Activated sludge of a landfill leachate treatment plant at a completed landfill (Tokyo, Japan) was collected on 20 July 2018. The landfill was in operation for several decades, during which time the leachate was treated by an activated sludge system in a sequencing batch mode. The sequencing batch reactor (SBR) had an effective volume of 740 m³ and received 74.2 m³/day of leachate on average. One cycle of the SBR operation lasted 2 h and consisted of 10 min for inflow, 45 min for mixing to ensure anoxic conditions, 15 min for aeration, and 50 min for settling and outflow of the treated water. Prior to entry of the influent to the SBR, the leachate was mixed with groundwater at a volumetric ratio of 3 (leachate): 1 (groundwater). Tables 1, 2 summarize the operating conditions, and the leachate characteristics and SBR performance, respectively. The organic carbon concentration was low (BOD < 15 mg/L and COD < 50 mg/L), whereas the ammonium concentration in the leachate was as high as 150 mg-N/L. Because of the imbalance of the carbon/nitrogen ratio, about 5.3 L of 50% v/v methanol was added during each inflow period as an external electron donor for denitrification. The methanol amount allowed a relative loading of COD/T-N (kg-COD/kg-N) of 4.38 ± 1.80, which is higher than the denitrification stoichiometry with methanol.

Twenty liters of activated sludge were taken from the SBR and immediately transferred to the laboratory. After thoroughly mixing the activated sludge, 15 mL of wastewater containing 22.2 mg-wet weight-activated sludge were used for DNA extraction by a DNA extraction kit (Fast DNA^TM Spin Kit for Soil, MP-Biomedicals, Santa Ana, CA, United States) according to the manufacturer’s protocol. The DNA concentration was measured using a spectrophotometer (NanoDrop 2000c, Thermo Scientific, Wilmington, DE, United States).

The sequencing and metagenomic assembly have been described in detail in a previous study (Yasuda et al., 2020). Briefly, sequencing was performed with an Ion Torrent (ION Torrent Ion S5) using an Ion 530^TM Chip. Torrent Suite v 4.4.2 (Thermo Fisher Scientific, Germany) was used for base calling and to run multiplexing according to the manufacturer’s protocol. After quality filtering of metagenomic reads and trimming both sides of the reads to ensure a Q20 quality score, 20.24 million reads were processed, giving 13,005,150 reads. Megahit v 1.1.3 (Li et al., 2015) was used at the minimum contig length of 1,000 bp for assembly, resulting in acquisition of 104.86 Mbp contigs. Mapping the reads to the assembly was accomplished with Bowtie2 v 2.3.4.2 (Langmead and Salzberg, 2013), and the subsequent identification of single-copy bacterial and archaeal genes was achieved using HMMER v 3.2 (Finn et al., 2011). Taxonomy was assigned to contigs using Centrifuge v 1.0.3 (Kim et al., 2016). The latest taxonomic information was used for classification (Waite et al., 2017, 2020; Parks et al., 2018). Default parameters were used for all software unless otherwise specified.

Amino-acid sequence data obtained by the HMMER procedure were used to query the Kyoto Encyclopedia of Genes and Genomes (KEGG; Kanehisa and Goto, 2000) with GhostKOALA v 2.1 (Kanehisa et al., 2016). The existence of functional genes and metabolic pathways associated with the production of valuable products by either microbial communities or a single species of indigenous bacteria were confirmed. Based on a literature survey for valuable products microbiologically synthesized from C1 compounds, the metabolic pathways and related functional genes for the production of ectoine (Pastor et al., 2010), PHB (Orita et al., 2014), carotenoids (astaxanthin and zeaxanthin; Ye and Bhatia, 2012; Zhang, 2018), and acetoin (Nielsen et al., 2010) were selected. Carotenoids are synthesized from acetyl-CoA to terpenoid backbone synthesis via either a mevalonate pathway or a non-mevalonate pathway (Supplementary Figure 1 and Supplementary Table 1). The majority of bacteria possess the latter, whereas some use the former pathway (Hoshino and Gaucher, 2018; Rico et al., 2019). Therefore, this study investigated both pathways.

This research assumed that all eubacteria possess central metabolisms, e.g., glycolysis and the TCA cycle. Essential functional genes and enzymes regulating biosynthesis of each valuable product are listed in Table 3. In addition to these compounds that C1-assimilating bacteria are potentially able to synthesize, 2,3-butanediol, which is a promising fuel that is chemically produced from acetoin (Nielsen et al., 2010; Ji et al., 2011), and lutein, which is a carotenoid mainly produced by plants and is known to play a role in vision enhancement (Ye and Bhatia, 2012; Nwachukwu et al., 2016; Zhang, 2018), were also selected.

Taxonomic classification results (Figures 1, 2 and Supplementary Figure 1) and repertories of metabolic pathways (Figure 3) were visualized with R (R Core Team, 2020), Rstudio (RStudio_Team, 2020), and Tidyverse package (Wickham et al., 2019).

The sequencing data generated by metagenomic analysis have been deposited in DDBJ with the following accession numbers: DRA009549, BioProject: PRJDB9252, and BioSample: SAMD00202538.

The SBR-based activated sludge system in this study intermittently received methanol as an external electron donor to replenish the limited amount of organic carbon constituents (average COD concentration of 43.9 ± 9.2 mg/L) from the closed landfill. The leachate contained a chloride concentration of 7,693 ± 1,767 mg/L, and the salinity was calculated at 1.23% by the correlation with electron conductivity and temperature (APHA et al., 1992), which is classified as brackish (Table 2). The average removal efficiencies of COD (including the added methanol) and nitrogen were 95.5 ± 1.6 and 97.2 ± 2.1%, and the volumetric removal rates of COD and T-N were 26.1 ± 8.8 g-COD/m³/day and 16.6 ± 3.1 g-N/m³/day, respectively. Nitrification was nearly complete in the SBR, resulting in an effluent ammonium concentration of below 0.05 mg N/L on average.

The class-level taxonomic composition of the activated sludge biomass is summarized in Figure 1. The class Gamma-proteobacteria was most abundant, accounting for up to 35% of the total population. This was followed by Alpha-proteobacteria (23%), and Actinobacteria (19%). The top 50 most abundant microbes at the genus and species levels and the corresponding relative abundances are shown in Figure 2 and Supplementary Figure 1, respectively. Methyloceanibacter caenitepidi, which is a marine methanol-assimilating bacterium, was the most abundant in the biomass (Supplementary Figure 1). Based on read numbers, its relative abundance was 1.3 times (1.2%) higher than that of the second most predominant species, Rhodopseudomonas palustris (0.9%). As for major guilds for ammonia oxidation, the genera Candidatus Nitrosopumilus and Nitrospira, known as ammonia-oxidizing archaea and complete ammonia-oxidizing bacteria, respectively, were detected (Figure 2 and Supplementary Figure 1).

The results of mapping our sample to reference metabolic pathways in the KEGG are summarized in Figure 3. The highest relative gene abundance was associated with carbohydrate metabolism (12%) and genetic information processing (12%), followed by protein families: genetic information processing (11%), and protein families: signaling and cellular processes (10%). The biosynthesis pathways for each valuable product after mapping with mixed culture data are illustrated in Figure 4. The mixed culture biomass in the activated sludge possessed a full set of functional genes to synthesize six of the valuable products, ectoine, PHB, zeaxanthin, astaxanthin, acetoin, and 2,3-butanediol (Figure 4 and Table 4). Comparing the hit numbers of the functional genes for ectoine, acetoin, and 2,3-butanediol for each enzymatic reaction step (Figures 4A,D,E) revealed a substantial difference in hit numbers. Specifically, the lowest and highest gene hit numbers for ectoine biosynthesis were ectA (177) and lysC (12,563) for conversion of L-2,4-diaminobutyrate into N-γ-acetyl-L2,4,-diaminobutyrate and of aspartate into aspartyl phosphate, respectively. These hit numbers were used to designate the degree of the rate-limiting reaction for biosynthesis of a valuable product (e.g., 0.014 = 177/12,563). Acetoin and 2,3-butanediol had the common functional gene alsD, which had a considerably lower hit number (115) than the others, leading to a relative abundance of the lowest/highest gene hit number being 0.0045. In contrast, PHB, zeaxanthin, and astaxanthin had relatively evenly-distributed gene hit numbers for their enzymatic reactions. For PHB biosynthesis, the hit numbers of phaA (2,532) and phaB (2,476), which encode acetyl-CoA acetyltransferase and acetoacetyl-CoA reductase, respectively, were lower than the one of phaC (14,491), which encodes poly (3-hydroxyalkanoate) polymerase. These hit numbers resulted in relative abundances of the lowest/highest gene hit number of 0.175 (phaA/phaC) and 0.171 (phaB/phaC), respectively. The same trend appeared with zeaxanthin and astaxanthin syntheses, resulting in a relative abundance of the lowest/highest hit number of 0.11. Terpenoid backbone biosynthesis pathways after gene-mapping are shown in Supplementary Figure 2. In addition, the results of gene-mapping of each valuable product to the KEGG reference pathways are shown in Supplementary Figures 3–5 and those of the terpenoid backbone biosynthesis are shown in Supplementary Figure 6. The mapping results shown in the figures indicate that the pathways for the production of the compounds are connected.

Linking the phylogeny with functions revealed that PHB, acetoin, and 2,3-butanediol can be genetically produced by a single bacterial species (Table 4). However, individual species of bacteria lack certain functional genes for the production of ectoine, lutein, zeaxanthin, and astaxanthin, implying that syntrophic interactions between different microbial species are needed for synthesis of these compounds. M. caenitepidi, Janthinobacterium sp. Marseille, and Ca. Filomicrobium marinum W harbored phaABC genes encoding enzymes for PHB biosynthesis, while Ca. promineofilum breve, Caldilinea aerophila, Lonsdalea britannica, and Orrella dioscoreae, possessed genes essential to acetoin production, i.e., ilvM, alsD, and butA (Figure 4). Among the bacteria capable of producing acetoin, Caldilinea aerophila, Lonsdalea britannica, and Orrella dioscoreae harbor butA and BDH, which genetically produce 2,3-butanediol from acetoin.

The MRM concept (Verstraete, 2007) paves the way for a paradigm shift in which wastewater can be regarded as a resource for valuable products. Applying a bottom-up approach based on gene-centric analysis to the MRM concept is essential for the feasibility of valuable product biosynthesis based on metabolic network information. Therefore, feasibility assessments based on the metagenome will help to establish innovative bioprocesses for valuable product biosynthesis. However, this approach has not been applied to a bioreactor for nitrification and denitrification in landfill leachate containing high concentrations of ammonia (e.g., several hundred mg N/L) and salt (above > 1%; Díaz et al., 2019; Shu et al., 2019). This study, therefore, advocates a rapid screening procedure based on the gene-centric analysis from metagenomic data in a simple manner. To the best of our knowledge, this is the first metagenomics study that investigates the metabolic potential for valuable product biosynthesis by activated sludge used for leachate treatment at a closed landfill site. The application of gene-centric metagenomic analysis confirmed that the activated sludge possessed the metabolic potential for the biosynthesis of six targeted products. The holistic and exclusive screening of metabolic pathway networks for valuable products based on gene-centric metagenomic analysis revealed (i) the potential for valuable product biosynthesis (Table 4) and (ii) inferred potential rate-limiting steps within the metabolic pathways (Figure 4). Although the taxonomic and functional information attained by the gene-centric analysis is not as profound as that by the genome-centric counterpart, the information provides the implication if a biomass sample possesses metabolic potentials for valuable product production or not. The gene-centric approach implemented in this study will facilitate the development of upcycling bioprocesses from landfill leachates in which organic carbon and nitrogen are abundant.

The gene hit number analysis suggests a potential rate-limiting step during the bio-transformations for value-added compounds, e.g., the conversion from L-2,4-diaminobutylate to N-γ-acetyl-L2,4-diaminobutyrate in a series of ectoine production from aspartate (Figure 4). A challenge in this study stems from the authenticity of the gene hit number as an indicator for rate-limiting steps in the metabolic pathways because the differences in gene expression affecting RNA and protein abundances, and the existence of pseudogenes may also determine enzymatic reactions. Although this challenge is under debate, comparing the gene hit numbers may provide rough measures of bottleneck enzymes in metabolic reactions. The credibility of quantifying gene hit numbers warrants future investigation.

This study demonstrated that the mixed microbial community in a landfill leachate treatment site harbored functional genes associated with the syntheses of ectoine, PHB, zeaxanthin, astaxanthin, acetoin, and 2,3-butanediol (Table 4). The presence of these genes suggests the potential for use of this activated sludge biomass in biosynthesis. As previously demonstrated, microbial consortia in activated sludge have been shown to produce short-chain fatty acids (Li et al., 2014), enzymes (Yu et al., 2009), polyhydroxyalkanoate (PHA; Lee and Yu, 1997), bio-flocculants (Wang et al., 2007), and fuels (Salama et al., 2017). As for ectoine synthesis, the analysis of the hit numbers of each functional gene suggested the conversion of L-2,4-diaminobutylate into N-γ-acetyl-L2,4-diaminobutyrate mediated by L-2,4-diaminobutyric acid acetyltransferase (EctA) as a rate-limiting step. Previous studies of EctA purified from halophilic bacteria revealed that suitable conditions for this process were pH 8.2–9.5, a temperature of 20°C, and a salt concentration of 2.3% (Ono et al., 1999; Richter et al., 2020), which do not coincide with the SBR conditions in this study (pH 7.35 ± 0.08, temperature of 28.8°C ± 1.4°C, and salinity of 1.23%). Moreover, caution should be taken when considering these findings because the conditions required to activate EctA in vitro and in vivo may not be the same. Accordingly, future studies to identify suitable environmental conditions for bacteria possessing EctA are needed.

The metagenomic analysis of the activated sludge used to treat landfill leachate allowed classification of the production potential for valuable products by (i) individual species of bacteria or (ii) mixed microbial communities (Table 4). The results demonstrated that individual species of bacteria present in the activated sludge of the landfill leachate treatment had the genetic potential to synthesize PHB, acetoin, and 2,3-butanediol among the seven products listed in Table 4.

Our metagenomic analysis unraveled that the species proximal to M. caenitepidi, Ca. Filomicrobium marinum W, and Janthinobacterium sp. Marseille were all found to possess the entire set of genes encoding for PHB synthesis. The metagenomic analysis of the taxonomy and functions revealed for the first time that M. caenitepidi, a marine methanol-assimilating bacterium, was dominant in the activated sludge of the landfill leachate treatment facility (Supplementary Figure 1). The presence of M. caenitepidi aside from the ocean has not been reported, except for one study (Yasuda et al., 2020). M. caenitepidi Gela4 was isolated from marine sediment near a hydrothermal vent containing methane in Kagoshima Bay, Japan (Takeuchi et al., 2014) and from a Belgian North Sea sediment (Vekeman et al., 2016). The genus Filomicrobium was detected in a methanol-fed denitrifying bioreactor from saline water (Rissanen et al., 2016). Ca. Filomicrobium marinum W (Henriques and De Marco, 2015), which was first isolated from North Atlantic surface seawater, is a methylotroph capable of potentially synthesizing PHB. Janthinobacterium isolates acquired from the Antarctic possess functional genes for PHB biosynthesis. Although they thrive in marine environments (Tan et al., 2020), the detection of Janthinobacterium sp. as a PHB producer in the landfill leachate treatment site makes sense as the presence of methanol and salinity provide a comparable environment. These species require greater ecophysiological descriptions to leverage their PHB synthesis potentials, which will be a follow-up aim.

Bacteria harboring phaABC store PHB intracellularly under feast-famine conditions (Inoue et al., 2018). Because the SBR in this study may induce feast–famine conditions over one cycle, these three bacteria accumulated PHB to attain energy for their preferential growth. Accordingly, the conditions likely exerted high selection pressure to preferentially enrich the sludge with these three potential PHB producers. Identification of their presence and functions paves the way for developing a PHB-producing process in a water/wastewater treatment system. However, this preliminary study was designed to identify the functions and compositions of an activated sludge microbiome treating landfill leachate, not to quantify water and intracellular constituents in the SBR. Therefore, future investigations to evaluate long-term SBR operation and the associated rates of production of valuable products in accordance with microbial population dynamics in the activated sludge are warranted.

This study also revealed that the bacteria listed in Table 4 harbor genes capable of producing acetoin and 2,3-butanediol. However, the ecological niches of taxonomic proxies for these bacteria are not consistent. Ca. Promineofilum and Caldilinea aerophile, which belong to the phylum Chloroflexi, are commonly found in municipal wastewater treatment plants (McIlroy et al., 2016) and hot spring sulfur turf mats (Sekiguchi et al., 2003), while Lonsdalea Britannica and Orrella dioscoreae are Gamma-proteobacteria isolated from the leaves of Quercus robur (Li et al., 2017) and Dioscorea sansibarensis (Carlier et al., 2017), respectively. Despite their genetic potentials to produce valuable products, in-depth physiological activity tests for actual production have not been conducted (Sekiguchi et al., 2003; Brady et al., 2012; McIlroy et al., 2016; Carlier et al., 2017; Li et al., 2017). Future research will be undertaken to confirm the product availability, yield, and production rate by bacterial species harboring genes related to PHB, acetoin, and 2,3-butanediol production.

In-depth analysis of mixed microbial community compositions and functions by gene-centric metagenomic analysis, proposed in this study, could facilitate a primary screening strategy for exploration and leveraging of promising microbial communities or species. Our gene-centric analysis confirmed that the activated sludge investigated did not harbor the functional genes encoding enzymes to synthesize lutein (Supplementary Figure 5). Moreover, inoculating the activated sludge would not succeed in enrichment with a microbial guild producing lutein. Pre-screening by gene-centric metagenomic analysis allows determination of whether a single microbe (I), a mixed microbial community (II), or no microbes (III) synthesize a valuable product (Figure 5). Such classifications may enable identification of the conditions required for a key mixed microbial community or an individual microbe to be enriched (Cases I and II Figure 5), ultimately allowing the development of an innovative microbial process. Accordingly, implementing bioaugmentation strategies (Laocharoen et al., 2015; Lin et al., 2018) or retrofitting treatment processes for valuable product biosynthesis, analogous to the concept of bioaugmentation technology where nitrifying bacteria are enriched in a tank installed in a sludge recycle line (Salem et al., 2003), are potential applications of this screening process (Figure 5). These findings will also enable research toward isolation or selective enrichment of promising microbes using suitable environmental conditions identified from metagenomic data. Moreover, if screening reveals no microbes have the desired properties, the method developed here will save a great deal of time and labor. The advocated scheme could be beneficial especially for engineers to bridge fundamental and applied research toward resource recovery and upcycling.