HasB, the Serratia marcescens TonB Paralog, Is Specific to HasR

Abstract

Serratia marcescens possesses two functional TonB paralogs, TonB_Sm and HasB, for energizing TonB-dependent transport receptors (TBDT). Previous work had shown that HasB is specific to heme uptake in the natural host and in Escherichia coli expressing the S. marcescens TBDT receptor HasR, whereas the S. marcescens TonB and E. coli TonB proteins function equally well with various TBDT receptors for heme and siderophores. This has raised the question of the target of this specificity. HasB could be specific either to heme TBDT receptors or only to HasR. To resolve this question, we have cloned in E. coli another S. marcescens heme receptor, HemR, and we show here that this receptor is TonB dependent and does not work with HasB. This demonstrates that HasB is not dedicated to heme TBDT receptors but rather forms a specific pair with HasR. This is the first reported case of a specific TonB protein working with only one TBDT receptor in one given species. We discuss the occurrence, possible molecular mechanisms, and selective advantages of such dedicated TonB paralogs.

Iron is an essential ion for most living organisms. Although abundant on earth, it is not easily available due to its insolubility under aerobic conditions. To retrieve environmental iron from ground and water, most microbes excrete siderophores that solubilize iron and return it to cell surface receptors (4). For microbes developing in the living host, and particularly in vertebrates, heme is the most abundant source of iron (35). Because of its high reactivity, it is usually not free but rather is bound to proteins intracellularly, mostly to hemoglobin and myoglobin and extracellularly to hemopexin and albumin. Bacteria have cell surface receptors for heme, for each type of hemoprotein, or both. It is not uncommon for one species to have several heme/hemoprotein receptors (34). In addition, several bacteria secrete hemophores, which are small proteins that scavenge free or protein-bound heme due to their high affinity for heme and return it to surface receptors (6).

In gram-positive bacteria, cell-wall-anchored surface proteins bind heme and/or hemoproteins and transfer heme to ABC transporters that internalize it through the plasma membrane using ATP hydrolysis as an energy source (19).

In gram-negative bacteria, heme sequentially crosses the two membranes. Both steps are energy-dependent processes. Transport through the outer membrane involves specific surface-exposed receptors powered by the proton motive force supplied by an inner membrane machine: the TonB-ExbB-ExD multiprotein complex (11). Transport through the inner membrane involves periplasmic binding protein-dependent ABC transporters energized by ATP hydrolysis (17, 31).

Whereas gram-positive heme surface receptors are heterogeneous, gram-negative heme outer membrane receptors belong to a family of TonB-dependent transport (TBDT) receptors that comprises heme, hemoproteins, hemophore, ferrisiderophore, B₁₂ vitamin, transferrin, and lactoferrin receptors. X-ray crystal structures have been determined only for a few siderophore receptors, but three-dimensional modeling has predicted that other receptors belonging to this family may have a similar folded structure. Known three-dimensional structures show a plug inside a β-barrel organization. The N-terminal plug domain closes the receptor pore and is exposed to the periplasm, where it can make contact with TonB (10). Binding of the ferrisiderophores triggers conformational changes in the plug domain, leading to formation of a channel for the ligand.

In addition, many outer membrane receptors involved in iron uptake (for citrate, pyoverdin, heme, and hemophore) are positively regulated by their ligand via extracytoplasmic function sigma factors and membrane-bound anti-sigma factors (5). These receptors have an extra N-terminal domain ranging from 50 to 100 amino acids, referred to as the signaling domain. The signaling cascade also requires the TonB complex. Ligand binding to its cognate receptor induces a TonB-dependent conformational change in the signaling domain that releases extracytoplasmic function sigma from anti-sigma inhibition and thereby leads to transcription of the receptor gene-containing operon (5).

Whereas receptors have a low degree of primary sequence homology, reflecting the diversity of ligands, the TonB/ExbB-ExbD complex is highly conserved and is able to energize several receptors. The TonB protein interacts directly with the receptors, whereas ExbB and ExbD make contact between themselves and with the TonB membrane anchor, playing a role in TonB stabilization and energy conveying to the outer membrane (27). The interactions between TonB and the TBDT receptors involve the carboxy terminus of TonB and several regions of the receptors. Genetic and structural studies have revealed the crucial role of contact between the third and fourth TonB β strands and a short peptide at the N terminus of the receptors named the TonB box, which is conserved on all receptors. Such an interaction has been visualized in two Escherichia coli TonB-receptor cocrystal structures as a β-strand exchange between the two protein segments (26, 30). Other interactions between TonB and the TBDT receptors involve several periplasmic loops on the receptors (26).

Genes encoding TonB, ExbB, and ExbD are present in all thus-far-sequenced gram-negative genomes. They display diverse types of genetic organization. Genes exbB and exbD are usually linked and located in one iron-regulated operon. Occasionally, tonB is clustered with exbB-exbD or with genes encoding ferrisiderophore or heme uptake systems. Unlike E. coli, which has a unique set of tonB-exbB-exbD genes, some bacteria have several paralogs, not all of them functional. Vibrio cholerae has one TonB-ExbB-ExbD system encoded by the larger replicon which is able to complement an E. coli tonB mutation for most TBDT receptor-mediated functions (23). A second TonB-ExbB-ExbD machine is encoded by the smaller V. cholerae replicon, which is located in the heme permease hutBCD operon (23). Both systems exhibit redundant functions for heme and siderophore uptake, with some distinct specificities. The TonB1 system is required for schizokinen uptake, and TonB2 is required for enterobactin utilization (29). The TonB1 system allows the use of heme at a wide range of medium osmolarities, whereas TonB2 promotes heme uptake only at low osmolarity (21). In Vibrio anguillarum, two TonB systems function in iron uptake, but only one is essential for virulence (32). In V. cholerae and V. anguillarum, the TonB2 complex requires an additional protein named TtpC (32). In Pseudomonas aeruginosa, several TonB-ExbB-ExbD complexes have been characterized. However, most heme and iron uptake functions rely on a TonB1 set (33). A third TonB was shown to be required for normal twitching mobility and extracellular assembly of type IV pili (14). In Actinobacillus pleuropneumoniae, two TonB systems are also present, and one was shown to be more important for virulence (1).

Serratia marcescens DNA sequence analysis has indicated the presence of several tonB-exbB-exbD genes (six tonB paralogs and two exbB-exbD pairs). Among their products, two TonB proteins have been shown to be functional: S. marcescens TonB (TonB_Sm) and HasB. TonB_Sm is encoded by a gene unlinked to any exbB-exbD genes or to any iron/heme transport genes. When expressed in E. coli, it complements most E. coli TonB (TonB_Ec) natural functions (12). HasB is encoded by a gene located in the has operon dedicated to hemophore-bound heme acquisition (25). This operon encodes HasR, the hemophore outer membrane receptor which is also a heme receptor; HasA, the hemophore; and HasD and HasE, inner membrane components of the type I secretion system required for HasA secretion and HasB. The has operon is negatively regulated by iron and positively regulated by a sigma anti-sigma (HasI-HasS) signaling cascade triggered by heme-loaded hemophore binding to HasR. The Has system is functional in an E. coli recombinant strain for free or hemophore-bound heme uptake and for hemophore-mediated positive regulation (2, 13). Both heme transport and heme signaling are dependent on the host TonB-ExbB-ExbD complex (13, 28). TonB_Sm complements the TonB_Ec protein for heme transport and heme signaling. In a previous work, we showed that HasB6, a HasB allele mutated in its transmembrane domain so as to be functional in E. coli, complements the TonB_Ec protein for heme transport and heme signaling via the Has system (25). However, HasB6 does not complement other TonB functions such as iron siderophore uptake, φ80 infection, or killing by colicin B. These results showed that HasB6 was specific to heme uptake and raised the question of the target of this specificity. HasB6 could be specific to heme TBDT receptors. At least one heme receptor other than HasR is present and functional in S. marcescens, since hasR mutants are still able to use heme, albeit less efficiently than a wild-type strain. Alternatively, HasR-HasB could form a specific pair. In S. marcescens, the iron regulation of HasR and HemR receptors and the presence of several TonB paralogs complicate the study of each TonB function. Thus, experiments were performed in E. coli.

In the present study, we have cloned, in E. coli, another heme receptor, HemR. We show that this receptor is functional and TonB dependent and does not function with HasB6, thereby demonstrating that HasB is not dedicated to heme TBDT receptors but rather forms a specific pair with HasR.

MATERIALS AND METHODS

Bacterial strains and plasmids.

The bacterial strains and plasmids used in the present study are listed in Table 1.

TABLE 1.

Strains and plasmids used in this study

Strain or plasmid	Genotypea	Source or reference
Strains
S. marcescens
SM365	Wild type	Laboratory collection
SMH3	Hem− Strr	28
SMH3 (hasR)	Hem−hasR::aphA Strr Kmr	28
SMH3 (hemR)	Hem−hemR::cat Strr Cmr	This study
SMH3/pKOBEGA	Hem− λ red functions; Strr Ampr	28
SMH3 (hasR hemR)	Hem−hasR::aphA hemR::cat Strr Kmr Cmr	This study
SMH3 (hasR hemR)/pAM239	Kmr Strr Spcr	This study
SMH3 (hasR hemR)/pAM239-hemR	Kmr Strr Spcr	This study
E. coli
XL1-Blue	F−supE44 hdsR17 recA1 endA1 gyrA46 thi relA1 Δlac F′[ΔproAB lacIqlacZΔM15 Tn10] Tetr	Laboratory collection
C600 (ΔhemA)	C600 (F−thr leu lacY thi supE ΔhemA::Km) Kmr	13
C600 (ΔhemA fur)	C600 (ΔhemA::Km fur::cat) Kmr Cmr	13
C600 (ΔhemA fur tonB)	C600 (ΔhemA::Km fur::cat tonB trp::Tn10) Kmr Cmr Tetr	13
C600 (ΔhemA fur exbB)	C600 (ΔhemA::Km fur::cat exbB::Tn10) Kmr Cmr Tetr	This study
C600 (ΔhemA)/pAM239	C600 (ΔhemA::Km)/pAM239 Kmr Spcr	This study
C600 (ΔhemA)/pAM238-hasR	C600 (ΔhemA::Km)/pAM238-hasR Kmr Spcr	This study
C600 (ΔhemA)/pAM239-hemR	C600 (ΔhemA::Km)/pAM239-hemR Kmr Spcr	This study
C600 (ΔhemA fur)/pAM239-hemR	C600 (ΔhemA::Km fur::cat)/pAM239-hemR Kmr Cmr Spcr	This study
C600 (ΔhemA fur tonB)/pAM238-hasR/pUC19	C600 (ΔhemA::Km fur::cat tonB trp::Tn10)/pAM238-hasR/pUC19 Kmr Cmr Tetr Spcr Ampr	This study
C600 (ΔhemA fur tonB)/pAM238-hasR/pUCtonBEc	C600 (ΔhemA::Km fur::cat tonB trp::Tn10)/pAM238-hasR/pUCtonBEc Kmr Cmr Tetr Spcr Ampr	This study
C600 (ΔhemA fur tonB)/pAM238-hasR/pUChasB6	C600 (ΔhemA::Km fur::cat tonB trp::Tn10)/pAM238-hasR/pUChasB6 Kmr Cmr Tetr Spcr Ampr	This study
C600 (ΔhemA fur tonB)/pAM239-hemR/pUC19	C600 (ΔhemA::Km fur::cat tonB trp::Tn10)/pAM239-hemR/pUC Kmr Cmr Tetr Spcr Ampr	This study
C600 (ΔhemA fur tonB)/pAM239-hemR/pUCtonBEc	C600 (ΔhemA::Km fur::cat tonB trp::Tn10)/pAM239-hemR/pUCtonBEc Kmr Cmr Tetr Spcr Ampr	This study
C600 (ΔhemA fur tonB)/pAM239-hemR/pUCtonBSm	C600 (ΔhemA::Km fur::cat tonB trp::Tn10)/pAM239-hemR/pUCtonBSm Kmr Cmr Tetr Spcr Ampr	This study
C600 (ΔhemA fur tonB)/pAM239-hemR/pUChasB6	C600 (ΔhemA::Km fur::cat tonB trp::Tn10)/pAM239-hemR/pUChasB6 Kanr Cmr Tetr Spcr Ampr	This study
Plasmids
pAM239	Spcr	Laboratory collection
pAM238-hasR	pAM238-hasR Spcr	15
pAM238-hasRΔ45-125	pAM238-hasRΔ45-125 Spcr	2
pAM239-hemR	pAM239-hemR Spcr	This study
pUC19	Ampr	Laboratory collection
pUCtonBEc	pUC19tonBEc Ampr	25
pUCtonBSm	pUC19tonBSm Ampr	25
pUChasB6	pUC19hasB6 Ampr	25

^aCm^r, chloramphenicol resistance; Str^r, streptomycin resistance; Tet^r, tetracycline resistance; Amp^r, ampicillin resistance; Spc^r, spectinomycin resistance; Km^r, kanamycin resistance.

Media and growth conditions.

Hemin, bovine hemoglobin, 2,2′dipyridyl (Dip), and δ-aminolevulinic acid (ALA) were obtained from Sigma Chemical. The hemoglobin concentration was calculated on the basis of the heme monomer. Hemin and hemoglobin solutions were filter sterilized with 0.45-μm-pore-size Millipore filters. Strains were grown on LB medium (22) aerobically at 30 or 37°C. When required, ALA was used at a concentration of 20 μg ml⁻¹. Iron-depleted medium was obtained with the addition of Dip at a 0.2 mM final concentration. Antibiotics were added to the following final concentrations (μg ml⁻¹): ampicillin, 100; spectinomycin, 100; chloramphenicol, 20; kanamycin, 25; streptomycin, 100; and tetracycline, 10. Solid media contained 1.5% Difco agar. Soft LB agar medium contained 0.7% Difco agar.

Growth assays on agar plates supplemented with purified HasA.

A 100-μl sample of an overnight culture of the tested strain was mixed with 4 ml of soft agar and poured onto LB plates supplemented with 10⁻⁷ hemoglobin. Wells (5 mm in diameter) were cut in the agar and filled with 100 μl of sterile HasA (10⁻⁵, 10⁻⁶, or 10⁻⁷ M) prepared from a C600(pSYC34) culture supernatant as described previously (28). Growth around the wells was recorded after overnight incubation at 37°C.

Genetic techniques.

P1 lysates and transductions were performed as described by Miller (22). Cells were transformed by the calcium chloride method (18) or electroporation (8).

DNA manipulations.

Chromosomal DNA was isolated by using the Wizard genomic DNA purification kit (Promega catalog no. A1120). Large-scale plasmid DNA preparations were made by using a Plasmid Maxi kit (Qiagen, Inc., Germany) as recommended by the manufacturer. Small-scale plasmid DNA preparations were done by using a QIAprep Spin Miniprep kit (Qiagen). Restriction, modification, and ligation were carried out according to the manufacturer's recommendations. DNA fragments of interest were separated by gel electrophoresis and isolated using a QIAquick gel extraction kit (Qiagen). DNA fragments were amplified in a Hybaid PCR thermocycler, using LA Taq (TaKaRa, Japan).

Nonpolar deletion of hemR in S. marcescens by red linear DNA gene inactivation.

A nonpolar mutation that deletes the entire hemR gene was created by allelic exchange using a method already described (7). Briefly, plasmid pKOBEGA (an ampicillin-resistant derivative of pKOBEG [see Table 1]) was introduced into the target strain, and electrocompetent cells were prepared at 30°C after induction of the λ red system carried by pKOBEGA with 0.2% arabinose. A three-step PCR procedure was used to produce a PCR product in which the cat gene from pHP45Ω (9) is flanked by 500-bp homology arms corresponding to DNA regions located upstream and downstream from the hemR start and stop codons, respectively. The following primers were used: for the left 500-bp hemR homology arm, hemR.500-5 (5′-ACCGCATCGAAAGCTTGCTGAAATAAC-3′) and hemR.cat.L-3 (5′-GCGGATGAATGGCAGAAATTGGTGCAACTCTCCGTATGTA-3′), and for the right 500-bp hemR homology arm, hemR.500-3 (5′-ACCTCTTACGTGCCGATCAATTGACTATCTGGCGGACAACCCGG-3′) and hemR.cat.L-5 (5′-ACCAGCGCCATCCAGGCCGGC-3′).

The cat gene cassette (0.9 kb) was amplified from strain E. coli FB8 (ttdA::Cm) (24) using the primers cat-5 (5′-AATTTCTGCCATTCATCCGC-3′) and cat-3 (5′-TTGATCGGCCGTAAGAGGT-3′). The PCR product resulting from the three-step procedure was introduced into S. marcescens SMH3/pKOBEGA using electroporation, and chloramphenicol-resistant deletion mutants produced by allelic exchange were selected at 37°C (to eliminate the thermosensitive plasmid pKOBEGA). Correct chromosomal insertion was checked by PCR amplification using the cat primers cat-3 and cat-5 in combination with hemR.500-3 and hemR.500-5, respectively.

Cloning of hemR.

The S. marcescens strain DB11 hemR DNA sequence was obtained from the S. marcescens genomic sequence database (http://www.sanger.ac.uk/Projects/S_marcescens/). A 2,202-bp DNA fragment corresponding to the hemR gene on the S. marcescens DB11 chromosome was amplified using forward and reverse primers hemR-5 (5′-GGCAGCCAGCAAAGCTTTTGATATGTTT-3′) and hemR-3 (5′-ATTTTGGCCTGTTGGTACCGTTCGTATAG-3′). PCR products were digested with HindIII and KpnI and ligated to pAM239 digested with HindIII and KpnI to create pAM339-hemR. The sequence of hemR inserted in plasmid pAM239-hemR was compared to the S. marcescens genomic sequence database.

RESULTS

In silico search for potential heme uptake systems in the S. marcescens genome.

The S. marcescens strain SM365 genomic sequence is not available, but the genome of S. marcescens strain DB11, shown to be closely related to SM365, as evidenced by comparison of the sequences of genes in the two organisms (V. Braun and C. Wandersman, unpublished data), was sequenced and is available in a database (http://www.sanger.ac.uk/cgi-bin/BLAST/submitblast/s_marcescens). The search for putative heme transporters other than the has system in the DB11 genomic sequence led to the identification of two coding sequences exhibiting homologies with outer membrane heme receptors. The first (SMA 2302) encoded a polypeptide sharing strong identity with putative heme receptors from Erwinia carotovora (71% identical), Photorhabdus luminescens (72% identical), and Bordetella bronchiseptica (68% identical). The other putative outer membrane heme receptor gene (SMA 1443) belonged to an operon encoding polypeptides sharing strong identity (in a range from 40 to 74%) with HemP, HemR, HemS, HemT, HemU, and HemV from Yersinia enterocolitica (31). A Fur box (GATAATTGATTCTCAGT) located upstream from hemR in the hemP coding sequence suggested iron-dependent regulation of hemR, as observed for most operons coding for heme acquisition systems (31). In Y. enterocolitica, HemR was demonstrated to be an outer membrane heme receptor. Similarly, HemR from S. marcescens contains FRAP and NPNL amino acid boxes present in all heme receptors that transport heme into the periplasm (3). The two conserved histidines involved in heme uptake were also present at positions 134 and 467 in HemR from S. marcescens.

HemR is functional for heme uptake in S. marcescens.

As already described, disruption of the hemophore-dependent heme receptor hasR in S. marcescens decreased but did not abolish the capacity for heme uptake (28). We hypothesized that hemR was responsible for has-independent uptake of heme. We disrupted hemR in an S. marcescens heme auxotroph SMH3 hasR mutant using a nonpolar gene substitution method as described in Materials and Methods. The SMH3 (hasR hemR) mutant obtained was tested for growth in the presence of exogenous hemoglobin at a final concentration of 10⁻⁵ M. As shown in Table 2, the SMH3 strain and its hasR derivative were able to take up heme and to grow. In contrast, S. marcescens SMH3 (hasR hemR) was unable to take up heme or to grow. This demonstrates that hemR coded for the alternative heme outer membrane receptor of S. marcescens. The hemR structural gene from S. marcescens was amplified, cloned in low-copy-number plasmid pAM239, and introduced into strain SMH3 (hasR hemR). The resulting strain, SMH3 (hasR hemR)/pAM239-hemR, was able to take up heme and to grow in the presence of hemoglobin (10⁻⁵ M) (Table 2). The latter validated the identification of HemR as a heme outer membrane receptor.

TABLE 2.

Heme uptake in S. marcescens strain SMH3 and its derivatives impaired in heme uptake systems^a

Strain	Phenotype	Growth on:
		LB	LB ALA	LB Hb Dip
SM365	Wild type	+++	+++	+++
SMH3	Hem−	−	+++	+++
SMH3 (hasR)	Hem−	−	+++	+++
SMH3 (hemR)	Hem−	−	+++	+++
SMH3 (hasR hemR)	Hem−	−	+++	−
SMH3 (hasR hemR)/ pAM239	Hem−	−	+++	−
SMH3 (hasR hemR)/ pAM239-hemR	Hem−	−	+++	+++

^aThe SMH3 heme auxotroph of S. marcescens and isogenic hasR::Km, hemR::Cm, and hasR::Km hemR::Cm mutants were plated on LB agar alone (LB), LB agar supplemented with 20 μM ALA (LB ALA), or LB agar supplemented with 0.15 mM Dip to chelate iron and 10⁻⁵ M hemoglobin (LB Hb Dip). Single colonies of each tested strain were reisolated on the indicated media to obtain isolated colonies. Groups of about 20 isolated colonies were then compared. Plates were incubated for 24 h at 30°C. “+++” corresponds to colony size equivalent to that obtained on LB plus ALA. All experiments were repeated three times.

Expression of the hemR gene from S. marcescens in E. coli enables heme uptake.

The study of heme uptake systems in S. marcescens can be fastidious. Thus, we introduced plasmid pAM239-hemR into E. coli strain C600 (ΔhemA) and tested growth stimulation of the transformant strains with heme or hemoglobin. Deletion of hemA, which abolishes ALA biosynthesis, enabled E. coli growth only in the presence of ALA. Addition of exogenous heme did not restore the growth of strain C600 (ΔhemA), since E. coli was unable to use exogenously supplied heme in the absence of a heterologous outer membrane heme transporter. As shown in Table 3, growth stimulation by exogenous hemoglobin was observed with strain C600 (ΔhemA) harboring a plasmid expressing hemR. Similar results were obtained with free heme (data not shown). These results show that HemR alone allowed heme transport across the outer membrane in E. coli and its use as a porphyrin source.

TABLE 3.

Heme uptake in E. coli strain C600 (ΔhemA) expressing hemR from S. marcescens^a

Strain	Growth on:
	LB	LB ALA	LB Hb Dip	LB Hb
C600 (ΔhemA)/pAM239	−	+++	−	−
C600 (ΔhemA)/pAM239-hemR	−	+++	+++	+++

^aStrains were plated on LB agar alone (LB), LB agar supplemented with 20 μM ALA (LB ALA), or LB agar supplemented with 10⁻⁵ M hemoglobin in the absence (LB Hb) or presence (LB Hb Dip) of 0.15 mM Dip as indicated to chelate iron. Single colonies of each tested strain were reisolated on the indicated media to obtain isolated colonies. Groups of about 20 isolated colonies were then compared. Plates were incubated for 24 h at 37°C. “+++” corresponds to the colony size equivalent to that obtained on LB plus ALA. All experiments were repeated three times.

HemR-dependent heme uptake functions with TonB but not HasB6.

Disruption of tonB or exbB in strain C600 (ΔhemA)/pAM239-hemR abolished growth in the presence of hemoglobin (Table 4). This indicated that heme transport through HemR was dependent on the TonB-ExbB-ExbD complex, since it is also the case for HasR. As shown in Table 4, introduction of a plasmid expressing either TonB_Ec or TonB_Sm in strain C600 (ΔhemA tonB)/pAM239-hemR or strain C600 (ΔhemA tonB)/pAM238-hasR restored heme uptake. In contrast, a plasmid expressing HasB6 enabled the heme utilization of C600 (ΔhemA tonB)/pAM238-hasR but not of C600 (ΔhemA tonB)/pAM239-hemR. These results showed that HasB6 was specific to the HasR-dependent heme uptake system and not to heme receptors in general.

TABLE 4.

Complementation of a tonB mutation for HemR- and HasR-dependent heme uptake in E. coli strain C600 (ΔhemA tonB) harboring pUC19 derivatives expressing tonB_Ec, tonB_Sm, or hasB6^a

Strain	Growth on:
	LB	LB ALA	LB Hb
C600/pAM239	−	+++	+++
C600 (ΔhemA)/pAM239	−	+++	−
C600 (ΔhemA)/pAM239-hemR/pUC19	−	+++	+++
C600 (ΔhemA exbB)/pAM239-hemR/pUC19	−	+++	−
C600 (ΔhemA tonB)/pAM239-hemR/pUC19	−	+++	−
C600 (ΔhemA tonB)/pAM239-hemR/pUCtonBEc	−	+++	+++
C600 (ΔhemA tonB)/pAM239-hemR/pUCtonBSm	−	+++	+++
C600 (ΔhemA tonB)/pAM239-hemR/pUChasB6	−	+++	−
C600 (ΔhemA)/pAM238-hasR/pUC19	−	+++	+++
C600 (ΔhemA tonB)/pAM238-hasR/pUC19	−	+++	−
C600 (ΔhemA tonB)/pAM238-hasR/pUCtonBEc	−	+++	+++
C600 (ΔhemA tonB)/pAM238-hasR/pUCtonBSm	−	+++	+++
C600 (ΔhemA tonB)/pAM238-hasR/pUChasB6	−	+++	+++

^aStrains were plated on LB agar alone (LB), LB agar supplemented with 20 μM ALA (LB ALA), or LB agar supplemented with 10⁻⁵ M hemoglobin (LB Hb). Single colonies of each tested strain were reisolated on the indicated media to obtain isolated colonies. Sizes of about 20 isolated colonies were then compared. Plates were incubated for 24 h at 37°C. “+++” corresponds to the colony size equivalent to that obtained on LB plus ALA. All experiments were repeated three times.

The HasR N-terminal signaling domain is not involved in HasB specificity.

HasR belongs to the autoregulated subgroup of TBDT receptors that have an extra N-terminal domain. To test whether this signaling domain was involved in the specificity of HasR-HasB recognition, a HasR mutant with its signaling domain deleted (HasRΔ45-125) was used. For these experiments, we used fur isogenic strains to avoid the deleterious effect of dipyridyl addition on tonB mutant growth. The mutant HasRΔ45-125 has a wild-type phenotype for heme uptake but is unable to initiate the has signaling cascade in the presence of a heme-loaded hemophore (2). The results presented in Table 5 show that this mutant receptor enabled hemophore-dependent and hemophore-independent heme acquisition as efficiently as did the wild-type receptor in the presence of HasB6. This clearly demonstrated that the specificity of HasB6 toward HasR was not related to the presence of a signaling domain. Surprisingly, with TonB_Ec, HasRΔ45-125 slightly stimulated hemophore-dependent heme uptake, whereas entire HasR had no effect (Table 5).

TABLE 5.

Complementation of the tonB mutation for free and hemophore-dependent heme uptake in E. coli strain C600 (ΔhemA) expressing HasR and HasRΔ45-125^a

Strain	Growth on:
	LB	LB ALA	LB Hb (colony size)	LB Hb HasA (radius around well)
C600 (ΔhemA fur tonB)/pAM238-hasR/pUC19			−	−
C600 (ΔhemA fur tonB)/pAM238-hasR/pUCtonBEc			+++	−
C600 (ΔhemA fur tonB)/pAM238-hasR/pUChasB6	−	+++	+++	+++
C600 (ΔhemA fur tonB)/pAM238-hasRΔ45-125/pUC19	−	+++	−	−
C600 (ΔhemA fur tonB)/pAM238-hasRΔ45-125/pUCtonBEc	−	+++	+++	+
C600 (ΔhemA fur tonB)/pAM238-hasRΔ45-125/pUChasB6	−	+++	+++	+++

^aFor free heme uptake, strains were plated on LB agar alone (LB) or LB agar supplemented with 10⁻⁵ M hemoglobin (LB Hb). Plates were incubated for 24 h at 37°C. “+++” corresponds to the colony size equivalent to that obtained on LB plus ALA (LB ALA). For hemophore-dependent heme uptake, a 100-μl sample of an overnight culture of the tested strain was mixed with 4 ml of soft agar and poured onto LB plates supplemented with 10⁻⁷ M hemoglobin. Wells (5 mm in diameter) were cut in the agar and filled with 100 μl of sterile HasA at 10⁻⁵ M (LB Hb HasA). Growth around the wells was recorded after overnight incubation at 37°C. “+++” corresponds to a thick growth radius of 10 mm around the well. “+” corresponds to a thin growth radius of 2 mm around the well. All experiments were repeated three times.

DISCUSSION

HasB6, a mutant HasB functional in E. coli, complements E. coli TonB function for heme acquisition via HasR but not other TonB functions, such as siderophore uptake via TBDT. In the present study, we addressed the question of HasB6 specificity toward heme uptake: is HasB specific to heme receptors or to HasR?

We identified a heme receptor paralog in the S. marcescens genomic sequence (hemR). We inactivated hemR with a nonpolar insertion in S. marcescens. Using single and double hemR-hasR mutants, we showed that, whereas each hasR or hemR single mutant is still able to acquire heme, the hemR hasR double mutant has completely lost this uptake capacity, demonstrating that hemR is functional in S. marcescens. When cloned into E. coli, hemR enables heme transport through the outer membrane in a TonB-dependent manner. HasB6 is unable to replace TonB for heme uptake via HemR. HasB is not dedicated to heme receptors in general but is specific to HasR. In fact, heme receptors do not form a homogeneous family. Amino acid comparison between them shows that heme receptors share homologies ranging from 20 to 90% identity. The HasR subfamily is only distantly related to other heme receptors. S. marcescens HemR and HasR share 23% identity and 37% similarity, values that are not significantly higher than the homologies shared with TBDT receptors involved in other types of iron source uptake. However, HasR and HemR strongly differ in their N termini, since HasR, which belongs to the family of autoregulated receptors, possesses an N-terminal extension, the signaling domain. We show here that HasR lacking the entire signaling domain is equally functional with TonB and HasB6. This demonstrates that the N-terminal extension is not required for the activation of HasR by HasB. Nevertheless, we observed that truncated HasR lacking its signaling domain, when energized by TonB, is more efficient at heme-hemophore uptake than the entire HasR protein. This discrepancy is not visible for heme uptake, which requires lower TonB expression (16). On the other hand, the Δ45-125 deletion does not enhance HasB6 activity. These results suggest that the N terminus extension restricts the access of the energizing protein. The absence of an effect of the HasR signaling domain upon HasB6 activity can be explained either by other contacts between the two proteins or by a higher affinity of HasB6 for HasR, resulting in full HasB6 activity even with fewer HasR binding sites available.

Interactions between TonB proteins and TBDT receptors involve their TonB C-terminal domain. This domain, poorly conserved between HasB and the other TonB proteins, was shown to contact the TonB boxes in BtuB-TonB and FhuA-TonB structures. The putative TonB boxes of HemR (DETMTVVA) and HasR (DSLTVLGA) are equally similar to the E. coli TBDT receptor TonB box consensus (DTLVVTA). It is thus unlikely that HasB specificity relies on the interaction between HasB and the HasR TonB box. Moreover, TonB boxes have been shown to be essential for ligand transport and ligand-mediated signaling but not for specificity. TonB box switching between an E. coli receptor (ChuA) and the V. cholerae receptor HutA, each of which functions only with the TonB protein from their original host, does not switch TonB dependency (20). Sequence comparison reveals that HasB is ca. 27% identical to TonB_Sm and 29% identical to TonB_Ec. HasB (263 amino acids) is longer than TonB_Sm (247 amino acids) and TonB_Ec (239 amino acids), indicating that HasB has a longer central proline-rich spacer. Also, the third “TonB β-strand,” located between amino acids 252 and 258 in HasB, is poorly conserved (with an identity of two out of seven). These differences might be responsible for the specificity of HasB toward HasR.

The genetic association of hasB with the gene encoding its dedicated TBDT receptor HasR in the has operon enables their coregulated expression. HasB is therefore positively regulated by the signaling cascade and is thus produced at a higher level when a heme-loaded hemophore is available. Such a high HasB level might be required for heme-hemophore uptake. Since HasB is specific to HasR, it might either not energize other TBDT receptors or not interact with them, preventing wasting of energy or a loss of the HasB molecular pool. We have previously shown that hemophore-dependent heme uptake in E. coli expressing the Has system is a high-energy-consuming process. Whereas free heme uptake via HasR functions at basal TonB-ExbB-ExbD levels, full TonB-ExbB-ExbD expression is required for recycling empty hemophore from the cell surfaces (16). HasB6 efficiently substitutes for TonB in this process (S. Létoffé, unpublished data). Thus, we hypothesize that TonB has only weak activity toward HasR, either because of low affinity or because of an incomplete interaction. HasB is the only described case of a TonB protein specific for only one TBDT receptor. However, the rapidly increasing DNA sequence data for microbial genomes frequently show the presence of potential TonB homologs in the vicinity of TBDT receptors, suggesting that the existence of TonB-TBDT receptor-specific pairs might not be an exception. In the case of the Has system, TonB can be considered a poor energy transmitter, allowing a basal function for the Has system, which works fully only in the presence of its own specific TonB ortholog, HasB.