Molecular notes on the Mediterranean species of the genus Celtis L. (Celtidaceae) O. DE CASTRO 1 & G. MAUGERI 2 1 Dip. delle Scienze Biologiche, Sez. Biologia Vegetale, Universita ` degli Studi Federico II, Napoli, Italy, and 2 D.A.C.P.A., Sez. Biologia ed Ecologia Vegetale, Catania, Italy Abstract To understand the molecular variation existing between Mediterranean species belonging to the genus Celtis, the Internal Transcribed Spacer and the trnL (UAA) intron were sequenced from multiple accessions. The analyses revealed that the species complex appears to be a rather homogeneous group with low molecular variability, and with the presence of exclusive haplotypes. Key words: Celtis, Internal Transcribed Spacer, low molecular variability, Mediterranean flora, trnL (UAA) intron Introduction The genus Celtis L. (Celtidaceae) contains about 70 species in the temperate and tropical zones (Engler, 1964). The history of the genus in the Mediterranean area is documented by several palaeontological records (Manchester, 1989; Palmarev, 1989). At present, in this area, there are few species belonging to this genus: C. aetnensis (Tornabene) Strobl, C. aspera (Ledeb.) Steven, C. australis L., C. caucasica Willd., C. glabrata Steven ex Planchon and C. tournefortii Lam. (Tutin, 1993). C. aetnensis and C. aspera may be also referred to C. tournefortii (Tutin, 1993). The aim of this research was to examine the molecular variability in these Mediterranean species (with the exception of C. caucasica, which was not available for analysis). For this purpose, two different segments of DNA were analysed: the Internal Transcribed Spacers 1 and 2 (ITS1 and 2) of the nuclear ribosomal DNA, and the intron of the trnL (UAA) gene from chloroplast DNA. Material and methods The specimens employed, their origins, and se- quence accession numbers are reported in Table I. Total DNA was extracted following the procedure described by Doyle & Doyle (1990). ITS1 and 2, and the trnL (UAA) intron, were amplified by using the primers reported by Aceto et al. (1999) and by Taberlet et al. (1991), respectively. PCR reactions were carried out as in Aceto et al. (1999), and sequenced using a 310 Automated DNA Sequencer (Applied Biosystems). Alignments were carried out by using the BioEdit software, version 5.0.6 (Hall, 2001). An unweighted maximum parsimony (MP) was conducted using the heuristic search strategy in PAUP, version 4.0b10 (Swofford, 1998). A statis- tical parsimony network was performed among C. tournefortii ITS haplotypes with the TCS software, version 1.18 (Clement et al., 2000). Results and discussion Lengths and Genbank accessions for the trnL (UAA) intron, and for ITS1 and 2 are given in Table I. The trnL (UAA) intron sequences were excluded from the analysis because they were identical and, therefore, yielded no information. Alignment of ITS sequences was unequivocal. The differences in ITS sequence length are related to the variation in the number of repetitions of polyA in ITS1 and polyC in ITS2; therefore, the heuristic search in MP analysis was conducted by coding each gap as missing character. Correspondence: Olga De Castro, Dip. delle Scienze Biologiche, Sez. Biologia Vegetale, Via Foria 223, I-80139 Napoli, Italy. Tel.: þ39 0812538530. Fax: þ39 081450165. E-mail: odecastr@unina.it Plant Biosystems, Vol. 140, No. 2, July 2006, pp. 171 – 175 ISSN 1126-3504 print/ISSN 1724-5575 online ª 2006 Societa ` Botanica Italiana DOI: 10.1080/11263500600756421