Agricultura 16: No 1-2:1-10(2019) https://doi.org/10.18690/agricultura.16.l-2.l-10.2019 University of Maribor Press Determination of Sambucus Interspecific Hybrid Structure using Molecular Markers Metka ŠIŠKO*, Anja IVANUŠ, Anton IVANČIČ University of Maribor, Faculty of Agriculture and Life Sciences, Pivola 10, 2311 Hoče, Slovenia ABSTRACT Phenotypic and genotypic variations within the genus Sambucus are limited. They could be efficiently increased by genetic recombination involving different species. The aim of the presented investigation was to assess the possibility of using molecular approach (i.e., microsattelites) in determination of unknown hybrid structures. The study involved 47 Sambucus genotypes (parental species and intespecific hybrids), and six microsatellite loci were analysed. The clustering method grouped the analysed genotypes into four main groups. The first main group involved two sub-groups: one with taxons and hybrids involving S. racemosa (in broad sense) and the other involving two botanical varieties of S. nigra. In the second main group, there was a sub-group involving hybrids between S. javanica and S. nigra, and a sub-group involving F1 hybrids between S. javanica and S. ebulus. The third main group contained a sub-group with hybrids between S. javanica and S. nigra 'Black Beauty', a sub-group with hybrids involving S. javanica, S. nigra and S. racemosa (miquelii), a sub-group with backcrosses S. javanica x (S. javanica x S. ebulus), and un unknown hybrid. The fourth main group included a subgroup with F1 hybrids S. javanica x S. ebulus, a sub-group involving various taxons of S. racemosa (in broad sense), and a sub-group involving hybrids between S. cerulea and S. javanica, with and without S. nigra. Our study shows that molecular analysis can be helpful in determining some of the unknown but simple interspecific hybrids of Sambucus. In the cases of complex hybrid combinations, the use of SSRs is most probably not the best solution. Key words: Sambucus, SSR markers, interspecific hybrids, clustering method, molecular markers INTRODUCTION Elderberries belong to the genus Sambucus and the family Adoxaceae (Bolli,1994). Since ancient times, they have been considered as very useful, especially as medicinal and food plants. The most valuable are their fruits and inflorescences, although useful substances can be found also in roots, bark, leaves and shoots (Shokrzadeh et al., 2009, Atkinson and Atkinson, 2002, Charlebois et al., 2010, Vlachojannis et al., 2010, Mikulic-Petkovsek et al., 2015a, Mikulic-Petkovsek et al., 2015b, Todorovic et al., 2017). Elderberries can be found in almost all regions of the world, except deserts and extremely cold areas. They grow as small deciduous trees or shrubs of various shapes, or herbs characterised by pinnately compound leaves borne oppositely along stems, and flat to roundish clusters of small white-yellowish or sometimes pinkish-purple flowers and, during maturity, by small brown-black, blue, red, orange or yellow berries (Fernald, 1950, Bolli, 1994). The taxonomy of the genus Sambucus is highly sophisticated due to enormous phenotypical diversity and geographical distribution. The genus includes from less than 10 to more than 30 species, depending on the taxonomical approach. Von Schwerin (1920) recognised 28 species and several varieties. R. Bolli (1994), in his PhD dissertation, reduced the number of species to only nine: Sambucus ebulus L., S. wightiana Wall. ex Wight et Arnott, S. adnata DC., S. *Correspondence to: E-mail: metka.sisko@um.si Published in 2020 i Determination of Sambucus Interspecific Hybrid Structure using Molecular Markers gaudichaudiana DC., S. australasica (Lindley) Fritsch, S. javanica Blume, S. nigra L., S. australis Cham. et Schlecht. and S. racemosa L. Several, previously independent species became subspecies or botanical varieties. Five taxa, formerly considered as distinct species (i.e., S. canadensis, S. cerulea, S. peruviana, S. maderensis and S. palmensis, are now being considered as subspecies within S. nigra. His reduction was most probably too drastic and probably needs several corrections (Applequist, 2015). Proper taxonomy is especially important for breeders conducting interspecific crosses. Following the traditional taxonomy, the most important species are S. cerulea Raf. (blue elder), S. javanica (Javanese or Chinese elder), and S. nigra (common or black elder/ elderberry). All elderberries are characterised by extremely small flowers, forming relatively dense inflorescences. Due to small sizes of flowers (when they are ready for emasculation, they measure 1.8-2.2 mm in diameter), artificial hybridisation of elderberries is very difficult. The emasculation usually takes place late in the afternoon, one day before anthesis, or in early hours in the following morning, when the flowers are still closed. The breeder has to remove all (5) anthers, together with corolla, without damaging the pistil. The pollination usually takes place immediately after all selected flowers of an inflorescence have been emasculated and other flowers carefully removed. For pollination, one can use fresh or properly stored pollen. To promote faster germination of pollen, low concentrated sugary-water solutions can be used. The pollinated inflorescences have to be protected from uncontrolled pollination by semi-transparent bags, which are removed after approximately a week (Ivancic, personal experience). Elderberry breeding is relatively new and most of the present cultivars are probably direct selections from locally grown populations, or hybrids between locally grown genotypes. To our knowledge, most of the breeders have been using intraspecific hybridisation in order to create genetic variation. Interspecific hybridization, so far, has been rarely used. The earliest systematic and documented attempts to produce hybrid plants were reported by Bocher (1941), Winge (1944), Chia (1975), Koncalova et al. (1983) and Nilsson (1987). The main problems were difficulties in hybridisation technique, low number of successful fertilisations and sterility among progenies. Higher level of fertility was achieved when using closely related species, e.g., S. nigra x S. canadensis, compared to the cross between S. nigra and S. ebulus. The large scale elderberry breeding programme based on interspecific hybridisation, at the Faculty of Agriculture and Life Sciences, University of Maribor, Slovenia, began in 2003 by the lead author of this article. As elderberries, in Slovenia, are not considered to be important fruit bearing plants, the programme has never been officially funded. The first experimental crosses based on normal hand emasculation technique took place several years earlier. Those crosses were based on classical manual emasculation and resulted in a limited number of successful interspecific hybrids. In 2001, a self-incompatible genotype of S. javanica was introduced from the Island of Espiritu Santo, Vanuatu, and two years later, included in the hybridisation programme. Due to its almost complete self-incompatibility, emasculation of the female parent was not necessary. Our hybridisation programme involved 5 crucial species (S. cerulea, S. javanica, S. ebulus, S. nigra, S. racemosa) and two botanical varieties (S. nigra L. var. laciniata L., S. nigra L. var. viridis Weston). S. racemosa was represented by various taxons which had been previously recognised as species, but in the revised taxonomy of Bolli (1994) they are all included to S. racemosa (i.e., S. koreana, S. miquelii, S. sibirica, S. tigranii). The two above mentioned botanical varieties are probably the oldest botanical varieties of elderberries documented in the literature. The scientific name of S. nigra subsp. nigra var. laciniata (parsley-leaved or cut-leaved variety) appeared for the first time in Species Plantarum (Linaeus, in 1753, vol.1, p. 270). Few years later, it was also mentioned by Richard Weston (1775), p. 39, who was also the author of the scientific name of the S. nigra subsp. nigra var. viridis. For this variety, he also used the common name 'green-berried elder tree'. Parsley-leaved genotypes were also described in other species (or subspecies), e.g., Sambucus canadensis L. var. laciniata A. Gray. They were also observed among our interspecific hybrids (Ivancic, personal observations). The main objective of this programme was to recombine the most important positive morphological and physiological characteristics of S. cerulea, S. javanica, S. ebulus, S. nigra and S. racemosa. The final aim was to create dwarf or semi-dwarf, herbaceous or semi-herbaceous plants which would enable plantations with a higher density of plants, simple maintenance and mechanical harvest (Simonovik, 2007, Simonovik et al., 2007). S. ebulus was considered as a source of genes for dwarf herbaceous growth, while S. nigra and S. cerulea were representing the main genetic resources for productivity and desired chemical composition of fruits and inflorescences. Because of dwarf herbaceous growth harvesting would be easy and pruning would not be needed. At the beginning of each growth season, plants would be cut to the ground (mulched) and new stems would regrow from underground shoots. S. javanica from Espiritu Santo, due to its tropic (or sub-tropic) origin, was mainly used as a source of self-incompatibility. Later, additional genotypes of this species were introduced from Africa, but they were not self-incompatible. New reports (e.g., Wenga et al., 2019), however, indicate that stem ethanol extracts of S. formosana Nakai (which is one of the synonyms of S. javanica) could be a valuable source of substances fighting human coronavirus NL63 (HCoV-NL63), one of the main circulating HCoVs worldwide, which causes difficulties in respiratory tract such as runny nose, cough, bronchiolitis and pneumonia. The breeding programme at the Faculty of Agriculture and Life Sciences followed the scheme of modified phenotypic recurrent selection adapted to specific characteristics of different elderberry species. The programme was divided in cycles and each cycle consisted of three phases: (1) genetic recombination of selected genotypes, (2) morphological and chemical evaluation of the offspring individuals and (3) selection of superior individuals for new genetic recombinations (crosses) in order to obtain a new generation of hybrids, belonging to a new cycle. The number of hybrid combinations per cycle varied between 100 and 120, while the number of seeds per cycle generally exceeded 4,000. The average germination rate was approximately 45% and 2 Determination of Sambucus Interspecific Hybrid Structure using Molecular Markers individual cycles were completed within 4 to 6 years (Ivancic, unpublished). Chemical analyses of the obtained interspecific hybrids and their parental species revealed tremendous variations of chemical composition and many of the hybrids were superior when compared to their parental species (Mikulic-Petkovsek et al., 2014, Mikulic-Petkovsek et al., 2015a, Mikulic-Petkovsek et al., 2015b, Mikulic-Petkovsek et al., 2016, Todorovic et al., 2017, Imensek et al., 2020). The first molecular evaluation of hybrids (i.e., hybrid origin) took place in 2006 and were published by Simonovik et al. (2007) and was based on the sequenced cpDNA region, genome size differences and PCR-RFLP analysis of the nrDNA ITS region. At the beginning (in the first cycle), molecular evaluation of the interspecific structure was not necessary. The hybrids had several intermediate characteristics and used to be more winter hardy when compared to the tropical female parent (S. javanica). It was not difficult to see the differences in growth types, leaf shapes and inflorescences. The examples were the hybrids S. javanica x S. ebulus, S. javanica x S. cerulea, S. javanica x S. racemosa and S. javanica x S. nigra. The difficulties, however, arose later when complex interspecific hybrids were formed. The hypothetical example can be: ((S. javanica x S. cerulea) x (S. javanica x S. racemosa)) x ((S. cerulea x S. nigra) x (S. javanica x S. nigra)). This complex hybrid involves 4 species and it is practically impossible to determine its hybrid structure using morphological traits. Following the comments published by Applequist (2015), we consider S. cerulea as an independent species. The main objective of this paper was to check the possibility of using SSRs for determining the complex hybrid structure in various complex elderberry interspecific hybrids. We assumed that molecular markers could probably be very helpful for finding out which species were incorporated in an unknown, complex interspecific hybrid. The exact hybrid structure (i.e., in which way individual species were incorporated in the hybrid structure) most probably could not be defined. MATERIALS AND METHODS Plant material (elderberry genotypes) The list of the species, varieties and interspecific hybrids is presented in Table 1. Samples of all analysed hybrids were collected in May 2016 from plants growing in the Plant Gene Bank of the Faculty of Agriculture and Life Sciences of the University of Maribor. Samples of the majority of parental species (i.e., S. cerulea, S. javanica, two genotypes of S. ebulus, two genotypes of S. nigra and S. racemosa -including its taxons named as koreana, miquelii, sibirica, tigranii), however, were collected in SE Slovenia where the initial hybridisation had taken place. Almost all hybrid plants and some of the parental genotypes, originated directly from seed. The exceptions were two botanical varieties (S. nigra var. laciniata, S. nigra var. viridis) and ornamental cultivar 'Black Beauty' (BB)), all belonging to S. nigra. There was no clonal multiplication of hybrids; the only exception was JAxCER No3 which had been clonally propagated for experimental purposes. This hybrid was found to have several very useful and attractive traits such as medium height, high yield, large infructescences, stable wine-red colour of juice, better taste when compared to black elderberry, pleasant odour of inflorescences (different from other genotypes) and plants were less affected by aphids. The analysed hybrids originated from the first three cycles of crossings (first three cycles of the modified phenotypic recurrent selection). In each cycle, the selection of the parental material for the following series of crosses was based on plant vigour, growth characteristics, and fruit and floral characteristics (fruit size and colour, taste and chemical composition). Since the parental material used in interspecific crosses was assumed to be highly heterozygous, each offspring individual originating from the same cross most probably represented a different genotype. This means that two or more hybrids with the same hybrid structure (e.g., JAx(JAxEB), Table 1) should be considered as genetically different. Molecular analysis DNA isolation DNA was extracted from fresh, young leaves using the CTAB protocol (Doyle and Doyle, 1987). To approximately 2-3 square centimetres of fresh leaf tissue, one ml of preheated (68 °C) CTAB extraction buffer [2 % (w/v) CTAB, cetyltriammonium bromide (Sigma), 1.4 NaCl, 20 mM EDTA, 100mM Tris-HCl (pH 8.0), 0.2 % 2-mercaptoethanol] was added and well homogenized with a mortar and pestle and transferred to a 2 ml tube. Samples were incubated for 1.5 h at 68°C in a water bath. After incubation, 600 (l of chloroform: isoamyl alcohol in a 24:1 proportion were added, and the samples were thoroughly mixed. The mixtures were centrifuged at 11.000 rev./min for 10 min. After centrifugation, the supernatant was transferred to a fresh 1.5 ml tube and the DNA was precipitated by the addition of 0.1 vol. of 3 M sodium acetate and 1 vol. of ice cold isopropanol and kept at -20 °C for 30 min. Samples were again centrifuged at 11.000 rev./min for 10 min. The pellet was washed in 70 % ethanol for 20 min, air dried and rehydrated in 100 (l of TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 8.0). The DNA concentration was estimated by DNA fluorometer (Hoefer, TKO 100). Two replicate extractions per sample were performed. Microsatellites Six microsatellite loci (Table 2) developed earlier by Clarke and Tobutt (2006) were used: EMSn002, EMSn003, EMSn010, EMSn019, EMSn023 and EMSn025. Ten (l of PCR mixture contained 2 ng DNA (0.5 (l), 5 (l Qiagen Master Mix Kit, 0.5 (l of each primer (forward and reverse), 2 (l H2O and 1.5 (l Q solution. PCR condition consisted of an initial denaturation at 95 °C for 5 min, followed by 30 cycles of 94 °C for 30 s, Ta (different annealing temperatures for different loci) for 60 s, and 72 °C for 60 s, and a final step 10 min at 72 °C. The annealing temperatures were different for the loci EMSn019, EMSn002, EMSn003 (Ta = 60°C), EMSn010 (Ta = 59°C) and EMSn023, EMSn025 (Ta = 58 °C). The polymerase chain 3 Determination of Sambucus Interspecific Hybrid Structure using Molecular Markers 155 (JAxNI)x(CER+MIQ) JANI/(CER+MIQ) 156 (JAxNI)x(RAC+TIG) 2 JANI/(RAC+TIG) 2 157 JAxMIQ 2 JAMIQ 2 158 (JAxNI)xMIQ JANI/MIQ 162 JAxMIQ S JAMIQ S 165 S. javanica 2 JA 2 166 Hybrid 5 Hyb 5 167 Hybrid S Hyb S 168 S. nigra var. laciniata LAC 170 Hybrid 8 Hyb 8 171 Hybrid 4 Hyb 4 172 NI(Bg) NI(Bg) 17S Hybrid 2 Hyb 2 174 Hybrid 6 Hyb 6 175 Hybrid 7 Hyb 7 176 Hybrid 1 Hyb 1 177 Hybrid 9 Hyb 9 179 S. ebulus 2 EB 2 180 S. javanica S JA S 182 JAxEB 7 JAEB 7 18S JAx(JAxEB) 1 JA/JAEB 1 184 JAx(JAxEB) 2 JA/JAEB 2 186 JAx(JAxEB) 4 JA/JAEB 4 187 JAx(JAxEB) 5 JA/JAEB 5 194 JAxCER No. S JACER No. S 195 (JAxNI)xCER JANI/CER 197 S. sibirica SIB reaction (PCR) was performed using a Whatman Biometra T-Gradient thermocycler (Goettingen, Germany). Capillary electrophoresis of PCR products was performed on Beckman Coulter CEQ8000 according to manufacturer's instructions. Fragment size analysis was done with the in-build software. A fluorescent-labeled size marker (Beckman Coulter DNA Size Standard Kit 400 bp) was used as a molecular weight reference. Data analysis All unambiguous fragments were scored for the presence (1) or absence (0) of each band. Only clear and reproducible fragments were taken for data analysis. The binary data matrix was used to calculate Dice's similarity coefficients (Dice, 1945). Values of Dice's coefficients are between 0 (there is no common band) and 1 (two genotypes have identical markers, so they are identical). Dice similarity coefficients were calculated using the DARWIN computer package (Perrier and Jacquemond-Collet, 2006). For each microsatellite locus, the number of alleles per locus (n), allele frequencies, observed heterozygosity (HO), expected heterozygosity (HE) and polymorphic information content (PIC) were calculated using the Cervus 3.0.7 computer program (Marshall et al., 1998; 2014 version). The average distance between pairs of accessions was obtained by taking into account microsatellite data and a neighbor-joining tree was constructed using the DARWIN computer package (Perrier and Jacquemond-Collet, 2006). A matrix of Dice similarity coefficients was used for assessing relationships among 47 genotypes using the neighbour-joining algorithm developed by Saitou and Nei (1987). Table 1: Plant materia! included in molecular analysis Sample No. Plant Material Abbreviation m S. racemosa RAC 1S2 S. nigra 1 NI 1 m S. miquelii MIQ m S. javanica 1 JA 1 1S6 S. nigra 2 NI 2 1S7 S. cerulea CER 1S8 S. tigranii TIG 1S9 S. koreana KOR 140 S. nigra 3 NI 3 141 c.v. Black Beauty BB 142 JAxEB 9 JAEB 9 14S JAxEB 10 JAEB 10 144 (JAxNI)xVIR JANI/VIR 145 JAxVIR JAVIR 146 (JAxCER)x(MIQ+TIG)a JACER/(MIQ+TIG) 147 JAxMIQ 1 JAMIQ 1 148 JAxTIG JATIG 150 (JAxNI)x(RAC+TIG) 1 JANI/(RAC+TIG) 1 152 (JAxNI)xBB 3 JANI/BB 3 15S JAxCER JACER Note: a - (JAxCER)x(MIQ+TIG) means (S. javanica x S. cerulea) x (S. racemosa - miquelii + S. racemosa - tigranii); + is indicating pollination with a mixture of pollen of 'MIQ' and 'TIG'. The names S. nigra (NI), S. cerulea (CER) and S. racemosa (RAC) correspond to the names S. nigra subsp. nigra, S. nigra subsp. cerulea and S. racemosa subsp. racemosa, respectively, according to the revised classification of Bolli (1994). S. racemosa subsp. racemosa also includes the taxa named as "miquelii" (MIQ), "sibirica" (SIB) and "tigranii" (TIG). The taxon "koreana" (KOR) is included in S. racemosa subsp. kamtschatica. S. javanica (JA) was represented by 3 genotypes. BB = S. nigra 'Black Beauty'; LAC = S. nigra subsp. nigra var. laciniata; VIR = S. nigra subsp. nigra var. viridis; Bg = plants growing in the Maribor University Botanical Garden; Hyb = interspecific hybrids having partly known or unknown parental species. RESULTS AND DISCUSSION The number of alleles (Table 2) detected per locus ranged from 11 (EMSn025) to 18 (EMSn010, EMSn019), with an average of 15.17 alleles per locus. The observed heterozygosity ranged between 0.511 (locus EMSn023) and 0.787 (locus EMSn010), with an average of 0.656. The expected heterozygosity ranged between 0.791 (locus EMSn023) and 0.901 (locus EMSn003), with an average of 0.850. The differences between the observed and expected heterozygosity were observed on all loci. The largest difference was observed on locus EMSn003 (0.284) and the lowest on locus EMSn010 (0.084). The number of microsatellite markers sufficient for reliable variety identification depends on the nature and discriminating power of each marker (Tessier et al., 1999), normally six markers are sufficient for differentiating between genotypes (Zulini et al., 2002). 4 Determination of Sambucus Interspecific Hybrid Structure using Molecular Markers Table 2. SSR loci analyzed and parameters of genetic variability calculated for different microsatellite loci of the 47 Sambucus genotypes: number of alleles (n), effective number of alleles (n ), observed (H ) and expected (H ) heterozygosity, and polymorphic information content (PIC) Locus n n H H PIC EMSn002 15 8.87 0.702 0.838 0.809 EMSn003 16 9.22 0.617 0.901 0.882 EMSn010 18 7.21 0.787 0.871 0.850 EMSn019 18 9.00 0.745 0.898 0.880 EMSn023 13 4.60 0.511 0.791 0.758 EMSn025 11 4.86 0.574 0.803 0.774 Average 15.167 6.79 0.656 0.850 0.825 The reliability of microsatellite markers in genotyping of varieties was determined on the basis of the following criteria: the complexity of the banding pattern, the amplification of quality PCR products, the stability of the microsatellite repeated structure, and the polymorphic information content The molecular analysis indicates that the dendrogram (Fig. 1) encompasses four main groups of genotypes. The first main group involves two sub-groups. The first sub-group includes five taxons which, according to Bolli (1994), belong to S. racemosa (i.e. S. koreana (KOR), S. miquelii (MIQ), S. racemosa (in narrow sense - RAC), S. sibirica (SIB) and S. tigranii (TIG)), and three interspecific hybrids. The interspecific hybrid JANI/(CER+MIQ) (the accession 155), which is the result of pollination with a mixture of CER and MIQ pollen, is obviously JANI/MIQ. For the hybrid JACER/(MIQ+TIG)) (the accession 146), it is not possible to find out which of the of markers. On the basis of PIC values, all the microsatellite loci were classified as very informative loci (PIC > 0.5), and all loci proved suitable for mapping (PIC > 0.7). The allele sizes, frequencies and variability parameters calculated for each locus are shown in Table 3. parents (MIQ or TIG) participated in fertilisation. Following Bolli (1994), we can write JACER/RAC (meaning RAC in a broad sense). The second sub-group involves LAC (S. nigra subsp. nigra var. laciniata), two undetermined hybrids (most probably JANI/LAC or JALAC, and two hybrids involving VIR (S. nigra subsp. nigra var. viridis) - JAVIR and JANI/ VIR. In the second main group (Fig. 1), there are two subgroups. The first one involves NI 1 and its three hybrids, most probably with one of the S. javanica genotypes. Considering the background of these two hybrids (they originate from the first two cycles) of crosses, their parental structure is Table 3. Allele size (bp) and allele frequencies (in parenthesis) of the 47 Sambucus genotypes. at six microsatellite loci Locus EMS n002 EMS n003 EMS n010 EMS n019 EMS n023 EMS n025 1 70 (0.170) 135 (0.064) 189 (0.021) 142 (0.043) 182 (0.011) 127 (0.011) 2 100 (0.255) 138 (0.011) 193 (0.032) 187 (0.021) 186 (0.053) 142 (0.096) 3 110 (0.011) 191 (0.138) 245 (0.021) 189 (0.160) 218 (0.011) 161 (0.117) 4 121 (0.011) 193 (0.149) 248 (0.085) 191 (0.213) 249 (0.192) 165 (0.032) 5 253 (0.032) 195 (0.106) 250 (0.287) 193 (0.043) 253 (0.383) 179 (0.032) 6 255 (0.011) 197 (0.053) 253 (0.096) 197 (0.021) 255 (0.032) 183 (0.032) s i> 7 265 (0.021) 200 (0.181) 260 (0.170) 201 (0.043) 258 (0.011) 185 (0.043)