Malus sylvestris (Mill.) is the only indigenous wild apple species in Central Europe. Agriculture, forestry and urbanization increasingly endanger Malus sylvestris natural habitats. In addition, the risks of cross-hybridization associated with increase in the cultivation of the domesticated apple Malus ×domestica (Borkh.), threatens the genetic integrity of M. sylvestris.

The present study investigated the number of hybrids, genetic diversity and genetic structure of 292 putative M. sylvestris that originate from five different natural M. sylvestris populations in Saxony, Germany. All samples were genetically analyzed using nine nuclear microsatellite markers (ncSSR) and four maternally inherited chloroplast markers (cpDNA) along with 56 apple cultivars commonly cultivated in Saxony.

Eighty-seven percent of the wild apple accessions were identified as pure M. sylvestris. The cpDNA analysis showed six private haplotypes for M. sylvestris, whereas three haplotypes were present in M. sylvestris and M. ×domestica. The analysis of molecular variance (AMOVA) resulted in a moderate (ncSSR) and great (cpDNA) variation among pure M. sylvestris and M. ×domestica individuals indicating a low gene flow between both species. The genetic diversity within the pure M. sylvestris populations was high with a weak genetic structure between the M. sylvestris populations indicating an unrestricted genetic exchange between these M. sylvestris populations.

The clear distinguishing of M. sylvestris and M. ×domestica confirms our expectation of the existence of pure M. sylvestris accessions in this area and supports the argument for the implementation of preservation measures to protect the M. sylvestris populations in Saxony. 

Sampling

Two hundred and ninety-two potential M. sylvestris trees sampled from five different areas within Saxony, Germany formed the basis of this study. The mapped trees were mainly found in sparse forests, along forest edges, stonewalls or other open landscape structures. The selection of trees was based on several morphological characters described by Wagner (1996) including fruit diameter < 3.5 cm, no fruit over color, absence of hairiness in the leaf lower surface, for example. In addition, 56 apple cultivars (M. ×domestica Borkh.) often cultivated in Germany were included as control genotypes. Leaf material of these 56 apple cultivars was provided by the Fruit genebank of the Institute of Breeding Research on Fruit Crops of the Julius Kühn-Institute (JKI) at Dresden-Pillnitz. Fresh leaf materials were collected in 2 ml tubes and dried using silica gel according to a modified protocol of (Chase & Hills, 1991). Leaf materials were stored at room temperature until DNA isolation. LGC Genomics (Berlin, Germany) performed DNA isolation and quantification. All samples were diluted to 10 ng/ µl.

Genetic analysis

Nine ncSSR primer pairs (CH01H10, CH04C07, CH01H01, Hi02C07, CH01F03b, GD147, CH01F02, CH02C09, GD12) developed for M. ×domestica (Liebhard et al., 2002) were combined in three multiplexes with three primers per multiplex. . Four cpDNA primer pairs located at the matK_dup (NCBI GenBank: AF309231), rps16 intron (NCBI GenBank: JQ391664) region and the intergenic spacer region rps16_trnQ (NCBI GenBank: AB604694) and rpl2_trnH (NCBI GenBank: JQ392141), respectively, were developed based on Malus sequence data obtained from the NCBI database (www.ncbi.nlm.nih.gov). Multiplex PCR was performed using the Type-It kit (Qiagen, Germany) according to the manufacturer’s protocol with three microsatellites per PCR in a total volume of 10 µl. PCR fragments were analyzed on a CEQ 8000XL Genetic Analyzer System (Beckman Coulter, Germany), for which forward primers were labelled with BMN-5, BMN-6 and DY751 (Biomers, Germany). PCR products were then diluted 1:100  using sterile water and 2 µl of this dilution was mixed with 29.9 µl of sample loading solution (Beckman Coulter) and 0.1 µl of size standard.

Data analysis

First, samples were grouped into pure M. sylvestris, M. ×domestica or hybrids using the model-based clustering method with STRUCTURE software version 2.3.4. (Pritchard, Stephens, & Donnelly, 2000). In order to improve the accuracy of the inference, the analysis was performed with prior information on the population (POPINFO model) in which each individual was defined either as ‘M. sylvestris’ or ‘M. ×domestica’. To accurately classify the individuals, 5 runs with K = 2 was performed with the program. The parameters were 50,000 burn-in periods and 50,000 Markov Chain Monte Carlo repetitions using the admixture model with correlated allele models. All M. sylvestris individuals with a probability of lower than 90 % that the assigned population is right were classified as hybrid. The probability of membership to the M. ×domestica genepool for the 56 reference M. ×domestica cultivars was > 90 % indicating almost a zero chance of introgression from M. sylvestris.   

Secondly, the number of natural M. sylvestris populations (K) were estimated with STRUCTURE. For this, we run the program again including all pure M. sylvestris accessions with K varying from 2 to 5 with 5 runs for each K value without POPINFO model. The remaining parameters were described as above. STRUCTURE HARVESTER (Earl & Vonholdt, 2012) was used for detecting the most likely value for K based on Evanno’s ΔK method (Evanno, Regnaut, & Goudet, 2005).

Statistical analysis

The mean number of alleles by locus (N_a), effective number of alleles (N_e), observed heterozygosity (H_o), expected heterozygosity (H_e) and number of private alleles (PA) were calculated for each ncSSR loci within the pure M. sylvestris individuals and the apple cultivars using the software GENALEX ver. 6.5 (Peakall & Smouse, 2006, 2012). Additionally, the allele frequencies of the chloroplast DNA markers were compared among the M. sylvestris individuals and the apple cultivars.

An ‘Analysis of Molecular Variance’ (AMOVA) was performed based on nuclear and chloroplast marker data. The molecular variance (ɸ_PT, an analogue of F_st) and the migration rate (N_m) was calculated among the pure M. sylvestris genotypes and the apple cultivars as well as among the different M. sylvestris populations using GENALEX ver. 6.5.

The correspondence (rxy) between geographic and genetic distance was performed by Mantel test with statistical testing by 9,999 permutations using the software GENALEX ver. 6.5 (Mantel, 1967).

Missing genetic data are indicated with -9.