HM Medical Clinic

 

Doi:10.1016/j.ympev.2004.09.009

Molecular Phylogenetics and Evolution 34 (2005) 106–117 Phylogenetic analysis of the genus Thymallus (grayling) based on mtDNA control region and ATPase 6 genes, with inferences on control region constraints and broad-scale Eurasian phylogeography E. Froufea,b, I. Knizhinc, S. Weissd,¤ a CIBIO/UP, Campus Agrário de Vairão, 4480-661, Vairão, Portugal b Faculdade de Ciências, Universidade do Porto, Praça Gomes Teixeira, 4000-002 Porto, Portugal c Irkutsk State University, Biological Faculty, ul K. Marksa 1, Irkutsk, 664003, Russia d Institute of Zoology, Karl-Franzens University Graz, Universitätsplatz 2, A-8010 Graz, Austria Received 21 March 2004; revised 20 September 2004 Available online 5 November 2004 We present Wrst insights into the molecular phylogeny of the grayling genus Thymallus (Salmonidae) using sequences from the mitochondrial control region and ATPase6 genes. A suite of analytical approaches were applied for each gene separately and for thecombined data. The ATPase6 gene is shown to have a mean divergence rate across the genus of 2.46 times faster than the completecontrol region. Based on the combined data, four major (internal) clades, presumably originating in the Pliocene, were resolved withhigh support in all analyses and represented two distinct lineages in the Amur basin, one lineage in all remaining Siberian and Mon-golian drainages, and one lineage corresponding to European grayling Thymallus thymallus. The resolution of multiple lineages,from both additional internal and terminal clades, within each major drainage basin underscores the complexity and eVects thatPleistocene hydrological dynamics have had on the distribution of biodiversity in Siberia.
 2004 Elsevier Inc. All rights reserved.
Keywords: mtDNA; Control region; ATPase6; Thymallus; Phylogeny; Salmonidae question of how many species the genus contains, or thebreadth of diversity within each species, remains unan- Fishes of the family Salmonidae hold the highest eco- swered. Depending on the source, four or Wve species of nomic, ecological, and evolutionary interest among all grayling are generally recognized to be living in the cold temperate freshwater Wshes. Among them, the genus freshwaters of Europe, Asia and North America. Euro- Thymallus (some authors list as a distinct family Thym- pean grayling, Thymallus thymallus are found from southern France (Loire basin) (ast s received increasing attention, especially in to the Balkans as far south as the Luba River in Monte- Eurasia, both through advances in basic research on nat- negro ( north, European grayling ural selection ( to con- are found from Great Britain across most of Scandina- cerns of seriously declining populat via east to the Urals (see Fig. 1 in). The Nonetheless, even the most basic polytypic Arctic grayling Thymallus arcticus are cur-rently found from the Urals in the west, across all ofnorthern Eurasia and into N. America as far east as the * Corresponding author. Fax: +43 316 380 9875.
Hudson bay, with an extant disjunct population in E-mail address: (S. Weiss).
Montana, USA, and extinct populations as far east as 1055-7903/$ - see front matter  2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.ympev.2004.09.009 E. Froufe et al. / Molecular Phylogenetics and Evolution 34 (2005) 106–117 Michigan (see further range descriptions in anol) from 16 primarily remote Asian populations, three locations in Europe and two in N. America, The Amur grayling, Thymallus grubii, are found covering the distribution of the genus and representing throughout the Amur drainage and some adjacent rivers the Wve noted systematic units, several subspecies, and Xowing into Okhotsk Sea and Sea of Japa distinct major basins (see . At least two individuals were sequenced for each putative taxon or . Mongolian grayling Thymallus brevirostris are major basin (except that there is only one individual limited to lakes of the closed central Asian basin in from the Ob basin). Outgroup taxa included two Western Mongolia and border regions of Kazakhstan sheeWsh (Stenodus leucichthys), one whiteWsh (Coreg- and the Tuva Republic, and a Wfth taxon, Thymallus onus lavaretus) and two mountain whiteWsh (Prosopium nigrescens, listed as a species in williamsoni). Sample locations are also provided in as a subspecies T. arcticus nigrescens in s present only in Lake Chovsgul,Mongolia AmpliWcation and sequencing The systematic status of these species, or additional Whole genomic DNA was extracted using a stan- putative taxa has repeatedly come into question, with dard high-salt protocol (). Two recent genetic data playing a central role. For European mtDNA genes, ATP6 and the CR, were ampliWed grayling, only one taxon has been recognized ( using the polymerase chain reaction (PCR). The ), however suggest that Adriatic ATP6 gene was ampliWed in all individuals using populations represent distinct lineages worthy of species the L8558 and H9208 primers described in status and reveal an equally divergent The CR (including a partial segment of lineage in the Loire basin, France. For Arctic grayling, both Xanking tRNA genes) was ampliWed in 13 Thy- some N. American lineages were originally given a species mallus individuals, using the LRBT-25 and LRBT-1195 status, such as Thymallus signiWer or Thymallus montanus primers described in. The remain- but current views have supported a single species (see ing CR sequences were taken from previously pub- ; and references therein). In lished research ( Asia, there is a long history of subspeciWc and even infra- ; GenBank Accession Nos.
subspeciWc designations for Arctic grayling, including dis- tinct taxa in Lake Baikal, the Lena River, and PCR conditions (25
l reactions) were as follows: northeastern Russia. Within the Amur basin, genetic and each reaction contained 19
l H O, 2.5
l of 10 morphological data presented in d mega BuVer B, 0.5
l of 10 mM of each primer, 1.5
l of clearly support the presence of mul- 25 mM MgCl , 0.5
l of 10 mM dNTP's, 0.1
l Promega tiple taxa, including the putative "large-scale" taxon Taq DNA polymerase, and 0.5
l of 100 ng/
l DNA tem- (). Mongolian grayling, was once thought plate. The cycle parameters were as follows: initial dena- to represent the primitive member of the genus turation at 94 °C for 3 min, denaturation at 94 °C (40 s), annealing at 53 °C for ATP6 and 55 °C for CR (40 s), and ) buow that its position extension at 72 °C (40 s) repeated for 30 cycles. AmpliWed reXects a relatively recent evolutionary event within the DNA templates were puriWed using the NucleoSpin T. arcticus complex. T. nigrescens, which is closely related Extract Kit (Machery-Nagel) and approximately 100 ng to populations in Lake Baikal, also underscores the para- of puriWed PCR product was used in cycle sequencing phyletic status of the T. arcticus complex. There has been reactions following ABI (Applied Biosystems) PRISM no systematic or comprehensive attempt at describing the BigDye Terminator protocols. Sequences were visual- phylogenetic relationships of major lineages within the ized on either an ABI-3100 or ABI-310 genotyping grayling genus. Thus, this paper presents the Wrst insights into the molecular phylogeny of Thymallus, using themitochondrial gene sequences ATPase6 (ATP6) and the Sequence alignment and analysis complete control region (CR).
The ATP6 sequences were aligned by eye based on the amino acid sequences using the standard mtDNA Materials and methods
code for vertebrates in BioEdit programThis gene was selected to have a coding region that could be easily aligned to outgroup taxa, and followingthe presumption that it would have a lower substitu- Grayling (N D 35) were collected by angling or nets tion rate than the CR, based on general knowledge of in 1998–2003 (a small Wn clip was preserved in 96% eth- Wsh mtDNA () and comparative data for E. Froufe et al. / Molecular Phylogenetics and Evolution 34 (2005) 106–117 Table 1Sample locations including major river basins, population code with the number of individuals sampled shown in brackets, geographical coordinates,and GenBank accession numbers for Thymallus and the three outgroups used in this study: Coregonus lavaretus, Stenodus leucichthys, and Prosopiumwilliamsoni Genbank accession GenBank accession numbers control region Shilka R. ! Amur R. ! 110° 25⬘ AY778972 Tatar Strait, PaciWc Ocean Sypchergurka River Olengui R. ! Ingoda R.! 113° 26⬘ AY778973 Shilka R. ! Amur R. ! Tatar Strait, PaciWc Ocean Amur R. ! Tatar Strait, 134 ° 53⬘ AY778974–AY778977 AY779007–AY779010 Tatar Strait, PaciWc Ocean 139° 21⬘ AY778978–AY778979 AY779011–AY779012 Khutu R.! Tumnin R. ! 138° 56⬘ AY778980–AY778982 AY246417–AY246419 Tatar Strait, PaciWc Ocean Amgun River ! Amur ! 134° 47⬘ AY778983–AY778984 AY246410–AY246411 Tatar Strait, PaciWc Ocean Irkutsk Reservoir ! Angara! 104° 36⬘ AY778985–AY778986 AY168361; AY168363 Lake Baikal (Central) 108° 37⬘ AY778987 South Fork Terkhyn R. Selenga River ! Lake Baikal Enisey River (upper) AY778990–AY778991 AY168387; AY168389 Lena River–Laptev Sea Dvina River ! White Sea Yukon Territory, PaciWc, North Am. Pac (1) British Columbia, PaciWc, North Am. Pac (1) 122° 20⬘ 55° 25⬘ Anui River ! Amur–Tatar Strait, 137° 55⬘ AY778998–AY778999 AY246404–AY246389 Selenga River ! Lake Baikal 100° 39⬘ AY779000–AY779001 AY168348; AY168349 Olekma River ! Lena River 119° 33⬘ AY779002–AY779003 AY779017 Soca, Isonzo ! Adriatic Loire ! Atlantic, France Coregonus (1) — Stenodus (Yukon Stenodus (2) AY778968–AY778969 — Prosopium Clearwater Drainage Prosopium (2) — AY778970–AY778971 — several other salmonid genera ( . The CR was chosen based on ourprevious work and the availability of GenBank Sequences were imported into PAUP*4.0b10 ( sequences. The usefulness of the CR for the outgroup for phylogenetic analyses as well as obtaining was examined using one complete Coregonus CR the observed pairwise sequence divergence (uncorrected p (GenBank Accession No. AB034824), whereby both distances) and the number of transitions and transversions.
the entire CR as well as several highly conserved Between-group variation (corrected for within-group vari- regions alone were aligned with Thymallus. Sequences ation) was calculated using the net nucleotide divergence were aligned with Clustal W ( (Da) in MEGA version 2.sess applying several diVerent gap costs, and the alignment the degree of saturation in each codon position of the was chosen that produced the shortest maximum ATP6 sequences, the number of transitions and transver- parsimony tree. Another approach to using the CR was sions were plotted against the uncorrected pairwise dis- followed by Wrst identifying several of the most tances, for Wrst, second, and third positions. Base conserved sequence blocks (50–200 bp) in our data set, composition homogeneity was tested using a chi-square and adding these blocks to the ATP6 sequences in an (
2) test for equal base frequencies across taxa, examining attempt to use additional phylogenetic information.
each coding position separately for the ATP6 data.

E. Froufe et al. / Molecular Phylogenetics and Evolution 34 (2005) 106–117 Maximum parsimony (MP), maximum-likelihood likelihood values plotted against generation time (ML), and Bayesian analysis were used for phyloge- reached a stable mean equilibrium value; "burn-in" netic reconstruction with the ATP6 and CR data sets data sampled from generations preceding this point alone, and for both genes together. Modeltest 3.0 were discarded. All data collected at stationarity were () was used to choose the used to estimate posterior nodal probabilities and a most likely of 56 models of nucleotide evolution for summary phylogeny. Two independent replicates were each data set alone and for the combined data set. The conducted and inspected for consistency to check for best-Wt models estimated by Modeltest 3.0 were used to local optima (). We estimate a tree using ML and Bayesian inference. A used the Shimodaira–Hasegawa test (SH-test) ( heuristic search (10 replicates) was used to estimate the ) as implemented in most likely topology for ML and MP methodologies.
PAUP*4.0b10 to compare the topologies of trees Heuristic searches started with stepwise addition trees within each data set.
and were replicated 10 times, with each replicate begin-ning with a random order of sequences. Branch swap-ping was performed by the tree-bisection-reconnection (TBR) method using default parameters. For MP anal-yses of the CR data, results were compared when gaps were treated as "missing data" and when coded as aWfth base. Bootstrap analysis ( The Wnal alignment of the ATP6 gene yielded 661 bp used to estimate support for the resulting topologies, (12 codons short of the complete gene) for 40 individu- with 1000 replicates, each with 100 random additions of als. There was no signiWcant diVerence in base frequen- sequences. Full heuristic search algorithms were cies across taxa and plots of the number of substitutions applied for the MP and the "fast" stepwise addition against uncorrected p distances revealed no saturation method for the ML analysis.
for transversions or transitions for both total positions The Bayesian analysis was implemented using and third codon positions alone). Pair- wise sequence divergence within the ingroup ranged which calculates Bayesian posterior probabilities using from 0 to 9.2%. A total of 200 variable sites were found a Metropolis-coupled, Markov chain Monte Carlo of which 191 were parsimony informative and there were (MC3) sampling approach. Analyses were carried out no stop codons. Within the ingroup the transition/trans- assuming the optimal model determined by Modeltest.
version ratio was 3.8, and there were 3 amino acid Chains were run for 1 £ 106 generations, and sampled changes (base pair position Wve: Ile-Val; 13: Leu-Ser- every 100. In both searches, stationarity of the Markov Thr; 14: Leu-Val). There were also 3 amino acid changes chain was determined as the point when sampled log between the ingroup and outgroup (position 3: Leu-Met; Fig. 1. The number of transitions and transversions plotted against uncorrected p distances. (A) Total number of substitutions; (B) third codon posi-tions alone.
E. Froufe et al. / Molecular Phylogenetics and Evolution 34 (2005) 106–117 60: Leu-Met; 183: Ile-Leu) and one within the outgroup The net divergence between the monophyletic (position 173: Leu-Met). Considering these results, there ingroup and outgroup taxa ranged from 14.3% for Pros- is no reason to suspect nuclear copies or mtDNA pseud- opium to 13.7% for one Stenodus sequence (Stenodus_2) and 14.0% for Coregonus. The other Stenodus sequence The MP analysis resulted in 33 equally parsimonious grouped with Coregonus with high-bootstrap support, trees of 400 steps (CI D 0.655; RI D 0.8752). Based on diVering by only 0.9%.
Modeltest, the HKY model with an estimate of invari-able sites (0.568) and a discrete approximation of the gamma distribution (1.147) was chosen. Using thismodel, we obtained one most likely tree with ML analy- Highly divergent sequence for the CR was noted sis. Bayesian analysis produced a similar topology, between Coregonus and Thymallus, and the shortest MP which diVered primarily in varying levels of node sup- tree obtained resulted in no additional internal nodes port. The strict consensus tree derived from MP gave (with or without support) and thus the complete CR was nearly identical results for the well-supported clades determined not to be useful for the outgroup taxa with obtained in ML, and likewise no further resolution in our aim of resolving relationships within the ingroup.
branch order (). The SH-test revealed no signiWcant The alignment for the CR across Thymallus included diVerences between tree topologies based on diVerent 1107 bp (including 87 bp of tRNA proline and 11 bp of phylogenetic methodologies. In all analyses the most tRNA phenylalanine genes). There were a total of 16 internal nodes were short with little or no support. How- indels of 1–2 bp in length and the transition/transversion ever, four major clades (most internal well-supported ratio was 2.0. Including indels there were a total of 32 nodes) as well as a number of more terminal clades were unique sequences among the 34 individuals analyzed.
identiWed with high-node support values (ese Pairwise sequence divergence ranged from 0 to 5.7%.
clades will be characterized below, considering as well The MP analysis resulted in 18 equally parsimonious the results from the CR.
trees of 308 steps (CI D 0.6461; RI D 0.8584) when gaps Fig. 2. Strict consensus tree of the 33 most-parsimonious trees for the ATPase6 sequences (see text for details). Bootstrap values (over 50%) for MP(above, left), for ML (HKY + G + I model) (above, right) and Bayesian probabilities (below the nodes). 100* means that all bootstrap values arehigher than 95. The tree was rooted using Coregonus, Stenodus, and Prosopium.

E. Froufe et al. / Molecular Phylogenetics and Evolution 34 (2005) 106–117 were treated as "missing data" and in 28 equally parsi- CR and ATP6 comparison and combined analysis monious trees of 334 steps (CI D 0.6617; RI D 0.8653)when gaps were treated as Wfth base. Based on Model- Pairwise sequence divergence estimates for the ATP6 test, the HKY model with an estimate of invariable sites gene were clearly higher than for the CR (). To (0.5938) and a discrete approximation of the gamma dis- compare divergence rates we calculated the mean diver- tribution (0.573) was chosen. Similar topologies were gence ratio for the two genes. This comparison was done revealed for all methodologies and the SH-test revealed excluding three OTU's for which the same individual no signiWcant diVerences between them. Thus only the was not sequenced for both genes. Based on 32 OTU's unrooted ML tree is presentether with sequenced at the individual level for both genes, mean major node support values for MP and Bayesian analy- pairwise divergence of ATP6 sequences was 2.46 greater sis. Description of the clades will be provided below (SD D 0.99) than for the CR. Plotting divergence esti- when considering both genes.
mates for the two genes reveals a nonlinear relation with Fig. 3. (A) A mid-point rooted tree derived from a ML search using the HKY + G + I model for the combined CR and ATP6 sequences (see text fordetails). This rooted tree is shown for pictorial purposes only and represents the same topology shown in the unrooted tree (B). All analyses (ML,MP, and Bayesian) gave similar estimates of relationships. For the major clades, bootstrap values (over 50%) are shown for ML (above); MP (with-out gaps) (below, left) and Bayesian probabilities (below, right). 100* means that all bootstrap values are higher than 95. The symbol and 䊏 referto two pairs of taxa that are found in sympatry in the Amur basin. (C) An unrooted tree derived from a ML search using the HKY + G + I model forCR sequences only, (see text for details). All analyses (ML, MP, and Bayesian) gave similar estimates of relationships. For the major clades, boot-strap values (over 50%) are shown for ML (above); MP (gaps treated as "missing" data) (below, left) and Bayesian probabilities (below, right). 100*means that all values are higher than 95.
Table 2Net nucleotide divergence (Da) between groups (uncorrected p distance) (corrected for within-group variation) for the four major clades (shown in derived from each gene data set separately as well as together Left table, ATP6 values in upper diagonal; CR values in lower diagonal. Right table, values for both genes together.

E. Froufe et al. / Molecular Phylogenetics and Evolution 34 (2005) 106–117 CR divergence estimates, which level oV in relation to est divisions within the genus for which there is reliable Preliminary analysis with the ATP6 gene combined Within clade 1 (Amur) there is high-node support for with several 50–200 bp blocks of the most conserved samples representing Amur grayling T. grubii. A sister regions of the CR, revealed no further resolution in clade to Amur grayling consists of samples from three branch order within the ingroup, and thus outgroup diVerent populations of a recently described but analysis was abandoned with CR sequence, and analysis unnamed taxon ("orange-spot") ( of the combined data set was carried out without root- ). Clade 2 consists of two ing. The Wnal alignment for both genes yielded a 1768 bp samples from the lower Amur, representing a distinct fragment. Pairwise sequence divergence among all taxa taxon ("lower Amur form") sympatric to the "orange- ranged from 0 to 6.5%. The MP analysis resulted in 1 spot" grayling characterized in tree of 541 steps treating gaps as "missing data" (250 n clade 3 there is support parsimony informative characters; CI D 0.6525, (from all four methodologies) for a Wrst split between the RI D 0.8652) and three trees of 567 steps when gaps were haplotypes from the Bureya River representing "large- treated as Wfth base (262 parsimony informative charac- ters; CI D 0.6220, RI D 0.8686). For the combined data and all other OTU's. These OTU's further form 3 well- set, the HKY model was again chosen, with an estimate supported lineages, one of which includes N. American, of invariable sites (0.593) and a discrete approximation Ural, and lower Lena populations of T. arcticus; one sin- of the gamma distribution (0.573). One unrooted ML gle haplotype from the upper Lena basin, and a diverse tree was found (). For pictorial purposes only, group of T. arcticus populations residing in the Baikal this tree is also shown with mid-point rooting (no basin, as well as the upper Ob and Enisey rivers, but also change in branch order occurs) together with the node T. nigrescens and T. brevirostris. Thus, as shown in support values for ML, MP (without gaps), and Bayes- T. arcticus is paraphyletic, rela- ian analysis (The SH-test revealed no signiW- tive to T. brevirostris and T. nigrescens. The forth clade cant diVerences between tree topologies based on consists of European grayling T. thymallus.
diVerent phylogenetic methodologies.
Beginning with the most internal nodes, four major lineages are resolved with consistently high support across all methodologies (wo lineages repre-sent taxa from within the Amur basin, including the Phylogenetic relationships among grayling lineages population sampled from the Okhotsk basin, adjacent tothe Amur. One lineage includes all N. American and For comprehensive characterization of the phyloge- Siberian samples outside the Amur basin in addition to netic relation of all grayling lineages, we use the com- two highly divergent haplotypes from the Bureya River bined data set, noting that there are no signiWcant (Amur basin). The fourth lineage consists of European conXicts in well-supported nodes between any of the grayling T. thymallus. These lineages represent the earli- trees for either gene. However, discussion of divergencesand divergence times is based on CR sequences only,relying on the 1%/MY calibration, which provides con-sistency with previously reported divergenceAlthough we demonstrate the lack of linearity in CRdivergence (rejection of the molecular clock), the calibra-tion is based on lineages diverging in the past 110–450,000 years, and was compatible with secondarypaleo-hydrological assumpsdiverging in the past 2.5 MYThus, the molecular clock heterogeneity that we observe(due to CR constraints) results in underestimates for themost divergent lineages, but perhaps have little eVect ondivergence estimates involving lineages evolving withinat least the last 2.5 MY.
The most fundamental and consistent result of the phylogenetic reconstruction was the resolution of fourmajor lineages, which, relative to the resolution of our Fig. 4. Plot of sequence divergences for ATP6 against the CR for Thy- sequence data, have split contemporaneously. This can mallus, Salmo, Salvelinus, and Oncorhychus (see text for details).
be seen in the outgroup-rooted trees ( a poly- E. Froufe et al. / Molecular Phylogenetics and Evolution 34 (2005) 106–117 chotomy (as nodes more internal than those supporting explained by river capture of north Xowing rivers (Zeya these four lineages are poorly supported). These lineages and Amgun/Bureya) by the Amur basin following the have presumably evolved in isolation betwee melting of the Stanovoy glacier complex and Okhotsk 4.8 Mya, predominately in the Pliocene epoc ice sheet as depicted in hus, there The maximum divergence (4.8%) is seen between two lin- appears to be two sympatric taxa in the Bureya River eages within the Amur basin, and the minimum between with 3.4% CR divergencegests that European grayling and the diverse clade 3 (1.6%). This there are as many as 3 taxa in the Bureya River, though underscores the historical depth of major phylogeo- our limited number of samples cannot yet support this graphic events for Thymallus within far-eastern Siberia, hypothesis. Clearly, as suggested in compared to the probable Pleistocene/Pliocene bound- and further characterized by detailed morphological ary colonization of Europe. This also contrasts with description ( Amur basin is a much shallower divergences and simpler phylogeo- complex system of multiple sympatric and allopatric graphic patterns for several other salmonid taxa in far- lineages of grayling, even based on our relatively sparse eastern Siberia ( A more detailed examination of the OTU's within Among the remaining lineages of clade 3 is a subclade these four mitochondrial lineages reveals considerable representing all samples within the Baikal/Enisey basin.
complexity in terms of the distribution of haplotypes These individuals (Irt, Uib, Sft, and Chv) represent a within and among drainages as well as their concor- shallow diverse clade described in dance with existing systematics and taxonomy. Clade 1 , presumed to have colonized Lake Baikal and its () consists of two individuals of Amur grayling T. tributaries in the mid-Pleistocene (110–450,000 years grubii sampled from tributaries of the upper Amur ago) following the "break-out" of the rising Lake Baikal drainage, corresponding roughly to the original type into the Angara River. Populations from the Angara specimen locality for the species (). A River (Irt), Lake Baikal (Uib), and the upper Selenga highly supported sister group to these upper Amur sam- basin (Sft) all represent T. arcticus whereby the popula- ples, are two individuals sampled from the Bureya River tion from ancient Lake Chovsgul in Mongolia (Chv), (an Amur tributary), which also represent T. grubii, part of the upper Selenga basin, represents T. nigrescens.
though from a divergent population as they are ca.
These populations await morphological analysis aimed 2000 km apart. For "orange-spot" grayling, the three at supporting or refuting various subspeciWc or intrasub- populations originate from presently disjunct but specWc designations. A sister clade from Baikal/Enisey adjoining basins (Merek–lower Amur; Buta–Okhotsk basin lineages contains two sub-clades one of which con- Sea; Botchi–Sea of Japan). We hypothesize that this tains one sample from the uppermost Enisey basin in taxon has a monophyletic origin with its present Mongolia (Jb1) and one sample from the uppermost Ob distribution reXecting the dynamics of the postglacial basin (Biy) in the Altai region, both considered to repre- hydrological landscape. Clade 2 consists of two sent T. arcticus. As show the specimens from the lower Amur ("lower Amur form") uppermost Enisey contains lineages distantly related found in sympatry with "orange-spot" grayling as from populations of the mid or lower Enisey basin, described in Froufe et al. (2003a) which are now directly connected to Baikal via its only hile in sympatry, we emphasize that no hybrid- outXow, the Angara River. The second sub-clade con- ization was documented between "orange-spot" and the tains two individuals of Mongolian grayling T. breviros- "lower Amur form" and sequence divergence (CR) is tris (Kn), from Khökh Nuur, a small lake at the 4.6% (rayling sampled from the easternmost range of the species. Divergence between lower Amur were previously described as T. arcticus the Baikal/Enisey clade and its sister clade is only 0.9% underscoring relatively recent phenotypic diversiWcation Clade 3 also contains samples from the Amur basin marked by the highly morphologically distinct Mongo- (Bureya) representing "large-scale" grayling. These hapl- lian grayling.
otypes are highly divergent from all other OTU's in One individual represents a highly divergent lineage Clade 3 (2.4–3.7%), and is 3.2–5.0 % divergent from all from the upper Lena basin (Chi3) and is poorly deWned other Amur basin OTU's (including samples from the within clade 3, as its position is never well supported and adjoining Okhotsk sea and Sea of Japan). The hypothe- changes depending on the methodology applied. This siuggesting that these "large- single sample was taken as a representative (data not scale" grayling are more closely related to Siberian shown) of the Lena basin clade depicted lineages (i.e., T. arcticus of Kolyma and lower Lena viduals from the lower Lena delta drainages referred to as the pallasi subspecies) north of region (Del) cluster together with one individual sam- the Amur basin is supported by all four phylogenetic pled from the Dvina River drainage west of the Ural approaches when both genes are considered ( mountains (Sja5), and these three individuals form a sis- The existence of this lineage in the Amur can be ter group to two individuals from PaciWc drainage popu- E. Froufe et al. / Molecular Phylogenetics and Evolution 34 (2005) 106–117 lations in North America (Pac). The Sja5 sample is the typic and morphological characters together with identical specimen sequenced for the NADH 5 gene in genetic data to provide the basis for comprehensive sys- nd apparently comes from tematic revision.
within the hybrid zone of T. thymallus and T. arcticus The authors are engaged in such a work which involves the collection of 12 meristic and 46 morphologi- the population is more than 500 km west of the Urals.
cal characters, sequencing of the complete CR, and The divergence between the upper Lena individual where necessary screening of multiple microsatellite loci (Chi3) and the latter described clade is 3.1%. While only (to determine reproductive isolation of putative sympat- a few individuals are used in this analysis, extensive sam- ric taxa), throughout the Eurasian distribution of the pling of the Lena basin demonstrates that these clades genus. It is our opinion that, especially for a popular represent two highly distinct monophyletic lineages well-studied group of Wshes like salmonids, such compre- (Weiss et al. unpublished data). While all of these popu- hensive and exhaustive work is necessary to provide lations are considered to represent T. arcticus, opinions both a clear understanding of evolutionary diversity and diVer concerning subspeciWc status of phenotypically to develop a broadly acceptable and practical systematic diverse grayling of Arctic draining rivers.
and taxonomic system.
In summary, clade 3 (exclusive of "large-scale" gray- ling) presents a confounding picture of a paraphyletic T. arcticus complex, containing several distinct species, aswell as a number of controversial subspecies. European The use of a coding gene allowed us to incorporate grayling, T. thymallus are represented by clade 4.
multiple outgroups, and validate the monophyletic sta- Though simplistically depicted here, the species is indeed tus of Thymallus. However, the use of the closest avail- phylogeographically complex), exhib- able outgroups, representing the three ancestral genera its 4% CR sequence divergence across its range, and may within Salmonoidei ( contain at least one additional taxon, from Adriatic and references therein), provided no resolution of draining rivers (), represented the Wrst splits within the ingroup, the outgroup taxon did here by the Soc18 sample.
not support any branch ordering among the four major This phylogenetic framework provides the Wrst lineages described above. Preliminary analysis (not broad-scale overview of the genus Thymallus, reporting shown) adding more derived genera (Salvelinus, Salmo, on several locations and or taxa that have not been pre- and Oncorhynchus) as outgroups, demonstrated that viously analyzed, expanding on the most recent Wndings they were even more distant from Thymallus, and again, in the Amur basin, and addressing the substantial provided no phylogenetic information for resolving the remaining systematic uncertainties particularly concern- branch order of the most internal nodes of the ingroup.
ing the T. arcticus complex. Based on our knowledge of Within the outgroup, we further note the relative prox- the within drainage genetic and phenotypic diversity in imity (0.9%) of one Stenodus sequence with Coregonus the Amur, Lena, and even Danube River in Europe, it is whereas the other is 3.3% divergent. As both Stenodus clear that few conclusions can be drawn from a basin individuals are from the sample population in the with few samples. For instance, while some authors Yukon drainage, and hybridization between these two assume that Lena river T. arcticus represent a subspecies genera is frequently reported ( (pallasi), we have found two highly divergent lineages in presume this to be the signal of past introgres- this drainage. For the Ob, another enormous hydrologi- cal system, we cannot assume that only one lineage ispresent based on our single mtDNA sequence. In fact, all Comparison of divergence rates between ATP6 and CR other major Siberian drainages (Lena, Enisey andAmur) that have been sampled more extensively, exhibit The ATP6 gene evolves faster than the complete CR multiple lineages with presumably deep and most likely within Thymallus. This is surprising consider the results allopatric origins. Thus, while signiWcant advances have within a study of brown trout Salmo trutta, where 11 been recently made in the description of systematic haplotypes were found screening ATP6 in 24 individuals diversity in Thymallus populations throughout their whereby 17 haplotypes were found using the Wrst 310 bp range, there are some important gaps in our knowledge.
of the . In a comparative study Together with the poorly sampled Ob, areas of both the of four taxa of PaciWc salmon (Oncorhynchus Enisey and Amur remain unexplored; some taxa (T. showed that the combined fragment of brevirostris, T. nigrescens) are only represented in our ATPase 8 and 6, COIII, and ND3 evolved slower in 3 of data by single populations; and some of the peripheral 4 taxa compared to the combined fragment of cyto- range of the T. arcticus complex have not been sampled.
chrome b (cyt b) and the CR. In a well-calibrated frame- Additionally, we stress the need for both population work, erive a 1.3%/MY rate level sampling and parallel collection of reliable pheno- for the ATP6 gene across a diverse range of teleost taxa, E. Froufe et al. / Molecular Phylogenetics and Evolution 34 (2005) 106–117 and report conspeciWc maximum (i.e., saturation) whereas pairwise divergences for ATP6 divergence rates in the CR nearly 3 times higher than for reach over 9%. This suggests species-speciWc constraints ATP6 in the characiform genus, Prochilodus.
on the non-coding CR within the genus Thymallus.
We evaluated the relative divergence rate of ATP6 While results for Salvelinus and Oncorhynchus are con- and the CR in available salmonid GenBank sequences.
gruent with our comparative rates within Thymallus, it is Choosing individuals for which both genes were clear that there is no overall constraint on the salmonid sequenced, the relative divergence rate across two species CR as sequences between genera are indeed diYcult to of charrs (Salvelinus alpinus, S. fontinalis; GenBank align. Thus, it appears that the CR's usefulness in salmo- Accession Nos. AF154851, AF154850), two species of nid phylogenetics is indeed limited to within genus com- Oncorhynchus (O. mykiss and O. tshawytscha, GenBank parisons. In another study, the CR was one of the only Accession Nos. L29771, AF392054), and two species of gene segments that did not support the monophyly of Salmo (S. salar, GenBank Accession No. NC_001960, the large genus Oncorhynchus, so even this generaliza- and S. trutta, GenBank Accession Nos. X74240– tion must be taken with caution ( X74245; M97962, M97966, M97972–M97984) was assessed. Calculating the ratio of mean divergences, as While the vertebrate CR contains both hyper-variable above, the ATP6 evolves 1.57 times faster than the CR in and conserved sequence blocks, the blocks are limited in Salvelinus and 1.93 times faster in Oncorhynchus ( length and are not necessarily conserved across all tele- For the Salmo comparison, only partial sequences were ost lineages (). Additionally, the con- available (315 bp for ATP6 and 241 alignable bp for CR) served blocks cannot explain the relaxation of our and for this more limited comparison the CR evolved hypothesized species-speciWc constraints among salmo- 1.98 times faster than ATP6. Thus, in 3 of the 4 salmonid nid genera. A related Wnding demonstrates both among genera evaluated, ATP6 is shown to evolve considerably and within genus-speciWc diVerences in CR versus cyt b faster than the CR. The ATP6 is thus another coding divergences in a survey of 68 avian specie mtDNA segment, together with NADH subunits 1, 2, 5, Whereby many avian lineages were shown to have more rapidly evolving CR's (in one genus ) that evolves faster than the CR in salmonid 5.14–21.65 times faster, dependant on species), at least two genera displayed CR/cytb ratios of less than 1 The CR in salmonid Wshes thus exhibits substantial (0.46:0.94 and 0.36:0.81). Such lineage-speciWc con- evolutionary constraints. It is noteworthy that some straints suggest yet-to-be identiWed selectively advanta- authors, inXuenced by the obviously higher substitution geous functions for the CR.
rates found in other Wshes or endothermic vertebrates,have entertained divergence rates as high as 10%/MY in a salmonid ( have not found anyempirical evidence for such CR divergence rates in sal- More frequent and improved calibrations of diver- monid Wshes, and reiterate that the approximate 1% rate gence estimates, have suggested that the origins of given for Thymallus is inferred from multiple indepen- many extant lineages reach into the Pliocene, con- dent sources (see discussions in tradicting earlier notions of the importance of the and ). This rate is also equal to that esti- Pleistocene epoch in generating present biodiversity.
mated for the entire mtDNA molecule for the salmonid This is true for the N. American avian fa ), diverse organisms on the Iberian Pen- While we calculated a mean relative rate for the ATP6 insula (, as well as for N.
and CR (2.46:1) the substitution rate of one or both American freshwater Wshes (Avise et al., 1998; Avise genes is clearly not constant across all lineages, and thus and Walker, 1998). For European freshwater Wshes, calculations of divergence estimates (Da) among groups suggested a pre-Pleistocene spread for both genes are not consistent. For example, both CR of T. thymallus, wed a (2.7%) and ATP6 (6.6%) divergence estimates corre- similar result for barbel Barbus barbus and spond to ca. 2.7 MY between European grayling and r chub Leuciscus cephalus, and maximum upper Amur grayling (clade 1), and results are similar divergences among major lineages of Salmo trutta between European grayling and the diverse cl point to at least a Pliocene/Pleistocene origin ( for CR D 1.6 MY vs. 4.8% for ATP6 D 1.9 MY) ( ). In a broader phylogenetic work on Euro- However, for all other comparisons, the pean cyprinids, port faster mutating ATP6 gene results in lower divergence that most speciation events within the genera Barbus estimates (50% or more) between the major lineages.
and Luciobarbus occurred during the Pliocene when the Thus, our mean relative estimate is very approximate as major existing European river drainages Wrst formed.
the relation between the two genes is not linear, with the In northern Asia, the overall understanding of glacial CR exhibiting constraints after about 5.8% divergence dynamics is far less clear than for N. America and E. Froufe et al. / Molecular Phylogenetics and Evolution 34 (2005) 106–117 Europe, and, due to the paucity of studies, the phylogeo- Apostolidis, A.P., Triantaphyllidis, C., Kouvatsi, A., Economidis, P.S., graphic patterns of most Asian freshwater organisms is 1997. Mitochondrial DNA sequence variation and phylogeographyamong Salmo trutta L. (Greek brown trout) populations. Mol.
unknown. The historical perspective of Siberian paleo- Ecol. 6, 531–542.
history was that the region was too dry to accumulate Berg, L.S., 1949. Freshwater Fishes of the USSR and adjacent coun- the amounts of precipitation needed to build glaciers as tries, fourth ed. Part 3. USSR Academy of Sciences, Moscow- extensive as in N. America and Europe. Most recent Leningrad, pp. 929–1382 (in Russian, Translation available, 1965, advances, however, repeatedly stress that Siberian glaci- Smithson. Inst. by Israel Prog. Sci. Transl., Jerusalem).
Bermingham, E., McCaVerty, S., Martin, A.P., 1997. Fish biogeography ations were at least as extensive, and corresponding and molecular clocks: perspectives from the Panamanian Isthmus.
Pleistocene dynamics at least volatile as elsewhere (see In: Kocher, T.D., Stepien, C.A. (Eds.), Molecular Systematics of ferences therein). While this new Fishes. Academic Press, San Diego, pp. 113–128.
paradigm is being more generally accepted, its eVects on Bernatchez, L., 2001. The evolutionary history of brown trout (Salmo Siberian fauna and particularly freshwater Wshes is little trutta L.) inferred from phylogeographic, nested clade, and mis-match analysis of mitochondrial DNA variation. Evolution 55 (2), appreciated. At least for the genus Thymallus, the overall pattern of Pliocene origins of major lineages and exten- Brunner, P.C., Douglas, M.R., Osinov, A., Wilson, C.C., Bernatchez, L., sive Pleistocene paleohydrological dynamics resulting in 2001. Holarctic phylogeography of arctic charr (Salvelinus alpinus a mosaic of highly divergent lineages in each of the L.) inferred from mitochondrial DNA sequences. Evolution 55 (3), major river basins investigated has been clearly pre- Churikov, D., Matsuoka, M., Luan, X., Gray, A.K., Brykov, V.L.A., Gharrett, A.J., 2001. Assessment of concordance among genealogi- cal reconstructions from various mtDNA segments in three speciesof PaciWc salmon (genus Oncorhynchus). Mol. Ecol. 10, 2329–2339.
Crespi, B.J., Fulton, M.J., 2004. Molecular systematics of Salmonidae: combined nuclear data yields a robust phylogeny. Mol. Phyl. Evol.
31 (2), 658–679.
Durand, J.D., Persat, H., Bouvet, Y., 1999. Phylogeography and post- We thank D.J. Harris, C. Sturmbauer, K. Sefc, S.
glacial dispersion of the chub (Leuciscus cephalus) in Europe. Mol.
Koblmüller, and N. Duftner for comments, suggestions Ecol. 8, 987–989.
or help with some analysis; P. Spruel and T. Otis for pro- Dybowski, B., 1869. VorläuWge Mitteilungen über die Fischfauna des viding the Prosopium and Stenodus samples, respec- OnonXusses und des Ingoda in Transbaikalien. Wiener Mitteilung,946–951 (in German).
tively; C. Primmer for use of the Sja5 (Ural) sample; E.
Felsenstein, J., 1981. Evolutionary trees from DANN sequences: a Taylor for the N. American samples; and A. Snoj for the maximum likelihood approach. J. Mol. Evol. 17, 368–376.
Drina River sample, and for various Russian samples, Froufe, E., Knizhin, I., Koskinen, M.T., Primmer, C.R., Weiss, S., A.F. Kirillov. We also give thanks for Weld assistance to 2003a. IdentiWcation of reproductively isolated lineages of Amur B.E. Bogdanov, A.L. Antonov, G.P. Safronov, A.N. Mat- grayling (Thymallus grubii Dybowski 1869): concordancebetween phenotypic and genetic variation. Mol. Ecol. 12 (9), veev, and V.P. Samusenok. Financial support for this study was provided by the Portuguese Ministry of Sci- Froufe, E., Alekseyev, S., Knizhin, I., Alexandrino, P., Weiss, S., 2003b.
ence and Technology, "Fundação para a Ciência e Tecn- Comparative phylogeography of salmonid Wshes (Salmonidae) ologia" Grant POCTI/BSE/33364/99 to P. Alexandrino reveals late to post-Pleistocene exchange between three now-dis- and S. Weiss, as well as a Ph.D. grant to E. Froufe junct rivers basins in Siberia. Diver. Distrib. 9, 269–282.
GiuVra, E., Bernatchez, L., Guyomard, R., 1994. Mitochondrial CR (SFRH/BD/11377/2002); Russian Ministry of Education and protein coding genes sequence variation among phenotypic (Grant A03-2.12-265), Russian program "Universities of forms of brown trout Salmo trutta from northern Italy. Mol. Ecol.
Russia" Grant No. UR.07.01.009.
3, 161–1171.
Gómez, A., Lunt, D.H., in press. Refugia within refugia: patterns of phylogeographic concordance in the Iberian Peninsula. In: Weiss,S., Ferrand, N. (Eds.), Phylogeography of Southern European Refugia, Kluwer Academic Press, Dordrecht, Netherlands.
Grosswald, M.G., 1998. New approach to the ice age paleohydrology Avise, J.C., Walker, D., Johns, G.C., 1998. Speciation durations and of Northern Eurasia. In: Benito, G., Baker, V.R., Gregory, K.J.
Pleistocene eVects on vertebrates phylogeography. Proc. R. Soc.
(Eds.), Paleohydrology and Environmental Change. Wiley, Chich- Lond. B 265, 1707–1712.
ester, England, pp. 199–214.
Avise, J.C., Walker, D., 1998. Pleistocene phylogeographic eVects on Hall, T.A., 1999. BioEdit: a user-friendly biological sequence alignment avian populations and the speciation process. Proc. R. Soc. Lond. B editor and analysis program for Windows 95/98/NT. Nucleic Acids 265, 457–463.
Res. Symp. Ser. 41, 95–98.
Alt, K.T., 1971. Occurrence of hybrids between Inconnu, Stenodus leuc- Hansen, M.M., Loeschcke, V., 1996. Genetic diVerentiation among ichthys nelma (Pallas), and Humpback whiteWsh, Coregonus pids- Danish brown trout populations, as detected by RFLP analysis of chian (Linnaeus) in Chatanika River, Alaska. Trans. Am. Fish Soc.
PCR ampliWed mitochondrial DNA segments. J. Fish Biol. 48, 422– 100, 362–365.
Antonov, A.L., 2000. New salmonoidei Wshes from the Amur basin. In: Huelsenback, J.P., Bollback, J.P., 2001. Empirical and hierarchical Proceedings Biodiversity and Dynamics of Ecosystem in North Bayesian estimation of ancestral states. Syst. Biol. 50, 351–366.
Eurasia (BDENE) Novosibirsk, vol. 3. Part 1, August 21–26, RAS Huelsenback, J.P., Ronquist, F., 2001. MR-BAYES: Bayesian inference SB, IC and G Novosibirsk, pp. 120–122.
of phylogeny. Bioinformatics 17, 754–755.
E. Froufe et al. / Molecular Phylogenetics and Evolution 34 (2005) 106–117 Jankovic, D., 1962. Synopsis of biological data on European grayling, Sambrook, J., Fritsch, E.F., Maniatis, T., 1989. Molecular Cloning: A Thymallus thymallus (L. 1758): FAO Fisheries Biology Synopses, Laboratory Manual, second ed. Cold Spring Harbor Press, Cold vol. 24. FAO, Rome.
Spring Harbour, NY.
Klicka, J., Zink, R.M., 1997. The importance of recent ice ages in speci- Sanford, C.P.J., 1990. Phylogenetic relationships of salmonid Wshes.
ation: a failed paradigm. Science 277 (12), 1666–1669.
Bull. Br. Mus. Nat. Hist. (Zool.) 56 (2), 145–153.
Knizhin, I., Weiss, S., Antonov, A.L., Froufe, E., 2004. Morphological Scott, W.B., Crossman, E.J., 1998. Freshwater Fishes of Canada, Wfth and genetic diversity of Amur graylings (Thymallus; Thymallidae).
ed. Galf House Publications, Oakville, Ont., Canada.
J. Ichthyol. 44, 59–76.
Shatunovskiy, M.I., 1983. Fishes of Mongolian People Republic Mos- Kotlik, P., Berrebi, P., 2001. Phylogeography of the barbel (Barbus bar- cow. Nauka Press p. 273.
bus) assessed by mitochondrial DNA variation. Mol. Ecol. 10, Shimodaira, H., Hasegawa, M., 1999. Multiple comparisons of log-like- lihoods with applications to phylogenetic inference. Mol. Biol.
Koskinen, M.T., Ranta, E., Piironen, J., Veselov, A., Titov, S., Haugen, Evol. 16, 1114–1116.
T.O., Nilsson, J., Carlstein, M., Primmer, C.R., 2000. Genetic lin- Shubin, P., Zakharov, A., 1984. Hybridization between European gray- eages and postglacial colonization of the grayling (Thymallus thy- ling, Thymallus thymallus, and Arctic grayling, Thymallus arcticus, mallus, Salmonidae) in Europe, as revealed by mitochondrial DNA in the contact zone of the species. J. Ichthyol. 4, 159–163.
analyses. Mol. Ecol. 9, 1609–1624.
Skurikhina, L.A., Mednikov, B.M., Tugarina, P.Ya., 1985. Genetiches- Koskinen, M., Haugen, T.O., Primmer, C.R., 2002a. Contemporary kaya divergentziya khariusov (Thymallus Cuvier) Evrazii i seti Wsherian life-history evolution in small salmonid populations.
vidov. Zool. Zh. 64 (2), 245 (in Russian).
Nature 419, 826–830.
Sivasundar, A., Bermingham, E., Orti, G., 2001. Population structure Koskinen, M.T., Knizhin, I., Primmer, C.R., Schlötterer, C., Weiss, S., and biogeography of migratory freshwater Wshes (Prochilodus : 2002b. Mitochondrial and nuclear DNA phylogeography of Thy- Characiformes) in major South American rivers. Mol. Ecol. 10, mallus spp. (grayling) provides evidence of ice-age mediated envi- ronmental perturbations in the world's oldest body of freshwater, Smith, G.R., 1992. Introgression in Wshes: signiWcance for paleontol- Lake Baikal. Mol. Ecol. 11, 2599–2611.
ogy, cladistics, and evolutionary rates. Syst. Biol. 41, 41–57.
Kottelat, M., 1997. European fresh water Wshes. An heuristic checklist Sunnik, S., Snoj, A., Dovc, P., 1999. Microsatellites in grayling (Thymal- of the fresh water Wshes of Europe (exclusive of former USSR), lus thymallus): Comparison of two geographically remote popula- with an introduction for non-systematists and comments on tions from the Danubian and Adriatic river basin in Slovenia. Mol.
nomenclature and conservation. Biology 52 (Suppl. 5), 1–271.
Ecol. 8, 1756–1758.
Kumar, S., Tamura, K., Jakobsen, I.B., Nei, M., 2001. MEGA2: molec- Sunnik, S., Snoj, A., Dovc, P., 2001. Evolutionary distinctiveness of ular evolutionary genetics analysis software. Bioinformatics 17, grayling (Thymallus thymallus) inhabiting the Adriatic river system, as based on mtDNA variation. Biol. J. Lin. Soc. 74, 375– Lee, W.J., Conroy, J., Howell, W.H., Kocher, T.D., 1995. Structure and evolution of teleost mitochondrial control regions. J. Mol. Evol. 41, Svetovidov, A.N., 1936. Graylings (genus Thymallus, Cuvier) of Europe and Asia. In: Proceedings of Institute of Zoology Academy Meyer, A., 1993. Evolution of mitochondrial DNA of Wshes. In: of Sciences USSR, III, pp. 183–301 (in Russian).
Hochackka, P.W., Mommsen, P. (Eds.), Molecular Biology Fron- SwoVord, D.L., 2002. PAUP* ver 4.0.b10. Phylogenetic analysis using tiers, Biochemistry and Molecular Biology of Fishes. Elsevier Press, parsimony and other methods. Sinauer Associates, Sunderland, Amsterdam, pp. 1–38.
Osinov, A.G., Lebedev, V.S., 2002. Genetic divergence and phylogeny of Thompson, J.D., Higgins, D.G., Gibson, T.J., 1994. CLUSTAL W: the Salmoninae based on allozyme data. J. Fish Biol. 57 (2), 354–381.
Improving the sensitivity of progressive multiple sequence align- Persast, H., Pattee, E., Roux, A.L., 1978. In: Origine et caractérstiques ment through sequence weighting, position-speciWc gap penalties, de la distrubtion de l'Ómbre commun, Thymallus thymallus L., and weight matrix choise. Nucleic. Acids Res. 22, 4673–4680.
(1758), en Europe et en France, vol. 20. International Association of Tugarina, P.Ya., Khramtzova, V.S., 1980. Morphizioloogicheskaya Theoretical and Applied Limnology, Verhandlugnen, pp. 217–2121.
kharakteristka amurskogo khariusa Thymalllus grubei Dyb. J. Ich- Persat, H., 1996. Threatened populations and conservation of the thyol. 20 (N4 (123)), 590–605 (in Russian).
European grayling, Thymallus thymallus (L., 1758). In: Hefti, A., Tugarina, P.Ya., Anudarin, D., 1972. Mongolian grayling Thymallus Kirchofer, D. (Eds.), Conservation of Endangered of Europe, pp.
brevirostris (Kessler) of Dzabakhan River Basin. J. Ichthyol. 12 (N5), 843–856 (in Russian).
Posada, D., Crandall, K.A., 1998. Modeltest: testing the model of DNA Uiblein, F., Jagsch, A., Honsig-Erlenburg, W., Weiss, S., 2001. Status, substitution. Bioinformatics 14, 817–818.
habitat use, and vulnerability of the European grayling in Austrian Ramsden, S.D., Brinkmann, H., Hawryshyn, C.W., Taylor, J.S., 2003.
waters. J. Fish Biol. 59 (A), 223–247.
Mitogenomics and the sister of Salmonidae. TREE vol. 18, No. 12.
Weiss, S., Uiblein, F., Jagsch, A., 2001. Conservation of salmonid Wshes Redenbach, Z., Taylor, E.B., 1999. Zoogeographical implications of in Austria: current status and future strategies. In: Collares-Pereira, variation in mitochondrial DNA of Arctic grayling (Thymallus arc- M.J., Coelho, M.M., Cowx, I.G. (Eds.), Freshwater Fish Conserva- ticus). Mol. Ecol. 8, 23–25.
tion: Options for the Future. Fishing News Books, Blackwell Sci- Reist, J.D., Vuorinen, J., Bodaly, R.A., 1992. Genetic and morphologi- ence, Oxford, pp. 72–78.
cal identiWcation of Coregonid hybrids Wshes from Arctic Canada.
Weiss, S., Persat, H., Eppe, R., Schlötterer, C., Uiblein, F., 2002. Com- Pol. Arch. Hydro. 39, 551–561.
plex patterns of colonization and refugia revealed for European Reshetnikov, Yu.S., Popova, O.A., Sokolov, L.I., Tsepkin, E.A., Sidel- grayling Thymallus thymallus, based on complete sequencing of the eva, V.G., Dorofeeva, E.A., Chereshnev, I.A., Moskal'kova, K.I., mtDNA control region. Mol. Ecol. 11, 1393–1407.
Dgebuadze, Yu.Yu., Ruban, G.I., Korolev, V.V., 2002. Atlas pres- Zardoya, R., Doadrio, I., 1999. Molecular evidence on the evolutionary novodnykh ryb Rossii v dvukh tomakh (Atlas of Russian Freshwa- and biogeographical patterns of European Cyprinids. J. Mol. Evol.
ter Fishes in Two Volumes), vol. 1, Moscow, Nauka, 379 pp. (in 49, 227–237.
Zinoviev, E.A., 1980 Parallelizm izmenchivosti u evropeiskogo i sibi- Roukonen, M., Kvist, L., 2002. Structure and evolution of the avian rskogo khariusov//Lososevidniye rybi. ZIN AN SSSR, Leningrad, mitochondrial control region. Mol. Phylogenet. Evol. 23, 422–432.
pp. 69–80 (in Russian).

Source: http://old.biosoil.isu.ru/Library/KnizhinPDF/MPE2005Phylog.pdf