Coral reefs are one of the most diverse ecosystems on the planet (Connell, 1978). Scleractinian corals that secrete calcium carbonate layers through the epidermis of the lower half of the column and the pedal disc are principal reef-builders (Drake et al., 2020). Other organisms such as coralline algae and hexactinellid sponges can contribute to reef formation when they leave their silica and carbonate skeletons. Despite occupying about 0.2% of the seafloor, coral reefs are responsible for sheltering and feeding at least one quarter of marine life (Knowlton and Jackson, 2001).

The coral food source comes from zooxanthellae, microscopic algae that live inside of it in a symbiotic relationship, and supply 90% of the coral’s nutritional needs. Corals prefer shallow, warm, and oligotrophic waters to fix carbon by oxygenic photosynthesis using Calvin-Benson-Bassham (CBB) cycle (Stanley Jr, 2006). Muddy waters of large tropical rivers often inhibit the formation of coral reefs as sediment covers them and consequently decrease the light availability for the zooxanthella, compromising its source of nutrients. Despite the fact that tropical rivers exhibit these features, some biogenic reefs can be found at the Amazon River mouth.

Recently, scientists recognized a new biogenic reef system of 56,000 km² in the Amazon River mouth close to offshore oil companies (i.e., BHP-Billiton, Queiroz Galvão, Ecopetrol, Total, BP, and Petrobras) - the Great Amazon Reef System (GARS) (Moura et al., 2016). A mud layer called “plume”, formed by organic and inorganic suspended material that spreads across the Atlantic Ocean, influences the formation of this reef. Instead of corals, coralline algae built the GARS foundation with small patches, platforms, and walls covered by scleractinians, sponges, octocorals, and black corals. Sponges that are adapted to this environment are the major reef epifauna. A total of 61 sponge taxa were identified in this ecosystem, with the most commonly found being Agelas spp., Aplysina spp., Callyspongia vaginalis, Clathria nicolea, Monanchora arbuscula, Oceanapia bartschi, and Geodia spp. (Moura et al., 2016). Sponges are known to filter water by choanocytes cells whose flagellar movement creates a water flux into their pores to capture dissolved organic matter (DOM), digest it (i.e., phagocytosis), and release particulate organic matter (POM) to higher trophic levels via the sponge loop (de Goeij et al., 2013).

Sponges have co-evolved with microbes for over 800 million years (Turner, 2021). The sponges’ microbiota can represent approximately 35% of the holobiont biomass (Webster and Thomas, 2016). More than 47 sponge-inhabiting phyla were described, some of which were found exclusively in these invertebrates (Reveillaud et al., 2014). Their microbiota provides the holobionts the capacity to fix carbon, metabolize nitrogen, sulfur, and phosphate (Pita et al., 2018). Moreover, sessile organisms such as sponges rely on chemical compounds produced by their microbiota to defend against predators, competitors, and invading organisms such as bacteria, viruses, and microeukaryotes. In many cases, compounds isolated from sponge’s microbiota are used as drug precursors and bioactive molecules in the pharmaceutical industry - being the most promising source for new bioactive products in the marine environment (Blunt et al., 2017). However, linking a specific microorganism from the sponge microbiota to a specific metabolic function and taxonomic group is a challenge, since many microorganisms are hard to cultivate in the laboratory, due to unknown growth requirements (Lewis and Ettema, 2019). Culture-independent techniques such as genome-resolved metagenomics allow reconstruction of draft genomes from the environment, enabling metabolic inferences for microorganisms that are little known or do not have yet cultivated representatives (Tyson et al., 2004).

In this context, the GARS sponge microbiome study provides insights into sponge microbiota’s role and provides information for ecological surveillance, and prospecting focused on biotechnological research. In this work, we used metagenome-assembled genomes (MAGs) to investigate the potential metabolic contribution of the sponges microbiome (Hügler and Sievert, 2011) to sustain this holobiont in a reduced-light environment.

We found that the sponge’s microbiome from the Great Amazon Reef System (GARS) is a unique community. The Amazon water turbidity-related microbiome influences the GARS sponge microbiome. However, data suggest more influences due to environmental features than to the shared microbiota. Significantly, and in contrast to other Reef systems, photosynthetic microorganisms (Cyanobacteria) were only present in low proportion in both 16S rRNA and MAG community analysis. GARS microbiome’s high diversity suggests that the plume layer does not restrict or enrich the microbiota for a specific taxonomic group. Moreover, 32 taxa on the class level were statistically different between the sponge microbiomes from GARS and traditional Caribbean reefs, indicating muddy water contributes to an adapted microbiota. Microbiome comparisons between the Caribbean (i.e., Curacao, Cayman Island, and Belize) and the Amazon region using sponges from the same genera, and two different genera, show that the findings are more general than they would be if we had made this comparison using a single genus and Caribbean location. Light in the GARS floor is faint but sufficient; a study shows the presence of photosynthetically available radiation (PAR) between 0.01 and 19.3 µE m^-2 s^-1 (Francini-Filho et al., 2018). Such ratio can support photosynthesis, and coralline algae reef-builders have been present in the GARS for at least 4487 years before today (Moura et al., 2016). Studies suggested high sedimentation as a principal stress factor in GARS, which results in total or partial GARS coverage in sediment (Moura et al., 2016; Omachi et al., 2019).

The data indicate a predominant heterotrophic lifestyle for the GARS sponge microbiota. The large amount of organic matter carried by the Amazon River may explain the sponge’s success in this ecosystem (Medeiros et al., 2015; Moura et al., 2016). A study revealed that dissolved organic carbon in the GARS region varies from 102.3 to 165.69 µM, and particulate organic carbon from 4.1 to 6.21 µM (unpublished data). The sponges’ microbiota also recycles and benefits from DOM (Azam et al., 1983) and contribute to more than half of the sponges’ diet (Kazanidis et al., 2018; Rix et al., 2020; Bart et al., 2021). Estimates that one sponge kilogram can filter 50,000 L of water daily and remove more than 90% of microbes in the water (Hill and Hill, 2009). Studies have demonstrated that photosynthetic net reduction and heterotrophic capacity enhancement are adaptive responses to overcome extreme conditions in turbid reefs (Burt et al., 2020).

Autotrophic microorganisms have an auxiliary role in the GARS sponge microbiome. MW-score indicates Calvin–Benson–Bassham (CBB) as the least relevant carbon fixation pathway present in the plume-adapted Amazon sponge microbial community, in stark contrast to the open ocean, where Cyanobacteria and algae are the major organisms responsible for carbon fixation through the CBB pathway (Raven, 2009; Bowler et al., 2010). Our MAGs analysis predicted CBB to occur in Rhodobacterales and Arenicellales. Rubisco type IV detected in a Chloroflexota MAG are related to functions other than carbon fixation (Tabita et al., 2007a; Tabita et al., 2007b). The presence of CBB cycle genes, especially Rubisco, in Proteobacteria and Rhodobacterales, has already been described, including the sponge microbiome (Badger and Bek, 2008; Asplund-Samuelsson and Hudson, 2021).

Metabolic predictions indicated 3HP/4HB as the most relevant carbon fixation pathway in the GARS sponge’s microbiome, significantly more abundant than CBB and rTCA (Wilcoxon test, P < 0.05). We predicted 3HP/4HB to occur in Thermoproteota and one Desulfobacterota MAG. The 3HP/4HB capability for the Desulfobacterota MAG remains uncertain as one of the two marker genes, the malonyl-CoA/succinyl-CoA reductase (NADPH) gene, could not be identified in this MAG. Additionally, the bacterial 4-hydroxybutyryl-CoA dehydratase is involved in the fermentation of aminobutyrate (Buckel and Golding, 2006). These findings were based on high-quality MAGs gene content and coverage observations - not expression data. Furthermore, it is important to note that genome-resolved metagenomics recovers just a fraction of the genomes present in the microbial community. The purpose of this study was to use in silico analyses to access the microbial functional capacity of samples for which there is no experimental information available.

Thaumarchaeota’s (Thermoproteota) variation in the 3HP/4HB pathway is well-described and considered the most energy-efficient aerobic carbon fixation pathway (Walker et al., 2010; Pester et al., 2011; Konneke et al., 2014). Studies have demonstrated the presence of Thaumarchaeota in the Amazon plume water and the sponge microbiome (Engelberts et al., 2020; Pinto et al., 2020). Our data indicate that these microorganisms play multiple roles in the sponge-microbiome holobiont physiology; besides carbon fixation, we predicted ammonia oxidation, sulfur oxidation, and sulfate reduction functions. Ammonia is the sponge’s main waste product, which is toxic at high levels. Ammonia oxidation can prevent its accumulation within sponge tissues. The same happens for sulfur oxidation, which prevents hydrogen sulfide levels to become high (Taylor et al., 2007).

Another predicted carbon fixation pathway was the 3-hydroxypropionate bicycle (3HP). This pathway was first described in Chloroflexus aurantiacus, a Chloroflexaceae member able to perform anoxygenic photosynthesis (Herter et al., 2001; Zarzycki et al., 2009). In our taxonomic analysis, we identified the presence of Chloroflexota MAGs related to Anaerolineae, Dehalococcoidia, and UBA223 families, not previously described as having this pathway. The MAGs predicted to encode this pathway in our GARS sponge microbiome were classified as Rhodobacteraceae. One study suggested Dinoroseobacter shibae a member of Rhodobacteraceae with mixotrophic metabolism, assimilating glucose by Entner-Doudoroff and fixing carbon by 3-hydroxypropionate bicycle (Tomasch et al., 2011). Alternative CO₂ fixation pathways may facilitate the assimilation of simple organic substances, conferring a competitive advantage over obligate autotrophs or heterotrophs (Hügler and Sievert, 2011). The genomic comparison of 3HP genes between our MAGs and D. shibae showed that they are similar.

The same uncertainty occurs for rTCA genes encoding MAGs, where hmmsearch detected the ATP-citrate lyase alpha-subunit only when a less restrictive cutoff was used. The rTCA pathway was predicted to be the most taxonomically wide-spread pathway in the GARS sponge microbiome within Acidobacteriota, Gammaproteobacteria, Latescibacterota, Bacteroidota, Chloroflexota, Actinobacteriota, Verrucomicrobiota, and Spirochaetota. An incomplete rTCA cycle was found in phylum Acidobacteriota (Fernandez et al., 2020), Bacteroidota Salinibacter ruber (Bagheri et al., 2019) and other phyla like Actinobacteriota and Chloroflexota (Vikram et al., 2016). Among our low-quality MAGs, we have recovered members of Aquificota, which like Chlorobiota and Campylobacterota are often described to participate in rTCA (Hügler and Sievert, 2011). Due to the oxygen sensibility of its enzymes, the rTCA pathway is usually found in anaerobes associated with deep-sea hydrothermal vents and dark oceans and in microorganisms found in microaerophilic conditions (Berg, 2011; Erb, 2011). These data corroborate with dissolved oxygen (i.e., 3.25-3.81 mg. L-1) values reported in the GARS (unpublished data).

Prediction of the complex metabolic pathways was the main limitation of this study. Those pathways have many promiscuous genes, which makes it difficult to establish their involvement in any specific pathway. Furthermore, database profiles are incomplete. Sometimes genes that have already been shown to participate in a pathway have not yet been included in the database. An example is the 3HP/4HB pathway that lacks genes from Thaumarchaeta on the KEGG profile – making it difficult to identify non-classic pathways. Crenarchaeal Sulfolobales genes were used to construct the 3HP/4HB pathway profile found in the database used in this study. Additionally, the metabolic characterization of organisms based on their metabolic potential can not be used to predict their in vivo metabolism. Hence, further experimental validation is necessary, either culture independently through, e.g., Meta-Transcriptomics and -Proteomics, or through cultivation followed by the characterization of the culture.

In summary, our data indicate that although the plume induces a low light stress on the GARS sponge microbiome, it does not completely prevent Cyanobacteria occurrence. Specifically, we found a predominance of heterotrophic compared to autotrophic metabolism in the plume-adapted Amazon sponge microbial community. We propose that in GARS, sponges accumulate dissolved organic matter (DOM), which is assimilated by themselves as well as its heterotrophic microbiota. In addition, the sponges’ microbiota might act in the detoxification of toxic compounds released by the sponge.

Other turbid reefs present similar conditions of restricted luminosity or sedimentation to those found by GARS sponges (Zweifler (Zvifler) et al., 2021). However, none of these environments receive such a large water flux. The Amazon River discharges 20% of global freshwater into the Atlantic Ocean (Richey et al., 1990; Dai and Trenberth, 2002). This huge water flux into the ocean drastically changes environmental parameters such as salinity, pH, luminosity, oxygenation, and sedimentation, thus creating the unique environment of GARS. This study shows a different biomass accumulation strategy for GARS sponge microbiome from that of traditional sponge microbiomes. The uniqueness of the GARS sponge microbiome is ripe for biotechnological exploration, as new genes and drug precursors may be discovered. In addition, from an ecological perspective, the GARS microbiome data may be useful for surveillance efforts of potential impacts of offshore oil and gas operations near this region.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: https://www.ncbi.nlm.nih.gov/, PRJNA796108.

OP and RK designed the study. Data analyses were performed by OP, TB, AP, TF, and RO. Data collection was performed by FT and MF. The manuscript was drafted by OP and revised by all authors, BQ critically reviewed and substantially edited the manuscript. All authors have read and approved the manuscript.

This study was financially supported by the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES), which also provided financial support for the first author (OP), Fundação de Amparo à Pesquisa do Estado do Rio de Janeiro (FAPERJ), and Fundação de Apoio à Pesquisa do Distrito Federal (FAP-DF). TB and AP were supported by the Ministry of Culture and Science of North Rhine-Westphalia (Nachwuchsgruppe “Dr. Alexander Probst”).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.