Bacterioplankton Communities in Dissolved Organic Carbon-Rich Amazonian Black Water

ABSTRACT The Amazon River basin sustains dramatic hydrochemical gradients defined by three water types: white, clear, and black waters. In black water, important loads of allochthonous humic dissolved organic matter (DOM) result from the bacterioplankton degradation of plant lignin. However, the bacterial taxa involved in this process remain unknown, since Amazonian bacterioplankton has been poorly studied. Its characterization could lead to a better understanding of the carbon cycle in one of the Earth’s most productive hydrological systems. Our study characterized the taxonomic structure and functions of Amazonian bacterioplankton to better understand the interplay between this community and humic DOM. We conducted a field sampling campaign comprising 15 sites distributed across the three main Amazonian water types (representing a gradient of humic DOM), and a 16S rRNA metabarcoding analysis based on bacterioplankton DNA and RNA extracts. Bacterioplankton functions were inferred using 16S rRNA data in combination with a tailored functional database from 90 Amazonian basin shotgun metagenomes from the literature. We discovered that the relative abundances of fluorescent DOM fractions (humic-, fulvic-, and protein-like) were major drivers of bacterioplankton structure. We identified 36 genera for which the relative abundance was significantly correlated with humic DOM. The strongest correlations were found in the Polynucleobacter, Methylobacterium, and Acinetobacter genera, three low abundant but omnipresent taxa that possessed several genes involved in the main steps of the β-aryl ether enzymatic degradation pathway of diaryl humic DOM residues. Overall, this study identified key taxa with DOM degradation genomic potential, the involvement of which in allochthonous Amazonian carbon transformation and sequestration merits further investigation. IMPORTANCE The Amazon basin discharge carries an important load of terrestrially derived dissolved organic matter (DOM) to the ocean. The bacterioplankton from this basin potentially plays important roles in transforming this allochthonous carbon, which has consequences on marine primary productivity and global carbon sequestration. However, the structure and function of Amazonian bacterioplanktonic communities remain poorly studied, and their interactions with DOM are unresolved. In this study, we (i) sampled bacterioplankton in all the main Amazon tributaries, (ii) combined information from the taxonomic structure and functional repertory of Amazonian bacterioplankton communities to understand their dynamics, (iii) identified the main physicochemical parameters shaping bacterioplanktonic communities among a set of >30 measured environmental parameters, and (iv) characterized how bacterioplankton structure varies according to the relative abundance of humic compounds, a by-product from the bacterial degradation process of allochthonous DOM.

Suppl. Figure 2: Rarefaction plots of the samples for each sampling site, for the global bacterioplankton. The rarefaction analysis was based on the Shannon diversity for each sample group, according to the sequencing depth (number of sequences used). Figure 3: Shannon diversity plots of the samples for the taxonomic structure of global and transcriptionally-active bacterioplankton. Samples are grouped according to their water color, and are colored according to their sampling site of origin.

Suppl.
Suppl. Figure 4: Rarefaction plots of the samples for each sampling site, for transcriptionallyactive bacterioplankton. The rarefaction analysis was based on the Shannon diversity for each sample group, according to the sequencing depth (number of sequences used) Suppl. Figure 5: Fluorescence Excitation Emission Scans (FEEMs) of the FDOM from all sites.
Suppl. Figure 6: Average relative abundance of Polynucleobacter ASVs in each water type for global bacterioplankton (a) and for transcriptionally-active bacterioplankton (b).

Relative abundance
Relative abundance Suppl. Figure 8: Flowchart of the pipeline used to produce the functional reference database.

Supplementary Tables
Suppl. The groups used for the PERMANOVA tests consisted of "Lake" and "River" ecosystems.
Here, "Water color" and "DNA or RNA" variables characterize the subset of samples that were used for the tests. "df res" means the number of degrees of freedom for residuals.  20 19.00 6.00 79.10 *1: "Chl a" means the concentration of chlorophyll a; "Phaeopig." means the concentration of phaeopigments; "Chla/DOC" is a ratio of the concentration of chlorophyll a divided by the concentration of DOC; "Temp. °C" means the temperature in ° Celsius; "Cond. uS" means the conductivity in microsiemens; "% O2" means the percentage of saturation of dissolved oxygen.

Water residence time
Our results suggest that water residence time significantly influenced the taxonomic structure and transcriptional activity of the bacterioplankton communities, both in black and white water environments (results in Suppl. Table 1). Although water residence time was not specifically measured in this study, ecosystems were classified as lakes (longer residence time) and rivers (shorter residence time) and enabled us to perform PERMANOVA analyses based on the type of ecosystem that was studied (see Table 1 In the Amazonian River system, water residence time varies according to the intense seasonality experienced by these ecosystems: Between the rainy season (January to June) and the dry season (July to December) the Amazon water level can vary of several meters -a variation of 29 meters was recorded in June 2021 (Espinoza et al. 2022). The temporal variability modulates the connectivity between environments and thus the water residence time. Overall, the effects of this seasonality on watercourse residence time, connectivity, and chemical profile likely interferes with bacterioplankton communities and merits further investigation.

Pathways of humic compounds' degradation
The set of genes that was detected in Amazonian Polynucleobacter, Methylobacterium and Acinetobacter (Fig. 7) suggests that they possess the genomic potential to be involved in the degradation of humic acids or their by-products, via a derivative of the b-aryl ether degradation pathway for diaryl residues.
Polynucleobacter: The Polynucleobacter detected contained the glutathione S-transferases ligF/ligG (GST, K00799), performing one of the main reactions of this funneling pathway leading to the production of vanillate. The O-demethylation of vanillate is still unresolved based on the gene set detected, but could involve a demethylase similar to ligM (K15066), since we detected genes coding for enzymes associated with the metabolism of protocatechuate (PCA), the product of the ligM reaction, such as 3-oxoadipate enol-lactonase/4-carboxymuconolactone decarboxylase (pcaL, K14727). The degradation of these substances appears to conclude in an extradiol 4,5-PCA ring meta cleavage as suggested by the presence of genes coding for enzymes ligI (a 2-pyrone-4,6dicarboxylate lactonase, K10221) and ligK (a 4-hydroxy-4-methyl-2-oxoglutarate aldolase, K10218) associated with this pathway (de Gonzalo et al. 2016).

Methylobacterium:
The set of genes found in Methylobacterium suggests that, like Polynucleobacter, this clade could degrade humic substances via a derivative of the b-aryl ether degradation pathway for diaryl residues, and concludes in an extradiol 4,5-PCA ring meta cleavage producing pyruvate and oxaloacetate.