INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Tryptophan tryptophylquinone (TTQ) is the prosthetic group of methylamine dehydrogenase (MADH) from several methylotrophic and autotrophic bacteria (13, 14, 20) and aromatic amine dehydrogenase from Alcaligenes faecalis (18). TTQ is formed by the posttranslational modification of two tryptophan residues of the polypeptide chain. Two atoms of oxygen are incorporated into the indole ring of one of the tryptophan residues, and a covalent bond is formed between the indole rings of the two tryptophan residues (22) (Fig. 1). The mechanism by which MADH biosynthesis occurs is not known, but it does require the action of other proteins, the expression of which is subject to the same genetic regulation as the structural genes for the enzyme. The methylamine utilization (mau) gene cluster of Paracoccus denitrificans contains several genes (Table 1), at least four of which encode proteins that are required for biosynthesis of MADH, in addition to the two MADH structural genes (10, 28). Thus, the heterologous expression of recombinant MADH cannot be accomplished by simply cloning the MADH structural genes and introducing them into another host. Homologous expression of a single mau structural gene in P. denitrificans after inactivation of the corresponding chromosomal gene is also problematic because of the nature of the native mau promoter. MADH is expressed only when P. denitrificans is grown on methylamine as a sole source of carbon. The mau genes are repressed in the presence of any other carbon source, such as succinate or methanol (20, 28). Since a primary reason for developing an expression system for MADH is to perform site-directed mutagenesis to study structure-function relationships, it is not feasible for the expression system to require active MADH for growth. It would then be impossible to produce any MADH mutants which lacked sufficient activity to support cell growth. These concerns necessitated construction of an expression system which lacked the native MADH regulatory element and included all genes required for the complete synthesis of MADH under the control of a different promoter which was not controlled by methylamine levels.

View larger version (9K): [in this window] [in a new window]   FIG. 1.   Structure of TTQ. The residue numbers for the two gene-encoded tryptophan residues which are modified to form TTQ are for the enzyme from P. denitrificans.

As a first step toward elucidating the molecular details of the mechanism of TTQ biosynthesis and performing structure-function studies of MADH by site-directed mutagenesis, we have developed a system for the heterologous expression of correctly assembled MADH that is constitutive with respect to carbon source. Rhodobacter sphaeroides was chosen as a host because it is phylogenetically similar to P. denitrificans (32) but does not utilize methylamine as a carbon source (23). Since this organism has not routinely been used for heterologous expression of foreign proteins, a novel expression vector was designed. The coxII promoter for subunit II of cytochrome oxidase from R. sphaeroides was used to control the expression of the necessary components of the P. denitrificans mau gene cluster. This promoter is regulated by oxygen concentration and is unaffected by carbon source (7). We report here that expression of P. denitrificans genes mauFBEDACJG in R. sphaeroides yields the synthesis of active MADH which is indistinguishable from the native MADH of P. denitrificans. When a similar construct with mauFBEDACJG was placed in Escherichia coli, amicyanin, the mauC gene product, was expressed but MADH was not detected. The likely reasons for this are discussed.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Growth of cells and media. The bacterial strains and plasmids used in this study are listed in Table 2. For molecular biology experiments, E. coli cells were grown in liquid culture or on agar plates in LB medium at 37°C, and R. sphaeroides and P. denitrificans cells were grown in Sistrom's medium A (26) at 30°C. For expression of recombinant proteins, the following growth media were used. E. coli BL21(DE3) cells were grown aerobically at 30°C in LB medium, supplemented with 100 µM CuSO₄. When expression of proteins was under the control of the lac promoter, 1.6 mM isopropyl--D-thiogalactopyranoside (IPTG) was added 4 h prior to harvesting. These conditions have previously been shown to optimize plasmid-driven amicyanin production in these cells (15). P. denitrificans and R. sphaeroides were grown aerobically at 30°C in mineral salts medium (8), supplemented with 0.6 mM NaHCO₃, 1.6 µM CuSO₄, 0.5 g of yeast extract per liter, and 50 mM succinate as a carbon source. When appropriate, antibiotics were added to the following final concentrations: ampicillin, 100 µg/ml; spectinomycin, 50 µg/ml; streptomycin, 50 µg/ml; kanamycin, 25 µg/ml; and tetracycline, 10 µg/ml for E. coli and 1 µg/ml for P. denitrificans and R. sphaeroides.

Molecular biology methods. Plasmid DNA was isolated from cell cultures by using miniprep kits from Qiagen or Gibco BRL. Larger cell culture volumes were processed by using the Promega midiprep system. Restriction and modification enzymes were purchased from New England Biolabs. DNA fragments were isolated from agarose gels by using a Bio-Rad Prep-a-Gene or Gibco BRL Concert kit and were ligated by using the Pharmacia Ready-To-Go system. Competent E. coli JM109 cells were purchased from Promega. E. coli S17-1 cells were made competent by the method of Chung and Miller (11). DNA sequencing was performed by the dideoxy method, using an Amersham T7 Sequenase Quick-Denature kit. Site-directed mutagenesis was performed with a Stratagene Quik-Change kit. Primers for sequencing and site-directed mutagenesis were obtained from Cybersyn. DNA sequence information was analyzed with the DNA Strider 1.0 and Clone Manager 4.0 computer programs.

Assaying coxII promoter activity in E. coli. To assess the ability of the coxII promoter from R. sphaeroides to function in E. coli, plasmid-driven expression of -galactosidase activity was assayed in E. coli DH5F' cells. The negative control, pRKK200 (Table 2), contains the lacZ gene with no promoter. The positive control, pRK415-1, contains the lacZ gene with its normal lac promoter. Plasmid pRKKcoxII contains the coxII promoter from R. sphaeroides directly upstream of the lacZ gene. DH5F' cells have a partial deletion of the chromosomal -galactosidase gene and so produce no background -galactosidase activity. Cells were grown aerobically in LB media to an optical density at 600 nm of approximately 0.4. IPTG was added to the positive control to induce the lac promoter. -Galactosidase activity was assayed in 0.1-ml aliquots of cells cultures as described in reference 24.

Construction of plasmids for the expression of MADH. The strategy for constructing plasmids for the expression of MADH is summarized in Fig. 2. The source of the genes for MADH expression was pMAU1K, a plasmid containing the entire mau locus that was originally isolated from the chromosome of P. denitrificans (28). The mau locus has been completely sequenced, and the functions of many of the gene products either are known or have been inferred from their sequences (9, 28-31) (Table 1). A 6.6-kb fragment containing the genes mauFBEDACJG was cut out of pMAU1K by using BsaAI and ApaI and inserted into both pBluescript II (pBSII) SK and pBSII KS, each of which had been opened with SmaI and ApaI. The BsaAI site occurs 164 bases upstream of the MauF protein start codon. During the cloning manipulations, 81 bases on the 5' end of this fragment were lost, fortuitously leaving another blunt end which could be ligated with the SmaI site of the pBSII vectors. Since BsaAI and SmaI produce blunt ends, this ligation resulted in loss of the SmaI site. Sequencing confirmed that the final constructs contained 83 bases of the intergenic region on the 5' side of mauF. The pBSII KS construct pMEG994 was used for further cloning manipulations. The pBSII SK construct pMEG993, which contains the mau genes in an orientation downstream of the lac promoter, was transformed into E. coli BL21(DE3) cells to test for expression of MADH.

View larger version (18K): [in this window] [in a new window]   FIG. 2.   Construction of plasmids for the heterologous expression of MADH.

Purification of proteins. With each bacterium, the periplasmic fraction of the cells was prepared as described previously (12) for P. denitrificans. To further fractionate cells, the spheroplasts which remained after the periplasm preparation were sonicated and then centrifuged to yield a supernatant that contained the cytoplasmic fraction and a pellet that contained the membrane fraction. MADH was purified from the periplasmic extracts by ion-exchange chromatography followed by gel filtration chromatography as described previously (12).

Biochemical characterization of recombinant MADH. The steady-state kinetic activity of MADH was assayed as previously described (12). Previously described methods were also used for transient kinetic studies in which stopped-flow spectroscopy was used to determine the binding constants as well as microscopic reaction rate constants for the reduction of MADH by methylamine (6) and oxidation of MADH by amicyanin (4, 5). The redox potential of MADH was determined spectrochemically as described previously (33). Polyacrylamide gel electrophoresis was performed by standard methods, and Western blot analyses were performed with polyclonal antibodies that had been raised against purified MADH from P. denitrificans.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Expression of recombinant MADH. Plasmid pMEG986 was introduced via conjugation with E. coli S17-1 cells into R. sphaeroides and P. denitrificans to test for MADH expression in these bacteria. The relative efficiency of heterologous expression of MADH in R. sphaeroides and homologous expression of MADH in P. denitrificans are summarized in Table 3. The highest level of expression of recombinant MADH was observed with R. sphaeroides, which normally does not synthesize MADH. The level of MADH produced by R. sphaeroides containing pMEG986, when grown on succinate as a carbon source, was approximately 35 mg/100 g (wet weight) of cells. This is about 20% of the level of MADH normally produced by wild-type P. denitrificans when grown on methylamine as a sole source of carbon and possibly an indication of the relative strengths of the P. denitrificans mau and R. sphaeroides coxII promoters. When P. denitrificans with no plasmid was grown on succinate as a carbon source, no MADH was detected. Interestingly, in P. denitrificans containing pMEG986, a very low level of homologous expression of MADH was observed during growth on succinate. The actual level was approximately 3% of that produced by R. sphaeroides containing pMEG986, when grown on succinate. This result suggests that the coxII promoter from R. sphaeroides does not function nearly as well in P. denitrificans as it does in R. sphaeroides.

Localization of MADH in the periplasm. Native MADH is normally localized in the periplasm of P. denitrificans. The export of MADH across the bacterial inner membrane most likely does not occur via the standard Sec protein export pathway. The product of the mauB gene, which codes for the small TTQ-bearing subunit of MADH, contains a relatively unusual 57-residue-long signal sequence (10). This sequence includes a recently characterized double-arginine motif (3) that is translocated by the gene products of a special tatABCD operon (25). It is also possible that at least one of the other mau gene products is required to assist in this specialized export and processing of the MADH small subunit. When MADH was expressed in R. sphaeroides, 100% of the detectable activity of MADH was present in the periplasmic fraction. The spheroplasts which remained after separation of the periplasmic fraction were subjected to sonication and further analysis. No MADH activity was detected in the cytoplasmic or membrane fractions. For comparison with a cytoplasmic marker, malate dehydrogenase activity (1) was assayed in each fraction. The distribution of this cytoplasmic marker was 34, 64, and 2%, respectively, in the periplasmic, cytoplasmic, and membrane fractions. Thus, while it appears that the fractionation procedure does release some cytoplasmic protein into the periplasmic fraction, the absence of any MADH in the cytoplasmic fraction clearly indicates that it is localized exclusively in the periplasm.

Characterization of recombinant MADH. During purification of recombinant MADH, the positions at which it was eluted during ion-exchange chromatography and gel filtration chromatography were the same as observed during the purification of the native MADH from P. denitrificans, suggesting no significant perturbation of structure. The purified recombinant MADH was analyzed by polyacrylamide gel electrophoresis under denaturing and nondenaturing conditions and compared with MADH from P. denitrificans. The recombinant and natural proteins were indistinguishable from each other.

Lack of MADH expression in E. coli. Plasmids pMEG987 and pMEG993 were introduced into E. coli BL21(DE3) cells by transformation to test for expression of MADH. No MADH activity was detected in E. coli cells which harbored pMEG987 (Table 3). However, these cells did express amicyanin, the mauC gene product. This is of interest because in this construct, synthesis of amicyanin indicates that transcription and translation of at least mauF through mauC must have occurred (Fig. 2). Antibodies to the MADH subunits were used in Western blot analyses to test for their presence in the periplasmic, cytoplasmic, and membrane fractions of these cells. No significant levels of either subunit in any of these cell fractions were detected. The lack of significant cross-reactive protein in the cytoplasm rules out the possibility that inclusion bodies had formed, and the absence of significant cross-reactive protein in the membrane fraction indicates that it was not accumulating there due to a defect in translocation. These results suggest that if the mauB and mauA gene products are synthesized, then they are rapidly degraded.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Our goal was to construct a vector which contained all genes necessary for the heterologous expression, biosynthesis, and assembly of MADH and which was under the control of a foreign promoter that was neither dependent on methylamine nor repressed by other carbon sources. This was achieved by placing mauFBEDACJG from the mau gene cluster of P. denitrificans under the control of the coxII promoter from R. sphaeroides in a broad-range-host vector and expressing these genes in R. sphaeroides. In the course of this study, we gained important information about the assembly of MADH and the potential utility of R. sphaeroides and the coxII promoter for the heterologous expression of foreign proteins.

These results prove that the genes mauFBEDACJG are sufficient for the complete biosynthesis of MADH and that the genes mauM and mauN are not required. While the genes mauFBEDACJG provide the genetic information necessary for MADH biosynthesis, certain specific features of the host bacterium are also necessary for correct biosynthesis and assembly of MADH. This is illustrated by the fact that these genes in pMEG987 did not support synthesis of a functional MADH in E. coli, while amicyanin was produced. The reason for this is not clear, but several aspects of the biosynthesis of MADH appear unusual. The product of the mauB gene, which codes for the small TTQ-bearing subunit of MADH, possesses an unusual 57-residue-long signal sequence (9, 10) which exhibits a double-arginine motif (3) that is translocated by the gene products of the special tatABCD operon (25). It has been proposed that signal sequences such as this may be characteristic of periplasmic proteins which possess complex redox cofactors (3). E. coli does possess the tatABCD operon (25) and in principle should be able to process the mauB signal sequence. Thus, the inability to synthesize a functional and stable MADH in E. coli must be related to other unusual features of MADH biosynthesis. From their deduced sequences, the mauE product appears to be a membrane protein (29), the mauD product may be involved in disulfide bond processing (29), and the mauG product may be a heme-bearing peroxidase (10, 29). These gene products may play essential roles in translocation of the MADH small subunit across the membrane, formation of the six disulfide bonds which are present in that subunit, and oxygenation reactions that are required for TTQ biosynthesis. If synthesis, assembly and proper localization of any of these accessory gene products are not accomplished by E. coli, then functional MADH expression will be unsuccessful. It was previously shown (29) by Western blot analysis that P. denitrificans cells which possessed mutations in either mauE or mauD were devoid of the MADH small subunit and had reduced levels of the large subunit. Amicyanin levels in these mutants, however, were comparable to wild-type levels. It may be that E. coli lacks certain endogenous proteins, cofactors, or specific features of its membrane that prevent the functional expression of one of these accessory gene products, which in turn prevents the correct posttranslational processing and assembly of MADH. The inability to detect significant amounts of either MADH subunit in E. coli extracts which contained amicyanin expressed from pMEG987 is probably due to the rapid degradation of incorrectly assembled or modified subunits. Such special requirements for the proper processing of accessory gene products are apparently satisfied by R. sphaeroides, which does allow the expression of the functional P. denitrificans MADH.

The mau genes allow P. denitrificans to grow on methylamine as a sole source of carbon (28-30). R. sphaeroides is incapable of growth with methylamine as its sole source of carbon. This raises the question of whether providing this bacterium with the ability to synthesize MADH would be sufficient to enable it to utilize methylamine as a metabolic carbon source. R. sphaeroides containing pMEG986 was unable to grow with methylamine as a sole carbon source under the growth conditions used in this study (data not shown). It was previously shown that R. sphaeroides can use methanol as a sole carbon source, but only under anaerobic photosynthetic growth conditions (2, 23). Anaerobic conditions are necessary to induce adequate levels of formaldehyde dehydrogenase (2), which is needed to convert the methanol-derived formaldehyde to formate, which is further oxidized to CO₂ to support autotrophic growth. Unfortunately, it was not possible to determine whether MADH could support methylamine-dependent growth under these anaerobic conditions because the coxII promoter used in our constructs will function only under aerobic conditions.

This work demonstrates the potential value of R. sphaeroides as a host for heterologous expression of proteins, especially complex proteins such as MADH that cannot be expressed in E. coli. The coxII promoter in constructs similar to pMEG986 should be suitable for heterologous expression of other gene products in R. sphaeroides. While the levels of MADH synthesis in R. sphaeroides were lower than those in the native P. denitrificans when induced by methylamine, this is more likely a reflection of the strength of the native mau promoter than a weakness of the coxII promoter. The level of protein expression which was obtained in this way was sufficient for biochemical characterization of MADH, and it should be sufficient for characterization of other heterologous proteins that cannot be expressed in standard systems.