Molecular and Functional Characterization of theRhodopseudomonas palustris No. 7 Phosphoenolpyruvate Carboxykinase Gene

The bacterial strains and plasmids used in this study are listed in Table1.

All chemicals were of the highest available purity and were purchased from Wako Pure Chemical Industries (Osaka, Japan) or Sigma (St. Louis, Mo.).

E. coli strains were grown aerobically at 37°C in Luria-Bertani (LB) medium (36).R. palustris No. 7 was cultivated aerobically in the dark at 35°C in van Niel’s complex medium (45) or in synthetic minimal medium (PMM) without yeast extract (10) and supplemented with a given 20 mM carbon source. Anaerobic growth in the light was at 35°C in synthetic medium containing 20 mM carbon source. When appropriate, media were supplemented with antibiotics. Final concentrations for E. coli were (per milliliter) 50 μg of ampicillin, 50 μg of kanamycin, 20 μg of tetracycline, and 10 μg of gentamicin; for R. palustris No. 7, the final concentrations were (per milliliter) 300 μg of gentamicin, 100 μg of streptomycin, and 200 μg of kanamycin.

Plasmid DNA was isolated by the alkaline lysis procedure (36). Restriction endonucleases and T4 ligase were obtained from Takara (Kyoto, Japan) and used according to the manufacturer’s instructions. E. coli strains were transformed by the CaCl₂ method (36). R. palustris No. 7 was transformed by electroporation (7). Restriction fragments were isolated, when required, from agarose gels (1% [wt/vol]) with the Prep-a-gene matrix (Bio-Rad, Richmond, Calif.) according to the manufacturer’s instructions.

All PCRs were carried out in a Perkin-Elmer (Foster City, Calif.) DNA thermal cycler model 480 according to the manufacturer’s instructions. For amplification of the pckAgene of R. palustris No. 7, degenerate primers were synthesized. Primer 1 (5′-TCTCGGATCCYTIATIGGIGAYGAYGARCA-3′ [I, inosine; R, A or G; Y, C or T]) and primer 2 (5′-TCTCGGATCCGGYTCIGTIAYICCIYKYTCIGTICCIGC-3′ [K, G or T]) were made for the amino acid sequences of a highly conserved domain of PEPCK proteins, LIGDDEH and AGTE(R/Q/K)G(I/V)T, respectively. PCR fragments were purified with the QIAquick PCR purification kit (Qiagen, Hilden, Germany). The PCR fragments were cloned in the pGEM-T vector (Promega Biotech, Madison, Wis.).

pGEM-T clones were sequenced by the dideoxy chain termination method as described by Sanger et al. (38) with a model 373A DNA sequencer (Applied Biosystems–Perkin-Elmer, Foster City, Calif.). Suitable λFIXII clones were identified by hybridization with probes derived from pGEM-T clones. These clones were the sources of restriction fragments, which were then cloned into pUC118 and sequenced as above. The nucleotide sequences of both strands were determined. DNA sequence data were analyzed with the INHERIT programs (Applied Biosystems–Perkin-Elmer) and the Genetyx programs (Software Development, Tokyo, Japan).

Total RNA was extracted from R. palustris No. 7 cells as previously described (18). Primers were synthesized with 5′ ends at positions 620, 640, and 660, complementary to the mRNA sequence within the structural gene, which begins at sequence position 585. Figure 1A shows the experimental design. The primers were labelled with [γ-³²P]ATP (36). Primer extension reactions were assembled as described in Babst et al. (1). Primer extension products were separated on 6% DNA sequencing gels, and the sequence position was determined by alignment with an [α-³²P]dATP-labelled DNA sequencing ladder obtained from supercoiled pMG200 DNA, containing the pckA gene, by the dideoxy method (U.S. Biochemicals Sequenase protocol) and with the same primer. Sequencing reactions were performed in the presence of 5% dimethyl sulfoxide. The gels were cut in half, dried, and exposed in two separate Fujix Image cassettes. Autoradiography was performed with a Fujix BAS2000 image analyzer system (Fuji, Tokyo, Japan).

β-Galactosidase was expressed as a hybrid pckA′-′lacZfusion protein, using the pckA transcriptional and translational signals. The protein fusion was constructed by inserting the 1.9-kb SmaI-HincII fragment (Fig.1A) encompassing the pckApromoter region into the pMC1871 SmaI site. Pale blueE. coli transformants harbored the recombinant plasmid with the insert in the correct orientation. Subsequently, thepckA-lacZ fusion was subcloned as a 5.0-kb SalI fragment into the SalI site of shuttle vector pMG102 (17). The resulting plasmid, pMG202, was introduced intoR. palustris No. 7 by electroporation. Transformants appeared as blue clones on van Niel’s agar plates containing 200 μg of kanamycin/ml and 40 μg of X-Gal (5-bromo-4-chloro-β-d-galactopyranoside)/ml.

A 1.3-kb HincII kanamycin cassette from pUC4K was inserted into the unique CpoI site within thepckA gene coding region carried on pMG200, which had been blunt ended with a Takara DNA-blunting kit according to the manufacturer’s instructions. The 2.8-kb KpnI PCR-amplifiedpckA::Km fragment from this construct was further subcloned into the KpnI site of pMG300, which is a gentamicin derivative of the suicide vector pGP704 (18), yielding pMG312. This plasmid can replicate only in a strain that produces the R6K-specified π protein and was therefore unable to replicate in R. palustris No. 7, which does not produce π. Plasmid pMG312 was transferred from E. coli SM10 λpir, which can produce π, to R. palustrisNo. 7 by conjugation. Diparental matings between E. coliSM10 λpir(pMG312) and R. palustris No. 7 were performed by using a filter mating technique as described before (18). Both the absence of vector sequences and the presence of pckA::Km cassette sequence were confirmed for several candidates by Southern hybridization analysis. Similarly, aR. palustris No. 7 pckA ppc double mutant was constructed by introducing a copy of the pckA gene disrupted by a tetracycline resistance cassette into a ppc mutant (18).

A crude enzyme extract was prepared as follows. Exponentially growing cells (25 ml) were harvested by centrifugation (10 min; 8,000 × g ; 4°C) and washed with STE buffer (100 mM NaCl, 10 mM Tris-HCl [pH 8.0], 1 mM EDTA). The cells were resuspended in 1 ml of the same buffer solution and disrupted by sonication (Astrason model XL2020) in an ice water bath for three 2-min periods, interrupted by 2-min cooling intervals. Cell debris was removed by centrifugation (20 min; 10,000 × g; 4°C), and the supernatant was used as a crude extract for enzyme assays.

Protein concentrations were measured with a Bio-Rad protein assay kit in either the standard or detergent-compatible format as appropriate, with bovine serum albumin as a reference standard.

PEPC, PEPCK, MAEA, and MAEB activities were determined as described previously (18, 19).

Pyruvate orthophosphate dikinase (PPDK) and PPS activities were measured as described by Østerås et al. (29), except that imidazole (pH 6.6) was substituted for Tris-HCl (pH 8.5) to obtain optimum enzyme activities.

β-Galactosidase assays were performed according to the method of Miller (26). A β-galactosidase unit is defined as 1 nmol of o-nitrophenyl-β-d-galactoside hydrolyzed per min. The data represent averages of at least three experiments.

The sequence data reported here have been deposited in the DDBJ-EMBL-GenBank database under accession no. AB015618 .

Based on the observation that amino acid sequences of known PEPCK proteins share several highly conserved domains, two oligonucleotide primers were designed in order to amplify a portion of the pckA gene from R. palustris No. 7. The degenerate oligonucleotide primers (primer 1 and primer 2 [see Materials and Methods]) were synthesized and used in a PCR withR. palustris No. 7 chromosomal DNA as a template. The 0.4-kb PCR product was subsequently cloned into the pGEM-T vector and sequenced. The resulting amino acid sequence appeared to be greater than 58% homologous to parts of the pckA gene products ofRhizobium meliloti (28), Saccharomyces cerevisiae (21), Staphylococcus aureus(41), E. coli (25), and T. cruzi (22).

The PCR fragment was labelled with α-³²P and used as a probe to screen a λFixII R. palustris No. 7 library (18). Screening of 10⁴ plaques yielded 13 positive clones. Restriction analysis demonstrated that all of the positive clones had overlapping inserts and contained a common 3.5-kbPstI fragment which hybridized to the PCR fragment (data not shown). Southern blotting of R. palustris No. 7PstI-digested chromosomal DNA probed with the same labelled PCR fragment DNA demonstrated that the 3.5-kb PstI fragment had not undergone any major deletions or rearrangements during cloning (data not shown) and was present in only a single copy.

A 2.5-kb SmaI-PstI fragment derived from the 3.5-kb PstI fragment was isolated and subcloned into plasmid pUC118 (Fig. 1A). The nucleotide sequence of this fragment and the deduced amino acid sequence showed homology to known pckA genes and polypeptides.

The coding region containing the R. palustris No. 7pckA gene consists of an open reading frame of 1,614 nucleotides, corresponding to 537 amino acids. The NH₂-terminal amino acid sequence of the purified PEPCK protein from R. palustris was determined to be MQETGVHNGAYG. This result corresponds to the deduced NH₂-terminal amino acid sequence determined by DNA sequence analysis, and GTG at position 585 was identified as the start codon of the pckA gene (Fig. 1B). At 9 bp upstream of the start codon there is a potential ribosome binding site, AGGAGG (Fig.1B). An inverted repeat downstream of the stop codon (TAA), centered at position 2237, showed a G+C-rich stem-loop structure followed by five T residues (Fig. 1D), the characteristics of a rho-independent terminator (34). The G+C content of the R. palustris No. 7 pckA coding region was high (64.6%), and the third positions of pckA gene codons showed an extreme preference (89.0%) for G or C, as expected in genes of PNSB (16). The deduced amino acid sequence is 68.2, 50.4, 46.2, 46.0, and 44.3% identical to that of the pckA gene products from R. meliloti, S. aureus, S. cerevisiae, T. cruzi, and E. coli, respectively. Comparative amino acid sequence alignment of ATP-dependent PEPCKs shows that all of the putative functional regions are conserved among this group of enzymes (Fig.2). The PEPCK-specific domain (22), which is conserved in all ATP- and GTP-dependent PEPCKs analyzed to date, is also found in the R. palustrisNo. 7 enzyme (residues 192 to 201). In addition, kinase-1a (phosphate and Mg²⁺ binding site), kinase-2 (Mg²⁺ binding site) (23), and the covalent metal ion binding sites (22) are essentially identical among all ATP-dependent PEPCKs. On the other hand, more variability was present in the adenine binding site (43), although the key residues, arginine and threonine (R₄₄₄ and T₄₅₀ in R. palustris No. 7 PEPCK), which in E. coli interact with the adenine base of ATP, are identical among these sequences.

The functionality of the R. palustris No. 7pckA gene was tested by complementation of the E. coli pckA maeA maeB mutant strain 1321, which cannot grow on minimum medium containing succinate as a sole carbon source (14). Plasmid pMG200, which contains the R. palustris No. 7pckA gene, was able to restore growth of an E. coli pckA maeA maeB mutant on minimum medium containing 0.4% succinate. This observation indicates that the R. palustrisNo. 7 PEPCK can function as a gluconeogenic enzyme in E. coli and suggests that the pckA promoter might function in E. coli.

To investigate the possibility of multicopy enhancement of PEPCK activity in E. coli and R. palustris No. 7, PEPCK assays were performed with crude extracts from E. coli andR. palustris No. 7 cells bearing either pMG201, which contains the R. palustris No. 7 pckA gene inserted into the shuttle vector pMG102, or vector alone. Plasmid pMG102 has ColE1 and Rhodopseudomonas sp. origins and has a copy number of about five molecules per cell in R. palustrisNo. 7 (17). Both E. coli and R. palustris No. 7 harboring pMG201 exhibited a PEPCK activity four times greater than that of the control strain, and sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis of pMG201-bearing cells showed enhanced production of a polypeptide of approximately 60 kDa in both strains (data not shown), in agreement with the protein size predicted by sequence analysis.

In order to identify the pckA promoter, the transcriptional start site of the pckA gene was determined by primer extension with 20-mer oligonucleotides (P1, P2, and P3) having 5′ ends at positions 620, 640, and 660, complementary to the coding region. Template RNA was extracted from a R. palustris No. 7 culture in late exponential phase. One major extension product was observed with all three primers, corresponding to a 5′ end at +72 bases upstream of translation initiation codon. The result with the primer at position 620 is shown in Fig. 1C.

Putative −10 and −35 promoter sequences, similar to the E. coli ς⁷⁰ promoter consensus (15) separated by a typical 18-nucleotide interregion spacing, were identified upstream of the transcriptional start point (Fig. 1B). This result could explain the multicopy enhancement of PEPCK activity observed with plasmid pMG201 in the E. coli host.

To facilitate study of the regulation of expression of the pckA gene, we constructed plasmid pMG202, in which lacZ expression is controlled by the transcriptional and translational regulatory signals of the pckA gene. The effect of growth conditions onpckA expression was studied by measuring β-galactosidase levels in R. palustris No. 7 harboring pMG202.

When cells were grown in PMM with gluconeogenic carbon sources, such as succinate (Fig. 3A and E) or malate (data not shown), under either anaerobic light or aerobic dark conditions,pckA expression was strongly induced during log phase and was 10- to 20-fold reduced at the onset of stationary phase. In cells growing with pyruvate (Fig. 3B and F) or lactate (data not shown), either anaerobically in the light or aerobically in the dark, log-phase inductions were also observed; however the magnitude of induction was lower for growth on pyruvate. Under anaerobic light conditions with ethanol and CO₂ (Fig. 3C), which is a CO₂-fixing condition in R. palustris No. 7 (2), or aerobic dark conditions with ethanol (Fig. 3G),pckA expression was also specifically induced at log phase. In addition, expression gradually increased again during stationary phase under anaerobic light conditions whereas no stationary-phase expression was observed under aerobic dark conditions. These data suggest that log-phase induction of expression of the R. palustris pckA gene does not depend on the carbon source.

Regulation of pckA gene expression was also studied during growth in van Niel’s complex medium. Cells grown in van Niel’s medium under anaerobic light (Fig. 3D) and aerobic dark (Fig. 3H) conditions showed the same log-phase induction of the pckA gene as in minimal media with other carbon sources, whereas in cells grown aerobically in the dark, pckA expression was reduced by half at the onset of stationary phase and gradually increased again during stationary phase. Cells grown anaerobically in the light with ethanol and CO₂ showed a weaker stationary-phase induction.

To confirm that the growth phase induction observed represented induction of the pckA gene, we measured PEPCK activity levels in wild-type R. palustris. When cells were grown in PMM with pyruvate under either anaerobic light (Fig.4A) or aerobic dark (Fig. 4C) conditions, log-phase inductions of PEPCK were observed with magnitudes similar to the β-galactosidase levels of the pckA-lacZ fusion, and the temporal expression patterns of PEPCK and β-galactosidase fusion protein activities corresponded well.

Cells grown in van Niel’s medium either anaerobically in the light (Fig. 4B) or aerobically in the dark (Fig. 4D) exhibited the same log-phase induction of PEPCK seen for the pckA-lacZ fusion experiments. However, in cells grown aerobically in the dark, PEPCK activity was gradually reduced during stationary phase, and a stationary-phase induction which corresponded to β-galactosidase levels in the pckA-lacZ fusion was not observed. These differences are considerable and suggest that PEPCK enzyme may be differentially unstable due to proteases expressed during stationary phase. High-level expression of the β-galactosidase fusion during stationary phase might represent increased transcription to compensate for PEPCK protein degradation. However, we have not examined PEPCK protein turnover under different growth conditions.

Thus, the log-phase induction of the pckA-lacZ fusion resembles the induction of the PEPCK enzyme seen in wild-type R. palustris and suggests that PEPCK is regulated at the transcriptional level.

In addition, cells grown anaerobically in the light with pyruvate (Fig.3B) showed three- to fivefold-higher activities than those grown with succinate (Fig. 3A) and ethanol-CO₂ (Fig. 3C), respectively. Furthermore, pyruvate-grown cells attained a higher final cell density (optical density at 660 nm [OD₆₆₀], 5.8) and showed a faster doubling time (5 h) than cells grown on succinate (final OD₆₆₀, 2.2; doubling time, 11 h) and ethanol-CO₂ (final OD₆₆₀, 3.5; doubling time, 14 h), suggesting that pckA gene expression may also depend on the cell growth rate.

To study the role of PEPCK in R. palustrisNo. 7 carbon metabolism, we constructed pckA mutants andpckA ppc double mutants. Four clones were identified aspckA mutants that had lost the vector mediating gentamicin resistance (out of 700 Km^r transconjugants). Similarly, 15 clones were isolated as pckA ppc double mutants from 800 Tc^r transconjugants. The integration of an inactivated copy of the pckA gene via a double-crossover event was confirmed by Southern hybridization experiments (data not shown). The PEPCK-specific activities in cell extracts of one of each of the recombinant strains and the parental strain were determined (Fig.5). The recombinant strains showed no detectable PEPCK activity, indicating that the pckA gene was inactivated.

The effect of PEPCK deficiency on the growth of R. palustrisNo. 7 was analyzed. We compared the growth of the pckAmutant strain with that of the wild-type R. palustris No. 7 in synthetic PMM under both aerobic dark and anaerobic light conditions with various carbon sources. When the cells were grown with succinate, malate, pyruvate, lactate, or ethanol, either anaerobically in the light or aerobically in the dark, the growth of the pckAmutant showed no significant difference from that of the wild-type strain (data not shown). We also measured MAEA-, MAEB-, PPS-, and PPDK-specific activities in both the wild-type and pckAmutant strains (Fig. 5). A schematic representation of the metabolic pathways and enzymic reactions also referred to in this paper is shown in Fig. 5. Although MAEB activity was undetectable, both strains show similarly high levels of MAEA activity, which can synthesize pyruvate from malate. Moreover, both PPS and PPDK activities were detected in both strains. Combined activities of MAEA, PPS, and PPDK, therefore, could contribute to the growth of the pckA mutant. This data and the comparative growth experiments indicate that R. palustris, like E. coli, has alternative gluconeogenic pathways and that PEPCK is dispensable as a gluconeogenic enzyme in this organism.

In order to determine if PEPCK played a role as an anaplerotic enzyme, we also compared the growth of the pckA ppc double-mutant strain with those of the wild type, the ppc mutant, and thepckA mutant in PMM under both aerobic dark and anaerobic light conditions with pyruvate. In cells grown with pyruvate aerobically in the dark, a R. palustris ppc mutant showed a doubling time slower than that of the wild-type strain, as described before (18), while the ppc mutant and pckA ppc double mutant showed identical growth. These data imply that PEPC may play an important role in the anaplerotic pathway but that PEPCK does not function similarly and that other anaplerotic enzymes besides PEPC may exist (Fig. 6A).

When grown anaerobically in the light, the ppc mutant differs from the parental strain in exhibiting slower growth, as previously described (18). However, six independentpckA ppc double mutants showed longer lag times before starting to grow and slower growth than the ppc mutant, while the wild type and the pckA mutant showed only a negligible difference in growth (Fig. 6B). Cells grown with lactate also showed the same growth characteristics compared with pyruvate-grown cells (data not shown). With other carbon sources, such as succinate, malate, or ethanol and CO₂, these mutants showed no significant difference from the wild-type strain (data not shown). These results suggest that PEPCK can function weakly as an anaplerotic enzyme in the absence of PEPC under anaerobic light conditions and that residual anaplerotic enzyme activities enable aR. palustris No. 7 pckA ppc double mutant to survive.