Phylogenetic Analysis of Wild and Garden Tulips Using Sequences of Chloroplast DNA R. Yanagisawa1, T. Kuhara2, T. Nishikawa3, D. Sochacki4, A. Marasek-Ciolakowska4 and K. Okazaki1 1 Graduate School of Science and Technology, Niigata University, Faculty of Agriculture, 2-8050 Ikarashi, Nishi-ku, Niigata 950-2181, Japan 2 Niigata Prefect. Botanical Garden, 186 Kanatsu, Akiba-ku, Niigata 956-0845, Japan 3 National Institute of Agrobiological Sciences, 2-1-2 Kannondai, Tsukuba, Ibaraki 305-8602, Japan 4 Research Institute of Horticulture, 96-100 Skierniewice, Konstytucji 3 Maja 1/3, Poland Keywords: cultivar origin, matK, molecular phylogeny, trnL, trnT-L, Tulipa Abstract Coding regions of trnL and matK and intergenic spacer (IGS) region of trnT-L in chloroplast were sequenced to analyze phylogenetic relationship of the genus Tulipa, including 27 cultivars, and 32 species tulips, by using Amana edulis as an outgroup. 12 variable sites out of 421 base pairs (bps) in the trnL, 19 variable sites out of 834 bps in the matK, and 62 variable sites out of 733 bps in the trnT-L are detected, indicating that the trnL has evolved more slowly and less informative for the classification. Five types of the sequences were identified in the trnL, 14 types in the matK and 20 types in the trnT-L IGS. The obtained results were summarized as follows. Clusianae subsection can be distinguished from Leiostemones and Eriostemones sections. Although T. praestans and T. tubergeniana are included in Eichleres section in the traditional classification, the two species are not sister relationships with Eichleres section in molecular phylogeny. Eichleres subsection except for T. praestans and T. tubergeniana is monophyletic and closely related to the species of Tulipa section. The garden tulips were divided the three groups. Group A included T. tubergeniana, T. schrenkii and ‘Duc van Tol’ tulips including ‘Keizerskroon’, indicating that T. schrenkii is contributed to the production of ‘Duc van Tol’ tulips. Group C included 10 cultivars such as ‘Faust’ and ‘Zulu’ which were recorded in the 19th century, indicating that cultivars included in Group C originated from a very old common progenitor of the ancestral cultivar and species. B group included 13 cultivars of T. gesneriana, T. acuminata, T. marjoretti and T. praecox, supporting a hypothesis that T. acuminata and T. marjoretti contributed to develop garden tulips. Our present data, while very informative, could not identify what species contributes to the large variability of late flowering tulips such as ‘Darwin’ and ‘Cottage’. INTRODUCTION Tulipa species are distributed from south-western Europe and North Africa to north-western China and the western Himalaya, where the species adapt to the cold and temperate conditions with dry summers (Wilford, 2006). A.G. Busbecquius travelled as envoy of the Holy Roman Empire to Turkey in 1554, and he saw the cultivation of tulips. Different providers including Busbecquius sent back seeds and bulbs to Europe around the 16-17th century, so that tulips became popular throughout Europe. Conrad Gesner saw tulips in Herwart’s garden in Augsburg, and depicted them in the book in 1561. Linnaeus (1753) designated a large number of tulips cultivars derived from complex crossing of wild species as Tulipa gesneriana, which is named after Gesner. The genus Tulipa includes approximately 100 species. So far many authorities attempted to classify tulip species based on various characteristics (Hall, 1940; Van Raamsdonk et al., 1997). The history of tulip classification was well reviewed in the literature written by Wilford (2006) and Zonneveld (2009). Eventually genus Tulipa is divided in two sections: Eriostemones and Tulipa (syn.: Leiostemones). Hall (1940) Proc. 24th Int. Eucarpia Symp. Section Ornamentals “Ornamental Breeding Worldwide” Ed.: T. Orlikowska Acta Hort. 953, ISHS 2012

103

divided Eriostemones into three subsections: Australes, Saxatiles and Biflores, and Leiostemones (syn.: Tulipa) into five: Clusianae, Gesnerianae, Oculus-solis, Eichleres and Kolpakowskianae. Hall (1940) classified the two species to section Orithyia, but now those two species can be excluded as belonging to the genus Amana. Van Raamsdonk et al. (1997) revised the two sections of genus Tulipa to the level of subgenus and divide each subgenus into sections. Subgenus Eriostemones was divided in the three sections: Australes, Saxatiles and Biflores, following Hall (1940). The subgenus Tulipa was divided into five sections: Clusianae, Tulipa, Tulipanum, Eichleres and Kolpakowskianae. Zonneveld (2009) modified the traditional classification based on nuclear genome sizes (DNA 2C value) of tulip species and established the new subgenus Clusianae because the species belonging to the subgenus Clusianae contain smaller nuclear DNA content than other tulips. Thus there is a general consensus concerning the primary division of genus Tulipa based on morphological characters, while infrageneric classifications of the genus Tulipa is still unclear. In addition, what species contributes to the large variability of garden tulips has been an issue of considerable dispute. Molecular phylogenetic analysis, especially using DNA sequences of chloroplast DNA (cpDNA), has succeeded in excluding such influences of morphological similarity from homoplasy (Soltis et al., 1997; Nishikawa et al., 2002). There are many different coding and intergenic regions in the chloroplast genome available for phylogenetic analysis (Taberlet et al., 1991). None has so far done phylogenetic analysis of the genus Tulipa. The purpose of the present study is two-fold. First, the study aims to evaluate the systematic status of infrageneric classifications of the genus Tulipa. Second, in order to know the origin and the domesticated process of cultivated tulips, the affinities and phylogenetic relationships between species tulips and cultivated tulips were analyzed using coding regions of trnL and matK and intergenic spacer (IGS) region of trnT-L in chloroplast for 27 cultivars, and 32 species tulips. MATERIAL AND METHODS Twenty-seven cultivars, and 32 species tulips of the genus Tulipa and one related species, Amana edulis, were used for the present study (Figs. 1-3). Young leaves were sampled from Niigata Prefectural Botanical Garden, Niigata Agricultural Research Institute Horticultural Research Center, Niigata University and Polish Research Institute of Horticulture. Leaves were stored at -80°C until DNA extraction. Total genomic DNA was extracted from silica-gel-dried leaves using the CTAB method of (Murray and Thompson, 1980). To obtain a coding region of trnL and a intergenic region of trnT-L, PCR was conducted using the two different universal primer sets: forward-c: (5’-CGAAATCGGTAGACGCTACG-3’) and reverse-d: (5’-GGGGATA GAGGGACTTGAAC-3’), and forward-a: (5’-CATTACAAATGCGATGCTCT-3’) and reverse-b: (5’-TCTACCGATTTCGCCATATC-3’) (Taberlet et al., 1991). To amplify a coding region of matK, two different primer sets were designed based on the matK reference sequences of tulip species (AB024386, AB040201, AF485324, AY624472, AY624473): the first set is forward-1F: (5’-TGCCTCTTGTCCTATGTCTC-3’) and reverse-1R: (5’-TTTCCAAATGGATAGGATGGGGT-3’), and the second set is forward-2F: (5’-TGGACTCCGTGGGCACATCAA-3’) and reverse-2R: (5’-ATGCA TC CTGTACGGTTGAGACC-3’). PCR was carried out in a thermal cycler (Takara Bio Inc., Otsu, Japan). Each reaction mixture was 25 µl, containing 10 ng of genetic DNA, 2.5 µl of 10× buffer [500 mM KCl, 100 mM Tris-HCl (pH 8.3), 15 mM MgCl2], 2 µl of dNTP mixture (a 2.5 mM concentration of each dNTP), 50 pmol of both primers, and 0.5 U of Takara Taq (Takara Bio Inc., Shiga, Japan). The thermal cycler program was run as follows: 35 cycles of 10 s at 98°C, 20 s at 57-61°C, depending on primer sequences, and 20 s at 72°C. After the last cycle, a final extension step (7 min, 72°C) was added. The amplified DNA was subjected to electrophoresis through 1% agarose gel. After electrophoresis, the target bands were excised from the gel and purified by MagExtractor-PCR and gel clean up kit (Toyobo Co., Ltd., Tokyo, Japan). The DNA was 104

re-suspended in 20 ml of TE (10 mmol/L Tris-HCl pH 8.0, 1 mmol/L EDTA). A cycle sequencing kit (Beckman Coulter) was used for preparation of samples for sequencing. The cycle sequencing involved 35 cycles of denaturation for 20 s at 96°C, annealing for 20 s at 50°C and extension for 2 min at 60°C. Reaction mixtures were subsequently stored at 4°C. The DNA sequencing with an autosequencer (CEQ8000, Beckman Coulter) was done according to the manufacturer’s instructions. The obtained sequences were visually aligned with ClustalW2 provided by DDBJ (http://www.ddbj.nig.ac.jp/). Phylogenetic analysis was performed using neighbor joining (NJ) method provided by Genetyx software (Genetyx Co., Tokyo, Japan). RESULTS AND DISCUSSION Partial sequences (421 bp) of trnL were determined for 14 species tulips, 14 garden tulips (T. gesneriana) and Amana edulis. A total of 11 variable nucleotide positions were detected and the sequences were analyzed by NJ method (Fig. 1). The results showed two major clades, one consisting of species belonging to subgenus Eriostemones and T. clusiana ‘Chrysantha’ of subgenus Tulipa and a second consisting of cultivars and species belonging to subgenus Tulipa such as T. gesneriana, T. schrenkii, T. praecox, etc. The first and second clades were significantly separated with bootstrap values of 71 and 85%, respectively. The separated location of subgenus Tulipa and subgenus Eriostemones in the molecular phylogenetic tree fits traditional classification, whereas although the position of T. clusiana ‘Chrysantha’ was controversial compared to the traditional classification, this is in agreement with a previous study by Zonneveld (2009), who reported that species of section Clusianae should separate from the two subgenera of Eriostemones and Tulipa. Similarly Blakey and Vosa (1982) reported that the species of section Clusianae has a distinct chromosome morphology and C-banding compared with the remaining species of genus Tulipa. Thirty two species tulips, 32 garden tulips (T. gesneriana) and Amana edulis were sequenced for the region of matK spanning from the initiation codon to the middle of the coding sequence (834 bp). A total of 19 variable nucleotide positions were detected. The phylogenetic relationship is shown in Figure 2. The phylogenetic analysis using matK showed two major clades, one (clade I) consisting of 6 species belonging to section Clusianae and a second (clade II) consisting of 48 species belonging to subgenus Tulipa. This result is consistent with that of trnL. It suggests the position of section Clusianae separated from the remaining of subgenus Tulipa. The clade II included 7 subclades, 21 isolated species and cultivars as follows. (i) including 8 cultivars of T. gesneriana such as Olaf; (ii) including 2 species of section Biflores (subgenus Eriostemones); (iii) including T. schrenkii (section Tulipa), T. tubergeniana (section Eichleres) and 10 cultivars (T. gesneriana); (iv) including 2 species of section Kolpakowskiana; (v) including T. orphanidea (section Australes) and T. praestans (section Eichleres), (vi) including 3 species of section Saxatiles; (vii) including 7 species of section Eichleres; (viii) 6 species (T. hageri, T. humilis, T. sylvestris, T. didieri, T. praecox and T. acuminata) and 14 cultivars of T. gesneriana were independent lineages. The notable finding of the clade II is as follows. 1) T. praestans and T. tubergeniana included in Eichleres section in the traditional classification are not sister relationships with species of section Eichleres in the matK tree. 2) The affinity of T. orphanidea (section Australes) and T. praestans (section Eichleres) was shown in the matK tree but this was not consistent with the result from the trnT-L tree (Fig. 3). The matK tree divided cultivars of T. gesneriana to three groups (A, B, C). The site-specific inversion (nt 595-600) at the middle of partial sequence of matK occurred in the group A consisting of ‘Olaf’, ‘Faust’, ‘Bonanza’, ‘Zulu’, ‘Queen of Night’, ‘Gander’, ‘Maureen’ and ‘Ile de France’. This inversion made the group A locate to a position distant from the rest of subgenus Tulipa. Intergenic sequences (749 bp) of trnT-L determined for 19 species tulips, 20 garden tulips (T. gesneriana) and Amana edulis. A total of 50 variable nucleotide positions were detected and the sequences were analyzed by NJ method (Fig. 3). As a result, 3 species belonging to section Clusianae were separated from the rest of genus 105

Tulipa. This result is consistent with the analyses using trnL and matK sequences. The major clade including most species of genus Tulipa was divided into two clades (I and II). The clade I corresponded to species belonging to subgenus Eriostemones and the clade II to subgenus Tulipa. The clade II was further divided 5 groups as follows. (i) including 2 species of section Kolpakowskiana and T. praestans (Section Eichleres); (ii) including T. eichleri, T. greigii and two cultivars of T. fosteriana, which corresponded to section Eichleres; (iii) including T. gesneriana cultivars, ‘Olaf’, ‘Faust’, ‘Zulu’, ‘Queen of Night’, ‘Gander’ and ‘Ile de France’, corresponding to group A in the matK tree; (iv) including T. schrenkii (section Tulipa), T. tubergeniana (section Eichleres) and 6 cultivars (T. gesneriana), which correspond to group B in the matK tree; (v) including T. marjoletii (section Tulipa), T. acuminata (section Tulipa) and 7 cultivars (T. gesneriana), corresponding to group C in the matK tree. ‘Oxford’, ‘Apeldoorn’ and ‘Keizerskroon’ included in the group C of the matK tree were allocated to the group B in the trnT-L tree. The subgenus Eriostemones includes 3 sections, Biflores, Australes and Saxatiles. T. tarda and T. biflora, belonging to section Biflores were monophyletic, which is in agreement with the traditional classification. On the contrary, the species of section Australes (T. sylvestris and T. humilis) and section Saxatiles (T. sylvestris and T. hageri) were closely related each other and the molecular phylogeny could not divide those species into the two taxa. 2 species of section Kolpakowskiana and T. praestans (Section Eichleres) were monophyletic but this relatedness should be confirmed because of low bootstrap value of this node (40). On the other hand, it was clear that at least T. praestans is distantly related to the species member of section Eichleres. This result is consistent with low crossability between T. praestans and the remaining species of section Eichleres (Van Raamsdonk et al., 1995). Similar to the analysis of the matK tree, cultivars of T. gesneriana were divided into three groups (A, B, C) in the trnT-L tree. The relatedness of T. schrenkii and other cultivars in Group B confirmed the generally-accepted belief that T. schrenkii is contributed to the production of the early-flowering tulips such as “Duc van Tol tulips”. T. tubergeniana was also included in Group B. There is a possibility that T. tubergeniana is one of progenitors in the production of garden tulip because T. tubergeniana was partially compatible to T. gesneriana so that some hybrids were obtained (Van Eijk et al., 1991). Group C included 10 cultivars such as ‘Faust’ and ‘Zulu’ which were recorded in the 19th century, indicating that cultivars included in Group C originated from a very old common cultivar or a wild progenitor. The group B included 13 cultivars of T. gesneriana, T. acuminata, T. marjoretti and T. praecox, supporting a hypothesis that T. acuminata and T. marjoretti contributed to develop garden tulips. The results of the molecular relatedness in this study have given us detailed information of relatedness of many species from genus Tulipa, while what species contributes to the large variability of late flowering tulips such as ‘Darwin’ and ‘Cottage’ could not be uncovered. More studies using nuclear coding genes as well as including more taxa in phylogenetic analysis are required to know origin of garden tulips. ACKNOWLEDGEMENTS The authors sincerely thank the Niigata Agricultural Research Institute Horticultural Research Center for kindly providing germplasm of tulips. Literature Cited Blakey, D.H. and Vosa, C.G. 1982. Heterochromatin and chromosome variation in cultivated species of Tulipa subg. Leiostemones (Liliaceae) Plant Syst. Evol. 139:163-178. Hall, A.D. 1940. The genus Tulipa. The Royal Horticultural Society, London. Murray, M.G. and Thompson, W.F. 1980. Rapid isolation of high molecular plant weight DNA. Nucleic Acids Res. 8:4321-4325. Nishikawa, T., Okazaki, K. and Nagamine, T. 2002. Phylogenetic relationships among Lilium auratum Lindley, L. auratum var. platyphyllum Baker and L. rubellum Baker 106

based on three spacer regions in chloroplast DNA. Breed. Sci. 52:207-213. Soltis, D.E., Gitzendanner, M.A., Strenge, D.D. and Soltis, P.S. 1997. Chloroplast DNA intraspecific phylogeography of plants from the Pacific Northwest of North America. Plant Syst. Evol. 206:353-373. Taberlet, P., Gielly, L. and Bouvet, J. 1991. Universal primers for amplification of three non-coding regions of chloroplast DNA. Plant Mol. Biol. 17:1105-1109. Van Eijk, J.P., Van Raamsdonk, L.W.D., Eikelboom, W. and Bino, R.J. 1991. Interspecific crosses between Tulipa gesneriana cultivars and wild Tulipa species: a survey. Sex. Plant Reprod. 4:1-5. Van Raamsdonk, L.W.D., Van Eijk, J.P. and Eikelboom, W. 1995. Crossability analysis in subgenus Tulipa of the genus Tulipa. Bot. J. Linn. Soc. 117(1):147-158. Van Raamsdonk, L.W.D., Eikelboom, W., De Vries, T. and Straathof, P.T. 1997. The systematics of the genus Tulipa L. Acta Hort. 430:821-828. Wilford, R. 2006. Tulips, species and hybrids for the gardener. Timber Press, Portland. Zonneveld, B.J.M. 2009. The systematic value of nuclear genome size for ‘‘all’’ species of Tulipa L. (Liliaceae). Plant Syst. Evol. 281:217-245.

107

Figures

Fig. 1. Phylogenetic tree constructed from trnK sequences of 14 species tulips, 14 garden tulips (T. gesneriana) and Amana edulis (outgroup) using the neighbor-joining method. Number above nodes indicate bootstrap estimate for 1000 replications.

108

Fig. 2. Phylogenetic tree constructed from matK sequences of 29 species tulips, 32 garden tulips (T. gesneriana) and Amana edulis (outgroup) using the neighbor - joining method. Number above nodes indicate bootstrap estimate for 1000 replications. The garden tulips were divided to three groups (A, B, C). Section Australes, section Saxatiles and section Biflores are included in subgenus Eriostemones. Section Clusianae and the remaining section are included in subgenus Tulipa.

109

Fig. 3. Phylogenetic tree constructed from trnT-L sequences of 19 species tulips, 20 garden tulips (T. gesneriana) and Amana edulis (outgroup) using the neighbor-joining method. Number above nodes indicate bootstrap estimate for 1000 replications. The garden tulips were divided to three groups (A, B, C). Section Australes, section Saxatiles and section Biflores are included in subgenus Eriostemones. Section Clusianae and the remaining section are included in subgenus Tulipa.

110