Similar Methanogenic Shift but Divergent Syntrophic Partners in Anaerobic Digesters Exposed to Direct versus Successive Ammonium Additions

Abstract

During anaerobic digestion (AD) of protein-rich wastewater, ammonium (NH₄⁺) is released by amino acid degradation. High NH₄⁺ concentrations disturb the AD microbiome balance, leading to process impairments. The sensitivity of the AD microbiome to NH₄⁺ and the inhibition threshold depend on multiple parameters, especially the previous microbial acclimation to ammonium stress. However, little is known about the effect of different NH₄⁺ acclimation strategies on the differential expression of key active microbial taxa. Here, we applied NH₄⁺ inputs of increasing intensity (from 1.7 to 15.2 g N-NH₄⁺ liters^-1) in batch assays fed with synthetic wastewater, according to two different strategies: (i) direct independent inputs at a unique target concentration and (ii) successive inputs in a stepwise manner. In both strategies, along the NH₄⁺ gradient, the active methanogens shifted from acetoclastic Methanosaeta to Methanosarcina and eventually hydrogenotrophic Methanoculleus. Despite shorter latency times, the successive input modality led to lower methane production rate, lower soluble chemical oxygen demand (sCOD) removal efficiency, and lower half maximal inhibitory concentration, together with higher volatile fatty acid (VFA) accumulation, compared to the independent input modality. These differential performances were associated with a drastically distinct succession pattern of the active bacterial partners in both experiments. In particular, the direct exposure modality was characterized by a progressive enrichment of VFA producers (mainly Tepidimicrobium) and syntrophic VFA oxidizers (mainly Syntrophaceticus) with increasing NH₄⁺ concentration, while the successive exposure modality was characterized by a more dynamic succession of VFA producers (mainly Clostridium, Sporanaerobacter, Terrisporobacter) and syntrophic VFA oxidizers (mainly Tepidanaerobacter, Syntrophomonas). These results bring relevant insights for improved process management through inoculum adaptation, bioaugmentation, or community-driven optimization. IMPORTANCE Anaerobic digestion (AD) is an attractive biotechnological process for wastewater bioremediation and bioenergy production in the form of methane-rich biogas. However, AD can be inhibited by ammonium generated by protein-rich effluent, commonly found in agro-industrial activities. Insights in the microbial community composition and identification of AD key players are crucial for anticipating process impairments in response to ammonium stress. They can also help in defining an optimal microbiome adapted to high ammonium levels. Here, we compared two strategies for acclimation of AD microbiome to increasing ammonium concentration to better understand the effect of this stress on the methanogens and their bacterial partners. Our results suggest that long-term cumulative exposure to ammonia disrupted the AD microbiome more strongly than direct (independent) ammonium additions. We identified bioindicators with different NH₄⁺ tolerance capacity among VFA producers and syntrophic VFA oxidizers.

Keywords: Methanoculleus; acclimation; acetate; adaptation; ammonia; disturbance; methane; methanogen disturbance; microbial diversity; perturbation; stepwise; stress adaptation; syntrophic; syntrophic acetate oxidation (SAO); syntrophs; volatile fatty acids (VFA).

FIG 1

Experimental setup for experiment 1 (independent ammonium additions) and experiment 2 (successive ammonium additions). The N-NH₄⁺ concentration corresponding to each condition is given in Table 1. At each NH₄⁺ level, three replicated vials were sacrificed for biochemical and microbiological analysis at the time of maximal methanogenic activity, while the others were maintained (experiment 1) or transferred to the next level (experiment 2).

FIG 2

(A) Methane production rate (MPR) in experiment 1 (independent N inputs) and experiment 2 (successive N inputs) for each ammonium condition. MPR was calculated as the maximal slope of methane production kinetics along time (Fig. S1). Mean and standard deviation were calculated on 6 and 6 to 24 replicates for experiments 1 and 2, respectively. (B) Volatile fatty acid (VFA) concentration in experiments 1 and 2 (the N-NH₄⁺ concentration corresponding to each sample number is given in Table 1). VFA concentrations were measured at the time of maximal methanogenic activity (end of exponential phase), and mean and standard deviation were calculated from triplicated incubation vials. (C) Abundances of present (DNA-based) bacteria and archaea for each N-NH₄⁺ level in experiment 1 and 2, represented as dots. Mean and standard deviations were calculated from technical duplicates on biological duplicates (i.e., n = 4). For each condition, the transcript-to-gene ratio (cDNA/DNA), indicative of active expression level, is represented as bars.

FIG 3

Taxonomic affiliation of the 11 most abundant phyla, obtained from the 16S rRNA gene (DNA) and transcripts (cDNA) sequences. The group “Other” contains all phyla with relative abundance lower than 0.5% of the community on average over all samples.

FIG 4

Principal coordinate analysis (PCoA) ordination of the present (DNA-based) and active (RNA-based) microbial communities from experiment 1 (independent N inputs) and experiment 2 (successive N inputs). Significant samples’ clustering was verified by nonparametric PERMANOVA (adonis function, 99 permutations, P < 0.001) and represented by ellipses built at 80% confidence interval level. All replicates are shown.

FIG 5

Principal coordinate analysis (PCoA) of the active microbial community in experiment 1 (independent ammonium inputs, panels A and C) and experiment 2 (successive ammonium inputs, panels B and D) based on Bray-Curtis dissimilarities. For each sample, the input ammonium concentration is indicated by the symbol color gradient. Duplicates are aggregated by the mean. (A and B) Correlation of the ordination with environmental and functioning variables (significance tested by envfit function, P value represented by the arrow color gradient). (C and D) Representation of the 50 most abundant ASVs from the active community of each experiment, aggregated at the family level and colored by phylum.

FIG 6

(A) Heatmap showing the normalized abundance (z-score, by row) of differentially expressed transcripts belonging to Halobacterota and Firmicutes phyla. The differentially expressed transcripts were selected by DESeq2 (adjusted P value of <0.05) and then aggregated at the genus level (or at the lowest available taxonomic level for unclassified genera, as identified by f_, family; o_, order; c_, class; p_, phylum). Each column represents the mean of biological duplicates. Samples and differentially expressed genera are arranged by hierarchical clustering based on their differential abundance patterns using Bray-Curtis distance. The vertical color code on the right side represents the taxonomic affiliation at the class level. The horizontal-colored bars at the top represent the ammonium concentration (from 1.7 to 15.3 g N-NH₄⁺ L¹) and the ammonium input modality (experiment1 and experiment 2). (B) Boxplot representing the distribution of the log-transformed relative abundances of each differentially expressed genus in the different samples.

FIG 7

Co-occurrence networks of the 50 most abundant ASV aggregated at the family level, for (A) experiment 1 (independent inputs) and (B) experiment 2 (successive inputs), based on Pearson correlations. The thickness of the lines is proportional to the Pearson correlation coefficient R. Only correlations with |R| of >0.8 and P value of <0.05 are shown. Positive and negative correlations are shown by red and green lines, respectively. The nodes are colored according to the taxonomic affiliation at the phylum level.