Genetic Diversity of Apple Clonal Rootstocks from the Collection of the Michurinsk State Agrarian University Based on SSR Markers

The Michurinsk State Agrarian University (Michurinsk SAU) is one of the leading centers for breeding apple clonal rootstocks. A diverse collection of apple rootstocks, founded in 1930s by V.I. Budagovsky, is maintained at the Michurinsk SAU. In the present study, 87 rootstocks from this collection were analyzed using 18 SSR markers to assess their genetic diversity and relatedness. The detected polymorphism level was rather high compared to the previous estimates of apple rootstock genetic variability. A total of 199 alleles were detected with an average of 11.1 alleles per locus. Among the detected alleles, 67 (33.67%) were rare and 43 (21.61%) were unique. The average PIC value was 0.73, and the expected and observed heterozygosity averaged 0.76 and 0.69, respectively. All the studied accessions except two could be identified with the used marker set. Cluster analysis revealed several groups according to the rootstocks’ pedigrees and genetic origin. Furthermore, Structure analysis revealed two main groups of the studied rootstock accessions. No significant differentiation of the studied sample according to dwarfing ability was detected, while weak differentiation was detected according to leaf color. SSR genotyping data can be used for rootstock fingerprinting and pedigree verification and will facilitate collection management. In addition, data on the genetic diversity and structure of the studied collection may be useful for further development of the Michurinsk SAU rootstock breeding program.


Introduction
The cultivated apple (Malus domestica Borkh.) is the main fruit crop in many regions of the world, including Russia. The choice of a scion-rootstock combination is of great importance for the commercial production of apples. Clonal rootstock affects the growth and development of scion, resistance to biotic and abiotic stresses, productivity, and also greatly facilitates the production of high-quality planting material.
The universally recognized world standards for apple clonal rootstocks are various 'M9' clones. However, significant differences in climatic conditions, biological factors (e.g., diseases and pests) and management practices in different apple-growing regions of the world determine the need for new resistant and adaptive rootstocks. There are several apple rootstock breeding programs in Russia, Europe, North America, Asia, New Zealand, etc. [1][2][3][4][5][6].
Due to harsh climatic conditions (first of all extremely low winter temperatures), apple orchards in Russia for several centuries have been planted on seedling rootstocks from the The observed heterozygosity (H o ) value ranged from 0.24 (CH03d01) to 0.86 (CH04f10) and averaged 0.69. The mean expected heterozygosity (H e ) was 0.76 and varied from 0.59 (CH03d01) to 0.88 (CH04f10). Polymorphism information content (PIC) was not lower than 0.57 (CH03d01) and averaged 0.73 (Table 1).
The average number of genotypes identified using one marker was 23.39, ranging from 12 (CH01f03b) to 46 (CH04f10). The 18 selected SSR markers allowed for the identification of 86 different genotypes among the 87 studied accessions. All the studied accessions had a unique SSR profile, except rootstocks '71-3-137' and '71-3-150' which had identical fingerprints. Moreover, three loci (CH02c02a, CH02c09 and CH04f10) are sufficient to identify each of the 86 genotypes.
The Dice genetic similarity coefficient was calculated for each pair of studied accessions. The mean Dice coefficient value was 0.36 and ranged from 0.04 to 1. The maximum similarity was found for a pair of accessions that had the same set of alleles ('71-3-137' and '71-3-150'). For 'B9' and '57-146', the Dice genetic similarity coefficient was also high (0.96). The lowest level of similarity was found for rootstocks '57-491' and '86-6-12'.
The second group (gray bars) consisted of 34 accessions. The composition of this group was similar to composition of the first four clusters on the dendrogram. This group included three foreign rootstocks ('M9 T337', 'G16' and 'MM106'), two accessions ('K-1' and 'B7-35'), which have 'M9' in their parentage, two local rootstocks obtained from wild Malus species ('14-1' and 'Babarabskaya yablonya'). Out of ten accessions with '82-27-6' in their parentage, nine were included in the second group. Other accessions from this group had different origins. The remaining 23 accessions were determined to be admixed, having components of both groups ( Figure 3).   . Probability of apple rootstock accession assignment to one of the groups (yellow colorgroup 1, gray color -group 2). Each accession is represented by a vertical bar partitioned into K = 2 segments. Under the Structure graph, the color of rootstock leaves (red or green) is shown.
The divergence between the two groups revealed by Structure analysis was evaluated by AMOVA (Table 2). This analysis revealed that a significant part of the variance (14%; p < 0.01) was ascribed to differences among the two detected groups.
A possible differentiation linked to phenotypic traits (leaf pigmentation and dwarf- had different origins. The remaining 23 accessions were determined to be admixed, having components of both groups (Figure 3).
The divergence between the two groups revealed by Structure analysis was evaluated by AMOVA (Table 2). This analysis revealed that a significant part of the variance (14%; p < 0.01) was ascribed to differences among the two detected groups. A possible differentiation linked to phenotypic traits (leaf pigmentation and dwarfing ability) was also investigated applying AMOVA (Table 2). Only 1% (p < 0.09) of the total variation occurred between groups with different dwarfing abilities (VD; VD + D; D/SD + SD; SD/I + I + I/V). Furthermore, the differentiation between groups with different leaf color (red or green) accounted for about 2% (p < 0.01) of the total variation (Table 2).

Discussion
The studied sample of 87 apple rootstocks from the collection of the Michurinsk State Agrarian University including 3 foreign rootstocks ('M9 T337', 'MM106' and 'G16') and 3 rootstocks from other Russian breeding centers ('B7-35', 'K-1' and '7-8-5' ('Ural 5')) was rather diverse: 199 alleles were detected with 11.1 alleles per locus and He = 0.76. Although there are few genetic diversity studies of apple rootstock collections, the polymorphism of the studied sample was still rather high, considering that mainly rootstocks from the Michurinsk SAU breeding program were studied.
Previously, a set of 66 rootstock clones of apples, representing random samples of rootstock varieties from major apple-growing regions in the world maintained at the apple gene bank of NIFTS, Morioka, Japan, was studied using seven SSR markers. As a result, 68 alleles were detected with an average of 9.7 alleles per locus and a mean heterozygosity value of 0.73 [28]. Another study of 41 rootstocks originating from Europe, North America and China from the Chinese germplasm centers using 62 SSR markers revealed 737 alleles with an average 11.9 alleles per locus [29].
All the 18 studied loci produced one or two discretely amplified fragments and allowed for the detection of 86 unique genotypes among 87 accessions. Rootstocks '71-3-137' and '71-3-150' derived from '58-257' × 'B9' had identical alleles in all studied SSR loci. These accessions are very similar phenotypically and may be the same genotype, or the resolving power of the chosen markers is not sufficient. Still, the rest of the studied sample can be successfully identified using a set of 18 markers, even rootstocks with a common pedigree.
Marker CH04f10 was the most polymorphic and allowed for the detection of 21 alleles and 46 genotypes (Table 1). Three loci (CH02c02a, CH02c09 and CH04f10) were sufficient to identify each of the 86 genotypes. The obtained SSR genotyping results can be used for future data comparison between different studies.
On the whole, the informativeness of the selected marker set (average PIC value 0.73) was comparable to the studies of Japanese (average PIC value 0.81) and Chinese (average PIC value 0.606) rootstock collections [28,29]. Though, in the study by Oraguzie et al. (2005) [28], only seven markers were used, while the other study used significantly more markers (62) but the results were visualized using silver-stained polyacrylamide gels [29]. In the studies of large apple germplasm collections using different SSR marker sets, the PIC value was 0.80 for apple cultivars from different French collections and repositories [21], 0.81 for local Italian cultivars and 0.80 for local and introduced accessions from north-eastern Italy [24,25].
The genetic diversity of apple clonal rootstocks is considered limited because the Malling rootstock series was the founding germplasm for all apple rootstock breeding programs and was the source of dwarfing and precocity. The Michurinsk SAU breeding program was no exception. The first rootstocks bred at the Michurinsk SAU were obtained from 'M8' (East Malling, UK) and local cold hardy cultivars [8]. , derived from species of Section Sorbomalus, were also included in the second Structure group. These accessions were the most genetically differentiated from all the other studied samples ( Figure 1) and had the largest number of unique alleles: nine were detected for rootstock 'G16' and seven for '14-1'.
In the Michurinsk SAU breeding program, Malus species (M. prunifolia, M. sieboldii and M. floribunda) were also used as sources of resistance to heat, high soil salinity and various leaf blots. The development of rootstocks with fire blight and woolly aphid resistance is also one of the objectives of the breeding program. Furthermore, in the Michurinsk SAU, much attention has always been paid to the high cold hardiness of the rootstock root system due to the unfavorable climatic conditions of many Russian apple-growing regions [8]. In this regard, cold resistant species and cultivars derived from the species M. baccata (cv. 'Tayezhnoye') and M. prunifolia (cvs. 'Kandil Kitayka' and 'Pepin Shafranny') were used as sources of high winter hardiness. As a result, most of the Michurinsk SAU rootstocks (e.g., '54-118', '57-490', '62-396', '67-5(32)', '70-20-20') withstand soil temperatures down to −16 • C, which is confirmed by field observations and controlled freezing tests.
A unique biological trait of Michurinsk SAU apple rootstocks is the purple-red color of young leaves and shoots, which are associated with anthocyanin accumulation. This trait is inherited from red-fleshed apple cultivars 'Krasniy Shtandart' ('Red Flag') and 'Rubinovoe' ('Ruby'), derived from M. niedzwetzkyana by I.V. Michurin. Red leaf color is common for the Michurinsk rootstocks; however, it does not seem to provide significant adaptive advantages, though there have been studies of the effect of anthocyanins on heat resistance and tree nutrition [35,36]. Still, it is used in breeding to confirm hybridity and is convenient when removing rootstock shoots from grafted trees in a nursery.
The differentiation of accessions according to leaf colour was studied. AMOVA analysis showed weak (2%) but statistically significant (p < 0.01) differentiation among the groups of accessions with or without antocyanin pigmentation ( Table 3). Out of 43 studied red-leaved accessions, 41 were included in the fifth cluster on the dendrogram. While the first four clusters were composed predominantly of green-leaved accessions (except '70-20-20' and '4-6-5') as well as the second group on the Structure graph (Figures 1 and 3). The differentiation of accessions with different growth habits was also analyzed. No clear differentiation of rootstocks according to dwarfing ability was revealed, although the loci (CH03a09 and CH02d08) associated with dwarfing were analyzed ( Table 2).
The present study provides the first insight into the genetic variation of apple rootstocks from the collection of the Michurinsk State Agrarian University, obtained over 90 years. The studied collection is rather diverse, which is the result of the use of wild Malus species and hybrid forms in the breeding program along with the local and foreign old and commercial apple cultivars. SSR genotyping data provide valuable information for the proper characterization of the plant material preserved in the collection, including rootstock identification and pedigree clarification, and for the further development of the Michurinsk SAU rootstock breeding program.

Plant Material and DNA Extraction
The plant material for the study included 87 apple rootstocks from the collection of the Michurinsk State Agrarian University with different dwarfing abilities and leaf colors (Table 3). Total genomic DNA was extracted from fresh young leaves using Quick-DNA Plant/Seed Miniprep Kit (Zymo Research, Irvine, CA, USA), according to the manufacturer's protocol. DNA samples extracted were quantified using a NanoDrop OneC (Thermo Scientific, Waltham, MA, USA) spectrophotometer.
PCR reactions were performed in T100 Thermal Cycler (BioRad, Hercules, CA, USA) in a final volume of 15 µL containing 20 ng of genomic DNA, 0.2 mM of each dNTP, 1.6 mM MgCl 2 , 1 × PCR buffer, 0.3 µM forward and reverse primers and 0.5 U of BioTaq DNA polymerase (Dialat Ltd., Moscow, Russia). Forward primers were labeled with four different fluorescent dyes (6-FAM, R6G, TAMRA and ROX). All the SSR loci were amplified as described in Gianfranceschi et al. [12] and Liebhard et al. [14] with minor modifications. All microsatellites were amplified separately and combined in multiplexes depending on the size range after PCR products were checked on 1.5% agarose gels in 1X TBE buffer and visualized by staining with ethidium bromide to test for the presence of PCR products.

Data Analysis
The frequencies of observed microsatellite alleles (rare-less than 5% of the accessions; and unique-less than 1%) and the expected (H e ) and observed (H o ) heterozygosity value of each microsatellite were measured using the GENALEX 6.41 software [38].
Based on the frequencies of observed microsatellite alleles, the polymorphism information content (PIC) was calculated as: where P i and P j are the population frequency of the ith and jth allele [39] in MS Excel. Dice coefficients were used for genetic similarity estimation and to visualize genetic relationships among the studied accessions by an UPGMA (unweighted pair group method with arithmetic mean) clustering method, using MEGA 11 [40].
Genetic structure analysis of the collection was performed using Structure v.2.3.4 software [41]. From 1 to 10 clusters (K) with 5 replicates for each K were tested. The number of possible clusters was found as a result of 1,000,000 iterations of the Markov chain Monte Carlo, taking into account genetic admixture and correlated allele frequencies.
The first 300,000 generations were eliminated (burn-in). The optimal number of clusters was determined as recommended by Evanno et al. [42] using the online program Structure Harvester [43]. The genotypes were assigned to one of the groups when the assigning probability was ≥80%.
The divergence between the groups revealed by Structure analysis and differentiation of accessions depending on leaf color (red/green) and dwarfing ability (VD; VD/D + D; D/SD + SD; SD/I + I + I/V) was investigated with Analysis of Molecular Variance (AMOVA) in the GENALEX 6.41 software [38]. The threshold for statistical significance was determined by running 999 permutations.