A NOTE ON THE STATUS OF GALBA OCCULTA JACKIEWICZ , 1959 ( GASTROPODA : HYGROPHILA : LYMNAEIDAE )

agreement no 534/P-DUN/2018 of April 4th, 2018 allocated to the activities for disseminating science: Task 1: Preparation of English versions of publications (sum funded by DUN 12,000 PLN) and Task 2: Digitalisation of publications and scientific monographs to enable their open access in the Internet (sum funded by DUN 11,070 PLN). https://doi.org/10.12657/folmal.026.029 Folia Malacol. 26(4): 231–247


INTRODUCTION
recognised a new species within the complex species Galba palustris (O.F. Müller, 1774) and named it G. occulta.The species was subsequently assigned to Stagnicola Jeffreys, 1830 treated as a subgenus within Lymnaea Lamarck, 1799 (jackiewicz 1993, 1998a, 2000) or as a separate genus (FaLkneR 1995, FaLkneR et al. 2001, GLöeR 2002, GLöeR & MeieR-bRook 2003).When taxonomic decisions based on the analysis of nucleotide sequences took on significance, MeieR-bRook & baRGues (2002), taking into account the length of ITS2 sequence in rDNA gene, included G. occulta Jackiewicz in a newly established genus Catascopia.They stated after waLteR (1969) that C. occulta was an American species introduced in Europe.Later VinaRski & GLöeR (2008) found that although jackiewicz (1992, 1998b) was aware of the occurrence of G. occulta in Siberia, she overlooked that the species discovered by her had been described earlier by westeRLund (1885) as Limnaea palustris var.terebra.VinaRski (2012) moreover argued that Ladislavella B. Dybowski, 1913 was the oldest available name for Catascopia Meier-Brook et Bargues, 2002.Finally, Galba occulta Jackiewicz, 1959 was assigned as a junior synonym of Ladislavella terebra (Westerlund, 1885) (andReyeVa et al. 2010, VinaRski 2012) and this name was used in subsequent publications (e.g.VinaRski 2012, Pieńkowska et al. 2014, 2015a, schniebs 2016, VinaRski et al. 2016a, Piechocki & wawRzyniak-wydRowska 2016, schniebs et al. 2018). hebeRt et al. (2003a, b) proposed that the nucleotide sequence of the cytochrome oxidase subunit 1 gene (COI) could be a marker that would allow to distinguish species, with suggestion that 3% genetic distance could be treated as a threshold between separate taxa at the species level.Usefulness of Hebert's barcoding in taxonomy was supported by many authors (e.g. tautz et al. 2003, GReGoRy 2005, PackeR et al. 2009, GoLdstein & desaLLe 2011) with some suggestion that the threshold should be higher for stylommatophoran gastropods (daVison et al. 2009, saueR & hausdoRF 2012).However schniebs et al. (2016) decided that COI sequences could not be used in stagnicoline lymnaeid taxonomy and excluded this gene from their molecular studies (e.g.schniebs et al. 2015, 2017, 2018, VinaRski et al. 2017).On the other hand aksenoVa et al. ( 2018) presented a deep revision of lymnaeid classification with a huge base of COI sequences for species identifications.
In this paper we compare COI and ITS2 sequences of L. terebra specimens from its Siberian and West Poland populations with the aim of restoring the validity of the taxon described by jackiewicz (1959) to commemorate outstanding achievements in lymnaeid taxonomy of this malacologist who died this year.
Sequences were prepared using the programme BIOEDIT, version 7.0.5.(haLL 1999).The alignments were performed using the CLUSTAL W programme (thoMPson et al. 1994) implemented in BIOEDIT.The ends of all sequences were trimmed to obtain four sets of equal length sequences: COI, ITS2 and COI+ITS2.The lengths of the sequences after cutting were 558 and 461 bp for COI, 488 positions for ITS2 and 1,014 positions for combined sequences of COI+ITS2 (558 bp + 456 positions).The sequences were collapsed to haplotypes (COI) and to common sequences (ITS2 and COI+ITS2) using the programme ALTER (Alignment Transformation EnviRonment) (GLez-Peña et al. 2010).
During analysis of the phylogenetic relationships, the sequences were analysed by the genetic distance Neighbour-Joining method (saitou & nei 1987) implemented in MEGA7 (kuMaR et al. 2016) using the Kimura two-parameter model (K2P) for pairwise distance calculations (kiMuRa 1980).Best-fit substitution models were calculated using algorithm implemented in MEGA 7 for every set of sequences independently: Tamura 3 parameter evolutionary model (taMuRa 1992) for COI alignments, Kimura 2-parameter model (kiMuRa 1980) for ITS2 set and a HKY substitution model for combined data set of COI+ITS2 (haseGawa et al. 1985).For all analyses we assumed a gamma distributed rate variation among sites.Maximum Likelihood analyses were performed using MEGA7.
The ML trees were tested by bootstrap analysis with 1,000 replicates (FeLsenstein 1985).In the case of combined alignment COI+ITS2 parallel Bayesian Interference was conducted using the programme MRBAYES 3.1.2(Ronquist & hueLsenbeck 2003).Four Monte Carlo Markov chains were run for 1 million generations, sampling every 100 generations (the first 250,000 trees were discarded as 'burn-in').This yielded a 50% majority rule consensus tree.Finally, calculated during ML analysis bootstrap values were mapped on the 50% majority rule consensus Bayesian tree.
To compare the above haplotypes with eighty COI sequences for Hinkleya caperata and one for L. elodes deposited in GenBank by MoRninGstaR et al. (2018) anddewaaRd et al. (2014), respectively (24 additional halotypes COI 54 -COI 77), they all had to be trimmed to 461 bp long haplotypes.The resulting ML tree (Fig. 2) showed similar clades as for longer haplotypes (Fig. 1) with one additional clade of haplotypes characteristic of H. caperata.The ITS2 sequences obtained from GenBank were attributed to 23 common sequences (Appendix 1).They clustered in seven clades on the ML tree (Fig. 3), three (S.corvus, S. palustris and L. stagnalis) clearly and four (L.occulta, L. terebra, L. elodes and L. tum-rokensis) less separated.The smallest K2P distances (Table 1) differentiated specimens within particular species (mean values up to 1.6%), except L. terebra with its intraspecific variation larger (5.3%).The largest K2P distances were between species included   2).However it is much larger than the number differences in ITS2 sequences between L. elodes and L. tumrokensis (Table 2) and much smaller than between L. occulta or L. terebra and species of Stagnicola or Lymnaea (data not shown).

DISCUSSION
No differences were found in the structure of the shell and reproductive system between topotypical material of Galba occulta Jackiewicz and Siberian Limnaea palustris var.terebra Westerlund in careful comparative studies (VinaRski 2003, 2012, VinaRski & GLöeR 2008).Moreover, specimens of both species were closely related based on their ITS2 sequences in rDNA gene (VinaRski et al. 2016a).Therefore VinaRski & GLöeR (2008) synonymised the taxa giving a priority to the older name.Somewhat later VinaRski (2012) assigned it to the genus Ladislavella B. Dybowski, 1913 as L. tere bra (Westerlund, 1885).
We found that the specimens from the Polish population in Gorzykowo near Gniezno (W.Poland) differed in their nucleotide sequences of COI and ITS2 fragments from the two Siberian populations (Tjumen Region and Altai Republic, Russia) (Figs 1-4).The K2P distances between the common sequences of ITS2 fragment suggest that these populations are closely related (as suggested by VinaRski et al. 2016a).However the differences of ITS2 sequences between the Altai and Tjumen populations require further in-depth research on a larger number of populations.On the other hand, K2P distances between the haplotypes COI 1 -COI 4 and COI 5 & COI 6, representing these two groups (Polish and Siberian), respectively, are much higher (12.2-12.8%)than the 3% threshold established by hebeRt et al. (2003a,  1).L. occulta or L. terebra were identified on the basis of shell and genital system features from several localities from Europe (West Poland, South Sweden, Czech Republic, Bosnia-Hercegovina, Ukraine) and Siberia (Yeniseysk and Selenga River near Baikal lake) by jackiewicz (1992, 1993, 1997, 1998a, b) (who used the name Lymnaea (Stagnicola) occulta); several localities in Germany and Rusia (especially in Siberia and Far East) were added to its distribution by VinaRski & GLöeR (2008) and recently from Ukraine (Khust district) by anistRatenko et al.
(2018) (using the name Ladislavella terebra).However L. occulta or L. terebra at these localities were identified on the basis of shell and genital system features.We report differences in COI and ITS2 sequences between one Polish (Gorzykowo, W. Poland) and two Russian (Siberia) localities.Further molecular research on L. occulta and L. terebra populations from other localities is necessary.Unfortunately all but one (Gorzykowo) Polish localities found by Jackiewicz and her co-workers (jackiewicz 1959, 1993, 1998a, 2000) were destroyed, so the Polish L. occulta lineage is threatened with extinction (Rybska et al. 2007).
Although this was not the aim of this study, we add a few remarks about lymnaeid taxa resulting from our analysis of the COI gene sequences:

Fig. 1 .
Fig.1.Maximum Likelihood (ML) tree of the 558-bp-long fragment of COI sequences of the studied lymnaeid species with the use of Lymnaea stagnalis as outgroup (see Appendix 1).Numbers on branches represent bootstrap support above 50%

Fig. 4 .
Fig. 4. Majority-rule consensus tree obtained from Bayesian Inference analysis (BI) of the combined data set of COI and ITS2 DNA sequences (see Table3) of the studied lymnaeid species.The tree was rooted with L. stagnalis combined sequences.Posterior probabilities (left) and bootstrap support above 50% from ML analysis (right) are marked on branches

Table 1 .
Ranges of K2P genetic distances for COI and ITS2 sequences analysed (mean values in parentheses) Fig. 2. Maximum Likelihood (ML) tree of the 461-bp-long fragment of COI sequences of the studied lymnaeid species with the use of L. stagnalis as outgroup (see Appendix 1).Shortening the length of the sequences within alingment allowed to add sequences of Hinkleya caperata to the phylogenetic analysis.Numbers on branches indicate bootstrap support above 50% Fig. 3. Maximum Likelihood (ML) tree of the 488-position-long fragment of ITS2 sequences of the studied lymnaeid species with the use of L. stagnalis as outgroup (see Appendix 1).Numbers on branches indicate bootstrap support above 50%

Table 3 .
Combined sequences of COI and ITS2 fragments for Bayesian analysis

Table 2 .
Number of differences in ITS2 nucleotide sequences between analysed species of genus Ladislavella in Stagnicola and Lymnaea clades and those grouped in four Ladislavella clades analysed in pairs (mean values 31.0-37.9%).Mean values of the K2P distance within Ladislavella species support their separation (L.occulta vs. L. terebra 4.4%, L. occulta vs. L. tumrokensis 6.8%, L. occulta vs. L. elodes 7.0%, L. terebra vs. L. tumrokensis 8.3%, L. terebra vs. L. elodes 8.1%), except a pair L. tumrokensis and L. elodes (mean K2P distance 0.6%).It is noteworthy that the number of differences between ITS2 sequences of L. occulta and L. terebra is smaller than those distinguishing L. occulta from L. elodes and L. tumrokensis as well as L. terebra from L. elodes and L. tumrokensis (Table 1. COI gene sequences well support the generic classification of the following taxa: Ladislavella B. Dybowski, 1913, Hinkleyia F. C. Baker, 1928, Stagnicola Jeffreys, 1830 and Lymnaea Lamarck, 1799.2. The results of this paper support the suggestion that L. tumrokensis and L. elodes are conspecific (according to VinaRski et al. 2017 they represent two subspecies of Ladislavella catascopium).3. The status of Stagnicola palustris should be verified.We found that this species was represented by two different COI lineages.schniebs et al.
(2016)pointed to the inconsistency of stagnicoline lymnaeid classification based on mitochondrial sequences vs. that resulting from the analysis of nuclear genes and reproductive system anatomy.Drawing conclusions on the basis of our results would be premature.Further research is needed on a larger number of S. palustris populations to determine if this difference in COI sequences is a result of interspecific hybridisation or speciation visible in the mitochondrial genome.4.Although Lymnaea stagnalis is a well defined species, it is also greatly diversified in COI nucleotide sequences which suggests that further studies on its populations would be necessary.