Entner-Doudoroff pathway for sulfoquinovose degradation in Pseudomonas putida SQ1

Abstract

Sulfoquinovose (SQ; 6-deoxy-6-sulfoglucose) is the polar head group of the plant sulfolipid SQ-diacylglycerol, and SQ comprises a major proportion of the organosulfur in nature, where it is degraded by bacteria. A first degradation pathway for SQ has been demonstrated recently, a "sulfoglycolytic" pathway, in addition to the classical glycolytic (Embden-Meyerhof) pathway in Escherichia coli K-12; half of the carbon of SQ is abstracted as dihydroxyacetonephosphate (DHAP) and used for growth, whereas a C3-organosulfonate, 2,3-dihydroxypropane sulfonate (DHPS), is excreted. The environmental isolate Pseudomonas putida SQ1 is also able to use SQ for growth, and excretes a different C3-organosulfonate, 3-sulfolactate (SL). In this study, we revealed the catabolic pathway for SQ in P. putida SQ1 through differential proteomics and transcriptional analyses, by in vitro reconstitution of the complete pathway by five heterologously produced enzymes, and by identification of all four organosulfonate intermediates. The pathway follows a reaction sequence analogous to the Entner-Doudoroff pathway for glucose-6-phosphate: It involves an NAD(+)-dependent SQ dehydrogenase, 6-deoxy-6-sulfogluconolactone (SGL) lactonase, 6-deoxy-6-sulfogluconate (SG) dehydratase, and 2-keto-3,6-dideoxy-6-sulfogluconate (KDSG) aldolase. The aldolase reaction yields pyruvate, which supports growth of P. putida, and 3-sulfolactaldehyde (SLA), which is oxidized to SL by an NAD(P)(+)-dependent SLA dehydrogenase. All five enzymes are encoded in a single gene cluster that includes, for example, genes for transport and regulation. Homologous gene clusters were found in genomes of other P. putida strains, in other gamma-Proteobacteria, and in beta- and alpha-Proteobacteria, for example, in genomes of Enterobacteria, Vibrio, and Halomonas species, and in typical soil bacteria, such as Burkholderia, Herbaspirillum, and Rhizobium.

Keywords: 6-deoxy-6-sulfoglucose; bacterial biodegradation; organosulfonate; sulfolipid; sulfur cycle.

Fig. 1.

Core enzyme reactions of the classical Entner–Doudoroff pathway for GP in Pseudomonas species in comparison to the newly discovered Entner–Doudoroff-type pathway for SQ, with its corresponding genes in Pseudomonas putida SQ1. (A) The cytosolic part of the Entner–Doudoroff pathway for glucose (19) is an alternative to classical glycolysis and generates two molecules of pyruvate from GP via PGL, PG, and KDPG, which are cleaved to pyruvate and GAP; the GAP is converted to pyruvate, as indicated. (B) Entner–Doudoroff pathway for SQ, which is a structural analog of GP, involves analogous enzyme reactions and the intermediates SGL, SG, and KDSG, which are cleaved to pyruvate and SLA, as demonstrated in this study. The pyruvate is used for growth, and the SLA is oxidized to SL, which is excreted. (C) Identified genes for the five core enzymes of the Entner–Doudoroff pathway for SQ (color-coded in B) are encoded in one gene cluster in P. putida SQ1, together with predicted genes for transport, for regulation, and presumably for funneling of other SQ derivatives into the pathway (main text). Note that the gene identifiers in this figure refer to IMG locus tags in the IMG annotation of the draft genome sequence of P. putida SQ1 (IMG Project ID Gp0039102) and that gene functions are specified according to their original IMG annotation in this figure.

Fig. 2.

Complete disappearance of SQ in cell-free extract of SQ-grown P. putida SQ1 cells concomitant with a transient formation of metabolites SGL, SG, and KDSG, and formation of the end-product SL. The reaction in soluble protein extract (soluble protein fraction) was started by addition of NAD⁺ (not shown) and followed in samples that were taken at intervals and analyzed by HPLC-MS; the time points of sampling are indicated. The total-ion chromatograms (TICs) recorded in the negative-ion mode from the MS/MS fragmentation of the quasi-molecular ions ([M-H]⁻) of SQ, SGL, SG, and KDSG, as well as SL, are shown. Note that SGL and KDSG were each observed as [M-H]⁻ ions of identical mass (m/z = 241) but that the compounds eluted at different HPLC retention times, as indicated for t = 5 min. Discussion of the MS/MS fragmentation of the metabolites SGL, SG, and KDSG, as well as SL, is provided in SI Materials and Methods.

Fig. S1.

MS/MS fragmentation of SGL, SG, and KDSG. (A) Fragment ions of the [M-H]⁻ ions of SGL. Fragmentation led to a loss of water (−18) and carbon dioxide (−44), and to the formation of HSO₃⁻ (81) ions. (B) Fragment ions of the [M-H]⁻ ions of SG. Fragmentation led to a loss of water (−18), CH₂O₂ (−46), C₂H₄O₃ (−76), and SO₃⁻ (−80), concomitant with the formation of HSO₃⁻ (81) ions. (C) Fragment ions of the [M-H]⁻ ions of KDSG. Fragmentation led mainly to a loss of water (−18) and carbon dioxide (−44), and to the formation of HSO₃⁻ (81) ions.

Fig. S2.

MS/MS fragmentation of SLA and SL. (A) Fragment ions of the [M-H]⁻ ions of SLA. Fragmentation led to a loss of water (−18) and to the formation HSO₃⁻ (81) and C₃H₃O₂⁻ (71) ions. (B) Fragment ions of the [M-H]⁻ ions of SL. Fragmentation led to a loss of water (−18) and to the formation HSO₃⁻ (81) ions.

Fig. 3.

Results of proteomic and transcriptional analyses indicating the involvement of a single SQ-degradation gene cluster in P. putida SQ1. A schematic representation of the gene cluster is illustrated in Fig. 1C. (A) Soluble proteins in SQ- and glucose-grown cells were separated by 2D-PAGE, and all prominent protein spots on the gels from SQ-grown cells that suggested inducibly produced proteins were identified by peptide PF-MS. The genes identified are indicated next to each protein spot (i.e., their locus tag number in the IMG annotation), with all genes located within the gene cluster labeled in red (0088–0090 and 0100); labeled in black are two prominent spots that identified genes considered unlikely to be involved in SQ degradation directly (0897, a translation elongation factor gene, and 4549, a gene for an ATP-binding cassette (ABC)-transport periplasmic binding protein). (B) Extract of the results of the total proteome (Orbitrap-MS) analyses of SQ-grown cells (blue bars) and glucose-grown cells (gray bars), illustrating the detected strong expression of proteins encoded within the newly identified gene cluster specifically during growth with SQ (0088–0091, 0094, and 0100), but not during growth with glucose. For comparison, the strong expression of enzymes of the Entner–Doudoroff pathway detected specifically during growth with glucose, but not during growth with SQ, is shown (Edd, Glk, Zwf-1, and Eda), as well as the expression level of enzymes for a further conversion of GAP and pyruvate (Gap and AceE) and of two constitutively produced proteins (SucC and AtpD) for each growth condition. (C) Differential transcriptional analysis (RT-PCR) of all genes encoded within the newly identified gene cluster (Fig. 1C), which indicated their strong and inducible transcription specifically during growth with SQ, but not during growth with glucose. A constitutively expressed gene (3160, for citrate synthase) served as a positive control, and the negative control was a PCR assay without RT (RNA) to confirm the absence of DNA contamination in the RNA preparations used. Each result in A–C was replicated with material from an independent growth experiment.

Fig. S3.

Growth experiment with P. putida SQ1 WT (A) and its insertion mutant (B) in gene 0090 (SQ dehydrogenase). Duplicate growth experiments are shown with glucose (open symbols) and SQ (solid symbols) as growth substrate in liquid cultures (6 mM each). Growth was monitored as optical density (OD₅₈₀).

Fig. S4.

Analysis of the purity of heterologously overproduced enzymes using SDS/PAGE. Marker proteins are indicated (kilodaltons). Lane A, SQ dehydrogenase (PpSQ1_00090); lane B, SGL lactonase (PpSQ1_00091); lane C, SG dehydratase (PpSQ1_00089); lane D, KDSG aldolase (PpSQ1_00100); lane E, SLA dehydrogenase (PpSQ1_00088); and lane F, GP dehydrogenase (Zwf-1, PpSQ1_03570). The protein bands with a lower molecular mass than expected in lane C represent C-terminally truncated versions of the recombinant His-tagged protein, as determined by PF-MS.

Fig. 4.

In vitro reconstitution of the Entner–Doudoroff pathway for SQ. The transformation of SQ to SGL, SG, KDSG, and SLA, and of SLA to SL, by successive addition of recombinantly produced pathway enzymes was followed by HPLC-MS. The initial substrate concentrations were 2 mM SQ and 3 mM NAD⁺, and 50 μg⋅mL⁻¹ of each enzyme was added. (A) Sample of SQ in reaction buffer (t = 0 min). (B) Sample taken 45 min after addition of SQ-dehydrogenase (gene PpSQ1_00090) (t = 45 min). (C) Sample taken 45 min after addition of SGL-lactonase (gene 0091) (t = 90 min). (D) Sample taken 45 min after addition of SG-dehydratase (gene 0089) (t = 135 min). (E) Sample taken 45 min after addition of KDSG-aldolase (gene 0100) (t = 180 min). (F) Sample taken 45 min after addition of SLA-dehydrogenase (gene 0088) and 2 mM NAD⁺ (t = 225 min). The TICs recorded in the negative-ion mode from the MS/MS fragmentation of the quasi-molecular ions ([M-H]⁻) of SQ, SG, SGL, and KDSG, as well as SLA and SL, are shown; note that SGL and KDSG were each observed as [M-H]⁻ ions of identical mass (m/z = 241) but that the compounds eluted at different HPLC retention times, as indicated in the duplicated panels for TIC MS² of m/z = 241. Pyruvate, as the second product of the KDSG-aldolase reaction (Fig. 1), could not be detected under the HPLC-MS conditions we used but was confirmed by other analytical methods.

Fig. 5.

Illustration of the SQ-degradation gene cluster in P. putida SQ1 with its five identified core genes (A) and of homologous, predicted SQ-degradation gene clusters found in genomes of other P. putida, other gamma-Proteobacteria (B), and in beta- and alpha-Proteobacteria (C and D, respectively). Homologous gene clusters in bacterial genomes were retrieved from the Joint Genome Institute’s IMG database by the Gene Cassette Search tool and the Gene Neighborhood viewer. Shown are gene clusters that contain gene homologs for at least four of the five core enzymes of the SQ Entner–Doudoroff pathway (genes are indicated by color coding). Also, candidate genes for SL export and for a predicted SQ-glyceride (SQG) alpha-glucosidase (main text) were frequently found to be conserved within these homolog clusters (gene symbols in gray). Note that other candidate genes in the clusters are not specified in this figure (gene symbols in white); for example, more variable genes were predicted for outer membrane porins, other transporters, or regulation (Fig. 1C).