Identification of functionally important regions of a ... - CiteSeerX

Microbiology (2003), 149, 3423–3435

DOI 10.1099/mic.0.26448-0

Identification of functionally important regions of a haemoglobin receptor from Neisseria meningitidis D. Perkins-Balding, M. T. Baer and I. Stojiljkovic Department of Microbiology and Immunology, Emory School of Medicine, 1510 Clifton Rd, Atlanta, GA 30322, USA

Correspondence D. Perkins-Balding [email protected]

Received 29 April 2003 Revised

9 July 2003

Accepted 9 September 2003

The HmbR outer-membrane receptor enables Neisseria meningitidis to use haemoglobin (Hb) as a source of iron. This protein functions by binding Hb, removing haem from it, and releasing the haem into the periplasm. Functionally important HmbR receptor domains were discerned using a series of HmbR deletions and site-directed mutations. Mutations exhibiting similar defective phenotypes in N. meningitidis fell into two groups. The first group of mutations affected Hb binding and were located in putative extracellular loops (L) L2 (amino acid residues (aa) 192–230) and L3 (aa 254–284). The second group of mutations resulted in a failure to utilize Hb but proficiency in Hb binding was retained. These mutations localized to the putative extracellular loops L6 (aa 420–462) and L7 (aa 486–516). A highly conserved protein motif found in all haem/Hb receptors, within putative extracellular loop L7 of HmbR, is essential for Hb utilization but not required for Hb binding. This finding suggests a mechanistic involvement of this motif in haem removal from Hb. In addition, an amino-terminal deletion in the putative cork-like domain of HmbR affected Hb usage but not Hb binding. This result supports a role of the cork domain in utilization steps that are subsequent to Hb binding.

INTRODUCTION Even though iron is the fourth most abundant element on the earth, bacterial pathogens are faced with the obstacle of obtaining enough of this essential element for colonization and growth (Griffiths, 1999). Gram-negative bacteria have evolved a variety of iron-assimilation systems that acquire iron from iron-carrier proteins either by directly binding to specific outer-membrane receptors or by release of bacterial siderophores (or siderophore-like proteins) that shuttle iron from host carrier proteins back to the bacterium (Wandersman & Stojiljkovic, 2000; Winkelmann, 2002). A valuable reservoir of iron in the body is iron protoporphyrin IX, or haem, a molecule which can be used by the bacterium for energy metabolism and oxidative catalysis (Stojiljkovic & Perkins-Balding, 2002; Wandersman & Stojiljkovic, 2000). Although abundant, haem is only found in limited amounts extracellularly, and most often it is tightly associated with host carrier proteins, especially globin (Genco & Dixon, 2001; Griffiths, 1999; Stojiljkovic & Perkins-Balding, 2002). Haem/haemoglobin (Hb) receptors have been identified in many bacterial species, both pathogens and commensals (Cope et al., 1994; Henderson & Payne, 1994; Letoffe et al., 1994; Stojiljkovic & Hantke, 1992). These receptors share a large degree of structural homology with the TonBdependent lactoferrin and transferrin receptors of Gramnegative bacteria (Gray-Owen & Schryvers, 1996; Schryvers Abbreviation: Hb, haemoglobin.

0002-6448 G 2003 SGM

Printed in Great Britain

& Stojiljkovic, 1999; Schryvers et al., 1998). The common theme of haem/Hb receptors is their ability to bind haem compounds, remove haem from them, and transport haem into the periplasm (Wandersman & Stojiljkovic, 2000). Haem/Hb receptors can be loosely grouped into three subfamilies based on the level of amino acid homology, overall design and substrate specificity. Haem ‘scavengers’ are a subfamily of enterobacterial haem/Hb receptors that share more than 60 % identical amino acid residues and have wide substrate specificity (Bracken et al., 1999; Wandersman & Stojiljkovic, 2000). Haem/Hb receptors from Haemophilus, Neisseria and Vibrio spp. share between 25 and 50 % identity and have narrow substrate specificity. The third subfamily of haem utilization systems is found in Serratia, Pseudomonas and Haemophilus spp., where special proteins called haemophores are secreted into the medium. Haemophores bring haem to the outer cell surface, where a TonB-dependent receptor takes haem and transports it into the periplasm (Cope et al., 1994; Letoffe et al., 1994, 1998). We have identified a high-affinity Hb outer-membrane receptor, HmbR, of Neisseria meningitidis (Stojiljkovic et al., 1995). HmbR enables meningococci to bind and use Hb as a source of iron and porphyrin (Stojiljkovic & Srinivasan, 1997; Stojiljkovic et al., 1996, 1995). HmbR-mediated Hb utilization has been reconstituted in Escherichia coli. As expected for a member of the TonB-dependent family of receptors, HmbR requires expression of the N. meningitidis Ton system to function as a Hb receptor in E. coli 3423

D. Perkins-Balding, M. T. Baer and I. Stojiljkovic

(Stojiljkovic & Srinivasan, 1997). Once HmbR transports haem into the periplasm, it is most likely delivered to the cytoplasm by HemTUV-like proteins, although homologous proteins have not been identified yet in N. meningitidis (Stojiljkovic & Hantke, 1994). In the cytoplasm, a newly identified haem oxygenase, HemO, is involved in the catabolism of haem in N. meningitidis to iron-biliverdin and CO (Ratliff et al., 2001; Zhu et al., 2000a, b). Despite a large volume of information on haem/Hb receptors collected in the last decade, important questions about their structure and mechanism of action remain largely unanswered. For example, very little is known about the receptor domains that participate in interactions with haem and haem–carrier complexes. In addition, these receptors must be able to extract haem from different protein carriers prior to the transport step, but the mechanism of haem removal is not understood. In the enterobacterial ‘scavenger’ subclass of haem/Hb receptors, amino acid residues that participate in the transport of haem through the receptor pore have been identified. Two histidine residues, His128 and His461, are essential for the function of the Yersinia enterocolitica HemR receptor (Bracken et al., 1999). To date, none of the haem/Hb receptor structures have been solved by crystallography. In order to gain information about the structure of HmbR, our approach followed the analysis of outer-membrane transporters, FepA, FhuA and LamB, using deletions to ascertain which surface-exposed loops of HmbR contribute to specific functions (Boulanger et al., 1996; Braun et al., 1999; Carmel & Coulton, 1991; Klebba et al., 1994; Newton et al., 1999). This approach exploits the tendency for deletions in loop regions of outermembrane proteins to have little effect, if any, on proper folding, assembly and localization (Benson et al., 1988; Klebba et al., 1994; Newton et al., 1999). Deletion of transmembrane regions, however, results most often in generating proteins that are either expressed at low levels or proteolytically degraded (Klebba et al., 1994; Lathrop et al., 1995; Newton et al., 1999). This approach is useful in narrowing down the functional regions of an extremely large outer-membrane protein, such as HmbR, so that a more targeted analysis can be done. In this communication, we propose a model for the structure of HmbR, which, like all of the characterized outermembrane transporters of Gram-negative bacteria, has an extended b-sheet that will putatively form an amphiphilic barrel in the outer membrane (Buchanan, 1999; Koebnik et al., 2000). Surface-exposed loop regions were predicted from this model and used as a guide for deletion mutagenesis. We identified HmbR mutants that were defective in Hb binding and/or Hb utilization. Further, we demonstrated the importance of the cork-like region, modelled after the plug domains of outer-membrane transporters FepA, FhuA and FecA and the highly conserved central motifs of Hb/haem receptors in the utilization of haem– protein complexes (Bracken et al., 1999; Wandersman & Stojiljkovic, 2000). 3424

METHODS Plasmids, bacteria and media. Strains and plasmids used and

constructed in this study are listed in Table 1. The meningococci were grown on GCB (Difco) agar containing the supplements as described by Kellogg et al. (1963), and incubated at 37 uC with 5 % CO2. E. coli was grown in LB (Luria broth). When necessary, the following antibiotics were used to grow E. coli: chloramphenicol, 30 mg ml21; tetracycline, 10 mg ml21; and ampicillin, 100 mg ml21. For N. meningitidis, streptomycin, 750 mg ml21; and erythromycin, 3 mg ml21 were used when needed. The construction of a two-dimensional model of the HmbR receptor. Several criteria were used in the construction of the twodimensional model of the HmbR b-barrel. (I) The propensity of amino acids to form b-strands was calculated using the method of

Schirmer & Cowan (1993) (see Fig. 1). This method calculates the mean hydrophobicity of the side of a putative transmembrane bstrand that is exposed to membrane lipids by averaging the hydrophobic indices of every second residue with a sliding window of 4 (Eisenberg et al., 1984; Pettersson et al., 1994; Urushibara et al., 1992). (II) The conserved residues between HmbR and its homologues, Vibrio cholerae HutA, Haemophilus ducreyi HgbA and N. meningitidis HpuB, were assigned to the membrane-embedded b-strands or tight b-turns. (III) Hypervariable regions of the HmbR receptors, regions that did not contain any conserved residues when compared with HpuB, HutA and HgbA, and regions having hydrophilic maxima in hydropathy profile were interpreted to form surface- or periplasmic-exposed loops (Baumler & Hantke, 1992; Jeanteur et al., 1991; Koebnik & Braun, 1993). (IV) possible b-turns were determined using the program PEPPLOT from the UWGCG software package (Gribskov et al., 1986). Crystallographic structures of FepA and FhuA siderophore receptors were used in the prediction of the putative cork of HmbR (Buchanan et al., 1999; Ferguson et al., 1998; Locher et al., 1998). However, the b-barrel structure of FepA and FhuA could not be used in the modelling due to the low level of homology between b-barrels of HmbR and siderophore receptors. The N-terminal sequence of the purified, native HmbR protein was determined using the Edman degradation method. Recombinant DNA techniques. Standard methods for plasmid

DNA preparation, restriction endonuclease analyses and ligations were carried out as described by Sambrook et al. (1989). PCR reactions were performed using high fidelity PfuI (Stratagene) or PwoI (Boehringer Mannheim) DNA polymerases according to the instructions of the suppliers. Oligonucleotides were custom-made by Gibco BRL. Construction of hmbR deletion and site-directed mutants.

Deletions in hmbR were constructed using restriction sites NotI, PstI and SalI, or by introducing BglII or EcoRI restriction sites at the position of the desired deletion end-points. The majority of inserted restriction sites did not introduce new amino acid residues into the HmbR sequence. By using convenient hmbR restriction sites, mutated receptor genes were reconstituted with the smallest possible PCR-generated hmbR fragment still carrying the desired deletion and cloned into pACYC184 vector (the largest PCR-amplified fragment was 1?2 kb, the smallest 0?1 kb). The nucleotide sequence of all PCR-generated DNA fragments was determined to rule out the existence of secondary mutations. Construction of an HmbR overexpression plasmid. Insertion

of a His tag into pIRS1173 was done using DNA primers: His-A: (59-GCATCATCACCATCACCATCTGCA-39) and His-B (59-GATGGTGATGGTGATGATGCTGCA-39). The primers were annealed and cloned into PstI-digested plasmid pIRS1173, to produce pSYB1724. The nucleotide sequence was determined to confirm that Microbiology 149

Mutagenesis of HmbR

Table 1. Strains and plasmids used in the study Strain

Relevant genotype

Source/reference

E. coli K-12 DH5a IR1532 IR1583 IR3758 IR3519 N. meningitidis IR1072 IR4130 IR3727 IR3728 IR3729 IR3730 IR3731 IR3732 IR3733 IR3769 IR3770 IR3771 IR3773 IR3778 IR3780 IR3781 IR3788 IR3792 IR3793 IR3794 IR3795 IR3796

DH5a but haem biosynthesis mutant IR1532 but [tonB, exbB, exbD]NM tetR IR1532 but [tonBEC : : IS10] DH5a with plasmid pSYB1724

Stratagene Stojiljkovic & Srinivasan (1997) Stojiljkovic & Srinivasan (1997) This study This study

Serogroup C, wild-type, hmbR+ DhpuAB As IR1072 but strR As IR1072 but hmbR : : rpsL ermC IR4130, but hmbRD[Q254–V261] IR4130, but hmbRD[K664–D689] IR4130, but hmbRD[A275–L286] IR4130, but hmbRD[G203–H231] IR4130, but hmbRD[L297–S325] IR4130, but hmbRD[V489–G502] IR4130, but hmbRD[H64–K66] IR4130, but hmbRD[H359–R369] IR4130, but hmbRD[H435–P454] IR4130, but hmbRD[G203S] IR4130, but hmbRD[V558–K577] IR4130, but hmbRD[V489–R527] IR4130, but hmbRD[Y389–P454] IR4130, but hmbRD[R192–Y230] IR4130, but hmbRD[D324–R369] IR4130, but hmbRD[K664–A669] IR4130, but hmbRD[H64–N76] IR4130, but hmbRD[N458–G465] IR4130, but hmbRD[N410–P454]

Stojiljkovic et al. (1995) This study This study This study This study This study This study This study This study This study This study This study This study This study This study This study This study This study This study This study This study This study

Plasmids pWKS30 pFLOB4300 pIRS626 pIRS1173 pSYB1724 pIRS1183 pIRS1184 pIRS1185 pIRS1192 pIRS1203 pIRS1325 pIRS1383 pIRS1384 pIRS1402 pIRS1420 pIRS1492 pIRS1561 pIRS1562 pIRS1565 pIRS1725 pIRS1751 pIRS1752 pIRS1753 pIRS1770

Low-copy-number plasmid, bla pUC18 but rpsL ermC cassette hmbR+, 6 kb ClaI pWKS30 hmbR+, 3?2 kb BamHI–HindIII pACYC184 pIRS1173 with hmbR-6his pIRS1173, but hmbRD[H64–K66] pIRS1173, but hmbRD[H359–R369] pIRS1173, but hmbRD[H435–P454] pIRS1173, but hmbRD[Q254–V261] pIRS1173, but hmbRD[K664–A669] pIRS1173, but hmbRD[G203S] pIRS1173, but hmbRD[V558–K577] pIRS1173, but hmbRD[K664–D689] pIRS1173, but hmbRD[H64–N76] pIRS1173, but hmbRD[A275–L286] pIRS1173, but hmbRD[N458–G465] pIRS1173, but hmbRD[D324–R369] pIRS1173, but hmbRD[R192–Y230] pIRS1173, but hmbRD[V489–R527] pIRS1173, but hmbRD[G203–H231] pIRS1173, but hmbRD[L297–S325] pIRS1173, but hmbRD[N410–P454] pIRS1173, but hmbRD[Y389–P454] pIRS1173, but hmbRD[V489–G502]

Wang & Kushner (1991) Johnston & Cannon (1999) Stojiljkovic et al. (1995) Stojiljkovic & Srinivasan (1997) This study This study This study This study This study This study This study This study This study This study This study This study This study This study This study This study This study This study This study This study

http://mic.sgmjournals.org

3425

D. Perkins-Balding, M. T. Baer and I. Stojiljkovic

Table 1. cont. Strain pAP1173 pAP1183 pAP1184 pAP1185 pAP1192 pAP1203 pAP1325 pAP1383 pAP1384 pAP1402 pAP1420 pAP1492 pAP1561 pAP1562 pAP1565 pAP1725 pAP1751 pAP1752 pAP1753 pAP1770 pAP1184-FLOB pBAD24 pBAD1352 pBAD1402 pBAD1725 pBAD1192 pBAD1770 pBAD1185 pBAD1492

Relevant genotype

Source/reference

hmbR , pWKS30 pAP1173, but hmbRD[H64–K66] pAP1173, but hmbRD[H359–R369] pAP1173, but hmbRD[H435–P454] pAP1173, but hmbRD[Q254–V261] pAP1173, but hmbRD[K664–A669] pAP1173, but hmbRD[G203S] pAP1173, but hmbRD[V558–K577] pAP1173, but hmbRD[K664–D689] pAP1173, but hmbRD[H64–N76] pAP1173, but hmbRD[A275–L286] pAP1173, but hmbRD[N458–G465] pAP1173, but hmbRD[D324–R369] pAP1173, but hmbRD[R192–Y230] pAP1173, but hmbRD[V489–R527] pAP1173, but hmbRD[G203–H231] pAP1173, but hmbRD[L297–S325] pAP1173, but hmbRD[N410–P454] pAP1173, but hmbRD[Y389–P454] pAP1173, but hmbRD[V489–G502] pAP1184, but hmbR : : rpsL ermC araC-regulated promotor araC-regulated hmbR pBAD1352, but hmbR1402 deletion pBAD1352, but hmbR1725 deletion pBAD1352, but hmbR1192 deletion pBAD1352, but hmbR1770 deletion pBAD1352, but hmbR1185 deletion pBAD1352, but hmbR1492 deletion

This study This study This study This study This study This study This study This study This study This study This study This study This study This study This study This study This study This study This study This study This study Guzman et al. (1995) This study This study This study This study This study This study This study

+

no unintended mutations had been generated. Hb and haem utilization assays were performed on E. coli strain IR1583 harbouring this plasmid to ensure that HmbR-6His was functional.

IR3780, IR3731, IR3732, IR3796, IR3781 and IR3733, respectively. The genotypes of the mutant hmbR alleles were checked by PCR and a restriction digest with the appropriate restriction endonuclease. Phenotypic testing of HmbR mutants. All hmbR constructs were

Gene replacement of hmbR mutant alleles in N. meningitidis IR1072. Mutant alleles of hmbR, demonstrating defective Hb

binding or Hb utilization phenotypes in E. coli, were used to replace the wild-type copy of hmbR in a serogroup C, clinical isolate of N. meningitidis, IR1072. First, a spontaneous streptomycin-resistant mutant of IR1072, IR4130, was isolated by plating 161010 c.f.u. of IR1072 onto GCB agar containing 750 mg streptomycin ml21. IR4130 was transformed with plasmid pAP1184-FLOB and an hmbR knockout transformant (IR3727) was selected on erythromycincontaining GCB agar plates. Plasmid pAP1184-FLOB is a derivative of pAP1184 (pWKS30 with hmbRD[H359–R369]), which has the rpsL ermC cassette from pFLOB4300 inserted into the unique BglII site in the hmbR gene (Johnston & Cannon, 1999). Strain IR3727 containing hmbR : : rpsL ermC mutation (streptomycin sensitive) was subsequently transformed with plasmids carrying hmbR mutant alleles (pAP1183, pAP1184, pAP1185, pAP1192, pAP1203, pAP1325, pAP1383, pAP1384, pAP1402, pAP1420, pAP1492, pAP1561, pAP1562, pAP1565, pAP1725, pAP1751, pAP1752, pAP1753 and pAP1770) and selected on GCB agar containing 750 mg streptomycin ml21, to generate strains IR3769, IR3770, IR3771, IR3728, IR3793, IR3773, IR3778, IR3729, IR3794, IR3730, IR3795, IR3792, IR3788, 3426

subcloned onto low-copy-number plasmid pWKS30 and transformed into the E. coli IR1583 strain expressing the TonB, ExbB and ExbD proteins of N. meningitidis. The ability of E. coli to use Hb and haem was tested on NBD plates (Nutrient Broth-dipyridyl) as described by Stojiljkovic & Srinivasan (1997). Haemin (bovine), human serum albumin and Hb (human) were obtained from Sigma. To test the ability of N. meningitidis strains to use haem-containing compounds as sole iron sources, a suspension of bacteria was plated onto GCB agar containing 50 mM desferrioxamine mesylate (Desferal, Ciba Geigy). Filter discs (0?25 inches, Schleicher & Schuell) impregnated with test compounds (10 ml of 5 mg ml21 stock solutions unless otherwise stated) were placed on these plates. Zones of growth around the discs were recorded after overnight incubation at 37 uC in the presence of 5 % CO2 (Stojiljkovic et al., 1995). Small-scale purification of HmbR. E. coli strain IR3519, contain-

ing plasmid pSYB1724, was inoculated at a 1 to 100 dilution into 200 ml L-broth containing 30 mg chloramphenicol ml21. Cells were grown to mid-exponential phase, and HmbR-6His expression was induced from the native promoter using 0?35 mM dipyridyl for 3 h at 37 uC. Outer membranes were prepared from spheroplasts using Microbiology 149

Mutagenesis of HmbR the method of Hantke (1981). Outer membranes were solubilized in 1 % ZW3-14/TBS (TBS is 50 mM Tris/HCl, 100 mM NaCl, pH 7?5). Solubilized outer membranes were diluted in 25 ml 20 mM phosphate buffer, pH 7?4, 0?5 M NaCl, 0?9 % N-octylglucoside, 0?25 mM PMSF. HmbR-6His was bound to Ni-NTA agarose (Qiagen), washed with 3 vols of the dilution buffer, and eluted in the dilution buffer containing 500 mM imidazole. The yield of purified HmbR was 0?25 mg per ml of culture. Outer membrane isolation and HmbR quantification. Prepara-

tions of neisserial outer membranes were made as described by Richardson & Stojiljkovic (1999). The total protein content of the outer membranes was determined by solubilization of membrane material followed by its quantification using a modified Lowry assay, the Bio-Rad DC (detergent-compatible) protein assay. The amount of HmbR and HmbR mutant proteins per outermembrane sample was determined by Western blotting as follows. Outer-membrane samples containing 140 mg total protein were electrophoresed on a 7?5 % SDS-polyacrylamide gel, and transferred to Hybond-P PVDF membrane (Amersham Pharmacia) using a Schleicher and Schuell electroblotting apparatus. The membrane was blocked in PBST (100 mM phosphate-buffered saline, pH 7?5, 0?5 % Tween 20) containing 5 % dry milk and incubated with rabbit polyclonal antiserum raised against a fusion protein consisting of HmbR and glutathione S-transferase (GST-HmbR). A secondary antibody, anti-rabbit IgG conjugated to alkaline phosphatase (Sigma), and fluorescent (ECF) substrate (Amersham Pharmacia) was used for detection, and visualized on a Molecular Dynamics Storm 860 Phosphorimager. ImageQuant software (Molecular Dynamics) was used to analyse the signal intensity. The amount of HmbR mutant protein per sample was extrapolated from a standard curve that was generated using known quantities of purified HmbR. The amount of the wild-type HmbR in 140 mg of total protein material was 350-fold higher than the lower limit of detection for HmbR by this method; therefore, those outer membranes that did not yield a detectable signal for HmbR were below this limit of detection and were not analysed further. Outer membranes that had detectable levels of HmbR were quantified by this method multiple times. Equal quantities of HmbR and HmbR mutant proteins in outer-membrane samples were then immobilized on nitrocellulose for Hb binding analysis as described below. Biotinylation of Hb. Human Hb (Sigma) was biotinylated using sulfo-NHS-LC-biotin (Pierce Endogen) and purified over a 10 ml desalting column. Biotin incorporation was determined by the HABA [2-(49-hydroxyazobenzene)benzoic acid] assay (Pierce Endogen), with 1?9 mol biotin incorporated per mol dissociated (a1b1) Hb dimers. Binding of biotinylated Hb to outer membranes immobilized on nitrocellulose. Outer-membrane samples from N. meningitidis,

each containing 200 ng HmbR (quantified as described above), were resuspended in 200 ml TBS and serially diluted twofold. The outer membranes were transferred onto Hybond-P (PVDF) membrane (Amersham Pharmacia) using a microsample filtration manifold (Minifold I, Schleicher & Schuell). The PVDF membrane was blocked with 5 % skim milk in TBST (TBS+0?1 % Tween 20) for 2 h at room temperature. The membrane was incubated with 50 nM biotinylated Hb (16 ml of 2 mg ml21 stock) in 10 ml 5 % skim milk/ TBST for 1 h at room temperature. The membrane was rinsed three times with TBST for 15 min. Hb binding was detected by incubation of the membrane with 5 mCi [35S]streptavidin (Amersham Pharmacia) diluted in 10 ml TBST solution for 1 h at room temperature. After rinsing the membrane with TBST three times for 15 min, Hb binding was quantified using a Molecular Dynamics Storm 860 Phosphorimager and by radiocounting in a Beckman LS6500 scintillation counter. Measurements using both methods were http://mic.sgmjournals.org

compared for consistency. Error in Hb binding values due to nonspecific binding of Hb to outer membranes was corrected by subtracting the radiocounts (phosphorcounts) of equivalent quantities of HmbR2 outer membranes. Construction of hmbR alleles expressed from an araCregulated promoter. The 3?3 kb BamHI–HindIII DNA fragment

carrying the hmbR gene was first cloned into pBAD24 from pIRS626 (Guzman et al., 1995; Stojiljkovic et al., 1995). The DNA fragment carrying the 59 end of the hmbR gene was PCR-amplified using the HmbR–EcoRI and internal HmbR primers. Amplified fragment was digested with EcoRI and NotI restriction endonucleases and cloned into hmbR-containing pBAD24. The resulting plasmid pBAD1352 contains the hmbR start codon positioned immediately downstream of the ribosome-binding site. hmbR mutant alleles were subcloned into pBAD1352 from pACYC184derivative plasmids, pIRS1402, pIRS1725, pIRS1192, pIRS1770, pIRS1185 and pIRS1492. The hmbR-internal NotI–SalI DNA fragment within each of these plasmids was gel purified and cloned into the NotI–SalI region of pBAD1352, replacing the wild-type sequence. Restriction analysis and sequencing was used to confirm the presence of mutant alleles. Outer membranes were prepared from 2?56109 cells of each strain and Western blotting using rabbit polyclonal antiserum against HmbR was performed on these samples to demonstrate outer-membrane localization of mutant HmbRs and equal protein expression. In vivo binding of biotinylated Hb to whole cells and flow cytometry. Hb binding to whole cells was measured using flow

cytometry on a FACS Calibur flow cytometer (Becton Dickinson). Bacteria were inoculated at a 1 : 100 dilution from an overnight culture into fresh Luria broth containing ampicillin (100 mg ml21) and d-aminolaevulinic acid (50 mg ml21). Cells were grown at 37 uC with shaking to an OD600 of 0?3–0?4, and then HmbR expression was induced for 1?5 h with 0?04 % arabinose (unless otherwise indicated). Approximately 0?66108 cells were harvested and washed with PBS-BSA (20 mM sodium phosphate buffer, pH 7?4, 150 mM NaCl, 10 mM MgCl2, 1 % BSA). Cells were resuspended in PBS-BSA with 1 mM biotinylated Hb and incubated on ice for 10 min. Cells were then diluted tenfold, fixed in 1 % formaldehyde PBS-BSA buffer, and washed. PE-conjugated streptavidin (Caltag Laboratories) was used to label cells with bound Hb. Data were collected for 26104 cells per sample as determined by forward light scatter and side scatter intensity, and were analysed using the FlowJo software package (Tree Star).

RESULTS Construction of altered HmbR proteins Model. A hypothetical two-dimensional model of HmbR was constructed as described in Methods. This model was created as a guide for predicting structural features of HmbR, since the actual structure of HmbR has not yet been solved. Based on this topological model, HmbR consists of 22 putative transmembrane b-strands and 11 surface exposed loops (assigned L1 to L11) (Fig. 1). Two putative disulfide bonds are located in L6 and L8. L7 contains an invariable His residue and amino acid motifs, FRAP (‘YRVP’ in HmbR) and NPNL, which are highly conserved in all haem/Hb receptors (Fig. 1) (Bracken et al., 1999; Wandersman & Stojiljkovic, 2000). The extensive amino-terminal region that precedes the b-barrel domain is predicted to form a cork or plug domain reminiscent of 3427

D. Perkins-Balding, M. T. Baer and I. Stojiljkovic

1 mg HmbR

0.1 mg HmbR

1770 (L7)

1725 (L2)

1492 (L6)

1402 (cork)

1192 (L3)

HmbR (WT)

the FepA, FhuA and FecA outer-membrane transporters of E. coli (Buchanan et al., 1999; Ferguson et al., 1998, 2002). The amino acid sequence of this amino-terminal region of HmbR was determined by Edman degradation (data not shown) and the processed form is depicted in Fig. 1.

1185 (L6)

Fig. 1. Topological model of HmbR. The size of each b-strand was arbitrarily taken to be 10 amino acid residues. L1–L11 indicate putative surface-exposed loops of HmbR. Arabic numerals indicate amino acid residues of HmbR. Deletions which affected HmbR function are highlighted in grey. A separate diagram of loop 7 indicates the location of conserved motifs described in the text.

(a) kDa 208

Mutations. Nineteen deletion and site-directed mutations

in HmbR were created to test the predictions of our topological model. Mutations were primarily generated to delete or partially delete putative surface loops; eight (of eleven) of the largest loops were targeted in this analysis. In addition, several deletions were made in the putative cork domain. Allelic exchange was used to replace the wild-type hmbR allele with these mutant alleles in the chromosome of N. meningitidis IR4130. Western blots indicated that HmbR mutant proteins were expressed in the outer membranes of N. meningitidis, which included deletions in the cork domain, L2, L3, L5, L6, L7, L8 and L10 (Fig. 2a). Most of the HmbR mutant proteins that were defective in outer-membrane export or assembly had deletions in predicted transmembrane regions, as expected from deletions of related ferric siderophore 3428

(b)

132 91

35

Fig. 2. Western blot using polyclonal rabbit anti-HmbR antibodies, showing the expression of HmbR and HmbR mutant proteins in outer-membrane samples of N. meningitidis (a), with outer membranes shown below on a Coomassie-stained SDS gel (b). Purified HmbR-His was included on the Western blot. Microbiology 149

Mutagenesis of HmbR

transporters (Klebba et al., 1994; Lathrop et al., 1995; Newton et al., 1999). Two independent deletions of L11 were attempted, but neither produced a mutant protein that could localize to the outer membrane. Ten additional mutants were tested in E. coli and were not transformed into N. meningitidis because they failed to localize to the outer membrane. Haem usage by HmbR deletion mutants in E. coli was assayed in order to provide a general criterion for outermembrane localization and conformation; it was not possible to examine HmbR-dependent haem utilization in N. meningitidis, since there exists an as yet unidentified haem uptake system besides that of Hb receptors HmbR and HpuB (Stojiljkovic & Srinivasan, 1997; Stojiljkovic et al., 1996). All strains containing HmbR mutant proteins which were detected in the outer membrane allowed haem usage by solid-phase growth assays. As expected, no strains were identified in this study that were unable to use haem but were proficient in Hb utilization. To assess whether this haem usage was physiologically relevant and not simply haem leakage through a disordered outer membrane, we looked at whether haem usage was TonB-dependent in E. coli. Plasmid pAP1173 expressing wild-type HmbR was placed in a tonB E. coli strain, IR3758 (strain IR1532, tonB : : IS10). This strain was unable to use haem as a source of iron, whereas isogenic strain IR1583 (tonBEC+, tonBNM+)

carrying the identical plasmid was proficient in haem usage (data not shown). Phenotypic characterization of hmbR deletion derivatives in N. meningitidis Six of the nineteen strains containing hmbR mutant alleles in N. meningitidis (hpuAB) strain IR1072 were functionally defective in either Hb binding or haem utilization and are shown in Fig. 1 and Table 2. The expression of HmbR mutant proteins under ironlimiting conditions in N. meningitidis was compared to the expression of wild-type HmbR by Western blot analysis (Fig. 2a). Equal amounts of outer membranes from N. meningitidis strains were loaded onto SDS gels, and stained with Coomassie blue (Fig. 2b) or immunoblotted with polyclonal antiserum against full-length HmbR (Fig. 2a). The amount of HmbR protein per sample was calculated using a standard curve generated with known amounts of purified HmbR. The expression of mutant proteins was typically less than that of wild-type HmbR, especially mutants 1185 (L6), 1192 (L3) and 1402 (cork), and therefore adjustments were made in the total amount of outer membranes used in in vitro Hb binding assays to normalize HmbR mutant proteins and wild-type HmbR. Immunodominant epitopes may have been deleted in the construction of HmbR mutants; this could subsequently lead to an

Table 2. Changes introduced into HmbR and resulting characteristics in N. meningitidis OM, outer membrane; tm, putative transmembrane region; +, wild-type expression or activity; +/2, reduced activity; 2, no expression or activity; ND, not done. The upper part of the table includes variant proteins which were expressed in the OM, but were functionally defective in either Hb binding or Hb utilization. Haem utilization was tested in E. coli strain IR1583 (see Results). Allele

Mutation

Putative location

OM location

Hb binding

Hb util.

Haem util. (E.c.)

1402 1725 1192 1185 1492 1770

D[H64–N76] D[G203–H231] D[Q254–V261] D[H435–P454] D[N458–G465] D[V489–G502]

cork L2 L3 L6 L6 L7

+ + + + + +

+ +/2 2 + + +

2 2 2 2 2 2

+ + + + + +

1183 1325 1562 1420 1751 1561 1184 1753 1752 1565 1383 1384 1203

D[H64–K66] D[G203S] D[R192–Y230] D[A275–L286] D[L297–S325] D[D324–R369] D[H359–R369] D[Y389–P454] D[N410–P454] D[V489–R527] D[V558–K577] D[K664–D689] D[K664–A669]

cork L2 L2 L3 L4/tm L4/tm L5 L5/tm/L6 L6/tm L7/tm L8 L10 L10

+ + 2 2 2 2 + 2 2 + + + +

+ +

+ + 2 2 2 2 + 2 2 2 + +/2 +

+ + 2 2 2 2 + 2 2 2 + + +

http://mic.sgmjournals.org

ND ND ND ND

+/2 ND ND

2 + + +

3429

D. Perkins-Balding, M. T. Baer and I. Stojiljkovic

underestimation of the amount of mutant protein used in the binding assay. Therefore, the defect in Hb binding observed by mutants may be greater than currently estimated. This would not change the conclusion that certain mutants are deficient in binding, only the severity of the defect. Hb utilization. The HmbR mutants 1402 (cork), 1725 (L2), 1192 (L3), 1185 (L6), 1492 (L6) and 1770 (L7) were unable to support growth of meningococci on Hb as a sole source of iron (Fig. 3b). However, several other HmbR mutants with lowered expression, on the basis of antibody binding, were fully functional in this assay. Hb binding. Hb binding was assayed in vitro by incubat-

ing biotinylated Hb with equal amounts of wild-type and mutant HmbR receptors expressed in N. meningitidis outer membranes attached to a solid support. The amount of Hb binding was quantified by measuring association to an 35S-labelled avidin probe. Two HmbR mutant proteins demonstrated dramatically reduced Hb binding activity when compared to wild-type HmbR. These mutants were 1192 (L3) and 1725 (L2), with 9 % and 33 % activity compared to the wild-type, respectively

(Fig. 3). Deletions in mutants 1192 and 1725 are located close to each other on putative L2 and L3 (Fig. 1). Hb binding to HmbR mutant proteins in vivo In the previous in vitro binding assay, N. meningitidis outer membranes were isolated by detergent extraction and were subject to drying on nitrocellulose. To confirm these Hb binding phenotypes seen in N. meningitidis and to address the possibility that the assay conditions may have affected the binding activity of HmbR mutants, an in vivo Hb binding analysis was performed using whole cells of E. coli expressing HmbR variants at high levels from an araCregulated promoter. The amount of Hb binding by whole cells was assayed by measuring association to fluorochrome PE-conjugated streptavidin using flow cytometry. The greatest sensitivity for the flow cytometry analysis of Hb binding was achieved in E. coli; when using N. meningitidis in this assay the difference between the fluorescence values was small (roughly tenfold) with the wild-type HmbRexpressing cells poorly labelled. Increasing HmbR expression using iron induction and IPTG induction, as well as using a non-encapsulated strain, did not improve labelling enough that differences in mutants could be evaluated with any level of confidence (data not shown). SDS-PAGE analysis and Western blotting confirmed that HmbR expression could be induced from the araCregulated promoter, and nutrition assays demonstrated the ability of E. coli IR1583 expressing HmbR to grow with Hb as a sole source of iron (data not shown). Different concentrations of the arabinose inducer were tested to determine the concentration required to achieve the maximum level of HmbR expression (data not shown). Outer membranes were prepared from E. coli strains expressing HmbR mutant proteins and examined using SDS-PAGE, stained with Coomassie blue (Fig. 4a), and immunoblotting using rabbit polyclonal antiserum against HmbR (Fig. 4b). HmbR mutant proteins overexpressed in E. coli IR1583 localized to the outer membrane and were expressed at high levels, according to a standard curve using purified HmbR. Relative to each other, only mutant 1402 (cork) seemed to be expressed at lowered levels.

Fig. 3. In vitro Hb binding activity of different HmbR mutant proteins relative to the wild-type HmbR. Binding of biotinylated Hb to N. meningitidis outer membranes expressing different HmbR mutant proteins was measured using [35S]avidin. Radioactivity associated with each sample was adjusted for nonspecific binding of Hb and/or avidin to outer-membrane material, and compared as a percentage to the radioactivity associated with the wild-type HmbR sample (the wild-type sample equal to 100 % Hb binding activity). The Hb binding activity of each sample was measured independently at least three times; standard deviations are shown with error bars. N. meningitidis expressing different HmbR mutant proteins were tested for their ability to grow on Hb by the disk diffusion assay described in Methods. +, Growth zones equivalent to the wild-type strain; ”, no growth zones. 3430

Cells that had the ability to bind Hb were easily distinguished by flow cytometry from those that could not (Fig. 5a). E. coli strain 1583 with pBAD24 had mean fluorescence values of 2?5±0?2, whereas the same cells expressing wild-type HmbR from pBAD1352 had a 66-fold greater mean fluorescence value of 165±2?6 (Fig. 5a, b). Fig. 5(b) depicts graphically the abilities of the cells expressing mutant HmbR proteins to bind Hb. Data obtained in vivo by this fluorescent method corresponded well to those obtained in vitro using N. meningitidis outer membranes. Again, HmbR mutant 1192 (L3) had the most severe defect in binding, with mutant 1725 (L2) binding poorly as seen in the previous assay conditions. It is important to note that mutants 1402 (cork), 1185 (L6), 1492 (L6) and 1770 (L7), which are completely unable to utilize Hb, bind Hb either Microbiology 149

(a)

1770 (L7)

1725 (L2)

1492 (L6)

1402 (cork)

1192 (L3)

1185 (L6)

HmbR+

HmbR

_

Mutagenesis of HmbR

kDa 210 135

82

39 (b)

210 135

82

39

Fig. 4. (a) SDS-PAGE of outer membranes from E. coli IR1583 expressing HmbR and HmbR mutant proteins from an araC-regulated promoter; (b) a Western blot of the identical samples, using polyclonal rabbit anti-HmbR antiserum.

very well or close to wild-type levels. Mutant protein 1402 was able to bind Hb proficiently even though the level of expression was lower than that of wild-type HmbR and other HmbR mutant proteins as determined by Western blotting.

DISCUSSION A simplified view of HmbR functions required for Hb utilization assumes that the receptor must bind Hb via a high-affinity site, remove haem from Hb, and transport it through the receptor. These functions should have corresponding receptor domains: a high affinity Hb-binding domain, and haem stripping and transport domains. Since the crystal structure of HmbR has yet to be solved, a working model of HmbR was designed in order to predict its secondary structure and to guide a mutational analysis aimed at locating functionally important domains of the receptor. The model of HmbR predicts 22 transmembranespanning regions form an amphiphilic b-barrel pore structure like the structures of the characterized bacterial outer-membrane proteins (Buchanan, 1999; Koebnik et al., http://mic.sgmjournals.org

Fig. 5. In vivo Hb binding activity of different HmbR mutant proteins and the wild-type HmbR using flow cytometry. (a) Binding of biotinylated Hb to whole cells was measured using streptavidin-PE. The left curve represents the total fluorescence of E. coli strain IR1583 with no HmbR; the right curve represents the total fluorescence of strain IR1583 with the wild-type HmbR expressed from pBAD1352 with arabinose induction. (b) Hb binding activity of cells expressing different HmbR mutant proteins (labelled below each column) from the araC-regulated PBAD promoter with arabinose induction. Mean fluorescence was calculated for each binding curve and the mean was derived from samples repeated in triplicate; 20 000 cells were counted per sample; standard deviations are shown with error bars.

2000). The model predicts eleven extracellular loops of varying sizes and ten simple reverse turns in the periplasmic space. The amino-terminal end of the protein is depicted in the periplasmic space, but probably extends into the pore of the barrel structure, like the plug domains of the ferric siderophore receptors (Buchanan et al., 1999; Clarke et al., 2001; Ferguson et al., 1998, 2002; Lambert et al., 1999). Our mutagenesis-based approach to identify the functional domains of HmbR followed the precedents of other outermembrane protein studies, including LamB and the ferric siderophore receptors (Klebba et al., 1994; Newton et al., 1999). A growing body of evidence indicates that surfaceexposed peptide loops of outer-membrane transporters confer specificity for a substrate (Barnard et al., 2001; 3431

D. Perkins-Balding, M. T. Baer and I. Stojiljkovic

Newton et al., 1999; Scott et al., 2001). Recently, the specificity of TonB-gated porins, FepA and FhuA, and their high affinity for ferric siderophores was attributed to their surface loops by equilibrium binding analyses (Scott et al., 2001). Both loop deletions and PCR mutagenesis of FepA point to substrate specificity function associated with loop domains (Barnard et al., 2001; Newton et al., 1999). Cocrystal structures of FhuA–ferrichrome and FecA–diferric– dicitrate demonstrate binding of ligand to external loops and cork domain (Ferguson et al., 2002; Lambert et al., 1999). Our purpose for targeting surface-exposed loops of HmbR was to reduce the formidable task of studying individual amino acids in such a large protein, and to identify the contributions of individual loops in Hb binding, haem stripping and transport. Previously, mutagenesis of FepA and FhuA has shown that, with some exceptions, deletions and insertions in surface-exposed loops are usually tolerated while deletions of transmembrane regions lead to defects in receptor assembly and/or export (Braun et al., 1999; Buchanan et al., 1999; Ferguson et al., 1998; Killmann et al., 1996; Koebnik & Braun, 1993; Lathrop et al., 1995; Liu et al., 1993; Locher et al., 1998; Newton et al., 1999; Postle, 1999). In addition, HmbR variant proteins described in this study were shown to retain TonB-dependent haem transport function. Productive interactions with the TonB/ ExbBD complex and a stable pore for haem to pass through are good indicators that gross changes in conformation have not resulted from loop deletions.

step and those deficient in Hb utilization steps that follow Hb binding. Deletion mutations which allowed HmbR to bind Hb well, but did not allow the use of Hb as an iron source, localized to the carboxyl-terminal region of HmbR in L6, L7 and the cork domain. These HmbR mutants were proficient in use of free haem, which suggests that the defect is limited to the removal of haem from Hb. The Hb binding of these variant proteins was not at the wild-type level, but they were not dramatically impaired. Since haem removal may be a complex process, it is possible that there are multiple sites of haem contact on several surface loops and the loss of any of these sites may compromise haem removal. This may explain how deletions in L6 exhibit slightly different Hb-binding activities. It is important to emphasize that the outer-membrane localization properties of these HmbR mutants were indistinguishable from those of the wild-type HmbR; again, however, the possibility of secondary effects of deletions on the HmbR structure cannot entirely be ruled out. Deletion of proximal loops, however, L5 and L8, had no effect on Hb binding or haem utilization, suggesting that they are not involved in either process.

This study has produced several important insights into HmbR structure–function relationships despite potential pitfalls of mutagenesis-based approaches. First, it separated the functions of haem from Hb utilization (in E. coli) since all HmbR mutants expressed in the outer membrane were efficient in haem use irrespective of their Hb utilization phenotypes. Haem usage was also shown to be a TonBdependent activity in E. coli, suggesting that haem is passing through the transporter in a physiologically relevant manner and not simply leaking through a disordered outer membrane. The lack of appropriate mutants in TonBindependent haem uptake in N. meningitidis prevented us from determining whether this also holds for HmbR expressed in meningococci. We assessed haem utilization as an indicator of proper conformation and outermembrane localization, because any mutations, especially deletions and insertions, can produce abnormalities in protein structure that may affect protein function in an irrelevant way. Since haem transport was shown to be a TonB-dependent process in E. coli, mutant receptors were also interacting with TonB in an appropriate manner. Only mutants that were proficient in haem utilization and present in the bacterial outer membrane by Western blot analysis were selected for further study. Loss of haem utilization was consistent with a change in tertiary structure of functional importance.

The region of HmbR deleted in L7 (mutant 1770, DV489– G502) is part of a highly conserved amino acid sequence in all haem/Hb receptors (Bracken et al., 1999). Deletion of this conserved domain resulted in a Hb-binding-proficient, but Hb-utilization-deficient phenotype. A portion of this conserved motif including the ‘NPNL’ box has homology to the more distantly related FepA transporter (Bracken et al., 1999). Much like HmbR, deletions in L7 (and L8) of FepA resulted in no binding or transport of ferric enterobactin, and resulted in what was described as an open channel based on increased susceptibility to large antibiotics (Newton et al., 1999). Structural studies of FecA and FecA–diferric– dicitrate complex have indicated substantial changes in the position of L7 and L8 of FecA upon ligand binding (11 A˚ and 15 A˚, respectively) (Van Der Helm et al., 2002). L7 in FecA does not appear to be involved in ligand binding, although L8 shows several interactions (Van Der Helm et al., 2002). This region is also homologous with the region of HemR that contains the functionally important His461 residue. Like the L7 deletion in HmbR, HemR His461 mutants were unable to use any haem–protein complex including Hb while retaining the ability to use free haem as an iron source (Bracken et al., 1999). Histidine residues often serve as axial ligands for haem binding; therefore, we targeted several histidine residues in L7 (and region) for mutagenesis (H448, H501 and H520). Replacement of these residues with alanine, however, resulted in Hb utilization phenotypes in N. meningitidis that were indistinguishable from the wild phenotype (data not shown). Ligand–receptor interactions may be multideterminant in HmbR, or the histidines in L7 may not directly interact with haem.

Second, the study revealed two types of Hb utilizationnegative mutant receptors, those deficient in the Hb binding

Mutant 1402 contains a deletion within the putative cork domain of HmbR (DH64–N76) and also has a

3432

Microbiology 149

Mutagenesis of HmbR

Hb-binding-proficient, but Hb-utilization-deficient phenotype. This amino-terminal deletion of HmbR does not remove the putative TonB box (residues E7–A13) which is conserved among the TonB-dependent transporters. The phenotype of this HmbR mutant suggests that the cork domain plays an active role in the use of haem associated with Hb. Indeed, a smaller deletion (three amino acids) in the same region (1183, DH64–K66) was also partially affected in usage of Hb when expressed in meningococci. Both mutants were fully capable of supporting growth on haem, when expressed in E. coli, suggesting that transport functions of HmbR are not grossly affected. The involvement of the putative cork domain of HmbR in Hb usage is in contrast to the many recent studies of the cork domains of ferric siderophore receptors which report no significant role of the cork in substrate binding or transport (Braun et al., 1999; Howard et al., 2001; Killmann et al., 2001; Scott et al., 2001; Usher et al., 2001). In fact without any contributions of the cork domain, FhuA has been reported to function as a TonB-dependent transporter, transporting at a rate of 45 % of wild-type FhuA (Braun et al., 1999). Recently, some controversy has arisen as to whether in vivo reconstitution of active transporters has occurred in cork-deletion studies of FhuA and FepA (Vakharia & Postle, 2002). Nonetheless, one difference between the ferric siderophore receptors and HmbR is that the whole siderophore traverses the outer membrane via the siderophore receptor, as free haem does with HmbR, but a second function, the removal of haem from Hb, must precede haem transport in HmbR. We propose an explanation where the cork domain of HmbR might be essential in the removal of haem from Hb, and that it is not necessarily involved in the binding or transport of free haem. A third important finding of our study is that deletions of L2 (1725) and L3 (1192) dramatically affected the ability of HmbR to bind and use Hb while preserving the use of haem as an iron source. Both in vitro analysis using purified outer membranes from N. meningitidis and in vivo measurements of Hb binding to E. coli cells expressing HmbR mutant proteins support the involvement of this region of HmbR in Hb binding. We recognize that indirect effects of deletion mutations in loops may perturb the tertiary structure of surface domains other than the deleted loop. Clearly, additional mutagenic studies focusing on smaller regions or on individual residues, which are not as likely to have an impact on structure, will be required. However, the initial target region for analysis has been narrowed by this study. Biochemical studies will also be designed to corroborate the validity of these predictions and explain the mechanism by which HmbR allows usage of Hb and free haem as sources of iron.

ACKNOWLEDGEMENTS We thank Klaus Hantke and Christoph Bo¨s for reading the manuscript and for their suggestions. This work was supported by Public Service Grant R01-AI42870. http://mic.sgmjournals.org

REFERENCES Barnard, T. J., Watson, M. E., Jr & McIntosh, M. A. (2001). Mutations

in the Escherichia coli receptor FepA reveal residues involved in ligand binding and transport. Mol Microbiol 41, 527–536. Baumler, A. J. & Hantke, K. (1992). Ferrioxamine uptake in Yersinia enterocolitica: characterization of the receptor protein FoxA. Mol Microbiol 6, 1309–1321. Benson, S. A., Occi, J. L. & Sampson, B. A. (1988). Mutations that

alter the pore function of the OmpF porin of Escherichia coli K12. J Mol Biol 203, 961–970. Boulanger, P., le Maire, M., Bonhivers, M., Dubois, S., Desmadril, M. & Letellier, L. (1996). Purification and structural and functional

characterization of FhuA, a transporter of the Escherichia coli outer membrane. Biochemistry 35, 14216–14224. Bracken, C. S., Baer, M. T., Abdur-Rashid, A., Helms, W. & Stojiljkovic, I. (1999). Use of heme-protein complexes by the Yersinia

enterocolitica HemR receptor: histidine residues are essential for receptor function. J Bacteriol 181, 6063–6072. Braun, M., Killmann, H. & Braun, V. (1999). The beta-barrel domain

of FhuADelta5–160 is sufficient for TonB-dependent FhuA activities of Escherichia coli. Mol Microbiol 33, 1037–1049. Buchanan, S. K. (1999). Beta-barrel proteins from bacterial outer

membranes: structure, function and refolding. Curr Opin Struct Biol 9, 455–461. Buchanan, S. K., Smith, B. S., Venkatramani, L., Xia, D., Esser, L., Palnitkar, M., Chakraborty, R., van der Helm, D. & Deisenhofer, J. (1999). Crystal structure of the outer membrane active transporter

FepA from Escherichia coli. Nat Struct Biol 6, 56–63. Carmel, G. & Coulton, J. W. (1991). Internal deletions in the FhuA

receptor of Escherichia coli K-12 define domains of ligand interactions. J Bacteriol 173, 4394–4403. Clarke, T. E., Tari, L. W. & Vogel, H. J. (2001). Structural biology of bacterial iron uptake systems. Curr Top Med Chem 1, 7–30. Cope, L. D., Thomas, C. E., Latimer, J. L., Slaughter, C. A., Muller-Eberhard, U. & Hansen, E. J. (1994). The 100 kDa

haem : haemopexin-binding protein of Hemophilus influenzae: structure and function. Mol Microbiol 13, 863–873. Eisenberg, D., Schwarz, E., Komaromy, M. & Wall, R. (1984).

Analysis of membrane and surface protein sequences with the hydrophobic moment plot. J Mol Biol 179, 125–142. Ferguson, A. D., Hofmann, E., Coulton, J. W., Diederichs, K. & Welte, W. (1998). Siderophore-mediated iron transport: crystal structure of

FhuA with bound lipopolysaccharide. Science 282, 2215–2220. Ferguson, A. D., Chakraborty, R., Smith, B. S., Esser, L., van der Helm, D. & Deisenhofer, J. (2002). Structural basis of gating by the

outer membrane transporter FecA. Science 295, 1715–1719. Genco, C. A. & Dixon, D. W. (2001). Emerging strategies in microbial haem capture. Mol Microbiol 39, 1–11. Gray-Owen, S. D. & Schryvers, A. B. (1996). Bacterial transferrin

and lactoferrin receptors. Trends Microbiol 4, 185–191. Gribskov, M., Burgess, R. R. & Devereux, J. (1986). PEPPLOT,

a protein secondary structure analysis program for the UWGCG sequence analysis software package. Nucleic Acids Res 14, 327–334. Griffiths, E. (1999). Iron and Infection: Molecular, Physiological and Clinical Aspects, 2nd edn. Chichester: Wiley. Guzman, L. M., Belin, D., Carson, M. J. & Beckwith, J. (1995).

Tight regulation, modulation, and high-level expression by vectors containing the arabinose PBAD promoter. J Bacteriol 177, 4121–4130. 3433

D. Perkins-Balding, M. T. Baer and I. Stojiljkovic Hantke, K. (1981). Regulation of ferric iron transport in Escherichia coli K12: isolation of a constitutive mutant. Mol Gen Genet 182, 288–292.

Newton, S. M., Igo, J. D., Scott, D. C. & Klebba, P. E. (1999). Effect of

Henderson, D. P. & Payne, S. M. (1994). Characterization of the

Pettersson, A., Klarenbeek, V., van Deurzen, J., Poolman, J. T. & Tommassen, J. (1994). Molecular characterization of the structural

Vibrio cholerae outer membrane heme transport protein HutA: sequence of the gene, regulation of expression, and homology to the family of TonB-dependent proteins. J Bacteriol 176, 3269–3277.

loop deletions on the binding and transport of ferric enterobactin by FepA. Mol Microbiol 32, 1153–1165.

gene for the lactoferrin receptor of the meningococcal strain H44/76. Microb Pathog 17, 395–408.

Howard, S. P., Herrmann, C., Stratilo, C. W. & Braun, V. (2001). In

Postle, K. (1999). Active transport by customized beta-barrels. Nat

vivo synthesis of the periplasmic domain of TonB inhibits transport through the FecA and FhuA iron siderophore transporters of Escherichia coli. J Bacteriol 183, 5885–5895.

Struct Biol 6, 3–6.

Jeanteur, D., Lakey, J. H. & Pattus, F. (1991). The bacterial porin

superfamily: sequence alignment and structure prediction. Mol Microbiol 5, 2153–2164. Johnston, D. M. & Cannon, J. G. (1999). Construction of mutant

strains of Neisseria gonorrhoeae lacking new antibiotic resistance markers using a two gene cassette with positive and negative selection. Gene 236, 179–184. Kellogg, D. S., Peacock, W. L., Deacon, W. E., Brown, L. & Pirkle, C. I. (1963). Neisseria gonorrhoeae. I. Virulence genetically linked to clonal

variation. J Bacteriol 85, 1274–1279. Killmann, H., Benz, R. & Braun, V. (1996). Properties of the FhuA

channel in the Escherichia coli outer membrane after deletion of FhuA portions within and outside the predicted gating loop. J Bacteriol 178, 6913–6920. Killmann, H., Braun, M., Herrmann, C. & Braun, V. (2001). FhuA

barrel-cork hybrids are active transporters and receptors. J Bacteriol 183, 3476–3487. Klebba, P. E., Hofnung, M. & Charbit, A. (1994). A model of

maltodextrin transport through the sugar-specific porin, LamB, based on deletion analysis. EMBO J 13, 4670–4675. Koebnik, R. & Braun, V. (1993). Insertion derivatives containing

segments of up to 16 amino acids identify surface- and periplasmexposed regions of the FhuA outer membrane receptor of Escherichia coli K-12. J Bacteriol 175, 826–839. Koebnik, R., Locher, K. P. & Van Gelder, P. (2000). Structure and

function of bacterial outer membrane proteins: barrels in a nutshell. Mol Microbiol 37, 239–253. Lambert, O., Moeck, G. S., Levy, D., Plancon, L., Letellier, L. & Rigaud, J. L. (1999). An 8-A˚ projected structure of FhuA, A ‘‘ligand-

gated’’ channel of the Escherichia coli outer membrane. J Struct Biol 126, 145–155. Lathrop, J. T., Wei, B. Y., Touchie, G. A. & Kadner, R. J. (1995).

Ratliff, M., Zhu, W., Deshmukh, R., Wilks, A. & Stojiljkovic, I. (2001).

Homologues of neisserial heme oxygenase in gram-negative bacteria: degradation of heme by the product of the pigA gene of Pseudomonas aeruginosa. J Bacteriol 183, 6394–6403. Richardson, A. R. & Stojiljkovic, I. (1999). HmbR, a hemoglobin-

binding outer membrane protein of Neisseria meningitidis, undergoes phase variation. J Bacteriol 181, 2067–2074. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). Molecular Cloning:

a Laboratory Manual, 2nd edn. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory. Schirmer, T. & Cowan, S. W. (1993). Prediction of membrane-

spanning beta-strands and its application to maltoporin. Protein Sci 2, 1361–1363. Schryvers, A. B. & Stojiljkovic, I. (1999). Iron acquisition systems in

the pathogenic Neisseria. Mol Microbiol 32, 1117–1123. Schryvers, A. B., Bonnah, R., Yu, R. H., Wong, H. & Retzer, M. (1998). Bacterial lactoferrin receptors. Adv Exp Med Biol 443,

123–133. Scott, D. C., Cao, Z., Qi, Z., Bauler, M., Igo, J. D., Newton, S. M. & Klebba, P. E. (2001). Exchangeability of N termini in the ligand-

gated porins of Escherichia coli. J Biol Chem 276, 13025–13033. Stojiljkovic, I. & Hantke, K. (1992). Hemin uptake system of Yersinia

enterocolitica: similarities with other TonB-dependent systems in gram-negative bacteria. EMBO J 11, 4359–4367. Stojiljkovic, I. & Hantke, K. (1994). Transport of haemin across the

cytoplasmic membrane through a haemin-specific periplasmic binding-protein-dependent transport system in Yersinia enterocolitica. Mol Microbiol 13, 719–732. Stojiljkovic, I. & Perkins-Balding, D. (2002). Processing of heme

and heme-containing proteins by bacteria. DNA Cell Biol 21, 281–295. Stojiljkovic, I. & Srinivasan, N. (1997). Neisseria meningitidis tonB,

exbB, and exbD genes: Ton-dependent utilization of protein-bound iron in Neisseriae. J Bacteriol 179, 805–812.

Sequences of the Escherichia coli BtuB protein essential for its insertion and function in the outer membrane. J Bacteriol 177, 6810–6819.

Stojiljkovic, I., Hwa, V., de Saint Martin, L., O’Gaora, P., Nassif, X., Heffron, F. & So, M. (1995). The Neisseria meningitidis haemoglobin

Letoffe, S., Ghigo, J. M. & Wandersman, C. (1994). Iron acquisition

receptor: its role in iron utilization and virulence. Mol Microbiol 15, 531–541.

from heme and hemoglobin by a Serratia marcescens extracellular protein. Proc Natl Acad Sci U S A 91, 9876–9880. Letoffe, S., Redeker, V. & Wandersman, C. (1998). Isolation and

characterization of an extracellular haem-binding protein from Pseudomonas aeruginosa that shares function and sequence similarities with the Serratia marcescens HasA haemophore. Mol Microbiol 28, 1223–1234. Liu, J., Rutz, J. M., Feix, J. B. & Klebba, P. E. (1993). Permeability

properties of a large gated channel within the ferric enterobactin receptor, FepA. Proc Natl Acad Sci U S A 90, 10653–10657. Locher, K. P., Rees, B., Koebnik, R., Mitschler, A., Moulinier, L., Rosenbusch, J. P. & Moras, D. (1998). Transmembrane signal-

ing across the ligand-gated FhuA receptor: crystal structures of free and ferrichrome-bound states reveal allosteric changes. Cell 95, 771–778. 3434

Stojiljkovic, I., Larson, J., Hwa, V., Anic, S. & So, M. (1996). HmbR

outer membrane receptors of pathogenic Neisseria spp. ironregulated, hemoglobin-binding proteins with a high level of primary structure conservation. J Bacteriol 178, 4670–4678. Urushibara, N., Kumazaki, T. & Ishii, S. (1992). Hemoglobin-

binding site on human haptoglobin. Identification of lysyl residues participating in the binding. J Biol Chem 267, 13413–13417. Usher, K. C., Ozkan, E., Gardner, K. H. & Deisenhofer, J. (2001).

The plug domain of FepA, a TonB-dependent transport protein from Escherichia coli, binds its siderophore in the absence of the transmembrane barrel domain. Proc Natl Acad Sci U S A 98, 10676–10681. Vakharia, H. L. & Postle, K. (2002). FepA with globular domain

deletions lacks activity. J Bacteriol 184, 5508–5512. Microbiology 149

Mutagenesis of HmbR

Van Der Helm, D., Chakraborty, R., Ferguson, A. D., Smith, B. S., Esser, L. & Deisenhofer, J. (2002). Bipartite gating in the outer

Winkelmann, G. (2002). Microbial siderophore-mediated transport. Biochem Soc Trans 30, 691–696.

membrane protein FecA. Biochem Soc Trans 30, 708–710.

Zhu, W., Wilks, A. & Stojiljkovic, I. (2000a). Degradation of heme in

Wandersman, C. & Stojiljkovic, I. (2000). Bacterial heme sources: the

gram-negative bacteria: the product of the hemO gene of Neisseriae is a heme oxygenase. J Bacteriol 182, 6783–6790.

role of heme, hemoprotein receptors and hemophores. Curr Opin Microbiol 3, 215–220. Wang, R. F. & Kushner, S. R. (1991). Construction of versatile low-

copy-number vectors for cloning, sequencing and gene expression in Escherichia coli. Gene 100, 195–199.

http://mic.sgmjournals.org

Zhu, W., Hunt, D. J., Richardson, A. R. & Stojiljkovic, I. (2000b).

Use of heme compounds as iron sources by pathogenic neisseriae requires the product of the hemO gene. J Bacteriol 182, 439–447.

3435