Do syntopic host species harbour similar symbiotic communities? The case of Chaetopterus spp. (Annelida: Chaetopteridae)

Hosts characteristics

The two Vietnamese host species of Chaetopterus are morphologically similar. However, one of them is significantly bigger than the other, both in terms of tube length (1.4:1, on average) and body volume (2.7:1, on average) (t-test p < 0.0001, Table 3). They also differ in the number of chaetigers of region A (9 and 9–11, respectively) (Figs. 2A and 2D) and in the denticles of the neuropodial uncini of region C (25–35 and 9, respectively) (Figs. 3A–3D), as well as in tube structure. Tubes of the bigger species are covered by silt, have a parchment-like appearance and the inner lining is iridescent, silver or golden in colour, showing distinct transverse annulations (Fig. 2C). In the smaller species, tubes are covered by sand and small coral and shell fragments, have a paper-like appearance with a semi-transparent inner lining, whitish or brownish in colour and lacking distinct annulations (Fig. 2F).

We tentatively identified the bigger host species as Chaetopterus cf. appendiculatus because, according to the original description, this species has a sand-coloured inner tube surface, showing dense transverse annulations. Additionally, it is the only described species of Chaetopterus possessing neuropodial uncini from region C with more than 20 small denticles. Chaetopterus appendiculatus was already reported as host of O. pettiboneae from the Banda Sea (Indonesia) by Petersen & Britayev (1997). Petersen (1997) proposed the redescription of C. appendiculatus as a valid species, based on the type material from Ceylon. The fact that formal redescription has never been published does not prevent us from considering the species as valid, whose formal redescription is far beyond the scope of this paper. However, the long geographical distance between Ceylon/Indonesia and Vietnam prevents us in fully assigning the Vietnamese specimens to C. appendiculatus, and we refer to the species as C. cf. appendiculatus in this paper. The smaller host is likely  undescribed.

The phylogenetic analysis including the COI sequences of the Vietnamese hosts (Fig.  4) showed low bootstrap values that did not allow us to fully resolve the phylogeny of Chaetopterus. However, it clearly revealed that the two Vietnamese hosts are different species included within two separate monophyletic clades (with 100% bootstrap support), thus confirming our morphological inference. Although with low support, the closest clades to those of the two Vietnamese Chaetopterus belong to C. variopedatus (Renier, 1804). However, the specimens joining the Chaetopterus sp. clade (42% bootstrap support) originate from the Mediterranean, while those joining the C. cf. appendiculatus clade (54% bootstrap support) originate from the Atlantic. As indicated by Martin et al. (2008), our results support the inference that the two populations of C. variopedatus belong to different species, with the Mediterranean species described and the Atlantic species still undescribed. The results also confirm that C. variopedatus sensu Hartman (1959) is not a single cosmopolitan species, but a complex including more than 20 different species (Bhaud, 1998; Osborn et al., 2007; Petersen, 1984a; Petersen, 1984b; Petersen, 1997). As is the case for C. appendiculatus, some of these species have not yet been formally redescribed. However, as many as nine species have recently been described, and five have been redescribed in the recent literature (Nishi, 2001; Nishi, Arai & Sasanuma , 2000; Nishi, Hickman Jr & Bailey-Brock, 2009; Osborn et al., 2007; Sun & Qiu, 2014).

The two Vietnamese Chaetopterus host species were found at the same localities, with their tubes deeply embedded in silty sand sediments. Chaetopterus cf. appendiculatus outnumbered Chaetopterus sp. in all samples, yet their proportion varied depending on the locality, with Chaetopterus sp. being relatively more abundant at Mot Islands (St. 2, 25.9%) and substantially less abundant at the other stations (5.0–9.1%) (Table 1).

The density of Chaetopterus ranged from 0.6 to 1.0 individuals per 100 m² in the transects, while the number of collected worms per diving hour was lower at St. 4 in Dam Bay and higher at St. 1 in Point Nam, Tre Island (Table 1).

Taxonomic composition of the symbiotic communities

91 individuals of seven species of animals occurred in association with the two host species of Chaetopterus. Among them, the polynoid polychaete Ophthalmonoe pettiboneae (Fig. 5C), the tergipedid nudibranch Phestilla sp. (Fig. 6G), the carapid fish Onuxodon fowleri (Smith, 1964) (Fig. 5D), and four species of decapods, three porcellanids, Eulenaios cometes (Walker, 1887) (Figs. 5A and 5B), Polyonyx cf. heok (Osawa & Ng, 2016) (Figs. 6A and 6B) and Polyonyx sp. (Figs. 6E and 6F), and the pinnotherid Tetrias sp. (Figs. 6C and 6D) (Table 4).

Four and three species were found inside the tubes of Chaetopterus sp. and C. cf. appendiculatus, respectively. Surprisingly, the symbiotic communities associated with the two hosts did not have any species in common, with the only similarity at a higher taxonomic level being the presence of porcellanid crabs (Table 4). Despite the lower sample size of Chaetopterus sp., the diversity of its associated community was higher than that of C. cf. appendiculatus. Accordingly, it may be expected that the number of species associated with Chaetopterus sp. would increase with an increasing number of analysed host individuals. Conversely, the diversity of the community associated with C. cf. appendiculatus showed an almost saturated species accumulation curve (Fig. 7A).

Ophthlmonoe pettiboneae is the single symbiotic species previously known from Vietnamese waters and from the same host species. The other six are here reported for the first time from the Vietnamese coasts. Moreover, O. fowleri is herein reported as a symbiont of chaetopterids for the first time, as well as Tetrias sp., Polyonyx sp., P. cf. heok and Phestilla sp., which are new to science and will be described at later dates in specialized papers. The tergipedid nudibranch is also, to the best of our knowledge, the first known nudibranch living in symbiosis with a polychaete host. It shows a posterior end functioning as a sucker (Fig. 6F) allowing it to attach to the smooth inner surface of the host tube, while the rest of the body can move freely (https://www.researchgate.net/publication/310159685_Phestilla_sp). Its flattened body, together with the lack of cnidosacs and the uniserial radula with long lateral denticles on the rachidian tooth clearly place it within the genus Phestilla. However, it differs from all known species of this genus by having a small central denticle of the radula, a wider foot and cerata arranged one per row only (Y Deart & I Ekimova, pers. comm., 2016). Moreover, its appearance and colouring (Fig. 6F) mimics, to some extent, that of the very posterior end of the chaetopterid host.

Structure of the symbiotic communities

A total of 61 (73.5%) out of 83 individuals of C. appediculatus and all 11 (100%) Chaetopterus sp. were infested. Among the species associated with C. appediculatus, O. pettiboneae showed a higher prevalence than the two other symbionts (Table 4). Among the associates with Chaetopterus sp., the prevalence ranged from 11% to almost 90%, with the maximum corresponding to P. cf. heok (Table 4).

The number of species inhabiting the same tube varied from zero to two in C. cf. appendiculatus, and from one to two in Chaetopterus sp. However, the species richness was significantly higher in the latter (p > 0.001, Figs. 7A and 7B) due to the common coexistence of two symbiont species in the same host tube. In fact, the small-sized Polyonyx sp. and Phestilla sp. (Figs. 6E and 6G) were found in all observed cases living with other crab species, mostly with the large sized P. cf. heok (Figs. 6A and 6B). Only in one case, Polyonyx sp. shared the host tube with Tetrias sp. (Table 5). In contrast, most tubes of C. appediculatus were occupied by a single symbiotic species, either O. pettiboneae or O. fowleri. Only in one case two species of symbionts (O. pettiboneae and E. cometes) were present sharing the same host tube (Table 5).

The number of symbiont individuals infesting one host varied from zero to three in C. cf. appediculatus, and from one to five in Chaetopterus sp., while the mean intensity was nearly three times lower in the former than in the later (p > 0.001, Table 6). Accordingly, most C. cf. appediculatus were infested by one symbiotic individual, while multiple infestation (two, three, or even four symbionts) was common in Chaetopterus sp. (Fig. 7B).

The distribution pattern of the symbionts among their hosts was primarily regular, as all. O. pettiboneae and most O. fowleri lived solitary inside their host tubes. In turn, Polyonyx cf. heok, Polyonyx sp., E. cometes, Tetrias sp. were found in host tubes usually as male/female pair. The number of nudibranchs varied from one to three per hosts (Table 6) and, in one case, a couple was observed near to a recently spawned egg-mass attached to the inner side of the host tube (Fig. 6H).

The component communities differed also in the relative abundance of a particular species. In the community associated with C. cf. appendiculatus, O. pettiboneae was dominant in terms of both prevalence and abundance. In the community associated with Chaetopterus sp., the dominance of the most abundant symbiont, P. cf. heok is less distinctive, with the role that other species had in the community structure being more relevant (Fig. 7C).

The average length of infested and non-infested tubes of C. appediculatus does not differ significantly (41.2–42.3, t-test, p = 0.65). The number of both symbiotic species and individuals do not show any significant correlation with host tube length in both chaetopterid species.

Community dissimilarity

Our results demonstrate a strict segregation in species composition of the communities associated with the two Vietnamese syntopic species of Chaetopterus, which had no species in common. However, at higher taxonomic levels (i.e., family, order and class), they were similar to each other and also resembled the symbiotic communities associated with other species of Chaetopterus and, even, echiuran worms in harbouring scale-worms, pocellanid and pinnotherid decapods and fishes (Anker et al., 2005; Gray, 1961; Ng & Sasekumar, 1993; Petersen & Britayev, 1997).

The two Vietnamese species of Chaetopterus are very similar in body morphology and tube shape, as well as in their trophic-functional characteristics. Thus, no reasons linked to host morphology were evident allowing us to explain the dissimilarity in symbiotic species composition. We suggest therefore that community composition appears to be determined by historical events rather than by the physical or biological habitat characteristics. We may hypothesize that both host species were originally allopatric. Thus, the evolutionary establishment of the respective specialized symbiotic associations would likely occur in different, separated geographical areas, with the posterior disappearance of geographical barriers leading to the current overlapping distribution. Once established, the respective symbiotic communities would be maintained by interspecific competition, leading to symbiont specialization to their respective host species as well as to preventing the exchange of symbionts between hosts when becoming sympatric, even being as closely related as is the case for these two species of Chaetopterus. However, our hypothesis does not exclude the possible existence of differences in host physiology or behaviour that would enhance the ability of the specialized symbionts to compete with possible invaders, thus contributing to maintain the differences in community composition.

Further assessment of this hypothesis would require an experimental approach to analyse the possible existence of a host-factor allowing the respective symbiont to recognize their own hosts, as well as to check the ability of the symbionts from one host to infest the other. In parallel, the regularity of the community segregation would have to be checked by more extensive field sampling addressed to discard (or reveal) the presence of additional symbionts on the alternative host species. This is particularly relevant for Chaetopterus sp. whose accumulation curve (Fig. 7A) supports an expected increase in the respective number of symbiotic species with sampling size. An additional, but not less pertinent question would be to assess the degree of specialization of the symbionts found in the two species of Chaetopterus, either based on previously published data or on our own observations. Therefore, it would be particularly relevant to consider whether they are obligatory or facultative and, in the case of obligatory symbionts, to further assess their degree of specificity (which may range from species-specific to opportunistic).

Symbionts’ specialization

Among the seven species of macroinvertebrates associated with C. cf. appendiculatus and Chaetopterus sp., four (i.e., one polychaete O. pettiboneae, one crab, E. cometes, and one fish, O. fowleri) are known as obligatory symbionts of chaetopterids and other benthic organisms. Onuxodon fowleri also lives in the mantle cavity of bivalves and inside holothurians (Markle & Olney, 1990; Parmentier, Chardon & Vanderwalle, 2002). In our samples, five of six individuals were juveniles, which allows us to suggest that they are employing C. cf. appendiculatus as temporal or intermediate hosts. The porcellanid E. cometes was reported from shallow waters off Singapore, living in association with a species of Chaetopterus, identified as C. variopedatus but this identification is undoubtedly incorrect (Ng & Nakasone, 1993). In turn, the scale-worm O. pettiboneae was first reported from Ambon Island (Indonesia) and later from the coasts of Vietnam, always in association with C. cf. appendiculatus (Britayev & Martin, 2005; Petersen & Britayev, 1997; this paper).

The four other species appear to be undescribed and are now being analysed by the corresponding specialists. However, we may infer some considerations on their degree of specialization based on existing papers dealing with the ecology and morphology of some closely related taxa. Concerning the symbiotic crabs, the porcellanid Polyonyx sp. belongs to the “Polyonyx sinensis” species complex, usually obligatory associates of tubicolous polychaetes, mainly with species of Chaetopterus (B Werding, 2016, unpublished data), while Polyonyx cf. heok belongs to the “Polyonyx pedalis” complex and the same or a very similar species has been recently reported from Singapore living in association with Chaetopterus cf. pacificus (Osawa & Ng, 2016). This suggests that both porcellanids are obligate and, probably, specialized symbionts of Chaetopterus sp. In turn, Tetrias sp. belongs to the Pinnotheridae, a family that mainly includes symbiotic species living as endo- or ectosymbionts in mollusc mantle cavities, polychaete burrows, echinoid integuments or tunicate branchial sacs (Drake et al., 2014). Among them, two species of Tetrias are currently known. Tetrias fischerii (Milne-Edwards, 1867) has been reported as symbiont of bivalves and annelids, while the host of Tetrias scabripes Rathbun, 1898 is unknown (Schmitt, McCain & Davidson, 1973). Although this cannot be assessed from our data, we suggest that the species associated with the Vietnamese Chaetopterus sp. is a specialized obligatory symbiont.

The third undescribed species, the nudibranch Phestilla sp., has several behavioural (i.e., two or more individuals sharing the same host, egg-masses attached to the inner tube surface) and morphological (i.e., posterior end working as a sucker, overall body shape mimicking that of the host) features clearly pointing toward a specialized symbiotic mode of life. This lead us to consider the species as the first know nudibranch living as a symbiont with a marine annelid host. Nudibranchs are well known as aposematic or mimetic organisms (Edmunds, 1987; Gosliner & Behrens, 1989; Rudman, 1991), some of them being considered as true symbionts. Among them, there are some species of Phestilla, which lives in association with corals and are highly specialized predators (Faucci, Toonen & Hadfield, 2007; Robertson, 1970), while the exact nature of the association of the Vietnamese Phestilla sp. and Chaetopterus sp. would need further research to be defined. The single related association occurred between the goniodorid nudibranch Lophodoris scala Marcus & Marcus, 1970 and the innkeeper echiurid Lissomyema exilii (Muüller, 1883). In this case, the nudibranch lives inside the host burrow, sometimes creeping along the host trunk and feeding, possibly exclusively, on Loxosomella spp., an entoproct that colonizes the burrow walls (Ditadi, 1982; Marcus & Marcus, 1970), which seems not to be the case for the Vietnamese species.

Accordingly, all symbionts found in association with Chaetopterus sp. and C. cf. appendiculatus have to be considered as obligatory symbionts. Among them, the less specialized is O. fowleri, which is known to infest hosts belonging to different types of animals (molluscs and polychaetes). The porcellanid crabs E. cometes and P. cf. heok are probably genus-specific symbionts, while the scale-worm O. pettiboneae, together with the other porcellanid crab Polyonyx sp. and the tergipedid nudibranch Phestilla sp., must be considered as species-specific symbionts. The specificity of the pinnotherid crab Tetrias sp. is not clear at this time. However, taking into account the relative abundance of pinnotherids among symbionts of Chaetopterus species (Petersen & Britayev, 1997; Schmitt, McCain & Davidson, 1973), we could also propose that it should be considered as a specialized symbiont, at least at family level.

Therefore, we consider all symbionts found in association with the two Vietnamese species of Chaetopterus as being, or tending to be, specialized symbionts, the single exception being the carapid fish.

Possible causes of observed differences in community structure

We found substantial differences in the structure of the symbiotic communities associated with Chaetopterus sp. and C. cf. appendiculatus. The first shows a significantly higher species richness and mean abundance than the second, while the second was clearly dominated by the presence of a single species, both in terms of abundance and frequency (Table 6, Fig. 6C). Taking into account that body size and tube length of Chaetopterus sp. are significantly lower than those of C. cf. appendiculatus, this situation is particularly unexpected. Usually, species richness and abundance increase with the increasing host size (e.g., Abele & Patton, 1976; Ribeiro, Omena & Muricy, 2003). Thus, the sitiuation of the Vietnamese partnership requires specific considerations.

We suggest that several factors are shaping the differences in the structure of the symbiotic communities associated with Chaetopterus sp. and C. cf. appendiculatus. Despite both host species having low population densities in Nhatrang Bay, that of Chaetopterus sp. was significantly lower, which would likely force the associated symbionts to use (and share) the few available hosts. This would possible explain the higher number of species in its associated community, as well as the fact that all host individuals of Chaetopterus sp. found in Nhatrang Bay harboured symbionts, in contrast to C. appenciculatus whose maximum prevalence was around 70%. Alternatively, the low density of both host populations may impede the secondary dispersion of the symbionts, which has been considered as a key mechanisms shaping the establishment and functioning of marine symbiotic communities (Mekhova et al., 2015) and raises the question on the adults’ ability of long-distance migration.

Based on species and individual’s distributions among hosts, we also hypothesized that another factor determining the observed differences in symbiotic community structure could be the existence of inter- and intraspecific competition. In fact, most tubes of Chaetopterus sp. were occupied by a minimum of two symbiotic species, this co-occurence being very rare in C. cf. appendiculatus, where each host individual was infested by one symbiotic species (Fig. 7B). The single exception was a host tube shared by O. pettiboneae and E. cometes. Accordingly, we suggest that the main driving factors may be resource partitioning between symbiotic species having different sizes, in the case of Chaetopterus sp., and strong interspecific interactions, in the case of C. cf. appendiculatus as previously reported for holothurian hosts (Lyskin & Britayev, 2005).

The characteristics of the symbiotic community structure associated with C. cf. appendiculatus, in which one host was usually occupied by one symbiotic species (Table 5), suggest the existence of interspecific competition between the polychaete and fish symbionts. In turn, the fact that there was a single symbiont per host (Table  5) supports the existence of intraspecific competition among polychaetes and fish individuals, respectively. At least for the polychaete, this hypothesis was supported by our direct observations in experimental aquaria, where individuals of O. pettiboneae were found to fight when trying to occupy the same host tube, as well as by the high frequency of body traumas present (TA Britayev & D Martin, 2016, unpublished data). In contrast, the bulk of Chaetopterus sp. symbionts were crabs (Fig. 7C). Territorial defence is a well-known phenomenon in symbiotic decapods too (Baeza, Stotz & Thiel, 2002; Huber, 1987; Vannini, 1985). However, their behaviour has a sexual component, as they often form heterosexual pairs consisting of gravid males and females co-inhabiting the same host (Castro, 2015; Patton, 1994), which was exactly the case of most porcellanid and pinnotherid crabs inhabiting the tubes of the two Vietnamese species of Chaetopterus (Table 5). This behaviour lead to a significant increase in the abundance of symbiotic individuals in the community associated with Chaetopterus sp. Therefore, while interspecific interactions seemed to affect both species richness and abundance, the intraspecific ones only affected the abundance.

Our observations support two main factors determining the structure of symbiotic communities associated with Chaetopterus sp. and C. cf. appendiculatus in Nhatrang Bay: the density of host populations and competition (both inter- and intraspecific). Moreover, the observed differences in community structure appear to be strictly related to the respective species composition. Accordingly, living in pairs and resource partitioning led to species coexisting in the tubes of Chaetopterus sp. and establish a tightly packed component community showing high species richness and mean intensity, together with a low species dominance. In contrast, the aggressive, strictly territorial species associated with C. cf. appendiculatus established a component community strongly dominated by host being inhabited by a single species and, thus, low species richness and mean intensity.

The existence of two closely related host species with overlapping distributions but harbouring very different symbiotic communities seems to be unusual. We suggest that it may probably be related with the scarcity of data currently available on the structure of symbiotic communities in marine environments. However, the situation is certainly not unique, as at least an additional example has been recently reported from Nhatrang Bay. In this case, the hosts were comatulid crinoids Comanthus gisleni Rowe, Hoggett, Birtles & Vail, 1986 and C. parvicirrus (Müller, 1841) (Mekhova & Britayev, 2012). Consequently, we expect further worldwide studies to discover more syntopic hosts harbouring symbiotic communities with contrasted composition and structure.