Genetic variation among trout in the River Neretva

(Excerpt from: "Genetic variation among trout in the River Neretva
basin, Bosnia and Herzegovina" by A. Razpet, S. Sušnik, T. Jug and A. Snoj - Balkan Trout Restoration Group http://www.balkan-trout.com)

Mitochondrial DNA haplotypes, characteristic of the Adriatic, Danubian and Atlantic lineages of brown trout Salmo trutta and of Salmo obtusirostris were found in trout inhabiting the River Neretva basin. With the exception of the one associated with softmouth trout, haplotypes were not correlated with operational taxonomic units based on phenotype. Nuclear DNA analysis identified four genetic assemblages corresponding to S. obtusirostris, different geographically confined autochthonous S. trutta populations, introduced S. trutta and a genetically heterogeneous group located between S. obtusirostris and S. trutta in the dendrogram of individuals, indicating the existence of hybrid swarms in the Neretva basin. Genetic assemblages corresponding to Salmo marmoratus and the recently proposed Salmo cf. montenigrinus were not detected. The presence of genetic intermediates indicates that the studied taxa are not completely reproductively isolated and that genetic stability has been either anthropogenically interrupted or not yet achieved among Neretva trout. This finding should be considered in management decisions since such an unstable community must be particularly susceptible to breakdown in genetic population structure as a result of hybridization between native and nonnative introduced trout stocks.  

INTRO 

The present coastal areas of the south-western Balkans were never glaciated during the Pleistocene Ice Ages and have as such served as major European refugia for organisms. The existence of such refugia has been confirmed by phylogeographic studies of several non-fish taxa (Hewitt, 1996, 1999; Taberlet et al., 1998) and fish taxa (Bianco, 1990; Durand et al., 1999; Kotlik & Berrebi, 2001), including Salmo trutta L. (Garcia-Marin et al., 1999). The co-existing incipient and invading representatives have thus created an opportunity for hybridization and generation of hybrid swarms (Seehausen, 2004). Consequently, a high level of genetic diversity is expected in this region.

The rivers of the south-western Balkans are grouped into three major catchments draining into the Black, Aegean and Adriatic seas. River Neretva and its tributaries represent the second largest (after River Drin) river system in the east Adriatic watershed and a central habitat for ichthyofauna in this area. Salmonids
from the Neretva basin show considerable variation in many aspects of morphology, ecology and behaviour. It is therefore not surprising that on the basis of these characteristics, several species, i.e. S. trutta (brown trout), Salmo marmoratus Cuvier (marble trout), Salmo obtusirostris (Heckel) (softmouth trout), Salmo farioides (Karaman) and Salmo dentex (Heckel) were identified in earlier studies (Heckel, 1852; Karaman, 1937). The existence of S. trutta and S. obtusirostris in the Neretva basin has never been disputed (Karaman,
1926; Vuković , 1982; Kosorić et al., 1983). Similarly, S. marmoratus has been widely recognized as occurring in the basin, but no genetic data have been presented to verify this taxon as synonymous with S. marmoratus of the northern Adriatic.

The status of S. dentex and S. farioides has always been questionable due to inadequate original descriptions, rare sightings and absence of type material for S. farioides (Kottelat, 1997). On the basis of a recent morphological study (Delling, 2003), another new species that in morphology appears intermediate between S. trutta and S. obtusirostris’ has been spotted in River Neretva and tentatively designated Salmo cf. montenigrinus [referring to Trutta montenigrina (Karaman) from the River Morača, Montenegro]. For detailed chronology of descriptions of the Neretva trout and changes in their nomenclature, see Kottelat (1997) and Delling (2003). Natural hybrids such as S. obtusirostris S. trutta and S. marmoratus S. trutta have been observed and reported for the Neretva basin (Vuković , 1982). The hybridization between autochthonous taxa was confirmed also in a hatchery experiment performed in the fish farm located in the River Buna, a tributary in the lower part of the River Neretva (Kosorić & Vuković , 1969). However, introduction of non-native brown trout has also been practised in River Neretva (Vuković , 1982). For instance, several transfers of brown trout from different locations in Slovenia into the River Neretva were carried out several times during the period of the former Yugoslavia (J. Ocvirk, pers. comm.). According to local fishermen, the stocking of River Neretva with trout of unknown lineage from various fish farms in Bosnia and Herzegovina and elsewhere is still practised (S. Mustafić , pers. comm.). Moreover, stocking activities have never been well documented, especially in the disturbed post-war years, and no official information is therefore available regarding the purpose of stocking and the origin or the numbers of stocked fish. Despite the apparent complexity and disputed taxonomy of Salmo in the system, no systematic study has been undertaken to verify the various operational taxonomic units (OTUs) in the Neretva basin. Specifically, no phylogeographic comparisons with the commonly accepted evolutionary lineages of brown trout (Bernatchez, 2001; Suarez et al., 2001) have been performed and no studies have been undertaken in this river system in terms of possible interactions between native and introduced trout populations. Thus, it is clear that the status of Neretva trout is insufficiently determined. The main goal of the present study was, therefore, to answer the following questions. (1) Which mtDNA lineages exist in the Neretva basin? (2) Are there any population substructures? (3) Is there congruence among mtDNA lineages and any genetically determined substructures and operational taxonomic groups based on previous phenotypic inferences (i.e. with proposed species names)? Four microsatellite loci, which emerged as informative in the previous studies of Balkan trout (unpubl. data), were chosen for analysis along with the mtDNA control region (CR) and two nuclear gene regions [part of lactate dehydrogenase LDH-C1 and the C intron of growth hormone 2 (GH2C)], which have already been used in similar studies (Oakley & Phillips, 1999; Snoj et al., 2002; Phillips et al., 2004) and which therefore offer a suitable comparative experimental system. On the basis of lineage-specific mtDNA haplotypes and nuclear DNA data processed with a Bayesian approach, the authors also aimed to distinguish between indigenous and introduced trout and thus examine the impact, if any, of stocking.

EVOLUTIONARY HISTORY CONSIDERATIONS

Structured genetic composition along with high levels of morphological variation and low haplotype diversity within the native Neretva trout (S. obtusirostris is excluded from this consideration) may indicate relatively recent evolution of the detected assemblages. The maximum time of their independent evolution is connected with the proposed population expansion of the Adriatic lineage that probably took place c. 155 000 years ago in the western Mediterranean as a response to dynamic climate and sea level fluctuations (Cortey et al., 2004; Sušnik et al., 2007a). Although the haplotypes, AdN and Ad-s3 belong to the same evolutionary lineage, it is evident from the haplotype network (Fig. 2) that they do not represent contiguous haplotypes on the haplotype network. The absence of intermediate haplotypes and the prevalence of apomorphy in AdN additionally refute a common ancestry and rather indicate unique arrivals into the Neretva basin, probably as a consequence of successive colonization events. A bimodal mismatch distribution based on mtDNA CR  haplotypes has been observed for the Adriatic drainage (Sušnik et al., 2007a), supporting the notion that the diversity of brown trout in the whole Adriatic drainage has been influenced by two colonization events. As the divergence between AdN and Ad-s3 exceeds even the net divergence between the Adriatic and Mediterranean clades (Cortey et al., 2004), it is assumed that the two haplotypes diverged in the initial phase of the Adriatic lineage formation. The haplotype AdN strongly outnumbers Ad-s3 and is more distant from the central haplotype on the network. It is therefore plausible that AdN represents a relict from the first wave of colonization, which followed the expansion within the Adriatic lineage, with Ad-s3 reaching the Neretva basin at a later stage. Wide distribution of the haplotype Ad-s3 throughout the Mediterranean [from the  Adriatic rivers in Slovenia (Snoj, 2004) to the islands of Corsica and Sardinia (Bernatchez, 2001)] additionally supports the hypothesis of a more recent dispersal of brown  trout bearing this haplotype.  However, the indigenous state of haplotype AdN to the territory of the south-western Balkans has been supported by a recent finding of closely related haplotypes (AdPrz, DQ318129, Marić et al., 2006, Fig. 2; and AD-C1, DQ381567, Sušnik et al., 2007a) present in a neighbouring part of the Adriatic drainage. Since hybrids often exhibit novel or extreme characters compared with parental taxa (Rieseberg et al., 1999; Seehausen, 2004), hybridization between successively invading lineages may be one of the important factors that have contributed to the heterogeneous phenotypes of trout in the Neretva basin.

CONSERVATION ASPECTS

The results presented here clearly show that stocking of non-native trout lineages has been performed in the River Neretva basin; moreover, there are indications that introgression with the introduced brown trout alleles has already occurred. From a conservation point of view, this is an important issue as balance between the different sympatric genetic assemblages present in the basin must be, as a result of establishment of only weak reproductive barriers, very unstable and sensitive. Any abundant introductions, reflected in introgression of exotic alleles into indigenous genomes, may create a bridge for hybridization with native trout resulting in overall panmixia and extinction of the extant genetic assemblages. The harmful effects of such hybridization  have already been demonstrated in Adriatic rivers, in the Soča basin in Slovenia and Italy, where S. marmoratus and Adriatic grayling (Thymallus thymallus L.) have almost been replaced by exotic brown trout and grayling introduced from River Sava (Berrebi et al., 2000; Susˇnik et al., 2004). Since the trout genealogy
and demography in the Neretva basin are far more complicated than those in the River Soča, extreme caution in management decisions is recommended and any introductions of non-native fishes are to be emphatically discouraged.

(Balkan Trout Restoration Group http://www.balkan-trout.com)