Gene 456 (2010) 45–53
Contents lists available at ScienceDirect
Gene j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / g e n e
Scottish Mytilus trossulus mussels retain ancestral mitochondrial DNA: Complete sequences of male and female mtDNA genomes Małgorzata Zbawicka a,⁎, Artur Burzyński a, David Skibinski b, Roman Wenne a a b
Department of Genetics and Marine Biotechnology, Institute of Oceanology, Polish Academy of Sciences, Powstańców Warszawy 55, 81-712 Sopot, Poland School of Medicine, Swansea University, SA2 8PP, Wales, Swansea, UK
a r t i c l e
i n f o
Article history: Received 19 November 2009 Received in revised form 1 February 2010 Accepted 17 February 2010 Available online 3 March 2010 Received by N. Okada Keywords: Mitogenomics Complete mitochondrial genome DUI CR duplication Phylogeography
a b s t r a c t Mytilus trossulus mussels occur in North America and in the Baltic Sea. Recently genetic markers for the three Mytilus subspecies M. edulis, M. galloprovincialis, and M. trossulus, have been detected at Loch Etive in Scotland suggesting mixed ancestry for this population. Of particular interest is the evidence that M. trossulus occurs at Loch Etive because it had not previously been reported in the British Isles. In the present study, analysis of subspecies-specific diagnostic nuclear DNA markers confirms the presence of a high frequency of mussels with M. trossulus ancestry at Loch Etive. The genetic structure suggests hybridisation at an intermediate stage compared with North American populations, where there is little hybridisation, and Baltic populations where there is extensive introgression. This points strongly against a Baltic origin for Loch Etive M. trossulus. The F and M mitochondrial DNA (mtDNA) genomes of Baltic M. trossulus are similar in sequence to the corresponding genomes in M. edulis and believed to be derived by introgression from that subspecies. Both F and M mtDNA genomes are observed at Loch Etive consistent with the presence of doubly uniparental inheritance. Here we provide the complete sequences of the three M. trossulus mtDNA genomes (one F and two M) from Loch Etive. These genomes are extremely similar to the corresponding genomes from ancestral M. trossulus in America but divergent from the genomes for Baltic M. trossulus. This is the first report of ancestral M. trossulus mtDNA genomes in Europe. The F and M genomes are diverged by 26% in nucleotide sequence, similar to other Mytilus F and M genomes. The gene arrangement in the sequenced genomes is also similar to that in other sequenced Mytilus mtDNA genomes. However the two sequenced M genomes differ by 960 bp which is caused by a duplication in the main noncoding region (CR). This duplication has not so far been observed in North American populations of M. trossulus. The coding regions of the Loch Etive genomes have no features suggesting that they are other than functional genomes and have Ka/Ks values in coding regions less than one indicative of purifying selection. Estimates of divergence times were made for both genomes and are consistent with invasion of Loch Etive by M. trossulus towards the end of the last glacial period. © 2010 Elsevier B.V. All rights reserved.
1. Introduction Abbreviations: atp6, ATP synthase subunit 6 gene; atp8, ATP synthase subunit 8 gene; bp, base pair; CD, constant domain; cob, apocytochrome b gene; cox1, cytochrome c oxidase subunit 1 gene; cox2, cytochrome c oxidase subunit 2 gene; cox3, cytochrome c oxidase subunit 3 gene; CR, control region; DUI, doubly uniparental inheritance; FEL, Fixed Effects Likelihood; Gln, glutamine; ITS, internal transcribed spacer; lrrna, large ribosomal RNA gene; mtDNA, mitochondrial DNA; MYA, million years ago; nad1, NADH dehydrogenase subunit 1 gene; nad2, NADH dehydrogenase subunit 2 gene; nad3, NADH dehydrogenase subunit 3 gene; nad4, NADH dehydrogenase subunit 4 gene; nad4l, NADH dehydrogenase subunit 4L gene; nad5, NADH dehydrogenase subunit 5 gene; nad6, NADH dehydrogenase subunit 6 gene; PARRIS, a Partitioning approach for Robust Inference of Selection; PCR, polymerase chain reaction; REL, Random Effects Likelihood; rRNA, ribosomal RNA; SLAC, Single Likelihood Ancestor Counting; srrna, small ribosomal RNA gene; tRNA, transfer RNA; trnA, transfer RNA Alanine; trnQ, transfer RNA glutamine; trnY, transfer RNA Tyrosine; VD1, variable domain 1; VD2, variable domain 2. ⁎ Corresponding author. Tel.: +48 58 551 72 81; fax: +48 58 551 21 30. E-mail address:
[email protected] (M. Zbawicka). 0378-1119/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.gene.2010.02.009
There are three recognised subspecies of mussels of the genus Mytilus in Europe: M. edulis, M. galloprovincialis and M. trossulus (McDonald et al., 1991; Gosling, 1992). M. trossulus originated in the Pacific, and colonised the northern Atlantic after the Bering Strait opened 3.5 MYA. It there gave rise to M. edulis and M. galloprovincialis (Riginos and Cunningham, 2005). Until recently it was believed that after a more recent invasion of Pacific mussels into the Atlantic, M. trossulus colonised two regions only: the Canadian Maritimes in North America and the Baltic Sea in Europe (Riginos and Cunningham, 2005). However, mussels with M. trossulusspecific alleles have been found in the Netherlands (Śmietanka et al., 2004) and allozyme data suggest that M. trossulus is also possibly present in Norway, in fjords near Bergen (Ridgway and Nævdal, 2004). Pure ancestral M. trossulus mussels have been found
46
M. Zbawicka et al. / Gene 456 (2010) 45–53
only in American populations (Riginos and Cunningham, 2005). M. trossulus mussels in Europe are generally localised at sites with low salinity and North American M. trossulus is also more tolerant to low salinity than M. edulis (Qiu et al., 2002). M. edulis and M. galloprovincialis occur at higher salinity. M. edulis is located in the northern part of the Atlantic and European seas, whereas M. galloprovincialis is distributed mainly in the Mediterranean Sea and along the Atlantic coastline of Western Europe. The hybrid zone between these subspecies spans more than 1000 km and occurs along the western coast of Europe from the Bay of Biscay along the French coast to Great Britain and Ireland (Skibinski et al., 1978; Coustau et al., 1991; Gosling, 1992; Bierne et al., 2003a). Hybrid zones of M. trossulus and M. edulis are now well documented in the Danish straits separating the Baltic and North Seas (Riginos and Cunningham, 2005), and in North America (Saavedra et al., 1996; Comesaña et al., 1999). There is recent evidence for the occurrence of M. trossulus mussels and M. trossulus/M. edulis hybrids in Loch Etive, Scotland (Beaumont et al., 2008) on the basis of morphology, diagnostic allozyme and a nuclear DNA marker. It was suggested that these M. trossulus mussels are a post glacial relict rather than a recent introduction associated with shipping activity. The gene arrangement of Mytilus mtDNA is different from that of other studied animals. It lacks atp8 and contains an extra tRNA for methionine. There are 12 proteins, 2 rRNAs, 23 tRNAs and the main noncoding control region (CR) which consists of three parts, two variable domains (VD1, VD2) and one conserved domain (CD) (Cao et al., 2004). Mytilus has an unusual type of mitochondrial DNA inheritance (referred to as doubly uniparental inheritance, DUI). Females have a maternally transmitted mtDNA (genome F), and males possess an additional mtDNA (genome M) which is transmitted paternally (Skibinski et al., 1994a,b; Zouros et al., 1994a,b). The F and M genomes in Mytilus are very similar to each other in gene content, arrangement, nucleotide composition and codon usage bias despite being highly diverged (up to 23%). The complete F and M mitochondrial genomes of M. edulis and M. galloprovincialis have been sequenced (Boore et al., 2004; Mizi et al., 2005; Breton et al., 2006; Venetis et al., 2007; Burzyński and Śmietanka, 2009). The F genomes of M. edulis and M. galloprovincialis are very closely similar, the main differences occurring in the noncoding regions. The ancestral M. trossulus F and M mtDNA genomes have been observed to date only in North American populations (Rawson and Hilbish, 1995; Stewart et al., 1995; Rawson, 2005; Cao et al., 2009). A genome from Nova Scotia which has both Mlike and F-like sequences in the CR, but is otherwise like the F genome, has been completely sequenced (Breton et al., 2006). There is now evidence that this is transmitted maternally (Cao et al., 2009) and it will be referred to here as the North American F genome of M. trossulus. For the ancestral M genome of M. trossulus only the sequence of three fragments of mtDNA are known: one covering a fragment of lrrna (Rawson and Hilbish, 1995; Rawson, 2005), the entire CR and a cob fragment (Rawson, 2005; Cao et al., 2009), the second covering part of cox1 (Riginos et al., 2004) and the third covering part of cox3 (Riginos et al., 2004). The complete mtDNA sequences of the F and M genomes from Baltic M. trossulus (Zbawicka et al., 2007) have low genetic distance from the respective F and M genomes of M. edulis suggesting that they derive through recent introgression from M. edulis (Zbawicka et al., 2007; Kijewski et al., 2006). The present study focuses on the question as to whether M. trossulus identified at Loch Etive in Scotland (Beaumont et al., 2008) derives from North American populations or from the Baltic. The entire sequences of both F and M mtDNA M. trossulus genomes in Loch Etive are sequenced and shown to have high affinity with the corresponding ancestral M. trossulus genomes from North America. The organisation of these genomes is discussed and compared with that of other sequenced Mytilus genomes.
2. Materials and methods/experimental procedures 2.1. Samples and DNA extraction Adult, reproductively mature mussels (40–50-mm shell length) were obtained from Loch Etive, Scotland at locations 56° 26′ 55″ N and 5° 11′ 47″ W, 2–4 m in depth. Mussels were sexed by microscopic examination of gonadal tissue. DNA was isolated from gametes from 39 females and 62 males as described previously (Burzyński et al., 2006) and analysed by PCR. 2.2. Taxonomic identification of mussels To determine the ancestry of the mussels studied, the following three nuclear DNA markers were used: Me 15/16 which is diagnostic for all three taxa M. edulis, M. trossulus, and M. galloprovincialis (Inoue et al., 1995), ITS which differentiates M. trossulus from other Mytilus taxa (Heath et al., 1995), and Efbis which is diagnostic for all three taxa (Bierne et al., 2003b; Kijewski et al., 2006). A sample of 105 mussels from Loch Etive were analysed using these three markers. 2.3. PCR and sequencing strategy Identification of M. trossulus mtDNA was performed by PCR amplification of the diagnostic fragments of the CR with a set of subspecies- and gender-specific primers (Supplementary Table 1). The successful amplification of a 900 bp fragment with the TRO5– TRO6 primers and absence of amplification with the primers AB23– TRO2 indicate the presence of an M. trossulus F genome. The successful amplification of a 1200 bp fragment with the AB23–TRO2 primers and absence of amplification with the TRO5–TRO6 primers indicate the presence of the M genome of M. trossulus. One female having the F haplotype (mussel 34LE) and two males with the M haplotype (mussels 117LE and 149LE) were chosen for complete mtDNA sequencing. The complete genome sequencing strategy had two stages: first the sequencing of the CR and second the sequencing of the coding part of the genome. PCR primers were designed based on GenBank records: NC_006161, AY823625, DQ198225, and DQ013366. For sequencing of the CR of the F genome three overlapping PCR products were used AB52–TRO4 for the 5′ end, TRO5–TRO6 for the 3′ end, and AB40–AB41 for the central region. For the CR of the M genome, three pairs of primers were used: AB52–TRO2 for the 5′ end, AB30–TRO1 for the 3′ end and CBM1–CBM2 for the central region. For sequencing of the coding part of the genome, two PCR steps and direct sequencing was used. A long-range PCR reaction was performed for each genome. For the F genome the universal AB23 forward and AB33 reverse primer gave a fragment of approximately 18 kb. For the M genome, the M specific TRO1 and universal AB33 reverse primers gave a fragment of approximately15 kb. AB23 and AB33 are localised in lrrna and the TRO1 specific primer in the CR (VD1) region. Subsequently, a set of 40 primers was used for nested PCR allowing re-amplification of twenty overlapping fragments for approximately 2× coverage. The PCR fragments were sequenced directly, as described previously (Burzyński et al., 2003; Zbawicka et al., 2007). The location, sequence, annealing temperature (optimized in gradient PCR) and other information for all primers are given in Supplementary Table 1. PCR amplifications were carried out in 15 µl reaction volumes containing 20 ng of template DNA, 0.4 µM of each primer, 200 µM nucleotides, 1.5 mM magnesium chloride, 0.5 U of high-fidelity DyNAzymeEXT2 DNA polymerase and appropriate reaction buffer from Finnzymes. After an initial denaturation for 5 min at 95 °C, 30 cycles were used with denaturation at 94 °C for 1 min (but 3 min for first cycle), annealing for 30 s, and extension at 72 °C for 55 s with a final 5 min extension at 72 °C. For long-range PCR, Phusion Pfx (Finnzymes Oy) polymerase was used according to
M. Zbawicka et al. / Gene 456 (2010) 45–53
the manufacturer protocol. For re-amplifications, a 1:1500 dilution of the long-range PCR product was used as template. All re-amplifications were performed as described previously (Zbawicka et al., 2007). All PCR reactions were performed in a T-gradient cycler from Biometra, Tampa, FL. PCR products (2 µl of each amplification) were visualized on 1% agarose gels stained with ethidium bromide. 2.4. Bioinformatic analysis Sequence assembly was facilitated by Phred (Ewing et al., 1998), Gap4 (Bonfield et al., 1995) and Staden (Staden et al., 2001) computer programs. Gene annotations were done semi-automatically. De novo prediction of all protein coding genes was attempted using a set of algorithms implemented in CRITICA (Badger and Olsen, 1999), Glimmer3 (Delcher et al., 1999) and wise2 (Birney et al., 2004). For prediction of RNA genes, the Cove implemented covariance model was employed (Eddy and Durbin, 1994). All predictions have been inspected and critically evaluated after comparison with the closest RefSeq annotations. Sequences were aligned using the ClustalX program version 1.83 (Thompson et al., 1997) with equal “pairwise” and “multiple” alignment parameter values (gap opening penalty = 15, gap extension penalty = 6.66). Amino acid sequences for protein coding genes were obtained using the genetic code of Drosophila mtDNA (as in Hoffmann et al., 1992). Genetic distance (K) based on Kimura's two-parameter model (1980), and distance for protein genes for synonymous (Ks) and nonsynonymous (Ka) substitutions, calculated by the modified Nei– Gojobori method (Nei and Gojobori, 1986) with Jukes–Cantor correction, were calculated using the computer program MEGA3 version 3.1 (Kumar et al., 2004), with standard error (SE) computed over 1000 bootstrap replicates. Sliding window plots were constructed using DnaSP version 4.10 (Rozas et al., 2003). Nonsynonymous and synonymous changes at individual codons were estimated in order to identify amino acid sites under positive or negative selection using likelihood based methods implemented in a software package available at http://www.datamonkey.org/ (Kosakowsky Pond and Frost, 2005). A further maximum likelihood approach for identifying positive selection (Scheffler et al., 2006) and implementation in this package was also used. The nature of size variation in the CR sequence of the genomes studied was investigated using BLAST searches and DotPlot comparisons. The DotPlot comparisons were made using the dotmatcher program from EMBOSS (Rice et al., 2000), with a threshold of 65% sequence similarity and a sliding window of 15 bp. Neighbor-joining phylogenetic trees for three mtDNA regions (parts of lrrna, cox1 and cob) were constructed using Mega version 3.1 (Kumar et al., 2004). Bootstrap support was assessed with 1000 replicates. For concatenated proteins, neighborjoining phylogenetic trees were constructed using the PAM Matrix (Dayhoff). The complete mtDNA sequences of F (34LE) and M (149LE and 117LE) variants obtained in this study have been deposited in GenBank (accession numbers GU936625–7). 3. Results and discussion 3.1. Taxonomic status of mussels studied The Loch Etive population has a high frequency of mussels heterozygous for alleles diagnostic for the three different subspecies, thus indicating considerable hybrid ancestry (Table 1). The frequencies of the T allele are 0.39 for Me 15/16, 0.38 for ITS and 0.50 for Efbis, and are significantly heterogeneous using Fisher's exact test (P = 0.024). The cause of the difference, in particular the high frequency for Efbis is not clear. It is not known whether or not these markers are linked, however differential introgression which has been reported and discussed previously in Mytilus (Bierne et al., 2003b) would be a plausible explanation. The frequency distribution of the
47
score for a hybrid index giving the number of M. trossulus gene copies out of a maximum of six for alleles at the loci Me15/16, ITS, and Efbis is given in Fig. 1. This confirms a high frequency of mussels having M. trossulus ancestry. These results support those of previous studies which provided evidence of mixed genetic composition of M. edulis, M. trossulus and M. galloprovincialis markers at Loch Etive reflecting introgression (Beaumont et al., 2008; Dias et al., 2008). The genetic structure of the Loch Etive population can be compared with that for Baltic and North American M. trossulus populations where studies using three nuclear markers have also been carried out (Riginos and Cunningham 2005; Kijewski et al., 2006). All three regions have high frequencies of E (M. edulis) and T (M. trossulus) diagnostic alleles, but Loch Etive is unique in having mussels with G (M. galloprovincialis) diagnostic alleles though at a low frequency. Both European populations have a high frequency of mussels that are heterozygous or have intermediate hybrid index values, whereas such mussels occur at low frequency in North America. This suggests extensive hybridisation in Europe but not in North America. Of most interest is that the multiple heterozygote with one T allele at each locus, consistent with F1 hybrid status, occurs in 11% of Loch Etive mussels but has not been observed in North America or the Baltic. This suggests that introgressive hybridisation is absent or at a low level in North America, at an intermediate stage at Loch Etive and extensive in the Baltic. This excludes the possibility that Loch Etive mussels derive from the Baltic and is consistent either with two separate transAtlantic invasions, or that Baltic M. trossulus are derived from another European population at an earlier stage in the introgression process, such as that at Loch Etive. In this latter circumstance the invasion of the Baltic would have been accompanied by the loss of the native North American mtDNA genomes. In the present study, the female from which the F genome was isolated had the genotype ET, TT, and TT for Me15/16, ITS and Efbis respectively. Males from which the M genome was isolated were ET, XX, TT (117LE) and TT, XT and TT (149LE), where X could be E or G. Morphologically the female belonged to the previously defined “fragile” mussels whereas both males to the “intermediate” group, reflecting certain similarity to M. trossulus (Beaumont et al., 2008). Thus all three mussels demonstrated some similarities to M. trossulus both in terms of marker genotype and morphology, though because of the introgressed nature of the population further comment on their provenance cannot be made. 3.2. Taxonomic affinity of mtDNA genomes sequenced To establish the taxonomic affinity of the mtDNA genomes studied, neighbor-joining trees for sequenced fragments of the lrrna, cox1, and cob were constructed (Fig. 2). For all three gene regions a very similar topology is obtained. The F genome sequenced in this study (from mussel 34LE) is always within a clade with F sequences from American M. trossulus, while the sequenced M genomes (from 117LE and 149LE) are always located within a clade with M sequences from American M. trossulus. This is thus clear evidence that the Loch Etive mtDNA genomes and the Loch Etive mussels from which they were obtained have American ancestry. This is thus the first demonstration of the presence of ancestral M. trossulus F and M mtDNA genomes in Europe. This should be contrasted with the situation for Baltic M. trossulus. In the Baltic, both mtDNA genomes have high similarity with M. edulis genomes consistent with recent introgression from M. edulis (Zbawicka et al., 2007; Kijewski et al., 2006). The distance (K) between American and Loch Etive genomes of M. trossulus was calculated for the lrrna, cob and cox1 fragments used for the phylogenetic tree construction. K values for the F genome range from 0.000 (SE = 0.000) for some lrrna fragments (T-2, AK4-F2 and AK12-F3 versus 34LE, Fig. 2A) and cob fragments (AM3 versus 34LE, Fig. 2C) to 0.014 (SE = 0.009) for one cob fragment comparison
48
M. Zbawicka et al. / Gene 456 (2010) 45–53
Table 1 Genotype frequencies of three nuclear markers Me 15/16, ITS and Efbis for Loch Etive mussels and for a previous study of Baltic mussels (Kijewski et al., 2006). For Me 15/16 and Efbis, E, T, and G refer to alleles diagnostic for M. edulis, M. trossulus, and M. galloprovincialis respectively. For ITS, X and T refer to an alleles diagnostic for edulis/galloprovincialis and trossulus respectively. N, number of individuals.
Loch Etive Baltic
N
Me 15/16
ITS
EE
ET
TT
EG
TG
XX
XT
TT
EE
ET
TT
EG
TG
105 189
0.4 0.42
0.24 0.46
0.25 0.12
0.08 0.00
0.04 0.00
0.41 0.42
0.42 0.57
0.17 0.01
0.36 0.00
0.16 0.05
0.4 0.95
0.04 0.00
0.04 0.00
(AK12-F3 versus 34LE, Fig. 2C). Corresponding F genome K values are higher for some comparisons between American M. trossulus fragments (0.0071, SE = 0.0038 between AY636152 and AY636153 for lrrna). This supplements the evidence from the phylogenetic trees confirming close similarity of the Loch Etive and American F genomes. K values for M genome comparisons range from 0.002 (SE = 0.002) for some lrrna fragments (T4 versus 149LE, Fig. 2A) to 0.021 (SE = 0.011) for cob fragments (SH18 versus 149LE, Fig. 2C). For the cox1 fragment, a higher K value is observed between the two Loch Etive M genomes (0.012, SE = 0.006) than between Loch Etive and some American fragments (0.0035, SE = 0.003, between 149LE and 737NF, Fig. 2B). Thus as for the F genome, the K values confirm the close similarity of the Loch Etive and American M genomes. 3.3. Mitochondrial genome organisation The complete sequence of the Loch Etive M. trossulus F genome described here is 18,653 bp long. It is one nucleotide longer than the North American F genome of M. trossulus (Breton et al., 2006; Cao et al., 2009), caused by a single nucleotide insertion in the CR, another between the nad5 and nad6 genes and a single nucleotide deletion in the lrrna gene. The sequenced M genomes of the Loch Etive M. trossulus are 16,578 bp (149LE) and 17,538 bp (117LE) long. The 960 bp length difference between these two genomes is caused by a 931 bp duplication in the major noncoding region and by a longer homoadenine run in the 5′ part of the CR. The shorter M genome (149LE) is about 50 nucleotides shorter than the M genome of M. edulis (Breton et al., 2006) and the M. edulis derived M genome of Baltic M. trossulus (Zbawicka et al., 2007). The two Loch Etive M genomes have the same gene order in the coding part as the F genome, and are similar in gene arrangement to previously published M. edulis and M. galloprovincialis genomes. Between the Loch Etive F and M genomes, distance (K) = 0.260 (SE = 0.005) on average, similar to the values calculated for congeneric pairs of genomes in M. edulis, M. galloprovincialis (Mizi et al., 2005) and Baltic M. trossulus (Zbawicka et al., 2007) varies along the genome (Fig. 3). There is generally good agreement in the pattern of K for the two subspecies. There are four regions of particularly high K and three of low K. The highest value occurs in noncoding regions: the CR and the part between nad3 and cox1 (UR4), and in some coding regions: the end of cob, the end of nad5 and the centre of nad6. The lowest value occurs in the rRNA genes, in two clusters of tRNA: between nad2 and nad3 and between cox2 and nad1. Some slight differences occur between subspecies, a higher K value at the beginning of nad1 for M. edulis and higher in nad6 for M. trossulus. However the general similarity in the pattern of change in the two subspecies suggests that evolutionary forces acting to cause differences between the F and M genomes are not markedly different in the two subspecies.
Efbis
lrrna in the studied M. trossulus F genome are similar to that in M. edulis, M. galloprovincialis and Baltic M. trossulus F genomes (Hoffmann et al., 1992; Mizi et al., 2005; Zbawicka et al., 2007). In the M genomes of Loch Etive M. trossulus, lrrna is three to five nucleotides shorter than in the other M genomes (Mizi et al., 2005; Breton et al., 2006; Zbawicka et al., 2007). Srrna in the Loch Etive M genome of M. trossulus is shorter by 52 (for 117LE) and 53 (for 149LE) nucleotides than the Loch Etive F genome, and 30 nucleotides shorter than in M. edulis and the Baltic M. trossulus M genome. The F and M genomes of Loch Etive M. trossulus contain 23 functional tRNA genes, and their order is the same as in M. edulis and M. galloprovincialis except that a trnQ gene occurs within the CR between an M-like and Flike part as previously reported for other genomes (Cao et al., 2009). This gene is functional, replacing the trnQ normally present in the coding region of most mtDNA genomes. There are in addition six tRNA pseudogenes, including one for Gln, in the CR. Both the functional trnQ and the pseudogenes have the same K value (0.015) when compared against the corresponding genes in North American M. trossulus. The K values between trnQ and its pseudogene within mussel 34LE and within the North American F genome are much higher at 0.168 and 0.207 respectively. For Loch Etive M. trossulus tRNA genes, the K value between the F and M genomes is 0.127 on average, and is higher than for other reported Mytilus comparisons. The most divergent of the tRNA genes is trnA, for which K = 0.379 on average between the M genome of Loch Etive M. trossulus and other reported Mytilus M genomes. Several differences in gene length occur between Loch Etive M and F genomes, caused by single deletions or insertions. Four protein coding genes differ in length: cox1, nad1, nad3 and nad5, by 20, 42, 1 and 6 amino acids respectively. Most protein coding genes have the same lengths in the North American M. trossulus F genome and in the F genome (34LE) reported here. There are three exceptions. In the Loch Etive F genome, nad1 starts earlier, for an additional 162 bp in length and finishes one nucleotide later with the incomplete
3.4. RNA and protein genes Information on the length, base composition and distance values for genes and noncoding regions of F (34LE) and M (117LE, 149LE) genomes of M. trossulus are given in Supplementary Table 2. Salient features are referred to below. The length and nucleotide sequence of
Fig. 1. The frequency distribution of the score for a hybrid index giving the number of M. trossulus gene copies out of a maximum of six for alleles at the Me15/16, ITS, and Efbis loci. A score of zero is either a pure M. edulis or M. edulis/M. galloprovincialis hybrid, whereas a score of six is a pure M. trossulus. Data from the Baltic Sea are from Kijewski et al. (2006). Scottish population is labelled in black, Baltic Sea population is grey.
M. Zbawicka et al. / Gene 456 (2010) 45–53 Fig. 2. Neighbor-joining unrooted phylogenetic trees for three mtDNA fragments of M. edulis, M. trossulus and M. galloprovincialis based on (A) 430-bp alignment of lrrna gene sequence fragments, (B) 324-bp alignments of cox1 gene sequences, and (C) 149-bp alignments of cob gene sequences. MtDNA genomes from the present study are marked in bold. Fragments of whole genomes: GB1—NC_006161 (Hoffmann et al., 1992), GR1—AY497292, GR2—AY363687 (Mizi et al., 2005), AM1—AY823623, AM2—AY823624, AM3—AY823625 (Breton et al., 2006), C—DQ399833 (Venetis et al., 2007), 39mc10—DQ198231, 87mc10—DQ198225 (Zbawicka et al., 2007). Other mtDNA fragments: E-4—U22867, G-4—U22872, G-5—U22873, G-10—U22878, GM-2 —U22885, T-2—U22880, T-4—U22882 (Rawson and Hilbish, 1995), AK3—DQ013366, AK4-F2—AY636150, AK12-F3—AY636151, SH18—DQ013367 (Rawson, 2005), 42Ori—EF434638 (Filipowicz et al., 2008), IMT1—AF242027 (Wares and Cunningham, 2001), 737NF—AY101432, 764WA—AY101435, 382WH—AY101414 (Riginos et al., 2004), MedFSR7—EU332504, PL1—AF527535, and SYD11021—DQ864424 unpublished. The bars indicate mtDNA lineages.
49
50
M. Zbawicka et al. / Gene 456 (2010) 45–53
Fig. 3. The distance (K) between the F and M genomes. Both genomes are linearised at the first nucleotide position of the major noncoding region and scanned with a sliding window of 150 bp moving in 50 bp steps. Comparison of M. trossulus genomes (34LE, AY823625—F and 117LE, 149LE—M) in black and comparison of M. edulis-like genomes (DQ198231 NC_006161, AY497292—F, DQ198225, and AY823623—M) in grey.
stop codon TA, cox3 is 143 bp longer at 936 bp, and one additional adenine TA(A)ATG occurs at position 15,433 between nad5 and nad6 so that these genes no longer overlap as in all other Mytilus genomes. For the M genome of M. trossulus, three protein coding genes differ from those in the other two Mytilus subspecies. Nad1 is much shorter than in M. edulis, M. galloprovincialis and Baltic M. trossulus F and M genomes because of a N100 bp long deletion in the initial part of the gene. There is a deletion in the initial part of cox1 which results in it being 24 bp shorter than in M. edulis, M. galloprovincialis and Baltic M. trossulus M genomes. Nad3 gene is three nucleotides longer at 354 bp than in previously described Mytilus genomes. This is the consequence of the mutation T to C in the stop codon, transforming it into a glutamine codon, the following three nucleotides forming the new stop codon. No differences are observed between the Loch Etive and North American F genomes of M. trossulus with respect to usage of initiation and termination codons. However, there are two differences when compared to the F genome of M. edulis, M. galloprovincialis and
Baltic M. trossulus: TAA is used instead of TAG for cox3 and cob, and the start codon GTG used instead of ATG for cob and nad4L, a characteristic of the M genome of M. trossulus. Five termination codons are different in the Loch Etive M genome compared with the M genome of M. edulis, M. galloprovincialis and Baltic M. trossulus: TAA instead of TAG for cox1 and cob, and TAG instead TAA for cox2, nad2 and nad4 gene. All mitochondrial proteins for the genomes sequenced here have been concatenated. This was also done for other published mtDNA genomes. The sequences were then assembled into groups of two sequences and distance comparisons made within and between these groups. The grouping allows calculation of net between group distance values (between subspecies, geographic regions, or genomes) in which gross between group distance is corrected for within group variation. The number of completely sequenced Mytilus genomes is limited but sufficient to allow the construction of groups to two sequences, the minimum number needed to carry out this correction. The groups coincide as far as possible with distinct Mytilus
Table 2 Net genetic distance (K, Ks, and Ka) and Ka/Ks for comparisons made within and between groups of two sequences of concatenated mitochondrial proteins. Superscripts in the first column indicate the sequences used in each comparison and referenced in the footnote. Taxon comparison
Origin of sequences
K
Ks
Ka
Ka/Ks
Within F genome M. trossulusa1, d1 M. edulisc1/M. gallo.b1, c1 M. trossulusa1, d1 versus M. edulisc1/M. gallo.b1
Amer + Eur Eur (Amer + Eur) versus Eur
0.0082(0.0006) 0.0083(0.0009) 0.1763(0.0048)
0.0201(0.0026) 0.0234(0.0027) 0.6770(0.0236)
0.0031(0.0007) 0.0017(0.0005) 0.0267(0.0019)
0.16 0.0726 0.0394
Within M genome M. trossulusd2, d3 M. edulisa2, c2 M. edulisa2, a3 M. trossulusd2, d3 versus M. edulisa2, M. trossulus versus M. edulisa2, a3
Eur Amer + Eur Amer Eur versus (Amer + Eur) Eur versus Amer
0.0155(0.0010) 0.0644(0.0024) 0.0132(0.0011) 0.2393(0.0060) 0.2649(0.0062)
0.0396(0.0031) 0.1889(0.0076) 0.0343(0.0031) 0.7531(0.0274) 0.8346(0.0278)
0.0052(0.0009) 0.0138(0.0014) 0.0039(0.0007) 0.0979(0.0038)) 0.1021(0.0040)
0.1326 0.0730 0.1137 0.1299 0.1223
(Amer + Eur) versus Eur Eur versus (Amer + Eur) Eur versus Amer
0.2657(0.0060) 0.2450(0.0053) 0.2546(0.0070)
0.8717(0.0324) 0.7702(0.0255) 0.8554(0.0265)
0.0962(0.0037) 0.0815(0.0036) 0.0854(0.0038)
0.1104 0.1058 0.0998
F versus M genome M. trossulusa1, d1 M. edulisc1/M. gallo.b1 M. edulisc1/M. gallo.b1
c2
a1—AY823625, a2—AY823623, a3—AY823624, Breton et al. (2006); b1—AY497292, Mizi et al. (2005); c1—DQ198231, c2—DQ198225, Zbawicka et al. (2007); d1—34LE, d2—117LE, d3—149LE, present study; Amer—American sequence, Eur—European sequence. Values in brackets are standard errors (SE).
M. Zbawicka et al. / Gene 456 (2010) 45–53
51
Fig. 4. Neighbor-joining phylogeny for concatenated proteins from whole genomes of Mytilus. MtDNA genomes from the present study are marked in bold. Other genomes: GB1—NC_006161 (Hoffmann1992), GR1—AY497292, GR2—AY363687 (Mizi et al., 2005), AM1—AY823623, AM2—AY823624, AM3—AY823625 (Breton et al., 2006), C—DQ399833 (Venetis et al., 2007), 39mc10—DQ198231, and 87mc10—DQ198225 (Zbawicka et al., 2007). The bars indicate mtDNA lineages.
taxa, thus allowing comparisons to be made between and within subspecies. The net distance values together with within group values are shown in Table 2. The lowest distance values are obtained for both F and M genomes within M. trossulus and M. edulis and within the M. edulis/M. galloprovincialis group as expected. Interspecific values are higher for both genomes, again as expected. Distances between the F and M genomes are similar for M. trossulus and the M. edulis/M. galloprovincialis group. Thus the sequenced M. trossulus genomes from Loch Etive do not exhibit discrepant features leading to unexpected large or small distance values. A neighbor-joining phylogenetic tree for the whole genomes based on the concatenated proteins is shown in Fig. 4. Although this tree does not include a sequence for American M. trossulus which has not yet been fully sequenced, it supports in its other general features the phylogenetic trees of Fig. 2 and with very high bootstrap support values. The Ka/Ks values (Table 2) are all low, characteristic of genomes under purifying selection (see also Supplementary Table 2 where Ka b Ks for all protein genes). Values involving M. trossulus from Loch Etive are not high, in the sense that they do not approach a value of one, thus there is no reason to assume that there has either been any loss of function of the genomes within Loch Etive or of any marked relaxation of purifying selection in this environment. The availability of the sequenced Loch Etive genomes also allows comparison of Ka/ Ks in F and M genomes, comparing M. trossulus with the other two subspecies, although the population provenance does not correspond exactly for the two genomes. The value for the F genome is much lower (0.0.0394, row three) than the M genome (0.1299 and 0.1223, rows seven and eight). This confirms previously reported observations of a greater selective constraint on the F than M genome (Stewart et al., 1995). The nine complete sequences providing data for Table 2 were analysed codon by codon to search for positive and negative selection using three different methods designated SLAC (Single Likelihood Ancestor Counting), FEL (Fixed Effects Likelihood) and REL (Random Effects Likelihood) (Kosakowsky Pond and Frost, 2005). Out of 3846 codons, 6 showed evidence of positive selection and 1216 showed evidence of negative selection by at least one of the methods (significance levels of 0.05 for SLAC and FEL and Bayes factor criterion of 50 for REL). Given the number of codons analysed there is thus clear evidence for negative but not positive selection. The negatively selected sites are not obviously clumped within the coding region. The PARRIS (a Partitioning approach for Robust Inference of Selection) method (Scheffler et al., 2006) confirmed the absence of positive selection. However one site (codon 837 in cox2) had a Bayes factor value exceeding the REL criterion. This site also differentiated the F and M genomes used in the analysis and might be a candidate for analysis in any future studies of adaptive divergence of the M and F genomes.
Of great interest is the question of the divergence time of the mtDNA lineages of Loch Etive and American mussels as this could throw light on the time of invasion of American M. trossulus. Pertinent to this is the evidence that salinity is correlated with distribution of M. trossulus mussels in Europe (Riginos and Cunningham, 2005). In Loch Etive, M. trossulus mussels occur more frequently in lowered salinity conditions (Beaumont et al., 2008). Furthermore juvenile North American M. trossulus mussels appear to be more tolerant to low salinity than M. edulis mussels (Qiu et al., 2002). Thus lower salinity seawater resulting from the melting of glaciers at the end of the last glacial period and more recent lower salinity of river-derived surface water due to melting ice could have favoured colonisation of Loch Etive by M. trossulus. An estimate of the divergence time of North American and Loch Etive mtDNA lineages has been obtained using population expansion theory following a previous application (Śmietanka et al., 2009). In this approach, the mismatch distribution can be used to obtain the value of τ = 2μt, where μ = substitution rate and t is time since expansion. Assuming that expansion of the range of M. trossulus occurred at the end of the most recent glacial period about 10 kya, μ can be estimated. An estimate of τ for Mytilus subspecies from European locations has been made previously for an ND2-COIII mtDNA fragment generating μ estimates for 4-fold degenerate sites of 1 and 2 million years for the F and M genomes respectively (Śmietanka et al., 2009). Employing the same DNA region, and comparing Loch Etive and American M. trossulus genomes, there are 143 4-fold degenerate sites in which there are 4 and 5 substitutions for the F and M genomes respectively. This gives divergence times of (4 / 143) / (2 × 1) = 0.014 and (5 / 143) / (2 × 2) = 0.009 million years for the F and M genomes respectively. Additional estimates made using the concatenated sequences and the divergence point of the M. trossulus–M. edulis/M. galloprovincialis lineage at the time of opening of the Bering Seaway 3.5 MYA (Vermeij 1991) give more ancient divergence values of 0.43 million years and 0.20 million years. The former values are likely to be more reliable as divergence times may be systematically overestimated in molecular clock calculations using more ancient events in calibration (Ho et al., 2005). However all estimates provide good evidence against the possibility of a recent invasion occurring within the last 1–2 thousand years and associated with recent human activity. 3.5. Major noncoding region The major noncoding region of the F genome of Loch Etive M. trossulus is 3063 bp long and is characteristic of the F2 type M. trossulus F genome that varies in length from 3060 to 3063 (Rawson, 2005; Cao et al., 2009). The region is very similar to the CR of the North American F genome of M. trossulus and consists of M-like and F-like
52
M. Zbawicka et al. / Gene 456 (2010) 45–53
3.6. Conclusion
Fig. 5. Dotplot comparison of M mtDNA genomes 117LE and 149LE from Loch Etive. A window of 15 bp and similarity threshold of 65% are used. The lower single headed arrow marks the end of lrrna, the upper single headed arrow marks the start of cob. The double headed arrows mark the position of the duplication.
Analysis of three nuclear DNA diagnostic markers confirms the presence of M. trossulus mussels at Loch Etive in Scotland. The observed higher frequency of multiple heterozygotes for diagnostic alleles at Loch Etive compared with North America or the Baltic, suggests either separate M. trossulus invasions of Loch Etive and the Baltic or that European populations such as that at Loch Etive have acted as a stepping stone for an invasion of the Baltic, accompanied by loss of the native North American mtDNA genomes. Complete sequencing of Loch Etive M. trossulus F and M mtDNA genomes, confirms the presence of DUI, and provides evidence that these genomes have close affinity with those in ancestral American M. trossulus rather than M. trossulus from the Baltic supporting the scenario of direct invasion of Loch Etive from North American populations rather than from the Baltic. These results thus provide the first evidence of ancestral M. trossulus mtDNA genomes in Europe. Estimates of divergence times were made for both genomes and are consistent with invasion of Loch Etive by M. trossulus towards the end of the last glacial period rather than in more recent human history. The similarity of the Loch Etive F and M genomes to the corresponding genomes in North America, despite the duplication of part of the CR in one of the Loch Etive M genomes, does not provide new insights into the mechanism of DUI. Acknowledgements
parts (Rawson, 2005; Breton et al., 2006; Cao et al., 2009) separated by a sequence with weak similarity to the fragment of lrrna and a few tRNA genes, an organisation resulting from duplication and deletion (Cao et al., 2009) rather than from nonhomologous recombination as originally postulated by Rawson (2005). It has been suggested that the M-like part of the North American F genome might be nonfunctional in failing to confer a paternal role (Cao et al., 2009), and the same may be true for the Loch Etive F genome. The CR of the Loch Etive M. trossulus M genome is also similar to the CR described for the North American M genome of M. trossulus (Rawson, 2005; Cao et al., 2009). Its length without duplication is 1063 bp a little shorter than the North American M genome which varies in length from 1067 bp to 1083 bp (Rawson, 2005; Cao et al., 2009). The differences are mainly caused by the different lengths of the homoadenine region. The longer of the two Loch Etive M genomes has a 931 bp long duplication with the break point in the 3′ part of the CR, position 984 in the CR (Fig. 5). The two copies of the duplicated region are not identical but very similar (K = 0.011). This duplication in the CR of the Loch Etive M. trossulus M genome is not observed in the M genome of American M. trossulus (Rawson, 2005; Cao et al., 2009). The CR of the Loch Etive M genome is similar in sequence to that from North American individuals (K = 0.009, SE = 0.003 and K = 0.019, SE = 0.004 for M genomes from Nova Scotia (AY515231) and the Pacific (DQ013366) respectively). The CRs from the two Loch Etive M genomes are also alike (K = 0.014). If the CR of the North American M. trossulus M genome, or part of it, does confer the paternal role on this genome (Cao et al., 2009), then it can be hypothesised that the possession of two similar copies of part of this, as in the one of the Loch Etive M genomes, does not impair this function. The observation of these two variant M genomes in Loch Etive is in line with previous observations of frequent duplications in the CR regions of Baltic mussels (Burzyński et al., 2006; Cao et al., 2009), whereas such variation has not become evident in M. edulis or M. galloprovincialis. Perhaps the M. trossulus nuclear background (for example with respect to mtDNA processing enzymes), or cytonuclear incompatibility in hybrid populations might contribute to mtDNA instability, frequent duplication and the phenomenon of role-reversal (Hoeh et al., 1997). This is speculation, and awaits accumulation of mtDNA genome sequences to establish majority consensus differences between the sequences derived from male and female gametes and having different transmission routes.
This study was partially funded from the SPB 16/PRIMUS/2006/0 and SPB NoE MGE 51/6.PR UE/2007/7. Institute of Oceanology PAS, Sopot, Poland and Swansea University, UK have been members of the EU: GOCE-CT-2004-505403, Network of Excellence on Marine Genomics Europe (MGE) 2004–2008. AB was partially supported by grant no. NN303418336 of the Polish Ministry of Higher Education. We thank Walter Speirs of Muckairn Mussels for supplying samples of mussels from Loch Etive. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.gene.2010.02.009. References Badger, J.H., Olsen, G.J., 1999. CRITICA: coding region identification tool invoking comparative analysis. Mol. Biol. Evol. 16, 512–524. Beaumont, A.R., Hawkins, M.P., Doig, F.L., Davies, I.M., Snow, M., 2008. Three species of Mytilus and their hybrids identified in a Scottish Loch: natives, relicts and invaders? J. Exp. Mar. Biol. Ecol. 367, 100–110. Bierne, N., et al., 2003a. Introgression patterns in the mosaic hybrid zone between Mytilus edulis and M. galloprovincialis. Mol. Ecol. 12, 447–461. Bierne, N., Daguin, C., Bonhomme, F., David, P., Borsa, P., 2003b. Direct selection on allozymes is not required to explain heterogeneity among marker loci across a Mytilus hybrid zone. Mol. Ecol. 12, 2505–2510. Birney, E., Clamp, M., Durbin, R., 2004. GeneWise and Genomewise. Genome Res. 14, 988–995. Bonfield, J.K., Smith, K.F., Staden, R., 1995. A new DNA sequence assembly program. Nucleic Acids Res. 23, 4992–4999. Boore, J.L., Medina, M., Rosenberg, L.A., 2004. Complete sequences of the highly rearranged molluscan mitochondrial genomes of the scaphopod Graptacme eborea and the bivalve Mytilus edulis. Mol. Biol. Evol. 21, 1492–1503. Breton, S., Burger, G., Stewart, D.T., Blier, P.U., 2006. Comparative analysis of gender-associated complete mitochondrial genomes in marine mussels (Mytilus spp.). Genetics 172, 1107–1119. Burzyński, A., Śmietanka, B., 2009. Is interlineage recombination responsible for low divergence of mitochondrial nad3 genes in Mytilus galloprovincialis? Mol. Biol. Evol. 26, 1441–1445. Burzyński, A., Zbawicka, M., Skibinski, D.O.F., Wenne, R., 2003. Evidence for recombination of mtDNA in the marine mussel Mytilus trossulus from the Baltic. Mol. Biol. Evol. 20, 388–392. Burzyński, A., Zbawicka, M., Skibinski, D.O.F., Wenne, R., 2006. Doubly uniparental inheritance is associated with high polymorphism for rearranged and recombinant control region haplotypes in Baltic Mytilus trossulus. Genetics 174, 1081–1094. Cao, L.Q., Kenchington, E., Zouros, E., Rodakis, G.C., 2004. Evidence that the large noncoding sequence is the main control region of maternally and paternally transmitted mitochondrial genomes of the marine mussel (Mytilus spp.). Genetics 167, 835–850.
M. Zbawicka et al. / Gene 456 (2010) 45–53 Cao, L.Q., et al., 2009. The control region of maternally and paternally inherited mitochondrial genomes of three species of the sea mussel genus Mytilus. Genetics 181, 1045–1056. Comesaña, A.S., Toro, J.E., Innes, D.J., Thompson, R.J., 1999. A molecular approach to the ecology of a mussel (Mytilis edulis–M. trossulus) hybrid zone on the east coast of Newfoundland. Can. Mar. Biol. 133, 213–221. Coustau, C., Renaud, F., Delay, B., 1991. Genetic characterisation of the hybridisation between Mytilus edulis and M. galloprovincialis on the Atlantic coast of France. Mar. Biol. 111, 87–93. Delcher, A.L., Harmon, D., Kasif, S., White, O., Salzberg, S.L., 1999. Improved microbial gene identification with GLIMMER. Nucleic Acids Res. 27, 4636–4641. Dias, P.J., Sollelis, L., Cook, E.J., Piertney, S.B., Davies, I.M., Snow, M., 2008. Development of a real-time PCR assay for detection of Mytilus species specific alleles: application to a sampling survey in Scotland. J. Exp. Mar. Biol. Ecol. 367, 253–258. Eddy, S.R., Durbin, R., 1994. RNA sequence analysis using covariance models. Nucleic Acids Res. 22, 2079–2088. Ewing, B., Hillier, L., Wendl, M.C., Green, P., 1998. Base-calling of automated sequencer traces using phred. I. Accuracy assessment. Genome Res. 8, 175–185. Filipowicz, M., Burzyński, A., Śmietanka, B., Wenne, R., 2008. Recombination in mitochondrial DNA of European mussels Mytilus. J. Mol. Evol. 67, 377–388. Gosling, E., 1992. The Mussel Mytilus: Ecology, Physiology, Genetics and Culture. Elsevier, Amsterdam. Heath, D.D., Rawson, P.D., Hilbish, T.J., 1995. PCR-based nuclear markers identify alien blue mussel (Mytilus spp.) genotypes on the west coast of Canada. Can. J. Fish. Aquat. Sci. 52, 2621–2627. Ho, S.Y.W., Phillips, M.J., Cooper, A., Drummond, A.J., 2005. Time dependency of molecular rate estimates and systematic overestimation of recent divergence times. Mol. Biol. Evol. 22, 1561–1568. Hoeh, W.R., Stewart, D.T., Saavedra, C., Sutherland, B.W., Zouros, E., 1997. Phylogenetic evidence for role-reversals of gender-associated mitochondrial DNA in Mytilus (Bivalvia: Mytilidae). Mol. Biol. Evol. 14, 959–967. Hoffmann, R.J., Boore, J.L., Brown, W.M., 1992. A novel mitochondrial genome organization for the blue mussel, Mytilus edulis. Genetics 131, 397–412. Inoue, K., Waite, J.H., Matsuoka, M., Odo, S., Harayama, S., 1995. Interspecific variations in adhesive protein sequences of Mytilus edulis, M. galloprovincialis, and M. trossulus. Biol. Bull. 189, 370–375. Kijewski, T.K., Zbawicka, M., Väinölä, R., Wenne, R., 2006. Introgression and mitochondrial DNA heteroplasmy in the Baltic populations of mussels Mytilus trossulus and M. edulis. Mar. Biol. 149, 1373–1385. Kosakowsky Pond, S.L., Frost, S.D.W., 2005. Not so different after all: a comparison of methods for detecting amino acid sites under selection. Mol. Biol. Evol. 22, 1208–1222. Kumar, S., Tamura, K., Nei, M., 2004. MEGA3: integrated software for molecular evolutionary genetics analysis and sequence alignment. Brief. Bioinform. 5, 150–163. McDonald, J.H., Seed, H.R., Koehn, R.K., 1991. Allozyme and morphometric characters of three species of Mytilus in the northern and southern hemispheres. Mar. Biol. 111, 323–335. Mizi, A., Zouros, E., Moschonas, N., Rodakis, G.C., 2005. The complete maternal and paternal mitochondrial genomes of the Mediterranean mussel Mytilus galloprovincialis: implications for the doubly uniparental inheritance mode of mtDNA. Mol. Biol. Evol. 22, 952–967. Nei, M., Gojobori, T., 1986. Simple methods for estimating the numbers of synonymous and nonsynonymous nucleotide substitutions. Mol. Biol. Evol. 3, 418–426. Qiu, J., Tremblay, R., Bourget, E., 2002. Ontogenetic changes in hyposaline tolerance in the mussels Mytilus edulis and M. trossulus: implications for distribution. Mar. Ecol. Progr. Ser. 228, 143–152.
53
Rawson, P.D., 2005. Nonhomologous recombination between the large unassigned region of the male and female mitochondrial genomes in the mussel, Mytilus trossulus. J. Mol. Evol. 61, 717–732. Rawson, P.D., Hilbish, T.J., 1995. Evolutionary relationships among the male and female mitochondrial DNA lineages in the Mytilus edulis species complex. Mol. Biol. Evol. 12, 893–901. Rice, P., Longden, I., Bleasby, A., 2000. EMBOSS: the European Molecular Biology Open Software Suite. Trends Genet. 16, 276–277. Ridgway, G., Nævdal, G.N., 2004. Genotypes of Mytilus from waters of different salinity around Bergen. Norway Helgol. Mar. Res. 58, 104–109. Riginos, C., Cunningham, C.W., 2005. Local adaptation and species segregation in two mussel (Mytilus edulis × Mytilus trossulus) hybrid zones. Mol. Ecol. 14, 381–400. Riginos, C., Hickerson, M.J., Henzler, C.M., Cunningham, C.W., 2004. Differential patterns of male and female mtDNA exchange across the Atlantic Ocean in the blue mussel, Mytilus edulis. Evolution 58, 2438–2451. Rozas, J., Sánchez-DelBarrio, J.C., Messeguer, X., Rozas, R., 2003. DnaSP, DNA polymorphism analyses by the coalescent and other methods. Bioinformatics. 19, 2496–2497. Saavedra, C., Stewart, D.T., Stanwood, R.R., Zouros, E., 1996. Species-specific segregation of gender-associated mitochondrial DNA types in an area where two mussel species (Mytilus edulis and M. trossulus) hybridize. Genetics 143, 1359–1367. Scheffler, K., Martin, D.P., Seoighe, C., 2006. Robust inference of positive selection from recombining coding sequences. Bioinformatics 20, 2493–2499. Skibinski, D.O.F., Ahmad, M., Beardmore, J.A., 1978. Genetic evidence for naturally occurring hybrids between Mytilus edulis and Mytilus galloprovincialis. Evolution 32, 354–364. Skibinski, D.O.F., Gallagher, C., Beynon, C.M., 1994a. Mitochondrial-DNA inheritance. Nature 368, 817–818. Skibinski, D.O.F., Gallagher, C., Beynon, C.M., 1994b. Sex-limited mitochondrial DNA transmission in the marine mussel Mytilus edulis. Genetics 138, 801–809. Śmietanka, B., Zbawicka, M., Wołowicz, M., Wenne, R., 2004. Mitochondrial DNA lineages in the European populations of mussels (Mytilus spp.). Mar. Biol. 146, 79–92. Śmietanka, B., Burzyński, A., Wenne, R., 2009. Molecular population genetics of male and female mitochondrial genomes in European mussels Mytilus. Mar. Biol. 156, 913–925. Staden, R., Judge, D.P., Bonfield, J.K., 2001. Sequence assembly and finishing methods, In: Baxevanis, A.D., Ouellette, B.F.F. (Eds.), Bioinformatics, Second Edition. Wiley, pp. 303–322. Stewart, D.T., Saavedra, C., Stanwood, R.R., Ball, A.O., Zouros, E., 1995. Male and female mitochondrial lineages in the blue mussel (Mytilus edulis) species group. Mol. Biol. Evol. 12, 735–747. Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F., Higgins, D.G., 1997. The CLUSTAL_X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res. 25, 4876–4882. Venetis, C., Theologidis, I., Zouros, E., Rodakis, G.C., 2007. A mitochondrial genome with a reversed transmission route in the Mediterranean mussel Mytilus galloprovincialis. Gene 406, 79–90. Vermeij, G.J., 1991. The anatomy of an invasion—the trans-arctic interchange. Paleobiology 17, 281–307. Wares, J.P., Cunningham, C.W., 2001. Phylogeography and historical ecology of the North Atlantic intertidal. Evolution 55, 2455–2469. Zbawicka, M., Burzyński, A., Wenne, R., 2007. Complete sequences of mitochondrial genomes from the Baltic mussel Mytilus trossulus. Gene 406, 191–198. Zouros, E., Ball, A.O., Saavedra, C., Freeman, K.R., 1994a. Mitochondrial DNA inheritance— reply. Nature 368, 818. Zouros, E., Oberhauser Ball, A.O., Saavedra, C., Freeman, K.R., 1994b. An unusual type of mitochondrial DNA inheritance in the blue mussel Mytilus. Proc. Natl. Acad. Sci. USA 91, 7463–7467.
