Interaction of Fe-Protoporphyrin IX and Heme Analogues with Purified Recombinant Heme Oxygenase-2, the Constitutive Isozyme of the Brain and Testes*

Heme oxygenase-2 (HO-2) is the predominant form of heme oxygenase in the brain and testes. The enzyme is not readily amenable to isolation from mammalian tis-sues and has not been characterized for its kinetic prop- erties and interaction with metalloporphyrins. isolation of HO-2 apparent molecular at near unity to a complex with the absorption maxi- mum at 403 The has a blue shift to 430 is reduced, with distinct a and p bands at 485 and 550 nm, respectively. The Soret band of the CO complex of ferrous heme-HO-2


Interaction of Fe-Protoporphyrin IX and Heme Analogues with
Purified Recombinant Heme Oxygenase-2, the Constitutive Isozyme of the Brain and Testes* (Received for publication, June 20, 1994, and in revised form, August 9, 1994) Inna Rublevskaya and Mahin D. Maine& Heme oxygenase-2 (HO-2) is the predominant form of heme oxygenase in the brain and testes. The enzyme is not readily amenable to isolation from mammalian tissues and has not been characterized for its kinetic properties and interaction with metalloporphyrins. Pres- A multistep protocol developed for isolation of HO-2 resulted in a homogeneous protein with a specific activity up to 6,500 nmol of bilirubidmgh. Based on SDS-polyacrylamide gel electrophoresis and Western blot analyses, the protein had an apparent molecular mass of -34 kDa. HO-2 binds Fe-protoporphyrin (heme) at near molar unity to give a complex with the absorption maximum at 403 nm. The Soret band has a blue shift to 430 nm when heme iron is reduced, with distinct a and p bands at 485 and 550 nm, respectively. The Soret band of the CO complex of ferrous heme-HO-2 is at 420 nm, and a and fi bands are at 540 and 572 nm, respectively. The apparent K,,, for Fe-protoporphyrin is 0.33 p~, with a V,,, of 0.45 mmol of bilirubidmgh. Zn-protoporphyrin is a strong mixed inhibitor of enzyme activity, whereas Co-protoporphyrin is a poor competitive inhibitor of activity. When HO-2 was preincubated (10 min at 4 "C) with Feprotoporphyrin, the cobalt complex did not inhibit enzyme activity, whereas the Zn-protoporphyrin effectively inhibited activity. Calorimetric measurements suggest that HO-Zheme interaction involves one type of association producing a single heat absorption peak upon melting of the complex and that the unfolding is not reversible. The association increases the enthalpy of HO-2 (130 kcaYmol versus 184 kcaYmo1) and increases the stability to heat denaturation by 9 "C. Heat duration of zinc complex involves at least two stages of unfolding.
The isomer-specific cleavage of the heme molecule (Fe-protoporphyrin IX, heme b, hemin, protoheme) at the o-mesobridge and formation of biliverdin, CO, and iron is catalyzed by the microsomal heme oxygenase system (Ref. 1 and reviewed in Ref. 2). The system requires the concerted activity of heme * This study was supported by National Institutes of Health Grants ES03968, ES01247, and ES04066 and Burroughs Wellcome Fund. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "uduertisernent" in accordance with 18  oxygenase isozymes and NADPH-cytochrome P-450 reductase; the reductase permits the reducing potential of NADPH to be utilized for activation of molecular oxygen. Two distinct forms of heme oxygenase, referred to as HO-1' and HO-2 (3,4), have been identified in several mammalian species, including human (5-111, and an avian form of heme oxygenase also has been described (12). The mammalian heme oxygenases are encoded by different genes; when rat HO-1 and HO-2 gene are compared, little resemblance in structure, regulation, or tissue distribution is observed (13)(14)(15)). The encoded HO-1 and HO-2 have in common a conserved sequence of 23 amino acids, which is suspected to be the heme-binding domain (9,13,16). The conserved domain is encoded by the only exon in rat HO-1 and HO-2 genes that shares some degree of similarity (-50%) (14,15). At the regulation level, HO-1, which was the first metalinducible enzyme characterized (171, is now known as hsp32 (18)(19)(20)(21) and is induced by a host of chemical stimuli and pathological conditions (reviewed in Ref. 21, whereas HO-2 is refractory to all agents that induce HO-1 (3, 4); consistent with this is the absence of any identifiable regulatory elements except for a glucocorticoid response element in the promotor region of the HO-2 gene (15). In brain the glucocorticoid response element is responsive to adrenal steroids (22).
At the protein level, unlike HO-1, HO-2 is not resistant to inactivation by heat and free radicals (3,4,23). This most likely reflects both the primary composition of the enzymes (HO-1 is devoid of sulfhydryl residues, whereas KO-2 has 3 cysteines (13)) as well as higher order differences in structure/ conformation of the proteins. Moreover, unlike HO-1, which can be purified with relative ease from rat organs (3,4), it has been difficult to purify HO-2; hence, little is known about its biochemical and kinetic properties. In fact, to our knowledge, only our laboratory, using purification conditions quite different from those that are of utility for HO-1 isolation, has succeeded in obtaining a homogeneous preparation of HO-2, albeit at an exceedingly low yield, from rat testes (4). HO-2 is a membranebound protein, and apparently it is anchored to the endoplasmic reticulum by a hydrophobic domain at the carboxyl terminus of the protein (8). This domain is not essential for heme catalysis (16) and also contributes to dificulty in purification efforts.
HO-1 and HO-2 display differential tissue distribution, with brain and testes being the predominant HO-2-expressing organs (4, 19, 24) and the spleen being the HO-1 expressing tissue (25). The curious finding that heme-degrading activity of the brain nearly matches that of the spleen (211, which is the main site of hemoglobin heme degradation, has led to the recent identification of CO as a potential signaling molecule for cGMP production (26)(27)(28)(29)(30). The heme analogue Zn-protoporphy-' The abbreviations used are: HO-1, -2, heme oxygenase-1, -2; PAGE, polyacrylamide gel electrophoresis. rin, an inhibitor of heme oxygenase activity (31), has been used to establish this concept; also various synthetic heme analogues have been used to suppress i n vivo hemoglobin heme degradation (31-33).
Presently we have used a cDNA construct in which the hydrophobic carboxyl-terminal amino acids are replaced by neutral residues for high level expression of HO-2; the nearly fulllength expressed recombinant protein is active and provides, for the first time, information on the kinetic properties of the homogenous enzyme and its interaction with heme and synthetic heme analogues.

EXPERIMENTAL PROCEDURES
Materials-Oligonucleotides for sequencing and mutagenesis were obtained from Midland Certified Reagent Co. (Midland, TX). Nitrocellulose for Western blot analysis was from Schleicher and Schuell (Keene, NH). Sequenase version 2.0, restriction enzymes, and other DNA modification enzymes were purchased from U. S. Biochemical Corp. NADPH, sodium cholate, Triton X-100, sodium dodecyl sulfate, and colloidal Coomassie Brilliant Blue G were purchased from Sigma. Fe-, Zn-, and Co-protoporphyrins were purchased from Porphyrin Products (Logan, UT). Reagents for protein determination were obtained from Bio-Rad. Rainbow-colored protein molecular weight markers were purchased from Amersham Corp. All other reagents were of the highest quality commercially available. Adult male Sprague-Dawley rats and New Zealand White rabbits were obtained from Harlan Industries (Madison, WI).
Purification of Enzymes and Production of Antibody to HO-2-Rat testes HO-2 was purified as described before (4). Biliverdin reductase was purified using the modifications (34) of an earlier method (35). NADPH-cytochrome P-450 reductase was purified as described by Yasukochi and Masters (36). Polyclonal antiserum to rat testes HO-2 was prepared in New Zealand White rabbits as previously described (13).
Mutagenesis of Rat HO-2-The HO-2 insert (Sal1 to EcoRI) of the expression clone IT-1 (previously designated pUC 18B#43; Ref. 8) was ligated into vector PBS+ (Stratagene, La Jolla, CA) to generate the plasmid pRHO/RS (16). This plasmid was used to generate the carboxyl terminus mutant HO-2 clone, pRHOP. The presence of a unique PstI site near the end of the coding region allowed the exchange of the hydrophobic carboxyl terminus with predominantly hydrophilic residues, and the resultant plasmid was transformed into Escherichia coli XL-1 Blue. The sequence of the carboxyl terminus of this construct, pRHOP, was confirmed by sequencing using the method of Chen and Seeburg (37). pRHOP encodes a protein in which 19 amino acids of the COOH terminus with hydrophobic character (297-315, Leu-Ile-Leu-Ala-Ala-Ser-Val-Ala-Leu-Val-Ala-Gly-Leu-Leu-Ala-~p-~r-~r-Met) (8) are replaced by 15 amino acids (Val-Asp-Ser-Arg-Gly-Ser-Pro-Gly-Thr-Glu-Leu-Glu-Phe-Pro-Leu) of neutral character (16).
Expression of Carboxyl Terminus-mutated HO-2 in E. coli a n d Cell Extract Preparation-E. coli XL-1 Blue (recAl, lac-(F'proAB, lac ZqZAMl5,TnlO(tetR))) carrying HO-2 plasmid pRHOP was grown to saturation overnight at 37 "C in 2 x YT medium (1.6% tryptone, 1% yeast extract, 0.5% NaCI) containing 50 pg/ml ampicillin. Cultures were diluted 1:IOO in the same medium and grown to a n A,, of 0.1. Isopropyl-1-thio-@-D-galactopyranoside was added to a final concentration of 3 mM, and induction was allowed to proceed until the cultures attained an A,,, of -1. Cell extract was obtained by modifying the procedure described by Scopes (38). Disruption of the bacterial cells was obtained by three freeze/thaw/sonication cycles. Lysates were digested (25 "C, 15 min) with deoxyribonuclease (10 pg/ml) in the presence of MgCI, prior to centrifugation (20,000 x g, 20 min).
Purification of Expressed HO-2-The following protocol was developed for the purification of recombinant HO-2 from E. coli because neither the protocol we have developed for purification of HO-2 from the rat testes (4) nor any other protocol used for the purification of HO-1(3, 39) yielded a suficiently purified preparation. All operations were performed a t 4 "C. With exceptions as noted, all buffers contained 0.4% (v/v) Triton X-100, 0.1% (w/v) sodium cholate, 10% (v/v) glycerol, and 0.1 mM EDTA. The supernatant after centrifugation of the cell extract was diluted with 3 volumes of 20 mM Tris-HC1 buffer, pH 7.0 (buffer A) and loaded onto a DEAE---Gel blue column equilibrated with buffer A, and unabsorbed proteins were eluted with about 3 column volumes of the same buffer. Elution of HO-2 from the column was performed with a linear gradient of KC1 (0-0.5 M) prepared in buffer A. Those fractions that contained heme oxygenase activity were collected, concentrated, and loaded onto a Sephacryl 5-200 column equilibrated with 20 mM Tris-HC1 buffer, pH 8.0, containing 80 mM KC]. HO-2 was eluted with the same buffer. The catalytically active fractions were pooled and applied onto a DEAE-Sepharose FF column equilibrated with 20 mM Tris-HCI buffer, pH 8.0. Thereafter, the column was washed with the same buffer containing 100 mM KCl. HO-2 was subsequently eluted from the column with a linear gradient of KC1 (100-250 mM) that was prepared with the wash buffer. Fractions exhibiting heme oxygenase activity were pooled; diluted with 1 volume of 20 mM potassium phosphate buffer, pH 6.8; adjusted to the same pH by the addition of I M potassium phosphate buffer, pH 6.8; and loaded onto a hydroxylapatite column equilibrated with the same buffer. The column was washed with 20 m M potassium phosphate buffer, pH 6.8, containing 0.2% (v/v) Triton X-100, 20% (v/v) glycerol, 0.1 mM EDTA and was eluted with a linear gradient of potassium phosphate (20-200 mM) prepared with the wash buffer. The preparation of HO-2 was homogeneous when subjected to SDSpolyacrylamide gel electrophoresis and had a specific activity of up to 6,500 units/mg of protein. One unit of activity was defined as the amount of the enzyme that catalyzed the formation of 1 nmol of bilirubinh.
Polyacrylamide Gel Electrophoresis of HO-2-SDS-polyacrylamide gel electrophoresis (PAGE) was performed according to the procedure described by Laemmli (40). The separating gels contained 10.0% T and 2.67% C, and the stacking gels contained 3.0% T and 257% C. The gels were stained with colloidal Coomassie Brilliant Blue G.
Western Immunoblotting-Protein samples were subjected to SDS-PAGE on a 1.5-mm slab. The separated proteins were electrophoretically transferred from the gel to a nitrocellulose membrane according to the procedure of Towbin et al. (41). Primary and secondary antibody treatment followed by peroxidase staining with 4-chloro-I-naphthol was performed by previously described methods (4).
Assay Procedures a n d Kinetic Analysis-Heme oxygenase activity was assessed following conversion of biliverdin to bilirubin by addition of purified biliverdin reductase to a reconstituted HO-2 assay system. The standard reaction mixture (1 ml) contained varying concentrations of Fe-protoporphyrin (hemin) as denoted in appropriate experiments; excess purified biliverdin reductase (0.75 unit), each unit being the amount of protein that catalyzes formation of 1 nmol of bilirubin; and NADPH-cytochrome P-450 reductase (0.5 unit) in 0.1 M potassium phosphate buffer, pH 7.4, and varying amounts of expressed enzyme a s indicated for each experiment. The reaction was initiated by addition of 0.1 volume of 2.75 mM NADPH. Both the blank and the test reaction mixtures were incubated in the dark for 12 min at 37 "C with constant shaking. The difference absorption spectra between 470 and 530 nm (42) were obtained using a n Aminco DW 2C spectrophotometer, and the amount of bilirubin was calculated from an extinction coefficient of 40 m"' cm" (17). Formation of 1 nmol of bilirubinh was defined as 1 unit of activity. The protein concentration was determined by the method of Lowry et al. (42). The procedure of Dulley and Grieve (43) was used when detergents were present in the sample. The kinetics of inhibition of Fe-protoporphyrin oxidation were examined following addition of 0.25 PM Co-protoporphyrin or 0.25 PM Zn-protoporphyrin to the standard assay using different Fe-protoporphyrin concentrations indicated in Fig. 5. Linear regression analysis was used for obtaining the best fit line.
Spectral Analysis and Calorimetric Measurements-Spectra1 analysis with Fe-protoporphyrin was carried out using a n Aminco DW2C spectrophotometer. Experimental details are provided in legends to appropriate figures. A Microcal MC-2 differential scanning calorimeter was used for calorimetric measurements. A protein concentration of 25 pf in a 25 mM phosphate buffer, pH 7.4, was used. The scan rate was 60 "Ch. All curves were corrected for the instrumental baseline obtained by filling both cells with the buffer used. The reversibility of unfolded protein was examined by scanning samples that had been cooled prior to completion of scanning. Integration procedures were carried out after normalization of the data for protein concentration and subtraction of the chemical base line, i.e. the straight line connecting the initial and final temperatures of the overall transition. Calculation of the denaturational enthalpy was performed as described by Privalov and Potekhin (44). Kinetic analyses were performed using best fit regression analysis. testes (4) was employed but did not yield a sufficient amount of homogenous protein. Thus, information we had gathered in a previous study (aimed at delineating domains of HO-21, which had shown that the carboxyl terminus of the protein does not contribute to its heme degrading activity and serves only as a membrane anchor (16), was used to advantage. The construct, pRHOP, has the sequence and hydrophobicity profile at its COOH terminus shown in Fig. 1.

Purification
This construct was used to express HO-2 in E. coli XL-1 Blue. To obtain a high yield pure preparation of expressed HO-2 we also tried a number of different separation media and permutations of methods described for purification of microsomal HO-1(3) or HO-2 (4) without success. The protocol outlined in Table I was designed and found to meet the criteria of a high yield homogenous preparation. Table I shows a typical purification from 3 liters of cells. As noted at the final step, 15% of the activity present in the cell lysate was recovered. The specific activity of the protein, which typically was about 6,500 nmol of bilirubidmg of proteinh, was nearly twice that we have obtained from rat testes (4). The highly laborious and time-consuming process of tissue processing likely accounts for a certain degree of inactivation of HO-2 in the course of purification (4). After we had established a working scheme for purification of HO-2, Wilks and Ortiz de Montellano (45) and Takahashi et al. (46) reported on purification of a truncated form of HO-1 that lacks the 26-amino acid COOH-terminal segment of the protein. We investigated whether HO-2 could be purified using those protocols and found them not applicable.
The expressed HO-2, which eluted as a rather sharp peak from hydroxylapatite, was analyzed for purity and immunoreactivity. The SDS-polyacrylamide gel electrophoresis analysis of fractions forming the peak is shown in Fig. 2 A . As noted, even when large amounts of protein were loaded, no trace of contaminating proteins could be detected in peak fractions (8 lanes). Based on Western blotting using rat testes HO-2 antiserum, as shown in Fig. 2B, the recombinant HO-2 shares similar antigenic properties with the rat testes HO-2. The apparent molecular mass of HO-2 purified from rat testes is -36 kDa (4), and that of the recombinant HO-2 is -34 kDa. The recombinant HO-2 is 4 amino acids shorter than native HO-2; this difference plus the potential for processing of the protein in mammalian tissue could account for the difference in apparent molecular weight. Sequence analysis did not indicate any changes in the NH, terminus of the encoded protein.
Metalloporphyrin Binding Properties of Recombinant HO-2-Spectrophotometric titration of HO-2 with Fe-protoporphyrin (Fig. 3) was carried by stepwise addition of 10 pl aliquots of TAFILE I Purification of recombinant rat HO-2 A representative purification scheme using 3-liter cell growth is "Experimental Procedures" by measuring formation of bilirubin from shown. Heme oxygenase activity was measured as described under the change in absorption between 470 and 530 nm using an extinction coefficient of 40 mM" cm" (17). One unit of activity is defined as the amount of the enzyme that catalyzes formation of 1 nmol of bilirubinh.   (Table I) displaying heme oxygenase activity were subjected to electrophoresis along with molecular weight standards. The proteins were detected by colloidal Coomassie Blue G staining. Molecular mass markers were bovine serum albumin (66 kDa), ovalbumin (45 kDa), carbonic anhydrase (29 kDa), and trypsinogen (24 m a ) . B, Westem immunoblot analysis of expressed HO-2. The purified preparation of expressed HO-2 was subjected to SDS-PAGE and, thereafter, electroblotted onto a nitrocellulose sheet and subjected to Western blot analysis. Purified rat testes HO-2 and rainbow molecular weight markers were used as standards. The blot was probed with rabbit antibody to rat HO-2. Lanes 1 and 6, purified rat testis HO-2; lanes 2 5 , increasing amounts of purified recombinant HO-2. a 9.05 p~ hemin solution to HO-2 in phosphate buffer (pH 7.4). The absorption maximum at 403 nm and the increase in absorption were recorded. As shown in Fig. 3, the increase in absorbance was linear with the increase in molecular ratio of Fe-protoporphyrin to HO-2 up to a ratio of 1.l:l; thereafter, a plateau level was reached. This finding suggests that Fe-pro-toporphyrifl0-2 interaction follows a 1:l stoichiometric ratio.
Subsequently, we examined the spectral characteristics of recombinant HO-2; results are shown in Fig. 4 inhibition of Fe-protoporphyrin oxidation by these complexes under different assay conditions described below.
A standard reaction mixture containing HO-2 was preincubated for 10 min at 0 "C with a constant amount of Fe-protoporphyrin, final concentration 2 p~, and thereafter varying amounts of Coor Zn-protoporphyrin ranging from 0 to 2.5 p~ were added. As shown in Fig. 6, under these conditions the addition of Co-protoporphyrin (0) to the assay system in 50% excess of Fe-protoporphyrin did not inhibit bilirubin formation. In contrast, Zn-protoporphyrin (A) had a strong inhibitory effect; at a 1:l molar ratio of Zn-protoporphyrin:Fe-protoporphyrin a 50% inhibition of bilirubin formation was noted.
Calorimetric Analysis-Interaction of recombinant HO-2 with Fe-protoporphyrin was further examined by measuring thermodynamic stability of the enzyme. As shown in Fig. 7  of Fe-protoporphyrin.HO-2 complex was not reversible. Calorimetric analysis of Zn-protoporphyrin.HO-2 is shown in the bottom panel. The pattern of the heat absorption peak upon melting of the complex was significantly different from that of the Fe-protoporphyrin complex of HO-2; a broad transition state followed by a sharp transition peak at 61.5 2 0.5 "C were obtained. As with the Fe-protoporphyrin.HO-2 complex, the unfolding of the zinc complex was not reversible.

DISCUSSION
Oxidation of heme in biological systems is catalyzed by two distinct proteins, HO-1 and HO-2, distributed differentially in mammalian organs (reviewed in Ref. 2). Brain and testes are the major HO-2-expressing organs, and HO-1 is not readily detected in the brain under normal conditions (20, 26, 47). HO-2 has little similarity in its primary structure and regulation with HO-1, nor do the conditions required for the purification of the proteins from rat organs bear any similarity (3,4). It follows that the ease of purification of HO-1 has been a reason why its characterization and identification as a distinct microsomal entity (17,48,49) preceded that of HO-2 by more than a decade. Accordingly, although a substantial amount of information has been gathered about HO-1, both its native (3) and recently its truncated forms (45,46) little is known about the kinetic properties of HO-2 and the kinetics of its reaction with metalloporphyrins.
The current structural model of HO-2, which is based on deletion and mutation analyses of recombinant HO-2, defines two domains essential for activity (16). One domain is conserved in all forms of heme oxygenase characterized to date with only one conservative substitution, which distinguishes the HO-1 and HO-2 isozymes; in all mammalian HO-1 species and the avian form the methionine present in HO-2 is replaced by leucine (9). This domain is made up of 23 amino acids and lies amid a conserved hydrophobic region containing a histidine residue (HidS1 in rat HO-2); it is postulated to be the enzyme interaction site with the heme molecule (21). The second domain is the amino terminus of HO-2; HO-2 has a stretch of 20 amino acids in its amino terminus that are not present in HO-1 and account for the larger size of HO-2. The deletion of this stretch results in loss of activity (16). This domain could be involved in correct folding and maintaining of an active configuration of the enzyme, and/or in interaction with heme or NADPH-cytochrome P-450 reductase. An additional hydrophobic domain, at the carboxyl terminus is dispensable for hemedegrading activity (16), and its function appears to be solely to anchor the protein to the endoplasmic reticulum. The model of the HO-2 was extended by the present spectral, kinetic, and thermodynamic studies using a recombinant protein in which the hydrophobic domain of the carboxyl terminus was converted to a neutral one.
The recombinant HO-2.heme complex-binding spectrum resembles that described for mammalian tissue HO-1 (39) and the truncated expressed HO-1 (45). The Soret bands for HO-2.heme in ferric, ferrous, and ferrous CO heme are at 403,430, and 420 nm for HO-2 uersus 405,429, and 421 nm for HO-1, respectively. For HO-2 the observed heme binding molar ratio is close to unity (l.l:l), whereas for HO-1 the observed ratio is 0.7 (39) (it is extrapolated to unity). The kinetic properties of HO-1 and HO-2, with heme as the substrate, closely resemble one another; the apparent I C , of HO-1 is 0.24 (41, and that of HO-2 is 0.33. At this time a comparison of calorimetric properties of recombinant HO-2 and HO-1 cannot be made because such data are not available for HO-1. Previous studies with HO-1 purified from spleen (50) or spleen microsomes (31, 51,52) have shown that the enzyme does not have an absolute specificity toward Fe-protoporphyrin and that various metalloporphyrins, including Zn-and Co- protoporphyrin, competitively inhibit Fe-protoporphyrin oxidation. The kinetic analyses conducted in the present study show that, as with HO-1, Co-protoporphyrin competitively inhibits the oxidation of Fe-protoporphyrin by HO-2. This finding is consistent with the role of the conserved domain of 23 amino acids in interaction with metalloporphyrins in both HO-1 and HO-2 species and serving as a heme-binding domain. Znprotoporphyrin, on the other hand, inhibits degradation of heme in a mixed manner. This finding suggests the possible differences in interaction of the two HO enzymes with Znprotoporphyrin. The finding, in turn, reflects dissimilarities in the primary composition and mass and likely differences in molecular organization and structure of the two proteins. Alternatively, it is possible that previous studies with HO-1 require reexamination.
Furthermore, in contrast to previous assumptions, the results obtained in the present study further suggest that the binding affinity of metalloporphyrins to HO-2 is greatly influenced by its central chelated metal. We postulate that the reactivity of the chelated metal with histidine residues, particularly histidine 151, the residue present in the putative heme binding domain, may have a major influence on binding affln-ity. This suggestion is supported by the well known reactivity of heme iron with the histidine imidazole group, e.g. in hemoglobin chains and myoglobin. The suggestion is also consistent with the previous observation (50, 51) that Go-protoporphyrin is a poor substrate for heme oxygenase. Furthermore, the observation that the kinetics of heme oxygenase activity are influenced by the conditions under which the enzyme is exposed to the metalloporphyrins suggest the influence of the central chelated metal of metalloporphyrins on interaction with HO-2. For instance, Co-protoporphyrins displayed inhibitory effects when added simultaneously with increasing concentrations of Fe-protoporphyrin to the HO-2 assay system (Fig. 5); however, when Fe-protoporphyrin (at a high concentration) was preincubated with HO-2 and an extended time (10 min at 4 "C) for interaction with HO-2 was allowed (presumably allowing for Fe-ligand interaction to occur), Co-protoporphyrin was not inhibitory to Fe-protoporphyrin oxidation (Fig. 6). The fact that Zn-protoporphyrin was inhibitory to heme oxidation under all conditions suggests that zinc complex interaction with the heme binding domain of the protein is stronger than that of CO-protoporphyrin. This observation also raises the possibility that Zn-protoporphyrin interacts with other sites on HO-2. This possibility is consistent with the observed mixed type of inhibition of enzyme activity by Zn-protoporphyrin (Fig. 5) as well as with calorimetric data discussed later.
Heat denaturation of apo-HO-2 is accompanied by a heat absorption peak at 49.5 "C. Binding of Fe-protoporphyrin to the protein leads to changes in the temperature of melting profile and the temperature of denaturational transition. The complex has an increased stability (by 9 "C) to heat denaturation and a change in enthalpy of 54 to 184 kcal/mol. Based on calorimetric measurements, it is apparent that interaction of the Fe-protoporphyrin with HO-2 involves one type of association profile as suggested by the presence of a single heat absorption peak. It would be reasonable to suggest that the association involves interaction of the metalloporphyrin with the heme binding site.
The increase in enthalpy and in T,,, indicate that HO-2 is stabilized to an appreciable extent by Fe-protoporphyrin. The increasing enthalpy upon binding Fe-protoporphyrin is not unprecedented. Wu and Morgan (53) have reported on increasing enthalpy when hemopexin binds heme. The thermal unfolding of HO-2 and Fe-protoporphyrin.HO-2 are apparently irreversible under the conditions used, since no endotherm is seen on rescanning. The complexes of HO-2 with Zn-protoporphyrin in general have a broader transition state than those of apo-HO-2 or Fe-protoporphyrin.HO-2, suggesting that more than one type of association with HO-2 may occur with concomitant differences in thermostability. Furthermore, the form of the heat absorption peak upon melting of the Zn-protoporphyrin complex indicates that the denaturation process involves more than one individual transition stage of melting, and at least one of these stages is strongly affected by the metalloporphyrin. It seems likely that Zn-protoporphyrin may bind strongly to HO-2 a t the postulated heme binding site, as suggested by a shift in the melting point of one of the transition states to 61.5 "C; the metalloporphyrin may also bind to one or more additional sites with weaker affinity. The identification of specific domains that form the additional interaction sites with the metalloporphyrin is beyond the scope of this study. However, because HO-2 has 3 cysteine residues and zinc has a high affinity for S H groups, i t is reasonable to propose that cysteine residues may constitute sites for formation of additional complexes. of this manuscript; Dr. William McCoubrey, Jr., for preparation of