Lymnaeid snail material

The snail specimens studied (n = 8572 specimens) were collected in the field along a ten-year period, from 20 lymnaeid populations present in geographical areas with human and/or animal fascioliasis endemicity of eight countries (Figure 1):

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Maps of the Old and New Worlds showing localities where Lymnaea schirazensis was collected: A) Old World: 1 = Taleb Abad river, Bandar Anzali, Gilan province, Iran; 2 = Medicine Faculty, Rasht, Gilan province, Iran; 3 = El Kazza, Hosh Esa district, Behera governorate, Egypt; 4 = Tiba, Delengate district, Behera governorate, Egypt; 5 = Boulin El Aly, Kafr El Dawar district, Behera governorate, Egypt; 6 = Albufera of Valencia, Valencia province, Spain; 7 = Nules-Moncofar, Castellon province, Spain; B) New World: 8 = Constanza, Departamento de La Vega, the Dominican Republic; 9 = Río Grande, Constanza, Departamento de La Vega, the Dominican Republic; 10 = Los Molinos, subcuenca Nexapa, Atlixco, Puebla, Mexico; 11 = Escuela A. Obrego, La Trinidad Tepango, Atlixco, Puebla, Mexico; 12 = Xalpatlaco, Atlixco, Puebla, Mexico; 13 = Jiutepec, Morelos, Huauchinango, Mexico; 14 = Laguna de Fe y Alegria, El Valle, Estado de Merida, Venezuela; 15 = Hotel Valle Grande, El Valle, Estado de Merida, Venezuela; 16 = Guarandauco, Chillogallo, Ecuador; 17 = La Buena Esperanza, Cayambe, Ecuador; 18 = Machachi, Santo Domingo, Ecuador; 19 = Baños del Inca, Cajamarca, Peru; 20 = Rio Lurin, Lima, Peru.

For a higher resolution situation of each locality, see respective coordenates in text.

https://doi.org/10.1371/journal.pone.0024567.g001

1. - Iran: 1) Taleb Abad river, Bandar Anzali, Gilan province (37°27′46″ N; 49°37′07″ E; −23 m below sea level-b.s.l.): 243 living specimens collected and analysed to assess potential natural infection by Fasciola through cercarial shedding verification; 2) garden of the Medicine Faculty, Rasht, Gilan province (37°11′39″ N; 49°38′04″ E; 26 m above sea level-a.s.l.): 81.
2. - Egypt: 3) El Kazza, Hosh Esa district, Behera governorate (30°50′54″ N; 30°16′16″ E; 1 m a.s.l.): 1956; 4) Tiba, Delengate district, Behera governorate (30°50′34″ N; 30°29′11″ E; 4 m a.s.l.): 887; 5) Boulin El Aly, Kafr El Dawar district, Behera governorate (31°10′42″ N; 30°10′54″ E; 0 m a.s.l.): 369.
3. - Spain: 6) Albufera of Valencia, Valencia province (39°17′42′′N; 0°20′41′′ W; −1 m b.s.l.): 1291; 7) Nules-Moncofar, Castellon province (39°50′01″ N; 0°06′28″ W; 34 m a.s.l.): 2077.
4. - the Dominican Republic: 8) Constanza, Departamento de La Vega (18°54′25′′ N; 70°44 25′′ W; 1,184 m a.s.l): 265; 9) Río Grande, Constanza, Departamento de La Vega (18°52′37′′ N; 70°43′38′′ W; 1,182 m a.s.l.): 32.
5. - Mexico: 10) Los Molinos, subcuenca Nexapa, Atlixco, Puebla (18°56′57″N; 98°23′25″W; 1,952 m a.s.l.): 149; 11) Escuela A. Obrego, La Trinidad Tepango, Atlixco, Puebla (18°51′46″ N; 98°26′33″ W; 1,774 m a.s.l.): 213; 12) Xalpatlaco, Atlixco, Puebla (18°55′58″ N; 98°26′22″ W; 1,903 m a.s.l.): 56; 13) Jiutepec, Morelos, Huauchinango (18°53′01″ N; 99°11′12″ W; 1,377 m a.s.l.): 31.
6. - Venezuela: 14) Laguna de Fe y Alegría, El Valle, Estado de Merida (8°37′29″ N; 71°08′30″ W; 1,837 m a.s.l.): 38; 15) Hotel Valle Grande, El Valle, Estado de Merida (8°40′28″ N; 71°06′03″ W; 2,174 m a.s.l.): 114.
7. - Ecuador: 16) Guarandauco, Chillogallo (0°17′48″ S; 78°38′55″ W; 3,158 m a.s.l.): 75; 17) La Buena Esperanza, Cayambe (0°03′39″ N; 78°08′25″ W; 2,821 m a.s.l.): 283; 18) Machachi, Santo Domingo (0°27′23″ S; 78°33′43″ W; 2,810 m a.s.l.): 317.
8. - Peru: 19) Baños del Inca, Cajamarca (07°12′54″ N; 78°26′08″ W; 2,611 m a.s.l.): 72; 20) Rio Lurin, Lima (12°12′28″ S; 76°52′00″ W; 154 m a.s.l.): 23.

Lymnaeid materials of the species G. truncatula from Albufera of Valencia, Spain, and Qued Tiout, Essaouira, Marrakesh, Morocco were used for comparative analyses.

Snail laboratory cultures

Lymnaeids were transported under isothermal conditions to the laboratory of Valencia. The possible natural infection by fasciolids was always individually verified prior to the launch of laboratory cultures. This was performed by keeping each lymnaeid specimen isolated in a Petri dish containing a small amount of natural water. After 24 h, the presence or absence of motionless metacercarial cysts or moving cercariae was verified in each Petri dish. A few non-infected specimens from each population were used for species classification of each population by means of DNA sequencing processes.

Afterwards, non-infected lymnaeids were arranged in standard breeding boxers containing 2000 ml fresh water according to lymnaeid species classification, to assure pure specific cultures. Finally, snails were adapted to and maintained under experimentally controlled conditions of 20° C, 90% relative humidity and a 12 h/12 h light/darkness photoperiod in precision climatic chambers (Heraeus-Vötsch VB-0714 and HPS-500). The water was changed weekly and lettuce added ad libitum.

Voucher specimens and taxonomic nomenclature aspects

Voucher specimens of L. schirazensis, both experimentally raised and molecularly classified as well as from molecularly classified populations collected in the field, have been deposited in the Department of Parasitology, Faculty of Pharmacy, University of Valencia, Spain, and the Staatliche Naturhistorische Sammlungen Dresden, Museum für Tierkunde, Dresden, Germany. Other field-collected specimens are kept at Laboratorio de Biología de Parásitos y Vectores, Escuela de Biología, Benemérita Universidad Autónoma de Puebla, Puebla, México.

The list of species (and subspecies) of molluscs and trematode parasites included in this study, according to the zoological nomenclature, is included in Table S1. The generic name Lymnaea is used here throughout in the broad taxonomic meaning (sensu lato = s. l.).

Molecular techniques

DNA haplotype nomenclature

The codes for the sequences obtained follow the standard nomenclature previously proposed for lymnaeid snails [12], [16], [39]. Note that haplotype codes are only definitive in the case of complete sequences. When dealing with fragments or incomplete sequences, haplotype codes are provisional.

Software programs

Phylogenetic inference

Phylogenies were inferred from DNA sequences using maximum-likelihood (ML) estimates with PAUP [54]. ML parameters such as model, base frequencies, transition/transversion ratio (ts/tv), the shape parameter for the gamma distribution, and the proportion of invariant sites, were optimized using the hierarchical likelihood ratio test (hLRT) and the Akaike information criterion (AIC) [61,62), implemented in Modeltest 3.7 [63]. Starting branch lengths were obtained using the least-squares method with ML distances.

To provide an assessment of the reliability of the nodes in the ML tree, three methods were used. First, a bootstrap analysis using 1000 replicates was made with heuristic search in PAUP. Second, a distance-based phylogeny using the neighbour-joining (NJ) algorithm [64] with ML pairwise distances was obtained. Statistical support for the nodes was evaluated with 1,000 bootstrap replicates, with and without removal of gapped positions. Third, a Bayesian phylogeny reconstruction procedure was applied to obtain posterior probabilities (BPP) for the nodes in the ML tree, by using the same evolutionary model as above, implemented in MrBayes 3.1 [65] with four chains for 1,000,000 generations and trees were sampled every 100 generations. The first 1,000 trees sampled were ruled out (“burn-in”), and clade posterior probabilities (PP) were computed from the remaining trees. Alternative methods of phylogenetic reconstruction allowing for an evaluation of the support for each node were also applied. A distance-based phylogeny using the NJ algorithm with LogDet distances was obtained. Statistical support for the nodes was evaluated with 1,000 bootstrap replicates.

Due to different limitations recently shown by mtDNA markers for interspecific sequence analyses in invertebrates [46], [66], [67], phylogenetic reconstruction by combined sequences data sets was made from ribosomal and mitochondrial markers separately. Combined sets analysed were: a) ITS-1 and ITS-2, considered as the markers to fit best for the analysis of relationships between species, as has already been verified in Lymnaeidae [16]; b) 18S, ITS-1 and ITS-2, to increase the support for evolutionarily older divergence nodes; c) 16S and cox1, to evaluate mtDNA genome information; and d) cox1 gene, by using both the three and only the first two codon positions to assess potential saturation.

Phylogenetic analyses were performed using reference sequences of lymnaeid DNA (see species and Acc. Nos. in list noted above in chapter of sequence comparisons), after adding the following sequences stored at databases: ITS-2: Radix auricularia H1 (AJ319628) and R. balthica H1 (AJ319633); ITS-1: R. auricularia HA (JF922878) and R. balthica HA (JF922879); 16S: L. (L.) stagnalis (AF485659), L. (S.) palustris (U82082), G. truncatula (HQ283236), L. viatrix (HQ283239), L. humilis HA and HB (FN182195, FN182196), L. cousini (HQ283237), L. meridensis HA, P. columella (U82073) and R. auricularia (AF485659). cox1: L. (L.) stagnalis (EU818795), L. (S.) palustris (EU818801), R. auricularia (EU818800). The intergenic region sequence (AY030361) [68] including both ITSs of a planorbid species, Biomphalaria pfeifferi, was used as outgroup. For the combined sets using 18S and 16S, B. alexandrina 18S (BAU65225), ITSs (AY030371) and 16S (AY030204) [68] were used for outgroups, similarly as B. alexandrina cox1 (AF199110).

Phenotypic study

Only specimens from pure laboratory cultures experimentally maintained in climatic chambers and with previous molecularly assessed lymnaeid species classification were used for snail description and intraspecific variability studies.

Shells and egg clusters of lymnaeids were measured, according to traditional malacological methods [19], using a computerised image-analysis system (CIAS) [69]. This system is based on a DXC-930P colour video camera (Sony, Tokyo) fitted to a stereomicroscope, and connected to computer running image analysis software (ImageProH Plus 4.5; Media Cybernetics Inc., Silver Spring, MD).

For anatomical studies, adult lymnaeids were collected in the field and allowed to relax overnight in water containing menthol. They were then immersed in hot water (70° C) for 40 s before transfer to water at room temperature. The soft parts were drawn from the shells with forceps applied to the cephalopedal mass, and fixed in slightly modified Railliet–Henry's fluid (930 ml distilled water, 6 g NaCl, 50 ml 40% formalin, and 20 ml glacial acetic acid). The fixed snails were then dissected under a stereomicroscope, so that drawings of the reproductive system could be made using a camera lucida.

Egg clusters were obtained from living lymnaeids experimentally maintained in climatic chambers. Living egg clusters were measured with CIAS. For egg clusters, cluster roundness (CR = CP²/4πCA) measurements were used to quantify the cluster shape. It is a measure of how circular an object is (the expected perimeter of a circular object divided by the actual perimeter). A circular object will have a roundness of 1.0, while more irregular objects will have larger values [70].

Experiments for selfing verification and characterisation

To verify the capacity of snails to give rise to offspring by selfing ( = autofecundation), egg clusters laid inside the breeding containers, in which pure laboratory cultures were kept within the climatic chambers, were isolated in Petri dishes. Immediately after egg hatching, each newborn snail specimen was isolated in a small Petri dish provided with water and microalgae (Oscilatoria formosa) as food. The growth of the snail and its possible egg cluster laying were thereafter followed on a daily basis. Small portions of fresh lettuce were added to complement the food diet when lymnaeid size was sufficient. This follow-up was continued not only until the first egg cluster appeared in the Petri dish, but expanded for up to several weeks until snail death. The purpose was to obtain numerous egg clusters from each isolated snail and to analyse their laying capacity. The following characteristics were assessed: snail life span ( = days elapsed from day of hatching to day of death); prelaying period ( = days elapsed from day of hatching to day when first cluster was laid); laying period ( = days of sexual activity elapsed from first to last day when clusters were laid by a snail specimen, inclusive); postlaying period ( = days elapsed between last laying day and death); total laying capacity ( = number of clusters/life span in days); laying rate in the sexually active period ( = number of clusters/laying period in days). Shape and size of the living clusters were measured with CIAS, and egg numbers per cluster were noted to verify a potential correlation with the size and age of the respective snail individual at the day of laying.

Liver fluke experimental infection assays

Isolates of F. hepatica from Poland and Peru, and F. gigantica from Egypt and Vietnam, maintained in the laboratory according to previously described methods [35], were used for experimental infection assays of the laboratory-reared lymnaeids.

Liver fluke eggs in fresh water were maintained in complete darkness at 20° C to start the embryogenic process. Embryogenesis was followed at intervals of four days until fully embryonated containing a developed miracidium. Developed miracidia were forced to hatch by putting fully embryonated eggs under light and used for the experimental infection of snails. Snails of the geographical strains from El Kazza and Tiba in Egypt and from Albufera of Valencia in Spain were used for experiments. These lymnaeid strains were selected due to their higher adult survival rates under experimental conditions. Only laboratory-borne specimens were used. Snails of different size and age within the length range of 3.5–7.5 mm were used to assess infection susceptibility.

Lymnaeids were mono- or trimiracidially infected by the aforementioned F. hepatica and F. gigantica isolates, by exposing each snail to one or three miracidia for four hours in a small Petri dish containing 2 ml of fresh water. During this short period, snail specimens were forced to stay inside water and the disappearance of the miracidium was taken as verification of its successful penetration into the snail. Therefafter, snails were returned to the same standard conditions in the climatic chamber (2000 ml containers, 20° C, 90% r.h., 12 h/12 h light/darkness, dry lettuce ad libitum) until day 30 post-infection (dpi). In that day, they were once again isolated in Petri dishes to permit daily monitoring of potential cercarial shedding by individual snails. Lettuce was provided ad libitum to each snail in a Petri dish during this monitoring period until death of the snail.

DNA sequences

Sequence characteristics are noted in Table 1. Data includes length and GC/AT content, obtained for each one of the five DNA markers analysed in the populations of L. schirazensis from the twenty localities of eight countries (Figure 1) with their corresponding haplotype codes and GenBank Accession Nos.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Nuclear ribosomal DNA and mitochondrial DNA marker sequences obtained from populations of Lymnaea schirazensis and respective GenBank accession numbers.

https://doi.org/10.1371/journal.pone.0024567.t001

18S rRNA gene

All of the 20 populations of L. schirazensis presented the same 18S sequence (Table 1). When comparing this 18S sequence with that of G. trucatula, a total of 19 variable positions appeared. They included 6 ts, 2 tv and 11 insertions/deletions (indels) in a 1853 bp-long pairwise alingment, most differences being located in the variable Helix E-10 of the V2 region according to the secondary structure (Figure 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. Nucleotide differences in a total of 64 variable positions found in the complete 18S rDNA sequence of the lymnaeid species compared and their location in the secondary structure.

Helix, Position and Difference number = numbers to be read in vertical. Position = numbers refer to positions obtained in the alignment made with MEGA 4.0. Identical = .; Indel = −. Shaded area corresponds to Helix E10-1 of the variable area V2 where differences in the 18S rRNA gene of Lymnaeidae are concentrated [33]. GenBank Accession Nos. = Z73980−Z73985 [33]; Z83831 [34]; AM4122222 [15]; FN182190 [45]; FN598151- FN598152 [55]; L. schirazensis from present paper. Sequence correspondences: * L. cubensis, L. viatrix, L. humilis and L. cousini without definitive genus ascription; 18S identical in L. viatrix and L. neotropica [15]; 18S identical in L. cousini and L. meridensis [55].

https://doi.org/10.1371/journal.pone.0024567.g002

A multiple 18S sequence alignment was 1867 bp long. It included: several Galba/Fossaria vector species such as L. cubensis, L. viatrix (the sequence being identical to that of L. neotropica), G. truncatula, L. humilis, and L. cousini (sequence identical to that of L. meridensis); the peculiar species P. columella; three representatives of stagnicolines (L. (L.) stagnalis, L. (S.) palustris and O. glabra); and two of the Radix group (R. auricularia and R. balthica). This alignment showed a total of 64 variable nucleotide positions (3.43% nucleotide divergence). Thirty-two of these 64 polymorphic sites appeared grouped in the short sequence between positions 206 and 266, which corresponds to the helix E10-1 of the variable area V2 of the secondary structure (Figure 2).

Second internal transcribed spacer rDNA ITS-2

Two different ITS-2 haplotypes were found among the L. schirazensis populations analysed (Table 1). Their pairwise comparison showed only 8 polymorphic sites, corresponding to 8 indels caused by the tetranucleotide microsatellite repeat TGCT, present twice in haplotype 1 between positions 128 and 135 of the alignment but absent in haplotype 2.

The very high number of nucleotide differences detected in the pairwise comparisons of these two ITS-2 sequences with the three ITS-2 haplotypes of G. truncatula available at GenBank is noteworthy. In the different alignments performed, with a bp length varying between 469 to 477, a total of 130–139 differences appeared, including 18–19 ts, 11 tv and 101–109 indels (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Sequence differences detected in pairwise comparisons of ITS-2 and ITS-1 between Lymnaea schirazensis and the most morphologically similar species Galba truncatula.

https://doi.org/10.1371/journal.pone.0024567.t002

When comparing the L. schirazensis ITS-2 sequences with the other species of the Galba/Fossaria group, the pairwise ITS-2 distance matrix obtained with PAUP (only parsimony informative sites considered) showed that the numbers of nucleotide differences were unexpectedly high in all cases. The lowest of these numbers appeared when comparing with G. truncatula, although rather high (64–66). Regarding L. humilis, L. cousini and L. meridensis, it ranged from 99 to 118. It appeared to be similarly high (121–136) with regard to L. cubensis, L. viatrix and L. neotropica. Nucleotide differences were also numerous when L. schirazensis was compared with stagnicolines and Pseudosuccinea (Table S2).

First internal transcribed spacer rDNA ITS-1

Two different ITS-1 haplotypes were also found among the L. schirazensis populations analysed (Table 1). The pairwise comparison of these two sequences showed only 3 polymorphic sites, corresponding to two insertions in positions 296 and 297 (AA or --) and to one mutation in position 443 (A or T) in haplotypes A and B, respectively.

A very high number of nucleotide differences also appeared in the pairwise comparisons with the three ITS-1 haplotypes of G. truncatula. In the different alignments performed, with a bp length varying from 558 to 560, a total of 134–138 differences appeared, including 31–32 ts, 21–23 tv and 81–83 indels (Table 2).

When comparing L. schirazensis ITS-1 sequences with the other species of the Galba/Fossaria group, the pairwise ITS-1 distance matrix obtained with PAUP showed that the numbers of nucleotide differences were also very large. The comparison with G. truncatula furnished the second lowest number (81–84), after that of the comparison with L. meridensis (77–80). Regarding L. humilis and L. cousini, it ranged from 87 to 92, and it appeared to be similarly high (83–92) with regard to L. cubensis, L. viatrix and L. neotropica. The number of nucleotide differences increased considerably when L. schirazensis was compared with stagnicolines (132–143) and Pseudosuccinea (129–130) (Table S3).

16S rRNA gene of the mtDNA

Two different sequences of the 16S fragment were found (Table 1). They differred by only 2 mutations (C/A in position 85 and A/G in position 305 of their pairwise alignment, in HA/HB, respectively).

Nucleotide comparison with other Galba/Fossaria species showed a total of 46 variable positions (10.75%), of which 31 were mutations (7.24%) and 15 were indels (3.50%). This alignment demonstrated that L. schirazensis was different from any other Galba/Fossaria at the level of this mtDNA gene (Figure 3). In pairwise comparisons, minimum differences were 18 mutations (4.2%) when compared with L. humilis, and a maximum of 22 mutations (5.14%) appeared in comparisn with L. cubensis. Nucleotide differences were more numerous when comparing with the stagnicolines S. elodes and S. bonnevillensis. The four species alignment showed a total of 56 variable positions (12.93%), 47 of which were mutations (10.85%) and 9 indels (2.08%) (alignment not shown).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 3. Variable positions showed by the mtDNA 16S sequence fragment in a 433-bp-long alignment including the two haplotypes of Lymnaea schirazensis and other Galba/Fossaria species.

Numbers (to be read in vertical) refer to positions obtained in the alignment made with MEGA 4.0. Identical = . ; Indel = − . Haplotype codes only provisional due to incomplete sequences of the gene. L. cubensis 16S-HB (FN182204), L. humilis 16S-HA (FN182195) and L. humilis 16S-HB (FN182196) [45]; F. bulimoides (AF485657) and F. obrussa (AF485658) [43]; * the sequence ascribed to the stagnicoline C. elodes (EU038305) [56] concerns in fact a Galba/Fossaria species-see analysis in [45].

https://doi.org/10.1371/journal.pone.0024567.g003

mtDNA cytochrome c oxidase subunit I cox1

Four different cox1 sequences were obtained (Table 1). When aligned, a total of 661 positions were conserved and only 11 were variable, comprising 2 parsimony informative and 9 singleton sites (Table 3). When comparing with species of the Galba/Fossaria group and other proximal lymnaeid species available at GenBank, whose cox1 fragment sequences were similar in length to those obtained in the present study, the higher number of nucleotide differences appeared evident in a pairwise cox1 distance matrix (Table S4). A comparison with more distant species such as stagnicolines (S. elodes) and Radix group (Austropeplea tomentosa) was not necessary.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Nucleotide differences found in the sequences of the 672-bp-long mtDNA cox1 gene fragment of the four L. schirazensis haplotypes described.

https://doi.org/10.1371/journal.pone.0024567.t003

The amino-acid sequence of that COX1 gene fragment was 224 aa long. A pairwise comparison showed a 100% identity between the three L. schirazensis haplotypes cox1-a, cox1-b and cox1-c, and only one amino-acid difference (isoleucine/threonine, respectively) in position 96 with L. schirazensis haplotype cox1-d from Baños del Inca, Peru. The provisional haplotypes COX1-I and COX1-II were assigned, respectively. The comparison with the amino acid sequence of other species of the Galba/Fossaria group and other proximal lymnaeid species showed that COX1-I of L. schirazensis was identical to that of other species, such as L. viatrix, L. humilis and L. meridensis. However, L. schirazensis COX1-II appeared to be unique (Figure 4).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 4. COX1 amino acid sequence differences detected in the alignment of the haplotypes of the lymnaeid species studied, together with species of the Galba/Fossaria group and other proximal lymnaeid species available in GenBank.

Only cox1 sequence fragments of a lenght similar to that of sequences obtained in present paper are included. Variable positions = Numbers (to be read in vertical) refer to positions obtained in the alignment made with MEGA 4.0. − = position not sequenced; ? = undetermined amino acid. Haplotype codes only provisional due to incomplete sequences of the gene. * Sequences somewhat shorter and including a few undetermined amino acids; ** sequences somewhat shorter although presumably identical to haplotype cox1a of the same species.

https://doi.org/10.1371/journal.pone.0024567.g004

Phylogenetic analysis

Combined ITS-1 and ITS-2 sets.

This single data-set generated a robust tree, indicating phylogenetic accordance between the two spacers. The ML model best fitting this data-set was HKY85+I, using a ts/tv ratio of 2 (kappa = 3.8979086), base frequencies for A, C, G and T of 0.21534, 0.26862, 0.22661 and 0.28943, respectively, and a proportion of invariable sites = 0.087. To assess the reliability of the nodes in the ML tree (Figure 5 A), a bootstrap analysis using 1,000 replicates was made using branch-swapping algorithm (tree-bisection-reconnection TBR) with heuristic search and the neighbour-joining (NJ) algorithm with the ML pairwise distances in PAUP. Finally, a Bayesian phylogeny reconstruction procedure was applied to obtain posterior probabilities (BPP) for the nodes in the ML tree with MrBayes 3.1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 5. Phylogenetic trees of lymnaeid species studied based on maximum-likelihood (ML) estimates: A) data set of ITS-1 and ITS-2, with the planorbid B. pfeifferi as outgroup (−Ln = 10016.27013); B) same data set of ITS-1 and ITS-2 with B. pfeifferi as outgroup, after adding Radix species (−Ln = 10078.46520).

Scale bar indicates the number of substitutions per sequence position. Support for nodes a/b/c: a: bootstrap with NJ reconstruction using PAUP with ML distance and 1000 replicates; b: bootstrap with ML reconstruction using PAUP with 1000 heuristic replicates; c: Bayesian posterior probability with ML model using MrBayes. See Table S1 for systematic-taxonomic notes.

https://doi.org/10.1371/journal.pone.0024567.g005

In the ML tree obtained (Figure 5 A), P. columella appeared basal to two large groupings, namely stagnicolines (including both Palaearctic and Nearctic species), and the Galba/Fossaria clade which comprised the F. hepatica main vector species. Lymnaea schirazensis appeared included within the latter. However, contrary to what would be expected, L. schirazensis did not cluster together with G. truncatula, of Old World origin, or other morphologically similar lymnaeids such as the New World L. cubensis, L. viatrix and L. neotropica. Surprisingly, it appeared linked to the Nearctic L. humilis and the two Neotropical L. cousini and L. meridensis, within a relatively well supported branch (82/78/88 in NJ/ML/BBP).

The topology obtained with the NJ algorithm using LogDet distances (figure not shown) was identical to that shown by the ML tree (Figure 5 A).

When adding two Radix species (data matrix of 24 taxa with 1,767 characters), the ML model best fitting was HKY85+G+I, using a ts/tv ratio of 1.45 (kappa = 2.872483), base frequencies for A, C, G and T of 0.22764, 0.25354, 0.23176 and 0.28705, respectively, a shape parameter (alpha) = 0.99, and a proportion of invariable sites = 0.13. In the ML tree obtained (Figure 5 B), Radix species clustered independently in a branch basal to all the remaining lymnaeids. Unexpectedly, the location of L. schirazensis changed, now becoming basal to the other Galba/Fossaria members which clustered together although with very low supports. The topology of the NJ LogDet tree was similar (figure not shown), but with L. schirazensis in a subclade with G. truncatula showing low bootstrap (only 59).

Combined 18S, ITS-1 and ITS-2 set.

This three ribosomal marker data set (matrix of 17 taxa and 3,540 characters) also generated a robust tree. The ML model best fitting was HKY85+I, using a ts/tv ratio of 1.48 (kappa = 2.9601982), base frequencies for A, C, G and T of 0.22150, 0.25100, 0.25820 and 0.26930, respectively, and a proportion of invariable sites = 0.44. In the ML tree obtained (Figure 6 A), L. schirazensis appeared clustering together with G. truncatula (79/81/- in NJ/ML/BPP), and the branch including L. humilis, L. cousini and L. meridensis appeared as the sister group inside the Galba/Fossaria clade. The topology obtained with the NJ algorithm using LogDet distances (figure not shown) was identical.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 6. Phylogenetic trees of lymnaeid species studied based on maximum-likelihood (ML) estimates: A) data set of 18S, ITS-1 and ITS-2, with B. alexandrina as outgroup (−Ln = 13171.38533); B) data set of 16S and cox1, with B. alexandrina as outgroup (−Ln = 5282.96177).

Scale bar indicates the number of substitutions per sequence position. Support for nodes a/b/c: a: bootstrap with NJ reconstruction using PAUP with ML distance and 1000 replicates; b: bootstrap with ML reconstruction using PAUP with 1000 heuristic replicates; c: Bayesian posterior probability with ML model using MrBayes. See Table S1 for systematic-taxonomic notes.

https://doi.org/10.1371/journal.pone.0024567.g006

Diagnostic description of Lymnaea (s. l.) schirazensis Küster, 1862

Shell.

The shell was brownish to reddish light brown, thin-walled, elongated conical, usually with four regular convex whorls and up to 5.5 whorls in the longer specimens (Figure 7). The whorls were somewhat inflated, slightly shouldered, with silky and longitudinally striated surface and separated by a deep, well-marked suture, increasing rather slowly in diameter. The columella was straight, unfolded, and the umbilicus open. The last or body whorl was almost ¾ times as high as the shell height, presenting a slight twisting trend along its basal part visible in the biggest shells when viewed dorsally, and which was due to the enlargement of the basal lip of the peristome (in the way of Pseudosuccinea columella). The spire was pointed. The aperture was elongatedly oval, slightly oblique, mid-sized and wider at the base. The thin peristome was patent throughout, the umbilicus partially covered by a widened columellar lip. The shell showed a tendency to be approximately one and a half to two times as long as it is wide, and its aperture tends to be slightly less than half as long as the shell. Measurements and calculated ratios of natural and experimental populations were noted (see Table 4 and Table S5, respectively).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 7. Shells of Lymnaea schirazensis in ventral, dorsal and from-below views, showing intraspecific variability: A,B) specimen (7.20 mm high) from Tiba, Delengate district, Behera governorate, Egypt; C) specimen (7.00 mm) from Albufera of Valencia, Valencia province, Spain; D–F) specimen (7.10 mm) from Albufera of Valencia, Valencia province, Spain; G,H) specimen (7.80 mm) from Nules-Moncofar, Castellon province, Spain; I,J) specimen (7.84 mm) from Laguna de Fe y Alegria, El Valle, Estado de Merida, Venezuela. Scale bars: A−E,G−J = 4 mm; F = 2 mm.

https://doi.org/10.1371/journal.pone.0024567.g007

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Lymnaeid shell measurement comparison between different natural populations of Lymnaea schirazensis and Galba truncatula from different countries.

https://doi.org/10.1371/journal.pone.0024567.t004

External morphological characteristics.

The cephalopedeal mass was pale greyish. The eyes were black, relatively big in size (Figure 8 A–C). Tentacles were elongate, slender, pyramidal, with narrow base (Figure 8 A–E). The mantle roof was dark, from dark brown to blackish throughout, with small unpigmented white-greyish round spots including several tiny circles at the beginning of the border of the pulmonary region and a few scattered further away in between the initial large round spots (Figure 8 F–I). The border of the mantle was light grey. The black pigmentation of the hypopeplear region of the mantle roof gave a dark appearance to the shell of living specimens by transparency (Figure 8 D,E). This dark appearance did not depend on the characteristics of the natural habitat, as it was maintained across the different laboratory-reared snail generations (quite the opposite of what happens with several darkish populations of other lymnaeid species under experimental conditions).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 8. External aspect of Lymnaea schirazensis: A–E) living specimens showing (i) large, round, black eyes, (ii) long, slender tentacles and (iii) dark shell (A = lighted from down; B–D = lighted from above; F = epi− and infralighted simultaneously); F–I) dark brown to blackish mantle roof of specimens from Spain (F, G) and Mexico (H, I) showing small unpigmented white-greyish round spots, including several tiny circles (artificially remarked in white with computer effects in I) at the beginning of the border of the pulmonary region (I = yellow rectangle in H).

https://doi.org/10.1371/journal.pone.0024567.g008

Anatomy.

Morphoanatomical features were studied (see Figure 9 and Figure S1). The renal tube extended straight from the pericardial region toward the mantle collar, diagonally across the roof of the palial cavity and parallel to the pulmonary vein and renal vein. It was a straight tube, lined with a folded epithelium, white to white-yellowish, that tapered, smoothed and became transparent distalward until the level of the septum between pulmonary and hypopeplear cavities. In that place, it turned with an almost right-angle to form a short ureter (which lacked the two distinct flexures present in other close species) and finally opened behind the pneumostome. A very slender muscle thread appeared lengthwise on the ventral surface of the renal tube. A layer of very thin transverse parallel muscle fibers gave the inner side of the pulmonary wall a densely striated appearance (Figure 9 A,B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 9. Aspects of soft part anatomy of Lymnaea schirazensis: A, B) renal tube and ureter in renal region extending between pericardium and mantle collar; C) carrefour in detail, with arrows indicating ducts to albumen gland (ag), spermiduct (spd) and nidamental gland (ng); D) oviducal crown turned to show detail of the region of oviduct pouch; E, F) reproductive system in two ventral views; G) female complex in dorsal view; H) prostate section showing absence of internal folds. Scale bars: A = 1.5 mm; B = 1.2 mm; C = 0.5 mm; D = 0.9 mm; E = 0.8 mm; F = 1 mm; G = 0.6 mm; H = 250 µm (drawings R. Rojas; plate configuration S. Mas-Coma).

https://doi.org/10.1371/journal.pone.0024567.g009

The ovotestis showed a lobulate surface formed by several acini, with a collecting canal which continued into the ovispermiduct. The latter presented a short and very thin proximal segment emptying into the seminal vesicle, an expanded portion of bosselated surface, which narrowed down into a slender distal segment ending in the carrefour surrounded by the albumen gland, oviduct and spermiduct (Figure 9 C–F).

The voluminous albumen gland covered the carrefour and the origin of a bosselated, transverse tubular oviduct which followed a somewhat convolute course, described a nearly complete circle in contact with the albumen gland, and continued into a striated, oblong nidamental gland (Figure 9 C,D and Figure S1 A,B). The initial part of the nidamental gland increased in width to keep the diameter throughout its length or even slightly enlarged in its distal part, to subsequently narrow to give rise to a wrinkled-walled uterus followed by a short vagina. The latter shows a first conical narrowing part and a final short tubular part (without any bulbous swelling or sphincter-like thickening present in other close species) (Figure 9 E–G and Figure S1 A,B), opening into the female genital pore with a slightly thickened lip. All parts of the female reproductive system proximal to the vagina itself showed a pronounced width.

The yellowish spermatheca (Figure S1 A,B) had an oval body and gave rise to a thin, relatively long spermathecal duct which emerged laterally from the spermatheca instead of from its terminal extremity. The almost uniformly thin, smooth-walled spermathecal duct extended diagonally between the nidamental gland and the prostate until joining the final part of the uterus and the beginning of the vagina. The spermathecal duct was not dilated at its distal end (Figure 9 E–G and Figure S1 B).

Emerging from the carrefour, the spermiduct showed an initial slightly dilated part with a granular outer surface slendering into a uniformly thin, straight duct which ran on the ventral surface of the nidamental gland. The distal portion of the spermiduct widened to form a prostate of granular surface, relatively small size, narrower than the nidamental gland. The small, oblong, light grey prostate gland varied in shape from inverted-pear-like to elongated-potato-like, without ventral lengthwise fissure (Figure 9 E,F and Figure S1 C). In cross section, the prostate showed a simple, slit-like lumen (lacking internal folds known in other lymnaeids) (Figure 9 H and Figure S1 D). The prostate gave distally rise to a smooth-walled vas deferens which ran shortly in contact with the vagina and extended until looping caudalward. After this loop, the vas deferens became a uniformly thin, long duct to finally merge into a curved, penial sheath or phallotheca (Figure 9 E,F and Figure S1 E,F).

The penis sheath was regularly cylindrical, with a very slightly thicker proximal part (which showed no patent circlet of minute knobs or ring of papillae described in close species). The penis sheath was about a little bit more than half as long than the praeputium. The maximum width of the praeputium was around two or even three times thicker than the penis sheath. The shape of the praeputium was usually elongate-conical, gradually narrowing to terminate in the male genital pore, although it also appeared cylindrical thoughout most of its length and only tapering distally (Figure 9 E,F and Figure S1 E,F). This conical shape of the praeputium appeared to be an exception among the typical conical praeputium shape in Lymnaeidae. The length of the praeputium measured 0.73–1.96 mm (mean 1.13 mm) while that of the penis sheath was 0.36–0.94 mm (mean 0.72 mm), with a praeputium length/penis sheath length ratio of 1.20–2.23 (mean 1.60).

The radula was characterised by a tricuspid central tooth, presenting the large central cusp, a mid-sized right cusp and a small left cusp. The first lateral teeth were usually bicuspid, with two large cusps. In Mexico and also Iran there was a faint tendency to display a small associated denticle appearing at the inner border of the endocone, thus appearing tricuspid. The subsequent four to six lateral teeth were invariably bicuspid, the intermediate teeth were tricuspid, and the marginal ones multicuspidate.

Egg clusters.

Shape variability of egg clusters and eggs was studied (Figure 10), and measurements and calculated ratios were noted (Table 5). Egg clusters were saccular, transparent, with a very thin outer membranous wall. The shape of these egg clusters showed an evident trend from kidney- to banana-like; the more curved and relatively more elongated and narrow the more numerous were the eggs inside. Small clusters containing only very few eggs (2–5) appeared rounded to oval, but the elongate-curved shape became soon apparent when the egg number increased. Eggs were from spherical to slightly ovoid, with a uniformly thick outer shell, usually appearing well separated from each other.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 10. Gradual evolution of egg cluster lays in experimentally raised Lymnaea schirazensis and Galba truncatula: A–O) L. schirazensis: note trend to kidney shape in early lays (E–K) and final trend to banana-shape in late trends (L–O); only in very early lays, when shape is still round-oval (A, B) and may sometimes become elongate (C, D), it can be confused with clusters of G. truncatula. P–Z) G. truncatula: note general trend to round-oval shape (P–R, V–X) and occasional variation to elongate shape (S–U, Y, Z).

Materials of L. schirazensis from strains originally collected in Xalpatlaco, Mexico (A, B, F, I, J), Jiutepec, Mexico (C), Albufera of Valencia, Spain (D, E, G, L, O), Tiba, Egypt (H), Escuela Obrego, Mexico (K). Materials of G. truncatula from strains originally collected in Qued Tiout, Marrakesh, Marocco (P–T, Y, Z) and Albufera of Valencia, Spain (U–X). Scale bar = 2 mm.

https://doi.org/10.1371/journal.pone.0024567.g010

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Egg cluster and egg measurement comparison between different experimentally-maintained populations of Lymnaea schirazensis and Galba truncaula from different geographical origins.

https://doi.org/10.1371/journal.pone.0024567.t005

Habitat

The ecological characteristics of L. schirazensis appeared to be very peculiar, mainly due to three aspects: (i) its marked amphibious characteristics, (ii) its frequency in antropophilic habitats, and (iii) the wide range of altitudes at which it is present.

Its pronounced terrestrial behaviour was surprising. In the laboratory, snails of this species appeared crawling on or attached to the lateral walls of the containers outside water. When taken with forceps and forced deep into water, they quickly moved out of it again.

In nature, this species was almost never found in water, nor even close to water edges, sometimes up to 1 m or farther from it (Figure 11 I). Outside water, it was usually found on the soil surface under grass besides slow running large rivers (Figure 11 B) or even under relative high bushes completely away from sunhine (Figure 11 H). The smallest amount of water to keep humidity was sufficient for this species to maintain populations, as in ditches of rural paths and dirt tracks (Figure 11 L,N) or even artificial cement canals in gardens (Figure 11 A) and drainage canalizations around buildings and dwellings (Figure 11 K). Man-made small irrigation canals around plant cultures also appeared to offer appropriate conditions for its development (Figure 11 C,D), as well as large drainage cement canals with almost no vegetation (Figure 11 E). Worth mentioning was its presence on mud and livestock footprints in and around animal farms (Figure 11 M).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 11. Environments of localities where Lymnaea schirazensis was collected: A) Garden of the Medicine Faculty, Rasht, Gilan province, Iran; B) Taleb Abad river, Bandar Anzali, Iran; C) El Kazza, Behera governorate, Egypt; D) Nules-Moncofar, Castellon province, Spain; E) Albufera of Valencia, Valencia province, Spain; F) Constanza, Departamento de La Vega, the Dominican Republic; G) Río Grande, Constanza, Departamento de La Vega, the Dominican Republic; H) Xalpatlaco, Atlixco, Mexico; I) Escuela A. Obrego, La Trinidad Tepango, Atlixco, Puebla, Mexico; J) Laguna de Fe y Alegría, El Valle, Merida, Venezuela; K) Hotel Valle Grande, El Valle, Merida, Venezuela; L) Guarandauco, Chillogallo, Ecuador; M) Machachi, Santo Domingo, Ecuador; N) Baños del Inca, Cajamarca, Peru.

https://doi.org/10.1371/journal.pone.0024567.g011

Besides this link to antropophilic habitats and livestock-inhabited places, it was also found, however, in more silvatic habitats. Examples were specimens under the dense vegetation of the edges of fast running waters in mountain rivers and streams, where populations undoubtedly derived from dragging and floods by the water from upper areas (Figure 11 F,J,G).

Its distribution from lowlands to highlands was also worth mentioning. Thus, it was present below sea level, e.g. −23 m at the river edge near the Taleb Abad river mouth at the Caspian Sea, near Bandar Anzali city, in the northern Gilan province of Iran, or the very low altitude of localities such as El Kazza in the Nile Delta region of Egypt, the Mediterranean coast between the villages of Nules and Moncofar in Castellon province of Spain, and Rio Lurin, near Lima, Peru. On the other extreme, this lymnaeid also inhabited high altitude areas such as the 2,821 m of La Buena Esperanza, Cayambe, the 3,158 m of Guarandauco, Chillogallo, both in Ecuador, or the 2,611 m of Baños del Inca, Cajamarca, Peru. Intermediate altitudes were those of the 1,184 m altitudes around Constanza city in the central mountains of the Dominican Republic, those around 1,770–1,950 m of the localities close to Puebla in Mexico, and the 1,800–2,170 m range along the El Valle valley in Merida, Venezuela.

Interestingly, in none of the aforementioned geographical zones did this snail species appear to be the only lymnaeid present in the area. Its populations may appear mixed or close to populations of other lymnaeids (its smaller size may be a competing handicap with regard to other species). Accompanying lymnaeids were other morphologically and ecologically very similar species of the Galba/Fossaria group inhabiting the same place or water body, or neighbouring localities within the same fascioliasis endemic areas. Thus, Galba truncatula was found in all the areas studied in the Old World (Iran, Egypt, Spain). In the New World, Lymnaea cubensis shared the same areas in the Caribbean (the Dominican Republic) and Mexico, L. humilis in Mexico, and G. truncatula, L. cubensis, L. cousini and L. neotropica in South America (Venezuela, Ecuador, Peru).

Worth mentioning was that specimens of L. schirazensis sometimes appeared so mixed or close to one another with specimens of G. truncatula, that one was convinced to deal with a population of only one species, e.g. in Spain (Figure 11 D,E).

Selfing and egg cluster laying

Egg cluster laying was verified to occur in all of the specimens from different populations maintained individually isolated after hatching (Table 6). Age and size when laying began differed somewhat from one individual to another within a population, less in populations of the New World (Mexico) than in those of the Old World (Egypt, Spain). Despite range overlap, those from Mexico proved to be markedly sexually precocious in age when compared to those from Egypt and Spain. Snail size when laying began did not show a similar correlation, suggesting that development speed and reaching a sufficient size may be more important for sexual activity.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 6. Characteristics of the first and five first egg clusters laid by experimentally-maintained Lymnaea schirazensis specimens of four populations from different geographical origins, followed up from the day of their individual isolation immediately after hatching.

https://doi.org/10.1371/journal.pone.0024567.t006

Comparison of the measurements of the first and first five clusters laid by each individual snail also showed that slight differences may be found according to the populations. Results indicated that there is no gradual increase in cluster size nor in the number of eggs per cluster within the laying of a snail specimen. Thus, sometimes the first egg cluster laid by a snail individual was larger and contained more eggs than those laid afterwards (2nd to fifth clusters) by the same specimen (Table 6). However, somewhat larger clusters including a slightly larger number of eggs/cluster were laid by older and larger snail specimens (up to 20 and 22 eggs/cluster in specimens from Egypt and Spain, respectively).

Snail life span was relatively short, of around 2.5–6.8 months. It includes a sexually active period of cluster laying lasting a total of 39–161 days, sometimes even only very few days before death. However, the final postlaying period may reach slightly more than 1.5 months in given specimens (Table 6 and Table 7; Figure 12). Both life span and laying period appeared to be longer in Old World (Egypt, Spain) specimens than in the New World (Mexico) specimens.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 12. Comparison showing interpopulational differences of mean characteristics of life span and selfing reproduction capacity of isolatelly-maintained Lymnaea schirazensis specimens of four populations from different geographical origins, experimentally followed up from the day of their individual isolation immediately after hatching.

Populations studied: Albufera of Valencia, Valencia, Spain; Tiba, Delengate district, Egypt; Escuela A. Obrego, La Trinidad Tepango, Atlixco, Puebla, Mexico; and Jiutepec, Morelos, Huauchinango, Mexico. n = 10 snail specimens followed per population (for details on intrapopulational variability ranges during the laying period, see Table 6 and Table 7); tlc = mean total laying capacity (number of clusters/life span in days); lrsp = mean laying rate in sexually active period (number of clusters/laying period in days); tnld = mean total non-laying days within laying period.

https://doi.org/10.1371/journal.pone.0024567.g012

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 7. Life span, selfing reproduction capacity and egg-cluster laying characteristics of isolatelly-maintained Lymnaea schirazensis specimens of four populations from different geographical origins, experimentally followed up from the day of their individual isolation immediately after hatching.

https://doi.org/10.1371/journal.pone.0024567.t007

This correlated with the number of clusters laid per snail, which was pronouncedly higher in Old than in New World specimens. Similar differences between Old and New World specimens were observed in the total laying capacity and the laying rate in the sexually active period.

Usually, snails only laid one egg cluster on the days they laid. Two clusters were sometimes laid by a snail specimen in the same day, although this appeared to be very rare in Old World (Egypt, Spain) specimens and occurred almost never (only once) in the New World (Mexico) specimens. Throughout the laying period, the number of days when no cluster was laid was larger in Old World (Egypt, Spain) specimens than in New World (Mexico) specimens. This fact seemed to be mainly related to the longer laying period in Old World specimens, as the differences in the maximum number of non-laying days between two laying days when comparing Old and New World specimens were too small (Table 7; Figure 12).

Natural infection studies and experimental transmission assays

Along the ten-year study period, none of the 8,572 snail specimens from the 20 localities of the 8 countries belonging to this lymnaeid species proved to be naturally infected with fasciolid larval stages. This is the opposite of what happened with other lymnaeid vector species also collected in the same endemic areas. In other species, naturally infected specimens were detected through cercarial shedding even despite the usual very low snail infection prevalences.

Several experimental infection assays including a total of 338 snail specimens from different geographical origins with miracidia from experimentally-reared strains of F. hepatica and F. gigantica, also from different geographical origins, did not allow for successful fluke larval development. None of the infected specimens appeared to become parasitised nor reached cercarial shedding (Table 8). Only in two specimens from Albufera of Valencia, Spain, infected with miracidia of F. hepatica from Poland, could a few rediae be observed by transparency through the snail shell at 23 dpi, although neither shed cercariae thereafter. In a subsequent experiment with the same snail- and fluke-strains for verification, no infection and not even rediae were obtained in spite of the higher number of lymnaeid specimens infected.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 8. Characteristics and results of experimental infection assays of Lymnaea schirazensis from different geographical origins with miracidia from experimentally-reared strains of Fasciola hepatica and F. gigantica also from different geographical origins.

https://doi.org/10.1371/journal.pone.0024567.t008

Results on survivorship of snails after infection showed pronounced differences according to snail strains and fluke species and strains used. There were experiments in which no snail infected by either F. hepatica or F. gigantica was able to survive after 30 dpi, up to other experiments in which snails survived for more than 100 or even 150 dpi, thus suggesting no negative effects. Nonetheless, results suggested a lower survival rate in infections by F. gigantica than by F. hepatica. Furthermore, in F. hepatica infection assays, a faster and higher snail mortality rate appeared when using trimiracidial instead of monomiracidial infection (Table 8).

Genetic distances

The twenty lymnaeid populations found in the eight countries of Asia, Africa, Europe, the Caribbean, North America and South America, show DNA sequences proving that all of them belong to the same species. The fact that this species presents an 18S rRNA sequence only is easily understandable given the slow evolution characteristics of this highly conserved gene [33], [34]. The low number of nucleotide differences at rDNA ITS-2 and ITS-1 level fits within the known intraspecific variability ranges of these two species markers [1], . However, it poses a question mark when considering the very wide geographical distribution (Figure 1), markedly variable habitats (Figure 11), and ecology (e.g., large altitudinal range) this lymnaeid species shows. Even the low number of only two and four haplotypes in the mtDNA 16S and cox1 gene fragments, respectively, including but a very few variable positions in both genes, is also surprising given the faster evolution of these mtDNA genes when compared to nuclear rDNA genes and spacers in invertebrates in general [46] and in lymnaeids in particular [45].

The interspecific comparison of the sequences of the snail species here in question demonstrates that it is a unique lymnaeid. It is genetically well different from G. truncatula and other species of the Galba/Fossaria group, as well as from stagnicolines and P. columella. At 18S rRNA level, L. schirazensis presents peculiar nucleotides in the helix E10−1 of the variable V2 area, which clearly differs from all other species (Figure 2). This fact in a well conserved gene such as 18S is worth mentioning [34]. At ITS-2 and ITS-1 level, the two main markers for species distinction [1], [16], [39], the very high number of nucleotide differences L. schirazensis shows when compared to all other lymnaeids must be highlighted. This is surprising when considering the pronounced morphological similarity of L. schirazensis with other lymnaeid species, mainly of the Galba/Fossaria group.

Concerning the sequences of the 16S and cox1 gene fragments, nucleotide differences between L. schirazensis and all other lymnaeids are also numerous (Figure 3 and Table S4), although less pronounced than those appearing in both ITSs despite the faster evolution of mtDNA. This may be related to the well known saturation of nucleotide positions in these two mitochondrial genes, a phenomenon previously described in Lymnaeidae [45].

The aforementioned surprising genetic distance results shown by the nuclear rDNA and mtDNA markers used can only be explained if the following conclusions are accepted:

1. - the very high interspecific variation suggests an old evolutionary divergence of L. schirazensis from other lymnaeid species of the Galba/Fossaria group as well as from stagnicolines and P. columella;
2. - the morphological similarity of L. schirazensis with other lymnaeid species of the Galba/Fossaria group indicates it to be a cryptic species by evolutionary convergence, probably as a consequence of adaptation to a similar amphibious way of life and similar habitats;
3. - the very low intraspecific variability of L. schirazensis suggests a very recent worldwide spread from only one original strain and geographical source.

Galba truncatula is the lymnaeid species on which the largest number of studies has focused for decades, and one of the snail species with the largest body of literature and longer list of synonyms among molluscs in general. This very wide multidisciplinary body of literature is related to its applied importance as the main vector of F. hepatica throughout many countries [12], [29], in a life cycle which was the first to be elucidated among trematodes [72]. Thus, it is surprising that another widely distributed lymnaeid species, similar in morphology, biology and ecology, has always been confused with and kept masked under G. truncatula until now.

Species ascription and specimen classification confusion

The very long list of species and varieties described long ago and later synonymised with G. truncatula [73] has been reviewed to analyse whether a species description published in the past could fit the characteristics of the lymnaeid species here in question. The older species name, fully fitting our material, appears to be Lymnaea schirazensis, described in 1862–1863 [71] from material collected in the locality of Shiraz, Iran, by von dem Busch who never published the description of his snail material (see Figure S2 and Figure S3). Consequently, the decision has been taken to ascribe the material described herein to the binomium Lymnaea schirazensis Küster, 1863. This species cannot be found in the literature published in recent decades. This ‘disappearance’ was undoubtedly due to its synonymy with G. truncatula, a proposal of Hubendick [73] apparently accepted by other specialists. Lymnaea schirazensis appears to have been given systematic validity and reported separately from G. truncatula only in Russia [74].

An exhaustive analysis of the old literature (see numerous references in [73]) suggests that several old species and variety names also synonymised with G. truncatula probably referred to L. schirazensis instead of to G. truncatula. The following items, in chronological order, are worth listing given the geographical locality the type materials came from:

1. - the four G. truncatula varieties distinguished in Algeria [75] suggest that at least one of them could be the species here in question: major (not that of Moquin-Tandon) from Mostaghanem, minutissima from Alger, Géryville and Bône, submalleata from Djelfa, and lanceolata from Alger;
2. - L. persica from the same type locality of Shiraz, Iran [76] is a synonym already previously proposed [77];
3. - L. hordeum from River Euphrates, Mesopotamia [78];
4. - L. delaunayi from Pasajes, northern Spain [79];
5. - the variety thiesseae from Euboea, Greece [80] is very probably synonymous; this synonymy was already aptly discussed [81];
6. - the two varieties neapolitana from Sebeto, near Naples, Italy [82] and telouetensis from Mount Atlas [83] merit further analysis.

The insufficient original description of L. zrmanjae from Zrmanja river, Dalmatian Croatia [84] may also lead to suspect a synonymy, although erroneously, as the latter indeed belongs to the family Hydrobiidae, as verified by the same author somewhat later (Brusina, 1902 in [85]) (see also Table S1).

These old names proposed in Europe and the Near East should be reviewed, but also more recent ones proposed for similar Galba/Fossaria lymnaeids in Europe, Asia and North America, so that further valuable information about the geographical distribution and worldwide spread of L. schirazensis can be obtained.

The coexistence with other very similar lymnaeid species of the Galba/Fossaria group within the same fascioliasis endemic areas, such as G. truncatula in the Old World, and L. cubensis, L. humilis, L. cousini and L. neotropica additionally to G. truncatula in the New World, suggests that a wider distribution of L. schirazensis may have been masked by lymnaeid specimen misclassifications. All of the aforementioned lymnaeids are vectors of fascioliasis and very easily confused when only classified from external characteristics or by non-expert scientists, mainly when specimens are small or mid-sized [15]. A size larger than the one of L. schirazensis is reached by Galba truncatula [29], L. cubensis [86], L. humilis [87], and L. cousini [88], [89]. Only L. neotropica is of a size similar to that of L. schirazensis [15]. Galba truncatula is present in the Gilan province of Iran [89], in the Nile Delta region [90], [91], along the Mediterranean coast of the Valencian Community, Spain [1], [92], and in Merida State of Venezuela [93]. Lymnaea cubensis is known throughout Hispaniola, including both the Dominican Republic [94] and Haiti [95], Puebla province of Mexico [96] and Merida State of Venezuela [93], [97]. Lymnaea humilis is the species most frequently mentioned to be involved in fascioliasis transmission in Mexico (see review in [45]), sometimes under the names of L. modicella and L. obrussa [28], [98], both synonymised with L. humilis [73], and also reported in Puebla State [96]. Lymnaea cousini has been reported in fascioliasis endemic Andean areas of Ecuador [55], [73], [99], and also in Merida State of Venezuela [88], [93]. However, this Venezuelan material later proved to belong to another species, namely L. meridensis [55]. Lymnaea neotropica has been reported, under the name of L. viatrix [15], in several fascioliasis endemic Andean valleys of Peru, as is the case of Cajamarca [100], [101].

Genotypic and phenotypic differentiation between L. schirazensis and G. truncatula

Henceforth, when DNA sequencing techniques are used, the molecular differentiation of L. schirazensis and G. truncatula can be easily made by comparison with the following sequences already available at GenBank/EMBL:

1. - L. schirazensis: specific classification can be based on the nucleotide sequences of ITS-2 rDNA, ITS-1 rDNA, mtDNA 16S, and mtDNA cox1; for supraspecific classification, the nucleotide sequences of the 18S rDNA and of both ITSs can be used (Table 1); the amino-acid sequence of the mtDNA COX1 protein does not appear to be helpful for any discrimination.
2. - G. truncatula: specific classification can be based on the nucleotide sequences of ITS-2 rDNA [1], [15], [35]; ITS-1 rDNA [15], [35], [39], and mtDNA cox1 [15], [57]; for supraspecific classification, the nucleotide sequences of the 18S rDNA [33] and of both ITSs can be used; the mtDNA COX1 amino-acid sequences [15], [57] do not appear to be useful at any level.

When DNA sequencing techniques are not available, to avoid further confusion between L. schirazensis and G. truncatula given the evident difficulty in distinguishing both species due to the morphological variability of their very similar shell, the following differential phenotypic characteristics ought to be considered:

1. - Shell size: smaller in L. schirazensis (Table 4), reaching a shell length of up to 12.00 mm in G. truncatula [29], whereas it shows a maximum of only 8.06 mm in L. schirazensis; other shell characteristics may be helpful, such as the regularly convex whorls and straight columella in L. schirazensis (Figure 7 A–D) when compared to the stepped whorls and folded columella in G. truncatula (Figure 13 A–F); however, shell variability may easily give rise to confusion when based only on these aspects;
   

   Download:

   • PPT
     PowerPoint slide
   • PNG
     larger image
   • TIFF
     original image

   Figure 13. Galba truncatula: A–G) Shells in ventral and dorsal views of specimens from Albufera of Valencia in Spain (A,B: 9.33 mm high), Sachsen in Germany (C: 6.20 mm), Nules in Spain (D,E: 6.00 mm), and Qued Tiout in Morocco (F,G: 8.29 mm). H) Eyes and tentacles in living specimen. I) Comparison of living L. schirazensis (left) and G. truncatula (right) showing differences in (i) eyes, (ii) tentacles and (iii) mantle roof colour through the shell (by infralighting). J,K) Mantle roof in specimens from Nules, Spain (J) and Iran (K). L) Part of reproductive system in ventral view. M,N) Male terminal organs in specimens from Nules, Spain (M) and Iran (N).

   Scale bars: A−G = 3 mm; H = 1.2 mm; I = 1.5 mm; J,K = 2 mm; L = 2.5 mm; M,N = 1 mm.

   https://doi.org/10.1371/journal.pone.0024567.g013

2. - Tentacles: of different form, being elongate, slender and with a narrow base in L. schirazensis (Figure 8 A–E), whereas they are wider and with a wide base in G. truncatula (Figure 13 G);
3. - Eyes: of different size, being relatively big and markedly larger in L. schirazensis than in G. truncatula (Figure 13 H,I);
4. - Egg clusters: of different shape trend, showing an evident trend from kidney- to banana-like, the more curved, elongated and narrow the more numerous are the eggs inside, in L. schirazensis (Figure 10 A–O); clusters tend to keep a rounded to oval shape even when containing more eggs, in G. truncatula (Figure 10 P–Z);
5. - Egg number per cluster: although no significant differences were found between both species in our study, according to the literature it appears that a higher mean egg number per cluster of 12–15 may be the usual rule in G. truncatula, with exceptions of a lower number depending on the soils of the different areas and sometimes even on the season [29]; this mean number shows a lower level range of around 6–14 in L. schirazensis (Table 5);
6. - Mantle colour: mantle roof from dark brown to blackish throughout, with unpigmented white-greyish round spots, this black pigmentation giving a dark appearance to the shell of living specimens by transparency in L. schirazensis (Figure 8 D–I); in G. truncatula the hypopeplear region of the mantle roof shows larger unpigmented whitish spots giving a pale appearance to the shell of living specimens by transparency (Figure 13 I–K);
7. - Radula: first bilateral teeth bicuspid in most populations, although in given populations a small associated denticle appears thus becoming tricuspid in L. schirazensis, whereas it appears tricuspid in G. truncatula;
8. - Male organs: with different praeputium length/penis sheath length ratio, which is of 1.20–2.23 (mean 1.60) in L. schirazensis (Figure 9 E,F and Figure S1 E,F); in G. truncatula the length of the praeputium is 0.82–2.80 mm (mean 1.41 mm) and that of the penis sheath is 0.22–0.64 mm (mean 0.37 mm), with a praeputium length/penis sheath length ratio of 2.50–5.90 (mean 3.44), according to data obtained by measuring specimens from Öland, (Sweden), Kühren, Ettenheim and Reusten (Germany), and Scharflinge (Austria) (Figure 13 M,N);
9. - Ecology: with different water dependence, showing marked amphibious characteristics with a pronounced terrestrial behaviour in L. schirazensis (see habitat types in Figure 11 A–M), whereas aquatic inside freshwater borders or amphibious but more water-dependent in G. truncatula.

Phylogenetic relationships and supraspecific classification

The phylogenetic analyses performed (Figure 5 A,B and Figure 6 A,B) confirms the ascription of L. schirazensis to the Galba/Fossaria group, as already suggested by the morphoanatomic characteristics of this species. Thus, its transfer to the genus Radix as proposed by some researchers (e.g., http://clade.ansp.org/malacology/collections/search.php?mode=browsetypes&targetfamily=Lymnaeinae) does not appear supported.

Unfortunately, there is great confusion around the generic names proposed for truncatula and proximal species in the Americas (the Central American cubensis, the North American humilis and bulimoides, and the South American viatrix and diaphana, only to mention the most important), all of which share the capacity of transmitting F. hepatica [1]. Besides the genus Galba Schrank, 1803 (type species: truncatula) [1], [14], [20], [34], [102], [103], in which it was included [50], other genera having frequently been used to include the New World Galba/Fossaria species, such as cubensis and viatrix, are: Fossaria Westerlund, 1885 (type species: truncatula) [104]–[107], Nasonia Baker, 1928 (type species: cubensis) (see [73]), the subgenus Fossaria (Bakerilymnaea) Weyrauch, 1964 (type species: cubensis) [104], [106], [108], and the subgenus Lymnaea (Afrogalba) Kruglov et Starobogatov, 1985 (type species: misassigned to cubensis; correct type species is mweruensis Connolly, 1929) [109].

In the many phylogenetic reconstruction assays performed, L. schirazensis appeared in three different positions:

1. - In the same clade with L. humilis, L. cousini and L. meridensis: That situation was the one obtaining higher support values, appearing in a higher number of phylogenetic trees, and supported by both rDNA (ITSs) and mtDNA (16S−cox1) sets. This may support the classification of L. schirazensis within the genus Pseudogalba Baker, 1913 (new name proposed for Simpsonia Baker, 1911; type species: humilis) (see [73]). However, the node branching L. schirazensis together with L. humilis, L. cousini and L. meridensis never appeared with sufficiently high support values.
2. - Basal to all other Galba/Fossaria members: This situation does not appear to fit any of the genera erected within Galba/Fossaria so far.
3. - Clustering together with G. truncatula: This relationship supporting its inclusion in the genus Galba is the one appearing in fewer trees.

Consequently, prudence suggests to keep L. schirazensis within the genus Lymnaea sensu lato for the time being.

Ecological adaptability, migration capacity and geographical distribution

Ecological characteristics of L. schirazensis are peculiar when compared to other Lymnaeidae. The amphibious behaviour differ between lymnaeid groups. Radix and stagnicoline species prefer permanent, larger and deeper water bodies, which they leave only sporadically. On the opposite, Galba/Fossaria species may be found on mud or wet sites very close to water in both temporary and permanent water bodies [1], [5]. This approaches L. schirazensis to Galba/Fossaria. However, no other lymnaeid shows such a preferential terrestrial behaviour.

Its amphibious characteristic and wide habitat range, mainly its presence in habitats with very small amounts of freshwater, suggest a great adaptation capacity. Additionally, its frequency in antropophilic and man-made habitats and its survival capacity outside water suggests that this snail can be passively transported by livestock and/or human activities, as demonstrated in G. truncatula [12], which shows similar habitat characteristics [29].

Selfing in L. schirazensis remembers the autofecundation preference in G. truncatula [29], [110]–[113] and other Galba/Fossaria species (Khoubbane et al., unpublished). Its laying capacity (Table 6 and Table 7) appears to be greater than in G. truncatula [29], with a higher number of clusters/month laid per specimen along the laying period. Both selfing and high laying capacity will undoubtedly facilitate its capacity to colonise neighbouring or distant places and expand geographically. Selfing, high egg laying, short life span, great spreading and colonisation capacity are typical of r-strategist organisms (see review on r/K selection in [114]).

This suggests that L. schirazensis may be far more widespread. Such a wider distribution may not only concern the eight countries noted, but also neighboring countries and others further north or south. However, studies are needed to assess whether its terrestrial life trend, outside water, may restrict its capacity to colonise colder areas, contrary to G. truncatula which inhabits high latitude regions [29], [73]. In Asia, L. schirazensis has been noted to have a distribution covering Iran, Afghanistan, mid-Asia and the Caucasus [74]. Although the geographical proximity suggests that it may be present further away from Iran, the more northern, colder latitudes indicate that a molecular verification of these Asian lymnaeids is indispensable.

Phylogeography and continental spread during the historical postdomestication period

The wide distribution of L. schirazensis in different continents is bewildering. The very low variability in its DNA sequences indicate that such a distribution was reached relatively recently. Molluscs are known to be able to follow long-distance dispersal, transported by migrating birds but mainly by human activities. Certain freshwater snail vectors have been reported even very far away from their area of origin (e.g., different continents) [115], including examples of lymnaeids [12], [35], [36]. A few bird species cross the Atlantic Ocean [116], [117], but the probability of a lymnaeid transport by a bird is very low owing to the low number of bird individuals undertaking this journey every year. Yet, proof of an incredible transequatorial dispersal of snails by birds is there [118].

Evidence suggests that G. truncatula may remain in dried mud stuck to the feet of ruminants, then go into hibernation or estivation, and be able to reactivate once in a new location following contact with water or sufficient humidity [12]. Thus, livestock import/export appears related to the continental spread of G. truncatula. Results from our field and laboratory work suggest that L. schirazensis has spread following the same passive transport way of G. truncatula: (i) similarity of livestock-frequented habitats (usually found in mud poached by livestock hooves), (ii) long survival outside freshwater (allowing for long distance transport), (iii) selfing (facilitating invasion of new areas), and (iv) high and fast egg laying capacity (facilitating establishment in a new area). Hence, a review of human history, commercial routes and livestock import/export between the countries inhabited by L. schirazensis offers a likely way of understanding its recent spread (Figure 14 and Figure 15). This chronological spread shows a clear parallelism with the origin and spread of F. hepatica and also in part with the spread followed by G. truncatula [12].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 14. Old World spread of Lymnaea schirazensis combined haplotype ITS-2 H1-ITS-1 HA-16S HA-cox1 Ha during the livestock postdomestication 10,000-year period from the “Fertile Crescent” region of origin in the Near East, including expansion route into the New World from southern Spain during the early period of colonization about 500–400 years ago (green lines and red dots).

For historical and archeological data supporting this recent spread, see text. Green dot of Shiraz, Iran = type locality of L. schirazensis; black lines = spreading routes into Europe and Africa to cover the geographical distribution suggested by the presumed species and variety synonyms (see text) and “resistant” Galba truncatula populations described in France [29]; brown lines = spreading routes into Asia according to the geographical distribution noted by Kruglov [50].

https://doi.org/10.1371/journal.pone.0024567.g014

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 15. New World spread of the different combined haplotypes of Lymnaea schirazensis from remote Old World source(s) in Europe (only or mainly southern Spain) during the early period of colonization about 500–400 years ago (green lines and red dots), illustrating the dispersal role played by Hispaniola island.

Combined haplotype ITS-2 H2-ITS-1 HB: (i) origin whether in Mexico itself after introduction from Hispaniola (green broken line) or from an unknown European source (yellow ? broken line), and (ii) subsequent introduction into Ecuador and Peru by a sea route from Mexico. Introduction of ITS-2 H1 in northern Peru whether by a maritime route through southern Central America (Panama isthmus) or by a terrestrial route from Venezuela (alternative dark broken lines). For historical and archeological data supporting this recent spread, and further details on mtDNA gene isolated haplotypes (16S HB and cox1 Hb, c and d), see text.

https://doi.org/10.1371/journal.pone.0024567.g015

Implications for fascioliasis