Introduction

Escherichia coli transport and use diverse C4-dicarboxylates (succinate, malate, aspartate or fumarate) in antiport manner or symport with H+ during aerobic or anaerobic growth. Among known C4-dicarboxylate transporters are DctA as well as the Dcu family DcuA, DcuB, DcuC and the putative DcuD transporter1. It is well established that DctA is important for aerobic growth on C4-dicarboxylates. Dcu carriers are different from DctA and form a separate group. It has been suggested that DcuA, encoded by dcuA gene, catalyzes the uptake of succinate or fumarate and is active either in aerobic or anaerobic conditions. The other carriers (DcuB, DcuC) are expressed only under anaerobic conditions1,2. It was clearly shown that DcuB is the major C4-dicarboxylate carrier under anoxic conditions. DcuC, encoded by the dcuC gene, is synthesized under anaerobic conditions and during glucose fermentative conditions is suggested to function preferably as an efflux carrier1,3. Gene expression data showed that fumarate or other C4-dicarboxylates might increase the gene expression level of several carriers4. But substitution of glucose by glycerol did not affect dcuC expression, thus it can be assumed that dcuC is not subject to catabolite repression and DcuC is needed for succinate efflux during glucose fermentation1,3. To be critical, it must be mentioned that glycerol substituted to glucose was used in the medium with the presence of fumarate, and glucose fermentation cannot be compared to glycerol fermentation, as fumarate respiration takes place. Moreover, these carrier proteins are dependent on external pH and lack of Dcu function in the cells resulted in aerobic growth on succinate when external pH was below 6.01.

The fourth DcuD carrier, encoded by dcuD gene (formerly yhcL), is not expressed under most of the conditions tested and its physiological role is still unknown5. Lately, a work demonstrated that in dcuD mutant the product yields of molecular hydrogen H2 and ethanol are improved6. Moreover, by deletion dcuB and dcuC but not dcuA and dcuD genes resulted in the increase of succinate production by 34%3. In addition, during glucose fermentation the deletions of dcuB and dcuC resulted in 90% decrease of succinate titer suggesting that DcuB and DcuC are responsible for succinate efflux under the latest conditions3.

A decade ago it was shown that glycerol can be fermented by E. coli under anaerobic conditions at different pH values7,8,9. Depending on external pH fermentation end products are various, and key bioenergetics parameters such as membrane potential, pH gradient and thus proton motive force (ΔµH+) values are also different, compared to glucose fermentative conditions10,11,12,13. One of the key enzymes for E. coli growth under anaerobic conditions is the proton translocating FOF1-ATPase, which is the main ΔµH+ generator. It has been experimentally shown that the FOF1-ATPase activity is necessary for the activity of membrane bound [Ni-Fe] hydrogenase (Hyd) enzymes, which are responsible for H2 metabolism and potassium (K+) transport enzymes such as Trk or others13,14. FO subunit of proton FOF1-ATPase is located inside the cytoplasmic membrane and contains a, b, and c subunits15,16. The extra-membranous F1 subunit is attached to the FO part, and in F1 ATP hydrolysis takes place under fermentative conditions15. Particularly, during glucose or glycerol fermentation Hyd-1 or Hyd-2 depend on the active FOF1-ATPase. Moreover, this link or metabolic cross-talk depends on external pH and other conditions17. The results were obtained by inhibiting the proton FOF1-ATPase with N,N’-dicyclohexylcarbodiimide (DCCD), a specific inhibitor of the E. coli FOF1-ATPase under anaerobic conditions18, or applying atp (DK8) mutant which do not have FOF1-ATPase19.

During glycerol fermentative conditions, the role of different carriers such as Dcu is not known because when the experiments were carried out with glycerol and fumarate1,2,3 the metabolism goes to fumarate respiration but not to glycerol fermentation. At that time glycerol fermentation was not known yet. So the current work describes novel properties of Dcu carriers and, especially previously unknown role of DcuD during glycerol fermentation at pH 7.5 and 5.5.

Results and Discussion

ATPase activity and H+ fluxes of E. coli wild type and dcu mutants at pH 7.5 and pH 5.5

In 2006 Gonzalez group7 experimentally demonstrated that E. coli can ferment glycerol at slightly acidic and further by our group at slightly alkaline pHs9. Moreover, responsible Hyd enzymes and relationship between these enzymes with main enzyme of bioenergetic relevance – the proton FOF1-ATPase during fermentation have been determined11. To understand what is the role of C4-dicarboxylate carriers (Dcu) during glycerol fermentation the activity of proton FOF1-ATPase has been investigated. For this objective, total and DCCD inhibited ATPase activity at pH 7.5 and pH 5.5 has been defined. It is well established that DCCD is the specific inhibitor of FOF1-ATPase under anoxic conditions18.

The highest total ATPase activity was determined at pH 7.5 in dcuABCD mutant membrane vesicles resulting in 139.6 nMol Pi/(min µg protein) (Fig. 1A) compared to wild type and dcu mutants. At pH 7.5 in wild type cells the total ATPase activity was higher by 44% compared to the cells grown at acidic pH 5.5 (compare Fig. 1A,B). In order to indicate the contribution of proton FOF1-ATPase in the total ATPase activity DCCD-inhibited ATPase activity was determined (see Methods). It was shown that at pH 7.5 in wild type and dcuABC mutant membrane vesicles DCCD the ATPase activity was inhibited 2 and 1.9 fold, respectively, suggesting that FOF1 contributed to total ATPase activity by 50% under the conditions mentioned. Interestingly, it was shown that in dcuABCD mutant DCCD inhibited by 22% more the ATPase activity compared to wild type and dcuABC triple mutant. But surprisingly in dcuD single mutant the DCCD inhibition was much higher and 82% of total ATPase activity came from FOF1 (see Fig. 1A).

Figure 1
figure 1

ATPase activity of membrane vesicles of E. coli BW25113 wild type, dcuABC, dcuABCD, dcuD mutant strains at pH 7.5 (A) and pH 5.5 (B). The DCCD (0.1 mM) was added into the assay medium when indicated. K+ (100 mM) was added in the assays when shown. The assays pH was the same as growth pH. Bacteria were grown at pH 7.5 or pH 5.5 in the presence of 10 g L−1 glycerol as carbon source at 37 °C. For the strains see Table 1; for others, see Methods.

Full size image

The data suggest that there might be some relationship or interaction between DcuD protein and the FOF1-ATPase at pH 7.5. Particularly, it had been reviewed1 that during respiration with the transport of succinate2-3H+ are symported but under fermentative conditions the amount of H+ that are symported during succinate efflux is not known yet. The interrelationship between DcuD and FOF1 can take place if the DcuB and DcuC efflux succinate as well as the DcuD symport protons via interacting with FOF1. Similar interaction of E. coli potassium transport Trk system with FOF1 had been shown before12,13,14,20. In addition, it was shown that Na+ ions were important for the transport of C4 dicarboxylates in W. succinogenes and the absence of Na+ ions during fumarate respiration disturbs generation of proton motive force21. Moreover, it is possible that DcuD interact with FOF1 via thiol groups by having dithiol-disulfide interchange and by this way translocate protons. The possibility of the involvement of thiol groups in the above mentioned reaction was reviewed before14. In addition, recently “hydrogenase complex” idea as “proton sensor” had been proposed and in this model Dcu carriers might interact to regulate the proton gradient22.

The role of DcuD in proton translocation suggested was also confirmed by the H+ flux determination in this mutant (Fig. 2). It was clearly demonstrated that when cells were grown on glycerol, and in the assays glycerol was added no significant differences had been detected in the dcuD and dcuABCD mutants compared to wild type regarding the H+ efflux (see Fig. 2). Only in dcuABC mutant, where only dcuD is present, the H+ efflux was decreased by 26%. This suggests that DcuD might be involved in H+ efflux thus regulating the transmembrane pH gradient via interacting with other systems.

Figure 2
figure 2

H+ efflux by whole cells of E. coli wild type and dcu mutants during glycerol fermentation at pH 7.5 and pH 5.5. In assays glycerol was used in the same concentration as in growth medium and succinate was added in concentration of 5 mM. For mutant strains, see Table 1; for the others, see the Methods section and the legend to Fig. 1.

Full size image

But when succinate was added in the assays the H+ fluxes in dcuD and dcuABCD mutants were higher by 1.6 and 2.07 folds, respectively, compared to wild type at pH 7.5. From the data obtained it is suggested that when DcuD is absent other Dcu carriers or membrane systems pump H+ out. There might be therefore some compensatory H+ pumping mechanism in Dcu carriers, which must be more deeply investigated to understand the likely compensatory function. Similar compensatory uptake or producing functions have been suggested for Hyd enzymes or for formate uptake or export Foc channels23,24. In addition, externally added succinate enters the cell in the absence of Dcu family carriers, which could indicate, in accordance with previously shown possibilities, that other C4 uptake systems can take the role of Dcu or other carriers1,2,6. But during anaerobic conditions there might be limited possibility of compensatory uptake functions of C4 uptake, an issue that must be further analyzed.

When cells were grown at pH 5.5 the highest total ATPase activity was obtained in dcuABC mutant membrane vesicles resulting in 122.84 nMol Pi/(min µg protein) (Fig. 1B) compared to that obtained in wild type and the dcu mutants. In dcuD and dcuABCD mutants the total ATPase activity was decreased by 29% and 35%, respectively, compared to wild type (see Fig. 1B). DCCD inhibited ATPase activity mainly in dcuABC and dcuD mutants by 89% and 48%, respectively. No inhibition was determined in wild type and DCCD inhibited the ATPase activity by 31% in dcuABCD mutant. These results suggest that the mutations might affect the FOF1-ATPase conformational change or there must be a direct or indirect link with FOF1 at this pH also. Moreover, the H+ flux measured in whole cells showed that at pH 5.5 when glycerol was added the flux was very low in wild type and mutant cells. Addition of succinate in assays resulted in increase of H+ flux reaching 0.9 mM/min in wild type but not mutant cells compared to the assays supplemented with glycerol where the H+ flux was 0.19 mM/min (see Fig. 2). In all dcu mutants the H+ flux was lowered but only in dcuD single mutant it was absent which suggest that at low pH, DcuD mainly contribute to H+ efflux across the membrane during glycerol fermentation. As the role of DcuD is not clear at all, it might be possible that it is involved in inter-membrane proton translocation and depending on pH either it pumps H+ out of the cell or translocates it to other membrane bound enzymes or transport systems.

Role of potassium ions in ATPase activity of E. coli wild type and dcu mutants membrane vesicles at pH 7.5 and pH 5.5

Earlier in many papers it has been described that K+ have significant role in FOF1-ATPase activity20,25. In addition, as stated above and model was proposed according to which K+ transport Trk system, Hyd-4 interacts with FOF1-ATPase and forms protein-protein complex at pH 7.5 during glucose fermentation14. Moreover, recently, it was shown that during mixed carbon (glucose, glycerol and formate) fermentative conditions formate dehydrogenase (Fdh) has some link or relationship with FOF1 depending on K+ at pH 5.526. All this suggests that there must be some effect of K+ on C4-dicarboxylate carriers such as the anaerobic Dcu family.

In order to reveal the role of K+, cells membrane vesicles total and DCCD inhibited ATPase activity was determined in the presence of K+ (100 mM) (see Fig. 1). It was shown that K+ had stimulatory effect on wild type cells at pH 7.5 but not at pH 5.5. Especially, at pH 7.5 total ATPase activity was stimulated by 40% and DCCD inhibited more FOF1-ATPase in the presence of K+ which is in good conformity with stimulatory effect of K+ on FOF1 during glucose fermentation20,25,27. But when analyzing the mutants only in dcuD mutant, K+ stimulated the total ATPase activity by 17% but not DCCD inhibited one (see Fig. 1). DCCD totally inhibited ATPase activity only in dcuABCD mutant. The data clearly demonstrate that K+ affects C4-dicarboxylate Dcu carrier system and there is a link between FOF1, Dcu and K+ transport system.

Specific growth rate and H2 production in E. coli wild type and dcu mutants at pH 7.5 and pH 5.5

For revealing the role of some proteins in bacterial cell physiology during fermentation it is important to determine one of the key physiological parameters of bacteria; the specific growth rate (µ).

During glycerol fermentation E. coli wild type cells µ was higher 1.86 fold at pH 5.5 than at pH 7.5 and yielded µ of 0.58 h−1. But interestingly DCCD inhibited growth at pH 7.5 but not at pH 5.5 (Fig. 3). The data are in good conformity with DCCD inhibited ATPase activity and H+ efflux at pH 5.5 where no inhibition was determined in wild type cells (see Figs 1B and 2). At pH 7.5 in all mutants µ was the same, as in wild type except dcuD one. In dcuD mutant µ was higher 2.3 fold, compared to wild type, whereas at pH 5.5 no such differences were found (see Fig. 3). Janausch and Unden5 stated that deletion of dcuD gene did not affect cell growth in any conditions tested, but they used glycerol plus fumarate and the metabolism goes to fumarate respiration but not to glycerol fermentation and that is why we suggest in this work that DcuD protein is presumably active under glycerol fermentative conditions. In general, it must be mentioned that mainly Dcu carriers were investigated for aerobic growth on succinate1,2 but fermentative conditions were not deeply studied in anaerobic conditions due to it no evidence for gene expression and thus activity of DcuD protein have been detected. Note, at pH 5.5 in all mutants µ was 1.5 fold lower than in wild type suggesting the role of Dcu carriers in cell growth.

Figure 3
figure 3

Specific growth rate (µ) of E. coli wild type and dcu mutants at pH 7.5 and pH 5.5. A parallel experiment with 0.2 mM DCCD has been performed. For strains see Table 1, for others see legends to Fig. 1.

Full size image

As it was stated that deletions in Dcu system have some effects on H+ flux and ATPase activity and it was established before that FOF1-ATPase has some link or relationship with Hyd enzymes, H2 production in dcu mutants was detected. When cells were grown on glycerol and in the assays glycerol was added, no differences were found in mutants compared to wild type at both pHs, except dcuD mutant in which H2 production was higher at pH 5.5 (Fig. 4). Moreover, no H2 generation was detected when succinate was added in the assays (data not shown). Note, that H2 production was inhibited by DCCD in E coli wild type and in all of the dcu mutants at both pH values. This suggests that FOF1 interacts with H2 producing Hyd enzymes but the role of Dcu C4-dicarboxylate carriers in H2 generation is absent.

Figure 4
figure 4

H2 production in E. coli wild type cells at pH 7.5 and pH 5.5. Cells were harvested and assayed with same concentration of glycerol as used in growth medium. The DCCD (0.2 mM) was added into the assay medium when indicated. For others see Methods section and legends to Fig. 1.

Full size image

Methods

Bacteria, growth conditions, membrane vesicles

The E. coli strains used in this study are listed in Table 1. Bacteria were grown under anaerobic conditions at 37 °C for 24 h in highly buffered peptone medium (20 g L−1 peptone, 15 g L−1 K2HPO4, 1.08 g L−1 KH2PO4, 10 g L−1 NaCl) with glycerol (10 g L−1) at pH 7.5 and pH 5.5. To achieve anaerobic conditions were achieved in glass bottles with plastic press-caps were used; O2 was removed from the medium by autoclaving, after which the bottles were closed by press-caps and residual O2 was rapidly removed by the inoculum. as described elsewhere17,28,29,30. The growth medium pH was measured by a pH-meter with pH-electrode (HJ1131B, Hanna Instruments, Portugal) and adjusted using of 0.1 M HCl or 0.1 N NaOH.

Table 1 Characteristics of E. coli wild type and mutant strains used.
Full size table

The bacterial specific growth rate (μ), presented as lg2/ doubling time, was calculated, as described31.

Membrane vesicles isolated from bacteria, which were treated with lysozyme and ethylenediaminetetraacetic acid and prepared by the osmotic lysis of spheroplasts32, as described previously20,25,27,28.

Membrane vesicles ATPase assay

ATPase activity was determined by the amount of inorganic phosphate (Pi) liberated in the reaction of membrane vesicles with 5 mM ATP (pH 7.5 and 5.5)20,25,27,28 in the assay mixture (50 mM Tris–HCl buffer (pH 7.5 and 5.5) containing 1 mM MgSO4) at 37 °C thermostated chamber. Note, ATPase in right-side-out vesicles can be reached by ATP due to membrane peculiarities of cells grown under the above mentioned conditions20,28. In contrast ATPase in in-side-out vesicles was easily reachable in the preparations, as suggested33. The ATPase activity was expressed in nMol Pi (min µg protein)−1. Pi was determined spectrophotometrically (UV–VIS spectrophotometer, Cary 60, Agilent Technologies, USA), as described20,25,27. Membrane vesicles were incubated with 0.1 mM DCCD (ethanol solution) for 10 min prior assays; ethanol in the final concentration of 0.5% was used, as a blank; no effect on growth and ATPase activity was observed.

The DCCD-sensitive (inhibited) ATPase activity was calculated as a difference between activities in the absence and in the presence of the inhibitor (DCCD). In the assays KCl as a source for potassium was added in the concentration of 100 mM.

Redox potential determination and hydrogen production assay

Redox potential (Eh) in bacteria was determined using two different redox, titanium-silicate (Ti-Si) (EO-02, Gomel State Enterprise of Electrometric Equipment (GSEEE), Gomel, Belarus) and platinum (Pt) (EPB-1, GSEEE, or PT42BNC, Hanna Instruments, Portugal) glass electrodes9,23,34,35. The Ti-Si-electrode measures the overall Eh, whereas the Pt-electrode is sensitive to H2 under anoxic conditions23,34,35,36. H2 production rate (VH2) was calculated as the difference between the initial rates of decrease in Pt- and Ti-Si-electrodes readings and expressed in mV of Eh per min per mg of cell dry weight (mV. min−1 mg−1 CDW). This type of electrochemical determination of H2 is similar to the Clark-type electrode used by Fernandez36 and other researchers37. As a control experiment, cells without any addition of carbon sources were used where H2 production was absent.

The H2 production determination was done in the assay buffer solution (150 mM Tris-phosphate, at the indicated pH, including 0.4 mM MgSO4, 1 mM NaCl and 1 mM KCl) upon glycerol addition. Glycerol, as a carbon source, was added for assays at similar concentrations, as used for cell cultivation.

H2 generation was also verified chemically23,35,38 and with Durham test tubes13.

Measurement of H+ fluxes

H+ fluxes by whole cells were determined by registering the changes in H+ activities in the medium using selective pH electrodes (Hanna Instruments)28,39. The electrode readings were calibrated by titration of the medium with 0.01 M HCl. Ion fluxes were expressed in mmol/min per 109 cells in 1 unit of volume.

Protein determination, chemicals and data processing

Protein concentration was measured by the method of Lowry40 using bovine serum albumin (BSA), as a standard.

In experiments agar, peptone, glycerol, Tris (Carl Roth GmbH, Germany), ATP (Tris salt), BSA, DCCD, lysozyme (Sigma, USA) and other chemicals of analytical grade were used.

Data obtained from 3 independent assays are averaged, and standard deviations of values are calculated and they do not exceed 3% if not shown. Student criteria (p-value) is applied to validate the statistical differences in average data between various series of experiments, as described previously29,30; the difference is valid when p < 0.05 or less if not given.