INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Alcaligenes eutrophus H16 is a gram-negative, strictly respiratory bacterium with a facultatively lithoautotrophic lifestyle. The organism grows on a wide range of sugars and organic acids. In the absence of such substrates, it can utilize H₂ and CO₂ as the sole sources of energy and carbon, respectively (reviewed in references 6, 17, and 18). Two biochemically and cytologically distinct enzymes catalyze the oxidation of molecular hydrogen in A. eutrophus. A heterodimeric membrane-bound hydrogenase (MBH) couples hydrogen oxidation to electron transport phosphorylation in a membrane-bound respiratory complex (42). The MBH is attached to the periplasmic surface of the inner membrane (14, 23). This enzyme is representative of a widespread type of [NiFe] hydrogenase, examples of which have been found in many different groups of gram-negative bacteria (17). The other hydrogenase of A. eutrophus is a heterotetrameric soluble hydrogenase (SH). Like the MBH, it is a nickel metalloenzyme (21). The SH couples hydrogen oxidation to NAD reduction, supplying the cell with reducing power under lithoautotrophic conditions (44). Homologous enzymes have been found in both gram-negative and gram-positive bacteria (26).

Both the SH and the MBH undergo a complex maturation process requiring an ensemble of specialized accessory proteins (4, 10, 29, 30, 35, 48). The enzymes themselves and their respective accessory proteins are encoded in neighboring gene clusters on the 450-kb endogenous megaplasmid pHG1 (11, 29, 48, 51). Altogether, 31 hydrogenase-related genes have been identified to date. Specialized maturation proteases (4, 48), metal-center-assembly proteins (10), a type b cytochrome (3), and a high-affinity nickel transporter (15) are among the products of the hydrogenase gene cluster.

Although similar hydrogenases are found in other lithoautotrophs, the pattern of hydrogenase regulation in A. eutrophus H16 is exceptional. Even in the closest relatives, such as A. hydrogenophilus, hydrogenase expression is strictly H₂ dependent (33). In contrast, A. eutrophus H16 synthesizes both hydrogenases not only in the presence of H₂ but also during growth on poor carbon sources. Hydrogenase synthesis is blocked during growth on preferentially utilized carbon sources, such as succinate and pyruvate. Thus, it appears that the signal which triggers derepression of the hydrogenase system is not a particular substrate but rather is a physiological cue related to the energy status of the cell (19, 22).

Early studies showed that the expression of the hydrogenases of A. eutrophus H16 is coordinate, implying the existence of a central regulatory function or functions (20). Indeed, subsequent genetic analysis led to the identification of a locus, designated hoxA, encoding a factor required for the synthesis of both hydrogenases (41). Sequencing revealed that the hoxA gene product is a member of the NtrC group of the superfamily of transcriptional activators (12). Somewhat later, a gene encoding the cognate histidine kinase was discovered (18). However, the product of this gene is inactive (18, 32). Among the hydrogenase null mutants was a strain with a defect in a chromosomal locus. The defective gene product turned out to be an rpoN homolog, indicating that the expression of at least some of the hydrogenase genes is dependent on the minor sigma factor ⁵⁴ (39, 53).

The finding that an NtrC-like activator is required for hydrogenase synthesis in A. eutrophus H16 suggests that key genes of the hydrogenase system are regulated at the level of transcription. The present study, focusing on the promoters controlling the genes for the catalytic subunits of the SH (hoxFUYH) and the MBH (hoxKG), provides the first decisive evidence for this assumption. We identified the probable promoter sites by means of primer extension mapping. Plasmid-borne reporter constructs and a quantitative assay for hydrogenase transcripts were used to demonstrate that the expression of the hydrogenase genes is regulated at the level of promoter activity and to monitor the course of induction. We show that the activities of the SH and the MBH promoters in vivo are dependent on HoxA and ⁵⁴. Furthermore, a heterologous expression system was used to measure the HoxA-mediated activation of the SH and the MBH promoters in the absence of other A. eutrophus gene products.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Strains and plasmids. Bacterial strains and plasmids are listed in Table 1. A. eutrophus H16 is the wild-type strain harboring the endogenous megaplasmid pHG1. Strains HF09 (20) and HF18 (41) are derivatives of H16. Escherichia coli S17-1 (46) served as a donor in conjugative transfers.

Media and growth conditions. Strains of A. eutrophus were grown in modified Luria broth containing 0.25% sodium chloride and 0.4% fructose or in mineral salts medium as described previously (12). Synthetic media for heterotrophic growth contained 0.4% fructose (FN medium), 0.4% succinate (SN medium), or 0.2% fructose and 0.2% glycerol (FGN medium). FGN medium contained 1 µM NiCl₂ in place of standard trace elements mixture SL6 (12). Lithoautotrophic cultures were grown in mineral salts medium under an atmosphere of hydrogen, carbon dioxide, and oxygen (8:1:1, vol/vol/vol). Strains of E. coli were grown in Luria-Bertani medium or in M9 medium containing glycerol (36). Solid media contained 1.5% agar. Antibiotics were added as appropriate (for A. eutrophus: kanamycin, 350 µg/ml; tetracycline, 15 µg/ml; for E. coli: kanamycin, 25 µg/ml; tetracycline, 15 µg/ml; ampicillin, 100 µg/ml). For induction of trc promoter-driven expression constructs, isopropyl--D-thiogalactopyranoside (IPTG) was added to cultures to a final concentration of 1 mM.

Conjugative plasmid transfer. Mobilizable plasmids were transferred from E. coli S17-1 to A. eutrophus by a spot mating technique (46). Transconjugants were selected on FN medium plates containing the appropriate antibiotics.

DNA techniques. Standard DNA techniques were used in this study (2). Large-scale isolation of plasmid DNA was carried out by the alkaline lysis procedure followed by ethidium bromide-cesium chloride gradient centrifugation. Smaller amounts of plasmid DNA were isolated with QIAGEN tip-20 columns (QIAGEN Inc.) according to the manufacturer's instructions. DNA fragments used in plasmid constructions were isolated from agarose gels with the GlassMAX spin column system (Life Technologies, Inc.).

Isolation of RNA. Total cellular RNA was isolated from 2-ml samples of cell suspension by a hot-phenol method (24).

Primer extension analysis. 5' Ends of in vivo mRNAs were mapped by a primer extension protocol (24). The synthetic oligonucleotide BF153 (5'-GTATGTCGATCAGCCGTGTACGG-3') was used as a specific primer for SH transcripts. BF174 (5'-TTCAGGAAACTTCGTCGCGA-3') and BF185 (5'-TGCCTGCGCATGACTTCATA-3') were used for mapping MBH transcripts. Primer extension reaction mixtures included 10 µg of RNA, 0.2 pmol of ³²P-labelled oligonucleotide, and 200 U of Moloney murine leukemia virus reverse transcriptase (Life Technologies). Following phenol-chloroform extraction and ethanol precipitation, extension products were separated in 6% sequencing gels together with sequence ladders of the corresponding regions and detected by autoradiography. For quantitative transcript determinations, the radioactivity of excised bands was determined with a Canberra-Packard 1600TR liquid scintillation counter.

RNase protection assays. Riboprobes were synthesized with a MAXIscript kit (Ambion, Inc.) and ³²P-labelled UTP (800 Ci/mmol; Dupont NEN). NdeI-linearized plasmid pCH185 and DdeI-linearized plasmid pCH292 served as templates for the generation of the MBH- and SH-specific probes, respectively. The sizes of the riboprobes were 204 and 225 nucleotides, respectively. The efficiency of incorporation of the radioactive label was monitored by trichloroacetic acid precipitation. Subsequently, the in vitro transcripts were purified by two rounds of ethanol precipitation. Total RNA (5 to 20 µg) was added to 30 µl of hybridization buffer (40 mM piperazine-N,N'-bis(2-ethanesulfonic acid [PIPES] [pH 6.4], 0.4 M NaCl, and 1 mM EDTA in a 1:4 (vol/vol) mixture of water-deionized formamide) containing 10⁵ to 10⁶ cpm of the appropriate riboprobe. After an initial denaturation step (5 min at 85°C), hybridization proceeded for at least 8 h at 45°C. RNase digestion cocktail (10 mM Tris-HCl [pH 7.5], 300 mM NaCl, 5 mM EDTA, 40 µg of RNase A per ml, 2 µg of RNase T₁ per ml) (350 µl) was added, and the mixture was incubated for 30 min at 30°C. Treatment with proteinase K (10 µl of 20% [wt/vol] sodium dodecyl sulfate [SDS], 2.5 µl of proteinase K (20 mg/ml); 37°C for 15 min) was followed by phenol extraction and precipitation in the presence of 10 µg of yeast tRNA. The pellet was dissolved in 3 to 5 µl of formamide loading buffer (95% formamide, 20 mM EDTA, 0.05% bromophenol blue, 0.05% xylene cyanol FF), and the mixture was applied to a 6% sequencing gel. In vitro transcripts of known lengths served as size standards. Quantitation of the protected fragments was done either by counting the radioactivity of excised bands with a Canberra-Packard 1600TR liquid scintillation counter or by analyzing scanned images obtained with a Molecular Dynamics 445 SI storage PhosphorImager by use of IPLab Gel software (Signal Analytics).

Enzyme assays. For enzyme assays, independent single colonies were picked from plates and inoculated into liquid media. Precultures were incubated for 15 to 20 h at 35°C. Since the hydrogenase system is repressed at temperatures above 33°C, this step ensured that the cells were uniformly devoid of hydrogenase at the beginning of an experiment. SH (hydrogen:NAD⁺ oxidoreductase; EC 1.12.1.2) activity was assayed by spectrophotometric determination of H₂-dependent NAD reduction in detergent-treated cells (16). MBH (ferredoxin:H⁺ oxidoreductase; EC 1.18.99.1) activity was determined by measurement of H₂-dependent methylene blue reduction in isolated membranes (42). One unit of hydrogenase activity was the amount of enzyme which catalyzed the formation of 1 µmol of product per min. -Galactosidase was assayed as described previously (57), and the activity (in units) was calculated according to Miller (36) except that cell optical density was measured at 436 nm (OD₄₃₆). Unless otherwise indicated, enzyme activities were assayed in mid-log-phase cells, i.e., cells grown to optical densities of 3 in SN medium, 5 in FN medium, 7 in FGN medium, and 1 under lithoautotrophic conditions. Protein determinations were done according to the method of Lowry et al. (34).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Primer extension mapping. Previous complementation studies with cloned DNA fragments carrying the SH and the MBH loci identified the 5' boundaries of the upstream sequences sufficient for high-level expression of the two enzymes. Sequencing showed that the complementing segments extended 259 bp upstream of the initiation codon of hoxF and 544 bp upstream of the initiation codon of hoxK, respectively (29, 51). In order to further localize the respective promoters, the 5' ends of in vivo transcripts of the two enzymes were mapped by means of the primer extension technique. For the SH, we detected a single extension product with a 5' terminus corresponding to bp 654 of the published sequence (51 bp upstream of the ATG of hoxF) (Fig. 1A). This finding confirms the previously published result (51) and suggests that the SH transcript starts at this site. The sequence 5'-TTGGCGCACATCCTGC-3', located a short distance upstream, is a likely candidate for the SH promoter (P_SH). Primer extension analysis of MBH transcripts gave unique signals for the two primers used. The sizes of the extension products indicated a common 5' end corresponding to bp 456 of the published sequence (94 bp upstream of the ATG of hoxK), suggesting a transcription start site at this position (Fig. 1B). Located 13 bp upstream is a sequence resembling a 24/12 promoter: 5'-CAGGCATAGATCTTGT-3'. This sequence (P_MBH) differs from the canonical sequence for RpoN-dependent promoters (49) at the second position of the second conserved dinucleotide (T instead of C). However, a similar sequence has been reported for the promoter of the psp operon (55).

View larger version (28K): [in this window] [in a new window]   FIG. 1.   Nucleotide sequences of the PSH (A) and the PMBH (B) regions. The start codons of hoxF and hoxK are highlighted, and the deduced N-terminal amino acid sequences are given below the nucleotide sequences. Shine-Dalgarno (S.D.) homologies are underlined. The 5' ends of in vivo mRNAs determined by primer extension are indicated by wavy lines. The putative PSH and PMBH are boxed, and the conserved dinucleotides of the 24/12 motifs are emphasized. Sequence elements resembling the consensus E. coli IHF binding site (9, 25) are indicated by asterisks. Palindromic sequence motifs are indicated by pairs of arrows above the sequence. Numbering is based on respective sequence publications (29, 51). A "G" was added to the sequence of the hoxF upstream region between positions 472 and 473 to correct a sequence error. The GenBank entry (accession no. 55230) has been amended accordingly.

Expression of the genes for the SH and the MBH enzymes is regulated at the transcriptional level. In order to study the activity of P_SH and P_MBH in vivo, we inserted fragments carrying the respective regions upstream of the promoterless lacZ gene in low-copy vector plasmid pEDY305 and monitored the -galactosidase activities produced by the resulting plasmids in cells of A. eutrophus H16 growing under hydrogenase-repressing and hydrogenase-inducing conditions. Succinate is a preferred substrate for A. eutrophus H16 and supports rapid growth. Both hydrogenases are tightly repressed in cultures grown on succinate (19). Succinate-grown cells harboring the indicator plasmids contained only small amounts of -galactosidase, indicating very weak transcription from the two promoters (Fig. 2A and B). Under autotrophic conditions with H₂ as the energy source, the organism relied on the hydrogenases to generate energy, and both enzymes were synthesized at high levels (Fig. 2E and F). The test-plasmid-harboring strains produced high levels of -galactosidase during growth on H₂, indicating strong transcription from P_SH and P_MBH (Fig. 2A and B). The hydrogenases are also derepressed to various degrees during growth on suboptimal carbon sources, i.e., carbon sources which support lower growth rates than does succinate. Cells grown on fructose, for instance, produce intermediate levels of SH and MBH (19). The test-plasmid-harboring strains produced intermediate levels of -galactosidase when cultivated on fructose (Fig. 2A and B). A. eutrophus H16 grows very slowly on glycerol and synthesizes large quantities of active SH and MBH (19). We also measured -galactosidase activities in test strains grown on glycerol. For convenience, the latter experiments were done with diauxic cultures supplemented with a mixture of fructose and glycerol. High levels of -galactosidase were present in the test strains following the transition to glycerol (Fig. 2A and B). Taken together, these results reveal a pattern of transcription approximately reflecting the patterns of SH and MBH activities (Fig. 2E and F). We conclude that transcriptional control is the major component in the regulation of these enzymes.

View larger version (32K): [in this window] [in a new window]   FIG. 2.   Comparison of promoter activities, transcript pool sizes, and hydrogenase activities for the SH and the MBH. Single colonies of A. eutrophus H16 or of A. eutrophus H16 harboring the indicator plasmids pGE319 and pGE320 were picked, inoculated into fructose medium, and grown for 15 to 20 h at 35°C. Fresh media were seeded from the precultures to an OD436 of 0.1, and the cultures were grown to the mid-log phase. Samples were taken for determination of -galactosidase or hydrogenase activity or for preparation of RNA. For heterotrophic cultures, the carbon and energy sources were succinate (SN medium), fructose (FN medium), and a mixture of fructose and glycerol (FGN medium). Autotrophic cultures (H2/CO2) were treated with a mixture of hydrogen, carbon dioxide, and oxygen gases. Specific activities of the hydrogenase enzymes are given as micromoles of product formed per minute per milligram of protein. The values for transcript abundance for the lithoautotrophic cultures were arbitrarily taken as 100%. Bars represent the means of five independent determinations. Except in the case of the transcript determinations, standard errors were too small for graphic representation. Rel., relative.

Cellular levels of the SH and the MBH transcripts reflect the activities of P_SH and P_MBH. The experiments with the promoter assay plasmids reported above showed that the differential activity of P_SH and P_MBH is the basis of the derepression of the hydrogenase enzymes. In this test system, however, differences in translational efficiency can skew promoter activity measurements. Therefore, quantitative comparisons are necessarily limited to values obtained under identical conditions. We therefore assayed a second transcriptional parameter, transcript abundance, by the RNase protection technique. A. eutrophus was grown under the standard repressing and derepressing conditions described above, and samples were taken from mid-log-phase cultures for the isolation of total cellular RNA. In RNA from succinate-grown cells, SH and MBH transcripts were not detectable (Fig. 2C and D). High levels of both transcripts were detected in glycerol- and H₂-grown cells. Cells grown on fructose contained intermediate amounts of SH and MBH transcripts. Thus, the patterns of transcript abundance for the SH and the MBH mRNAs reflected the profiles of apparent P_SH and P_MBH activities, confirming the transcriptional regulation of the hydrogenase genes.

SH and MBH transcript levels increase abruptly upon derepression. We used quantitative primer extension assays to monitor changes in the SH and the MBH transcript pools in A. eutrophus during diauxic growth on a mixture of fructose and glycerol and during lithoautotrophic growth on H₂ (Fig. 3A and B). Diauxic and lithoautotrophic cultures were seeded with fructose-grown cells. In the diauxic cultures, constant transcript levels were found during late exponential growth on fructose. After this substrate was exhausted, the cells entered a lag phase before resuming growth (not visible at the scale of the graph in Fig. 3A). An increase in hydrogenase transcript levels began shortly after the onset of this lag phase and continued for 10 to 20 h. After this period, transcript pool levels began declining. Transfer from a heterotrophic culture with a preferred substrate, such as fructose, to lithoautotrophic growth conditions also led to an increase in transcript levels (Fig. 3B). This increase continued during exponential growth. The mRNA levels peaked at the onset of the stationary phase (at an OD₄₃₆ of 9), whereupon a rapid decline set in. It appears that in both the lithoautotrophic and the diauxic cultures, transcription is triggered in response to exhaustion of a preferentially utilized substrate.

View larger version (16K): [in this window] [in a new window]   FIG. 3.   Kinetics of SH and MBH transcript pools following derepression of the hydrogenase system. (A) Relative transcript abundance in diauxic fructose-glycerol-grown cells. (B) Relative transcript abundance in lithoautotrophic cells. Cultures were grown at 30°C and sampled at various intervals. After determination of the OD436, total cellular RNA was isolated from 2-ml samples, and 10 µg each of the resulting RNA samples was added to primer extension reactions containing the specific 32P-labeled primers. After separation in 6% sequencing gels, the radioactive bands were excised and counted in a liquid scintillation counter. Radioactivity (in counts per minute) and cell density (OD436) were plotted against time. Symbols: , radioactivity of extension products obtained with the SH-specific primer;  radioactivity of extension products obtained with the MBH-specific primer;  cell density.

Transcription from P_SH and P_MBH is dependent on ⁵⁴ and the hydrogenase regulator HoxA. Genetic studies revealed that ⁵⁴ and the positive regulator HoxA are absolutely required for hydrogenase synthesis (12, 39). In order to directly test the requirement for these gene products for the transcriptional activities of P_SH and P_MBH, we introduced indicator plasmids into A. eutrophus HF09 and HF18 (Table 1), which are null mutants for rpoN and hoxA, respectively. Comparison of the -galactosidase activities produced by the mutant and wild-type strains under hydrogenase-derepressing conditions indicated that the promoter activities were marginal in both the ⁵⁴- and the HoxA-deficient backgrounds (Fig. 4). RNase protection assays with RNAs from strains HF09 and HF18 failed to detect SH and MBH transcripts (data not shown). Taken together, these data show that P_SH and P_MBH require both gene products for activation.

View larger version (14K): [in this window] [in a new window]   FIG. 4.   Activities of PSH and PMBH in rpoN and hoxA mutants. A. eutrophus H16 and transconjugants harboring plasmids pGE319 (open bars) and pGE320 (closed bars) were grown on FGN medium to an OD436 of 8. Samples were assayed for -galactosidase activity (see the legend to Fig. 2 for details). Bars represent the means of five independent determinations. Standard errors were too small for graphic representation. wt, wild type.

Induction of P_SH and induction of P_MBH proceed at similar rates. Promoter activity measurements with the two test-plasmid-harboring strains showed that -galactosidase activities remained constant during the course of logarithmic growth on succinate, fructose, and H₂, as was expected when the rates of transcription, translation, transcript decay, and enzyme degradation were in equilibrium (data not shown). In contrast, transcriptional measurements in cells grown on glycerol revealed prolonged induction kinetics (data not shown). Thus, under the latter conditions, singular measurements at arbitrary time points do not provide a meaningful measure for quantitating relative promoter strength. In this situation, the rate of induction of a promoter is a more reliable representation of its transcriptional activity. Therefore, we monitored the increase in -galactosidase activity during the initial phase of logarithmic growth on glycerol and plotted the data points against cell density (Fig. 5). Both promoters showed linear induction kinetics. The induction of P_SH and the induction of P_MBH proceeded at the same rates during growth on glycerol.

View larger version (14K): [in this window] [in a new window]   FIG. 5.   Induction rates for PSH and PMBH during exponential growth on glycerol. A. eutrophus H16 harboring plasmids pGE320 (open circles) and pGE319 (closed circles) was grown on FGN medium. Samples were taken at 30-min intervals from the mid-log-phase cultures and assayed for -galactosidase activity (see the legend to Fig. 2 for details). Linear regression plots show -galactosidase activities as a function of OD436. Representative results from two experiments are shown.

HoxA activates transcription from P_SH and P_MBH in a heterologous system. HoxA is a response regulator-type transcriptional activator (12). Regulatory proteins of this class are typically paired with sensory proteins of the histidine kinase family. Interaction between the two alters the phosphorylation status of the regulator, thereby altering its capacity to activate cognate promoters (54). Some response regulators are also capable of activating transcription in the absence of their specific kinase (27, 40). In order to confirm the role of HoxA in the activation of P_SH and P_MBH and to determine whether other gene products unique to A. eutrophus are required, we introduced a plasmid-borne copy of hoxA under the control of a regulatable promoter into E. coli strains carrying indicator plasmids and measured -galactosidase activities after the induction of HoxA (Fig. 6). The -galactosidase levels of the induced cultures increased over the course of the experiment, whereas the levels of the uninduced controls remained constant. The -galactosidase activities of the induced cultures were relatively low; nevertheless, the difference between the induced and the uninduced cultures was significant. Thus, HoxA mediates weak activation of cognate promoters in a heterologous system. The low level of transcription suggests that additional regulatory components specific to A. eutrophus are necessary for efficient activation.

View larger version (18K): [in this window] [in a new window]   FIG. 6.   HoxA-mediated activation of PSH and PMBH in E. coli. E. coli JM109(pCH618/pGE320) and E. coli JM109(pCH618/pGE319) were inoculated into fresh minimal medium containing glycerol and grown at 30°C to an OD600 of 0.3. The cultures were then split, and one duplicate of each was induced by the addition of IPTG to a final concentration of 1 mM (0 h). Samples were taken immediately and thereafter at 2-h intervals for 8 h and assayed for -galactosidase activity. Final samples were taken at 18 h. Symbols:  and  induced cultures;  and  uninduced cultures;  and , JM109(pCH618/pGE320);  and , JM109(pCH618/pGE319). The experiment was done twice, with comparable results.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Previous studies suggested that the expression of the SH and the MBH genes in A. eutrophus is controlled by a transcriptional mechanism similar to the glnAp2 paradigm (54). The above results provide extensive evidence in support of this hypothesis. Two independent lines of evidence confirm transcriptional control. First, in vivo assays of promoter activity revealed high levels of transcription under hydrogenase-derepressing conditions. Second, determination of the relative abundances of the SH and the MBH transcripts by a physical method showed an approximate correlation between transcript pools and the respective hydrogenase activities. The latter method permitted a direct quantitative comparison between data collected under different growth conditions for a given transcript. It should be noted that, under our experimental conditions, a quantitative comparison of the SH and the MBH mRNA pools relative to each other was not possible. On the other hand, the promoter activity data allowed conclusions to be drawn about the relative strengths of P_SH and P_MBH. The activity of P_SH was higher under all conditions except growth on glycerol. The correlation between the two data sets suggests that differential transcript stability does not play a major role in the regulation of SH and MBH expression.

The hydrogenase activities reflect the transcriptional data, with an obvious exception. The activity of the SH in lithoautotrophically grown cells was disproportionately low. This result may have been due to the inactivation of the enzyme in the presence of O₂ and electron donors, e.g., H₂ (43). This inactivation has been shown to take place in vivo and may be caused by superoxide radicals produced by the hydrogenase itself.

The kinetics of the transcript pools of cells growing on glycerol and on H₂ revealed a dramatic increase in SH and MBH mRNAs upon derepression. On the whole, these kinetics resembled the kinetics of enzyme activities for the two hydrogenases. Furthermore, the increase in transcript levels was coordinate, as was the appearance of the enzyme activities (19). This finding suggests that derepression is synchronized by a common mechanism. Our data showing that both P_SH and P_MBH are HoxA dependent indicate that HoxA is the synchronizing agent. Earlier studies on carbon- and oxygen-limited continuous cultures revealed a link between the derepression of hydrogenase synthesis and limitation of energy (19, 22). The transcriptional data presented here are compatible with the physiological findings. Transcript levels began rising during the lag phase after the exhaustion of fructose in the fructose-glycerol cultures or during the initial lag phase in the lithoautotrophic cultures.

Primer extension analysis identified putative transcription start points upstream of hoxF and hoxK. Although data of this type are notoriously subject to artifacts due to, e.g., transcriptase stalling, transcript degradation, and promiscuous priming, they are an invaluable aid in identifying promoters. Sequence elements resembling the 24/12 consensus sequence of ⁵⁴-dependent promoters are located just upstream of the putative transcription start points. The presumptive SH promoter, 5'-TTGGCGCACATCCTGC-3', contains the typical GG-N₁₀-GC motif. The candidate for the MBH promoter, 5'-CAGGCATAGATCTTGT-3', contains a T at the 12 position. This exception is rare but has been documented in at least one case (55). The finding that transcription from the SH and the MBH upstream regions is ⁵⁴ dependent supports the assignment of these sequences as the SH and the MBH promoters. RNA polymerase bound to an ⁵⁴-dependent promoter requires a transcriptional activator to form an open complex. Our data show that the activities of both P_SH and P_MBH are absolutely dependent on the NtrC homolog HoxA.

Studies of the glnAp2 promoter led to a molecular model for transcriptional activation (31). Key features of this model are the binding of an activator protein at a specific site upstream of a ⁵⁴-dependent promoter and looping of the DNA to permit direct contact of the activator and the polymerase (47). The spacing of the regulatory sites is important. Typically, the binding site for the activator lies between 120 and 160 relative to the transcription start site (8). In the hoxF upstream region, tandem palindromes consisting of the motif 5'-CAAG-N₁₀-CTTG-3' are centered at 159 and 202. Deletion analysis showed that this region contains a signal essential for high-level SH expression (57). A similar sequence motif is found in the hoxK upstream region. This motif consists of the sequence elements 5'-CATG-N₁₁-ATTG-3' and 5'-CAGG-N₉-CTTG-3' centered at 187 and 210, respectively. These distances are atypical but are within the range of published values (8). For the nifF promoter, for instance, the NifA binding site is located between 250 and 270 (37).

An important feature of the glnAp2 activation mechanism is the participation of the DNA-bending protein integration host factor (IHF) (28). IHF binds to the DNA at a site between the promoter and the activator binding site and facilitates DNA looping. E. coli IHF binds to fragments containing the hoxF promoter in vitro (57). It remains to be shown whether an A. eutrophus IHF homolog plays a role in the activation of the hoxF and/or hoxK promoters.

Among the best-studied hydrogen-oxidizing bacteria are Rhodobacter capsulatus, Bradyrhizobium japonicum, and Rhizobium leguminosarum. The genes encoding the dimeric hydrogenases of these organisms have been identified, and the controlling promoters have been characterized. These studies revealed remarkable similarities. In all three cases, the hydrogenase genes are transcribed from ⁵⁴-dependent promoters under the control of NtrC-like transcriptional activators (7, 38, 52). Furthermore, in vivo and/or in vitro data document the role of IHF-like proteins in promoter activation (5, 7, 50).

Experiments with a heterologous system revealed that, outside the context of the A. eutrophus cell, HoxA is capable of mediating only weak activation of its cognate promoters P_SH and P_MBH. This low-level transcription may be an indication that other regulatory components and/or modification of HoxA are required for full activation.

This study represents the first stage of a systematic investigation of the regulation of the hydrogenase genes. A detailed molecular analysis of the two promoters identified here is now under way. Screening of the hydrogenase gene cluster has, to date, identified four additional promoters, which are presently being characterized (45). The ultimate goal of these studies is to understand the workings of a complex multigene system.