Springer-VerlagTokyo102650918-94401618-086030669031Journal of Plant ResearchJ Plant Res007610.1007/s10265-003-0076-8 J Plant Res (2003) 116:115–132 © The Botanical Society of Japan and Springer-Verlag Tokyo 2003 Digital Object Identifier (DOI) 10.1007/s10265-003-0076-8 ORIGINAL ARTICLE M. Pfosser • W. Wetschnig • S. Ungar • G. Prenner Phylogenetic relationships among genera of Massonieae (Hyacinthaceae) inferred from plastid DNA and seed morphology Received: June 26, 2002 / Accepted: December 5, 2002 / Published online: February 22, 2003 Abstract The tribe Massonieae Baker (Hyacinthaceae- Key words Hyacinthaceae · Massonieae · Molecular phy- Hyacinthoideae) presently consists of about 19 genera and logeny · Plastid DNA sequences · Scanning electron micros- 230 species distributed from Africa (south of the Sahara) to copy · Seed morphology Madagascar and India. Based on atpB and trnL-F DNA sequences the tribe is monophyletic only when the genus Pseudoprospero is excluded from Massonieae. In most trnL-F trees, this genus occupies a basal position within Introduction subfamily Hyacinthoideae and is sister to the rest of the subfamily. Molecular data suggest that the remaining gen- Previous analyses using atpB, rbcL and trnL-F data have era of Massonieae do not share common ancestry with the shown that the family Hyacinthaceae is monophyletic Eurasian/North-African tribe Hyacintheae Dumort. (Scilla, and is nested within Asparagales. Its closest relatives are Hyacinthus and allies), and thus a narrow concept of the the family Themidaceae and Aphyllanthes monspeliensis essentially Eurasian genus Scilla is supported. Members of (Chase et al. 2000; Fay et al. 2000; Pfosser and Speta 1999). well-supported clades in Massonieae usually show similari- Based on molecular, morphological, karyological and ties in seed characteristics as determined by scanning chemotaxonomical data the family can be split into the four electron microscopy. Phylogenetic position and seed mor- subfamilies Oziroeoideae, Urgineoideae, Ornithogaloideae phology indicate that Massonia angustifolia and M. zeyheri and Hyacinthoideae (Pfosser and Speta 1999, 2001). do not belong to the genus Massonia but fall into a clade Taxa from the fifth subfamily Chlorogaloideae previously together with Daubenya, Androsiphon and Amphisiphon. included in Hyacinthaceae (Speta 1998a, 1998b) have been The genus Whiteheadia appears paraphyletic in the 50% shown to have affinities to the families Anthericaceae, majority rule trnL-F tree and occupies a basal position next Funkiaceae and Agavaceae but not to Hyacinthaceae to Massonia. However, in the strict consensus tree neither (Pfosser and Speta 1999, 2001). Within Hyacinthoideae a monophyly nor polyphyly can be excluded for this genus. distribution of genera into the tribes Massonieae Baker and Seed appendages are documented for members of the Hyacintheae Dumort. has been proposed (Speta 1998a). genera Ledebouria and Lachenalia. Within the genera of The two tribes show a strong geographic pattern. According Massonieae there is a tendency towards bending of the seed to Speta, the Massonieae consist of genera with a distribu- axis. This phenomenon is most obvious within the genus tion range from Africa (south of the Sahara) to the Arabian Lachenalia. Delimitation of genera based on seed morphol- peninsula, Madagascar and India, whereas the tribe ogy largely agrees with the results of molecular studies. Hyacintheae is exclusively North Hemispheric and consists Correlation between number, size and color of seeds, of genera with a predominantly Eurasian/North-African geographical distribution and phylogenetic position of the distribution. No phylogenetic investigations have been genera are discussed. undertaken to infer relationships among genera of Masso- nieae so far. Recently, genetic diversity of South African species of Scilla sensu lato has been investigated by RAPD analysis (van Staden and Pan 2001). However, the absence M. Pfosser (*) · S. Ungar of any genus outside of Massonieae makes it impossible to Department of Higher Plant Systematics and Evolution, Institute of draw phylogenetic conclusions from their work. Botany, Rennweg 14, 1030, Vienna, Austria Tel. +43-1-427754150; Fax +43-1-42779541 The inventory of species within Massonieae is not yet e-mail: [email protected] complete and delimitation of genera is still under dispute; W. Wetschnig · G. Prenner for example, in 1997, Müller-Doblies and Müller-Doblies Institute of Botany, Karl-Franzens-University, Graz, Austria published a partial revision of Massonieae in which they 116 described the new genus Namophila and recognized 15 CTAB, 100 mM Tris, 1.4 M NaCl, 20 mM EDTA, 0.2% mer- genera (Müller-Doblies and Müller-Doblies 1997). Speta captoethanol, pH 8.0) for 30 min at 60°C; 500 ml chloro- (1998a) described the new genera Merwilla, Pseudopros- form/isoamylalcohol (24/1) was added and the extraction pero and Avonsera and also recognized fifteen, partly dif- mix incubated for 15 min at 4°C. After centrifugation, the ferent, genera. Pfosser and Speta (1999) questioned the DNA was precipitated with 500 ml isopropanol. The pellet independent generic status of Daubenya, Androsiphon and was washed with 70% ethanol and dissolved in 100 ml TE Amphisiphon. Goldblatt and Manning (2000) finally trans- buffer. ferred the monotypic genera Androsiphon and Amphisi- phon as well as Neobakeria namaquensis and Massonia angustifolia to Daubenya. Jessop (1970) and Stedje (1998) DNA sequencing grouped several members of Massonieae within a broadly defined genus, Scilla. However, such a classification inevita- Two non-coding regions and one coding region of the plas- bly results in a highly polymorphic genus (Pfosser and Speta tid genome were sequenced. The trnL(UAA) intron and the 1999, 2001, 2003). intergenic spacer (IGS) between the trnL(UAA)-3¢exon Morphology of seeds is rather diverse in Hyacinthaceae. and the trnF(GAA) gene were amplified together in a single Seed characteristics have been shown to be useful for sub- PCR reaction (Pfosser and Speta 1999). The atpB gene familial delimitation (Speta 1998a) and generic grouping was amplified using the primers S2 and 1493R (Hoot et al. (Jessop 1975). Recently, a survey of seed morphology 1995). Amplified double-stranded DNA fragments were among genera of Massonieae was presented (Wetschnig et sequenced directly on an ABI377 automated sequencer al. 2002). (Perkin Elmer, Beaconsfield, UK) following the DYEnam- The aim of the present study was to investigate the icET cycle sequencing protocol (Amersham Pharmacia, monophyly of the tribe Massonieae, the relationship among Piscataway, N.J.). Both strands were sequenced using the genera and to the sister tribe Hyacintheae. The study is nested sequencing primers described for the trnL-F region based on plastid DNA and seed morphology using material (Pfosser and Speta 1999). Sequencing primers for the atpB from Africa, Madagascar, India, Eurasia, North Africa and gene were the same as those used by Hoot et al. (1995) East Asia. One coding region of plastid DNA (atpB) and except for two internal sequencing primers which were two non-coding regions, the trnL(UAA) intron and the designed specifically for the sequencing of Asparagalean trnL(UAA)-trnF(GAA) intergenic spacer, were investi- taxa (primer 385x: 5¢-GCG CAG ATC TAT GAA TAG gated. The atpB gene has been shown to be useful for higher GTG ATG T-3¢, primer 766x: 5¢-TAA CAT CCC GGA level systematic studies in monocots (Chase et al. 2000; AAT ATT CCG CCA T-3¢). On average, less than 1% of Savolainen et al. 2000), whereas the non-coding regions data matrix cells were scored as missing data. have been used successfully in Hyacinthaceae and many other groups to reveal subfamilial relationships (Pfosser and Speta 1999, 2001, 2003; Stedje 1998, 2000). Seed mor- phology data based on light and electron microscopic exam- Phylogenetic analysis inations were used as additional characters for generic delimitation within Massonieae. Correlation between seed Sequence manipulations were performed on a Digital morphology, geographical distribution and phylogenetic Alpha 1000A 5/400 server under the operating system position of genera are discussed. Digital Unix V.4.0D. DNA sequences were pre-aligned using the PileUp program of the GCG software package (Genetics Computer Group 1994). Final alignment of DNA sequences was performed by visual inspection. The Materials and methods sequences have been trimmed on both ends to exclude ambiguous positions in close proximity to the sequencing Taxa sampled primers. All sequences have been deposited in the EMBL database (for accession numbers refer to Appendix). Phy- This analysis is based on material of Hyacinthoideae sam- logenetic analysis using the maximum parsimony method pled from Africa, Madagascar, India, Eurasia, North Africa was performed with the computer program PAUP* version and East Asia. Voucher information for all plant accessions, 4.0b10 (Swofford 2000). Most parsimonious trees were geographic origin, and EMBL database accession numbers obtained by 1,000 replicates of random sequence addition are provided in the Appendix. Nomenclature follows that using tree bisection-reconnection branch swapping under of Speta (1998a, 1998b). the Fitch criterion (Fitch 1971). Ten thousand fast bootstrap replicates (Felsenstein 1985) were used to assess confidence limits for the resulting tree topologies. Indels in the data matrix were coded as additional characters (indels longer DNA extraction than one nucleotide position