic plants of Western European mountains bear moderate to high levels of genetic diversity (Jiménez-Mejías et al. 2015), suggesting that divergence in genetic diversity between endemic and widespread species could not occur in this region (Forrest et al. 2017; and references therein). We are not aware, however, of any previous analysis addressing comparatively the genetic diversity in several congeneric pairs of endemic and widespread species at a local scale (i.e. simultaneously accounting for species relatedness and geographic location) in the Mediterranean or elsewhere.

In this study, the average polymorphism in AFLP markers per population ranged between 52.9 % (D. laureola, widespread) and 75.7 % (E. cazorlanum , restricted). All species studied here thus exceeded the population average value (41.6 % ± 21.5) reported by Reisch and Bernhardt-Römermann (2014) for 75 AFLP studies worldwide. These data suggested that our study region might be particularly rich not only in species numbers but also in the genetic diversity they preserve, which provides an extra value for future persistence (Jump et al. 2009). Previous studies in the Sierra de Cazorla region have likewise revealed consistently high levels of genetic diversity in plants and associated microfungi (Jordano and Godoy 2000; Medrano and Herrera 2008; Cánovas et al. 2015; Herrera et al. 2011, 2014), thus suggesting that this might be a distinctive feature of this biodiversity hotspot and highlighting the interest of conducting additional research in the framework of the ‘genetic hotspot’ concept (Souto et al. 2015).

Genetic distinctiveness of populations is also a relevant parameter from a conservation perspective (Jump et al. 2009; López-Pujol et al. 2009; Allendorf 2017). Except for one outlier population of D. oleoides , the PPRIV was <25 % in all populations and <10.7 % in half of them. We did not find significant differences between average PPRIV per population or RI of endemic and widespread plants at the local scale investigated here. Altogether our results supported the hypothesis that plant endemics from Western Mediterranean mountains do not necessarily harbour lower levels of genetic diversity (Forrest et al. 2017), although similar analyses in other geographic locations are required to verify the universality of this pattern. Furthermore, a wider geographic sampling should be conducted to conclude about population differentiation at the species level for conservation purposes. First, because genetic distinctiveness may appear only in a few populations that might be particularly valuable for the survival of endemic and widespread species (e.g. Hampe and Petit 2005; Forrest et al. 2017). And second, because in widespread species the few populations sampled in this study mostly occur at the rear-edge border of their distribution ranges, where genetic diversity could be particularly reduced or skewed and non-representative for the species (Hampe and Petit 2005; Eckert et al. 2008; Reisch and Bernhardt-Römermann 2014).

### Relationships between genetic and epigenetic diversity

Despite the heterogeneity of our study populations in the features mentioned above, we were still able to find significant relationships between genetic and epigenetic diversity, that were clearest for M-MSAP markers. Our results supported the expectation of higher epigenetic (M-MSAP) than genetic (AFLP) polymorphism in the whole sample, which was consistent with the idea that substantial epigenetic variation at shorter spatial distances might contribute significantly to population epigenetic diversity (see also Medrano et al. 2014; Herrera et al. 2016; and references therein). An excess of DNA methylation variation relative to genetic variation has been found in populations of the great majority of plant species examined to date and epigenetic variation has been linked to functional phenotypic variation (e.g. Gao et al. 2010; Herrera and Bazaga 2010; Lira-Medeiros et al. 2010; Richards et al. 2012; Medrano et al. 2014; Jueterbock et al. 2019; Shi et al. 2019; Wang et al. 2020), and also in most wild animals (for a recent review, see Hu and Barret 2017). In particular, Wang et al. (2020) found that the introduced populations of the clonal herb Hydrocotyle vulgaris exhibited extremely low genetic diversity but variation in specific leaf area and other leaf traits were all positively related to intra-population polymorphism in MSAP markers. Also, in clonal meadows of Zostera marina methylation variation promotes variation in fitness-related traits of ecological relevance, specifically photosynthetic performance and heat stress resilience, and contributes to the long-term survival of its genetically depauperate populations by increasing mitotically heritable and ecologically relevant phenotypic variation (Jueterbock et al. 2019). Such findings suggest that extensive epigenotypic variations may support phenotypic variation and have a key role in the evolutionary potential of wild populations, as predicted by theoretical models (e.g. Feinberg and Irizarry 2010; Furrow and Feldman 2014; Smithson et al. 2019). Moreover, for the group of endemic plants studied here the difference between epigenetic and genetic diversity was greater at populations with low genetic polymorphism, and the two variables were positively related, suggesting that at least in part epigenetic variation could be dependent on genetic variation. However, in widespread species, epigenetic polymorphism was consistently high regardless of the genetic polymorphism observed, and the two variables were not significantly related. Such contrasting patterns were partly consistent with our expectation that a higher epigenetic diversity could possibly alleviate the lower genetic diversity expected in populations of endemic plants. However, the relationship between the two variables remained positive and held at the population but not at the species level, suggesting that intraspecific variation was as relevant as interspecific variation in understanding the association between genetic and epigenetic diversities of wild plants (Herrera et al. 2016). A more complete understanding of the evolutionary relevance of epigenetic variation and the contribution of epigenetics to phenotypic differentiation and plant adaptation to environmental changes will require additional research on the extent to which epigenetically induced responses are also under genetic control and contribute to population and species persistence (Richards 2006).

## Conclusions

We found that all populations studied exhibited moderate to high levels of genetic polymorphism. However, contrary to our expectations, populations of restricted endemic and widespread species did not differ in average genetic or epigenetic diversity in our study region. The concurrent estimation of genetic and epigenetic diversity in replicated population samples of each species confirmed that levels of polymorphism at frequently methylated cytosines (M-MSAP markers) were always higher than polymorphism at genetic (AFLP) markers in both restricted endemic and widespread species. Importantly, correlation analyses showed that only in the group of species with restricted distribution did populations with higher epigenetic diversity tend to have also higher levels of genetic variation, although the difference between epigenetic and genetic diversity was greater at populations with low genetic polymorphism. This result is partly consistent with our expectation that a higher epigenetic diversity could possibly alleviate the lower genetic diversity expected in populations of endemic plants. Further work is needed to understand the effects that other intrinsic plant traits (e.g. lifeform) and population features (e.g. size, isolation) may have in comparative multispecies patterns of genetic and epigenetic co-variation in this Mediterranean hotspot region or elsewhere.

## Sources of Funding

Financial support for this study was provided by grants CGL2013-43352-P and CGL2016-76605-P from the Spanish Ministry of Economy and Competitiveness.

## Contributions by the Authors

C.A., C.M.H. and M.M. conceived the idea and designed the study; C.A. and M.M. gathered plant material from natural populations; E.L. and P.B. isolated all DNAs and performed all AFLP and MSAP analyses; M.M. performed data analysis and prepared all figures and tables. C.A., C.M.H. and M.M. wrote the manuscript. All the authors read and approved the manuscript.

None declared.

## Acknowledgements

We thank Noelia Zarza and Elena Villa for assistance in the field work; Abelardo Aparicio and Alfredo Benavente for plant identification; the reviewers for their constructive comments and suggestions which greatly improved this work; and the associate editors for invitation to contribute to this special issue. Consejería de Medio Ambiente, Junta de Andalucía, authorized plant collections and provided invaluable facilities in Sierra de Cazorla.

## Literature Cited

1

Aedo C, Castroviejo S2005Anthos. Sistema de información sobre las plantas de Españahttp://www.anthos.es/ (July 2019).

2

Allendorf FW2017. Genetics and the conservation of natural populations: allozymes to genomes. Molecular Ecology26:, pp.420–430.

3

Allendorf FW, Luikart G2007Conservation and the genetics of populations. Oxford: Blackwell.

4

Alonso C, Medrano M, Pérez R, Canto A, Parra-Tabla V, Herrera CM2019. Interspecific variation across angiosperms in global DNA methylation: phylogeny, ecology and plant features in tropical and Mediterranean communities. The New Phytologist224:, pp.949–960.

5

Alonso C, Pérez R, Bazaga P, Medrano M, Herrera CM2016. MSAP markers and global cytosine methylation in plants: a literature survey and comparative analysis for a wild-growing species. Molecular Ecology Resources16:, pp.80–90.

6

Bates D, Maechler M, Bolker B, Walker S2015. Fitting linear mixed effects models using lme4. Journal of Statistical Software67:, pp.1–48.

7

Blanca G, Cabezudo B, Cueto M, Salazar C, Morales Torres C2011Flora Vascular de Andalucía Oriental, 2nd edn. Granada, Spain: Universidades de Almería, Granada, Jaén y Málaga.

8

Bolker BM2015. Linear and generalized linear mixed models. In: Fox GA, Negrete-Yankelevich S, Sosa VJ, eds. Ecological statistics: contemporary theory and application. Oxford: Oxford University Press, , pp.309–333.

9

Bolker BM, Brooks ME, Clark CJ, Geange SW, Poulsen JR, Stevens MH, White JS2009. Generalized linear mixed models: a practical guide for ecology and evolution. Trends in Ecology & Evolution24:, pp.127–135.

10

Bonin A, Ehrich D, Manel S2007. Statistical analysis of amplified fragment length polymorphism data: a toolbox for molecular ecologists and evolutionists. Molecular Ecology16:, pp.3737–3758.

11

Cánovas JL, Jiménez JF, Mota JF, Gómez PS2015. Genetic diversity of Viola cazorlensis Gand., an endemic species of Mediterranean dolomitic habitats: implications for conservation. Systematics and Biodiversity13:, pp.571–580.

12

Castroviejo S(coord. gen). 1986–2012Flora iberica 1–8, 10–15, 17–18, 21. Madrid, Spain: Real Jardín Botánico, CSIC.

13

Cole CT2003. Genetic variation in rare and common plants. Annual Review of Ecology Evolution and Systematics34:, pp.213–237.

14

Dowen RH, Pelizzola M, Schmitz RJ, Lister R, Dowen JM, Nery JR, Dixon JE, Ecker JR2012. Widespread dynamic DNA methylation in response to biotic stress. Proceedings of the National Academy of Sciences of the United States of America109:, pp.E2183–E2191.

15

Eckert CG, Samis KE, Lougheed SC2008. Genetic variation across species’ geographical ranges: the central-marginal hypothesis and beyond. Molecular Ecology17:, pp.1170–1188.

16

Ehrich D2006. AFLPdat: a collection of R functions for convenient handling of AFLP data. Molecular Ecology Notes6:, pp.603–604.

17

Ellegren H, Galtier N2016. Determinants of genetic diversity. Nature Reviews. Genetics17:, pp.422–433.

18

Ellstrand NC, Elam DR1993. Population genetic consequences of small population size: implications for plant conservation. Annual Review of Ecology and Systematics24:, pp.217–242.

19

Feinberg AP, Irizarry RA2010. Stochastic epigenetic variation as a driving force of development, evolutionary adaptation, and disease. Proceedings of the National Academy of Sciences of the United States of America107:, pp.1757–1764.

20

Foll M, Fischer MC, Heckel G, Excoffier L2010. Estimating population structure from AFLP amplification intensity. Molecular Ecology19:, pp.4638–4647.

21

Forrest A, Escudero M, Heuertz M, Wilson Y, Cano E, Vargas P2017. Testing the hypothesis of low genetic diversity and population structure in narrow endemic species: the endangered Antirrhinum charidemi (Plantaginaceae). Botanical Journal of the Linnean Society183:, pp.260–270.

22

Foust CM, Preite V, Schrey AW, Alvarez M, Robertson MH, Verhoeven KJ, Richards CL2016. Genetic and epigenetic differences associated with environmental gradients in replicate populations of two salt marsh perennials. Molecular Ecology25:, pp.1639–1652.

23

Frankham R2005. Genetics and extinction. Biological Conservation126:, pp.131–140.

24

Fulneček J, Kovařik A2014. How to interpret methylation sensitive amplified polymorphism (MSAP) profiles?BMC Genetics15:, pp.1–9.

25

Furrow RE, Feldman MW2014. Genetic variation and the evolution of epigenetic regulation. Evolution68:, pp.673–683.

26

Gao L, Geng Y, Li B, Chen J, Yang J2010. Genome-wide DNA methylation alterations of Alternanthera philoxeroides in natural and manipulated habitats: implications for epigenetic regulation of rapid responses to environmental fluctuation and phenotypic variation. Plant, Cell & Environment33:, pp.1820–1827.

27

Gáspár B, Bossdorf O, Durka W2019. Structure, stability and ecological significance of natural epigenetic variation: a large-scale survey in Plantago lanceolata. The New Phytologist221:, pp.1585–1596.

28

Gitzendanner MA, Soltis PS2000. Patterns of genetic variation in rare and widespread plant congeners. American Journal of Botany87:, pp.783–792.

29

Gort G, van Eeuwijk FA2010. Codominant scoring of AFLP in association panels. Theoretical and Applied Genetics121:, pp.337–351.

30

Grativol C, Hemerly AS, Ferreira PC2012. Genetic and epigenetic regulation of stress responses in natural plant populations. Biochimica et Biophysica Acta1819:, pp.176–185.

31

Guarino F, Cicatelli A, Brundu G, Improta G, Triassi M, Castiglione S2019. The use of MSAP reveals epigenetic diversity of the invasive clonal populations of Arundo donax L. PLoS One14:, pp.e0215096.

32

Guevara MA, de María N, Sáez-Laguna E, Vélez MD, Cervera MT, Cabezas JA2017. Analysis of DNA cytosine methylation patterns using methylation-sensitive amplification polymorphism (MSAP). In: Kovalchuk I, ed. Plant epigenetics. Methods and protocols. New York: Humana Press, , pp.99–112.

33

Hampe A, Petit RJ2005. Conserving biodiversity under climate change: the rear edge matters. Ecology Letters8:, pp.461–467.

34

Herden J, Eckert S, Stift M, Joshi J, van Kleunen M2019. No evidence for local adaptation and an epigenetic underpinning in native and non-native ruderal plant species in Germany. Ecology and Evolution9:, pp.9412–9426.

35

Herrera CM, Alonso C, Medrano M, Pérez R, Bazaga P2018. Transgenerational epigenetics: inheritance of global cytosine methylation and methylation-related epigenetic markers in the shrub Lavandula latifolia. American Journal of Botany105:, pp.741–748.

36

Herrera CM, Bazaga P2010. Epigenetic differentiation and relationship to adaptive genetic divergence in discrete populations of the violet Viola cazorlensis. The New Phytologist187:, pp.867–876.

37

Herrera CM, Medrano M, Bazaga P2016. Comparative spatial genetics and epigenetics of plant populations: heuristic value and a proof of concept. Molecular Ecology25:, pp.1653–1664.

38

Herrera CM, Medrano M, Bazaga P2017. Epigenetic and genetic spatial structure of the perennial herb Helleborus foetidus: isolation by environment, isolation by distance, and functional trait divergence. American Journal of Botany104:, pp.1195–1204.

39

Herrera CM, Pozo MI, Bazaga P2011. Clonality, genetic diversity and support for the diversifying selection hypothesis in natural populations of a flower-living yeast. Molecular Ecology20:, pp.4395–4407.

40

Herrera CM, Pozo MI, Bazaga P2014. Nonrandom genotype distribution among floral hosts contributes to local and regional genetic diversity in the nectar-living yeast Metschnikowia reukaufii. FEMS Microbiology Ecology87:, pp.568–575.

41

Hu J, Barrett RDH2017. Epigenetics in natural animal populations. Journal of Evolutionary Biology30:, pp.1612–1632.

42

Jiménez-Mejías P, Fernández-Mazuecos M, Amat ME, Vargas P2015. Narrow endemics in European mountains: high genetic diversity within the monospecific genus Pseudomisopates (Plantaginaceae) despite isolation since the late Pleistocene. Journal of Biogeography42:, pp.1455–1468.

43

Johannes F, Schmitz RJ2019. Spontaneous epimutations in plants. The New Phytologist221:, pp.1253–1259.

44

Jordano P, Godoy JA2000. RAPD variation and population genetic structure in Prunus mahaleb (Rosaceae), an animal-dispersed tree. Molecular Ecology9:, pp.1293–1305.

45

Jueterbock A, Boström C, Coyer JA, Olsen JL, Kopp M, Dhanasiri AK, Smolina I, Arnaud-Haond S, Van de Peer Y, Hoarau G2019. Methylation variation promotes phenotypic diversity and evolutionary potential in a millenium-old clonal seagrass meadow. bioRxiv: 787754. doi:10.1101/787754

46

Jump AS, Marchant R, Peñuelas J2009. Environmental change and the option value of genetic diversity. Trends in Plant Science14:, pp.51–58.

47

Karron JD1987. A comparison of levels of genetic polymorphism and self-compatibility in geographically restricted and widespread plant congeners. Evolutionary Ecology1: , pp.47–58.

48

Kilvitis H, Alvarez M, Foust C, Schrey A, Robertson M, Richards C2014. Ecological epigenetics. In: Landry CR, Aubin-Horth N, eds. Ecological genomics. Dordrecht, The Netherlands: Springer, , pp.191–210.

49

Kronholm I, Collins S2016. Epigenetic mutations can both help and hinder adaptive evolution. Molecular Ecology25:, pp.1856–1868.

50

Kruckeberg AR, Rabinowitz D1985. Biological aspects of endemism in higher plants. Annual Review of Ecology and Systematics16:, pp.447–479.

51

Lämke J, Bäurle I2017. Epigenetic and chromatin-based mechanisms in environmental stress adaptation and stress memory in plants. Genome Biology18:, pp.124.

52

Latzel V, Allan E, Bortolini Silveira A, Colot V, Fischer M, Bossdorf O2013. Epigenetic diversity increases the productivity and stability of plant populations. Nature Communications4:, pp.2875.

53

Lauria M, Echegoyen-Nava RA, Rodríguez-Ríos D, Zaina S, Lund G2017. Inter-individual variation in DNA methylation is largely restricted to tissue-specific differentially methylated regions in maize. BMC Plant Biology17:, pp.52.

54

Lavergne S, Thompson JD, Garnier E, Debussche M2004. The biology and ecology of narrow endemic and widespread plants: a comparative study of trait variation in 20 congeneric pairs. Oikos107:, pp.505–518.

55

Leimu R, Mutikainen P, Koricheva J, Fischer M2006. How general are positive relationships between plant population size, fitness and genetic variation?Journal of Ecology94:, pp.942–952.

56

Lenth R2018emmeans: estimated marginal means, aka least-squares meanshttps://CRAN.R-project.org/package=emmeans. (July 2019)

57

Lira-Medeiros CF, Parisod C, Fernandes RA, Mata CS, Cardoso MA, Ferreira PC2010. Epigenetic variation in mangrove plants occurring in contrasting natural environment. PLoS One5:, pp.e10326.

58

Liu S, Sun K, Jiang T, Feng J2015. Natural epigenetic variation in bats and its role in evolution. The Journal of Experimental Biology218:, pp.100–106.

59

López-Pujol J, Bosch M, Simon J, Blanché C2009. Patterns of genetic diversity in the highly threatened vascular flora of the Mediterranean basin. In: Columbus AM, Kuznetsov L, eds. Endangered species: new research. New York: Nova Science Publishers, , pp.45–79.

60

Lynch M, Milligan BG1994. Analysis of population genetic structure with RAPD markers. Molecular Ecology3:, pp.91–99.

61

Médail F, Diadema K2009. Glacial refugia influence plant diversity patterns in the Mediterranean basin. Journal of Biogeography36:, pp.1333–1345.

62

Médail F, Quézel P1999. Biodiversity hotspots in the Mediterranean basin: setting global conservation priorities. Conservation Biology13:, pp.1510–1513.

63

Medrano M, Herrera CM2008. Geographical structuring of genetic diversity across the whole distribution range of Narcissus longispathus, a habitat-specialist, Mediterranean narrow endemic. Annals of Botany102:, pp.183–194.

64

Medrano M, Herrera CM, Bazaga P2014. Epigenetic variation predicts regional and local intraspecific functional diversity in a perennial herb. Molecular Ecology23:, pp.4926–4938.

65

Melendo M, Giménez E, Cano E, Gómez-Mercado F, Valle F2003. The endemic flora in the south of the Iberian Peninsula: taxonomic composition, biological spectrum, pollination, reproductive mode and dispersal. Flora198:, pp.260–276.

66

Meudt HM, Clarke AC2007. Almost forgotten or latest practice? AFLP applications, analyses and advances. Trends in Plant Science12:, pp.106–117.

67

Moler ERV, Abakir A, Eleftheriou M, Johnson JS, Krutovsky KV, Lewis LC, Ruzov A, Whipple AV, Rajora OP2018. Population epigenomics: advancing understanding of phenotypic plasticity, acclimation, adaptation and diseases. In: Rajora OP, ed. Population genomics: concepts, strategies and approaches. Cham, Switzerland: Springer, , pp.179–280.

68

Mota JF, Pérez-García FJ, Jiménez ML, Amate JJ, Peñas J2002. Phytogeographical relationships among high mountain areas in the Baetic Ranges (South Spain). Global Ecology and Biogeography11:, pp.497–504.

69

Nybom H2004. Comparison of different nuclear DNA markers for estimating intraspecific genetic diversity in plants. Molecular Ecology13:, pp.1143–1155.

70

Paun O, Schönswetter P2012. Amplified fragment length polymorphism (AFLP) - an invaluable fingerprinting technique for genomic, transcriptomic and epigenetic studies. Methods in Molecular Biology862:, pp.75–87.

71

Pérez-Figueroa A2013. MSAP: a tool for the statistical analysis of methylation-sensitive amplified polymorphism data. Molecular Ecology Resources13:, pp.522–527.

72

Preite V, Snoek LB, Oplaat C, Biere A, van der Putten WH, Verhoeven KJ2015. The epigenetic footprint of poleward range-expanding plants in apomictic dandelions. Molecular Ecology24:, pp.4406–4418.

73

Quadrana L, Colot V2016. Plant transgenerational epigenetics. Annual Review of Genetics50:, pp.467–491.

74

R Development Core Team2019R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computinghttps://www.R-project.org/ (July 2019).

75

Reisch C, Bernhardt-Römermann M2014. The impact of study design and life history traits on genetic variation of plants determined with AFLPs. Plant Ecology215:, pp.1493–1511.

76

Reyna-López GE, Simpson J, Ruiz-Herrera J1997. Differences in DNA methylation patterns are detectable during the dimorphic transition of fungi by amplification of restriction polymorphisms. Molecular & General Genetics253:, pp.703–710.

77

Richards EJ2006. Inherited epigenetic variation – revisiting soft inheritance. Nature Reviews Genetics7:, pp.395–401.

78

Richards EJ2008. Population epigenetics. Current Opinion in Genetics & Development18:, pp.221–226.

79

Richards CL, Schrey AW, Pigliucci M2012. Invasion of diverse habitats by few Japanese knotweed genotypes is correlated with epigenetic differentiation. Ecology Letters15:, pp.1016–1025.

80

Schmid MW, Heichinger C, Coman Schmid D, Guthörl D, Gagliardini V, Bruggmann R, Aluri S, Aquino C, Schmid B, Turnbull LA, Grossniklaus U2018. Contribution of epigenetic variation to adaptation in Arabidopsis. Nature Communications9:, pp.4446.

81

Schönswetter P, Tribsch A2005. Vicariance and dispersal in the alpine perennial Bupleurum stellatum L. (Apiaceae). Taxon54:, pp.725–732.

82

Schrey AW, Alvarez M, Foust CM, Kilvitis HJ, Lee JD, Liebl AL, Martin LB, Richards CL, Robertson M2013. Ecological epigenetics: beyond MS-AFLP. Integrative and Comparative Biology53:, pp.340–350.

83

Schulz B, Eckstein RL, Durka W2013. Scoring and analysis of methylation-sensitive amplification polymorphisms for epigenetic population studies. Molecular Ecology Resources13:, pp.642–653.

84

Schulz B, Eckstein RL, Durka W2014. Epigenetic variation reflects dynamic habitat conditions in a rare floodplain herb. Molecular Ecology23:, pp.3523–3537.

85

Searle SR, Speed FM, Milliken GA1980. Population marginal means in the linear model: an alternative to least squares means. American Statistician34:, pp.216–221.

86

Shi W, Chen X, Gao L, Xu CY, Ou X, Bossdorf O, Yang J, Geng Y2019. Transient stability of epigenetic population differentiation in a clonal invader. Frontiers in Plant Science9:, pp.1851.

87

Smithson MW, Dybdahl MF, Nuismer SL2019. The adaptive value of epigenetic mutation: limited in large but high in small peripheral populations. Journal of Evolutionary Biology32:, pp.1391–1405.

88

Souto CP, Mathiasen P, Acosta MC, Quiroga MP, Vidal-Russell R, Echeverría C, Premoli AC2015. Identifying genetic hotspots by mapping molecular diversity of widespread trees: when commonness matters. The Journal of Heredity106:, pp.537–545.

89

Spens AE, Douhovnikoff V2016. Epigenetic variation within Phragmites australis among lineages, genotypes, and ramets. Biological Invasions18:, pp.2457–2462.

90

Stuessy TF, Takayama K, López-Sepúlveda P, Crawford DJ2014. Interpretation of patterns of genetic variation in endemic plant species of oceanic islands. Botanical Journal of the Linnean Society174:, pp.276–288.

91

Thiebaut F, Hemerly AS, Ferreira PCG2019. A role for epigenetic regulation in the adaptation and stress responses of non-model plants. Frontiers in Plant Science10:, pp.246.

92

Thompson JD, Lavergne S, Affre L, Gaudeul M, Debussche M2005. Ecological differentiation of Mediterranean endemic plants. Taxon54:, pp.967–976.

93

Totté A, Delgado A, Navarro T, Meerts P2015. Narrow endemics of the Almeria Province (Andalusia, Spain) differ in their traits and ecological niche compared to their more widespread congeners. Folia Geobotanica50:, pp.137–150.

94

Verhoeven KJ, Jansen JJ, van Dijk PJ, Biere A2010. Stress-induced DNA methylation changes and their heritability in asexual dandelions. The New Phytologist185:, pp.1108–1118.

95

Vos P, Hogers R, Bleeker M, Reijans M, van de Lee T, Hornes M, Frijters A, Pot J, Peleman J, Kuiper M, Zabeau M1995. AFLP: a new technique for DNA fingerprinting. Nucleic Acids Research23:, pp.4407–4414.

96

Wang MZ, Li HL, Li JM, Yu FH2020. Correlations between genetic, epigenetic and phenotypic variation of an introduced clonal herb. Heredity124:, pp.146–155.

97

Weising K, Nybom H, Wolff K, Kahl G2005DNA fingerprinting in plants: principles, methods, and applications, 2nd edn. Boca Raton, FL: CRC Press.

98

Wibowo A, Becker C, Marconi G, Durr J, Price J, Hagmann J, Papareddy R, Putra H, Kageyama J, Becker J, Weigel D, Gutierrez-Marcos J2016. Hyperosmotic stress memory in Arabidopsis is mediated by distinct epigenetically labile sites in the genome and is restricted in the male germline by DNA glycosylase activity. eLife5:, pp.e13546.

99

Wilschut RA, Oplaat C, Snoek LB, Kirschner J, Verhoeven KJ2016. Natural epigenetic variation contributes to heritable flowering divergence in a widespread asexual dandelion lineage. Molecular Ecology25:, pp.1759–1768.

100

Zuur AF, Ieno EN, Walker NJ, Saveliev AA, Smith GM2009Mixed effects models and extensions in ecology with R. New York: Springer.

Citing articles via
https://www.researchpad.co/tools/openurl?pubtype=article&doi=10.1093/aobpla/plaa013&title=Comparative genetic and epigenetic diversity in pairs of sympatric, closely related plants with contrasting distribution ranges in south-eastern Iberian mountains&author=Mónica Medrano,Conchita Alonso,Pilar Bazaga,Esmeralda López,Carlos M Herrera,Mohamed Abdelaziz,&keyword=AFLP,DNA methylation,endemism,epigenetic diversity,genetic diversity,Mediterranean mountains,MSAP,population epigenetics,&subject=Special Issue: The Ecology and Genetics of Population Differentiation in Plants,Studies,AcademicSubjects/SCI01210,