INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Heme compounds represent a ubiquitous, iron-containing subclass of biological tetrapyrroles. In association with specific apoproteins, they serve a wide range of important functions including electron transport (e.g., cytochromes), binding and transport of O₂ (e.g., hemoglobin), and oxidative catalysis (e.g., peroxidases). The universal tetrapyrrole precursor -aminolevulinic acid (ALA) is synthesized by either one or the other of two distinct pathways (for reviews, see references 9 and 41). Plants, algae, archaea, and most eubacteria (including enteric bacteria) use the so-called glutamate (C₅) pathway, which includes reduction of tRNA-bound glutamyl by glutamyl-tRNA reductase. Nonplant eukaryotes and members of the  subgroup of proteobacteria (including the rhizobia) synthesize ALA via condensation of glycine and succinyl coenzyme A (CoA) by ALA synthase (Shemin pathway). Subsequently, ALA is converted to protoheme in seven successive enzymatic reactions which are identical in both classes of organisms. Depending on the cellular oxygen conditions, the fifth reaction of this sequence, namely, the oxidative decarboxylation of coproporphyrinogen III to protoporphyrinogen IX, is catalyzed by different enzymes in facultatively aerobic bacteria. Under aerobic conditions, the oxidation is performed by coproporphyrinogen III oxidase (e.g., the product of the Escherichia coli hemF gene), which uses molecular oxygen as an electron acceptor. In the absence of oxygen, the reaction is catalyzed by the structurally unrelated coproporphyrinogen III dehydrogenase, whose activity requires NADP⁺, ATP, Mg²⁺, and L-methionine. Corresponding genes (commonly named hemN) have been cloned from a number of bacteria such as E. coli (70), Salmonella enterica serovar Typhimurium (73), Pseudomonas aeruginosa (63), Rhodobacter sphaeroides (18), and Ralstonia eutropha (46).

Bacterial heme synthesis is subject to multiple controls by the availability of oxygen, iron, and/or heme. Regulation in response to cellular oxygen conditions was observed predominantly at two steps, namely, ALA synthesis and oxidative decarboxylation of coproporphyrinogen III. Regardless of the pathway used for ALA synthesis, expression of the respective genes was elevated in various bacterial species when they were grown under low-oxygen conditions (20, 40, 56). Similarly, expression of the hemN genes from E. coli, R. eutropha, and P. aeruginosa is increased under anaerobic conditions, which is understandable in the light of the oxygen-independent reaction catalyzed by the HemN protein (46, 63, 70).

In our laboratories, we are studying the gram-negative soil bacterium Bradyrhizobium japonicum, which can exist as a free-living organism or as a nitrogen-fixing root nodule symbiont of its soybean host plant. The switch between the different lifestyles requires a high degree of physiological flexibility. For example, B. japonicum induces the synthesis of a cbb₃-type, high-affinity terminal oxidase (FixNOQP) during the infection process in order to cope with the microaerobic conditions prevailing in the infected host plant tissue (for a review, see reference 36). Coordinated expression of symbiotic genes is brought about by the concerted action of two largely independent regulatory cascades which both respond to cellular oxygen conditions (8, 52; for reviews, see references 22 and 23). In the FixLJ-FixK₂ cascade, the low-oxygen signal is detected and transduced by the FixLJ two-component regulatory system, whereas NifA fulfills the same function in the RegSR-NifA cascade. The signal for the superimposed RegSR two-component regulatory system is not known at present. Transcriptional activation of target genes by NifA involves a specialized RNA polymerase holoenzyme containing the alternative  factor RpoN (⁵⁴ or ^N), which enables recognition of 24/12-type promoters (for a review, see reference 21).

In a recent study, Nienaber and coworkers used a DNA-RNA hybridization approach for a global survey of the regulatory function of NifA with regard to the so-called symbiotic gene region of the B. japonicum chromosome (53). This DNA region of about 400 kb comprises a large number of genes concerned with nodulation and symbiotic nitrogen fixation (45). Using this strategy, we have identified several new NifA-regulated genes, including nrgA and nrgBC, whose functions are unknown, and, interestingly, a hemN-like gene. Thus it appeared that regulation of symbiotic genes and a key heme biosynthetic gene was linked in B. japonicum via NifA, which would make perfect sense in view of the de novo synthesis of heme proteins during the transition from the free-living to the symbiotic state. However, we present evidence here that the previously identified hemN gene (hemN₁) is not functional under the conditions tested and that its regulation also involves the FixLJ-FixK₂ cascade in addition to NifA. Moreover, we report on the identification and characterization of a second hemN gene (hemN₂) which is essential for anaerobic heme biosynthesis in B. japonicum.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and plasmids. The bacterial strains and plasmids used in this work are listed in Table 1.

Media and growth conditions. Luria-Bertani (LB) medium (48) was used for growth of E. coli. For growth of serovar Typhimurium TE3006 cells, LB medium supplemented with 20 µg of hemin/ml was used. Where appropriate, antibiotics were used at the following concentrations (in micrograms per milliliter): ampicillin, 200 (50 for serovar Typhimurium); kanamycin, 30 (10 for serovar Typhimurium); tetracycline, 10. Peptone-salts-yeast extract (PSY) medium (62) supplemented with 0.1% L-arabinose was used for routine aerobic and microaerobic cultures of B. japonicum, while yeast extract-mannitol (YEM) medium supplemented with 10 mM KNO₃ (19) was used for anaerobic B. japonicum cultures. Where appropriate, 15 µg of hemin/ml and 0.01% (vol/vol) Tween 80 (to prevent hemin precipitation) were added to anaerobic cultures of B. japonicum cells. Microaerobic cultures (10-ml volumes) were kept under a gas mixture consisting of 0.5% O₂ and 99.5% N₂ in rubber-stoppered serum bottles (volume, 500 ml), and the gas phase was periodically replaced (approximately every 12 h) by flushing the bottles with the same gas mixture. Anaerobic cultures were kept under nitrogen in serum bottles. Cells used for heme protein analysis were grown aerobically in YEM medium lacking KNO₃ and anaerobically in KNO₃-supplemented YEM medium. Concentrations of antibiotics for use in B. japonicum cultures were as follows (in micrograms per milliliter): spectinomycin, 100; kanamycin, 100; streptomycin, 50; tetracycline, 50 (solid media) or 25 (liquid media).

Routine DNA work and sequence analysis. Recombinant DNA work and Southern blotting were performed according to standard protocols (64). Probes for Southern blot hybridizations were generated by PCR and labeled with digoxigenin (DIG). Single- or double-stranded plasmid DNA was cycle-sequenced by the chain termination method of Sanger et al. (65) with DNA sequencers (models 373A and 377; Perkin-Elmer Applied Biosystems, Foster City, Calif.). For computer-assisted analyses of DNA and protein sequences, we used the software package (version 10.0) of the Genetics Computer Group of the University of Wisconsin (UWGCG) (Madison, Wis.). Homology searches were performed by using the National Center for Biotechnology Information's BLAST network server (http://www.ncbi.nlm.nih.gov/BLAST/).

Cloning of two B. japonicum hemN genes. The hemN₁ gene was identified previously and cloned on an 8.1-kb EcoRI fragment in plasmid pRJ8224 (Fig. 1A) (53). A hemN₁-specific DIG-labeled probe (1.11 kb) was generated by PCR with the following hemN₁-internal primers: hemN2, 5'-A₂₃₄TCCCGTGTCGCTATACATC-3' (forward primer; 20 nucleotides [nt]), and hemN3, 5'-C₁₃₄₃CTCTTCAATGTCCACTATCCC-3' (reverse primer; 22 nt) (numbering refers to the hemN₁ sequence deposited in GenBank). A partial B. japonicum genomic plasmid library was then hybridized with this probe under low-stringency conditions (5× SSC [1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate], 58°C). A strongly hybridizing clone was found to contain a 3.2-kb EcoRI-HindIII fragment present in plasmid pRJ8267 and shown by sequence analysis to comprise a second hemN-like gene (Fig. 1B).

Transcript analysis. Transcripts of both hemN genes were analyzed by primer extension experiments according to previously described protocols (5, 53). Oligonucleotide hemN1 (28 nt; 5'-C₁₃₅GAGCGTCGCAATACTTGTTCAGGAGAG-3') was used as a primer for detection of hemN₁ transcripts, and oligonucleotides 8267-3 (28 nt; 5'-C₅₈₂ACTGCGAGATCAGCTCTCATATCTTAC-3') and 8267-4 (23 nt; 5'-A₆₂₁TAGCTCGTGTAGCGAGGCAGTC-3') were used for hemN₂ transcripts. RNA was isolated as described previously (5) from B. japonicum strains 110spc4 (wild type), A9 (nifA), and 9043 (fixK₂) grown aerobically, microaerobically, or anaerobically (where applicable).

Construction of hemN mutants. B. japonicum hemN₁ mutant strain 8244 originated from a random Tn5 mutagenesis performed with pRJ8224 in E. coli cys::Tn5. The mutated EcoRI fragment containing Tn5 at nucleotide position 383 (corresponding to codon 96 of hemN₁) was subcloned into vector pSUP202 and used for marker exchange mutagenesis in B. japonicum as described previously (33, 34). The hemN₂ mutant GRN307 was constructed by insertion of a 2-kb SalI fragment from pHP45:: (Sm^r Sp^r) into the hemN₂-internal SalI site at position 1132 via marker exchange mutagenesis. In mutant 8275, the hemN₂ gene was disrupted by cointegration of plasmid pRJ8275 containing a 447-bp hemN₂-internal SacI-SalI fragment. The hemN₁ hemN₂ double mutant 8275-44 was obtained by applying the same mutagenesis procedure used for construction of strain 8275 to the hemN₁ mutant 8244. The correct genomic structure of all mutant strains was confirmed by Southern blot analysis of genomic DNAs.

Construction of chromosomally integrated hemN'-'lacZ fusions. For construction of the hemN₁'-'lacZ fusion, a 4.25-kb EcoRI-HindIII fragment of pRJ8224 was subcloned into pNM482, resulting in a fusion of lacZ to the 149th codon of hemN₁. This fusion was subcloned as a 9.2-kb EcoRI-StuI fragment into pSUP202pol4 and integrated into the B. japonicum chromosome by conjugation and homologous recombination (Tc^r selection), resulting in strain 8246 (Fig. 1A). Similarly, a 545-bp StuI-Ecl136II fragment was cloned into SmaI-digested pSUP480 for construction of a hemN₂'-'lacZ fusion, which was chromosomally integrated via conjugation and homologous recombination, yielding strain 8281 (Fig. 1B). Both hemN fusions were also integrated into the fixK₂ mutant 9043. Furthermore, the hemN₁ fusion was introduced into the nifA mutant A9.

-Galactosidase assays. -Galactosidase activity assays were performed as described previously (25).

Plant infection test. The symbiotic phenotype of hemN mutants was determined in soybean plant infection tests as described elsewhere (31, 33).

Cell fractionation, protein gel electrophoresis, and heme staining. Cells of B. japonicum grown aerobically in YEM medium (500 ml) were harvested by centrifugation (8,000 × g for 10 min), washed twice with YEM, and finally resuspended in 10 ml of the same medium. After resuspension, 5 ml of the cell suspension was recentrifuged, and the pellet was stored at 20°C until use. The remaining 5 ml of cell suspension was inoculated into YEM medium (1 liter) supplemented with 10 mM KNO₃ and was incubated anaerobically at 28°C for 72 h. Aerobically and anaerobically cultured cells were washed with 50 mM sodium phosphate buffer (pH 7.0) containing 1 mM MgCl₂, 0.1 mM CaCl₂, and 0.9% NaCl and then resuspended in 3 ml of the same buffer containing 1 mM 4-amidinophenylmethanesulfonyl fluoride, 20 µg of DNase I/ml, and 20 µg of RNase A/ml. Cells were disrupted by three passages through an ice-cold French pressure cell (SLM-Aminco) at a pressure of about 120 MPa. The cell extract was centrifuged at 10,000 × g for 10 min to remove unbroken cells, and the supernatant was then centrifuged at 140,000 × g for 2 h. The pellet was resuspended in 2 ml of the same buffer and centrifuged at 140,000 × g for 2 h. The final pellet was resuspended in 100 µl of loading buffer (124 mM Tris-HCl [pH 7.0], 20% glycerol, and 4.6% sodium dodecyl sulfate [SDS]) and electrophoresed on an SDS-12% polyacrylamide gel at 4°C. Proteins were transferred to a nitrocellulose filter and stained for heme-dependent peroxidase activity by chemiluminescence as described previously (71). The protein concentration was estimated by using the Bio-Rad assay (Bio-Rad Laboratories, Richmond, Calif.) with bovine serum albumin as the standard.

Complementation tests with Salmonella serovar Typhimurium TE3006 and B. japonicum GRN307. Plasmids pRJ8262 and pRJ8276 (Fig. 1) carrying transcriptional fusions of the vector-borne lac promoter to hemN₁ and hemN₂, respectively, were used for heterologous complementation tests with the serovar Typhimurium hemN hemF mutant TE3006. They were constructed by cloning a 1.45-kb BglII fragment (hemN₁) and a 1.87-kb NsiI-EcoRI fragment (hemN₂) into BamHI-digested pUC18 (pRJ8262) and PstI-EcoRI-digested pUCBM21 (pRJ8276), respectively. Plasmids pRJ8262 and pRJ8276 and the control vector pUCBM21 were transformed into serovar Typhimurium TE3006. Growth of the resulting strains was tested on solid LB medium plus 0.2% glucose with (control) or without 20 µg of hemin/ml under aerobic and anaerobic conditions.

Nucleotide sequence accession number. The nucleotide sequences of the B. japonicum hemN₁ and hemN₂ genes have been deposited in GenBank under accession numbers AF276709 and AJ002517, respectively.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Identification and sequence analysis of two B. japonicum hemN genes. The B. japonicum hemN₁ gene was identified on an 8.1-kb EcoRI fragment (Fig. 1A) of cosmid 14 in the course of screening the symbiotic genome region for segments that are transcribed in a NifA-dependent manner (53). DNA sequence analysis of the hybridizing region on this EcoRI fragment, which was cloned in plasmid pRJ8224 (Fig. 1A), revealed the presence of an open reading frame (hemN₁) whose predicted protein product showed significant similarity to anaerobic coproporphyrinogen III dehydrogenases from different bacteria (see below and Discussion). When a PCR-generated, hemN₁-internal DNA fragment was used to probe Southern blots of restricted B. japonicum genomic DNA under low-stringency conditions, the observed hybridization signals suggested the presence of a second hemN-like gene (data not shown). Based on this observation, we cloned a 3,163-bp HindIII-EcoRI fragment (insert of plasmid pRJ8267) whose sequence analysis eventually confirmed the presence of the presumed second hemN gene, which we named hemN₂ (Fig. 1B).

View larger version (23K): [in this window] [in a new window]   FIG. 1.   Physical maps of the B. japonicum (Bj) hemN1 (A) and hemN2 (B) regions and plasmids used for complementation experiments with B. japonicum GRN307 (C). Shown are the inserts of relevant plasmids, which are specified by their designations on the right. Vertical arrows mark the location of Tn5 in the hemN1 mutant 8244 and of the  cassette in the hemN2 mutant GRN307. The genetic structure of mutant 8275 is depicted in panel B, with the hemN2-disrupting vector pSUP202pol4 symbolized by a double line (not drawn to scale). Horizontal arrows indicate the orientation of the lac promoter present on the vector of the corresponding plasmids. The bottom parts of panels A and B show the structures of the chromosomally inserted lacZ fusions present in strains 8246 and 8281, respectively. Panel C shows the inserts of plasmids pRJ8294 and pRJ8295, which carry transcriptional fusions between the hemN2 promoter region (PhemN2) and hemN1 and between the hemN1 promoter region (PhemN1) and hemN2, respectively. Also shown is the insert of the control plasmid pRJ8296. Positions of start and stop codons of the hemN genes, of the Tn5 insertion in strain 8244, and of restriction sites are indicated. A, AscI; Bg, BglII; E, EcoRI; Ec, Ecl136II; H, HindIII; K, KpnI; N, NsiI; Nd, NdeI; S, SalI; Sa, SacI; Sm, SmaI; St, StuI. Restriction sites in parentheses were lost during cloning procedures, and NdeI sites in brackets were introduced by PCR amplification with mutagenic primers (see Materials and Methods). Position numbers refer to the hemN1 and hemN2 sequences deposited in GenBank.

Transcript analyses of hemN₁ and hemN₂. Primer extension experiments were performed to analyze hemN transcripts in cells of B. japonicum wild-type, fixK₂ mutant, and nifA mutant strains grown under different conditions (Fig. 2). No transcript was detected from either hemN gene when RNA from aerobically grown cells was used. By contrast, reverse transcription with RNA isolated from microaerobically or anaerobically grown cells revealed hemN₁- and hemN₂-specific transcripts starting at G₇₄ and A₅₀₇, respectively, 23 and 54 nt upstream of the respective translational start codons. The synthesis of both hemN transcripts was dependent on the FixK₂ activator protein, as deduced from their absence in fixK₂ mutant cells. In line with this finding, potential FixK boxes (22) are located around positions 41.5 (hemN₁; T₂₆TGN₇CAA) and 40.5 bp (hemN₂; T₄₆₀TGN₈CAA) upstream of the respective transcription start sites (see also Discussion). The putative 35 and 10 regions associated with both hemN genes exhibited only limited similarity with the consensus sequence defined for B. japonicum housekeeping promoters (10). Because our previous competitive DNA-RNA hybridization studies suggested that hemN₁ is under the control of NifA (53), we extended the hemN₁ transcript analysis to anaerobically grown nifA mutant cells. As shown in Fig. 2A, the hemN₁ transcript was present but less abundant in the nifA mutant than in the wild type. Moreover, a potential 24/12-type promoter was found upstream of the hemN₁ transcriptional start site (G₅₃GCAN₇GCT; the most invariant nucleotides at positions 24 and 12 are underlined); this promoter, in concert with RNA polymerase containing ⁵⁴, may direct NifA-dependent hemN₁ transcription.

View larger version (71K): [in this window] [in a new window]   FIG. 2.   Mapping of the transcription start sites of B. japonicum hemN1 (A) and hemN2 (B) by primer extension. Total RNA was isolated from aerobically (+), microaerobically (±), or anaerobically () grown cells of wild-type (wt) B. japonicum or mutant strain A9 (nifA) or 9043 (fixK2) and was used for primer extension experiments with the hemN1-specific primer hemN1 (A) and the hemN2-specific primer 8267-4 (B). The sequence ladders shown were generated with pRJ8262 (A) and pRJ8267 (B) DNAs and the same primers which were used for the respective transcript mapping. Note that panel B is a composite of relevant lanes of a single gel and that the "T" lane is missing in the sequence ladder. The dominant transcription start sites are marked by "+1" and solid arrowheads. The relevant nucleotide sequences of the hemN1 and hemN2 promoter regions are shown at the bottoms of panels A and B, respectively. Elements of putative 35/10- and 24/12-type promoters are shaded and highlighted on a solid background, respectively. Conserved nucleotides of presumptive FixK boxes (5'-TTGATN4ATCAA-3') (22) are italicized and shaded. The transcription start site is indicated by "+1," and the start codon of each hemN gene is underlined. The numbering of nucleotide positions refers to that given in Fig. 1 and in the sequences deposited in GenBank.

Analysis of hemN₁ and hemN₂ expression with lacZ fusions. To quantify hemN expression, translational lacZ fusions to both genes were constructed as described in Materials and Methods and integrated into the chromosome of wild-type B. japonicum. The same fusions were also integrated into the chromosomes of suitable regulatory mutants in order to confirm the results of the transcript studies, which suggested a dependence of both hemN genes on FixK₂ and, in addition, a dependence of hemN₁ on NifA. The resulting strains were cultivated under aerobic, microaerobic, and, where applicable, anaerobic conditions, and hemN expression was monitored by assaying -galactosidase activity (Table 2). None of the hemN genes was significantly expressed in aerobically grown cells, but expression of both genes was strongly induced (20-fold) under reduced oxygen conditions. Microaerobic induction of both hemN genes was strictly dependent on FixK₂. Furthermore, it appeared as if the postulated 24/12-type promoter indeed contributed to hemN₁ expression under anaerobic conditions, as inferred from the reduced hemN₁ expression in the nifA mutant background. Since fixK₂ mutants do not grow anaerobically with nitrate (52), it was not possible to directly assess the likely function of FixK₂ in anaerobic hemN expression.

Characterization of hemN mutants. The physiological importance of the B. japonicum hemN genes was studied by phenotypic analysis of suitable hemN mutant strains. First, we tested growth of the mutant strains under different oxygen conditions. None of the mutants differed from the wild type with regard to aerobic and microaerobic (0.5% O₂) growth (data not shown). By contrast, the hemN₂ mutant strain 8275 was unable to grow under anaerobic conditions with nitrate as the terminal electron acceptor whereas growth of the hemN₁ mutant 8244 under these conditions was comparable to that of the wild type (Fig. 3). Anaerobic growth of the hemN₂ mutant was largely restored when the medium was supplemented with 15 µg of hemin per ml. Thus, an intact hemN₂ gene is essential for anaerobic growth, and its function cannot be replaced by the other hemN homolog, hemN₁.

View larger version (18K): [in this window] [in a new window]   FIG. 3.   Anaerobic growth of wild-type B. japonicum (solid circles) and hemN mutant strains 8244 (hemN1 mutant) (solid squares) and 8275 (hemN2 mutant) (triangles). Washed cells were inoculated into YEM medium plus 10 mM KNO3 lacking (solid symbols) or containing (open triangles) 15 µg of hemin/ml, and growth of the cells under anaerobic conditions was measured by monitoring the optical density at 600 nm.

Analysis of heme proteins. Membrane protein fractions from wild-type 110spc4 and mutant GRN307 cells were separated by SDS-polyacrylamide gel electrophoresis and stained for covalently bound heme proteins. In the membrane fraction from aerobically grown cells, two stained bands with M_rs of 28,000 and 20,000 were detected (Fig. 4, left two lanes). These two proteins have been identified previously as the membrane-bound c-type cytochromes c₁ (M_r, 28,000) and CycM (M_r, 20,000) (12, 68). After incubation of the cells under anaerobic conditions with nitrate, the membrane fraction from wild-type cells showed, in addition to cytochrome c₁ and CycM, at least four c-type cytochromes with M_rs of 32,000, 25,000, 24,000, and 16,000 (Fig. 4, second lane from the right). As described by Preisig and coworkers (58), there is even a fifth protein with an M_r of 28,000 comigrating with cytochrome c₁. The induced proteins with M_rs of 32,000 and 28,000 detected in anaerobic cultures of wild-type cells have been identified previously as the B. japonicum FixP and FixO proteins, respectively, of the cbb₃-type, high-affinity cytochrome oxidase encoded by the fixNOQP operon (58, 60). Similarly, previous N-terminal amino acid sequencing of the 16-kDa protein identified this protein as a nitric oxide reductase homolog (NorC; R. Zufferey, unpublished data). The identities of the 24- and 25-kDa proteins are not known. In contrast to the membrane fraction from anaerobic cultures of the wild type, that of mutant strain GRN307 contained only two major c-type cytochromes, the 28-kDa cytochrome c₁ and the 20-kDa CycM protein (Fig. 4, rightmost lane), which were also present in aerobically grown cells. Occasionally, trace amounts of the 32-kDa FixP protein were also detected in the mutant membranes. Similar results were obtained with the hemN₂ mutant 8275, whereas the hemN₁ mutant 8244 showed the same pattern of heme-stained proteins as the wild type (data not shown). Taken together, it is evident that synthesis of anaerobically inducible c-type cytochromes is severely impaired in the hemN₂ mutants (see also Discussion).

View larger version (48K): [in this window] [in a new window]   FIG. 4.   Heme-stained proteins in membrane fractions prepared from cells of B. japonicum strains 110spc4 (wt) and GRN307 (hemN2) grown aerobically in YEM medium or anaerobically in YEM medium supplemented with 10 mM KNO3. Proteins were separated by SDS-polyacrylamide gel electrophoresis, transferred to nitrocellulose, and stained for covalently bound heme proteins. Each lane contains about 25 µg of membrane proteins. Heme-stained c-type cytochromes identified previously (58) are specified at the right margin. The 16-kDa protein was identified as a NorC homolog (nitric oxide reductase) by N-terminal sequencing of the protein (R. Zufferey, unpublished data). The identities of unmarked bands are not known. Apparent molecular masses of the proteins are shown at the left margin.

Complementation of hemN mutants. The phenotypic analysis of the hemN mutants suggested that hemN₁ does not functionally compensate for a missing hemN₂ gene. We were interested in verifying this by complementation tests. When plasmids pRJ8262 and pRJ8276 (Fig. 1) were used to complement the Salmonella serovar Typhimurium hemF hemN double mutant TE3006, functional complementation for growth on heme-deficient medium was observed only with the hemN₂ plasmid pRJ8276, not with the hemN₁ plasmid pRJ8262 (Fig. 5). For complementation experiments in the homologous B. japonicum background, plasmids pRJ8294, pRJ8295, and pRJ8296 (Fig. 1C) (see Materials and Methods for detailed descriptions of these plasmids) were integrated into the chromosome of the hemN₂ mutant GRN307, and the resulting strains were tested for the ability to grow anaerobically in YEM medium with nitrate. B. japonicum strain 8294-1, which contains the transcriptional P_hemN2::hemN₁ fusion, was unable to grow under these conditions, indicating that hemN₁ could not substitute for hemN₂ even when it was transcribed from the hemN₂ promoter (data not shown). By contrast, strain 8295-4, harboring the P_hemN1::hemN₂ fusion, grew as well as the control strain 8296 (hemN₂ with its own promoter), thus demonstrating that hemN₁ promoter activity is not limiting in this test. Taken together, the complementation results strongly support the notion that the product of hemN₁ is not a functional substitute for the HemN₂ protein.

View larger version (79K): [in this window] [in a new window]   FIG. 5.   Complementation of the heme auxotrophic mutant Salmonella serovar Typhimurium TE3006 (hemF hemN). The TE3006 derivatives harboring either pUCBM21 (vector control), pRJ8262 (B. japonicum hemN1), or pRJ8276 (hemN2) were grown overnight in liquid LB medium containing ampicillin (50 µg/ml) and hemin (20 µg/ml) at 30°C. After washing and adjustment to an optical density at 600 nm of 0.8, 1-µl volumes of the suspensions and serial dilutions thereof (indicated between the panels) were spotted onto LB agar plates containing ampicillin, glucose (0.2% [wt/vol]), and either 20 µg of hemin/ml (left panel) or 0.2 mM isopropyl--D-thiogalactopyranoside (IPTG) (right panel). After incubation of the plates at 30°C for 40 h, the plates were photographed. Note that comparable background growth was observed on hemin-free medium with TE3006 harboring pUCBM21 or pRJ8262.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Starting out from a competitive DNA-RNA hybridization approach to identify NifA-dependent genes in the symbiotic gene region, we have identified two distinct hemN genes in B. japonicum. Unexpectedly, expression of hemN₁ was only partially dependent on NifA but predominantly dependent on FixK₂, which is also the mediator of microaerobic induction of hemN₂. Thus, the subtractive hybridization method which initially led to the discovery of hemN₁ (53) is apparently sensitive enough to detect genes whose expression varies only by a factor of 2. Unfortunately, it was not possible to determine the contribution of FixK₂-dependent hemN₁ expression under anaerobic conditions due to the inability of fixK₂ mutants to grow under these conditions. The seemingly absent hemN₁ expression under microaerobic conditions in the fixK₂ mutant strain K₂8246 can be explained by the impaired activity of the oxygen-sensitive NifA protein under these conditions (26). Moreover, NifA-mediated activation of the hemN₁ promoter is probably weak because of the absence of a consensus NifA binding site (TGTN₁₀ACA) (13). The core elements of the promoters directing NifA- and FixK₂-dependent transcription of hemN₁ seem to overlap, as indicated by the comparable lengths of the primer extension products detected in the wild-type, fixK₂ mutant, and nifA mutant strains. This situation is reminiscent of the B. japonicum fixR-nifA operon, which is also preceded by two overlapping promoters that are recognized by distinct RNA polymerase holoenzymes (6, 7). Obviously, the complex transcriptional organization of hemN₁ would require further analysis of the proposed promoters. However, in the absence of any functional evidence for HemN₁, such studies are of little significance; hence, they were not performed.

FixK₂-dependent, microaerobic induction of both hemN genes is in line with the presence of putative FixK boxes (TTGN₇CAA; TTGN₈CAA) (Fig. 6) 41.5 and 40.5 bp upstream of the transcriptional start of hemN₁ and hemN₂, respectively. Not surprisingly, the associated promoter regions are only poorly conserved, as is the case in numerous other FixK₂-dependent genes and operons, such as fixNOQP (52, 57) and fixGHIS (59), and also in many Fnr-activated promoters (32). In agreement with our findings with the B. japonicum hemN genes, induction of hemN-like genes under low-oxygen conditions has been reported for E. coli (70), R. eutropha (46), and B. subtilis (39), but detailed information about the regulatory mechanism is available only for P. aeruginosa (63). In this organism, anaerobic induction of hemN expression is dependent on the dual action of two Fnr-like redox response regulators, Anr and Dnr, and an Anr box centered around position 41.5 upstream of the hemN transcriptional start site. Unlike the B. japonicum hemN genes, hemN of P. aeruginosa is also expressed under aerobic conditions, and Anr (but not Dnr) is required for this type of control.

View larger version (25K): [in this window] [in a new window]   FIG. 6.   FixK boxes associated with B. japonicum heme biosynthetic genes which are subject to oxygen control. A deduced consensus sequence (cons) is given at the bottom of the figure. The vertical dotted line marks a dyad symmetry axis with respect to the most conserved nucleotides, indicated by white letters on a solid background. Numbers refer to the distance of the marked nucleotide from the transcription start site, except in the case of hemB, where the distance relative to the translation start site is indicated because the hemB transcription start site is not known (14). The function of the FixK box of hemA has been substantiated by deletion analysis (56).

With the newly identified hemN genes, a total of four different heme biosynthesis genes are now known in B. japonicum. The three previously characterized genes include hemA (encoding ALA synthase [47]), hemB (encoding aminolevulinic acid dehydratase [14]), and hemH (encoding ferrochelatase [29]). While the hemH gene appears not to be regulated, expression of hemA and hemB is under dual control by the cellular oxygen and iron conditions (15, 16, 55, 56). Microaerobic induction of both hemA and hemB was shown to require fixJ. It is very likely that this type of control is mediated via the subordinate fixK₂ gene, because both genes are preceded by a well-conserved FixK box as in the case of the two hemN genes (Fig. 6) (see reference 52). By contrast, iron control involves two different regulatory proteins, namely fur for hemA and irr for hemB (35). Whether the hemN genes of B. japonicum are subject to iron control is not presently known, but results of similar studies in E. coli suggest that iron may indeed affect heme biosynthesis at the level of hemN expression (70).

The B. japonicum hemN genes were functionally characterized by mutation and complementation experiments with a heme-auxotrophic Salmonella serovar Typhimurium hemF hemN double mutant. Both sets of experiments clearly indicated that under the conditions tested only HemN₂ is functional and that it is required for anaerobic growth and efficient symbiotic nitrogen fixation but not for aerobic or microaerobic growth. Hence, it is likely that the "coproporphyrinogenase" activity measured in extracts of anaerobically grown B. japonicum cells by Keithly and Nadler (43) originated from hemN₂ expression. Hemin supplementation restored anaerobic growth of the hemN₂ mutant, indicating that heme is taken up by the cells and plays an essential role under these conditions.

Evidence for two genes specifying (putative) oxygen-independent coproporphyrinogen III dehydrogenases was reported previously for other bacteria such as R. sphaeroides (HemF and HemZ) (18, 74), Helicobacter pylori (69), Synechocystis sp. strain PC6803 (42) (accession numbers BAA18218 and BAA17272), and B. subtilis (HemN and HemZ) (37, 39). The existence of a second hemN gene (hemZ) was also predicted in P. aeruginosa, but the corresponding gene has not yet been identified (63). In B. subtilis, both proteins were shown to be functional by complementation of the Salmonella serovar Typhimurium hemF hemN double mutant. Moreover, the presence of a third anaerobic coproporphyrinogen III dehydrogenase in this species was postulated because a hemN hemZ double mutant showed a wild-type phenotype (39).

Heme biosynthesis in B. japonicum is particularly crucial during free-living, microaerobic growth and in symbiosis when the specialized respiratory chain comprising the FixNOQP oxidase is synthesized to support microaerobic respiration (3, 4; for a recent review, see reference 36). We showed that the c-type cytochromes FixO and FixP, whose synthesis is induced under low-oxygen conditions, are missing in membranes from anaerobically incubated mutant strains lacking a functional hemN₂ gene. In addition, a heme-stainable band of approximately 16 kDa, which is very prominent in anaerobic wild-type membranes, is also absent in the hemN₂ mutant. Somewhat unexpectedly, cytochromes c₁ and CycM were still detectable in the mutant membranes. These proteins, however, are synthesized under aerobic conditions also, and thus, they possibly persisted in the cells after the shift to anaerobiosis. Thus, the lack of a functional coproporphyrinogen III dehydrogenase is manifested only at the level of proteins synthesized de novo under anaerobic conditions. Nitrogen fixation activity was not completely abolished in hemN₂ mutants. It is unlikely that the mutant bacteroids were supplemented with heme by the soybean host plant, because a B. japonicum hemB mutant, which is blocked at the second step in heme biosynthesis, could not be rescued by the plant (14). Thus, small amounts of heme may be synthesized in the hemN₂ mutant bacteroids, possibly due to residual activity of the (yet unidentified) aerobic coproporphyrinogen III oxidase, which may be partially functional even at the low-oxygen conditions in nodules. The observation that hemN₂ mutants grew normally under microaerobic conditions (0.5% oxygen) provides further support for this hypothesis.

Amino acid sequence comparison of different HemN proteins cannot provide a clear clue as to the basis of the likely nonfunctionality of HemN₁ (Fig. 7). The characteristic glycine-rich H(F/W)GGGTPT₁₁₂ motif (numbers in italics refer to B. japonicum HemN₁ and HemN₂) is perfectly conserved in both B. japonicum proteins. The distal cysteine of the H₅₃XPFCX₃CX₂CXC motif, which is characteristic for anaerobic coproporphyrinogen III dehydrogenases (74), is replaced by a phenylalanine in HemN₁. However, the same exchange is also observed in other HemN-like proteins, e.g., those of B. subtilis (HemN and HemZ) (37, 39), M. tuberculosis (17), and Haemophilus influenzae (28). Similarly, F and Q of the LXRNFQGY₃₀₉ motif, which is conserved in the C-terminal portion of HemN proteins from gram-negative bacteria, are replaced by S₃₀₆ and L₃₀₇ in HemN₁, but HemN proteins from gram-positive organisms also deviate from the consensus sequence at these positions.

View larger version (103K): [in this window] [in a new window]   FIG. 7.   Amino acid sequence alignment of the B. japonicum (Bj) HemN proteins with HemN-like proteins of Pseudomonas sp. strain G-179 (Ps) (accession number AAC79446) (11), R. sphaeroides (Rs) (HemN accession number, P95651; HemZ accession number, P51008) (18, 74), E. coli (Ec) (P32131) (70), B. subtilis (Bs) (HemN accession number, P54304; HemZ) (38, 39), and M. tuberculosis (Mt) (P71756) (17). Residues conserved in all nine proteins are highlighted on a solid background and represented by uppercase letters in the consensus line; residues conserved in at least seven proteins are highlighted on a shaded background and represented by lowercase letters in the consensus line.

In conclusion, one is left with the question of the role of the seemingly nonfunctional (but expressed) hemN₁ gene in B. japonicum. It appears unlikely that hemN₁ becomes functional under conditions which were not tested in this study. A hypothetical role of HemN₁ under aerobic conditions as proposed for HemN of P. aeruginosa (63) can be excluded because the hemN₁ gene was not expressed in B. japonicum under these conditions. After all, families of homologous genes (paralogs) are not uncommon in B. japonicum, as documented by two rpoN genes (44), three rpoH genes (50, 51), two fixK genes (2, 52), and five groESL operons (25, 27). However, unlike the newly identified hemN genes, individual members of the previously characterized B. japonicum gene families were shown to be at least partially functional.