- Email: [email protected]
Isolation of the cDNA for human prostaglandin H synthase
PROSTAGLANDINS
ISOLATION OF THE eDNA FOR HUMAN PROSTAGLANDIN H SYNTHASE ,
Timothy Hla, Michael Farrell, Ajit Kumar and J. Martyn Bailey Biochemistry Department, George Washington University School of Medicine and Health Sciences, Washington D.C. 20037
ABSTRACT Prostaglandin H Synthase (PGHS, cyclooxygenase) is a 67 kd protein which catalyzes the first step in prostaglandin synthesis. The primary amino acid sequence and the molecular mechanisms regulating expression are unknown. We report here isolation of a eDNA clone for the enzyme from human vascular endothelial cells for use in such studies. High titre, polyclonal antiserum against PGHS was developed in rabbits. The antiserum was monospecific, reacted with cyclooxygenase on Western blots at a limiting dilution of 1:500,000 and immunoprecipitated cyclooxygenase synthesized by in vitro translation of PGHS messenger RNA. It was used to screen a lambda gt11 eDNA expression library from human endothelial cells. Three positive clones were isolated. Following plaque purification, one clone reacted strongly with two other polyelonal antisera independently raised against highly purified cyclooxygenase and the aspirin-acetylated enzyme. Western blot analysis confirmed production of a l a r g e ~ 180 kd fusion protein of cyclooxygenase and beta-galactosidase. The eDNA insert of approximately 2.2 kilo base pairs was excised and subeloned into plasmid pUC8. A 24 nucleotide DNA probe, synthesized according to the amino acid sequence of the aspirin-aeetylation site of cyclooxygenase, hybridized strongly with the 2.2 kbp cDNA insert. It is concluded that the 2.2 kbp eDNA insert represents a eDNA clone for human cyclooxygenase, which also expresses the aspirin-acetylation site. This is the first reported isolation of the eDNA for this enzyme, and will facilitate further studies on the primary sequence and on the regulation of the enzyme at the molecular level. INTRODUCTION Prostaglandin H synthase (PGHS), E.C. 1.14.99.1, also known as cyclooxygenase, is the rate-limiting enzyme in the biosynthesis of active prostanoids, including the prostaglandins, thromboxanes and prostacyclins. In the cardiovascular system, prostacyclin and thromboxane are two important prostanoids, synthesized by vascular
* To whom correspondence and reprint requests should be addressed.
DECEMBER 1986 VOL. 32 NO. 6
829
PROSTAGLANDINS
wall (endothelial and smooth muscle cells) and blood platelets respectively. Prostacyclin (PGI 2) is the most potent inhibitor of platelet aggregation yet discovered and a vasodilator whereas thromboxane A 2 mediates the opposite effect (I). Refractoriness to prostanoid releasing stimuli can result from irreversible self-inactivation of cyclooxygenase. This phenomenon was first described by Smith and Lands using the isolated enzyme preparation (2). Subsequently it was confirmed that cyclooxygenase activity was also self-limited or 'down-regulated' in various other systems including human vascular endothelial cells (3), isolated aortic segments (4) and cultured smooth muscle cells (5). The mechanism is thought to involve inactivation of the apoprotein moiety of the enzyme by reactive free-radical species generated during catalysis (I). Non-steroidal anti-inflammatory drugs (NSAID) such as aspirin and indomethacin irreversibly inhibit PGHS (I) and induce an analogous refractory state. It was shown that aspirin acetylates an internal serine residue in the enzyme protein (7) and that both drugs cause a time-dependent conformational change, leading to permanent inactivation of enzymatic activity (6). Recovery of cyclooxygenase activity from these refractory states has been characterized in a number of cultured cell systems (8-11). In all cases recovery depended on synthesis of new protein since it was blocked by cycloheximide. The half-life of PGHS protein was found to be one of the shortest in vivo , consistent with the idea that regulatory enzymes have a high turnover rate (12). In addition, we (13) and others (11) have found that PGHS activity is stimulated by serum factors such as epidermal and platelet derived growth factors. The molecular mechanisms responsible for the regulation of expression this enzyme however, remain to be elucidated. Purification and characterization of PGHS protein has been successfully achieved by a number of groups (14-18). PGHS has been purified from various sources including ram seminal vesicles (RSV), bovine seminal vesicles and human lung cancer cell line Lu-65. The RSV enzyme, which has been most thoroughly characterized was shown to be a dimer of 67 kd subunits. The N-terminal 14 amino acids and a 22 amino acid stretch at the aspirin acetylation site have been sequenced (7). An isolated eDNA for PGHS would be an ideal probe to follow expression of PGHS at the mRNA level. In addition, it would provide, for the first time, primary sequence information for the enzyme, which might provide insights into its mechanism of catalysis. In this paper, we report the cloning of a eDNA for human endothelial cell cyclooxygenase using an ~m~,nological approach. METHODS Purification of PGHS Ram seminal vesicles (RSV) were used for purification of pure PGHS because of the high abundance of the enzyme. Modification of
830
DECEMBER 1986 VOL. 32 NO. 6
PROSTAGLANDINS
the procedure by van der Oderra (14) was used to purify the enzyme for antibody production. In brief, 40 g of frozen RSV (Pel Freez Biologicals, Arkansas) was homogenized in 40 ml of homogenizing buffer (O.05M Tris-HCl pH 8.0, 10 mM EDTA I mM diethyldithiocarbamate and 0.1% Tween-20). Post-mitochondrial supernatant was obtained by centrifugation in a Sorvall SS-34 rotor for 10 min at 13000 rpm. Microsomal pellet was obtained (130,000 x g, I hr) and washed in homogenizing buffer containing 0 . 1 M sodium perchlorate. Washed microsomal pellets were solubilized in homogenizing buffer + I% Tween-20. Insoluble material was pelleted and enzymically active PGHS was obtained in solubilized form. This fraction was then further purified on preparative 10% SDS-polyacrylamide gel electrophoresis (SDS-PAGE) using the buffer system of Lamelli (19). Preparation of Antiserum to Purified PGHS and Western Blotting The pure, denatured PGHS antigen was obtained following brief Coomassie staining and excising the band from preparative SDS-PAGE gels. It was then ground in an equal volume of 100 mM Tris 7.4 buffer and drawn into an 18 gauge needle. Following dilution with I volume of Freund's Complete adjuvant it was injected subcutaneously into 4-5 sites on the backs of 2 female New Zealand White Rabbits. Injection schedule was as follows: Day I, 200 ug of cyclooxygenase + Freund's Complete adjuvant; Day 14, 100 ug of cyclooxygenase + Freund's Incomplete adjuvant; Day 28, 100 ug of cyclooxygenase + Freund's Incomplete adjuvant. The rabbits were bled on day 40 via the ear vein and the serum was obtained. Rabbit antiserum to PGHS was characterized by Western Blotting as described by Towbin et al. (20). Proteins were separated on 10% SDS-PAGE as before and transferred onto nitrocellulose paper (0.45 micron) in a Biorad Electroblotting apparatus in 25 mM Tris/192 mM Glycine pH 8.3 buffer containing 20% methanol. Nitrocellulose paper was blocked with I% albumin solution and was reacted with the indicated dilution of anti-PGHS serum, 1:200 dilution of goat-anti-rabbit antiserum (Sigma) and 1:2000 dilution of rabbit peroxidase antiperoxidase (Sigma) in sequence. After washing the unbound antibodies, the filter was developed in Tris buffered saline solution containing 0.02% imidazole, 0.02% diaminobenzidine and 0.01% H202. Preparation of mRNA and In Vitro Translation RNA was prepared from RSV by the guanidinium isothiocyanate cesium chloride procedure (21). The mRNA fraction was purified by affinity chromatography on oligo-dT cellulose as described (22) and was translated in the message-depende~ rabbit reticulocyte in vitro translation system supplemented with ~ S-methionine (NEN, > 800 Ci/mmol) (23). The polypeptides obtained directly from the translation mix or as immunoprecipitates were resolved by 10% SDS-PAGE and the dried gel was exposed to Kodak X-ray film for autoradiography.
DECEMBER 1986 VOL. 32 NO. 6
831
PROSTAGLANDINS
Immunoprecipitation The reticulocyte lysates obtained as described above were diluted to 200 ul in 50 mM Tris buffered saline solution and incubated with 5 ul of test antisera for 2 hours. They were then immunoprecipitated by adding 10 ul of a 50% suspension of protein A sepharose (Pharmacia) and incubated overnight. Immunoprecipitates were washed in the same buffer 3 times and SDS-PAGE sample buffer was used to release the bound proteins, which were then analyzed by 10% SDS-PAGE. Molecular Biological Techniques A lambda gt11 cDNA library from human venous endothelial cells (HUVEC) was a gift from Drs. Maciag and Rioca of Meloy Laboratories, Inc., Rockville, MD. This library was prepared by reverse transcribing mRNA from HUVEC with dT19 Ig primer (24). Growth of lambda gt11, ~m~nosc@~efilng, lysogenization and plaque purification were performed as detailed (25), except that detection of antibodies was carried out by the indirect peroxidase antiperoxidase technique as described in the Western blotting section. Lambda DNA was isolated from the lysogenic E.Coli according to the polyethylene glycol precipitation method described in (26). The 24 nucleotide DNA probe, corresponding to the sequence of the aspirin-acetylation site, was synthesized in the Applied Biosystems Model 380B DNA synthesizer using phosphoramidite chemistry, as described in (33). Degeneracy in the nucleotide codons was minimized by the inosine substitution procedure described in Figure 6 and the Results section. Routine molecular biology procedures such as kinasing, restriction digestion, Southern blotting and subcloning were done as described by the general methods of Maniatis et al. (27). RESULTS Purification of PGHS and Antiserum Development RSV is one of the richest sources for PGHS and we used the microsomal preparations from this tissue to prepare the antigen. As described in the methods section, the 67 kd band from solubilized microsomal proteins on 10% SDS-PAGE was identified as PGHS based on the following criteria: I) Immunoprecipitatlon with 2 monoclonal antibodies raised against PGHS cyo-1 and cyo-5 (28) and 2) Rw value comparison with pure PGHS obtained from Oxford BiomediCal Inc. From each preparative SDS-PAGE, we obtained approximately 50 ug of PGHS protein by staining and cutting the gel sections corresponding to 67 kd PGHS band.
832
DECEMBER 1986 VOL. 32 NO. 6
PROSTAGLANDINS
The monospecificity of the antiserum was established by Western (Immuno) Blotting technique. Solubilized RSV microsomal proteins or pure PGHS standard was subjected to 10% SDS-PAGE and electroblotted onto nitrocellulose sheets. The nitrocellulose sheets were then reacted with indicated dilutions of antisera and immuno-stained by the indirect peroxldase method. As shown in Figure I, anti-PGHS antiserum reacted only with the 67 kd PGHS band from total microsomal proteins at the dilution of 1:500,000. It reacted with the same intensity with pure PGHS standard as well whereas non-immune rabbit serum did not give detectable staining. Thus, the antiserum preparation contained high concentration of antibodies directed at denatured epitopes of PGHS, suitable for screening recombinant fusion proteins in the lambda gt11 system.
~ ~
92 kd
~
~ ~.
~
66
~
45
~
2L "
I
2
3
4
5
6
7
8
25
9
i0
Figure I. Proteins were fractionated o n 10% SDS-PAGE and transferred onto a nitrocellulose paper. Subsequently they were stained for total protein by imido black or immunostained. Lane (1,10) molecular weight markers stained with imido black (2,3) anti-PGHS antiserum 1:500,000, antigen is same as lane 5 (4) control non-immune rabbit serum 1:500,000, antigen is same as lane 5 (5) authentic PGHS standard stained with imido black (6,7) anti-PGHS antiserum 1:500,000, antigen is same as lane 9 (8) control non-immune rabbit serum 1:500,000, antigen is same as lane 9 (9) RSV solubilized microsomal proteins stained with imido black.
DECEMBER 1986 VOL. 32 NO. 6
833
PROSTAGLANDINS
Immunoprecipitation of Nascent PGHS from In Vitro Translated Polypeptides To test the ability of anti-PGHS antiserum to precipitate PGHS in solution, we prepared mRNA from ram seminal vesicles and translated it in vitro in the presence of ~ S-methionlne in rabbit reticulocyte lysates. As shown in Figure 2, analysis of the product on 10% SDS-PAGE showed that anti-PGHS antiserum specifically precipitated a 67 kd band from 35 S-labelled peptides derived from translation of RSV mRNA fraction. The smaller molecular weight polypeptides are from non-speciflc adsorption to the immunopreclpitate since they are present in both the control non-immune serum lane 3 as well as anti-PGHS antiserum.
92 kd 66
45
25
t,
I
2
3
4
5
Figure 2. mRNA from RSV was translated ~ vitro in rabbit reticulocyte lysates in the presence of ~ S-methionine. The proteins were either analyzed directly or immunoprecipitated with test antiserum and protein A sepharose and analyzed. They were resolved on a 10% SDS-PAGE gel and autoradiography was performed. Lane (I) total proteins synthesized from RSV mRNA (2) immunoprecipitate of lane I with anti-PGHS antiserum (3) immunoprecipitate of lane I with non-immune rabbit serum (4) control translation mix without exogenous mRNA (5) positive control translation mix with brome mosaic virus mRNA.
834
DECEMBER 1986 VOL. 32 NO. 6
PROSTAGLANDINS
To determine the detection limit of anti-PGHS antiserum, different amounts of pure PGHS standard were spotted on nitrocellulose sheets in a fixed volume of I ul. They were then reacted with 1:100 dilution of anti-PGHS antiserum and immunostained as in Western Blots. The antiserum stained the PGHS protein up to a detection limit of 0.1 ng. This is below the amount of protein usually found in lambda gt11 plaques making this serum a feasible screening agent to isolate the eDNA clone for PGHS in the lambda gt11 system. Screening of Lambda gt11 cDNA Library from Human Endothelial Cells In the lambda gt11 system, the eDNA directed proteins are expressed in a fused form with E-Coli beta-galaetosidase protein when lambda gt11-infected plaques on E. Coli host YI090 are induced with isopropylthiogalactoside (IPTG). Using this system, we infected E. Coli host YI090 cells with lambda gt11 containing eDNA inserts from human endothelial cells to initiate a lytic cycle. After IPTG induction on nitrocellulose sheets, the proteins were detected with anti-PGHS antiserum (1:100 dilution). First round screening of approximately 300,000 plaques produced 19 positives which were further purified. Second and subsequent rounds of screening produced three immunoreactive clones which were then plaque purified. They were termed as lambda gtCox8, lambda gtCox18 and lambda gtCox21. Lambda gtCox18 gave the strongest signal, 21 being intermediate and 8 being the weakest in immunoreactivity (Figure 3). Lambda gtCox's were lysogenized into an E.Coli strain YI089 and the resulting lysogens were grown up to liter quantities at 32 C. Since YI089 contains a temperature sensltlve lambda repressor, temperature shlft to 42 C causes the lysogenlzed phage to initiate the lytic cycle of growth. This method allowed us to obtain large quantities of recombinant protein as well as DNA (25). Lysogenic E. Coli, containing lambda gtCox18 were grown up in the presence of IPTG to induce the production of fusion protein encoded by the eDNA. The proteins obtained were analyzed by Western blotting. A large fusion protein ( 4/180 kd) composed of beta-galactosidase and the eDNA encoded protein was detected by anti-PGHS antiserum only when induced by IPTG. This confirms that the immunological reactivity observed on the plaques was indeed from the expression of recombinant lambda gt11. O
.
.
•
.
O
DECEMBER 1986 VOL. 32 NO. 6
835
PROSTAGLANDINS
lambda gtCoxl8
lambda gtCox8
lambda gtCox21 f
Figure 3. Plaque purified lambda gtCox's were plated on LB am~icillin plates containing YI090 cells. Following phage growth at 42vC for 5 hours, nitrocellulose filters, pre-soaked in 10 mM IPTG were placed on top of the phage plaques and allowed to produce the recombinant protein overnight at 37°C. The presence of the protein was detected by immunostainlng with anti-PGHS antiserum similar to the procedure in the Western blots. 8, 18 and 21 are 3 different lambda gtCox clones. A, B and C represent 1:100 dilution of phage stocks.
836
DECEMBER 1986 VOL. 32 NO. 6
PROSTAGLANDINS
Cross-Reactivity of Lambda gtCox's with Different Anti-PGHS Antisera
We obtained two independently raised anti-PGHS antisera against the ram enzyme. One such antiserum from R. Bockman was a rabbit polyclonal antiserum against enzymatically active PGHS purified by conventional column chromatographic techniques (18). Another antiserum from G. Roth, also against the ram PGHS~ was prepared by a different method which entail~d acetylation with ~H-labelled aspirin and isolation of the ~H-labelled protein to be used as an antigen (29). These antisera were teated against the lambda gtCox clones at indicated dilutions. As shown in Figure 4, lambda gtCox18 reacted strongly with all three anti-PGHS antisera. This clone was therefore selected for further characterization.
8-3 8-4
8-2 18-1 8-1 18-4
18-2 18-3
21-3
21-2
21-4~,
21-1
Figure 4. Lambda gtCox's were plated on YI090 cells and Jmm)nostained with different anti-PGHS serum as in Figure 3. 8, 18 and 21 represent 3 different lambda gtCox clones. Antiserum I is obtained from R. Bockman and is directed against pure PGHS. Antiserum 2 from G. Roth is directed against pure aspirinated PGHS. Antiserum 3 is control non-immune rabbit serum and antiserum 4 is anti-PGHS antiserum prepared by us. All antlsera were used at the dilution of 1:100 except for #2, which was at 1:200 dilution.
D E C E M B E R 1986 VOL. 32 NO. 6
837
PROSTAGLANDINS
Analysis of cDNA Lysogenized lambda gtCox18 was used to prepare large quantities of DNA. The lambda gtCox18 DNA, when out with the restriction enzyme EcoRI produced a cDNA insert of approximately 2.2 kbp as shown in Figure 5. This insert was then subcloned into the EcoRI site of plasmid pUC 8 yielding a recombinant plasmid called pCoxTH18. This plasmid contains 2.7 kbp sequences from pUC 8 and 2.2 kbp from the insert.
23.1 9.4 6.5 2.3 2.0
I
I
2
Figure 5. DNA from lambda gtcoxa8 was prepared as described. It was then cut with EcoRI and was subjected to I% agarose gel electrophoresis in Tris-Acetate EDTA (TAE) buffer. The gel was stained with ethidium bromide. Lanes: (I) lambda DNA Hind III cut molecular weight markers (2) lambda gtCox18 EcoRI cut DNA.
838
DECEMBER 1986 VOL. 32 NO. 6
PROSTAGLANDINS
Design of a Synthetic DNA Probe Specific for the Aspirin-Acetylation Site of Cyclooxygenase The N-terminal 14 amino acids and a more internal 22 amino acid stretch around the anpirin-acetylation site have been sequenced (7). A synthetic 24-nucleotide DNA probe corresponding to an eight amino-acid sequence in the aspirin-acetylation site was synthesized using an Applied Systems 380B DNA synthesizer. In doing so, all 3 or 4 fold degenerate eodons (Figure 6) were replaced with inosine (30) to reduce the overall degeneracy and the anti-sense probe was synthesized as shown. The p ~ b e is a 24~mer, in which the degeneracy is reduced from 2 -~ to only 2- by the inosine substitution technique and is designated PGHS2.
PGHS2 Asn-Pro-Ile-Glu-Ser-Pro-Glu-Tyr 5'-AAU-CCU-AUU-GAA-AGU-CCU-GAA-UAU-3' C C C G C C G C A A A G G 5'-ATA-TTC-IGG-ACT-TTC-IAT-IGG-ATT-3' G C G C G
Figure 6. The 8 amino acid stretch of PGHS from the aspirin-acetylation site of PGHS was used to design the DNA probe. All possible codons as indicated were made less degenerate by replacing with inosine. The antisense probe (bottom) was synthesized as shown. 39 The synthetic DNA probe was labelled at its 5' ~gd with -P-phosphate using polynucleotide kinase and gamma-~-P-ATP.
DECEMBER 1986 VOL. 32 NO. 6
839
PROSTAGLANDINS
It was then hybridized on a Southern blot to the DNA from pCoxTH18 cut with several diagnostic restriction enzymes. As shown in Figure 7, restriction enzymes BamHI and Sstl each cut once in the insert region of pCoxTH18 whereas no restriction sites for any of the other enzymes were contained in the insert. ~ e n the same gel was blotted on nitrocellulose and hybridized with J~P-labelled PGHS2, strong hybridization with the intact plasmid DNA was observed. In addition BamHI cut out a 1.5 kbp fragment from 5' region of the insert which also hybridized strongly and left a 3.4 kbp fragment containing 2.7 kbp pUC 8 vector sequences and 0.7 kbp 3' region sequences from the cDNA insert. The B.4 kbp fragment did not give appreciable hybridization. These results indicate that the cloned 2.2 kbp cDNA represents the cDNA for cyclooxygenase and that the aspirin-acetylation site is located in the 1.5 kbp 5' moiety.
,,irt~
x
~
n
-I- m
0
co
EtBr stained 2.2 Kb
~mMW~ ~
5.0 4.0 3.0 20
~ ~Q
32 -PGHS2
16
hybrldlsed ~ =
-122 Kb
f _
-
i
.
-
-
5.0 3 . 0
.1~ 1.6
Figure 7. pCoxTH18 DNA was prepared and digested with a battery of restriction enzymes. The DNA was then subjected to I% ~ a r o s e gel electrophoresis in TAE buffer and stained with ethidium bromide to locate the bands. It was then blotted o n ~ a sheet of nitrocellulose paper and hybridized with ~-P-labelled PGHS2. Hybridization was done at room temperature overnight. It was then washed in 0.2 x SSC buffer at room temperature to remove unbound probe and exposed for autoradiography.
840
DECEMBER 1986 VOL. 32 NO. 6
PROSTAGLANDINS
DISCUSSION Modification of an established procedure for purification of PGHS on polyacrylamide gel electrophoresis was used to prepare large amounts of enzyme protein. Denatured proteins from SDS-PAGE are a good source of antigen for library screening purposes since the antigenic epitopes obtained from a lambda gt11 expression library will probably differ from the conformation of the native enzyme. Also polyacrylamide is known to be a good adjuvant. As shown by Western Blotting assay, the antiserum obtained was specific only for the 67 kd PGHS band and in addition, reacted with enzymatically active, pure PGHS from Oxford Biomedical, Inc., at the same signal intensity. Ram seminal vesicles contain high levels of PGHS and high levels of the mRNA for the enzyme so that the 67 kd band constituted a major component in the total protein profile obtained from RSV mRNA translation (Figure 2). The PGHS antiserum selectively ~ u n o p r e c i p i t a t e d the 67 kd band from the in vitro translated, S-methionine labelled, nascent proteins d~rec-~edby the mRNA fraction from ram seminal vesicles. This further confirms the selectivity of the antiserum preparation. The PGHS antiserum had a detection limit of 0.1 ng by our peroxidase detection method. This indicated that it could be used to screen a lambda gt11 expression vector library since more than 0.1 ng of protein can be synthesized by the phage plaques in this system (25). We used a eDNA library from human endothelial cells in lambda gt11 to isolate PGHS clones with the PGHS antiserum. These cells produce significant quantities of prcstanoids (8) and the enzyme protein itself has been localized by immunofluorescent techniques
(32). Screening with anti-PGHS antiserum to a library of 300,000 recombinant phage yielded 3 positive clones which were then plaque purified. Lysates containing the fusion proteins directed by these phage were analyzed by Western blotting with the anti-PGHS antiserum. Immuno-reactive, large ( ~ 180 kd) proteins were found only when induced with the lac-operon inducer IPTG, indicating that the immunoreactivity observed in the plaques was eDNA directed. Of the three positive clones, lambda gtCox18 reacted strongly with two other independently raised anti-PGHS antiserum and was therefore selected for further characterization whereas the other two did not react appreciably. The DNA isolated from lambda gtCox18 was cut with EcoRI, which cleaved-off the eDNA insert. It was approximately 2.2 kbp indicating that it could code for a protein containing about 700 amino acids. Since cyclooxygenase contains approximately 600 amino acids, the 2.2 kbp DNA could potentially code for the full length PGHS. However, since we have no information on the extent of the 3' and 5' untranslated regions of the cyclooxygenase messenger RNA, and if the 3' sequences were large, then the eDNA obtained would not be of full length. No PGHS enzymatic activity was detected in lysates of lambda gtCox18 from E. Coli which is not unexpected because the
DECEMBER 1986 VOL. 32 NO. 6
841
PROSTAGLANDINS
cyclooxygenase sequences are expressed as part of the large aJ 180 kd fusion protein with beta-galactosidase. The eDNA insert from lambda gtCox18 was subcloned into the EcoRI site of plasmid pUC 8 yielding the recombinant plasmid pCoxTH18. From the known restriction information about the pUC 8 vector, several restriction enzymes were tested on the DNA from pCoxTH18 to identify restriction sites in the insert. There are no restriction sites in pUC 8 for BstEII, EcoRV, XhoI, SstI and ClaI, whereas PstI, HindIII, BamHI and SalI have unique sites. The BamHI site is 6bp downstream from the EeoRI site. Thus BamHI cut pCoxTH18 and produced a large 5' insert fragment of 1.5 kbp and a 3.4 kbp ~ a g m e n t of fused vector and 3' insert sequences. The P-labelled synthetic DNA probe synthesized according to the aspirin-acetylation site of PGHS hybridized strongly to the intact plasmid DNA and to the large BamHI cut 5' insert fragment of 1.5 kbp, but not to the vector and 3' sequences. Based upon the antisera reactivities and the nucleic acid hybridization data, the results indicate that the isolated 2.2 kbp DNA represents the eDNA for human Prostaglandin H Synthase and that the aspirin acetylation site in the enzyme is associated with the 1.5 kbp 5' moiety of the eDNA. ACKNOWLEDGEMENTS We thank Drs. Macing and Rices of Meloy Laboratories, Inc., Rockville, MD for the human endothelial cell eDNA library and Dr. R. Bockman of Memorial Sloan-Kettering Cancer Centre, N.Y. and Dr. Gerald Roth for two independently prepared antisera against PGHS. We are grateful to Mr. M. Yadven and Mr. C. Siegall for assistance in making the synthetic DNA probe and to Miss G. Tillisch for manuscript preparation. KEY WORDS Cyclooxygenase, Prostaglandin Synthase, eDNA Cloning, Human Endothelial Cells REFERENCES
I.
Pace-Aciak, C.R. and W.L. Smith. In:The Enzymes Vol. 16 (P. Boyer, ed.) Academic Press, N.Y., 1983. p. 543.
2.
Smith, W.L. and W.E.M. Lands. Oxygenation of polyunsaturated fatty acids during prostaglandin biosynthesis by sheep vesicular gland. Biochemistry1_! :3276. 1972.
.
842
Brotherton, A. and J.C. Hoak. Prostacyclin synthesis in human vascular endothelial cells is limited by irreversible deactivation of cyclooxygenase. J. Clin. Invest. 72:1255. 1983.
DECEMBER 1986 VOL. 32 NO. 6
PROSTAGLANDINS
4.
Goldsmith, d.C., C.T. Jafvert, P. Dollar, W.G. Owen and J.C. Hoak. Prostacyclin release from cultured and ex vivo bovine vascular endothelium. Lab. Invest. 45:197. 1980.
5.
Bailey, J.M., B. Muza, T. Hla and J. Pash. Role of epidermal growth factor in cyclooxygenase synthesis. In:Adv. PG, TX. and LT. Res. Vol. 15. (0. Hayaishi and S. Yamamoto, eds.) Raven Press, N.Y., 1985. p. 141.
. Kulmacz, R.J. and W.E.M. Lands. Stoichiometry and kinetics of the interaction of PG H synthase with anti-inflammatory agents. J. Biol. Chem. 260:12572. 1985. 7.
Roth, G.J., E.T. Machuga and J. Ozols. Isolation and covalent structure of the aspirin-modified, active-site region of prostaglandin synthetase. Biochemistry 22:4672. 1983.
8,
Jaffe, E.A. and B.B. Weksler. Recovery of endothelial cell prostacyclin production after inhibiton by low doses of aspirin. J. Clin. Invest. 63:532. 1979.
. Whiting, J., K. Salata and J.M. Bailey. Aspirin: an unexpected side effect on prostacyclin synthesis of cultured vascular smooth muscle cells. Science 210:663. 1980. 10.
Baenziger, N.L., P.R. Becherer and P.W. Majerus. Characterization of prostacyclin synthesis in cultured human arterial smooth muscle cells, venous endothelial cells and skin fibroblasts. Cell 16:967. 1979.
11.
Habernicht, A.S.R., M. Goerig, J. Grulich, D. Rothe, R. Gronwald, V. Loth, G. Schettler, B. Kommerell, R. Ross. PDGF stimulates PG synthesis by rapid and denovo induction of cyclooxygenase. J. Clin. Invest. 75:1381. 1985.
12.
Fagan, J.M. and A.L. Goldberg. Inhibitors of protein and RNA synthesis cause a rapid block in prostaElandin production at the prostaglandin synthase step. Proc. Nat'l. Acad. Sci., U.S.A. 83:2771. 1986.
13.
Bailey, J.M., B. Muza, T. Hla and K Salata. Restoration of prostacyclin synthase in cultured vascular smooth muscle cells after aspirin treatment: regulation by epidermal growth factor. J. Lipid Res. 26:54. 1985.
14.
van der Ouderra, F.J., M. Buytenhek, D.H. Nugteren and D.A. van Dorp. Purification and characterization of prostaglandin endoperoxide synthase from sheep vesicular glands. Biochim. Biophys. Acta. 487:315. 1977.
D E C E M B E R 1986 VOL. 32 NO. 6
843
PROSTAGLANDINS
15.
Miyamotc, T., N. Ogino, S. Yamamoto and O. Hayaishi. Purification of prcstaglandin endcpercxide synthase from bovine vesicular gland miercsomes. J. Biol. Chem. 251:2629. 1976.
16.
Hemler, M., W.E.M. Lands and W.L. Smith. Purification of the cyclooxygenase that forms prostaglandins. J. Biol. Chem. 251 :5575. 1976.
17.
Roth, G.J., C.J. Sick and J. Ozols. Structural characteristics of prostaglandin synthetase from sheep vesicular gland. J. Biol. Chem. 255:1301. 1980.
18.
Hiekok, N.J., G. Chin and R.S. Beckman. Characterization of prcstaglandin endoperoxide synthase from a human cell line. Bioehim. Bicphys. Aeta 877:79. 1986.
19.
Laemmli, U.K. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680. 1970.
20.
Towbin, H., T. Stachelin and J. Gordon. Electrophoretie transfer of proteins from SDS-PAGE to nitrocellulose filters for detection with antibodies. Prec. Nat'l. Aead. Sci. U.S.A. 76:4350. 1979.
21.
Chirgwin, J.M., A.E. Przybyla, R.T. McDonald and W.J. Rutter. Isolation of RNA from sources enriched in ribcnuclease. Biochemistry 18:5294. 1979.
22.
Aviv, H. and P. Leder. Purification of mRNA by affinity chromatography on olige-dT cellulose. Prec. Nat'l. Acad. Sei. U.S.A. 69:1408. 1972.
23.
Pelham, H.R. and R.E. Jackson. Translation of mRNA in rabbit reticulocyte lysates. Eur. J. Biochem. 67:247. 1976.
24.
Tong, B.D., S. Levine, M. Jaye, G. Rieca, W. Drohan, T. Maciag and T. Deuel. Isolation and sequencing of a eDNA clone homologous to the v-sis oncogene from human endothelial cell. Mol. Cell Biol. 6:3018. 1986.
25.
Huynh, T.V., R. Young and R. Davis. In:DNA cloning: a practical approach. (D.M. Glever, ed.) I.R.L. Press, Oxford., 1985. p. 49.
26.
Kaslow, D. A rapid method for obtaining 14:6767. 1986.
27.
Maniatis, T., E.F. Fritsch and J. Samhrook. In:Molecular cloning: a laboratory manual. Cold Spring Harbor Press, N.Y., 1982.
844
DNA. Nue. Acid. Res.
DECEMBER 1986 VOL. 32 NO. 6
PROSTAGLANDINS
28.
Dewitt, D.L., T.M. Rollins, J.S. Day, J.D. Gauger and W.L Smith. Orientation of active site and antigenic determinants of PG synthase in the endoplasmic reticulum. J. Biol. Chem. 256 :10375. 1981.
29.
Roth, G. Radioimmunoassay for PG H synthase. In:Methods in Enzymology. Vol. 86. Academic Press, N.Y., 1982. p. 222.
3O.
Ohtsuka, E., S. Matsuki, M. Ikehara, Y. Takahashi and K. Matsobara. An alternative approach to deoxyribonucleotides as hybridization probes by insertion of deoxyinosine at ambiguous codon positions. J. Biol. Chem. 260:2605. 1985.
31.
Williams-Ashman, H.G. Regulatory features of seminal vesicle development and function. Curr. Top. Cell Reg. Vol. 22. Academic Press, N.Y., 1983. P. 201.
32.
Dewitt, D.L., J.S. Day, W.K. Sonnenburg and W.L. Smith. Concentrations of PG H synthase and PGI 2 synthase in the endothelium and smooth muscle. J. Clin. Invest. 72:1882. 1983.
33.
Beaucage, S . L . and M.H. C a r u t h e r s .
Editor:
P.W. Ramwell
Tet. Lett. 22:1859.
Received:
DECEMBER 1986 VOL. 32 NO. 6
11-3-86
1981.
Accepted: 11-4-86
845
ISOLATION OF THE eDNA FOR HUMAN PROSTAGLANDIN H SYNTHASE ,
Timothy Hla, Michael Farrell, Ajit Kumar and J. Martyn Bailey Biochemistry Department, George Washington University School of Medicine and Health Sciences, Washington D.C. 20037
ABSTRACT Prostaglandin H Synthase (PGHS, cyclooxygenase) is a 67 kd protein which catalyzes the first step in prostaglandin synthesis. The primary amino acid sequence and the molecular mechanisms regulating expression are unknown. We report here isolation of a eDNA clone for the enzyme from human vascular endothelial cells for use in such studies. High titre, polyclonal antiserum against PGHS was developed in rabbits. The antiserum was monospecific, reacted with cyclooxygenase on Western blots at a limiting dilution of 1:500,000 and immunoprecipitated cyclooxygenase synthesized by in vitro translation of PGHS messenger RNA. It was used to screen a lambda gt11 eDNA expression library from human endothelial cells. Three positive clones were isolated. Following plaque purification, one clone reacted strongly with two other polyelonal antisera independently raised against highly purified cyclooxygenase and the aspirin-acetylated enzyme. Western blot analysis confirmed production of a l a r g e ~ 180 kd fusion protein of cyclooxygenase and beta-galactosidase. The eDNA insert of approximately 2.2 kilo base pairs was excised and subeloned into plasmid pUC8. A 24 nucleotide DNA probe, synthesized according to the amino acid sequence of the aspirin-aeetylation site of cyclooxygenase, hybridized strongly with the 2.2 kbp cDNA insert. It is concluded that the 2.2 kbp eDNA insert represents a eDNA clone for human cyclooxygenase, which also expresses the aspirin-acetylation site. This is the first reported isolation of the eDNA for this enzyme, and will facilitate further studies on the primary sequence and on the regulation of the enzyme at the molecular level. INTRODUCTION Prostaglandin H synthase (PGHS), E.C. 1.14.99.1, also known as cyclooxygenase, is the rate-limiting enzyme in the biosynthesis of active prostanoids, including the prostaglandins, thromboxanes and prostacyclins. In the cardiovascular system, prostacyclin and thromboxane are two important prostanoids, synthesized by vascular
* To whom correspondence and reprint requests should be addressed.
DECEMBER 1986 VOL. 32 NO. 6
829
PROSTAGLANDINS
wall (endothelial and smooth muscle cells) and blood platelets respectively. Prostacyclin (PGI 2) is the most potent inhibitor of platelet aggregation yet discovered and a vasodilator whereas thromboxane A 2 mediates the opposite effect (I). Refractoriness to prostanoid releasing stimuli can result from irreversible self-inactivation of cyclooxygenase. This phenomenon was first described by Smith and Lands using the isolated enzyme preparation (2). Subsequently it was confirmed that cyclooxygenase activity was also self-limited or 'down-regulated' in various other systems including human vascular endothelial cells (3), isolated aortic segments (4) and cultured smooth muscle cells (5). The mechanism is thought to involve inactivation of the apoprotein moiety of the enzyme by reactive free-radical species generated during catalysis (I). Non-steroidal anti-inflammatory drugs (NSAID) such as aspirin and indomethacin irreversibly inhibit PGHS (I) and induce an analogous refractory state. It was shown that aspirin acetylates an internal serine residue in the enzyme protein (7) and that both drugs cause a time-dependent conformational change, leading to permanent inactivation of enzymatic activity (6). Recovery of cyclooxygenase activity from these refractory states has been characterized in a number of cultured cell systems (8-11). In all cases recovery depended on synthesis of new protein since it was blocked by cycloheximide. The half-life of PGHS protein was found to be one of the shortest in vivo , consistent with the idea that regulatory enzymes have a high turnover rate (12). In addition, we (13) and others (11) have found that PGHS activity is stimulated by serum factors such as epidermal and platelet derived growth factors. The molecular mechanisms responsible for the regulation of expression this enzyme however, remain to be elucidated. Purification and characterization of PGHS protein has been successfully achieved by a number of groups (14-18). PGHS has been purified from various sources including ram seminal vesicles (RSV), bovine seminal vesicles and human lung cancer cell line Lu-65. The RSV enzyme, which has been most thoroughly characterized was shown to be a dimer of 67 kd subunits. The N-terminal 14 amino acids and a 22 amino acid stretch at the aspirin acetylation site have been sequenced (7). An isolated eDNA for PGHS would be an ideal probe to follow expression of PGHS at the mRNA level. In addition, it would provide, for the first time, primary sequence information for the enzyme, which might provide insights into its mechanism of catalysis. In this paper, we report the cloning of a eDNA for human endothelial cell cyclooxygenase using an ~m~,nological approach. METHODS Purification of PGHS Ram seminal vesicles (RSV) were used for purification of pure PGHS because of the high abundance of the enzyme. Modification of
830
DECEMBER 1986 VOL. 32 NO. 6
PROSTAGLANDINS
the procedure by van der Oderra (14) was used to purify the enzyme for antibody production. In brief, 40 g of frozen RSV (Pel Freez Biologicals, Arkansas) was homogenized in 40 ml of homogenizing buffer (O.05M Tris-HCl pH 8.0, 10 mM EDTA I mM diethyldithiocarbamate and 0.1% Tween-20). Post-mitochondrial supernatant was obtained by centrifugation in a Sorvall SS-34 rotor for 10 min at 13000 rpm. Microsomal pellet was obtained (130,000 x g, I hr) and washed in homogenizing buffer containing 0 . 1 M sodium perchlorate. Washed microsomal pellets were solubilized in homogenizing buffer + I% Tween-20. Insoluble material was pelleted and enzymically active PGHS was obtained in solubilized form. This fraction was then further purified on preparative 10% SDS-polyacrylamide gel electrophoresis (SDS-PAGE) using the buffer system of Lamelli (19). Preparation of Antiserum to Purified PGHS and Western Blotting The pure, denatured PGHS antigen was obtained following brief Coomassie staining and excising the band from preparative SDS-PAGE gels. It was then ground in an equal volume of 100 mM Tris 7.4 buffer and drawn into an 18 gauge needle. Following dilution with I volume of Freund's Complete adjuvant it was injected subcutaneously into 4-5 sites on the backs of 2 female New Zealand White Rabbits. Injection schedule was as follows: Day I, 200 ug of cyclooxygenase + Freund's Complete adjuvant; Day 14, 100 ug of cyclooxygenase + Freund's Incomplete adjuvant; Day 28, 100 ug of cyclooxygenase + Freund's Incomplete adjuvant. The rabbits were bled on day 40 via the ear vein and the serum was obtained. Rabbit antiserum to PGHS was characterized by Western Blotting as described by Towbin et al. (20). Proteins were separated on 10% SDS-PAGE as before and transferred onto nitrocellulose paper (0.45 micron) in a Biorad Electroblotting apparatus in 25 mM Tris/192 mM Glycine pH 8.3 buffer containing 20% methanol. Nitrocellulose paper was blocked with I% albumin solution and was reacted with the indicated dilution of anti-PGHS serum, 1:200 dilution of goat-anti-rabbit antiserum (Sigma) and 1:2000 dilution of rabbit peroxidase antiperoxidase (Sigma) in sequence. After washing the unbound antibodies, the filter was developed in Tris buffered saline solution containing 0.02% imidazole, 0.02% diaminobenzidine and 0.01% H202. Preparation of mRNA and In Vitro Translation RNA was prepared from RSV by the guanidinium isothiocyanate cesium chloride procedure (21). The mRNA fraction was purified by affinity chromatography on oligo-dT cellulose as described (22) and was translated in the message-depende~ rabbit reticulocyte in vitro translation system supplemented with ~ S-methionine (NEN, > 800 Ci/mmol) (23). The polypeptides obtained directly from the translation mix or as immunoprecipitates were resolved by 10% SDS-PAGE and the dried gel was exposed to Kodak X-ray film for autoradiography.
DECEMBER 1986 VOL. 32 NO. 6
831
PROSTAGLANDINS
Immunoprecipitation The reticulocyte lysates obtained as described above were diluted to 200 ul in 50 mM Tris buffered saline solution and incubated with 5 ul of test antisera for 2 hours. They were then immunoprecipitated by adding 10 ul of a 50% suspension of protein A sepharose (Pharmacia) and incubated overnight. Immunoprecipitates were washed in the same buffer 3 times and SDS-PAGE sample buffer was used to release the bound proteins, which were then analyzed by 10% SDS-PAGE. Molecular Biological Techniques A lambda gt11 cDNA library from human venous endothelial cells (HUVEC) was a gift from Drs. Maciag and Rioca of Meloy Laboratories, Inc., Rockville, MD. This library was prepared by reverse transcribing mRNA from HUVEC with dT19 Ig primer (24). Growth of lambda gt11, ~m~nosc@~efilng, lysogenization and plaque purification were performed as detailed (25), except that detection of antibodies was carried out by the indirect peroxidase antiperoxidase technique as described in the Western blotting section. Lambda DNA was isolated from the lysogenic E.Coli according to the polyethylene glycol precipitation method described in (26). The 24 nucleotide DNA probe, corresponding to the sequence of the aspirin-acetylation site, was synthesized in the Applied Biosystems Model 380B DNA synthesizer using phosphoramidite chemistry, as described in (33). Degeneracy in the nucleotide codons was minimized by the inosine substitution procedure described in Figure 6 and the Results section. Routine molecular biology procedures such as kinasing, restriction digestion, Southern blotting and subcloning were done as described by the general methods of Maniatis et al. (27). RESULTS Purification of PGHS and Antiserum Development RSV is one of the richest sources for PGHS and we used the microsomal preparations from this tissue to prepare the antigen. As described in the methods section, the 67 kd band from solubilized microsomal proteins on 10% SDS-PAGE was identified as PGHS based on the following criteria: I) Immunoprecipitatlon with 2 monoclonal antibodies raised against PGHS cyo-1 and cyo-5 (28) and 2) Rw value comparison with pure PGHS obtained from Oxford BiomediCal Inc. From each preparative SDS-PAGE, we obtained approximately 50 ug of PGHS protein by staining and cutting the gel sections corresponding to 67 kd PGHS band.
832
DECEMBER 1986 VOL. 32 NO. 6
PROSTAGLANDINS
The monospecificity of the antiserum was established by Western (Immuno) Blotting technique. Solubilized RSV microsomal proteins or pure PGHS standard was subjected to 10% SDS-PAGE and electroblotted onto nitrocellulose sheets. The nitrocellulose sheets were then reacted with indicated dilutions of antisera and immuno-stained by the indirect peroxldase method. As shown in Figure I, anti-PGHS antiserum reacted only with the 67 kd PGHS band from total microsomal proteins at the dilution of 1:500,000. It reacted with the same intensity with pure PGHS standard as well whereas non-immune rabbit serum did not give detectable staining. Thus, the antiserum preparation contained high concentration of antibodies directed at denatured epitopes of PGHS, suitable for screening recombinant fusion proteins in the lambda gt11 system.
~ ~
92 kd
~
~ ~.
~
66
~
45
~
2L "
I
2
3
4
5
6
7
8
25
9
i0
Figure I. Proteins were fractionated o n 10% SDS-PAGE and transferred onto a nitrocellulose paper. Subsequently they were stained for total protein by imido black or immunostained. Lane (1,10) molecular weight markers stained with imido black (2,3) anti-PGHS antiserum 1:500,000, antigen is same as lane 5 (4) control non-immune rabbit serum 1:500,000, antigen is same as lane 5 (5) authentic PGHS standard stained with imido black (6,7) anti-PGHS antiserum 1:500,000, antigen is same as lane 9 (8) control non-immune rabbit serum 1:500,000, antigen is same as lane 9 (9) RSV solubilized microsomal proteins stained with imido black.
DECEMBER 1986 VOL. 32 NO. 6
833
PROSTAGLANDINS
Immunoprecipitation of Nascent PGHS from In Vitro Translated Polypeptides To test the ability of anti-PGHS antiserum to precipitate PGHS in solution, we prepared mRNA from ram seminal vesicles and translated it in vitro in the presence of ~ S-methionlne in rabbit reticulocyte lysates. As shown in Figure 2, analysis of the product on 10% SDS-PAGE showed that anti-PGHS antiserum specifically precipitated a 67 kd band from 35 S-labelled peptides derived from translation of RSV mRNA fraction. The smaller molecular weight polypeptides are from non-speciflc adsorption to the immunopreclpitate since they are present in both the control non-immune serum lane 3 as well as anti-PGHS antiserum.
92 kd 66
45
25
t,
I
2
3
4
5
Figure 2. mRNA from RSV was translated ~ vitro in rabbit reticulocyte lysates in the presence of ~ S-methionine. The proteins were either analyzed directly or immunoprecipitated with test antiserum and protein A sepharose and analyzed. They were resolved on a 10% SDS-PAGE gel and autoradiography was performed. Lane (I) total proteins synthesized from RSV mRNA (2) immunoprecipitate of lane I with anti-PGHS antiserum (3) immunoprecipitate of lane I with non-immune rabbit serum (4) control translation mix without exogenous mRNA (5) positive control translation mix with brome mosaic virus mRNA.
834
DECEMBER 1986 VOL. 32 NO. 6
PROSTAGLANDINS
To determine the detection limit of anti-PGHS antiserum, different amounts of pure PGHS standard were spotted on nitrocellulose sheets in a fixed volume of I ul. They were then reacted with 1:100 dilution of anti-PGHS antiserum and immunostained as in Western Blots. The antiserum stained the PGHS protein up to a detection limit of 0.1 ng. This is below the amount of protein usually found in lambda gt11 plaques making this serum a feasible screening agent to isolate the eDNA clone for PGHS in the lambda gt11 system. Screening of Lambda gt11 cDNA Library from Human Endothelial Cells In the lambda gt11 system, the eDNA directed proteins are expressed in a fused form with E-Coli beta-galaetosidase protein when lambda gt11-infected plaques on E. Coli host YI090 are induced with isopropylthiogalactoside (IPTG). Using this system, we infected E. Coli host YI090 cells with lambda gt11 containing eDNA inserts from human endothelial cells to initiate a lytic cycle. After IPTG induction on nitrocellulose sheets, the proteins were detected with anti-PGHS antiserum (1:100 dilution). First round screening of approximately 300,000 plaques produced 19 positives which were further purified. Second and subsequent rounds of screening produced three immunoreactive clones which were then plaque purified. They were termed as lambda gtCox8, lambda gtCox18 and lambda gtCox21. Lambda gtCox18 gave the strongest signal, 21 being intermediate and 8 being the weakest in immunoreactivity (Figure 3). Lambda gtCox's were lysogenized into an E.Coli strain YI089 and the resulting lysogens were grown up to liter quantities at 32 C. Since YI089 contains a temperature sensltlve lambda repressor, temperature shlft to 42 C causes the lysogenlzed phage to initiate the lytic cycle of growth. This method allowed us to obtain large quantities of recombinant protein as well as DNA (25). Lysogenic E. Coli, containing lambda gtCox18 were grown up in the presence of IPTG to induce the production of fusion protein encoded by the eDNA. The proteins obtained were analyzed by Western blotting. A large fusion protein ( 4/180 kd) composed of beta-galactosidase and the eDNA encoded protein was detected by anti-PGHS antiserum only when induced by IPTG. This confirms that the immunological reactivity observed on the plaques was indeed from the expression of recombinant lambda gt11. O
.
.
•
.
O
DECEMBER 1986 VOL. 32 NO. 6
835
PROSTAGLANDINS
lambda gtCoxl8
lambda gtCox8
lambda gtCox21 f
Figure 3. Plaque purified lambda gtCox's were plated on LB am~icillin plates containing YI090 cells. Following phage growth at 42vC for 5 hours, nitrocellulose filters, pre-soaked in 10 mM IPTG were placed on top of the phage plaques and allowed to produce the recombinant protein overnight at 37°C. The presence of the protein was detected by immunostainlng with anti-PGHS antiserum similar to the procedure in the Western blots. 8, 18 and 21 are 3 different lambda gtCox clones. A, B and C represent 1:100 dilution of phage stocks.
836
DECEMBER 1986 VOL. 32 NO. 6
PROSTAGLANDINS
Cross-Reactivity of Lambda gtCox's with Different Anti-PGHS Antisera
We obtained two independently raised anti-PGHS antisera against the ram enzyme. One such antiserum from R. Bockman was a rabbit polyclonal antiserum against enzymatically active PGHS purified by conventional column chromatographic techniques (18). Another antiserum from G. Roth, also against the ram PGHS~ was prepared by a different method which entail~d acetylation with ~H-labelled aspirin and isolation of the ~H-labelled protein to be used as an antigen (29). These antisera were teated against the lambda gtCox clones at indicated dilutions. As shown in Figure 4, lambda gtCox18 reacted strongly with all three anti-PGHS antisera. This clone was therefore selected for further characterization.
8-3 8-4
8-2 18-1 8-1 18-4
18-2 18-3
21-3
21-2
21-4~,
21-1
Figure 4. Lambda gtCox's were plated on YI090 cells and Jmm)nostained with different anti-PGHS serum as in Figure 3. 8, 18 and 21 represent 3 different lambda gtCox clones. Antiserum I is obtained from R. Bockman and is directed against pure PGHS. Antiserum 2 from G. Roth is directed against pure aspirinated PGHS. Antiserum 3 is control non-immune rabbit serum and antiserum 4 is anti-PGHS antiserum prepared by us. All antlsera were used at the dilution of 1:100 except for #2, which was at 1:200 dilution.
D E C E M B E R 1986 VOL. 32 NO. 6
837
PROSTAGLANDINS
Analysis of cDNA Lysogenized lambda gtCox18 was used to prepare large quantities of DNA. The lambda gtCox18 DNA, when out with the restriction enzyme EcoRI produced a cDNA insert of approximately 2.2 kbp as shown in Figure 5. This insert was then subcloned into the EcoRI site of plasmid pUC 8 yielding a recombinant plasmid called pCoxTH18. This plasmid contains 2.7 kbp sequences from pUC 8 and 2.2 kbp from the insert.
23.1 9.4 6.5 2.3 2.0
I
I
2
Figure 5. DNA from lambda gtcoxa8 was prepared as described. It was then cut with EcoRI and was subjected to I% agarose gel electrophoresis in Tris-Acetate EDTA (TAE) buffer. The gel was stained with ethidium bromide. Lanes: (I) lambda DNA Hind III cut molecular weight markers (2) lambda gtCox18 EcoRI cut DNA.
838
DECEMBER 1986 VOL. 32 NO. 6
PROSTAGLANDINS
Design of a Synthetic DNA Probe Specific for the Aspirin-Acetylation Site of Cyclooxygenase The N-terminal 14 amino acids and a more internal 22 amino acid stretch around the anpirin-acetylation site have been sequenced (7). A synthetic 24-nucleotide DNA probe corresponding to an eight amino-acid sequence in the aspirin-acetylation site was synthesized using an Applied Systems 380B DNA synthesizer. In doing so, all 3 or 4 fold degenerate eodons (Figure 6) were replaced with inosine (30) to reduce the overall degeneracy and the anti-sense probe was synthesized as shown. The p ~ b e is a 24~mer, in which the degeneracy is reduced from 2 -~ to only 2- by the inosine substitution technique and is designated PGHS2.
PGHS2 Asn-Pro-Ile-Glu-Ser-Pro-Glu-Tyr 5'-AAU-CCU-AUU-GAA-AGU-CCU-GAA-UAU-3' C C C G C C G C A A A G G 5'-ATA-TTC-IGG-ACT-TTC-IAT-IGG-ATT-3' G C G C G
Figure 6. The 8 amino acid stretch of PGHS from the aspirin-acetylation site of PGHS was used to design the DNA probe. All possible codons as indicated were made less degenerate by replacing with inosine. The antisense probe (bottom) was synthesized as shown. 39 The synthetic DNA probe was labelled at its 5' ~gd with -P-phosphate using polynucleotide kinase and gamma-~-P-ATP.
DECEMBER 1986 VOL. 32 NO. 6
839
PROSTAGLANDINS
It was then hybridized on a Southern blot to the DNA from pCoxTH18 cut with several diagnostic restriction enzymes. As shown in Figure 7, restriction enzymes BamHI and Sstl each cut once in the insert region of pCoxTH18 whereas no restriction sites for any of the other enzymes were contained in the insert. ~ e n the same gel was blotted on nitrocellulose and hybridized with J~P-labelled PGHS2, strong hybridization with the intact plasmid DNA was observed. In addition BamHI cut out a 1.5 kbp fragment from 5' region of the insert which also hybridized strongly and left a 3.4 kbp fragment containing 2.7 kbp pUC 8 vector sequences and 0.7 kbp 3' region sequences from the cDNA insert. The B.4 kbp fragment did not give appreciable hybridization. These results indicate that the cloned 2.2 kbp cDNA represents the cDNA for cyclooxygenase and that the aspirin-acetylation site is located in the 1.5 kbp 5' moiety.
,,irt~
x
~
n
-I- m
0
co
EtBr stained 2.2 Kb
~mMW~ ~
5.0 4.0 3.0 20
~ ~Q
32 -PGHS2
16
hybrldlsed ~ =
-122 Kb
f _
-
i
.
-
-
5.0 3 . 0
.1~ 1.6
Figure 7. pCoxTH18 DNA was prepared and digested with a battery of restriction enzymes. The DNA was then subjected to I% ~ a r o s e gel electrophoresis in TAE buffer and stained with ethidium bromide to locate the bands. It was then blotted o n ~ a sheet of nitrocellulose paper and hybridized with ~-P-labelled PGHS2. Hybridization was done at room temperature overnight. It was then washed in 0.2 x SSC buffer at room temperature to remove unbound probe and exposed for autoradiography.
840
DECEMBER 1986 VOL. 32 NO. 6
PROSTAGLANDINS
DISCUSSION Modification of an established procedure for purification of PGHS on polyacrylamide gel electrophoresis was used to prepare large amounts of enzyme protein. Denatured proteins from SDS-PAGE are a good source of antigen for library screening purposes since the antigenic epitopes obtained from a lambda gt11 expression library will probably differ from the conformation of the native enzyme. Also polyacrylamide is known to be a good adjuvant. As shown by Western Blotting assay, the antiserum obtained was specific only for the 67 kd PGHS band and in addition, reacted with enzymatically active, pure PGHS from Oxford Biomedical, Inc., at the same signal intensity. Ram seminal vesicles contain high levels of PGHS and high levels of the mRNA for the enzyme so that the 67 kd band constituted a major component in the total protein profile obtained from RSV mRNA translation (Figure 2). The PGHS antiserum selectively ~ u n o p r e c i p i t a t e d the 67 kd band from the in vitro translated, S-methionine labelled, nascent proteins d~rec-~edby the mRNA fraction from ram seminal vesicles. This further confirms the selectivity of the antiserum preparation. The PGHS antiserum had a detection limit of 0.1 ng by our peroxidase detection method. This indicated that it could be used to screen a lambda gt11 expression vector library since more than 0.1 ng of protein can be synthesized by the phage plaques in this system (25). We used a eDNA library from human endothelial cells in lambda gt11 to isolate PGHS clones with the PGHS antiserum. These cells produce significant quantities of prcstanoids (8) and the enzyme protein itself has been localized by immunofluorescent techniques
(32). Screening with anti-PGHS antiserum to a library of 300,000 recombinant phage yielded 3 positive clones which were then plaque purified. Lysates containing the fusion proteins directed by these phage were analyzed by Western blotting with the anti-PGHS antiserum. Immuno-reactive, large ( ~ 180 kd) proteins were found only when induced with the lac-operon inducer IPTG, indicating that the immunoreactivity observed in the plaques was eDNA directed. Of the three positive clones, lambda gtCox18 reacted strongly with two other independently raised anti-PGHS antiserum and was therefore selected for further characterization whereas the other two did not react appreciably. The DNA isolated from lambda gtCox18 was cut with EcoRI, which cleaved-off the eDNA insert. It was approximately 2.2 kbp indicating that it could code for a protein containing about 700 amino acids. Since cyclooxygenase contains approximately 600 amino acids, the 2.2 kbp DNA could potentially code for the full length PGHS. However, since we have no information on the extent of the 3' and 5' untranslated regions of the cyclooxygenase messenger RNA, and if the 3' sequences were large, then the eDNA obtained would not be of full length. No PGHS enzymatic activity was detected in lysates of lambda gtCox18 from E. Coli which is not unexpected because the
DECEMBER 1986 VOL. 32 NO. 6
841
PROSTAGLANDINS
cyclooxygenase sequences are expressed as part of the large aJ 180 kd fusion protein with beta-galactosidase. The eDNA insert from lambda gtCox18 was subcloned into the EcoRI site of plasmid pUC 8 yielding the recombinant plasmid pCoxTH18. From the known restriction information about the pUC 8 vector, several restriction enzymes were tested on the DNA from pCoxTH18 to identify restriction sites in the insert. There are no restriction sites in pUC 8 for BstEII, EcoRV, XhoI, SstI and ClaI, whereas PstI, HindIII, BamHI and SalI have unique sites. The BamHI site is 6bp downstream from the EeoRI site. Thus BamHI cut pCoxTH18 and produced a large 5' insert fragment of 1.5 kbp and a 3.4 kbp ~ a g m e n t of fused vector and 3' insert sequences. The P-labelled synthetic DNA probe synthesized according to the aspirin-acetylation site of PGHS hybridized strongly to the intact plasmid DNA and to the large BamHI cut 5' insert fragment of 1.5 kbp, but not to the vector and 3' sequences. Based upon the antisera reactivities and the nucleic acid hybridization data, the results indicate that the isolated 2.2 kbp DNA represents the eDNA for human Prostaglandin H Synthase and that the aspirin acetylation site in the enzyme is associated with the 1.5 kbp 5' moiety of the eDNA. ACKNOWLEDGEMENTS We thank Drs. Macing and Rices of Meloy Laboratories, Inc., Rockville, MD for the human endothelial cell eDNA library and Dr. R. Bockman of Memorial Sloan-Kettering Cancer Centre, N.Y. and Dr. Gerald Roth for two independently prepared antisera against PGHS. We are grateful to Mr. M. Yadven and Mr. C. Siegall for assistance in making the synthetic DNA probe and to Miss G. Tillisch for manuscript preparation. KEY WORDS Cyclooxygenase, Prostaglandin Synthase, eDNA Cloning, Human Endothelial Cells REFERENCES
I.
Pace-Aciak, C.R. and W.L. Smith. In:The Enzymes Vol. 16 (P. Boyer, ed.) Academic Press, N.Y., 1983. p. 543.
2.
Smith, W.L. and W.E.M. Lands. Oxygenation of polyunsaturated fatty acids during prostaglandin biosynthesis by sheep vesicular gland. Biochemistry1_! :3276. 1972.
.
842
Brotherton, A. and J.C. Hoak. Prostacyclin synthesis in human vascular endothelial cells is limited by irreversible deactivation of cyclooxygenase. J. Clin. Invest. 72:1255. 1983.
DECEMBER 1986 VOL. 32 NO. 6
PROSTAGLANDINS
4.
Goldsmith, d.C., C.T. Jafvert, P. Dollar, W.G. Owen and J.C. Hoak. Prostacyclin release from cultured and ex vivo bovine vascular endothelium. Lab. Invest. 45:197. 1980.
5.
Bailey, J.M., B. Muza, T. Hla and J. Pash. Role of epidermal growth factor in cyclooxygenase synthesis. In:Adv. PG, TX. and LT. Res. Vol. 15. (0. Hayaishi and S. Yamamoto, eds.) Raven Press, N.Y., 1985. p. 141.
. Kulmacz, R.J. and W.E.M. Lands. Stoichiometry and kinetics of the interaction of PG H synthase with anti-inflammatory agents. J. Biol. Chem. 260:12572. 1985. 7.
Roth, G.J., E.T. Machuga and J. Ozols. Isolation and covalent structure of the aspirin-modified, active-site region of prostaglandin synthetase. Biochemistry 22:4672. 1983.
8,
Jaffe, E.A. and B.B. Weksler. Recovery of endothelial cell prostacyclin production after inhibiton by low doses of aspirin. J. Clin. Invest. 63:532. 1979.
. Whiting, J., K. Salata and J.M. Bailey. Aspirin: an unexpected side effect on prostacyclin synthesis of cultured vascular smooth muscle cells. Science 210:663. 1980. 10.
Baenziger, N.L., P.R. Becherer and P.W. Majerus. Characterization of prostacyclin synthesis in cultured human arterial smooth muscle cells, venous endothelial cells and skin fibroblasts. Cell 16:967. 1979.
11.
Habernicht, A.S.R., M. Goerig, J. Grulich, D. Rothe, R. Gronwald, V. Loth, G. Schettler, B. Kommerell, R. Ross. PDGF stimulates PG synthesis by rapid and denovo induction of cyclooxygenase. J. Clin. Invest. 75:1381. 1985.
12.
Fagan, J.M. and A.L. Goldberg. Inhibitors of protein and RNA synthesis cause a rapid block in prostaElandin production at the prostaglandin synthase step. Proc. Nat'l. Acad. Sci., U.S.A. 83:2771. 1986.
13.
Bailey, J.M., B. Muza, T. Hla and K Salata. Restoration of prostacyclin synthase in cultured vascular smooth muscle cells after aspirin treatment: regulation by epidermal growth factor. J. Lipid Res. 26:54. 1985.
14.
van der Ouderra, F.J., M. Buytenhek, D.H. Nugteren and D.A. van Dorp. Purification and characterization of prostaglandin endoperoxide synthase from sheep vesicular glands. Biochim. Biophys. Acta. 487:315. 1977.
D E C E M B E R 1986 VOL. 32 NO. 6
843
PROSTAGLANDINS
15.
Miyamotc, T., N. Ogino, S. Yamamoto and O. Hayaishi. Purification of prcstaglandin endcpercxide synthase from bovine vesicular gland miercsomes. J. Biol. Chem. 251:2629. 1976.
16.
Hemler, M., W.E.M. Lands and W.L. Smith. Purification of the cyclooxygenase that forms prostaglandins. J. Biol. Chem. 251 :5575. 1976.
17.
Roth, G.J., C.J. Sick and J. Ozols. Structural characteristics of prostaglandin synthetase from sheep vesicular gland. J. Biol. Chem. 255:1301. 1980.
18.
Hiekok, N.J., G. Chin and R.S. Beckman. Characterization of prcstaglandin endoperoxide synthase from a human cell line. Bioehim. Bicphys. Aeta 877:79. 1986.
19.
Laemmli, U.K. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680. 1970.
20.
Towbin, H., T. Stachelin and J. Gordon. Electrophoretie transfer of proteins from SDS-PAGE to nitrocellulose filters for detection with antibodies. Prec. Nat'l. Aead. Sci. U.S.A. 76:4350. 1979.
21.
Chirgwin, J.M., A.E. Przybyla, R.T. McDonald and W.J. Rutter. Isolation of RNA from sources enriched in ribcnuclease. Biochemistry 18:5294. 1979.
22.
Aviv, H. and P. Leder. Purification of mRNA by affinity chromatography on olige-dT cellulose. Prec. Nat'l. Acad. Sei. U.S.A. 69:1408. 1972.
23.
Pelham, H.R. and R.E. Jackson. Translation of mRNA in rabbit reticulocyte lysates. Eur. J. Biochem. 67:247. 1976.
24.
Tong, B.D., S. Levine, M. Jaye, G. Rieca, W. Drohan, T. Maciag and T. Deuel. Isolation and sequencing of a eDNA clone homologous to the v-sis oncogene from human endothelial cell. Mol. Cell Biol. 6:3018. 1986.
25.
Huynh, T.V., R. Young and R. Davis. In:DNA cloning: a practical approach. (D.M. Glever, ed.) I.R.L. Press, Oxford., 1985. p. 49.
26.
Kaslow, D. A rapid method for obtaining 14:6767. 1986.
27.
Maniatis, T., E.F. Fritsch and J. Samhrook. In:Molecular cloning: a laboratory manual. Cold Spring Harbor Press, N.Y., 1982.
844
DNA. Nue. Acid. Res.
DECEMBER 1986 VOL. 32 NO. 6
PROSTAGLANDINS
28.
Dewitt, D.L., T.M. Rollins, J.S. Day, J.D. Gauger and W.L Smith. Orientation of active site and antigenic determinants of PG synthase in the endoplasmic reticulum. J. Biol. Chem. 256 :10375. 1981.
29.
Roth, G. Radioimmunoassay for PG H synthase. In:Methods in Enzymology. Vol. 86. Academic Press, N.Y., 1982. p. 222.
3O.
Ohtsuka, E., S. Matsuki, M. Ikehara, Y. Takahashi and K. Matsobara. An alternative approach to deoxyribonucleotides as hybridization probes by insertion of deoxyinosine at ambiguous codon positions. J. Biol. Chem. 260:2605. 1985.
31.
Williams-Ashman, H.G. Regulatory features of seminal vesicle development and function. Curr. Top. Cell Reg. Vol. 22. Academic Press, N.Y., 1983. P. 201.
32.
Dewitt, D.L., J.S. Day, W.K. Sonnenburg and W.L. Smith. Concentrations of PG H synthase and PGI 2 synthase in the endothelium and smooth muscle. J. Clin. Invest. 72:1882. 1983.
33.
Beaucage, S . L . and M.H. C a r u t h e r s .
Editor:
P.W. Ramwell
Tet. Lett. 22:1859.
Received:
DECEMBER 1986 VOL. 32 NO. 6
11-3-86
1981.
Accepted: 11-4-86
845