Oxygen-Mediated Regulation of Porphobilinogen Formation in Rhodobacter capsulatus

Abstract

A Rhodobacter capsulatus hemC mutant has been isolated and used to show that oxygen regulates the intracellular levels of porphobilinogen. Experiments using a hemB-cat gene fusion demonstrated that oxygen does not transcriptionally regulate hemB transcription. Porphobilinogen synthase activity is not regulated by oxygen nor is the enzyme feedback inhibited by hemin or protoporphyrin IX. It was demonstrated that less than 20% of [¹⁴C]aminolevulinate was incorporated into bacteriochlorophyll, suggesting that the majority of the aminolevulinate is diverted from the common tetrapyrrole pathway. Porphobilinogen oxygenase activity was not observed in this organism; however, an NADPH-linked aminolevulinate dehydrogenase activity was demonstrated. The specific activity of this enzyme increased with increasing oxygen tension. The results presented here suggest that carbon flow over the common tetrapyrrole pathway is regulated by a combination of feedback inhibition of aminolevulinate synthase and diversion of aminolevulinate from the pathway by aminolevulinate dehydrogenase.

It has been almost a half-century since the work of Cohen-Bazire et al. (13) demonstrated that oxygen regulates bacteriochlorophyll synthesis in the purple nonsulfur photosynthetic bacteria. However, the mechanism by which oxygen regulates this pathway has been difficult to elucidate. For Rhodobacter capsulatus, it has been shown that a bchH mutant accumulates much more protoporphyrin IX when grown under low oxygen tension than under high oxygen tension (6), indicating that oxygen regulates some step in the common tetrapyrrole pathway (Fig. 1). Rebeiz and Lascelles (39) proposed that oxygen inhibits the chelation of magnesium into the protoporphyrin ring, the first step in the branch leading to bacteriochlorophyll. The accumulating protoporphyrin IX would be diverted to heme, which would feedback inhibit aminolevulinate synthase. Inhibition of this enzyme would reduce carbon-flow over the common tetrapyrrole pathway.

FIG. 1.

The common tetrapyrrole pathway in R. capsulatus. Dashed lines denote multistep pathways.

This theory was tested in R. capsulatus by growing a bchH mutant (AJB456) under high and low oxygen concentrations in the presence of exogenous aminolevulinate (5). Surprisingly, protoporphyrin IX accumulation was still regulated by oxygen. However, when exogenous porphobilinogen, the second intermediate in the pathway, was added to a culture of the bchH mutant, protoporphyrin IX accumulation was no longer regulated by oxygen (5). This observation indicated that the major oxygen-regulated control point in the common tetrapyrrole pathway is not the synthesis of aminolevulinate. Rather, it appears that oxygen regulates the second step in the pathway, the conversion of aminolevulinate to porphobilinogen. This is reminiscent of the situation in Bradyrhizobium japonicum, where transcription of the hemB gene is regulated by oxygen (10).

The lack of a hemC mutant has made it difficult to confirm and extend these results. The most straightforward way to determine whether oxygen regulates porphobilinogen formation in R. capsulatus would be to measure porphobilinogen levels in a hemC mutant. Such a mutant would lack porpobilinogen deaminase and would therefore accumulate porphobilinogen. In this study, the construction and characterization of a hemC mutant is detailed. Using this mutant, it was determined that the intracellular porphobilinogen level is regulated by oxygen tension.

Since the situation in R. capsulatus appears to be similar to that of B. japonicum (10), transcriptional regulation of the hemB gene was investigated. We report here that hemB transcription is not regulated in response to oxygen tension nor is the activity of porphobilinogen synthase feedback inhibited by either hemin or protoporphyrin IX. We did observe, however, that intracellular hemin levels appear to influence carbon flow over the common tetrapyrrole pathway. Additionally, R. capsulatus has a novel enzyme, aminolevulinate dehydrogenase, which reduces aminolevulinate to aminohydroxyvalerate. The regulation of porphobilinogen levels appears to be mediated by a combination of feedback inhibition of aminolevulinate synthase by hemin and by diverting aminolevulinate from the common tetrapyrrole pathway.

MATERIALS AND METHODS

Strains, growth media, and reagents.

Experiments were performed on R. capsulatus strain PAS100 (45) and on the derivative strains AJB456 [Φ(bchH"-lacZ⁺)700 crtD233 hsd-1 str-2] and AJB503 [Φ(bchD"-lacZ⁺)713 crtD233 hsd-1 str-2] constructed by Mu d1 insertion (6). R. capsulatus cells were routinely grown either in a malate-minimal salts medium (RCV; 48) or in 0.3% Difco Bacto-Peptone-0.3% Difco yeast extract (PYE). Amino acids, when used, were added to yield a final concentration of 0.5 mM. Antibiotics, when used, were added to yield the following final concentrations: kanamycin, 10 μg/ml, and streptomycin, 75 μg/ml. When solid media were employed, 1.5% agar (Difco) was added to the above media. When hemin was included in media, it was dissolved in a minimal amount of 1 N NaOH and added to the medium to a final concentration of 0.5 mM. Aminohydroxyvalerate was synthesized by reduction of aminolevulinate with sodium borohydride as previously described (20).

Measurements.

Crude extracts were made by harvesting cells and suspending them in 1 ml of the appropriate buffer per 100 ml of culture. The cell suspension was sonicated six times for 20 s with 30-s cooling periods between bursts. The suspension was centrifuged at 27,000 × g for 20 min. Chloramphenicol acetyltransferase activity was measured as previously described (40). Aminolevulinate synthase activity was measured as previously described (9). Aminolevulinate dehydrogenase activity was measured by following the oxidation of NADPH at 340 nm. The reaction mixture contained 100 μmol of Tris-Cl, pH 7.5, 5 μmol of aminolevulinate, and 0.1 μmol of NADPH in a volume of 1.0 ml. The reaction was carried out in a dual beam spectrophotometer with water instead of aminolevulinate in the reference cuvette. A molar extinction coefficient of 6.22 × 10³ was used and specific activity was determined as nanomoles of NADPH oxidized per minute per milligram of protein. Porphobilinogen synthase activity was measured as previously described (41). Porphobilinogen deaminase activity was measured as previously described (28). Porphobilinogen concentration was determined in sonic extracts by the method described by Shemin (41). The protein concentration of sonic extracts was determined by the dye binding method described by Bradford (7). Bacteriochlorophyll concentration was determined by harvesting a 1-ml sample and extracting the pellet with acetone-methanol (7:2), as described by Cohen-Bazire et al. (13). A millimolar extinction coefficient of 76 (770 nm) was used (11). Protoporphyrin IX was extracted from a 1-ml sample of culture with ethyl acetate-acetic acid (3:1). The porphyrin was then extracted from the organic layer into 3 N HCl. The protoporphyrin IX concentration was determined by comparing the fluorescence (excitation wavelength = 407 nm, emission wavelength = 605 nm) with a standard curve developed using commercially obtained protoporphyrin IX (Porphyrin Products, Logan, Utah). The protein concentration of cells extracted with acetone-methanol was determined as previously described (4). Due to day-to-day variations in porphyrin levels of the inoculum and growth conditions, all measurements reported are representative values from several trials.

Protein characterization.

Determination of molecular weight using electrophoresis under nondenaturing conditions was as previously described (8, 14). Porphobilinogen deaminase activity was localized within the gel by incubation with 0.1 mg of porphobilinogen per ml for 30 min at 37°C with shaking. The hydroxymethylbilane formed was converted to uroporphyrin I by incubation for 20 min in the dark in 1 N HCl-0.01% iodine. Sodium metabisulfite (0.1%) was added to decolorize the iodine, and the gel was viewed under shortwave UV light. Two-dimensional gel electrophoresis was as previously described (37). Proteins separated by two-dimensional gel electrophoresis were blotted to an Immobilon-P^SQ membrane (Millipore Corp., Bedford, Mass.), and the N-terminal amino acid sequence was determined by the Protein Chemistry Core Facility, Baylor College of Medicine (Houston, Tex.).

DNA manipulations and sequencing.

DNA sequence was determined using the CircumVent Kit (New England Biolabs, Inc., Beverly, Mass.). R. capsulatus chromosomal DNA was isolated using the Puregene DNA Isolation Kit (Gentra Systems, Inc., Minneapolis, Minn.). Procedures associated with Southern analysis were as previously described (50).

Nucleotide sequence accession number.

The DNA sequence presented here is accessible from the GenBank database under accession number U16796.

RESULTS

Purification of porphobilinogen deaminase.

The hemC mutant was isolated by first purifying porphobilinogen deaminase, using the amino acid sequence to clone the hemC gene and inserting a kanamycin resistance cartridge into that gene, followed by recombination of the mutated gene into the R. capsulatus chromosome. As a first step in cloning the hemC gene, the R. capsulatus porphobilinogen deaminase was purified following a procedure similar to that used for the purification of this enzyme from R. sphaeroides (28). PAS100 cells from 8 liters of culture grown in PYE medium were harvested and suspended in 0.1 M Tris-Cl, pH 8.0. A crude extract was prepared from these cells and brought to 45% saturation with ammonium sulfate. The precipitate was removed by centrifugation, and the supernatant was brought to 60% saturation with ammonium sulfate. After centrifugation, the pellet was resuspended in 0.1 M Tris-Cl (pH 8.0)-1 mM DTT-1 mM EDTA, dialyzed, and heated for 10 min at 60°C. After removal of denatured proteins by centrifugation, the suspension was subjected to chromatography through DEAE-cellulose, hydroxylapatite, and Sephadex G-75. This protocol resulted in 138-fold purification (Table 1). Nondenaturing polyacrylamide gel electrophoresis revealed the presence of one major band (36 ± 2 kDa) and three minor bands (data not shown). The gel was incubated with porphobilinogen, and the uroporphyrin I produced was viewed under short-wave UV light. Only the major band produced uroporphyrin I (data not shown). The R. capsulatus porphobilinogen deaminase was separated from the other proteins remaining in the sample by two-dimentional gel electrophoresis and blotted onto Immobilon-P^SQ. The sequence of the first 20 amino acids was determined and found to be identical to a translation of an open reading frame located upstream of the R. capsulatus hemE gene, which we had previously cloned and sequenced (27).

TABLE 1.

Purification of porphobilinogen deaminase

Procedure	Volume (ml)	Total protein (mg)	Sp acta	Fold purification	Yield (%)
Crude extract	102	1,050	193	1.0	100
(NH4)2SO4	21.5	323	465	2.4	74
Heated	20.0	276	428	2.2	55
DEAE-cellulose	25.0	60.0	1,470	7.6	43
Hydroxylapatite	12.5	13.8	4,730	25	32
Sephadex G-75	5.0	1.93	26,700	138	25

^aPorphobilinogen synthase specific activity is expressed as nanomoles of porphobilinogen formed/hour/milligram of protein.

Sequencing of hemC gene.

In the process of cloning the hemE gene, plasmid pCAP154, which contains 1.2 kb of R. capsulatus DNA upstream of the hemE gene, was isolated (Ineichen and Biel, unpublished data). The nucleotide sequence of the R. capsulatus DNA from this plasmid was determined (GenBank accession number U16796). A 951-bp open reading frame was found which would encode a protein with a molecular mass of 34 kDa. The open reading frame and the hemE gene are divergently transcribed. The derived amino acid sequence of the R. capsulatus enzyme shows 47% identity with E. coli, 47% identity with Pseudomonas aeruginosa, and 43% identity with B. subtilis porphobilinogen deaminases (1, 23, 34). The consensus porphobilinogen deaminase cofactor binding site is completely conserved in this sequence (33). This strongly suggested that the open reading frame was the hemC coding region.

Expression of the hemC gene in Escherichia coli.

In order to verify that this sequence coded for a functional porphobilinogen deaminase, the region was cloned into E. coli. A 1.1-kb HindIII-BamHI fragment beginning 38 bp upstream of the hemC start codon was cloned into pUC19 (47) so that transcription from P_lac on the plasmid would extend into the hemC gene. This plasmid, pCAP178, was electroporated into E. coli strain NM522 (22), and porphobilinogen deaminase activity was determined. As a control, enzyme activity was also determined for NM522(pUC19). Both strains were grown with 5 μg of IPTG/ml to stimulate transcription from P_lac. The porphobilinogen deaminase activity in NM522(pCAP178) was fivefold higher than the activity in NM522(pUC19) (specific activities of 49 and 10, respectively). These results confirm that the open reading frame is indeed the hemC gene.

Isolation of a R. capsulatus hemC mutant.

The cloning of the hemC gene provided us with a straightforward method of isolating a hemC mutant. The 4.0-kb BamHI fragment containing the hemC gene was inserted into the vector pSUP202 (44). This vector can be mobilized into R. capsulatus but will not replicate there (44). The 1.3-kb kanamycin cartridge from pUC-4K (47) was inserted into the StuI site within the hemC gene (nucleotides 282 to 287). The resulting plasmid, pCAP181, contains approximately 0.9 kb of R. capsulatus DNA on one side of the kanamycin cartridge and 3.1 kb on the other side (Fig. 2).

FIG. 2.

Diagram of pCAP181. B, BamHI; H, HincII; S, StuI.

The plasmid was mobilized into R. capsulatus strain PAS100. Kanamycin-resistant colonies were selected on media containing hemin, cysteine, methionine, and kanamycin. A kanamycin-resistant, hemin-requiring strain, designated AJB565, was isolated and further characterized. A hemC mutant should be unable to synthesize siroheme or vitamin B₁₂ and therefore should require cysteine and methionine. As expected, AJB565 required both of these amino acids in addition to hemin. The hemC mutant will grow on solid media containing hemin but will not grow in liquid media containing hemin. This is reminiscent of the hemA mutant, AJB529, which shows very limited growth in broth containing hemin (49).

Physical characterization of AJB565.

In order to confirm that the kanamycin cartridge had recombined into the chromosome within the hemC gene, Southern blot analyses were performed. Blots were made with EcoRI-digested chromosomal DNA from strains PAS100 and AJB565. One blot was probed with pUC-4K, which contained the kanamycin cartridge. Plasmid pUC-4K hybridized to a 9.0-kb fragment of AJB565 DNA but did not hybridize to PAS100 DNA (Fig. 3, lanes 3 and 4), indicating that the 1.3-kb kanamycin cartridge was inserted into a chromosomal EcoRI fragment of approximately 7.7 kb. A second blot was probed with pCAP168, which contains a 0.5-kb HincII-EcoRI fragment containing the first half of the hemC gene (nucleotides 111 to 632). This probe hybridized to a 7.5-kb fragment of PAS100 DNA and to a 9.0-kb fragment of AJB565 DNA (Fig. 3, lanes 5 and 6), confirming the insertion of the kanamycin cartridge into the hemC gene of AJB565. No hybridization occurred in a third blot, probed with pSUP202, demonstrating that the entire plasmid had not inserted into the chromosome of AJB565 (Fig. 3, lanes 1 and 2).

FIG. 3.

Southern blots of strain PAS100 and AJB565 DNAs. Chromosomal DNAs were digested with EcoRI. Lane 1, PAS100 DNA probed with pSUP202; lane 2, AJB565 DNA probed with pSUP202; lane 3, PAS100 DNA probed with pUC-4K, which carries the kanamycin cartridge; lane 4, AJB565 DNA probed with pUC-4K; lane 5, PAS100 DNA probed with pCAP168, which carries the first half of the hemC gene; lane 6, AJB565 DNA probed with pCAP168.

Bacteriochlorophyll and enzymatic activities in AJB565.

The level of bacteriochlorophyll in the hemC mutant AJB565 was compared with the levels in the parental strain and in the hemA mutant AJB529 (Table 2). Since AJB565 will not grow in liquid media, all measurements were performed on cultures grown on RCV-hemin agar plates. The bacteriochlorophyll level in cultures of strain PAS100 grown on agar plates was similar to the level observed in broth-grown cultures (50). The level of bacteriochlorophyll in AJB529 was approximately 10% of the level in the parental strain PAS100, confirming the results obtained with broth-grown cultures (50). As expected, AJB565 has no detectable bacteriochlorophyll.

TABLE 2.

Bacteriochlorophyll and porphobilinogen levels and enzymatic activities

Strain	Media	Growth conditiona	Bchlb	PBGc	ALASd	PBGSe	PBGDf
PAS100 (hem+)	Broth	Aerobic			74	220	130
	Agar	Aerobic	0.28		78	220	110
	Agar	Low oxygen	7.8
AJB529 (hemA)	Broth	Aerobic			9.0	210	160
	Agar	Aerobic	0.40		9.1	180	120
AJB565 (hemC)	Agar	Aerobic	NDg	1.5	91	210	ND
	Agar	Low oxygen		29

^aAerobic growth in broth was obtained by sparging the cultures with compressed air (6). Aerobic and low oxygen growth on agar were as described in the text.

^bBacteriochlorophyll concentration expressed as nanomoles/milligram of protein.

^cPorphobilinogen concentration expressed as nanomoles/milligram of protein.

^dAminolevulinate synthase specific activity expressed as nanomoles of aminolevulinate formed/hour/milligram of protein.

^ePorphobilinogen synthase specific activity expressed as nanomoles of porphobilinogen formed/hour/milligram of protein.

^fPorphobilinogen deaminase specific activity expressed as nanomoles of porphobilinogen consumed/hour/milligram of protein.

^gND, not detectable. Minimum detectable levels are 0.01 nmol of bacteriochlorophyll/mg of protein and 4.0 nmol of porphobilinogen consumed/h/mg of protein.

The results shown in Table 2 demonstrate that growth in broth versus agar plates does not alter the activities of any of the enzymes measured in strains PAS100 and AJB529. Aminolevulinate synthase activity in AJB529 was approximately 10% of that in PAS100, while the activities of the other two enzymes were the same in the two strains. The residual aminolevulinate synthase activity in AJB529 is most likely due to a small amount of transcription initiating within the Tn 5 in this strain (49). While R. sphaeroides has two hemA isozymes (hemA and hemT) (35), multiple studies suggest that R. capsulatus has only a single hemA isozyme (17, 25, 50).

Strain AJB565 had normal levels of aminolevulinate synthase and porphobilinogen synthase, but had no detectable porphobilinogen deaminase activity. These results confirm that the kanamycin cartridge interrupted and inactivated the hemC gene. The observation that the specific activity of the first two enzymes in the pathway are normal in this strain indicates that this mutation does not have pleotropic effects on the pathway.

Regulation of porphobilinogen formation in AJB565.

Since the hemC mutant cannot be grown in liquid media, experiments to determine whether porphobilinogen formation is regulated by oxygen had to be performed on solid media. In performing the experiments described above, we noted that when enough cells were spread onto the agar surface to form a lawn, very little bacteriochlorophyll was produced. If fewer cells were spread onto the agar surface so that individual colonies formed, the centers of the colonies were highly pigmented. This observation allowed us to develop a method for growing cultures under high and low oxygen tensions on solid media. Growth under low oxygen tension was obtained by spreading approximately 10⁶ cells on an agar slant in a 3- by 12-cm tube. The tube was flushed for 30 min with a mixture of 92% N₂, 5% CO₂, and 3% O₂ and capped with a rubber stopper. Aerobic growth was obtained by spreading the same number of cells either on a slant capped with a foam stopper or on an agar plate. Cultures were grown for five to six days. Using this protocol, the bacteriochlorophyll concentration in PAS100 grown under low oxygen tension was 28-fold higher than it was when the strain was grown under high oxygen tension (Table 2). These results demonstrated that this procedure could be used to measure differences in porphyrin levels in cultures grown under high and low oxygen tensions on solid media.

The experiment was repeated with the hemC mutant AJB565, which accumulates porphobilinogen. The porphobilinogen concentration in strain AJB565 grown under low oxygen tension was 19-fold higher than in the high-oxygen tension culture (Table 2). This confirms previous results (5) and directly demonstrates that oxygen regulates porphobilinogen accumulation in R. capsulatus.

Transcriptional regulation of the hemB gene.

Regulation of hemB transcription by oxygen was examined using a hemB-cat transcriptional fusion. The internal EcoRV fragment of the hemB gene (GenBank accession number U14593; nucleotides 585 to 789) (14) was replaced with the cat gene coding region from pCM7 (12). This placed the cat gene under the control of the hemB promoter. This construct was cloned into the PstI site of the mobilizable plasmid pRK404 (15). To prevent transcriptional interference from the lac promoter on pRK404, the omega cartridge (38) was cloned into the HindIII site immediately upstream of the hemB-cat fusion. This plasmid was designated pCAP148. A second plasmid was constructed as a control for differences in plasmid copy number and supercoiling. This plasmid, pCAP153, has the cat gene fused to the lac promoter of pRK404.

After moving the plasmids into R. capsulatus, cultures were grown under high and low oxygen tensions as previously described (6). The chloramphenicol acetyltransferase activity in the control strain containing pCAP153 was 1.4-fold higher when the strain was grown under low oxygen tension than when it was grown under high oxygen tension (specific activities of 46 and 33, respectively). Chloramphenicol acetyltransferase activity in the strain containing pCAP148 was also 1.4-fold higher when grown under low oxygen tension than when grown under high oxygen tension (specific activities of 23 and 16, respectively). These results indicate that oxygen does not regulate transcription of the hemB gene.

Regulation of porphobilinogen synthase activity.

While oxygen does not regulate hemB transcription, it is possible that the activity of porphobilinogen synthase is regulated either directly or indirectly by oxygen. The specific activity of porphobilinogen synthase was measured in cultures of PAS100 grown aerobically and photosynthetically (Table 3). While the level of bacteriochlorophyll was over 14-fold higher in the photosynthetically grown culture, porphobilinogen synthase specific activity was unchanged. This enzyme was not inhibited by either hemin or protoporphyrin IX.

TABLE 3.

Bacteriochlorophyll levels and enzymatic activities in PAS100

Strain	Growth conditiona	Additionb	Bchlc	PBGSd	ALADe
PAS100 (hem+)	PS		26	158	29
	Aerobic		1.8	160	117
	PS	Hemin		163
	PS	Proto		184

^aAerobic growth was obtained by sparging broth cultures with compressed air (6). Photosynthetic (PS) cultures were grown in bottles completely filled with media and illuminated using 100-W incandescent bulbs.

^bPorphyrins were added to the reaction mixtures to a final concentration of 100 μM. Proto, protoporphyrin IX.

^cBacteriochlorophyll concentration expressed as nanomoles/milligram of protein.

^dPorphobilinogen synthase specific activity expressed as nanomoles of porphobilinogen formed/hour/milligram of protein.

^eAminolevulinate dehydrogenase specific activity expressed as nanomoles of NADPH oxidized/minute/milligram of protein.

Influence of heme on carbon flow over common tetrapyrrole pathway.

It was previously demonstrated that increasing the level of ferrochelatase in a strain reduced the specific activity of aminolevulinate synthase in that strain and resulted in lowered carbon flow over the common tetrapyrrole pathway (30). It was suggested that the increased ferrochelatase resulted in higher levels of heme which in turn feedback inhibited aminolevulinate synthase (30). If this supposition is correct, then bchD and bchH mutants, which are unable to convert protoporphyrin IX to Mg-protoporphyrin monomethyl ester, should also have lower than normal carbon flow over the common tetrapyrrole pathway. These mutants would be expected to have increased levels of heme which should feedback inhibit aminolevulinate synthase. The protoporphyrin IX levels in bchD and bchH mutants grown under low oxygen tension in broth were compared to the bacteriochlorophyll level in PAS100 grown under similar conditions. As expected, the protoporphyrin IX levels were much lower than the corresponding level of bacteriochlorophyll in PAS100 (Table 4). This indicates that mutations in either bchD or bchH caused carbon flow over the common tetrapyrrole pathway to be reduced. If the reduced carbon flow was due to feedback inhibition of aminolevulinate synthase, then the addition of aminolevulinate to the culture should increase the protoporphyrin IX level. Table 4 shows that the addition of aminolevulinate resulted in a more than 10-fold increase in protoporphyrin IX accumulation, indicating that feedback inhibition of aminolevulinate synthase influences carbon flow over the common tetrapyrrole pathway. It was therefore of interest to determine whether protoporphyrin IX levels in the bchH mutant are subject to oxygen-mediated regulation. When strain AJB456 was grown with an initial oxygen tension of 23%, the intracellular level of protoporphyrin IX was 18 nM, which increased to 48 nM when grown with an initial oxygen tension of 3%. The experiment was repeated in the presence of 1 mM aminolevulinate. Under these conditions, the intracellular level of protoporphyrin IX was 82 nM when grown with an initial oxygen tension of 23%, and 200 nM when grown with an initial oxygen tension of 3%. While aminolevulinate increased the protoporphyrin IX levels, the levels were still regulated by oxygen. This is consistent with earlier results which suggest that feedback inhibition of aminolevulinate synthase cannot be the sole mechanism by which the common tetrapyrrole pathway is regulated (5).

TABLE 4.

Porphyrin levels in bch⁺, bchD, and bchH strains

Strain	Addition	Proto- porphyrina	Bacterio- chlorophyllb	Total amt of porphyrinc
PAS100 (bch+)	None	0.3	7.0	7.3
AJB503 (bchD)	None	0.3	0.0	0.3
AJB456 (bchH)	None	0.2	0.0	0.2
AJB456 (bchH)	ALAd	3.0	0.0	3.0

^aProtoporphyrin IX concentration expressed as nanomoles/milligram of protein.

^bBacteriochlorophyll concentration expressed as nanomoles/milligram of protein.

^cTotal of protoporphyrin IX and bacteriochlorophyll.

^dAminolevulinate (ALA) was added to a final concentration of 1 mM.

Metabolism of [¹⁴C]aminolevulinate.

It has been demonstrated in Rhodospirillum rubrum that aminolevulinate is not committed to tetrapyrrole biosynthesis (43). The possibility that aminolevulinate or porphobilinogen might be diverted from the common tetrapyrrole pathway was investigated by measuring the conversion of [¹⁴C]aminolevulinate to bacteriochlorophyll. Strain PAS100 was grown in the presence of 10 μM aminolevulinate (containing 2 μCi of [¹⁴C]aminolevulinate). After 1 h, 91% of the aminolevulinate had been consumed, measured as ALA-pyrrole (32). A 1.0-ml portion (3.8 × 10⁵ cpm) of the culture was extracted with acetone-methanol (7:2) and spotted on an LK6D thin-layer chromatography plate (Whatman, Inc., Clifton, N.J.). The plate was developed with 70% ethanol, and the radioactivity in the bacteriochlorophyll-containing band was measured. The band contained only 6.6 × 10⁴ cpm, which corresponds to 19% of the aminolevulinate consumed in the 1-h incubation. When this experiment was repeated with PAS100 cells grown photosynthetically, 14% of the incorporated aminolevulinate was converted to bacteriochlorophyll. These results indicate that over 80% of the aminolevulinate was being diverted from the pathway.

Enzymatic reduction of aminolevulinate.

Shigesada et al. (43) discovered that much of the aminolevulinate in R. rubrum is reduced to aminohydroxyvalerate (Fig. 4). To determine whether such an activity is also present in R. capsulatus, a crude extract was made from strain PAS100 grown photosynthetically. When aminolevulinate was added to the reaction mixture, the extract was able to oxidize NADPH, indicating that R. capsulatus does indeed possess aminolevulinate dehydrogenase. In order to determine if aminolevulinate was being reduced to aminohydroxyvalerate, the reaction product was compared by thin-layer chromatography to chemically synthesized aminohydroxyvalerate. After a 1-h incubation, a 10-ml reaction mixture was terminated by addition of enough 1 M acetic acid to reduce the pH to 2.5. The precipitate was removed by centrifugation, and the supernatant passed over a 0.8- by 10-cm column of Dowex 2 × 8, acetate form. The column was washed with 15 ml of water. The eluates were combined and evaporated to dryness. The resulting material was dissolved in a minimal amount of water and subjected to thin-layer chromatography on a Silica Gel HL plate (Analtech, Newark, Del.). The plate was developed with benzene-acetic acid-water (8:2:2) and stained with 1% ninhydrin in acetone. After heating, aminolevulinate formed a yellow spot with an R_f of 0.20. Aminohydroxyvalerate formed a purple spot with an R_f of 0.15. The reaction product formed a purple spot that comigrated with authentic aminohydroxyvalerate.

FIG. 4.

Reaction catalyzed by aminolevulinate dehydrogenase.

The crude extract from photosynthetically grown PAS100 was capable of oxidizing 29 nmol of NADPH/min/mg of protein (Table 3). This specific activity (equal to 1,700 nmol/h/mg of protein) is approximately 10 times higher than the specific activity of porphobilinogen synthase and might well account for the observation presented above that greater than 80% of aminolevulinate is diverted from the common tetrapyrrole pathway. It is possible of course, that there are other enzymes in R. capsulatus in addition to aminolevulinate dehydrogenase and porphobilinogen synthase that use aminolevulinate.

If aminolevulinate dehydrogenase plays a role in regulating carbon flow over the common tetrapyrrole pathway, it would be expected that the specific activity of the enzyme might be higher in aerobically grown cells than in photosynthetically grown cells. When PAS100 was grown aerobically, by sparging the culture with compressed air, the specific activity was more than fourfold higher than when grown photosynthetically (Table 3). This observation is intriguing and suggests that aminolevulinate dehydrogenase might play a role in regulating carbon flow over the common tetrapyrrole pathway.

Porphobilinogen oxygenase.

Several organisms have been shown to contain porphobilinogen oxygenase, an enzyme capable of oxidizing the pyrrole ring of porphobilinogen (18). To test for the presence of porphobilinogen oxidase in R. capsulatus, a crude extract from PAS100 was incubated with porphobilinogen in the presence or absence of dithionite (18). In the absence of dithionite, porphobilinogen consumption is due to porphobilinogen deaminase. In the presence of dithionite, porphobilinogen consumption would be due to the activities of both porphobilinogen deaminase and porphobilinogen oxygenase, if present. There was no difference in porphobilinogen consumption whether or not dithionite was present (14.4 nmol of porphobilinogen consumed versus 13.2 nmol consumed, respectively). Since the presence of dithionite did not increase porphobilinogen consumption in crude extracts of PAS100, it appears that R. capsulatus does not contain a porphobilinogen oxygenase activity.

DISCUSSION

It was surprising to find the hemC gene next to and divergently transcribed from the hemE gene. There are numerous reports of hem gene clusters, and hemCD operons are especially common (1, 19, 23, 29, 46). However, no other cases of clustering of hemC and hemE were found in a search of the GenBank database. The unique relationship of these two genes in R. capsulatus is especially interesting in light of the presence of a palindrome in the region between hemC and hemE (nucleotides 98 to 115). This palindrome has been found upstream of several R. capsulatus genes involved in photosynthesis (2, 31) and is similar to a consensus sequence found in the recognition sites of many positive and negative regulatory genes (21). Site-directed mutagenesis of the palindrome upstream of the bchC promoter suggested a role for this sequence in oxygen-mediated transcriptional regulation (31). The role that this palindrome might play, if any, in regulating transcription of the hemC and hemE genes is unclear, as neither of these genes appear to be transcriptionally regulated by oxygen (G. Ineichen, K. Canada, and A. Biel, unpublished observations).

The lack of mutants in the common tetrapyrrole pathway has been a major stumbling block in studying the regulation of this pathway in photosynthetic bacteria. This is the first report of a mutant which has a complete block in the common tetrapyrrole pathway after aminolevulinate synthase. It should be possible to use this method to isolate R. capsulatus mutants with insertions in hemB, hemE, and hemH, the other hem genes which have been cloned (26, 27, 30). The mutant, AJB565, has the expected profile of enzymatic activities (Table 2) and requires cysteine and methionine, presumably due to an inability to produce siroheme and vitamin B₁₂. It is unclear why AJB565 will not grow in broth containing hemin; however, it has been previously noted that the hemA mutant also shows very limited growth in broth containing hemin (50).

The isolation of a hemC mutant allowed us to look at regulation of porphobilinogen levels. The results presented in Table 2 demonstrate that porphobilinogen levels are regulated by oxygen tension. Previous results demonstrated that the major regulatory point for oxygen-mediated control of the common tetrapyrrole pathway was after aminolevulinate formation (5). From this, it would seem most likely that oxygen would control the conversion of aminolevulinate to porphobilinogen. In some organisms, the conversion of aminolevulinate to porphobilinogen has been determined to be the regulated step of the common tetrapyrrole pathway. For example, in B. japonicum, oxygen controls the common tetrapyrrole pathway by regulating transcription of the hemB gene (10). In other organisms, however, the pathway is regulated by controlling the synthesis of aminolevulinate. In E. coli, for example, the pathway is regulated by feedback inhibition of aminolevulinate formation by heme (3).

The possibility that the hemB gene is transcriptionally regulated by oxygen was tested by constructing a hemB-cat fusion vector and measuring chloramphenicol acetyltransferase in cultures grown under high and low oxygen tensions. There was no evidence of transcriptional regulation of the hemB gene. This is consistent with the finding that the activity of porphobilinogen synthase is not regulated by oxygen (Table 3). Additionally, this enzyme does not appear to be feedback inhibited by either hemin or protoporphyrin IX (Table 3).

The formation of aminolevulinate, being the first step in the common tetrapyrrole pathway, is the most logical place for the pathway to be regulated. Regulation of this step by heme in E. coli has been demonstrated (3) and has been proposed as the mechanism by which this pathway is regulated in Rhodobacter sphaeroides (39). Indeed, we have previously demonstrated that the intracellular level of heme does influence carbon flow over this pathway in R. capsulatus (30). However, measurements of hemA transcription, aminolevulinate synthase activity and the effect of exogenous aminolevulinate on protoporphyrin accumulation all suggest that regulation of aminolevulinate formation is unlikely to be the sole mechanism by which oxygen controls the common tetrapyrrole pathway. Transcription of hemA is only two- to threefold higher when cultures are grown under low oxygen tension than when they are grown under high oxygen tension (24, 49). Aminolevulinate synthase activity is the same in cultures grown under high and low oxygen tensions (5, 24), although a sixfold increase in enzyme activity has been reported in photosynthetically grown cultures (24). Additionally, it has been demonstrated that protoporphyrin accumulation in a bchH mutant is regulated by oxygen even in the presence of exogenous aminolevulinate, while the presence of exogenous porphobilinogen abolishes oxygen-mediated regulation, indicating that oxygen must regulate some step later in the pathway (5).

The results presented in this study confirm the previous work by showing that bchD and bchH mutants, which accumulate protoporphyrin IX and, therefore, presumably heme, have reduced carbon flow over the common tetrapyrrole pathway (Table 4). Thus, it appears that in R. capsulatus, as in R. sphaeroides, the intracellular level of heme plays a role in carbon flow over the common tetrapyrrole pathway (39). However, while the addition of exogenous aminolevulinate increased carbon flow, it did not alleviate oxygen-mediated regulation of the pathway, suggesting that while feedback inhibition of aminolevulinate synthase undoubtedly plays a role in controlling carbon flow over this pathway, it cannot be the sole mechanism by which oxygen regulates the pathway.

It is interesting to note that B. japonicum can obtain its aminolevulinate from the host plant, and therefore that the conversion of aminolevulinate to porphobilinogen is the first essential step in the common tetrapyrrole pathway (36). In contrast, in E. coli, aminolevulinate is committed to tetrapyrrole biosynthesis, and its formation is regulated by oxygen tension (3). The metabolism of [¹⁴C]aminolevulinate in R. capsulatus suggests that in this organism aminolevulinate is not committed to tetrapyrrole biosynthesis. In fact, only a small fraction of the aminolevulinate is used in tetrapyrrole biosynthesis. Similar results have been observed in R. rubrum, for which it was noted that only approximately 10% of [¹⁴C]aminolevulinate was incorporated into tetrapyrroles and that the majority of the aminolevulinate was converted to aminohydroxyvalerate (42, 43). Shigesada was able to demonstrate conversion of [¹⁴C]aminolevulinate to aminohydroxyvalerate by a cell extract of R. rubrum (42). We have demonstrated that this enzyme also exists in R. capsulatus and that the very high specific activity of this enzyme could account for the low level of incorporation of [¹⁴C]aminolevulinate into bacteriochlorophyll. The observation that the specific activity of aminolevulinate dehydrogenase in R. capsulatus crude extracts is much higher in aerobically grown cultures than in photosynthetically grown cultures suggests a mechanism by which the common tetrapyrrole pathway might be regulated.

The findings presented here suggest that carbon flow over the common tetrapyrrole pathway is controlled both by regulating the rate of aminolevulinate synthesis and by regulating the amount of aminolevulinate that is diverted from the pathway. It seems probable that neither factor alone could account for the large difference in protoporphyrin IX levels seen under high and low oxygen tensions. Based on this, we envision that at high oxygen tensions, the higher specific activity of aminolevulinate dehydrogenase diverts more aminolevulinate from the common tetrapyrrole pathway, resulting in a decrease in protoporphyrin IX synthesis. Since the formation of Mg-protoporphyrin monomethyl ester from protoporphyrin IX is completely shut off under high oxygen tensions (5), all the protoporphyrin IX is converted to heme, which feedback inhibits aminolevulinate synthase, resulting in a further decrease in carbon flow over the common tetrapyrrole pathway.

There is presently no evidence to suggest what the ultimate fate of the aminohydroxyvalerate might be, although it is interesting that a link between the common tetrapyrrole pathway and poly-β-hydroxybutyrate synthesis has been demonstrated with R. sphaeroides (16).

Isolation of a mutant lacking aminolevulinate dehydrogenase is obviously very important to determining the role this enzyme plays in regulation of the common tetrapyrrole pathway. Isolation of such a mutant would also help in determining the presence of other enzymes which use aminolevulinate. Additionally, since this appears to be the quantitatively major use of aminolevulinate, it is of interest to investigate the fate of the aminohydroxyvalerate produced.