in

Meta-analysis cum machine learning approaches address the structure and biogeochemical potential of marine copepod associated bacteriobiomes

CAB diversity between the copepod genera

Calanus spp. are filter feeders and mostly herbivores, but do feed on ciliates and other heterotrophic protists during reproduction and energy shortfall38,39. This may be the reason for their high H index. Most of the gene sequences used for this meta-analysis were from Calanus finmarchicus; however, Centropages sp. feeds on different sources, from microalgae to fish larvae40. Acartia spp. are primarily omnivorous (with a high degree of carnivore behaviour), feeding on phytoplankton, rotifers, and occasionally ciliates41, whereas Temora spp. frequently switches its feeding behaviour, i.e., from omnivore to herbivore, based on season and on food availability42. The bacterial alpha diversity analysis in the Temora spp. revealed a significantly lower Shannon diversity. However, in an earlier study, no difference was reported in alpha diversity between the Temora sp. and Acartia sp.37. This can be explained based on the source of copepods involved for the study by Wega et al.37, which was based only on a single source, i.e., the central Baltic sea; however, in our case the CAB sequences for Acartia spp. were from the central Baltic sea37 as well as the Gulf of Maine10. The occurrence of high Faith’s_PD in Pleuromamma spp. may be due to their range distribution in the water column, and few species within Pleuromamma spp. are known to migrate vertically11,43, or possibly due to their food uptake, which includes phytoplankton, microzooplankton (ciliates and flagellates) and detritus11,44.

The consensus phylogram revealed that, at the genera level, Calanus spp. was phylogenetically closer to Pleuromamma spp. and formed two distinct clusters in the PCoA plot. Furthermore, the difference in dissimilarity percentage of CAB between Pleuromamma spp. and Calanus spp. may be attributed to the difference in vertical migration, life stages and feeding behaviour between the two copepod genera. Pleuromamma spp., an omnivorous feeder11,44, can migrate vertically up to 1000 m11,43 whereas Calanus sp., mostly herbivores but occasional omnivores36,37, can migrate up to 600 m45,46. This may also be due to the difference in the life stage of Calanus sp. (the microbial communities varied between diapausing and active feeding)2.

ANCOM

In an early report, bacterial members belonging to the Gammaproteobacteria were observed to be dominant in Calanus finmarchicus, followed by members of Alphaproteobacteria10. However, in the present ANCOM, the presence of Gamma and Alphaproteobacteria were equal (three genera each) in Calanus spp. (Fig. 3). Similar to our results, the unclassified genus of Rhodobacteraceae was reported to be abundant in Acartia longiremis10. Colwelliaceae was reported to be abundant in Calanus finmarchicus10; however, in the present analysis, family Colwelliaceae was found in a high percentage in Centropages sp.. An abundance of Flavobacteriaceae was observed, along with phytoplankton and diatoms in the gut of Calanus finmarchicus containing food2, whereas Sedinimicola sp. (Flavobacteriaceae) was observed to be dominant in Acartia longiremis, Calanus finmarchicus and Centropages hamatus10. In addition, Dorosz et al.47 reported that Flavobacterium was more dominant in Temora longicornis than in Acartia tonsa, whereas, in our case, Flavobacteriaceae was found in a high percentage in Calanus spp.. Upon comparison of the present ANCOM and previous reports, Pseudoalteromonas sp. appeared in high percentage not only within Centropages sp.10 but also in consistent and abundant bacteria in Acartia sp., and Calanus sp. The prevalence of Pseudomonas has been observed in Pleuromamma sp.11, whereas this was not the case in our analysis (Fig. 3). Similarly, Cregeen11 analysed the bacteriobiome of Pleuromamma sp. and observed the dominance of Alteromonas, but, from our meta-analysis, a higher abundance of Alteromonas was observed in Centropages sp. compared to five other genera, including Pleuromamma spp. (Fig. 3).

From our analysis, Nitrosopumilus was observed contain a high amount of Temora spp., but the abundance of Nitrosopumilus was reported to show no difference between the particle-associated in the water column and within Temora sp.37; thus, the high percentage observed in our analysis may be due to the exchange of Nitrosopumilus from seawater. Vibrionales was identified as a core member in the gut of Pleuromamma spp.1, similar to the present analysis, wherein Vibiro percentage was found to be high in the CAB of Pleuromamma spp.. The copepods were reported to have a selective niche of Vibrio capable of degrading chitin1,48. In the present analysis, seven bacterial taxa were found to be in high percentages in Centropages sp. and, among those seven, four taxa belong to the Gammaproteobacteria. A high proportion of Gammaproteobacteria in Centropages sp. was also reported previously10.

Machine learning-based prediction

The masking effect of the abundant bacterial community associated with the copepod diet and ambient water column should not hinder the detection of core OTUs, as evidenced by previous studies1,2. QIIME2 core_abundance algorithms used in the present study did not predict single bacterial s-OTUs (data not presented). Hence, we used machine learning approaches to detect important core s-OTUs specific to copepod genera.

From our SML classifier results, the important s-OTUs predicted in Calanus spp. and Pleuromamma spp. were found to have high prediction accuracy (area under the curve (AUC) = 1.00). Therefore, we discuss the important s-OTUs predicted for these two copepod genera (Calanus spp. and Pleuromamma spp.). To begin with, among the important s-OTUs predicted in Calanus spp. from the present analysis (both SML models: RFC and GBC), Gammaproteobacteria was a dominant member (15 and 9 s-OTUs from RFC and GBC, respectively) followed by Alphaproteobacteria, which represents 6 and 3 s-OTUs from RFC and GBC, respectively. This observation was similar to that in an earlier study, where Gammaproteobacteria and Alphaproteobacteria were reported as core OTUs in Calanus finmarchicus2. In addition, within the Gammaproteobacteria, seven (RFC) and five (GBC) s-OTUs representing the Acinetobacter (Moraxellaceae) were predicted as important s-OTUs in the present study, similar to an earlier study in which Moraxellaceae was reported to be closely associated with Calanus finmarchicus10. Moreover, four s-OTUs of Acinetobacter (Moraxellaceae) were also reported as core OTUs in Calanus finmarchicus2. In addition to the present analysis, three s-OTUs from both SML classifiers (RFC and GBC) belonging to Vibrio shilonii were predicted as important s-OTUs in Calanus spp.. Comparably, four OTUs of Vibrionaceae (three OTUs of Vibrio sp. and one similar to Vibrio harveyi) were observed in Calanus finmarchicus2.

In the present SML analysis, one genus Bradyrhizobium (order Rhizobiales), was predicted as an important s-OTU in Pleuromamma spp. by GBC classifiers. Moreover, in the present ANCOM, Bradyrhizobium was found in a high percentage within Pleuromamma spp.. This Bradyrhizobium is also known to contain nifH gene, as they usually occur in seawater49 and SML-GBC also predicted this genus as an important s-OTU in Calanus spp.. Bradyrhizobiaceae was also found to be the most abundant OTU, contained in 79 of the total 137 sequences in the negative control in a similar analysis1. Thus, in the case of Bradyrhizobium, a further investigation is required in order to come to a meaningful conclusion.

Moreover, in a previous study, order Vibrionales was also predicted as a core member (based on presence/absence) in Pleuromamma spp.1. The genus Pseudoalteromonas was also already reported as occurring in high abundance in Pleuromamma sp.11. However, in the present analysis, GBC predicted five s-OTUs of Pseudoalteromonas as important s-OTUs in Pleuromamma spp., whereas RFC predicted two s-OTUs of Pseudoalteromonas as important s-OTUs in Acartia spp., Calanus spp., and Centropages sp. (Fig. 4e). This is similar to Pseudoalteromonas, which is reported as a constant and stable OTU in Acartia sp.37, Calanus sp.2 and Centropages sp.10. Thus, it is unwise to consider Pseudoaltermonas as being specific to one copepod genera.

In the present study, the GBC model predicted three s-OTUs of Alteromonas and two s-OTUs of Marinobacter as important ones in Pleuromamma spp., and ANCOM also showed that the genus Marinobacter proportion was high in Pleuromamma spp.. Comparably, both Alteromonas and Marinobacter were reported as common in Pleuromamma sp.11. Though the abundance of genus Sphingomonas was low, it was reported to appear consistently in Pleuromamma sp.11, and our analysis predicted this genus as an important s-OTU of Pleuromamma spp. (from GBC) (Fig. 4f).

In the present study, the GBC model predicted Limnobacter as an important s-OTU in Pleuromamma spp., and ANCOM also showed that the proportion of genus Limnobacter was high in Pleuromamma spp.. Moreover, in a previous study, Limnobacter was reported to occur in high abundance in, as well as being unique to, copepods (Pleuromamma spp.)11. Also, the genera Methyloversatilis was reported to be low in abundance in Pleuromamma spp., whereas the SML-GBC model in this study predicted this genus to be an important s-OTU in Pleuromamma spp. (Fig. 4f). The order Pseudomonadales was reported as a core member in Pleuromamma spp.1; however, our GBC model predicted the bacterial genera Enhydrobacter (Pseudomonadales) as an important s-OTU in Pleuromamma spp. (Fig. 4f). In addition, from ANCOM, this genus Enhydrobacter was found in high percentage in Pleuromamma spp., but was also reported to be high in proportion in calanoid copepods6. One another important s-OTU predicted in Pleuromamma spp. by our GBC model was Desulfovibrio, and ANCOM also showed that the proportion of genus Desulfovibrio was found to be high in Pleuromamma spp..

HTCC2207 (Gammaproteobacteria) was predicted as an important s-OTU in Calanus spp. by both SML models. Also, from ANCOM, HTCC2207 was found in a high percentage in Calanus spp.. HTCC2207 is usually more abundant in seawater, and has been reported as present in Acartia longiremis., Calanus finmarchicus and Centropages hamatus with a full gut10. Due to their known proteorhodopsin gene and being free water—living bacteria50, the probability of detecting this bacterium in the copepod gut may be determined by food ingestion.

Sediminibacterium (Chitinophagaceae) was reported to be in low abundance but regularly present in Pleuromamma sp.11. However, in the present analysis, the RFC model predicted Sediminibacterium as important s-OTUs in Acartia spp., Calanus spp. and Temora spp. (Fig. 4e,f), whereas the GBC model predicted Sediminibacterium as important s-OTUs in Acartia spp. and Temora spp. (Fig. 4). Chitinophagaceae was reported to be associated with calanoid copepods in the North Atlantic Ocean6. Earlier studies showed that the genus Photobacterium (Phylum: Proteobacteria) was abundant in Pleuromamma sp.11, Centropages sp.10, and Calanus finmarchicus2. Herein, Photobacterium was detected as an important s-OTU in Calanus spp. by the RFC model only. Furthermore, in the present analysis, Nitrosopumilus was predicted as an important s-OTU in Acartia spp. and Temora spp. by both the SML models, and this genus was also reported to be in high percentage in Acartia sp. and Temora sp.37.

Furthermore, RFC predicts Pelomonas as an important s-OTU in Acartia spp., Centropages sp. and Calanus spp.. However, in a previous study, Pelomonas was ruled out as a core OTU in Calanus spp.2. The GBC predicted two s-OTUs of RS62 and one s-OTUs of Planctomyces as important ones in Acartia spp., and Temora spp.. RS62 belongs to the order Burkholderiales, and though this order was reported to be abundant, abundance varied between individual copepods (Acartia sp. and Temora sp.)37. Burkholderiales was also reported as a main copepod-associated community9. However, in the present study, the genus Comamonas belonging to Burkholderiales was predicted as an important s-OTU in Acartia spp., and Temora spp. by both SML models.

Approximately 25 taxa detected by the RFC approach were also found in high percentages from ANCOM. Among them, five s-OTUs, viz., Anaerospora, Micrococcus, Micrococcus luteus, Vibrio shilonii and Methylobacteriaceae, were predicted as important s-OTUs in Calanus spp. in our report, for the first time (Fig. 4e). From the 28 taxa detected by the GBC model, four s-OTUs, viz., Phaeobacter, Acinetobacter johnsonii, Vibrio shilonii, and Piscirickettsiaceae, were predicted as important s-OTUs in Calanus spp. in our report, for the first time (Fig. 4f). In addition, eight s-OTUs, viz., Marinobacter, Limnobacter. Methyloversatilis, Desulfovibrio, Enhydrobacter, Sphingomonas, Alteromonas and Coriobacteriaceae, were predicted as important s-OTUs in Pleuromamma spp. in the GBC model, for the first time.

Potential biogeochemical genes of CAB and their variation and abundance

Bacterial communities exploit copepods as microhabitat by colonising copepods’ internal and external surfaces, and mediate marine biogeochemical processes9. CABs also metabolise organic compounds, such as chitin, taurine, and other complex molecules in and around the copepod, which may be a hotspot for the biogeochemical process9. In an earlier analysis, potential functional genes in the water column of the Southern Ocean were processed using Parallel-Meta3 software51; herein, we have used a more advanced PICRUSt2 analysis to screen for the potential functional genes.

Methanogenesis

In the present analysis, the bacterial taxa involved in methane production, viz. methanogenesis, methylphosphonate, DMSP and DMSO, were observed in all copepod genera but relative proportion varied between genera. A similar observation in Acartia sp. and Temora sp. has been reported37.

In the present analysis, we found that CAB has a complete set of aerobic methanogenesis genes (PhnL, M, J, H and G) which convert methylphosphonate (MPn) to methane (CH4)52. Some copepods, like Acartia sp. and Temora sp., were reported to associate with bacteria involved in CH4 production from MPn37. De Corte et al.9 suggested that different copepod species have different CAB, and only some copepods have the specific CAB for methanogenesis and other biogeochemical cycles.

A previous study (with 14 C-labelled experiments) observed high methane production in Temora longicornis compared to Acartia spp.53. In addition, the methanogenic archeae i.e., Methanobacterium bryantii-like sequences, Methanogenium organophilum, Methanolobus vulcani-like sequences and Methanogenium organophilum were noted in Acartia clausi and Temora longicornis faecal pellets54. In the present study, we observed that Pleuromamma spp. has a high proportion of the mcrA gene (Fig. S2).

T. longicornis fed with a high content of TMA-/DMA-rich phytoplankton produced the maximum amount of CH4, suggesting that this production may be due to the micro-niches inside the copepods55. However, in our analysis, CAB of Pleuromamma spp. was found to have a high proportion of the dmd-tmd gene.

In our meta-analysis, Acartia spp. was found to have a high proportion of the dmdA gene. The taxa detected in the present study, such as Pelagibacteraceae, some Alpha and Gammaproteobacteria, are known to have dmdA genes56.

Copepods feeding on phytoplankton liberate DMSP, which, in turn, is utilised by the DMSP-consuming bacteria in the gut (Acartia tonsa), leading to methane production57. Moreover, the methane enrichment in the Central Baltic Sea is due to the dominant zooplankton Temora longicornis feeding on the DMSP-/DMSO-rich Dinophyceae, resulting in methane release53.

Instead of analysing faecal pellets57 and anaerobic incubation experiments58, further research should also consider CAB-mediated aerobic methanogenesis as one factor with which to solve the ‘ocean methane paradox’.

Methanotrophic potential of CAB

The present analysis showed that the CABs of Pleuromamma spp. and Centropages sp. were had a high proportion of methanol dehydrogenase genes (mxaF and mxaI) (Fig. S2). This may be due to the presence of Proteobacteria that involves methane oxidation, viz., Beijerinckiaceae, Methylococcaceae, Methylocystaceae and Verrucomicrobia (Supplementary File Table S3)59.

Assimilatory sulphate reduction

A relative abundance of taxa such as Synechococcus and the Deltaproteobacterial family (unclassified genera in Desulfovibrionaceae), Rhodobacteraceae and Flavobacterium (Supplementary File Table S3) were observed in the CAB of Temora spp., which may be responsible for the ASR pathway, as these taxa are known to have ferredoxin-sulphite reductase activity (Supplementary File Table S3).

Nitrogen fixation

A high abundance of nifH gene was reported in copepods collected from the coastal waters of Denmark (Øresund) (mostly contributed by Acartia spp.), with Vibrio spp. as dominant members16. However, in the present study, the nifH gene was found to be high in the CAB of Pleuromamma spp. (Fig. S4), and one should note that this may be due to the high abundance of genus Vibrio in the CAB of Pleuromamma spp. (Supplementary File Table S3). Vibrio attached to the exoskeleton and gut lining of copepods60 using chitin as both a carbon and energy source was previously reported10. Furthermore, copepods are reported to be a hotspot for nitrogen fixation at a rate of 12.9–71.9 μmol N dm−3 copepod biomass per day16. The abundance of nifH gene in the CAB of Pleuromamma spp. may be due to the presence of genera including Synechococcus, Prochlorococcus, Bradyrhizobium, Microcystis, and Trichodesmium (Supplementary File S3).

Denitrification

In our analysis, the CAB of Temora spp. were found to have the highest proportion of napA and napB genes (Fig. S4), followed by Pleuromamma spp., whereas an abundance of napA and narG genes were reported in North Atlantic copepods contributed by Calanus sp. and Paraeuchaeate sp.9. However, in the present analysis, the CAB of Temora spp. was found to have a high proportion of narG (Fig. S4). Bacterial genera including Pseudoalteromonas, Actinobacterium and Shewanella also contain the nirS gene, as reported in both live and dead Calanus finmarchicus14. Likewise, from our analysis, both Pseudoalteromonas and Actinobacteria were found in Calanus spp.. A metagenome analysis of copepod-associated microbial community reported them having genes responsible for denitrification and DNRA9.

Anaerobic nitric oxide reduction

Families including Aeromonadaceae and Enterobacteriaceae were observed in the CAB of Pleuromamma spp. and Calanus spp., in relatively higher proportion than in other copepods. The genera Aeromonas (family Aeromonadaceae)61 and Escherichia coli (family Enterobacteriaceae)62 are known to contain norV genes. The presence of these bacterial taxa in Pleuromamma spp. may be due to feeding of ciliates, flagellates, and detritus particles11,44. This may be one reason for a high proportion of norV and norW genes in these copepods (Fig. S4).

Carbon processes

Bacterial taxa like Colwelliaceae10,63Flavobacterium, Arthrobacter, Serratia, Bacillus, Enterobacter, Vibrio64, Pseudoalteromonas63 and Achromobacter65 produce chitinase. The presence of chitinase gene in CAB is unsurprising, as their foregut and hindgut are both made up of chitin11. The overall outline of CAB-mediated biogeochemical pathways is represented in Fig. 6.

Figure 6

Overall representation of the potential functional genes of CAB involved in biogeochemical cycles. The circle and colour represent the copepod genera contained in high proportion for that particular biogeochemical process.

Full size image

Role of CAB in iron remineralization

Pleuromamma spp. carries a similar proportion of ferric iron reductase (fhuF) and ferrous iron transport protein A (feoA) genes (Fig. S6a,b). The presence of a high proportion of ferric iron reductase gene fhuF in Pleuromamma spp. requires detailed investigation. It was reported that acidic and low-oxygen conditions in the copepod gut may assist iron dissolution and remineralisation, forming soluble Fe(II)13,66. This increases the iron bioavailability in the surroundings, promoting phytoplankton growth66. In addition, bacterial community associated with the zooplankton, such as Bacteroidetes, Alphaproteobacteria and Gammaproteobacteria, are known to carry genes involved in iron metabolism9.

In an early study on Thalassiosira pseudonana fed to Acartia tonsa, iron was found in the faecal pellets67. However, in the present analysis, Acartia spp. was found to have a lower proportion of the feoA gene compared to Temora spp. and Pleuromamma spp.. Moreover, genes involved in iron metabolism were reported to be high in zooplankton-associated microbiome9.

The differential iron contributions of different copepod genera were unknown until now. For organisms that must combat oxygen limitation for their survival (Pleuromamma spp.), pathways for the uptake of ferrous iron are essential. Nevertheless, the meta-analysis performed here showed that Pleuromamma spp. may be a significant contributor to both iron bioavailability and nitrogen fixation.

CAB as a source of cyanocobalamin-synthesising prokaryotes

Organisms within all domains of life require the cofactor cobalamin (vitamin B12), which is usually produced only by a subset of bacteria and archaea68. Previous studies reported that the cobalamin in ocean surface water is due to de novo synthesis by Thaumarchaeota. Moreover, few members of Alphaproteobacteria, Gammaproteobacteria and Bacteroidetes genomes were reported to contain the cobalamin-synthesising gene68. In our analysis, the CAB of Temora spp. was found to have a high proportion of Thaumarchaeota, whereas Alpha-gammaproteobacteria content was found to be high in the CAB of Acartia spp., Calanus spp. and Pleuromamma spp.. In this regard, further studies on CAB diversity from different ocean realms would shine a light on the actual potential of CAB in global biogeochemical cycles.


Source: Ecology - nature.com

Reducing inequality across the globe and on campus

George Shultz PhD ’49, renowned statesman and former professor, dies at 100