Host-dependent morphology of Isthmiophora melis (Schrank, 1788) Luhe, 1909 (Digenea, Echinostomatinae) – morphological variation vs. molecular stability

Echinostomes are cosmopolitan digenean parasites which infect many different warm-blooded hosts. Their classification is extremely confused; the host spectrum is wide, and morphological similarities often result in misidentification. During our long-term studies on the helminth fauna of rodents and carnivores we have collected 27 collar-spined echinostomes which differ in morphology to an extent that suggests the presence of more than one species. Here, we describe this material, and the extent of host-related variation in this parasite. Specimens of Isthmiophora isolated from four host species (badger, American mink, hedgehog, striped field mouse) were subject to morphological and molecular examination; the data were statistically analysed. Our results show that genetically all the Isthmiophora specimens obtained from all the examined hosts are conspecific and represent I. melis. On the other hand, the individuals isolated from Apodemus agrarius are morphologically distinct and, based on this criterion alone, should be described as a new species. The morphological traits of Isthmiophora melis are much variable and host-dependent; without molecular analysis they would suggest a necessity to describe a new species or even genus. Such a high level of intraspecific variability may be affected by the host’s longevity.


Background
Since molecular techniques became commonly used in taxonomic studies, the list of valid taxa in different groups of organisms has been changing, and in many cases the results of molecular investigations are radically different from those obtained with classical methods. However, while museum collections dating from the pre-molecular period remain the cornerstone of taxonomy, morphology must continue to provide a starting point for molecular studies [1]. Molecular taxonomy has also contributed to revealing the common occurrence of cryptic species in nature, in virtually all major taxa.
Although such species are genetically distinct from each other, they are morphologically very similar [2,3].
On the other hand, free-living organisms and parasites can adjust their life-history strategies and a given genotype may produce a variety of phenotypes under different environmental conditions [4]. Due to their exposure to widely differing environmental conditions (i.e. different host species, host's immune system), parasites often display a phenotypic plasticity which is expressed as differences in body size or fecundity [4]. In the case of Digenea, most species-diagnostic features are the body proportions or the shape and location of internal organs. Phenotypic variation may be induced by differences in the intensity of infection ("crowding effect") and in the host's identity ("host-induced variation") [5]. These phenotypic effects may lead to species-specific variation resulting in misidentification [6].
Echinostomes are cosmopolitan digenean parasites which mainly infect many different warm-blooded hosts [7]. The taxonomy and classification of the echinostomes is highly confused. The wide host spectrum of echinostomes is a result of phylogenetic, physiological, and ecological adjustments between the parasite and the host in a dynamic evolutionary process, where the main factor influencing the host specificity is the host's behaviour, particularly the feeding habits of vertebrate hosts [7]. Species misidentifications have arisen because of similarities in morphology and because of the lack of isolates in molecular databases [8]. Most studies have focused on the genus Echinostoma, especially the "revolutum" group, e.g. [7][8][9][10][11][12][13][14], see Kostadinova and Gibson [11] for review. Despite numerous studies, two recent papers [15,16] showing that even by the use of molecular tools the taxonomy of this group is still not straightforward. However, taxonomic difficulties are also known in other groups of the Echinostomatinae [17,18]. One of the remaining interesting issues concerns the 27-collarspined echinostomes of the genus Isthmiophora Luhe, 1909. The type-species, I. melis, is a parasite reported mainly from European, Asian and American carnivores. However, Radev et al. [19], in his review of literature, lists c. 30 species as definitive hosts of this parasite. Host-induced morphological variation within this digenean, which apparently lacks host specificity, should be clearly visible.
During our long-term studies on the helminth fauna of rodents and carnivores we have collected 27 collarspined echinostomes which differ in morphology to an extent suggesting the presence of more than one species. Molecular studies, on the other hand, suggested that these worms belong to Isthmiophora. Here we describe this material, and the extent of host-related variation in this parasite.

Parasite sampling
Representatives (N = 148) of Isthmiophora used for the morphological analysis were collected from four host species: striped field mouse (Apodemus agrarius, N = 37), European badger (Meles meles, N = 13), American mink (Neovison vison, N = 64) and European hedgehog (Erinaceus europaeus, N = 34), all captured during parasitological and faunistic studies carried out by the Department of Parasitology in cooperation with the Polish Academy of Sciences. The rodents were captured in Lower Silesia (Dolina Baryczy, Nature Reserve "Stawy Milickie", 51°31′56″ N/17°20′12″E) in 2010permission 46/2008 issued by the Second Local Commission for Animal Experiments, worms form the mink (N. Poland; Marzęcino, 54°13′ 1.54″N 19°13′20.43″E) captured in 2010 were obtained from the Polish Academy of Sciences, trematodes from the hedgehog and badger (2010) were obtained from the Czech Republic (Zahlinice, 49°17′06″N/17°28′41″ E). After washing in tap water, the worms were fixed in 70 % ethanol. Some of the collected trematodes were stained in iron-aceto-carmine [20], dehydrated in a graded ethanol series, cleared in clove oil, mounted in Canada balsam and identified according to Kostadinova and Gibson [17]. The voucher specimens of trematodes obtained from each hostare deposited in the polish helminthological collection of Natural History Museum of Wroclaw University (MNHW).

Statistical analysis
All the examined specimens of I. melis were subject to detailed morphological and morphometric analysis, including the following measurements: body length (L), maximum body width (W), body area (BA), maximum body width as a proportion of body length (BW), forebody length (FB), forebody as a proportion of body length (FO), hindbody length (HB), hindbody as a proportion of body length (H), post-testicular region length (PTR), posttesticular region length as a proportion of body length (T), oral sucker area (OSA), ventral sucker area (VSA), anterior testis area (ATA), posterior testis area (PTA), ovary area (OA), gonad area/body area (GA/BA), ventral sucker to ovary distance as a proportion of body length (U), egg length (EL), egg width (EW). The body and gonad areas were calculated using the following equations: body area = π*(body length/2)*(body widith/2); gonads area = π*r 2 . All the measurements were expressed in micrometers and proportions as percentage. Prior to the analysis the data were log-transformed (log 10 ). The mean (M), minimum/ maximum values and coefficients of variation (CV %; defined as the ratio of standard deviation to the mean) were calculated for all the variables. One-way analysis of variance (ANOVA) was carried out to test if the particular morphological features of Isthmiophora differed between the host species. In the next step we performed discriminant analysis. To avoid the size effect of the worms (Isthmiophora spp. isolated from the badger was much bigger than those from the other hosts) only variables expressed as ratios (BW, FO, H, T, U, GA/BA) were included in this analysis. Moreover, according to the literature data, the major diagnostic characters in this taxon are based on ratios (i.e. BW, FO, T and U). All the analyses were conducted using Statistica 10.0 software.

Molecular analysis
Molecular analysis was performed for I. melis collected from four host species studied, from which a set of two worms was used for the analysis (N = 8). DNA was extracted using the DNeasy Blood and Tissue Kit (Qiagen), and amplified using PCR specific for 2 nuclear markers (internal transcribed spacers 1 and 2 [ITS1, ITS2] and a fragment of mitochondrial cytochrome oxidase I [CO1] gene) ( Table 1). Two additional molecular markers (SSU and LSU of rDNA) were amplified for the specimens from A. agrarius. PCR conditions included initial denaturation in 95°C for 5 min, followed by 35 cycles: 45 s denaturation (95°C), 30 s annealing (52°C for SSU, LSU, ITS 1, ITS 2 and 48°C for COI), 30 s elongation (72°), and a 5 min step of final elongation (72°C). PCR products were sequenced using the same primer pairs, and chromatograms inspected visually for ambiguities. In order to elucidate any homologies with the previously deposited sequences in GenBank, we conducted a BLAST search (http://blast.ncbi.nlm.nih.gov/Blast.cgi?CMD = Web&PA-GE_TYPE = BlastHome). Multiple alignment was done using CLUSTAL W in MEGA 5.0 package [21]. The sequences obtained in this study were deposited in GenBank under the following accession numbers:

Molecular analysis
The morphological distinctness of I. melis from the striped field mouse did not permit unambiguous identification of the parasite to specific or even generic level. Two markers 18S rDNA (1078 bp) and 28S rDNA (1352 bp) were therefore used for preliminary identification. BLAST analysis showed 99 % similarity with the sequences of I. melis

Morphological analysis
The means, ranges and CV % values of I. melis from the four host species are shown in Table 2 The results of post hoc Tukey test showed that the differences were mostly associated with the trematodes from A. agrarius (Table 3). These specimens were characterised by a relatively smaller body size, higher values of maximum body width, expressed as proportion of body length (BW), a very short post-testicular region and therefore low values of the post-testicular region as a proportion of body length (T) (Fig. 1). The worms from A. agrarius displayed the highest values of relative gonad area/body area (GA/BA) (Fig. 1). In discriminant analysis (Table 4) the model was generated by the use of 6 variables. The chi-square test showed that the first three roots were required to separate I. melis among the four host species. The roots accounted for 91.5 % (Root 1), 96.6 % (Root 2) and 100 % (Root 3) of the overall variation. Root 1 separated the worms from A. agrarius based on the following variables, in order of descending importance: GA/BA, T, FO and U. The analysis also revealed that according to these criteria all the specimens of I. melis from A. agrarius were classified correctly (Table 5). Roots 2 and 3 separated trematodes from M. meles based on GA/BA and U, however this explained only 8.5 % of the variation. These results are also visible in the plot of canonical scores (Fig. 2) where I. melis from the striped field mouse are clearly separated from those isolated from the remaining hosts.

Discussion
The history of the genus Isthmiophora Luhe, 1909, especially in relation to the genus Euparyphium Dietz, 1909, is long and complicated, but both genera were established as valid by Kostadinova and Gibson [17]. The main characteristic features of Isthmiophora are: anterior position of the testes (proportion of length of post-testicular region to body length = 30-50 %), short forebody (FO = 10-20 %), presence of an armed cirrus, small head collar with 27 collar spines, varied size of dorsal spines (oral longer than aboral), short uterus and large eggs [17]. The life cycle of Isthmiophora includes lymnaeid snails, tadpoles and fish as intermediate hosts and carnivores as definitive hosts. Six species (I. melis, I. hortensis (= Echinostoma hortense), I. beaveri, I. citellicola, I. inermis, I. lukjanovi) are currently regarded as valid [17]. I. melis is widespread in Europe, Asia and North America and uses more than 30 species of vertebrates as definitive hosts [19], including humans and rodents: Apodemus agrarius, A. sylvaticus, Rattus norvegicus and Mus musculus [22][23][24]. In Poland the species has been reported from fox, marten, badger, hedgehog and rodents [22,24]. The specimens of I. melis from A. agrarius collected in Lower Silesia did not fully correspond to the description of I. melis [17], and two of the key features: forebody as a proportion of body length (FO) and post-testicular field as a proportion of body length (T), were distinct. According to Kostadinova [18], Isthmiophora possessed an intestinal bifurcation just anterior to the ventral sucker, the cirrus was armed and T = 30-50 % while Euparyphium was characterised by the intestinal bifurcation located halfway between the pharynx and the The data are presented pairwise for particular host species (Aa -A. agrarius, Ee -E. europaeus, Nv -N. vison, Mm -M. meles) and indicate statistical significance (+) or its lack (n/s) ventral sucker, unarmed cirrus and T = 20-30 %. The worms from A. agrarius had a short post-testicular field as a proportion of body length (T = 26.6 %), an armed cirrus and the intestinal bifurcation situated halfway between the pharynx and the ventral sucker. These features did not permit unambiguous identification of the trematodes as Isthmiophora. Additionally, Radev et al. [19] showed that in experimental infections of hamsters, specimens of I. melis still corresponded to the general description of the species, and the key features did not change significantly. Molecular identification, based on SSU and LSU of rDNA, of I. melis from the striped field mouse definitively confirmed their identity as Isthmiophora, while the less conservative markers (ITS1/ITS2 of rDNA and COI of mtDNA) pointed to a specific identity as I. melis. Additional specimens of I. melis, isolated from different hosts (M. meles, N. vison and E. europaeus), shared this molecular identity, with minimal (1.4 %) variation within the COI gene. Based on this molecular analysis, we must conclude that the echinostomatids collected from A. agrarius did represent I. melis. Nolan and Cribb [6] presented an extensive discussion of the role of ITS sequences in digenean taxonomy. Internal transcribed spacers in this group in general showed a small intraspecific variation, which was however sufficient to explore the validity of species boundaries in the group [6]. Morgan and Blair [12] also investigated the taxonomic position of eight 37-collarspined echinostomatid species using ITS sequences and found that these spacer regions provided sufficient variation to distinguish 5 of the 8 nominal species examined, and the level of interspecific variation ranged between 1.1 % and 19.2 %. The remaining three species had identical ITS sequences and were indistinguishable. The same authors [13] also re-examined the same material using mitochondrial markers (CO1 and ND1), which allowed for unambiguous identification of the analysed material. Based on the reliability of the combination of nuclear and mitochondrial markers in these studies [6,12,13], we are confident that our specimens from A. agrarius do indeed represent I. melis. The observed morphological variation must therefore be host-induced phenotypic variation, and its scale affects the diagnostic features at both generic and specific level. There is an extensive literature on the influence of population density on the echinostome morphology (e.g. the "crowding effect" of Fried and Freeborne [25,26]), but we suspect an effect of host longevity as well. In general, the lifespan of carnivorous species is considerably longer than that of small rodents. The lifespan of A. agrarius in the wild approximates a few months (5)(6)(7)(8) only, while, for example, the lifespan of M. meles is up to 15 years. The growth of body size and internal organs of trematodes is correlated at the initial phase. At a later period, when the gonads are fully developed, the body continues to grow with a simultaneous slower growth rate of gonads. For example, in experimental studies on the development of E. revolutum, Franco et al. [27] observed that gonads were fully developed 20-25 days post infection while full body size was only attained 55 days post infection. In our studies the highest values for the coefficient of relative gonad area to body area was observed in the trematodes from A. agrarius, indicating that the growth of the body had ceased; the values of this coefficient in the striped field mouse were almost identical as those observed in the badgerthe type host for I. melis.
Genetic markers constitute a powerful tool in the studies on intraspecific variation in many taxa, including helminths, but the morphology still plays a crucial part in species descriptions [28]. However it is evident that morphology alone may not provide adequate taxonomic resolution and may lead to misidentifications. The phenotypic plasticity of helminths has been reported frequently in the literature, e.g. [29][30][31][32]. For example, Boyce et al. [30] explain the differences in the morphology of Notocotylus malhamensis Boyce et al. 2012 as a result of the presence of young adults in one of the hosts, i.e. the specimens of N. malhamensis in Microtus agrestis have not fully developed. The second possible reason of host-induced morphological differences in N. malhamensis is crowding effect. The wide host range of I. melis combined with the very different sizes of the hosts (e.g. badger vs. field striped mouse) makes the phenotypic plasticity even more spectacular. Our studies suggest that species identification is very subjective and,  when descriptions of new species or even higher taxa are based on few specimens, misidentification is very likely. Thus the combination of morphology with molecular analysis and studies on life histories is most desirable when identifying parasites.

Conclusions
The morphological traits of Isthmiophora melis are highly variable and host-dependent, and without molecular analysis they might lead to a description of a new species or even genus. Such a high level of intraspecific variation may be affected by the host's longevity.
Submit your next manuscript to BioMed Central and take full advantage of: