Cryptic diversity in Telestes pleurobipunctatus (; ) as a consequence of historical biogeography in the Ionian Freshwater Ecoregion (Greece, Albania) Ivana Buj, Radek Šanda, Stamatis Zogaris, Jörg Freyhof, Matthias F. Geiger & Jasna Vukić

Hydrobiologia The International Journal of Aquatic Sciences

ISSN 0018-8158

Hydrobiologia DOI 10.1007/s10750-019-3935-6

1 23 Your article is protected by copyright and all rights are held exclusively by Springer Nature Switzerland AG. This e-offprint is for personal use only and shall not be self- archived in electronic repositories. If you wish to self-archive your article, please use the accepted manuscript version for posting on your own website. You may further deposit the accepted manuscript version in any repository, provided it is only made publicly available 12 months after official publication or later and provided acknowledgement is given to the original source of publication and a link is inserted to the published article on Springer's website. The link must be accompanied by the following text: "The final publication is available at link.springer.com”.

1 23 Author's personal copy

Hydrobiologia

https://doi.org/10.1007/s10750-019-3935-6 (0123456789().,-volV)(0123456789().,-volV)

PRIMARY RESEARCH PAPER

Cryptic diversity in Telestes pleurobipunctatus (Actinopterygii; Leuciscidae) as a consequence of historical biogeography in the Ionian Freshwater Ecoregion (Greece, Albania)

Ivana Buj . Radek Sˇanda . Stamatis Zogaris . Jo¨rg Freyhof . Matthias F. Geiger . Jasna Vukic´

Received: 14 August 2018 / Revised: 6 March 2019 / Accepted: 8 March 2019 Ó Springer Nature Switzerland AG 2019

Abstract It was reported that the complex geolog- corroborates the complexity of historical geologic ical history of the southern Balkans caused diversifi- events in the area. Telestes pleurobipunctatus com- cation of terrestrial species. On the other hand, prises two main clades (northern and southern), which knowledge of the evolutionary history of freshwater have been separated from each other for about seven biota is scarce. In order to contribute to an under- million years and that are further divided. The Louros standing of the historical biogeography of the region’s and Arachthos Rivers were denoted as ancestral freshwaters, as well as to describe intraspecific geographic ranges of the northern clade. The distri- diversity and structure of a species distributed in the bution range of the ancestor of the southern clade Ionian Ecoregion, we have analysed the population seems to comprise rivers flowing into the Patraikos genetic structure and evolutionary history of Telestes Gulf. Structures within the two main clades, as well as pleurobipunctatus. Analyses were based on the mito- timing of diversification events are different, implying chondrial gene for cytochrome b from samples separate evolutionary courses. covering the whole distribution range of this species. A high level of intraspecific structuring was revealed Keywords Conservation priorities Á Geological inside T. pleurobipunctatus. Its evolutionary history history Á Mediterranean biodiversity Á Population genetic structure Á Telestes

Handling editor: Diego Fontaneto

I. Buj (&) J. Freyhof Department of Zoology, Faculty of Science, University of German Centre for Integrative Biodiversity Research Zagreb, Rooseveltov trg 6, 10000 Zagreb, Croatia (iDiv) Halle-Jena-Leipzig, Deutscher Platz 5a, e-mail: [email protected] 04103 Leipzig, Germany

R. Sˇanda M. F. Geiger National Museum, Va´clavske´ na´meˇstı´ 68, 115 79 Prague, Zoologisches Forschungsmuseum Alexander Koenig, Czech Republic Adenauerallee 160, 53113 Bonn, Germany

S. Zogaris J. Vukic´ Hellenic Centre for Marine Research, Institute of Marine Department of Ecology, Faculty of Science, Charles Biological Sciences and Inland Waters, Athens-Sounio, University, Vinicˇna´ 7, 128 44 Prague, Czech Republic 19013 Anavissos, Attiki, Greece

123 Author's personal copy

Hydrobiologia

Introduction might have allowed connections among land masses (presumably between Peloponnese, Corfu and other Historical isolation of the Balkan Peninsula from the islands with the mainland), as well as among adjacent remaining parts of Europe, starting from the Oligocene river basins due to lowering of the sea level (Maurakis with the uplift of the Dinarides–Hellenides range, and et al., 2001). continuing during most of the Neogene (Ro¨gl, The Ionian Freshwater Ecoregion, described as 1998, 1999), induced the emergence of many endemic containing high levels of heterogeneity and harbour- species (Economidis & Banarescu, 1991). Diversifi- ing many endemic species (Oikonomou et al., 2014), cation events were furthermore provoked by a com- comprises an area from the extreme southern tip of plex geological history. The southern Balkan Albania to western Greece and most of the Pelopon- Peninsula has been exposed to significant tectonic nese (Zogaris & Economou, 2017). It is characterized events, volcanism and sea level changes from the by typical karstic features, canyons, small lakes and Oligocene onwards (Ro¨gl, 1998). As a consequence of many small river basins (Oikonomou et al., 2014). movements of the African and Anatolian plates during Economidis and Banarescu (1991), as well as the early Miocene (around 23 million years ago; Banarescu (2004) considered the region of the Ionian ¨ MYA) a continuous landmass (so-called Aga¨is) watershed to be a part of the wider Adriatic-Ionian appeared at the northern coast of the Tethys Ocean. Ecoregion; however, Zogaris et al. (2009a) and In the Middle and Late Miocene (around 12–8 MYA) Oikonomou et al. (2014) denote it as a separate tectonic disturbances provoked the break-up of the ecoregion, based on the freshwater fish diversity. The southern Aegean landmass and initiated the modern evolutionary history of the Ionian Ecoregion was biogeographic history of that area (Poulakakis et al., complex, with palaeo-lakes existing in several areas 2005b). In the same time period, the Mid-Aegean (Oikonomou et al., 2014). Some of the boundaries of Trench (MAT) was formed, beginning from the east of this region have still not been defined with certainty Crete and dividing the Aegean into Asian and (Zogaris & Economou, 2017). European sides (Poulakakis et al., 2005b). During There are several studies showing that the complex the latest part of the Messinian (MSC; 5.96–5.3 geological history of the southern Balkans contributed MYA), the level of the Mediterranean Sea dropped to the diversification and shaping of the distribution (so-called Messinian Salinity Crisis; MSC). Accord- areas of many terrestrial (Beerli et al., 1996; ing to Bianco (1990), it was subsequently filled with Sfenthourakis & Legakis, 2001; Poulakakis et al., freshwater from the Paratethys so it might have 2003, 2005a) and plants (Dimopoulos et al., 2016). allowed dispersal of freshwater fishes. Even though However, knowledge on the evolutionary history of there is a lack of evidence on the freshwater phase of freshwater fishes, which have different dispersal ways the Mediterranean Sea during the MSC (Roveri et al., than terrestrial animals or plants, is scarce (Economou 2014), some authors recognized MSC as the most et al., 2007). The Balkan Peninsula, with more than important event causing a great level of endemism in 193 native freshwater fish species, has the highest southern Greece (Zardoya & Doadrio, 1999). How- ichthyofaunal species richness in Europe (Oikonomou ever, there is evidence that older events played even a et al., 2014). In the Aegean-Anatolian realm, there are more important role (Perea et al., 2010). Uplift of the evidences of rich fish community from the Early Pindos Mountains was further reported to mark Miocene (Vasileiadou et al., 2017). In the area of division of zoogeographic regions on land, but also Lapsarna (on the Levsos Island), fossils of three the separation of the Aegean and the Ionian water- cyprinid species (Paleoleuciscus sp., Barbus sp. and sheds (Economou et al., 2007). Geologic isolation of Cyprinion sp.) were discovered in lacustrine sedi- the Peloponnese by the sea, presumably also signif- ments and marls of an Early Miocene origin icant for the evolutionary history of various animals, (Vasileiadou et al., 2017). Existence of palaeo-lakes probably occurred 4–3 MYA (Sola et al., 2013)or in the Ionian Ecoregion (Oikonomou et al., 2014), as about 2.5 MYA (Zardoya & Doadrio, 1999; Dubut well as survival of many old fish lineages (Economou et al., 2012). Another event proposed to have shaped et al., 2007) have been reported. recent diversity of the Greek fauna were the Pleis- The fish Telestes belongs to the Leuciscidae tocene glaciations (2.58–0.12 MYA). Glaciations family and subfamily (Scho¨nhuth et al., 123 Author's personal copy

Hydrobiologia

2018) and represents a model for investigating the knowledge is important for conservation planning as historical biogeography of European freshwater drai- well. nage systems (Buj et al., 2017). In the south-western Balkans, it is represented with two species: T. pleurobipunctatus (Stephanidis, 1939) and T. beoticus Materials and methods (Stephanidis, 1939) (Kottelat & Freyhof, 2007). Together with Chondrostoma s.l., Phoxinellus and Analyses of T. pleurobipunctatus were based on 114 Pseudophoxinus, Telestes belongs to the linage VI in individuals from 13 localities (number of individuals the molecular phylogeny of the subfamily Leuciscinae per locality = 4–13), covering the whole geographic (Perea et al., 2010). This lineage diverged around range of the species (Table 1, Fig. 1) and were based 20.86 MYA (Perea et al., 2010), whereas divergence on the mitochondrial gene for cytochrome b (cyt b). of the genus Telestes was estimated to have started The cyt b gene was chosen due to its suitability for about 16 MYA, with the majority of recent species taxonomic and population genetic studies as proven in occurring in the Adriatic basin (Buj et al., 2017). several previous investigations on various vertebrates Telestes pleurobipunctatus, known as the Epiros (e.g. Tsigenopoulos & Berrebi, 2000; Castresana, Riffle Dace, is a small freshwater fish distributed from 2001; Farias et al., 2001; Tobe et al., 2010). Moreover, southern Albania, throughout western Greece to the it was reported as the most useful marker in recovering central western part of the Peloponnese Peninsula, phylogenetic relationships among closely related taxa from Butrint to Alfios drainages, including Corfu (Castresana, 2001; Patwardhan et al., 2014), therefore (Kerkyra) Island (Kottelat & Freyhof, 2007). Its range being adequate for purposes of this investigation. The is entirely within the Ionian Freshwater Ecoregion rate of mitochondrial DNA evolution is several-fold (Zogaris et al., 2009a; Barbieri et al., 2015). This higher than of typical nuclear DNA (Avise, 1986), species was already identified as an example of providing a larger number of parsimony informative allopatric subclades with pronounced geographic sites and a lower consistency index in mitochondrial structuring of haplotypes (Geiger et al., 2014), imply- genes. ing its adequacy for investigating historical biogeog- The sampling was conducted by electrofishing. A raphy. Previous studies that included T. small piece of fin tissue was preserved in ethanol and pleurobipunctatus (e.g. Perea et al., 2010; Geiger vouchers are stored in the National Museum in Prague et al., 2014; Buj et al., 2017) investigated wider fish and the Zoologisches Forschungsmuseum A. Koenig, groups (perimediterranean Leuciscinae, Mediter- Bonn. Extraction of the total genomic DNA was ranean freshwater fishes, genus Telestes, respectively) performed using a standard extraction product and did not focus on T. pleurobipunctatus and (DNeasy tissue kit, Qiagen). For polymerase chain thereafter, did not reveal its intraspecific diversity in reaction (PCR) amplifications of the cyt b gene we detail or implications for the biogeography (either used primers GluF and ThrR (Machordom & Doadrio, historical or recent) of the region. 2001), whereas internal primers CB4-GLU (50 CCT Revelation and description of the intraspecific GAA AYA TYG GYG TRGT 30) and PHOX-THR (50 structure and diversity of T. pleurobipunctatus,as AGG AGG AAR TGR AAT GCG AA 30) were well as a detailed explanation of its historical origin, is employed for sequencing. The PCR amplification a principal goal of this study. Due to its distribution protocol and PCR product purification follow Sˇanda range concordant with the Ionian Ecoregion and et al. (2008). Obtained cyt b sequences were aligned already observed structuring of haplotypes, the manually and visually checked for gaps and stop obtained results may contribute to our understanding codons. In order to test whether all mutations were of the historical biogeography of this region and selectively neutral, statistical tests D* and F* (pro- denotation of fine scale biogeographical subdivisions. posed by Fu & Li, 1993) and that of Tajima (Tajima, Furthermore, revealing the evolutionary history of a 1989) were performed using DnaSP v.5. (Librado & certain species and reconstructing how it responded to Rozas, 2009). environmental changes in the past can help predict its Phylogenetic reconstruction was conducted using response to anthropogenic changes in the future. This three methods: maximum parsimony (MP) imple- mented in PAUP (version 4.0b10; Swofford, 2002), 123 Author's personal copy

Hydrobiologia

Table 1 Sampling localities, geographic coordinates and GenBank accession numbers of Telestes pleurobipunctatus samples included in the investigation Sampling locality (river or lake) Locality code (as Country Coordinates Number of Haplotype Accession on the Fig. 1) samples codes numbers

Pavllo 1 Albania 39°4103400N 6 ALB1 MK585354 20°0605500E ALB2 MK585355 Bistrice 2 Albania 39°5402700N 4 ALB3 MK585356 20°0603100E ALB4 MK585358 LOU1 MK585357 Corfu 3 Greece 39°2803700N 5 KER1 MK585379 19°5404200E Kalamas 4 Greece 39°4103900N 10 KAL1 MK585359 20°3504600E KAL2 MK585360 KAL3 MK585361 KAL4 MK585362 KAL5 MK585363 KAL6 MK585364 LOU1 MK585357 Arachthos 5 Greece 39°0905600N 13 ARAC1 MK585373 20°5903600E ARAC2 MK585374 ARAC3 MK585375 ARAC4 MK585376 ARAC5 MK585377 LOU2 MK585378 Louros (near Aghios Georgios village) 6 Greece 39°1601600N 7 LOU1 MK585357 20°5005400E LOU7 MK585402 LOU8 MK585403 LOU9 MK585404 Louros (Kampi spring) 6 Greece 39°130600N 9 LOU2 MK585378 20°540800E LOU3 MK585398 LOU4 MK585399 LOU5 MK585400 LOU6 MK585401 LOU10 MK585397 Acheron 7 Greece 39°2605000N 9 ACHR1 MK585365 20°300040’E ACHR2 MK585366 ACHR3 MK585367 KAL5 MK585363 Acheloos 8 Greece 39°0000600N 11 ACHL1 MK585368 21°3402200E ACHL2 MK585369 ACHL3 MK585370 ACHL4 MK585371 ACHL5 MK585372

123 Author's personal copy

Hydrobiologia

Table 1 continued Sampling locality (river or lake) Locality code (as Country Coordinates Number of Haplotype Accession on the Fig. 1) samples codes numbers

Pinios 9 Greece 37°5003800N 12 PIN1 MK585386 21°3201900E PIN2 MK585387 PIN3 MK585388 PIN4 MK585389 PIN5 MK585390 PIN6 MK585391 PIN7 MK585392 PIN8 MK585393 Evinos 10 Greece 38°2701000N 10 EVIN1 MK585380 21°4202900E EVIN2 MK585381 EVIN3 MK585382 EVIN4 MK585383 EVIN5 MK585384 EVIN6 MK585385 Kotichi 11 Greece 37°5900300N 8 KOT1 MK585394 21°2502500E KOT2 MK585395 KOT3 MK585396 Alfios 12 Greece 37°3802000N 10 ALF1 MK585405 21°3405200E ALF2 MK585406 ALF3 MK585407

Bayesian inference (BI) implemented in MrBayes Namely, MJ method constructs a phylogenetic net- (version 3.1.2; Huelsenbeck & Ronquist, 2001), and work that enables visualization of genealogical rela- median-joining (MJ) using Network 4.5.1.6. software tionships among closely related haplotypes together (Fluxus Technology Ltd.). For MP analysis, a heuristic with alternative scenarios and considers presence of search mode with 100 replicates was used, with ancestral haplotypes in the recent genetic structure. randomized input orders of taxa, and TBR branch Divergence time estimation was conducted using swapping with all codon sites and nucleotide substi- the Bayesian MCMC coalescent method implemented tutions types weighted equally. Nonparametric boot- in BEAST v1.7.0. (Drummond et al., 2012). The strict strapping (1000 pseudo-replicates, 10 additional molecular clock was applied, based on the results of sequence replicates) was used to assess branch support the log-likelihood ratio test (LRT) performed using (BS). Each BI analysis consisted of two simultaneous PAUP* (the likelihood scores were the same with and runs. For each, Markov Chain Monte Carlo was run without molecular clock enforcement). The molecular four times for three million generations with trees clock calibration was based on the divergence rate of sampled every 100 generations. The first 20% of the cyt b gene in Leuciscinae of 0.4% per lineage per sampled trees were discarded and Bayesian posterior million years, as concluded by Perea et al. (2010). probabilities (BPP) were estimated from the 50% Branch rates were drawn from an uncorrelated majority-rule consensus tree of the retained trees. For lognormal distribution and a Yule speciation prior rooting of phylogenetic trees, sequences of T. beoticus with random starting tree. Substitution model applied and Squalius squalus (Bonaparte, 1837) were used as was HKY with Gamma site heterogeneity model. We outgroups. Median-joining (MJ) algorithm was used default prior distributions for kappa, frequencies employed as an additional method, since we have and alpha, whereas substitution rate parameters were analysed phylogenetic relationships among haplo- unlinked across codon positions. The number of types presumably belonging to a single species. MCMC steps (the length of chain) was three millions. 123 Author's personal copy

Hydrobiologia

for ancestral range reconstruction was obtained by BI analysis, and seven recent geographic ranges of T. pleurobipunctatus were denoted, as follows: • A rivers in southern Albania, the northernmost part of T. pleurobipunctatus’ range (Pavllo, Bistrice) • B Corfu island • C northern part of the Ionian watershed in Greece (Acheron, Kalamas) • D central part of the Ionian watershed (Louros, Arachthos) • E Acheloos • F Evinos R. with the mouth in the Patraikos Gulf • G Peloponnese Rivers (Kotichi, Pinios, Alfios). Pairwise comparison of uncorrected sequence divergence (p-distances) of the investigated cyt b gene was analysed using MEGA version 3.1 (Tamura et al., 2007). In order to check whether interpopulational p-distances are significantly higher than intrapopulational, we have performed Permuta- tion test of variance (PERMANOVA), using PAST 3.22 software, with 9999 permutations. Using DnaSP we calculated the following measures of intrapopulational genetic diversity: number of polymorphic sites (S), haplotype diversity (Hd), average number of nucleotide differences (K), nucleo- tide diversity (p). Employing the same software estimation of the level of genetic differentiation between populations was conducted, by calculating 2 statistics based on haplotypes (v test and HST) and the * * ones based on nucleotide sequences (KST, KST, KS, Z and Z*). All calculated statistics are widely used in Fig. 1 a Collection sites and river basins hosting Telestes population genetic investigations, for detecting pleurobipunctatus (1 Pavllo, 2 Bistrice, 3 Corfu, 4 Kalamas, 5 genetic differentiation among populations and sub- Arachthos, 6 Louros (two collection sites are on the Louros R., populations (e.g. Ballard, 2000;Bo¨hme et al., 2006; see Table 1), 7 Acheron, 8 Acheloos, 9 Pinios, 10 Evinos, 11 Kotichi, 12 Alfios). The dotted line defines the Valtos- Hall, 2006; Brazee & Lindner, 2013; Buj et al., 2017). Akarnanika-Cephalonia watershed boundary where there has The traditional method for testing genetic differenti- been a long-term separation of major basins and the two distinct ation of populations is v2 test (Hudson et al., 1992). It clades. b General orientation map was proposed for detecting genetic variation by testing allele frequencies in samples from different localities In order to estimate the geographic range of (Nei, 1983). It was later adapted to use with nucleotide ancestral populations, Statistical Dispersal-Vicariance variation by treating each distinct haplotype as an Analysis implemented in S-DIVA software (Yu et al., allele (Hudson et al., 1992) and as such it was used in 2010) was employed. This method reconstructs the this investigation (for mathematical equation see ancestral distribution in a phylogeny by optimizing a Hudson et al., 1992). Another haplotype statistic three-dimensional cost matrix, in which extinctions employed is HST, which is based on the weighted and dispersals ‘cost’ more than vicariance, and it average of estimated haplotype diversities in popula- determines statistical support for ancestral range tions and estimate of the haplotype diversity in the reconstruction (Yu et al., 2010). An input set of trees total sample (for mathematical equation see Hudson 123 Author's personal copy

Hydrobiologia

2 et al., 1992). The above-mentioned v test and HST cyt b haplotypes, with overall haplotype diversity of statistics are based on haplotype frequencies in the 0.98. Neutrality tests suggested no deviation from sample and do not account the number of differences mutation-drift equilibrium in all populations and the between sequences. The remaining statistics (so- overall data set. called sequence statistics), on the other hand, use the For the purposes of this investigation, we named the information on the number of differences between smallest phylogenetic groups inside the structure of T. sequences (Hudson et al., 1992). Statistic KST incal- pleurobipunctatus as lineages, whereas divisions of culates weighted average of the number of differences higher order are named clades. MP and BI phyloge- between sequences from different localities divided by netic methods yielded trees with similar overall the average number of differences between sequences, topology (Fig. 2). Separation of samples from north * * regardless of their locality; statistics KST and KS are and south parts of T. pleurobipunctatus distribution similar to KST and KS (statistic that gives weighted range into two well-supported clades is obvious. The average of the average number of differences between northern clade comprises samples from populations sequences from two localities), but give less weighting located in southern Albania and the northwest part of to large numbers of differences; Z and Z* statistics Greece, namely from river basins of Pavllo, Bistrice, take into consideration a weighted sum of averages of Messongi (on the Corfu Island), Kalamas, Acheron, the ranks of the function that denotes the number of Louros and Arachthos. The southern clade comprises differences between sequences from different locali- samples from the river basins Acheloos, Evinos, ties (for mathematical equations and applications Pinios, Kotichi (including the Vergas stream) and considerations see Hudson et al., 1992). The null Alfios. The northern clade is further separated into two hypothesis that there is no genetic diversity between lineages (N1 and N2), whereas inside the southern two populations was rejected based on the permutation clade each population forms an independent lineage. test, if P \ 0.05. Since most estimates of genetic Even though two northern lineages (N1 and N2) are differentiation are not useful test statistics because genetically distinct and do not overlap in large parts of their distribution under the null hypothesis is not their distribution ranges, it is noteworthy that fishes known, permutation-based method using Monte Carlo belonging to both lineages were found in the Louros simulations to estimate significance levels is usually River, but at different places within the river basin. employed (Hudson et al., 1992), as was also done in Phylogenetic network obtained by MJ algorithm this investigation. Interactions among populations revealed the same structuring (Fig. 3). Overall, there were further investigated using maximum likelihood is a large proportion of private haplotypes, found only approach (Beerli, 1998; Beerli & Felsenstein, 2001) in a single river (50/54). Only four haplotypes (all implemented in MIGRATE 3.2.1 (Beerli, 2009). We belonging to the northern clade) are found on more estimated immigration rates as mutation-scaled effec- than one locality: haplotype ALB2 is shared between tive immigration rates (M), which measures the two Albanian populations, LOU1 is the only haplotype importance of immigrants vs. mutation when bringing found on more than two localities (Pavllo, Bistrice, new variants into the population. Kalamas, Louros), LOU2 exists in Louros and Since our investigation revealed a high level of Arachthos, and KAL5 was found in Kalamas and genetic structuring inside T. pleurobipunctatus,we Acheron. identified fixed differences between lineages using Timing of diversification events, as well as ances- SITES (Hey & Wakeley, 1997) that might be useful in tral geographic range reconstructions, are marked on molecular diagnostics. the Fig. 4. The separation of the two main T. pleurobipunctatus clades can be dated back to the late Miocene, whereas other diversification events Results occurred during the Pliocene and Pleistocene. All S-DIVA analyses identified the central part of the Obtained sequences were 1140 base pairs (bp) long, Ionian watershed (the Louros and Arachthos Rivers) and neither insertions nor deletions were recorded. as containing ancestral T. pleurobipunctatus popula- Inside the total sample of 114 sequences belonging to tions. However, besides in that area, it is very likely T. pleurobipunctatus, we have recognized 54 different that ancestral population persisted also in the 123 Author's personal copy

Hydrobiologia

Fig. 2 MP phylogram of cyt b haplotypes from watersheds covering distribution range of Telestes pleurobipunctatus. Numbers at nodes represent MP BS and BPP values 123 Author's personal copy

Hydrobiologia

Fig. 3 MJ network presenting genealogical relationships among investigated haplotypes. Circle size is proportional to haplotype frequency, the number of mutations is displayed along the branches in cases of more than five mutations

Peloponnese Rivers and possibly also in Acheloos Tests of interpopulational genetic differentiation, (lower probability). The two lineages belonging to the as well as gene flow estimations, revealed genetic northern clade are of Pliocene origin (their separation distinctiveness of the majority of populations and occurred about 3.23 MYA), and their ancestor was restricted or absent gene flow. The null hypothesis present in the area of the Louros and Arachthos Rivers. stating the absence of genetic distinctiveness between The ancestor of the southern clade seems to have had a populations was accepted only for two Albanian wider distribution inside the Peloponnese Rivers, as populations (Pavllo and Bistrice; P = 0.24-0.38). All well as the Acheloos River. Inside the southern clade remaining populations are genetically distinct one independent evolutionary course of lineages from from another based on all tests of genetic differenti- Acheloos and Alfios also started in Pliocene. ation employed (P \ 0.05). Samples from each river Interpopulational p-distances are significantly were considered as a single population. Only in the higher than intrapopulational (PERMANOVA, Louros River, samples from two localities are also P \ 0.05). Differences between northern and southern genetically distinct and belong to two phylogenetic clade are especially pronounced (3.7–5.8%). Further- lineages. The results for Corfu population might not be more, the population from Acheloos is also distinct completely reliable due to low sampling number. In from the remaining southern populations (p-distances accordance with the results of the genetic differenti- range between 2.8 and 4.0%). Intrapopulational ation tests, we detected gene flow among the Albanian p-distances, on the other hand, range between 0.1 populations, from the Pavllo to Bistrice (maximum and 0.9%, with highest values observed for Pinios and likelihood estimation of M = 7480). Furthermore, Evinos populations. Pavllo and Kalamas seem to have been connected with migration events (M = 4780), with the Pavllo

123 Author's personal copy

Hydrobiologia

Fig. 4 Divergence time and ancestral geographic range esti- watershed in Greece (Acheron, Kalamas), D central part of the mations of Telestes pleurobipunctatus lineages. Timing of Ionian watershed (Louros, Arachthos), E Acheloos River, F the splitting events is presented at nodes by mean value and 95% Evinos River, G Peloponnese rivers (Kotichi, Pinios, Alfios). credibility range in brackets (in million years). Ancestral ranges The proportion in the pie denotes probability of a certain are denoted by circles by nodes; C northern part of the Ionian ancestral range population being the source of immigrants and Kala- differences between the Alfios population and the mas the receiving population. Our results further Evinos, Kotichi and Pinios populations. Among imply that the migrations were an important event for populations from rivers Pinios, Evinos and Kotichi, entering genetic variation into the Louros population the number of fixed differences is much smaller (2–6). with Bistrice and Arachthos populations identified as Measures of intrapopulational genetic polymor- sources of immigrants (M = 2050 and M = 2420, phism revealed various levels of diversity (Table 4). respectively), even though those rivers are not con- Besides the population from Corfu, in which a single nected any more. No gene flow was recorded among haplotype was found, lower genetic diversity is populations belonging to the southern clade (Table 2). characteristic for the Kotichi River basin population Concordant with all previous results are SITES (Hd = 0.464; K = 0.679). Populations from Acheron, outcomes (Table 3). Between lineages of the two main Arachthos, Acheloos and Alfios Rivers express mod- clades there are 40–58 fixed differences, whereas the erate genetic diversity, whereas the highest amount of two lineages of the northern clade differ in 17 fixed intrapopulational genetic polymorphism is contained nucleotide substitutions. The Acheloos population inside the Pavllo, Bistrice, Kalamas, Pinios and Evinos differs in 31–36 nucleotide sites from the remaining River populations. lineages in the southern clade. There are 17–19 fixed

123 Author's personal copy

Hydrobiologia

Table 2 Ranges and mean values (in brackets) of the p-distances among T. pleurobipunctatus populations and phylogenetic groups, as well as the intraspecific p-distances N1 N2 Acheloos Evinos Kotichi Pinios Intrapop.

N1 0.1–0.8 (0.4) N2 1.9–2.5 (2.2) 0.1–0.5 (0.3) Acheloos 4.6–5.4 (5.0) 5.1–5.8 (5.4) 0.1–0.4 (0.2) Evinos 4.6–5.5 (4.9) 5.3–6.1 (5.5) 3.3–4.0 (3.5) 0.1–0.9 (0.4) Kotichi 4.5–5.1 (4.8) 5.2–5.7 (5.4) 2.9–3.2 (3.1) 0.4–1.1 (0.6) 0.1–0.2 (0.1) Pinios 4.3–5.1 (4.7) 4.9–5.5 (5.2) 3.2–3.7 (3.5) 0.7–1.5 (0.9) 0.4–0.9 (0.7) 0.1–0.7 (0.2) Alfios 3.7–4.4 (4.0) 4.0–4.6 (4.3) 2.8–3.1 (2.9) 1.8–2.5 (2.0) 1.6–1.9 (1.8) 1.5–2.0 (1.8) 0.1–0.3 (0.2) N1 the first lineage inside the northern clade (comprises samples from Pavllo, Bistrice, Louros, Kalamas, Corfu and Acheron); N2 the second lineage of the northern clade (comprising samples from Arachthos and Louros)

Table 3 Number of fixed differences in pairwise comparison also separated; the northern clade is distributed in of identified lineages inside T. pleurobipunctatus southern Albania and north-western Greece and N2 Acheloos Evinos Kotichi Pinios Alfios comprises rivers flowing to the northern part of the Ionian Sea, whereas the southern clade corresponds to N1 17 50 47 48 45 40 south-western Greek river basins and rivers surround- N2 58 57 58 55 45 ing the Patraikos Gulf (Peloponnese rivers and main- Acheloos 35 33 36 31 land rivers flowing into the Patraikos Gulf). Previous Evinos 2 6 19 investigation (Buj et al., 2017) proved that no other Kotichi 5 18 currently recognized Telestes species is represented in Pinios 17 the Ionian sea watershed, all observed clades and lineages belong to T. pleurobipunctatus, although, based on the results of this investigation, it contains a very high amount of diversity and pronounced Discussion structuring. Separation of the two main T. pleurobipunctatus All results obtained reveal a high level of intraspecific clades might have been provoked by the uprising of structuring of T. pleurobipunctatus and the existence the western branch of the Pindos Mountains, which, of several evolutionary lineages. High intraspecific thereafter, defines two biogeographic subregions diversity is most likely due to a long independent inside the Ionian Freshwater Ecoregion: northern and evolution of several lineages (originating from the southern. The border between them is located just Miocene) and corroborates the complexity of histor- south to the Amvrakikos Gulf, between the Arachthos ical tectonic activity in the area. and Acheloos river basins and continuing to the near shore continental islands of Lefkada and Cephalonia Genetic structure and biogeography of T. (the division marked on the Fig. 1). The genetic pleurobipunctatus with implications for defining differentiation that can be connected with biogeo- freshwater ecoregions in Greece graphic subregions as denoted here was found also inside some other freshwater fishes, implying the Phylogeographic structure of T. pleurobipunctatus and possibility that the observed pattern is more widely its division into two independent clades are corrobo- present in this ecoregion’s freshwater fauna. Results of rated by all analyses: p-distance between two clades is genetic investigations of another Leuciscidae genus high, they are genetically distinct based on tests of Squalius from western Greece are complementary genetic differentiation, gene flow between them is with our finding. Squalius populations from the rivers absent, and their ancestors were distributed in different Kalamas, Acheron, Louros and Arachthos belong to areas. Recent distribution ranges of the two clades are

123 Author's personal copy

Hydrobiologia

Table 4 Intrapopulational genetic polymorphism of T. pleurobipunctatus populations Nh S HdK p f (%)

Pavllo 6 4 4 0.8 1.733 0.00152 ALB1: 17; ALB2: 50; ALB3: 17; LOU1:17 Bistrice 4 3 3 0.833 1.667 0.00146 ALB2: 25; ALB4: 25; LOU1: 50 Corfu 5 1 0 0 0 0 KER1: 100 Kalamas 10 7 9 0.867 2.867 0.00251 KAL1: 10; KAL2: 10; KAL3: 10; KAL4: 10; KAL5: 40; KAL6: 10; LOU1:10 Arachthos 13 6 7 0.795 1.718 0.00151 ARAC1: 8; ARAC2: 8, ARAC3: 38; ARAC4: 8; ARAC5:31; LOU2: 8 Louros 16 10 32 0.9 13.858 0.01216 LOU1: 31; LOU2: 13; LOU3: 6; LOU4: 6; LOU5: 6; LOU6. 6; LOU7: 13; LOU8: 6; LOU9: 6; LOU10: 6 Acheron 9 4 8 0.583 2.778 0.00244 ACHR1: 67; ACHR2: 11; ACHR3: 11; KAL5: 11 lineage N1 43 19 24 0.918 3.838 0.00337 ALB1: 2; ALB2: 9, ALB3: 2; ALB4: 2; LOU1: 21; KAL1: 2; KAL2: 2; KAL3: 2; KAL4: 2; KAL5: 11; KAL6: 2; ACHR1: 14; ACHR2: 2; ACHR3: 2; KER1: 11, LOU7: 5; LOU8: 2; LOU9: 2 lineage N2 20 11 13 0.9 2.516 0.00221 ARAC1: 5; ARAC2: 5; ARAC3: 20; ARAC4: 5; ARAC5: 20; LOU2: 15; LOU3: 5; LOU4: 5; LOU5: 5; LOU6: 5; LOU10: 5 Acheloos 11 5 6 0.764 1.345 0.00118 ACHL1: 27; ACHL2: 9; ACHL3: 46; ACHL4: 9; ACHL5: 9 Pinios 12 8 9 0.924 2.136 0.00187 PIN1: 8; PIN2: 8; PIN3: 25; PIN4: 8; PIN5: 17; PIN6: 8; PIN7: 8; PIN8: 17 Evinos 10 6 11 0.889 3.422 0.003 EVIN1: 10; EVIN2: 10; EVIN3: 20; EVIN4: 20; EVIN5: 10; EVIN6: 30 Kotichi 8 3 2 0.464 0.679 0.0006 KOT1: 75; KOT2: 13; KOT3: 13 Alfios 10 3 3 0.622 1.067 0.00094 ALF1: 60; ALF2: 20; ALF3: 20 N number of sequences, h number of haplotypes, S number of polymorphic sites, Hd haplotype diversity, K average number of nucleotide differences, p nucleotide diversity, f haplotype frequency one mitochondrial lineage, whereas populations from where both of them are found. Dubut et al. (2012) Acheloos, Evinos and the rivers of western Pelopon- concluded that the two populations in the Louros River nese to another lineage (Durand et al., 2000; Geiger are reproductively isolated and explained observed et al., 2014; Perea et al., 2016). Similar geographic situation by parapatric speciation. Our results are division of genetic diversity was also observed in ambiguous regarding divergence mode of the two Luciobarbus albanicus (Steindachner, 1870) (Cypri- lineages found in the Louros River. We revealed that nidae) by Geiger et al. (2014), where samples from the ancestral populations of both northern lineages as well Kalamas and Arachthos rivers form one mitochondrial as the whole northern clade were distributed in an area lineage, whereas samples from the rivers Acheloos comprising the Louros River, which implies the and Pinios (on Peloponnese) belong to another possibility of parapatric speciation. However, we lineage. Furthermore, analysis of genetic variability demonstrated that two lineages, whose representatives of freshwater killifish genus Valencia (Valenciidae, were found in the Louros River, have a broader Cyprinodontiformes) (Vogiatzi et al., 2014) also geographic distribution, so the secondary contact pointed to similar biogeographic subdivisions. It is scenario is also possible. Further evolutionary inves- important to mention that, due to lower number of tigations are required to resolve this conundrum. sampling locations included in Vogiatzi et al. (2014), Each population in the southern clade forms a results are not completely comparable. However, they separate lineage; different river basins are inhabited by point to similar biogeographic pattern, with the first evolutionary distinct populations. Especially pro- clade of Vogiatzi et al. (2014) falling within our nounced is the differentiation of the Acheloos (main- northern clade, and second clade being similar to our land Greece) and the Alfios (Peloponnese) populations southern clade. from the remaining ones. The two lineages of the northern clade of T. Most of T. pleurobipunctatus populations contain pleurobipunctatus occur in separate drainages, with a only private haplotypes, a result that corroborates their single exception being the Louros River drainage, distinctiveness and isolation. Gene flow and 123 Author's personal copy

Hydrobiologia divergence time estimations also corroborate the Ketmaier et al., 2004) or the opening of the Strait of presumption of longer isolation of the majority of T. Gibraltar, caused diversification inside T. pleuro- pleurobipunctatus populations. The only exceptions bipunctatus. This may be explained by the remarkably are some populations that belong to N1 and N2 steep Hellenic trench that lies immediately off the lineages and that are geographically located closer to coast of western Greece, but it is also possible that the each other (Pavllo and Bistrice; Louros, Arachthos, Messinian Salinity Crisis did not have the proposed Kalamas and Acheron). effect on the freshwater ichthyofauna, which is Even though the high diversity and levels of concordant with other recent investigations of fresh- endemism have already been described in the fresh- water fishes in the Mediterranean area (Perea et al., waters of the western Balkans (Oikonomou et al., 2010; Buj et al., 2015, 2017). Furthermore, we did not 2014), our results acknowledge the existence of great find evidence of Pleistocene migrations among Ach- diversity among populations that are considered to eloos and the remaining populations, as previously belong to a single species. This hidden diversity was proposed (Guinand et al., 2001). Based on our results, revealed on a molecular genetic level in this investi- the isolation of Acheloos occurred in the Pliocene. It gation, whereas detail analysis of morphology is is, nevertheless, probable that Evinos River and the required to test whether this molecular diversity is Peloponnesian basins of Pinios and Kotichi came in reflected by morphological variation as well. Molec- contact during Pleistocene glaciations, when sea levels ular uniqueness and the ancient origin of several were approximately 200 m lower than today (Peris- lineages comprised under T. pleurobipunctatus imply soratis & Conispoliatis, 2003). the possibility that T. pleurobipunctatus is actually a Geological events that have contributed to the complex of species and invoke the necessity of a vicariance and isolation in the southern part of the comprehensive taxonomic investigation. Ionian region (reviewed in Economou et al., 2007) include the early isolation of the southern part of the Impact of the described geologic events Peloponnese by mountain barriers and a deep marine on the evolutionary history of T. pleurobipunctatus trench close to the coast; the separation of the Peloponnese from the mainland by the opening of The separation of the two T. pleurobipunctatus clades the Corinthian Gulf during the early part of the Late is dated back to around 7 MYA (5.6-8.7), a period after Pliocene; the progressive uplift of the Ionian islands the formation of the Mid-Aegean Trench, with throughout the Pliocene, and the sea level regressions pronounced tectonic activity on the Greek mainland in the Patraikos Gulf during the Holocene. Since two (Jolivet & Faccenna, 2000; Poulakakis et al., 2005b). Peloponnese populations (from the Kotichi and Pinios The uplift of the main part of the Pindos Mountains, Rivers) are more similar to the population from the proposed for the evolution of herpetofauna (Lymber- mainland Evinos River than to the population from the akis & Poulakakis, 2010) and freshwater planarians Alfios River, it is probable that during the Pleistocene (Sola et al., 2013), apparently did not provoke glacial sea regression, when there was land in today’s diversification inside T. pleurobipunctatus. Yet, this Patraikos Gulf, these rivers were connected (Perisso- event denoted the eastern border of the distribution ratis & Conispoliatis, 2003). This is also corroborated range of this species. However, based on the phylo- by the distribution of two other endemic Ionian fishes geographic structure and distribution of two main (Economou et al., 2007) that both exist in the Acheloos clades, it is possible that the uplift of the Valtos and Pinios, but not in the Alfios R. (i.e. Luciobarbus Mountains (western branch of the Pindos Mountains) albanicus and Tropidophoxinellus hellenicus and its extension western towards the Ionian islands (Stephanidis, 1971)). Complex changes took place in (marked on the Fig. 1) occurred in that period and the southern part of the Ionian Ecoregion and some are created a barrier to gene flow between two clades, not connected to freshwater fish dispersal; our results provoking their isolation. We did not find any imply that the formation of the Corinthian Gulf is not a evidence that the Messinian Salinity Crisis, proposed vicariant event for the genus Telestes and therefore as particularly significant for evolution of Greek cannot be used for molecular clock calibration. herpetofauna (Lymberakis & Poulakakis, 2010) and Moreover, the calibration based on the Corinthian diversification of freshwater taxa (Durand et al., 1999; Gulf opening seems to be responsible for incorrect 123 Author's personal copy

Hydrobiologia dating of diversification events in some previous rather long stretches from cool water springs near sea- papers (e.g. Dubut et al., 2012). level up to about 800 m elevation in the Acheloos Finally, contrary to the situation with many terres- River (Economou et al., 1999; Economou et al., 2006; trial species (for example see Lymberakis & Poula- Koutsikos et al., 2012). Since it prefers cool flowing kakis, 2010), the distribution of T. pleurobipunctatus water conditions with diverse habitats and it under- seems to be a result of its evolutionary history only, goes reproductive migrations, the general impairment and it is not likely that it was affected by anthro- of ecological integrity in springs, streams and rivers is pogenic translocations. This is concordant with find- posing threats to its viability. Furthermore, T. pleuro- ings of previous authors (Ketmaier et al., 2004). bipunctatus populations are potentially threatened by a variety of anthropologic pressures, including surface Genetic diversity and conservation and underground water abstraction, barriers to fish recommendations movement, dams, habitat alteration and pollution (Economou et al., 1999; Barbieri et al., 2002). Its If we consider the two northern lineages as separate habitats in the uplands are better conserved, but there entities, both of them express high levels of genetic are increasing impacts from various development diversity, based on all calculated measures, as well as projects, such as hydroelectric dams (Zogaris et al., high number and uniform frequencies of haplotypes. 2009b). Hybridization with other cyprinids in polluted Since intrapopulational and intraspecific genetic and degraded lowland stream reaches has been diversity is usually positively correlated with popula- documented (Bianco, 1988). Future climate change tion/species size, but also with mean fitness (Frank- and constituent warming of streams may impact ham et al., 2010), populations/species with greater habitats as is foreseen in ecohydrologic modelling genetic diversity have greater chance to adapt to for the Acheloos River (Papadaki et al., 2016). changed environmental conditions and survive. Con- Based on the obtained results, we were able to cordantly, our data are important for conservation identify units that are of exceptional importance for planning. Even though genetic diversity measures future evolution and survival of the whole species (or imply stability of two northern lineages, the level of species complex) and propose them as conservation intrapopulational diversity of almost all populations is priorities. Conservational units were not chosen as smaller. The causes of the extremely low genetic populations declining and facing the greatest risk of diversity on Corfu (only one haplotype found) should extinction, but rather as entities that possess genetic be further investigated. As already mentioned, results attributes significant for the present and future gener- obtained for the Corfu population might not be ations of species in question (Ryder, 1986), and are completely reliable, due to low number of samples. most likely to ensure the unconstrained evolutionary On the other hand, the highest genetic diversity found history of T. pleurobipunctatus. Such genetic attri- in the Louros River is a consequence of the presence of butes, employed here in the conservational units two well-differentiated lineages. Inside the southern identification, are: I. genetic distinctiveness—at least clade, low levels of genetic diversity are characteristic one population inside each genetically distinct unit for Kotichi and Alfios populations. These results should be protected; II. intrapopulational genetic imply that any anthropogenic impact will have greater diversity—populations with the highest level of effects on those populations, because their ability to genetic polymorphism inside each lineage should be cope with environmental changes is reduced. protected, because they are more likely to ensure Even though T. pleurobipunctatus is assessed as a viability of the lineage under changing conditions; Least Concern species in the IUCN Red List of and, III. populations of particular importance for Threatened Species (Crivelli, 2006), conservation lineage stability, e.g. sources of immigrants. actions should focus on distinctive populations. This According to these principles, the following pop- is based on the high level of independence of the ulations should be considered as distinct conserva- Telestes pleurobipunctatus populations and tional units: notable cryptic differentiation revealed in this inves- (1) Inside the northernmost lineage N1, Bistrice and tigation. All rivers inhabited by T. pleurobipunctatus Pavllo populations should be recognized and are relatively short, but the species is found along 123 Author's personal copy

Hydrobiologia

protected as conservation priorities, since they Krysˇtufek & J. M. Reed (eds), Balkan Biodiversity. are identified as sources of immigrants and also Springer, Dordrecht. Barbieri, R., A. N. Economou, M. Stoumboudi & P. S. Econo- contain the highest levels of genetic diversity. midis, 2002. Freshwater fishes of Peloponnese (Greece): (2) We propose the Arachthos population as a distribution, ecology and threats. Conservation of Fresh- conservation priority, since it is the only river water Fishes: Options for the Future. pp. 55–64. where only N2 lineage was observed. Barbieri, R., S. Zogaris, E. Kalogianni, M. Stoumboudi, Y. Chatzinikolaou, S. Giakoumi, Y. Kapakos, D. Kommatas, (3) Each population inside the southern clade N. Koutsikos, V. Tachos, L. Vardakas & A. N. Economou, should be considered as distinct conservational 2015. Freshwater fishes and lampreys of Greece: an unit, because all southern populations fulfil two annotated checklist. Monographs on Marine Sciences No. proposed criteria—they are independent and 8. Hellenic Center for Marine Research, Athens. ISBN: 978-960-9798-06-8. genetically distinct from the remaining T. pleu- Beerli, P., 1998. Estimation of migration rates and population robipunctatus populations; and they are partic- sizes in geographically structured populations. In Car- ularly important for the viability of the lineage, valho, G. (ed.), Advances in Molecular Ecology., NATO- since each of the southern lineages is composed ASI Workshop Series IOS Press, Amsterdam: 39–53. Beerli, P., 2009. How to use Migrate or why are markov chain of a single population. monte carlo programs difficult to use? In Bertorelle, G., M. Finally, in order to ensure the unconstrained W. Bruford, H. C. Hauffe, A. Rizolli & C. Vernese (eds), Population Genetics for Conservation, Vol. 17., evolutionary development of T. pleurobipunctatus, Conservation Biology Cambridge University Press, regardless whether we consider it a single species or a Cambridge. species complex, viability of all observed lineages Beerli, P. & J. Felsenstein, 2001. Maximum likelihood estima- should be estimated. The genetic screening supports tion of a migration matrix and effective population sizes in n subpopulations by using a coalescent approach. Pro- the promotion of careful taxonomic work for delin- ceedings of the National Academy of Sciences of the USA eating evolutionary significant units and future devel- 98: 4563–4568. opments towards the better description of the Beerli, P., H. Hotz & T. Uzzell, 1996. Geologically dated sea constituent cryptic species or species complex. barriers calibrate a protein clock for Aegean water frogs. Evolution 50: 1676–1678. Bianco, P., 1990. Potential role of the paleohistory of the Acknowledgements This work was supported by the Czech Mediterranean and Paratethys basin on the early dispersal Science Foundation (Project GA15-19382S), institutional of Euro-Mediterranean freshwater fishes. Ichtyological resources of the Ministry of Education, Youth, and Sports of Exploration Freshwaters 1: 167–184. the Czech Republic (JV) and Ministry of Culture of the Czech Bianco, P. G., 1988. cephalus (Linnaeus), with Republic (DKRVO 2019-2023/VI.4.a, National Museum, records of fingerling adult males, Leuciscus pleuro- 00023272) (RS). IB received support from the SYNTHESYS bipunctatus (Stphanidis) and their hybrids from western Programme project CZ-TAF-4228 at the National Museum Greece. Journal of Fish Biology 32: 1–16. Prague, financed by European Community Research Bo¨hme, M. U., N. Schneeweis, U. Fritz, J. Moravec, I. Majla´th, Infrastructure Action under the FP7 ‘‘Capacities’’ Programme. V. Majla´thova & T. U. Berendonk, 2006. Genetic differ- Support came also from the Project FREDIE (www.fredie.eu) entiation and diversity of Lacerta viridis viridis (Laurenti, funded by the Joint Initiative for Research and Innovation 1768) within the northern part of its range: an investigation (PAKT) program of the Leibniz Association (JF, MFG). We are using mitochondrial haplotypes. Salamandra 42: 29–40. grateful to D. Kommatas and L. Vardakas for assistance in the Brazee, N. J. & D. L. Lindner, 2013. Unravelling the Phellinus field, as well as to A. N. Economou for discussions on the pini s.l. complex in North America: a multilocus phylogeny evolutionary history of T. pleurobipunctatus. and differentiation analysis of Porodaedalea. Forest Pathology 43: 132–143. Buj, I., M. C´ aleta, Z. Marcˇic´,R.Sˇanda, J. Vukic´ & M. Mra- References kovcˇic´, 2015. Different histories, different destinies-im- pact of evolutionary history and population genetic structure on extinction risk of the Adriatic spined loaches Avise, J. C., 1986. Mitochondrial DNA and the evolutionary (genus Cobitis; , Actinopterygii). PLoS genetics of higher animals. Philosophical Transactions of ONE 10(7): e0131580. https://doi.org/10.1371/journal. the Royal Society B 312: 325–342. pone.0131580. Ballard, J. W. O., 2000. Comparative genomics of mitochon- Buj, I., Z. Marcˇic´,M.C´ aleta, R. Sˇanda, M. F. Geiger, J. Freyhof, drial DNA in Drosophila simulans. Journal of Molecular A. Machordom & J. Vukic´, 2017. Ancient connections Evolution 51: 64–75. among the European rivers and watersheds revealed from Banarescu, P. M., 2004. Distribution patterrn of the aquatic the evolutionary history of the genus Telestes fauna of the Balkan Peninsula. In Griffiths, H. I., B. (Actinopterygii; Cypriniformes). PLoS ONE 12(12): e0187366. https://doi.org/10.1371/journal.pone.0187366. 123 Author's personal copy

Hydrobiologia

Castresana, J., 2001. Cytochrome b phylogeny and the taxon- affects barcoding accuracy of its freshwater fishes. omy of great apes and mammals. Molecular Biology and Molecular Ecology Resources 14: 1210–1221. Evolution 18: 465–471. Guinand, B., J. D. Durand & J. Laroche, 2001. Identifying main Dimopoulos, P., T. Raus, E. Bergmeier, T. Constantinidis, G. evolutionary mechanisms shaping genetic variation of Iatrou, S. Kokkini, A. Strid & D. Tzanoudakis, 2016. Leuciscus cephalus L. 1758 () in Western Vascular plants of Greece: an annotated checklist. Greece: discordance between methods. Evolution, Life Willdenowia 46: 301–347. Sciences 324: 1045–1060. Drummond, A. J., M. A. Suchard, D. Xie & A. Rambaut, 2012. Hall, G., 2006. Selective constraint and genetic differentiation in Bayesian phylogenetics with BEAUti and the BEAST 1.7. geographically distant barley yellow dwarf virus popula- Molecular Biology and Evolution 29: 751–761. tion. Journal of General Virology 87: 3067–3075. Dubut, V., A. Fouquet, A. Voisin, C. Costedoat, R. Chappaz & Hey, J. & J. Wakeley, 1997. A coalescent estimator of the A. Gilles, 2012. From late Miocene to Holocene: processes population recombination rate. Genetics 145: 833–846. of differentiation within the Telestes genus (Actinoptery- Hudson, R. R., D. D. Boos & N. L. Kaplan, 1992. A statistical gii: Cyprinidae). PlosOne 7(3): e34423. test for detecting geographic subdivision. Molecular Durand, J. D., A. R. Templeton, B. Guinand, A. Imsidirou & Y. Biology and Evolution 9: 138–151. Bouvet, 1999. Nested clade and phylogeographic analyses Huelsenbeck, J. P. & F. Ronquist, 2001. MRBAYES: bayesian of the chub, Leuciscus cephalus (Teleostei, Cyprinidae), in inference of phylogenetic trees. Bioinformatics 17: 754–755. Greece: implications for Balkan Penninsula biogeography. Jolivet, L. & C. Faccenna, 2000. Mediterranean extensions and Molecular Phylogenetecs and Evolution 13: 566–580. the Africa-Eurasia collision. Tectonics 19: 1095–1106. Durand, J. D., E. Unlu, I. Doadrio, S. Pipoyan & A. R. Tem- Ketmaier, V., P. G. Bianco, M. Cobolli, M. Krivokapic, R. pleton, 2000. Origin, radiation, dispersion and allopatric Caniglia & E. de Matthaeis, 2004. Molecular phylogeny of hybridization in the chub Leuciscus cephalus. Proceedings two lineages of Leuciscinae cyprinids (Telestes and Scar- of the Royal Society London B 267: 1687–1697. dinius) from the peri-Mediterranean area based on cyto- Economidis, P. S. & P. M. Banarescu, 1991. The distribution chrome b data. Molecular Phylogenetics and Evolution 32: and origins of freshwater fishes in the Balkan Peninsula, 1061–1071. especially in Greece. Internationale Revue der gesamten Kottelat, M. & J. Freyhof, 2007. Handbook of European fresh- Hydrobiologie und Hydrographie 76: 257–283. water fishes. Freyhof, Berlin. Economou, A. N., R. Barbieri, C. H. Daoulas, T. Psarras, M. Koutsikos, N., S. Zogaris, L. Vardakas, V. Tachos, E. Kalo- Stoumboudi, I. Bertahas, S. Giakoumi & A. Patsias, 1999. gianni, R. Sˇanda, Y. Chatzinikolaou, S. Giakoumi, P. Endangered Freshwater Fish of Western Greece and the S. Economisid & A. N. Economou, 2012. Recent distri- Peloponnese: Distribution, Abundance, Threats and Mea- butional contributions to the freshwater ichthyofaunal of sures for Protection. Final Technical Report, National Greece. Mediterranean Marine Science 13: 268–277. Centre for Marine Research, 341 pp. (In Greek). Librado, P. & J. Rozas, 2009. DnaSP v5: a software for com- Economou A. N., Y. Chatzinikolaou, U. Dussling, K. Basel, S. prehensive analysis of DNA polymorphism data. Bioin- Giakoumi & S. Zogaris, 2006. Distribution patterns of fish formatics 25: 1451–1452. in mountain streams of western Greece. 10th International Lymberakis, P. & N. Poulakakis, 2010. Three continents Congress on the Zoogeography and Ecology of Greece and claiming an archipelago: the evolution of Aegean’s her- Adjacent Regions, Book of Abstracts, Patras, p. 30. petofaunal diversity. Diversity 2: 233–255. Economou, A. N., S. Giakoumi, L. Vardakas, R. Barbieri, M. Machordom, A. & I. Doadrio, 2001. Evidence of a Cenozoic Stoumboudi & S. Zogaris, 2007. The freshwater ichthy- Betic–Kabilian connection based on freshwater fish phy- ofaunal of Greece: an update based on a hydrographic basin logeography (Luciobarbus, Cyprinidae). Molecular Phy- survey. Mediterranean Marine Science 8: 91–168. logenetics and Evolution 18: 252–263. Farias, I. P., G. Ortı´, I. Sampaio, H. Schneider & A. Meyer, Maurakis, E. G., M. K. Pritchard & P. S. Economidis, 2001. 2001. The cytochrome b Gene as a phylogenetic marker: Historical relationships of mainland river drainages in the limits of resolution for analyzing relationships among Greece. BIOS 6: 109–124. cichlid fishes. Journal of Molecular Evolution 53: 89–103. Nei, M. & R. K. Chesser, 1983. Estimation of fixation indices and Frankham, R., J. D. Ballou & D. A. Briscoe, 2010. Introduction gene diversities. Annals of Human Genetics 47: 253–259. to Conservation Genetics. Cambridge University Press, Oikonomou, A., F. Leprieur & I. D. Leonardos, 2014. Bio- Cambridge. geography of freshwater fishes of the Balkan Peninsula. Fu, Y. X. & W. H. Li, 1993. Statistical tests of neutrality of Hydrobiologia 738: 205–220. mutations. Genetics 133: 693–709. Papadaki, C., K. Soulis, R. Mun˜oz-Mas, F. Martinez Capel, S. Geiger, M. F., F. Herder, M. T. Monaghan, V. Almada, R. Zoragis, L. Ntoanidis & E. Dimitriou, 2016. Potential Barbieri, M. Bariche, P. Berrebi, J. Bohlen, M. Casal- impacts of climate change on flow regime and fish habitat Lopez, G. B. Delmastro, G. P. J. Denys, A. Dettai, I. in mountain rivers of the southwestern Balkans. Science of Doadrio, E. Kalogianni, H. Ka¨rst, M. Kottelat, M. Kovacˇic´, the Total Environment. https://doi.org/10.1016/j.sctotenv. M. Laporte, M. Lorenzoni, Z. Marcˇic´,M.O¨ zulug˘,A. 2015.06.134. Perdices, S. Perea, H. Persat, S. Porcelotti, C. Puzzi, J. Patwardhan, A., S. Ray & A. Roy, 2014. Molecular markers in Robalo, R. Sˇanda, M. Schneider, V. Sˇlechtova´,M. phylogenetic studies: a review. Journal of Phylogenetics & Stoumboudi, S. Walter & J. Freyhof, 2014. Spatial Evolutionary Biology 2: 131. https://doi.org/10.4172/ heterogeneity in the Mediterranean biodiversity hotspot 2329-9002.1000131.

123 Author's personal copy

Hydrobiologia

Perea, S., M. Bohme, P. Zupancic, J. Freyhof, R. Sanda, M. (Platyhelmintes, Tricladida, Dugesiidae). Molecular Phy- Ozulug, A. Abdoli & I. Doadrio, 2010. Phylogenetic rela- logenetics and Evolution 66: 877–888. tionships and biogeographical patterns in circum- Swofford, D. L., 2002. PAUP*: Phylogenetic Analysis Using Mediterranean subfamily Leuciscinae (Teleostei, Cypri- Parsimony (*and Other Methods), Version 4 [Computer nidae) inferred from both mitochondrial and nuclear data. software and manual]. Sinauer Associates, Sunderland, BMC Evolutionary Biology 10: 265. MA. Perea, S., J. Vukic´,R.Sˇanda & I. Doadrio, 2016. Ancient Tajima, F., 1989. Statistical method for testing the neutral mitochondrial capture as factor promoting mitonuclear mutation hypothesis by DNA polymorphism. Genetics discordance in freshwater fishes: a case study in the genus 123: 585–595. Squalius (Actinopterygii, Cyprinidae) in Greece. PloS Tamura, K., J. Dudley, M. Nei & S. Kumar, 2007. MEGA4: ONE 11(12): e0166292. https://doi.org/10.1371/journal. molecular evolutionary genetics analysis (MEGA) soft- pone.0166292. ware version 4.0. Molecular Biology and Evolution 24: Perissoratis, C. & N. Conispoliatis, 2003. The impacts of sea- 1596–1599. level changes during latest Pleistocene and Holocene times Tobe, S. S., A. C. Kitchener & A. M. Linacre, 2010. Recon- on the morphology of the Ionian and Aegean seas (SE structing mammalian phylogenies: a detailed comparison Alpine Europe). Marine Geology 196: 145–156. of the cytochrome B and cytochrome oxidase subunit I Poulakakis, N., P. Lymberakis, A. Antoniou, et al., 2003. mitochondrial genes. PLoS ONE 5: e14156. Molecular phylogeny and biogeography of the wall lizard Tsigenopoulos, C. S. & P. Berrebi, 2000. Molecular phylogeny Podarcis erhardii (Squamata: Lacertidae). Molecular of North Mediterranean freshwater barbs (genus Barbus: Phylogenetics and Evolution 28: 38–46. Cyprinidae) Inferred from cytochrome b sequences: bio- Poulakakis, N., P. Lymberakis, C. S. Tsigenopoulos, A. geographic and systematic implications. Molecular Phy- Magoulas & M. Mylonas, 2005a. Phylogenetic relation- logenetics and Evolution 14: 165–179. ships and evolutionary history of snake-eyed skink Able- Vasileiadou, K., M. Bo¨hme, T. A. Neubauer, G. L. Georgalis, G. pharus kitaibelii (Sauria: Scincidae). Molecular E. Syrides, L. Papadopoulou & N. Zouros, 2017. Early Phylogenetics and Evolution 34: 245–256. Miocene gastropod and ectothermic vertebrate remains Poulakakis, N., P. Lymberakis, E. Valakos, P. Pafilis, E. Zouros from the Lesvos Petrified Forest (Greece). PalZ 91: & M. Mylonas, 2005b. Phylogeography of Balkan wall 541–564. lizard (Podarcis taurica) and its relatives inferred from Vogiatzi, E., E. Kalogianni, B. Zimmerman, S. Giakoumi, R. mitochondrial DNA sequences. Molecular Ecology 14: Barbieri, P. Paschou, A. Magoulas, D. Tsaparis, N. Pou- 2433–2443. lakakis & C. S. Tsigenopoulos, 2014. Reduced genetic Ro¨gl, F., 1998. Palaeographic consideration for Mediterranean variation and strong genetic population structure in the and Paratethys seaways (Oligocene to Miocene). Annalen freshwater killifish Valencia letourneuxi (Valenciidae) des Naturhistorischen Museums in Wien 99A: 279–310. based on nuclear and mitochondrial markers. Biological Ro¨gl, F., 1999. Mediterranean and Paratethys. Facts and Journal of the Linnean Society 111: 334–349. hypotheses of an Oligocene to Miocene paleogeography. Yu, Y., A. J. Harris & X. He, 2010. S-DIVA (Statistical Dis- Geologica Carpathica 50: 339–349. persal-Vicariance Analysis): a tool for inferring biogeo- Roveri, M., R. Flecker, W. Krijgsman, J. Lofi, S. Lugli, V. graphic histories. Molecular Phylogenetic and Evolution Manzi, F. J. Sierro, A. Bertini, A. Camerlenghi, G. De 56: 848–850. Lange, R. Govers, F. J. Hilgen, C. Hu¨bscher, P Th Meijer & Zardoya, R. & I. Doadrio, 1999. Molecular evidence on the M. Stoica, 2014. The Messinian Salinity Crisis: past and evolutionary and biogeographical patterns of European future of a great challenge for marine sciences. Marine cyprinids. Journal of Molecular Evolution 49: 227–237. Geology 352: 25–58. Zogaris, S., A. Economou & P. Dimopoulos, 2009a. Ecoregions Ryder, O. A., 1986. Species conservation and systematics: the in the Southern Balkans: should they be revised? Journal of dilemma of subspecies. Trends in Ecology and Evolution 1: Environmental Management 43: 682–697. https://doi.org/ 9–10. 10.1007/s00267-008-9243-y. Sˇanda, R., J. Vukic´, L. Choleva, J. Krˇ´ızˇek, A. Sˇediva´, S. Shumka Zogaris, S., Y. Chatzinikolaou & P. Dimopoulos, 2009b. & I. F. Wilson, 2008. Distribution of loach fishes (Cobiti- Assessing environmental degradation of montane riparian dae, Nemacheilidae) in Albania, with genetic analysis of zones in Greece. Journal of Environmental Biology 30: populations of Cobitis ohridana. Folia Zoologica 57: 719–726. 42–50. Zogaris, S. & A. N. Economou, 2017. The biogeographic Scho¨nhuth, S., J. Vukic´,R.Sˇanda,L.Yang&R.L.Mayden,2018. characteristics of the river basins of Greece. In Skoulikidis, Phylogenetic relationships and classification of the Holarctic N., I. Karaouzas & E. Dimitriou (eds), The Rivers of family Leuciscidae (Cypriniformes: Cyprinoidei). Molecular Greece: Evolution, Current Status and Perspectives. Phylogenetics and Evolution 127: 781–799. Springer, NewYork: 53–95. Sfenthourakis, S. & A. Legakis, 2001. Hotspots of endemic terrestrial invertebrates in southern Greece. Biodiversity Publisher’s Note Springer Nature remains neutral with and Conservation 10: 1387–1417. regard to jurisdictional claims in published maps and Sola, E., R. Sluys, K. Gritzalis & M. Riutort, 2013. Fluvial basin institutional affiliations. history in the northeastern Mediterranean region underlies dispersal and speciation patterns in the genus Dugesia

123