Phylogenetic position of the relict South American genus Idiopyrgus Pilsbry, 1911 (Gastropoda, Truncatelloidea), with the description of two new cave species

Idiopyrgus is a relict genus of freshwater snails from Brazil traditionally classified in the family Pomatiopsidae. Herein, we use mo lecular data from newly acquired specimens to test that classification through Bayesian inference phylogenetic analysis. We conclude that Idiopyrgus belongs in the Gondwanan family Tomichiidae, together with the African genus Tomichia and the Australian genus Coxiella . Furthermore, we reassess currently synonymized genus- and species-level names in Idiopyrgus . The genera Hydracme and Aquidauania are considered synonymous with Idiopyrgus . The species I. brasiliensis and I. pilsbryi are restored as accepted species; I. walkeri is considered a taxon inquirendum. Two new species from caves in Bahia state are described herein: Idiopyrgus adamanteus sp. nov. and Idiopyrgus minor sp. nov.


Introduction
Idiopyrgus Pilsbry, 1911 is a relict genus of operculated freshwater snails from Brazil, which, as presently understood, includes only two species (Simone 2006). Idiopyrgus (Fig. 1) is the single South American representative of the family Pomatiopsidae, which is mostly an Asian family, albeit with representatives in North America (Pomatiopsis Tryon, 1862), southern Australia (Coxiella E. A. Smith, 1894) and South Africa (Tomichia Benson, 1851) (Wilke 2019). The other potential South American Pomatiopsidae genus, Spiripockia Simone, 2012, was recently transferred to Cochliopidae based on new anatomical data .
This disjunct distribution has been difficult to explain and competing biogeographic hypotheses have proved impossible to be fully untangled (Wilke 2019). Phylogenetic works involving the family have mostly been concerned with its Asian branch, which is by far the most diverse (e.g., Attwood et al. 2003;Guan et al. 2008;Kameda and Kato 2011;Liu et al. 2014). The only study to include both Tomichia and Coxiella was Wilke et al. (2013), which recovered both genera as sister taxa, separated from Pomatiopsidae and hence recognized as the family Tomichiidae. Nevertheless, this classification scheme has not yet gained widespread acceptance and is absent from later works (e.g., Wilke 2019).
As such, the inclusion of the South American Idiopyrgus, largely ignored so far, is the final piece of this puzzle. Herein we use newly collected material to include Idiopyrgus in a molecular phylogenetic framework of freshwater Truncatelloidea, in an attempt to resolve the matters of phylog-
New specimens of Idiopyrgus spp. have been collected in Bahia state by the team at LES as part of cave surveys in the region. They come from the following localities (and caves): Central municipality (Gruta da Machadinha, Toca Bonita, Toca da Mulher, Toca de Candeias); Coribe municipality (Gruna Bem Bom); Feira da Mata municipality (Gruna da Pingueira II, Gruna do Engrunado). A paper is in preparation with a full account of the localities surveyed during those expeditions and the land and freshwater gastropods found.
The The holotype of Idiopyrgus brasiliensis (MZSP 14598; MS, Campo Grande) could not be found in the MZSP collection and is therefore considered lost. Pilsbry (1924) did not indicate the type material of I. walkeri, which is a very unusual situation for that author; the specimen(s) could not be located in the ANSP collection, where most of his types are housed.

Molecular data
One juvenile and one adult specimen (paratypes of each of the new Idiopyrgus species) preserved in ethanol 98% were used for DNA extraction. With the exception of the holotypes and other paratypes of these species, all other available specimens of Idiopyrgus spp. are dry shells only. DNA extraction followed the standard protocol of the QIAGEN DNEasy® Blood & Tissue Kit, with the addition of a repetition of the final step in order to increase yield. The following markers were target herein: (1) the barcoding fragment of the mitochondrial COI gene (c. 650 bp; primers LCO/HCO of Folmer et al. 1994); (2) the mitochondrial 16S rRNA gene (c. 450 bp; primers 16SarL/16SbrH of Simon et al. 1994); (3) the ITS2 region and the 5′ end of the 28S rRNA gene (c. 1,400 bp; primers LSU-1/LSU-3 and LSU-2/LSU-5 of Wade et al. 2006). The protocol for PCR were as follows. COI and 16S: initial denaturation at 96 °C (3 min); 35 cycles of denaturation at 95 °C (30 s), annealing at 48 °C (1 min), and extension at 72 °C (2 min); final extension at 72 °C (5 min). ITS2+28S: initial denaturation at 95 °C (3 min); 40 cycles of denaturation at 95 °C (30 s), annealing at either 50 °C (ITS2 section) or 45 °C (28S section) (1 min), and extension at 72 °C (5 min for ITS2 section; 2 min for 28S section); final extension at 72 °C (4 min).
The PCR products were assessed via agarose gel electrophoresis, cleaned (following the manufacturer's protocol) with ExoSAP-IT™ (Affymetrix Inc.), and sent to Massey Genome Service (Massey University, Palmerston North, New Zealand) for Sanger sequencing. Sequences were quality-proofed and assembled in Geneious Prime (v.2020.2.2, Biomatters Ltd.). The consensus sequence of each marker was uploaded to GenBank (registration numbers are listed under each species entry below).
The DNA sequences for the phylogenetic analysis were extracted from GenBank (Table 1), with a focus on Pomatiopsidae and Tomichiidae. Representatives of other closely-related families were also selected, including those families and genera in which Idiopyrgus has been historically classified or been considered related to, such as Amnicolidae, Assimineidae and Tateidae (e.g., Pilsbry 1924;Morretes 1949;Rey 1959). A Littorinidae species was chosen as outgroup (Table 1).
Sequence alignment was conducted in Geneious Prime via the MAFFT plugin (v.7.450;Katoh et al. 2002;Katoh and Standley 2013) with default settings. The resulting alignment of each marker (COI, 16S, and ITS+28S) was visually proofed for inconsistencies and then run through Gblocks (Castresana 2000;Talavera and Castresana 2007) with the least restrictive settings to eliminate poorly-aligned or data-deficient positions that could hamper the analysis. The resulting post-Gblocks alignments were concatenated (2,610 bp in total) for a single Bayesian inference phylogenetic analysis, conducted with MrBayes (v.3.2.7, Ronquist et al. 2012) via the CIPRES Science Gateway (v. 3.3, Miller et al. 2015). Two concurrent analyses were run, each with 4 Markov chains of 40 million generations (the first 20% discarded as 'burn-in'), the default priors, nst = 6, rates = invgamma, temperature parameter = 0.1, sampling every 1,000 generations, and with substitution model parameters unlinked across the markers (COI, 16S, ITS+28S). MCMC convergence was assessed by examining the standard deviation of split frequencies (<0.01) and the potential scale reduction factor (PSRF ~1.0), as well as trace plots in Geneious (Ronquist et al. 2009).

Results and discussion
The specimens used herein for DNA analysis were identified as two new species of Idiopyrgus, described below: Idiopyrgus adamanteus sp. nov. and Idiopyrgus minor sp. nov.
Distribution. Known from four caves in Central municipality.
Etymology. From Latin, meaning 'of diamond'; an allusion to the region in Bahia known as Chapada Diamantina, where all the caves in the species' known distribution are located.
Diagnosis. Elongated shell spire; narrow body whorl; more circular aperture when compared to congeners.
Description. Shell small, truncatelloid, of translucent beige to whitish color. Protoconch bulbous and translucent, of fragile appearance (it is broken off in many specimens). Spire narrow. Shell smooth except for fine growth lines; growth lines can be more marked in some specimens, forming a light axial ribbing pattern (Fig. 2E, G). Suture deep; subsutural cord observed in one specimen (Fig. 2E). Whorl profile markedly convex, rounded. Whorls initially increasing regularly in size with growth, but final whorls are typically similar in size ( Fig. 2A, D-G). Aperture small in proportion to body whorl; oval to circular, with a more acuminate end towards the upper lip insertion. Peristome complete; reflexed but not thickened. Shell rimate. Operculum translucent, of same color as the shell. Soft body white.
Remarks. Idiopyrgus adamanteus sp. nov. is most similar in shell shape to I. souleyetianus and I. brasiliensis, but it can be diagnosed from them by the fewer number of whorls and the much narrower body whorl, as well as a more circular aperture (Fig. 1). It can be distinguished from I. rudolphi by the deeper suture, more convex whorl profile and more circular aperture. It differs from the remaining congeners by its narrower shell, more elongated spire and smaller aperture (in proportion to body whorl).
Idiopyrgus adamanteus sp. nov. displays a fair amount of morphological variation in the shell, even in specimens Table 1. Species used in the phylogenetic analysis, with GenBank registration number of each marker, locality data, and reference. The taxa are arranged alphabetically for ease of use; type species of the genera are indicated by an asterisk after their name. from the same cave ( Fig. 2A-D, F-H), like what is observed in many other freshwater truncatelloids. The shell and the spire of some specimens can be shorter (Fig. 2C) or taller and more elongated (Fig. 2D), the body whorl can be slighter wider (Fig. 2C, H), the shell sculpture can be more prominent (Fig. 2E, G), and a subsutural cord can be present (Fig. 2E). Given its occurrence in four distinct caves of Central region (Bahia state), Idiopyrgus adamanteus sp. nov. is considered a troglophilic species rather than a strict troglobitic one. The region of Central is characterized by a dry vegetation type (Caatinga) and the caves represent good shelters for these animals, given their high-humidity microhabitats. Few collections were conducted outside the caves and, as such, this species occurrence outside subterranean habitats was not tested. Nevertheless, the lack of body pigmentation and the translucent shell of some specimens ( Fig. 2A-H) are possible troglomorphisms, i.e., morphological traits typically observed in cave-dwelling gastropods (Christiansen 2012;Trontelj et al. 2012;Salvador et al. 2022).
The caves of Central municipality and their surroundings are impacted by deforestation caused by subsistence agriculture, mining activities, and wind energy installations in the karst landscapes (Bichuette, pers. obs.).
Distribution. Known only from two caves in Feira da Mata municipality.
Etymology. From Latin, meaning 'little'.  Diagnosis. Smaller shell than congeners; spire not as elongated and wider than most congeners; aperture proportionately larger in relation to preceding whorl.
Description. Shell minute, truncatelloid, of translucent white color. Protoconch not too prominent. Shell smooth except for fine growth lines. Suture deep, rendering whorls lightly shouldered in holotype (Fig. 2I, J). Whorl profile convex. Whorls increasing regularly in size with growth. Aperture large in proportion to body whorl; oval to circular, with a more acuminate end towards the upper lip insertion. Peristome complete; reflexed but not thickened. Shell rimate. Operculum translucent, of same color as the shell. Soft body white.
Remarks. Idiopyrgus minor sp. nov. can be distinguished from its congeners by its small size. It also has a less elongated and wider spire than most of its congeners, the exceptions being I. pilsbryi and I. walkeri (Fig. 1). Nevertheless, it can also be easily diagnosed from them: I. minor sp. nov. is much smaller and has a narrower shell than I. walkeri, and has a larger and wider aperture than I. pilsbryi. Finally, there is a reasonably large genetic distance separating I. minor sp. nov. from I. adamanteus sp. nov. (Fig. 3), comparable to the distance between the species of Truncatellidae and Stenothyridae present on the tree and larger than the distances between species of Pomatiopsidae. This species is considered troglobitic, given its occurrence in a single cave of the Serra do Ramalho karst area. The specimens were collected in pools resulting from infiltration water and seem to be restricted to this habitat. Besides the translucent shell and lack of body pigmentation, the smaller body size of I. minor sp. nov. can also be interpreted as a possible troglomorphism, given that miniaturization is common in cave organisms, including snails (Christiansen 2012;Salvador et al. 2022).
The Serra do Ramalho region has been historically exploited for agriculture, charcoal production, and mining, from small to large scales (Gallão and Bichuette 2018).

Phylogenetics and biogeography
The phylogenetic analysis included a total of 51 species (counting the outgroup) and all family level taxa were well-supported (posterior probability, PP = 1 in all cases except Falsicingulidae, for which PP = 0.98). As such, even though there is an unsolved polytomy in the tree ( Fig. 3; in all likelihood a result of the exclusion of northern hemisphere families not immediately related to our question), the membership of each species to a particular family can be readily and reliably assessed.
In that regard, it can be seen that Idiopyrgus does not belong to Pomatiopsidae; rather, it is grouped together with Tomichia and Coxiella. These three genera together form the family Tomichiidae, which recovers the results of Wilke et al. (2013) and fully establishes this family as a Gondwanan clade. The most basal branch in Tomichiidae appears to be the Australian Coxiella, with the South African Tomichia and the South American Idiopyrgus spp. forming a well-supported monophyletic clade (PP = 0.97; Fig. 3). This represents a typical Gondwanan scenario, with the Australian branch being the most basal one. Furthermore, it can be surmised that all three genera are relicts in their respective continents, given their present low diversity and restricted distributions (Wilke 2019).
Nevertheless, the diversity of the Brazilian branch of this family might be underestimated. Specimens of I. souleyetianus have been reported from several localities in Brazil, including the aphotic zones of caves (e.g., Thiengo et al. 2005;Salvador et al. 2017;Salvador and Simone 2021). Shell morphology and the apparent isolation of some populations in cave systems indicate, however, that "I. souleyetianus" from the literature is likely a species complex (see also the discussion below). The two new species described herein from caves also point towards this conclusion. Future morphological and molecular studies might thus shed light on the real diversity of this group in Brazil.
It is also noteworthy that Tomichiidae is absent from Zealandia, especially considering that the family Tateidae is likewise a Gondwanan clade and has representatives in South America, Australia and New Zealand ( Fig.  3; Zielske et al. 2016;Ponder 2019). Given the more basal position of Tateidae in relation to Tomichiidae, it can be hypothesized that it represents an earlier branch of Truncatelloidea and that Tomichiidae therefore arose after Zealandia had split from Gondwana (circa 80 Ma).

Taxonomy
Based on the study of type material and further specimens available to us, it was possible to reassess those species of Idiopyrgus that are presently considered synonyms (as per Simone 2006) of I. souleyetianus. This assessment was done based on conchological characters of the type specimens, topotypes, and additional voucher specimens, as no additional DNA-grade material could be acquired. Therefore, we are aware that our proposed arrangement may change in the future when more material becomes available.
I. souleyetianus (Fig. 1A, B) is here restricted to populations from Espírito Santo state (Southeast Atlantic hydrographic region), which includes the type locality, and Bahia state (São Francisco hydrographic region; specimens reported by Salvador et al. 2017) (Fig. 4). The specimens from northern Minas Gerais (São Francisco hydrographic region; Fig. 4) assigned by Salvador and Simone (2021) to I. souleyetianus display much smaller shells, with fewer whorls and a weak axial teleoconch sculpture. As such, they might represent a still unrecognized species. Likewise, the records from Goiás state assigned to I. souleyetianus by Thiengo et al. (2005: not illustrated) stem from an area belonging to yet another hydrographic region (Tocantins-Araguaia; Fig. 4) and should be reassessed; they are treated here as Idiopyrgus sp.
Idiopyrgus pilsbryi (Fig. 1C, D) has a much smaller shell than I. souleyetianus, having circa 2/3 of the latter's shell length. The whorls of I. souleyetianus are taller than those of I. pilsbryi, and it has a taller spire with more whorls. I. pilsbryi inhabits Rio Grande do Norte state, ca. 1,500 km north of the type locality of I. souleyetianus, in a completely different climate and biome. It is geographically isolated from I. souleyetianus, whose northernmost occurrence is in central Bahia state (Fig. 4), and in a different hydrographic region (Eastern Northeast Atlantic region). As such, we reinstate I. pilsbryi as an accepted species.
Idiopyrgus walkeri was described from Minas Gerais state and its type specimen could not be located in the ANSP collection during the present study. By its description and published illustrations (Pilsbry 1924, reproduced here as Fig. 1E), it has a more conical shell and wider and more convex whorls than I. souleyetianus (being of similar size to it), alongside a shorter spire with fewer whorls. It inhabits the São Francisco hydrographic region (Fig. 4), but on a different area from where I. souleyetanus is known. In absence of type material, I. walkeri is here considered potentially distinct from I. souleyetanus, but it is regarded as a taxon inquirendum until the types come to light or topotypes are collected. Three specimens collected from caves in Coribe municipality, southern Bahia state (Fig. 4), have wider shells and could represent specimens of I. walkeri.
Idiopyrgus rudolphi (Fig. 1F, G) is the most immediately diagnosable species, as its shell has a straighter spire profile, almost subulinid-like, with whorls only slightly convex in profile. This feature was deemed sufficient by Haas (1938) to describe the genus Hydracme to house this species. Considering the morphological variation known in the shell shape of truncatelloids, we prefer to follow the most conservative approach for now and to maintain Hydracme as a synonym of Idiopyrgus until molecular studies on I. rudolphi can be conducted and its position in the group's phylogeny assessed. I. rudolphi is known only from its type locality in Pernambuco state (São Francisco hydrographic region), seemingly isolated from I. souleyetianus to the south and I. pilsbryi to the north (Fig. 4).
The type material of Idiopyrgus brasiliensis, despite being databased in the MZSP collection, could not be located during the present study. According to the species description and published illustrations (Rey 1959, reproduced here as Fig. 1H), this species is very similar in shell shape and size to I. souleyetianus, but it has a taller body whorl, a more vertically positioned aperture, and a less reflected lip. It is also geographically isolated from other Idiopyrgus spp. (Fig. 3). Therefore, we reinstate it as an accepted species. It inhabits Mato Grosso do Sul state (Paraguay hydrographic region) and, given its occurrence in Corumbá municipality (Rey 1959) close to the border, it should likely be found in Bolivia and perhaps Paraguay as well. The record from Mato Grosso state (Simone 2006;Birckolz et al. 2016) is in all likelihood a misinterpretation of the records of Rey (1959), since back then the state of Mato Grosso still contained the area that later became Mato Grosso do Sul state. The genus Aquidauania, which was erected to house I. brasiliensis (erroneously described as belonging to Oncomelania Gredler, 1881, an Asian genus; Rey 1959), is therefore a synonym of Idiopyrgus¸ as already indicated by Simone (2006). The two new species described herein, I. adamanteus sp. nov. and I. minor sp. nov., are from Bahia state, belonging to the Middle São Francisco hydrographic region (Fig. 4). Idiopyrgus adamanteus sp. nov. is apparently isolated from I. souleyetianus in the south and I. rudolphi in the north (Fig. 4). The distribution of I. minor sp. nov., however, overlaps with I. souleyetianus and potentially with I. walkeri as well (Fig. 4). It is thus tempting to consider I. souleyetianus and/or I. walkeri as widespread species whose isolated populations in caves might have given rise to new species such as I. minor sp. nov.

Conclusion
Based on our results, we transfer Idiopyrgus (and Coxiella) from Pomatiopsidae to Tomichiidae, and reinstate previously synonymized species-level taxa. Hence, we propose the following arrangement for family Tomichiidae and, in particular, the genus Idiopyrgus.