A Sulfoglycolytic Entner-Doudoroff Pathway in Rhizobium leguminosarum bv. trifolii SRDI565

Sulfonate sulfur is a major form of organic sulfur in soils but requires biomineralization before it can be utilized by plants. Very little is known about the biochemical processes used to mobilize sulfonate sulfur. We show that a rhizobial isolate from soil, Rhizobium leguminosarum SRDI565, possesses the ability to degrade the abundant phototroph-derived carbohydrate sulfonate SQ through a sulfoglycolytic Entner-Doudoroff pathway. Proteomics and metabolomics demonstrated the utilization of this pathway during growth on SQ and provided evidence for gluconeogenesis. Unexpectedly, off-cycle sulfoglycolytic species were also detected, pointing to the complexity of metabolic processes within cells under conditions of sulfoglycolysis. Thus, rhizobial metabolism of the abundant sulfosugar SQ may contribute to persistence of the bacteria in the soil and to mobilization of sulfur in the pedosphere.

ABSTRACT

Rhizobia are nitrogen-fixing bacteria that engage in symbiotic relationships with plant hosts but can also persist as free-living bacteria in the soil and rhizosphere. Here, we show that free-living Rhizobium leguminosarum SRDI565 can grow on the sulfosugar sulfoquinovose (SQ) or the related glycoside SQ-glycerol using a sulfoglycolytic Entner-Doudoroff (sulfo-ED) pathway, resulting in production of sulfolactate (SL) as the major metabolic end product. Comparative proteomics supports the involvement of a sulfo-ED operon encoding an ABC transporter, sulfo-ED enzymes, and an SL exporter. Consistent with an oligotrophic lifestyle, proteomics data revealed little change in expression of the sulfo-ED proteins during growth on SQ versus mannitol, a result confirmed through biochemical assay of sulfoquinovosidase activity in cell lysates. Metabolomics analysis showed that growth on SQ involves gluconeogenesis to satisfy metabolic requirements for glucose-6-phosphate and fructose-6-phosphate. Metabolomics analysis also revealed the unexpected production of small amounts of sulfofructose and 2,3-dihydroxypropanesulfonate, which are proposed to arise from promiscuous activities of the glycolytic enzyme phosphoglucose isomerase and a nonspecific aldehyde reductase, respectively. The discovery of a rhizobium isolate with the ability to degrade SQ builds our knowledge of how these important symbiotic bacteria persist within soil.

IMPORTANCE Sulfonate sulfur is a major form of organic sulfur in soils but requires biomineralization before it can be utilized by plants. Very little is known about the biochemical processes used to mobilize sulfonate sulfur. We show that a rhizobial isolate from soil, Rhizobium leguminosarum SRDI565, possesses the ability to degrade the abundant phototroph-derived carbohydrate sulfonate SQ through a sulfoglycolytic Entner-Doudoroff pathway. Proteomics and metabolomics demonstrated the utilization of this pathway during growth on SQ and provided evidence for gluconeogenesis. Unexpectedly, off-cycle sulfoglycolytic species were also detected, pointing to the complexity of metabolic processes within cells under conditions of sulfoglycolysis. Thus, rhizobial metabolism of the abundant sulfosugar SQ may contribute to persistence of the bacteria in the soil and to mobilization of sulfur in the pedosphere.

INTRODUCTION

Sulfur is essential for plant growth and is the fourth most important macronutrient after nitrogen, phosphorus, and potassium. Up to 10 kg/ha/year of sulfur is deposited in rain, especially near industrialized areas (1). However, sulfur dioxide emissions from industrial sources have decreased in recent decades as a result of pollution mitigation and the move to low-sulfur fuels and renewable energy sources, and quantities received from atmospheric sources are now at levels below that required by most crops (2). Sulfur deficiency in soils is primarily combated by application of sulfur-containing fertilizers, such as superphosphate, ammonium sulfate, and gypsum (3), which are applied across all major crop-growing and pasture areas worldwide (4). Soils contain significant amount of sulfur, yet plants can use sulfur only in the form of sulfate, and it has been shown that 95 to 98% of sulfur in soils is in the form of unavailable biological sulfur (4). Thus, effective microbial cycling of sulfur from biological to inorganic forms within the soil is important (5) and has the potential to enhance crop yields and reduce reliance on fertilizers.

X-ray absorption near-edge spectroscopy measurements have led to estimates that approximately 40% of sulfur within various sediments and humic substances exists as sulfonate (6). Chemical methods of analysis applied to a range of forest soils revealed that sulfonate sulfur accounted for 40% of the total organic sulfur pool in the majority of cases (7). Little detail is known about the speciation of organic sulfonates in soils, but one important input is phototroph-derived litter. It is estimated that around 10 billion tonnes of the sulfosugar sulfoquinovose (SQ) is produced annually by photosynthetic organisms, including plants, cyanobacteria, and algae (8). SQ is primarily found as the glycerolipid sulfoquinovosyl diacylglycerol (SQDG), and land plants can contain as much as 10% SQDG in their thylakoid membrane glycerolipids (9). Very little is known about how SQ is metabolized within soils, although it has been shown to undergo very rapid mineralization to inorganic sulfate (10).

Bacteria are likely to be primarily responsible for the biomineralization of SQ, possibly by using SQ as a carbon source and catabolizing it via a modified version of glycolysis, termed sulfoglycolysis (11). Two sulfoglycolytic processes have been described: the sulfoglycolytic Embden-Meyerhof-Parnas (sulfo-EMP) pathway (12), and the sulfoglycolytic Entner-Doudoroff (sulfo-ED) pathway (Fig. 1) (13). The sulfo-ED pathway was first reported in Pseudomonas putida strain SQ1, a bacterium isolated from freshwater sediment that catabolizes SQ with excretion of equimolar amounts of sulfolactate (SL) (13). The sulfo-ED operon of P. putida SQ1 contains 10 genes, including those corresponding to a transcriptional regulator, an SQ importer and SL exporter, a sulfoquinovosidase, SQ mutarotase, SQ dehydrogenase, SL lactonase, SG dehydratase, KDSG (2-keto-3,6-dideoxy-6-sulfogluconate) aldolase, and SLA (3-sulfolactaldehyde) dehydrogenase. Based on genome-wide annotation studies, the sulfo-ED pathway is predicted to occur in a range of alpha-, beta-, and gammaproteobacteria (13). However, no direct evidence for this pathway has been reported for any organism other than P. putida SQ1. Other members of the microbial community can catabolize SL and 2,3-dihydroxypropanesulfonate (DHPS; the product of the sulfo-EMP pathway) to inorganic sulfur (14), completing the biomineralization of SQ.

FIG 1.

Proposed sulfoglycolytic Entner-Doudoroff (sulfo-ED) pathway in Rhizobium leguminosarum bv. trifolii SRDI565. (a) Operon encoding the sulfo-ED pathway. (b) Proposed sulfo-ED pathway. (c) Comparison with the Entner-Doudoroff pathway.

Rhizobium leguminosarum bv. trifolii SRDI565 (syn. N8-J) was isolated from a soil sample collected in western New South Wales, Australia, but has the capacity to colonize Trifolium subterraneum subsp. subterraneum (subterranean clover) and other Trifolium spp. (15). Colonization of Trifolium species with SRDI565 results in suboptimal nodulation and nitrogen fixation in some species and ineffective nitrogen fixation in others, leading to reduced shoot nitrogen content relative to other commercial strains (16). Interestingly, the genome of SRDI565 contains all the genes needed for a functional sulfo-ED pathway (17), although there is no evidence to show that this is operative and/or that SRDI565 can use SQ as a major carbon source.

Rhizobia participate in sophisticated symbiotic relationships with leguminous host plants that allow them to fix atmospheric dinitrogen to provide a growth advantage to the host (18). Symbiosis is triggered by molecular communication between the bacterium and the host, resulting in nodule formation on the root and colonization by the bacterium. Within nodule bacteroids, the energy-intensive fixation of nitrogen is supported by C₄-dicarboxylates (primarily malate, fumarate, and succinate) obtained from glycolysis of sucrose photosynthate within the plant host (18). Owing to the importance of biological nitrogen fixation for input of nitrogen into the biosphere, the symbiosis of rhizobia and leguminous hosts has been well studied. However, rhizobia can also exist as free-living bacteria within the soil and rhizosphere (19). Here, like other soil bacteria, they adopt a saprophytic and oligotrophic lifestyle, where they utilize a variety of alternative carbon sources, including a wide range of carbohydrates (20). Most likely, the ability of various rhizobia to persist in the pedosphere depends upon their ability to utilize diverse carbohydrate and noncarbohydrate substrates and establish an appropriate niche. SQ or its glycosides are likely to be a common soil constituent and nutrient, given their ubiquitous production by plants. Possibly, the sulfo-ED pathway in SRDI565 might provide it with the capacity to survive on plant-derived SQ or SQDG in the rhizosphere and in the soil.

Here, we investigated whether the sulfo-ED pathway is active in SRDI565 and its potential role in utilizing plant-derived SQ or SQDG in the rhizosphere and in the soil. We show that SRDI565 can grow on SQ and sulfoquinovosyl glycerol (SQGro) as the sole carbon source. Growth on SQ leads to excretion of SL into the growth medium, indicating active sulfoglycolysis. This was supported by proteomic analyses, which showed that several proteins encoded by the sulfo-ED operon exhibit increased expression when bacteria are grown on SQ, while metabolomic analyses confirmed the presence of characteristic intermediates of the sulfo-ED pathway, as well as the unexpected production of intracellular DHPS. Overall, we show that SRDI565 has an active pathway for SQ utilization which may support growth of this bacterium in the environment; thus, this strain constitutes a new model organism for the study of the sulfo-ED pathway.

RESULTS

Analysis of the genome of SRDI565 revealed a sulfo-ED operon that had the same genes as but no synteny with the P. putida SQ1 operon (Fig. 1). Genes with high sequence identity to the P. putida genes included those encoding a putative SQase, SQ dehydrogenase, SL lactonase, SG dehydratase, KDSG aldolase, and SLA dehydrogenase and an SL exporter (see Fig. S1 to S6 in the supplemental material). The SRDI565 operon contains some important differences compared to that of P. putida SQ1. In particular, it lacks a putative SQ mutarotase gene (21) and appears to use an ABC transporter to import SQ/SQGro in place of an SQ/SQGro importer/permease. The putative sulfo-ED pathway in SRDI565 is consistent with the proposed protein functions outlined in Fig. 1b, with a comparison to the classical ED pathway in Fig. 1c.

Initial attempts were made to grow SRDI565 in completely defined medium, such as M9 minimal medium containing 125 μg ml⁻¹ biotin (22), to allow assessment of the effects of different carbon sources on bacterial growth. However, optimal growth could be achieved only by using a yeast extract-based medium (16). In particular, robust growth was achieved using a 5% dilution of 1 g liter⁻¹ yeast extract (Y_5% medium) containing 5 mmol mannitol (Y_5%M), while no detectable bacterial growth was observed on Y_5% medium alone. Significantly, SRDI565 also grew robustly on Y_5% medium containing 5 mM SQ (Y_5%SQ) and reached the same final optical density at 600 nm (OD₆₀₀) as in Y_5%M (Fig. 2a). SRDI565 also grew on Y_5% medium containing glucose, although to a lower final OD₆₀₀ than in Y_5%M or Y_5%SQ. ¹³C nuclear magnetic resonance (NMR) spectroscopic analysis of the culture medium of stationary-phase SRDI565 grown in Y_5%SQ revealed the presence of three major signals corresponding to SL (Fig. 2b). A fourth signal was also observed but not assigned and was also present in stationary-phase medium of cells grown on Y_5%M, suggesting that it is derived from other carbon sources in the yeast extract. SRDI565 also grew on Y_5% containing SQGro, but less robustly than on SQ.

FIG 2.

Growth of Rhizobium leguminosarum bv. trifolii SRDI565 on SQ produces SL as the major terminal metabolite. (a) Growth of SRDI565 on 5% yeast extract medium containing 5 mM SQ (solid circles) or 5 mM mannitol (open circles). The data are representative of 2 independent experiments. (b) ¹³C NMR (126 MHz) spectra of SQ (top), 5 mM SQ in 5% yeast extract medium (middle), and spent culture medium from growth of SRDI565 on 5 mM SQ (bottom). (c) ¹³C NMR (126 MHz) spectrum of spent culture medium from growth of SRDI565 on 5 mM [¹³C₆]SQ. The signal at δ 38.7 ppm is present in control experiments with SRDI565 grown on mannitol and is believed to derive from yeast extract. (d) Tabulated ¹³C NMR (126 MHz) data for [¹³C₃]SL from panel c. All samples contain 10% D₂O, added to allow frequency lock. (e) Quantitative proteomics was undertaken to identify proteins associated with sulfoquinovose catabolism versus mannitol. Examination of proteins observed to increase in abundance more than 4-fold revealed 17 proteins, including alpha-dehydro-beta-deoxy-d-glucarate aldolase (WP_017967308.1), highlighted in blue. (f) Growth in sulfoquinovose leads to the increase of multiple proteins associated with the TCA cycle, including NAD(P)-dependent oxidoreductase (WP_017965793.1), NADH-quinone oxidoreductase subunit NuoH (WP_017963854.1), NAD-dependent succinate-semialdehyde dehydrogenase (WP_017967313.1), and citrate synthase/methylcitrate synthase (WP_017964386.1).

We next examined changes in the proteome of SRDI565 cultivated on mannitol versus SQ. Label-free quantitative proteomic analysis of five experimental replicates of SRDI565 cultivated on each carbon source identified 2,954 proteins, with 1,943 proteins quantified in at least 3 experimental replicates under each growth condition (Table S1). Expression levels of 17 proteins potentially associated with SQ metabolism were significantly elevated [−log₁₀(P) of >2 and a fold change greater than 2 log₂ units] in bacteria cultivated in Y_5%SQ (Fig. 2e and f). In particular, a suspected KDSG aldolase (annotated as alpha-dehydro-beta-deoxy-d-glucarate aldolase; WP_017967308.1), a member of the proposed sulfo-ED pathway, was significantly increased [−log₁₀(P) of 4.74429 and a fold change of 2.38 log₂]. Consistent with the involvement of this pathway, we also observed a significant yet less dramatic increase in the proposed SQase (annotated alpha-glucosidase; WP_017967311.1) [−log₁₀(P) of 1.43643 and a fold change of 1.02 log₂]. Additional members of the predicted pathway expressed at higher levels in SQ-fed bacteria included the suspected SQ dehydrogenase (annotated as SDR family oxidoreductase; WP_017967310.1), identified by tandem mass spectrometry (MS/MS) events in 4 of 6 SQ experiments compared to 1 mannitol experiment, and the suspected SG dehydratase (annotated as dihydroxy-acid dehydratase, WP_017967307.1), identified by MS/MS events in 3 of 6 SQ experiments compared to 0 mannitol experiments; however, owing to their low abundance, they could not be accurately quantified (Fig. S7).

Other proteins that were significantly increased in SQ-fed bacteria included a NAD(P)-dependent oxidoreductase (WP_017965793.1), the NADH-quinone oxidoreductase subunit NuoH (WP_017963854.1), a NAD-dependent succinate-semialdehyde dehydrogenase (WP_017967313.1), and a citrate synthase/methylcitrate synthase (WP_017964386.1), supporting an alteration of the tricarboxylic acid (TCA) cycle and oxidative phosphorylation under conditions of growth on SQ (Fig. 2f).

To demonstrate activity for a representative sulfo-ED enzyme from SRDI565, we cloned and expressed the gene encoding the putative SQase. To support future structural studies, we expressed the N-terminal hexahistidine-tagged K375A/K376A variant, termed RlSQase*, a mutant enzyme whose design was guided by the Surface Entropy Reduction prediction (SERp) server (Fig. S8) (23). Size exclusion chromatography–multiple-angle light scattering (SEC-MALS) analysis of RlSQase* revealed that the protein exists as a dimer in solution (Fig. S8). Enzyme kinetics were analyzed using the chromogenic SQase substrate 4-nitrophenyl α-sulfoquinovoside (PNPSQ). PNPSQ was designed as an analogue of the natural substrate SQGro, and its hydrolysis results in release of the chromophore 4-nitrophenolate, which can be detected using UV-visible spectrophotometry with high sensitivity at 400 nm or at the isosbestic point, 348 nm (24, 25). RlSQase* exhibited a bell-shaped pH profile with an optimum at pH 7 to 8 and consistent with titration of catalytically important residues with a pK_a1 value of 6.5 ± 0.4 and a pK_a2 value of 8.6 ± 0.3. The enzyme displayed saturation kinetics with Michaelis-Menten parameters: k_cat = 1.08 ± 0.17 s⁻¹, K_m = 0.68 ± 0.25 mM, and k_cat/K_m = (1.59 ± 0.83) × 10³ M⁻¹ s⁻¹ (Fig. 3a and b). For comparison, the kinetic parameters for Agrobacterium tumefaciens SQase are as follows: k_cat = 22.3 ± 0.6 s⁻¹, K_m = 0.21 ± 0.03 mM, and k_cat/K_m = (1.1 ± 0.1) × 10⁵ M⁻¹ s⁻¹. Those for Escherichia coli SQase YihQ are as follows: k_cat = 32.7 ± 0.6 s⁻¹, K_m = 0.15 ± 0.01 mM, and k_cat/K_m = (2.2 ± 0.2) × 10⁵ M⁻¹ s⁻¹ (24).

FIG 3.

Rhizobium leguminosarum SRDI565 produces a functional sulfoquinovosidase that can be detected in cell lysates. (a) pH profile of RlSQase*. Specific activities were determined for hydrolysis of PNPSQ at the isosbestic point, 348 nm. (b) Michaelis-Menten plot of kinetic parameters for RlSQase* for hydrolysis of PNPSQ at 400 nm. (c) Analysis of sulfoquinovosidase activity of SRDI565 lysate grown on sulfoquinovose and mannitol. Cell lysates of soluble proteins derived from growth on SQ or mannitol was standardized for equal protein and SQase activity, measured using the chromogenic substrate PNPSQ at 400 nm. SQase activity was confirmed by inhibition by the azasugar inhibitor IFGSQ. Error bars denote standard errors of the means.

Direct evidence for enzymatic activity associated with the sulfo-ED operon in SRDI565 was obtained by measuring SQase enzyme activity in cell lysates. SRDI565 was grown to mid-logarithmic phase in Y_5%M and Y_5%SQ media, and the harvested cells were used to prepare a cell-free lysate containing soluble proteins. Incubation of both Y_5%M- and Y_5%SQ-derived lysates with PNPSQ resulted in production of 4-nitrophenolate at similar rates. The activity in the Y_5%SQ-derived lysate was inhibited by the addition of isofagomine-SQ (IFGSQ), an azasugar inhibitor of SQases that carries out key interactions in the active site that mimic those required for substrate recognition (Fig. 3c) (24). The similar levels of activity of SQase in both mannitol- and SQ-grown SRDI565 are consistent with the abundance of the putative SQase WP_017967311.1 detected by proteomic analysis.

To further confirm that a sulfo-ED pathway was operative in cells, a targeted metabolomics approach was used to detect expected intermediates in bacteria grown on Y_5%SQ medium. Detected intermediates were identified based on their liquid chromatography-MS/MS (LC-MS/MS) retention time and mass spectra with authentic reference standards of the sulfo-EMP and sulfo-ED pathway that were synthesized in-house. Sulfogluconate (SG) was synthesized by oxidation of SQ with iodine (26) (Fig. S9), while SQ, SF, sulfofructose-1-phosphate (SFP), DHPS, SLA, and SL were prepared as previously reported (27). SRDI565 was grown to mid-log phase in Y_5%M or Y_5%SQ and metabolically quenched, and extracted polar metabolites were analyzed by LC-MS/MS. SQ-grown bacteria contained SQ, SF, SG, SL, and DHPS, while SFP and SLA could not be detected (Fig. 4a to e). The detection of SG is characteristic of a sulfo-ED pathway and presumably arises from the action of the putative SQ dehydrogenase and SGL lactonase. The identification of DHPS and SF was unexpected, as these are intermediates or products of the sulfo-EMP pathway (12). BLAST analysis of the genome of SRDI565 did not identify putative genes for the sulfo-EMP pathway. SF may therefore be formed by the action of phosphoglucose isomerase (PGI), while DHPS could be the product of a promiscuous aldehyde reductase. SRDI565 was unable to utilize DHPS or SL as the sole carbon source in Y_5% medium, supporting the absence of an alternative pathway of sulfoglycolysis that utilizes these intermediates. Unexpectedly, cytosolic levels of DHPS were 20-fold higher than SL, suggesting that cells may lack a membrane transporter to export accumulated DHPS, in contrast to the SL transporter.

FIG 4.

Detection of sulfoglycolytic intermediates and end products in cytosolic extracts of SRDI565. SRDI565 was grown on Y_5%SQ medium and metabolically quenched by rapid cooling to 4°C, followed by extraction of cellular metabolites and LC-MS analysis. Sulfoglycolytic and glycolytic/neoglucogenic intermediates SQ (a), SG (b), SL (c), SF (d), and DHPS (e) were detected. In each panel, the upper portion corresponds to the collision-induced dissociation mass spectrum of chemically synthesized standard, while the lower portion is the equivalent mass spectrum for the metabolite identified in the cytosolic extract. (f) Relative mass spectrometric intensities of metabolites from cells grown on Glc or SQ.

NMR and LC-MS/MS analysis of the culture supernatant of both unlabeled and ¹³C₆-labeled SQ-cultivated SRDI565 confirmed that the substrate is almost completely consumed by the time bacteria reach stationary growth (final concentration of 0.006 ± 0.001 mM, compared to 5.0 ± 0.5 mM SQ in the starting medium) (Fig. S10). When a highly sensitive cryoprobe was used, ¹³C NMR spectroscopic analysis revealed that both DHPS and SG were present in culture supernatant of [¹³C₆]SQ-cultivated SRDI565. Quantitative LC-MS/MS analysis showed that consumption of SQ was associated with production of SL (5.70 ± 0.12 mM) and low levels of DHPS (0.081 ± 0.010 mM), SG (0.172 ± 0.006 mM), and SF (0.002 ± 0.0001 mM) (Table 1). This experiment was repeated to assess the effect of growth of SRDI565 but with SQGro as the carbon source. As noted previously, SRDI565 grows inconsistently on SQGro, and complete consumption of SQGro could not be achieved. However, the results of partial consumption broadly agreed with the results for growth on SQ, namely, that SL is the major terminal metabolite detected in the culture medium, with much smaller amounts of SF, SG, and DHPS (Table 1).

TABLE 1.

Analysis of sulfonate metabolites detected in spent culture medium of SRDI565 grown on 5.0 ± 0.5 mM SQ or SQGro

Metabolite	Metabolite concn (mM)a after growth on:
	SQ	SQGrob
SL	5.70 ± 0.12	3.14 ± 0.03
DHPS	0.081 ± 0.010	0.116 ± 0.002
SQ	0.006 ± 0.001	0.215 ± 0.001
SF	0.002 ± 0.0001	0.003 ± 0.0001
SG	0.172 ± 0.006	0.200 ± 0.008
SQGro		2.16 ± 0.06

^aMeasurements were performed in triplicate using LC-MS/MS. Values are means ± standard errors of the means (standard error estimate).

^bGrowth on SQGro was incomplete.

DISCUSSION

We demonstrate here that SRDI565 has a functional sulfo-ED pathway that allows this bacterium to utilize SQ as its major carbon source. Catabolism of SQ is primarily or exclusively mediated by a sulfo-ED pathway, with production of SL as the major end product, similar to the situation in P. putida SQ1, the only other experimentally described exemplar of this pathway (13). In contrast to P. putida SQ1, SRDI565 also produces trace amounts of DHPS, which could reflect the presence of enzymes which exhibit promiscuous activities similar to those in the conventional sulfo-EMP pathway. This observation is reminiscent of Klebsiella sp. strain ABR11, isolated from soil (28), which is also able to grow on SQ with production of both SL and DHPS. Klebsiella sp. strain ABR11 possesses an NAD⁺-specific sulfoquinovose-dehydrogenase activity (29), suggesting that it has an operative sulfo-ED pathway.

Various bacteria that can metabolize SQ have been isolated from soil, including Agrobacterium sp. (29), Klebsiella sp. (29), and Flavobacterium sp. (30), as well as P. putida SQ1 (13), which was isolated from a freshwater littoral sediment. These bacteria may work cooperatively with organisms, such as Paracoccus pantotrophus NKNCYSA, that can convert SL to mineral sulfur, leading to stoichiometric recovery of sulfite/sulfate (14). Together these bacterial communities achieve the complete mineralization of SQ to sulfate, which is available for use by plants.

Proteomic and biochemical evidence suggests that the sulfo-ED pathway is constitutively expressed in SRDI565, with only relatively small increases in protein expression, as shown by statistically significant increases in only KDSG aldolase and SQase in the presence of SQ. As SRDI565 in the soil is likely to be oligotrophic, constitutive expression of the sulfo-ED pathway may allow simultaneous usage of multiple nonglycolytic substrates without requirement for significant transcriptional changes. Consistent with this view, the proteomic abundance of the putative substrate-binding domain-containing protein LacI-type regulator WP_157386381.1 was unchanged between mannitol- and SQ-grown SRDI565. The sulfo-ED operon in SRDI565 differs from that described for P. putida SQ1 through the absence of a putative SQ mutarotase. SQ undergoes mutarotation with a half-life of approximately 6 h, which is much longer than that of the glycolytic intermediate Glc-6-P, which has a half-life of just seconds (21). Aldose mutarotases are often relatively nonspecific, and it is possible that a constitutive mutarotase not encoded in the sulfo-ED operon expressed by the cell provides this catalytic capacity. Alternatively, the SQ dehydrogenase may not be stereospecific, with the ability to act on both anomers of SQ, or it may even act on α-SQ (the product released from SQGro by an SQase) at a high rate, such that mutarotation to β-SQ is insignificant. A second difference in the sulfo-ED operon is the presence of an ABC transporter gene. ABC transporters are the most common solute transporters and can translocate their substrates in either a forward or reverse direction (31). While we propose that the ABC transporter operates in the forward direction, based on the presence of a signal sequence in the putative solute binding domain targeting it to the periplasm and consistent with a wide range of sugar import systems, the directionality of transport and thus the choice of substrate (SQ/SQGro versus SL) may depend on the relative abundance of these metabolites intra- and extracellularly.

Sulfoglycolysis in SRDI565 leads to production of pyruvate and the excretion of the C₃-organosulfonate SL (Fig. 5). In order to satisfy the demands of the pentose phosphate pathway and cell wall biogenesis, sulfoglycolytic cells must synthesize glucose-based metabolites, such as glucose-6-phosphate and glucose-1-phosphate. Gluconeogenesis has been studied in Rhizobium leguminosarum strain MNF3841 and operates through a classical pathway involving fructose bisphosphate aldolase (32). Action of phosphoglucose isomerase on SQ might lead to production of SF, thereby explaining the observation of this metabolite in SRDI565. This is not likely to be consequential, as the reversibility of this reaction should allow complete consumption of any SF through isomerization back to SQ. The formation of DHPS may result from a promiscuous aldehyde reductase. Analysis of spent culture medium reveals that the production of DHPS is minor in terms of total carbon balance. However, within the cytosol, DHPS accumulates to much higher levels than SL, presumably because of the absence of a dedicated exporter for the former. Possibly, reduction of SLA to DHPS is reversible and enables conversion of this metabolite to SL and subsequent excretion from the cell. The observation of SG, SF, and DHPS in the spent culture medium at low levels is suggestive of low levels of leakage of these metabolites from the cell, through either cell lysis or leaky export systems.

FIG 5.

Proposed pathway for SQ metabolism in Rhizobium leguminosarum SRDI565.

Given that SQ contains a significant portion of organic sulfur within plants, the pathways of SQ catabolism leading to release of its sulfur may be important to enable recycling of this important macronutrient. Plants can use only sulfate, which is poorly retained by most soils. Biomineralization of organic sulfur to sulfate is important to allow plants to access this element. As one of just two known pathways for the catabolism of SQ, the sulfo-ED pathway is likely to be an important part of environmental breakdown of SQ and may contribute to the persistence of symbiotic rhizobia within the pedosphere. The present work lays the groundwork for a more detailed investigation of sulfoglycolysis in a well-characterized bacterium with an established capability for symbiosis with a leguminous plant host.

MATERIALS AND METHODS

Reagents.

SQ, [¹³C₆]SQ, SF, SFP, SLA, SL, and DHPS were chemically and chemoenzymatically synthesized as described previously (27). IFGSQ was chemically synthesized as described previously (24).

Bacteria and culture conditions.

Rhizobium leguminosarum bv. trifolii SRDI565 was a gift from Ross Ballard (South Australian Research and Development Institute, Adelaide, South Australia, Australia). Minimal salts medium consists of 0.5 g liter⁻¹ K₂HPO₄, 0.2 g liter⁻¹ MgSO₄, 0.1 g liter⁻¹ NaCl, and 0.33 g liter⁻¹ CaCl₂, adjusted to pH 7.0. Y_5%M consists of minimal salts medium plus 50 mg liter⁻¹ yeast extract and 5 mM mannitol. Y_5%SQ consists of minimal salts medium plus 50 mg liter⁻¹ yeast extract and 5 mM SQ.

Growth curves were determined in a MicrobeMeter built in-house according to published plans (33) and blueprints available at https://humanetechnologies.co.uk/download-microbemeter/. The MicrobeMeter was calibrated by performing serial 2-fold dilutions across the detection range of the MicrobeMeter (0 to 1,023 U), starting with a culture of SRDI565 with an OD₆₀₀ of approximately 1. OD₆₀₀ measurements were made with a UV-visible spectrophotometer and plotted against the reading of the MicrobeMeter. The data were fitted to a polynomial to obtain a calibration curve.

(i) Proteomic sample preparation.

Cells were washed 3 times in phosphate-buffered saline (PBS), collected by centrifugation at 10,000 × g at 4°C, and then snap-frozen. Frozen whole-cell samples were resuspended in 4% sodium dodecyl sulfate (SDS)–100 mM Tris (pH 8.0)–20 mM dithiothreitol (DTT) and boiled at 95°C with shaking at 2,000 rpm for 10 min. Samples were then clarified by centrifugation at 17,000 × g for 10 min, the supernatant was collected, and protein concentration was determined using a bicinchoninic acid assay (Thermo Scientific Pierce). One hundred micrograms of protein from each sample was cleaned using SP3-based purification according to previously published protocols (34). Briefly, reduced samples were cooled and then alkylated with 40 mM 2-chloroacetamide (CAA) for 1 h at room temperature in the dark. The alkylation reactions were quenched with 40 mM DTT for 10 min, and then samples were precipitated onto SeraMag Speed Beads (GE Healthcare, USA) with ethanol (final concentration, 50% [vol/vol]). Samples were shaken for 10 min to allow complete precipitation onto beads and then washed three times with 80% ethanol. The precipitated-protein-covered beads were resuspended in 100 mM ammonium bicarbonate containing 2 μg trypsin (1/50 [wt/wt]) and allowed to digest overnight at 37°C. Upon completion of the digests, samples were centrifuged at 14,000 × g for 5 min to pellet the beads, and the supernatant was collected and desalted using homemade C₁₈ stage tips (35). The eluted material was dried and stored until LC-MS analysis.

(ii) Proteomics analysis using reversed-phase LC-MS.

Purified peptides prepared were resuspended in buffer A* (2% acetonitrile [ACN], 0.1% CF₃CO₂H) and separated using a two-column chromatography setup composed of a PepMap100 C₁₈ 20-mm by 75-μm trap and a PepMap C₁₈ 500-mm by 75-μm analytical column (Thermo Fisher Scientific). Samples were concentrated onto the trap column at 5 μl/min for 5 min and infused into an Orbitrap Elite mass spectrometer (Thermo Fisher Scientific). Gradients of 120 min were run in which the buffer composition was altered from 1% buffer B (80% ACN, 0.1% formic acid) to 28% B over 90 min, from 28% B to 40% B over 10 min, and from 40% B to 100% B over 2 min, held at 100% B for 3 min, dropped to 3% B over 5 min, and held at 3% B for another 10 min. The Elite Orbitrap mass spectrometer was operated in a data-dependent mode, automatically switching between the acquisition of a single Orbitrap MS scan (120,000 resolution) and a maximum of 20 MS/MS scans (collision-induced dissociation [CID] normalized collision energy, 35; maximum fill time, 100 ms; automatic gain control [AGC], 1 × 10⁴).

(iii) Mass spectrometry data analysis.

Proteomic comparison of growth with and without sulfoquinovose was accomplished using MaxQuant (v1.5.5.1) (36). Searches were performed against Rhizobium leguminosarum bv. trifolii SRDI565 (NCBI taxonomy ID 935549; downloaded 8 January 2019; 6,404 entries) with carbamidomethylation of cysteine set as a fixed modification. Searches were performed with trypsin cleavage, allowing 2 miscleavage events and the variable modifications of oxidation of methionine and acetylation of protein N termini. The precursor mass tolerance was set to 20 ppm for the first search and 10 ppm for the main search, with a maximum false discovery rate (FDR) of 1.0% set for protein and peptide identifications. To enhance the identification of peptides between samples, the “match between runs” option was enabled, with a precursor match window set to 2 min and an alignment window of 10 min. For label-free quantitation, the MaxLFQ option within MaxQuant (37) was enabled in addition to the requantification module. The resulting peptide outputs were processed within the Perseus (v1.4.0.6) (38) analysis environment to remove reverse matches and common protein contaminants with missing values imputed.

Enzyme kinetics of RlSQase*. (i) Michaelis-Menten plot.

Kinetic analysis of RlSQase* was performed using PNPSQ as the substrate, using a UV-visible spectrophotometer to measure the release of the 4-nitrophenolate (λ = 348 nm). Assays were carried out in 50 mM sodium phosphate–150 mM NaCl (pH 7.2) at 30°C using 212 nM RlSQase* at substrate concentrations ranging from 0.05 μM to 4 mM. Using an extinction coefficient for 4-nitrophenolate of 5.125 mM⁻¹cm⁻¹, kinetic parameters were calculated using Prism.

(ii) pH profile.

For the determination of pH profile, specific activities of RlSQase* were monitored by measuring absorbance changes at a wavelength of 348 nm in the presence of sodium acetate buffer (pH 5.6), sodium phosphate buffer (pH 6.0 to 8.5), and glycine NaOH buffer (pH 8.8 to 9.2). The assays were performed at 30°C in duplicate, and specific activities were determined using an extinction coefficient for p-nitrophenol (PNP) of 5.125 mM⁻¹cm⁻¹ at the isosbestic point (348 nm). One unit of SQase activity is defined as the amount of protein that releases 1 μmol PNP per min.

Cloning, expression, and kinetic analysis of RlSQase*.

The gene sequence coding for the RlSQase* SERp mutant was synthesized with codon optimization for expression in E. coli and was cloned within a pET-28a(+) vector with C-terminal His tag through GenScript. The plasmid His₆-RlSQase*-pET-28a(+) containing the gene for target RlSQase* was transformed into E. coli BL21(DE3) cells for protein expression. Precultures were grown in LB medium (5 ml) containing 30 μg/ml for 18 h at 37°C and 200 rpm. Cultures (1 liter of LB medium supplemented with 30 μg/ml kanamycin) were inoculated with the preculture (5 ml) and incubated at 37°C and 200 rpm until an OD₆₀₀ of 0.6 to 0.8 was reached. Protein expression was induced by addition of IPTG (isopropyl-β-d-thiogalactopyranoside; 1 mM), and shaking was continued overnight (20 to 22 h) at 18°C and 200 rpm. The cells were harvested by centrifugation (5,000 rpm, 4°C, 20 min), resuspended in 50 mM Tris–300 mM NaCl (pH 7.5) buffer, and subjected to further cell lysis. Cells were disrupted using a French press under a pressure of 20,000 lb/in², and the lysate was centrifuged at 50,000 × g for 30 min.

The N-terminal His₆-tagged protein was purified by immobilized metal ion affinity chromatography, followed by size exclusion chromatography (SEC) (Fig. S8). The lysate was loaded onto a pre-equilibrated nickel-nitrilotriacetic acid (Ni-NTA) column, followed by washing with loading buffer (50 mM Tris-HCl, 300 mM NaCl, 30 mM imidazole [pH 7.5]). The bound protein was eluted using a linear gradient with buffer containing 500 mM imidazole. Protein-containing fractions were pooled, concentrated, and loaded onto a HiLoad 16/600 Superdex 200 gel filtration column pre-equilibrated with 50 mM Tris-HCl–300 mM NaCl (pH 7.5) buffer. The protein was concentrated to a final concentration of 60 mg ml⁻¹ using a Vivaspin 6 with a 300-kDa molecular weight cutoff membrane for characterization and enzyme assays.

SEC-MALS analysis.

SEC-MALS experiments were conducted on a system comprising a Wyatt HELEOS-II multiangle light scattering detector and a Wyatt rEX refractive index detector linked to a Shimadzu high-performance liquid chromatography (HPLC) system (SPD-20A UV detector, LC20-AD isocratic pump system, DGU-20A3 degasser, and SIL-20A autosampler). Experiments were conducted at room temperature (20 ± 2°C). Solvents were filtered through a 0.2-μm filter prior to use, and a 0.1-μm filter was present in the flow path. The column was equilibrated with at least 2 column volumes of buffer (50 mM Tris, 300 mM NaCl [pH 7.5]) before use, and buffer was infused at the working flow rate until baselines for UV, light scattering, and refractive index detectors were all stable. The sample injection volume was 100 μl RlSQase* at 6 mg/ml in 50 mM Tris buffer–300 mM NaCl (pH 7.5). Shimadzu LC Solutions software was used to control the HPLC, and Astra V software was used for the HELEOS-II and rEX detectors (Fig. S8). The Astra data collection was 1 min shorter than the LC Solutions run to maintain synchronization. Blank buffer injections were used as appropriate to check for carryover between sample runs. Data were analyzed using the Astra V software. Molar masses were estimated using the Zimm fit method with degree 1. A value of 0.158 was used for the protein refractive index increment (dn/dc).

Detection of SQase activity in cell lysates.

SRDI565 was grown in 50 ml Y_5%M and Y_5%SQ media at 30°C to mid-log phase, equivalent to an OD₆₀₀ of approximately 0.2, measured using a Varian Cary50 UV-visible spectrophotometer. Cells were harvested by adding a 3× volume of ice-cold PBS to metabolically quench the samples and then centrifuged at 2,000 × g and 4°C for 10 min. The supernatant was discarded, and the cells were washed 3 times with ice-cold PBS, with each wash involving resuspension and centrifugation at 2,000 × g and 4°C for 10 min. The cells were collected once more by centrifugation at 10,000 × g and 4°C for 1 min, snap-frozen in liquid nitrogen, and stored at –80°C.

Cells were lysed by addition of 1,000 μl prechilled PBS, 1 μl RNase A, 1 μl DNase, 1 μl hen egg white lysozyme (100 mg ml⁻¹; Sigma), and a 1× final concentration of Complete EDTA-free protease inhibitor cocktail (Roche) to the cell pellet. The cells were gently resuspended and mixed at 4°C for 10 min. The suspension was placed on ice, irradiated with a Sonoplus HD3200 MS 73 sonicator probe (Bandelin) at a frequency of 20 kHz, 20% amplitude, and a pulse of 2 s on and 8 s off, repeated for a total time of sonication to 150 s, and then incubated on ice for 5 min. The suspension was clarified by centrifuging at 14,000 × g and 4°C for 1 min, and the supernatant was filtered through a Nanosep mini-centrifugal spin column with a 0.2-μm filter (Pall) into a 1.5-ml Eppendorf tubes and stored at 4°C. Protein concentration was determined using a bicinchoninic acid assay.

SQase activity was measured in triplicate using PNPSQ and an Agilent Cary UV workstation (G5191-64000) at 30°C. Reaction mixtures contained buffer consisting of 50 mM NaP_i and 150 mM NaCl (pH 7.4) and 2.5 mM PNPSQ. Reactions were initiated by addition of SQ- or mannitol-derived lysate to a final concentration of 43.7 μg ml⁻¹ protein, and absorbance was monitored at 400 nm for 3 h. After 3 h, IFGSQ was added to a final concentration of 6.25 mM to the SQ-lysate sample, and absorption was monitored for 3 h.

Metabolite analysis of R. leguminosarum cell extracts. (i) Metabolic quenching and extraction.

SRDI565 was grown on Y_5%SQ or Y_5% medium containing 35 mM glucose to mid-logarithmic phase (OD₆₀₀ of approximately 0.15), calculated based on the OD₆₀₀ measured by a Cary 50 UV-visible spectrophotometer, and was rapidly quenched in a prechilled 15-ml Falcon tube containing phosphate-buffered saline (PBS) at 4°C. Ice-cold PBS (11 ml) was infused into cell culture medium (4 ml). The Falcon tubes were mixed by inversion, incubated in an ice-water slurry for 5 min, and then centrifuged at 2,000 × g at 1°C for 10 min. The supernatant was removed by aspiration, and cell pellets were washed twice with 1 ml of ice-cold PBS (with resuspension each time) and transferred into 1.5-ml Eppendorf tubes. Cells were pelleted by centrifugation at 14,000 rpm, and residual solvent was carefully removed. Cell pellets were stored at –80°C until extraction. Cells were extracted in 200 μl of extraction solution (methanol-water, 3:1 [vol/vol]) containing an internal standard, 5 μM [¹³C₄]aspartate (Cambridge Isotopes), and subjected to 10 freeze-thaw cycles to facilitate cell lysis (30 s in liquid nitrogen followed by 30 s in a dry ice-ethanol bath). Debris was pelleted by centrifugation at 14,000 rpm (1°C for 5 min), and cell lysate was transferred into an HPLC vial insert for LC-MS analysis.

(ii) LC-MS analysis and identification of sulfonate metabolites.

Separation and detection of polar metabolites were performed using an Agilent Technologies 1200 series high-performance liquid chromatograph (HPLC) coupled to a quadrupole time-of-flight mass spectrometer (6545 QTOF; Agilent Technologies) using a method modified from that of Masukagami et al. (39). Metabolite extracts were transferred into 2-ml autosampler vials with glass inserts and placed in the autosampler and held at 4°C prior to analysis. Metabolite separation was performed by injecting 7 μl of the extract into a SeQuant ZIC-pHILIC PEEK (polyetheretherketone)-coated column (150 mm by 4.6 mm, 5-μm polymer; Merck Millipore) maintained at 25°C, with a gradient of solvent A (20 mM ammonium carbonate [pH 9.0]; Sigma-Aldrich) and solvent B (100% acetonitrile, hypergrade for LC-MS [LiChrosolv; Merck]) at a flow rate of 0.3 ml/min. A 33.0-min gradient was set up as follows: 0 min, 80% B; 0.5 min, 80% B; 15.5 min, 50% B; 17.5 min, 30% B; 18.5 min, 5% B; 21.0 min, 5% B; 23.0 min, 80% B.

The LC flow was directed into an electrospray ionization (ESI) source with a capillary voltage of 2,500 V operating in negative ionization mode. Drying nitrogen gas flow was set to 10 liters/min, and sheath gas temperature and nebulizer pressure were set to 300°C and 20 lb/in², respectively. The voltages of fragmentor and skimmer were set at 125 V and 45 V, respectively. Data were acquired in MS and MS/MS mode, with scan ranges of 60 to 1,700 m/z and 100 to 1,700 m/z, respectively, at a rate of 1.5 spectra/s. MS/MS acquisition was performed with four collision energies (0, 10, 20, and 40 V). The mass spectrometer was calibrated in negative mode prior to data acquisition, and mass accuracy during runs was ensured by a continuous infusion of reference mass solution at a flow rate of 0.06 ml/min (API-TOF reference mass solution kit; Agilent Technologies). Data quality was ensured by multiple injections of standards (each at a 1.5 μM concentration) and a pooled biological sample (a composite of cell extracts) used to monitor the instrument performance. Samples were randomized prior to metabolite extraction and LC-MS analysis.

(iii) Standard preparation.

Standards of selected metabolites (Table S1) were prepared at 10 μM in 80% acetonitrile (hypergrade for LC-MS [LiChrosolv; Merck]) and injected separately into a column connected to the mass spectrometer interface. Retention time and detected molecular ion were used to create a targeted MS/MS acquisition method. The spectra, mass-to-charge ratio (m/z), and retention time were imported into a personal compound database and library (PCDL Manager, version B.07.00; Agilent Technologies) used in the data processing workflow.

(iv) Data analysis.

Data were analyzed using MassHunter qualitative and quantitative analysis software (version B.07.00; Agilent Technologies). Identification of metabolites was performed in accordance with metabolite identification (Metabolomics Standard Initiative) level 1 based on the retention time and molecular masses matching authentic standards included in the personal database and library. Peak integration was performed with MassHunter quantitative software (version B.07.00; Agilent Technologies) on the spectra from identified metabolites.

Chemical synthesis of 6-deoxy-6-sulfo-d-gluconate.

NaOH in methanol (4% [wt/vol]; 4 ml) was added dropwise to a stirred solution of sulfoquinovose (100 mg; 0.410 mmol) and iodine (209 mg; 1.65 mmol) in water (1 ml) and methanol (4 ml) held at 40°C. As the sodium hydroxide was added, the color of iodine dissipated. The solvent was evaporated under reduced pressure, and the crude residue was subjected to flash chromatography (ethyl acetate [EtOAc]-methanol [MeOH]-H₂O, 4:2:1 to 2:2:1, then water) to give the 6-deoxy-6-sulfogluconate sodium salt (89.2 mg). An aqueous solution of the sodium salt was eluted through a column of Amberlite IR120 (H⁺ form) resin. The acidic eluate was collected and concentrated under reduced pressure to give SG (71.3 mg, 67%). ¹H NMR (400 MHz, D₂O): δ 4.23 to 4.15 (1H, m, H2), 4.13 (1 H, d, J = 3.3 Hz, H3), 4.05 (1 H, t, J = 3.2 Hz, H5), 3.74 (1 H, dd, J = 6.5, 3.4 Hz, H4), 3.35 (1 H, d, J = 14.6 Hz, H6a), 3.05 (1 H, dd, J = 14.6, 9.7 Hz, H6b); ¹³C (¹H) NMR (100 MHz, D₂O) δ 178.7 (C1), 74.2 (C4), 73.8 (C2), 70.8 (C3), 67.8 (C5), 53.4 (C6); HRMS (ESI^–) for C₆H₁₁O₉S [M^–]: calculated, 259.0129; found, 259.0131.

Quantitation of metabolite levels in spent culture medium.

The metabolites (DHPS, SF, SQ, SL, and SG) present in spent culture medium were quantified against standard solutions of pure metabolites by LC-ESI-MS/MS. Quantification was done with the aid of calibration curves generated by dissolving the pure standards in spent medium from SRDI565 grown on Y_5%M. Spiked spent medium was diluted 100-fold with water and then analyzed by LC-MS/MS with α-MeSQ as the internal standard. For quantitation of metabolites, spent culture medium from SRDI565 grown in Y_5%SQ or Y_5%SQGro was diluted 100-fold with water and analyzed by LC-MS/MS with α-MeSQ as the internal standard.

LC-ESI-MS/MS analysis was performed using a TSQ Altis triple-quadrupole mass spectrometer (Thermo Fisher Scientific) coupled with a Vanquish Horizon UHPLC system (Thermo Fisher Scientific). The column was a ZIC hydrophilic interaction liquid chromatography (HILIC) column (5 μm, 50 by 2.1 mm; Merck). The HPLC conditions were as follows: from 90% B to 40% B over 15 min, 40% B for 5 min, and back to 90% B over 1 min (solvent A, 20 mM NH₄OAc in 1% acetonitrile; solvent B, acetonitrile); flow rate, 0.30 ml min⁻¹; injection volume, 1 μl. The mass spectrometer was operated in negative ionization mode. Quantitation was done using the MS/MS selected reaction monitoring (SRM) mode using Thermo Scientific Xcalibur software and normalized with respect to the internal standard, α-MeSQ. Prior to analysis, for each analyte, the sensitivity for each SRM-MS/MS transition was optimized. Results were as follows: for DHPS, ESI–MS/MS m/z of [M-H]⁻, 155; product ions, 137, 95; retention time, 4.91 min; for α-MeSQ (internal standard), ESI-MS/MS m/z of [M-H]⁻, 257; product ions, 166, 81; retention time, 6.31 min; for SF, ESI-MS/MS m/z of [M-H]⁻, 243; product ions, 207, 153; retention time, 6.81 min; for SQ, ESI-MS/MS m/z of [M-H]⁻, 243; product ions, 183, 123; retention time, 7.58 and 7.89 min for α/β; for SL, ESI-MS/MS m/z of [M-H]⁻, 169; product ions, 107, 71; retention time, 9.26 min; for SG, ESI-MS/MS m/z of [M-H]⁻, 259; product ions, 241, 161; retention time, 9.66 min; and for SQGro, ESI-MS/MS m/z of [M-H]⁻, 317; product ions, 225, 165; retention time, 7.15 min.

Data availability.

The mass spectrometry proteomics data have been deposited with the ProteomeXchange Consortium via the PRIDE partner repository with the data set identifier PXD015822.