Direct Hemin Transfer from IsdA to IsdC in the Iron-regulated Surface Determinant (Isd) Heme Acquisition System of Staphylococcus aureus*

The iron-regulated surface determinants (Isd) of Staphylococcus aureus, including surface proteins IsdA, IsdB, IsdC, and IsdH and ATP-binding cassette transporter IsdDEF, constitute the machinery for acquiring heme as a preferred iron source. Here we report hemin transfer from hemin-containing IsdA (holo-IsdA) to hemin-free IsdC (apo-IsdC). The reaction has an equilibrium constant of 10 ± 5 at 22 °C in favor of holo-IsdC formation. During the reaction, holo-IsdA binds to apo-IsdC and then transfers the cofactor to apo-IsdC with a rate constant of 54.3 ± 1.8 s–1 at 25 °C. The transfer rate is >70,000 times greater than the rate of simple hemin dissociation from holo-IsdA into solvent (ktransfer = 54.3 s–1 versus k–hemin = 0.00076 s–1). The standard free energy change, ΔG0, is –27 kJ/mol for the formation of the holo-IsdA-apo-IsdC complex. IsdC has a higher affinity for hemin than IsdA. These results indicate that the IsdA-to-IsdC hemin transfer is through the activated holo-IsdA-apo-IsdC complex and is driven by the higher affinity of apo-IsdC for the cofactor. These findings demonstrate for the first time in the Isd system that heme transfer is rapid, direct, and affinity-driven from IsdA to IsdC. These results also provide the first example of heme transfer from one surface protein to another surface protein in Gram-positive bacteria and, perhaps most importantly, indicate that the mechanism of activated heme transfer, which we previously demonstrated between the streptococcal proteins Shp and HtsA, may apply in general to all bacterial heme transport systems.

Bacterial pathogens have evolved acquisition machinery for heme as a preferred source of essential iron. In Gram-negative pathogens, specific outer membrane receptors (1, 2) sequester heme from heme-hemophore complexes or host hemoproteins and bring it into the periplasmic space in a TonB-dependent process (3). Specific ATP-binding cassette (ABC) 2 transporters then transport heme across the cytoplasmic membrane (4). A heme-specific ABC transporter is also a component of the heme acquisition machinery in Gram-positive pathogens such as Corynebacterium diphtheriae (5), Streptococcus pyogenes (6,7), Staphylococcus aureus (8), and Streptococcus equi (9). In addition, cell surface heme-binding proteins have been identified in S. pyogenes (10), S. aureus (11), S. equi (9), and Bacillus anthracis (12), suggesting that, besides ABC transporters, cell surface heme-binding proteins are required for heme acquisition by Gram-positive bacteria.
The S. aureus heme uptake system consists of the iron-regulated surface determinants (Isd), including the surface proteins IsdA, IsdB, IsdH or HarA, and IsdC, and ABC transporters Isd-DEF and HtsABC (8,11). The S. pyogenes heme uptake machinery consists of the surface proteins, Shr and Shp, and ABC transporter HtsABC (6,7,10,(13)(14)(15)(16). The surface protein components, Shr and Shp of the GAS acquisition machinery and IsdA, IsdB, IsdH, and IsdC of the S. aureus system, do not share significant sequence homology. Thus, the S. pyogenes and S. aureus systems represent two distinct heme acquisition pathways in Gram-positive bacteria. The S. pyogenes system has a homologue in S. equi (9), whereas the homologue of the S. aureus system is present in B. anthracis (12). These surface proteins, but not Shp, contain the NEAr transporter (NEAT) domains (17), although the structure of the Shp heme-binding domain (18) is similar to that of the NEAT domains of IsdA (19) and IsdC (20).
Previous characterization of the Isd system from S. aureus indicated that the system takes up hemin as an iron source (11). IsdB is a hemoglobin receptor and is required for uptake of hemin from methemoglobin (21). IsdH binds haptoglobin-hemoglobin, although it appears not to be critical for using methemoglobin hemin as an iron source in vitro (22). IsdA and the B. anthracis homologue of IsdC are also important for hemin uptake (12,19). IsdB, IsdA, and IsdC bind heme (11,19,20,23,24), and structural studies show that IsdA and IsdC bind hemin in a pentacoordinate complex with a tyrosine residue as the only axial ligand (19,20), in contrast to the hexacoordination of the heme iron in Shp and HtsA (14,15). It has been proposed that IsdH and IsdB capture haptoglobin-hemoglobin and hemoglobin, respectively, and heme is transferred from bound hemoglobin to IsdA, then IsdC, and finally to ABC transporters IsdDEF and/or HtsABC (25,26). This hypothesis has not been experimentally tested, and little is known about the biochemical mechanism of heme acquisition by this system. This study describes the rapid and direct hemin transfer from holo-IsdA to heme-free IsdC (apo-IsdC). This affinity-driven transfer occurs in two steps, the formation of a holo-IsdA-apo-IsdC complex and subsequent hemin transfer. This study not only unveils the first hemin transfer reaction of the Isd system but also documents the first example of hemin transfer from one surface protein to another surface protein in Gram-positive bacteria.
Protein Purification-Both IsdA and IsdC were expressed in Escherichia coli BL21 (DE3) containing pISDA and pISDC, respectively. Bacteria were grown at 37°C in 6 liters of Luria-Bertani broth supplemented with 100 mg/liter ampicillin to an absorbance at 600 nm of about 1.0, and the production of the proteins were induced by 0.4 mM isopropyl ␤-D-thiogalactopyranoside for 6 h. Bacteria were harvested by centrifugation. The bacterial pellet was suspended in 60 ml of 20 mM Tris-HCl, pH 8.0, and sonicated on ice for 15 min. IsdA and IsdC in the lysate were purified according to the manufacturer's protocol using a nickel-nitrilotriacetic acid-agarose column.
Preparation of Apo-IsdA and Apo-IsdC-Purified recombinant IsdC was a mixture of apo-form and a complex with ironfree protoporphyrin. To prepare apo-IsdC, the mixture in 20 mM Tris-HCl, pH 8.0, was loaded onto a DEAE-Sepharose column (1.5 ϫ 5 cm). Apo-IsdC was eluted with 30 mM NaCl in Tris-HCl, dialyzed against 3 liters of Tris-HCl. Apo-IsdA was prepared using the methyl ethyl ketone method (27).
Preparation of Holo-IsdA and Holo-IsdC-Holo-IsdC was obtained by reconstitution of its apo-form with hemin. One ml of apo-IsdC was incubated with excess hemin, loaded onto a Sephadex G-25 column (0.5 ϫ 30 cm), and eluted with 10 mM Tris-HCl, pH 8.0. The holoprotein without free hemin was collected. Purified IsdA was a mixture of apo-and holo-forms. Homogeneous holo-IsdA was similarly prepared by reconstituting apo-IsdA in the mixture with hemin.
Determination of Protein Concentration and Heme Content-Protein concentration was determined using a modified Lowry protein assay kit purchased from Pierce with bovine serum albumin as a standard. Heme content was measured using the pyridine hemochrome assay with the extinction coefficient ⑀ 418 ϭ 191.5 mM Ϫ1 cm Ϫ1 (28).
Rates of Hemin Association to and Dissociation from IsdA and IsdC-The rates of hemin dissociation from holo-IsdA and holo-IsdC were measured using H64Y/V68F whale sperm apomyoglobin as a hemin scavenger according to the previous method (29) with slight modification. Each protein (3 M) was incubated with 48 M apomyoglobin in 1 ml of 20 mM Tris-HCl, pH 8.0, and the changes in absorbance at 600 and 554 nm were monitored. Sucrose in the original method was not used in this study. The ⌬(A 600 -A 554 ) time courses were fit to a single exponential equation to obtain the first-order rate constants for hemin dissociation.
The rates of hemin binding to the apoproteins were measured using a stopped-flow spectrophotometer equipped with a photodiode array detector (SX20; Applied Photophysics). Hemin (2 M) in one syringe was mixed with apoprotein at Ն3 times [hemin] in another syringe. Entire spectra were recorded over time in each reaction. Changes in absorbance at the indicated wavelengths were analyzed as described under "Results." Hemin Transfer-Holo-IsdA and apo-IsdC or holo-IsdC and apo-IsdA at indicated concentrations were incubated in 0.1 ml of 20 mM Tris-HCl, pH 8.0, at room temperature (22°C) for 2 or 20 min. Each mixture was then loaded onto a small DEAE-Sepharose column (0.2 ml of resin), and the column was first eluted with 0.3 ml Tris-HCl to recover IsdA, which did not bind to the column, and then washed with 1 ml of Tris-HCl. IsdC bound to the column was eluted with 0.3 ml of 0.2 M NaCl in Tris-HCl. Separation of the two proteins was confirmed by SDS-PAGE analysis. The concentrations of holo-IsdA and holo-IsdC in the isolated samples were calculated from A 406 using the extinction coefficients of 9.5 ϫ 10 4 and 1.1 ϫ 10 5 M Ϫ1 ⅐cm Ϫ1 , respectively. The concentrations of apo-IsdA and apo-IsdC in the same samples were calculated from A 280 after subtracting the contribution from the holo-form using the extinction coefficients of 1.6 ϫ 10 4 and 1.8 ϫ 10 4 M Ϫ1 ⅐cm Ϫ1 , respectively.
Kinetics of Hemin Transfer from Holo-IsdA to Apo-IsdC-The spectral changes associated with hemin transfer from holo-IsdA to apo-IsdC were used to measure the transfer rate using the stopped-flow spectrophotometer at the indicated temperatures. Holo-IsdA (2.4 M) in one syringe was mixed with apo-IsdC at Ͼ5 times [holo-IsdA] in another syringe. Entire spectra were recorded over a range of 250 -800 nm against time in each reaction. Changes in absorbance at the appropriate wavelengths were analyzed.

RESULTS
IsdA and IsdC Proteins-Purified recombinant IsdA was a mixture of apo-and holo-forms. Homogeneous holo-IsdA was prepared by reconstitution of apo-IsdA with hemin, and 95% apo-IsdA was prepared by extraction with methyl ethyl ketone. Holo-IsdA exhibits absorption peaks at 406, 503, 529, and 622 nm in the visible region (Fig. 1A). Reduction of holo-IsdA with dithionite shifted the Soret peak from 406 to 430 nm and abolished the A 622 peak but did not result in the resolved ␣ and ␤ absorption bands seen in hemochrome (Fig. 1A). These absorp-tion features are consistent with the previous observation that the heme iron in IsdA is pentacoordinate (19,24).
Purified recombinant IsdC was also a mixture of apoprotein and a complex that displays absorption peaks at 407, 511, 547, 570, and 625 nm (data not shown) in the visible region. This feature is almost identical to that of the GST-IsdC fusion protein, which was found to bind iron-free protoporphyrin (23), indicating that the chromophore associated with IsdC was protoporphyrin. Apo-IsdC was readily separated from the IsdCprotoporphyrin complex by anion exchange chromatography. Holo-IsdC reconstituted from apo-IsdC and hemin displays the broad Soret peak at 402 nm and charge transfer bands at 501, 532, and 626 nm (Fig. 1B). Holo-IsdC treated with dithionite possesses the Soret peak at 417 nm and lacks the dominant ␣ and ␤ bands, and the reduction of holo-IsdC appears not to be complete (Fig. 1B). These spectral characteristics are consistent with the pentacoordination of the heme iron in the crystal structure of IsdC (20).
Relative Affinity of IsdA and IsdC for Hemin-To determine the affinity of IsdA for hemin, the rates of hemin binding and dissociation from IsdA were measured. When hemin was mixed with apo-IsdA, the spectrum of the reaction shifted from that of free hemin to those of holo-IsdA ( Fig. 2A). The time course of ⌬(A 406 -A 374 ) could be described by a two-exponential equation, resulting in two observed rate constants. The fast and slow phases contributed to approximate 70 and 30% of the total spectral change, respectively (Fig. 2C). The value of observed rate constant in the fast phase (k obs1 ) varies hyperbol-ically with [apo-IsdA], whereas the observed rate constant in the slow phase (k obs2 ) does not display a clear dependence on [apo-IsdA] (Fig. 2D). Using the method of de Villiers et al. (30), monomeric Fe(III)-protoporphyrin IX was ϳ60% of total hemin at 2 M in 20 mM Tris-HCl, pH 8.0. Thus, the fast phase is most likely the reaction of apo-IsdA with monomeric hemin. If this is true, the results of the fast-phase reaction suggest a two-step binding process involving the formation of a heminapoprotein intermediate that is followed by axial coordination to form the final holoprotein as proposed in Scheme 1, where k 1 and k 2 are the rate constants for bimolecular formation and unimolecular dissociation of the initial apoprotein-hemin complex, respectively, and k coordination and k Ϫhemin are the internal first-order rate constants for iron coordination to and dissociation from the final protein ligands, respectively.
If k 1 and k 2 are much greater than k coordination , and k Ϫhemin is much smaller than k coordination , the observed constant in the fast phase (k obs1 ) is described by Equation 1, where K d is k 2 /k 1 . The values for K d and k coordination from fitting the k obs1 data in Fig. 2D To determine the rate of dissociation of hemin from IsdA, IsdA was mixed with excess H64Y/V68F apomyoglobin, and dissociated hemin was captured by apomyoglobin. At 6 h after mixing holo-IsdA and apo-Mb, the mixture displayed the spectrum of H64Y/V68F holomyoglobin, as evidenced by the formation of the A 600 peak (Fig. 3A), indicating the loss of hemin from IsdA. ⌬(A 600 -A 544 ) time course associated with the hemin dissociation was fitted to a single exponential  Because the dissociation of hemin is extremely slow, k Ϫhemin in Scheme 1 must be Ͻ Ͻk 1 or k 2 . Therefore, the rate con-stants of IsdA obtained from the time course in Fig. 3 are directly equal to k Ϫhemin in Scheme 1.
The association equilibrium constants (K hemin ) for hemin binding to apo-IsdA can be estimated by the ratio of the apparent second-order association rate constant (kЈ hemin ϭ k coordination /K d ) and the hemin dissociation rate constant k Ϫhemin . The K hemin value for IsdA is 3.8 ϫ 10 11 M Ϫ1 (Table 1).
An attempt was also made to determine the affinity of IsdC for hemin. The spectral changes of the hemin/apo-IsdC mixture indicate that most of hemin already bound to IsdC when the first spectrum was recorded in a stopped-flow spectrophotometer (Fig. 2B). Because the reaction is too fast, the k coordination and K d of hemin association with apo-IsdC could not be reliably measured. Unlike the reaction of holo-IsdA and apomyoglobin, the 3.2 M holo-IsdC, 48 M apomyoglobin mixture at 12 h after mixing had a spectrum that was still more close to that of holo-IsdC than to that of holomyoglobin (Fig. 3B), and ⌬(A 600 -A 544 ) associated with the holo-IsdC/apomyoglobin reaction was 20% of the expected change (Fig. 3C). These results indicate that IsdC has higher affinity for hemin than both IsdA and H64Y/V68F myoglobin.
Hemin Transfer from IsdA to Apo-IsdC-We next tested whether IsdA transfers its hemin to apo-IsdC. Holo-IsdA (36 M) and 47 M apo-IsdC were incubated at room temperature for 2 min, and the two proteins were separated on a small DEAE column. SDS-PAGE analysis confirmed the separation of the two proteins (Fig. 4A). The normalized spectra of IsdA and IsdC before and after the reaction are shown in Fig. 4, B and C, respectively. The ratio of A 406 /A 280 of the treated IsdA was 29% that of the starting holo-IsdA. A 406 is the absorption by the bound hemin, and A 280 is mainly absorbance of the protein moiety. The lower A 406 /A 280 ratio of treated IsdA indicates that IsdA lost hemin in its reaction with apo-IsdC. Consistent with this result, holo-IsdC was present in treated IsdC based on the presence of the absorbance of bound hemin (Fig. 4C). Measurements of protein and hemin contents of these samples indicated that 90% of the holo-IsdA and 65% of the apo-IsdC lost and gained hemin, respectively. These results indicate that IsdA efficiently transfers its hemin to apo-IsdC.
Equilibrium Constant of the Holo-IsdA/Apo-IsdC Reaction-To demonstrate whether the hemin transfer reaction is reversible, 15 M holo-IsdC was incubated with 56 M apo-IsdA for 2 min, and the two proteins were separated as described above. Based on the spectra of the separated proteins (Fig. 5) and the extinction coefficients, 24% of holo-IsdC transferred hemin to apo-IsdA. Thus, the reaction is reversible. To estimate the equilibrium constant of this reversible reaction, 33 M holo-IsdA was incubated with apo-IsdC at 12, 24, 48, 71, or 95 M at room temperature (22°C) for 20 min, and the two proteins were separated. The concentrations of apo-and holo-forms of each isolated protein were calculated using the corresponding extinction coefficients, which were then used to determine the concentrations of apo-and holo-forms of each protein in the reaction mixture on the assumption that the separation did not  Heme Acquisition in S. aureus MARCH 14, 2008 • VOLUME 283 • NUMBER 11 shift the equilibrium. Fig. 6 represents the concentrations of holo-IsdA, apo-IsdA, and holo-IsdC as a function of initial [apo-IsdC]/[holo-IsdA]. Based on these data, the mean value Ϯ S.D. of the equilibrium constant of the reaction was found to be 10 Ϯ 5. The results indicate that the equilibrium is in favor of holo-IsdC formation, being consistent with the higher affinity of IsdC for hemin than IsdA.
Kinetics of Hemin Transfer from IsdA to Apo-IsdC-Although oxidized holo-IsdA and holo-IsdC have similar Soret absorption peaks, holo-IsdC does have higher absorbance than holo-IsdA in the region between 350 and 405 nm (Fig. 7A). Because holo-IsdA transfers its hemin to apo-IsdC, there should be increase in absorbance in the region of 350 -405 nm during the transfer, and this increase could be used to determine the kinetics of the hemin transfer reaction. Thus, limited holo-IsdA was reacted with apo-IsdC at various concentrations in a stopped-flow spectrophotometer, and spectra were recorded over time. The absorbance in the region of 350 -405 nm was indeed rapidly increased after mixing holo-IsdA and apo-IsdC (Fig. 7B). The difference in absorbance between 374 and 344 nm (A 374 -A 344 ) was used to determine the kinetics of the transfer. A 374 -A 344 increases rapidly and fits a single exponential equation (Fig. 7C), resulting in an observed first-order rate constant (k obs ). The k obs value depends hyperbolically on [apo-IsdC] (Fig. 7D).
An attempt was also made to kinetically characterize hemin transfer from holo-IsdC to apo-IsdA. However, the spectral change was too little to be monitored because the reverse reaction is not efficient. Thus, the kinetics of the reverse reaction could not be determined by the analysis used in the forward reaction. The data can be interpreted by a minimal model for hemin transfer from holo-IsdA to apo-IsdC given in Scheme 2. In this model, holo-IsdA forms a complex with apo-IsdC, and hemin is then reversibly and directly transferred to apo-IsdC to yield apo-IsdA:holo-IsdC, which is subsequently dissociated into apo-IsdA and holo-IsdC. The reversible reaction can be neglected under the conditions in Fig. 7. When the initial [apo-IsdC] is Ն5 [holo-IsdA] and k transfer presumably Ͻ Ͻ k 1 , k 2 , and k 3 , the hemin transfer from holo-IsdA to apo-IsdC is a pseudo first-order process. The observed rate constant, k obs , is given by Equation 2, where K d is the dissociation constant of the holo-IsdA-apo-IsdC complex, and k 1 , k 2 , and k transfer are the rate constants of the individual reactions proposed in Scheme 2. Fitting of the data in Fig. 7D    temperatures from 17 to 25°C were determined as described above. The k transfer could not be measured at Ն30°C because the reaction was too fast. The data were analyzed according to the Eyring equation (Equation 3).
where k B , h, and R are Boltzman's, Planck's, and the gas con-stants, respectively. The values of the activation entropy (⌬S ‡ ) and enthalpy (⌬H ‡ ) for the first-order transfer of hemin were 15 Ϯ 7 J/(K mol) and 67 Ϯ 12 kJ/mol, respectively. The free energy of activation, ⌬G ‡ calculated as ⌬H ‡ Ϫ T⌬S ‡ , was 63 kJ/mol at 25°C.

Enthalpy and Entropy Changes for Formation of the Holo-IsdA-Apo-IsdC Complex-
The values of the equilibrium dissociation constant for formation of the transient holo-IsdA-apo-IsdC complex at different temperatures were also determined as described above and analyzed by the van't Hoff equation (Equation 4).
where K association is the equilibrium association constant for formation of the holo-IsdA-apo-IsdC complex and equal to 1/K d from Table 2. The standard enthalpy (⌬H 0 ) and entropy (⌬S 0 ) changes were Ϫ68 Ϯ 11 kJ/mol and Ϫ136 Ϯ 39 J/(K mol) for the formation of holo-IsdA:apo-IsdC, respectively. The standard free energy change, ⌬G 0 , is Ϫ27 kJ/mol for the formation of the holo-IsdA-apo-IsdC complex.

DISCUSSION
The Isd proteins have been extensively characterized in terms of their function in hemin acquisition, hemin binding, and structures of the proteins or heme-binding domains. However, little is known about the biochemical mechanism of the heme acquisition by the Isd system. In this study, we demonstrate the rapid, direct, and affinity-driven hemin transfer from IsdA to IsdC, documenting the first proteinto-protein hemin transfer in the Isd system and providing the first example of hemin transfer from one surface protein to another surface protein in Gram-positive bacteria. In addition, the kinetic and thermodynamic analyses suggest that the hemin transfer is through an activated holo-IsdAapo-IsdC complex.
Apo-IsdA and holo-IsdC are formed during the brief incubation of holo-IsdA with apo-IsdC. The spectrum of the holo-IsdA/apo-IsdC mixture shifts from the spectrum of holo-IsdA to that of holo-IsdC. Thus, holo-IsdA and apo-IsdC lose and gain hemin, respectively, during the incubation of holo-IsdA with apo-IsdC, and the spectral change can be used to follow the hemin transfer.
Kinetic analyses of the transfer and hemin dissociation from holo-IsdA strongly indicate that holo-IsdA directly transfers its hemin to apo-IsdC. In indirect transfer, hemin is first dissociated into solvent from the donor and then captured by the acceptor. If the IsdA-to-IsdC hemin transfer were indirect, the rate of the transfer should be close to the rate of the hemin dissociation from IsdA. However, the rate constant of the transfer is Ͼ70,000 times greater than the rate of simple hemin dissociation from holo-IsdA into solvent (k transfer ϭ 54.3 s Ϫ1 versus k Ϫhemin ϭ 0.00076 s Ϫ1 ), ruling out an indirect transfer mechanism. Furthermore, the values of observed k transfer depend hyperbolically on [apo-IsdC]. This result indicates that holo-IsdA forms a complex with apo-IsdC prior to hemin transfer, further supporting a direct transfer mechanism.  Because hemin aggregates in aqueous solution, it has been a challenge to estimate the affinity of hemoproteins for hemin using kinetic analysis of hemin association to apoproteins and dissociation from holoproteins. The secondary order rate constants of hemin association have been estimated as the rate constants of association of CO-heme complex with apoproteins (32). The accuracy of this estimation is unknown. Thus, we performed the kinetic analysis of hemin association with apo-IsdA, which fits a two-exponential equation with ϳ70% of spectral change attributable to the fast phase. de Villiers et al. (30) recently reported that aqueous hemin solution contains only monomers and dimers. Using the method of de Villiers et al. (30), percentile of monomeric hemin in the hemin solution before mixing with apo-IsdA in our stopped-flow measurements was estimated to be ϳ60%. Our data could be interpreted as follows. The fast phase is the reaction of monomeric hemin with apo-IsdA, whereas the slower phase is the dissociation of dimeric hemin followed by the rapid association of the resulting monomeric hemin with apo-IsdA. Thus, we could use the fast phase of the reaction to obtain the secondary rate constant of the hemin association to apo-IsdA.
Although we can estimate the affinity of IsdA for hemin, the affinity of IsdC for hemin could not be determined under the same conditions, because the values of k coordination in hemin association and k Ϫhemin in hemin dissociation could not be determined. However, our data strongly indicate that IsdC has a much higher affinity for hemin than IsdA. First, hemin association to apo-IsdC is faster than that to apo-IsdA. Second, the majority of hemin in holo-IsdA, but not holo-IsdC, was lost to the apomyoglobin under the same conditions. Third, the hemin transfer from holo-IsdA to apo-IsdC is reversible, and the equilibrium constant of the reaction is about 10 in favor of holo-IsdC formation. If the interaction between IsdA and IsdC does not shift the equilibrium, IsdC would have an affinity for hemin 10-fold as that of IsdA. The Trp-77 residue unique to IsdC "interlocks" hemin into its pocket in its crystal structure, and this effect has been proposed to account for assumed higher hemin affinity of IsdC versus that of all other Isd proteins containing the NEAT domain(s) (20). Our data are consistent with this hypothesis but also suggest that the faster rate of hemin association to apo-IsdC also contributes to the higher affinity of IsdC for hemin. Taken together, our data indicate that efficient hemin transfer from IsdA to IsdC is at least partially driven by the higher affinity of IsdC for hemin compared with that of IsdA.
It has previously been hypothesized that hemin is transferred in the following order: IsdH-haptoglobin-hemoglobin/IsdBhemoglobin 3 IsdA 3 IsdC 3 IsdDEF and/or HtsABC (25,26). Rapid and direct hemin transfer from IsdA to IsdC provides the first piece of experimental evidence supporting this hypothesis. It will be necessary to examine hemin transfer reactions and kinetic mechanisms for the various protein couples of these   proteins to fully establish a heme acquisition model in the Isd system.
The Shr/Shp/HtsABC and Isd systems are two distinct heme acquisition pathways composed of surface proteins and ABC transporter in Gram-positive pathogens. Shr likely binds host hemoproteins (7) and has two NEAT domains (17). The protein binds hemin and transfers it to apo-Shp but not to apo-HtsA (16). Although Shp was not included in the NEAT family by Andrade et al. (17), the homology in amino acid sequence between the heme-binding domain of Shp and the heme-binding or NEAT domain of IsdA or IsdC (17-19% identity) is similar to that between the IsdA and IsdC NEAT domains (19% identity) (18), suggesting that Shp is at least a distant member of the NEAT family that includes Shr, IsdH, IsdA, IsdB, and IsdC (17). We have shown that Shp rapidly and directly transfers its hemin to apo-HtsA. These observations suggest a heme flow model of hemoglobin 3 Shr 3 Shp 3 HtsA in the S. pyogenes system. It appears that there are parallel functions of the components in the two systems, i.e. Shr functions like IsdH/B, Shp like IsdA/C, and HtsA like IsdE and/or S. aureus HtsA.
The affinities of IsdA and IsdC for hemin are greater than those of Shp and HtsA (Table 3). IsdA-hemin and IsdC-hemin complexes are pentacoordinate and use Tyr as their only axial ligand (19,20), whereas Shp-and HtsA-hemin complexes are hexacoordinate using two Met and Met/His residues as the axial ligands, respectively (15,18). The Tyr-hemin ligation may be a major factor for the higher hemin affinities of IsdA and IsdC. The axial Met residues of the Shp heme iron are both required for rapid Shp-to-HtsA hemin transfer (15), and the hemin acceptor in each couple has a higher affinity for hemin, suggesting that the axial ligands and relative hemin affinity may have been evolved for efficient hemin transfer in each system.
IsdA apparently transfers its hemin to apo-IsdC through an activated transfer mechanism. Holo-IsdA forms a complex with apo-IsdC, releasing free energy. The free energy released is used to weaken hemin binding in holo-IsdA and thus facilitate transfer to apo-IsdC. This mechanism is similar to that of hemin and heme transfer from streptococcal Shp to apo-HtsA (14). Protein interaction also is critical for hemin transfer from Serratia marcescens hemophore HasA to HasA receptor HasR (1). The mechanism of activated heme transfer may apply in general to all direct heme transfers in bacterial heme transport systems. a Data were estimated from the value of K hemin of IsdA and the equilibrium constant (K eq ϭ 10) of the reaction of holoIsdA with apoIsdC on the assumption that the K eq value is solely dependent on the relative affinity of IsdA and IsdC for hemin.