Morphological Modiﬁcations and Injuries of Corals Caused by Symbiotic Feather Duster Worms (Sabellidae) in the Caribbean

: Some coral-associated invertebrates are known for the negative impact they have on the health of their hosts. During biodiversity surveys on the coral reefs of Curaçao and a study of photo archives of Curaçao, Bonaire, and St. Eustatius, the Caribbean split-crown feather duster worm Anamobaea sp. (Sabellidae) was discovered as an associate of 27 stony coral species (Scleractinia spp. and Millepora spp.). The worm was also found in association with an encrusting octocoral ( Erythropodium caribaeorum ), a colonial tunicate ( Trididemnum solidum ), various sponge species, and thallose algae (mainly Lobophora sp.), each hypothesized to be secondary hosts. The worms were also common on dead coral. Sabellids of the genera Bispira and Sabellastarte were all found on dead coral. Some of them appeared to have settled next to live corals or on patches of dead coral skeleton surrounded by living coral tissue, forming pseudo-associations. Associated Anamobaea worms can cause distinct injuries in most host coral species and morphological deformities in a few of them. Since Anamobaea worms can form high densities, they have the potential to become a pest species on Caribbean coral reefs when environmental conditions become more favorable for them.


Introduction
As foundation species, reef corals provide a habitat to a large diversity of marine invertebrates, which represent a variety of phyla [1][2][3][4]. A large proportion of these invertebrate taxa use these corals as living hosts, whereas others only need dead coral as a rocky substrate for settlement and growth. The first category mostly contains species that live in strict symbiotic relations with their host corals (be it commensalistic, mutualistic, or parasitic) and are generally known as coral-associated fauna [5,6]. Due to their vulnerability to disturbance, their presence is supposed to be indicative of reef health [7,8]. These relations may vary because, in some studies, coral-associated species are reported as beneficial to their host by offering protection against predators and diseases [9][10][11][12][13] or cleaning services [14]. In other hosts, associated species are shown to be harmful by causing coral injuries or by obstructing the host's growth [15][16][17][18][19][20].
It is not precisely known if some reef-dwelling invertebrates, such as feather duster worms (fan worms) of the family Sabellidae, live in symbiosis with corals. Sabellids are tube-forming, solitary, or colonial sedentary polychaetes occurring in benthic environments. The protective tube is usually flexible and predominantly buried in sediment or attached to

Results
Twenty-seven host-coral species, consisting of 25 scleractinians (Anthozoa) and two milleporids (Hydrozoa), were recorded for the coral-associated feather duster worm, divided over 10 families and 16 genera (Table 1; Figures 2-7). In addition, the species was found in association with the encrusting octocoral Erythropodium caribaeorum ( Figure 8A,B), the colonial tunicate Trididemnum solidum ( Figure 8C,D), phaeophyceaen algae, in particular Lobophora sp. (Figure 8E,F), and various sponge species (Figure 9). The records were from the southern Caribbean islands of Bonaire and Curaçao and the Eastern Caribbean island of St. Eustatius (Table 1).
The symbiotic worms, identified as split-crown feather duster worms of the genus Anamobaea Krøyer, 1856 [22,27,29], showed some variation in coloration, ranging from white to dark red and various combination patterns of these colors (Figures 2, 3 and 9A). Two species from the Caribbean have been described, which can be distinguished by two morphological characters [27,29,32] that are not clearly visible in the photographs: Anamobaea phyllisae Tovar-Hernández & Salazar-Vallejo, 2006 has two dorsal kidney-shaped shields over the anterior margin of the base of its crown and smooth flanges (without papillae) and Anamobaea orstedi Krøyer, 1856 does not have such shields, and its flanges are wrinkled (with papillae). The former species has so far only been reported from the type locality in the British Virgin Islands, whereas the latter has a wider geographic range [29,32]. Because we are not sure about the identity of the associated worms, we refer to them as Anamobaea sp.
Most observed worms were withdrawn in their tubes; only a few of them were observed with extended radioles protruding from the tube (Figures 2 and 3). Some extended worms appeared to be shy and quickly retracted into the tube when their pictures were taken (Figure 3). On some occasions, the worms showed high densities, either inside a living host (Figures 2A,B,D, 4E and 6A) or on dead coral ( Figure 3C,D). Some host coral species showed peak-shaped deformities around the worm tubes ( Figure 4). In the foliaceous coral Agaricia lamarcki, the deformity resembles a sleeve that continues to grow upward and in thickness around the worm's tube, allowing the top to remain free ( Figure 4A,B). Peak-shaped deformities in various sizes were most abundantly found in Pseudodiploria strigosa ( Figure 4C-E) and less commonly in Orbicella annularis and O. franksi ( Figure 4F,G). When the largest peak found in P. strigosa ( Figure 4C) was removed, the worm tube appeared to be at least 8 cm long and deeply embedded inside the remaining part of the host coral ( Figure 5). The symbiotic worms, identified as split-crown feather duster worms of the genus Anamobaea Krøyer, 1856 [22,27,29], showed some variation in coloration, ranging from white to dark red and various combination patterns of these colors (Figures 2, 3 and 9A). Two species from the Caribbean have been described, which can be distinguished by two morphological characters [27,29,32] that are not clearly visible in the photographs: Anamobaea phyllisae Tovar-Hernández & Salazar-Vallejo, 2006 has two dorsal kidney-shaped shields over the anterior margin of the base of its crown and smooth flanges (without papillae) and Anamobaea orstedi Krøyer, 1856 does not have such shields, and its flanges are wrinkled (with papillae). The former species has so far only been reported from the type locality in the British Virgin Islands, whereas the latter has a wider geographic range [29,32]. Because we are not sure about the identity of the associated worms, we refer to them as Anamobaea sp.
Most observed worms were withdrawn in their tubes; only a few of them were observed with extended radioles protruding from the tube (Figures 2 and 3). Some extended worms appeared to be shy and quickly retracted into the tube when their pictures were taken ( Figure 3). On some occasions, the worms showed high densities, either inside a living host (Figures 2A,B,D, 4E and 6A) or on dead coral ( Figure 3C,D). Some host coral species showed peak-shaped deformities around the worm tubes ( Figure 4). In the foliaceous coral Agaricia lamarcki, the deformity resembles a sleeve that continues to grow upward and in thickness around the worm's tube, allowing the top to remain free ( Figure 4A,B). Peak-shaped deformities in various sizes were most abundantly found in Pseudodiploria strigosa ( Figure 4C-E) and less commonly in Orbicella annularis and O. franksi ( Figure 4F,G). When the largest peak found in P. strigosa ( Figure 4C) was removed, the worm tube appeared to be at least 8 cm long and deeply embedded inside the remaining part of the host coral ( Figure 5).
Coral injuries were abundant around worm tubes in various coral species (Figures 6  and 7). The wounds, visible as dead lesions, were either at the periphery of live coral tissue ( Figure 6A,B) or more toward the middle and surrounded by live coral tissue (Figures 6C,D and 7). Some dead patches were used as substrates by algae and sponges (Figures 6 and 7). In some coral species, the live tissue around the gash showed a discoloration, suggesting that it was spreading from the wound centered around the worm ( Figure 7C,D,F).       Orbicella franksi at St. Eustatius (2015) with a small worm peak (arrow), next to a serpulid Christmas tree worm (S. giganteus). The maximum width of each sabellid tube is ca. 5 mm.

Figure 5.
Tube of a split-crown feather duster (Anamobaea sp.) after removal of the peak-shaped deformation in a Pseudodiploria strigosa coral (see Figure 4C). The visible part of the tube is 8 cm long.
Feather duster worms of two other species were not observed inside living corals but in dead skeleton directly next to a living coral or in a patch of dead coral surrounded by healthy coral tissue. They are the magnificent feather duster Sabellastarte magnifica (Shaw, 1800) ( Figure 10) and the social feather duster Bispira brunnea (Treadwell, 1917) ( Figure 11). Sabellastarte magnifica was found in or next to live coral colonies of the corals Diploria labyrinthiformis, Madracis auretenra, Meandrina meandrites, Millepora alcicornis, Orbicella annularis, Pseudodiploria strigosa, and Stephanocoenia intersepta. Their tubes reached diameters of nearly 2 cm and could therefore be distinguished from the tubes of Anamobaea sp., which reached up to 0.5 cm in width. Bispira brunnea was only found on dead patches of Montastraea cavernosa and Orbicella annularis ( Figure 11). A published photograph from the Mexican Caribbean shows B. brunnea in a coral injury on top of a colony of Siderastrea siderea [54]. This worm species can be distinguished from the other two because it occurs as colonies instead of single individuals and because its tubes and radioles are much smaller than those of the others. Because all Bispira and Sabellastarte worms appeared to live on dead coral skeleton, near live coral, or at a distance, it is unclear whether they were symbionts or part of pseudo-associations. Figure 5. Tube of a split-crown feather duster (Anamobaea sp.) after removal of the peak-shaped deformation in a Pseudodiploria strigosa coral (see Figure 4C). The visible part of the tube is 8 cm long.
Feather duster worms of two other species were not observed inside living corals but in dead skeleton directly next to a living coral or in a patch of dead coral surrounded by healthy coral tissue. They are the magnificent feather duster Sabellastarte magnifica (Shaw, 1800) ( Figure 10) and the social feather duster Bispira brunnea (Treadwell, 1917) ( Figure 11). Sabellastarte magnifica was found in or next to live coral colonies of the corals Diploria labyrinthiformis, Madracis auretenra, Meandrina meandrites, Millepora alcicornis, Orbicella annularis, Pseudodiploria strigosa, and Stephanocoenia intersepta. Their tubes reached diameters of nearly 2 cm and could therefore be distinguished from the tubes of Anamobaea sp., which reached up to 0.5 cm in width. Bispira brunnea was only found on dead patches of Montastraea cavernosa and Orbicella annularis ( Figure 11). A published photograph from the Mexican Caribbean shows B. brunnea in a coral injury on top of a colony of Siderastrea siderea [54]. This worm species can be distinguished from the other two because it occurs as colonies instead of single individuals and because its tubes and radioles are much smaller than those of the others. Because all Bispira and Sabellastarte worms appeared to live on dead coral skeleton, near live coral, or at a distance, it is unclear whether they were symbionts or part of pseudo-associations.

Discussion
This report presents, for the first time, extensive evidence for the association of feather duster worms with corals, other sessile invertebrates, and algae in the Caribbean. This discovery is remarkable because of the strikingly large wounds and deformities inflicted by them on their host corals. Photographs of the worms indicate that these associations have been present at least since 2014 on the coral reefs of Curaçao (Southern Caribbean) and since 2015 at St. Eustatius (Eastern Caribbean). Prior to that, they may have remained unnoticed because of the worm's withdrawal behavior, because it was perhaps less abundant in the past, or because scientists studying the worms did not pay much attention to the hosts.        Coral deformities around sabellid worms embedded in the host's skeleton appear to be limited to a few scleractinian species of which Pseudodiploria strigosa appears to be the most common. Because coral-dwelling sabellids have been observed deep inside the coral skeleton, and the life span of sabellids may be over 10 years [55], these deformations have taken several years to develop. Morphological anomalies are not exceptional among corals inhabited by associated fauna. For example, the sabellid Perkinsiana anodina lives in short tube-shaped protuberances on the surface of an encrusting mushroom coral, which are part of the host's coral skeleton [32,51]. Coral gall crabs have received much attention because of the crescent-, canopy, slit-, and basket-shaped pits inside various coral species [44,[56][57][58][59]. Coral cysts and pits made by other crabs and by shrimps in stony corals have also been described [60,61], which should not be confused with gall-shaped excavations made by shrimps [62,63], bivalves [64][65][66][67], and gastropods [34,68]. Copepods are also known to induce the forming of dwellings in corals, either as galls [69,70] or as tubular outgrowths [71,72]. Ascothoracidan crustaceans of the genus Petrarca Fowler, 1889 are known to form conspicuous galls in shallow-water and deep-sea corals [73][74][75]. Diversity 2022, 14, x FOR PEER REVIEW 13 of 20

Discussion
This report presents, for the first time, extensive evidence for the association of feather duster worms with corals, other sessile invertebrates, and algae in the Caribbean. This discovery is remarkable because of the strikingly large wounds and deformities inflicted by them on their host corals. Photographs of the worms indicate that these associations have been present at least since 2014 on the coral reefs of Curaçao (Southern Caribbean) and since 2015 at St. Eustatius (Eastern Caribbean). Prior to that, they may have remained unnoticed because of the worm's withdrawal behavior, because it was perhaps less abundant in the past, or because scientists studying the worms did not pay much attention to the hosts.
Coral deformities around sabellid worms embedded in the host's skeleton appear to be limited to a few scleractinian species of which Pseudodiploria strigosa appears to be the most common. Because coral-dwelling sabellids have been observed deep inside the coral skeleton, and the life span of sabellids may be over 10 years [55], these deformations have taken several years to develop. Morphological anomalies are not exceptional among corals inhabited by associated fauna. For example, the sabellid Perkinsiana anodina lives in short tube-shaped protuberances on the surface of an encrusting mushroom coral, which are part of the host's coral skeleton [32,51]. Coral gall crabs have received much attention because of the crescent-, canopy, slit-, and basket-shaped pits inside various coral species [44,[56][57][58][59]. Coral cysts and pits made by other crabs and by shrimps in stony corals have also been described [60,61], which should not be confused with gall-shaped excavations made by shrimps [62,63], bivalves [64][65][66][67], and gastropods [34,68]. Copepods are also known to induce the forming of dwellings in corals, either as galls [69,70] or as tubular outgrowths [71,72]. Ascothoracidan crustaceans of the genus Petrarca Fowler, 1889 are known to form conspicuous galls in shallow-water and deep-sea corals [73][74][75].
Tube-dwelling gammarid amphipods and chaetopterid polychaete worms have been reported to induce the forming of densely distributed finger-like structures in Montipora corals [76,77]. Coral barnacles usually become embedded in the coral skeleton and become partly overgrown by coral tissue [44,60]. Some alterations in the coral skeleton morphology are microscopic and hardly visible, such as those caused by coral-dwelling hydroids of the genus Zanclea [78][79][80]. In contrast, vermetid snails that live inside branching Stylophora and massive Porites corals are known to modify the host's morphology on a larger scale by flattening its surface relief, which is attributed to growth inhibition caused by the snail's toxic mucus webs [81][82][83]. In contrast, large growth alterations in massive, branching, and encrusting corals consisting of deep fissures can be Tube-dwelling gammarid amphipods and chaetopterid polychaete worms have been reported to induce the forming of densely distributed finger-like structures in Montipora corals [76,77]. Coral barnacles usually become embedded in the coral skeleton and become partly overgrown by coral tissue [44,60]. Some alterations in the coral skeleton morphology are microscopic and hardly visible, such as those caused by coral-dwelling hydroids of the genus Zanclea [78][79][80]. In contrast, vermetid snails that live inside branching Stylophora and massive Porites corals are known to modify the host's morphology on a larger scale by flattening its surface relief, which is attributed to growth inhibition caused by the snail's toxic mucus webs [81][82][83]. In contrast, large growth alterations in massive, branching, and encrusting corals consisting of deep fissures can be formed by Pedum scallops embedded in corals [34,66,67]. Some aggressive coral-dwelling sponges are not considered long-term associated fauna because they usually tend to overgrow and kill their hosts, but in some foliose corals, they evoke a morphological response, which is visible as the growth of flap-like protrusions that overlap the approaching sponges [84]. A modified morphology is also seen in other foliose corals that overgrow sponges as if the coral shape is molded by that of the sponge [85]. All these examples indicate that some corals may adapt their shape to resist the presence of potentially harmful associated fauna or competitors for space.
Coral injuries caused by feather duster worms have not been reported before. These appear to be much larger than those caused by coral-dwelling Spirobranchus worms [18,19,86]. On the other hand, large densities of serpulid worms overgrowing live coral may eventually cause partial coral mortality [87]. Many feather duster worms in the present study were found on dead coral ( Figure 3A,B), and some of them formed clusters ( Figure 3C,D). In some cases, the worm-infested dead-coral area was next to live coral, suggesting that the worms contributed to partial coral mortality ( Figures 3E,F and 5A). A few patches of dead coral were surrounded by discolored live-coral tissue ( Figure 6C,D,F). This may represent a reaction to stress as seen in some massive Porites corals in which polyps in contact with algae or epifauna show pink or purple pigmentation [18,66,[88][89][90]. The difference is that there may not be extra pigmentation in the examples of the present study.
Some split-crown feather dusters at Curaçao were not hosted by stony corals but by other invertebrates. These invertebrates may have either colonized dead coral or overgrown living corals and became secondary hosts when the worms were able to resist becoming overgrown as well. The last scenario has been shown by serpulid Christmas tree worms of the genus Spirobranchus. The encrusting octocoral Erythropodium caribaeorum is recognized as an aggressive competitor for space in the Caribbean [91], which is able to overgrow corals but apparently not their symbiotic Spirobranchus worms [92], similar to some feather duster worms at Curaçao in the present study ( Figure 8A,B). Similarly, the colonial tunicate Trididemnum solidum is notorious for overgrowing Caribbean corals [93,94], except for their associated Spirobranchus [95] and seemingly also individuals of symbiotic Anamobaea sp. (Figure 8C,D). Sponges are also able to overgrow corals with the exception of symbiotic Spirobranchus [96,97] and apparently also Anamobaea sp. (Figure 9). The feather duster worm was also observed in association with algae, in particular the brown algae Lobophora sp. (Figure 8E,F). Lobophora has increased in abundance over the last decades at Curaçao and is able to overgrow live coral [98,99]. It is likely that it is able to overgrow dead and live coral containing Anamobaea sp., but apparently the worm tubes protrude too far to become outcompeted.
The cause of the injurious effect of the feather duster worms is unclear. The size of the wounds suggests that the worms produce toxins, but there is limited information on toxicity as a defense mechanism in Sabellidae [100]. The use of toxins can perhaps prevent worms from becoming overgrown by their hosts, as seen in Pedum scallops [66]. The mucus secreted by some sabellid species proves to have antibacterial properties [101][102][103]. According to a recent review paper on polychaete toxins, no relevant information appears to be available on the negative effect of sabellid mucus on other organisms [104]. In contrast, coral-dwelling worm snails, which occupy the same ecological niche as the feather duster worms of the present study [20], are well known for their venomous mucus and the damage this may inflict on the host corals [105,106].
Unlike Anamobaea sp., the relation of Sabellastarte magnifica and Bispira brunnea to corals is unclear because they were never found in living coral tissue (Figures 10 and 11). A close proximity to live corals shown by these two species may be unusual since they were also commonly found at a distance from live corals. Therefore, it may be more appropriate to use the term "pseudo-association" for this kind of unclear relation. On the other hand, it is also possible that these worms cause damage to corals and are responsible for coral mortality in their proximity.
The present study shows that the Caribbean feather duster worm Anamobaea sp. is more common and harmful to corals than previously known. The species has a symbiotic relation with a large range of corals and other invertebrates, which was also unknown before. It is unclear if the species has increased in abundance recently. Because this worm has the potential to become a pest species, future research should focus on its population dynamics, its settling behavior on live corals (as done with larvae of symbiotic barnacles [107,108]), and the cause, growth, and extension of coral wounds around its tubes. The larval settlement behavior of S. magnifica and B. brunnea also needs to be investigated in order to find out whether these species prefer to live in close proximity to corals or not.