Genome-centric metagenomic insights into the role of Chloroflexi in anammox, activated sludge and methanogenic reactors

Applying a series of molecular biology tools, the diversity and metabolic potential of the phylum Chloroflexi was studied in three different wastewater treatment systems: a methanogenic reactor (anaerobic, C-removal), an activated sludge reactor (aerobic, C-removal) and an anammox reactor (anaerobic, N-removal).

Diversity of Chloroflexi members in the different wastewater reactors according to 16S rRNA gene amplicon sequencing

The phylum Chloroflexi was detected in all reactor samples. Regarding the relative abundance higher values were detected in the anammox reactor samples compared to the methanogenic and aerobic reactors samples (26.9—33.5% for anammox, 4.9 -22% for activated sludge and 0.8—10.7% for methanogenic reactors) (Fig. 1).

Fig. 1figure 1

Boxplots showing the relative abundance of the 16 most abundant phyla in the total community according to the 16S rRNA gene amplicon sequence in each reactor type

MiDAS database, specifically designed to analyze microbial communities from activated sludge systems and methanogenic reactors [63], was used to classify sequences at different taxonomic levels (non-described microorganisms are labeled with a number according to the MiDAS taxonomy). The results confirmed that Anaerolineae class widely dominated the Chloroflexi community in all reactors (Fig. 2a), which was in accordance with results previously reported in aerobic, anaerobic and anammox reactors [2, 4, 17, 26, 27, 29, 36, 64,65,66,67,68]. Most of the genera detected were specifically associated with one type of reactor.

Fig. 2figure 2

Heatmaps showing the relative abundance within the phylum Chloroflexi according to the 16S rRNA gene amplicon sequence: A) at class level, B) at genus level (30 most abundant genus) C) at species levels (30 most abundant species). Non-classified microorganisms are labeled with a number according to the MiDAS taxonomy. Black triangles in panels A and B mark genus and species belonging to Dehalococcoidia class, all the remaining microorganisms belonging to the class Anaerolineae. Red triangles in panel A mark ‘Ca. Villigracilis’

One exception was ‘Ca. Villigracilis’, which was shared by all reactors with a similar relative abundance (Fig. 2b). This filamentous bacterium has been reported as widely distributed in activated sludge reactors, where it probably contributes to the matrix supporting floc formation [69]. Although ‘Ca. Villigracilis’ has been metabolically characterized by FISH-MAR and reported as the only facultative aerobic member described within Anaerolineae class so far, the genome is not yet available [69].

Several genera classified according to the MiDAS taxonomy, and therefore with no cultured representative, presented high abundance within the Anaerolineaceae and Caldilineaceae families in all reactors. Specifically, genera belonging to the family Caldilineaceae were the most abundant in the activated sludge reactor (Fig. 2b). Meanwhile, genera belonging to the family Anaerolineaceae were the most abundant in anammox and methanogenic reactors.

From all genera detected in our samples, only five harbor cultured representatives: Flexilinea, Bellilinea, Longilinea, Leptolinea and Anaerolinea. All of them have strains isolated from anaerobic reactors which were described as strictly anaerobic chemoorganotroph involved in carbohydrates fermentation [9, 10, 12]. In pure culture, their growth was enhanced in co-cultivation with a hydrogenotrophic methanogens. This could explain why sequences belonging to these genera were almost exclusively detected in the methanogenic reactor.

Moreover, all genera were represented by few dominant species which in most cases were not shared between reactors (Fig. 2c). In the samples studied, all Chloroflexi members presented filamentous morphology except for the anammox reactor samples (these results are described in the Supplementary material, section Morphology determined by FISH).

Several ASVs determined using the 16S rRNA gene sequences, could not be classified at class (e.g., ASV6009, ASV6008, ASV6010), genera (ASV6634) or species level (e.g., ASV6007, e.g., ASV6634, e.g., ASV6400, etc.). Even though MiDAS database represents a great advance in understanding the microbiomes of wastewater treatment systems, there are still undescribed microorganisms not included and more effort is needed to expand this database.

General genomic features of Chloroflexi metagenomic assembled genomes

We used differential coverage binning [39] to obtain metagenomic assembled genomes (MAGs) from the three different reactors. A total of 75, 48 and 56 MAGs with > 50 completeness and < 10% contamination were successfully retrieved from MO, AMX and RH reactors samples, respectively. Seventeen MAGs belonging to the phylum Chloroflexi (9 AMX, 4 MO and 4 RH) were analyzed further. Recovered genomes completeness ranged from 85.59% and 96.79%, and contamination after the reassembly and polishing (manually curated) ranged between 0% and 3.36% (Table 2).

Table 2 Statistics summary for the 17 Chloroflexi MAGs recovered in this study

Three of these MAGs (AMX47, AMX55 and AMX56) were > 90% complete, had less than 5% contamination and importantly, included the full-length 16S, 23S, and 5S rRNA genes, and > 18 tRNA genes (Table 2), satisfying the criteria for high-quality (HQ) draft MAGs, according to the minimum information about a MAG (MIMAG) standard [70].

Phylogenomic analysis of Chloroflexi MAGs

To determine the phylogenetic position of the MAGs, a phylogenomic tree based on concatenated alignments of 120 single copy marker genes was constructed using 105 reference Chloroflexi genomes retrieved from NCBI (February 2021) and the 17 Chloroflexi MAGs obtained in our work. Of the 17 MAGs, six had sufficiently high degree of similarity (> 95% ANI to a representative genome [71]) to be classified at species level. The remaining MAGs presented ANI values below 95% with any reported genome and could not be classified at species level. According to the phylogenomic analysis, 7 were classified to genus level and 4 to family level (Table 2). The results showed that most of the MAGs (16 MAGs) were positioned into the Anaerolineae class while only one genome was positioned within the Dehalococcoidia class (Fig. 3a).

Fig. 3figure 3

Phylogenomic and phylogenetic analysis of Chloroflexi based on the metagenomics results and the 16S rRNA gene sequence analysis: A) Phylogenomic tree constructed with the 17 genomes from this work and reference genomes retrieved from the NCBI, B) Phylogenetic tree constructed with the most abundant ASVs retrieved from the 16S rRNA gene amplicon sequences (with abundance higher than 5% in at least one sample) and the sequences of the 16S rRNA gene from 10 MAGs from this work, and reference sequences of the phylum Chloroflexi retrieved from NCBI. The tree was reconstructed using the ML method and the GTR model. ML bootstrap values greater than or equal to 70% are shown at each node. Bar indicates 0.01 substitutions per site. The ASV and MAGs are colored by reactor: blue for RH, orange for MO and green for AMX. Sequences from the phylum Thermotogota were used as outgroup for rooting trees

At the order level, MAGs were distributed within the orders Anaerolineales, SBR1031, B4-G1, Caldilineales, UBA3071, UCB3 and UBA2991 (Table 2).

In addition, a phylogenetic tree was constructed based on the 16S rRNA gene sequences retrieved from 10 MAGs and the ASV sequences retrieved from the amplicon sequencing analysis (Fig. 3b, Fig. S2). Because of the few genomes available in the databases, it was not possible to construct both phylogenetic trees (based on 16S rRNA gene sequences and based on genomes sequences) with the exact same members of Chloroflexi. For example, ‘Ca. Villigracilis’ and ‘Ca. Sarcinatrix’, do not have representative genomes in databases so far, and were therefore not added to the genome-based phylogenetic tree. Moreover, although frequently used for phylogenetic identification, 16S rRNA gene sequences are notoriously difficult to assemble from metagenomes [72]. Thus, 16S rRNA genes were present only in 10 of the 17 MAGs (Table 2). Taking all this into consideration, the phylogenetic placement of 10 MAGs based on the 16S rRNA gene (Fig. 3b) was consistent with the phylogenomic tree based on 120 marker genes genes (Fig. 3a).

Interestingly, five MAGs from the anammox reactor (AMX55, AMX15, AMX39, AMX14 and AMX68), formed a monophyletic clade separated from other genomes in the phylogenomic tree (Fig. 3a). The 16S rRNA gene from AMX55 clustered with the sequence of ‘Ca. Villigracilis’ in the phylogenetic tree, with a 95.68% of identity according to the MiDAS taxonomy (Fig. 3b). Therefore, the retrieved MAG represents the first recovered genome within the ‘Ca. Villigracilis’ genus [73].

Within the ‘Ca. Villigracilis’ cluster, an ASV was classified within the UTCFX1 genus according to MiDAS. UTCFX1 represents a MAG retrieved from a nitritation-anammox sequencing batch reactor. The authors detected the presence nitrate reduction genes in this genome which were actively transcribed according to the metatranscriptomics analysis suggesting an important interaction with anammox microorganisms [26].

New genera and species candidates

In total, three MAGs (AMX47, AMX55 and AMX56) satisfied the criteria for high-quality recovered genomes [70] and met the minimal suggested required standards to be proposed as Candidates of new genera or species [74] (Table 2, Table S1). The full-length 16S rRNA gene sequence from AMX47, AMX55 and AMX56 showed identity values between 90.14% and 96.99% with their closest relatives according to Silva and MiDAS database comparisons (Table S2). The three MAGs presented ANI values < 79.2% with genomes from described species, and the AAI values for AMX47, AMX55 and AMX56 were 63.1, 57.6 and 56.6%, respectively (Table S1). Given the lack of close relatives, AMX47 and AMX56 genomes represent two different novel genera in the phylum, meanwhile, AMX55 represents the first representative genome of the genus ‘Ca. Villigracilis’ (based on the taxonomic sequence identity threshold recommendations of [75].

General analysis of metabolic pathways and genes

Different Chloroflexi genera and species predominate in the different wastewater treatment systems studied but, is their function redundant? For the 17 Chloroflexi MAGs, genes were annotated using a variety of protein databases to infer their metabolic potential. Because most of these MAGs were estimated to be between 85 and 93% complete, genes for additional pathways, might not be identified in this study. To overcome this problem, we also annotated genomes from described species closely related to our assembled genomes according to the phylogenomic tree (Fig. 3a). This allowed us to infer if a metabolic pathway of interest was not present in the genome or if its absence might be due to an incomplete genome recovery.

The metabolic pathway analysis was focused on answering the following questions:

1-Are they potentially capable of performing aerobic and/or anaerobic respiration?

2-Do they have the potential capability to hydrolyze different compounds?

3-Which carbon compound degradation pathways do they have?

4-Do they have N-removal potential?

5-Do they have genes related to biomass adhesion properties?

Aerobic and anaerobic respiration pathways

Reactors were operated in anaerobic (anammox and methanogenic reactors) or aerobic (activated sludge reactor) conditions and therefore, genes related to aerobic and anaerobic respiration, or fermentative metabolisms were searched against the assembled MAG sequences.

Regarding respiration pathways, the genomic analysis showed that most MAGs (10 out of 17) had an incomplete oxidative phosphorylation pathway. Five MAGs (RH43, RH52, AMX55, AMX14 and AMX56) harbor the genes to use O2 and different nitrogenous compounds (N2O, NO3− and/or NO2−) as final electron acceptors, and two of them presented only the genes to use NO3− or NO2− (MO66 and AMX68, respectively) (Fig. 4, Fig. S3, Table S1).

Fig. 4figure 4

Heatmap showing the completeness and the incompleteness of each metabolic pathway for the 17 MAG. Colors indicate pathways level completeness: dark brown (complete), light brown (1 block missing), beige (2 block missing), light green (more than 2 blocks missing) and white (not present)

The complete phosphorylation pathway and the presence of nitrate and nitrite reductases in AMX55 was in accordance with the in situ characterization of ‘Ca. Villigracilis’, as this genus has the ability to take up substrates under anoxic conditions in presence and absence of nitrate/nitrite [69]. As we mentioned, the metatranscriptomics analysis showed that nitrate reduction genes from UTCFX1 (closely related to AMX55 in the phylogenetic tree, Fig. 3b) were expressed in an anammox reactor [26].

The tolerance to aerobic conditions is determined by the presence of oxidative stress protection genes. The presence of these genes was expected in genomes retrieved from the activated sludge reactor as these microorganisms are continuously exposed to oxygen (RH21 and RH38). Nevertheless, these genes were found in all MAGs (Fig. 4, Table S1). This is in concordance with the results obtained for ‘Ca. Brevefilum fermentans’ retrieved from an anaerobic digester [17].

Metabolic pathways involved in polymers hydrolysis, carbon respiration and fermentation

To test the hypothesis that Chloroflexi members are capable of recycling soluble microbial products acting as hydrolytic bacteria, we performed the annotation of genes for glycosyl hydrolases such as cellulases, endohemicellulases, amylases, amino sugar-degrading enzymes, oligosaccharide-degrading enzymes, and also extracellular peptidases. The presence of all these genes indicated that most MAGs have the potential of hydrolyzing cellulose, starch, protein and/or peptides (Fig. 4, Supplementary Data 1). This is in accordance with previous in situ studies, which revealed high level of surface associated hydrolytic enzymes and their involvement in the breakdown of complex organic compounds [14, 76]. This result indicates that the Chloroflexi phylum may play an active role in hydrolyzing complex organic matter in activated sludge, as well as in methanogenic and anammox reactors.

Regarding central carbon metabolism, all MAGs contained multiple transporters for different organic compounds (including sugars, amino acids, proteins and fatty acids) indicating that each species has alternative routes for incorporating and recycling dissolved organic matter (scavenge macronutrients) (Fig. 4, Supplementary Data 1). The Embden-Meyerhof-Parnas (EMP) pathway for glycolysis was complete in all MAGs except in RH21, MO66 and AMX47. RH21 and MO66 are phylogenetically related to Aggregatilinea lenta and Litorilinea aerophila, respectively, which have the complete glycolysis pathway. Thus, incomplete glycolysis pathways in RH21 and MO66 genomes may be due to incomplete genome recoveries. On the other hand, AMX47 was closely related to Dehalococcoides mccartyi, which does not have the complete glycolysis pathway suggesting that AMX47 does not perform glycolysis. The potential of AMX55 to consume glucose was consistent with the results showed for ‘Ca. Villigracilis’ [69]. Genes for pyruvate oxidation to acetyl-CoA were annotated in all MAGs (Fig. 4, Supplementary Data 1). Beta-oxidation was annotated in most of the MAGs and might represent an important metabolic route to obtain carbon and reducing equivalents for all Chloroflexi species. Another interesting finding was that genes for N-acetylglucosamine transportation and metabolism (PTS, nagAB) were annotated in RH43 (Fig. 4, Supplementary Data 1). These results are in accordance with other reports where Chloroflexi members appear to retrieve N-acetylglucosamine from lysed cells revealed by micro-autoradiography or FISH studies [3, 14, 23, 77,78,79,80,81,82]. In addition to the N-acetylglucosamine metabolism, we suggest that scavenging occurs through hydrolysis of complex organic compounds outside the cell, which are then transported to the cytoplasm, and are then metabolized via glycolysis or beta-oxidation (degradation of fatty acids and branched-chain amino acids) pathways. These results support the previous hypothesis that Chloroflexi has an important beneficial role in degradation of lysed bacterial cell debris and EPS.

Fermentation pathways including genes for acetate, ethanol, lactate, acetoin, formate and/or propionate production were present in all MAGs (Fig. 4, Supplementary Data 1). These results were in accordance with previous information of the Anaerolineae class (isolates and MAGS) retrieved from wastewater treatment system which were involved in sugar, amino acid or protein fermentation with different end products [8,9,10, 12].

Metabolic pathways for amino acids degradation and propionate formation (methylmalonyl-CoA pathway and aldehyde:ferredoxin oxidoreductase) was found in five MAGs (RH38, AMX14, AMX47, AMX55 and AMX68). These genes were also annotated in other Chloroflexi members [17, 83,84,85].

Most of the MAGs (13 MAGs) contained a formate dehydrogenase H (fdhF) to convert formic acid decomposition into CO2. Only AMX55 had the ability to convert formic acid into H2 and CO2 by the presence of both fdhF and hydrogenases, under anaerobic conditions in the absence of exogenous electron acceptors as was observed in previous studies [82, 86]. In addition, the formate transporter (focA) was annotated in most of these MAGs. Synergistic interaction of Anaerolineae members with the methanogenic archaea, was previously reported in anammox [87] and methanogenic reactors [2, 17, 88]. This could be a common scenario in these systems where the excess of electrons from organic carbon oxidation by Anaerolineae members could be transferred to methanogenic archaea.

Potential for polymers and lipids storage

The potential to store glycogen, polyhydroxyalkanoates and lipids was more common among MAGs retrieved from activated sludge than from methanogenic or anammox reactors. RH21, RH38, RH43, MO66 and AMX9 had complete glycogen biosynthesis and degradation pathways, suggesting that this polysaccharide may serve as a possible storage compound under unbalanced growth conditions (e.g., when C and/or N is temporally limited) (Fig. 4, Supplementary Data 1) [89, 90]. This result is in accordance with the information reported for ‘Ca. Amarolinea aalborgensis’ and ‘Ca. Promineofilum breve’ both retrieved from activated sludge [7, 16].

In addition to glycogen as storage compound, RH43 and RH52, encodes a potential biosynthesis pathway for polyhydroxyalkanoates (PHA) (Fig. 4, Supplementary Data 1), which are usually formed as carbon and energy storage compounds [91] under conditions of carbon excess and nitrogen or phosphate limitation [92]. Thus, the potential for PHA storage could favor these organisms with intermittent carbon availability present in these systems. Also, the annotation of a putative acyl-CoA:DAG acyltransferases (atfA), which catalyzes the final step in the synthesis of triacylglycerols, indicates the potential for lipid storage [93] in RH21, RH38, AMX56 and AMX57. This capability was also reported for ‘Ca. Promineofilum breve’ [16].

Presence of Nitrogen metabolism pathways

Different genes related to the nitrogen cycle such as the dissimilatory nitrate reduction to ammonia (DNRA) and partial denitrification were mostly annotated in Chloroflexi MAGs from anammox reactors. DNRA is part of the nitrogen cycle, and it has been annotated in several Chloroflexi genomes [

留言 (0)

沒有登入
gif