| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 277, Issue 9, 7191-7200, March 1, 2002
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
From the Department of Pharmaceutical Chemistry, School of Pharmacy, University of California, San Francisco, California 94143-0446
Received for publication, October 2, 2001, and in revised form, December 21, 2001
 |
ABSTRACT |
|---|
 TOP
 ABSTRACT
 INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES
|
|---|
The heme in lactoperoxidase is attached
to the protein by ester bonds between the heme 1- and 5-methyl groups
and Glu-375 and Asp-275, respectively. To investigate the cross-linking
process, we have examined the D225E, E375D, and D225E/E375D mutants of bovine lactoperoxidase. The heme in the E375D mutant is only partially covalently bound, but exposure to H2O2
results in complete covalent binding and a fully active protein.
Digestion of this mutant shows that the heme is primarily bound through
its 5-methyl group. Excess H2O2 increases the
proportion of the doubly linked species without augmenting enzyme
activity. The D225E mutant has little covalently bound heme and a much
lower activity, neither of which are significantly increased by the
addition of heme and H2O2. The heme is linked to this protein through a single bond to the 1-methyl group. The D225E/E375D mutant has no covalently bound heme and no activity. A
small amount of iron 1-hydroxymethylprotoporphyrin IX is
obtained from the wild-type enzyme along with the predominant
dihydroxylated derivative. The results establish that a single covalent
link suffices to achieve maximum catalytic activity and suggest that the 5-hydroxymethyl bond may form before the 1-hydroxymethyl bond.
The mammalian peroxidases, notably
LPO,1 MPO, EPO, and thyroid
peroxidase, are distinguished from the plant and fungal peroxidases in
that the prosthetic heme group of the mammalian enzymes is covalently
attached to the protein. The x-ray crystal structure of MPO, the only
mammalian peroxidase for which a structure is available, shows that the
heme group is attached via three links, (a) an ester bond
between the 1-methyl group of the heme and the carboxyl of Glu-242,
(b) an ester bond between the 5-methyl group and the
carboxyl of Asp-94, and (c) a thioether sulfonium bond between the No crystal data are available for LPO, but the prosthetic group
isolated from the digested protein has been identified by NMR and mass
spectrometric methods as the 1,5-dihydroxymethyl derivative of
heme (5, 6). The structure of the isolated prosthetic group indicates
that in LPO only two ester bonds covalently bind the heme to the
protein. Difference Fourier transform infrared studies provide
independent evidence that the heme is linked to the protein via ester
bonds (7). A structure model of LPO, constructed by sequence alignment
with the MPO template, suggests that Asp-225 is covalently linked to
the heme 5-methyl and Glu-375 to the 1-methyl (8). The involvement of
these two residues in cross-linking to the heme has been confirmed by
mass spectrometric and sequencing analysis of the fragments obtained
from protein digests (9, 10). The absence of a third cross-link
analogous to that in MPO between the vinyl group and a methionine
agrees with the fact that the spectrum of LPO, unlike that of MPO, is not shifted due to conjugation of the sulfonium-substituted vinyl group
with the porphyrin (2, 3, 11).
In an earlier communication (6) we demonstrate that the heme in the LPO
heterologously expressed in a baculovirus system is only partially
covalently attached. Furthermore, we demonstrate that incubation of the
recombinant protein with low amounts of H2O2
increased both the extent of covalent heme binding and the catalytic
activity (6). These results provide strong evidence that the
heme-protein bonds are formed through a self-processing mechanism.
Subsequent studies with EPO showed that the light and heavy chains were
not cross-linked in the majority of the isolated protein (10). In EPO
the heme is linked to Asp-93 and Glu-241, one of which is in the heavy,
and the other, in the light chain of the mature enzyme (10). Treatment
of the protein with H2O2 increased
cross-linking of the chains, in accord with the hypothesis that the
heme had only formed a link to one of the two non-cross-linked chains
rather than the usual link to both. The formation of the interchain
cross-link thus provided evidence that in EPO the heme cross-linking
reaction was also an autocatalytic process. Finally, although
cross-linking of the heme in thyroid peroxidase has not been
unambiguously demonstrated, it has been shown that exposure to
H2O2 is important for the formation of mature,
active protein (12). This provides indirect evidence that autocatalytic
formation of the heme-protein cross-links probably also occurs in
thyroid peroxidase.
In the present paper, we address three questions. (a) How
stringent are the conformational requirements for proper formation of
the ester cross-links? (b) Is the double covalent link
important for catalytic activity? (c) Is formation of the
two cross-links a sequential or random process? The results shed light
on the processing of LPO that is relevant to formation of the mature forms of all the mammalian peroxidases.
General--
Sf9 cells (Invitrogen) were grown in Excell
420TM (JRH Biosciences), and Hi5 cells were grown in Express FiveTM
(Invitrogen) supplemented with glutamine according to the
manufacturer's instructions. Both cell lines were kept in suspension
at 28 °C and maintained at densities between 0.5 × 106 and 2 × 106 cells/ml. All reactants
were purchased from Sigma unless otherwise noted, and all experiments
were performed at room temperature unless otherwise stated. Iron(III)
8-hydroxyprotoporphyrin IX was isolated from phenylhydrazine-treated
horseradish peroxidase as reported (13). UV-visible spectra and
ABTS-oxidizing activities were recorded on a Hewlett Packard 8452A
diode array spectrophotometer, and HPLC analyses were performed on a
Hewlett Packard 1090 instrument. The HPLC was equipped with a
UV-visible diode array detector, and the reactions were monitored at
215, 280, and 400 nm. Glass reaction vessels were siliconized when
needed using SigmacoteTM. After allowing the vessels to stand for 2-3
min with the manufacturer's solution, the vessels were rinsed twice
with hexane and once with acetone and, finally, air-dried.
Cloning and Expression--
The bovine LPO gene (2.1 kilobases) including the secretion signal sequence was cloned in the
EcoRI cloning site of the baculovirus pACGP67B expression
vector (Pharmingen, San Diego, CA) through introduction of this
restriction site at positions 120 and 2262 of the bovine LPO gene. The
D225E and E375D mutants were obtained by overlap extension PCR, the
first by substituting a C for a G at position 797 and the second by
substituting a G for a C at position 1247. The double mutant was
obtained by two successive overlap extension PCR steps. The resulting
constructs were amplified in the DH5 Purification--
Cell suspensions were harvested 2.5 days
post-infection and centrifuged at 4 °C for 1 h at 10,000 rpm.
The supernatant was diluted with an equal volume of double-distilled
water, after which protease inhibitors were added at the following
final concentrations: phenylmethylsulfonyl fluoride, 200 mg/liter;
N- Spectrophotometric Measurements--
Activities were followed at
414 nm in 50 mM acetate buffer (pH 4.5) in the presence of
250 µM ABTS and 125 µM
H2O2, both from freshly prepared stock
solutions. Specific activities were calculated using the values
Reaction with H2O2--
The appropriate
enzyme was diluted to a final concentration of 1 mg/ml in 0.1 M bis-Tris propane buffer (pH 8.2), and the required amount
of freshly diluted H2O2 solution
(10 Pronase Digestion--
Protein samples, pretreated with
H2O2 or not, at a 1 mg/ml concentration in 0.1 M bis-Tris propane buffer (pH 8.2) were made 1 mM in CaCl2. A freshly prepared solution of
trypsin (20 mg/ml, 1:5, w/w) was then added, and the mixture was
incubated for ~3 h at 37 °C. The CaCl2 concentration
was then raised to 3 mM, a freshly prepared 20 mg/ml
Pronase solution was added (1/5, w/w), and the incubation was prolonged
for a further 20 h. Samples were then made 15% in acetonitrile
and 0.5% in trifluoroacetic acid and were injected onto the HPLC. For
the experiment described in Fig. 6, the whole digestion procedure was
performed in siliconized glass microvessels.
HPLC Analysis--
Protein and porphyrin separations were
achieved on a 214TP5415 reversed phase C4 analytical column
(Vydac, Hesperia, CA) using 0.1% trifluoroacetic acid in
double-distilled water and 0.085% trifluoroacetic acid in acetonitrile
(Fisher) as mobile phase components at a flow rate of 1 ml/min. A
linear gradient of 0.85%/min starting at t = 3 min and
25% (undigested samples) or 15% (Pronase-digested samples) was used.
In the latter case, the compounds eluting from the column were
repetitively collected in siliconized glass tubes and concentrated
through gentle heating under a stream of N2. These
solutions were never allowed to dry out. For separation of the 5- and
8-hydroxyhemes, the same column and mobile phase were used, except that
a gradient of 0.2%/min starting at 30% was employed. Under these
conditions, 5-hydroxyheme eluted at 10.4 min and 8-hydroxyheme at 10.7 min.
Liquid Chromatography/Mass Spectrometry Analysis--
Molecular
masses were determined by liquid chromatography coupled with
electrospray ionization mass spectrometry. Liquid chromatography was
performed on an Applied Biosystems HPLC instrument (solvent delivery
system 1408, programmable absorption detector 785A) equipped with a
214TP51 reversed phase C4 capillary column (VydacTM,
Hesperia, CA) with 1% HCO2H in double-distilled water and
1% HCO2H in acetonitrile (Fisher) as the mobile phase
components (flow rate 40 µl/min). A 2%/min linear gradient
(20-60%) was used, and detection was performed at 215 nm. In-line
mass measurements were undertaken on a Perseptive Biosystems Mariner
Biospectrometry Work station in a positive ion polarity mode with the
following settings: spray potential, 2600-2700 V; nozzle potential,
125 V; 20 scans/min; flow rate 5 µl/min; acquisition range 300-2000
mass units. Calibration was achieved using a gramicidin standard.
Molecular Modeling--
Representations of the LPO molecular
model are based on the coordinates of De Gioia et al. (8)
and were obtained with RasMol.
Expression and Purification--
The enzymes were produced with
the baculovirus system previously developed for expression of the
wild-type enzyme (6). The viral particle yield after three
amplification stages was essentially independent of the mutation in the
protein, and titrating the final suspensions by plaque assay repeatedly
gave titers of ~ 1 × 108 plaque-forming
units/ml. However, although the E375D mutant was expressed as
efficiently as wild-type rLPO, both the D225E and D225E/E375D mutants
were consistently obtained in lower yields (Table
I) and were more sensitive to spontaneous
degradation during expression and purification. The addition of
protease inhibitors during the entire purification procedure helped to
improve both the yield and final purity of the proteins. The wild-type
and recombinant enzymes could be stored at
Polyacrylamide gel electrophoresis of the purified recombinant enzymes
indicated that they all have a molecular mass of ~73 ± 2 kDa
(Fig. 1A), a value
considerably lower than the 78 kDa of nLPO (Fig. 1A,
lane 1). The difference in these values indicates that the
recombinant proteins are glycosylated to a much lower extent than the
native enzyme. A lower extent of glycosylation has also been reported
for LPO expressed in Chinese hamster ovary cells (15) and in a
baculovirus-insect cell system (16) as well as for the heterologous
expression of MPO (16, 17). A small impurity at ~50 kDa in our
recombinant proteins was difficult to remove but did not interfere with
the subsequent studies. Further purification to remove the impurity was
therefore not carried out on a routine basis. In the case of the D225E
and D225E/E375D mutants, a low molecular weight
(Mr = 14.5 kDa) contaminant reproducibly co-purified with the protein (not seen in Fig. 1A, indicated
by the asterisk in Figs. 2 and
3). This low molecular weight contaminant contained a covalently
bound heme but did not derive from LPO and was probably a fragment of
cytochrome c from the insect cells. When isolated, this
fragment showed no detectable peroxidatic activity. The addition of 0.5 mM CaCl2 during the whole purification procedure eliminated this component but also significantly decreased the yield of the proteins.
As shown in Table I, rLPO and its mutants exhibited an absorption
maximum at 412 nm identical to that of nLPO except for the D225E
mutant, which had a slightly red-shifted Soret band at 415 nm. The
unperturbed absorption maxima indicate that the mutations had little
effect on the heme coordination sphere and environment.
In contrast, the RZ value, a measure of the iron porphyrin content of
the proteins, was very sensitive to mutation of the residues to which
the heme is attached. Wild-type rLPO gives an RZ value of 0.85 that is
very similar to that of nLPO and, thus, compares favorably with that
obtained with other LPO expression systems (14, 15), but the E375D
mutant has an RZ value reduced by 35% and the D225E mutant by 44%.
The severely depressed RZ = 0.09 value for the double mutant
indicates that this protein has very little heme bound at all, in
agreement with the fact that its Soret band is barely above the base
line. As expected and as indicated by the sharpness of the Soret bands,
the purification procedure, which includes two cation-exchange
chromatographic steps and one ultrafiltration, removes nonspecifically
bound heme from the proteins.
The ABTS oxidation activities qualitatively follow the same pattern as
the RZ values. The wild type has a specific activity somewhat lower
than nLPO, the E375D mutant ~65% that of the wild-type, and the
D225E an activity some 60-fold lower than that of the wild type. The
D225E/E375D mutant has almost no activity, in agreement with its very
low heme content.
HPLC Analysis--
The purified recombinant proteins and nLPO were
injected onto a C4 HPLC column under conditions that
separate non-covalently bound porphyrins from those covalently attached
to the protein (Fig. 2). Under these conditions, free heme derivatives
and heme-bound LPO peptides have an absorption maximum that is shifted
to ~400 nm, as reported (9). In the case of nLPO, the trace of the protein at 280 nm and of the porphyrin at 400 nm match very closely with a single peak at 29.5 min (HPLC RZ = 1.3). The heme cofactor in the native enzyme, as expected, is quantitatively bound in a
covalent manner to the protein. In the case of wild-type rLPO, the
major peak eluting at 29.7 min also absorbs at both wavelengths (HPLC
RZ ~ 0.9). In agreement with our previous observations (6), a
free peak attributable to unmodified heme was detected at 20.1 min.
These results indicate that in the recombinant wild-type enzyme only a
portion of the cofactor is covalently bound to the protein. The ratio
of the areas at 400 nm of the peaks at 20.1 and 29.7 min indicates that
approximately one-third of the heme is not covalently bound. The E375D
mutant, in addition to the free heme peak at 20.1 min, gives a
protein-bound heme peak with the slightly shifted retention time of
28.5 min. Approximately 65% of the heme was not covalently bound, and
the HPLC RZ for the protein-bound peak was ~0.5.
The elution profile of the D225E mutant is somewhat more complex. Apart
from a minor peak at ~14 min due to the already-mentioned insect
cytochrome c-like protein and a small free heme peak, the protein as measured at 280 nm reproducibly eluted as two partially resolved peaks at 30.0 and 31.5 min. The earliest peak, hence, had a
retention time similar to those for nLPO and wild-type rLPO but had
very little 400-nm absorbance associated with it, whereas the second
peak had an HPLC RZ of ~1.1, similar to that obtained with nLPO,
but eluted 1.5 min later. The D225E mutant is thus present as
two separable enzyme populations, one fully loaded with the cofactor
and the other almost completely devoid of it.
HPLC analysis of the double mutant indicates the presence
of a weak free-heme peak and a single protein peak at 28.5 min with virtually no covalently attached heme, as expected from the low RZ
value for the protein. The retention time of the double mutant protein
is close to that of the E375D single mutant.
H2O2 Treatment--
In earlier work we
demonstrated that treating wild-type rLPO with
H2O2 increases its content of covalently bound
heme, a finding that implicated an autocatalytic mechanism for covalent
attachment of the prosthetic group (6). We have therefore investigated the behavior of the recombinant proteins toward mild
H2O2 treatment. Preliminary experiments showed
that treating the wild type and E375D mutants with four eq of
H2O2 effectively increased the extent of
covalent heme binding (data not shown). The four recombinant enzymes
and nLPO were therefore incubated with H2O2
under these conditions, and the changes in the extent of covalent heme
attachment, iron porphyrin pattern, and activity were determined.
SDS-PAGE shows that the H2O2-treated proteins
did not undergo any detectable modification as a result of reaction
with the peroxide (Fig. 1B, compare with Fig.
1A). Mild oxidative treatment clearly did not result in
protein cleavage or other gross structural alterations. Under the
conditions of these studies, a small amount of oligomerization was
observed with nLPO and a trace with wild-type rLPO but none with the
mutant proteins. The low extent of oligomerization of nLPO and rLPO
contrasts with the extensive oligomerization observed under other
experimental conditions (18).
The specific activities of wild-type and E375D rLPO are increased by
preincubation with H2O2 by 27 and 370%,
respectively (Table II). Analogous
pretreatment of the D225E mutant does not alter its specific activity.
In contrast, the very low specific activity of the double mutant is
decreased by 40% and that of the native enzyme by 20%, following
peroxide pretreatment. It is notable that the final specific activities
of the wild-type and E375D mutant are comparable with each other and
only a little lower than that of the native enzyme.
HPLC analysis of the H2O2-treated samples
showed that the retention times of the proteins remained unchanged
(Fig. 3), in accord with their unaltered
migration on SDS-PAGE (Fig. 1B). However, for both the wild
type and E375D mutant, the 400-nm absorbance associated with the
protein peak increased markedly at the expense of the absorbance
associated with the free-heme peak, which virtually disappears. The
HPLC RZ values for the eluted proteins were similar to the value of
~1.2 obtained earlier for nLPO. This observation demonstrates that
the amount of non-covalently bound heme present in the wild-type and
E375D proteins is sufficient to give a final stoichiometry of
approximately one molecule of heme per molecule of protein.
In contrast, no significant changes were observed in the HPLC traces
for the D225E and D225E/E375D mutants as a result of H2O2 pretreatment. This is not unexpected, as
the HPLC analysis of the untreated enzymes indicated that they
contained very little non-covalently bound heme and, therefore, very
little heme that could be covalently incorporated into the active site.
The H2O2 treatment was therefore repeated in
the presence of increasing amounts of added hemin, but no increase in
catalytic activity was observed, and the slight increase in covalently
bound heme that was detected resulted from nonspecific covalent binding
(data not shown).
Iron Porphyrin Distribution--
To further study the effect of
mutating Glu-375 and Asp-225 on cross-linking of the heme to the
protein, the number and position of the covalent links to the heme in
each protein was investigated. Previous work has shown that digestion
with Pronase leads to partial liberation of the prosthetic group from
nLPO (6, 19, 20), but a later study reported higher efficiency with the
combined use of two proteases (9). Optimization of the conditions (not shown) led us to the combined use of trypsin and Pronase at pH 8.2 in
the presence of CaCl2 but the absence of added reductant. These conditions, which result in a very efficient liberation of
peptide-free heme derivatives, were employed with all the proteins except the D225E/E375D mutant, which had virtually no covalently bound
heme. The digested solutions were then directly injected onto the HPLC,
and species with a porphyrin-like UV-visible spectrum were collected
and analyzed by electrospray mass spectrometry.
The HPLC traces and the UV-visible spectra of the collected peaks are
presented in Fig. 4, A and
B, respectively. The slight solvent-dependent
red shift already mentioned was also seen here. Digestion of the native
enzyme gives a single species at 17.5 min with a molecular weight of
648 atomic mass units. Previous work has identified this compound as
1,5-dihydroxy heme or, more precisely, iron(III)
1,5-dihydroxymethylprotoporphyrin IX (6, 7, 9). Similar digestion of
wild-type rLPO before exposure to H2O2 yielded
three compounds. The major one, like that from nLPO, eluted at 17.5 min
and had a molecular mass of 648 atomic mass unit. It is, therefore,
also iron(III) 1,5-dihydroxyprotoporphyrin IX. Of the two minor peaks,
that at 31.1 min (m/z = 616) co-eluted with
heme and was unmodified iron(III) protoporphyrin IX. The second minor
peak, at 23.8 min, had a hydrophobicity intermediate between that of
heme and 1,5-dihydroxyheme and co-migrated with the heme derivative
isolated from the D225E mutant. It is, therefore, a monohydroxylated
heme derivative (see below). Upon treatment of the wild-type
enzyme with H2O2, both minor peaks virtually disappeared, and the peak at 17.5 min increased in intensity.
The pattern observed on proteolysis of the E375D mutant was quite
different. The major component was now free heme (Rt = 31.1 min, 616 atomic mass unit, 70% of total 400 nm absorbance), although two additional peaks were also detected. One of these eluted
at 17.5 min (648 atomic mass unit, 5% relative intensity) and
corresponded to the 1,5-dihydroxyheme peak obtained with nLPO. The
second minor peak eluted at 25.6 min (632 atomic mass unit, 25%
relative intensity) and had all the characteristics of a
monohydroxylated heme derivative. However, its elution time is clearly
different from that of the monoxyhydroxy derivative obtained from
wild-type rLPO (25.6 versus 23.8 min), so the two
monohydroxylated hemes are distinct entities. Upon
H2O2 treatment of the E375D mutant, the
monohydroxylated heme species increased at the expense of the free heme
peak, whereas the 1,5-dihydroxyheme peak only increased slightly.
Digestion of the D225E mutant, whether treated with
H2O2 or not, yielded a single detectable heme
species at 23.8 min. The intensity of this peak was not altered by
incubation with H2O2. Its retention time and
molecular mass (632 atomic mass unit) identify it as the same
monohydroxy derivative as was obtained from the wild-type protein. The
peak eluting at 15.7 min in the chromatogram (denoted by the
asterisk in Fig. 4A) is the cytochrome
c-like contaminant. Isolation and separate digestion of this
contaminant under the same conditions only gave back the same 15.7-min
peak, so the contaminant did not interfere with the proteolytic
analysis. As a further test of the product structures, co-injection of
the two monohydroxy derivatives with elution times of 23.8 and 25.6 min
with authentic iron(III) 8-hydroxymethylprotoporphyrin IX showed that
the hydroxyl group was not on the 8-methyl group in either of the
monohydroxylated heme derivatives. The 8-hydroxylated standard was
prepared as previously reported (13). Thus, four heme
derivatives are obtained from the recombinant proteins; they are heme
and 1,5-dihydroxyheme (rLPO and E375D mutant) and two different
monohydroxylated hemes that elute, respectively, at 23.8 min (wild type
and D225E mutant) and 25.8 min (E375D mutant), neither of which is the
8-hydroxyheme derivative.
To better define the relationship between the nature of the heme group
and peroxide treatment, the E375D mutant was treated with various
amounts of H2O2, the activity of the protein
was measured, and the heme species present in the protein were
quantitatively determined by digestion and HPLC analysis (Fig.
5). H2O2
treatment of the E375D mutant triggered a rapid disappearance of the
parent heme that is matched by a sharp initial increase in the
monohydroxylated derivative and a parallel rise in catalytic activity.
The proportion of 1,5-dihydroxyheme also increased but to a much lower
extent. As the number of added H2O2 equivalents
exceeds 2, the proportion of the dihydroxylated species increased as
that of the monohydroxylated heme decreases. The surprising finding,
however, is that the specific activity actually decreases at the higher
peroxide concentrations, where the 1,5-dihydroxyheme content increases
at the expense of that of the monohydroxylated heme. When the enzyme
treated with 10 eq of H2O2, which exhibited a
Soret band at 416 nm just after treatment, was treated with a limited
amount of potassium ferrocyanide, the Soret band shifted back to 412 nm. The sample thus obtained, when treated again with 10 eq of
H2O2, showed a further increase in the
proportion of the dihydroxylated heme derivative to 40% of the total
heme. However, this increase in double cross-linked heme was not
associated with a change in the catalytic activity (data not
shown).
Although the recombinant wild-type protein is less glycosylated
than the native protein and therefore has a lower molecular mass (Fig.
1A), the similarities in the spectroscopic, structural and
catalytic properties of the recombinant and native enzymes indicate
that the glycosylation pattern is not critical for enzyme function. The
recombinant wild-type enzyme has an RZ value comparable with that of
nLPO but a lower specific activity (Table I). This result is paralleled
by the fact that only 65-70% of the total heme content in wild-type
rLPO is covalently bound to the protein and is in accord with previous
results (6). Complete digestion of the enzyme, which releases the
prosthetic group, shows that the major component in rLPO is the same
1,5-dihydroxyheme that is found in nLPO. Two minor components were also
detected and were identified as unmodified heme and a monohydroxylated
heme. As reported previously (6, 14), mild H2O2
treatment causes a ~27% increase in the specific activity. We
establish here that this treatment also converts virtually all of the
heme and monohydroxyheme species into a single covalently bound
derivative that is released by proteolysis as 1,5-dihydroxyheme. This
finding suggests that the monohydroxyderivative is either the 1- or
5-hydroxylated derivative of iron(III)-protoporphyrin IX and is either
an intermediate or side product of the double cross-linking process. No
significant amount of any other heme derivative has been detected in
the wild-type rLPO hydrolysates. The observation of only one of the two
expected monohydroxy heme derivatives suggests that the cross-linking
pathway follows an ordered sequence in which one methyl group of the
heme is covalently bound (hydroxylated) before the other (Scheme
1).
INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES
-carbon of the 2-vinyl group and Met-243 (1, 2). Chemical, spectroscopic, and mutagenesis studies have provided supporting evidence for the sulfonium ion link (3, 4).
EXPERIMENTAL PROCEDURES
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES
Escherichia coli
strain, and the nucleotide sequence was verified by complete sequencing
of the insert. Co-transfection was performed using the BaculoGoldTM
(Pharmingen) transfection kit according to the manufacturer's
instructions. Viral stocks were generated and amplified from single
virus populations according to established procedures (14). Hi5 cells
were infected at a density of 2.106 cells/ml using a
multiplicity of infection of ~4. At the time of infection, a freshly
prepared hemin solution was added (final concentration 6 µM) along with penicillin (final concentration 100 units/ml) and streptomycin sulfate (final concentration 100 µg/ml).
-p-tosyl-L-lysine chloromethyl
ketone (TLCK), 75 mg/liter; benzamidine, 380 mg/liter; antipain and
leupeptin, 5 mg/liter. After stirring for 10 min, batch purification
was carried out as follows. Amberlite CG50 cation exchange resin
pre-equilibrated in 50 mM Tris.HCl (pH 7.4) was added in
three portions over 2 h (~20 ml of settled resin/liter of
supernatant). After decantation, the slurry was transferred to a
column, and the settled beads were washed first with ~5 volumes of 50 mM Tris.HCl and then with ~3 volumes of 150 mM Tris.HCl. The pH of both buffers was 7.4 and did or did
not contain 0.5 mM CaCl2. The CaCl2
was only employed in one purification of the D225E mutant, used to show
that it eliminates the cytochrome c-like contaminant. The
protein was finally eluted with 500 mM Tris.HCl at the same
pH. The fractions containing material absorbing at 412 nm were pooled
and diluted 10-fold with double-distilled water containing or not, as
indicated in the text, 0.5 mM CaCl2. The resulting solution, which was sometimes turbid, was chromatographed over a CL-6B cation exchange column (2-3 ml of resin/liter of culture). At the end of the loading, the column was washed with ~50
volumes of 50 mM Tris.HCl then ~50 volumes of 100 mM Tris.HCl, both at pH 7.8, containing or not 0.5 mM CaCl2 as indicated in the text.
Elution was achieved with a linear gradient of KCl (0-0.3 M, ~8 column volumes). Fractions were checked for
activity and absorbance at both 280 and 412 nm. Even in the case of the
double mutant, residual activities were high enough to allow
unequivocal identification of the LPO-containing fractions. Selected
fractions were then pooled and concentrated to at least 2 mg/ml by
ultrafiltration over AmiconTM YM30 membranes.
= 38,000 liters/mol/cm (oxidation product of ABTS) and
= 113,000 liters/mol/cm at 412 and 280 nm (LPO),
Mr = 78 kDa for nLPO and 73 kDa for all rLPOs,
independently of the number of ester bonds between the apo-enzyme and
the cofactor. Activities were expressed in µmol of ABTS/min/nmol of
enzyme. In the case of highly active samples (wild type and E375D
mutant), the enzymes were diluted to the necessary concentration in 5 mg/ml bovine serum albumin to limit enzyme losses through nonspecific adsorption. UV-visible spectra of whole samples were recorded in 0.1 M bis-Tris propane buffer (pH 8.2) at a concentration where absorbances were 1 absorbance units or lower. RZ values were calculated as A412/A280 in bis-Tris
propane 0.1 M, pH 8, as well as in the assay buffer. The
HPLC RZ values were calculated as the ratio of the absorbances at 400 and 280 nm.
4) was added in single equivalent portions every 3 min. The reaction media was further incubated for 10 min and was then
analyzed. For the experiment described in Fig. 6, the
H2O2 treatment was carried out in siliconized
reaction glass microvessels.
RESULTS
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES
20 °C for prolonged
periods without serious loss of activity.
Properties of the recombinant LPO proteins and commercial nLPO
View larger version (10K):
[in a new window]
Fig. 1.
SDS-PAGE (7.5%) analysis of the recombinant
LPOs. A, before H2O2 treatment.
B, after treatment with 4 eq of
H2O2. Samples (1.5 mg/ml) were incubated with
successive equivalents at 3-min intervals. Lane 1,
commercial sample; lane 2, wild-type LPO; lane 3,
E375D; lane 4, D225E; lane 5, D225E/E375D.
Samples were run under denaturing and reducing conditions (7.5% gel).
The arrow points at likely dimers.
View larger version (24K):
[in a new window]
Fig. 2.
HPLC analysis of nLPO and the four rLPO
proteins. Traces at 400 nm (dotted line, left
scale) and 280 nm (solid line, right scale)
of the commercial and the purified recombinant enzymes are presented.
The corresponding UV-visible spectra are shown to the right
of each trace. The asterisk indicates the peak
due to the cytochrome c-like impurity (see under
"Results"). For the D225E mutant, the spectrum in the dotted
line is for the species eluting at 30.0 min, and that in the
solid line is for the species eluting at 31.5 min. The
rising absorbance at the end of the run is due to material eluted in
the column wash.
Specific activities before and after treatment with four equivalents of
H2O2
View larger version (22K):
[in a new window]
Fig. 3.
HPLC analysis of nLPO and the four rLPO
proteins after treatment with four single equivalents of
H2O2. The HPLC traces at 400 nm
(dotted line, left scale) and 280 nm (solid
line, right scale) for the commercial and purified
recombinant enzymes are shown, and the corresponding UV-visible spectra
are presented to the right of each trace. The
asterisk indicates the peak due to the cytochrome
c-like impurity (see under "Results"). For the
D225E mutant, the spectrum in the dotted line is for the
species eluting at 30.0 min, and that in the solid line is
for the species eluting at 31.5 min. The rising absorbance at the end
of the run is due to material eluted in the column wash.
View larger version (12K):
[in a new window]
Fig. 4.
A, HPLC chromatograms of the
digested native and recombinant enzymes. Traces were recorded at 400 nm
for enzymes digested with or without prior pretreatment with
H2O2. Absorbances (milliabsorbance units
(mAU)) were normalized in each case to the highest peak for
the sake of clarity. No change was observed when nLPO was treated with
H2O2. B, UV-visible spectra of the
compounds isolated in panel A. wt, wild
type.
View larger version (19K):
[in a new window]
Fig. 5.
Variation of the activity and the relative
proportion of each heme derivative versus the number
of H2O2 equivalents with which the E375D mutant
was incubated, monohydroxy heme mono(OH), di(OH),
dihydroxyheme (heme b). A second experiment
showed the same trends but the absolute values differed slightly.
DISCUSSION
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES
View larger version (17K):
[in a new window]
Scheme 1.
Possible reaction pathways for the
formation of the covalent bonds between the protein and the heme in
LPO. R = (CH2)2COOH.
Replacement of an aspartate by a glutamate or vice versa is a conservative substitution with respect to side-chain hydrophobicity, pKa, and charge. The only substantive change in such a replacement is the difference of one methylene unit in the length of the side chain. Nevertheless, the D225E and E375D mutations differentially alter the properties of rLPO. The HPLC traces of rLPO and nLPO were very similar, but reproducible differences were observed in the elution times of the three rLPO mutants. The E375D and D225E/E375D mutants eluted ~1 min earlier than the wild-type enzyme, whereas the D225E mutant eluted as two distinct protein peaks, one without heme, that had the same elution time as the wild-type, and a second heme-bound species ~1.5 min later. The observation of these two D225E protein peaks is notable in view of the fact that only one HPLC peak is observed for both heme-free and heme-bound proteins with rLPO and its E375D mutant. It is unclear why covalent heme binding alters the protein properties to a larger extent in the D225E than E375D mutant, although we cannot exclude the possibility that two forms of the protein are present, only one of which has bound the heme group. Whatever differences are reflected in the HPLC elution times of the denatured D225E mutant components, the fraction that binds the heme group does not appear to be greatly perturbed in the folded state, as its Soret maximum differs by only 3 nm from that of the other proteins (Table I). It is perhaps interesting in this context that mutating Asp-94 in MPO also resulted in the observation of two protein species (21).
The D225E mutation arrested ester bond formation after a single cross-link was forged. Only one monohydroxyheme derivative was obtained after digestion of the protein, and this derivative was the same as the one isolated as a minor component from wild-type rLPO. No other hydroxylated heme was detected in the proteolytic digestion. If one assumes that a mutation will impair ester formation at the mutated site rather than at the alternative carboxyl site, mutating Asp-225 hinders formation to the 5-hydroxyheme cross-link. The single monohydroxylated derivative isolated from the D225E mutant can therefore be identified as Fe(III) 1-hydroxymethylprotoporphyrin IX, an assignment consistent with its spectrum, polarity, mass spectrometric molecular mass, and non-identity with the 8-hydroxymethyl derivative.
The sensitivity of the protein to Asp-225 mutations may reflect the critical position of this residue in the protein sequence. His-226, the adjacent residue, is the critical acid-base catalyst, and Asp-227, the subsequent residue, has been proposed to be part of the LPO calcium binding site (22, 23). The catalytic action of the protein and formation of the double heme cross-link may therefore be particularly sensitive to even small structural perturbations in this region of the protein. The formation of a single bond between the protein and the heme has a precedent in EPO, for which it has been reported that roughly two-thirds of the native enzyme has the heme attached through a single ester bond, and the other third, through two ester bonds (10). Exposure of EPO to H2O2, as first demonstrated for rLPO and confirmed here (6), results in complete double cross-linking of the heme prosthetic group.
Mutating Glu-375 less dramatically decreased the RZ value and specific
activity (~65 and ~35%, respectively) than mutating Asp-225.
Furthermore, incubation of the enzyme with 2 eq of
H2O2 yielded an enzyme with a heme content and
activity similar to that of the wild-type enzyme (Table II). The LPO
active site is, thus, more tolerant to structural perturbation in the
vicinity of the 1- than 5-methyl group of the heme. The porphyrin
distribution pattern before and after treatment with
H2O2 differed from that of the wild-type
enzyme. In the untreated E375D mutant, the porphyrin species were
unmodified heme (70%), a small amount of 1,5-dihydroxyheme (5%), and
a new monohydroxy heme derivative (25%). The new monohydroxylated species is 5-hydroxyheme, because it differs from the 1-hydroxyheme species from the D225E mutant, differs from authentic 8-hydroxyheme, and is the ester product expected when Glu-375 is mutated in the LPO
structure model. Incidentally, the non-identity with 8-hydroxyheme shows that the heme in the E375D mutant is not inserted in an orientation that is flipped by 180° around its meso
-
axis. In contrast to the D225E mutant, incubating the E375D
mutant with 2 eq of H2O2 converted the
unmodified heme in the enzyme almost completely to the monohydroxy
derivative (Fig. 4), and incubation with excess
H2O2 partially converted the monohydroxy heme
into the dihydroxylated species (Fig. 5). These observations imply that
the monohydroxyheme is an intermediate in formation of the double
cross-link. Clearly, with the E375D mutant, the first bond formed is
that between Asp-225 and the 5-methyl of the heme. The leveling off of
the proportion of the dihydroxylated heme formed at ~30% of the
total when more than 10 eq of H2O2 were added
can be explained by conversion of the LPO with a single heme cross-link to a metastable compound II or III species by reaction with the excess
H2O2 in the absence of any reductant. Indeed,
the shift in the absorption maximum from 416 to 412 nm upon reduction
shows that the H2O2-treated E375D mutant was in
an oxidized state. Thus, in the absence of a reductant that can return
the enzyme to the ferric state, no further reaction with
H2O2 and, consequently, no further ester bond
formation is possible. The observation of a further increase in the
proportion of the monohydroxy heme derivative when the enzyme was
reduced before exposure to additional H2O2 is
consistent with this proposal.
The vast majority of the active site residues are conserved among LPO,
MPO, and EPO, which suggests that the active site architecture is
strongly conserved in this family of proteins (8). A notable exception
is the three residues after Glu-375; in bovine LPO they are QIL, in
human EPO they are TPK, and in human MPO they are MPE. In the LPO
structural model, Glu-375 is located at the top of a loop, pointing
toward the 1-methyl carbon of the cofactor and appears to be in a less
congested and rigid environment. The only other catalytically important
residue in its immediate vicinity is Arg-372, which is three residues
away. The other elements surrounding the 1-methyl substituent are two
short helixes (492-498 and 529-539) and a longer helix (549-560)
connected by a loose loop to an additional very short helix (residues
491-498). These helices should provide a very hydrophobic environment
above the plane of the heme. In contrast, the B-C edge of the
tetrapyrrole ring is sandwiched between two roughly parallel helixes
spanning residues 214-229 and 457-469 that contain Asp-227, which is
a part of the putative calcium binding site, His-226, the catalytic
histidine, His-468, the proximal iron ligand, and Arg-465. The
equivalent residue to Arg-465 in MPO is Arg-499, a residue in direct
ionic contact with the C-ring propionate side chain. This very compact
and probably fairly rigid assembly leaves little room around Asp-225
and supports the conclusion that formation of the ester bond at this
site is relatively sensitive to mutations (Fig.
6). In general, the environment surrounding Glu-375 in the structural model appears to be more tolerant
than that around Asp-225.
|
The data in Fig. 5 show that in the E375D mutant, unmodified heme is present in the mutant, and maximum activity is obtained upon treatment with H2O2 when the proportion of the singly bound 5-hydroxyheme is highest. Increasing the H2O2 to 20 eq increases the amount of doubly bound heme at the expense of the singly bound while decreasing the catalytic activity. In contrast, only a limited amount of heme is bound in the D225E mutant and that through a single ester bond to the 1-methyl group. Furthermore, treatment with increasing amounts of H2O2 does not give rise to any doubly linked heme. These results suggest that covalent bond formation to the 5-methyl has less stringent requirements than to the 1-methyl (Scheme 1). It is therefore possible that the 5-methyl bond is normally formed faster. The 5-hydroxyheme bond is formed first in the E375D mutant, and then the second cross-link is readily formed with excess H2O2, even though this requires attachment to the shortened carboxyl chain at position 375. In contrast, in the D225E mutant, some heme is bound as a monoester to the 1-methyl group, but it is unable to form the second cross-link on exposure to excess H2O2.
Three conclusions follow from these studies. First, a non-covalently
bound heme in the LPO active site, whether mutated or not, can react
with H2O2 to give a reactive species that
promotes reaction of the 1- and 5-methyl groups with nearby carboxylic side chains. A covalent bond need not be present to allow the heme to
be regio- and chemo-selectively oxidized. Second, a single link between
the enzyme and the heme cofactor is enough to produce a fully
functional protein. Third, because the singly bound heme in the D225E
mutant is not able to forge the second link, its formation represents a
dead-end process in the mutant. Failure to form this second link could
be due either to an inherent difficulty per se in forming
that bond after the Glu-375 bond or/and a difficulty introduced by
extending the carboxylate side chain at position 225 in the D225E mutant.
| |
ACKNOWLEDGEMENTS |
|---|
We thank David Maltby for help in obtaining the mass spectrometric data. Mass spectrometry was carried out in the Bioorganic and Biomedical Mass Spectrometry Facility (A. Burlingame, Director), supported by National Institutes of Research Resources Grants BRTP PR01614 and 5P30 DK26743.
| |
FOOTNOTES |
|---|
* This work was supported by National Institutes of Health Grant GM32488.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
To whom correspondence should be addressed: School of Pharmacy,
S-926, University of California, San Francisco, CA 94143-0446. Tel.:
415-476-2903; Fax: 415-502-4728; E-mail: ortiz@cgl.ucsf.edu.
Published, JBC Papers in Press, December 26, 2001, DOI 10.1074/jbc.M109523200
| |
ABBREVIATIONS |
|---|
The abbreviations used are: LPO, lactoperoxidase; nLPO, native LPO; rLPO, recombinant LPO; MPO, myeloperoxidase; heme, iron protoporphyrin IX regardless of the iron oxidation and ligation state; ABTS, 2,2'-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid); EPO, eosinophil peroxidase; HPLC, high pressure liquid chromatography; RZ, Reinheit Zahl.
| |
REFERENCES |
|---|
|
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES
|
|---|
| 1. | Fenna, R., Zeng, J., and Davey, C. (1995) Arch. Biochem. Biophys. 316, 653-656 |
| 2. | Fiedler, T. J., Davey, C. A., and Fenna, R. E. (2000) J. Biol. Chem. 275, 11964-11971 |
| 3. | Taylor, K. L., Strobel, F., Yue, K. T., Ram, P., Pohl, J., Woods, A. S., and Kinkade, J. M., Jr. (1995) Arch. Biochem. Biophys. 316, 635-642 |
| 4. | Kooter, I. M., Moguilevsky, N., Bollen, A., van der Been, L. A., Otto, C., Dekker, H. L., and Wever, R. (1999) J. Biol. Chem. 274, 26794-26802 |
| 5. | Rae, T. D., and Goff, H. M. (1996) J. Am. Chem. Soc. 118, 2103-2104 |
| 6. | DePillis, G. D., Ozaki, S., Kuo, J. M., Maltby, D. A., and Ortiz de Montellano, P. R. (1997) J. Biol. Chem. 272, 8857-8860 |
| 7. | Kooter, I. M., Pierik, A. J., Merkx, M., Averill, B. A., Moguilevsky, N., Bollen, A., and Wever, R. (1997) J. Am. Chem. Soc. 119, 11542-11543 |
| 8. | De Gioia, L., Ghibaudi, E. M., Laurenti, E., Salmona, M., and Ferrari, R. P. (1996) J. Biol. Inorg. Chem. 1, 476-485 |
| 9. | Rae, T. D., and Goff, H. M. (1998) J. Biol. Chem. 273, 27968-27977 |
| 10. | Oxvig, C., Thomsen, A. R., Overgaard, M. T., Sørensen, E. S., Højrup, P., Bjerrum, M. J., Gleigh, G. J., and Sottrup-Jensen, L. (1999) J. Biol. Chem. 274, 16953-16958 |
| 11. | Wever, R., and Plat, H. (1981) Biochim. Biophys. Acta 661, 235-239 |
| 12. | Fayadat, L., Niccoli-Sire, P., Lanet, J., and Franc, J.-L. (1999) J. Biol. Chem. 274, 10533-10538 |
| 13. | Ator, M. A., David, S. K., and Ortiz de Montellano, P. R. (1987) J. Biol. Chem. 262, 14954-14960 |
| 14. | O'Reilly, D. R., Miller, L. K., and Luckow, V. A. (1992) Baculovirus Expression Vectors , pp. 128-130, W. H. Freeman and Co., New York |
| 15. | Watanabe, S., Varsalona, F., Yoo, Y.-C., Guillaume, J.-P., Bollen, A., Shimazaki, K., and Moguilevsky, N. (1998) FEBS Lett. 441, 476-479 |
| 16. | Shin, K., Hayasawa, H., and Lönnerdal, B. (2000) Biochem. Biophys. Res. Commun. 271, 831-836 |
| 17. | Moguilevsky, N., Garcia-Quintana, L., Jacquet, A., Tournay, C., Fabry, L., Piérard, L., and Bollen, A. (1991) Eur. J. Biochem. 197, 605-614 |
| 18. | Lardinois, O. M., and Ortiz de Montellano, P. R. (1999) J. Biol. Chem. 274, 35441-35448 |
| 19. | Sievers, G. (1979) Biochim. Biophys. Acta 579, 181-190 |
| 20. | Nichol, A. W., Angel, L. A., Moon, T., and Clezy, P. S. (1987) Biochem. J. 247, 147-150 |
| 21. | Kooter, I. M., Moguilevsky, N., Bollen, A., Sijtsema, N. M., Otto, C., Dekker, H. L., and Wever, R. (1999) Eur. J. Biochem. 264, 211-217 |
| 22. | Booth, K. S., Kimura, S., Lee, H. C., Ikeda-Saito, M., and Caughey, W. S. (1989) Biochem. Biophys. Res. Commun. 160, 897-902 |
| 23. | Shin, K., Hayasawa, H., and Lönnerdal, B. (2001) Biochem. Biophys. Res. Commun. 281, 1024-1029 |
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  M. Zederbauer, P. G. Furtmuller, M. Bellei, J. Stampler, C. Jakopitsch, G. Battistuzzi, N. Moguilevsky, and C. Obinger Disruption of the Aspartate to Heme Ester Linkage in Human Myeloperoxidase: IMPACT ON LIGAND BINDING, REDOX CHEMISTRY, AND INTERCONVERSION OF REDOX INTERMEDIATES J. Biol. Chem., June 8, 2007; 282(23): 17041 - 17052. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Huang, G. Wojciechowski, and P. R. O. de Montellano Role of Heme-Protein Covalent Bonds in Mammalian Peroxidases: PROTECTION OF THE HEME BY A SINGLE ENGINEERED HEME-PROTEIN LINK IN HORSERADISH PEROXIDASE J. Biol. Chem., July 14, 2006; 281(28): 18983 - 18988. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Colas and P. R. O. de Montellano Horseradish Peroxidase Mutants That Autocatalytically Modify Their Prosthetic Heme Group: INSIGHTS INTO MAMMALIAN PEROXIDASE HEME-PROTEIN COVALENT BONDS J. Biol. Chem., June 4, 2004; 279(23): 24131 - 24140. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
O. M. Lardinois and P. R. O. de Montellano Intra- and Intermolecular Transfers of Protein Radicals in the Reactions of Sperm Whale Myoglobin with Hydrogen Peroxide J. Biol. Chem., September 19, 2003; 278(38): 36214 - 36226. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. A. LeBrun, U. Hoch, and P. R. Ortiz de Montellano Autocatalytic Mechanism and Consequences of Covalent Heme Attachment in the Cytochrome P4504A Family J. Biol. Chem., April 5, 2002; 277(15): 12755 - 12761. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||
