A mitochondrial phylogeography of Brachidontes variabilis (Bivalvia: Mytilidae) reveals three cryptic species

 2007 The Authors Journal compilation  2007 Blackwell Verlag, Berlin

Accepted on 25 January 2007 J Zool Syst Evol Res doi: 10.1111/j.1439-0469.2007.00421.x

1 Dipartimento di Biologia Animale, Universita` di Palermo, Palermo, Italy; 2Department of Ecology and Evolutionary Biology, University of Colorado, Boulder, CO, USA

A mitochondrial phylogeography of Brachidontes variabilis (Bivalvia: Mytilidae) reveals three cryptic species M. Sirna Terranova1, S. Lo Brutto1, M. Arculeo1 and J. B. Mitton2

Abstract This study examined genetic variation across the range of Brachidontes variabilis to produce a molecular phylogeography. Neighbour joining (NJ), minimum evolution (ME) and maximum parsimony (MP) trees based on partial mitochondrial DNA sequences of 16S-rDNA and cytochrome oxidase (COI) genes revealed three monophyletic clades: (1) Brachidontes pharaonis s.l. from the Mediterranean Sea and the Red Sea; (2) B. variabilis from the Indian Ocean; (3) B. variabilis from the western Pacific Ocean. Although the three clades have never been differentiated by malacologists employing conventional morphological keys, they should be ascribed to the taxonomic rank of species. The nucleotide divergences between Brachidontes lineages (between 10.3% and 23.2%) were substantially higher than the divergence between congeneric Mytilus species (2.3–6.7%) and corresponded to interspecific divergences found in other bivalvia, indicating that they should be considered three different species. Analysis of the 16S-rDNA sequences revealed heteroplasmy, indicating dual uniparental inheritance (DUI) of mtDNA in the species of Brachidontes collected in the Indian Ocean, but not in the species in the Pacific nor in the species in the Red Sea and the Mediterranean Sea. When we employed the conventional estimate of the rate of mitochondrial sequence divergence (2% per million years), the divergence times for the three monophyletic lineages were 6–11 Myr for the Indian Ocean and Pacific Ocean Brachidontes sp. and 6.5–9 Myr for the Red Sea and Indian Ocean Brachidontes sp. Thus, these species diverged from one another during the Miocene (23.8–5.3 Myr). We infer that a common ancestor of the three Brachidontes species probably had an Indo-Pacific distribution and that vicariance events, linked to Pleistocene glaciations first and then to the opening of the Red Sea, produced three monophyletic lineages. Key words: Brachidontes variabilis – B. pharaonis Mytilidae – phylogeography – 16S-rDNA – COI

Introduction The Indo-Pacific mussel, Brachidontes variabilis Krauss, 1848, by virtue of its high fecundity, early maturity, planktonic larvae, and its tolerance of high salinity and high temperature, occupies diverse habitats across a wide geographic range. It has been reported in high-salinity lagoons, on open coasts, and in polluted waters. This small mussel, belonging to the family Mytilidae, is an euryhaline, eurythermal, diet generalist. It anchors itself to hard substrates in the mid-littoral zone with byssal threads, forming Ômytilid bedsÕ, which may reach high densities and completely cover rocky shores when wave exposure and sedimentation conditions are optimal (Safriel et al. 1980). It is widely reported that the geographical range of B. variabilis includes the western Pacific Ocean, the Indian Ocean, the Red Sea and the Mediterranean Sea (Taylor 1971; Sasekumar 1974; Barash and Danin 1986; Morton 1988), but the occurrence of this species in some areas is controversial. According to Barash and Danin (1986), B. variabilis occurs along eastern African coasts, from the Red Sea to southern Africa, in the Indian Ocean except for the Persian subregion and Malaysia, and in the western Pacific Ocean. In contrast, Sasekumar (1974) reported B. variabilis in Malaysia, while Arcidiacono and Di Geronimo (1976) supported the occurrence of B. variabilis along the western African coasts. Identification of B. variabilis relies primarily on shell shape and shell dimensions, colour (externally dark brown-black, internally tinged violet-black), the presence of numerous fine radial, bifurcating ribs which become coarser posteriorly, a crenulate margin, dysodont teeth and the absence of a septum beneath the beaks. The equivalve and inequilateral shell is variable in shape and height/length ratio and has terminal umbones, sometimes greatly expanded posteriorly, sometimes arcuate, occassionally subcylindrical with beaks not quite

J Zool Syst Evol Res (2007) 45(4), 289–298

terminal (Fischer 1870). This high degree of morphological variation makes identification difficult, hinders systematic studies, and potentially hides cryptic species, compelling a search for systematic characters not influenced by environmental variations. Different interpretations of the high variability of the shell characters led systematists to use different names, all of them considered as synonyms with B. variabilis (Krauss, 1848), adding to the confusion; Mytilus exustus Linnaeus, 1827, Mytilus (or Hormomya) variabilis Krauss, 1848, Brachidontes semistriatus Krauss, 1848, Mytilus (or Brachidontes pharaonis) (Fischer, 1870), Mytilus senegalensis Lamarck, 1889, Mytilus arabicus Jousseaume ms. in Lamy 1919 are the ones found in literature. The ambiguity is the greatest for the Mediterranean Sea and the Red Sea regions, where many authors (Arcidiacono and Di Geronimo 1976; Chemello and Oliverio 1995; Gianguzza et al. 1997; Rilov et al. 2002) have used B. pharaonis (Fischer, 1870) as a synonym of B. variabilis (Krauss, 1848), but never as a different species. Physiological and ecological studies have been conducted on Mediterranean (Safriel et al. 1980; Sara` et al. 2000) and IndoPacific B. variabilis (Taylor 1971; Sasekumar 1974; Barash and Danin 1986; Morton 1988), whereas few genetic studies have been conducted, and genetic data have been too limited for systematic studies. A few populations of B. pharaonis from the Red Sea and the western Mediterranean Sea (at that time variabilis), were analysed by starch gel electrophoresis (Lavee 1981; Safriel and Ritte 1983); the authors concluded that 8% of the ÔMediterraneanÕ alleles, with moderate frequencies, were ÔuniqueÕ and not found in the Red Sea. Brachidontes variabilis from the Red Sea was also used in a revision of the Pteriomorphia classification, analysing 18S rDNA sequences of several Bivalves species (Steiner and Hammer 2000). Recently, the

290

Sirna Terranova, Lo Brutto, Arculeo and Mitton

mtDNA variation was used to describe the population structure of B. pharaonis and to make inferences about its invasion of the Mediterranean Sea; these studies did not reveal any geographical pattern or differentiation among local populations (Shefer et al. 2004; Sirna Terranova et al. 2006). Recently, an analysis of the nuclear and mitochondrial variation in Brachidontes exustus in Florida and the Caribbean revealed cryptic species (Lee and Foighil 2004), emphasizing the need to conduct more thorough genetic analyses of B. variabilis. To resolve this taxonomic confusion attributable to conflicting morphological descriptions and the lack of historical biogeography for B. variabilis, we performed a range-wide survey of mtDNA sequences to infer the historical biogeography of B. variabilis and to estimate the genetic distances and taxonomic relationships among the populations in different geographical areas. We collected sequences from two mitochondrial genes, 16S-rDNA and subunit I of cytochrome oxidase (COI). Variability at 16S-rDNA is generally useful as an interspecific taxonomic marker and it resolved taxonomic relationships in the family Mytilidae (Rawson and Hilbish 1995), and COI has been used in many phylogeographic studies (Avise 1994). The sequence data provide new markers to test conventional morphological classifications, and to measure phylogenetic relationships among the species (Davis and Nixon 1992; Stepien and Kocher 1997).

Methods Sample collection Thirty-four individuals of B. pharaonis (synonymous with B. variabilis) were collected from Stagnone di Marsala, Italy and Tel Aviv, Israel, both 20°

0°

in the Mediterranean Sea, and from Safaga Bay, Egypt in Red sea (Fig. 1). A total of 20 B. variabilis were also collected from Reunion Rocks, Madagascar and Kwa Zulu-Natal, Pennington, South Africa; both these sample sites are in the Indian Ocean. Seven individuals were sampled from Ting Kok, Hong Kong, in the western Pacific Ocean (Fig. 1). A sample collected in Malaysia, identified as B. variabilis by the local malacolagists, was identified by us as a species belonging to a different genus, Xenostrobus. One specimen of Mytilaster minimus and one of Mytilus galloprovincialis were collected in Italy; these were used as outgroups. Samples were kept in ethanol at )20C until tissue removal. In our sampling, we recorded the sex of each individual (when possible) and recorded the tissue used for the DNA extraction, because most bivalves have separate male and female mtDNA lineages, a bizarre condition referred to as doubly uniparental inheritance (DUI) (Zouros et al. 1994). Most species belonging to the families Unionidae and Mytilidae exhibit a maternally inherited mitotype expressed in all tissues of the female offspring, while males express a paternally inherited mitotype in gonad tissues, but the female mitotype is predominant in somatic tissues (Skibinski et al. 1994a; Hoeh et al. 1996; Liu et al. 1996; Quesada et al. 1996). The sex of each individual was determined by microscopic inspection (except for the individuals from Hong Kong, whose gonads were completely empty) by removing a part of the gonad and examining it under a microscope to detect either sperm or eggs. A piece of mantle tissue, full of gametes, suitable for mtDNA analysis, was stored in ethanol at 4C.

PCR and sequencing Total DNA was extracted from a very small (about 10 mg) piece of ethanol-preserved mantle tissue using the Dneasy Tissue Kit (QIAGEN, Qiagen, Valencia, CA, USA) and was stored at )20C. A fragment of the mitochondrial 16S ribosomal mtDNA gene (16S-rDNA) was amplified with the universal primers 16sar-L (5¢CGCCTGTTTAACAAAAACAT3¢) and 16sbr-H (5¢CCGGTCTGAACTCAATCACG3¢)

40°

45°

45°

40°

40°

1

35°

Mediterranean Sea

35°

2

Suez Canal

30°

30°

3 Red Sea

25° 0°

20°

km 0 200400

25° 40°

0°

30°

60°

90°

120° 40°

40°

20°

6

Pacific 20° Ocean

0°

0°

Indian Ocean

4 –20°

–20°

5 –40°

–40° 0°

30°

60°

90°

120°

Fig. 1. Sampling sites of Brachidontes sp. Mediterranean Sea: 1, Stagnone di Marsala, Southern Italy; 2, Tel Aviv, Israel. Red Sea: 3, Safaga Bay, Egypt. Indian Ocean: 4, Reunion Rocks, Madagascar; 5, Pennington, South Africa. Pacific Ocean: 6, Ting Koc, Hong Kong

 2007 The Authors J Zool Syst Evol Res (2007) 45(4), 289–298 Journal compilation  2007 Blackwell Verlag, Berlin

Evolutionary history of Brachidontes variabilis

291

(Palumbi 1996). A fragment of the mitochondrial cytocrome oxidase I gene (COI) was amplified with the universal primers: LCO1490 (5¢GGTCAACAAATCATAAAGATATTGG3¢) and HCO2198 (5¢TAAACTTCAGGGTGACCAAAAAATCA3¢) (Folmer et al. 1994). Double stranded DNA was amplified in 50 ll reaction volumes which contained two units of Taq polymerase, 1x reaction buffer [200 mM (NH4)SO4, 100 mM Tris HCl pH 8.8, 0.1% (v/v) Tween], 2 mM MgCl2, 40 mM dNTPs, 10 pmol of each primer and approximately 200 ng of DNA. The polymerase chain reaction (PCR) amplification conditions were as follows: after hot start (95C for 9 min), 30 cycles of denaturation at 93C for 1 min, annealing at 50C (COI) or 54C (16S-rDNA) for 1 min, and extension at 70C for 1 min. The amplification was completed with a final extension at 70C for 10 min. Polymerase chain reaction product for both genes was separated in 2% w/v horizontal agarose gels stained with ethidium bromide. A single band was cut from the standard agarose gel; the PCR fragment was extracted and purified using a QIAquick Gel Extraction Kit (QIAGEN, Hilden, Germany) and the DNA was sequenced in an abi prism 3700 (Applied Biosystems, Foster City, CA, USA) automated sequencer. Sequences were obtained for a 500 bp fragment of the 16SrDNA gene and a 618 bp fragment of the COI gene. In the following text, we present DNA sequences in the 5¢-3¢ orientation with respect to the light strand of mtDNA. Sequencing results were manually edited with the program chromas version 1.45 (Chromas 2006) (http://www.technelysium.com.au/ chromas14x.html). Sequences have been submitted to GenBank for COI (accession numbers: DQ836012, corresponds to specimen F3IT in Fig. 2; DQ836013, F1IT; DQ836019, F1MA; DQ836020, M3MA; DQ836021, 1HK; DQ836022, M. minimus; DQ836023, M. galloprovincialis) and for 16S-rDNA (AN: DQ833530, corresponds to specimen M2IT in Fig. 2; DQ833531, M3IS; DQ836014, F1MA; DQ836015, M3MA; DQ836016, 5HK; DQ836017, M. minimus; DQ836018, M. galloprovincialis).

Data analysis Phylogeographical and molecular evolutionary analyses were conducted with both loci using mega, v. 2.1 (Kumar et al. 2001) and DNAsp, v. 3.51 (Rozas and Rozas 1999). Genetic heterogeneity within populations was estimated as haplotype diversity (h), indicating the probability that two randomly chosen haplotypes are different in a population, and nucleotide diversity (p), indicating the mean percent differences between all pairs of haplotypes in a population. These two indices were calculated following Nei (1987) with DNAsp, v. 3.51. Mytilaster minimus and M. galloprovincialis were used as outgroups in the phylogeographic analysis. All the sequences were aligned using the clustal-w program (Thompson et al. 1994). Mutational saturation of 16S-rDNA and COI sequences was evaluated by plotting pairwise nucleotide divergence versus transitions (Ti) and transversions (Tv) of the Brachidontes sequences; p-distance was calculated as the proportion of nucleotide sites at which the two compared sequences were different, obtained by dividing the number of nucleotide differences by the total number of nucleotides compared, as implemented in mega software. Evolutionary distances between the samples of Brachidontes assemblage were calculated using the Tamura-Nei model (Tamura and Nei 1993) for 16S-rDNA and COI, with mega 2.1 software (Kumar et al. 2001). Brachidontes sp., M. minimus and M. galloprovincialis sequences were aligned against 16S-rDNA and COI sequences of other species of bivalves extracted from the GenBank database (16S: Mytilus edulis, accession number (AN), U22865; Geukensia demissa, AN, U68772; Saccostrea cuccullata, AN, AF458907; COI: M. edulis, AN, U68773; G. demissa, AN, U56844; S. cuccullata, AN, AY038076). For this portion of the phylogenetic analysis, a single consensus sequence represented each of the major geographic regions of B. variabilis: the Mediterranean–Red Sea, the Indian Ocean and the western Pacific Ocean. For this analysis, the COI and 16S-rDNA consensus sequences were joined into a single sequence. Nucleotide sequences were used to construct neighbour joining (NJ) (Saitou and Nei 1987), minimum evolution (ME) (Rzhetsky and Nei 1992) and maximum parsimony (MP) (Nei and Kumar 2000) trees

using mega 2.1 software (Kumar et al. 2001). Support for nodes in all trees was assessed with the bootstrap confidence levels using 1000 replicates (Felsenstein 1985).

Results 16S-rDNA sequences Partial mitochondrial 16S-rDNA gene sequences, 500 bp long, were obtained from 39 specimens of B. variabilis, including those from the Mediterranean and the Red Sea, which were also named B. pharaonis. A total of 22 haplotypes were detected in the 39 mussels, and 16 were unique haplotypes or haplotypes found in only one collection locality. Haplotypes were shared among localities within the Mediterranean Sea and the Red Sea, and among localities within the Indian Ocean or within Pacific Ocean. But none of the haplotypes was shared between any of the three large geographic areas (Mediterranean Sea and Red Sea, Indian Ocean, Pacific Ocean; Fig. 1). As a consequence of the differences among these three geographical areas, the full alignment revealed indels at four nucleotide sites. Within the alignment containing all of the sequences, 108 of the nucleotide sites were variable, and 103 were parsimoniously informative (Table 1). Among all the 139 mutations, 69 were transitions and 70 were transversions (0.99 : 1). Summing across geographic areas, 61 mutations were detected at 57 segregating sites. The mean base composition in the Brachidontes samples was 32.7% thymine, 15.6% cytosine, 28.9% adenine and 22.8% guanine. The pattern of base composition was homogenous among all Mytilidae taxa and revealed biases for thymine and against cytosine. The 16S-rDNA sequences were used to calculate genetic diversity within the Mediterranean Sea, the Indian Ocean and the Pacific Ocean for all samples of B. variabilis. Indian Ocean populations showed the highest variability in term of nucleotide diversity (p) with a value of 0.040, higher than the values of 0.005 and 0.002 for the Mediterranean–Red Sea and Pacific ocean areas, respectively. Values of haplotype diversity (h) varied from 0.733 to 0.955 (Table 1). The high nucleotide diversity found in the Indian Ocean samples was due to the presence of two mitochondrial forms, with 4% sequence divergence, represented in the two sister groups detected in the Indian Brachidontes cluster of Fig. 2. The two forms of mtDNA, M-type and F-type, were detected in males and females, respectively, revealing dual uniparental inheritance (DUI) in the Brachidontes clade in the Indian Ocean. Weighted and unweighted NJ, ME, and MP trees exhibited the same topology and high bootstrap values (Fig. 2). All trees discriminated the Brachidontes sp. of the three areas: Mediterranean and Red Sea–Indian Ocean–Pacific Ocean, and illustrated the heteroplasmy detected in the Indian Ocean. The matrix of pairwise Tamura-Nei genetic distances is presented in Table 2, together with the pairwise ratios of Ti/Tv. The minimum pairwise nucleotide divergence value among all specimens was observed between the Mediterranean and the Red Sea samples (0.005). Genetic distances between samples from the Mediterranean and the Red Sea and all other samples of B. variabilis ranged from 0.134 to 0.161. Saturation, assessed by plotting the numbers of transitions and transversions against p-distances was detected for transitions at 16S-rDNA, which formed a plateau between 25 and 30 (Fig. 3). However, transversions at 16S-rDNA increased

 2007 The Authors J Zool Syst Evol Res (2007) 45(4), 289–298 Journal compilation  2007 Blackwell Verlag, Berlin

292

 2007 The Authors J Zool Syst Evol Res (2007) 45(4), 289–298 Journal compilation  2007 Blackwell Verlag, Berlin

Sirna Terranova, Lo Brutto, Arculeo and Mitton

Evolutionary history of Brachidontes variabilis

293

Table 1. Genetic diversity of 16S-rRNA gene sequences detected in Brachidontes samples separated in the three major areas Area/clade

Mediterranean–Red Sea Brachidontes pharaonis

Locality

Italy

Israel

Egypt

Indian Brachidontes variabilis Together

Sequence length (bp) Sample size Number of haplotypes Haplotype diversity

499 499 499 499 6 6 2 14 3 4 2 5 0.733 0.800 1.000 0.725 (±0.155) (±0.172) (±0.500) (±0.104) Nucleotide diversity 0.005 0.006 0.008 0.005 (±0.0015) (±0.0012) (±0.0040) (±0.0006) Mutations/segregating sites 5/5 6/6 4/4 6/6 Ti/Tv 5/0 6/0 4/0 6/0 Pi/S 5/0 4/2 4/0 5/1

Madagascar South Africa 500 12 10 0.955 (±0.057) 0.039 (±0.0062) 50/46 39/11 42/4

500 6 4 0.800 (±0.172) 0.045 (±0.0087) 42/42 35/7 34/7

Pacific Brachidontes variabilis

Together

Hong Kong

500 18 13 0.902 (±0.066) 0.040 (±0.0039) 52/48 41/11 43/5

495 7 4 0.810 (±0.130) 0.002 (±0.0005) 3/3 3/0 1/2

Ti, transitions; Tv, transversions; Pi, parsimony sites; S, singleton sites.

Table 2. Tamura-Nei model (1993) nucleotide distances of 16S-rDNA sequences (below the diagonal) and pairwise numbers of transitions/ transversions (above the diagonal) between the Brachidontes samples of the major biogeographic areas Mediterranean and Red Sea

Italy Israel Egypt Madagascar South Africa Hong Kong

Indian Ocean

Pacific Ocean

Italy

Israel

Egypt

Madagascar

South Africa

Hong Kong

* 0.005 0.005 0.137 0.135 0.161

6/0 * 0.005 0.137 0.134 0.160

5/0 6/0 * 0.136 0.134 0.160

58/46 53/45 51/45 * 0.041 0.124

47/47 48/44 47/45 41/11 * 0.123

33/47 35/45 33/46 55/36 52/33 *

regularly with p-distance. P-distances between 0.0 and 0.08 were found between populations within the large geographic areas, and p-distances of 0.06 were detected between the M and F types in the Indian Ocean. COI sequences The COI sequences, each 618 bp, revealed 47 haplotypes in 59 Brachidontes sp (Table 3). As with the 16S-rDNA data, none of the haplotypes was common to all of the six samples, and none was found in more than one of the large geographic areas, although some were widely distributed within a geographical area. A total of 231 mutations was detected at 187 variable sites, of which 151 were transitions and 80 were transversions (Ti : Tv ratio 1.89 : 1) and 168 were parsimonyinformative (Table 3). These latter were differently distributed in the codon positions, for 24 were at the first position, two at the second, and 161 at the third.

Two substitutions at the second and one at the first codon position produced amino acid changes. The ones at the second codon position included: a substitution at the 500th nucleotide producing an amino acid substitution, Gly fi Val, and a nucleotide substitution at the 425th position which produced a Leu fi Try substitution at the 142nd amino acid in a mussel from Hong Kong sample. The amino acid change of the first codon position, occurring at the 88th bp position, divided Mediterranean– Red Sea individuals into two groups, one characterized by a leucine (L-form) and another by a methionine (M-form) at the 30th amino acid position. Both of these amino acids are apolar, so this does not appear to be a drastic substitution; we are unaware of the physiological consequences of this polymorphism. The coexistence of the two forms, equally distributed among localities and genders, has been discussed in a previous article (Sirna Terranova et al. 2006), in which we presented data suggesting that the L-form was the ancestral character and that the M-form was derived. The more extensive genetic and geographic data here reveal that the M and L-forms are also differentiated at16S-rDNA, and that all Indo-Pacific populations had leucine at the 30th amino acid (L-form), supporting the hypothesis that the M-form in Mediterranean and Red Sea populations is a derived character (Figs 2 and 4). Haplotype diversity was very high in all the three areas: 0.973 for the Mediterranean Sea and the Red Sea, 0.968 for the Indian Ocean and 1.000 for the Pacific Ocean. In contrast, nucleotide diversity decreased markedly from the Mediterranean Sea and the Red Sea (0.039) to the Indian (0.021) and the Pacific Ocean (0.006) (Table 3). The number of transitions and transversions within and between the three geographic areas are presented in Tables 3 and 4, and, in all cases, transitions greatly exceeded transversions,

Fig. 2. Neighbour joining (NJ) tree of Brachidontes assemblage obtained from evolutionary distances calculated using the Tamura-Nei model (Tamura and Nei 1993) for the 16S-rDNA gene (on the left) and COI (on the right). On the right: The Mediterranean–Red Sea clade is divided into two groups, one characterized by a leucine (L-form; specimens in italic) and the other by a Methionine (M-form; specimens in bold) at the 30th amino acid position. Individuals within the Indian Ocean clade are divided into two sister groups, named group I, specimens in grey boxes, and group II, specimens in black boxes. On the left: note the presence of two sister groups within Mediterranean–Red Sea clade, which discriminate the COI forms (L-form, specimens in italics, and M-form, specimens in bold). Also note the two mitochondrial forms (M-type and F-type) supporting the existence of sex-linked heteroplasmy detected in the Indian Ocean Brachidontes lineage; only within the female lineage group I, specimen in grey box, and group II, specimen in black box, of COI tree can be discriminated. Bootstrap values are reported in percentages in the principal branches nodes. Notes. Each ID-specimen is composed by the following code: F or M, when identified, for female or male + number of the individual + abbreviation of locality: IS, Israel; IT, Italy; EG, Egypt; MA, Madagascar; SA, South Africa; HK, Hong Kong

 2007 The Authors J Zool Syst Evol Res (2007) 45(4), 289–298 Journal compilation  2007 Blackwell Verlag, Berlin

294

Sirna Terranova, Lo Brutto, Arculeo and Mitton

(a) 40 35

N Ti. Tv

30 25 20 15 10 5 0 0.00

0.02

0.04

0.06

(b)

0.08 p-distance

0.10

0.12

0.14

0.16

90 80 70 N Ti. Tv

60 50 40

Fig. 3. (a) 16S-rDNA and (b) COI nucleotide divergence, from pairwise comparison between single individuals, plotted versus the number (N) of Transitions (Ti, triangles) and Transversions (Tv, circles)

30 20 10 0 0.00

0.02

0.04

0.06

0.08

0.10 0.12 p-distance

0.14

0.16

0.18

0.20

0.22

Table 3. Genetic diversity of the mtDNA COI sequences detected in Brachidontes samples in the three major areas Area/clade Locality Sequence length (bp) Sample size Number of haplotypes Haplotype diversity Nucleotide diversity Mutations/segregating sites Ti/Tv Pi/S

Mediterranean–Red Sea Brachidontes pharaonis

Indian Ocean Brachidontes variabilis

Pacific Ocean Brachidontes variabilis

Italy

Israel

Egypt

Together

Madagascar

South Africa

Together

Hong Kong

618 17 11 0.941 (±0.036) 0.038 (±0.004) 52/52 47/5 48/4

618 15 14 0.990 (±0.028) 0.042 (±0.004) 57/56 51/6 48/8

618 2 2 1.000 (±0.500) 0.073 (±0.036) 45/45 41/4 0/45

618 34 25 0.973 (±0.015) 0.039 (±0.002) 66/65 60/7 51/14

618 17 14 0.956 (±0.044) 0.021 (±0.002) 44/44 38/6 23/21

618 3 3 1.000 (±0.272) 0.024 (±0.010) 22/22 19/3 0/22

618 20 17 0.968 (±0.033) 0.021 (±0.001) 48/48 41/7 23/25

618 5 5 1.000 (±0.126) 0.006 (±0.001) 9/9 8/1 1/8

Ti, transitions; Tv, transversions; Pi, parsimony sites; S, singleton sites.

L M Mediterranean-Red Sea Brachidontes pharaonis Indian Ocean Brachidontes variabilis L Pacific Ocean Brachidontes variabilis L Mytilaster minimus L Geukensia demissa P Mytilus galloprovincialis P Mytilus edulis 100 Outgroup 87

99 96 91 L

Fig. 4. Phylogenetic relationships derived from an NJ analysis of 16S-rDNA and COI sequences concatenated under Tamura-Nei model. On the tree, the amino acid carried at the 30th position for each lineage is shown. It can be noted that Mediterranean–Red Sea clade presents methionine (M), a derived character in respect to the leucine (L), which is present in Mytilaster minimus and Geukensia demissa. Another derived character is Phenylalanine (P) carried in the same amino acid position by the other two species of Mytilus. Bootstrap values are reported in percentage in the principal branches nodes. Saccostrea cucculata, belonging to the family Ostreidae, was used as outgroup in the construction of trees

as expected. Plots of the number of transitions and transversions (Fig. 3) versus genetic distance revealed, in contrast to the plots for 16S-rDNA, that neither transitions nor transi-

tions reached saturation. Genetic distances between 0.06 and 0.08 include the comparisons of the L and M forms in the Mediterranean Sea and the Red Sea.

 2007 The Authors J Zool Syst Evol Res (2007) 45(4), 289–298 Journal compilation  2007 Blackwell Verlag, Berlin

Evolutionary history of Brachidontes variabilis

295

Table 4. Tamura-Nei model (1993) nucleotide distances of COI sequences (below the diagonal) and pairwise numbers of transitions/ transversions (above the diagonal) between the Brachidontes samples from the major biogeographic areas Mediterranean and Red Sea

Indian Ocean

Pacific Ocean

Sample

Italy

Israel

Egypt

Madagascar

South Africa

Hong Kong

Italy Israel Egypt Madagascar South Africa Hong Kong

* 0.041 0.043 0.180 0.185 0.230

56/7 * 0.044 0.182 0.187 0.232

50/5 54/6 * 0.177 0.182 0.232

110/44 114/42 102/42 * 0.022 0.220

102/42 104/40 96/37 41/7 * 0.224

107/48 111/52 104/48 96/55 82/53 *

The pairwise nucleotide distance between samples ranged from 0.041 to 0.232 (Table 4), indicating substantial differentiation among the three geographic areas. The number of nucleotide differences varied from 58 to 99, and sequence divergence ranged from 15.7% to 19.2%. Greater distances between the Mediterranean Sea and the Pacific than between the Mediterranean Sea and the Indian Ocean suggested the evolutionary hypothesis that the Pacific and the Indian Ocean populations diverged first, and then Red Sea populations diverged from populations in the Indian Ocean. As in the 16S-rDNA data, NJ, MP and ME trees had the same topology, with sequences from the Mediterranean Sea and the Red Sea, the Indian Ocean, and the Pacific Ocean, defining three major clades. Within the Mediterranean Sea and the Red Sea clade the two sister groups, the L-form and the M-form, were found in both COI and 16S-rDNA data. Within the Indian Ocean clade, 16SrDNA exhibited M and F forms, differentiated by 4% sequence divergence (Fig. 2), revealing dual uniparental inheritance. Two clades (I and II) were also detected at COI (Fig. 2), but each of the clades was detected in both males and females, so these clades do not reflect the DUI revealed by 16S-rDNA. Phylogenetic analysis using 16S-rDNA and COI In addition to the phylogeographic analysis of the Brachidontes (Fig. 2), a phylogenetic analysis was extended to other Mytilidae species: Mytilaster minimus, Mytilus galloprovincialis, M. edulis and G. demissa.

Pairwise nucleotide distances between Mytilidae species for the 16S-rDNA and COI genes were obtained with the TamuraNei method, and they are reported in Table 5. This comparison employed the three Brachidontes clades, since the previous data confirmed that the Mediterranean–Red Sea, the Indian Ocean and the Pacific Ocean were occupied by three highlydifferentiated clades of Brachidontes. Both genes gave congruent results: the matrix showed increasing values of genetic distances from Mediterranean– Red Sea Brachidontes to the genus Mytilus and finally to the outgroup Saccostrea cucculata (Table 5). Mytilaster minimus was genetically closer to the Brachidontes species than G. demissa. Phylogenetic trees (NJ, ME and MP) were constructed with the two mitochondrial genes sequences combined. The three trees were concordant; Fig. 4 presents the NJ tree. The tree presents the genus Brachidontes as the most derived and, within this genus, the most recent bifurcation differentiated the Mediterranean B. pharaonis from the B. variabilis in the Indian Ocean. A robust node (bootstrap value was 99%) supported the monophyly of the Brachidontes genus. It was noted that the sequence divergences between Brachidontes clades (ranging from 16.7% to 20.4%) were always higher than the divergence in the congeneric Mytilus species (4.6%), indicating that they could be considered as three different species. The tree exhibits the amino acid at the 30th position of COI, and it clearly shows that the M-form is derived, for it is present, along with the L-form, in just one of the species resulting from the most recent bifurcation. The L-form is the only form found in the other congeners, and the most closely related species (Fig. 4). Mytilus edulis and M. galloprovincialis have phenylalanine (P) at the 30th amino acid of COI.

Discussion Cryptic species Genetic analyses on samples throughout the range (Fig. 1) revealed three well-differentiated clades identifying three cryptic species. A recent study of genetic variation in Brachidontes in the Caribbean reported a similar result; the taxon previously recognized as B. exustus is composed of four cryptic species (Lee and Foighil 2004). In the pairwise comparison of nucleotide distances of 16SrDNA and COI genes between the three areas, the values of divergence (ranging between 0.123–0.161 and 0.177–0.232,

Table 5. Nucleotide distances of 16S-rDNA gene (below the diagonal) and COI (above the diagonal) using Tamura-Nei method Mediterranean– Red Sea, Brachidontes pharaonis Mediterranean–Red Sea Brachidontes pharaonis Indian Brachidontes variabilis Pacific Brachidontes variabilis Mytilaster minimum Geukensia demissa Mytilus galloprovincialis Mytilus edulis Saccostrea cucculata

Indian, Brachidontes variabilis

Pacific, Brachidontes variabilis

Mytilaster minimum

Geukensia demissa

Mytilus galloprovincialis

Mytilus edulis

Saccostrea cucculata

*

0.185

0.232

0.345

0.283

0.455

0.448

0.602

0.118 0.151 0.278 0.327 0.428 0.438 0.624

* 0.103 0.254 0.319 0.406 0.417 0.636

0.219 * 0.260 0.368 0.398 0.407 0.674

0.358 0.314 * 0.396 0.445 0.435 0.655

0.257 0.261 0.335 * 0.396 0.395 0.649

0.422 0.459 0.492 0.454 * 0.067 0.663

0.413 0.448 0.501 0.444 0.023 * 0.649

0.557 0.566 0.588 0.525 0.633 0.617 *

Mytilidae species: Brachidontes species, Mytilaster minimum, Mytilus galloprovincialis, Mytilus edulis, Geukensia demissa and Ostreidae species: Saccostrea cucculata.

 2007 The Authors J Zool Syst Evol Res (2007) 45(4), 289–298 Journal compilation  2007 Blackwell Verlag, Berlin

296 respectively (Tables 3 and 4) are substantially greater than the distances [0.067 for 16S-rDNA and 0.023 for COI, Table 5) between Mytilus edulis and M. galloprovincialis, and correspond to interspecific distances in other Bivalvia (Shearer et al. 2002)]. Nucleotide distances within the geographic areas (0.005–0.041, Table 2, and 0.02–0.044, Table 4, respectively) are comparable with distances reported in other species of mussels (Shearer et al. 2002). Because M. edulis and M. galloprovincialis are recognized as different species, and the clades reported here have substantially greater distances among them, each Brachidontes clade should be recognized as a distinct species. A systematic revision of the taxon B. variabilis is needed. The name B. pharaonis is most appropriate for the species in the Mediterranean Sea and the Red Sea, but it must be determined whether B. variabilis should be applied to the species in the Indian Ocean, or the species in the Pacific Ocean. The nucleotide distances (Tables 2, 4 and 5) and the phylogenetic analysis (Fig. 4) suggest that the most ancient divergence was between the clades currently in the Pacific and Indian Oceans. Subsequently, the clade currently in the Mediterranean Sea and the Red Sea diverged from the clade in the Indian Ocean. An approximate date of divergence can be estimated if we assume that divergence accumulated at a relatively constant rate through time. A conventional estimate of the rate of the mitochondrial sequence divergence is 20 · 10)9 per site per year per evolutionary line, or 2% sequence divergence per million years, Myr, between pairs of lineages (Brown et al. 1979). The sequence divergence between the Pacific Ocean sequences and those in the Indian Ocean are about 10.3% for 16Sr-DNA and 21.9% for COI (Table 5), so estimates for this first divergence range from 5.6 to 10.5 Mya. Thus, the three Brachidontes diverged during the Miocene (23.8–5.3 My). During the Miocene, Asian and Australasian elements collided, restricting the marine exchange between the northeastern Pacific and the south-western Indian Ocean, and creating Wallace’s Line. Perhaps a common ancestor of the three Brachidontes species had an Indo-Pacific distribution, and reduced marine exchange permitted the divergence of populations in the Pacific Ocean from those in the Indian Ocean. The common ancestor in the Indian Ocean would have migrated naturally into the Red Sea. Sea level fluctuations caused by glacial cycles might have isolated the Red Sea from the Indian Ocean. This could have occurred at Bab el Mandeb, a narrow and shallow sill between the Red Sea and the Gulf of Aden. Currently, the eastern channel is 3 km wide and 30 m deep, and the western channel is 25 km wide and has a maximum depth of 310 m. The family Mytilidae (Rafinesque, 1815) is thought to be an ancient group with roots reaching to the Devonian, and it was supposed to contain an estimated 250 species in 33 genera (Boss 1971) divided among four subfamilies (Newell 1969). It has been supposed that the genus Brachidontes (Swainson, 1840) appeared in the late Cretaceous. This genus belongs to the subfamily of Mytilinae, composed of an uncertain number of species. Currently, 27 of them are in the collection of the American Museum of Natural History, but this is certainly an underestimate, since there are recent discoveries of cryptic species (Lee and Foighil 2004). The systematic classification and phylogeny of the whole family, Mytilidae, are problematic and are not completely developed due to several reasons: (1) many mytilid species and

Sirna Terranova, Lo Brutto, Arculeo and Mitton even genera inhabit deep-water habitats which makes it very difficult to collect them; (2) the are many endemic species; (3) shell shape may be extremely dependent on environmental conditions and on the age of the animal (Seed 1968), it is not always evident which polymorphic characters provide diagnostic characters; (4) in the Bivalvia, the homology of many morphological characters is difficult to determine due to convergent and/or parallel evolution (Steiner and Hammer 2000). Until today, subfamilies and genera of Mytilidae have been assigned largely according to the shape and characteristics of the valves, sculpture, hinges and dentition; these assignments were rarely, if ever guided by genetic data. We conducted a survey of the morphological variation to determine whether the morphological characters used by Krauss to describe B. variabilis could distinguish among mussels collected in the three major geographic areas. Shape, the number of teeth, and colour varied greatly within every population sample, while there were only moderate differences among samples from the Mediterranean Sea and the Red Sea, the Indian Ocean, and the Pacific Ocean. The morphological variability fell within the range of variability attributed to this species, and conformed to the morphologic descriptions by Krauss (1848) and subsequently by Fischer (1870). Further morphological studies should be initiated to identify diagnostic characters for these cryptic species, even though DNA barcoding is coming next to traditional methods as a bioidentification system of animal taxa (Hebert et al. 2003). Evidence of subgroups Evidence for DUI was detected only in the clade in the Indian Ocean, and curiously, at 16S-rDNA but not at COI. This species, like others of Mytiloidea, Heterodonta and Unioida (Skibinski et al. 1994b; Zouros et al. 1994; Hoeh et al. 1996; Liu et al. 1996; Quesada et al. 1996), possesses the unusual feature of DUI of mitochondrial DNA: the F-type has maternal inheritance, and the M-type has paternal inheritance. Occasionally, a maternally transmitted molecule may invade the paternal route, a phenomenon called Ômasculinization of the F-moleculeÕ (Hoeh et al. 1997; Saavedra et al. 1997). If the invading F-molecule becomes fixed in the paternally inherited line, then the divergence between the F-type and the M-type drops to zero, and consequently, evidence of DUI disappears. Thus, it is not presently known whether DUI is truly lacking in the clades in the Pacific and the Mediterranean Sea and the Red Sea, or whether these clades have recently experienced masculinization of the F-molecule. Another possibility is that our primers preferentially amplified the F-type sequences, causing us to miss evidence for DUI. This might explain the curious result in the Indian Ocean clade, in which we find clear evidence for DUI in 16S-rDNA sequences, but not in COI sequences (Fig. 2). MtDNA COI sequences also revealed an ancient amino acid polymorphism in the cytochrome oxidase subunit I and concomitant differentiation in the 16S-rDNA gene. These two forms have been named the L-form and the M-form, referring to either leucine or methionine at the 30th amino acid (Sirna Terranova et al. 2006). Within each of these forms, subsequent substitutions have accumulated so that the L-form and the M-form define clades that are differentiated by 7.3% sequence divergence. The L and M forms are equally distributed in the sexes, and all populations within the Mediterranean contained both forms. This amino acid polymorphism is

 2007 The Authors J Zool Syst Evol Res (2007) 45(4), 289–298 Journal compilation  2007 Blackwell Verlag, Berlin

Evolutionary history of Brachidontes variabilis

297

restricted to the clade in the Mediterranean Sea and the Red Sea. The M-form is the derived form, for the M-form clade has less variation than the L-form clade, and leucine is fixed in the other cryptic species, and it is also in G. demissa and M. minimus, the closest of the outgroup species (Fig. 4). Patterns of the molecular variation suggested that this ancient polymorphism is exposed to some form of balancing selection (Sirna Terranova et al. 2006).

Acknowledgements We thank B. Galil and B. Morton for collecting samples. Partial support was provided by University of Palermo (Fondi di Ateneo ex 60% and CORI project).

Riassunto Lo studio filogeografico di Brachidontes variabilis (Bivalvia: Mytilidae) rivela tre specie criptiche Lo studio filogeografico e` stato condotto su tutto l’areale di Brachidontes variabilis (Krauss, 1848) attraverso l’analisi di sequenze mitocondriali (16S-rDNA e COI) che hanno separato i campioni in tre cladi monofiletici. Diversi algoritmi (NJ, ME e MP) hanno elaborato alberi con la stessa topologia, in cui e` possibile riconoscere: (1) Brachidontes pharaonis s.l. dell’area Mar Mediterraneo – Mar Rosso; (2) Brachidontes variabilis dellÕ Oceano Indiano; (3) Brachidontes variabilis dell’Oceano Pacifico. Il loro grado di divergenza e` sufficientemente alto da potere ascrivere al rango di specie i singoli cladi, nonostante non siano stati ancora individuati i caratteri tassonomici distintivi, a causa della grande variazione morfologica. La divergenza nucleotidica tra le tre linee di Brachidontes era compresa tra 10.3% e 23.2%, in un range di valori superiori a quelli trovati nel confronto tra specie congeneriche di Mytilus sp (2.3–6.7%). Utilizzando il tasso evolutivo, che convenzionalmente viene applicato ai valori di divergenza genetica di geni mitocondriali (2% per milioni di anni), si sono ricavati tempi di divergenza corrispondenti a 6–11 milioni di anni tra Oceano Indiano e Pacifico, e a 6.5–9 milioni di anni tra Mar Rosso e Oceano Indiano. Le tre linee evolutive sembrano essersi separate durante il Miocene. Probabilmente un comune antenato con distribuzione Indo-Pacifica puo` essere andato incontro a processi di vicarianza e/o di dispersione legati alle glaciazioni pleistoceniche prima e all’apertura del Mar Rosso dopo.

References Arcidiacono A, Di Geronimo I (1976) Studio biometrico di alcuni campioni di Brachidontes variabilis (Krauss). Conchiglie 12:61–74. Avise JC (1994) Molecular Markers, Natural History and Evolution. Chapman and Hall, New York, NY, 511pp. Barash A, Danin Z (1986) Further additions to the knowledge of Indo-Pacific Mollusca in the Mediterranean Sea. Spixiana 9:117– 141. Boss KJ (1971) Critical Estimate of the Number of Recent Mollusca. Occas. Pap. Mollusks Mus. Comp. 2001. Harv. Univ. v3, no 40., pp 81–135. Brown WM, Gilbert TL, Wilson AC (1979) Rapid evolution of animal mitochondrial DNA. Proc Nat Ac Sci U S A 76:1967–1971. Chemello R, Oliverio M (1995) Lessepsian migrations: a theoretical Ôisland-jumpingÕ model. Biol Mar Medit 3:444–446. Chromas (2006) http://www.technelysium.com.au/chromas14x.html. Davis JI, Nixon KC (1992) Populations, genetic variation, and the delimitation of phylogenetic species. Syst Biol 41:421–435. Felsenstein J (1985) Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39:783–791. Fischer P (1870) Sur la Faune conchyliologique marine des baies de Suez et de l’Akabah. J Conchyliol XIII:178–179. Folmer O, Black M, Hoen F, Lutz R, Vrijenhoek R (1994) DNA primers for amplification of mitochondrial cytochrome c oxidase

subunit I from diverse metozoan invertebrates. Mol Mar Biol Biot 3:294–299. Gianguzza P, Chemello R, Riggio S (1997) Segnalazione di Brachidontes pharaonis (P. Fischer, 1870) (Bivalvia, Mytilidae) nella salina di Marsala e considerazioni sulla distribuzione della specie in Mediterraneo. Boll Malac 33:169–172. Hebert PDN, Cywinska A, Ball SL, de Waard JR (2003) Biological identifications through DNA barcodes. Proc R Soc Lond B Biol Sci 270:313–321. Hoeh WR, Stewart DT, Saavedra C, Sutherland BW, Zouros E (1996) Cytochrome c Oxidase sequence comparisions suggest an unusually high rate of mitochondrial DNA evolution in Mytilus (Mollusca: Bivalvia). Mol Biol Evol 13:418–421. Hoeh WR, Stewart DT, Saavedra C, Sutherland BW, Zouros E (1997) Phylogenetic evidence for role-reversals of fender-associated Mitochondrial DNA in Mytilus (Bivalvia: Mytilidae). Mol Biol Evol 14:959–967. Kumar S, Tamura K, Nei M (2001) MEGA: Molecular Evolutionary Genetic Analysis, version 2.1. Pennsylvania State University, University Park, PA. Lavee D (1981) Genetic variability in a colonizatory system: comparison between the mussels Brachidontes variabilis and Modiolus Auriculatus (Bivalvia: Mytilidae). PhD Dissertation, The Hebrew University of Jerusalem, Jerusalem, Israel, 80pp. Lee T, Foighil D (2004) Hidden Floridian biodiversity: mitochondrial and nuclear gene trees reveal four cryptic species within the scorched mussel, Brachidontes exustus, species complex. Mol Ecol 13:3527– 3542. Liu HP, Mitton JB, Wu SK (1996) Paternal mitochondrial DNA differentiation far exceeds maternal DNA and allozyme differentiation in the freshwater mussel, Anodonta grandis grandis. Evolution 50:952–957. Morton B (1988) The population dynamics and reproductive cycle of Brachidontes variabilis (Bivalvia: Mytilidae) in a Hong Kong Mangrove. Malacological Review 21:109–117. Nei M (1987) Molecular Evolutionary Genetics. Columbia University Press, New York, NY. Nei M, Kumar S (2000) Molecular Evolution and Phylogenetics. Oxford University Press, New York. Newell ND (1969) Subclass Palaeoheterodonta. In Moore RC (ed.), Treatise on invertebrate paleontology: mollusca 6. Geological Society of America, University of Kansas, Boulder, Colorado & Lawrence, Kansas, KS, pp. N393–N489. Palumbi SR (1996) Nucleic acids II: the polymerase chain reaction. In: Hillis DM, Moritz C, Mable BK (eds), Molecular Systematics. Sinauer Associates, Inc., Sunderland, pp 205–248. Quesada H, Skibinski DAG, Skibinski DOF (1996) Sex-biased heteroplasmy and mitochondrial DNA in the mussel Mytilus galloprovincialis. Lmk Cur Gen 29:423–426. Rawson PD, Hilbish TJ (1995) Evolutionary relationships among the male and female Mitochondrial DNA lineages in the Mytilus edulis species complex. Molec Biol Evol 12:893–901. Rilov G, Gasith A, Benayahu Y (2002) Effect of an exotic prey on the feeding pattern of a predatory snail. Mar Envir Res 54:85–98. Rozas J, Rozas R (1999) DnaSP version 3: an integrated program for molecular population genetics and molecular evolution analysis. Bioinformatics 15:174–175. Rzhetsky A, Nei M (1992) A simple method for estimating and testing minimum-evolution trees. Mol Biol Evol 9:945–967. Saavedra C, Reynero M, Kouros E (1997) Male-dependent doubly uniparental inheritance of mitochondrial DNA and female-dependent sex-ratio in the mussel Mytilus galloprovincialis. Genetics 145:1073–1082. Safriel UN, Ritte U (1983) Universal correlates of colonizing ability. In: Swingland IR, Greenwood PJ (eds), The Ecology of Animal Movement. Clarendon Press, Oxford, pp 215–239. Safriel UN, Gilboa F, Felsenburg D (1980) Distribution of rocky intertidal mussels in the Red Sea coasts of Sinai, the Suez Canal and the Mediterranean coast of Israel, with special reference to recent colonizers. J Biog 7:39–62. Saitou N, Nei M (1987) The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol Biol Evol 6:514–525.

 2007 The Authors J Zool Syst Evol Res (2007) 45(4), 289–298 Journal compilation  2007 Blackwell Verlag, Berlin

298 Sara`, G, Romano C, Caruso M, Mazzola A (2000) The new lessepsian entry Brachidontes pharaonis (Fischer P., 1870) (Bivalvia, Mytilidae) in the western Mediterranean: a physiological analysis under varying natural conditions. J Shell Res 19:967–977. Sasekumar A 1974) Distribution of macrofauna on a Malayan mangrove shore. J Anim Ecol 43:51–69. Seed R (1968) Factors influencing shell shape in the mussel Mytilus edulis. J Mar Biol Ass U K 48:561–584. Shearer TL, Van Oppen MJH, Romano SL, Wo¨rheide R (2002) Slow mitochondrial DNA sequence evolution in the Anthozoa (Cnidaria). Mol Ecol 11:2475–2487. Shefer S, Abelson A, Mokady O, Geffen E (2004) Red to Mediterranean Sea bioinvasion: natural drift through the Suez Canal, or anthropogenic transport? Mol Ecol 13:2333–2343. Sirna Terranova M, Lo Brutto S, Arculeo M, Mitton JB (2006) Population structure of Brachidontes pharaonis (P. Fischer, 1870) (Bivalvia, Mytilidae) in the Mediterranean Sea, and evolution of a novel mtDNA polymorphism. Mar Biol 159:89–101. Skibinski DOF, Gallagher C, Beynon CM (1994a) Sex-limited mitochondrial DNA transmission in the marine mussel Mytilus edulis. Genetics 138:801–809. Skibinski DOF, Gallagher C, Beynon CM (1994b) Mitochondrial DNA inheritance. Nature 368:817–818. Steiner G, Hammer S (2000) Molecular phylogeny of the Bivalvia inferred from 18S rDNA sequences with particular reference to the Pteriomorphia. In: Harper EM, Taylor JD, Crame JA (eds), The

Sirna Terranova, Lo Brutto, Arculeo and Mitton Evolutionary Biology of the Bivalvia. Geological Society, London, Special Publications 177, pp. 11–29. Stepien CA, Kocher TD (1997) Molecules and morphology in studies of fish evolution. In: Kocher TD, Stepien CA (eds), Molecular Systematics of Fishes, Ch. 1. Academic Press, San Diego, CA, pp. 1–11. Tamura K, Nei M (1993) Estimation of the number of nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees. Molec Biol Evol 10:512–526. Taylor JD (1971) Reef associated molluscan assemblages in the Western Indian Ocean. Symp Zool Soc Lond 28:501–534. Thompson JD, Higginsm DG, Gibson TJ (1994) CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, positions-specific gap penalties and weight matrix choice. Nucl Ac Res 22:4673–4680. Zouros E, Saavedra AO, Freeman KR (1994) An unusual type of mitochondrial DNA inheritance in the blu mussel Mytilus. Proc Nat Ac Sci U S A 91:7463–7467. AuthorsÕ addresses: Margherita Sirna Terranova, Sabrina Lo Brutto and Marco Arculeo (for correspondence), Dipartimento di Biologia Animale, Universita` di Palermo, via Archirafi 18, 90123 Palermo, Italy. E-mail: [email protected], [email protected], marculeo@ unipa.it; Jeffry B. Mitton, Department of Ecology and Evolutionary Biology, University of Colorado, Boulder, Colorado, 80309, USA. E-mail: [email protected]

 2007 The Authors J Zool Syst Evol Res (2007) 45(4), 289–298 Journal compilation  2007 Blackwell Verlag, Berlin