Anaerobic Faecalicatena spp. degrade sulfoquinovose via a bifurcated 6-deoxy-6-sulfofructose transketolase/transaldolase pathway to both C2- and C3-sulfonate intermediates

Abstract

Plant-produced sulfoquinovose (SQ, 6-deoxy-6-sulfoglucose) is one of the most abundant sulfur-containing compounds in nature and its bacterial degradation plays an important role in the biogeochemical sulfur and carbon cycles and in all habitats where SQ is produced and degraded, particularly in gut microbiomes. Here, we report the enrichment and characterization of a strictly anaerobic SQ-degrading bacterial consortium that produces the C₂-sulfonate isethionate (ISE) as the major product but also the C₃-sulfonate 2,3-dihydroxypropanesulfonate (DHPS), with concomitant production of acetate and hydrogen (H₂). In the second step, the ISE was degraded completely to hydrogen sulfide (H₂S) when an additional electron donor (external H₂) was supplied to the consortium. Through growth experiments, analytical chemistry, genomics, proteomics, and transcriptomics, we found evidence for a combination of the 6-deoxy-6-sulfofructose (SF) transketolase (sulfo-TK) and SF transaldolase (sulfo-TAL) pathways in a SQ-degrading Faecalicatena-phylotype (family Lachnospiraceae) of the consortium, and for the ISE-desulfonating glycyl-radical enzyme pathway, as described for Bilophila wadsworthia, in an Anaerospora-phylotype (Sporomusaceae). Furthermore, using total proteomics, a new gene cluster for a bifurcated SQ pathway was also detected in Faecalicatena sp. DSM22707, which grew with SQ in pure culture, producing mainly ISE, but also 3-sulfolacate (SL) 3-sulfolacaldehyde (SLA), acetate, butyrate, succinate, and formate, but not H₂. We then reproduced the growth of the consortium with SQ in a defined co-culture model consisting of Faecalicatena sp. DSM22707 and Bilophila wadsworthia 3.1.6. Our findings provide the first description of an additional sulfoglycolytic, bifurcated SQ pathway. Furthermore, we expand on the knowledge of sulfidogenic SQ degradation by strictly anaerobic co-cultures, comprising SQ-fermenting bacteria and cross-feeding of the sulfonate intermediate to H₂S-producing organisms, a process in gut microbiomes that is relevant for human health and disease.

Keywords: 3-sulfolactaldehyde; 6-deoxy-6-sulfofructose transaldolase; anaerobic microbial metabolism; carbon and sulfur cycle; isethionate; transketolase.

Figure 1

(A) Illustration of the known 6-deoxy-6-sulfofructose transaldolase (sulfo-TAL) pathway. SQ is imported, most likely by transporter SftA, and isomerized to 6-deoxy-6-sulfofructose (SF) by the SQ isomerase SftI. The non-sulfonated C₃-(glycerone) moiety of SF is transferred to glyceraldehyde 3-phosphate (G3P) by the SF transaldolase SftT, resulting in fructose-6-phosphate (F6P) and 3-sulfolactaldehyde (SLA); the former is used for energy generation and regeneration of the G3P acceptor molecule (not shown, see Frommeyer et al., 2020 for details). The SLA is oxidized to 3-sulfolactate to generate additional NADH for aerobic respiration in aerobic Priestia [formerly Bacillus] aryabhattai SOS1, whereas in strictly anaerobic Clostridium spp. and Eubacterium rectale it is employed as internal electron acceptor for NAD⁺ regeneration (fermentation) and reduced to 2,3-dihydroxypropane-1-sulfonate (DHPS); both degradation products are not further converted but excreted (Frommeyer et al., 2020). (B) Illustration of the two major 6-deoxy-6-sulfofructose transketolase (sulfo-TK) pathway variants. SQ is imported, most likely by an ABC transporter (Arumapperuma et al., 2024), and isomerized to SF by isomerase SqvD (Liu et al., 2021; Chu et al., 2023). A non-sulfonated C₂-(ketol) moiety of SF is transferred to G3P by SF transketolase (SqwGH), yielding xylulose-5-phosphate (X5P) and the sulfo-sugar 4-deoxy-4-sulfoerythrose (SE). The SE is then isomerized to 4-deoxy-4-sulfoerythrulose (SEu) by the SEu isomerase (SqwI), and the SEu is converted by SqwGH to yield another molecule of X5P and 2-sulfoacetaldehyde (SAA). In Clostridium sp. MSTE9 (Liu et al., 2021), the SAA is reduced to isethionate (ISE; 2-hydroxyethanesulfonate) by the NADH-dependent SAA reductase SqwF, which is then excreted (SqwE). Another variant of this pathway has been described for Acholeplasma sp. using recombinant enzymes (Chu et al., 2023), in which SAA is oxidized further to 3-sulfoacetate (SAc) via a NAD⁺-dependent, CoA-acylating sulfoacetaldehyde dehydrogenase (SqwD), producing sulfoacetyl-CoA (SAc-CoA), and a CoA ligase (SqwKL); the SAc is exported by the exporter SqwE. Isomerases are indicated in blue, reductases in yellow, dehydrogenases in orange, ligase in light blue, transaldolase in light green, and transketolase in dark green. Importers and exporters are indicated in grayscale.

Figure 2

Microscopic appearance of two different cell types present in the highly enriched, stable anaerobic consortium that degraded SQ via ISE to H₂S. The cells were embedded on agarose-coated slides and imaged by phase-contrast microscopy. The two different cell types are indicated by arrows, rod-shaped cells (indicated by white-filled arrows) and filamentous cells (white open arrows) (A). In some instances, the filamentous cells carried terminal spores (black arrows in panel B). Scale bars, 10 μm.

Figure 3

Phylogenetic composition of the SQ-degrading consortium as determined by Illumina amplicon sequencing of a 570 bp fragment spanning the V3 to V5 region of the 16S rRNA gene. Relative abundances were calculated from the sequence reads obtained after quality assessment. Only three genera were detected within two families.

Figure 4

Growth experiments with the highly enriched consortium and with Faecalicatena sp. DSM22707 and B. wadsworthia 3.1.6 in pure culture, each with SQ as the sole carbon source. (A) The consortium completely utilized SQ concomitant with growth and ISE and acetate production; small amounts of DHPS were also detectable (not shown, see text). Furthermore, hydrogen gas (H₂) was produced (see text and Supplementary Figure S1), which disappeared when the consortia were incubated for additional weeks, while some ISE was degraded (see text and Supplementary Figure S2). (B) When the headspace of outgrown consortium-cultures was spiked with additional H₂ gas after 7 d of incubation (indicated by the red dashed line), the ISE was completely utilized, concomitant with additional acetate production; sulfide was also detected in higher amounts (see text). (C) Faecalicatena sp. DSM22707 completely utilized SQ with concomitant growth and production of ISE and acetate and produced significant amounts of butyrate. Furthermore, small amounts of SL, SLA and SAA, and succinate and formate were detected, but no H₂ in the headspace (see text). (D) After growth of Faecalicatena sp. DSM22707 with SQ, the culture medium was filter-sterilized and inoculated with B. wadsworthia 3.1.6 (indicated by the red dashed line); additionally, the headspace was spiked with H₂ gas. The ISE disappeared completely while additional acetate and sulfide were produced. Each growth experiment was performed at least in triplicate; mean values are given, and error bars correspond to the standard deviation.

Figure 5

Extract of the data on gene expression levels in the consortium, as determined by transcriptomics and proteomics, showing candidate genes for SQ (A) and ISE degradation, including genes for sulfite reduction (B) and housekeeping genes (C) for comparison. The IMG metagenome annotation was used to build the total-proteomics database and IMG Gene IDs (prefix: Ga0499732_) are indicated as gene identifiers. The gene clusters on scaffolds Ga0499732_056 and _143 for SQ utilization were assigned to the Faecalicatena MAG, and scaffolds _441, _699 and _700 for ISE utilization and sulfite respiration were assigned to the Anaerospora MAG. Transcript abundance is given in transcripts per million (TPM) and protein abundance is expressed as the mean peak area of all peptides identified by mass spectrometry. Transcriptomic sequencing was performed in duplicate and total proteomics in triplicate when starting with consortium cultures harvested each at late exponential growth phase; error bars correspond to standard deviation. DMT, drug/metabolite transporter; BMC, bacterial microcompartment.

Figure 6

Volcano plot illustrating the differential proteomics results obtained for pure cultures of Faecalicatena sp. DSM22707 grown with either 10 mM SQ or glucose. Proteins that were found to be significantly more abundant (up-regulated, as defined by a log₂ fold change (FC) cutoff of 3 and a p-value cutoff of 5×10⁻⁵) are indicated as red dots, shifted to the right if they were more abundant in SQ-grown cells. In this figure, the proteins encoded by the SQ-metabolism gene cluster are labeled by the last two digits of their IMG Gene IDs (prefix 29568592xx); consecutive numbers indicate that the genes are encoded next to each other.

Figure 7

Illustration of the gene clusters for SQ and ISE metabolism as identified by transcriptomics and/or proteomics in this study. (A) Comparison of the SQ degradation gene clusters found in the Faecalicatena-phylotype of the consortium and in the pure culture of Faecalicatena sp. DSM22707. The genes are arranged in the same order and the encoded proteins show ≥94% identical amino acid sequence (id-aa). Note that the predicted SQase and the ABC transporter genes were located on a different scaffold for the consortium metagenome. (B) Comparison of the SQ degradation gene cluster of strain DSM22707 with those of P. aryabhattai SOS1 (sulfo-TAL pathway) and Clostridium sp. MSTE9 (sulfo-TK pathway). Gene 2956859258 was predicted to be NAD⁺-dependent SQase, because of its homology to Arthrobacter sp. strain AK01 (see main text). (C) Gene cluster identified for ISE desulfonation in the Anaerospora-phylotype of the consortium in comparison to the gene cluster of B. wadsworthia 3.1.6. IMG gene IDs are indicated for single genome sequences and IMG locus tags (scaffolds) for the consortium metagenome. The digits shown in the gene arrows complete the respective IMG gene IDs (i.e., replace the XX). Sequence identity is indicated in grayscale (numbers, % identity).

Figure 8

Structural superimposition of the Alphafold-predicted heterotetramer of Faecalicatena sp. DSM22707 SF transketolase (N-terminal in teal, C-terminal in cyan) with the crystal structure of the single-gene fructose 6-phosphate transketolase from Scheffersomyces stipitis (PDB code: 5XU2; depicted in yellow) containing the bound cofactor thiamine pyrophosphate (TPP, purple), a calcium ion (Ca²⁺, black) and the sugar-phosphate substrate of S. stipitis transketolase (F6P, pink) (A). A close-up of the superimposed catalytic centers is also shown (B), in which the differing amino-acid residues are depicted [coloring according to the global structure in (A)]. For Faecalicatena transketolase, the residue H466 is exchanged for D108, and S383 is exchanged for A36. Residues R525/166 and R356/9 are conserved between the two enzymes (B).

Figure 9

Superimposition of the Alphafold predicted structures of the two Faecalicatena sp. DSM22707 transaldolase candidates expressed during SQ utilization (F1, IMG Gene ID: 2956859250; F2, 2956859251) with the cryo-EM structure of P. megaterium SF transaldolase (BmSF-TAL, PDB code: 8 BC4) including the SF ligand of BmSF-TAL (A). Also shown are the superimposed catalytic sites of transaldolase candidates F1 and F2 with the BmSF ligand (ligand QC9) as seen in the cryo-EM structure (B). For comparison, the catalytic triad, sulfonate-orienting, and sugar-orienting residues are shown as predicted by Snow et al. (2023), showing that transaldolase F1 contains the same residues as the BmSF enzyme but that the transaldolase F2 differs in its residues (C). Transaldolase BmSF-TAL, green; Faecalicatena transaldolase candidates F1 and F2, yellow and blue, respectively.

Figure 10

Illustration of gene clusters found in genomes of cultured bacteria (isolates) or MAGs of uncultured bacteria, each highlighted for a co-localization of valid candidate genes for SF transaldolases, SF transketolases, and SQ and SEu isomerases. The pictograms indicate the origin of the isolate/MAG which are animal feces or intestinal contents (paw prints), human feces (humans), engineered environments (cogwheels), or natural environments (leaves). For further details on their origin, see Supplementary Table S7.

Figure 11

Illustration of the proposed bifurcated sulfo-TK and sulfo-TAL pathway in the Faecalicatena species (A) and the desulfonation pathway for the major excreted product, ISE, in the Anaerospora phylotype of the consortium and B. wadsworthia 3.1.6 of the defined co-culture (B), as examined in this study. The growth physiology experiments showed that SQ is degraded predominantly to ISE, but that SLA and SL (pure culture) or DHPS (consortium) are also produced and excreted in small amounts, providing a strong indication that the pathway bifurcates at the cleavage of the C₆-backbone of SQ to yield both C₂- and C₃ intermediates. Corresponding SF transketolase and SF transaldolase candidate enzymes were found to be encoded in a gene cluster that was transcribed and/or translated along with valid SF isomerase candidates for the upstream pathway, and SEu isomerase and SAA and SLA reductase and dehydrogenase candidates for the pathways downstream of the bifurcation. The generated F6P and X5P entry into either glycolysis or pentose phosphate pathway (PPP) is also indicated (A). The ISE is utilized for organosulfonate respiration by the Anaerospora phylotype or Bilophila wadsworthia 3.1.6 when employing the GRE sulfite-lyase reaction in bacterial microcompartments (BMC) (light blue), as indicated by the transcriptional/proteomics data and as previously described (Peck et al., 2019; Burrichter et al., 2021) (B). AA, acetaldehyde, AcCoA, acetyl-CoA, AcP, acetyl phosphate, Ac, acetate, Dsr, dissimilatory sulfite reductase.