ARCHIVES
OF BIOCHEMISTRY
AND
BIOPHYSICS
Vol. 280, No. 2, August 1, pp. 325-330, 1990
lmmunoaffinity Purification Thromboxane Synthase R. Niising, Faculty
S. Schneider-Voss,
of Human
and V. Ullrichi
of Biology, P.0. Box 5560, University
of Konstanz,
D-7750 Konstanz, Federal Republic of Germany
Received October 26,1989, and in revised form March 5,199O
A recently produced monoclonal antibody against human thromboxane synthase was used to purify the enzyme from platelets in a one-step procedure with good yields. The isolated protein exhibited a single band of about 58 kDa by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and contained one heme/mol. Although the visible spectrum of the oxidized enzyme displayed a peak at 418 nm like the previously isolated enzyme after dithionite reduction and CO addition, it shifted to 419 nm but not to 450 nm where only a small shoulder could be detected. Its catalytic activity was only l-5% of the previous preparations, but with the same K, of about 10 MM and a ratio of thromboxane B2: 12-hydroxyheptadecatrienoic acid of 1:l. Studies with EPR spectrometry and inhibitors confirmed that only a minor part of the enzyme was in its native heme-thiolate conformation, whereas the major part had been converted to the inactive P420 form by the elution procedure. The amino acid analysis revealed 46% hydrophobic residues. According to the sequence of 26 amino acids from the N-terminus and two tryptic peptides no homology to one of the cytochrome P450 monooxygenases, or to cyclooxygenase, or to prostacyclin synthase was detected. (c 1990 Academic Press, Inc.
Thromboxane A, (TxA,)’ has been implicated in various pathophysiological conditions as a proaggregatory and vasoconstricting mediator (1,2). Its biosynthesis involves arachidonate liberation, prostaglandin endoper’ To whom correspondence should be addressed. * Abbreviations used: SDS, sodium dodecyl sulfate; TX, thromboxane; HHT, 12.I-hydroxy-5,8,10-heptadecatrienoic acid; PAGE, polyacrylamide gel electrophoresis; HPETE, hydroperoxyeicosatetraenoic acid; CNBr, cyanogen bromide; Chaps, (3.[(3-cholamidopropyl)dimethylammonio]-l-propansulfonate; CO, carbon monoxide; UK37248, 4-(2.(lH-imidazol-l-yl)ethoxy)benzoic acid; OKY-1581, sodium (E)-3-(4-(3-pyridylmethyl)phenyl)-2-methylacrylate; MDA, malondialdehyde; PGH,, prostaglandin H,; DTT, dithiothreitol. Copyright co 1990 hy Academic Press, Inc. All rights of reproduction in any form reserved.
oxide formation, and isomerization to TxA2. The latter reaction which was first demonstrated in 1976 by Needleman et al. (3) in the microsomal fraction of human and horse platelets is catalyzed by Tx-synthase. Its activity was later found to be present in many other tissues such as lung, kidney, and spleen (4), as well as in macrophages (5) and lung fibroblasts (6). Several groups of investigators have attempted to purify Tx-synthase without much success. In 1985 Haurand and Ullrich (7) reported the purification of Tx-synthase with a molecular weight of 58.8 kDa for the monomeric enzyme. One heme per polypeptide chain was present and a close analogy to the group of cytochrome proteins was established by optical and EPR spectroscopy. The enzyme formed 12-L-hydroxy-5,8,10-heptadecatrienoic acid and malondialdehyde together with TxAz in a 1:l:l ratio. A mechanism explaining this product pattern has been proposed (8). Quantitative structure-activity relationships with inhibitors have confirmed this mechanism and have allowed us to propose a structure for the active site (9). In order to obtain more enzyme for further studies, we have attempted to replace our time-consuming and lowyield purification protocol by a more efficient procedure. Since monoclonal antibodies against this enzyme had been raised in our laboratory, we employed immunoadsorption for the isolation of Tx-synthase. The same approach had been previously undertaken by Shen et al, (10) for the enzyme from swine lung, but the protein obtained had different properties with regard to molecular activity, spectral characteristics, and molecular weight. The aim of the work described here was not only for a faster isolation procedure but also toward an explanation of these,discrepancies. MATERIALS
AND
METHODS
Materials. Culture media and supplements were purchased from Gibco Laboratories (Grand Island, NY) and fetal calf serum from Biochrom KG (Berlin, FRG). All materials for column chromatography, including protein A, were supplied by Pharmacia (Uppsala, Sweden). 325
326
NUSING.
SCHNEIDER-VOSS.
Nitrocellulose paper was purchased from Millipore (Bedford, MA). Enzyme-labeled antibodies were obtained from Jackson Immunoresearch Lab. (Avondale, U.S.A.). All solvents and other substances used were of analytical grade and obtained from E. Merck (Darmstadt, FRG), Roth (Karlsruhe, FRG), or Sigma. PGH,, HHT, and 5-, 12-, and 15-HPETE were biosynthesized according to published procedures (11, 12). OKY-1581 was a gift from ON0 Pharmaceutical Co., Ltd. (Osaka, Japan), and UK-37248 was kindly provided by Pfizer, Central Research (Sandwich, England). The monoclonal antibody Tue 300 described recently (13) was obtained from roller bottle cultures and purified by protein A chromatography according to Ey et al. (14). For immobilization 30 mg puriImmunoafinity chromatography. fied antibodies in 0.1 M NaHC03, 0.5 M NaCl, pH 8.3, were coupled to CNBr-activated Sepharose CL-4B according to the instructions of the supplier. For the precolumn, bovine IgG (200 mg) was equilibrated in 0.1 M NaHCO,, 0.5 M NaCl, pH 8.3. Cyanogen bromide activation of Sepharose CL-4B was performed according to the basic method outlined by Fuchs and Sela (15). After coupling, the column material was washed with 0.1 M glycine, pH 2.5, and equilibrated in 10 mM KzHPO,, 10% glycerol (v/v), 0.7% Chaps, pH 8.3. Microsomes of human platelets were prepared as described (7). Solubilization of Tx-synthase was achieved by homogenizing microsomal protein in 10 mM K,HPO,, 1 mM DTT, 1 mM EDTA, 10% (v/v) glycerol, 0.7% Chaps (solubilization buffer). After stirring on ice for 30 min, the mixture was centrifuged at 100,OOOgfor 90 min. The resulting supernatant was used directly for enzyme purification. Soluhilized human platelet microsomes were pumped at 7 ml/h over the lo-ml precolumn linked in series to the 5-ml column of Tue 300. Sepharose CL-4B. After loading onto the affinity column, the precolumn was removed and the antibody column was washed with the following buffers at 30 ml/h; solubilization buffer (30 ml); 50 mM K,HPO,, 10% glycerol, 1% Chaps, and 1 M NaCI, pH 7.4 (40 ml); 50 mM K,HPO,, 10% glycerol, and 0.1% Chaps (20 ml). Tx-synthase was then eluted with 0.1 M glycine-HCl, 10% glycerol, 0.1% Chaps, pH 2.7, or by conditions as indicated and the eluted fraction immediately neutralized by the addition of 1 M TrisHCl, pH 8.0. Optical spectroscopy. Optical difference spectra with Tx-synthase inhibitors were recorded with a Uvikon spectrophotometer between 500 and 300 nm with 500 ~1 affinity-purified Tx-synthase (120 pg) in both cuvettes. Optical absolute spectra were measured with 500 pl affinity purified Tx-synthase. For the CO-binding, the sample was gassed with CO for about 1 min, and then a few crystals of sodium dithionite were added to the sample cuvette. EPR spectroscopy. Enzyme-containing column material (500 ~1) was transferred to a quartz tube and frozen in liquid nitrogen. After the EPR spectrum was recorded, the sample was carefully thawed, 20 nmol UK-37248 was added, and a new EPR spectrum was recorded. Spectra were evaluated as described in (16). Enzyme activity was determined by Assay of Tx-synthase activity. measuring HHT formation at 234 nm versus time as described recently (17). The K, value was calculated from a Lineweaver-Burk plot. Polyacrylamide gel electrophoresis and immunoblotting of Txsynthase. Proteins were fractionated by 10% SDS-PAGE according to the method of Laemmli (18). The gel was silver-stained as described by Merril et al. (19). Mohilities were measured relative to the band of the tracking dye. For molecular weight estimation, the R, values of the protein markers were plotted using linear regression analysis. For immunoblotting, electrophoretic transfer of proteins from polyacrylamide gel to nitrocellulose paper was performed according to the procedure of Towbin et al. (20). Immunostaining was carried out as described (13). Amino acid and peptide sequence. Amino acid analysis was performed as described in (21). Tryptic peptides were separated by re-
AND
ULLRICH TABLE
I
Different Elution Conditions of ImmunoafIinity Chromatography
Condition pH 2.7 pH 12.0
4M
Urea
1 M NaSCN 1.5 M MgCl* 20% DMF 40% DMF
Protein be) 1.2 1.2 n. d”. n.d. n.d. n.d. 0.6
Activity (nmol/min/mg) 255 128 n.d. n.d. n.d. n.d. nd.
Note. After Tx-synthase was bound to the antibody column (1 ml bed vol), the column was washed as described in Fig. 1, eluted by the following indicated conditions, and dialyzed against 50 mM K2HP04, 10% glycerol, 0.1% Chaps, pH 7.4. The activity was assayed as described under Materials and Methods. ’ Not detected. versed-phase chromatography using a Vydac Cl8 column and an aqueous TFA/acetonitril gradient system. N-terminal sequence determination by automated Edman degradation was performed by a gasphase sequenator (Applied Biosystems, Type 470). Protein was estimated as described in (22) with bovine serum albumin as a standard.
RESULTS Previous investigations on the specificity of three monoclonal antibodies had revealed that only one, Tue 300, reacted with native Tx-synthase (13) and, hence, proved to be suitable for immunoaffinity purification. The enzyme activity was not influenced by monoclonal antibody binding (13). Therefore, we concluded that the epitope is independent of the active site. By the use of protein A-Sepharose, Tue 300 was able to precipitate the enzyme activity and the immunoprecipitate from lz51-labeled solubilized microsomes revealed a single band of 58 kDa. Larger amounts of monoclonal antibodies were obtained from roller bottle cell cultures purified on protein A-Sepharose chromatography and coupled to CNBr-activated Sepharose CL-4B. The column (5 ml bed vol) bound 30 mg of purified monoclonal antibody. One hundred sixty milliliters of solubilized human platelet microsomes was applied to a precolumn (10 ml bed vol, containing 180 mg bovine IgG coupled to CNBr-activated Sepharose CL-4B) and then directly to the antibody column. The precolumn was removed and the affinity column was washed as described under Materials and Methods. The tight binding of the enzyme to the monoclonal antibody Tue 300 allowed us to wash the column under very stringent conditions before final elution. The use of a precolumn to which unspecific bovine IgG was coupled made a partial purification of solubilized microsomes unnecessary. The column material was separated in aliquots of 1 ml bed vol and different elution conditions were tested (Table I).
IMMUNOAFFINITY
PURIFICATION
OF HUMAN
THROMBOXANE
327
SYNTHASE
MW WASH
,‘ASH I
II
v
v
(kD), 92-
ELUTION
Q
67-
451 31-
21-,
Id-1
,4--p--
A I
10
I
20 FRACTION
I
I
30
BO
NUMBER
of Tx-synthase on antibody affinity column. FIG. 1. Purification The column was washed first with 30 ml wash I buffer (50 mM KzHPOI, 10% glycerol, 0.7% Chaps, 1 mM DTT, 1 mM EDTA, pH 7.4) and then with 40 ml wash II buffer (50 mM K,HPO,, 10% glycerol, 1% Chaps, 1 M N&l, pH 7.4). The bound enzyme was eluted by 0.1 M glycine-HCl, pH 2.7, containing 10% glycerol and 0.1% Chaps into one-tenth volume of 1 M Tris-HCl, pH 8.0. The fractions were monitored at the wavelengths 280 and 405 nm.
The bound protein could be eluted only at extreme pH conditions such as pH 2.7 or pH 12.0 from the antibody column, with poorer recovery of activity for the latter. Other conditions such as 1.5 M MgC12, 20% dimethylformamide, or 1 M NaSCN failed to elute the enzyme. Forty percent dimethylformamide partly eluted the enzyme but totally destroyed the enzymatic activity. Figure 1 shows the elution profile of Tx-synthase from the antibody column, eluted with 0.1 M glycine-HCl, pH 2.7, containing 10% glycerol and 0.1% Chaps. In preliminary experiments, compared to Lubrol PX (7), Chaps improved separation and sharpened the peak of activity. These elution conditions were used throughout for immunoaffinity purification of Tx-synthase. The proteins applied were mostly eluted by the washing procedure while 80% of the applied enzyme, as determined by an immunometric assay, were retained. The purity of the enzyme from different stages of purification was assessed by SDS-PAGE and silver-staining, as shown in Fig. 2A. A single polypeptide having a molecular weight of 58 kDa was eluted from the antibody column. To verify the identity of the eluted protein as Tx-synthase, immunoblotting was carried out with solubilized platelet microsomes and the purified enzyme, probed with polyclonal antibody directed to conventionally purified human Tx-synthase. In each lane one band of 58 kDa, corresponding to thromboxane synthase (Fig. 2B), appeared. These results were strong evidence that the eluted protein was indeed Tx-synthase. A summary of the overall purification procedure is presented in Table II.
12
3
6
4
1
2
FIG. 2. (A) SDS-PAGE of column fractions from antibody column. Separation was performed in 10% slab gel. After electrophoresis, the gel was silver stained. Lane 1, solubilized microsomes (10 pg); lane 2, proteins not bound by the antibody column (10 Kg); lane 3, affinity purified enzyme (0.1 pg); lane 4, lysis buffer. (B) Immunoblotting of eluted protein. Enzyme samples were electrophoresed and transferred to nitrocellulose paper and incubated with polyclonal antibodies directed to Tx-synthase. Lane 1, solubilized microsomes (50 fig); lane 2, affinity-purified enzyme (50 ng).
Activity calculation from Table II reveals that immunopurified Tx-synthase has only about 3% of the original activity (7). However, the measured activity is nearly identical to the reported activity of immunopurified Txsynthase from swine lung by Shen et al. (10). In order to study the discrepancy between our earlier reported activity and the poor recovery of the immunopurified material we first used spectroscopic methods. The visible spectrum of affinity-purified Tx-synthase is shown in Fig. 3. The enzyme exhibited a Soret absorbance band at 418 nm. After gassing with CO and reduction with dithionite, the Soret band shifted to 419 nm with a light shoulder at 450 nm. This was in contrast to native Tx-synthase, which shifted completely to 450 nm after reduction in the presence of CO (7). The catalytic properties of the affinity-purified enzyme was further investigated by difference spectra with UK-37248. This compound binds selectively to the active site of Tx-synthase, causing a shift in visible spectra, which was well detected by difference spectroscopy (9). As shown in the inset of Fig. 3, the incubation of Tx-
TABLE
II
Purification of Thromboxane Synthase from Human Platelets
Purification
step
Microsomes Solubilized microsomes Affinity purified thromboxane synthase
Volume (ml)
Protein (mg)
80 82
1360 544
4
1.6
Specific activity (nmol/min/mg) 15.0 33.2 255.0
328
NUSING,
SCHNEIDER-VOSS.
AND
ULLRICH
015
s 6 D
01
:
i 9 005
MAGNETIC
0i
Loo
500
600
nm
FIG. 3. Visible spectrum of affinity-purified Tx-synthase. The spec120 +g of trum was monitored in 50 mM K,HPO,, pH 7.4,‘containing the enzyme against the blank buffer. Inset, difference spectrum of Txsynthase treated with UK-37248. Affinity-purified enzyme (120 pg) was present in the sample and reference cuvette. Tx-synthase inhibitor was added to the sample cuvette 5 min prior to recording.
synthase (120 pug) with UK-37248 yielded a difference spectrum with a minimum at 399 nm and a maximum at 427 nm. A rough calculation suggests that the activity could correlate with the light shoulder at 450 nm seen in the visible spectrum. No typical P450 signal could be seen (data not shown) from the EPR spectrum performed with about 10 nmol of enzyme. When, however, the immunoabsorbed fraction from the column, together with the Sepharose support, was recorded in the EPR spectrometer, the typical g tensor of cytochrome P450 with g values at 1,89, 2,25, and 2,45 appeared, although the g, and g, values were broadened (data not shown). Addition of the pyridinebased specific inhibitor UK-37248 sharpened the signal and resulted in the spectrum shown in Fig. 4, which verifies the protein nature as Tx-synthase, as shown earlier (7). The split signal at g = 2 does not belong to the P450 protein and probably indicates a partially denatured enzyme or a metal contamination. Due to the extensive washings, the material bound to the immuno-Sepharose matrix was about 95% pure Tx-synthase, as judged from the SDS-gel electrophoresis. The activity for HHT formation was inhibited in a dose-dependent manner by specific Tx-synthase inhibitors of the pyridine type as well as of the imidazole type. OKY-1581 inhibited the activity by 50% at 0.5 nM and UK-37248 was active at 1 nM. The K, for the enzyme with PGH, was 10 PM and the ratio of TXB, to HHT was determined as 0.9:1 by the incubation of Tx-Synthase with [14C]PGHZ, in agreement with our recent re-
FIELD
STRENGTH
IG1
FIG. 4. EPR spectra of immunoadsorbed Tx-synthase. Enzymecontaining column material, after extensive washing free of other proteins, was recorded after incubation with 20 nmol UK-37248 in an EPR spectrometer as described under Materials and Methods. The sample contained approximately 2 nmol of Tx-synthase.
ports (7, 8). The heat-denaturated enzyme did not produce HHT or TXB2. The activity of affinity-purified Tx-synthase was dramatically reduced by multiple additions of PGH, (Fig. 5). The inactivation effect was dependent on the substrate concentration used in relation to the amount of enzyme present. The initial control activity also increased with increasing PGH2 concentration but, with higher substrate levels, the fall in activity after multiple additions was more pronounced. This inactivation process was further investigated by the addition of 10 and 50 PM TXB,, HHT, or MDA together with PGHz. Neither of these enzyme products had an inhibitory effect on Tx-synthase (Table III). Hydroperoxy fatty acids have been reported to inhibit Tx-synthase in platelet microsomes, but the mechanism
2W QJM 20pM
t t t t Addition of mdlcated PGHz concentration
FIG. 5. Inactivation of Tx-synthase by PGH,. Tx-synthase activity was assayed in the presence of the indicated PGH, concentration and again after every 2 min (arrows) with the addition of fresh PGH, of the indicated concentration.
IMMUNOAFFINI’I’Y TABLE
PUHIFICA’l’IUN
VP HUMAN
III
activity
Concentration (PM)
TxB,
% of control
10 50 10 50 10 50 10 50
HHT MDA 5-HPETE 12.HPETE
10
15-HPETE
50 10 50
96.3 92.8 93.3 87.1 100.0 93.3 58.3 11.1 47.2 2.8 83.3 111.1
Note. Different metabolites in the indicated concentration were added simultaneously with PGH, to affinity-purified Tx-synthase and the activity was measured as described under Materials and Methods. Each value is the mean of duplicate assay.
of this inhibition was not established (23). As listed in Table III, 5- and 12-HPETE inhibited the enzyme in a dose-dependent manner after coincubation of Tx-synthase with PGHZ. 15HPETE had no inhibitory effect on Tx-synthase activity, even at high concentrations. Since the rapid purification protocol and the good yields made larger amounts of protein available, the amino acid composition of Tx-synthase could be investigated. This investigation revealed that the enzyme is a moderately hydrophobic protein containing approximately 46% hydrophobic residues (Table IV). Because the amino terminus was not blocked, we were able to sequence the N-terminus of Tx-synthase with a length of 26 amino acids (Table V). The sequences of two tryptic peptides are also presented. DISCUSSION Our attempts to purify human Tx-synthase by immunoaflinity were successful as far as yields and amounts TABLE
IV
Amino Acid Composition of Tx-Synthase Amino acid Asx Ser Pro Ala Val Ile Tyr His Arg
45 23 41 46 27 16 14 10 33
Amino acid Thr Glx Gly cys Met Leu Phe Lys
19 44 31 l-2 17 60 34 22
329
V
Partial Protein Sequences of Tx-Synthase NH,-terminus
Metaholite
b Y N 1 HA3b
TABLE
of Different Eicosanoid Metabolites on Tx-Synthase Activity
Effect
‘l’HKUMl5UKANr;
Peptides Peptide 1 Peptide 2
MettGlu-Ala-Leu-Gly-Phe-Leu-Lys-Leu-GluVal-Asn-Gly-Pro-Met-Val-Thr-Val-AlaLeu ?“-Val-Ala-Leu-Leu-Ala Ile-Lys-Gln-Val-Leu~Val-Glu-Asn-Phe-SerAsnPhe-Thr-Asn-Arg VallProoLeu-Ala-Arg-Ile-LeuPro-AsnLys Asn-Arg-Asp-Glu-Leu-Asn-Gly-Phe-Phe-AsnLys-Leu-Ile-Arg-Asn~Val-Ile-Ala-Leu-ArgAsp-Asn-Asn-Ala-Ala-Glu-Glu-Arg-Arg-?-GluPhe-Leu-Glu
DAmino acid not identified.
of enzyme were concerned. Whereas molecular weight, heme content, and antigenic properties were in agreement with those of the conventionally purified Tx-synthase, the molecular activity reached only a few percent. From our results we can conclude that this was a consequence of the rather rigid conditions of elution of the antibody-adsorbed enzyme since the EPR spectrum of the immunoadsorbed enzyme showed a typical P450 g tensor, whereas with the eluted enzyme, this spectrum which is typical for an intact Fe(III)-thiolate structure (24) had disappeared. In addition, the reduced carbon monoxide spectrum revealed most of the enzyme to be present in the P420 form with only a weak shoulder at 450 nm. We had previously confirmed that as in the cytochrome P450 monooxygenase systems, a conversion to the P420 form was paralleled by loss of Tx-synthase activity (25). In the immunopurified enzyme preparation, the inhibitors OKY-1581 and UK-37248 blocked the enzyme with the same high affinity as for the native enzyme. Little binding was observed spectrally, which agreed with only a small fraction of the hemoprotein reacting with the inhibitor. Also, the affinity for the substrate PGH, was unchanged. Thus, it seemed likely that we were dealing with a few percent of active protein and with a bulk of denatured enzyme unable to bind substrate or inhibitor. These results may also explain the data of Shen et al. (10) on their immunoaffinity purification of Tx-synthase from swine lung, which also indicated a pure but less active protein. Since similar elution conditions were employed, the enzyme was probably largely inactive and in a native state may have exhibited a similar high molecular activity as did the human enzyme from platelets. In contrast to the data by Shen et al. (lo), human Txsynthase exhibited a molecular weight of 58 kDa as did the enzyme from human lung and swine platelets, as verified by immunoblot technique (data not shown). The fact that the P420 form of Tx-synthase is devoid of activity supports our hypothesis that an intact iron(III)-thiolate bond is required for Tx-synthase ac-
330
NikING,
SCHNEIDER-VOSS,
tivity (7,26). A loss in heme content during affinity purification is unlikely because our measured ratio of 1:l for heme to protein is in agreement with earlier results (7) and the residual activity was not affected by exogenous heme (data not shown). As also reported by others (10,27), we observed a rapid self-inactivation of Tx-synthase by its substrate PGHz. This effect was concentration-dependent and was not mediated by the products TxB,, HHT, or malondialdehyde. Since 5-and 12-HPETE led to a similar inactivation, one may assume that oxidative processes are responsible. In the late stage of massive platelet aggregation, 12-HPETE may accumulate. Therefore, for platelets the inactivation of Tx-synthase may even be a physiological event. In our experiments, 15-HPETE was not effective in destruction, in contrast to reports by Shen et al. (10). Although our preparation did not turn out to be suitable for activity measurements, it made available relatively large quantities of protein which could be used for a protein chemical characterization. The amino acid composition after acid hydrolysis gave about 46% hydrophobic residues, which corresponds well with other microsomal proteins and also with the 44% of the closely related prostacyclin synthase (28). The low content of cysteine is remarkable and may indicate that only one cysteine is allowed to be present in the polypeptide. This would then have to be the heme-interacting cysteine at the active site. The N-terminus also contained a majority of hydrophobic residues but, as for the two presented peptides, no striking homology was found between Txsynthase and other cytochrome P450 proteins. We found homology neither with cyclooxygenase (29) nor with reported peptides of prostacyclin synthase (28). We expect that Tx-synthase is synthesized without a cleaved leader sequence, since the amino terminus is methionine. The high content of hydrophobic amino acids suggests a membrane spanning region of the presented sequence. Efforts for sequencing Tx-synthase with molecular biological methods are now being undertaken and will then allow us to conclusively compare Tx-synthase with related enzymes. ACKNOWLEDGMENTS We amino by Dr. ported
thank Dr. F. Lottspeich (MPI Munich, FRG) for performing the acid and sequence analyses. EPR spectra were kindly recorded P. Kroneck (University of Konstanz, FRG). The study was supby the Deutsche Forschungsgemeinschaft (SFB 156/A4).
REFERENCES 1. Needleman, P., Turk, J., Jakschik, B. A., Morrison, A. R., and Lefkowith, J. B. (1986) Annu. Reu. Biochem. 55,699102.
AND
ULLRICH
2. Granstrom, E., Diczfalusy, U., Hamberg, M., Hansson, G., Malmsten, C., and Samuelson, B. (1982) in Prostaglandins and the Cardiovascular System (Oates, J. A., Ed.), pp. 15-58, Raven Press, New York. 3. Needleman, P., Moncada, S., Bunting, S., Vane, J. R., Hamberg, M., and Samuelson, B. (1976) Nature (London) 261,558-560. 4. Hamberg, M. (1976) Biochim. Biophys. Acta 431,651-654. 5. Morley, J., Bray, M. A., Jones, R. W., Nugteren, D. H., and van Dorp, D. A. (1979) Prostaglandins 17,730-736. 6. Hopkins, N. K., Sun, F. F., and Gorman, R. R. (1978) Biochem. Biophys. Res. Commun. 85,827-836. 7. Haurand, M., and Ullrich, V. (1985) J. Biol. Chem. 260, 15,05915,067. 8. Hecker, M., and Ullrich, V. (1989) J. Biol. Chem. 264,141-150. 9. Hecker, M., Haurand, M., Ullrich, V., and Terao, S. (1986) Eur. J. Biochem. 157,217-223. 10. Shen, R. F., and Tai, H. H. (1986) J. Eiol. Chem. 261, 11,59211,599. A., and Ullrich, V. (1987) Biochem. 11. Hecker, M., Hatzelmann, Pharmacol. 36,851-855. 12. Porter, N. A., Logan, J., and Kontoyiannidou, V. (1979) J. Org. Chem. 44,3177-3181. 13. Niising, R., Wernet, M. P., and Ullrich, V. (1989) Submitted for publication. 14. Ey, P. L., Prouze, S. J., and Jenkin, C. R. (1978) Immunochemistry 15,429-436. Im15. Fuchs, S., and Sela, M. (1978) in Handbook of Experimental munology (Weir, D. M., Ed.), Vol. 10.1-10.6, Blackwell Scientific, Philadelphia, PA. 16. Marchesini, A., and Kroneck, P. (1979) Eur. J. Biochem. 101,6576. 17. Hecker, M., Haurand, M., Ullrich, V., Diczfalusy, U., and Hammarstrom, S. (1987) Arch. Biochem. Biophys. 254,124-135. 18. Laemmli, U. K. (1970) Nature (London) 227,680-685. 19. Merril, C. R., Goldman, D., Sedman, S. A., and Ebert, M. H. (1981) Science 2 11,1437-1438. 20. Towbin, H., Staehlin, T., and Gordon, J. (1979) Proc. Natl. Acad. Sci. USA 76,4350-4354. 21. Lottspeich, F., Kellermann, J., Henschen, A., Rauth, G., and Miiler-Esterl, W. (1984) Eur. J. Biochem. 142,227-232. 22. Pierce Chemical Co., Rockford, IL, instruction manual. S., and Falardeau, P. (1977) Proc. Natl. Acad. Sci. 23. Hammarstrom, USA 74,3691-3695. 24. Murray, R., Fisher, M., Delorummer, P., and Sligar, S. (1985) in Metalloproteins (Harrison, P., Ed.), Vol. I, pp. 157-206, Macmillan & Co, London. 25. Ullrich, V., Haurand, M., and Hecker, M. (1989) in Cytochrome P-450: Biochemistry and Biophysics (Schuster, I., Ed.), pp. 41-48, Taylor & Francis, London/New York/Philadelphia. 26. Hanson, L. K., Eaton, W. A., Sligar, S. G., Gunsalus, I. C., Gouterman, M., and Connell, C. R. (1976) J. Amer. Chem. Sot. 98,26722674. 27. Hall, R. E., Tuan, W., and Venton, D. L. (1986) Biochem. J. 233, 637-641. 28. Inoue, M., Smith, W. L., and Dewitt, D. L. (1987) in Advances in Prostaglandin, Thromboxane, and Leukotriene Research (Samuelsson, B., Paoletti, R., Ramwell, P. W., Eds.), Vol. 17, pp. 2933, Raven Press, New York. 29. Merlie, J. P., Fagan, D., Mudd, J., and Needleman, P. J. (1988) J. Biol. Chem. 263,3550-3553.
OF BIOCHEMISTRY
AND
BIOPHYSICS
Vol. 280, No. 2, August 1, pp. 325-330, 1990
lmmunoaffinity Purification Thromboxane Synthase R. Niising, Faculty
S. Schneider-Voss,
of Human
and V. Ullrichi
of Biology, P.0. Box 5560, University
of Konstanz,
D-7750 Konstanz, Federal Republic of Germany
Received October 26,1989, and in revised form March 5,199O
A recently produced monoclonal antibody against human thromboxane synthase was used to purify the enzyme from platelets in a one-step procedure with good yields. The isolated protein exhibited a single band of about 58 kDa by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and contained one heme/mol. Although the visible spectrum of the oxidized enzyme displayed a peak at 418 nm like the previously isolated enzyme after dithionite reduction and CO addition, it shifted to 419 nm but not to 450 nm where only a small shoulder could be detected. Its catalytic activity was only l-5% of the previous preparations, but with the same K, of about 10 MM and a ratio of thromboxane B2: 12-hydroxyheptadecatrienoic acid of 1:l. Studies with EPR spectrometry and inhibitors confirmed that only a minor part of the enzyme was in its native heme-thiolate conformation, whereas the major part had been converted to the inactive P420 form by the elution procedure. The amino acid analysis revealed 46% hydrophobic residues. According to the sequence of 26 amino acids from the N-terminus and two tryptic peptides no homology to one of the cytochrome P450 monooxygenases, or to cyclooxygenase, or to prostacyclin synthase was detected. (c 1990 Academic Press, Inc.
Thromboxane A, (TxA,)’ has been implicated in various pathophysiological conditions as a proaggregatory and vasoconstricting mediator (1,2). Its biosynthesis involves arachidonate liberation, prostaglandin endoper’ To whom correspondence should be addressed. * Abbreviations used: SDS, sodium dodecyl sulfate; TX, thromboxane; HHT, 12.I-hydroxy-5,8,10-heptadecatrienoic acid; PAGE, polyacrylamide gel electrophoresis; HPETE, hydroperoxyeicosatetraenoic acid; CNBr, cyanogen bromide; Chaps, (3.[(3-cholamidopropyl)dimethylammonio]-l-propansulfonate; CO, carbon monoxide; UK37248, 4-(2.(lH-imidazol-l-yl)ethoxy)benzoic acid; OKY-1581, sodium (E)-3-(4-(3-pyridylmethyl)phenyl)-2-methylacrylate; MDA, malondialdehyde; PGH,, prostaglandin H,; DTT, dithiothreitol. Copyright co 1990 hy Academic Press, Inc. All rights of reproduction in any form reserved.
oxide formation, and isomerization to TxA2. The latter reaction which was first demonstrated in 1976 by Needleman et al. (3) in the microsomal fraction of human and horse platelets is catalyzed by Tx-synthase. Its activity was later found to be present in many other tissues such as lung, kidney, and spleen (4), as well as in macrophages (5) and lung fibroblasts (6). Several groups of investigators have attempted to purify Tx-synthase without much success. In 1985 Haurand and Ullrich (7) reported the purification of Tx-synthase with a molecular weight of 58.8 kDa for the monomeric enzyme. One heme per polypeptide chain was present and a close analogy to the group of cytochrome proteins was established by optical and EPR spectroscopy. The enzyme formed 12-L-hydroxy-5,8,10-heptadecatrienoic acid and malondialdehyde together with TxAz in a 1:l:l ratio. A mechanism explaining this product pattern has been proposed (8). Quantitative structure-activity relationships with inhibitors have confirmed this mechanism and have allowed us to propose a structure for the active site (9). In order to obtain more enzyme for further studies, we have attempted to replace our time-consuming and lowyield purification protocol by a more efficient procedure. Since monoclonal antibodies against this enzyme had been raised in our laboratory, we employed immunoadsorption for the isolation of Tx-synthase. The same approach had been previously undertaken by Shen et al, (10) for the enzyme from swine lung, but the protein obtained had different properties with regard to molecular activity, spectral characteristics, and molecular weight. The aim of the work described here was not only for a faster isolation procedure but also toward an explanation of these,discrepancies. MATERIALS
AND
METHODS
Materials. Culture media and supplements were purchased from Gibco Laboratories (Grand Island, NY) and fetal calf serum from Biochrom KG (Berlin, FRG). All materials for column chromatography, including protein A, were supplied by Pharmacia (Uppsala, Sweden). 325
326
NUSING.
SCHNEIDER-VOSS.
Nitrocellulose paper was purchased from Millipore (Bedford, MA). Enzyme-labeled antibodies were obtained from Jackson Immunoresearch Lab. (Avondale, U.S.A.). All solvents and other substances used were of analytical grade and obtained from E. Merck (Darmstadt, FRG), Roth (Karlsruhe, FRG), or Sigma. PGH,, HHT, and 5-, 12-, and 15-HPETE were biosynthesized according to published procedures (11, 12). OKY-1581 was a gift from ON0 Pharmaceutical Co., Ltd. (Osaka, Japan), and UK-37248 was kindly provided by Pfizer, Central Research (Sandwich, England). The monoclonal antibody Tue 300 described recently (13) was obtained from roller bottle cultures and purified by protein A chromatography according to Ey et al. (14). For immobilization 30 mg puriImmunoafinity chromatography. fied antibodies in 0.1 M NaHC03, 0.5 M NaCl, pH 8.3, were coupled to CNBr-activated Sepharose CL-4B according to the instructions of the supplier. For the precolumn, bovine IgG (200 mg) was equilibrated in 0.1 M NaHCO,, 0.5 M NaCl, pH 8.3. Cyanogen bromide activation of Sepharose CL-4B was performed according to the basic method outlined by Fuchs and Sela (15). After coupling, the column material was washed with 0.1 M glycine, pH 2.5, and equilibrated in 10 mM KzHPO,, 10% glycerol (v/v), 0.7% Chaps, pH 8.3. Microsomes of human platelets were prepared as described (7). Solubilization of Tx-synthase was achieved by homogenizing microsomal protein in 10 mM K,HPO,, 1 mM DTT, 1 mM EDTA, 10% (v/v) glycerol, 0.7% Chaps (solubilization buffer). After stirring on ice for 30 min, the mixture was centrifuged at 100,OOOgfor 90 min. The resulting supernatant was used directly for enzyme purification. Soluhilized human platelet microsomes were pumped at 7 ml/h over the lo-ml precolumn linked in series to the 5-ml column of Tue 300. Sepharose CL-4B. After loading onto the affinity column, the precolumn was removed and the antibody column was washed with the following buffers at 30 ml/h; solubilization buffer (30 ml); 50 mM K,HPO,, 10% glycerol, 1% Chaps, and 1 M NaCI, pH 7.4 (40 ml); 50 mM K,HPO,, 10% glycerol, and 0.1% Chaps (20 ml). Tx-synthase was then eluted with 0.1 M glycine-HCl, 10% glycerol, 0.1% Chaps, pH 2.7, or by conditions as indicated and the eluted fraction immediately neutralized by the addition of 1 M TrisHCl, pH 8.0. Optical spectroscopy. Optical difference spectra with Tx-synthase inhibitors were recorded with a Uvikon spectrophotometer between 500 and 300 nm with 500 ~1 affinity-purified Tx-synthase (120 pg) in both cuvettes. Optical absolute spectra were measured with 500 pl affinity purified Tx-synthase. For the CO-binding, the sample was gassed with CO for about 1 min, and then a few crystals of sodium dithionite were added to the sample cuvette. EPR spectroscopy. Enzyme-containing column material (500 ~1) was transferred to a quartz tube and frozen in liquid nitrogen. After the EPR spectrum was recorded, the sample was carefully thawed, 20 nmol UK-37248 was added, and a new EPR spectrum was recorded. Spectra were evaluated as described in (16). Enzyme activity was determined by Assay of Tx-synthase activity. measuring HHT formation at 234 nm versus time as described recently (17). The K, value was calculated from a Lineweaver-Burk plot. Polyacrylamide gel electrophoresis and immunoblotting of Txsynthase. Proteins were fractionated by 10% SDS-PAGE according to the method of Laemmli (18). The gel was silver-stained as described by Merril et al. (19). Mohilities were measured relative to the band of the tracking dye. For molecular weight estimation, the R, values of the protein markers were plotted using linear regression analysis. For immunoblotting, electrophoretic transfer of proteins from polyacrylamide gel to nitrocellulose paper was performed according to the procedure of Towbin et al. (20). Immunostaining was carried out as described (13). Amino acid and peptide sequence. Amino acid analysis was performed as described in (21). Tryptic peptides were separated by re-
AND
ULLRICH TABLE
I
Different Elution Conditions of ImmunoafIinity Chromatography
Condition pH 2.7 pH 12.0
4M
Urea
1 M NaSCN 1.5 M MgCl* 20% DMF 40% DMF
Protein be) 1.2 1.2 n. d”. n.d. n.d. n.d. 0.6
Activity (nmol/min/mg) 255 128 n.d. n.d. n.d. n.d. nd.
Note. After Tx-synthase was bound to the antibody column (1 ml bed vol), the column was washed as described in Fig. 1, eluted by the following indicated conditions, and dialyzed against 50 mM K2HP04, 10% glycerol, 0.1% Chaps, pH 7.4. The activity was assayed as described under Materials and Methods. ’ Not detected. versed-phase chromatography using a Vydac Cl8 column and an aqueous TFA/acetonitril gradient system. N-terminal sequence determination by automated Edman degradation was performed by a gasphase sequenator (Applied Biosystems, Type 470). Protein was estimated as described in (22) with bovine serum albumin as a standard.
RESULTS Previous investigations on the specificity of three monoclonal antibodies had revealed that only one, Tue 300, reacted with native Tx-synthase (13) and, hence, proved to be suitable for immunoaffinity purification. The enzyme activity was not influenced by monoclonal antibody binding (13). Therefore, we concluded that the epitope is independent of the active site. By the use of protein A-Sepharose, Tue 300 was able to precipitate the enzyme activity and the immunoprecipitate from lz51-labeled solubilized microsomes revealed a single band of 58 kDa. Larger amounts of monoclonal antibodies were obtained from roller bottle cell cultures purified on protein A-Sepharose chromatography and coupled to CNBr-activated Sepharose CL-4B. The column (5 ml bed vol) bound 30 mg of purified monoclonal antibody. One hundred sixty milliliters of solubilized human platelet microsomes was applied to a precolumn (10 ml bed vol, containing 180 mg bovine IgG coupled to CNBr-activated Sepharose CL-4B) and then directly to the antibody column. The precolumn was removed and the affinity column was washed as described under Materials and Methods. The tight binding of the enzyme to the monoclonal antibody Tue 300 allowed us to wash the column under very stringent conditions before final elution. The use of a precolumn to which unspecific bovine IgG was coupled made a partial purification of solubilized microsomes unnecessary. The column material was separated in aliquots of 1 ml bed vol and different elution conditions were tested (Table I).
IMMUNOAFFINITY
PURIFICATION
OF HUMAN
THROMBOXANE
327
SYNTHASE
MW WASH
,‘ASH I
II
v
v
(kD), 92-
ELUTION
Q
67-
451 31-
21-,
Id-1
,4--p--
A I
10
I
20 FRACTION
I
I
30
BO
NUMBER
of Tx-synthase on antibody affinity column. FIG. 1. Purification The column was washed first with 30 ml wash I buffer (50 mM KzHPOI, 10% glycerol, 0.7% Chaps, 1 mM DTT, 1 mM EDTA, pH 7.4) and then with 40 ml wash II buffer (50 mM K,HPO,, 10% glycerol, 1% Chaps, 1 M N&l, pH 7.4). The bound enzyme was eluted by 0.1 M glycine-HCl, pH 2.7, containing 10% glycerol and 0.1% Chaps into one-tenth volume of 1 M Tris-HCl, pH 8.0. The fractions were monitored at the wavelengths 280 and 405 nm.
The bound protein could be eluted only at extreme pH conditions such as pH 2.7 or pH 12.0 from the antibody column, with poorer recovery of activity for the latter. Other conditions such as 1.5 M MgC12, 20% dimethylformamide, or 1 M NaSCN failed to elute the enzyme. Forty percent dimethylformamide partly eluted the enzyme but totally destroyed the enzymatic activity. Figure 1 shows the elution profile of Tx-synthase from the antibody column, eluted with 0.1 M glycine-HCl, pH 2.7, containing 10% glycerol and 0.1% Chaps. In preliminary experiments, compared to Lubrol PX (7), Chaps improved separation and sharpened the peak of activity. These elution conditions were used throughout for immunoaffinity purification of Tx-synthase. The proteins applied were mostly eluted by the washing procedure while 80% of the applied enzyme, as determined by an immunometric assay, were retained. The purity of the enzyme from different stages of purification was assessed by SDS-PAGE and silver-staining, as shown in Fig. 2A. A single polypeptide having a molecular weight of 58 kDa was eluted from the antibody column. To verify the identity of the eluted protein as Tx-synthase, immunoblotting was carried out with solubilized platelet microsomes and the purified enzyme, probed with polyclonal antibody directed to conventionally purified human Tx-synthase. In each lane one band of 58 kDa, corresponding to thromboxane synthase (Fig. 2B), appeared. These results were strong evidence that the eluted protein was indeed Tx-synthase. A summary of the overall purification procedure is presented in Table II.
12
3
6
4
1
2
FIG. 2. (A) SDS-PAGE of column fractions from antibody column. Separation was performed in 10% slab gel. After electrophoresis, the gel was silver stained. Lane 1, solubilized microsomes (10 pg); lane 2, proteins not bound by the antibody column (10 Kg); lane 3, affinity purified enzyme (0.1 pg); lane 4, lysis buffer. (B) Immunoblotting of eluted protein. Enzyme samples were electrophoresed and transferred to nitrocellulose paper and incubated with polyclonal antibodies directed to Tx-synthase. Lane 1, solubilized microsomes (50 fig); lane 2, affinity-purified enzyme (50 ng).
Activity calculation from Table II reveals that immunopurified Tx-synthase has only about 3% of the original activity (7). However, the measured activity is nearly identical to the reported activity of immunopurified Txsynthase from swine lung by Shen et al. (10). In order to study the discrepancy between our earlier reported activity and the poor recovery of the immunopurified material we first used spectroscopic methods. The visible spectrum of affinity-purified Tx-synthase is shown in Fig. 3. The enzyme exhibited a Soret absorbance band at 418 nm. After gassing with CO and reduction with dithionite, the Soret band shifted to 419 nm with a light shoulder at 450 nm. This was in contrast to native Tx-synthase, which shifted completely to 450 nm after reduction in the presence of CO (7). The catalytic properties of the affinity-purified enzyme was further investigated by difference spectra with UK-37248. This compound binds selectively to the active site of Tx-synthase, causing a shift in visible spectra, which was well detected by difference spectroscopy (9). As shown in the inset of Fig. 3, the incubation of Tx-
TABLE
II
Purification of Thromboxane Synthase from Human Platelets
Purification
step
Microsomes Solubilized microsomes Affinity purified thromboxane synthase
Volume (ml)
Protein (mg)
80 82
1360 544
4
1.6
Specific activity (nmol/min/mg) 15.0 33.2 255.0
328
NUSING,
SCHNEIDER-VOSS.
AND
ULLRICH
015
s 6 D
01
:
i 9 005
MAGNETIC
0i
Loo
500
600
nm
FIG. 3. Visible spectrum of affinity-purified Tx-synthase. The spec120 +g of trum was monitored in 50 mM K,HPO,, pH 7.4,‘containing the enzyme against the blank buffer. Inset, difference spectrum of Txsynthase treated with UK-37248. Affinity-purified enzyme (120 pg) was present in the sample and reference cuvette. Tx-synthase inhibitor was added to the sample cuvette 5 min prior to recording.
synthase (120 pug) with UK-37248 yielded a difference spectrum with a minimum at 399 nm and a maximum at 427 nm. A rough calculation suggests that the activity could correlate with the light shoulder at 450 nm seen in the visible spectrum. No typical P450 signal could be seen (data not shown) from the EPR spectrum performed with about 10 nmol of enzyme. When, however, the immunoabsorbed fraction from the column, together with the Sepharose support, was recorded in the EPR spectrometer, the typical g tensor of cytochrome P450 with g values at 1,89, 2,25, and 2,45 appeared, although the g, and g, values were broadened (data not shown). Addition of the pyridinebased specific inhibitor UK-37248 sharpened the signal and resulted in the spectrum shown in Fig. 4, which verifies the protein nature as Tx-synthase, as shown earlier (7). The split signal at g = 2 does not belong to the P450 protein and probably indicates a partially denatured enzyme or a metal contamination. Due to the extensive washings, the material bound to the immuno-Sepharose matrix was about 95% pure Tx-synthase, as judged from the SDS-gel electrophoresis. The activity for HHT formation was inhibited in a dose-dependent manner by specific Tx-synthase inhibitors of the pyridine type as well as of the imidazole type. OKY-1581 inhibited the activity by 50% at 0.5 nM and UK-37248 was active at 1 nM. The K, for the enzyme with PGH, was 10 PM and the ratio of TXB, to HHT was determined as 0.9:1 by the incubation of Tx-Synthase with [14C]PGHZ, in agreement with our recent re-
FIELD
STRENGTH
IG1
FIG. 4. EPR spectra of immunoadsorbed Tx-synthase. Enzymecontaining column material, after extensive washing free of other proteins, was recorded after incubation with 20 nmol UK-37248 in an EPR spectrometer as described under Materials and Methods. The sample contained approximately 2 nmol of Tx-synthase.
ports (7, 8). The heat-denaturated enzyme did not produce HHT or TXB2. The activity of affinity-purified Tx-synthase was dramatically reduced by multiple additions of PGH, (Fig. 5). The inactivation effect was dependent on the substrate concentration used in relation to the amount of enzyme present. The initial control activity also increased with increasing PGH2 concentration but, with higher substrate levels, the fall in activity after multiple additions was more pronounced. This inactivation process was further investigated by the addition of 10 and 50 PM TXB,, HHT, or MDA together with PGHz. Neither of these enzyme products had an inhibitory effect on Tx-synthase (Table III). Hydroperoxy fatty acids have been reported to inhibit Tx-synthase in platelet microsomes, but the mechanism
2W QJM 20pM
t t t t Addition of mdlcated PGHz concentration
FIG. 5. Inactivation of Tx-synthase by PGH,. Tx-synthase activity was assayed in the presence of the indicated PGH, concentration and again after every 2 min (arrows) with the addition of fresh PGH, of the indicated concentration.
IMMUNOAFFINI’I’Y TABLE
PUHIFICA’l’IUN
VP HUMAN
III
activity
Concentration (PM)
TxB,
% of control
10 50 10 50 10 50 10 50
HHT MDA 5-HPETE 12.HPETE
10
15-HPETE
50 10 50
96.3 92.8 93.3 87.1 100.0 93.3 58.3 11.1 47.2 2.8 83.3 111.1
Note. Different metabolites in the indicated concentration were added simultaneously with PGH, to affinity-purified Tx-synthase and the activity was measured as described under Materials and Methods. Each value is the mean of duplicate assay.
of this inhibition was not established (23). As listed in Table III, 5- and 12-HPETE inhibited the enzyme in a dose-dependent manner after coincubation of Tx-synthase with PGHZ. 15HPETE had no inhibitory effect on Tx-synthase activity, even at high concentrations. Since the rapid purification protocol and the good yields made larger amounts of protein available, the amino acid composition of Tx-synthase could be investigated. This investigation revealed that the enzyme is a moderately hydrophobic protein containing approximately 46% hydrophobic residues (Table IV). Because the amino terminus was not blocked, we were able to sequence the N-terminus of Tx-synthase with a length of 26 amino acids (Table V). The sequences of two tryptic peptides are also presented. DISCUSSION Our attempts to purify human Tx-synthase by immunoaflinity were successful as far as yields and amounts TABLE
IV
Amino Acid Composition of Tx-Synthase Amino acid Asx Ser Pro Ala Val Ile Tyr His Arg
45 23 41 46 27 16 14 10 33
Amino acid Thr Glx Gly cys Met Leu Phe Lys
19 44 31 l-2 17 60 34 22
329
V
Partial Protein Sequences of Tx-Synthase NH,-terminus
Metaholite
b Y N 1 HA3b
TABLE
of Different Eicosanoid Metabolites on Tx-Synthase Activity
Effect
‘l’HKUMl5UKANr;
Peptides Peptide 1 Peptide 2
MettGlu-Ala-Leu-Gly-Phe-Leu-Lys-Leu-GluVal-Asn-Gly-Pro-Met-Val-Thr-Val-AlaLeu ?“-Val-Ala-Leu-Leu-Ala Ile-Lys-Gln-Val-Leu~Val-Glu-Asn-Phe-SerAsnPhe-Thr-Asn-Arg VallProoLeu-Ala-Arg-Ile-LeuPro-AsnLys Asn-Arg-Asp-Glu-Leu-Asn-Gly-Phe-Phe-AsnLys-Leu-Ile-Arg-Asn~Val-Ile-Ala-Leu-ArgAsp-Asn-Asn-Ala-Ala-Glu-Glu-Arg-Arg-?-GluPhe-Leu-Glu
DAmino acid not identified.
of enzyme were concerned. Whereas molecular weight, heme content, and antigenic properties were in agreement with those of the conventionally purified Tx-synthase, the molecular activity reached only a few percent. From our results we can conclude that this was a consequence of the rather rigid conditions of elution of the antibody-adsorbed enzyme since the EPR spectrum of the immunoadsorbed enzyme showed a typical P450 g tensor, whereas with the eluted enzyme, this spectrum which is typical for an intact Fe(III)-thiolate structure (24) had disappeared. In addition, the reduced carbon monoxide spectrum revealed most of the enzyme to be present in the P420 form with only a weak shoulder at 450 nm. We had previously confirmed that as in the cytochrome P450 monooxygenase systems, a conversion to the P420 form was paralleled by loss of Tx-synthase activity (25). In the immunopurified enzyme preparation, the inhibitors OKY-1581 and UK-37248 blocked the enzyme with the same high affinity as for the native enzyme. Little binding was observed spectrally, which agreed with only a small fraction of the hemoprotein reacting with the inhibitor. Also, the affinity for the substrate PGH, was unchanged. Thus, it seemed likely that we were dealing with a few percent of active protein and with a bulk of denatured enzyme unable to bind substrate or inhibitor. These results may also explain the data of Shen et al. (10) on their immunoaffinity purification of Tx-synthase from swine lung, which also indicated a pure but less active protein. Since similar elution conditions were employed, the enzyme was probably largely inactive and in a native state may have exhibited a similar high molecular activity as did the human enzyme from platelets. In contrast to the data by Shen et al. (lo), human Txsynthase exhibited a molecular weight of 58 kDa as did the enzyme from human lung and swine platelets, as verified by immunoblot technique (data not shown). The fact that the P420 form of Tx-synthase is devoid of activity supports our hypothesis that an intact iron(III)-thiolate bond is required for Tx-synthase ac-
330
NikING,
SCHNEIDER-VOSS,
tivity (7,26). A loss in heme content during affinity purification is unlikely because our measured ratio of 1:l for heme to protein is in agreement with earlier results (7) and the residual activity was not affected by exogenous heme (data not shown). As also reported by others (10,27), we observed a rapid self-inactivation of Tx-synthase by its substrate PGHz. This effect was concentration-dependent and was not mediated by the products TxB,, HHT, or malondialdehyde. Since 5-and 12-HPETE led to a similar inactivation, one may assume that oxidative processes are responsible. In the late stage of massive platelet aggregation, 12-HPETE may accumulate. Therefore, for platelets the inactivation of Tx-synthase may even be a physiological event. In our experiments, 15-HPETE was not effective in destruction, in contrast to reports by Shen et al. (10). Although our preparation did not turn out to be suitable for activity measurements, it made available relatively large quantities of protein which could be used for a protein chemical characterization. The amino acid composition after acid hydrolysis gave about 46% hydrophobic residues, which corresponds well with other microsomal proteins and also with the 44% of the closely related prostacyclin synthase (28). The low content of cysteine is remarkable and may indicate that only one cysteine is allowed to be present in the polypeptide. This would then have to be the heme-interacting cysteine at the active site. The N-terminus also contained a majority of hydrophobic residues but, as for the two presented peptides, no striking homology was found between Txsynthase and other cytochrome P450 proteins. We found homology neither with cyclooxygenase (29) nor with reported peptides of prostacyclin synthase (28). We expect that Tx-synthase is synthesized without a cleaved leader sequence, since the amino terminus is methionine. The high content of hydrophobic amino acids suggests a membrane spanning region of the presented sequence. Efforts for sequencing Tx-synthase with molecular biological methods are now being undertaken and will then allow us to conclusively compare Tx-synthase with related enzymes. ACKNOWLEDGMENTS We amino by Dr. ported
thank Dr. F. Lottspeich (MPI Munich, FRG) for performing the acid and sequence analyses. EPR spectra were kindly recorded P. Kroneck (University of Konstanz, FRG). The study was supby the Deutsche Forschungsgemeinschaft (SFB 156/A4).
REFERENCES 1. Needleman, P., Turk, J., Jakschik, B. A., Morrison, A. R., and Lefkowith, J. B. (1986) Annu. Reu. Biochem. 55,699102.
AND
ULLRICH
2. Granstrom, E., Diczfalusy, U., Hamberg, M., Hansson, G., Malmsten, C., and Samuelson, B. (1982) in Prostaglandins and the Cardiovascular System (Oates, J. A., Ed.), pp. 15-58, Raven Press, New York. 3. Needleman, P., Moncada, S., Bunting, S., Vane, J. R., Hamberg, M., and Samuelson, B. (1976) Nature (London) 261,558-560. 4. Hamberg, M. (1976) Biochim. Biophys. Acta 431,651-654. 5. Morley, J., Bray, M. A., Jones, R. W., Nugteren, D. H., and van Dorp, D. A. (1979) Prostaglandins 17,730-736. 6. Hopkins, N. K., Sun, F. F., and Gorman, R. R. (1978) Biochem. Biophys. Res. Commun. 85,827-836. 7. Haurand, M., and Ullrich, V. (1985) J. Biol. Chem. 260, 15,05915,067. 8. Hecker, M., and Ullrich, V. (1989) J. Biol. Chem. 264,141-150. 9. Hecker, M., Haurand, M., Ullrich, V., and Terao, S. (1986) Eur. J. Biochem. 157,217-223. 10. Shen, R. F., and Tai, H. H. (1986) J. Eiol. Chem. 261, 11,59211,599. A., and Ullrich, V. (1987) Biochem. 11. Hecker, M., Hatzelmann, Pharmacol. 36,851-855. 12. Porter, N. A., Logan, J., and Kontoyiannidou, V. (1979) J. Org. Chem. 44,3177-3181. 13. Niising, R., Wernet, M. P., and Ullrich, V. (1989) Submitted for publication. 14. Ey, P. L., Prouze, S. J., and Jenkin, C. R. (1978) Immunochemistry 15,429-436. Im15. Fuchs, S., and Sela, M. (1978) in Handbook of Experimental munology (Weir, D. M., Ed.), Vol. 10.1-10.6, Blackwell Scientific, Philadelphia, PA. 16. Marchesini, A., and Kroneck, P. (1979) Eur. J. Biochem. 101,6576. 17. Hecker, M., Haurand, M., Ullrich, V., Diczfalusy, U., and Hammarstrom, S. (1987) Arch. Biochem. Biophys. 254,124-135. 18. Laemmli, U. K. (1970) Nature (London) 227,680-685. 19. Merril, C. R., Goldman, D., Sedman, S. A., and Ebert, M. H. (1981) Science 2 11,1437-1438. 20. Towbin, H., Staehlin, T., and Gordon, J. (1979) Proc. Natl. Acad. Sci. USA 76,4350-4354. 21. Lottspeich, F., Kellermann, J., Henschen, A., Rauth, G., and Miiler-Esterl, W. (1984) Eur. J. Biochem. 142,227-232. 22. Pierce Chemical Co., Rockford, IL, instruction manual. S., and Falardeau, P. (1977) Proc. Natl. Acad. Sci. 23. Hammarstrom, USA 74,3691-3695. 24. Murray, R., Fisher, M., Delorummer, P., and Sligar, S. (1985) in Metalloproteins (Harrison, P., Ed.), Vol. I, pp. 157-206, Macmillan & Co, London. 25. Ullrich, V., Haurand, M., and Hecker, M. (1989) in Cytochrome P-450: Biochemistry and Biophysics (Schuster, I., Ed.), pp. 41-48, Taylor & Francis, London/New York/Philadelphia. 26. Hanson, L. K., Eaton, W. A., Sligar, S. G., Gunsalus, I. C., Gouterman, M., and Connell, C. R. (1976) J. Amer. Chem. Sot. 98,26722674. 27. Hall, R. E., Tuan, W., and Venton, D. L. (1986) Biochem. J. 233, 637-641. 28. Inoue, M., Smith, W. L., and Dewitt, D. L. (1987) in Advances in Prostaglandin, Thromboxane, and Leukotriene Research (Samuelsson, B., Paoletti, R., Ramwell, P. W., Eds.), Vol. 17, pp. 2933, Raven Press, New York. 29. Merlie, J. P., Fagan, D., Mudd, J., and Needleman, P. J. (1988) J. Biol. Chem. 263,3550-3553.
