Molecular analysis of polymorphic species of the genus Marshallagia (Nematoda: Ostertagiinae)

Background The genus Marshallagia (Family Haemonchidae, subfamily Ostertagiinae) contains multiple species of nematodes parasitising the abomasum (or duodenum) of ruminants, in particular of Caprinae. Male specimens have been described to be polymorphic with the frequent/major morphotype initially described in the genus Marshallagia while the minor/rare morphotype was initially often placed in the genus Grossospicularia. Due to common morphological features, certain pairs of morphotypes were suggested to belong to the same species such as Marshallagia marshalli/M. occidentalis. However, molecular evidence to confirm these pairs of morphotypes belonging to the same species is missing. Methods In the present study, Marshallagia sp. were collected from domestic sheep in Uzbekistan. Male specimens were morphologically described with particular emphasis on the structure of the bursa copulatrix. After DNA isolation from morphologically identified specimens, PCRs targeting the ribosomal internal transcribed spacer 2 (ITS2) and mitochondrial cytochrome c oxidase subunit 1 (cox1) regions were conducted. After Sanger sequencing, maximum likelihood phylogenetic analyses and pairwise identities between sequences were calculated. Results The major morphotypes of M. marshalli, M. schumakovitschi and M. uzbekistanica and the minor morphotypes M. occidentalis, M. trifida and M. sogdiana were identified and their morphology was documented in detail. ITS2 sequences showed little variation and did not allow diagnosing species. In contrast, phylogenetic analysis of cox1 sequences identified highly supported clusters and verified that M. marshalli, M. occidentalis and M. uzbekistanica are different morphotypes of the species M. marshalli while M. schumakovitschi and M. trifida represent distinct morphotypes of M. trifida. For M. sogdiana no corresponding major morphotype could be identified in the present study. Due to a large barcoding gap, comparison of cox1 sequences in terms of percent identity was sufficient to reliably assign the sequences to a particular species without phylogenetic analysis. Conclusions The data presented here create a framework that will allow the classification of other members of the genus in the future and underline that parallel morphological and molecular analysis of specimens is crucial to improve the taxonomy of polymorphic species.


Background
Nematodes of the genus Marshallagia Orloff, 1933 are parasites of the abomasum and duodenum of free-ranging and domesticated ruminants and are most often associated with Caprinae from the Holarctic region [1,2]. They have a considerable impact on the metabolism of host animals frequently leading to economic losses [3][4][5]. According to our data, the species of Marshallagia marshalli Ransom, 1907 is widespread in Uzbekistan and it is the most frequently detected member of this genus [6].
Drozdz [7,8] described that morphological dimorphism is common among males of several species of the subfamily Ostertagiinae, i.e. presence of "major" and "minor" morphotypes in the same host. Major morphotypes are named due to their higher frequency relative to minor morphotypes, not due to differences in size. In contrast to males, females are monomorphic, and this situation has often led to the description of distinct species for these male morphotypes.
In the genus Marshallagia, five dimorphic species have been described by Drozdz [9]. Later, 12 dimorphic species of Marshallagia were differentiated by morphological descriptions incorporating the synlophe in males, the structure of the spicules as well as the genital cone in major and minor morphotype males [2]. In particular, structural characters of the spicules and genital cone were used to distinguish between major and minor morphotypes. The major morphotypes were usually placed in the genus Marshallagia while minor morphotypes were often allocated to Grosspiculagia Orloff, 1933. Thus, major and minor morphotypes of the same species have historically been described as different nominal species, often in separate genera [9,10]. However, there are also morphological features that are shared between major and minor morphotypes but can be used to discriminate between species including in particular details of the morphology of esophagus and synlophe as well as the shape of the rays of the copulatory bursa [9]. For instance, M. occidentalis Ransom 1907 (minor morphotype) has clearly distinguishable morphological features discriminating it from M. marshalli (major morphotype), but there are clear hints that M. marshalli and M. occidentalis are synonyms and represent only different morphotypes of M. marshalli, whereas M. mongolica Shumakovitschi, 1950 appears to be a major morphotype while M. grossospiculum Li, Yin, Kong & Jang, 1987, which corresponds to Marshallagia sp. 1 of Drozdz [9], is most likely a minor morphotype of the same species. The same relationship presumably also applies to M. schumakovitschi Kadyrov, 1959 (major morpotype) and M. trifida (Guille, Marotel & Penisset, 1911) (= Marshallagia sp. 2 of Drozdz [9]) (minor morphotype) [2].
Even less is known about the occurrence of cryptic species in other frequently occurring representatives of the Ostertagiinae, especially in M. marshalli due to its presence in several host species [2,11]. Regarding nomenclature, taxonomy and phylogeny of the genus Marshallagia, Hoberg et al. [2] proposed that additional studies are required combining morphological and molecular analysis from individual specimens to confirm descriptions of the major species M. skrjabini as well as the minor species M. belockani and M. sogdiana. Moreover, it remains an open question as to whether or not M. marshalli/M. occidentalis is distributed in North America and in Eurasia. Marshallagia marshalli/M. occidentalis is the only species of the genus with a Holarctic distribution but detailed comparisons between North American and Central Eurasian populations have not been conducted. Therefore, it should be considered that there are two closely related species with more restricted geographic distribution patterns [2] and that reports of identification of M. marshalli from Eurasian ruminants are due to misidentification.
Based on studies by Drozdz [9], Hoberg et al. [2] and Wyrobisz et al. [12]  Within the Strongylida, species distinction supported merely by morphological features is difficult, and requires confirmation by means of molecular methods. For instance, recent data have suggested that Cooperia spatulata is just a morphotype of C. punctata [13] while there is evidence that the small strongyle morphospecies Cylicostephanus calicatus and C. minutus are in fact cryptic species complexes of at least two and three genospecies, respectively [14]. The taxonomy of Ostertagiinae is mainly complicated by complex relationship between species and morphotypes in the genus Teladorsagia, but complexity may also be expected among other Ostertagiinae (e.g. in the genera Ostertagia and Marshallagia) [12]. Studies devoted to the problems of taxonomy of strongylid nematodes have shown that noncoding regions of rRNA genes, in ITS1 and particularly ITS2, are well suited to allow taxonomic discrimination and this also applies to members of the Ostertagiinae [15][16][17][18][19][20][21][22][23][24]. These studies have led to important insights in the evolution and solved general questions of phylogeny of Ostertagiinae including confirmation of conspecificity for M. marshalli and M. occidentalis [25]. Furthermore, questions regarding conspecificity of supposed major and minor morphs and cryptic species within the genus of Orloffia Drozdz, 1965 were solved [26]. There is no unanimous opinion about the taxonomic independence and specific composition of the genus Orloffia. At the same time, there are several species and genera, for which important taxonomic and phylogenetic questions remain unresolved. Additional molecular analyses of ribosomal and mitochondrial DNA will allow understanding these problems in more depth.
Regarding species identification in terms of barcoding properties, mitochondrial DNA sequences were shown to be superior to ribosomal spacers [27]. Recent work on Cooperia spp. (Cooperiidae, Cooperinae) and C. minutus (Strongylidae, Cyathostominae) revealed that combined analyses of mitochondrial and nuclear marker sequences improved species identification and phylogenetic analyses [13,14].
This study aimed to provide sequence data on mitochondrial cytochrome c oxidase subunit 1 (cox1) and nuclear ribosomal intergenic spacer 2 (ITS2) DNA for some species or morphotypes of the genus of Marshallagia. The objective was to morphologically identify individual Marshallagia sp. specimens from Uzbekistan to the species and morphotype level followed by obtaining molecular ITS2 and cox1 data from the same specimens in order to clarify the taxonomic status of the local Marshallagia species/morphotypes. Male specimens were manually cut in two parts: the posterior region was cleared in phenol alcohol (80% melted phenol and 20% ethanol) for examination of the morphological features and the anterior and middle parts were fixed in 70% ethanol for the molecular studies (Table 1).

Morphological identification
Marshallagia species were identified according to morphological and morphometrical characters using literature data [1,2,9]. All adult male worms isolated from each sheep were morphologically analyzed to identify parasite species. The species identification was established based on caudal bursa according to the features proposed by above literature, especially morphological characters and measurements of spicules, dorsal ray and gubernaculum (Table 1).
An equal mixture of lactic acid and glycerin was used to enlightenment the posterior part of the studied nematodes without additional staining. Also included in the present paper are morphological studies made by the authors analyzing specimens (paratypes or sintypes) of species (M. marshalli Ransom, 1907 andM. schumakovitschi Kadyrov, 1959) in the Central Helminthological Museum FGBNU, Russian Institute of Parasitology, Animal and Plant named after K. I. Skrjabin, Moscow. For this purpose, a microscope ML 2000 equipped with a digital camera (Meiji, Saitama, Japan) was used.

DNA extraction
For DNA isolation, at least a single specimen of each species/morphotype was used. Before isolation of genomic DNA, the ethanol was removed and the adult nematodes were washed with sterile water and DNA was extracted using the NucleoSpin ® Tissue Kit (Macherey-Nagel, Düren, Germany) in accordance with the manufacturer's protocol. The DNA was eluted with 50 µl elution buffer provided in the kit and stored at − 20 °C until further use. The extracted DNA was quantified on a Take3 plate in an Epoch plate reader (Biotek, Berlin, Germany).

PCR and cloning
PCRs were conducted using (i) a combination of the forward and reverse primers flanking the complete ITS2 region [28] and (ii) a partial cox1 gene fragment [29] ( Table 2). PCR reactions contained 0.2 mM dNTPs, 250 nM of each primer, 0.4 U Phusion Hot Start II High-Fidelity DNA polymerase (Thermo Fisher Scientific, Darmstadt, Germany) and 2 μl template DNA in 20 μl 1× HF buffer. PCRs were performed on a C1000 or S1000 PCR cycler (Bio-Rad, Feldkirchen, Germany). PCR products were purified using DNA Clean & Concentrator TM -5 (Zymo Research, Freiburg, Germany) and amplification products were analyzed by agarose gel electrophoresis in 1.0-1.5% agarose gels. Purified fragments were ligated into the StrataClone Blunt PCR Cloning Vector pSC-Bamp/kan (Agilent, Waldbronn, Germany) and transformed into StrataClone SoloPack competent Escherichia coli cells according to the manufacturer's protocol. Plasmid DNA was purified using the EasyPrep1 Pro kit (Biozym, Hessisch Oldendorf, Germany) and sent for sequencing to LGC Genomics (Berlin, Germany).

Sequence comparisons and phylogenetic analyses
Sequences from the present study were analyzed together with sequences previously deposited in Gen-Bank. As an outgroup, two sequences per gene from the species Teladorsagia circumcincta, were included. Accession numbers of all sequences from specimens investigated in the present study are provided in Table 3. The ITS2 and cox1 sequences were aligned using MAFFT (multiple sequence alignment using fast Fourier transformation) in the Q-INS-I modus that takes predicted RNA secondary structures into account [30] and the M-COFFEE modus of T-Coffee (Treebased Consistency Objective Function for alignment Evaluation) [31], respectively. The cox1 alignment was manually edited to ensure that codons were not interrupted by gaps. For calculation of relative identity (%) between sequences, alignments were analyzed using the dist.dna function in the ape 4.0 (Analyses of Phylogenetics and Evolution) package [32] in R 4.0.0 statistics software [33]. Identities were calculated as "raw" identities and pairwise deletion of positions with gaps was turned on. Comparisons of sequences within the genus Marshallagia were sorted into the individual intraspecies comparisons and a single category containing all interspecies comparisons. The identity in percent for all these comparisons within the genus Marshallagia were compared using the Kruskall-Wallis test followed by a Conover-Iman post-hoc test with the function posthoc. kruskal.conover.test as implemented in the R package PMCMR 4.3 [34]. All P-values below 0.05 were considered to be statistically significant. Scatter plots were visualized using GraphPad Prism 5.03 (GraphPad, La Jolla, USA). Phylogenetic analyses were conducted on a single gene level. First, substitution saturation tests were conducted according to Xia et al. [35] using DAMBE 5 (Data Analysis in Molecular Biology and Evolution) software [36]. DAMBE 5 was also used to split the cox1 alignment into one partition for the first and second and another partition for the third codon position. Maximum likelihood phylogenetic trees were calculated using IQ-TREE [14] on the IQ-TREE server (http://iqtre e.cibiv .univi e.ac.at). Using the ModelFinder option of IQ-TREE [37], autodetermination of the best model applying the Bayesian information criterion was performed including models with FreeRate heterogeneity. Ultrafast bootstrapping (1000 bootstrapped replicates) [38] and the Shimodaira-Hasegawa approximate likelihood ratio test (SH-aLRT) (1000 replicates) [39] were used to obtain node support statistics.

Morphological identification
Six morphotypes of Marshallagia species were found within the present material. Major and minor Marshallagia species were isolated from domesticated sheep in Uzbekistan, separated according to their morphological identification (Fig. 1) and morphometric comparison (  [7,8] described the phenomenon of regular co-occurrence of rather rare (minor) species in pairs with the more numerous (major) species, which, together with subtle morphological features, lead to the hypothesis that they represent different morphotypes of the same species. Based on this hypothesis, the analyzed specimens of Marshallagia were divided into six separate morphotypes that were grouped into three species with co-existing morphotypes (Fig. 1). The original micrographs of the specimens that were used for molecular analysis are presented for one exemplary individual per morphotype: M. marshalli (Fig. 1a); M. schumakovitschi (Fig. 1b); M. uzbekistanica (Fig. 1c); M. occidentalis (Fig. 1d); M. trifida (Fig. 1e); and M. sogdiana (Fig. 1f ). A detailed morphological description of the different morphotypes is given in Additional file 1: Text S1. The most reliable characters for differentiation among species of Marshallagia, and specifically of the major morphotypes of the respective species, include the placement of the trifurcation of the spicule tips, the form of the dorsal and ventral processes (relative length, curved or straight), and the chitinized structure of the tip of respective processes.
The three major morphotypes designated as M. marshalli, M. schumakovitschi and M. uzbekistanica (Fig. 1ac) differ by the distal ends of spicule processes, strongly curved for the first one and with tubercles for the second and third species. All three major morphotypes are characterized by the absence of a gubernaculum (Fig. 1a-c).
Marshallagia marshalli (Ransom, 1907) is the typespecies for the genus. Ransom [40,41] described this species and the minor morphotype, M. occidentalis, based on specimens in domesticated sheep (O. aries L.) from North America. In specimens of the major morphotype of M. marshalli the spicules are strongly curved in lateral view, 210-310 µm in length, 42-54 µm in width; eyelet at trifurcation prominent; with dorsal and ventral process nearly equal in length; gubernaculum absent or strongly chitinized (Fig. 1a). The M. marshalli morphotypes studied by us corresponded to the morphology of the paratype of M. marshalli (CHM 14768) ( Table 1).
Marshallagia schumakovitschi differs from M. marshalli in the structure of the spicule. The most distal sixth part of the spicules is divided into three processes. Dorsal and ventral processes are nearly equal in length, the ventral process terminates in a simple tip, which may be bent; the dorsal process is weakly sclerotized, blunt and not strongly recurved and extends to near the termination of the main shaft (Fig. 1b).
Marshallagia uzbekistanica is a morphologically atypical form placed in the genus Marshallagia in the original description. Spicules are asymmetric (Fig. 1c). The characteristic of M. uzbekistanica turned out to be a peculiar structure of the spicules, which are weakly chitinized at the proximal end and granular in structure [42].
Minor morphs represented in the present study by the morphotypes M. occidentalis, M. trifida and M. sogdiana correspond to the diagnosis of the genus Grosspiculagia, which is now considered to be a synonym of Marshallagia. In contrast to the major morphs, all minor morphs have thick spicules that are split into three processes. The two more massive processes have cap-shaped distal ends and sometimes a hook-like outgrowth. In contrast to the major morphs, all minor morphs have a transparent, sometimes subtle gubernaculum (Fig. 1d-f ).
Marshallagia occidentalis represents the minor morphotype of M. marshalli [9]. Near the middle of the spicule length, spicules are divided into three processes: two ventral and one dorsal. A gubernaculum is present but to the rear its diameter is strongly reduced (Fig. 1d).
Marshallagia trifida is the minor morphotype of M. schumakovitschi and can be identified based on its spicule structure [2,9]. The ventral process of the spicules is strongly curved; the dorsal process extends to the tip of the main shaft of the spicule. The gubernaculum is fusiform (Fig. 1e).
Marshallagia sogdiana is the minor morphotype of M. skrjabini Asadov, 1954 and was transferred to that species as a new combination of morphs [2]. The proximal ends of the spicules are characterized by the presence of a peculiar, disc-like structure (Fig. 1f ). In the middle of the spicule it is divided into three processes: two ventral and one dorsal. The distal spicules have a membrane in the form of a sheath. A gubernaculum is present.
Thus, based on morphological characters and morphometric comparison of males of Marshallagia sp., two pairs of major and minor morphotypes were identified: M. marshalli/M. occidentalis, M. schumakovitschi/M. trifida while for the pair M. skrjabini/M. sogdiana only the minor morphotype was found (Table 4). In addition, M. uzbekistanica was identified as an unusual major morphotype. The data on the morphology of these species from samples collected in Uzbekistan presented here closely corresponds to the previously published data by Asadov [43], Ivashkin [1] and Hoberg et al. [2]. Herein, to scrutinize the identity of the morphs of the genus Marshallagia, molecular studies were conducted in particular to confirm which major and minor morphotypes belong to the same species.  nuclear ITS2 rRNA (321-325 bp excluding the primers) and mitochondrial partial cox1 (696 bp) genes were amplified, cloned and sequenced. All sequences were deposited in GenBank under the accession numbers provided in Table 3.

Internal transcribed spacer 2
Identity between all ITS2 sequences of the genus Marshalagia from the present study (n = 13) or from Gen-Bank revealed between 89.7% and 100% identity with 85% of the pairwise comparisons showing > 95% identity. A maximum likelihood phylogenetic tree calculated from ITS2 sequences from the present study plus those available in GenBank using two representative T. circumcincta sequences as an outgroup is shown in Additional file 2: Figure S1. The phylogram reveals that the ITS2 sequence contains virtually no phylogenetic signal since (i) there are barely any clusters of sequences showing high statistical support, (ii) sequences assumed to come from different species were virtually identical and (iii) the sequences assigned to the same species are found scattered all over the tree. This indicates that ITS2 sequences are not suitable to address taxonomic or even phylogenetic questions within the genus Marshallagia. This is in agreement with other recent studies analyzing closely related strongyle nematodes showing that ITS2 is an excellent marker to identify the genus, but that closely related species differ only minimally in their ITS2 sequence and that the phylogenetic signal obtained from ITS2 sequences of Marshallagia specimen was not reliable [13,14].

Cytochrome c oxidase subunit 1 gene
The phylogenetic tree calculated from cox1 sequences from the present study or downloaded from GenBank identified seven highly supported clusters (named I-VII from basal to distal operational taxonomic units (OTUs) in Fig. 2) with very low variability within the clusters. All available Marshallagia sequences come from only three studies. One from China [44] reporting 36 cox1 sequences, another Chinese study reporting a complete mitochondrial genome annotated as M. marshalli [45] and 13 sequences from the present study. The clusters I (Ostertagia lanceata erroneously assigned to the genus Marshallagia by Lv et al. [44] as indicated by Hoberg et al. [2], IV (M. occidentalis) and V (M. hsui) contained only sequences reported by Lv et al. [44].  [45]. The fact that two species sequences (M. marshalli and M. occidentalis) both occur in two different clusters clearly indicates that at least some of the published specimens were morphologically misidentified. Neither Lv et al. [44] nor Sun et al. [45] provided any morphological data for the specimens used to obtain their DNA sequences. Lv et al. [44] not even mentioned the criteria used for identification or any species identification key while Sun et al. [45] explicitly stated that morphological identification was difficult, and that molecular identification was used. They report that the ITS sequences obtained from their specimens were 99% identical to sequences KT428384 and HQ389231 deposited in Gen-Bank as M. marshalli. However, in the ITS2 phylogenetic analysis presented in Additional file 2: Figure S1, these two sequences are (i) not grouped closely together and (ii) also virtually identical to ITS2 sequences deposited in GenBank for other Marshallagia species. Based on the present findings it can be assumed that the complete mitochondrial genome reported by Sun et al. [45] does not represent a M. marshalli but instead a M. mongolica sequence.
Considering each of the highly supported major clusters in Fig. 2, identification of a valid species can be further confirmed by looking at the barcoding gap. Figure 3 shows raw percent identity for all pairwise comparisons between Marshallagia sp. sequences from the alignment used to calculate the cox1 phylogenetic tree in Fig. 2. It clearly exhibits that all intraspecies comparisons show  identities between 93.7% and 100% while all interspecies comparisons are in the range of 84.2-89.6%. Comparisons between different morphotypes of the same species are in the same range as comparisons between identical morphotypes. Since ranges of intra-and inter-species comparisons are not overlapping (barcoding gap), a simple calculation of identities between sequences will be sufficient to assign a sequence to a particular species allowing diagnosis without phylogenetic reconstruction.

Polymorphic Marshallagia species
The data presented here clearly show that M. marshalli is a polymorphic species and that M. occidentalis and M. uzbekistanica should be considered synonyms and that their descriptions detail the morphology of rare morphotypes of M. marshalli. In addition, the morphotypes M. schumakovitschi and M. trifida belong to the same species. Although M. schumakovitschi represents the major morphotype, this name should be considered a synonym of M. trifida since M. trifida was described earlier and this name has therefore priority according to the rules of zoological nomenclature. Data from Lv et al. [44] furthermore suggest that M. grossospiculum is a minor morphotype of the species M. mongolica and only the latter represents a valid species name. However, absence of morphological data in the publication by Lv et al. [44] suggests that further confirmation is warranted.
It remains unclear whether all Marshallagia species are polymorphic but the new data presented here confirm certain pairs of morphotypes that must be considered to be the same species. For other species, the availability of cox1 sequences will allow to group more morphotypes once new sequences from morphologically identified and documented material become available. Marshallagia DNA sequences from specimens of unclear origin The phylogenetic tree in Fig. 2 contains several GenBank entries from samples of doubtable origin. In particular, there is no available description for M. hsui and the name is considered to be a nomen nudum [2]. The sequences obviously belong to a Marshallagia species but without a detailed, publicly available morphological description, it is impossible to decide whether they represent an undescribed species or if they should be assigned to one of the many described species for which no cox1 reference sequences are currently available. The M. occidentalis specimens from China for which no morphological description was provided by Lv et al. [44] have presumably been identified incorrectly. The sequences apparently represent Marshallagia sequences but the assignment to M. occidentalis is in contrast to all data summarized by Hoberg et al. [2] and to the results of the present study, which all consider M. occidentalis as a minor morphotype of M. marshalli. Without voucher material deposited in a museum, the taxonomic value of these sequences will remain very limited and only after identical/highly similar sequences will be reported with a detailed morphological description and deposited voucher material, the sequences will be of epidemiological and taxonomic value.

Conclusions
Species limits remain poorly defined within the genus Marshallagia (and presumably other, particularly polymorphic Ostertagiinae as well) where subtle morphological differences, high morphological or genetic variability, incomplete descriptions, and circumscribed differential diagnoses hinder identification [2,43,46,47]. It was shown in the present study that at a number of described and named taxa from Eurasia are in fact only synonyms of previously established major or minor morphotypes. For holarctic species such as M. marshalli with its different morphotypes, it would be highly interesting to compare Eurasian and North American specimens regarding mitochondrial genotypes to determine if there is further population genetic structuring or even different genospecies on both continents. In any case, free access to species descriptions, including approaches to overcome language barriers such as very limited accessibility of original parasitic nematode descriptions in Russian for western parasitologists and vice versa, would be required to improve the situation. Missing of accurate and detailed figures and difficult access to representative type-specimens is especially problematic and complicates the possibility of complete and direct comparisons among otherwise similar species and respective morphotypes [2]. Accordingly, comprehensive revision of the genus Marshallagia appears warranted but is currently unfortunately unrealistic.