|Home | About | Journals | Submit | Contact Us | Français|
Transcription of the Corynebacterium diphtheriae hmuO gene, which encodes a heme oxygenase involved in heme iron utilization, is activated in a heme- or hemoglobin-dependent manner in part by the two-component system ChrA-ChrS. Mutation of either the chrA or the chrS gene resulted in a marked reduction of hemoglobin-dependent activation at the hmuO promoter in C. diphtheriae; however, it was observed that significant levels of hemoglobin-dependent expression were maintained in the mutants, suggesting that an additional activator is involved in regulation. A BLAST search of the C. diphtheriae genome sequence revealed a second two-component system, encoded by DIP2268 and DIP2267, that shares similarity with ChrS and ChrA, respectively; we have designated these genes hrrS (DIP2268) and hrrA (DIP2267). Analysis of hmuO promoter expression demonstrated that hemoglobin-dependent activity was fully abolished in strains from which both the chrA-chrS and the hrrA-hrrS two-component systems were deleted. Similarly, deletion of the sensor kinase genes chrS and hrrS or the genes encoding both of the response regulators chrA and hrrA also eliminated hemoglobin-dependent activation at the hmuO promoter. We also show that the regulators ChrA-ChrS and HrrA-HrrS are involved in the hemoglobin-dependent repression of the promoter upstream of hemA, which encodes a heme biosynthesis enzyme. Evidence for cross talk between the ChrA-ChrS and HrrA-HrrS systems is presented. In conclusion, these findings demonstrate that the ChrA-ChrS and HrrA-HrrS regulatory systems are critical for full hemoglobin-dependent activation at the hmuO promoter and also suggest that these two-component systems are involved in the complex mechanism of the regulation of heme homeostasis in C. diphtheriae.
Iron is an essential nutrient for nearly all organisms since it is the catalytic center for numerous enzymes involved in key cellular processes such as DNA synthesis and electron transport. In biological systems, iron is relatively scarce, since it is largely insoluble at physiological pH levels and most of the iron is sequestered by proteins and heme, which is also associated with protein (33). The dearth of free iron within an organism presents a challenge to pathogenic bacteria in obtaining the necessary iron from the host; thus, pathogens have developed a number of mechanisms for acquiring iron from host molecules such as heme. Corynebacterium diphtheriae, the causative agent of the severe upper respiratory infection diphtheria, has transport systems that are specific for various iron sources including heme (12, 21, 22, 35, 37, 38). In C. diphtheriae, an ATP-binding cassette transporter for heme, encoded by hmuTUV, and a heme oxygenase (HmuO), which releases iron from the protoporphyrin ring of heme, have been described previously (12, 40, 60). Because of the potential toxicity of both iron and heme, cellular levels of these compounds must be controlled. Therefore, in most organisms, expression of proteins involved in the uptake of iron and heme and in the degradation and synthesis of heme are often tightly regulated (6, 12, 18, 19, 20, 22, 27, 28, 30, 31, 34, 36, 41, 50, 51, 56, 57).
It was previously shown that expression of the hmuO promoter in C. diphtheriae is repressed by iron and activated by heme or hemoglobin (41). Transcription from the hmuO promoter is repressed under high-iron conditions, while under iron-depleted conditions, a low level of expression is observed (5, 42). Repression in the presence of iron is due to the activity of DtxR, a global iron-responsive regulator that controls the expression of at least 40 genes in C. diphtheriae, including the tox gene, which encodes the diphtheria toxin (8, 21, 23, 35, 37, 38, 45, 46). High-level hmuO promoter activity is observed only in the presence of a heme source, such as hemoglobin, and optimal hemoglobin-dependent activation is detected under low-iron conditions when DtxR-mediated repression is alleviated (5, 41). Hemoglobin-dependent activation of the hmuO promoter was shown to be mediated in part by the ChrA-ChrS two-component signal transduction system (42). Deletion of either chrS (encoding a sensor kinase) or chrA (encoding a response regulator) resulted in an approximately 10-fold decrease in hemoglobin-dependent activation at the hmuO promoter, as determined using a hmuO-lacZ reporter construct (5). However, neither of these mutations fully abolished hemoglobin-dependent activation at the hmuO promoter, which suggests that an additional activator is involved in the expression of the hmuO promoter.
In this study, we have identified a putative two-component signal transduction system encoded by DIP2267 and DIP2268 (designated hrrA and hrrS, respectively), which has significant amino acid sequence similarity to ChrA and ChrS. Analysis of various mutations in these regulatory systems demonstrates that both of these two-component systems are required for full hemoglobin-dependent activation at the hmuO promoter. We also show that these regulators are involved in hemoglobin-dependent repression of genes encoding enzymes involved in heme biosynthesis.
Bacterial strains used in this study are listed in Table Table1.1. For culturing Escherichia coli, Luria-Bertani (LB) medium (Difco, Detroit, MI) was used, and heart infusion broth (Becton Dickinson, Sparks, MD) containing 0.2% Tween 80 (HIBTW) was used for growing C. diphtheriae strains. Antibiotics were added to LB medium at concentrations of 34 μg/ml for chloramphenicol and 50 μg/ml for kanamycin. Antibiotics were also added to HIBTW at a concentration of 2 μg/ml for chloramphenicol and 50 μg/ml for kanamycin. Ethylenediamine di-(O-hydroxyphenylacetic acid) (EDDA) was added to HIBTW medium to chelate iron at 12.5 μg/ml (unless indicated otherwise). mPGT medium, a semidefined low-iron medium, was used for the heme utilization assays and has been described previously (53). Antibiotics, EDDA, Tween 80, and hemoglobin (human) were obtained from Sigma Chemical Co. (St. Louis, MO).
Plasmids used in this study are listed in Table Table1.1. C. diphtheriae strain C7(−) genomic DNA was used as the template for all PCRs. PCR-amplified DNA fragments were cloned first into a pCR-Blunt II-TOPO vector (Invitrogen). The promoter probe vector pCM502, which was used for the construction of the lacZ promoter fusion reporter plasmids, contains a promoterless lacZ gene, replicates at a low copy number in C. diphtheriae (one to two copies per cell ), and has been described previously (41). Plasmid pCP0-1 contains lacZ fused with the hmuO promoter and has been described previously (42). To construct plasmid phrrS-PO, which contains a promoterless lacZ gene fused to the putative hrrS promoter region, a 285-bp fragment encompassing the entire hrrS-DIP2269 intergenic region plus flanking sequence was amplified with primers 68/69-PO-1 and 68/69-PO-2 and cloned into pCM502. The hrrA promoter-lacZ fusion plasmid, phrrA-PO, was constructed with a 282-bp fragment that amplified the hrrS-hrrA intergenic region with the flanking sequence using primers hrrA-PO-F and hrrA-PO-R. The plasmid that contains the hemA promoter region was made by amplifying a 538-bp fragment that contains 30 bp of the 5′ region of hemA and a 508-bp upstream sequence. This fragment was cloned into the pCM502 vector to create a hemA promoter-lacZ fusion (phemA-PO). While it is possible that in some instances, the measurement of promoter activity from the low-copy-number pCM502 plasmid may not precisely reflect activity of the native chromosomal promoters, our previous experience with pCM502 suggests that differences in expression between plasmid and chromosomally encoded promoters are minimal.
Plasmid pECK18mob2, which was used to construct plasmids utilized in complementation studies, has a copy number of 37 ± 4 per cell and has been previously described (54). To create plasmid pECK-chrS, an 1,851-bp fragment was amplified using primers chrS-U1 and chrA-U2 and was initially cloned into pCR-BluntII-TOPO. This fragment was then moved into pECK18mob2 by excision of the insert with BamHI and PstI. To construct pECK-chrA, an 1,178-bp fragment was amplified with primers chrA-U1 and SDNR, and the remaining steps of the cloning scheme were similar to those described above for pECK-chrS. A 2,424-bp fragment was amplified with primers chrS-U1 and SDNR and cloned as described previously for the other pECK18mob2 derivatives to create pECK-chrSA. To make pECK-hrrS, a 1,536-bp fragment was amplified with primers 2268-pKn-F and 2268-pKn-R2, cloned into pCR-BluntII-TOPO, and then moved into pECK18mob2 after digestion with BamHI and SpeI. A similar scheme was utilized to construct pECK-hrrA, except that a 745-bp fragment was amplified with primers hrrA-F2 and hrrSA-pKn. Plasmid pECK-hrrSA was also constructed in a way similar to that described for pECK-hrrS; however, a 2,035-bp fragment was amplified with primers 2268-pKn-F and hrrSA-pKn. The construction of pKN-hrrS was identical to that described above for pECK-hrrS, except that the BamHI-SpeI fragment containing hrrS was moved into pKN2.6 instead of pECK18mob2. Primers used for plasmid constructions are listed in Table Table22.
Deletions in the C. diptheriae C7(−) genes chrSA, hrrS, and hrrSA were accomplished using an allele replacement technique that has been previously described (55). Mutant construction utilized PCR to clone DNA fragments located upstream and downstream of the region targeted for deletion. To construct the hrrS mutant (C7hrrSΔ), a 585-bp downstream fragment and a 561-bp upstream fragment were amplified and fused together at a common SalI site. The resulting mutant is predicted to encode a truncated HrrS product that has 48 amino acids from the amino terminus fused in frame to 34 amino acids of the carboxyl terminus, deleting 82% of HrrS. Construction of the hrrSA mutant (C7hrrSAΔ) utilized the same upstream fragment described above for C7hrrSΔ along with a 1,936-bp section downstream of hrrA. In this mutant, 48 amino acids of HrrS are predicted to be present, but all of HrrA is deleted. The chrSA mutant (C7chrSAΔ) contains a 1,811-bp deletion within the chrSA coding region, and the resulting product is predicted to contain eight residues at the amino terminus of ChrS fused in frame to four residues from the carboxyl terminus of ChrA. Primers used for mutant construction are listed in Table Table2.2. PCR was used to confirm the mutations in all of the deletion mutants (data not shown). To construct the hrrA mutant (C7hrrA−), a 408-bp internal fragment of hrrA was amplified using primers 2267-KO-F and 2267-KO-F. These primers contain stop codons at the putative phosphorylation site (D54) and within the second helix of the predicted helix-turn-helix motif (H186), respectively (indicated in bold in Table Table2).2). This fragment was inserted into pKN2.6, and the resulting plasmid was used to make a vector integration mutant of C7(−) as previously described (43).
Bacterial cultures were inoculated from frozen stocks into 1 ml of HIBTW broth and grown overnight at 37°C with shaking. The overnight cultures were then inoculated at an optical density at 600 nm (OD600) of 0.2 into fresh HIBTW containing 12.5 μg/ml EDDA. These cultures were grown at 37°C with shaking for 5 to 6 h, at which time the OD600 was measured. Cultures were washed and resuspended in mPGT and then inoculated at an OD600 of 0.03 into fresh mPGT containing 3.6 μg/ml EDDA and either 10 μg/ml hemoglobin or 50 μg/ml hemoglobin. These cultures were incubated at 37°C overnight (18 to 24 h), and the optical density was measured.
To determine the effects of hemoglobin on growth in liquid medium, cultures inoculated from frozen stocks were grown overnight at 37°C in 4 ml of HIBTW and then inoculated at an OD600 of 0.1 into fresh HIBTW that contained either hemoglobin at 140 μg/ml or no added hemoglobin. Cultures were grown at 37°C with shaking, and OD600 measurements of the bacterial culture were made at various time points as indicated.
Overnight cultures of C. diphtheriae strains containing lacZ fusion constructs were grown in HIBTW medium and then inoculated at an OD600 of 0.1 into fresh HIBTW medium that contained various supplements as indicated. After 5 to 6 h of growth at 37°C with shaking (at mid-log phase), LacZ activity was determined as previously described (44). Hemoglobin was used as a heme source instead of heme, since higher expression in the chrSA mutants is observed with hemoglobin, and hemoglobin is less toxic than heme (5).
A previous report indicated that in wild-type C. diphtheriae strain C7(−), expression of the hmuO promoter was repressed by iron and activated in the presence of a heme source, such as hemoglobin, with optimal expression occurring under low-iron conditions in the presence of hemoglobin (5, 41). Hemoglobin-dependent expression was significantly reduced in a strain that carried a deletion in either chrS or chrA (5). In this study, we constructed a mutant of C7(−) in which both chrS and chrA were deleted (C7chrSAΔ) and observed that the expression of the hmuO promoter in this double mutant was similar to that reported previously for the single-mutant strains C7chrSΔ and C7chrAΔ; therefore, only the results for C7chrSAΔ are reported here (5) (Table (Table33).
During growth in the absence of hemoglobin, expression from the hmuO promoter in the C7chrSAΔ mutant was similar to that in the wild-type strain (Table (Table3).3). In the presence of hemoglobin, a reduction in hemoglobin-dependent activation at the hmuO promoter was observed in C7chrSAΔ compared to that detected in the wild-type strain (Table (Table3).3). However, a considerable level of expression was maintained by the C7chrSAΔ mutant grown in low-iron medium containing hemoglobin (−Fe +Hb); this level of expression was sixfold higher than that under growth conditions in medium with low iron alone (−Fe), which suggests that a second hemoglobin-dependent regulator is involved in hmuO promoter expression (Table (Table33).
We conducted a BLAST search of the recently completed C. diphtheriae genome and identified a putative two-component system, encoded by the genes DIP2268 (a sensor kinase) and DIP2267 (a response regulator) (10), which shares similarity with ChrS and ChrA. The DIP2268 and DIP2267 genes are designated hrrS and hrrA (heme responsive regulator sensor and activator), respectively, based on the proposed functions of these genes.
ChrA-ChrS and HrrA-HrrS share significant sequence similarity to the NarX-NarL family of two-component systems (2). ChrS contains 417 amino acids, while HrrS is 459 amino acids long, and the two proteins are approximately 30% identical. Both ChrS and HrrS appear to have two distinct domains, an amino-terminal “sensor domain” in which ChrS is predicted to have five transmembrane regions, and HrrS has four putative membrane-spanning segments. The two proteins share almost 40% similarity throughout this amino-terminal domain, a region that typically shows little or no conservation among most sensor kinases. The carboxyl-terminal kinase domains of ChrS and HrrS exhibit significant similarity (~55%), and both kinases contain a conserved histidine residue, which is the predicted site for autophosphorylation.
The putative response regulators ChrA and HrrA contain 199 and 212 residues, respectively, and are approximately 50% identical. Both proteins contain a conserved aspartate residue, which is the putative site for phosphorylation, and a helix-turn-helix motif that is predicted to be required for DNA binding.
The genetic organization of hrrSA is similar to that of chrSA, with the sensor kinase gene proximal to the putative promoter region and directly upstream of the gene encoding the response regulator (5) (Fig. (Fig.1).1). While an 81-base pair intergenic region separates hrrS and hrrA, no gap is present between chrS and chrA. The start codon of chrA (ATG) overlaps the stop codon of chrS (TGA) (i.e., ATGA). Genes hrrS and hrrA are flanked by two open reading frames (ORFs), DIP2266 and DIP2269, which are divergently transcribed from that of hrrSA and have no significant database matches (Fig. (Fig.1).1). Because of the similarity between HrrA-HrrS and ChrA-ChrS, we sought to determine if the HrrA-HrrS two-component system is involved in the hemoglobin-dependent activation of the hmuO promoter.
To assess the effect of mutations at the hrrSA loci on the expression of the hmuO promoter, mutations in the hrrSA genes were constructed in both the wild-type and the chrSA mutant backgrounds (Fig. (Fig.11 and Table Table1).1). Growth in −Fe +Hb medium resulted in a statistically significant decrease (P < 0.05) in hmuO promoter expression in C7hrrA− (the hrrA mutant), relative to that in the wild-type strain (Table (Table3).3). This result suggests that HrrA contributes to hemoglobin-dependent hmuO promoter expression. Expression of the hmuO promoter was also analyzed in the hrrS deletion mutant C7hrrSΔ and in the C7hrrSAΔ mutant (from which both hrrS and hrrA were deleted). In contrast to the reduced expression observed for the mutant C7hrrA−, the levels of expression in mutants C7hrrSΔ and C7hrrSAΔ were not statistically significantly different from that of the wild type (Table (Table3).3). Surprisingly, when the C7hrrSΔ mutant strain was grown under −Fe medium conditions, there was a 14-fold increase in the level of expression compared to that of the wild-type or any of the other mutant strains (Table (Table3).3). This phenotype of C7hrrSΔ was complemented by the cloned hrrS gene on plasmid pECK18-hrrS, and expression was reduced to wild-type levels (data not shown). The reason for the increased expression of the hmuO promoter in low-iron medium is as yet unknown, but it requires a functional hrrA gene in addition to the chrSA genes, since deletion of the chrS or the chrA gene in a hrrS deletion background abolished activation (Table (Table3).3). The three mutants that were constructed with the hrrSA genes (C7hrrA−, C7hrrSΔ, and C7hrrSAΔ) exhibited different phenotypes with regard to the regulation of hmuO expression under various iron and hemoglobin conditions, suggesting that the functional relationship between HrrS and HrrA is more complex than that of a typical cognate sensor kinase-response regulator pair.
Expression of the hmuO promoter was also measured in strains that were defective in both sensor kinase genes chrS and hrrS (C7chrSΔ/hrrSΔ), in both response regulator genes chrA and hrrA (C7chrAΔ/hrrA−), and in the mutant C7chrSAΔ/hrrSAΔ, from which both chrS-chrA and hrrS-hrrA were deleted. In all three mutant strains, hemoglobin-dependent activation was abolished, and the level of activity measured for low-iron medium containing hemoglobin was similar to that observed for medium with low iron alone (Table (Table3).3). Expression under all other conditions was similar to that observed for the C7chrSAΔ mutant (Table (Table3).3). These results provide strong evidence that HrrA as well as HrrS contributes to hemoglobin-dependent hmuO expression by demonstrating that both HrrA and HrrS are required for the hemoglobin-dependent expression observed for C7chrSAΔ.
Since the previous results indicate that both ChrA-ChrS and HrrA-HrrS are required for full hemoglobin-dependent activation of the hmuO promoter, we sought to determine whether there is cross talk between these two signal transduction systems or, more specifically, whether the sensor kinase from one of the two-component systems activates the response regulator of the other system. Interestingly, the strain from which chrA and hrrS were deleted (C7chrAΔ/hrrSΔ) maintained hemoglobin-dependent activation of the hmuO promoter; a fourfold higher activity level was seen with the strain grown in −Fe +Hb medium than that grown in −Fe medium (Table (Table3).3). This strain has functional copies of the genes encoding both the sensor kinase ChrS and the response regulator HrrA. These findings suggest that ChrS has the ability to activate the HrrA response regulator. Mutant strains that had only a functional copy of chrS (C7chrAΔ/hrrSAΔ) or hrrA (C7chrSAΔ/hrrSΔ) did not exhibit hemoglobin-dependent activation (Table (Table3),3), which suggests that both ChrS and HrrA are needed for activation of the hmuO promoter. The mutant strain C7chrSΔ/hrrA−, which carries functional copies of both the sensor kinase HrrS and the response regulator ChrA, showed no hemoglobin-dependent activation, since the activity in low-iron medium containing hemoglobin was the same as that observed for medium with low iron alone (Table (Table3).3). This result suggests that HrrS does not activate ChrA.
In the strain from which both two-component systems were deleted, C7chrSAΔ/hrrSAΔ, no hemoglobin-dependent activation of hmuO promoter expression was observed (Table (Table3).3). To determine if the cloned gene chrSA or hrrSA could restore hemoglobin-dependent activation of the hmuO promoter in this double-mutant strain, we constructed various plasmids using the Corynebacterium-E. coli shuttle vector pECK18mob2 (54). These constructs were moved into the mutant C7chrSAΔ/hrrSAΔ carrying pCP0-1, and β-galactosidase activity was measured. The vector alone in C7chrSAΔ/hrrSAΔ has no effect on expression (Table (Table4).4). As shown in Table Table4,4, the cloned chrSA genes on pECK-chrSA restored hemoglobin-dependent activation in C7chrSAΔ/hrrSAΔ to levels similar to those observed for wild-type cells (Table (Table3).3). When the double mutant C7chrSAΔ/hrrSAΔ carried a plasmid containing the cloned hrrSA genes (pECK-hrrSA), an increase in hemoglobin-dependent activation at the hmuO promoter was observed, and in −Fe +Hb medium, expression was greater than 10-fold higher than that observed for −Fe medium (Table (Table4).4). The expression level observed for mutant C7chrSAΔ/hrrSAΔ with pECK-hrrSA was similar to that measured in the C7chrSAΔ mutant (Table (Table3),3), which suggests that HrrA-HrrS is responsible for the hemoglobin activation observed for C7chrSAΔ. These findings provide strong evidence that the HrrA-HrrS two-component system contributes to the hemoglobin-dependent activation at the hmuO promoter.
The transcriptional studies with the chrSA mutants indicate that the HrrA-HrrS system contributes about 20 to 25 units of the hemoglobin-dependent activity at the hmuO promoter (Table (Table33 and and4);4); however, the expression level in the hrrSA double mutant was not statistically significantly different from that in the wild type (Table (Table3).3). An inability to demonstrate a statistically significant reduction in hemoglobin activation in the hrrSA mutant may be due to the inherent variability of the LacZ assays, since only a 20- to 25-unit reduction would be expected.
No hemoglobin-dependent activation at the hmuO promoter was observed for the C7chrSAΔ/hrrSAΔ strain that carried a plasmid with only the chrS, hrrS, or hrrA gene (data not shown). However, approximately 20 units of activity was measured for the mutant strain carrying pECK-chrA (chrA+) grown in low-iron medium with or without hemoglobin (data not shown), which suggests that the ChrA response regulator alone, when expressed from a multicopy plasmid, can activate the hmuO promoter independently of a heme source. It was previously observed in E. coli that a high-copy number plasmid carrying chrA alone can activate transcription from the hmuO promoter in the absence of a heme source (42).
It was previously shown that strains with mutations in hmuO, which encodes a heme oxygenase, have a defect in the utilization of heme as an iron source (12). The low level of hmuO promoter activity in the chrSA hrrSA double mutants suggests that these strains may have a reduced ability to utilize heme or hemoglobin as an iron source due to a decreased production of HmuO. To test the capacity of various mutants to use heme iron, strains were grown in a low-iron semidefined medium (mPGT) in the presence or absence of hemoglobin. Neither the mutants nor the wild-type strains were able to grow in mPGT medium unless an iron source, such as hemoglobin, was provided (data not shown). When hemoglobin was added to the mPGT medium at 10 μg/ml, the mutant C7chrSΔ/hrrSΔ showed reduced growth relative to that of the wild-type strain but exhibited growth similar to that of a hmuO deletion strain (C7hmuOΔ) (Fig. (Fig.2A).2A). At an increased concentration of hemoglobin (50 μg/ml), the growth of mutant C7chrSΔ/hrrSΔ was similar to that of the wild type, while C7hmuOΔ grew to a lower density (Fig. (Fig.2B).2B). Results similar to that observed for the mutant C7chrSΔ/hrrSΔ were obtained using C7chrAΔ/hrrA− and C7chrSAΔ/hrrSAΔ (data not shown). No defect in the use of hemoglobin as an iron source was observed for the single mutants C7chrSΔ and C7hrrSΔ (Fig. (Fig.2A).2A). These findings indicate that there is a correlation between low expression of hmuO and a reduced ability to utilize hemoglobin as an iron source. At higher concentrations of hemoglobin (50 μg/ml), it is likely that in the mutant C7chrSΔ/hrrSΔ, sufficient HmuO is expressed to degrade heme and provide adequate amounts of iron for growth.
In routine studies, we observed that the mutant strain C7hrrSΔ, which carries a deletion of hrrS, has reduced growth in rich medium (HIBTW) compared to that of the wild-type C7(−) strain. However, it was observed that when heme or hemoglobin was added to the medium, mutant C7hrrSΔ cultures grew to densities that were similar to those of the wild type. To further characterize this phenotype, mutant C7hrrSΔ was inoculated in HIBTW medium with and without hemoglobin and growth was monitored over time. Wild-type C. diphtheriae strain C7(−) and the other mutants were also included as controls. The growth of wild-type cultures was not affected by the presence of hemoglobin in the HIBTW medium (Fig. (Fig.3A).3A). However, in comparison to that of the wild type, C7hrrSΔ cultures grown in the absence of hemoglobin reached lower densities at all time points (Fig. (Fig.3A).3A). In the presence of hemoglobin, the growth kinetics of the mutant C7hrrSΔ was nearly identical to that of the wild type (Fig. (Fig.3A),3A), which indicates that C7hrrSΔ requires an exogenous heme source to achieve wild-type levels of growth. In contrast to C7hrrSΔ, all other mutants tested for the hrrSA or chrSA genes (C7chrSΔ, C7chrSAΔ, C7hrrA−, C7hrrSAΔ, C7chrSΔ/hrrSΔ, and C7chrSΔ/hrrA−) did not exhibit a requirement for a heme source during growth in HIBTW medium (data not shown). The growth defect of C7hrrSΔ in rich medium is specific to the hrrS gene, since the cloned hrrS gene (pKN-hrrS) in C7hrrSΔ restored it to wild-type levels of growth in HIBTW medium (Fig. (Fig.3B3B).
The requirement of a heme source for optimal growth of mutant C7hrrSΔ may be due to a defect in the expression of heme biosynthesis genes or in heme metabolism. Biosynthesis could be affected in the mutant strain C7hrrSΔ if the HrrA-HrrS two-component system is involved in the expression of heme biosynthetic genes. The genome of C. diphtheriae encodes several predicted heme biosynthetic genes, and four of these genes, hemA, hemC, hemD, and hemB, appear to be organized as an operon (10). The products of these genes are required for early steps in the heme biosynthetic pathway, and the first gene in this group, hemA, encodes glutamyl-tRNA reductase, which acts at the first committed step in heme biosynthesis (1, 4, 17). To investigate whether the mutant C7hrrSΔ exhibits a reduction in the expression of these heme biosynthetic genes, a transcriptional reporter construct was made with the putative promoter region upstream of hemA fused to a lacZ cassette, and β-galactosidase expression in wild-type and mutant strains was assessed.
Expression of the hemA promoter in the wild-type strain is not iron regulated, since no differences in expression were observed between growth under high- and low-iron conditions (data not shown). However, transcription from the hemA promoter was repressed by hemoglobin in the wild-type C7(−) strain; expression of hemA was approximately threefold lower in the presence of hemoglobin than in the absence of hemoglobin (Table (Table5).5). In the mutant C7hrrSΔ, hemA promoter expression in medium without hemoglobin was twofold lower than that of the wild type (Table (Table5).5). The reduced expression from the hemA promoter in mutant C7hrrSΔ may render the strain unable to synthesize adequate levels of heme during growth in HIBTW medium and, thus, could account for the requirement of a heme source for optimal growth of this strain in HIBTW medium (Fig. (Fig.3).3). A growth requirement for heme was not observed for mutants C7hrrA− and C7hrrSAΔ, and in contrast to the reduced expression of the hemA promoter observed for mutant C7hrrSΔ, expression of the hemA promoter in mutants C7hrrA− and C7hrrSAΔ was approximately twofold higher than that observed for the wild type (Table (Table5).5). In mutants C7hrrA− and C7hrrSAΔ, the activity level measured in the hemoglobin-containing medium was approximately threefold lower than in the absence of hemoglobin, indicating that hemA promoter expression is still repressed in a hemoglobin-dependent manner (Table (Table5).5). Expression of the hemA promoter in various chrSA mutants (C7chrAΔ, C7chrSΔ, and C7chrSAΔ) was found to be similar to that of the wild type (Table (Table55 and data not shown).
Expression of the hemA promoter was also analyzed in several chrSA hrrSA double mutants. In strains with mutations in both sensor kinases (C7chrSΔ/hrrSΔ), both response regulators (C7chrAΔ/hrrA−), or both two-component systems (C7chrSAΔ/hrrSAΔ), expression of hemA was increased relative to that of the wild-type strain, both in the presence and absence of hemoglobin (Table (Table5),5), which indicates that hemoglobin-dependent repression was abolished in these double mutants (Table (Table5).5). These results indicate that both ChrA-ChrS and HrrA-HrrS contribute to the hemoglobin-dependent repression at the hemA promoter, and mutations in both systems are required to abolish repression. It should also be noted that the mutations that abolished hemoglobin-dependent repression of hemA expression (C7chrSΔ/hrrSΔ, C7chrAΔ/hrrA−, and C7chrSAΔ/hrrSAΔ) also resulted in a loss of hemoglobin-dependent activation of the hmuO promoter (Table (Table33).
To determine if there was cross talk between ChrA-ChrS and HrrA-HrrS at the hemA promoter, expression of hemA was analyzed in mutants C7chrAΔ/hrrSΔ (chrS+ hrrA+) and in C7chrSΔ/hrrA− (chrA+ hrrS+). Expression of the hemA promoter in both of these mutants in the presence of hemoglobin was approximately threefold lower than that in the absence of hemoglobin (Table (Table5),5), which indicates that hemoglobin-dependent repression occurs in these strains and, moreover, these results provide evidence for cross talk between ChrS and HrrA and between HrrS and ChrA.
In a recent study that examined DtxR-regulated genes in Corynebacterium glutamicum, a DtxR binding site was found in the intergenic region of genes encoding a sensor kinase (CgtS11) and a response regulator (CgtR11), which have extensive similarity to HrrS (56% identity) and HrrA (86% identity), respectively (59). Expression of the response regulator CgtR11 was shown to be regulated by iron and DtxR in C. glutamicum (59). In C. diphtheriae, hrrS and hrrA are separated by 81 base pairs; however, there are no significant sequence similarities in this intergenic region between C. diphtheriae and C. glutamicum, and there is no evidence for a DtxR binding site in C. diphtheriae in this region. To determine if there is promoter activity in the hrrS-hrrA intergenic region, a DNA fragment containing this region was amplified and cloned into the lacZ fusion reporter construct pCM502. In the wild-type C. diphtheriae strain, expression from this hrrA-specific promoter construct (phrrA-PO) was approximately 10-fold higher than that observed for the hrrS promoter construct (phrrS-PO), which harbors the region upstream of hrrS (Fig. (Fig.11 and and4).4). In contrast to the iron-dependent repression of CgtR11 in C. glutamicum, no iron regulation was observed for either the hrrS promoter or the hrrA promoter in C. diphtheriae (data not shown). However, in the presence of hemoglobin, hrrA promoter activity was reduced approximately twofold, suggesting that this specific hrrA promoter is repressed by hemoglobin (Fig. (Fig.4).4). The hemoglobin-dependent repression at the hrrA promoter is mediated, at least in part, by the HrrA-HrrS two-component system, since this repression was abolished in mutant strains C7hrrA−, C7hrrSAΔ, and C7hrrSΔ (Fig. (Fig.44 and data not shown). Mutations in the chrSA genes had no effect on expression at the hrrA promoter (data not shown). Examples of differential regulation for sensor kinase and response regulator pairs have been observed with other systems (13, 59). The findings reported here suggest that HrrA is needed at levels higher than those for HrrS, and since we have shown that HrrA activity can be stimulated by both HrrS and ChrS, it is possible that increased levels of HrrA are needed due to its interaction with multiple sensor kinases and its requirement to act at multiple promoters.
Two-component signal transduction systems provide a mechanism by which bacteria can rapidly adapt to changing environmental conditions. Environmental signals are integrated through the action of a histidine kinase and a partnered response regulator, which is often a transcription factor; thus, a common effect of the environmental stimulus is the up or down regulation of one or more genes. C. diphtheriae encodes 11 predicted two-component signal transduction systems (10), and until now, only the ChrA-ChrS system has been characterized. Although ChrA-ChrS was the first bacterial heme-responsive two-component system described, several heme-responsive transcriptional regulators have been reported (6, 18, 19, 20, 50, 56, 57). Moreover, two additional signal transduction systems with amino acid sequence similarities to ChrA-ChrS have been described recently, CgtR11-CgtS11 in C. glutamicum and SenR-SenS in Streptomyces reticuli (32, 60). An environmental signal involved in the activation of these systems has yet to be identified.
In this study, we have identified an additional C. diphtheriae two-component signal transduction system encoded by hrrA and hrrS, which is involved in the hemoglobin-dependent activation at the hmuO promoter and in the hemoglobin-dependent repression at putative promoters upstream of hemA and hrrA. While the contribution of HrrA-HrrS to hemoglobin activation at hmuO is severalfold lower than that of ChrA-ChrS, the HrrA-HrrS system is critical for wild-type levels of hemoglobin-activated expression under low-iron conditions.
An unusual finding from this study was that disparate results were observed among the hrrSA mutants, which suggests that the HrrA-HrrS system does not conform to the model of the typical two-component system, such as the ChrA-ChrS system. Observations similar to what we have shown in the HrrA-HrrS system, in which a mutation in a gene encoding a sensor kinase results in a different phenotype than that from a mutant that carries a defective cognate response regulator, have been reported previously (25, 52). In some of these cases, the disparate characteristics are attributed to cross talk with a second two-component system, and examples of cross talk between noncognate sensor kinase-response regulator pairs have been described (14, 15, 26, 58, 61). In the present study, evidence of cross talk was observed between the ChrA-ChrS and HrrA-HrrS systems at the hmuO and hemA promoters. Hemoglobin-dependent activation of the hmuO promoter was observed for C7chrAΔ/hrrSΔ, a strain with functional ChrS and HrrA proteins (Table (Table3).3). Cross talk between ChrS and HrrA provides a possible explanation for the wild-type levels of hmuO expression observed for the hrrS mutant C7hrrSΔ (Table (Table3);3); this could occur through activation of both ChrA and HrrA (via phosphorylation) by ChrS to promote wild-type levels of expression of hmuO.
We also provided evidence that cross talk between the HrrA-HrrS and ChrA-ChrS two-component systems occurred at the hemA promoter, and this phenomenon may account for some of the phenotypes observed for the hrrSA mutants, although alternative explanations are possible. The physiological relevance of the cross talk reported in this study is not known, and we have no direct evidence that cross talk occurs in the wild-type strain. Nevertheless, cross-regulation between these two-component systems could facilitate or “fine tune” the control of heme homeostasis by the bacteria. The hmuO gene encodes a heme oxygenase, an enzyme that is involved in heme degradation, while the hemA operon controls heme biosynthesis; it is perhaps not surprising that these two opposing activities (i.e., heme degradation and heme synthesis) exhibit some level of coordinated regulation.
The findings in this study suggest that HrrA and ChrA function as activators at the hmuO promoter and as repressors of hemA expression, and examples of two-component systems possessing this type of dual function have been described previously (7, 9, 11, 16, 24, 48). While DNA binding sites for ChrA and HrrA have not been identified, it was previously shown that hemoglobin-dependent regulation of hmuO requires a 50-bp sequence upstream of the promoter, a region which may harbor binding sites for ChrA and/or HrrA (42). A previous study showed that the cloned chrA gene is able to activate hmuO expression in E. coli, suggesting direct activation by ChrA at the hmuO promoter (42). However, similar studies with the cloned hrrA gene in E. coli failed to demonstrate activation of hmuO expression (data not shown), which suggests that HrrA does not act directly at the hmuO promoter, although alternative explanations are possible, such as poor expression of hrrA or a requirement for additional activating factors or environmental signals.
A model that describes the transcriptional regulation of the hmuO and hemA promoters by the HrrA-HrrS and ChrA-ChrS systems is presented in Fig. Fig.5.5. The model does not account for all the mutant phenotypes but rather focuses on the most significant results from this study to provide a description of the complex regulation observed at the hmuO and hemA promoters. The model predicts that in the absence of hemoglobin or any heme source, the HrrA-HrrS and ChrA-ChrS systems are inactive at the hmuO promoter (Fig. (Fig.5A).5A). Under high-iron conditions in the absence of heme, hmuO transcription is fully inhibited due to DtxR-mediated repression, while in low-iron medium, hmuO is expressed only at a low level, since ChrA and HrrA are not activated (Fig. (Fig.5A).5A). In the presence of a heme source, it is predicted that the detection of heme or hemoglobin (possibly at the cell surface) by the ChrS and HrrS sensor kinases results in the phosphorylation and subsequent activation of the response regulators ChrA and HrrA, respectively, such that under low-iron conditions, expression of hmuO is fully activated by ChrA and HrrA, either by direct binding of these activators upstream of the promoter (as shown in Fig. Fig.5B)5B) or through an indirect mechanism that may involve additional, but as-yet-unknown, regulatory factors. The mutant studies indicated that the ChrA-ChrS systems provides >80% of the activity under low-iron conditions in the presence of hemoglobin (Table (Table33 and and4).4). In high-iron medium containing hemoglobin, DtxR-mediated repression of hmuO appears to be partially reversed by the ChrA-ChrS system (Fig. (Fig.5B).5B). HrrA-HrrS appears to have no affect on expression under these conditions, since no activity is observed in high-iron medium in the chrSA mutants (Table (Table3,3, HrrAS+).
Since iron regulation was not observed at the hemA promoter in the wild-type strain, the model shown in Fig. 5C and D only considers hemA expression in the presence and absence of hemoglobin. This model predicts that both the HrrA-HrrS and the ChrA-ChrS systems contribute to the hemoglobin-dependent repression of hemA expression. This proposal is based on the analysis of the various mutants which shows that the deletion of either system alone had only a minimal effect on repression, while deletion of both systems abolished hemoglobin-dependent repression (Fig. (Fig.5D5D and Table Table5).5). Although in the absence of hemoglobin disparate results were observed between hrrS and hrrA hrrSA mutants, the data overall suggest that the HrrA-HrrS system is involved in the repression of hemA in the absence of a heme source (Fig. (Fig.5C).5C). It is unclear what role, if any, the ChrA-ChrS system has in regulating hemA expression in the absence of hemoglobin, since mutations in the chrSA genes had no effect on expression (Fig. (Fig.5C5C and Table Table55).
The findings from this study suggest that the ChrA-ChrS and HrrA-HrrS two-component systems of C. diphtheriae regulate heme homeostasis through the activation and repression of genes involved in heme degradation and heme biosynthesis. An understanding of the mechanism for ChrA-ChrS and HrrA-HrrS activation and repression at the various promoter regions will require additional studies that focus on putative interactions among the various purified activators and repressors (DtxR) and on the identification of putative binding sites for these response regulators.
We thank Olaf Schneewind for providing the allele exchange vector and E. coli strain S17-1. We also thank Karen Meysick and Scott Stibitz for critical reading and helpful comments on the manuscript.
Editor: V. J. DiRita
Published ahead of print on 12 March 2007.