Systematic evaluation of cryptic freshwater snails from central Chile, including the enigmatic Littoridina santiagensis (Gastropoda, Truncatelloidea)

. Walter Biese described Littoridina santiagensis Biese, 1944 (Cochliopidae) from Estero Dehesa in 1944 based exclusively on external shell features and a second allopatric population in Yeso Spring three years later. Since 2011 different samplings have been carried out at the type locality and have only provided specimens of the morphologically similar invasive mudsnail Potamopyrgus antipodarum Gray, 1843 (Tateidae), raising doubts about the identity of the species. The recent finding of two snail morphotypes in Yeso Spring, a thick shelled form congruent with type specimens of L. santiagensis and a slender one morphologically associable to P. antipodarum , allowed comparative studies, including the taxonomic analysis of additional populations with similar shell morphology occurring in central Chile. A DNA barcoding (COI) approach identified the slender form from Yeso Spring in Maipo Basin and a second population from the contiguous Rapel Basin indeed as the invasive P. antipodarum ; however,


Introduction
The taxonomic placement of minute Chilean freshwater snails solely described based on morphological characters, such as Littoridina santiagensis Biese, 1944, is rather controversial.This species was described in the genus Littoridina Souleyet, 1852 by Walter Biese from Estero Dehesa, a small tributary of the Mapocho River located east of Santiago in the Maipo Basin, based on characteristics of the shell and for which only a single drawing is available from the original description (Biese 1944) (Fig. 1).Along with similar forms, Biese related L. santiagensis to the Littoridina hatcheri Pilsbry, 1911 group, a species subsequently transferred to the genus Heleobia Stimpson, 1865 within the family Cochliopidae (Hershler & Thompson 1992), but now recognized as member of the family Tateidae, closely related to the genus Potamolithus Pilsbry, 1896 (Koch et al. 2015).Shortly after the original publication, Biese (1947) discovered a second allopatric population in a spring tributary of the Yeso River, 55 km south of the type locality, which he assigned to L. santiagensis using the same typological criteria.Subsequent finds of the species have not been made, but it was cited by Stuardo (1961) in one of the most important checklists of Chilean freshwater mollusks of the 20th century.Later, Hershler & Thompson (1992), in their influential work on Cochliopinae, placed all the Chilean species of Littoridina, including L. santiagensis, in the genus Heleobia, but without providing arguments for this translocation.This was perhaps one of the reasons why several authors continued to consider these species as members of Littoridina (e.g., Sielfeld 2001;Valdovinos 1999Valdovinos , 2006Valdovinos , 2008)).Biese, 1944. A. The only drawings of a specimen of the species taken from the original description by Biese (1944).B. Lectotype housed at the Museo Nacional de Historia Natural, Santiago, Chile (MNHNCL) (after Collado et al. 2011).

Fig. 1. "Littoridina" santiagensis
During a study of Chilean freshwater snails performed by the first author in 2011, specimens of L. santiagensis were not found at its type locality but specimens attributable to the species were found in Yeso Spring (Collado et al. 2011).Based on anatomical characters of the male copulatory organ, the authors moved the Chilean species of Littoridina to the genus Heleobia, although in the case of L. santiagensis this was not justified because no males were found among the specimens from that locality.More recently, a molecular phylogenetic analysis based on DNA sequences of the mitochondrial gene cytochrome oxidase subunit I (COI) obtained from snails collected from Estero Dehesa identified these specimens as the New Zealand mudsnail Potamopyrgus antipodarum (Gray, 1843) in the family Tateidae (Collado 2014), which is morphologically similar to L. santiagensis, suggesting either species displacement or an early misidentification of these snails as members of the genus Littoridina.The absence of males in snails examined from Yeso Spring (Collado et al. 2011) also contributed to raise doubts about the identity of this species, since the genera Littoridina and Heleobia contain dioecious species (e.g.Hubendick 1955;Gaillard & de Castellanos 1976;Cazzaniga 1980Cazzaniga , 1982aCazzaniga , 1982b;;Hershler & Thompson 1992;Collado et al. 2011;Ovando & De Francesco 2011;Collado 2012Collado , 2015)).
New collections at the type locality (Estero Dehesa) of L. santiagensis in 2014, 2015 and 2017 have only been successful for P. antipodarum despite intensive sampling effort performed in different microhabitats (pers.obs.).On the other hand, two morphotypes were found in Yeso Spring, a slender form conferred to P. antipodarum and another thicker one provisionally assigned to L. santiagensis, taking into account the previous record by Biese and its morphological similarity regarding the type specimens of this species studied previously (Fig. 1) (Collado et al. 2011).Given the new material, in the present study we performed comparative studies between these snails and evaluated their systematic position based on morphological and molecular data, including other morphologically similar snail populations of uncertain taxonomic status found in central Chile.
DNA extraction and amplification of partial sequences of the COI gene were performed following Collado (2014) using the new pair of primers ATCTGGTTTAGTRGGRACA (forward) and CCTGCYAAAACAGGYAAAG (reverse).The new sequences were deposited in GenBank (MH536524-MH536536).Genetic uncorrected p-distances were estimated in MEGA v.7.0 (Kumar et al. 2016).An initial BLASTn-analysis against the data base of GenBank confirmed that the species in question belonged to Tateidae.We also performed a Bayesian tree-based barcoding approach (Erickson & Driskell 2012) to assess the familial/generic allocation of the species analyzed together with representative species of these families retrieved from GenBank (Hershler et al. 1999;Haase 2005;Kroll et al. 2012;Hamada et al. 2013;Wilke et al. 2013;Koch et al. 2015;Collado et al. 2016a;de Lucía & Gutiérrez Gregoric 2017;Colgan & da Costa, unpublished data).As COI was highly saturated across both families as shown by the test of Xia et al. (2003), our analysis was not supposed to reflect accurate relationships, but it allowed us to assign the Chilean species to their respective families and genera.We constructed a Bayesian tree using MrBayes ver.3.2.2 (Ronquist et al. 2011) based on the HKY + I + Γ substitution model identified as best fitting by jModelTest ver.2.1.4(Darriba et al. 2012) according to all four criteria implemented in this program.MrBayes was run for 2 Mio generations with a burn-in of 25%.In order not to aggravate the saturation problem by introducing an outgroup, the tree was midpoint rooted in FigTree ver.1.4.3 (http://tree.bio.ed.ac.uk/software/figtree/).For further comparison, we also obtained sequences of two specimens of Potamolithus from Puerto Chico, Llanquihue Lake, Región de Los Lagos (41°19′37.06″S; 72°57′26.64″W), the type locality of P. australis Biese, 1944, the only species of the genus reported from Chile.However, as it has been considered a nomen dubium (López Armengol 1985), we treat these snails as Potamolithus sp. until a formal redescription may be established.

Species delimitation
The classification of specimens based on external shell features was difficult to perform in the field.Under magnification, two snail morphotypes could be recognized in Yeso Spring and Lo Carreño, a slender form morphologically associable to P. antipodarum and a thick-shelled one congruent with the type specimens of L. santiagensis, here transferred to Potamolithus based on anatomical characters and DNA barcode analysis (see below).The thicker-shelled specimens from El Colorado and Viña Casas del Maule were assigned to Potamolithus as well (Fig. 2).
FemaleS.With nuchal node, white lips, tentacles grayish or black (Fig. 3H-I).Some specimens with a white band at base of tentacles where eyes are located.Males were not observed.
Radula.Taenioglossate (formula 3-1-3), with two marginal teeth and a lateral tooth placed on each side of the central tooth (Fig. 3J-K).Central tooth trapezoidal, dorsally concave, ca.20 μm wide; basal tongue U-shaped; with 5-6 lateral cusps on each side of median cusp and 3-4 pairs of basal cusps, first basal cusps arise from tooth face and are larger than those on cutting edge.Median cusp of central tooth

Ecology
Potamolithus santiagensis comb.nov. is a herbivorous-detritivorous species that inhabits small water bodies like springs and small streams.

Distribution
Yeso Spring in Cajón del Maipo, Región Metropolitana, Central Chile (Biese 1947;Collado et al. 2011;present study).This spring is a small tributary of the Yeso River, which empties in the Maipo River.The species also inhabits small irrigation canals in the Región de O'Higgins (Lo Carreño) and Región del Maule (El Colorado).Snail collections made in 2011, 2014, 2015 and 2017 at the type locality Estero Dehesa have not provided specimens of the species.This ecosystem has been invaded by P. antipodarum (Collado 2014) and Physa sp.(unpublished data).In Yeso Spring, P. santiagensis comb.nov.coexists with P. antipodarum and snails of the genus Chilina Gray, 1828.

Molecular analysis
The COI phylogenetic analysis grouped the original sequences and those downloaded from GenBank into two main clades, Tateidae and Cochliopidae, respectively, both supported by posterior probabilities (p.p.) of 1.00 (Fig. 5).Our original sequences clustered in the first clade, either in the genus Potamolithus or Potamopyrgus, both groups inferred with high node support (0.97 and 1.00 p.p., respectively).The sequences of the slender morphotype from El Yeso Spring and Lo Carreño were identified as P. antipodarum, and those of the thicker one as representatives of Potamolithus (Fig. 5).The snails from El Colorado, Viña Casas del Maule, Puerto Chico, "Heleobia" sp.from Uspallata and "Heleobia" hatcheri from Aguas Negras in Argentina retrieved from GenBank were also recovered among Potamolithus species.The COI genetic distances among Tateidae species/populations from South America ranged between 0.2% and 17.0% (Table 3).The genetic distances between P. santiagensis comb.nov.and snails from Puerto Chico were estimated at 6.7-7.2%.In the Bayesian tree, they were placed in different subclades.The population from Viña Casas del Maule grouped with "Heleobia" sp.from Uspallata (1.00 p.p.); they differed genetically from P. santiagensis comb.nov.by about 5.0%.On the other hand, the sequence divergence between the population from El Colorado and P. santiagensis comb.nov.from Yeso Spring and Lo Carreño was low (0.0-0.6%).
The presence of a nuchal node, also observed in females of several species of this genus (Davis & Pons da Silva 1984;López Armengol 1996;Pons da Silva & Veitenheimer-Mendes 2004a), absent in Heleobia, is a morphological character that also supports this change, although it is not a diagnostic character of the genus (Núñez 2017).The three or four pairs of basal cusps on the rachidian tooth characteristic of the tateid radula also distinguish this group from Heleobia species ( The Bayesian analysis revealed that part of the material from Lo Carreño and the population from El Colorado grouped with sequences of P. santiagensis comb.nov.from El Yeso Spring, so we assigned these snails to this species.In Lo Carreño we also identified the New Zealand mudsnail P. antipodarum, extending its invasive range to a new hydrological basin further south of the previously known distribution in the country (Collado 2014(Collado , 2016)).Morphometric data subjected to PCA also separated Potamolithus populations and the New Zealand mudsnail.Taking into account that the presence of the New Zealand mudsnail has also been reported at several type localities of endemic species of Heleobia in the Región de Coquimbo (Biese 1944(Biese , 1947;;Collado 2014), where they were previously confused (Collado et al. 2011), these cryptic snails could potentially also be discriminated using this simple methodology.Single shell dimensions may or may not be operational, though, since a number of comparisons were statistically undifferentiable.The Bayesian analysis also inferred the sequences of "Heleobia" sp.Uspallata and "Heleobia" hatcheri from Argentina as members of the genus Potamolithus, representing other cases of familial and generic misidentification.
In the Southern Cone of South America, the close phylogenetic relationships found between samples of "Heleobia" hatcheri from the Central-West region of Argentina and Potamolithus sp. from Llanquihue Lake in the Chilean Patagonia suggests that dispersal across the Andes may have played a role in the distribution of these species.The same may be inferred regarding the Chilean Potamolithus populations from Viña Casas del Maule and "Heleobia" sp.from Uspallata, Argentina, and the lowland populations of P. santiagensis comb.nov., which exhibit disjunct distributions in Chile.
The New Zealand mudsnail has been considered as a highly invasive species worldwide (Son 2008;Alonso & Castro-Díez 2012;Butkus et al. 2012), potentially able to seriously affect the native snail fauna in invaded aquatic ecosystems (Richards 2002).After this species arrived in the Snake River drainage in North America in 1987 (Bowler 1991), five native species were recorded as either "threatened" or "endangered", in part due to the introduction of the species (Richards 2002).It is possible that something similar has happened in Estero Dehesa considering the failure to find specimens of P. santiagensis  comb.nov.and the current occurrence of New Zealand mudsnails at this locality.Although without being alarmist, perhaps a "chronicle of a death foretold" could be written for the El Yeso Spring population based on the noticeable abundance of New Zealand mudsnails in this habitat compared to P. santiagensis comb.nov.In this context, the finding of the new populations of this species in El Colorado, where the New Zealand mudsnail has not been introduced, acquires real significance giving a light of hope for the preservation of the species.

Fig. 5 .
Fig. 5. Bayesian tree based on COI sequences.Numbers at nodes indicate posterior probability values (only given if ≥ 0.95).Names in bold refer to new sequences reported in this paper.Numbers following taxa refer to GenBank sequences.

Table 1 .
Average shell dimensions (± standard deviation) of six linear variables used in the comparative study of Potamolithus populations and Potamopyrgus antipodarum (Gray, 1843) from central Chile.

Table 2 .
P-values of Kruskal-Wallis (in column Shell variable) and pairwise post-hoc Mann-Whitney U-tests of six linear variables.

Table 3 .
Percent of mean sequence variation (based on p-distance) of COI data among South American species/populations belonging to the Tateidae.