Sponge diversity in Eastern Tropical Pacific coral reefs: an interoceanic comparison

Sponges are an important component of coral reef communities. The present study is the first devoted exclusively to coral reef sponges from Eastern Tropical Pacific (ETP). Eighty-seven species were found, with assemblages dominated by very small cryptic patches and boring sponges such as Cliona vermifera; the most common species in ETP reefs. We compared the sponge patterns from ETP reefs, Caribbean reefs (CR) and West Pacific reefs (WPR), and all have in common that very few species dominate the sponge assemblages. However, they are massive or large sun exposed sponges in CR and WPR, and small encrusting and boring cryptic species in ETP. At a similar depth, CR and WPR had seven times more individuals per m2, and between four (CR) and five times (WPR) more species per m2 than ETP. Perturbation, at local and large scale, rather than biological factors, seems to explain the low prevalence and characteristics of sponge assemblages in ETP reefs, which are very frequently located in shallow water where excessive turbulence, abrasion and high levels of damaging light occur. Other factors such as the recurrence of large-scale phenomena (mainly El Niño events), age of the reef (younger in ETP), isolation (higher in ETP), difficulty to gain recruits from distant areas (higher in ETP), are responsible for shaping ETP sponge communities. Such great differences in sponge fauna between the three basins might have consequences for coral reef structure and dynamics.

Qualitative sampling. All studied ETP reefs were between 1 and 6 m depth. In each one, the sampling was undertaken by a 2 h random dive 55 , during which three divers searched for sponges in different areas of the reef, both exposed and cryptic habitats, which included the lower surfaces of live or dead corals, interstices of coral framework, and loose heads which were overturned and examined.
Fragments of specimens that we were not able to identify "in situ" were collected, fixed and preserved in 70% ethanol. Spicules were cleaned in boiling nitric acid followed by water rinse and dehydration in alcohol, then dried on a microscope slide or circular cover slip for SEM (scanning electron microscopy). Spicule measurements (30 for each type) were made by light microscopy.
Sampling in cryptic spaces is difficult and time consuming, and the differentiation of individuals of sponges to estimate biomass is virtually impossible. However, the simple presence/absence is enough for species richness [56][57][58] Therefore, similarity among reefs was established by means of a classification analysis, using species as variables. The similarity matrix for the classification was calculated by means of the Sørensen index based on presence/absence 59 . The results were then graphically described using dendrograms with the UPGMA (unweighted pair-group method using centroids) aggregation algorithm 60 . Chondrilla pacifica X X 2 10 Chondrosia tenochca X 4 20 Cinacyrella sp. X 1 5 Cladocroce reina X 2 10 Cliona amplicavata X X 12 60 Cliona californiana X X 2 10 Cliona euryphylle X X 6 30 Cliona flavifodina X X X 5 25 Cliona medinae X 2 10 Cliona microstrongylata X 2 10 Cliona mucronata X X X 10 50 Cliona pocillopora X X X 9 45 Cliona raromicrosclera X 1 5 Cliona sp. X 3 15 Cliona tropicalis X X X 14 70 Cliona vallartense X X 3 15 Cliona vermifera X X X 18 90 Cliothosa tylostrongilata X 1 5 www.nature.com/scientificreports www.nature.com/scientificreports/ Multidimensional scaling and ordination were used to detect community patterns, using the PRIMER (v 6.1.11) software program 61 , and a two-dimensional non-metric Multidimensional Scaling (MDS), based on the Sørensen similarity matrix, was used to visualize community patterns (Fig. 3). The adequacy of the MDS was assessed through the stress coefficient, which should be <0.15 in order to minimize misinterpretations 61 .
Quantitative sampling. A quantitative sampling (density and species per square meter) was undertaken only in two reefs, because is a very difficult and time-consuming process. Besides, it's also necessary brakes large coral pieces to detect cryptic sponges. However, the data can be considered as representative for the whole Mexican Pacific reefs. A previous study in Panama (Pacific side) also reached the same conclusion 62 . For that, five transects 18 m long were set up, and six quadrats of 1 m 2 were placed along each one, resulting in a sampling area of 6 m 2 per transect (a total of 30 m 2 per reef). Density (ind. m −2 ) was estimated by counting all patches found inside each quadrant and later average per square meter. Species richness was estimated in the total of the sampling area (30 m 2 ), and later average by square meter. In the case of boring sponges, their appearance in the samples was quantified as a unique patch due to the difficulty to differentiate among individuals.

Data from Caribbean (CR) and West Pacific reefs (WPR).
In order to compare the information gained from this study with those from CR and WPR, an exhaustive research of the literature was done. All the papers about coral reef sponges with information about number of species, abundance (density), diversity (species per surface unit) were utilized to obtain mean values per basin and depth (see for example Fig. 4).
The most diverse family was Clionaidae, which contained four genera and 18 species, and the most common genus was Cliona with 13 species. In Haplosclerida the most common genus was Haliclona with five species, in Tetractinellida Thoosa with four species, and in Poecilosclerida it was Mycale with three species (Table 1). The species not identified are potential new species and are currently under study.
There were no clear groups because reefs were mixed on the cluster analysis (cluster not figured). In agreement with cluster, MDS did also not show a clear gradient and the reefs were arranged regionally only partially (stress 0.16) (Fig. 3); for example, the sponge community of Playa Blanca (Isla Socorro), is close to that of Punta Mita, although they are geographically separated from each other. On the right side of the ordination, some reefs that appeared together, as for example Islote Zacatoso, Playa las Gatas and Caleta de Chon, are spatially next to each other. In the left corner, appears some locations from Revillagigedo archipelago such as Clarion and Roca Partida, which are close to each other, and presented the lowest sponge diversity.
Regarding the quantitative sampling, the abundance varied from 0.57 to 4.3 (1.69 in average) ind. m −2 . The number of species per m 2 varied from 0.06 to 0.66 species per m 2 (0.25 in average) (Fig. 4).
The overwhelming majority of the species was cryptic (Fig. 5), occurring as small encrusting patches underneath coral rubbles and dead corals, or boring, measuring in the order of centimeters, only six species were relatively large measuring in the order of decimeters. The latter have the capacity of overgrowing live coral: Callyspongia californica, Chalinula nematifera, Mycale cecilia, M. magnirhaphidifera, Haliclona caerulea, and Amphimedon texotli. The last two, are the only massive species in all the reefs (Fig. 6).
It's also important to note the rare occurrence of keratose sponges. Except for four species of Aplysina and one of Ircinia, most of the reefs, and indeed the surrounded areas, are devoid of horny species or, if they are present, they are very small. www.nature.com/scientificreports www.nature.com/scientificreports/

Discussion
Coral reefs are the largest structures created by any group of animals in the world. Their three-dimensional framework forms numerous habitats which are densely populated by an enormous variety of organisms 65 such as sponges, which represent the major trophic link in organic matter transfer from the pelagic to the benthic compartment in these ecosystems 30,31,66 . There are not many studies about diversity of sponges on coral reefs, most published papers focused on Caribbean (CR), and West Pacific (WPR) 67 , while ETP is practically unknown 68 . Previous to this study, we know only two works, one on coral reef sponges in Panama, which cited 22 species 62 , and other one from Colombia, which did not deal exclusively with coral reef sponges but other habitats as well, it recorded 21 species 69 . Thus, the present work is the first large-scale study devoted exclusively to coral reef sponges from ETP, which, despite of the high number of reefs studied, and the vast geographical area that they represent, showed a surprising low number of species (87 species). This difference is more evident if we compare the total diversity in the entire Mexican Pacific coast, with particular reefs from CR or WPR; e.g. in Thousands Islands reefs (North Western of Java), 118 species are reported 70 , in the Spermonde Archipelago (south western Sulawesi, Indonesia), 151 species are recorded 71,72 . Reefs in the Gulf of Mannar and Palk Bay region (India) harbor more than 319 species 73 (see Fig. 7) [74][75][76][77][78][79][80][81][82][83][84][85][86][87] . A similar situation is found in the Caribbean. To quote some examples; 92 species in Bonaire reefs 88 , 124 species in Belize reefs -counting only cryptic species- 22 , which reach more than 300 species when included also exposed one 26,89,90 , 156 species in Curaçao (Saba Bank), 160 species in Cuba 91 (Fig. 7). It is important to note that small cryptic, boring, and thinly encrusting (<4 cm in diameter) specimens were excluded from most of these studies, so, the inclusion of those, would increase dramatically the number of sponge species in CR and WPR. www.nature.com/scientificreports www.nature.com/scientificreports/ By regions, the differences are more impressive yet: 420 species in coral reefs from Indonesia (830 in total in the country) 92 , 486 in coral reefs from Indian waters 73,92,93 , or 1500 species for the Great Barrier from Australia 94 . The WPR, particularly the "coral triangle" region, support the most diverse sponge assemblages in the world, probably including a high number of yet undescribed species.
When we compared standardized measures, such as abundance and species per square meter, the difference was also remarkable, with ETP drastically much lower than others regions (Fig. 4). In ETP we have 1.6 ind. per m 2 and 0.25 species per m 2 , in average. At a similar depth (below 6 m), CR and WPR have seven times more ind. per m 2 (≈11 ind. per m 2 ), and between four (1.3 for CR) and five times (1.4 for WPR) more species per m 2 . These differences between ETP and the other two basins were significant (Tables 2 and 3).
Comparison at deepest depth is not possible because there is no data for ETP below 6 m depth. The difference between CR and WPR was only significant for abundance at the 6-10 m depth interval (highest values): 12 vs 22 ind. per m 2 ; respectively. Previous studies also showed that diversity (per unit-area) is similar in CR and WPR, but sponge biomass is greater in CR 95,96 . The decrease of diversity at 16-20 m is better explained in terms of the smaller number of papers that report information at this depth, rather than an inherently poor fauna. This increase of abundance (density and cover) along a depth gradient with highest values at intermediate depths seems to be a general pattern of coral reef sponges previously observed in CR 97,98 .
Explaining the differences between CR and WPR is beyond the goal of this research. However, previous studies showed that factors such as food limitation 99,100 , chemical defense 101 , and nutritional strategies, with CR sponges more heterotrophic, and WPR more autotrophic 95,96 could be responsible of the differences.
In ETP as in CR 72,102,103 [among others] and WPR 99,100 [among others] very few species dominate the sponge assemblages, with a high percentage of species recorded from only a single site. This seems to be a general pattern in coral reef sponges worldwide. However, in ETP, the species that dominate the assemblages are boring sponges such as Cliona vermifera, Thoosa mismalolli or Pione carpenteri, which have a wide ETP occurrence and very broad ecological distribution. The prevalence of boring sponges in Mexican reefs is very interesting and remarkable, since these sponges are highly resilient to anomalous temperature shifts 104,105 , especially when compared to tolerance thresholds found for corals 106,107 . Previous studies showed that high anomalous temperatures that were detrimental to corals, had no negative effect in abundance and reproduction of C. vermifera 108 . The resilience demonstrated by boring sponge species in the ETP to thermal shock supports ecological projections that sponges will become an increasing threat to coral and coral reef health. However, recently it has been shown that elevated temperature can disrupt the functionality of microbial symbionts of Cliona orientalis, which occurred at a lower temperature than the 32 °C threshold that induced sponge bleaching 109 .
In summary, ETP coral sponges are not only less diverse compares to CR and WPR, there are also striking differences in growth form and size, because they are mostly cryptic encrusting, and very small in size (generally less than a few square centimeters). No sponges can be seen, except by close inspection of the bases of corals or by breaking open the reef frame (Fig. 2). In contrast, in CR and WPR sponges are more diverse, also in morphology, more species are living on exposed surfaces, they reach larger sizes, to over 2 m in largest dimension, and they can constitute the most abundant animals on the reef. For example, in the Wakatobi Marine National Park (Sulawesi, Indonesia) more than 200 individuals per m 2 have been reported which occupy more space than the corals 110 . Similar to CR, in WPR many large, conspicuous sponges are present, such as the giant barrel sponge Xestospongia testudinaria or species of Lamellodysidea, Phyllospongia, and Carteriospongia. www.nature.com/scientificreports www.nature.com/scientificreports/ Explaining the interoceanic differences: the influence of local and large-scale factors. Different theories have arisen to explain the pattern observed in ETP, particularly considering the dominance of encrusting and cryptic species, and the low diversity of their assemblages 111 . Prevalence of small cryptic encrusting sponges has been traditionally explained by the predation pressure by spongivorous fishes 112 , since cavities provide some advantage to cryptobionts by excluding certain predators. However, the pressure from spongivores is a common factor in the three areas 100 for WPR 101,113,114 [among others for CR and 115,116 for ETP]. There are also other sponge predators than fishes such as mollusks, echinoderms, and crustaceans, all of which have cryptobiontic representatives [117][118][119] .
Perturbation, at local and large scale, rather than biologicals factors, seems to explain the low prevalence and characteristics of sponge assemblages in ETP reefs, which are very frequently located in shallow water, where turbulence is periodically very strong, which, together with abrasion provoked by particles in suspension, www.nature.com/scientificreports www.nature.com/scientificreports/ sedimentation 41,120 , and high levels of damaging light, limit sponge survival and shape ETP sponge assemblages. The low sponge abundance at shallow depth in the Caribbean too has been associated with turbulence and numerous studies [97][98][99][100]112,121 , among others concluded that three depth-related factors influence sponge community structure on most Caribbean reefs: turbulence, spatial competition and predation. The first two only influence sponge communities at shallow depths, mostly above 10 m, and competition mostly above 20 m. In WPR, the   Table 2. Summary of the two-way ANOVA for differences in abundance (ind. per m 2 ) in ETP, CR and WPR, at different depths (see Fig. 4). Bold denotes a statistically significant difference. All F-ratios are based on the residual mean square error.
www.nature.com/scientificreports www.nature.com/scientificreports/ exclusion of sponges from shallow waters was also attributed to excessive turbulence and possibly by high levels of damaging light 28,70,122 among others. Exposed reefs at Isla del Coco 123 , and Clipperton 124 , which present a similar pattern with a high proportion of thin encrustations, support the suggestion that the turbulence and abrasion shape shallow coral reef sponge communities.
Beside the influence of local abiotic variables that could explain the low diversity and the prevalence of small encrusting species in ETP reefs, and indeed in CR and WPR shallow reefs, it is important to highlight the recurrent large-scale phenomena in the ETP, such as frequent upwellings that bring cold water up onto the reefs 12 , and periods of high-water temperature during El Niño years, which cause death and destruction of corals that have not had the chance to reach the levels of development found in the CR 15 . Moreover, ocean conditions in CR are relatively constant providing an environment that is conducive to reef growth (the average age of Caribbean reefs is 5600 years old). ETP reefs are smaller, younger (varying from 200 to 5600 years old), and with variable conditions where disturbances are more pronounced 125 . Explaining the interoceanic differences: the influence of evolutionary history. A hypothesis, which serves to explain the impoverished nature of the ETP coral fauna, is based on the unstable composition of faunas in remote marginal regions, and in the low resilience of these faunas. Due mainly to physical perturbations commented above, species already living near their tolerance limits become locally extinct and are not soon replenished after disturbances because of their isolation from source populations 126 .
The separation of ETP and the CR, occurred 3.5 million years ago, stopping the flow of species from the CR to the ETP, and since then, ETP has been highly isolated by cool currents from the north and south, and the Eastern Pacific Barrier (EPB) to the west; a vast expanse of deep water 16,125 . The isolation of the ETP sponge assemblages is supported by the fact that sponge fauna of Clipperton Island has stronger affinities with the Central and West Pacific regions than with the East Pacific region with which it shares only two or three species 124 . The majority of Clipperton species appears to have invaded from the west, evidenced by shared distributions or occurrence of close relatives in Hawaii, Tuvalu, Indonesia, New Caledonia and Australia. A study of the corals of Clipperton 127 came to a similar conclusion.
High-diversity locations such as the Philippines, Indonesia, or the Great Barrier Reef show a greater resilience to recurrent disturbances 128-131 that depauperate, marginal sites [e.g., the Galapagos Islands, Panama, Hawaii 7,132 , and indeed, a larger capacity to recovery after disturbances. They are also evidences that show that the cryptobenthic fauna of the Gulf of California is highly vulnerable to natural and anthropogenic disturbance as a result of the high specificity in habitat use of dominant species, and its low diversity, which limits the potential functional redundancy of the system, compromising the ecosystem's functioning, resilience and stability 133 .
Unequal rates of speciation, extinction and migration have resulted in greater diversity in the Caribbean than in the Pacific since, ETP reefs are also impoverished with respect to coral diversity (130 species in ETP vs 240 in CR), gorgonians, zoanthids, calcareous green algae, and other sessile groups 111,134,135 .
All this, suggests that current patterns of biodiversity should be interpreted in light of both contemporary and historical processes, which have been hypothesized to be most important for taxonomic groups with poor dispersal abilities 128 .
In conclusion, factors such as isolation, difficulty to gain recruits from distant areas, perturbation, resilience, age of the reef, allowed processes like natural selection to change the species composition of each area 12 .  Table 3. Summary of the two-way ANOVA for differences in number of species in ETP, CR and WPR, at different depths (see Fig. 4). Bold denotes a statistically significant difference. All F-ratios are based on the residual mean square error. (