Heme Transfer from Streptococcal Cell Surface Protein Shp to HtsA of Transporter HtsABC

Abstract

Human pathogen group A streptococcus (GAS) can take up heme from host heme-containing proteins as a source of iron. Little is known about the heme acquisition mechanism in GAS. We recently identified a streptococcal cell surface protein (designated Shp) and the lipoprotein component (designated HtsA) of an ATP-binding cassette (ABC) transporter made by GAS as heme-binding proteins. In an effort to delineate the molecular mechanism involved in heme acquisition by GAS, heme-free Shp (apo-Shp) and HtsA (apo-HtsA) were used to investigate heme transfer from heme-containing proteins (holo proteins) to the apo proteins. In addition, the interaction between holo-Shp and holo-HtsA was examined using native polyacrylamide gel electrophoresis. Heme was efficiently transferred from holo-Shp to apo-HtsA but not from holo-HtsA to apo-Shp. Apo-Shp acquired heme from human hemoglobin, and holo-Shp and holo-HtsA were able to form a complex, suggesting that Shp actively relays heme from hemoglobin to apo-HtsA. These findings demonstrate for the first time complex formation and directional heme transfer between a cell surface heme-binding protein and the lipoprotein of a heme-specific ABC transporter in gram-positive bacteria.

Iron is an essential nutrient required for the growth and survival of most bacterial pathogens (27). There is insufficient free iron to support bacterial growth in mammalian hosts due to the extremely low solubility of ferric ion in water at physiological pH (28). To survive in hosts, bacterial pathogens must assimilate iron from host proteins (20). The majority of iron in mammalian hosts is present in the form of protein-bound heme, and heme is thus believed to be a major source of iron for bacterial pathogens. To acquire heme, bacteria produce at least four different types of heme-binding proteins. The first class includes substrate-binding proteins of heme-specific ATP-binding cassette (ABC) transporters, which transport heme across the cytoplasmic membrane in both gram-positive and gram-negative pathogens (5, 11, 14, 25, 26). The second category consists of secreted heme-binding proteins called hemophores, which are produced by certain gram-negative bacteria to scavenge free or protein-bound heme and deliver it to specific cell surface receptors (16, 17). The third group consists of heme-binding outer membrane receptors (3, 21, 22, 29), while the fourth includes the cell surface heme-binding proteins of gram-positive pathogens (13, 19).

In gram-negative bacteria, specific outer membrane receptors sequester heme from heme-hemophore complexes and/or host heme proteins and transport it into the periplasmic space in a TonB-dependent process (12). Specific ABC transporters then transport heme across the cytoplasmic membrane (10). The processes concerning heme acquisition in gram-positive pathogens are less well understood. Progress has recently been made in identifying the machinery for heme acquisition in certain gram-positive bacteria. ABC transporters are required for heme acquisition in Corynebacterium diphtheriae (5); Streptococcus pyogenes, or group A streptococcus (GAS) (1, 13, 14); and Staphylococcus aureus (23). Cell surface heme-binding proteins appear to play a role in heme acquisition in GAS (1, 13) and S. aureus (19). However, the mechanisms by which cell surface heme-binding proteins and heme-specific ABC transporters acquire heme from host proteins remain unknown.

GAS is a major human pathogen that causes a variety of diseases such as pharyngitis, necrotizing fasciitis, and streptococcal toxic shock syndrome (4). GAS can take up heme from hemoglobin and haptoglobin-hemoglobin complexes as a source of iron (6, 8). We recently identified two heme-binding proteins made by GAS, Shp (13) and HtsA (14). Shp is a cell surface protein, while HtsA represents the lipoprotein component of an ABC transporter (designated the heme transporter of group A streptococcus, or HtsABC). Bates et al. (1) reported that HtsABC (which they designated SiaABC for streptococcal iron acquisition transporter) is involved in the binding of hemoproteins and acquisition of iron. shp, htsABC, or siaABC genes, an upstream gene shr, and five other contiguous downstream genes are cotranscribed (1, 13). The purpose of this study was to examine how Shp and HtsA participate in heme acquisition. We investigated heme transfer from heme-containing proteins (holo-Shp, holo-HtsA, and human hemoglobin) to heme-free Shp and HtsA (apo-Shp and apo-HtsA). We found that apo-Shp acquires heme from hemoglobin and that holo-Shp forms a complex with holo-HtsA and rapidly transfers its heme to apo-HtsA, suggesting that Shp actively relays heme from hemoglobin to apo-HtsA. This is the first demonstration of complex formation and the efficient heme transfer between a cell surface heme-binding protein and the lipoprotein of a heme-specific ABC transporter in gram-positive bacteria.

MATERIALS AND METHODS

Materials.

Human hemoglobin was purchased from ICN Pharmaceuticals, Inc. (Aurora, OH). Sephadex G-25, SP-Sepharose, phenyl-Sepharose and DEAE-Sepharose were obtained from Amersham (Piscataway, NJ). Nickel-nitrilotriacetic acid agarose (Ni-NTA) was obtained from QIAGEN (Valencia, CA). Bovine hemin chloride was obtained from Sigma (St. Louis, MO). Solutions were buffered with 20 mM Tris-HCl, pH 8.0, unless otherwise specified.

Proteins.

Recombinant Shp, which was prepared as previously described (13), was tag free and contained amino acids 30 to 258 of its precursor. The protein lacked the presumed secretion signal sequence (amino acids 1 to 29) and the transmembrane domain and charged tail (amino acids 259 to 291). The recombinant Shp protein moiety had a calculated pI of 5.23 and a −4.6 charge at pH 7.0. This protein bound to both SP-Sepharose (cation exchange) and DEAE-Sepharose (anion exchange) columns at pH 8.0. His-tagged recombinant HtsA (14) and SPy0252 (15) were prepared as previously described and had 12 amino acids, MHHHHHHLETMG, fused to the second amino acid residue of their mature form at the amino terminus. The HtsA protein moiety had a calculated pI of 5.99 and a −6.29 charge at pH 7.0. SPy0252 had a calculated pI of 5.6 and a −6.9 charge at pH 7.0. Both HtsA and SPy0252 bound to DEAE-Sepharose but not to SP-Sepharose.

Preparation of apo-HtsA and apo-Shp.

HtsA was prepared as previously described (14) and contained 80% apo-HtsA and 20% holo-HtsA. To separate apo-HtsA from holo-HtsA, this sample in Tris-HCl was loaded onto a 2.5- by 20-cm DEAE-Sepharose column. The column was eluted with 110 mM NaCl in Tris-HCl. The fractions containing >98% apo-HtsA were pooled. The pooled sample was dialyzed against 3 liters Tris-HCl and concentrated using a Centricon Plus 20 filtration device (Millipore, Bedford, MA).

Apo-Shp was prepared by removal of the bound heme from holo-Shp with apo-HtsA. Apo-HtsA (>98% heme free, 0.28 mM) was incubated with 0.12 mM holo-Shp in 0.3 ml Tris-HCl for 1 h, loaded onto a 1- by 3-cm SP-Sepharose column, and then washed with 5 ml Tris-HCl. Apo-Shp was eluted with 100 mM NaCl in Tris-HCl. The sample obtained was passed through a Ni-NTA column (0.2 ml resin) to remove residual HtsA, resulting in HtsA-free apo-Shp in the flowthrough. A negative control experiment was similarly performed using SPy0252 in place of apo-HtsA. The Shp protein samples isolated did not have HtsA or SPy0252 based on sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis.

Reconstitution of holo-Shp.

Holo-Shp was reconstituted from bovine hemin and apo-Shp. Apo-Shp (100 μl of 40 μM) was incubated with 80 μM hemin in 100 mM Tris-HCl, pH 8.8, for 10 min, loaded onto a Sephadex G-25 column (0.5 by 30 cm), and then eluted with Tris-HCl. The holo-Shp protein collected was dialyzed against 2 liters Tris-HCl.

Determination of protein concentration and heme/hemin content.

Protein concentrations were determined with a modified Lowry protein assay kit from Pierce (Rockford, IL) using bovine serum albumin as a standard (18). A pyridine hemochrome assay (9) was used to assess the heme/hemin content of protein samples. Protein samples in 750 μl Tris-HCl were mixed with 175 μl pyridine, 75 μl of 1 N NaOH, and approximately 2 mg sodium hydrosulfite. The optical spectrum was then immediately recorded using a SPECTRAmax 384 Plus spectrophotometer (Molecular Devices, Sunnyvale, CA). Heme and hemin contents were determined by measuring the absorbance at 418 nm with the extinction coefficient ɛ₄₁₈ = 191.5 mM⁻¹ cm⁻¹ (9).

Heme transfer.

Heme transfer from holo-Shp to apo-HtsA or from hemoglobin to apo-Shp was monitored by following the Soret absorption peak shift of the protein mixtures. Holo-Shp had 22-nm and 16-nm red shifts in the Soret absorption peak compared with human hemoglobin and HtsA, respectively. Proteins at concentrations indicated in the figure legends were mixed in Tris-HCl and incubated at the indicated temperature. Absorption spectra of the mixtures were measured at the indicated times using a SPECTRAmax 384 spectrophotometer. Heme transfer from holo-Shp to apo-HtsA was also confirmed as follows. Apo-HtsA (10.5 μM) was incubated with 13 μM holo-Shp in 0.5 ml Tris-HCl at room temperature for 2 h. The sample was loaded onto a column containing 0.2 ml Ni-NTA resin, and the column was sequentially washed with 10 ml Tris-HCl, 2 ml of 5 mM imidazole in Tris-HCl, and 2 ml of 25 mM imidazole in Tris-HCl. The bound proteins were eluted with 250 mM imidazole in Tris-HCl. Heme associated with the recovered protein was detected by the presence of the Soret absorption peak and a yellow band following native PAGE and by the pyridine hemochrome assay. For native PAGE analysis, samples were mixed with an equal volume of 2× native sample buffer containing 62.5 mM Tris-HCl (pH 6.8), 40% glycerol, and 0.01% bromophenol blue (Bio-Rad, Hercules, CA), loaded onto a 12% polyacrylamide Tris-HCl gel, and electrophoresed using running buffer without sodium dodecyl sulfate at 70 V for 4 h.

Detection of the Shp/HtsA complex.

Holo-Shp and reconstituted holo-HtsA were incubated in Tris-HCl for 30 min and were resolved by native PAGE as described above. Gels were stained with GelCode blue (Pierce), and protein bands were analyzed by SDS-PAGE as follows. Gel pieces containing the indicated bands were excised from the gel and boiled in 20 μl of 1× Laemmli sample buffer (Bio-Rad) for 4 min and were then subjected to SDS-PAGE. The appearance of a band which did not correspond to that of each protein alone and contained both proteins was considered evidence for complex formation.

RESULTS

Heme transfer from holo-Shp to apo-HtsA.

If Shp forms part of the heme-acquisition machinery in GAS, we could expect the transfer of heme from holo-Shp to apo-HtsA. Apo-HtsA was needed to test this idea. We previously reported that 80% of recombinant HtsA was in a heme-free form (14). Anion exchange chromatography was used to further resolve the apo and holo forms, resulting in the preparation of apo-HtsA that was >98% heme free.

To ascertain whether holo-Shp could transfer its heme to apo-HtsA, the shift in the Soret absorption peak of bound heme in the holo-Shp/apo-HtsA mixture was monitored, since holo-Shp and holo-HtsA possess a 16-nm difference in the Soret peak. As shown in Fig. 1A, within 20 min after holo-Shp was mixed with apo-HtsA, the holo-Shp absorption peak at 428 nm shifted to 415 nm, which was close to the 412-nm peak of holo-HtsA. A dramatic shift in the spectrum occurred by the time the initial spectrum was collected immediately following mixing, and the shift was nearly completed within 20 min. Given that HtsA possesses the His tag, it could possibly interact with holo-Shp, thus resulting in the observed spectral change. To rule out this possibility, the control protein SPy0252 with the same His tag was mixed with holo-Shp. No spectral shift was observed even 1 h after the two proteins were mixed (Fig. 1B), indicating that the His tag did not contribute to the spectral shift in the holo-Shp/apo-HtsA mixture. These results suggested that holo-Shp rapidly transferred its heme to apo-HtsA.

FIG. 1.

Spectral evidence for heme transfer from holo-Shp to apo-HtsA. (A) The shift over time in the Soret absorption peak of a mixture containing 6.5 μM holo-Shp and 13.3 μM apo-HtsA at room temperature. The spectra of apo-HtsA alone (solid curve) and holo-Shp alone (dashed line) are included for comparison. (B) The Soret absorption peaks of 10 μM holo-Shp alone (solid circles) and a mixture of 10 μM holo-Shp and 18 μM SPy0252 (open circles). The spectrum of the holo-Shp/SPy0252 mixture was obtained 1 h after the proteins were mixed.

If heme was transferred from holo-Shp to apo-HtsA, HtsA obtained from the holo-Shp/apo-HtsA mixture should have been converted into holo-HtsA. Since HtsA contained a His tag and Shp did not, affinity chromatography with a Ni-NTA column was used to isolate HtsA from a mixture containing 10.5 μM apo-HtsA and 13 μM holo-Shp. HtsA isolated from the mixture contained no Shp, as determined by SDS-PAGE analysis (Fig. 2A, lane 3), and yielded a spectrum similar to that of holo-HtsA (14) (Fig. 2B). Protein and heme content measurements indicated that the HtsA sample had 0.95 heme per HtsA molecule. These results indicated that apo-HtsA acquires heme from holo-Shp during the incubation, consistent with the observed shift in the Soret peak shown in Fig. 1A. Heme-containing proteins appear as yellow bands in native PAGE, and this feature was used to further test whether treated apo-HtsA is converted into holo-HtsA. Native PAGE analysis showed that the HtsA protein isolated from the holo-Shp/apo-HtsA mixture yielded a yellow band that had the same mobility as reconstituted holo-HtsA (Fig. 2C). These results confirmed that holo-Shp transfers its heme to apo-HtsA.

FIG. 2.

Additional evidence for heme transfer from holo-Shp to apo-HtsA. Apo-HtsA (10.5 μM) was incubated with 13 μM holo-Shp in 0.5 ml of 20 mM Tris-HCl, pH 8.0, at room temperature for 2 h. The sample was then loaded onto a column with 0.2 ml Ni-NTA resin and eluted as described in the text. (A) SDS-PAGE analysis showing separation of HtsA from its mixture with Shp. Lanes: 1, control HtsA; 2, the incubated mixture of apo-HtsA and holo-Shp before separation; 3, HtsA isolated from the mixture using Ni-NTA; 4, control Shp. (B) The spectra of apo-HtsA (dashed curves) and HtsA (solid curve) isolated from the holo-Shp/apo-HtsA mixture. (C) Photograph showing the yellow band of the following samples resolved by native PAGE. Lanes: 1, reconstituted holo-HtsA; 2, HtsA isolated from the holo-Shp/apo-HtsA mixture.

Preparation of apo-Shp.

Apo-Shp could not be prepared by separating its bound heme from the protein in saturated urea solution (13). However, heme transfer from holo-Shp to apo-HtsA could be utilized to prepare apo-Shp. Holo-Shp was incubated with apo-HtsA for 2 h at a holo-Shp:apo-HtsA molar ratio of 1:2. The Shp protein was then separated from HtsA using SP-Sepharose and Ni-NTA chromatography. The Shp protein obtained lacked the Soret absorption peak (Fig. 3) and was 99% heme free on the basis of protein and heme content measurements, indicating that apo-Shp was successfully prepared following heme transfer from holo-Shp to apo-HtsA. A parallel control experiment was also performed using SPy0252, and Shp isolated from a mixture with SPy0252 retained its heme (Fig. 3), indicating that heme transfer was specific for apo-HtsA and that the His tag was not involved in the heme assimilation from holo-Shp by apo-HtsA.

FIG. 3.

Preparation of apo-Shp. Holo-Shp (0.12 mM) was incubated with 0.28 mM apo-HtsA or control protein SPy0252 in 0.3 ml Tris-HCl for 1 h, and Shp protein was then isolated from the mixtures as described in Materials and Methods. Presented are the spectra of the Shp samples isolated from their mixture with apo-HtsA (solid curve) and SPy0252 (dashed curve).

Reconstitution of holo-Shp.

To test whether apo-Shp retained its ability to bind heme, apo-Shp was incubated with hemin and then separated from free hemin by gel filtration using a Sephadex G-25 column. The spectra of Shp before and after the treatment are presented in Fig. 4, showing that treated Shp had the typical spectrum of hemin-binding proteins. As assessed by the pyridine hemochrome and protein assays, the protein bound 0.96 hemin per protein molecule. These results indicate that apo-Shp can bind hemin.

FIG. 4.

Reconstitution of holo-Shp. Apo-Shp was incubated with hemin, and the protein was separated from free hemin by gel filtration using a Sephadex G-25 column. Presented are the optical absorption spectra of apo-Shp (dashed curve) and reconstituted holo-Shp (solid curve) at 13.5 μM.

Heme transfer from holo-Shp to apo-HtsA was nearly irreversible.

Heme transfer from holo-Shp to apo-HtsA suggested that Shp relays heme from host and/or other streptococcal heme proteins to the HtsABC transporter. If so, the equilibrium between holo-Shp/apo-HtsA and apo-Shp/holo-HtsA should favor apo-Shp/holo-HtsA. To test this idea, the Soret peaks of holo-Shp/excess apo-HtsA and holo-HtsA/excess apo-Shp mixtures were compared with those of holo-Shp alone and holo-HtsA alone. Under these conditions, the Soret peaks of the mixtures were closer to that of holo-HtsA than that of holo-Shp (Fig. 5), indicating that the equilibrium favored apo-Shp/holo-HtsA. These results suggested that heme transfer occurs efficiently from holo-Shp to apo-HtsA but not from holo-HtsA to apo-Shp.

FIG. 5.

The equilibrium between holo-Shp/apo-HtsA and apo-Shp/holo-HtsA favored apo-Shp/holo-HtsA. Excess apo-Shp (21 μM) and 4.8 μM holo-HtsA (open circles) or excess apo-HtsA (9.2 μM) and 5.5 μM holo-Shp (solid circles) were incubated in 20 mM Tris-HCl, pH 8.0, at room temperature for 1 h. Holo-HtsA was prepared from the apo-HtsA/holo-Shp mixture. Absorption spectra of the mixtures in the Soret peak region were collected and compared with those of holo-HtsA alone (solid curve) and holo-Shp alone (dashed curve). The spectra were normalized for comparison.

Assimilation of heme from hemoglobin by apo-Shp.

If Shp plays a role in relaying heme from host and/or other streptococcal heme proteins to the HtsABC transporter, apo-Shp should be able to acquire heme from other heme proteins. Holo-Shp and hemoglobin had Soret absorption peaks at 428 nm and 406 nm, respectively. By monitoring the spectral change based on this difference, we examined whether apo-Shp could acquire heme from hemoglobin. Figure 6 shows the shift in the Soret absorption peak of the mixture over time. At time zero when human hemoglobin containing 7.5 μM heme was mixed with 20 μM apo-Shp, the absorption peak was at 406 nm, due to heme bound to hemoglobin (Fig. 6). The absorption decreased at 406 nm and increased around 425 nm over time. The peak eventually shifted to 425 nm, indicating acquisition of the heme from hemoglobin by apo-Shp. Hemoglobin was not degraded by Shp during heme acquisition, as determined by SDS-PAGE. Thus, apo-Shp can acquire heme from hemoglobin.

FIG. 6.

Spectral evidence for heme transfer from hemoglobin to apo-Shp. Apo-Shp (27 μM) was mixed with human hemoglobin containing 8.0 μM heme in 20 mM Tris-HCl, pH 8.0, at room temperature, and the absorption spectra of the sample were obtained at the indicated times after the proteins were mixed. Presented is the shift over time in the Soret absorption peak from that of hemoglobin to that of holo-Shp.

Effect of temperature on heme transfer.

Heme transfer from hemoglobin to apo-Shp was relatively slow compared to heme transfer from holo-Shp to apo-HtsA at room temperature. To investigate whether the kinetics of the heme exchange is altered at physiological temperature, heme transfer was compared at 24°C and 37°C. The rate of heme transfer from hemoglobin to apo-Shp was greatly enhanced at 37°C compared with that at 24°C (Fig. 7A). In contrast, there was no dramatic change in the rate of heme transfer from holo-Shp to apo-HtsA when temperature increased from 24°C to 37°C (Fig. 7B). These results indicated that heme transfer from hemoglobin to apo-Shp possesses a higher energy barrier than heme transfer from holo-Shp to apo-HtsA.

FIG. 7.

Effect of temperature on heme transfer. (A) Change in absorbance at 425 nm (ΔA₄₂₅) versus time monitoring heme transfer from human hemoglobin to apo-Shp at 24°C (open circles) and 37°C (solid circles). Apo-Shp (30 μM) was mixed with human hemoglobin containing 8.0 μM heme in 20 mM Tris-HCl, pH 8.0, at time zero at the indicated temperatures. (B) Change in absorbance at 410 nm (ΔA₄₁₀) versus time following heme transfer from holo-Shp to apo-HtsA at 24°C (open circles) and 37°C (solid circles). Holo-Shp (10.2 μM) was mixed with 18 μM apo-HtsA in 20 mM Tris-HCl, pH 8.0, at time zero at the indicated temperature.

Shp/HtsA complex.

Initial native PAGE of the holo-Shp/holo-HtsA mixture revealed unexpected shifts in the mobility of HtsA and Shp. We examined whether the shifts were due to the formation of a Shp/HtsA complex. Holo-HtsA was incubated with variable amounts of holo-Shp at holo-HtsA:holo-Shp molar ratios from 1:1 to 1:4 at room temperature for 30 min and then subjected to native PAGE analysis. As shown in Fig. 8A, the 1:1 sample showed a band which was not present in the samples consisting of holo-HtsA or holo-Shp alone, and the intensities of the HtsA and Shp bands of the 1:1 sample were less than expected. The intensity of the novel band slightly increased when higher levels of holo-Shp were used. An SPy0252/holo-Shp mixture control yielded only the expected SPy0252 and holo-Shp bands. These results suggested that the novel band is due to a specific complex of HtsA and Shp. To verify that the novel band was an HtsA/Shp complex, proteins present in the novel and other bands of the native PAGE gel of panel A were subjected to SDS-PAGE. The novel band contained similar levels of Shp and HtsA (Fig. 8B, lane 4), while the other bands contained only HtsA, Shp, or SPy0252 as expected. The HtsA sample did not possess significant amounts of HtsA at the location corresponding to the location of the complex band (Fig. 8B, lane 2). These results demonstrated that holo-HtsA and holo-Shp form a complex.

FIG. 8.

Formation of a holo-Shp/holo-HtsA complex. (A) Native PAGE analysis of a holo-HtsA/holo-Shp mixture. Holo-HtsA (60 μM) was incubated with holo-Shp at the indicated holo-HtsA:holo-Shp molar ratios in 20 mM Tris-HCl, pH 8.0, at room temperature for 30 min, mixed with an equal volume of 2× native sample buffer, and resolved by PAGE under native conditions. A mixture of 120 μM holo-Shp and 60 μM SPy0252 was also analyzed as a control. Holo-HtsA (rHtsA), holo-Shp (hShp), and SPy0252 (252) were included as references. The numbered boxes indicate the pieces of gels from which proteins were extracted for SDS-PAGE analysis shown in panel B. (B) SDS-PAGE analysis of selected bands from panel A. Pieces of gel indicated by the numbered boxes in panel A were excised. Proteins in these gel pieces were extracted by boiling in 20 μl of 1× Laemmli sample buffer for 4 min and resolved by SDS-PAGE through a 12% polyacrylamide gel. Lane numbers correspond to the numbers of the boxes in panel A for the bands where proteins were extracted. Both gels were stained using GelCode blue.

DISCUSSION

The major findings of this study are that apo-Shp can acquire heme from human hemoglobin and that holo-Shp forms a complex with holo-HtsA and rapidly transfers its bound heme to apo-HtsA. The heme transfer was firmly demonstrated by the shift in the Soret absorption peak and by the isolation of holo-HtsA and apo-Shp from a holo-Shp/apo-HtsA mixture. Heme transfer from holo-Shp to apo-HtsA was rapid, and the equilibrium between holo-Shp/apo-HtsA and apo-Shp/holo-HtsA in the mixture favored apo-Shp/holo-HtsA, suggesting a predominant one-way heme transfer from holo-Shp to apo-HtsA. The efficiency of heme transfer from holo-Shp to apo-HtsA was further confirmed by the fact that apo-Shp could be prepared by incubating holo-Shp with apo-HtsA. Native PAGE, which has been widely used to detect the formation of protein complexes, demonstrated the formation of a holo-Shp/holo-HtsA complex. Our findings indicate that Shp can relay heme from hemoglobin to the HtsABC transporter. Gram-positive pathogens have thick cell walls, and some have additional capsule. Heme-specific ABC transporter is likely buried inside the cell wall (26). Heme is usually associated with host proteins and thus may not freely diffuse into the cell wall to reach the ABC transporter. Therefore, cell surface heme-binding proteins such as Shp are needed to relay heme from host heme proteins to ABC transporters in gram-positive bacteria.

Recombinant HtsA contained the His tag at the amino terminus. The amino terminus is where the lipid moiety is attached in mature lipoproteins and should be exposed on the protein surface. Therefore, it is unlikely that the His tag changed the structure of HtsA. SPy0252 is also a putative lipoprotein (15). The His tag of recombinant SPy0252, which was the same as in HtsA, did not confer the ability of SPy0252 to assimilate heme from holo-Shp, nor was there a shift in the Soret absorption peak. This suggested that the His tag of HtsA was not involved in heme transfer from holo-Shp to apo-HtsA.

Gram-positive bacteria produce cell surface heme-binding proteins (13, 19), as well as heme-specific ABC transporters (1, 5, 13, 14, 19). It has been proposed that the cell surface proteins relay heme across the bacterial envelope (19). It is not known how heme passage through the cell surface proteins is linked to membrane-crossing heme transport through ABC transporters. Efficient Shp-to-HtsA heme transfer and formation of the Shp/HtsA complex demonstrated in this work now provide a direct link between the heme relay and membrane-crossing heme transport processes. All GAS proteins involved in heme acquisition may form a complex for the ordered and efficient heme transport. It is also possible that heme transport processes across the outer and cytoplasmic membranes are directly linked together in gram-negative pathogens.

Recombinant Shp from Escherichia coli is present in the holo form (13), indicating that Shp efficiently obtained heme from E. coli heme proteins. Additionally, heme was still associated with Shp protein in saturated urea solution (13). These results suggest that Shp has a high affinity for heme. In contrast, most recombinant HtsA is in the apo form (14), suggesting that HtsA can not efficiently obtain heme from E. coli heme proteins. These observations suggest that Shp functions as a relay protein to the efficient assimilation of heme from other heme proteins and transfer of the captured heme to apo-HtsA. This idea is further supported by the fact that apo-Shp is able to acquire heme from human hemoglobin. Alternatively, another streptococcal protein may be needed to capture heme from host proteins and relay it to Shp. One possible candidate for this type of protein might be the protein recently identified as streptococcal hemoprotein receptor (Shr), which could bind host heme proteins (1). The shr gene is located immediately upstream of shp and is cotranscribed with shp and htsABC or siaABC (1). However, whether Shr binds heme and whether Shr can obtain heme from host heme proteins remain unknown.

The heme uptake machinery of S. aureus consists of iron-regulated surface determinants (Isd), which include four cell-wall-anchored heme proteins and a transporter (19). While HtsA shares 40% identity in amino acid sequence with IsdE, Shp and Shr have no homologues in S. aureus. The cell-wall-anchored heme proteins of S. aureus have no homologues in the available GAS genomes (2, 7, 24), suggesting that GAS may use cell surface heme-binding proteins that structurally differ from those of S. aureus. Interestingly, a BLAST search of available bacterial genomes found that Streptococcus equi has Shp and Shr homologues (http://www.sanger.ac.uk/Projects/S_equi/). Although the cell surface heme proteins in different gram-positive bacteria may be structurally unrelated, the mechanisms involved in heme acquisition are expected to be similar; that is, all proteins involved form a complex, and heme is transferred sequentially from one protein to the next.

In summary, we found that apo-Shp acquires heme from hemoglobin and that holo-Shp transfers its heme to apo-HtsA. These findings suggest a direct link between the process mediated by cell surface heme-binding proteins and the membrane-crossing heme transport through the ABC transporter in GAS.