Asterinides sp. an endemic stygobitic seastar from an anchialine cave and its interactions among prokaryotic communities

Abstract

Anchialine caves house a vast variety of organisms that support complex ecological relationships among themselves and their environment. The following study was made in the anchialine karst cave El Aerolito, found on Cozumel Island, Quintana Roo, Mexico. It explores the relationship between wall microbial mats and the diet of Asterinides sp., an endemic stygobitic seastar. Wall microbial mats inside the cave were sampled and the stomach microbiome of Asterinides sp. was obtained through regurgitation. Asterinides sp. sampling was made through the Catcher Collection Chamber (CCC), an innovative technology for the exploration of these ecosystems. The obtained results suggest that microbial mats are part of the diet of Asterinides sp. The following results highlight the potential relevance of the microbial communities inside the trophic chain present in El Aerolito. Additionally, the methodology presented here provides a useful framework for future ecological research in El Aerolito cave.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-36065-5.

Introduction

Anchialine caves of the Yucatán Peninsula were formed millions of years ago through rainwater infiltration into the karstic soil. This process led to progressively dissolving and collapsing limestone rock due to the action of carbon dioxide naturally present in rainwater, thus forming these cave systems¹. The anchialine caves exhibit unique physicochemical conditions, characterized by the convergence of two water layers with different salinity levels. This stratification arises from porous, or subterranean connection to the sea, which allows seawater infiltration into the system, enabling halocline formation. These characteristics establish anchialine caves as extremophile environments; they are considered natural evolutive laboratories that have been little explored¹. Anchialine cave fauna is characterized by a high grade of endemism², that has developed specialist strategies throughout their lifespan, including a tight relationship with the present prokaryotic microbiome. It is expected that microbiomes play a crucial role in biogeochemical cycles and trophic chains on anchialine cave ecosystems^3,4.

The karstic cave system El Aerolito, found on Cozumel Island, Quintana Roo, Mexico (20º 27′ 58″ N, 86º 58′ 41″ W)⁵, houses an extraordinary faunal biodiversity, particularly the phylum Echinodermata, being the dominant group considering their density. The Asterinides sp. seastar population at El Aerolito is of particular interest since it is an undescribed species⁶ with a small population². The taxonomic evaluation is in progress (Solis-Marin, 2025, personal communication). Asterinides sp. can be found in the marine water layer (salinity = 37.5) and an average temperature of 27.5 °C in El Aerolito on cave Site-D and Wonderland. The commonly found feeding strategy within this genus is opportunism and omnivory; digestion is made through extrusion of the stomach, and assimilation of nutrients is possibly slow⁷. This seastar can be commonly found adhered to the walls of El Aerolito with their stomach placed on the microbial mats.

The importance of metabolic networks among bacterial communities plays fundamental roles in driving biogeochemical cycles, mediating the dynamic of carbon, nitrogen, and sulfur in ecosystems^8,9. Particularly, extreme subterranean aquatic environments are proposed as reservoirs for complex and new prokaryotic biodiversity with unique metabolic strategies¹⁰. Considering this, aphotic karst caves ecosystems are critical since they could foster sophisticated synergistic nutrient relations among bacterial communities and local fauna^11–14.

In aquatic karst caves, the metabolic functions of bacterial communities have been studied in different types of systems like blue holes, streams, submarine caves, sinkholes, anchialine caves and “cenotes”^15–18. However, scientific exploration in aphotic zones is scarce; the use of specialist methodologies like cave diving in remote, no-light and difficult access areas contributes to this scientific data limitation for the study of echinoderm fauna of anchialine caves in Cozumel Island, Mexico³.

Therefore, this study aims to explore the possible ecological relationship between the stygobiont seastar Asterinides sp., and the microbial communities of the wall mats present in El Aerolito, sustaining the hypothesis that these microbial mats serve as a food source for Asterinides sp.

Methods

The field collection at El Aerolito anchialine cave was conducted during the dry season of November 2021 at Site-D cave², and Wonderland; both sites are at the marine groundwater layer without direct access to the halocline. Four groups of samples were collected: water, microbial mat, one rock covered with microbial mat, and regurgitation of the seastar Asterinides sp. (Fig. 1). Five regurgitation samples of Asterinides sp. (S0R, S1R, S3R, S4R, S5R) were obtained at Site-D using a special method that included the technological development of the Catcher Collection Chamber (CCC), consisting of an anti-turbulence chamber that made possible the regurgitation precipitate¹⁹. Additionally, three regurgitation samples were obtained using a traditional method, which consists of collection with a sterile plastic bag: from Site-D (AR and BR) and Wonderland (W6R). Seven microbial mat samples were obtained from Site-D and one from Wonderland, but only one from Site-D (S2TM2) could be processed; as well as one rock covered with microbial mat collected at Wonderland. Although multiple matrices were collected, the analyses on this study were primarily focused on characterizing the regurgitation samples.

Fig. 1.

Fieldwork from this study. (a,b) In situ photographs of Asterinides sp. specimens adhered to the microbial mat on the wall of El Aerolito. (c) Illustrative diagram of the sampling site in El Aerolito, Cozumel. (d–f) photographs of the field collection. Images taken by: (a,b) Fernando Calderón Gutiérrez in Site-D, (c) Cindel Vergara Ovando, (d–f) Sarah Rubelowsky in Site-D.

DNA was extracted from three matrices: microbial mat, water, and regurgitation (suppl. Table 1). For the Illumina MiSeq sequencing process V6-V8 region of the 16S rRNA specific for Bacteria was used (B969F:5′-ACGCGHNRAACCTTACC-3′; BA1406R:5′-ACGGGCRGTGWGTRCAA-3 × )²⁰. Obtained sequences were preprocessed with DADA2²¹ in RStudio v4.2.2, which included removal of singletons (< 1% of the total ASVs). Sequences were identified to genus level and grouped in Amplicon Sequence Variants (ASVs) according to the SILVA v138.1 (2019) database²². Due to technical considerations, we only worked with forward sequences with ~ 300 bp. The Vegan²³, Phyloseq²⁴, Microbiome²⁵, and Tidyverse²⁶ packages of RStudio v4.2.2 were used for rarefaction analysis (suppl. Graph 1 and 2). Alpha diversity was calculated using Shannon and inverse Simpson indexes, the richness through average of observed ASV and Fisher index, and the dominance through the Simpson index. Beta diversity was assessed by using weighted and unweighted UniFrac²⁷ through an nMDS analysis²³, and a PERMANOVA²³ for weighted and unweighted UniFrac distances. Additionally, the functional prediction of carbon (C), nitrogen (N), and sulfur (S) were made with Tax4fun2²⁸ (suppl. Functional prediction Table 2). Lastly, the morphology of the microbial mats was examined by Scanning Electron Microscopy (SEM).

Results and discussion

Characteristics of sequence data

We obtained 143,983 sequences from twelve samples: four of water, seven of regurgitates, and one sample of microbial mat (suppl. Table 3). After quality control and chimera removal, 57% of the initial sequences were retained, resulting in 1528 ASVs (suppl. Table 4). Overall, only 0.73% of the sequences were identified as chimeras.

The average coverage amount of diversity in bacteria (31.75%) points out a high amount of undescribed diversity (47.2% unclassified at the genus level). The estimated richness with Fisher index (43.645) and the average number of the observed ASV (x̄ = 147.5 ± 84.87) are relatively high, suggesting a high potential species richness. Gini-Simpson (0.902) and Shannon (4.538) indexes point out the dominance of a few species in the community (suppl. Graph 3). The stress value for non-metric multidimensional scaling (nMDS) weighted was 0.1605715, and 0.1489075 for the unweighted. The PERMANOVA analysis for nMDS with weighted (p = 0.146) and unweighted (p = 0.164) UniFrac distances did not show significant differences. Principal coordinates analysis (PCoA) using weighted and unweighted UniFrac distances did not reveal groupings by passage (suppl. Graph 4), suggesting low beta diversity. However, it is important to consider that only one processable microbial mat was available from Site-D, and only one sample of regurgitation from Wonderland. Further studies on bacterial diversity in El Aerolito are necessary to explore microbial diversity.

Technical considerations related to cave diving at this aphotic, narrow, and aquatic cave could be an important factor influencing the sampling effort, as reflected by the non-asymptotic rarefaction curves and the low percentage of coverage diversity (suppl. Table 5). While current microbial databases lack representation from aquatic karst cave systems, and differences in genetic, molecular, and sequencing methodologies (e.g.,¹¹) present challenges for taxonomic comparisons, these limitations emphasize the opportunity for further exploration and expansion of reference data, particularly for understudied environments. These opportunity areas were reflected in the high number of unclassified species and suggest significant potential for yet undiscovered bacterial diversity and their possible role into the ecology and biogeochemical cycles in El Aerolito.

Microbial community composition

The dominant phylum was Proteobacteria (suppl. Graph 5 and Table 6); at genus level the HIMB11, Pseudoalteromonas, Reyranella, Sediminibacterium, and Pseudoalteromonas genera were present in all samples; HIMB11 and Pseudoalteromonas with greater abundance in the water sampling group. Psychrobacter was only detected in the water and regurgitation sampling groups and Nitrospira was detected in regurgitation and microbial mat but not in water. Hgcl, Arenimonas, Alishewanella and Rheinheimera were only found in regurgitation (Fig. 2; Table 1).

Fig. 2.

Relative abundance of bacterial composition at the genus level obtained from the analysis of sequences of three sampling groups: water, regurgitation, and microbial mat. (a) Bacterial composition at genus level. (b) Venn diagram based on relative abundance of Bacteria on the left.

Table 1.

Relative abundance of bacteria genera by sample.

Genus	S0-AGUA	S1-AGUA	S2-AGUA	S0R	S1R	S3R		S4R	S5R	AR	BR	W6R	S2TM2
Alishewanella	0	0	0	0	0	0		0	0	0	0	0.35077067	0
Arenimonas	0	0	0	0.09919571	0	0.01635688		0.00566907	0	0.06669641	0.08934426	0	0
Hgcl clade	0	0	0	0	0	0.06195787		0.03909295	0	0.01672987	0	0	0
HIMB11	0.28626776	0.01163101	0.26974063	0.02055407	0.14635535	0		0.04074643	0.14212548	0.04751283	0	0	0.06688963
Nitrospira	0	0	0	0.10947274	0	0.03890954		0.02940829	0	0.06558108	0.07704918	0	0.25035834
Other	0.64227292	0.05367337	0.62824207	0.5665773	0.75626424	0.76604709		0.77075706	0.47247119	0.72273031	0.67622951	0.28397945	0.57190635
Pseudoalteromonas	0.01851055	0.86640346	0.01268012	0.12868633	0.00455581	0.01982652		0.0148813	0.23303457	0.03948149	0.15737705	0.00700607	0.01863354
Psychrobacter	0.0232458	0.06301019	0	0.06344951	0	0.00619579		0.01677099	0.03072983	0	0	0.05838393	0
Reyranella	0.0038743	0.00085365	0.01383285	0.00223414	0.08029613	0.03420074		0.03342388	0.02944942	0.01851439	0	0	0.03631151
Rheinheimera	0	0	0	0	0	0		0	0	0	0	0.29985988	0
Sediminubacterium	0.0258286	0.00442832	0.07550432	0.00983021	0.01252847	0.05650558		0.04925003	0.0921895	0.02275262	0	0	0.05590062

Regurgitation sample W6R (Wonderland) was different in taxonomical and compositional characteristics compared to the samples from Site-D. Hgcl and Arenimonas were found at Site-D. Alishewanella and Rheinheimera were only found in sample W6R (Fig. 2; Table 1). Furthermore, the microbial mat sample (S2TM2) from Site-D appeared more similar in its microbial community composition to the regurgitate samples from the same site than to the regurgitated sample (W6R) from Wonderland, which showed a distinct community profile. In similar ecosystems, the variations in microbial community composition are linked to physicochemical conditions and distance between samples, connection with the sea, and the cave entrance^10,11,29. In El Aerolito, there are differences in physicochemical conditions between Site-D and Wonderland (suppl. Graph 6). Furthermore, seasonal rainfall influences groundwater flow and mixing of water, among other chemical characteristics³⁰, however, it is unknown how the bacterial community responds to these changes^8,31. It is necessary to deepen investigation to understand the community structure and diversity from Site-D and Wonderland.

The exclusive presence of Hgcl and Arenimonas in Site-D regurgitation samples, and Alishewanella and Rheinheimera in W6R sample from Wonderland, potentially indicate associations with Asterinides sp. Other studies have identified high microbial abundances in the digestive microbiome of seastars. For example, 27% of the microbial community abundance of Certonardoa semiregularis is represented by four bacterial groups: Psychrobacter sp. and Rheinheimera sp. (10.5% each), Pseudoalteromonas spp. (5%) and Arenimonas spp. (1%)³². Interestingly, Psychrobacter has been described previously as dominant in the intestinal microbiome of Asterias amurensis, suggesting a potential association with Asteroidea digestive systems³³.

The Hgcl clade is a ubiquitous taxon with high degree of diversification³⁴, associated with N-rich environments in photic estuaries^35,36 and commonly found in oligotrophic ecosystems like in the deep of high-mountain lakes³⁷. The widespread distribution of the Hgcl clade could explain its presence in the regurgitate of Asterinides sp.

Metabolic potential

S-biogeochemical cycle

According to the metabolic profiles, the dominant function in the water, regurgitation, and microbial mat samples is sulfate oxidation within the S-cycle (Fig. 4); which is consistent with the presence of sulfate on water, %DO and the relative abundance of soxB gene (suppl. Graph 6, Table 7 and Graph 7). Conducted studies in different sulfate-rich subaquatic anchialine caves^18,38 as well as analysis of the water column of other cavernous anchialine estuaries, provide similar conclusions: the dominant metabolic functions in the water layer located beneath the halocline are associated with sulfate-methane cycling and ammonia oxidation³⁹.

Fig. 4.

Heatmap of relative abundance in the metabolic routes, and a dendrogram based on the relative abundance of metabolism enzymes by sample-type (water, regurgitation and microbial mat (MM)), obtained with Tax4Fun2²⁸. Genes used for each metabolic category: carbon fixation on photosynthetic organism: rbc5; carbon fixation on bacterial organism: aclA, abfD, Mcl; methane metabolism: pmoA, cbiA, mcrA, mtrA, acsE; nitrogen metabolism: hzoA, nifH, amoA, nosZ; sulfur metabolism: dsrA, soxB (suppl. Table 2).

The structure of bacterial communities includes genera associated with sulfur-oxidizing functions in the three matrices (Fig. 2; Table 1). Supporting this observation, the genus Pseudoalteromonas has been reported participating in sulfur oxidation processes⁴⁰.

Among genera associated with the sulfur cycle, the oxidizing-sulfur Pseudoalteromonas has been described in the digestive microbiome of sea sponges⁴¹. Sediminibacterium has been related to sulfur oxidation in marine ecosystems, although its role is not described in karst systems⁴². Specifically in karst systems, Alishewanella was classified as an oxidizing-sulfur bacterium, and observations in the water column of Cenote Calia indicate this genus exhibits temporal variability, being more abundant during the dry season⁴³. Lastly, HIMB11 has a sox gene cluster which could reflect metabolic potential for sulfur oxidation⁴⁴.

At the microbial mats from El Aerolito it was also possible to observe filamentous macroscopic structures and Beggiatoa-like microscopic structures (Fig. 3a and suppl. Photography 1) which may indicate the presence of organisms of Beggiatoa, or a related genus (Fig. 3b-I). Similar macroscopic structures have been observed in different microbial mats under similar environmental conditions⁴⁵. SEM images of the microbial mat (Fig. 3b-III) presented a similar morphology with Thiotrix biofilms⁴⁶. Both Beggiatoa and Thiotrix are associated with sulfur oxidation, and their presence was reported in microbial mats in sulfate-rich marine aphotic environments¹². Otherwise, structures suggest cell-like morphologies similar to cocci or bacilli (Fig. 3b-II).

Fig. 3.

Scanning electron microscopy (SEM) of a sample taken from a microbial mat from the Wonderland passage at El Aerolito. (a) SEM in X2000 view. (b) Zoom of structures of interest in the microbial mat (not to scale). b-I indicates a similar structure of an organism Beggiatoa-like. b-II is similar to a conglomerate of coccoid- and bacillus-shaped cells. b-III morphology showing similarity with an organism Thiothrix-like. Images facilitated by the Laboratory of Microscopy of the Faculty of Sciences, UNAM, taken by Silvia Espinoza Matías in 2022.

C-biogeochemical cycle

The 3-hydroxypropionate (3HP) pathway is the second most abundant metabolic route for carbon fixation after sulfur oxidation in the functional prediction (Fig. 4). However, autotrophic carbon fixation in aphotic marine ecosystems is not limited to the 3HP pathway: there are chemoautotrophic bacteria (e.g., some Proteobacteria) that fix carbon using energy derived from processes such as sulfur oxidation or anaerobic processes, such as ANAMMOX^47–49.

Reyranella was detected across all three sample types (Fig. 2), has been reported in oxic karstic caves⁵⁰, and occasionally some isolates were characterized as manganese oxidizers⁵¹, while Rheinheimera has been found to be abundant in oxic marine waters, is considered to possess a flexible metabolic profile and potentially contribute to organic matter degradation⁵². However, the specific ecological role of both genera in karstic submarine caves remain unclear.

N-biogeochemical cycle

According to the functional prediction (Fig. 4), the metabolism associated with the cycle of N is the least abundant in the three sample matrices. The most significant function is the denitrification, represented by the relative abundance of nosZ (suppl. Graph 7). However, this process may occur alongside nitrification within microbial mats, as observed in Webbubie submarine cave, where nitrification and denitrification take place within microbial mat, creating microanoxic conditions that could also support ANAMMOX activity (which oxidizes ammonium anaerobically using nitrite)^11,53,54. This suggests that ANAMMOX could also occur in El Aerolito, although no specific bacterial group linked to this pathway was detected (Fig. 2; Table 1; suppl. Table 7).

Nitrospira is present only in regurgitation and microbial mat samples (Fig. 2; Table 1) and probably associated with the function of nitrification. Nitrospira has been identified into microbial mat from Weebubbie cave⁵⁴ and in the aphotic zone of the water column from Ox Bel Ha anchialine cave, where was linked to nitrogen oxidation³⁰.

Based on compositional analysis, no statistical differences were found in the community structure of the three sample matrices (Fig. 2). However, the functional metabolic dendrogram (Fig. 4) suggests important metabolic similarity between the microbial mat and the regurgitation.

Microbial Mats as a Dietary Component of Asterinides sp.

The family Asterinidae, to which Asterinides sp. belongs, is characterized as opportunistic omnivorous, using any available source of organic matter⁵⁵. Extreme environments similar to the conditions found in El Aerolito, coupled with the restricted distribution of Asterinides sp. to Site-D and Wonderland passages, and the low motility of the seastar suggest that the microbial mat represents the most probable energy source for these populations. This hypothesis is supported by the greatest similarity of microbial community composition and identified metabolic routes among regurgitation and microbial mats samples. This resemblance could be possible since microbial mats are part of the diet of Asterinides sp.; furthermore, functional predictions indicated a relative predominance of oxidative pathways in the regurgitate (Fig. 4 and suppl. Graph 7), including genes related to sulfur and carbon oxidation (e.g. soxB, mcl). The functional profile may suggest that the metabolic activity in the regurgitate is primarily aerobic, as has been proposed for the digestive microbiome of some shallow-water seastars⁵⁶.

The extracellular polymeric substances (EPS) present characteristics like adhesion and adsorption properties that make them naturally bind with particles, molecules, metals, heavy metals and environmental contaminants⁵⁷. The EPS could provide essential nutrients such as carbon, polysaccharides, proteins, and may serve as a calcium source, an important variable in Asterinides sp. tissue regeneration^58,59. The assimilation of nutrients of EPS was reported in other animals like the seastar Amphipholis gracillima⁵⁹.

The EPS has implications for immobilization⁶⁰, subsequently leading to ingestion and bioaccumulation of heavy metals in tissues of organisms that use EPS as a nutrient source⁶¹. This suggests the possibility of exposure to external pollutants—such as hydrocarbons like diesel derived from nearby vehicular traffic—which could be involved in EPS in microbial mats and lead to the bioaccumulation of contaminants in Asterinides sp.

The presence of Psychrobacter in water and regurgitate could raise concerns about potential contamination, since this genus has been associated with wastewater and microplastics^62–64. Contaminants probably originated from nearby hotels or by infiltration of hydrocarbons and other pollutants due to the Aerolito proximity to the South Coastal Highway.

The results suggest that microbial mats are part of the digestive microbiome of the fauna that lives in environments such as anchialine caves, and/or the taxonomic composition of the communities of microbial mats is an important variable in the localization and delimitation of populations of Asterinides sp. in the passages of El Aerolito.

This study provides a methodological and technological reference for the future studies of gut content of echinoderms, and ecological processes in groundwater environments. Future explorations in the El Aerolito and other anchialine caves can contribute to the generation of conservation instruments for these exceptional evolution biomes.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

F.A.S.M., C.V.O., M.R.O., F.C.G., G.M.T. and N.C. wrote the main manuscript text and C.V.O. prepared Figs. 1 and 2. F.A.S.M., M.R.O. and N.C. developed the protocol experiment and supervised the project. C.V.O. performed laboratory experiments . F.A.S.M. and N.C. supervised the sampling. F.C.G. made the diving sampling. N.C. prepared samples for preservation. M.R.O. supervised experiments in laboratory. G.M.T. developed the CCC (sampling unit). All authors reviewed the manuscript.

Funding

Research funding is a grant provided by Dirección General de Asuntos del Personal Académico, Universidad Nacional Autónoma de México, PAPIIT-IN207021, to financially support the scientific project; and by Universidad Nacional Autónoma de México, to support posgraduate grant.

Data availability

The sequence datasets that support this study are available on NCBI Sequence Read Archive (SRA), under BioProject number PRJNA1119047 (https://www.ncbi.nlm.nih.gov/sra/PRJNA1119047) ; individual accession numbers: SAM41635988, SAM41635989, SAM41635990, SAM41635991, SAM41635992, SAM41635993, SAM41635994, SAM41635995, SAM41635996, SAM41635997, SAM41635998, SAM41635999. The data generated during the current study are included in the supplementary material.

Declarations

Competing interests

The authors declare no competing interests.