Shangjin Tan

The sponge microbiome underpins host function through provision and recycling of essential nutrients in a nutrient poor environment. Genomic data suggest that carbohydrate degradation, carbon fixation, nitrogen metabolism, sulphur metabolism and supplementation of B‐vitamins are central microbial functions. However, validation beyond the genomic potential of sponge symbiont pathways is rarely explored. To evaluate metagenomic predictions, we sequenced the metagenomes and metatranscriptomes of three common coral reef sponges: Ircinia ramosa , Ircinia microconulosa and Phyllospongia foliascens . Multiple carbohydrate active enzymes were expressed by Poribacteria, Bacteroidota and Cyanobacteria symbionts, suggesting these lineages have a central role in assimilating dissolved organic matter. Expression of entire pathways for carbon fixation and multiple sulphur compound transformations were observed in all sponges. Gene expression for anaerobic nitrogen metabolism (denitrification and nitrate reduction) were more common than aerobic metabolism (nitrification), where only the I. ramosa microbiome expressed the nitrification pathway. Finally, while expression of the biosynthetic pathways for B‐vitamins was common, the expression of additional transporter genes was far more limited. Overall, we highlight consistencies and disparities between metagenomic and metatranscriptomic results when inferring microbial activity, while uncovering new microbial taxa that contribute to the health of their sponge host via nutrient exchange.

Marine invertebrates harbour a complex suite of bacterial and archaeal symbionts, a subset of which are probably linked to host health and homeostasis. Within a complex microbiome it can be difficult to tease apart beneficial or parasitic symbionts from nonessential commensal or transient microorganisms; however, one approach is to detect strong cophylogenetic patterns between microbial lineages and their respective hosts. We employed the Procrustean approach to cophylogeny (PACo) on 16S rRNA gene derived microbial community profiles paired with COI, 18S rRNA and ITS1 host phylogenies. Second, we undertook a network analysis to identify groups of microbes that were co‐occurring within our host species. Across 12 coral, 10 octocoral and five sponge species, each host group and their core microbiota (50% prevalence within host species replicates) had a significant fit to the cophylogenetic model. Independent assessment of each microbial genus and family found that bacteria and archaea affiliated to Endozoicomonadaceae, Spirochaetaceae and Nitrosopumilaceae have the strongest cophylogenetic signals. Further, local Moran's I measure of spatial autocorrelation identified 14 ASVs, including Endozoicomonadaceae and Spirochaetaceae, whose distributions were significantly clustered by host phylogeny. Four co‐occurring subnetworks were identified, each of which was dominant in a different host group. Endozoicomonadaceae and Spirochaetaceae ASVs were abundant among the subnetworks, particularly one subnetwork that was exclusively comprised of these two bacterial families and dominated the octocoral microbiota. Our results disentangle key microbial interactions that occur within complex microbiomes and reveal long‐standing, essential microbial symbioses in coral reef invertebrates.

Planktonic protists, including both autotroph and heterotroph, have been recognized as a major contributor to primary production and consumers of bacteria, archaea and picophytoplankton. However, the understanding of protistan diversity is typically impeded by the large amount of metazoans when employing universal primers to environmental samples. In this study, we developed 3 blocking primers to inhibit the amplification of metazoan DNA in PCR. First, we optimized the design and concentration of blocking primers by using 2 metazoans and 3 protists in quantitative PCR. No significant difference (q > 0.05) was found in protistan community structure at phylum, family or OTU levels and genetic diversity between samples amplified with and without adding blocking primers, indicating that the blocking primer does not alter the composition of protistan community. An application to samples with a high abundance of metazoans showed that the blocking primer can reduce 42.1–72.4% of metazoan sequences, resulting the retrieval of a higher protistan richness and diversity. The use of the blocking primer helped to identify protistan community composition from a large size fraction, which is usually dominated by metazoans. Protistan community from the small (0.2–10 μm) and large (> 10 μm) size fractions exhibited a low similarity of 36.6% and shared 14.7% OTUs. About 63.8% of the OTUs were unique to the large fraction. Species from groups, such as Lepidodinium, Warnowia, Kareniaceae, Torodiniales, Phaeocystis, Chrysochromulina and Chrysophyceae, were enriched in the large fraction, indicating that they could be largely underestimated in studies that exclude the large-sized cells. Blocking primers are a promising tool to increase the efficiency for the characterization of protistan diversity in aquatic planktonic ecosystems.

Viruses and dissolved organic matter (DOM) are major top–down and bottom–up factors, respectively, which control the abundance and diversity of bacterial communities. However, the relative strength and interaction between these two regulators in different marine ecosystems are not fully understood. We applied viral reduction and cross‐transplantation of eutrophic estuarine and pristine coastal waters to examine the separate and combined effects of viruses and DOM on the bacterial community composition. We showed that bacteria were more abundant in the virus‐depleted groups, particularly in those from the eutrophic estuarine water (i.e., > 7 times higher, when compared with the virus‐rich groups). Cluster analyses showed that the bacterial communities in the virus‐rich groups were most similar to the in situ populations. Bacterial community richness, evenness and diversity were reduced at high DOM concentrations but were sustained by viruses. This allowed bacteria that were less vulnerable to viral infection to thrive (such as those in the families Alteromonadaceae, Rhodobacteraceae, Flavobacteriaceae, and SAR11), while suppressing those capable of fast growth at elevated DOM concentrations but which were more vulnerable to viral attack (such as those from the family Vibrionaceae). Our results showed that viral regulation occurring at the strain level can affect the bacterial community structure at higher phylogenetic levels. In addition, some less abundant groups, such as Vibrio and Alteromonas, might actually win the competition for resources, providing viral control is removed. Our results also showed that the population dynamics of certain bacterial groups was controlled by viruses, and the strength of such control was modulated by the concentration of DOM. We suggest that the composition of the bacterial community in aquatic environments results from the interactive regulation by DOM and viruses among other regulators.