V. DOURIS, S. GIOKAS 1, R. LECANIDOU, M. MYLONAS 2 and G.C. RODAKIS

Transcription

1 J. Moll. Stud. (1998), 64,81-92 The Malacological Society of London 1998 PHYLOGENETIC ANALYSIS OF MITOCHONDRIAL DNA AND MORPHOLOGICAL CHARACTERS SUGGEST A NEED FOR TAXONOMIC RE-EVALUATION WITHIN THE ALOPIINAE (GASTROPODA: CLAUSILIIDAE) V. DOURIS, S. GIOKAS 1, R. LECANIDOU, M. MYLONAS 2 and G.C. RODAKIS Department of Biology, Section of Biochemistry and Molecular Biology, University of Athens, Athens, Greece. 'Department of Biology, Section of Ecology and Taxonomy, University of Athens, Athens, Greece. 'Department of Biology, University of Crete, P.O. Box 147, Iraclion, Crete, Greece (Received 1 October 1996; accepted 25 April 1997) ABSTRACT A combined approach of mitochondrial large subunit ribosomal RNA (lrrna) sequence analysis and cladistic analysis of morphological characters was used in an effort to test previously proposed hypotheses on the taxonomy and phytogeny of certain taxa within the land snail family Clausiliidae. Ten populations belonging to three clausiliid subfamilies were examined, focusing mainly on the Alopiinae. Special consideration was given to the Graciliaria-type clausiliar apparatus (GCA), whose homoplasious character is confirmed: the two approaches give congruent results after its exclusion. The present analysis indicates that phylogenetic relations, at least within the Alopiinae, have to be reexamined. In view of the above, a new classification for certain taxa is proposed. INTRODUCTION The clausiliid subfamily Alopiinae includes the genera Albinaria, lsabellaria, Sericata, Agathylla, Cristataria and Medora. From these, the taxonomic status of Albinaria, lsabellaria and Medora, is uncertain (Nordsieck, 1977; Giusti, Grappelli, Manganelli, Fondi & Bullini, 1986). Medora is distributed in Italy and along the Dalmatian coastline, while Albinaria and lsabellaria in the southern Balkan peninsula. The latter two genera exhibit extreme interspecific and intraspecific morphological differentiation, although their distribution, whether allopatric or sympatric, is restricted to the relatively small area shown in Figure 1 (Nordsieck, 1974; Zilch, 1977a, 1977b). Clarification of evolutionary processes is hindered by the absence of unequivocal criteria governing the taxonomic importance of the different conchological and genitalia characters usually used in Correspondence lo: G.C. Rodakis conventional taxonomy. Thus, discrimination of different species within each genus,.but also of different genera, is often ambiguous (Nordsieck, 1984; Gittenberger, 1987; Gittenberger & Schilthuizen 1996). The suggested polyphyly of the genus lsabellaria (Nordsieck, 1984; Gittenberger, 1987), recently reinforced by sequence analysis of the ITSI rrna region of some lsabellaria and Albinaria taxa (Schilthuizen, Gittenberger & Gultyaev, 1995) indicates that taxa currently classified as lsabellaria should be reconsidered. Especially in the Peloponnese, an area where both genera exhibit mosaic distributions, identification of Albinaria and lsabellaria species or subspecies is difficult. As a result certain forms have an uncertain systematic classification (as Albinaria or lsabellaria) or are considered as hybrids (Nordsieck, 1984; Gittenberger, 1987,1994). A special case of interest concerns the taxonomic significance of the so called "Graciliaria type" or "G-type" clausiliar apparatus (GCA) (Fig. 2), a feature typical within lsabellaria but uncommon in Albinaria (Gittenberger, 1987, 1994). The GCA consists of a set of morphological characters not found in the majority of the clausiliid taxa. Its presence in several but not closely related clausiliid genera (Idyla, Graciliaria, Armenica, Muticaria, Papillifera, Neostyriaca, Leucostigma, Ruthenica etc.) has been attributed to independent origins (Nordsieck, 1982; Gittenberger, 1987, 1994; Gittenberger & Schilthuizen, 1996). It has also been proposed that the GCA is an effective adaptation (Nordsieck, 1982), since its construction restricts the contact of the animal's soft body with the external environment. In that way it could serve as a mechanism of efficient resistance to high temperature and low humidity. It is widely accepted that the resolution of

2 82 V. DOURIS classification problems as well as the determination of phylogenetic relations among certain taxa could be facilitated by the parallel use of molecular and morphological approaches, although congruence between the two approaches is not guaranteed (see reviews by Hillis, 1987; Patterson, Williams & Humphries, 1993). A first attempt at using such combined methods in land snails, based on the cladistic analysis of mitochondrial DNA (mtdna) restriction site polymorphisms and conchology (Douris, Rodakis, Giokas, Mylonas & Lecanidou, 1995), has provided congruent results in Albinaria species, indicating that it is reasonable to employ comparative molecular and morphological approaches in a broader range of taxa and taxonomic problems. In the present paper, we studied populations from different clausiliid subfamilies, focusing mainly on Alopiinae genera, as well as populations from the same genus, inspecting them for an extensive set of morphological characters. Individuals from the same populations were also analysed by amplification and sequencing of mtdna segments, a procedure that has proved very informative in revealing phylogenetic histories (Kocher, Thomas, Meyer, Edwards, Paabo, Villablanca & Wilson, 1989). Our aim was to evaluate the taxonomic importance of certain conchological characters (like those concerning the type of clausiliar apparatus) and also to contribute to the resolution of the systematic status of the Peloponnesian socalled Isabellaria group. The concordance between morphological and molecular approaches is investigated, a new classification of some Peloponnesian Isabellaria is proposed, and possible evolutionary hypotheses are discussed. MATERIALS AND METHODS Biological material Specimens from 1 populations belonging to 3 clausiliid subfamilies (Clausiliinae, Alopiinae, Mentissoideinae) were collected from restricted areas (no more than 2 m 2 ). The species and collecting sites (see also map in Fig. 1) are recorded in Table 1. Morphological data analysis About ten specimens from each population were examined for a series of qualitative conchological characters (see Table 2 and Fig. 2). Genitalia qualitative characters have proved invalid in distinguishing different genera of the subfamily Alopiinae (Nordsieck, 1977; Kemperman, 1992), and were therefore excluded from the morphological data set. It should be noted that exclusion of genitalia characters may lead to an underestimation of the distance between taxa belonging to different subfamilies of Clausiliidae. However, the main scope of this work was the Table 1. List of Clausiliidae populations and collection sites used in the analysis. The collection sites are numbered as in Figure 1. Subfamily Species Collection Site Alopiinae Mentissoideinae Clausiliinae Albinaria coerulea (Rossmassler 1835) Albinaria discolor (L. Pfeiffer 1846) Albinaria grisea (Deshayes 1833) Isabellaria haessleini (Fauer 1978) Isabellaria campylauchen (Boettger 1883) Isabellaria butoti (Nordsieck 1984) Isabellaria saxicola (Pfeiffer 1848) Sericata sericata {Boettger 1878) Idyla bicristata (Rossmassler 1839) Clausilia bidentata (Strom 1765) Amorgos island, Hora, Cyclades (1) Monemvassia, Lakonia, Peloponnese (2) Mount Ymittos, Kessariani, Attica (3) cross-road to Geraki, on the road from Monemvassia to Gythio, Lakonia, Peloponnese (4) Monemvassia, Lakonia, Peloponnese (2) 2 km from Geraki to Kosmas, Lakonia, Peloponnese (5) Mount Ymittos, Kessariani, Attica (3) Mount Dirfy, Steni, Evia (6) Mount Dirfy, Steni, Evia (6) Sheffield, England

3 PHYLOGENETIC RELATIONS WITHIN THE ALOPIINAE 83 Table 2. Conchological characters and their states. 1. protoconch colour: = beige, 1 = black; 2. dorsal keel: = obsolete, 1 = intermediate height, 2 = prominent; 3. basal keel: = obsolete, 1 = intermediate height, 2 = prominent; 4. shell development: = gradual, 1 = abrupt; 5. shell shape: = spindle shape, 1 = club shape; 6. shell colour: = dark brown, 1 = brown, 2 = beige, 3 = dirty white, 4 = white; 7. shell transparency: = transparent shell, 1 = not transparent shell; 8. shell spots; = absent, 1 = dots, 2 = striped; 9. cervix sculpture: = uniform, 1 = not uniform; 1. cervix ribbing: 1 = weak, 2 = prominent; 11. teleoconch ribbing: = imperceptible, 1 = prominent; 12. protoconch ribbing: = imperceptible, 1 = weak, 2 = prominent; 13. sutures; = shallow, 1 = deep; 14. peristome detachment: = present, 1 = absent; 15. peristome reflection: = present, 1 = absent; 16. peristome lips: = thin, 1 = medium, 2 = thick; 17. lip colour: = white, 1 = beige; 18. apertural shape: = roundish, 1 = pear shape; 19. clausilial disc end: = pointed, 1 = rounded; 2. clausilial disc shape: - elongated. 1 = not elongated; 21. frontal upper palatal fold (FUP): = absent, 1 = present; 22. distal upper palatal fold (DUP) kyrtosis: = absent, 1 = present; 23. lunular position: = dorsal, 1 = lateral; 24. lunula - DUP: = not distinct, 1 = distinct; 25. lunula: = obsolete, 1 = prominent; 26. principalis - DUP: = principalis supersedes DUP, 1 = principalis does not supersede DUP; 27. DUP - principalis contact: = DUP touches principalis, 1 = DUP does not touch principalis; 28. columellaris position: = basal, 1 = high; 29. columellaris inclination: = horizontal, 1 = inclined, 2 = vertical; 3. subcolumellaris: = reaches the apertural lip, 2 = visible from outside. 3 = visible from lateral view; 31. columellaris-subcolumellaris end: = subcolumellaris stops before columellaris, 1 = subcolumellaris does not stop before columellaris; 32. columellaris - subcolumellaris start: : collumellaris starts below subcolumellaris. 1 = subcollumellaris starts below columellaris; 33. columellaris shape: = not forked, 1 = forked; 34. parietalis start: = reaches the lip, 1 = does not reach the lip; 35. parietalis length: = long, 1 = short; 36. parietalis height: = intermediate, 1 = prominent; 37. parallelis; = absent, 1 = present; 38. spiralis: = absent, 1 = obsolete, 2 = prominent; 39. fulcrans; = absent, 1 = present; 4. subclaustralis; = absent, 1 = present; 41. basalis: = absent, 1 = present; 42. suturalis; = absent, 1 = present determination of the intra-subfamily relations. Con- collapsed to yield polytomies and topological consequently, only conchological characters that were straints were not enforced. A second run was enforced not uninformative or difficult to be determined with with the same options after excluding characters that accuracy were used. Data were analysed with the are related with the formation of the GCA; these PAUP package (version 3.1.1, Swofford, 1993). Multi- characters are: 27 (DUP- principals contact), 37 state taxa were interpreted as polymorphic. Multi- (parallelis), 38 (spiralis), 39 (fulcrans) and 42 (sutustate characters were treated as ordered, because of ralis). the obvious succession of their states, and they were weighted according to the number of their states, rtna with base weight 6, so that each character had equal weight regardless of the number of states (Swofford, Total DNA was extracted from the foot of each ani- 1993). More specifically, we enforced the branch- mal, which was incubated in.5 ml of lysis buffer (2 and-bound search option with the furthest addition mm Tris.Cl ph 7.5, 2 mm NaCI, 2 mm EDTA, sequence. Branches having maximum length zero 2% SDS..5 mg/ml Proteinase K) as 5 C until it was

4 V. DOURIS

5 PHYLOGENETIC RELATIONS WITHIN THE ALOPIINAE 85 Figure I.A. Distribution of genus Albinaria (black areas) in the East Mediterranean. B. Distribution of genus Isabellaria (black areas) in the southern Balkan peninsula. Question marks (?) indicate a very poorly known northern distribution. C. Map of Greece indicating the collecting sites (black circles) of the populations used in this study. Each site is numbered as in Table 1. The collecting site of Clausilia (in Britain) is not shown. B D Cs Cer Ap Pr SuFUR PI Figure 2. Aspects of a Clausiliid shell showing the conchological characters used in the cladistic analysis. A. frontal view; B. dorsal view; C. schematic frontal view of the aperture; D. schematic dorsal view of the aperture; E. schematic representation of the normal type clausiliar apparatus; F. schematic representation of Graciliaria type clausiliar apparatus. Abbreviations: Ap, aperture; Ba, basalis; Bk, basal keel; Cer, cervix; Cl, clausilium disc; Co, columellaris; Cs, clausilium stalk; Dk, dorsal keel; DUP, distal upper palatalis fold; FUP, frontal upper palatalis fold; Fu, fulcrans; Lu, lunela; Pa, parietalis; PI, parallelis; Pr, principalis; Pro, protoconch; Sc, subcolumellaris; SI, sulcalis; Sp, spiralis; Su, suturalis; Tel, teleoconch, (modified from Nordsieck 1982). completely dissolved (usually 3-6 minutes). The lysate was extracted several times by Phenol/chloroform and the nucleic acids were pelleted with ethanol precipitation. The pellet was dissolved in 5 JJLI of TE buffer (1 mm Tris.Cl ph 7.5,.1 mm EDTA) and RNA was removed by RNase treatment (5u.l of 1 mg/ml RNase A were added and the mixture incubated at 37 C for 3 minutes). After the addition of 4 u.1 of TE buffer and 5 u.1 of 4M NaCl, the solution was extracted again with Phenol/chloroform, and DNA was pelleted with ethanol precipitation and dissolved in 4 u.1 of TE buffer in order to be used as template in amplification reactions. DNA concentration was determined spectrophotometrically. Amplification Primers for the amplification reaction were designed after comparison of the published sequences for Albinaria turrita (Lecanidou, Douris & Rodakis, 1994) and A. coerulea (Hatzoglou, Rodakis & Lecanidou, 1995) mitochondrial large ribosomal RNA genes with the homologous "universal" primers (Palumbi, Martin, Romano, McMillan, Stice & Grabowski, 1991). The designed primer pair (16s-l: 5'GGTCTGAACTCA- GATCATGT-3', where W stands for A or T) was used to amplify a region corresponding to bases of the A coerulea mitochondrial genome (Hatzoglou et al., 1995). Amplification was carried out in 1 u.1 of a solution containing 2 mm Tris.Cl ph 8.3, 5 mm KC1,.1% gelatin, 2 mm MgCl 2, 2 nm of each deoxynucleotide, 5 pmole of each primer, 1 u,g template DNA and 2 units of Taq DNA polymerase (Boeringher Mannheim), in an MJ Research thermal cycler. Each cycle of the reaction consisted of denaturation for 1 min at 94 C, hybridisation for 2 min at 48 C, and extension for 3 min at 72 C. This cycle was repeated 35 times, and after

6 86 V.DOURIS the last cycle there as afinalextension step for 2 min at 72 C. 1985), 5 times for parsimony and neighbor-joining trees, and 1 times for Maximum-likelihood trees. Sequencing The amplification reactions were electrophoresed in 2% agarose gels, the gel fragments containing the amplification products were excised from the gels and DNA was purified by electroelution (Sambrook, Fritsch & Maniatis, 1989) and column separation through a NACS PREPAC minicolumn (BRL, cat. No 1525NP). The purified DNA was sequenced from both ends by the dideoxynucleotide method according to the Sequenase ver. 2. protocol (United States Biochemical) as modified for sequencing doublestranded amplification products (McPherson, Jones & Gurr, 1991). The primers used for the amplification reaction also served as sequencing primers. Sequence Data Analysis Sequences were aligned with the aid of the computer programme CLUSTALW (Thompson, Higgins & Gibson, 1994). Nucleotide distances were calculated using the DNADIST programme of PHYLIP package (Felsenstein, 1989) under the assumptions of the Kimura 2-parameter model (Kimura, 198). Cladistic analysis was performed with the PAUP package ver (Swofford, 1993) with branch-and-bound search. Characters were treated as unordered, branches having maximum length zero collapsed to yield polytomies, topological constraints were not enforced and trees were unrooted. Neighbor-joining (Saitou and Nei, 1987) and Maximum-likelihood (Felsenstein, 1981) trees were constructed using the NEIGHBOR and DNAML programmes of PHYLIP, respectively. All trees were submitted to bootstrapping (Felsenstein, RESULTS The conchological character-state matrix is shown in Table 3. Cladistic analysis of these data produced a single most parsimonious tree (length = 594, CI =.57, RI =.52), shown in Figure 3A. The following observations can be made on that tree: a) The genera Clausilia and Idyla are very well separated from the group consisting of the genera Albinaria, Sericata and Isabellaria, even though genitalia characters were not examined b) Two subgroups are obvious in the latter group: one containing the closely related taxa /. saxicola (Ymittos) and S. sericata (Steni) together with /. campylauchen (Monemvassia) and A. grisea (Ymittos), and a second subgroup consisting of the closely related taxa /. butoti (Kosmas) and /. haessleini (Geraki) together with A. discolor (Monemvassia) and A. coerulea (Amorgos). Thus, it can be noted that the generally accepted distinction between the genera Isabellaria and Albinaria is not supported, since A. grisea clusters with Isabellaria (I. campylauchen, 1. saxicola), while /. butoti and /. haessleini are grouped with Albinaria {A. discolor, A. coerulea). Surprisingly, Sericata sericata and Isabellaria saxicola comprise a monophyletic group. The nucleotide sequence alignment of the Table 3. Conchological character-state matrix. Character numbers are as in Table 2. Taxon Clausilia bidentata Idyla bichstata Sericata sericata Isabellaria butoti Isabellaria campylauchen Isabellaria haessleini Isabellaria saxicola Albinaria coerulea Albinaria discolor Albinaria grisea characters

7 PHYLOGENETIC RELATIONS WITHIN THE ALOPIINAE 87 Cluiisilia Hih-iuatu Idyla biirisiata - Albinaria griscu Isabdlaria t tinipytttiuhcn - Clausilia b Idyla hicristaiu - Striciiiu scricaia Isabdlaria saxia>ta r Claiisilia bidcnlala - ld\ la bicrisniltt ^^ Isubdlaria saxicola I Sericata scricaia Isabdlaria saxicola I... Sericata scricaia 45 i isabcilariu Isabdlaria hacssldni hacxacim ' Albinuriti discolor Isabdlaria campylauchctt Albinaria grisea, ilbinariu cocrulca Albinaria UK-mica Albinaria cocrtdea r Albinaria discolor Isabdlaria bmoli Isabdlaria bmoli Isabdlaria biitnti L_ Isabdlaria hacx.sli.irii B Isabdlaria cimtpy Imu hen Albinaria grisca Isabdlaria hacsslcini Albinaria discolor Figure 3.A. Most parsimonious phylogenetic tree resulting from the cladistic analysis of all conchological characters of Table 2. B. Neighbor-joining tree of the IrRN A sequences, resulting from the distance matrix of Table 4. Numbers indicate bootstrap values. C. Most parsimonious tree from the analysis of the conchological characters, excluding the ones related to the GCA (see Materials and Methods for details). large rrna segment analysed is shown in Figure 4. As can be seen, sequences do not have the same length; for example Idyla (and sometimes also Clausilia) has additional sequence portions, compared to the Albinaria and Isabellaria lrrna sequence. Perhaps these portions form specific secondary structure domains which are not present in the Alopiinae lrrna molecule. Since no valid secondary structure prediction is available for snail lrrna, we chose to use only confidently aligned sequence portions for further analysis. Thus, we excluded regions with large gaps or questionable alignments (boxed areas in Fig. 4). This resulted in a set of 321 positions on which tree construction was based. Nucleotide divergence (number of substitutions per site, according to Kimura's model) between each pair of sequences was calculated. The matrix of these distances (Table 4) was used for construction of the Neighbor-joining tree shown in Figure 3B. This tree, similarly to that based on morphological characters, shows a clear separation of Clausilia and Idyla from Isabellaria and Albinaria (bootstrap value 1). Moreover, /. saxicola (Ymittos) and 5. sericata (Steni) are very well separated (bootstrap value 1) from all other Isabellaria and Albinaria. It is worth noting that the three Isabellaria from the Peloponnese (/. campylauchen (Monemvassia), /. butoti (Kosmas) and /. haessleini (Gerald) are grouped together with Albinaria. More specifically, A. grisea (Ymittos) and /. campylauchen are very close and clearly form a monophyletic group which is very well supported by bootstrap values, while /. haessleini and A. discolor also cluster together. In contrast to the morphology-based tree, /. saxicola and 5. sericata appear paraphyletic, but close to each other. However, bootstrap analysis does not favour any particular hypothesis about the branching order, paraphyly or monophyly for these taxa. The same observations can be made if the data set is analysed by maximum likelihood or parsimony methods. The maximum likelihood tree topology is almost identical to the topology of the Neighbor-joining tree, with slightly different branch lengths and bootstrap values. Parsimony analysis has resulted in two equally parsimonious trees (length = 21, CI =.717, RI =.667). One of these trees has exactly the same topology with the Neighbor-joining tree, while the other differs mainly in the branching order of /. saxicola and 5. sericata. Bootstrap analysis also indicates this instability. Comparison of the morphological and mtdna trees (Figs 3A and 3B) reveals some important similarities, a) Clausilia and Idyla in both trees are distant from each other and from the rest of the examined taxa, revealing the extent of evolutionary divergence of the Alopiinae from the other clausiliid subfamilies b) Isabellaria saxicola and Sericata sericata are

8 Figure 4. Nucleotide sequences and sequence alignment of the amplified segment of the lrrna of the ten Clausiliidae analysed. Boxed areas indicate regions with large gaps or ambiguous alignment that were excluded from the analysis. (389) (39) (389) (393) (392) (39) (397) (391) (43) (436) a o 73 A. A. A. I coerulea TTATCTGCCCAGTGA \A. AAirTTAACGGCCGC-AGTACCTTGACTGTGCAAAGGTAGCATAATAATTTGACTTTTAAATGGAGCCTAGAATGAAAGAAAGAACGTAT discolor TTATCTGCCCAGTGA TA ATT IT, AAACGGCCGC-AGTACCTTGACTGTGCAAAGGTAGCATAATCATTTGACTTTTAAATGGAGCCTAGAATGAAAGAAAGAACGTAG grisea TTGTCTGCCCAGTGA U3 TATrTTAACGGCCGC-AGTACCTTGACTGTGCAAAGGTAGCATAATCATTTGACTTTTAATTGGAGCCTAGAATGAAAGAAATAACGTAG campylauchenttgtctgcccggtga AATrTTAACGGCCGC-AGTACCTTGACTGTGCAAAGGTAGCATAATCATTTGACTTTTAAATGGAGCCTAGAATGAAGGAAGTAACGTAG I. haessleini I. butoti TTATCTGCCCAGTGA TA TTATCTGCCCAGTGA!A ACTrTTAACGGCCGC-AGTACCTTGACTGTGCAAAGGTAGCATAATCATTTGACTTTTAATTGGAGCCTAGAATGAAAGAAAGCACGTAG AA- rttaacggccgc-agtaccttgactgtgcaaaggtagcataatcatttgacttttaaatggagcctagtatgaatgaaagaacgtag Z. isabellina TTATCTGCCCAGTGA A 1 'AACGGCCGC-AGTACCCTGACTGTGCAAAGGTAGCATAATCATTTGACTTTTAATTGGAGTCTGGAATGAAAGAAAGCACGTAG S. sericata ATATCTGCCCAGTGA AATfTTTAACGGCCGC-AGTACCCTGACTGTGCAAAGGTAGCATAATCATTTGGCTTTTAATTGGAGTCTGGAATCAAAGAAGTTACGTAG c. bideatata CGCTCTGCCCAGTCA ITTGAAT"I FTTA 'AACGGCCGCTAGTACACTGACTGTGCTAAGGTAGCATAATAAATTGGCTTTTAATTGGAGTCTGGAATAAAAGAGTTCATGGGG I. bicristata ATTTCTGCCCGGTGA MAT-A' TT TTTA 'AACGGCCGC-AGTACACTGACTGTGCAAAGGTAGCATAATAAATTGGCTTTTAATTGAAGTCGGGAATGAAAGAAAATTTGGGA A. coerulea A. discolor GCAACTTGTCT :AT'-TGATATTACTTTA VATTTA ATGGTTGAGTGAAAATACTCATAATTT-TAATAATAGACGAGAAGACCCTTAGAATTTTTAAAA SCAATAAG ACAACTTGTCT ^T'-TAATAATACCTTA W^ATTA CCTCATGAGTGAAAATACTCTTGAATT-TAATATTAGACGAGAAGACCCTTAGAATTTTAATTA UVTATAAG-- A. grisea ACAATTTGTCT TAT.'-TCATATTATCTTAKATTTG CCGTATGAGTGAAAATGCTCATGGTTT-TAAAAATAGACGAGAAGACCCTTAGAATTTTAAAAAakTATTAGTA I. canpylauchenacaaattgtct ^TATGCCATTATCTTAKAATTA.--CCGTATGAGTGAAAATGCTCATGGGTT-TAAAAATAGACGAGAAGACCCTTAGAATTTTAAAAA:ATATTAGCA I. haessleini ACAACTTGTCT EAT'-TAATAATACTTTTKAATTG CTGCATGAGTGAAAATACTCTTGAATT-AAATAATAGACGAGAAGACCCTTAGAATTTTAAAAAa^AATAAG I. butoti ATAACTTGTCTtTAT-TAATATAAATTTA^AATTG CCAAATGAGTGAAAATGCTCATGCTTT-GAATAATAGACGAGAAGACCCTTAGAATTTTAAAAArATAATAG-- CAT I. isabellina ATTTCTTGTCT :GCGAATTAATATTTATkAATTT'--TTAAGGAAGTGAAAATACTTCTAATTA-AAATAAAAGACGAGAAGACCCTTAGAATTTTGATTA^CAGACTT- S. sexicata ATTAATTGTCT TACAGTAAAATAATTTAKAATTAi CTAAAGAAGTGAAAATGCTTCTGAATATAAATAATAGACGAGAAGACCCTTAGAATTTTTAAGA-TTATATAT- c. bidentata GGTACATGTCT ^VT I. bicristata TTTTACTGTClfaTAATTATTTTTATATA\AATTAATCTTAGAAAGTGAAAATACTTTCAGAAA-ATATAATAGACGAGAAGACCCTATAAATTTTAAATAUVGGTAACTT A. A. A. I. I. coerulea discolor grisea canpylauchen haessleini I. butoti I. isabellina S. ssricata c. bidentata I. bicristata TAATTCTTATTAGAT rttttgttggggcaacaatatttca \-A TAKTAAATATA TTAAT- T-GAA-AGTAATAAGTCGATTA- - \ATAAT -TCAATCTTGTAAATT rttttgttggggcaacagtatttca r-agc-ga MAAATATT TCCTT-GG CAGATTAA-TCGATTTT- CATAKTAAATATAkTATT-TA AAGAATACGTCGATTT --TATAAT --\ATAAT TTAC-TCTAGTACGAT rttttgttggggcaacaatatttcarj -ATACATCTAGTAAGATtTTTTGTTGGGGCAACAATATTTCA T-AACAAA ^TAAATATT ITATA A--TATTATATGTCGATTT-- TATAAT CCAATCTTATAAATT jtttttgttggggcaacagtatttca:-. - A A f k T A A A T A T T r A C A T - T A A C C G G A T T A A A T C G A T T T T - \ATAAT TCTGTCTGTTAAGAT rttttgttggggcaacaatatttca P -TATA-AT ATGAATATATCGATTT --VATAAT -ATGTTAGTTTGATAAlTTTTTGTTGGGGCAACAATATTTCAtr-A-TATAKTAAATATTKACCT-TAA-TAATATACGACGATTT-- --\ATAAT TAATAGATATAATAT)rTTTTGTTGGGGCAACATTATTTCA(r- - ^TAAATAAT UkAA TAT-TATTATGTAACGATTA \ATAAT TTTTAGTCCTTCG CTTTTGTTGGGGCAACAAGTTGACA 3 TTGA iaaaaaatt raactcaataaattttaaggaattcttcgagtttt \CTAAG TTTTAATAGTGGTATTACTAT rttttgttggggcaacaacttacca TTAGCTAA VTGAATAAT MAATGTAGGTAAAGTATTATAGTTACATATACTTTTTAAT \ATAAT A. coerulea A. discolor T- ^TAGAAAAATTACCTAAGGGATAACAGCATAATTT- TATTAATAAGCTTATGACCTCGATGTTGGACTAGGTACTATTAAGGCTAATCGTTTTAA ktagagaaattacctaagggataacagcataattt- TAATAATAAGCTTGTGACCTCGATGTTGGACTAGGTACTATTAGGGCTAATCGTTTTAA A. grisea T- MAATAAAATTACCTAAGGGATAACAGCATAATTT- TAGTAGTAAGTTTGTGACCTCGATGTTGGACTAGGTACTATTAAGGCTAATCGTTTTAA I. caopylauchent- VTAATTAAATTACCTAAGGGATAACAGCATAATTT- TACTAGTAAGTTTGTGACCTCGATGTTGGACTAGGTACTATTAAGGCTAATCGTTTTAA I. haessleini I. butoti I. isabellina S. sericata C. bidentata T- T- A- A G - \TAGAGAAATTACCTAAGGGATAACAGCATAATTT- TAATAATAAGCTTGTGACCTCGATGTTGGACTAGGTACTATTAAGGCTAATCGTTTTAA ^CAGTAAAATTACCTAAGGGATAACAGCATAATCT TAATAATATGTTTGTGACCTCGATGTTGGACTAGGTACTATTAAGGCTAATCGTTTTAA - - kaagtaaaattacctaagggataacagcataatttat TAATAAGTTTGTGACCTCGATGTTGGACTAGGTACTATTAAAGCTAACCGTTTTAA - VTAAATAAATTACTATAGGGATAACAGCATAATTTTTCATT-TAAGATTGTGACCTCGATGTTGGATTAGGGACTTATTAAATAAACCATTTAAA I. bicristata cbaagtattactattawtaaaaaaattactatagggataacagcataattt- T-TATTAAAAAGATTGTGACCTCGATGTTGGACTAGGGAATATTTAGGGCAG-CATTCTAA

9 PHYLOGENETIC RELATIONS WITHIN THE ALOPIINAE 89 closely related c) /. campylauchen and A. grisea are also closely related. In both trees, the traditional distinction between Albinaria and Isabellaria is not supported. The basic difference between the two trees lies in the grouping of the A. grisea - I. campylauchen pair and of the /. haessleini -1. butoti pair. Interestingly, if the morphological characters that are related to the type of clausiliar apparatus are excluded from the analysis, the result is a unique and more robust tree (length = 496, CI =.53, RI =.571) (Fig. 3C), which is very similar to the mtdna tree of Figure 3B as far as the major distinction between Isabellaria and Albinaria is concerned: all the Alopiinae form a monophyletic group with the Peloponnesian Isabellaria forming a monophyletic group with the Albinaria taxa. Moreover, Isabellaria saxicola remains closely related to Sericata sericata, while A. discolor and /. haessleini are now grouped together. DISCUSSION One of the major problems in estimating phylogenetic relations and testing evolutionary hypotheses is the discrimination between homologies and homoplasies. This problem is especially, obvious in the case of the so-called "G-type" clausiliar apparatus (GCA), since it is not clear whether it is a result of common origin or parallelism and, furthermore, whether it is an adaptation or not (Nordsieck, 1982; Gittenberger, 1987, 1994, Gittenberger & Schilthuizen, 1996). The GCA has been considered in the past (Nordsieck, 1979) as a character set of high taxonomic value, and has been widely used for discrimination between Albinaria and Isabellaria (Nordsieck, 1982; Gittenberger, 1987). On the other hand, the "independent origins" argument for the appearance of the GCA has already been established (Nordsieck, 1982; Gittenberger, 1987; Gittenberger & Schilthuizen, 1996). Phylogenetic analysis of the mtdna data among the taxa used in this study, does not support the "homology" hypothesis for the GCA: taxa with or without it (for example /. campylauchen - A. grisea, and /. haessleini - A. discolor) can be very closely related, while taxa sharing the GCA feature (such as Idyla and Isabellaria can be quite distant (Fig. 3A). Moreover, if the characters which are related to the GCA are excluded from the morphological data set (Fig. 3C), the results of the morphological and mtdna analysis almost coincide. Thus, our results indicate that the appearance of the GCA can be attributed to parallelism and that developmental constraints are relatively relaxed in the case of its formation. As far as the proposed "adaptive" function of the GCA is concerned, certain observations indicate that a "non-adaptive" hypothesis is more valid. For the clausiliid taxa distributed in the east Mediterranean region, summer is the adverse period. Several mechanisms are employed to increase viability during this period: the snails aestivate (aestivation is "the warm weather equivalent of hibernation" - Gould, 1985) forming a thick epiphragm, usually in clusters on the rock surfaces or in crevices or under stones (Gittenberger, 1991; Kemperman, 1992; Schilthuizen, 1994; Heller & Dolev, 1994, Arad, Goldenberg & Heller, 1995; Giokas, 1996). Aestivation is triggered by photoperiodism, while awakening by a combination of photoperiodism and autumn rainfalls (Heller & Dolev, 1994; Arad etal, 1995; Giokas, 1996). Albinaria taxa, which have a more southern distribution than those of Isabellaria (Fig. 1), lack the GCA and are generally exposed during aestivation, while certain Isabellaria taxa that have the GCA usually aestivate under stones. This may indicate that the GCA is not actually so important for viability of these land snails as far as resistance to water loss is concerned. An alternative suggestion is that the GCA could offer a selective advantage as a mechanism of resistance to unknown predators (Gittenberger & Schilthuizen, 1996). In any case, its appearance seems to follow a stochastic mode and thus its exclusion is justified, if a more accurate representation of the phylogenetic relations is desired. An interesting observation on the phylogenetic analysis shown in Figure 3 concerns the non-monophyletic origin of the genera Albinaria and Isabellaria. Indications for polyphyly of the genus Isabellaria have been reported previously; these are based on morphological and ecogeographical data (Nordsieck, 1984; Gittenberger, 1987), as well as on some molecular data (sequence comparison of the ITS1 nuclear rrna region of some Isabellaria and Albinaria taxa; Schilthuizen et al, 1995). A point that needs to be answered, however, is what this "polyphyly" actually represents. We propose that the observed "polyphyly" is an "artifact" arising from acceptance of the conventional systematic classification for these taxa. More specifically, the species /. butoti, I. campylauchen and /. haessleini should be considered

10 9 V. DOURIS Table 4. Matrix of pairwise Kimura distances (X1). A. coerulea A. discolor A. grisea 1. campylauchen 1. haessleini 1. butoti 1. saxicola S. sericata C. bidentata 1. bicristata as belonging to the genus Albinaria rather than Isabellaria. Our hypothesis is supported by the following observations: (1) The molecular and morphological approaches are congruent (Fig.3). It should be noted that, at least as far as the close relation of /. haessleini and A. discolor is concerned, our molecular approach, based on mitochondrial DNA, is in agreement with results obtained from nuclear DNA sequences (Schilthuizen et a/., 1995). (2) The so-called Isabellaria species (/. butoti, I. campylauchen, I. haessleini) are much closer to the Albinaria taxa than to Isabellaria saxicola, which is close to Sericata sericata. Moreover, the distances (nucleotide substitutions per site; x 1 as in Table 4) between /. campylauchen and A. grisea (3.6) and between /. haessleini and A. discolor (3.61), are much lower than the distances typically observed between Albinaria species (7.84 to 9.24 in Table 4) and falls in the range of distances observed between "subspecies" or allopatric populations of the same species (Douris, Rodakis & Lecanidou, in prep.). (3) Natural hybridisation phenomena observed between sympatric populations of /. butoti and A. grisea (Nordsieck, 1984) and between A. discolor and /. campylauchen at Monemvassia (Giokas, 1996) indicate a close relation between certain Isabellaria and Albinaria taxa. Based on the above, we propose certain alterations in the systematic classification of the analysed Peloponnesian Isabellaria taxa. Namely, Isabellaria butoti should be considered as Albinaria butoti; Isabellaria camplylauchen as Albinaria grisea campylauchen; and Isabellaria haessleini as Albinaria discolor haessleini. It must be noted that the proposed classification for /. campylauchen justifies Boettger (1883), who first described this species as A. grisea campylauchen, while for /. haessleini (A. discolor haessleini) one of Gittenberger's classifications (Gittenberger, 1994) is supported. In the present analysis, two taxa belonging to other clausiliid subfamilies (Clausilia bidentata: Clausiliinae and Idyla bicristata: Mentissoideinae) were used as outgroups. Using both types of data, a very clear separation is obtained both between Clausilia and Idyla, as well as between these two and the monophyletic group of the Alopiinae taxa (Fig. 3). The latter group includes the "Albinaria and Peloponnesian Isabellaria" cluster mentioned above, plus Sericata sericata and Isabellaria saxicola. Our analysis indicates that, unlike the Peloponnesian Isabellaria taxa, /. saxicola is very clearly separated from the Albinaria taxa (Fig. 3). In the morphological analysis, /. saxicola forms a monophyletic group with S. sericata (Fig. 3C), while in the mtdna results the two taxa appear closely related but paraphyletic, although this branching order is not well supported by bootstrap analysis (Fig. 3B). Furthermore, the nucleotide sequence divergence (X1) between /. saxicola and S. sericata (7.48) is in the same range with the distances typically observed between congeneric species (Table 4). Despite the fact that only one Sericata and one "true" Isabellaria are analysed here, our observations indicate that phylogenetic relations within the subfamily Alopiinae are more complicated than those reflected by conventional classification. Further comparative analysis of morphological and molecular data from all Alopiinae genera could provide a more informative representation of the intra-subfamily relations and probably lead to the determination of genus-specific morphological characters.

11 PHYLOGENETIC RELATIONS WITHIN THE ALOPIINAE 91 ACKNOWLEDGEMENTS We are grateful to Dr Michael Mindrinos for PCRprimer construction and to Evi Hatzoglou for collection of the population of A. coerulea from Amorgos. This work was supported by the University of Athens. REFERENCES ARAD, Z., GOLDENBERG, S. & HELLER, J Water balance and resistance to desiccation in rockdwelling snails. International Journal of' Biometeorology, 38: BOETTGER, O On new Clausiliae from the Levant, collected by Vice-Admiral T. Spratt R.N. Proceedings of the Zoological Society of London, 1883: DOURIS, V., RODAKIS, G.C., GIOKAS, S., MYLONAS, M. & LECANIDOU, R Mitochondrial DNA and morphological differentiation of Albinaria populations (Gastropoda: Clausiliidae). Journal of Molluscan Studies, 61: FELSENSTEIN, J Evolutionary trees from DNA sequences: a maximum likelihood approach. Journal of Molecular Evolution, 17: FELSENSTEIN, J Confidence limits on phylogenies: an approach using the bootstrap. Evolution, 39: FELSENSTEIN, J PHYLIP - Phylogeny Inference Package, version 3.2. Cladistics, 5: GIOKAS, S The differentiation of the genus Albinaria in the Hellenic region. PhD Thesis, University of Athens, Athens, Greece. GITTENBERGER, E Neue Taxa der sogenannten Gattung Isabellaria (Gastropoda- Pulmonata: Clausiliidae) vom Peloponnes. Basteria, 51: GITTENBERGER, E What about non-adaptive radiation? Biological Journal of the Linnean Society, 43: GITTENBERGER, E Five new Albinaria from the eastern Peloponnese, Greece; with notes on Isabellaria s. 1. (Gastropoda Pulmonata: Clausiliidae). Basteria, 58: GITTENBERGER, E., SCHILTHUIZEN, M. (1996) Parallelism in the origin of the G-type clausiliar apparatus (Gastropoda Pulmonata: Clausiliidae) In: Origin and evolutionary radiation of the mollusca (J. Taylor ed.), Oxford University Press, Oxford. GIUSTI, F., GRAPPELLI, C, MANGANELLI, G., FONDI, R., & BULLINI, L. (1986) An attempt of natural classification of the genus Medora in Italy and Yugoslavia, on the basis of conchological, anatomical and allozymic characters (Pulmonata: Clausiliidae). Lavori Societa Italiana Malacologica, 22: GOULD, S.J Theflamingo'ssmile - reflections in natural history. Norton, New York. HATZOGLOU, E., RODAKIS, G.C. & LECANIDOU, R Complete sequence and gene organization of the mitochondrial genome of the land snail Albinaria coerulea. Genetics, 14: HELLER, J. & DOLEV, A Biology and population dynamics of a crevice-dwelling land snail, Cristataria genezarethana (Clausiliidae). Journal of Molluscan Studies 6: HILLIS, D.M., Molecular versus morphological approaches to systematics. Annual Review of Ecology and Systematics, 18: KEMPERMAN, T.C.M Systematics and evolutionary history of the Albinaria species from the Ionian islands of Kephallinia and Ithaca (Gastropoda, Pulmonata, Clausiliidae) Backhuys, Leiden. KIMURA, M A simple method for estimating evolutionary rates of base substitutions through comparative studies of nucleotide sequences. Journal of Molecular Evolution, 16: KOCHER, T.D., THOMAS, W.K., MEYER, A., EDWARDS, S.V., PAABO, S., VILLABLANCA, F.X. & WILSON, A.C Dynamics of mitochondrial DNA evolution in animals: Amplification and sequencing with conserved primers. Proceedings of the National Academy of Sciences U.S.A., 86: LECANIDOU, R., DOURIS, V. & RODAKIS, G.C Novel features of metazoan mtdna as revealed from sequence analysis of three mitochondrial DNA segments of the land snail Albinaria turrita (Gastropoda, Clausiliidae). Journal of Molecular Evolution, 38: MCPHERSON, M.J., JONES, K.M. & GURR, S.J PCR with highly degenerate primers. In: PCR, a practical approach. (McPherson, M.J., Quirke, P. & Taylor, G.R. eds) IRL Press, Oxford. NORDSIECK, H Zur Anatomie und Systematik der Clausilien, XV. Neue Clausilien der Balkan - Halbinsel (mit taxonomischer Revision einiger Gruppen der Alopiinae und Baleinae). Archiv fur Molluskenkunde, 14: NORDSIECK, H Zur Anatomie und Systematik der Clausilien, XVII. Taxonomische Revision des Genus Albinaria Vest. Archiv fur Molluskenkunde, 17: NORDSIECK, H Zur Anatomie und Systematik der Clausilien, XXI. Das system der Clausilien, II; Die rezenten europaischen Clausilen. Archiv fur Molluskenkunde, 19: NORDSIECK, H Die Evolution des Verschlussapparats der Schliessmundschnecken (Gastropoda: Clausiliidae). Archiv fur Molluskenkunde, 112: NORDSIECK, H Neue Taxa rezenter europaischer Clausilien, mit Bemerkungen zur Bastardierung bei Clausilien (Gastropoda: Clausiliidae). Archiv fur Molluskenkunde, 114: PALUMBI, S., MARTIN, A., ROMANO, S., MCMILLAN, W.O., STICE, L., & GRABOWSKI, G The simple fool's guide to PCR, version 2. University of Hawaii, Honolulu. PATTERSON, C, WILLIAMS, D.M. and HUMPHRIES, C.J Congruence between molecular and morphological phylogenies. Annual Review of Ecology and Systematics, 24: SAITOU, N., & NEI, M The Neighbor-Joining method: A new method for reconstructing phylo-

12 92 V. DOURIS genetic trees. Molecular Biology and Evolution, 4: SAMBROOK, J., FRITSCH, E.F., & MANIATIS, T Molecular cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. SCHILTHUIZEN, M Differentiation and hybridisation in a polytypic snail. Pasmams, Den Haag. SCHILTHUIZEN, M., GITTENBERGER, E. & GULTYAEV, A.P Phylogenetic relationships inferred from the sequence and secondary structure of ITS 1 rrna in Albinaria and putative Isabellaria species (Gastropoda, Pulmonata, Clausiliidae). Molecular Phylogenetics and Evolution, 4: SWOFFORD, D.L PAUP Phylogenetic Analysis Using Parsimony, Version computer program distributed by the Illinois Natural History Survey, Champaign, Illinois. THOMPSON, J.D., HIGGINS, D.G. & GIBSON, T.J CLUSTALW: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position specific gap penalties and weight matrix choice. Nucleic Acids Research, 22: ZILCH, A. 1977a. Die Typen und Typoide des Natur - Museums Senckenberg, 57: Mollusca: Clausiliidae (4): Alopiinae (2): Alopiini (1). Archiv fur Molluskenkunde, 17: ZILCH, A. 1977b. Die Typen und Typoide des Natur - Museums Senckenberg, 58: Mollusca: Clausiliidae (5): Alopiinae (3): Alopiini (2). Archiv fur Molluskenkunde, 18: Note added in proof: DNA sequences have been deposited in the Gen- Bank (accession numbers: AF1275 to AF1283).