INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Various bacterial strains reduce azo dyes under anaerobic conditions. The most generally accepted hypothesis for this phenomenon is that many bacterial strains possess rather unspecific cytoplasmic enzymes, which act as "azo reductases" and under anaerobic conditions transfer electrons via soluble flavins to the azo dyes (35). We have recently suggested a different mechanism for the unspecific anaerobic reduction of azo dyes by Sphingomonas sp. strain BN6 and other bacteria. In this system, the reduction of the azo dyes is catalyzed extracellularly by the action of mediator compounds, which are either formed during the metabolism of certain substrates by the bacteria themselves or which may be added externally (e.g., compounds such as anthraquinonesulfonates). These mediators enable the transfer of redox equivalents from the cell membrane of the bacteria to azo dyes. In addition, it was demonstrated that Sphingomonas sp. strain BN6 also possesses cytoplasmic azo reductase activities, which could be assayed in vitro with cell extracts under anaerobic conditions (14, 15, 17). It was shown that the reduction rate of whole cells of Sphingomonas sp. strain BN6 for amaranth after growth with glucose was 0.3 µmol min¹ g of soluble cell protein¹ and could be increased by the addition of anthraquinone-2-sulfonate to 3 µmol min¹ g of soluble cell protein¹. In contrast, in vitro with cell extracts or solubilized membrane preparations, specific activities of 20 to 40 µmol min¹ g of soluble cell protein¹ were determined (15, 17). A possible explanation for the significantly lower reaction rates of the whole-cell preparations compared to those in the in vitro systems was that in vivo the cytoplasmic azo reductase activity does not function, because the highly polar sulfonated azo dyes are not able to penetrate through the cell membranes. In the present study, we attempted to clarify the question of whether in vivo the intracellular cytoplasmic azo reductase activity is involved at all in the metabolism of the azo compounds or whether this enzymatic activity is only active under in vitro conditions. Therefore, we increased the intracellular cytoplasmic azo reductase activity by using a genetic engineering approach and compared the resulting recombinant organism in vivo with the parent strain.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains. The isolation and characterization of Sphingomonas sp. strain BN6 have been described before (20). The strain has been deposited at the Deutsche Sammlung von Mikroorganismen und Zellkulturen (DSMZ) in Braunschweig, Germany, as DSM 6383. Escherichia coli K38 and JM109(DE3) were used for overexpression of the flavin reductase, and E. coli S17-1 (32) was used for the conjugative transfer of plasmids to strain BN6.

Culture conditions. Sphingomonas sp. strain BN6 was routinely grown in a mineral medium according to the method of Dorn et al. (7) with glucose (15 mM) and naphthalene-2-sulfonate (0.5 mM). For Sphingomonas sp. strain BN6(pRJR34), this medium was supplemented with 10 µg of tetracycline per ml. E. coli strains were routinely cultured in nutrient broth (NB) or Luria-Bertani (LB) medium. E. coli JM109(pRJR34) was grown in LB medium plus 12.5 µg of tetracycline per ml and isopropyl--D-thiogalactopyranoside (IPTG; 1 mM) to induce the flavin reductase.

Plasmids and DNA manipulation techniques. Plasmid DNA was isolated from E. coli by the method of Lee and Rasheed (18). Digestion of DNA with restriction endonucleases (Gibco BRL, Boehringer), electrophoresis, and ligation with T4 DNA ligase (Gibco BRL) were performed according to standard procedures (30). Transformation of E. coli was done by the method of Chung et al. (4).

Detection of the fre gene from E. coli in Sphingomonas sp. strain BN6(pRJR34) by PCR. For the PCR, oligonucleotide primers were custom synthesized according to the sequence of the gene described by Spyrou et al. (33). From the nucleotide sequence of the fre gene, bases 16 to 38 (forward primer) and 627 to 646 (reverse primer) were used as primers. PCR mixtures (50 µl) for the amplification of DNA contained 50 pmol of each primer, 10 to 20 ng of plasmid template DNA (isolated with Pharmacia Flexiprep kits), 0.1 mM (each) deoxynucleotide triphosphate, and 1 U of Taq DNA polymerase, polymerase reaction buffer, and MgCl₂ according to the manufacturer (GibcoBRL). The PCR was performed with a touchdown thermocycle program under the following conditions: initial denaturation (95°C, 3.75 min) before addition of the polymerase, 10 cycles with decreasing annealing temperature (60 to 55°C, 30 s), polymerization (72°C, 1 min), and denaturation (95°C, 45 s). Fifteen more cycles with 55°C as the annealing temperature and 72°C as the polymerization temperature were performed before the reaction was stopped, and the reaction products were analyzed by agarose gel electrophoresis. Thus, a fragment of the expected size (630 bp) was detected after amplification of the DNA from E. coli JM109(DE3)(pEE1001) and Sphingomonas sp. strain BN6(pRJR34), but not with DNA from the wild-type strain, Sphingomonas sp. strain BN6.

Anaerobic reduction of azo dyes by whole-cell preparations. Sphingomonas sp. strain BN6 was grown aerobically in a mineral medium to the late exponential growth phase (optical density at 546 nm of about 2 to 4) and then transferred to rubber-stoppered serum bottles (100 ml). Oxygen was removed from the medium by at least 15 2-min cycles of evacuation and flushing with nitrogen gas. Glucose (10 mM) and amaranth (1.3 mM) were added anaerobically, and samples were taken at different time intervals. The decrease in the dye concentration was determined by high-performance liquid chromatography as described previously (14).

Preparation of cell extracts. Cell suspensions in 50 mM Tris-HCl (pH 7.5) were disrupted by using a French press (Aminco, Silver Spring, Md.) at 80 MPa. Cell debris was removed by centrifugation at 100,000 × g for 30 min at 4°C. The protein concentration was determined by the method of Bradford (2), with bovine serum albumin as the standard.

Enzyme assays. One unit of enzyme activity was defined as the amount of enzyme that converts 1 µmol of substrate per min.

(i) Flavin reductase. The NAD(P)H:flavin oxidoreductase was measured by a modification of the method given by Fontecave et al. (9). In this aerobic assay, the flavin reductase catalyzes the reduction of riboflavin, and the reduced riboflavin is immediately reoxidized by oxygen. Cell extract was added to a solution (final volume, 1 ml) containing 10 µmol of Tris-HCl (pH 7.5), 0.25 µmol of NADPH, and 0.003 µmol of riboflavin. The decrease in A₃₄₀ was measured spectrophotometrically. Reaction rates were calculated by using a molar extinction coefficient of 6.3 mM¹ cm¹.

(ii) Azo reductase. Azo reductase activity was routinely measured by a modification of the spectrophotometric assay described previously by Haug et al. (14). The anaerobic reaction mixture contained (in 1.6 ml) 70 µmol of Tris-HCl buffer (pH 7.5), 1 µmol of NADH, 0.32 µmol of flavin adenine dinucleotide (FAD), and 0.1 µmol of amaranth. The stock solutions were made anaerobically by repeated gassing with N₂. The reaction was performed in gastight cuvettes and started by the addition of cell extracts. The decrease in the concentration of amaranth was determined spectrophotometrically at 520 nm, and reaction rates were calculated by using a molar extinction coefficient of 22.6 mM¹ cm¹. In certain experiments, azo dyes different from amaranth were used. The relevant wavelengths and extinction coefficients for these dyes are summarized in Table 1.

Chemicals. All chemicals were obtained from Merck, Fluka, Sigma, or Aldrich, except for Mordant yellow 3, which was kindly provided by Bayer AG.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

The flavin reductase from E. coli as azo reductase. It was previously suggested that, under anaerobic conditions, unspecific cytoplasmic azo reductases act via the intermediate formation of free reduced flavins (35). This hypothesis has been substantiated by the observation that photochemically generated reduced FAD (FADH₂) reduced the sulfonated azo compound amaranth (10). In this system, the only function of the "cytoplasmatic azoreductases" would be the supply of reduced flavins for the subsequent purely chemical reduction of the azo compounds by the reduced flavins. It was therefore reasoned that any flavin reductase [NAD(P)H:flavin oxidoreductase] could also act as an azo reductase. Recently, a flavin reductase gene (fre) was cloned. The gene product, which is naturally part of the ribonucleotide reductase complex of E. coli, was produced in large amounts with the expression plasmid pEE1001. In this construct, the flavin reductase gene is transcribed from the phage T7 promoter (26, 33). This system was applied to assay the flavin reductase for its azo reductase activity. E. coli JM109(DE3)(pEE1001), E. coli K38(pEE1001), and, as a negative control, E. coli JM109 were grown in LB medium, cell extracts were prepared, and the flavin reductase and the azo reductase activities were determined under anaerobic conditions. Thus, it was confirmed that the E. coli strains with the recombinant plasmid really produced the flavin reductase with very high specific activities, and it was demonstrated that the flavin reductase could act under anaerobic conditions as azo reductase (Table 2). Surprisingly, the flavin reductase activity was constitutively expressed in E. coli JM109(DE3)(pEE1001), and the activity could not be increased by the addition of IPTG (up to 2 mM). The proposed function of FAD as a mediator compound transferring redox equivalents from the flavin reductase to the azo dye was shown in a control experiment, in which the cell extract was incubated with NADPH and amaranth with or without added FAD. Thus, it was found that, in the absence of externally added FAD, less than 5% of the azo reductase activity could be detected.

Comparison of the cytoplasmic azo reductase activities in E. coli K38(pEE1001) and Sphingomonas sp. strain BN6 with different industrially relevant azo dyes. It has been previously reported that azo reductase activity was present in cell extracts of Sphingomonas sp. strain BN6 (14, 15, 17). Therefore, the azo reductase activities of cell extracts from E. coli K38(pEE1001) and Sphingomonas sp. strain BN6 were compared and tested for their ability to reduce different industrially relevant azo dyes (Table 3). Thus, it was found that the azo reductase activity in the cell extract of E. coli K38(pEE1001) for many azo dyes was more than 100-fold higher than the activities found in cell extracts of strain BN6. The cell extracts from both strains had high relative activities for those dyes which carried in the ortho position to the azo group a hydroxyl group and in general had somehow lower activities for dyes with a hydroxy group in the para position to the azo bond. The hydrogen of the hydroxy group in the ortho position can form an intramolecular hydrogen bridge with the nitrogen atom of the azo group. This reduces the electron density at the azo bond and thus facilitates a reductive cleavage. Furthermore, it was observed that an increasing number of sulfonic acid substituents also resulted in an increased reduction rate. This may be caused by the electron withdrawal effect of the sulfonic acid substituents, which also reduces the electron density at the azo bond.

Transfer of the flavin reductase gene to Sphingomonas sp. strain BN6. The plasmid pEE1001, which carries the fre gene, is a pBR322 derivative and has only a narrow host range. It has been previously reported that IncP1 plasmids can be conjugatively transferred to Sphingomonas sp. strain BN6 (28). Therefore, plasmid pRJR34 was constructed, which contains the fre gene cloned into the broad-host-range plasmid pRK415 (16). After the preparation of cell extracts from JM109(pRJR34), flavin reductase activities of 1.7 U/mg of protein were found. Thus, the flavin reductase was expressed in this system too. Finally, E. coli S17-1 (32) was transformed with plasmid pRJR34, and the resulting transformant was used to conjugatively transfer pRJR34 to Sphingomonas sp. strain BN6. After selection on a mineral medium with glucose (10 mM) plus tetracycline (5 µg/ml), a transfer frequency of plasmid (pRJR34) from E. coli S17-1 to strain BN6 of almost 100% was found.

Expression of the flavin reductase and the azo reductase activity in Sphingomonas sp. strain BN6. Sphingomonas sp. strain BN6(pRJR34) was grown in liquid culture with NB and tetracycline (10 µg/ml), and cell extracts were prepared. The flavin reductase activity in the recombinant strain was 0.7 U/mg of protein, compared to a flavin reductase activity in wild-type strain BN6 of 0.008 U/mg of protein. The presence of the fre gene from E. coli in Sphingomonas sp. strain BN6(pRJR34) was also shown by PCR (see Materials and Methods). These results demonstrated that the recombinant organism indeed expressed the flavin reductase gene originating from E. coli. Therefore, whether these cell extracts also showed an increased azo reductase activity under anaerobic conditions was also tested. Thus, it could be demonstrated that the azo reductase activity in the recombinant strain was almost 30-fold higher than that in the wild-type strain (Table 2).

1

1

View larger version (14K): [in this window] [in a new window]   FIG. 1.   Reduction of amaranth (A) or Mordant yellow 3 (B) under anaerobic conditions by whole cells of Sphingomonas sp. strain BN6 () or BN6(pRJR34) (). The cells were grown aerobically in a mineral medium with glucose (10 mM) and naphthalene-2-sulfonate (0.5 mM) [plus 10 µg of tetracycline per ml for BN6(pRJR34)]. When the cells reached the late-exponential growth phase, 25 ml of the cultures was transferred into serum bottles (100-ml volume) with a nitrogen atmosphere, and glucose (10 mM) and Mordant yellow 3 (1 mM) or amaranth (1.3 mM) were added. At the indicated intervals, aliquots were taken, cells were removed by centrifugation, and the concentrations of the azo dyes were determined by high-performance liquid chromatography.

Aerobic function of the flavin reductase as azo reductase. All of the experiments described above were performed under anaerobic conditions, because it is well known that reduced flavins react rapidly with molecular oxygen. Since the recombinant strains expressed the flavin reductase with high specific activities, we assumed that there could be an azo reductase activity of the flavin reductase as well in the presence of molecular oxygen. Therefore, the azo reductase assays with cell extracts were compared under anaerobic and aerobic conditions. The reactions were analyzed spectrophotometrically at  = 340 nm to measure the oxidation of NADH and at  = 520 nm to assay the reduction of amaranth (Fig. 2). Thus, it was found that, also under aerobic conditions, some decrease in the concentration of amaranth was observed, although the reaction rates were significant lower than those of the anaerobic control. In contrast, almost identical oxidation rates for NADH were observed under aerobic and anaerobic conditions (Fig. 2). This observation may explain the results of some sporadic reports about the presence of unspecific aerobic azo reductases in bacteria such as E. coli (11, 12).

View larger version (14K): [in this window] [in a new window]   FIG. 2.   Reduction of amaranth by cell extracts from E. coli K38(pEE1001) under aerobic (, ) and anaerobic conditions (, ). The reaction mixtures contained (per milliliter) 50 µmol of NaK phosphate buffer (pH 7.7), 0.4 µmol of NADH, 0.1 µmol of FAD, and 0.05 µmol of amaranth. The reactions were started by the addition of 20 µl of a cell extract from E. coli K38(pEE1001) (88 µg of protein), and the decrease in absorbance was simultaneously determined spectrophotometrically at  = 340 nm (, ) and  = 520 nm (, ). For the reaction under anaerobic conditions, all solutions were made anaerobically by repeated gassing with N2, and the reaction was performed in gastight cuvettes.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

There are numerous reports which describe the reduction of azo compounds by bacteria under anaerobic conditions. The main interest in this field has focused on bacteria from the intestine which are involved in the metabolism of azo dyes ingested as food additives, and there are some studies which refer to bacteria isolated from soil or sewage treatment systems (5, 6, 8, 27, 29, 31). The earlier studies were mainly performed with facultatively anaerobic bacteria (e.g., Proteus vulgaris, Streptococcus faecalis, or Bacillus sp.). It had been repeatedly suggested that, in these strains, cytosolic flavin-dependent reductases were responsible for this unspecific reaction and that the low permeability of the cell membranes for the highly polar sulfonated azo compounds would cause the significantly lower reduction rates observed with whole cells compared to those in cell-free systems (19, 27, 36). Because none of these cytosolic azo reductases has ever been purified, until now it has been speculative whether bacterial flavin reductases were responsible for the azo reductase activity observed with bacterial cell extracts. The present study proved that flavin reductases are indeed able to act as azo reductases and that cell extracts containing the cloned flavin reductase decolorized a broad range of industrially relevant azo dyes.

Our results suggested that the reduction of sulfonated azo dyes by reduced flavins formed by cytosolic flavin-dependent azo reductases is mainly observed in vitro and in vivo is of insignificant importance. This became evident with the recombinant strain BN6(pRJR34), which in vitro showed significantly increased azo reductase activities compared to those of the wild type, although in vivo only a very small increase in azo reductase activity was found. Theoretically, this could also have been caused by a lack of the required cofactors (NADH and/or free flavins) in the resting cells. This seems to be improbable, because the anaerobically incubated cells should possess a rather high concentration of NADH. There is only limited information available about the free flavin content of bacterial cells. However, it has been observed that recombinant FMN-dependent luciferases are fully functional in different bacterial backgrounds without the concurrent transfer of recombinant FMN reductases (34). This has also been shown recently for strain BN6 (J. Klein, personal communication). This suggests that reduced free flavins are present in various bacteria and also in strain BN6 in quantities which allow the support of enzymatic reactions which require reduced flavins.

A further indication that the flavin-dependent azo-reductases are almost completely laboratory artifacts was shown by the effects of externally added flavins on the rates of azo reduction under anaerobic conditions. Thus, it was found that the addition of FAD to a resting cell suspension of strain BN6 resulted in no significant increase in the rate of azo dye reduction. In contrast, the addition of the same concentration of anthraquinone-2,6-disulfonate resulted in an almost 10-fold increase in the dye reduction rate (17). Similar observations were also made with Sphingomonas sp. strain BN6(pRJR34) and E. coli JM109(DE3)(pEE1001) (unpublished results). This suggested that the bacterial membranes not only are efficient barriers for the uptake of sulfonated azo dyes, but also are hardly permeable for flavin-containing cofactors and therefore restrict the transfer of reduction equivalents by flavins from the cytosol of intact cells to sulfonated azo dyes in the culture medium. Thus, in living cells with intact cell membranes, other enzyme systems and/or other redox mediators are presumably responsible for the reduction of azo dyes. In bacteria which possess electron transport systems in their membranes, such as the purely aerobic or facultatively anaerobic bacteria studied in the present report, the transfer of the redox equivalents from the respiratory chain of the cell membranes to appropriate redox mediators could take place directly. If intracellular reductases should be involved in this process, it may be assumed that mediators different from flavin cofactors must be involved. A prerequisite for these mediators would be a higher ability to pass the bacterial membranes than that of flavins.

A different concept for the reduction of sulfonated azo compounds which also does not require a membrane transport of the dyes has been suggested for bacterial strains from the strictly anaerobic microflora of the intestine. Rafii and coworkers isolated different bacteria from the human intestine (e.g., Eubacterium sp., Clostridium sp., Butyrivibrio sp., or Bacteroides sp.), which decolorized sulfonated azo dyes during growth on solid or liquid complex media. It was shown that at least part of the azo reductase activities were extracellular, because the culture supernatants were able to decolorize the dyes under anaerobic conditions (24, 25). This extracellular azo reductase activity was due to the activity of specific proteins, because after nondenaturing polyacrylamide gel electrophoresis and activity staining under anaerobic conditions, usually only one distinct protein band in the gels was able to decolorize the dyes (21). In most isolates, flavin compounds were required for azo reductase activity, but in some clostridia, the azo reductase activity was independent of externally added flavins (24). Recently, a DNA fragment from Clostridium perfringens was cloned in phage gt11, which increased the azo reductase activity of lytic and lysogenic cultures of E. coli infected with the recombinant phage (23). The azo reductase from C. perfringens was detected by immunoelectron microscopy throughout the cytoplasm and in the vicinity of the cells (22). This suggested that the protein is not a typical extracellular enzyme, but that it is presumably released only from lysed cells. The azo reductase activity from C. perfringens is clearly different from the flavin reductase investigated in the present study, because the enzyme activity was described as being independent of that from added flavins, and furthermore, the enzyme was rapidly and irreversibly inactivated by oxygen (24). It is still unclear in this system how the supposed extracellular azoreductases should gain the NADH necessary for the reduction of the azo dyes in their extracellular environment, and there may be some effects from the complex growth media of the bacteria. Furthermore, there are some other publications about the reduction of azo compounds by whole-cell preparations of strictly anaerobic bacteria which suggested the involvement of redox mediators (3, 4). Therefore, it can't be excluded at present that extracellular mediator compounds also participate in the reduction of azo compounds by strictly anaerobic bacteria.

The third possibility for the extracellular reduction of azo dyes caused by microorganisms is the action of reduced inorganic compounds (Fe²⁺, H₂S), which are formed as end products of certain strictly anaerobic bacteria. Also, in these cases, mediator compounds may have an important function (13, 37).

It may be concluded that, depending on the environmental conditions (or the reaction conditions of a biotechnological process), different reactions catalyzed by bacteria can accomplish the reduction of azo compounds under anaerobic conditions, but that intracellular reactions have only insignificant importance in these reactions.