Genetic Diversity in Prunus spinosa L. and Challenges in the Use of Autochthonous Sources

Genetic diversity in natural stands of autochthonous blackthorn (Prunus spinosa L.) of different German provenances has been analyzed using a highly reproducible high-annealing-temperature random amplified polymorphic DNA (HAT-RAPD) protocol. The findings were compared to those from seedstocks of the same provenances, reported earlier. Generally, genetic diversity in the natural stands was even lower (Ho 0.099–0.116) compared to the corresponding seedstocks (Ho 0.118–0.133). Furthermore, genetic differentiation was found to be moderate between natural residential sources (pairwise Fst 0.138–0.184, 22.527% variation among populations), but higher than between the seedstocks (pairwise Fst 0.086-0.104, 7.782% variation among populations). The findings are discussed in respect to German conservation law and its practical implementation.


INTRODUCTION
Genetic diversity is generally agreed to be a main prerequisite for evolution providing raw material for selection (Rees et al. 2001, Crawford andWhitney 2010).Also natural genetic diversity provides raw material for breeding food and feed varieties (Hoisington et al. 1999, Esquinas-Alcázar 2005).While artificially increasing genetic diversity by mutation breeding or artificial crosses across breeding borders can increase genetic diversity, naturally occurring genetic diversity remains the largest pool of genes/allels to ensure healthy environments and secured food situation even during challenges like socioecological development and global climate change.Thus, conservation of genetic diversity (by protection of environment) was agreed upon as a common goal for humanity (Convention on biological diversity 1992).Those ageements have to be implementet into national law by the subscribing contries.In Germany, the Federal Conservation Act was accordingly amended in 2010 (BnatSchG) to (among other points) include conservation of genetic diversity.Thus, it also regulates the introduction of non-resident species into the open landscape (outside settled areas and not used in agri-, horti-, or silviculture) -which will be prohibited from 2020 on.Until then there is a period of transition during which use of residential sources for any species is required whenever possible.In the BnatschG, the term species is defined including sub-species levels like subspecies or even populations (BnatSchG 2010, §7(2)3).This regulation presents a practical problem for the use of plants in open landscape plantation.While an increasing number of publications are available for forest trees (Kremer et al. 2002, Petit et al. 2003, Magri et al. 2006) not much is known about genetic constitution of many other plants including shrubs, widely used in open landscape plantings.Despite this lack of knopwledge, 9 regions of origin were introduced for woody plants (shrubs), not regulated by forrestry or agricultural laws in 2003 (BMVEL 2003).Base for those regions were ecological basic units based on geographical classification of natural landscapes in Germany according to Schmidt and Krause (1997).This regulation does not take into account finer ecological structuring nor does it consider the different biologies of plants species.In 2012 the number of regions of origin was further reduced to 6 (BMU 2012).
Faced with this situation we aimed to broaden our knowledge on one of the species most commenly used for plantings in the open landscape -blackthorn (Prunus spinosa L.).Prunus spinosa L. (common blackthorn, sloe) is an insect-pollinated, animaldispersed shrub native to Europe, North Africa, and West Asia (Schütt et al. 1992).It is very wide-spread over Germany and most of Europe, and therefore it is often used in open landscape plantation and renaturation measures in Germany.Blackthorn is assumed to be allo-tetraploid (2n=4x=32, Reynders-Aloisi and Grellet 1994) and it also propagates strongly by root suckers (Guitian et al. 1993).
In the recent years, several publications emerged dealing with the genetic structure of blackthorn.Mohanty et al. (2000Mohanty et al. ( , 2002) ) analyzed cp-DNA to observe large scale genetic structure in Europe, including a few scattered samples from Germany.Isozyme analyses for several areas in Germany showed moderate to low nuclear genetic diversity within and among blackthorn populations (Leinemann et al. 2002, Fronia 2009)

Plant material
Fully expanded leaves were collected from 40 individuals of each of the source populations which provided the seedstocks analyzed earlier (Eimert et al. 2012).That may or may not include the actual shrubs from which those seeds were collected, as those were not labeled at that time.Populations sampled (Figure 1) were located in (at that time) two regions of origin (BMVL 2003).Populations "Hö" (near the town of Höxter in the State of North Rhine-Westphalia) and ''Fu'' (near the town of Fulda in the State of Hessia) belong to the designated region of origin 4, while population "Rh" (in the Rheingau Region of the State of Hessia ) was located in the then region of origin 6.For samples from population sampled in wild stands "WT" is added to the label.For the corresponding seedstocks, samples of the corresponding regions of origin from earlier studies (Eimert et al. 2012) were reanalyzed.
Those poluplation are additionally labelled "Nu" (nursery material) and "ud." in case undifferentiated material was used.(Jensen et al. 2005).Genetic structure within and among the sampled populations was also analyzed using a Bayesian approach implemented in the software Structure (Pritchard et al. 2000).Geographic distances beween populations were estimated using the Google EarthTM (Google 2009) ''ruler'' tool.

Population statistics
All 360 Individuals of the different sources were analysed by HAT-RAPD using 11 primers resulting in 390 scorable band classes.While a high number of those bands were polymorphic, the gene diversity within the populations remained rather low (Table 1).Fixed bands and private bands could be observed in every poulation.However, no fixed private bands could be identified.Also, it can be seen, that heterozygosity in all three wild populations is even lower than in the corresponding seedstock populations.
Using a stepwise mutation model with a 99.5% confidence intervall most of the loci used as markers in this study seem to behave selectivelly neutral in the wild autochthonous populations.Nevertheless, several Fst outliers could be detected identifying possible candidates for balancing (7 loci) and for positive (12 loci) selection (Figure 2).
Tajima's test of selective neutrality revealed no significant deviations from a neutral model in any of the tested populations (Table 2).While there are no larger clusters consisting uniformly of individuals of only one given population, one main clade consisted most of the individuals from sampled from the three wild stands, a second clade included most of the autochthonous seedstocks (except for Ba(Nu)) and a third one which combined most of the commercial seedstocks and the Bavarian autochthonous seedstock (Figure 3).The standard similarity was computed using Jaccard's similarity coefficient.Standard Jaccard distance transformation (d=1-s) was applied to calculate pairwise Fst values to access differentiation among populations (Table 3).Fst values were calculated for both neutral and non-neutral markers and no significant difference was observed (data not shown).
While the differentiation among the seedstock populations was usually weak there is mostly moderate to even strong differentiation between the wild populations and the seedstock populations and also among the wild populations themselves.Accordingly, the derived dendrogram apparently shows three larger clusters -a wider one consisting of all wild populations with the Rh(WT) population slightly removed and two tighter ones, one consisting of the autochthonous seedstocks (with the exception of Ba(Nu)) and a another one consisting of the two undifferentiated commercial seedstocks and the Bavarian autochthonous seedstock (Figure 4).Multivariate analysis (PCoA) reveals that the wild populations are most differentiated with the Rh (WT) beeing the most distant (Figure 5).The wild populations are losely grouped while the seedstock populations group tighter and the autochthonous seedstocks group together (except Ba(Nu)) and even farther then the commercial ones.An AMOVA revealed that, in all the seedstock populations the higher genetic diversity was mostly due to variation within the populations (92.218%) and less among them (7.782%).In the wild populations, the generally lower genetic diversity is shifted to 77.473% variation within and 22.527% among populations, respectively.However, using Fisher's exact test for population differentiation (Raymond and Rousset 1995) no differentiation between any of the analyzed populations could be detected at the 95% significance intervall (after 30 000 Markov transformations).

Genetic Structure -Bayesian Estimates
Possible genetic structures were also analyzed using the Bayesian approach based on the most likely k value (Evanno et al. 2005).Using the most sensitive settings a weak genetic structure (ΔK=48.33)was shown with the most likely k=6 (Figure 6).and only very few loci seemed to be under selective pressure as judged from the number of Fst outliers.Also, the fact that no fixed private bands could be detected in 390 markers seems to contradict an IBA (isolation by adaption) scenario.If, on the other hand, the differentiation was due to gene flow and drift assumed for the neutral theory of molecular evolution (Kimura 1968) one would expect an isolation by distance (IBD) effect for populations geographically and or ecologically seperated from each other.A Mantel test on that account did not confirm such effect.Furthermore, the amount of gene flow calculated from these weak Fst values is accordingly high with Nm=5.558 (Nm = 0.5(1 -Fst)/Fst) meaning that to explain this low differentiation between our aotuochthonous populations a migration of more than 5 individuals per generation into each population would be required.As the geographical distance between those populations is more than 130 -210 km this seems an unlikely event.On the other hand, the measurement of gene flow cannot distinguish between recent or historical events (Lowe et al. 2008).Humans have apparently used blackthorn fruits for consumption since the neolithic (Karg andMarkle 2002, Martin et al. 2008) and it was cultivated in Nothern Europe since the Roman Empire (Karnitsch 1953).Thus, anthropogenic distribution of blackthorn seems very plausible.
It also has been noted, that in many cases the impact of processes on genetic structuring can be obscured (Orsini et al. 2013).Thus, it can be difficult to distinguish the results of IBD and IBC (isolation by colonization).These authors propose to study the patterns of variation of both neutral and non-neutral markers in order to distinguish between the different driving forces of genetic differentiation.Using those parameters, we observe a pattern more akin to that described bei Orsini et al. (2013) as typical for IBC rather than that for IBD or IBA (isolation by adaption) in that the Fst values for neiter the neutral nor the non-neutral loci vary significantly over geographical distance.
Summarizing, we observe a weak pattern of differentiation between wild autochthonous populations of blackthorn in Germany.No obvious deviation from neutral evolution was detected and, thus, isolation by distance or by adaption seems unlikely.The observed patterns support an (weak) isolation by colonialization probably further driven by landscape fragmentation.
On the other hand, the situation in seedstock populations collected from those wild stands is different.That the genetic diversity within those populations is higher is not surprizing by itself, considering the almost ubiquitous occurrence of blackthorne in Germany (although mostly in fragmented smaller stands), its outcrossing biology and tetraploidy.At the same time, the genetic differentiation between the seedstocks populations is smaller than that between their corresponding source populations.That leads to the situation that, genetically, the wild autochthonous populations are more similar to each other than to their own more heterozygous offspring (Fig. 3).This situation presents some problems for conservation measures.If the main function of the Conservation Law is to maintain the genetic status quo of an existing population/sub-population the usage of residential seedstocks will not necessarily accomplish that goal in full.If the weak differentiation we observed in wild autochthonous blackthorn populations is indeed mainly due to colonization even the use of autochthonous seedstock will not maintain the same narrow genetic composition.That will obviously depend on the concerned species, its current genetic composition and its reproductive biology.Thus, from a biological point of view an accordingly adapted approach for each species would be appropriate.
For blackthorn it seems that, at least for the regions of origin 4 and 6, the observed genetic situation does not correspond with the proposed geographic differentiation, as has been shown for other regions (Leinemann et al. 2014).A more thorough analysis of populations from all regions of Germany would be required to propose corresponding and biologically meaningful regions of origin.A recently implemented further reduction of the proposed number of regions of origin for autochthonous plants from 9 to 6 (BMU 2012) might be nearer to the "genetic truth" for blackthorn.Nonetheless, the situation will have to be evaluated for each species separately.

Figure 1
Figure 1 Schematic geographical map of Germany showing the sources of autochthonous plant material used in the recent analysis.Populations: Hö (Höxter), Fu (Fulda), Rh (Rheingau); Numbers refer to the designated regions of origin (BMELV 2003); bar 100 km

Figure 2
Figure 2 Fst outliers detected by LOSITAN in autochthonous wild populations (red area -candidates for positive selection, yellow area -candidates for balancing selection).

Figure 3
Figure 3 Radial cladogram showing the clustering of all screened individuals (neighbor joining, Jaccard).

Figure 4
Figure 4 Dendrogram of the studied populations based on UPGMA analysis of pairwise distances using the Jaccard similarity index: Areas bordered by slashed lines indicate the three main clusters.

Figure 5
Figure 5 Multivariate (PCoA) analysis of seedstock and wild populations

Table 1
1=bands which are monomorphic in a given population; 2=bands which are found exclusively in one given population; 3=bands which are found exclusively in one given population and are monomorphic in that population; 4 h=Nei's (1978) unbiased gene diversity (= measure for heterozygosity); 5 ud.=undifferentiated.

Table 2 :
Tajima's test of selective neutrality