NEW EVIDENCE ON THE RELATIONSHIPS BETWEEN HYPNOPHILA BOURGUIGNAT, 1859 AND GOMPHROA WESTERLUND, 1902 (GASTROPODA: EUPULMONATA: AZECIDAE)

analysis of nucleotide sequences of mitochondrial COI and nuclear 5.8S+ITS2+28S gene fragments was performed on newly obtained specimens of Hypnophila pupaeformis (Cantraine). The results partially agree with previous morphological (shell and genitalia) analysis. They support separateness of H. pupaeformis from all species assigned to Gomphroa, Cryptazeca, Hypnocarnica and Azeca. They also show close relationships of H. pupaeformis with the Gomphroa group. Indeed Hypnophila and Gomphroa form a clade consisting of four subclades: Hypnophila and three lineages named provisionally Gomphroa A, Gomphroa B and Gomphroa C. However, more research is needed to determine their relationships and to establish whether Hypnophila and Gomphroa are two genera or two (or even four) subgenera of one genus. key words: Azecidae; genera; systematics; molecular features; COI; ITS2; nucleotide sequences

Division of Hypnophila s.l. into two genera is well supported by morphological analysis (shell features and genital anatomy). Molecular studies (analysis of nucleotide sequences of selected fragments of mitochondrial and nuclear genes) have confirmed that the species included in the genus Gomphroa form a closely related group, but molecular comparison with species of true Hypnophila has not hitherto been undertaken. Molecular analysis of newly obtained material of Hypnophila pupaeformis is presented in this paper.
DNA e x t r a c t i o n , a m p l i f i c a t i o n a n d s e q u e n c i n g Small foot tissue fragments of alcohol preserved snails were used for total DNA extraction with Tissue Genomic DNA extraction Mini Kits (Genoplast) according to the manufacturer's instructions. The purified total DNA was used as template for amplification by polymerase chain reaction (PCR) of partial sequences, using the following primers: for COI -two Folmer's "universal" primers LCO1490 (5'-GGTCAACAAATCATAAA-GATATTGG-3') and HC02198 All polymerase chain reactions were performed in a volume of 10 μl. The amplified COI fragments, consisting of 710 base pairs (bp), were obtained under the following thermal profile: 5 min at 95 °C followed by 35 cycles of 30 s at 95 °C, 1 min at 50 °C, 1 min at 72 °C, and finally 5 min at 72 °C using the Type-it Microsatellite PCR Kit (Qiagen). Amplification products of ITS2 with 5.8S and 28S flanking fragments of 944-945 bp (including 52-53, 573 and 319 bp for 5.8S, ITS2 and 28S, respectively) were obtained using the same cycling parameters. Two rounds of amplifications were performed: the first with the purified total DNA as template and the second with 1 μl of the 10× diluted product from the first round as template. The amplified H3 sequences consisted of 429 bp. PCR reactions (10 μl) were performed according to the procedure described by coLGan et al. (1998).
The PCR products were verified by agarose gel electrophoresis (1% agarose). Prior to sequencing, samples were purified with thermosensitive Exonuclease I and FastAP Alkaline Phosphatase (Fermentas, Thermo Scientific). Finally, the amplified products were sequenced in both directions with BigDye Terminator v3.1 on an ABI Prism 3130XL Analyzer (Applied Biosystems, Foster City, CA, USA) according to the manufacturer's protocols.
Sequences were edited by eye using the programme BIOEDIT, version 7.0.6 (Hall 1999). The alignments were performed using the CLUSTAL W programme (tHoMpson et al. 1994) implemented in MEGA 7 (kumar et al. 2016. The COI and H3 sequences were aligned according to the translated amino acid sequences. Gaps and ambiguous positions were removed from COI alignments prior to phylogenetic analysis. The ends of all sequences were trimmed. The lengths of the COI and H3 sequences after cutting were 476 and 252 bp, respectively. Sequences consisting of the 3'-end of 5.8S, ITS2 and 5'-end of 28S were aligned with sequences from GenBank. The alignment of all sequences was 1,052 positions (base pairs+indels) in length. In the analysis of ITS2 and 28S, treated separately, the alignments were 784 and 319 positions in length, respectively. The sequences were collapsed to haplotypes (COI) and to common sequences (5.8S+ITS2+28S) using the programme ALTER (Alignment Transformation EnviRonment) (GLez-Peña et al. 2010). Finally COI haplotypes and 5.8S+ITS2+28S common sequences were joined into concatenated sequences COI+(5.8S+ITS2+28S) and the resulting alignment was 1,318 positions in length (476 COI + 842 5.8S+ITS2+28S). P h y l o g e n e t i c i n f e r e n c e The sequences deposited in GenBank are shown in Table 1.

RESULTS
Two new COI, four 5.8S+ITS2+28S and four H3 sequences were obtained from the specimens of H. pupaeformis from Croatia and deposited in GenBank (Table 1). Partial sequences of mitochondrial COI and nuclear 5.8S+ITS2+28S gene fragments were compared with sequences of these genes deposited in GenBank by other authors (see : Table 1) (H3 sequences were not used in phylogenetic analysis   Table 1). Numbers next to branches indicate bootstrap support above 50% calculated for 1,000 replicates (Felsenstein 1985).  Table 1). Numbers next to branches indicate bootstrap support above 50% calculated for 1,000 replicates (Felsenstein 1985). The tree was rooted with Cochlicopa lubrica concatenated sequence of MF545160 and AY014019, deposited in GenBank by dewaard (2017)  because no reference sequences could be found in GenBank resources). ML trees with phylogenetic analysis of single locus datasets of COI, ITS2 and 28S (not shown) and the multilocus dataset of concatenated 5.8S+ITS2+28S sequences (Fig. 1) showed that H. pupaeformis sequences were grouped on distinct branches. The same result was obtained for concatenated COI+(5.8S+ITS2+28S) sequences in ML (Fig. 2) and BI (Fig. 3) analysis.

DISCUSSION
Phylogeny based on morphological characters, which was presented in a previous paper , showed that true Hypnophila species belong to a monophyletic group supported by two synapomorphies: the elongate ovoid-cylindrical shell and the cup-like initial por-tion of one of the two penial plicae bordering the vas deferens opening into the penis. This clade constituted the sister group of Azeca based on loss of the rows of pits on the protoconch. In turn, Azeca plus Hypnophila was the sister group of the lineage including Gomphroa species except G. boissii, based on  Table 1). Posterior probability values are indicated next to the branches. The tree was rooted with Cochlicopa lubrica concatenated sequence of MF545160 and AY014019, deposited in GenBank by dewaard ( (Giusti & ManGanelli 1984, weLter-scHuLtes 2012, ŠtaMol et al. 2018) should be divided into two separate genera: Gomphroa and Hypnophila s.str. The former included nine species of Gomphroa, all but one of which occur in the western Mediterranean (the one exception is the Dalmatian G. zirjensis). The latter comprised the remaining four species of Hypnophila with distribution in the western Balkan Peninsula, including the western Balkan islands. Two species of Gomphroa occurring in north western Africa (G. maroccana (Mousson, 1873), G. psathyrolena (Bourguignat, 1859)) were not included in the analysis because they were only known from the original description and very few other contributions (ŠtaMol et al. 2018. The new molecular data strongly support the separateness of H. pupaeformis from all other azecid species as well as its close relationships with the Gomphroa group. Indeed, each analysis of gene sequences obtained from H. pupaeformis, i.e. those concerning separate analysis of each gene (mitochondrial COI or nuclear ITS2 and 28S) as well as those of concantenated sequences (5.8S+ITS2+28S, Fig. 1; COI+(5.8S+ITS2+28S) , Figs 2-3), showed a clearly distinct branch for H. pupaeformis, separate from those of species belonging to Gomphroa, Hypnocarnica, Cryptazeca and Azeca, on the phylogenetic trees. The K2P distances of COI sequences found in this paper are similar to those published by , which are now supplemented by analysis of COI from H. pupaeformis (not previously available). Some differences in the results : table 4 and this paper: Table 2) are derived from the need to trim the COI sequences to 476 bp. However, the branch for H. pupaeformis sequences forms a subclade within the group of species assigned to Gomphroa in each tree, indicating that Gomphroa is paraphyletic. Gomphroa sensu Manganelli et al. (2019) may be divided into three subgroups named provisionally Gomphroa A, Gomphroa B and Gomphroa C (Figs 1-3). The K2P distances within and between these groups are similar, and similar K2P distances also distinguish all Gomphroa groups and H. pupaeformis (Table 3). Gomphroa A includes some species from Provence, Tuscany, the Tuscan Archipelago, Sardinia and the Pontine Archipelago; Gomphroa B includes the Sardinian G. dohrni and the Corsican G. remyi; Gomphroa C includes two Iberian species (cianFaneLLi et al. 2018b, ŠtaMol et al. 2018. The relationships between these groups and Hypnophila are still unclear: Hypnophila may be the sister group of Gomphroa A; in turn, this clade has unresolved relationships with Gomphroa B and Gomphroa C (Fig. 1); Hypnophila may have unresolved relationships with Gomphroa A and the clade consisting of Gomphroa B plus Gomphroa C (Fig. 2); Hypnophila may be a sister group of Gomphroa A; in turn, this clade is a sister group of Gomphroa B and in turn the last clade is a sister group of Gomphroa C (Fig. 3). The division of Gomphroa into three separate subgroups is not supported by any morphological feature. Although the Sardo-Corsican Gomphroa B may be distinct due to a proportionally smaller penis (ManGanelli et al. 2019), the Iberian Gomphroa C includes species with a "normal" penis (G. malagana) as well as species with a micropenis (G. boissii). On the contrary, the distinction between Gomphroa and Hypnophila is also supported by some shell and genital features (ManGanelli et al. 2019). We have always stressed (Pieńkowska et al. 2018) that molecular features alone are insufficient to make taxonomic conclusions but that they must be supported by morphological and anatomical features. Thus any taxonomic conclusion concerning the relationship between the genera Gomphroa and Hypnophila seems to be premature. At the moment we can only confirm the separateness of H. pupaeformis from all species assigned to Gomphroa, as well as to Cryptazeca, Hypnocarnica and Azeca. More research is needed to determine whether Hypnophila and Gomphroa represent two genera, or two (or even four) subgenera of one genus. Further research should include at least some of the other Greek Hypnophila species (H. polita, H. cyclothyra and H. zacynthia) and some other Gomphroa species such as the Dalmatian G. zirjensis and one or more Sicilian species. Nor is any division of Gomphroa into further subgenera possible at the present time. In our analysis, we again used sequences deposited in GenBank by cianFaneLLi et al. (2018b) for single specimens representing particular species assigned to Gomphroa . Consequently, more specimens of at least some of Gomphroa species need to