(Received for publication, December 18, 1996, and in revised form, February 12, 1997)
From the Departments of Biochemistry and Biophysics and Environmental Medicine, University of Rochester School of Medicine, Rochester, New York 14642
The heme oxygenase (HO) system degrades heme to
biliverdin and CO and releases chelated iron. In the primary sequence
of the constitutive form, HO-2, there are three potential heme binding sites: two heme regulatory motifs (HRMs) with the absolutely conserved Cys-Pro pair, and a conserved 24-residue heme catalytic pocket with a
histidine residue, His151 in rat HO-2. The visible
and pyridine hemochromogen spectra suggest that the Escherichia
coli expressed purified HO-2 is a hemoprotein. The absorption
spectrum, heme fluorescence quenching, and heme titration analysis of
the wild-type protein versus those of purified double
cysteine mutant (Cys264/Cys281 Ala/Ala)
suggest a role of the HRMs in heme binding. While the
His151
Ala mutation inactivates HO-2,
Cys264
Ala and Cys281
Ala mutations
individually or together (HO-2 mut) do not decrease HO activity. Also,
Pro265
Ala or Pro282
Ala mutation does
not alter HO-2 activity. Northern blot analysis of ptk cells indicates
that HO-2 mRNA is not regulated by heme. The findings, together
with other salient features of HO-2 and the ability of heme-protein
complexes to generate oxygen radicals, are consistent with HO-2, like
five other HRM-containing proteins, having a regulatory function in the
cell.
A great deal remains to be learned about the biochemical and physiological functions of heme oxygenase (HO)1 isozymes, HO-1 and HO-2 (1, 2), which have been traditionally viewed only in terms of heme catabolism. And some confusion persists about why there are two forms of an enzyme that, by all appearances, have the same catalytic activity and substrate specificity. More puzzling is the high level expression of the second isozyme, HO-2, in tissues and cells that have no role in hemoglobin heme turnover (3-6). The aim of this study was to learn more about HO-2.
HO-1 and HO-2 catalyze the conversion of heme (Fe-protoporphyrin IX) to
biliverdin and CO and release Fe in a reaction that utilizes 3 mol each
of oxygen and NADPH (7). All products of HO activity are suspected to
be physiologically active (4-6). The constitutive form, HO-2, and the
inducible form, HO-1, are different gene products (8-10). Except for
catalyzing heme and sharing a stretch of amino acids known as the "HO
signature" (GenBankTM), they have little resemblance to
one another. The "HO signature" is part of a 24-residue domain,
which forms the heme catalytic pocket and that, except for one residue,
is conserved (11) among all HO-1s and HO-2s characterized to date. The
HO signature motif has a conserved histidine residue, His-151 in HO-2,
that is essential for its activity (12). In HO-1 the conserved
histidine stabilizes a distal water ligand, and based on experimental
findings (13) it is placed in a position close to the ligand binding
site and plays a role in oxygen binding/activation in the distal heme
pocket. At the primary amino acid level, the similarity between rat
HO-1 and HO-2 is a mere 43% (14, 15). HO-1, also known as HSP32, is
responsive to an extensive array of chemical agents and stimuli (reviewed in Ref. 16), while HO-2 is constitutively expressed in all
cell types, and the only inducers of the enzyme identified to date are
the adrenal glucocorticoids (10, 17). Glucocorticoids cause increased
association of the protein with the nuclear envelope as visualized by
immunostaining (18). HO-2 is a single copy gene with multiple
transcripts (3, 19, 20) ranging in size between ~1.3 and 2.1 kilobase
pairs that differ in use of three different 5-untranslated regions and
two poly(A) signals (19, 20). Between the two poly(A) signals a
consensus sequence of 5
-TTTTTGCA-3
is found, which is 100% identical
to the oxygen/nitrogen-sensing sequence (21-23) and is found in the
erythroprotein gene (21, 22).
Aside from the overall differences in amino acid composition of HO-1 and HO-2, a major difference is the presence of two cysteines in all HO-2s and the absence of this residue in all HO-1s (11, 14, 15, 19, 24-28). This residue is the axial ligand for the heme prosthetic moiety in various hemoproteins, including all cytochrome P450s and nitric-oxide synthase isozymes (29-33). In HO-2 this residue is flanked downstream by a proline residue followed by phenylalanine (Cys-Pro-Phe); two copies of such arrangements are present in the predicted sequence of the protein. The Cys-Pro dipeptide, often flanked downstream by a hydrophobic residue, phenylalanine, is the absolutely conserved core of the recently identified motif called the heme regulatory element (HRM) (34). There is a tendency for a positively charged residue (arginine or lysine) to flank the core upstream. The HO-2 core and the surrounding residues are as follows: Val261-Arg-Lys-Cys-Pro-Phe-Tyr-Ala-Ala-Gln and Gly278-Ser-Asp-Cys-Pro-Phe-Arg-Thr-Ala-Met.
In the second (Cys281-Pro) dipeptide, the upstream residues
(Gly-Ser-Asp) are polar; polar residues (glycine and serine) also flank
two copies of heme lyase HRM (34). HO-2 is among only six proteins
identified to have this core motif and the characteristic flanking
residues (34). The others are Saccharomyces cerevisiae heme
lyase (35), human erythroid aminolevulinate synthase (36),
Escherichia coli catalase (37), rabbit heme-regulated initiator factor
kinase (38), and S. cerevisiae HAP1, a
transcriptional activator that responds to oxygen/heme (34, 39, 40).
The HRMs bind heme and confer regulation by heme to proteins; in fact, only one copy of HRM is adequate for such activity (34).
Given the fact that in HO-2, two copies of the core HRM are present, we undertook the present study to examine whether, in intact HO-2 protein, the HRMs are involved in heme binding and to investigate their function in heme catalysis. In the course of this investigation we have found that HO-2 is a hemoprotein and provide good evidence to suggest that HRMs are involved in binding of heme but not in heme catalysis. Furthermore, we confirm that the His151 in the 24-residue heme pocket is essential for HO-2 activity.
Oligonucleotides for sequencing and mutagenesis
were obtained from Midland Certified Reagent Co. (Midland, TX) or Life
Technologies, Inc. Two oligonucleotides 10 residues long encompassing
the HRMs of HO-2
(Val261-Arg-Lys-Cys-Pro-Phe-Tyr-Ala-Ala-Gln)
and
(Gly278-Ser-Asn-Cys-Pro-Phe-Arg-Thr-Ala-Met)
were synthesized and purified by high pressure liquid chromatography
(95%) by Primm Laboratories (Cambridge, MA). These peptides herein are
referred to as the Cys264 and Cys281 peptides,
respectively. Nitrocellulose for Western blot analysis and Nytran for
Northern hybridization were from Schleicher and Schuell (Keene, NH).
Sequenase, version 2.0, random primer labeling system, restriction
enzymes, and other DNA modification enzymes were purchased from U.S.
Biochemical Corp. [-32P]dCTP and
[
-35S] dATP were obtained from U.S. Biochemical Corp.
or DuPont NEN. Reagents for protein determination were obtained from
Bio-Rad. All other reagents were of the highest quality commercially
available. Adult male Harlan Sprague Dawley rats and New Zealand
rabbits were obtained from Harlan Industries (Madison, WI). Rat
biliverdin reductase was purified essentially by the method described
previously (41). NADPH-cytochrome P450 reductase was purified as
described by Yasokuchi and Masters (42).
E. coli Inv F
(F
end A1
rec A1 hsd R17 (rk
,
mk+) sup E44 thi-1
gyrA96 relA1
80 lac Z
M15
(lac ZYA-arg F) U169
) carrying
HO-2 plasmids were grown to saturation overnight at 37 °C in 2 × YT medium containing 50 µg/ml ampicillin. Cultures were diluted
1:100 in the same medium and grown to an A600 of ~1.0. One milliliter was removed from the culture and prepared for
SDS-polyacrylamide gel electrophoresis as previously detailed (19). To
assay HO-2 expression, 30-µl aliquots of bacterial lysates were
examined on a 12.5% polyacrylamide gel. The gel was electroblotted
onto nitrocellulose and probed with HO-2 polyclonal antibody as
described previously (8). Cell lysates were prepared essentially by the
method of Scopes (43) in a buffer containing 20 mM
Tris-HCl, pH 7.5, 1 mM EDTA, 20% (v/v) glycerol, and 0.4% (v/v) Triton X-100. Total extracted protein in bacterial lysates was
quantitated by the method of Bradford (44) using bovine serum albumin
as the standard. HO activity was measured as previously detailed (15)
in the presence of added purified rat liver biliverdin reductase and
NADPH-cytochrome P450 reductase. One unit of activity was defined as
producing 1 nmol of bilirubin/h.
Plasmid DNA from the rat HO-2
expression clone, pRHOP (12, 15) was utilized as the substrate to carry
out site-directed mutagenesis of cysteine residues using the mutagenic
primers indicated below as detailed previously (12), and the products
were transformed into E. coli XL-1 Blue (recA1,
lac(F
, proAB, lacIqZ
M15,
Tn10(tetR))) cells. The HO-2 mutagenesis primers were
5
-AGCATAAAAGGGGGCTTTACGTACATC-3
, complimentary to
nucleotides 778-804 (15) with a GC mismatch (boldface type) for CA to
convert Cys264 into alanine;
5
-CCGGAAGGGGGCGTTGCTGCCTCC-3
, complimentary to nucleotides
829-852 also with a GC for CA mismatch to convert Cys281
into alanine; and 5
-CGAGTGTAAGCCGCGGCCACCAG-3
,
complementary to nucleotides 442-464 with mismatches to convert
His151 to alanine. Transformants were screened initially
for the loss of the unique NdeI site and the gain of a
single NcoI site and were subsequently sequenced by the
method of Chen and Seeburg (45) to identify mutants. DNA from mutant
clones and the parental plasmid were separately transformed into
Inv
F
cells (Invitrogen, San Diego, CA) which were assayed for
immunoreactive protein and heme oxygenase activity as described above.
The double mutant Cys264/Cys281
Ala/Ala,
referred to as "HO-2 mut," was generated from the Cys264
Ala mutant using the Cys281
Ala
mutagenesis primer. The same method was also used to generate Pro265 and Pro282
Ala mutants using
oligonucleotide primers complimentary to nucleotides 778-810
(5
-CTGAGCAGCATAAAATGCGCATTTACGTACATC-3
) and
nucleotides 832-850
(5
-GGCTGTCCGGAATGCGCAGTTGCTGGC-3
), respectively
(mismatched nucleotides are in boldface type). Both mutagenic primers
introduce an FspI site, and transformants were initially
screened for the presence of the restriction site prior to
sequencing.
The wild-type and mutant HO-2 constructs were subjected
to an additional round of polymerase chain reaction-mediated
mutagenesis. DNA from the carboxyl-terminal substitution mutant, pRHOP
(12), and the double Cys Ala mutant above were utilized as template for the polymerase chain reaction utilizing the primers
5
-CGCAATTAACCCTCACTAAAG-3, which represents the sequence upstream of
the polylinker of the vector, pBS+, and
5
-GTCGACTAATGATGATGATGATGATGCTTATCGTCATCGTCCTGCAGGCTAGGCTTCCTG-3
, which represents the reverse complement of HO-2 nucleotides
870-888 plus a histidine "tag" of six codons (boldface type) and
an enterokinase cleavage site (underlined) to facilitate removal of the
tag from the fusion protein. The primer also introduces a stop codon
(double underline) and a SalI restriction site (italics)
following the histidine tag (5
to it in the primer). Polymerase chain
reaction products were cloned into the vector pCRII (Invitrogen
Corporation, San Diego, CA). Transformants were screened for the
presence of a 934-base pair SalI fragment representing the
tagged HO-2 coding region. The inserts were sequenced to confirm their
identity and subcloned into the SalI site of pBS+
(Stratagene, La Jolla, CA), and the resultant plasmids were transformed
into Inv
F
. The orientation of the insert was determined by
restriction analysis and confirmed by sequencing.
A similar construct was generated for HO-1. In this case first strand
cDNA was generated from 1 µg of testis poly(A) RNA using the
cDNA cycle kit (Invitrogen) priming with oligo(dT) and was used as
a template for polymerase chain reaction using the primers 5-GGGAAGCTTGGAGCGCCCACAGCTCG 3
, representing nucleotides
2-18 of HO-1 (14) and a HindIII restriction site (italics)
and
5
-AAGCTTATGATGATGATGATGATGCTTATCGTCATCGTCCATGGCATAAATTCCCACTG-3
, which contains the reverse complements of nucleotides 848
867, an enterokinase cleavage site (underlined), histidine tag (boldface type), stop codon (double underline), and HindI III site
(italics). The fragment was cloned into the HindIII site of
the pBS+.
Fusion proteins were purified from overnight bacterial cultures
utilizing ProBondTM (Invitrogen) columns in accordance with
the manufacturer's instructions. Pooled peak fractions from the
ProBondTM column elution were buffer-exchanged into 50 mM Tris-HCl, pH 8.0, containing 10% glycerol, 1 mM CaCl2, and 0.1% Tween 20 and concentrated
to approximately 1 mg/ml. When necessary, the histidine tag was removed
using enterokinase (EKMaxTM, Invitrogen) utilizing 1 unit/mg protein and digesting for 16 h at 14 °C. Removal of
enterokinase was accomplished using soybean trypsin inhibitor-agarose
(Sigma) as described by the manufacturer. Eluted proteins were then
buffer-exchanged into a buffer appropriate to the final application (20 mM Tris-HCl, pH 7.5, containing 1 mM EDTA and
20% glycerol for storage at 80 °C) using Centricon 10 microconcentrators. Preparations were judged to be >90% homogeneous as assessed by SDS-polyacrylamide gel electrophoresis followed by
staining with Coomassie Brilliant Blue. The SDS gel profile of final
preparations is shown in Fig. 1. (The lower molecular weight band in HO-2 mut is often observed in SDS gels of HOs and is due
to the cleavage of the proteins.)
Spectral Analysis of Heme Binding Using E. coli Expressed, Purified HO-2
The hemoprotein nature of HO-2 was established by
examining spectral properties of the purified protein preparations over the range of 350-650 nm at 2 nm/s. The reference was 0.1 M
Tris-HCl, pH 7.5, with 0.01% Tween 20, while the test cuvette
contained purified protein in the same buffer at the concentrations
indicated in the figure legends. Heme was quantitated by a pyridine
hemochromogen assay (46). The change in OD between 557 and 575 nm was
used to determine the heme concentration using an extinction
coefficient of 32.4 mM1 cm
1.
For heme binding studies, heme was prepared fresh by dissolving in a
1:1 (v/v) mixture of 1 M NH4OH/methanol, and
volume was adjusted by the addition of 0.1 M Tris-HCl (pH
7.5) containing 0.01% Tween 20. Heme binding was determined by
absolute absorption spectroscopy using buffer solution as the
reference. Reconstitution of HO-2 with heme was carried out by
incubating purified HO-2 with a 5-fold molar excess concentration of
heme. After incubation at 4 °C for 1 h, excess heme was removed
by chromatography through a G25 column. 0.5-ml fractions were
collected, and heme and protein concentrations were measured in each
fraction.
The heme binding of the purified wild-type HO-2 and HO-2 mut proteins was also examined by UV fluorescence quenching (47, 48), and for comparison that of rat HO-1 was also examined. The fluorescence of a solution of protein (0.5 µM) in 0.1 M Tris-HCl, pH 7.5, was measured using 280 nm as the excitation wavelength and scanning emissions from 300 to 450 nm. The emission spectrum had a maximum at ~330 nm. Subsequently, incremental additions of heme were made to the cuvette, and the fluorescence was measured following each addition.
Northern Blot AnalysisA full-length (1300-base pair) HO-2
cDNA isolated from a rat testis DNA library (15) was used as HO-2
hybridization probe. Mouse -actin and HO-2 cDNA probes were
labeled with [32P]dCTP according to the manufacturer's
instructions, using the random primer DNA labeling system, and further
purified by spin column chromatography.
ptk cells were maintained at 37 °C under 5% CO2 in Dulbecco's modified Eagle's medium containing 10% bovine calf serum supplemented with 4.1 mM L-glutamine and 100 units/ml penicillin and 100 µg/ml streptomycin sulfate. For RNA analysis, cells were grown to 60-70% confluence. The medium was replaced with serum-free medium supplemented with GMS-X (Life Technologies, Inc.) for 2.5 h and then subsequently replaced with fresh serum-free medium or the same medium containing 25 µM heme. After 1 or 4 h, total RNA was prepared from control or heme-treated cells. Poly(A) RNA was isolated by oligo(dT)-cellulose chromatography, fractionated on a 1.2% (w/v) agarose gel, and transferred to Nytran. Prehybridization, hybridization of the appropriate 32P-labeled cDNA, and posthybridization treatment of the blots were performed essentially as described earlier (3).
The
involvement of HRMs and HO signature domains of HO-2 in heme catalysis
and binding was examined using wild-type and mutant HO-2 proteins. The
oligonucleotide primers indicated under "Experimental Procedures"
were utilized to substitute alanine for Cys264 and
Cys281 of the HRM sequences or His151 in the
conserved 24-residue domain in bacterial plasmids, and HO-2 was
expressed as a LacZ fusion protein in E. coli. Mutations were confirmed by sequence analysis, and expressed proteins were analyzed for HO activity; data are presented in Fig. 2.
As shown in panel a, the bacterial strain expressing the
His151 mutant does not have detectable activity, confirming
the previous report (12). Neither the mutation of Cys264
nor that of Cys281, however, caused a notable effect on
activity. Also, substitution of alanine for Pro265 or
Pro282 had no discernible effect on enzyme activity, and
essentially the same results were obtained as with Cys264
Ala and Cys281
Ala mutants (data not shown). To
address the possibility that only a single copy of the HRM sequence is
required for activity, a construct was generated combining both
cysteine mutations. As is shown in the figure, the double mutant also
did not display a considerable difference in activity from either of
the single mutant or the wild-type constructs. To investigate whether
the absence of activity in the His151 mutant was due to
decreased expression of the mutated protein as well as whether there
was an overexpression of Cys
Ala mutants, Western blot analysis of
the same E. coli expression cultures used for activity
analysis was carried out using equal amounts of bacterial cell lysate.
Fig. 2b shows that the lack of a detectable activity of
His151
Ala clearly was not due to the absence of the
expressed protein (lane 3). Indeed, we consistently observe
a higher expression of His151
Ala mutant than of the
wild-type protein (lane 2). Another important observation is
that an overexpression of Cys mutants, particularly the Cys double
mutant (lane 6) was not the reason for the unaffected
heme-degrading activity of the expressed protein.
Heme Content of Expressed HO-2 Protein
The finding that HRMs
are not involved in heme degradation encouraged us to question whether
HO-2 is a hemoprotein. The inserts from the wild-type and the
Cys264/281 Ala double mutant plasmids were utilized to
generate plasmids expressing the same proteins with a histidine tag
(His6), which could be removed by enterokinase digestion,
at their carboxyl termini. Wild-type HO-2 and HO-2 mut (double cysteine
mutant) were expressed in E. coli and purified. The purified
proteins, which were >90% homogenous, as assessed by
SDS-polyacrylamide gel electrophoresis (see "Experimental
Procedures"), were used to assess heme content and for spectral
analysis; results are presented in Fig. 3. The purified
wild-type protein has an intense Soret band at 406 nm; upon reduction
with dithionite, the maximum shifts to 424 nm, and the peak is at 421 nm for the ferrous CO complex. The absorption in the visible region
(500-700 nm) of the ferrous heme (inset a) shows a 632-nm
absorption band. The 630-nm band is typical of high spin hemoproteins.
The values obtained for absorbance maxima for the oxidized form at 406 nm are close to that of hemoglobin (403-406), which, at neutral pH,
exists predominately as a six-coordinate form with water at the sixth position. In contrast, the double mutant, in which the Cys residues of
both HRMs have been replaced by alanine, does not have a discernible heme spectrum, suggesting that, while wild-type HO-2 is a hemoprotein, this property is dependent on the presence of HRMs. The hemoprotein nature of HO-2 was confirmed using the pyridine hemochromogen assay
(Fig. 3, inset b). While the HO-2 mut had negligible levels of heme, the wild-type protein contained 0.66 nmol of heme/mg of
protein or an approximately 1:50 molar ratio of heme:HO-2. Clearly the
ratio of heme to protein is low, but it is important to remember that
the bacterial system is expressing the enzyme that degrades
heme synthesized by the bacteria and that the expression system was
designed to express HO-2, not heme biosynthesis enzymes. In addition,
the purification procedure required prolonged exposure to an acidic pH
of 6.0 and 0.5 M imidazole; at acidic pH the propionate side chains of heme are protonated.
Next, we determined if the purified protein could bind additional heme. For this, heme in molar excess was added to the purified HO-2 protein (5:1). The solution was incubated 1 h at 4 °C with gentle mixing, and then the nonspecific heme bound and free heme were removed by gel filtration chromatography. The molar ratio of protein to heme (pyridine hemochromogen) in the fractions was determined, and a value of 1:3 was obtained. In a previous study (49), using a conventional method of protein purification (all steps carried out at pH 7.5) and utilizing spectrophotometric titration (50) when plotting increments of the increase in absorbance at the Soret band against the molar ratio of exogenously added heme, the increase in absorbance reached a maximum at a ratio of 1:1. The difference between the present results and the previous one likely reflects the duration of interaction between heme and HO-2 protein, which was immediate in the previous study as well as the purification procedures; the association of heme with apoprotein can be time-dependent (51). Also, at that time we were unaware that the enzyme is a constitutive hemoprotein.
To further test heme binding to HRMs, the following experiments were
conducted. In the first series, the visible spectrum of wild-type
HO-2-heme and HO-2 mut-heme complexes using a fixed protein
concentration and increasing heme concentrations was examined (Fig.
4). As noted, at a protein:heme ratio up to 1:1 a shift of the Soret band of heme from 389 nm (inset) to 406 nm was
recorded for both the wild-type HO-2 and the HO-2 mutant. At a 1:1
ratio, the wild-type HO-2-heme complex had an extinction coefficient of
85 mM1 cm
1, whereas the HO-2
mut-heme complex had an extinction coefficient of 65 mM
1 cm
1 for the Soret band. As
seen, up to a 1:2 protein:heme ratio, the absorbance for the wild-type
HO-2-heme complex was consistently higher than the HO-2 mut. Above 1:1
there is a blue shift for the HO-2 mut-heme complex as the heme to
protein ratio increases, contrasting with the absence of such a shift
for the wild-type protein-heme protein complex up to a 1:3 ratio. The
Soret band of the wild-type protein-heme complex shifts toward blue
when heme is present at or above a 4-fold molar excess. We infer from these observations that the HRMs present in the wild-type protein bind
heme and are responsible for the difference in spectral behavior of the
two heme-protein complexes at higher concentrations. Thus, the addition
of heme at concentrations exceeding the potential specific heme binding
sites of the proteins appears to result in a shift toward the 389-nm
maximum of heme and may be due to nonspecific interactions. From the
differences in the amount of heme required to cause this shift, the
presence of two additional binding sites in the wild-type protein
compared with the mutant may be inferred.
Additional evidence suggesting binding of heme to HRMs is provided by
analysis of UV fluorescence quenching (Fig. 5) and the visible hemochrome (Fig. 6). Although fluorescence of
the wild-type HO-2 is not completely quenched until a 4-fold molar
excess of heme is added, only a 2-fold molar excess completely quenches fluorescence of the HO-2 mut. For comparison, the fluorescence quenching of HO-1, which binds heme with 1:1 stoichiometry (50) is
included. Further evidence for the interaction of HRMs with heme was
obtained from a visible hemochromogen. Hemochromogens have a different
color than heme, which is visible to the naked eye. The results when
10-residue Cys264 or Cys281 peptides in water
were added at a 2:1 molar ratio to a solution of heme are shown in Fig.
6. As noted, in the presence of Cys264, the brown color of
the heme solution shows a dramatic change in color and becomes
red-brown; the color change in the presence of Cys281 is
more subdued. A visible change in color was also observed at a 1:1
molar ratio.
Northern Blot Analysis of Cultured Cells following the Addition of Heme
A possible consequence of heme binding by HO-2 could be
autoregulation of HO-2 transcription. To address this possibility, ptk
cells were examined by Northern hybridization following exposure to 25 µM heme. As shown in Fig. 7a,
there was no detectable change in the level of the single
~1.3-kilobase pair HO-2 homologous transcript in this cell line
1 h after the addition of heme (lane 1) compared with
untreated cells (lane 2). Equal loading was confirmed by
probing the same blot with an actin probe (panel b). The
same result was also obtained when RNA was prepared 4 h after
exposure to heme (data not shown).
Presently we report that HO-2 is a constitutive hemoprotein, and, as indicated by site-directed mutagenesis experiments, the hemoprotein nature appears linked to the HRMs of the protein; these motifs are not involved in the catalytic activity of the protein. At this time, we cannot assuredly predict whether, in the natural state of the enzyme, each HRM binds one heme or a single heme forms a "bridge" between the two HRMs either by coordination or electron interactions or, for that matter, whether HO-2 is associated with heme under all conditions. Such analyses await more detailed structural characterization of the HO-2-heme complex. Nonetheless, theoretically speaking, if each HRM were to bind one heme molecule, then, when fully occupied, HO-2 protein would have three bound heme molecules, the third site being the 24-residue conserved "heme pocket" domain. The findings of the heme binding/heme reconstitution analysis experiments showing that 3 mol of heme appear to specifically bind to 1 mol of protein would be consistent with the possibility of each HRM binding one heme. Indeed, the involvement of HRMs in heme binding is supported by UV quenching analysis (Fig. 5) in that the HO-2 mutant profile closely resembles that of HO-1, which has a single binding site, while that of wild-type HO-2 is multiphasic, suggesting multiple sites with differential affinity for heme. Because of this apparent differential affinity of the sites for heme, prediction of the number of binding sites based on the initial slope of the titration curve (52, 53) is not possible. Interaction with the heme would be anticipated to take place with the catalytic site. This is suggested by the fact that heme quenches HO-1. HO-2 and HO-1 have a conserved tryptophan upstream of the conserved domain, Trp101 in HO-1 and Trp120 in HO-2. There are also two tyrosines in the conserved domain, Tyr153 and Tyr156 in HO-2 and Tyr134 and Tyr136 in HO-1. As has been reported for various hemoproteins, interaction of the side chains of nearby aromatic amino acids with heme can influence protein fluorescence (52-54).
While the HRMs do not appear to have a role in catalysis of heme by
HO-2, revealed by the observation that Cys264 Ala,
Cys281
Ala, Pro265
Ala,
Pro282
Ala, or double Cys mutations have little effect
on heme oxidation, the single histidine (His151) in the
conserved domain of HOs is clearly indispensable for HO-2 activity. As
noted earlier, HO-2 shares this domain of 24 amino acids with HO-1,
which also has no HRMs and is not known to be a hemoprotein (50, 55);
therefore, it is not surprising that HO-2 mut functions effectively as
a catalytic enzyme but does not display hemoprotein characteristics.
His151 of HO-2 is equivalent to His 132 of HO-1. In HO-1
mutation of His132 results in up to 80% loss of activity
and is essential for H2O2-supported oxidation
of heme by the enzyme (13). As suggested by studies with HO-1
mutagenesis, other histidines also seem to be important for heme
catalysis activity; for instance, His25, which corresponds
to His44 in HO-2, is the proximal heme iron ligand in HO-1
(56, 57).
It seems reasonable to suspect that HO-2, like the other five HRM-containing proteins (34-40), has a regulatory role, linked to heme binding, in the cell. HO-2 is unique among HRM-containing proteins by having what appears to be two distinctly different kinds of relation with the heme molecule; it binds heme at the HRMs, without degrading the metalloporphyrin, and it interacts with heme at the "heme pocket" for catalysis. In fact, HO-2 is the only protein described to date for which such dual function can be ascribed. The possibility that two different sites of HO-2 may have different functions is not, however, unprecedented; for example, the FixL protein of Rhizobium meliloti has a separate heme-binding/oxygen-sensing domain and a functional kinase domain (58).
The regulatory function of HO-2 can be envisioned to ultimately result in controlling cellular heme concentration, with HO-2 functioning as a "heme sensor." In a capacity as a heme sensor, HO-2 could also be visualized to serve as a transcriptional regulator by fine tuning the activity of transcriptional factors and genes that are heme-responsive, including HO-1 (7, 59). One may reason that in such a capacity, heme bound by the HO-2 HRMs would be protected from degradation and that when the concentration of heme exceeds the binding capacity of HRMs then excess heme would become available to serve as substrate for catalysis through interactions with the "catalytic pocket." As such, HO-2 could control expression of genes that are responsive to heme, although the enzyme itself is not regulated by heme (Fig. 7). The unresponsiveness of HO-2 gene regulation to heme would be an important aspect of its regulatory activity. Another heme-related regulatory function of HO-2 could involve the ability of heme to activate molecular oxygen and form reactive oxygen radicals; catalyzing oxygen radical formation is an inherent activity of the heme molecule that is greatly potentiated when it is bound to proteins (60). If HO-2 were involved in oxygen radical generation, then, aside from regulatory mechanisms for control of stress protein gene expression, another system that likely could utilize the reactive oxygen-generating function of HO-2 would be the male reproductive system, specifically the function of the sperm cells. These cells depend on hydroxyl radicals for function and are also damaged by excess levels of the radicals (61). Sperms, as well as their progenitor cells, have high levels of HO-2 (5), and the protein is present in the mature sperm acrosomes and the flagella. These segments of the sperm are essential for its capacitation. And capacitation of sperm is mediated by hydroxyl radicals (61). Given the fact that mature sperm neither have detectable biliverdin reductase nor NADPH-cytochrome P450 reductase, then generation of oxygen radicals would appear to be a plausible function of HO-2.
We are grateful to R. Vulapalli for assistance with cell culture studies and S. Bono for preparation of the manuscript.