INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

As one of the most prominent low-molecular-weight products of microbial metabolism, acetate is well known for its cytotoxicity that includes retardation of growth and product formation at concentrations below 5 g/liter (2, 4, 18). These toxic effects are related to the weak lipophilic nature of the undissociated acid that enables the molecule to cross the cytoplasmic membrane. This diffusion is generally thought to dissipate ion gradients, increase the internal acetate concentration, and/or disrupt membrane processes, thereby poisoning the cell (1, 4, 5, 26).

While most microorganisms are sensitive to higher concentrations of acetate, a few are known to be relatively resistant. A prominent example of such resistant organisms are the so-called acetic acid bacteria (AAB), the genera Acetobacter and Gluconobacter. Species of the former have been used for millennia in the production of acetic acid as vinegar (6, 28). In industrial settings, Acetobacter aceti can grow at acetate concentrations of up to about 60 g/liter (23) and accumulate final concentrations exceeding 140 g/liter in semicontinuous processes (6). Survival under these hostile conditions is apparently sufficient, so that such cultures can serve as inocula for subsequent batch cultures. Oxygen deficiency of Acetobacter leads within seconds to a drop in energy charge (13) and, likely as a consequence, a rapid loss of viability (6, 19). This illustrates the importance of cellular energetics for acetate resistance. The molecular mechanism of resistance in AAB, however, remains essentially unknown.

The previously reported acetate resistance genes, aarABC, of a thermophilic A. aceti strain were important for resistance on solid media (7). The identified functions of the AarA and AarC gene products in citrate synthesis (7) and acetate uptake (8), respectively, show that these proteins confer resistance by acetate assimilation via a local reduction of acetate concentrations on solid media. In liquid media, however, assimilation of acetate is not a pertinent resistance mechanism, in particular when acetate is continuously produced by the organism. Thus, it appears that different resistance mechanisms operate in acetate-containing liquid media, such as during vinegar production.

Presumably as a consequence of their genetic variability, the AAB rapidly lose acetate resistance when removed from its presence (6). For this reason, industrial AAB are maintained at high acetate concentrations, and vinegar processes are operated for years without interruption so that the industrial strain remains adapted to high acetate concentrations. This adaptation appears to be a prerequisite for high acetate tolerance, because wild-type AAB do not exhibit pronounced acetate tolerance (16). In fact, when cultivated on appropriate carbon sources, Escherichia coli exhibits a similar tolerance of acetate (17).

Previous proteome analysis of AAB revealed eight so-called acetate stress proteins (Asps) that were induced specifically by challenging unadapted A. aceti and Gluconobacter suboxydans cultures with 10 g of acetate per liter (16).

Here we investigate the changes in global protein expression levels during long-term adaptation of A. aceti to high acetate concentrations, as a first step to the characterization of the molecular mechanisms underlying acetate resistance. For this purpose, A. aceti wild-type was exposed to stepwise increased acetate concentrations in either serial batch or continuous cultures and subsequently analyzed by two-dimensional protein electrophoresis (2DE). Detection and N-terminal sequences of proteins that are induced exclusively in response to acetate adaptation, as opposed to general stress responses, are described.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Strain and medium. A. aceti DSMZ 2002 was obtained from the Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH (DSMZ) and routinely grown on complex YPD medium containing 30 g of glucose, 2 g of Bacto peptone, and 5 g of yeast extract (pH 6.5) per liter at 30°C. Sterile glucose was added to the separately sterilized complex components from a 30% (wt/vol) stock solution. For acetate supplementation, YPD medium containing the desired concentration of potassium acetate was titrated to pH 6.5 by addition of YPD medium containing the same concentration of acetic acid and sterilized by filtration. All bioreactor media were supplemented with 0.05% (wt/vol) polypropyleneglycol 2000 to prevent foaming.

Culture conditions. Stress experiments were performed in 500-ml baffled flasks, filled with 100 ml of medium, at 30°C and 200 rpm. At an optical density at 600 nm (OD₆₀₀) of about 1, exponentially growing cells were diluted 1:1 in fresh medium containing either 0.5 M NaCl or 10 mM H₂O₂. After incubation at 30°C (or 42°C for heat stress) for 3 h, cells were harvested by centrifugation and prepared for 2DE.

1

Analytical procedures. Glucose and total protein concentrations were determined enzymatically (Synchron CX5CE; Beckman) with kits supplied by the manufacturer. Acetate concentrations were determined by high-pressure liquid chromatography (HPLC) (Series 200; Perkin Elmer) using an ion exclusion HPLC column (Supelcogel H; Supelco) with 0.2 N H₃PO₄ as the mobile phase. Oxygen and carbon dioxide concentrations in the feed and effluent gas of the bioreactor were determined with a mass spectrometer (Prima 600; Fissons Instruments).

Protein sample preparation for 2DE. Cells were harvested by centrifugation for 15 min at 7,000 × g and 4°C. The pellets were resuspended in deionized water, aliquoted in microcentrifuge tubes, and centrifuged again for 10 min at 10,000 × g and 4°C, and the pellets were stored at 80°C. Protein samples for 2DE were prepared according to the protocol of Tonella et al. (29). First, frozen cells were resuspended in 2 ml of low-salt washing buffer (3 mM KCl, 1.5 mM KH₂PO₄, 68 mM NaCl, and 9 mM NaH₂PO₄). Typically, the equivalent of 25 to 50 ml of culture at an OD₆₀₀ of 1 was used. The cells were washed four times with washing buffer and resuspended in storage buffer (10 mM Tris-HCl [pH 8], 1.5 mM MgCl₂, 10 mM KCl, 0.5 mM dithioerythritol, 0.5 mM Pefabloc SC, and 0.1% [wt/vol] sodium dodecyl sulfate [SDS]) at an estimated pellet-to-buffer volume ratio of 4:1. After cell disruption by vortexing for 10 min at the highest speed (Vortex Genie 2), cells were used directly for 2DE or stored at 80°C. Prior to 2DE analysis, sample quality and protein concentration were assayed by analytical one-dimensional (1D) SDS-polyacrylamide gel electrophoresis (SDS-PAGE) on a 12% minigel (14).

Analytical 2DE. Two-dimensional electrophoresis was adapted from the method of O`Farrell (21). Prior to isoelectric focusing, immobilized pH gradient (IPG) strips were rehydrated along with the protein samples. For this purpose, 2 to 3 µl of sample was mixed with rehydration buffer (5 M urea, 0.333% BioLyte pH 3 to 10, 0.167% BioLyte pH 5 to 7, 4% CHAPS [3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate], 2 M thiourea, and 65 mM dithiothreitol [DTT]) and equally dispensed onto 18-cm IPG strips (pH 3 to 10 NL; Pharmacia). The exact sample volume was chosen so that equal amounts of protein were loaded on each gel. The strips were covered with cover fluid (Plusone Dry Strip cover fluid; Pharmacia) and rehydrated overnight in a dry strip reswelling tray (Pharmacia).

2DE image analysis. Comparisons of spot intensities were performed with digitized images of the silver-stained gels. A first detection of major differences in spot intensities was achieved visually by fast flipping of negative gel images. Detailed detection of less obvious intensity changes was done with computer-aided image analysis. For this purpose, reproducible spots were detected and quantified on different gel images with the Melanie II software (Geneva Bioinformatics S.A). Estimation of molecular weight and isoelectric point (pI) was based on 2DE analysis of protein samples that were mixed with a 2DE protein standard (Bio-Rad).

Preparative 2DE and N-terminal sequence determination. Preparative 2DE was performed as described for analytical 2DE but with a sample volume of 20 to 30 µl. Proteins were transferred overnight from gels to Immobilon-P polyvinylidene difluoride (PVDF) membrane (Millipore) in a Trans-Blot cell (Bio-Rad) at 100 mA with an initial current of 150 mA for 3 h using transblot buffer (192 mM glycine, 20% methanol, and 25 mM Tris base). The membrane was stained with Coomassie blue, washed for 1 to 2 days in distilled H₂O, and air dried. Selected protein spots were excised and subjected to N-terminal sequencing on an automated peptide sequencer (Sequenator G1000A; Hewlett Packard).

Analytical one-dimensional gel electrophoresis of membrane proteins. Membrane proteins of A. aceti were isolated according to a modified method of Poole (24). The equivalent of 10 to 30 ml of culture (OD₆₀₀ of 1) of frozen cells from continuous cultures at different acetate concentrations was resuspended in 1 ml of buffer (30 mM Tris-HCl [pH 8.0], 20% [wt/vol] sucrose, and 50 µg/ml chloramphenicol). Potassium-EDTA (pH 7.0) and lysozyme were added to final concentrations of 10 mM and 2 mg/ml, respectively, and the cells were incubated for 1 h at 37°C under gentle agitation. Resulting spheroplasts were sedimented for 30 min at 13,000 × g and 4°C. The remaining pellet was dispersed in 1 ml of phosphate buffer (10 mM potassium phosphate [pH 6.6], 2 mM MgSO₄, and 10 µg/ml each of DNase and RNase), using a syringe and a needle (1.2 by 40 mm). The suspension was incubated for 30 min at 37°C under gentle agitation, followed by centrifugation at 800 × g and 4°C for 1 h. The supernatant was centrifuged at 100,000 × g and 4°C for 1 h, and the resulting membrane pellet was resuspended in 10 to 20 µl of 50 mM potassium phosphate buffer (pH 6.6). Equal amounts of membrane proteins, preestimated on a test gel, were then resolved by analytical 1D SDS-PAGE on a 10% minigel and visualized by silver staining.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Growth characteristics at increasing acetate concentrations. To gain insight into the protective response of AAB to the toxic effects of their major metabolic product acetate, the unadapted wild-type A. aceti strain DSMZ 2002 was cultivated at increasing acetate concentrations. In serial batch cultivation, the first batch without acetate supplementation was followed by three batches at 10 g/liter and three batches at 20 g/liter total acetate concentration. The general response to acetate challenge was a slower growth rate, but adaptation to the acetate challenge manifested itself in an approximately 60% increased growth rate in the second culture at 10 g of acetate per liter, compared to the first culture at 10 g of acetate per liter (Table 1). Upon an increase to 20 g of acetate per liter supplementation, the growth rate dropped to about 0.1 h¹ and no further adaptation was observed. In a similar experiment with E. coli at acetate concentrations between 5 and 8 g/liter, we could not detect any adaptation (data not shown).

Protein expression profile of acetate-adapted A. aceti To identify proteins that are upregulated during adaptation of A. aceti to high acetate concentrations, we examined exponentially growing cells from serial batch (Table 1) and continuous cultures (Table 2) by 2DE analysis. Because we intended to compare proteome patterns in different physiological steady states and not short-term stress responses, we investigated expression levels rather than synthesis rates. Generally, 2DE analysis allowed us to resolve up to 1,500 distinct protein spots within a size range of 15 to 100 kDa and a pH range of 4 to 9 (Fig. 1). The 2DE patterns of independent samples harvested from the same culture were highly reproducible. Despite important differences in expression patterns under batch and continuous-culture conditions, the overall 2DE patterns were very similar and thus allowed us to quickly identify the same spots on different gels. For the following detailed analysis, all 2DE analyses were divided into three groups: serial batch, continuous culture, and general stress.

View larger version (112K): [in this window] [in a new window]   FIG. 1.   Silver-stained 2DE images of exponentially growing A. aceti in serial batch cultures at 0 (A) and 20 g of acetate (B) per liter and in continuous culture at 0 (C) and 30 g of acetate (D) per liter. Diamonds indicate acetate adaptation proteins that are induced in acetate-containing serial batch and continuous cultures. Circles indicate microsequenced proteins that are induced by acetate in either serial batch or continuous culture. Triangles highlight the previously described Asps (16).

View larger version (74K): [in this window] [in a new window]   FIG. 2.   Relative expression levels of 19 identified Aaps from 2DE analysis of serial batch and continuous cultures at increasing acetate concentrations. Expression levels are normalized to the expression levels on reference gels without acetate supplementation, and seven experimental conditions are shown for each Aap in one box. These conditions are defined in the inset at the lower right. Hatched bars indicate visually estimated expression level. The three insets above each bar plot box show spot images of the corresponding Aap from 2DE of serial batches at 0, 10, and 20 g of acetate per liter, from left to right, respectively. Error bars indicate deviations of expression levels from two independent gels.

N-terminal sequencing of Aaps. To obtain first hints on protein function, 13 Aaps were isolated from preparative 2DE gels and subjected to N-terminal sequence analysis. Furthermore, we isolated seven protein spots that were significantly upregulated in continuous and repeated batch cultivations but not by any other stress condition. To verify the discrimination between general stress proteins and Aaps, five general stress-responsive proteins were also isolated and sequenced.

Analysis of expression of membrane proteins. Because the 2DE analysis employed is not ideally suited for resolving membrane proteins, we investigated the membrane protein composition by 1D SDS-PAGE. Membrane protein-enriched fractions of A. aceti were analyzed at increasing acetate concentrations in continuous culture (Fig. 3). These membrane protein-enriched fractions showed a pattern on 1D SDS-PAGE that was distinct from that of the total protein fraction, but abundant cytoplasmic proteins are probably also present in this fraction (data not shown). Most of the about 40 resolved proteins were within a size range of 30 to 100 kDa. The most abundant proteins that constitute the prominent bands of the membrane fractions were found not to be affected by increased acetate cocentrations.

View larger version (117K): [in this window] [in a new window]   FIG. 3.   Silver-stained SDS-polyacrylamide gel electrophoresis of membrane proteins. Lanes (from left to right): M, molecular size markers; 1 to 5, membrane fractions of cells from continuous culture at 0, 10, 20, 25, and 30 g of acetate per liter, respectively. Proteins that were at least twofold induced by acetate are designated M1 to M11.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Adaptation of A. aceti to high acetate concentrations clearly elicits a complex response at the level of protein expression, with a total of about 40 proteins induced. Some of these induced proteins presumably belong to the so-called general stress protein class because they were also induced by other stress conditions. This is a well-known phenomenon that has been described for many bacterial stress responses (12), including challenge with the weak organic acid benzoate (15). While about 35 proteins appeared to be specifically induced during acetate adaptation, only 19 exhibited the same behavior in batch and continuous culture. Only these 19 proteins that were not induced during other stress conditions were considered Aaps.

Few of the identified acetate adaptation proteins exhibited maximum expression levels at 10 g of acetate per liter, while most exhibited higher expression levels at higher acetate concentrations. Surprisingly, all Aaps were also detected in the absence of acetate, indicating that these proteins are also relevant during A. aceti's normal life. We cannot exclude that acetate challenge leads to coexisting subpopulations in our cultures (27). This, however, does not invalidate our conclusion on Aaps, since the relative concentration of these proteins increased in the whole culture, either because it was upregulated in all cells or because the proportion of a better adapted subpopulation increased.

Unlike the more common use of 2DE to analyze rapid cellular stress responses (9, 12, 15, 16), we investigated long-term adaptation to the weak acid acetate. This constitutes a very different condition, which is also evidenced by the absence of any significant induction of eight previously identified acetate stress proteins (16). These proteins are induced by a 10 g of acetate per liter challenge of unadapted A. aceti, but only one of them, Asp C, was also induced during acetate adaptation in continuous culture (Fig 1). While five other Asps (A, B, D, E, and H) were identified on our gels, the remaining two, Asps F and G, could not be identified, either for technical reasons or because of their transient expression profile.

Generally, N-terminal sequences of organisms without a sequenced genome do not allow us to draw firm conclusions on sequence homologies and thus on protein function. However, some of the identified homologies of our sequenced proteins indicate a role of membrane transport processes in acetate resistance. This view is supported by the observed induction of several bands in 1D gels of membrane proteins. The strongest homology to a membrane transport protein was found for Aap J, which is homologous to the potassium channel -subunit. The associated -subunit is responsible for K⁺ ion conduction and voltage-dependent gating (11), and thus might be induced by the presence of potassium, the acetate counterion in our study. Preliminary DNA sequence data show that the aapJ gene of A. aceti indeed encodes a protein that is homologous to the potassium channel -subunit (P. Steiner, U. F. Püntener, and U. Sauer, unpublished data). Less significant homology was found for Aap F to the putative formate transporter of E. coli and of protein 20 to an NADH:ubiquinone oxidoreductase. The latter homology also supports the notion that cellular energetics are important for acetate resistance, which was shown recently for Saccharomyces cerevisiae (22).

Overall, our data indicate that membrane functions, possibly including acetate transport and respiration, are involved in the acetate resistance mechanism of A. aceti. The relatively large number of identified Aaps, however, indicates that other resistance mechanisms may also be at work. Because the previously described acetate resistance proteins AarABC are involved in acetate assimilation (7, 8), it is unlikely that they contribute to resistance under the presently investigated conditions of growth in liquid culture. Thus, it is not surprising that neither of these proteins was significantly induced during acetate adaptation.