INTRODUCTION
The phenomenon of ‘species flocks’, that is the occurrence of numerous species of congeneric (or confamilial) parasites in the same host species, has been the focus of a number of studies, particularly of parasitic nematodes. One of the best-known examples of the phenomenon is that of the oxyuroid nematodes found in the colon of tortoises (Schad, Reference Schad1963; Petter, Reference Petter1966). In this instance, each species of tortoise harbours large numbers of congeneric or confamilial nematode species, all belonging to the family Pharyngodonidae (Petter, Reference Petter1966). However, there are differences in the longitudinal and radial distributions of these nematode species within the colon of the host (Schad, Reference Schad1963). Other examples of species flocks of parasitic nematodes are those found in the large intestines of equids, elephants and rhinoceroses, as well as those in the sacculated forestomachs of kangaroos and wallabies (Inglis, Reference Inglis1971). All of these nematodes have direct life-cycles with the ingestion of eggs or third-stage larvae from the environment (Anderson, Reference Anderson2000).
In the case of equids (horses, donkeys and zebras), 14 genera and 50 species of strongyloid nematodes belonging to the tribe Cyathostominea are currently recognized (Lichtenfels et al. Reference Lichtenfels, Kharchenko and Dvojnos2008), with the common co-occurrence of many species (Bucknell et al. Reference Bucknell, Hoste, Gasser and Beveridge1996; Anjos and Rodrigues, Reference Anjos and Rodrigues2003; Bu et al. Reference Bu, Niu, Gasser, Beveridge and Zhang2009; Kuzmina et al. Reference Kuzmina, Zvegintsova and Zharkikh2009), but again with differences in the distribution of species within the gastro-intestinal tract (Ogbourne, Reference Ogbourne1976; Mfitilodze and Hutchinson, Reference Mfitilodze and Hutchinson1985; Bucknell et al. Reference Bucknell, Gasser and Beveridge1995; Stancampiano et al. Reference Stancampiano, Mughini Gras and Poglayen2010). Comparably detailed studies on the strongyloid nematodes of elephants and rhinoceroces have not been conducted, but four genera and 49 species belonging to the related strongyloid tribes Kiluluminea, Murshidinea and Quiloninea are known to occur in their large intestines (Chabaud, Reference Chabaud1957; Round, Reference Round1968; Canaris and Gardner, Reference Canaris and Gardner2003; Beveridge and Jabbar, Reference Beveridge and Jabbar2013). In the case of kangaroos and wallabies (family Macropodidae) some 36 genera and 256 species of nematodes belonging to the sub-family Cloacininae occur in the sacculated forestomach (Beveridge and Chilton, Reference Beveridge and Chilton2001), frequently in large numbers (Beveridge and Arundel, Reference Beveridge and Arundel1979), with again, some degree of spatial separation within the stomach (Mykytowycz, Reference Mykytowycz1964; Arundel et al. Reference Arundel, Beveridge and Presidente1979; Pamment et al. Reference Pamment, Beveridge and Gasser1994).
In a critical review of the phenomenon of species flocks in parasitic helminths, Kennedy and Bush (Reference Kennedy and Bush1992) indicated several difficulties in the application of this appellation to the examples cited above. First of all, these authors noted that species flocks, according to classical definitions, could be defined either by ecological parameters such as co-occurrence (Mayr, Reference Mayr, Echelle and Kornfield1984) and endemism (Ribbink, Reference Ribbink, Echelle and Kornfield1984) or could be circumscribed phylogenetically, with a species flock being a monophyletic assemblage (Greenwood, Reference Greenwood, Echelle and Kornfield1984). In the former case, such associations of multiple congeners could develop through a number of host colonization events, while in the latter case, the communities could evolve through intra-host speciation. Intra-host speciation (Price, Reference Price1980; Poulin, Reference Poulin2007) or even sympatric speciation (Kunz, Reference Kunz2002) are considered to be potentially common modes of evolution in parasites.
The phylogenetic definition of species flocks has been successfully applied to cichlid fishes (Seehausen, Reference Seehausen2006) and to rock fishes (Alesandrini and Bernardi, Reference Alesandrini and Bernardi1999), but the paucity of rigorous phylogenetic studies of parasitic nematodes means that this definition is often not applicable. In the case of the cyathostomin nematodes of equids (Equus spp.), available molecular evidence suggests that they do indeed form a monophyletic assemblage (Hung et al. Reference Hung, Chilton, Beveridge and Gasser2000; McDonnell et al. Reference McDonnell, Love, Tait, Lichtenfels and Matthews2000). The same situation may apply to the strongyloid nematodes of elephants as most species belong to two related tribes, the Murshidiinea and Quiloninea (Lichtenfels, Reference Lichtenfels, Anderson, Chabaud and Willmott1980). This may well also apply to the Cloacininae in macropodids. However, appropriate morphological and molecular phylogenetic studies are lacking. A similar situation pertains, in the case of molecular studies, to the oxyuroid nematodes in tortoises for which molecular data are lacking. As a consequence, among nematodes, it appears that the phylogenetic definition of a species flock may only be applicable currently to the cyathostomin nematodes of equids. A molecular study of species of Onchocerca occurring in cattle also suggests within-host speciation in this nematode genus, but the result is dependent upon the inclusion of remaining congeners (Morales-Hojas et al. Reference Morales-Hojas, Cheke and Post2006).
In considering the ecological definition of a species flock, Kennedy and Bush (Reference Kennedy and Bush1992) pointed out that the criterion of ‘co-occurrence’ was also difficult to apply since, for several parasite groups, relevant data were not available on the co-occurrence of congeneric or confamilial parasites within an individual host; rather published data described the meta-community. This ecological difference is also reflected in phylogenetic studies in which the criterion of sympatry in potential examples of intra-host speciation can be problematical (McCoy, Reference McCoy2003). Data on the co-occurrence of species of nematodes are however available for the cyathostomins of equids (Bucknell et al. Reference Bucknell, Hoste, Gasser and Beveridge1996; Anjos and Rodrigues, Reference Anjos and Rodrigues2003; Bu et al. Reference Bu, Niu, Gasser, Beveridge and Zhang2009; Kuzmina et al. Reference Kuzmina, Zvegintsova and Zharkikh2009) as well as for the cloacinine nematodes of macropodids (Beveridge et al. Reference Beveridge, Chilton and Spratt2002), but not for the remaining nematode communities cited.
In spite of these potential difficulties, the phenomenon of species flocks in parasites clearly warrants further study. The genus Cloacina, found in the stomachs of macropodid marsupials may represent a suitable model for additional studies as it currently contains 116 described species (Beveridge, Reference Beveridge1998, Reference Beveridge1999, Reference Beveridge2002, Reference Beveridge2014; Beveridge and Speare, Reference Beveridge and Speare1999; Beveridge et al. Reference Beveridge, Jabbar and Shuttleworth2014a , Reference Beveridge, Ngyuen, Nyein, Cheng, Koehler, Shuttleworth, Gasser and Jabbar b ) with several additional species as yet undescribed (Chilton et al. Reference Chilton, Huby-Chilton, Johnson, Beveridge and Gasser2009). These nematodes also have a relatively high degree of host specificity (Beveridge et al. Reference Beveridge, Chilton and Spratt2002). In addition, the study of a single genus (although the monophyly of Cloacina has not yet been investigated using molecular methods) overcomes the difficulty of deciding whether species flocks should be considered as being composed of congeners or whether the concept should expand to con-sub- familiar or con-familiar taxa (see Kennedy and Bush, Reference Kennedy and Bush1992), a potential complication in the studies of the nematode assemblages of equids. Furthermore, studies to date of different species of kangaroos and wallabies (seven species of the currently recognized 54 host species were studied by Beveridge et al. Reference Beveridge, Chilton and Spratt2002) have shown that assemblages of nematode species range across a continuum, from three to 12. This spectrum of hosts with varying numbers of co-occurring congeners is not represented in any other host-parasite system as there remain only five extant species of equids (all with a very similar parasite fauna), three species of elephants and four species of rhinoceros, all relicts of formerly more diverse faunas (Franzen, Reference Franzen2010). In addition a number of the latter host species are endangered thereby imposing a limitation on parasitological studies.
In an attempt to address the question of whether the species flocks of Cloacina seen in macropodids are monophyletic, Beveridge et al. (Reference Beveridge, Chilton and Spratt2002) undertook a phylogenetic analysis based on morphological characters. However, given that only a limited number of morphological characters was available, a common phenomenon for parasitic nematodes, the resulting phylogeny exhibited a relatively low consistency index. In spite of this reservation, it was suggested that while several nematode species pairs could be identified in individual host species, there was no strong evidence for the existence of monophyletic species flocks (Beveridge et al. Reference Beveridge, Chilton and Spratt2002). It was therefore tentatively suggested that the assemblages might have arisen through host switching.
Given the difficulties encountered in the use of morphological characters to establish phylogenetic relationships among nematodes, it was decided to re-examine the problem using molecular methods. The combined first and second internal transcribed spacers (ITS-1 and ITS-2) of nuclear ribosomal DNA represent ideal markers as they have previously been used successfully to establish the phylogenetic relationships among some strongylid nematodes (Hung et al. Reference Hung, Chilton, Beveridge and Gasser2000; Gouÿ de Bellocq et al. Reference Gouÿ de Bellocq, Ferté, Depaquit, Justine, Tiller and Durette-Desset2001).
In the present study, we focus on macropodid host species which harbour large numbers of co-occurring nematode species. Based on the preliminary work of Beveridge et al. (Reference Beveridge, Chilton and Spratt2002), these hosts were the swamp wallaby, Wallabia bicolor, the black-stripe wallaby, Macropus (Notamacropus) dorsalis, the wallaroo, M. (Osphranter) robustus (and its close, sympatric relative M. (O.) antilopinus), the quokka, Setonix brachyurus and the rock wallabies, Petrogale spp. Other host species such as the red kangaroo, Macropus (O.) rufus, the red-legged pademelon, Thylogale stigmatica, the whiptail wallaby, M. (N.) parryi, the agile wallaby, M. (N.) agilis and the tammar wallaby, M. (N.) eugenii, were also included as sequence data were also available for species of Cloacina found in them.
MATERIALS AND METHODS
Nematodes were obtained from the stomachs of kangaroos and wallabies which had been shot commercially, collected as fresh road-kills or from road-kills frozen prior to examination. Nematodes were washed in saline and then frozen in liquid nitrogen and stored at −80° prior to examination. Additional samples of nematodes from each host were fixed in Berland's fluid (glacial acetic acid and formalin; Gibson, Reference Gibson1979) for morphological examination.
Frozen nematodes were thawed, the head and tail were removed from individuals, fixed in lactophenol and mounted permanently in polyvinyl lactophenol as voucher specimens, with the mid-body region being used for genetic analyses. Nematodes were identified following Beveridge (Reference Beveridge1998, Reference Beveridge1999) and Beveridge et al. (Reference Beveridge, Jabbar and Shuttleworth2014a , Reference Beveridge, Ngyuen, Nyein, Cheng, Koehler, Shuttleworth, Gasser and Jabbar b ). Voucher specimens (hologenophores) have been deposited in the South Australian Museum (SAM), Adelaide (Table 1). In some instances, the unique specimen used for genetic studies (the hologenophore) was not preserved. In these instances, fixed specimens of the same species from the same host individual (paragenophores) have been deposited in SAM (Table 1). Where possible, registration data for both hologenophores and paragenophores have been included. Codes for slide numbers and/or host identifications included in Table 1 correspond to entries in the SAM database.
Species of Cloacina and Arundelia included in this study with collection details, morphological voucher numbers (paragenophores in parentheses) and molecular sequence registration numbers
Genomic DNA was isolated from the remaining part of each nematode using a small-scale sodium-dodecyl-sulphate/proteinase K extraction procedure (Gasser et al. Reference Gasser, Chilton, Hoste and Beveridge1993), followed by purification using a mini-column (Wizard™ Clean-Up, Promega). The region of rDNA comprising the ITS-1, 5·8S rRNA gene, ITS-2 and flanking sequences (=ITS+) was amplified by the polymerase chain reaction (PCR) using primers NC16 (forward; 5′-AGTTCAATCGCAATGGCTT-3′) and NC2 (reverse; 5′-TTAGTTTCTTTTCCTCCGCT-3′). PCRs were performed in 50 µL volumes using the following conditions: 30 cycles at 94 °C for 30 s (denaturation), 55 °C for 30 s (annealing) and 72 °C for 30 s (extension), followed by one cycle at 72 °C for 5 min (final extension). Negative (no-DNA) controls were included in each set of reactions. Amplicons were purified using mini-columns (using Wizard™ PCR-Preps, Promega), and the ITS+ sequenced in both directions using the primers NC16 and NC2 in separate reactions. The sequences generated in the present study have been deposited in GenBank (Table 1).
Additional sequences already present in the GenBank database were utilized and are indicated in Table 1. In the case of published studies of the genetic variability within species of Cloacina (Shuttleworth et al. Reference Shuttleworth, Beveridge, Chilton, Koehler, Gasser and Jabbar2014, Reference Shuttleworth, Jabbar and Beveridge2016a , Reference Shuttleworth, Beveridge, Koehler, Gasser and Jabbar b ), a representative sequence from each species was selected. In these studies, although sequence variability was found within each species of Cloacina, individuals representing each species formed a distinctive clade. On this basis, additional species have been added as a single sequence. In instances in which the sequences available were from different host species (C loacina parva from Macropus (O.) robustus and Petrogale purpureicollis; C loacina phaedra from Macropus (N.) agilis and M. (N.) parryi), both sequences were included in the analyses. Due to the occurrence of hybridization between the related phascolostrongyline nematodes Paramacropostrongylus typicus and Paramacropostrongylus iugalis (Chilton et al. Reference Chilton, Beveridge and Andrews1997 Reference Chilton, Beveridge, Hoste and Gasser b ), all chromatograms were examined for the possible occurrence of hybrids within the genus.
Sequences were initially aligned using Muscle (Edgar, Reference Edgar2004) and alignments adjusted manually using the program Mesquite v.3.03 (Maddison and Maddison, Reference Maddison and Maddison2015). Analyses of sequence data were conducted by Bayesian inference (BI) using Monte Carlo Markov Chain analysis in the program MrBayes v.3.2.3 (Ronquist and Huelsenbeck, Reference Ronquist and Huelsenbeck2003). The likelihood parameters set for the BI analysis of sequence data were based on the Akaike Information Criteria test in jModeltest v.2.1.7 (Posada, Reference Posada2008). The alignment was partitioned into two datasets (ITS1 and ITS2). The number of substitutions was set at 6 (Nst = 6), with a gamma-distribution and a proportion of invariable sites. For the tree, posterior probability (pp) values were calculated by running 2 000 000 generations with four simultaneous tree-building chains. Trees were saved every 100th generation. At the end of each run, the standard deviation of split frequencies was <0·01, and the potential scale reduction factor approached one. For each analysis, a 50%-majority rule consensus tree was constructed based on the final 75% of trees produced by BI. Analyses were run three times to ensure convergence and insensitivity to priors. The ITS+ sequence of Arundelia dissimilis, a species within a related genus in the same tribe, Cloacininea, was used as the outgroup.
The parasite phylogeny was compared with a molecular phylogeny of the hosts based primarily on Meredith et al. (Reference Meredith, Westerman and Springer2008). As there is no comprehensive molecular phylogeny for the Macropodidae, any taxa missing from the latter study were interpolated based on the comprehensive dataset of Cardillo et al. (Reference Cardillo, Bininda-Edmonds, Boakes and Purvis2004) and the resulting tree is presented as a cladogram. For comparison with the host phylogeny, the parasite tree was also converted to a cladogram.
Macropus (O.) antilopinus and M. (O.) robustus share most of their parasites (Beveridge et al. Reference Beveridge, Chilton, Johnson, Smales, Speare and Spratt1998). Cloacina dirce and Cloacina dindymene from M. (O.) antilopinus, included in this study, are also both common parasites of M. (O.) robustus (Beveridge, Reference Beveridge1998). Similarly, the four species of Petrogale included in this study (Petrogale assimilis, Petrogale herberti, Petrogale persephone, Petrogale purpureicollis) have extremely similar helminth communities (Beveridge et al. Reference Beveridge, Spratt, Close, Barker and Sharman1989; Begg et al. Reference Begg, Beveridge, Chilton, Johnson and O'Callaghan1995) and the host genus is treated as a single entity in the results section. Cloacina robertsi occurs in all four species of rock wallaby, while Cloacina caenis and Cloacina pearsoni occur in all but P. persephone. Cloacina parva occurs in P. purpureicollis, but not in the remaining rock wallaby species (Beveridge, Reference Beveridge1998; Chilton et al. Reference Chilton, Huby-Chilton, Johnson, Beveridge and Gasser2009). Host nomenclature and distributions (in Fig. 1) follow van Dyck and Strahan (Reference van Dyck and Strahan2008).
Collecting localities for all host species examined and geographical ranges of the principal macropodid hosts included in this study: Macropus (O.) antilopinus, M. (N.) dorsalis, M. (O.) robustus, Petrogale spp., Setonix brachyurus and Wallabia bicolor. Collecting localities close to one another have been combined on the maps. The host species collected are indicated for each locality. A. Collecting localities: 1, Roebourne, Fortescue River Roadhouse (M. (O.) robustus); 2, Menzies (M. (O.) robustus); 3, Kalgoorlie (M. (O.) robustus); 4, Mulga Park Station (M. (O.) robustus, M. (O.) rufus); 5, Cloncurry (M. (O.) robustus); 6, Mount Surprise (M. (O.) antilopinus); 7, Kangaroo Hills Station (M. (O.) robustus); 8, Bowen (M. (N.) dorsalis); 9, Marlborough (M. (N.) dorsalis); 10, Rockhampton, Yeppoon (M. (N.) agilis, M. (N.) dorsalis, W. bicolor); 11, Warwick (M. (O.) robustus); 12, Kingstown (M. (O.) robustus); 13, Wollomombi (M. (O.) robustus). B. Collecting localities: 14, Wellington Dam (S. brachyurus); 15. Kangaroo Island (M. (N.) eugenii); 16, The Gurdies, Phillip Island (W. bicolor); 17, Healesville, Dixon's Creek (W. bicolor); 18, Bourke (M. (O.) rufus); 19 Miles (W. bicolor); 20, Darling Plains Station (M. (N.) parryi); 21, Mount Sebastopol (Petrogale herberti); 22, Sarina (Thylogale stigmatica); 23, Proserpine, Shute Harbour, Airlie Beach (P. persephone, W. bicolor); 24, Mount Louisa (P. assimilis); 25, Winton (P. purpureicollis); 26, Mount Isa, Mary Kathleen (P. purpureicollis).
In instances where sister species of Cloacina were identified within the same macropodid host species in the molecular phylogeny, additional data were sought to confirm whether the occurrences of the sister species were sympatric in order to provide evidence for or against the hypothesis that these were instances of within host speciation.
The ITS+ molecular tree was compared with the phenetic arrangement of taxa proposed by Beveridge (Reference Beveridge1998) in an attempt to correlate principal morphological features with the molecular phylogenetic data.
RESULTS
ITS+ sequence data were available for 59 species of Cloacina (Table 1). Of these, 25 are novel sequences while the remainder have been published previously and deposited on GenBank. Given that the Neighbour-Joining and BI trees constructed were similar in topology to one another, only the BI tree is presented here (Fig. 2). In the BI trees, multiple clades were identified with high posterior probabilities (>0·92).
Bayesian Inference (BI) phylogenetic tree of associations between species of Cloacina. Posterior probabilities >0·90 are shown at nodes. Clades identified are discussed in the text.
No evidence of hybrids was found among the nematodes included in the study.
The small basal clades in the phylogenetic tree consisted of a mixture of species found in W. bicolor (Cloacina annulata, Cloacina papillatissima and Cloacina galatea), M. (N.) dorsalis (Cloacina burnettiana and Cloacina polyxo) and T. stigmatica (Cloacina cloelia) (Fig. 2). In the remaining tree, three major clades were evident. One clade was relatively small containing five species from S. brachyurus (Cloacina cadmus, Cloacina ceres, Cloacina circe, Cloacina setonicis and Cloacina telemachus) together with single species (in a sister species relationship) from W. bicolor (Cloacina castor). A second clade consisted of 16 nematode species from multiple host species. Members of this clade characteristically exhibited long branches. A third clade consisted of 31 species, again from multiple host species but generally with shorter branch lengths than those seen in the second clade.
Of the principal host species included in the analysis (M. (N.) dorsalis, M. (O.) robustus (together with M. (O.) antilopinus), Petrogale spp., W. bicolor, S. brachyurus), that is those host species parasitised by several species of Cloacina, their nematode species were spread across multiple clades (Fig. 3). The 17 species of Cloacina from M. (O.) robustus (including M. (O.) antilopinus) were distributed across 11 clades, the nine species from M. (N.) dorsalis were distributed across four clades, the 14 species from W. bicolor were distributed across 11 clades, the eight species from S. brachyurus were distributed across two clades and the six species from Petrogale spp. were distributed across three clades (Fig. 4). Thus, none of the assemblages of Cloacina spp. in these macropodid hosts was monophyletic. Other host species included were represented by fewer parasite taxa, but in the case of species occurring in M. (O.) rufus and T. stigmatica, the two species of nematodes occurred in different clades (Fig. 2).
Cladogram of associations between species of Cloacina, with posterior probabilities >0·90 shown at nodes, on left, with the corresponding phylogram of host relationships, based on Meredith et al. (Reference Meredith, Westerman and Springer2008) but with interpolations of missing taxa based on Cardillo et al. (Reference Cardillo, Bininda-Edmonds, Boakes and Purvis2004) on the right. Clades a–d, f–k, indicate sister species occurring in the same macropodid host species; clade e indicates a series of morphologically related species.
Species of Cloacina from the principal hosts utilized in this study, the wallaroo, Macropus (O.) robustus (combined with M. (O.) antilopinus), the black-striped wallaby, M. (N.) dorsalis, the swamp wallaby, Wallabia bicolor, the quokka, Setonix brachyurus, and rock wallabies, Petrogale spp., showing the distribution of species from each host or host group by clade from Fig. 2. Clades identified by letters are the same as those shown in Fig. 3.
Sister species occurring in the same macropodid host species were identified in a number of instances (Figs 3 and 4). In M. (O.) robustus (included with (M. (O.) antilopinus), four groups of sister species were identified (clades c, d, g, h); in M. (N.) dorsalis, two groups of sister species were identified (clades a, k); in W. bicolor, two pairs of sister species were identified (clades f, i); Cloacina antigone and Cloacina io also formed a clade (Fig. 2) but with low statistical support; in S. brachyurus, one group of five sister species was identified (clade j) while in Petrogale spp., a single clade of four sister species was identified (clade b). Of the remaining host species included in the analysis, M. (O.) rufus had two species of Cloacina in different clades, while T. stigmatica also had two species of Cloacina in different clades. Other host species included were parasitised by a single species of Cloacina and comments on sister species relationships are consequently not relevant.
Specimens of Cloacina phaedra were included from two host species, M. (N). agilis and M. (N). parryi. However, no genetic differences were detected between these nematodes. Specimens of C. parva from P. purpureicollis and M. (O.) robustus showed some sequence differences but were closely related.
DISCUSSION
The principal question posed at the outset of this study was whether the assemblages of Cloacina species found in the different species of macropodid hosts represented monophyletic groups. Although the molecular tree presented herein for the 59 species of Cloacina is not considered to be a definitive phylogeny of the genus, it does provide insight into genetic associations among currently known taxa. The study of the nematodes of several host species harbouring multiple species of Cloacina presented here (M. (N.) dorsalis, M. (O.) robustus, Petrogale spp., W. bicolor) provides strong evidence that they do not each represent a monophyletic assemblage and that representatives of each assemblage belong to different clades within the genus. These data confirm the tentative suggestions of a previous morphological study that also suggested the assemblages of Cloacina spp. in macropodid hosts were not monophyletic (Beveridge et al. Reference Beveridge, Chilton and Spratt2002). Some caution may be needed in accepting this conclusion as only approximately 50% of species of this large nematode genus were included and extensive assemblages in the grey kangaroos (M. (M.) fuliginosus and M. (M.) giganteus) (sub-genus Macropus) as well as those from the New Guinea scrub wallabies (Dorcopsis spp.) (tribe Dorcopsini of Prideaux and Warburton, Reference Prideaux and Warburton2010), some of which clustered on a within-host basis in the preliminary study of Beveridge et al. (Reference Beveridge, Chilton and Spratt2002), remain to be examined.
As a consequence, based on the examples presented here, the assemblages of Cloacina spp. in kangaroos and wallabies do not appear to comply with the definition of a species flock based on monophyly, but, by contrast appear to represent assemblages in which host colonization has played a significant role in their evolution (Fig. 3). This outcome is consistent with previously published morphological studies (Beveridge and Chilton, Reference Beveridge and Chilton2001; Beveridge et al. Reference Beveridge, Chilton and Spratt2002) as well as recent molecular studies of three related cloacinine genera Cyclostrongylus, Rugopharynx and Pharyngostrongylus in which host colonization appears also to have played a major role in parasite evolution (Chilton et al. Reference Chilton, Huby-Chilton, Koehler, Gasser and Beveridge2016a , Reference Chilton, Huby-Chilton, Gasser, Koehler and Beveridge b , Reference Chilton, Huby-Chilton, Koehler, Gasser and Beveridge c ).
In a number of instances, pairs of sister species or multiple sister species were found in the same kangaroo or wallaby host species suggesting the possible occurrence of within-host speciation (Fig. 4). Within-host speciation needs to be distinguished from sympatric speciation as parasites may have diverged within allopatric populations of the same host species (McCoy, Reference McCoy2003). Consequently, in examining the sister-groups of species of Cloacina apparently exhibiting within-host speciation, their known geographical distribution and co-occurrence in individual hosts also need to be taken into consideration. Such data are available for the species of Cloacina (clade d) from M. (O.) robustus containing Cloacina epona, Cloacina feronia and Cloacina frequens (Shuttleworth et al. Reference Shuttleworth, Beveridge, Koehler, Gasser and Jabbar2016b ). M. (O.) robustus is distributed across virtually the entire Australian continent (Clancy and Croft, Reference Clancy, Croft, van Dyck and Strahan2008) (Fig. 1). In this instance, C. frequens exhibits a northern and western distribution while C. feronia exhibits a southern and eastern distribution, with overlap in central and Western Australia. Cloacina daveyi occurs in south and central Australia, overlapping with C. feronia and C. frequens. Cloacina epona, however, is restricted to north-western Queensland (Shuttleworth et al. Reference Shuttleworth, Beveridge, Koehler, Gasser and Jabbar2016b ). In the case of these species, it is possible that past allopatry of host populations has been involved in parasite speciation, although there are no phylogeographic studies of the host species currently available to test this hypothesis. By contrast, Cloacina johnstoni and Cloacina macropodis commonly co-occur in M. (O). r. erubescens, although C. macropodis has a wider geographical distribution than that of C. johnstoni, occurring as well in M. r. robustus, providing the possibility that they may have co-evolved in sympatry (Shuttleworth et al. Reference Shuttleworth, Jabbar and Beveridge2016a ).
The sister species pair C. dindymene and C. dirce are parasites of both M. (O.) robustus and M. (O.) antilopinus in northern Australia (Beveridge, Reference Beveridge1998). Their co-occurrence has been studied in only one part of the ranges of the two host species (i.e. in north-eastern Queensland) (Beveridge et al. Reference Beveridge, Chilton, Johnson, Smales, Speare and Spratt1998) but limited records indicate co-occurrence of these nematode species in the Northern Territory and the Kimberley region of Western Australia as well (Beveridge, Reference Beveridge1998). However, since two host species are involved, their interrelationships require additional investigation. Likewise, in clade g, C. typhon is a common parasite of M. (M.) giganteus as well as occurring at lower prevalences in other sympatric macropodids (Beveridge et al. Reference Beveridge, Chilton, Johnson, Smales, Speare and Spratt1998). Consequently, interpretations of the evolution of the C. typhon – C. ixion species pair may have involved host colonization.
Macropus (N.) dorsalis has a more restricted distribution in north-eastern Australia (Johnson, Reference Johnson, van Dyck and Strahan2008) (Fig. 1), but is host to two clades (a, k) of host-specific species, which have apparently evolved within the same host species. One additional species in clade a (C. clymene) is a parasite of M. (O.) robustus, suggesting a case of host colonisation. Studies of the co-occurrence of species of Cloacina in M. (N.) dorsalis are currently limited to the more northern parts of its range (Beveridge et al. Reference Beveridge, Chilton, Johnson, Smales, Speare and Spratt1998) with few additional records from more southern regions (Beveridge, Reference Beveridge1998). The current data suggest that these species have evolved within the host in sympatry, but additional collections are required to confirm this hypothesis.
In W. bicolor, two pairs of sister species (clades f, i) were identified. All four of these species occur throughout the geographical range of the host (Beveridge, Reference Beveridge2016) and consequently the possibility of allopatric differentiation of these species seems unlikely. They may represent instances of both within-host and sympatric speciation. However, the helminth community of W. bicolor exhibits distinct regional differentiation (Beveridge, Reference Beveridge2016), so that the possibility of allopatric speciation cannot be excluded.
The quokka, S. brachyurus, has an extremely limited distribution in Western Australia (de Torres, Reference de Torres, van Dyck and Strahan2008) (Fig. 1) and is phylogenetically basal to the remaining macropodid genera (Meredith et al. Reference Meredith, Westerman and Springer2008). In spite of this, it is host to five sister species (Clade j) suggesting instances of within-host and possibly sympatric speciation. The original descriptions of these species were from island populations (Beveridge, Reference Beveridge1999). The current molecular data are from mainland populations of the host in which the same nematode species were recovered. It seems unlikely that there are significant differences between island and mainland populations of these nematodes, an hypothesis supported by similar molecular studies of species of the related genus Rugopharynx from island and mainland populations of its hosts (Chilton et al. Reference Chilton, Huby-Chilton, Koehler, Gasser and Beveridge2016c ).
The rock wallabies, Petrogale spp., represent a distinctive clade within the Macropodidae (Meredith et al. Reference Meredith, Westerman and Springer2008) and their principal species of Cloacina, C. caenis, C. pearsoni and C. robertsi (Beveridge et al. Reference Beveridge, Spratt, Close, Barker and Sharman1989) form a distinctive clade (clade b) suggesting a degree of co-evolution with the host genus. However, because of the complexity of taxonomic relationships between species of Petrogale (Potter et al. Reference Potter, Cooper, Metcalfe, Taggart and Eldridge2012) and the possibility that each of the nematode species included in this study represents a species complex, with a different genetic form in each wallaby host species (Chilton et al. Reference Chilton, Huby-Chilton, Johnson, Beveridge and Gasser2009), any conclusions need to be guarded.
Attempts to correlate principal clades identified in the BI analysis with defining, autapomorphic characters as utilized by Beveridge (Reference Beveridge1998) were largely unsuccessful, suggesting that many of the morphological characters currently utilized for identification are homoplasious. This topic however is potentially the focus of an additional study. One clade (clade e) did however conform to the current phenetic classification with couplet 7 in the key to the genus (Beveridge, Reference Beveridge1998), which identified transverse folds in the lining of the anterior oesophagus as a taxonomic feature. This feature separated C. similis, C. communis, C. petronius, C. petrogale, C. phaeax and C. phaedra from congeners. All of these species are members of clade e, but occur in a wide range of host species, consistent with the proposal for speciation being primarily by host colonisation. Cloacina petrogale was shown to be a species complex using multilocus enzyme electrophoretic data (Chilton et al. Reference Chilton, Beveridge and Andrews1997a , Reference Chilton, Beveridge, Hoste and Gasser b ). The species was subsequently subdivided based on minor morphological differences by Beveridge (Reference Beveridge1998). The current molecular sequence data support both the electrophoretic data and the taxonomic decisions made on this basis. In this study, C. phaedra collected from both M. (N.) agilis and M. (N.) parryi was shown to be identical and therefore to be the same species. This species was found to occur in M. (N.) parryi at a prevalence of 36% but was found in only a single individual of M. (N.) agilis (3%) (Beveridge et al. Reference Beveridge, Chilton, Johnson, Smales, Speare and Spratt1998), suggesting that M. (N.) parryi was the primary host but that transfer to a sympatric host, M. (N.) agilis, was possible.
Cloacina parva was obtained from M. (O.) robustus and P. purpureicollis, with slight differences in the sequence data. Although originally described from M. (O.) robustus by Johnston and Mawson (Reference Johnston and Mawson1938) and is a common parasite of this host species (prevalence 51%, Beveridge et al. Reference Beveridge, Chilton, Johnson, Smales, Speare and Spratt1998), it is also common in Petrogale spp. (Beveridge et al. Reference Beveridge, Spratt, Close, Barker and Sharman1989; Bradley et al. Reference Bradley, Beveridge, Chilton and Johnson2000). The data presented here suggest that it originated in a clade of Petrogale specific taxa and therefore its occurrence in M. (O.) robustus is a colonization event. This hypothesis requires more detailed examination.
Speciation within Cloacina appears to have been primarily by host colonization but with additional examples of within-host speciation. The proposition of host colonization as a principal mode of speciation is not novel within this subfamily (Beveridge and Chilton, Reference Beveridge and Chilton2001 (morphological analyses); Chilton et al. Reference Chilton, Huby-Chilton, Koehler, Gasser and Beveridge2016a , Reference Chilton, Huby-Chilton, Gasser, Koehler and Beveridge b , Reference Chilton, Huby-Chilton, Koehler, Gasser and Beveridge c (molecular analyses)). The reasons for this mode of evolution probably relate to the life-cycles of the parasites and the ecology of their hosts (Beveridge and Spratt, Reference Beveridge and Spratt1996). The life-cycles of the parasites are presumed to be direct based primarily on the study of Labiosimplex eugenii by Smales (Reference Smales1977). Third stage larvae develop in the external environment and are ingested on herbage. As multiple species of macropodids may graze in the same environment (Jarman and Phillips, Reference Jarman, Phillips, Grigg, Jarman and Hume1989) (Fig. 1), ingestion of larvae deposited by a related species of macropodids is likely. An instance of this phenomenon in the present study is the occurrence of C. phaedra in its principal host M. (N.) parryi and its occurrence in a sympatric host species, M. (N.) agilis, but at a much lower prevalence. Such host transfers provide the basis for colonization of a new host species. A second factor identified by Beveridge and Spratt (Reference Beveridge and Spratt1996) was the voluminous forestomachs of macropodids, the site which these nematodes inhabit and therefore the lack of potential competition in occupying a novel niche. Hoste and Beveridge (Reference Hoste and Beveridge1993) were unable to establish any evidence of competition between the nematode species present in the forestomachs of the macropodid species, which they studied. Consequently, the system of multiple macropodid species grazing in the same environment, the direct life cycle of the parasites and the apparent lack of competition in the site of establishment in the host are likely to facilitate host colonization.
The examples presented here are of within-host speciation warrant additional scrutiny as possible examples of sympatric speciation. However, in order to establish that sympatric speciation has occurred, it is necessary to show that the initial stages of divergence occurred in sympatry (McCoy, Reference McCoy2003). Thus the possibility that a past host transfer from a now extinct host species is difficult to exclude. In the case of equids and rhinoceroses, the extant species represent relics of a much more diverse fauna present during the Pleistocene (2 million years ago) (Franzen, Reference Franzen2010) and similarly, the extant macropodids represent a fraction of the species that existed previously (Prideaux and Warburton, Reference Prideaux and Warburton2010). Consequently, establishing that macropodid nematode species evolved in sympatry represents a challenge. In addition, the phylogeographical history of most extant macropodids remains unknown.
Molecular phylogeographic data exist for some macropodid species (M. (O.) rufus, M. (M.) fuliginosus and M. (M.) giganteus (Clegg et al. Reference Clegg, Hale and Moritz1998; Neaves et al. Reference Neaves, Zenger, Prince, Eldridge and Cooper2009; Coghlan et al. Reference Coghlan, Goldizen, Thomson and Seddon2015)) but not for the host species included in this study. Evidence which may potentially support a hypothesis of sympatric speciation could be the demonstration of niche separation within the macropodid forestomach. While some degree of niche separation has been demonstrated among Cloacina spp. occurring in the forestomach of M. (M.) fuliginosus (Pamment et al. Reference Pamment, Beveridge and Gasser1994) there are no comparable studies on the parasites of the hosts included in the current molecular study.
The data presented here on host- switching and within-host speciation in the evolution of complex parasite communities of nematodes are mirrored in comparable studies on the monogenean parasites of fish (Šimková et al. Reference Šimková, Desdevises, Gelnar and Morand2000, Reference Šimková, Gelnar and Morand2001, Reference Šimková, Morand, Jobet, Gelnar and Verneau2004; Huyse and Volckaert, Reference Huyse and Volckaert2005) suggesting that these patterns may be common among the various groups of parasitic helminths.
ACKNOWLEDGEMENTS
Thanks are due to all of our collaborators who assisted in the collection of specimens used in this study, including Andrew Doube, Peter Johnson, Richard Norman, Shane Middleton, Lesley Warner and Ross Andrews. Specimens were collected under the following state-issued permits: Queensland National Parks and Wildlife Service (T00436, T1131), the Wildlife Service and the South Australian Department of Environment and Heritage (EO7358), the Northern Territory Department of Primary Industry (15747) and the Western Australian Department of the Environment and Conservation (SF007407).
FINANCIAL SUPPORT
Collecting was supported financially by the Australian Research Council and the Australian Biological Resources Study.
