- Email: [email protected]
Phylogeographic structure and demographic patterns of brown trout in North-West Africa
Molecular Phylogenetics and Evolution 61 (2011) 203–211
Contents lists available at ScienceDirect
Molecular Phylogenetics and Evolution journal homepage: www.elsevier.com/locate/ympev
Phylogeographic structure and demographic patterns of brown trout in North-West Africa Aleš Snoj a, Saša Maric´ b, Simona Sušnik Bajec a, Patrick Berrebi c, Said Janjani d, Johannes Schöffmann e,⇑ a
University of Ljubljana, Biotechnical Faculty, Department of Animal Science, Groblje 3, SI-1230 Domzˇale, Slovenia University of Belgrade, Faculty of Biology, Institute of Zoology, Studentski trg 16, 11001 Belgrade, Serbia c Université Montpellier II, Institut des Sciences de l’Evolution, UMR 5554 UM2-CNRS-IRD, CC 065, Place E. Bataillon, 34095 Montpellier Cedex 5, France d Faculté des Sciences Semlalia de Marrakech, Hydrobiology Laboratory, Avenue Prince Moulay Abdellah BP, 2390 Marrakech, Morocco e Lastenstrasse 25, 9300 St. Veit/Glan, Austria b
a r t i c l e
i n f o
Article history: Received 10 March 2011 Revised 13 May 2011 Accepted 14 May 2011 Available online 30 May 2011 Keywords: Salmo trutta Atlantic lineage Mitochondrial DNA Microsatellite DNA Morocco Atlas Mountains
a b s t r a c t The objectives of the study were to determine the phylogeographic structure of brown trout (Salmo trutta) in Morocco, elucidate their colonization patterns in North-West Africa and identify the mtDNA lineages involved in this process. We also aimed to resolve whether certain brown trout entities are also genetically distinct. Sixty-two brown trout from eleven locations across the Mediterranean and the Atlantic drainages in Morocco were surveyed using sequence analysis of the mtDNA control region and nuclear gene LDH, and by genotyping twelve microsatellite loci. Our study confirms that in Morocco both the Atlantic and Mediterranean basins are populated by Atlantic mtDNA lineage brown trout only, demonstrating that the Atlantic lineage (especially its southern clade) invaded initially not only the western part of the Mediterranean basin in Morocco but also expanded deep into the central area. Atlantic haplotypes identified here sort into three distinct groups suggesting Morocco was colonized in at least three successive waves (1.2, 0.4 and 0.2–0.1 MY ago). This notion becomes more pronounced with the finding of a distinct haplotype in the Dades river system, whose origin appears to coalesce with the nascent stage of the basal mtDNA evolutionary lineages of brown trout. According to our results, Salmo akairos, Salmo pellegrini and ‘‘green trout’’ from Lake Isli do not exhibited any character states that distinctively separate them from the other brown trout populations studied. Therefore, their status as distinct species was not confirmed. Ó 2011 Elsevier Inc. All rights reserved.
1. Introduction The brown trout (Salmo trutta, in this study considered as a single species except for Salmo salar, Salmo obtusirostris and Salmo ohridanus) is the most widely distributed Palaearctic freshwater fish. It is native across Europe, including Iceland, and parts of western and central Asia, with a natural range that extends southwards even to the Atlas Mountains of North-West Africa (Pellegrin, 1921, 1924a; Azeroual, 2003). Up until the Miocene, the inland water ichthyofauna of North-West Africa had a tropical character, after which it became Palaearctic in composition (Greenwood, 1974). Due to unsuitable tropical warm conditions in seas surrounding the Iberian Peninsula and northern Africa (Pérès, 1985), it is believed that brown trout did not arrive in this region before the onset of the Pleistocene cooling. During the warm ⇑ Corresponding author. Fax: +43 4212 214214. E-mail addresses: [email protected] (A. Snoj), [email protected] (S. Maric´), [email protected] (P. Berrebi), [email protected] (S. Janjani), [email protected] (J. Schöffmann). 1055-7903/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.ympev.2011.05.011
interglacial periods, brown trout in North-West Africa largely became extinct with some populations retreating to higher altitudes, while during cold periods, the distribution range advanced southwards. Thus brown trout from more northerly locations expanded their range along the Atlantic coast as far south as the Draa river basin at a latitude of about 28°N, and along the Mediterranean zone eastwards to Algeria and even Sicily (Schöffmann et al., 2007; Cortey et al., 2009). Due to cycling of glacial periods, the southward expansion of brown trout into the coastal rivers of the Iberian Peninsula and North-West Africa appears to have repeated several times (Cortey et al., 2009). Brown trout in Morroco of both the Atlantic and the Mediterranean basins have been assigned, on the basis of a limited sampleset (Bernatchez, 2001; Suarez et al., 2001), to the southern clade of the Atlantic mtDNA lineage, largely present in the Iberian peninsula (Cortey et al., 2009). This assignment is centered on the concept of major evolutionary lineages based upon mtDNA analysis, e.g., Atlantic (AT), Danube (DA), Mediterranean (ME), Adriatic (AD), marmoratus (MA) (Bernatchez et al., 1992) and Duero (DU) (Suarez et al., 2001).
204
A. Snoj et al. / Molecular Phylogenetics and Evolution 61 (2011) 203–211
In Morocco, autochthonous brown trout presently occur in the High Atlas above an altitude of 1500 m, in the Middle Atlas above 1200 m, and in the Mediterranean tributaries of the Rif Mountains, where they reside at an altitude of about 400 m (Azeroual, 2003; unpublished data). Yet in the middle years of the last century, the lower altitude limit of brown trout residence in the Middle Atlas was found to be about 900 m (Kerans, 1962). In the western High Atlas Mountains, the southernmost natural distribution of brown trout is nowadays confined to a latitude of 30° 500 N (River Nfiss of the Tensift basin), and 31° 020 (Lake Ifni of the Sous basin). Curiously, one brown trout population inhabits the headwaters of a tributary of the River Ziz, which rises on the south slope of the High Atlas and drains into the Sahara Desert (Ilahiane, 1996; Hammada, 2007). Isolated lake-dwelling brown trout still exist in two lakes of the High Atlas: the recently described ‘‘dwarf trout’’, Salmo akairos of Lake Ifni (Delling and Doadrio, 2005) and the ‘‘green trout’’ of Lake Isli (Vivier, 1948; Mouslih, 1987). Another lake-dwelling trout, also described as a distinct species Salmo pallaryi, was reported from the high mountain Lake Aguelmame Sidi Ali in the Middle Atlas (Pellegrin, 1924a,b) and from the nearby lake Tiguelmamine N’Ait Mahi (Vivier, 1948). However, this trout disappeared soon after the introduction of carp (Cyprinus carpio) in 1934 (Mouslih, 1987; Schöffmann, 1993). Salmo macrostigma (Duméril, 1858), first described from Oued el Abaïch in north-eastern Algeria, was described also in Morocco from the River Tigrigra of the Sebou basin, the upper River Moulouya with its tributaries Ansegmir and Berrem, and the River Melloulou, a tributary to the lower Moulouya river system (Pellegrin, 1924a; Joleaud, 1938). The name S. macrostigma has been applied also to several periMediterranean trout populations (Bianco, 1994; Kottelat, 1997) though without any demonstrative analysis. Another morphologically distinct species, Salmo pellegrini, was described from the stream Ourika, south of the city of Marrakech (Werner, 1931) and characterized by an appearance intermediate between S. pallaryi and S. macrostigma. In the course of the twentieth century, many native brown trout populations in Morocco became drastically reduced or extirpated due to human activities. The predominant threats to the remaining populations are increased water use, environmental degradation due to deforestation and overgrazing, road and dam constructions, water pollution and global warming. Some locations have been stocked with a domestic brown trout strain originating from River Oum Er Rbia upstream wild brown trout at the state hatchery of Azrou (Centre National d’Hydrobiologie et de Pisciculture, CNHP, unpublished report). The primary objectives of this study are to determine phylogeography of brown trout in Morocco, elucidate their colonization patterns and identify the lineages involved. In this context we examine the proposed hypothesis of colonization of Sicily with brown trout via North-West Africa.
A further aim is to genetically profile S. pellegrini (River Ourika), S. akairos (Lake Ifni) and ‘‘green trout’’ (Lake Isli) and determine if they represent genetically distinct entities. 2. Material and methods 2.1. Material Between 1998 and 2010, 62 brown trout from eleven locations across the Mediterranean and the Atlantic river basins in Morocco were fin-clipped (Table 1, Fig. 1). Allegedly unstocked locations were selected based upon information published in the literature (Schöffmann, 1993; Kottelat, 1997; Bernatchez, 2001; Delling and Doadrio, 2005) and the authors’ personal observations. The highly limited range of brown trout in Morocco—it is confined mainly to remote and inaccessible small streams in the Atlas and Riff Mountains—made sampling extremely difficult and, crucially, contributed to a relatively small number of sampling sites. Total DNA was isolated from fin tissue preserved in 96% ethanol following the protocol of Medrano et al. (1990). Brown trout samples from Sicily (see Schöffmann et al., 2007) were reanalyzed in order to obtain comparable sequences of mtDNA CR. 2.2. Mitochondrial DNA amplification, sequencing and data analysis Mitochondrial DNA CR was amplified by polymerase chain reaction (PCR) using primers 28RIBa (Sušnik et al., 2001) and HN20 (Bernatchez and Danzmann, 1993). Each 30 ll reaction included 1 lM of each primer, 0.2 lM dNTP, 1.5 lM MgCl2, 1 PCR buffer, 1 U Taq polymerase (Applied Biosystems, Foster City, CA, USA) and 50 ng of genomic DNA. Amplified DNA fragments were run on a 1.5% gel and subsequently dissected and isolated from the gel using the QIAEX II gel Extraction Kit (QIAGEN, Hilden, Germany). The conditions for PCR were initial denaturation (95 °C, 3 min) followed by 30 cycles of strand denaturation (94 °C, 45 s), primer annealing (52 °C, 45 s) and DNA extension (72 °C, 2 min). All PCR amplifications were performed in a programmable thermocycler GeneAmpÒ PCR System 9700 (Applied Biosystems). All sequencing reactions were prepared using a BigDye Terminator Ready Reaction Mix (Applied Biosystems) according to the manufacturer’s recommendations. The control region fragment was sequenced from both directions using PCR primers. Termination PCRs were performed in a programmable thermocycler under the following conditions: 10 s denaturation at 96 °C, 5 s annealing at 50 °C and 4 min extension at 60 °C, repeated for 30 cycles. The amplified, fluorescently labeled and terminated DNA was saltprecipitated and analysed with an ABI Prism 3130 xl automated sequencer.
Table 1 Sample sites and sizes (N), mtDNA haplotypes, and corresponding river system and drainages (AT, Atlantic; ME, Mediterranean). Sample numbers correspond to those in Fig. 1. For details about the samples from Sicily, see Schöffmann et al. (2007). Sample No.
Location
N
mtDNA haplotype
River system/basin
Drainage
1 2 3 4 5 6 7 8 9 10 11
Lake Isli Melloul Ourika Ait Mizane Lake Ifni M-Goun Dades Zaouia-Sidi-Hamza Berrem Kanar Adelma
4 4 5 12 7 2 12 5 4 3 4
ATM2 ATM6 ATM4 ATM1, ATcs25 ATM3 Dades Dades ATM7 ATM4 ATcs33 ATM5
Oued Melloul ? Oum Er Rbia Oum Er Rbia Tensift Reghaya ? Tensift Tifnout ? l’oued Souss Draa Draa Ziz Moulouya Medit. Sea basin Medit. Sea basin
AT AT AT AT AT AT AT AT ME ME ME
A. Snoj et al. / Molecular Phylogenetics and Evolution 61 (2011) 203–211
205
Fig. 1. Map of sampling sites. The numbers correspond to those in Table 1.
Sequences of 980 bp of mtDNA CR were aligned using computer program ClustalX (Thompson et al., 1997). To assign haplotypes of individuals to the mtDNA lineages previously identified for brown trout, data were compared to numerous haplotype sequences of brown trout found across the Iberian Peninsula and Mediterranean drainage (against all known haplotypes from the AT, DU, ME, MA and AD lineages and some from the DA lineage). The GenBank accession numbers of all the mtDNA CR sequences used in this study are reported in Table 2. Aligned haplotypes were imported into the program PAUP Version 4.0b10 (Swofford, 2000) for phylogenetic analysis. Maximum parsimony (MP) and maximum likelihood (ML) were used for phylogenetic reconstruction. For MP, insertions or deletions (indels) were included as a fifth character. A heuristic search (10 replicates) with Tree Bisection Reconnection (TBR) branch-swapping was employed to find the most parsimonious trees. For ML, a sequence evolution model was first chosen using the program Modeltest Version 3.7 (Posada and Crandall, 1998) incorporated into PAUP. After a model was chosen, a heuristic search (10 replicates) was used to estimate the most likely topology. Support values for the nodes were obtained with 1000 bootstrap replicates for MP or ML analysis, while the fast stepwise addition method was used for ML. In addition, the genealogical relation of haplotypes was depicted also using a 95% statistical parsimony network constructed from mtDNA CR sequences using program TCS 1.3 (Clement et al., 2000).
the single polymorphic site. To determine the state of this site, a 428 bp fragment of the LDH gene was amplified using primers Ldhxon3F and Ldhxon4F following the conditions reported in McMeel et al. (2001), and sequenced. Sequencing reactions were prepared as described above, using the Ldhxon3F primer. All the LDH sequences were deposited in GeneBank under Acc. Nos. JF297983–JF297990. 2.4. Microsatellite amplification and data analysis Twelve microsatellite loci were chosen for analysis using primers for PCR amplification and multiplex PCR conditions as described in Lerceteau-Köhler and Weiss (2006). Aliquots of amplified fluorescently labeled DNA were mixed with formamide and GENESCAN500 ROX Size Standard (Applied Biosystems) and genotyped on the ABI Prism 3130 xl with GeneMapperÒ Software v4.0 (Applied Biosystems). Genetic relationships among the populations and the individuals were estimated with factorial correspondence analysis (FCA) using GENETIX 4.04 (Belkhir et al., 2004). This analysis permits the plotting of each individual based on its complete genotype. The inertia values (i.e., the proportion of the total information contained by an axis, expressed as a percentage) along each axis were shown to be equivalent to linear combinations of the monolocus fixation index (Fst) values (Guinand, 1996). Pairwise Fst values were calculated in FSTAT (Goudet, 2001). 3. Results
2.3. LDH gene analysis 3.1. Mitochondrial DNA analysis To check for the presence of northern European hatchery strain of Atlantic lineage in the sampled populations, every individual was tested for the LDH 90 allele (McMeel et al., 2001), which is diagnostic for this strain and can be specifically determined upon
3.1.1. Haplotype determination DNA alignment of the sequences from the Moroccan samples collapsed into ten haplotypes (Table 1), eight of which had not
206
A. Snoj et al. / Molecular Phylogenetics and Evolution 61 (2011) 203–211
Table 2 GenBank accession numbers of the mtDNA CR sequences used in this study. Asterisks denote new sequences. Haplotype
Acc. number
Haplotype
Acc. number
JF297978 JF297979 JF297980 JF297975 JF297977 JF297976 JF297982 JF297981 JF297974 AF273086 AF273087 AF274574 AF274575 AF274576 AF274577 AY836327 AY836328 EF530476 EF530477 EF530478 EF530479 EF530480 EF530481 EF530483 EF530484 EF530485 EF530486 EF530487 EF530488 EF530489 EF530490 EF530491 EF530492 EF530493 EF530494 EF530495 EF530496 EF530497 EF530498 EF530499 EF530500 EF530501 EF530502 EF530503 EF530504 EF530505 EF530506 EF530507 EF530508 EF530509 EF530510 EF530511
ATcs52 AT11a AT11b DUcs1 DUcs2 DUcs3 DUcs4 DUcs5 DUcs6 DUcs7 DUcs8 DUcs9 DUcs10 DUcs11 DUcs12 DUcs13 DUcs14 DUcs15 DUcs16 DUcs17 DUcs18 DUcs19 DUcs20 DUcs21 DUcs22 DUcs23 DA1a DA3 DA9 DA22 DA24 ADcs1 ADcs2 ADcs4 ADcs5 ADcs8 ADcs9 ADcs14 ADcs15 ADcs19 ADcs20 haplo12 ADZ1 MEcs1 MEcs3 MEcs8 MEcs9 MAcs1 MA2a MA2b Salmo ohridanus Salmo salar
EF530512 AY185578 AY185579 EF530513 EF530514 EF530515 EF530516 EF530517 EF530518 EF530519 EF530520 EF530521 EF530522 EF530523 EF530524 EF530525 EF530526 EF530527 EF530528 EF530529 EF530530 EF530531 EF530532 EF530533 EF530534 EF530535 AY185568 AY185571 AY185572 AY185573 AY185576 AY836350 AY836331 AY836333 AY836334 AY836337 AY836338 AY836343 AY836344 AY836348 AY836349 AY926570 DQ381565 AY836350 AY836352 AY836357 AY836358 AY836365 DQ841189 DQ841190 AY926564 AF133701
AtM1 AtM2 AtM3 AtM4 AtM5 AtM6 AtM7 Dades AtSic ATcs1 ATcs2 ATcs3 ATcs4 ATcs5 ATcs6 ATcs11 ATcs12 ATcs14 ATcs15 ATcs16 ATcs17 ATcs18 ATcs19 ATcs21 ATcs22 ATcs23 ATcs24 ATcs25 ATcs26 ATcs27 ATcs28 ATcs29 ATcs30 ATcs31 ATcs32 ATcs33 ATcs34 ATcs35 ATcs36 ATcs37 ATcs38 ATcs39 ATcs41 ATcs42 ATcs43 ATcs45 ATcs46 ATcs47 ATcs48 ATcs49 ATcs50 ATcs51
yet been reported, while two (ATcs25 and ATcs33) had been previously described for the Iberian Peninsula. All the haplotypes belonged to the AT phylogenetic lineage, except for haplotype ‘‘Dades’’ (see below). This haplotype was found only in the Dades river system (Draa basin), where it appeared to be fixed. The other haplotypes were also found to be fixed in particular sampling locations, with the exception of Ait Mizane, where two haplotypes were found (Table 1). All haplotypes appeared to be specifically linked to sampling location except for ATM4, which was bound to brown trout from Berrem (Moulouya river system, Mediterranean basin) and Ourika (Tensift river system, Atlantic basin, which was expected to harbor S. pellegrini (Werner, 1931)). Sequences of brown trout from Sicily revealed a single and unique haplotype, ATSic.
3.1.2. Phylogenetic trees Phylogenetic tree reconstruction sorted all the haplotypes found in Morocco and Sicily into the AT clade, with the exception of haplotype Dades, whose phylogenetic resolution was not well resolved: on the ML tree (Fig. 2), this haplotype formed a polytomy with the Danubian and Atlantic–Duero clades, while on the MP tree it was basal to all Danubian and Atlantic haplotypes (100% bootstrap support; data not shown). The other haplotypes found in Morocco were more consistently distributed on the trees, appearing in three groups. The first group was represented by haplotypes ATM1 and ATM6, found to be closely related and forming a basal clade vs. the entire Atlantic-Duero clade. The second group (ATM3/4/5, ATcs33 and ATSic) assorted with haplotypes previously found in the Iberian Peninsula (i.e., clade AT 3-3 or so-called ‘‘southern At-clade’’; Cortey et al., 2009), and the third group consisted of haplotypes ATM2 and ATM7, which were associated with clade AT 3-2 (Cortey et al., 2009), found across the entire Atlantic basin. 3.1.3. Haplotype network The initial designed haplotype network—based on 116 haplotypes and representing the majority of S. trutta mtDNA variants— revealed a very tangled and looped structure. Following removal of approximately half of the uninformative haplotypes (either on tip positions or distant from Moroccan haplotypes or both), the resolution of the network improved considerably without changing the basic haplotype network organization, and displayed more clearly an interesting relationship among Moroccan, south Atlantic and Mediterranean haplotypes. This network (Fig. 3) clearly showed that all Moroccan haplotypes, with the exception of ‘‘Dades’’, endemic to the Draa basin, fell into the Atlantic haplotype cluster. Most of the Moroccan haplotypes were connected closely to haplotypes previously described from the Iberian Peninsula and were distributed among two groups with the central haplotypes ATcs33 and ATcs1 representing, respectively, plesiomorphic haplotypes for the so-called southern (AT 3-3) and northern Atlantic (AT 3-1, Cortey et al., 2009) clades. In the southern clade, haplotype ATcs33 was found to be connected to its recent Moroccan derivates with single-mutation branches including the haplotype ATSic, while in the northern clade, Moroccan haplotypes were more distant (four or six mutational steps) from the central one (ATsc1). Haplotypes ATM2 and ATM7 are derivates of ATsc25 (one or two mutation steps away) all of which belong to the group that was previously found in both the southern and northern Atlantic basins (Cortey et al., 2009). The haplotype ‘‘Dades’’ was found to be equally separated (12 mutational steps) from the Atlantic, Duero and Adriatic lineages. The haplotype distribution did not seem to follow any obvious geographical pattern. 3.2. LDH results Alignment of the obtained DNA sequences revealed that all individuals tested had alleles at the LDH locus other than LDH 90. In addition to the unique mutation site that specifically distinguishes between 90 and non-90 alleles (see McMeel et al., 2001), several other polymorphic sites were found, permitting construction of a Bayesian tree, which showed no phylogeographic or population organization (data not shown). 3.3. Microsatellites data analysis The FCA diagram (Fig. 4) indicated three main well differentiated groups that included brown trout collected in (1) Dades, (2) Kanar and (3) the remaining sampling sites.
A. Snoj et al. / Molecular Phylogenetics and Evolution 61 (2011) 203–211
207
Fig. 2. Maximum likelihood tree (HKY + I + G model, transition: transversion 2.3540; proportion of invariable sites (I) 0.6675; gamma distribution shape parameter 0.9157) for genus Salmo based on 980 bp of the mtDNA control region. Haplotypes representing all S. trutta mtDNA lineages were included in the analysis, the ones characteristic for trout in Morocco are shown in bold. Bootstrap support values refer to maximum likelihood (top) and maximum parsimony (bottom). Values <50 are marked with ‘‘/’’ or are not marked when there was no value above 50 in any of the analyses.
208
A. Snoj et al. / Molecular Phylogenetics and Evolution 61 (2011) 203–211
Fig. 3. Network relating haplotypes found in Morocco (in bold) with previously published haplotypes of other S. trutta phylogenetic lineages. Ellipses represent haplotypes; solid lines connect each haplotype (regardless of length) and represent single mutation events. Small black circles represent missing or theoretical haplotypes.
Fig. 4. Three-dimensional cluster analysis for all trout specimens from Morocco based on 12 microsatellite loci.
There were no groupings following specific geographic patterns (e.g., being associated with the Atlantic or Mediterranean drainage) nor corresponding with any mtDNA haplotype clusters. High pair-wise Fst values (Table 3) indicated considerable genetic differentiation corresponding to ancient isolation between the populations, particularly in relation to Dades (0.52; SD = ±0.08) and Kanar (0.57; SD = ±0.06).
4. Discussion 4.1. Dominant mtDNA lineages in Morocco Our study confirms that in Morocco, both the Atlantic and Mediterranean basins are populated by Atlantic mtDNA lineage brown trout previously described by Bernatchez (2001). In addition, the
209
A. Snoj et al. / Molecular Phylogenetics and Evolution 61 (2011) 203–211 Table 3 Fst values among the populations sampled in Morocco.
Lake Isli Melloul Ourika Ait Mizane Lake Ifni M-Goun Dades Zaouia-Sidi-Hamza Berrem Kanar Adelma
Lake Isli
Melloul
Ourika
Ait Mizane
Lake Ifni
M-Goun
Dades
Zaouia-Sidi-Hamza
Berrem
Kanar
Adelma
0.0000
0.4553 0.0000
0.6682 0.5099 0.0000
0.4700 0.2989 0.5261 0.0000
0.5768 0.4559 0.5040 0.5077 0.0000
0.5486 0.3904 0.6131 0.4587 0.4578 0.0000
0.5569 0.4553 0.5887 0.5102 0.5540 0.1402 0.0000
0.4778 0.2909 0.5277 0.2272 0.4852 0.4033 0.4609 0.0000
0.5743 0.3163 0.4931 0.3617 0.5384 0.4769 0.5116 0.2982 0.0000
0.6579 0.5274 0.6880 0.5750 0.5635 0.5774 0.5800 0.5209 0.6335 0.0000
0.4438 0.3453 0.5089 0.4546 0.3415 0.3621 0.4563 0.3667 0.4621 0.5075 0.0000
Atlantic haplotypes identified here sort into three distinct groups suggesting a heterogeneous demographic history of brown trout in Morocco. This notion becomes more pronounced with the finding of a distinct haplotype—Dades—whose origin appears to coalesce with the nascent stage of the basal mtDNA evolutionary lineages of brown trout. Of the ten haplotypes found in Morocco, eight are strictly endemic, while the other two (ATcs25 and 33) have been reported previously for the Iberian Peninsula. Given the prevalence in Morocco of endemic haplotypes and absence of the LDH 90 allele and haplotypes ATsc1, 2, 3 and 4, denoting the most widespread European domestic strain of brown trout, one can conclude that brown trout demography in north-west Africa, at least in the sites sampled, has not been affected by introduction of commercial domestic brown trout from Europe. According to unpublished information provided by Azrou fish farm (CNHP) on its stocking activity during the last 10 years, the brown trout strain it has used originates from the Oum Er Rbia river population, which bears haplotype ATM6. This domestic strain should therefore have introduced this haplotype wherever the stocking has been successful (genetic participation of introduced fish in the natural population). However, since this haplotype was not detected in any population analysed, we can conclude that the genetic structure described for each sampling location is entirely natural. 4.2. Population structure according to nuclear markers To detect signals of present or recent gene flow between populations, microsatellites were also included in the analyses. There was virtually no correlation between the mtDNA and microsatellite structures of brown trout in Morocco, apart from confirmation of long-term isolation of the Dades population. According to FCA analysis, Moroccan brown trout appear to be highly structured, with no overlapping populations. However, there is no similarity in the nuclear and mitochondrial geographic organizations. Mitochondrial DNA, because of clonal transmission, is a more reliable tool for reconstruction of ancient phylogenies than are rapidly evolving and recombining nuclear microsatellites. We therefore consider that the origin of the present Moroccan populations (affected by several migrations at several periods contributing to several lineages, a long period of isolation and probably a high rate of population extinction and recolonization) cannot be properly described by microsatellites, also partly because of possible genetic saturation leading to homoplasy, which is here very likely reflected in the distorted distribution of the samples plotted on the FCA graph (Fig. 4). 4.3. Origins of brown trout inhabiting Mediterranean drainages The present study suggests that no haplotype other than the Atlantic ones exists in the Mediterranean drainages in Morocco.
In Kanar, a single haplotype ATcs33 was observed that is reported to exist also in the Atlantic basin of the Iberian Peninsula. Despite reports of repeated stocking (I. Doadrio, personal communication), no such evidence was found from the present analysis. Although stocking activities appear not to have happened in the Morocco brown trout populations studied, the demography presents a different situation for the Berrem population, where the only haplotype detected was ATM4 (southern clade), which was detected also in Ourika (Tensift river basin, Atlantic). These two locations are geographically very remote and even belong to different basins, so it is highly unlikely that fish have dispersed naturally from one location to the other via inland migration (e. g., river capture or similar) in historical times. Recent stocking also seems an unlikely explanation as the two populations differ considerably on the basis of microsatellites (pair-wise Fst = 0.5). Moreover, Ourika is considered to be populated by distinct S. pellegrini. It is possible though that this haplotype reached both locations by the same lineage colonization that took place in both the Atlantic and Mediterranean basins, at a particular stage in the multi-wave establishment (see below) of the present Moroccan ichthyofauna, and that the presence of the same haplotype in both locations is a consequence of an ancestral polymorphism. The southern Atlantic clade is the only one recorded for the Mediterranean basin in Morocco (Fig. 1, sample sites 9–11) indicating that the brown trout population here is homogeneous, probably a consequence of a single wave of migration into this area. Further evidence of dispersal of the southern Atlantic lineage into the Mediterranean basin is seen in the presence of haplotype ATSic, which belongs to the southern Atlantic clade and has been found to occur naturally in Sicily (Schöffmann et al., 2007). This is an important observation and clearly demonstrates that the Atlantic lineage (especially its southern clade) invaded initially not only the western part of the Mediterranean basin in Morocco but also expanded deep into the central area. However, the same colonization wave appears to have reached also more southern locations in the Atlantic basin in Morocco, with the presence of S. pellegrini in River Tensift and of S. akairos in Lake Ifni. The evolution of two Atlantic populations into these two distinct species (at least at the morphological level) suggests an ancient invasion locally maintained and protected by strong isolation, permitting local adaptation and differentiation. 4.4. History of establishment of brown trout populations in Morocco: mtDNA The ‘‘Moroccan haplotypes’’ of southern and northern clades differ considerably with respect to the number of mutational steps separating them from their central haplotype. After transforming these mutational steps into molecular clock estimates (1% per 1 million years; Bernatchez, 2001), the three major groups of Moroccan haplotypes seem to have evolved over three different time periods: 400,000 (ATM1, ATM6), 200,000 (ATsc25, ATM2
210
A. Snoj et al. / Molecular Phylogenetics and Evolution 61 (2011) 203–211
and ATM7) and 100,000 (ATcs33, ATM3, ATM4 and ATM5) years ago. However, the ‘‘Dades’’ haplotype was difficult to place in any of the known mtDNA lineages. Its plesiomorphic state with respect to the Atlantic and the Danubian lineages, rather deep position and long branch on the phylogenetic tree (Fig. 2), and deep and equal divergence from the Atlantic, Adriatic and Duero lineages (12 mutational steps away from each; Fig. 3) demonstrate a long-lasting independent evolution of this haplotype. These observations suggest its existence and arrival at approximately the same time as the appearance of the major lineages of brown trout, or possibly even earlier, and which based upon the molecular clock took place some 1.2 million years ago. The Dades population can justifiably be considered a relict of the first established brown trout lineage in Morocco and perhaps a representative of the predifferentiated ancestor of all trout lineages. The results obtained in the present study suggest that the colonization of brown trout in Morocco took place in several waves, as has been hypothesized previously (Cortey et al., 2009). Based upon our results, Morocco was colonized in at least three successive waves (1.2, 0.4 and 0.2–0.1 MY ago), leading to a specific genetic signature of Moroccan brown trout comprising at least three unrelated groups of mtDNA haplotypes. Generally speaking, there is little phylogeographic structure found in patterns of haplotype distribution in Morocco. For instance, the Tensift River harbors alternatively southern clade haplotypes in its Ourika tributary (sampling location 3; haplotype ATM4 corresponding to S. pellegrini) and northern clade haplotype in its Ait Mizane tributary (sampling location 4; haplotype ATM1, as well as the ubiquitous Atlantic haplotype ATsc25). This incongruence is probably due to stochastic effects of turbulent climatic and geo-morphological events that took place during the repeated cool and warm Pleistocene periods, causing independent local extinctions and permitting foundation of new populations during the next wave of invasion. Such intra-basin differentiations demonstrate very limited dispersal of brown trout populations, even between close tributaries, most probably due to high water temperatures in the plains preventing downstream trout migration. Of the Moroccan brown trout that are believed to represent independent evolutionary entities, three of them were included in our study: S. pellegrini (River Ourika; Werner, 1931), S. akairos (Lake Ifni; Delling and Doadrio, 2005) and ‘‘green trout’’ (Lake Isli; Vivier, 1948; Mouslih, 1987). According to our results, none of these three populations exhibited any character states that would distinctively separate them from the other studied brown trout populations. Therefore, their status as distinct species was not confirmed at the molecular level. Nevertheless, in the extreme conditions encountered by brown trout in the periphery of the species range, special adaptation to specific conditions, together with long isolation in drainages, in some cases even without a marine estuary, has led to strong morphological differentiation, seen in for example the ‘‘green’’ or ‘‘dwarf’’ phenotypes, which do not necessarily correspond to distinct genetic signatures. In contrast to this observation, we identified a brown trout population from the Dades river system that exhibits a distinctive external appearance (see graphic abstract) and a highly specific genetic make-up reflected by both mitochondrial and microsatellite DNA. Given the estimated time of its origin, probably prior to the time of the constitution of the S. trutta complex we know today, this population deserves special attention in terms of its classification status. Due to the considerable genetic diversity reflected by several endemic mtDNA haplotypes, which appear unique to almost each investigated location, and the preserved integrity of indigenous populations, brown trout in Morocco should be attached a special importance in the frame of the S. trutta complex. Many threats, including climate warming and environmental degradation due
to the modernization of Morocco, seriously endanger these special entities of brown trout. These concepts have to be considered by local managers in order to enhance the survival of brown trout in Morocco through appropriate management strategies and careful monitoring of stocking activities, particularly by prevention of introduction of non-native strains and long-distance translocation of feral specimens. Only through such careful management can the preservation of this unique genetic pool of brown trout be ensured. Acknowledgments We would like to thank the Berber people of the Atlas Mountains and G. Pustovrh for helping us with the collecting of samples. Special thanks go to I. Wilson for editing and correcting the manuscript and B. Delling for useful comments. This work was supported by the Slovenian Research Agency. References Azeroual, A., 2003. Monographie des Poissons des Eaux Continentales du Maroc: Systématique, Distribution et Ecologie. Université Mohammed V – Agdal. PhD Thesis, Faculté des Sciences, Rabat University. Belkhir, K., Borsa, P., Chikhi, L., Raufaste, N., Bonhomme, F., 1996–2004. GENETIX 4.04, Logiciel sous Windows TM Pour la Genetique des Populations. Laboratoire Genome, Populations, Interactions, CNRS UMR 5000, Universite de Montpellier II, Montpellier, France. Bernatchez, L., 2001. The evolutionary history of brown trout (Salmo trutta L.) inferred from phylogeographic, nested clade, and mismatch analyses of mitochondrial DNA variation. Evolution 55, 351–379. Bernatchez, L., Danzmann, R.G., 1993. Congruence in control-region sequence and restriction-site variation in mitochondrial DNA of brook charr (Salvelinus fontinalis Mitchill). Mol. Biol. Evol. 10, 1002–1014. Bernatchez, L., Guyomard, R., Bonhomme, F., 1992. DNA sequence variation of the mitochondrial control region among geographically and morphologically remote European brown trout Salmo trutta populations. Mol. Ecol. 1, 161–173. Bianco, P.G., 1994. L’ittiofauna continentale dell’Appennino umbro-marchigiano, barriera semipermeabile allo scambio di componenti primarie tra gli opposti versanti dell’Italia Centrale. Biogeographia 17, 427–485. Clement, M., Posada, D., Crandall, K.A., 2000. TCS: a computer program to estimate gene genealogies. Mol. Ecol. 9, 1657–1659. Cortey, M., Vera, M., Pla, C., Garcia-Marin, J.L., 2009. Northern and Southern expansions of Atlantic brown trout (Salmo trutta) populations during the Pleistocene. Biol. J. Linn. Soc. 97, 904–917. Delling, B., Doadrio, I., 2005. Systematics of the trouts endemic to Moroccan lakes, with description of a new species (Teleostei: Salmonidae). Ichthyol. Explor. Freshw. 16, 49–64. Duméril, A., 1858. Note sur une truite d’Algérie (Salar macrostigma, A. Dum.). C. R. Acad. Sci. Paris 47, 160–162. Goudet, J., 2001. FSTAT, A Program to Estimate and Test Gene Diversities and Fixation Indices (Version 2.9.3).. Greenwood, P.H., 1974. Review of the Cenozoic freshwater fish faunas of Africa. Ann. Geol. Surv. Egypt 4, 211–232. Guinand, B., 1996. Use of a multivariate model using allele frequency distribution to analyse patterns of genetic differentiation among populations. Biol. J. Linn. Soc. 58, 173–195. Hammada, S., 2007. Etudes sur la Végétation des Zones Humides du Maroc. PhD Thesis, Rabat University. Ilahiane, H., 1996. Small-Scale Irrigation in a multiethnic oasis environment: the case of Zaouit Amelkis village, southeast Morocco. J. Political Ecol. 16, 89–106. Joleaud, L., 1938. Études de géographie zoologique sur la Berbérie: Les truites. Hesperis, Arch. berbères et Bull. Inst. Hautes Etudes Marocaines 25, 247–249. Kerans, H.J., 1962. Fish acclimatization in Morocco. Salmon Trout Mag 166, 193–197. Kottelat, M., 1997. European freshwater fishes. Biologia 52 (Suppl. 5), 1–271. Lerceteau-Köhler, E., Weiss, S., 2006. Development of multiplex PCR microsatellite assay in brown trout Salmo trutta, and its potential application for the genus. Aquaculture 258, 641–645. McMeel, O.M., Hoey, E.M., Ferguson, A., 2001. Partial nucleotide sequences, and routine typing by polymerase chain reaction–restriction fragment length polymorphism, of the brown trout (Salmo trutta) lactate dehydrogenase, LDHC190 and 100 alleles. Mol. Ecol. 10, 29–34. Medrano, F.J., Aasen, E., Sharrow, L., 1990. DNA extraction from nucleated red blood cells. Biotechniques 8, 43. Mouslih, M., 1987. Introductions de poissons et d’écrevisses au Maroc. Rev. Hydrobiol. Trop. 20, 65–72. Pellegrin, J., 1921. Les poissons des eaux douces de l’Afrique du Nord Française: Maroc, Algérie, Tunisie, Sahara. Mem. Soc. Sci. Natl. Maroc 1, 1–216. Pellegrin, J., 1924a. Les Salmonidés du Maroc. CR Acad. Sci. Paris 178, 970–972. Pellegrin, J., 1924b. Le Salmo pallaryi Pellegrin, poisson du Moyen Atlas marocain. Bull. Mus. Hist. Natl. Paris 30, 181–184.
A. Snoj et al. / Molecular Phylogenetics and Evolution 61 (2011) 203–211 Pérès, J.M., 1985. History of the Mediterranean biota and the colonization of the depth. In: Margalef, R. (Ed.), Western Mediterranean. Pergamon Press, London, pp. 200–234. Posada, D., Crandall, K.A., 1998. MODELTEST: testing the model of DNA substitution. Bioinformatics 14, 817–818. Schöffmann, J., 1993. Autochthone Forellen (Salmo trutta L.) in Nordafrika. Österr. Fisch 46, 164–169. Schöffmann, J., Sušnik, S., Snoj, A., 2007. Phylogenetic origin of Salmo trutta L 1758 from Sicily, based on mitochondrial and nuclear DNA analyses. Hydrobiologia 575, 51–55. Suarez, J., Bautista, J.M., Almodovar, A., Machordom, A., 2001. Evolution of the mitochondrial control region in Palaearctic brown trout (Salmo trutta) populations: the biogeographical role of the Iberian Peninsula. Heredity 87, 198–206.
211
Sušnik, S., Snoj, A., Dovcˇ, P., 2001. Evolutionary distinctness of grayling (Thymallus thymallus) inhabiting the Adriatic river system, as based on mtDNA variation. Biol. J. Linn. Soc. 74, 375–385. Swofford, D.L., 2000. PAUP Phylogenetic Analysis using Parsimony (and Other Methods). Version 4. Sinauer Associates, Sunderland. 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 tool. Nucleic Acids Res. 25, 4876–4882. Vivier, P., 1948. Note sur les eaux douces du Maroc et sur leur mise en valeur. B. Fr. Piscic. 150, 5–27. Werner, F., 1931. Ergebnisse einer zoologischen Forschungsreise nach Marokko. II.. Fische. Sitzb. Ak. Wiss. Wien 1, 261–270.
Contents lists available at ScienceDirect
Molecular Phylogenetics and Evolution journal homepage: www.elsevier.com/locate/ympev
Phylogeographic structure and demographic patterns of brown trout in North-West Africa Aleš Snoj a, Saša Maric´ b, Simona Sušnik Bajec a, Patrick Berrebi c, Said Janjani d, Johannes Schöffmann e,⇑ a
University of Ljubljana, Biotechnical Faculty, Department of Animal Science, Groblje 3, SI-1230 Domzˇale, Slovenia University of Belgrade, Faculty of Biology, Institute of Zoology, Studentski trg 16, 11001 Belgrade, Serbia c Université Montpellier II, Institut des Sciences de l’Evolution, UMR 5554 UM2-CNRS-IRD, CC 065, Place E. Bataillon, 34095 Montpellier Cedex 5, France d Faculté des Sciences Semlalia de Marrakech, Hydrobiology Laboratory, Avenue Prince Moulay Abdellah BP, 2390 Marrakech, Morocco e Lastenstrasse 25, 9300 St. Veit/Glan, Austria b
a r t i c l e
i n f o
Article history: Received 10 March 2011 Revised 13 May 2011 Accepted 14 May 2011 Available online 30 May 2011 Keywords: Salmo trutta Atlantic lineage Mitochondrial DNA Microsatellite DNA Morocco Atlas Mountains
a b s t r a c t The objectives of the study were to determine the phylogeographic structure of brown trout (Salmo trutta) in Morocco, elucidate their colonization patterns in North-West Africa and identify the mtDNA lineages involved in this process. We also aimed to resolve whether certain brown trout entities are also genetically distinct. Sixty-two brown trout from eleven locations across the Mediterranean and the Atlantic drainages in Morocco were surveyed using sequence analysis of the mtDNA control region and nuclear gene LDH, and by genotyping twelve microsatellite loci. Our study confirms that in Morocco both the Atlantic and Mediterranean basins are populated by Atlantic mtDNA lineage brown trout only, demonstrating that the Atlantic lineage (especially its southern clade) invaded initially not only the western part of the Mediterranean basin in Morocco but also expanded deep into the central area. Atlantic haplotypes identified here sort into three distinct groups suggesting Morocco was colonized in at least three successive waves (1.2, 0.4 and 0.2–0.1 MY ago). This notion becomes more pronounced with the finding of a distinct haplotype in the Dades river system, whose origin appears to coalesce with the nascent stage of the basal mtDNA evolutionary lineages of brown trout. According to our results, Salmo akairos, Salmo pellegrini and ‘‘green trout’’ from Lake Isli do not exhibited any character states that distinctively separate them from the other brown trout populations studied. Therefore, their status as distinct species was not confirmed. Ó 2011 Elsevier Inc. All rights reserved.
1. Introduction The brown trout (Salmo trutta, in this study considered as a single species except for Salmo salar, Salmo obtusirostris and Salmo ohridanus) is the most widely distributed Palaearctic freshwater fish. It is native across Europe, including Iceland, and parts of western and central Asia, with a natural range that extends southwards even to the Atlas Mountains of North-West Africa (Pellegrin, 1921, 1924a; Azeroual, 2003). Up until the Miocene, the inland water ichthyofauna of North-West Africa had a tropical character, after which it became Palaearctic in composition (Greenwood, 1974). Due to unsuitable tropical warm conditions in seas surrounding the Iberian Peninsula and northern Africa (Pérès, 1985), it is believed that brown trout did not arrive in this region before the onset of the Pleistocene cooling. During the warm ⇑ Corresponding author. Fax: +43 4212 214214. E-mail addresses: [email protected] (A. Snoj), [email protected] (S. Maric´), [email protected] (P. Berrebi), [email protected] (S. Janjani), [email protected] (J. Schöffmann). 1055-7903/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.ympev.2011.05.011
interglacial periods, brown trout in North-West Africa largely became extinct with some populations retreating to higher altitudes, while during cold periods, the distribution range advanced southwards. Thus brown trout from more northerly locations expanded their range along the Atlantic coast as far south as the Draa river basin at a latitude of about 28°N, and along the Mediterranean zone eastwards to Algeria and even Sicily (Schöffmann et al., 2007; Cortey et al., 2009). Due to cycling of glacial periods, the southward expansion of brown trout into the coastal rivers of the Iberian Peninsula and North-West Africa appears to have repeated several times (Cortey et al., 2009). Brown trout in Morroco of both the Atlantic and the Mediterranean basins have been assigned, on the basis of a limited sampleset (Bernatchez, 2001; Suarez et al., 2001), to the southern clade of the Atlantic mtDNA lineage, largely present in the Iberian peninsula (Cortey et al., 2009). This assignment is centered on the concept of major evolutionary lineages based upon mtDNA analysis, e.g., Atlantic (AT), Danube (DA), Mediterranean (ME), Adriatic (AD), marmoratus (MA) (Bernatchez et al., 1992) and Duero (DU) (Suarez et al., 2001).
204
A. Snoj et al. / Molecular Phylogenetics and Evolution 61 (2011) 203–211
In Morocco, autochthonous brown trout presently occur in the High Atlas above an altitude of 1500 m, in the Middle Atlas above 1200 m, and in the Mediterranean tributaries of the Rif Mountains, where they reside at an altitude of about 400 m (Azeroual, 2003; unpublished data). Yet in the middle years of the last century, the lower altitude limit of brown trout residence in the Middle Atlas was found to be about 900 m (Kerans, 1962). In the western High Atlas Mountains, the southernmost natural distribution of brown trout is nowadays confined to a latitude of 30° 500 N (River Nfiss of the Tensift basin), and 31° 020 (Lake Ifni of the Sous basin). Curiously, one brown trout population inhabits the headwaters of a tributary of the River Ziz, which rises on the south slope of the High Atlas and drains into the Sahara Desert (Ilahiane, 1996; Hammada, 2007). Isolated lake-dwelling brown trout still exist in two lakes of the High Atlas: the recently described ‘‘dwarf trout’’, Salmo akairos of Lake Ifni (Delling and Doadrio, 2005) and the ‘‘green trout’’ of Lake Isli (Vivier, 1948; Mouslih, 1987). Another lake-dwelling trout, also described as a distinct species Salmo pallaryi, was reported from the high mountain Lake Aguelmame Sidi Ali in the Middle Atlas (Pellegrin, 1924a,b) and from the nearby lake Tiguelmamine N’Ait Mahi (Vivier, 1948). However, this trout disappeared soon after the introduction of carp (Cyprinus carpio) in 1934 (Mouslih, 1987; Schöffmann, 1993). Salmo macrostigma (Duméril, 1858), first described from Oued el Abaïch in north-eastern Algeria, was described also in Morocco from the River Tigrigra of the Sebou basin, the upper River Moulouya with its tributaries Ansegmir and Berrem, and the River Melloulou, a tributary to the lower Moulouya river system (Pellegrin, 1924a; Joleaud, 1938). The name S. macrostigma has been applied also to several periMediterranean trout populations (Bianco, 1994; Kottelat, 1997) though without any demonstrative analysis. Another morphologically distinct species, Salmo pellegrini, was described from the stream Ourika, south of the city of Marrakech (Werner, 1931) and characterized by an appearance intermediate between S. pallaryi and S. macrostigma. In the course of the twentieth century, many native brown trout populations in Morocco became drastically reduced or extirpated due to human activities. The predominant threats to the remaining populations are increased water use, environmental degradation due to deforestation and overgrazing, road and dam constructions, water pollution and global warming. Some locations have been stocked with a domestic brown trout strain originating from River Oum Er Rbia upstream wild brown trout at the state hatchery of Azrou (Centre National d’Hydrobiologie et de Pisciculture, CNHP, unpublished report). The primary objectives of this study are to determine phylogeography of brown trout in Morocco, elucidate their colonization patterns and identify the lineages involved. In this context we examine the proposed hypothesis of colonization of Sicily with brown trout via North-West Africa.
A further aim is to genetically profile S. pellegrini (River Ourika), S. akairos (Lake Ifni) and ‘‘green trout’’ (Lake Isli) and determine if they represent genetically distinct entities. 2. Material and methods 2.1. Material Between 1998 and 2010, 62 brown trout from eleven locations across the Mediterranean and the Atlantic river basins in Morocco were fin-clipped (Table 1, Fig. 1). Allegedly unstocked locations were selected based upon information published in the literature (Schöffmann, 1993; Kottelat, 1997; Bernatchez, 2001; Delling and Doadrio, 2005) and the authors’ personal observations. The highly limited range of brown trout in Morocco—it is confined mainly to remote and inaccessible small streams in the Atlas and Riff Mountains—made sampling extremely difficult and, crucially, contributed to a relatively small number of sampling sites. Total DNA was isolated from fin tissue preserved in 96% ethanol following the protocol of Medrano et al. (1990). Brown trout samples from Sicily (see Schöffmann et al., 2007) were reanalyzed in order to obtain comparable sequences of mtDNA CR. 2.2. Mitochondrial DNA amplification, sequencing and data analysis Mitochondrial DNA CR was amplified by polymerase chain reaction (PCR) using primers 28RIBa (Sušnik et al., 2001) and HN20 (Bernatchez and Danzmann, 1993). Each 30 ll reaction included 1 lM of each primer, 0.2 lM dNTP, 1.5 lM MgCl2, 1 PCR buffer, 1 U Taq polymerase (Applied Biosystems, Foster City, CA, USA) and 50 ng of genomic DNA. Amplified DNA fragments were run on a 1.5% gel and subsequently dissected and isolated from the gel using the QIAEX II gel Extraction Kit (QIAGEN, Hilden, Germany). The conditions for PCR were initial denaturation (95 °C, 3 min) followed by 30 cycles of strand denaturation (94 °C, 45 s), primer annealing (52 °C, 45 s) and DNA extension (72 °C, 2 min). All PCR amplifications were performed in a programmable thermocycler GeneAmpÒ PCR System 9700 (Applied Biosystems). All sequencing reactions were prepared using a BigDye Terminator Ready Reaction Mix (Applied Biosystems) according to the manufacturer’s recommendations. The control region fragment was sequenced from both directions using PCR primers. Termination PCRs were performed in a programmable thermocycler under the following conditions: 10 s denaturation at 96 °C, 5 s annealing at 50 °C and 4 min extension at 60 °C, repeated for 30 cycles. The amplified, fluorescently labeled and terminated DNA was saltprecipitated and analysed with an ABI Prism 3130 xl automated sequencer.
Table 1 Sample sites and sizes (N), mtDNA haplotypes, and corresponding river system and drainages (AT, Atlantic; ME, Mediterranean). Sample numbers correspond to those in Fig. 1. For details about the samples from Sicily, see Schöffmann et al. (2007). Sample No.
Location
N
mtDNA haplotype
River system/basin
Drainage
1 2 3 4 5 6 7 8 9 10 11
Lake Isli Melloul Ourika Ait Mizane Lake Ifni M-Goun Dades Zaouia-Sidi-Hamza Berrem Kanar Adelma
4 4 5 12 7 2 12 5 4 3 4
ATM2 ATM6 ATM4 ATM1, ATcs25 ATM3 Dades Dades ATM7 ATM4 ATcs33 ATM5
Oued Melloul ? Oum Er Rbia Oum Er Rbia Tensift Reghaya ? Tensift Tifnout ? l’oued Souss Draa Draa Ziz Moulouya Medit. Sea basin Medit. Sea basin
AT AT AT AT AT AT AT AT ME ME ME
A. Snoj et al. / Molecular Phylogenetics and Evolution 61 (2011) 203–211
205
Fig. 1. Map of sampling sites. The numbers correspond to those in Table 1.
Sequences of 980 bp of mtDNA CR were aligned using computer program ClustalX (Thompson et al., 1997). To assign haplotypes of individuals to the mtDNA lineages previously identified for brown trout, data were compared to numerous haplotype sequences of brown trout found across the Iberian Peninsula and Mediterranean drainage (against all known haplotypes from the AT, DU, ME, MA and AD lineages and some from the DA lineage). The GenBank accession numbers of all the mtDNA CR sequences used in this study are reported in Table 2. Aligned haplotypes were imported into the program PAUP Version 4.0b10 (Swofford, 2000) for phylogenetic analysis. Maximum parsimony (MP) and maximum likelihood (ML) were used for phylogenetic reconstruction. For MP, insertions or deletions (indels) were included as a fifth character. A heuristic search (10 replicates) with Tree Bisection Reconnection (TBR) branch-swapping was employed to find the most parsimonious trees. For ML, a sequence evolution model was first chosen using the program Modeltest Version 3.7 (Posada and Crandall, 1998) incorporated into PAUP. After a model was chosen, a heuristic search (10 replicates) was used to estimate the most likely topology. Support values for the nodes were obtained with 1000 bootstrap replicates for MP or ML analysis, while the fast stepwise addition method was used for ML. In addition, the genealogical relation of haplotypes was depicted also using a 95% statistical parsimony network constructed from mtDNA CR sequences using program TCS 1.3 (Clement et al., 2000).
the single polymorphic site. To determine the state of this site, a 428 bp fragment of the LDH gene was amplified using primers Ldhxon3F and Ldhxon4F following the conditions reported in McMeel et al. (2001), and sequenced. Sequencing reactions were prepared as described above, using the Ldhxon3F primer. All the LDH sequences were deposited in GeneBank under Acc. Nos. JF297983–JF297990. 2.4. Microsatellite amplification and data analysis Twelve microsatellite loci were chosen for analysis using primers for PCR amplification and multiplex PCR conditions as described in Lerceteau-Köhler and Weiss (2006). Aliquots of amplified fluorescently labeled DNA were mixed with formamide and GENESCAN500 ROX Size Standard (Applied Biosystems) and genotyped on the ABI Prism 3130 xl with GeneMapperÒ Software v4.0 (Applied Biosystems). Genetic relationships among the populations and the individuals were estimated with factorial correspondence analysis (FCA) using GENETIX 4.04 (Belkhir et al., 2004). This analysis permits the plotting of each individual based on its complete genotype. The inertia values (i.e., the proportion of the total information contained by an axis, expressed as a percentage) along each axis were shown to be equivalent to linear combinations of the monolocus fixation index (Fst) values (Guinand, 1996). Pairwise Fst values were calculated in FSTAT (Goudet, 2001). 3. Results
2.3. LDH gene analysis 3.1. Mitochondrial DNA analysis To check for the presence of northern European hatchery strain of Atlantic lineage in the sampled populations, every individual was tested for the LDH 90 allele (McMeel et al., 2001), which is diagnostic for this strain and can be specifically determined upon
3.1.1. Haplotype determination DNA alignment of the sequences from the Moroccan samples collapsed into ten haplotypes (Table 1), eight of which had not
206
A. Snoj et al. / Molecular Phylogenetics and Evolution 61 (2011) 203–211
Table 2 GenBank accession numbers of the mtDNA CR sequences used in this study. Asterisks denote new sequences. Haplotype
Acc. number
Haplotype
Acc. number
JF297978 JF297979 JF297980 JF297975 JF297977 JF297976 JF297982 JF297981 JF297974 AF273086 AF273087 AF274574 AF274575 AF274576 AF274577 AY836327 AY836328 EF530476 EF530477 EF530478 EF530479 EF530480 EF530481 EF530483 EF530484 EF530485 EF530486 EF530487 EF530488 EF530489 EF530490 EF530491 EF530492 EF530493 EF530494 EF530495 EF530496 EF530497 EF530498 EF530499 EF530500 EF530501 EF530502 EF530503 EF530504 EF530505 EF530506 EF530507 EF530508 EF530509 EF530510 EF530511
ATcs52 AT11a AT11b DUcs1 DUcs2 DUcs3 DUcs4 DUcs5 DUcs6 DUcs7 DUcs8 DUcs9 DUcs10 DUcs11 DUcs12 DUcs13 DUcs14 DUcs15 DUcs16 DUcs17 DUcs18 DUcs19 DUcs20 DUcs21 DUcs22 DUcs23 DA1a DA3 DA9 DA22 DA24 ADcs1 ADcs2 ADcs4 ADcs5 ADcs8 ADcs9 ADcs14 ADcs15 ADcs19 ADcs20 haplo12 ADZ1 MEcs1 MEcs3 MEcs8 MEcs9 MAcs1 MA2a MA2b Salmo ohridanus Salmo salar
EF530512 AY185578 AY185579 EF530513 EF530514 EF530515 EF530516 EF530517 EF530518 EF530519 EF530520 EF530521 EF530522 EF530523 EF530524 EF530525 EF530526 EF530527 EF530528 EF530529 EF530530 EF530531 EF530532 EF530533 EF530534 EF530535 AY185568 AY185571 AY185572 AY185573 AY185576 AY836350 AY836331 AY836333 AY836334 AY836337 AY836338 AY836343 AY836344 AY836348 AY836349 AY926570 DQ381565 AY836350 AY836352 AY836357 AY836358 AY836365 DQ841189 DQ841190 AY926564 AF133701
AtM1 AtM2 AtM3 AtM4 AtM5 AtM6 AtM7 Dades AtSic ATcs1 ATcs2 ATcs3 ATcs4 ATcs5 ATcs6 ATcs11 ATcs12 ATcs14 ATcs15 ATcs16 ATcs17 ATcs18 ATcs19 ATcs21 ATcs22 ATcs23 ATcs24 ATcs25 ATcs26 ATcs27 ATcs28 ATcs29 ATcs30 ATcs31 ATcs32 ATcs33 ATcs34 ATcs35 ATcs36 ATcs37 ATcs38 ATcs39 ATcs41 ATcs42 ATcs43 ATcs45 ATcs46 ATcs47 ATcs48 ATcs49 ATcs50 ATcs51
yet been reported, while two (ATcs25 and ATcs33) had been previously described for the Iberian Peninsula. All the haplotypes belonged to the AT phylogenetic lineage, except for haplotype ‘‘Dades’’ (see below). This haplotype was found only in the Dades river system (Draa basin), where it appeared to be fixed. The other haplotypes were also found to be fixed in particular sampling locations, with the exception of Ait Mizane, where two haplotypes were found (Table 1). All haplotypes appeared to be specifically linked to sampling location except for ATM4, which was bound to brown trout from Berrem (Moulouya river system, Mediterranean basin) and Ourika (Tensift river system, Atlantic basin, which was expected to harbor S. pellegrini (Werner, 1931)). Sequences of brown trout from Sicily revealed a single and unique haplotype, ATSic.
3.1.2. Phylogenetic trees Phylogenetic tree reconstruction sorted all the haplotypes found in Morocco and Sicily into the AT clade, with the exception of haplotype Dades, whose phylogenetic resolution was not well resolved: on the ML tree (Fig. 2), this haplotype formed a polytomy with the Danubian and Atlantic–Duero clades, while on the MP tree it was basal to all Danubian and Atlantic haplotypes (100% bootstrap support; data not shown). The other haplotypes found in Morocco were more consistently distributed on the trees, appearing in three groups. The first group was represented by haplotypes ATM1 and ATM6, found to be closely related and forming a basal clade vs. the entire Atlantic-Duero clade. The second group (ATM3/4/5, ATcs33 and ATSic) assorted with haplotypes previously found in the Iberian Peninsula (i.e., clade AT 3-3 or so-called ‘‘southern At-clade’’; Cortey et al., 2009), and the third group consisted of haplotypes ATM2 and ATM7, which were associated with clade AT 3-2 (Cortey et al., 2009), found across the entire Atlantic basin. 3.1.3. Haplotype network The initial designed haplotype network—based on 116 haplotypes and representing the majority of S. trutta mtDNA variants— revealed a very tangled and looped structure. Following removal of approximately half of the uninformative haplotypes (either on tip positions or distant from Moroccan haplotypes or both), the resolution of the network improved considerably without changing the basic haplotype network organization, and displayed more clearly an interesting relationship among Moroccan, south Atlantic and Mediterranean haplotypes. This network (Fig. 3) clearly showed that all Moroccan haplotypes, with the exception of ‘‘Dades’’, endemic to the Draa basin, fell into the Atlantic haplotype cluster. Most of the Moroccan haplotypes were connected closely to haplotypes previously described from the Iberian Peninsula and were distributed among two groups with the central haplotypes ATcs33 and ATcs1 representing, respectively, plesiomorphic haplotypes for the so-called southern (AT 3-3) and northern Atlantic (AT 3-1, Cortey et al., 2009) clades. In the southern clade, haplotype ATcs33 was found to be connected to its recent Moroccan derivates with single-mutation branches including the haplotype ATSic, while in the northern clade, Moroccan haplotypes were more distant (four or six mutational steps) from the central one (ATsc1). Haplotypes ATM2 and ATM7 are derivates of ATsc25 (one or two mutation steps away) all of which belong to the group that was previously found in both the southern and northern Atlantic basins (Cortey et al., 2009). The haplotype ‘‘Dades’’ was found to be equally separated (12 mutational steps) from the Atlantic, Duero and Adriatic lineages. The haplotype distribution did not seem to follow any obvious geographical pattern. 3.2. LDH results Alignment of the obtained DNA sequences revealed that all individuals tested had alleles at the LDH locus other than LDH 90. In addition to the unique mutation site that specifically distinguishes between 90 and non-90 alleles (see McMeel et al., 2001), several other polymorphic sites were found, permitting construction of a Bayesian tree, which showed no phylogeographic or population organization (data not shown). 3.3. Microsatellites data analysis The FCA diagram (Fig. 4) indicated three main well differentiated groups that included brown trout collected in (1) Dades, (2) Kanar and (3) the remaining sampling sites.
A. Snoj et al. / Molecular Phylogenetics and Evolution 61 (2011) 203–211
207
Fig. 2. Maximum likelihood tree (HKY + I + G model, transition: transversion 2.3540; proportion of invariable sites (I) 0.6675; gamma distribution shape parameter 0.9157) for genus Salmo based on 980 bp of the mtDNA control region. Haplotypes representing all S. trutta mtDNA lineages were included in the analysis, the ones characteristic for trout in Morocco are shown in bold. Bootstrap support values refer to maximum likelihood (top) and maximum parsimony (bottom). Values <50 are marked with ‘‘/’’ or are not marked when there was no value above 50 in any of the analyses.
208
A. Snoj et al. / Molecular Phylogenetics and Evolution 61 (2011) 203–211
Fig. 3. Network relating haplotypes found in Morocco (in bold) with previously published haplotypes of other S. trutta phylogenetic lineages. Ellipses represent haplotypes; solid lines connect each haplotype (regardless of length) and represent single mutation events. Small black circles represent missing or theoretical haplotypes.
Fig. 4. Three-dimensional cluster analysis for all trout specimens from Morocco based on 12 microsatellite loci.
There were no groupings following specific geographic patterns (e.g., being associated with the Atlantic or Mediterranean drainage) nor corresponding with any mtDNA haplotype clusters. High pair-wise Fst values (Table 3) indicated considerable genetic differentiation corresponding to ancient isolation between the populations, particularly in relation to Dades (0.52; SD = ±0.08) and Kanar (0.57; SD = ±0.06).
4. Discussion 4.1. Dominant mtDNA lineages in Morocco Our study confirms that in Morocco, both the Atlantic and Mediterranean basins are populated by Atlantic mtDNA lineage brown trout previously described by Bernatchez (2001). In addition, the
209
A. Snoj et al. / Molecular Phylogenetics and Evolution 61 (2011) 203–211 Table 3 Fst values among the populations sampled in Morocco.
Lake Isli Melloul Ourika Ait Mizane Lake Ifni M-Goun Dades Zaouia-Sidi-Hamza Berrem Kanar Adelma
Lake Isli
Melloul
Ourika
Ait Mizane
Lake Ifni
M-Goun
Dades
Zaouia-Sidi-Hamza
Berrem
Kanar
Adelma
0.0000
0.4553 0.0000
0.6682 0.5099 0.0000
0.4700 0.2989 0.5261 0.0000
0.5768 0.4559 0.5040 0.5077 0.0000
0.5486 0.3904 0.6131 0.4587 0.4578 0.0000
0.5569 0.4553 0.5887 0.5102 0.5540 0.1402 0.0000
0.4778 0.2909 0.5277 0.2272 0.4852 0.4033 0.4609 0.0000
0.5743 0.3163 0.4931 0.3617 0.5384 0.4769 0.5116 0.2982 0.0000
0.6579 0.5274 0.6880 0.5750 0.5635 0.5774 0.5800 0.5209 0.6335 0.0000
0.4438 0.3453 0.5089 0.4546 0.3415 0.3621 0.4563 0.3667 0.4621 0.5075 0.0000
Atlantic haplotypes identified here sort into three distinct groups suggesting a heterogeneous demographic history of brown trout in Morocco. This notion becomes more pronounced with the finding of a distinct haplotype—Dades—whose origin appears to coalesce with the nascent stage of the basal mtDNA evolutionary lineages of brown trout. Of the ten haplotypes found in Morocco, eight are strictly endemic, while the other two (ATcs25 and 33) have been reported previously for the Iberian Peninsula. Given the prevalence in Morocco of endemic haplotypes and absence of the LDH 90 allele and haplotypes ATsc1, 2, 3 and 4, denoting the most widespread European domestic strain of brown trout, one can conclude that brown trout demography in north-west Africa, at least in the sites sampled, has not been affected by introduction of commercial domestic brown trout from Europe. According to unpublished information provided by Azrou fish farm (CNHP) on its stocking activity during the last 10 years, the brown trout strain it has used originates from the Oum Er Rbia river population, which bears haplotype ATM6. This domestic strain should therefore have introduced this haplotype wherever the stocking has been successful (genetic participation of introduced fish in the natural population). However, since this haplotype was not detected in any population analysed, we can conclude that the genetic structure described for each sampling location is entirely natural. 4.2. Population structure according to nuclear markers To detect signals of present or recent gene flow between populations, microsatellites were also included in the analyses. There was virtually no correlation between the mtDNA and microsatellite structures of brown trout in Morocco, apart from confirmation of long-term isolation of the Dades population. According to FCA analysis, Moroccan brown trout appear to be highly structured, with no overlapping populations. However, there is no similarity in the nuclear and mitochondrial geographic organizations. Mitochondrial DNA, because of clonal transmission, is a more reliable tool for reconstruction of ancient phylogenies than are rapidly evolving and recombining nuclear microsatellites. We therefore consider that the origin of the present Moroccan populations (affected by several migrations at several periods contributing to several lineages, a long period of isolation and probably a high rate of population extinction and recolonization) cannot be properly described by microsatellites, also partly because of possible genetic saturation leading to homoplasy, which is here very likely reflected in the distorted distribution of the samples plotted on the FCA graph (Fig. 4). 4.3. Origins of brown trout inhabiting Mediterranean drainages The present study suggests that no haplotype other than the Atlantic ones exists in the Mediterranean drainages in Morocco.
In Kanar, a single haplotype ATcs33 was observed that is reported to exist also in the Atlantic basin of the Iberian Peninsula. Despite reports of repeated stocking (I. Doadrio, personal communication), no such evidence was found from the present analysis. Although stocking activities appear not to have happened in the Morocco brown trout populations studied, the demography presents a different situation for the Berrem population, where the only haplotype detected was ATM4 (southern clade), which was detected also in Ourika (Tensift river basin, Atlantic). These two locations are geographically very remote and even belong to different basins, so it is highly unlikely that fish have dispersed naturally from one location to the other via inland migration (e. g., river capture or similar) in historical times. Recent stocking also seems an unlikely explanation as the two populations differ considerably on the basis of microsatellites (pair-wise Fst = 0.5). Moreover, Ourika is considered to be populated by distinct S. pellegrini. It is possible though that this haplotype reached both locations by the same lineage colonization that took place in both the Atlantic and Mediterranean basins, at a particular stage in the multi-wave establishment (see below) of the present Moroccan ichthyofauna, and that the presence of the same haplotype in both locations is a consequence of an ancestral polymorphism. The southern Atlantic clade is the only one recorded for the Mediterranean basin in Morocco (Fig. 1, sample sites 9–11) indicating that the brown trout population here is homogeneous, probably a consequence of a single wave of migration into this area. Further evidence of dispersal of the southern Atlantic lineage into the Mediterranean basin is seen in the presence of haplotype ATSic, which belongs to the southern Atlantic clade and has been found to occur naturally in Sicily (Schöffmann et al., 2007). This is an important observation and clearly demonstrates that the Atlantic lineage (especially its southern clade) invaded initially not only the western part of the Mediterranean basin in Morocco but also expanded deep into the central area. However, the same colonization wave appears to have reached also more southern locations in the Atlantic basin in Morocco, with the presence of S. pellegrini in River Tensift and of S. akairos in Lake Ifni. The evolution of two Atlantic populations into these two distinct species (at least at the morphological level) suggests an ancient invasion locally maintained and protected by strong isolation, permitting local adaptation and differentiation. 4.4. History of establishment of brown trout populations in Morocco: mtDNA The ‘‘Moroccan haplotypes’’ of southern and northern clades differ considerably with respect to the number of mutational steps separating them from their central haplotype. After transforming these mutational steps into molecular clock estimates (1% per 1 million years; Bernatchez, 2001), the three major groups of Moroccan haplotypes seem to have evolved over three different time periods: 400,000 (ATM1, ATM6), 200,000 (ATsc25, ATM2
210
A. Snoj et al. / Molecular Phylogenetics and Evolution 61 (2011) 203–211
and ATM7) and 100,000 (ATcs33, ATM3, ATM4 and ATM5) years ago. However, the ‘‘Dades’’ haplotype was difficult to place in any of the known mtDNA lineages. Its plesiomorphic state with respect to the Atlantic and the Danubian lineages, rather deep position and long branch on the phylogenetic tree (Fig. 2), and deep and equal divergence from the Atlantic, Adriatic and Duero lineages (12 mutational steps away from each; Fig. 3) demonstrate a long-lasting independent evolution of this haplotype. These observations suggest its existence and arrival at approximately the same time as the appearance of the major lineages of brown trout, or possibly even earlier, and which based upon the molecular clock took place some 1.2 million years ago. The Dades population can justifiably be considered a relict of the first established brown trout lineage in Morocco and perhaps a representative of the predifferentiated ancestor of all trout lineages. The results obtained in the present study suggest that the colonization of brown trout in Morocco took place in several waves, as has been hypothesized previously (Cortey et al., 2009). Based upon our results, Morocco was colonized in at least three successive waves (1.2, 0.4 and 0.2–0.1 MY ago), leading to a specific genetic signature of Moroccan brown trout comprising at least three unrelated groups of mtDNA haplotypes. Generally speaking, there is little phylogeographic structure found in patterns of haplotype distribution in Morocco. For instance, the Tensift River harbors alternatively southern clade haplotypes in its Ourika tributary (sampling location 3; haplotype ATM4 corresponding to S. pellegrini) and northern clade haplotype in its Ait Mizane tributary (sampling location 4; haplotype ATM1, as well as the ubiquitous Atlantic haplotype ATsc25). This incongruence is probably due to stochastic effects of turbulent climatic and geo-morphological events that took place during the repeated cool and warm Pleistocene periods, causing independent local extinctions and permitting foundation of new populations during the next wave of invasion. Such intra-basin differentiations demonstrate very limited dispersal of brown trout populations, even between close tributaries, most probably due to high water temperatures in the plains preventing downstream trout migration. Of the Moroccan brown trout that are believed to represent independent evolutionary entities, three of them were included in our study: S. pellegrini (River Ourika; Werner, 1931), S. akairos (Lake Ifni; Delling and Doadrio, 2005) and ‘‘green trout’’ (Lake Isli; Vivier, 1948; Mouslih, 1987). According to our results, none of these three populations exhibited any character states that would distinctively separate them from the other studied brown trout populations. Therefore, their status as distinct species was not confirmed at the molecular level. Nevertheless, in the extreme conditions encountered by brown trout in the periphery of the species range, special adaptation to specific conditions, together with long isolation in drainages, in some cases even without a marine estuary, has led to strong morphological differentiation, seen in for example the ‘‘green’’ or ‘‘dwarf’’ phenotypes, which do not necessarily correspond to distinct genetic signatures. In contrast to this observation, we identified a brown trout population from the Dades river system that exhibits a distinctive external appearance (see graphic abstract) and a highly specific genetic make-up reflected by both mitochondrial and microsatellite DNA. Given the estimated time of its origin, probably prior to the time of the constitution of the S. trutta complex we know today, this population deserves special attention in terms of its classification status. Due to the considerable genetic diversity reflected by several endemic mtDNA haplotypes, which appear unique to almost each investigated location, and the preserved integrity of indigenous populations, brown trout in Morocco should be attached a special importance in the frame of the S. trutta complex. Many threats, including climate warming and environmental degradation due
to the modernization of Morocco, seriously endanger these special entities of brown trout. These concepts have to be considered by local managers in order to enhance the survival of brown trout in Morocco through appropriate management strategies and careful monitoring of stocking activities, particularly by prevention of introduction of non-native strains and long-distance translocation of feral specimens. Only through such careful management can the preservation of this unique genetic pool of brown trout be ensured. Acknowledgments We would like to thank the Berber people of the Atlas Mountains and G. Pustovrh for helping us with the collecting of samples. Special thanks go to I. Wilson for editing and correcting the manuscript and B. Delling for useful comments. This work was supported by the Slovenian Research Agency. References Azeroual, A., 2003. Monographie des Poissons des Eaux Continentales du Maroc: Systématique, Distribution et Ecologie. Université Mohammed V – Agdal. PhD Thesis, Faculté des Sciences, Rabat University. Belkhir, K., Borsa, P., Chikhi, L., Raufaste, N., Bonhomme, F., 1996–2004. GENETIX 4.04, Logiciel sous Windows TM Pour la Genetique des Populations. Laboratoire Genome, Populations, Interactions, CNRS UMR 5000, Universite de Montpellier II, Montpellier, France. Bernatchez, L., 2001. The evolutionary history of brown trout (Salmo trutta L.) inferred from phylogeographic, nested clade, and mismatch analyses of mitochondrial DNA variation. Evolution 55, 351–379. Bernatchez, L., Danzmann, R.G., 1993. Congruence in control-region sequence and restriction-site variation in mitochondrial DNA of brook charr (Salvelinus fontinalis Mitchill). Mol. Biol. Evol. 10, 1002–1014. Bernatchez, L., Guyomard, R., Bonhomme, F., 1992. DNA sequence variation of the mitochondrial control region among geographically and morphologically remote European brown trout Salmo trutta populations. Mol. Ecol. 1, 161–173. Bianco, P.G., 1994. L’ittiofauna continentale dell’Appennino umbro-marchigiano, barriera semipermeabile allo scambio di componenti primarie tra gli opposti versanti dell’Italia Centrale. Biogeographia 17, 427–485. Clement, M., Posada, D., Crandall, K.A., 2000. TCS: a computer program to estimate gene genealogies. Mol. Ecol. 9, 1657–1659. Cortey, M., Vera, M., Pla, C., Garcia-Marin, J.L., 2009. Northern and Southern expansions of Atlantic brown trout (Salmo trutta) populations during the Pleistocene. Biol. J. Linn. Soc. 97, 904–917. Delling, B., Doadrio, I., 2005. Systematics of the trouts endemic to Moroccan lakes, with description of a new species (Teleostei: Salmonidae). Ichthyol. Explor. Freshw. 16, 49–64. Duméril, A., 1858. Note sur une truite d’Algérie (Salar macrostigma, A. Dum.). C. R. Acad. Sci. Paris 47, 160–162. Goudet, J., 2001. FSTAT, A Program to Estimate and Test Gene Diversities and Fixation Indices (Version 2.9.3).
A. Snoj et al. / Molecular Phylogenetics and Evolution 61 (2011) 203–211 Pérès, J.M., 1985. History of the Mediterranean biota and the colonization of the depth. In: Margalef, R. (Ed.), Western Mediterranean. Pergamon Press, London, pp. 200–234. Posada, D., Crandall, K.A., 1998. MODELTEST: testing the model of DNA substitution. Bioinformatics 14, 817–818. Schöffmann, J., 1993. Autochthone Forellen (Salmo trutta L.) in Nordafrika. Österr. Fisch 46, 164–169. Schöffmann, J., Sušnik, S., Snoj, A., 2007. Phylogenetic origin of Salmo trutta L 1758 from Sicily, based on mitochondrial and nuclear DNA analyses. Hydrobiologia 575, 51–55. Suarez, J., Bautista, J.M., Almodovar, A., Machordom, A., 2001. Evolution of the mitochondrial control region in Palaearctic brown trout (Salmo trutta) populations: the biogeographical role of the Iberian Peninsula. Heredity 87, 198–206.
211
Sušnik, S., Snoj, A., Dovcˇ, P., 2001. Evolutionary distinctness of grayling (Thymallus thymallus) inhabiting the Adriatic river system, as based on mtDNA variation. Biol. J. Linn. Soc. 74, 375–385. Swofford, D.L., 2000. PAUP Phylogenetic Analysis using Parsimony (and Other Methods). Version 4. Sinauer Associates, Sunderland. 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 tool. Nucleic Acids Res. 25, 4876–4882. Vivier, P., 1948. Note sur les eaux douces du Maroc et sur leur mise en valeur. B. Fr. Piscic. 150, 5–27. Werner, F., 1931. Ergebnisse einer zoologischen Forschungsreise nach Marokko. II.. Fische. Sitzb. Ak. Wiss. Wien 1, 261–270.