0013-7227/90/1261-0167$02.00/0 Endocrinology Copyright © 1990 by The Endocrine Society

Vol. 126, No. 1 Printed in U.S.A.

Prothrombin Levels Are Increased in the Estrogen Treated Immature Rat Uterus* KATHERINE P. HENRIKSON, ELENA E. JAZIN, JEFFREY A. GREENWOOD, AND HERBERT W. DICKERMAN Wadsworth Center for Laboratories and Research, New York State Health Department (K.P.H., E.E.J., H.W.DJ; State University of New York Graduate School of Public Health (K.P.H., E.E.J., K.W.D.); and Siena College (J.A.G.), Albany, New York 12201

kinetic constants similar to those of thrombin with tripeptide pnitroanilides as substrates, and it digests actin to give the same peptides as thrombin. We conclude that the uterine proenzyme is prothrombin. The time course of the prothrombin response to estrogen suggests that prothrombin enters the uterus as part of the transudation of plasma proteins that occurs after estrogen stimulation. A membrane-bound uterine procoagulant that activates uterine prothrombin also increases in response to estrogen stimulation. We propose that the simultaneous increase in these two activities results in a localized generation of thrombin, a well characterized mitogen in fibroblasts and epithelial cells. Our results suggest that thrombin may have a vital function as a mitogen in the early steps of the estrogen-stimulated hypertrophy and hyperplasia of the immature uterus. (Endocrinology 126: 167-175,1990)

ABSTRACT. An estrogen-responsive uterine proenzyme of a proteinase in the immature rat uterus has been known for some time. Its mol wt is 77,000, its N-terminal amino acid sequence is the same as prothrombin's for 15 residues, it contains 7carboxyl glutamate residues, its biosynthesis is prevented by warfarin, it cross-reacts with antibodies to human and rat prothrombin, and it can be activated by human factor Xa or a uterine procoagulant. The products of activation, when separated on sodium dodecyl sulfate-gels, react with antibodies to human or rat prothrombin to give bands that have mol wt corresponding to those of the products of activation of prothrombin. These activation intermediates hydrolyze synthetic substrates specific for thrombin and have the same mol wt as the activation products of prothrombin. The proteinase generated in the activation has the following properties of thrombin: it is inhibited by hirudin and PheProArg-chloromethyl ketone, it has

T

HE PHYSIOLOGICAL response of the immature rat uterus to estrogen stimulation results in the development of a reproductively competent organ. It is a complex process, consisting of distinct early and late responses. Early responses include increases in the activity of specific enzymes (1, 2) and a marked change in vascular permeability, with consequent transudation of plasma proteins and water imbibition (3). The transudation of serum albumin and plasminogen has been demonstrated in rat and mouse uteri (4-6). The idea that one of the plasma proteins could have an important role as a growth factor in the subsequent steps of uterine hypertrophy and hyperplasia has received little attention. We have described an arginine esteropeptidase activity in the immature rat uterus which increases to peak activity in a hormone- and tissue-specific response 3 h after estrogen stimulation of the animal (7). This activity

is a complex of an integral plasma membrane activator and a proenzyme which, after a calcium-requiring activation, is a serine proteinase (8, 9). Although our initial studies suggested that the proenzyme was uterine in origin (7), further characterization of the enzyme and proenzyme have revealed a distinct likeness of these components to thrombin and prothrombin. We now report that the proenzyme and enzyme are identical to prothrombin and thrombin. Thrombin is a mitogen in epithelial cells (10), muscle cells (11), and fibroblasts (12). A specific, high affinity plasma membrane thrombin receptor has been described in chick embryo fibroblasts (13). The increase in prothrombin in the uterus, derived from estrogen-stimulated plasma transudation, provides a source for the localized generation of thrombin. We suggest that thrombin serves a paracrine function in the hypertrophy and hyperplasia of the estrogen-stimulated immature uterus.

Received September 5, 1989. Address all correspondence and requests for reprints to: Katherine P. Henrikson, Ph.D., Wadsworth Center for Laboratories and Research, New York State Department of Health, Empire State Plaza, P.O. Box 509, Albany, New York 12201-0509. * This work was supported by NSF Grant DCB-8710087.

Materials and Methods Materials iV-(p-Tosyl)L-arginine-[3H]methyl ester (TAME) was supplied by Amersham Corp. (Arlington Heights, IL). Tri- and 167

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 November 2015. at 15:56 For personal use only. No other uses without permission. . All rights reserved.

168

ESTROGEN INCREASES UTERINE PROTHROMBIN ACTIVITY

tetra-peptide p-nitroanilides (pNAs) were purchased from Helena Laboratories (Beaumont, TX), Calbiochem (La Jolla, CA), Boehringer-Mannheim (Indianapolis, IN), and Cambridge Research Biochemicals (Valley Stream, NY). PheProArg-chloromethyl ketone (PPACK) was obtained from Calbiochem, rabbit antihuman prothrombin was from Dako-Patts (Santa Barbara, CA), and immunoblotting supplies were from Kirkegaard and Perry Laboratories (Gaithersburg, MD). Immobilon membranes were purchased from Millipore, Inc. (Bedford, MA). Prestained electrophoretic standard proteins were obtained from Bio-Rad Laboratories (Richmond, CA) or Diversified BioTech (Newton, MA). Hirudin was a generous gift from Dr. William Lawson of the Wadsworth Center. Warfarin, purified rat prothrombin, and highly purified human factor X were generous gifts from Dr. Michael J. Fasco of the Wadsworth Center. Dr. Robert E. Olson of the State University of New York at Stony Brook gave us the rabbit antirat prothrombin antibody. It was purified by chromatography on a protein-ASepharose column. Centricon microconcentrators (10,000 mol wt cut-off) were obtained from Amicon (Danvers, MA). Nineteen- to 21-day-old female Wistar rats were supplied by the Griffin Laboratory of the Wadsworth Center and maintained on rat chow. All other biochemicals and supplies came from Sigma Chemical Co. (St. Louis, MO) or Fisher Scientific (Rochester, NY). Enzyme preparation and assay Animals were injected ip with 4.5 /ug estradiol (E2) and killed 3 h later by cervical dislocation. Uteri were removed, cytosols were prepared, and enzyme was purified as previously described (7, 8) through ammonium sulfate precipitation, DEAE ion exchange chromatography, and fast liquid chromatography on a Superose 12 gel exclusion column. Active fractions were pooled and either used for experiments or further purified by ion exchange chromatography on a Mono Q 5/5 column. Pooled Superose 12 fractions (2 ml) were loaded onto the Mono Q column. The column was washed with 10 mM Tris-Cl buffer, pH 7.7-0.1 mM EGTA-0.1 M NaCl, then eluted with a linear gradient of 0.1-0.5 M NaCl in the same buffer. Fractions were collected in siliconized tubes which contained sufficient glycerol to result in a final concentration of 5%. The column was run at room temperature, and fractions were collected in an ice bath. The enzyme was assayed as previously described (8,14) with certain modifications. The activator was separated from the proenzyme by the ion exchange chromatographic step. The protein peak that was not retained by the DEAE column was the source of activator for assays of the proenzyme. The activator and proenzyme were incubated in 10 mM Tris, pH 7.7, with 2 mM CaCl2 for 30 min at 37 C to allow activation to occur. Aliquots were taken in triplicate into a reaction mixture which contained 0.1 M cyclohexylaminopropane sulfonic acid (CAPS) buffer (pH 9.5), 4 mM Tris buffer (pH 7.7), and 16 ftM [3H]TAME (150,000 cpm in ;00 (d). The final assay pH was 9.0. The capless Eppendorf tube containing the reaction mixture was placed in a vial at room temperature, which also held 10 ml of a toluene-based scintillation fluor and 20 n\ 0.05 M TAME in 10% acetic acid at room temperature. The reaction

Endo • 1990 Voll26«Nol

was terminated after 30 min by vigorous shaking of the vial. The product of the reaction, [3H]methanol, was extracted into the toluene-based scintillation fluid, while the unreacted substrate remained in the aqueous droplets and was not counted. A control that lacked proenzyme was carried through the procedure to provide a correction for background. One thousand counts in the assay were equivalent to 11 pmol product formed, and 1 TAME unit (U) of enzyme activity catalyzed the formation of 1 pmol [3H]methanol/min. Assays with pNA substrates Kinetic constants for three tripeptide-pNA substrates were determined spectrophotometrically at 20 C. Substrate solutions (1-25 HM) were prepared by dilution of a stock solution into 0.1 M CAPS, pH 9.0. One milliliter of substrate solution was placed in a cuvette, the initial absorbance at 405 fim was recorded, 62.7 TAME U enzyme were added, and the change in the absorbance at 405 ^m was recorded for at least 5 min. The rate of the reaction was determined at least four times for each substrate concentration. When the nominal substrate concentration was less than 6 nM, the true concentration was determined by allowing the reaction to go to completion and recording the final absorbance. A molar extinction coefficient of 10,700 (15) was used. Kinetic constants were calculated with ENZFITTER, a nonlinear regression data analysis program from Elsevier-Biosoft (Milltown, NJ). Gel and blotting procedures

Gel electrophoresis was performed by the method of Laemmli (16). Gels contained 10% acrylamide and 0.1% sodium dodecyl sulfate (SDS) unless otherwise specified. In some cases proteins were transferred electrophoretically from the gel to a nitrocellulose membrane in Tris-glycine-20% methanol buffer at pH 9.5 or to an Immobilon membrane in CAPS buffer at pH 11. Prestained standards were used when proteins were transferred. Since their migration rates may be altered by labeling with the dye (17), caution must be used in assigning precise mol wt to experimental protein bands based on their migration relative to that of prestained standards. However, the standards do provide a convenient estimate of the mol wt of the experimental protein bands. Immunoblotting was initiated with a primary antibody that was either a commercial rabbit antihuman prothrombin immunoglobulin G (IgG) fraction used at a concentration of 0.19 mg/ml or a rabbit antirat prothrombin antiserum purified by protein-A-Sepharose chromatography and used at a concentration of 0.11 mg/ml. The nitrocellulose membranes were blocked with 0.2% milk protein for 1-2 h at 37, washed with four changes of 10 mM Tris, pH 7.7-0.15 M NaCl-0.05% Tween-20 (TBST), incubated with primary antibody for 1-2 h at room temperature, washed with TBST, incubated for 1 h at room temperature with a 1:500 dilution of phosphatase-coupled goat antirabbit IgG, and washed again with TBST. Color was developed in a brief incubation with 5-bromo-4-chloro-3-indolyl phosphate and nitroblue tetrazolium at pH 9.5. Blots with nonimmune rabbit serum as the primary antibody gave no reaction.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 November 2015. at 15:56 For personal use only. No other uses without permission. . All rights reserved.

ESTROGEN INCREASES UTERINE PROTHROMBIN ACTIVITY

169

B

Other procedures

The Bradford microprocedure was used for protein determinations (18). Amino acid sequences were determined on an Applied Biosystems 477 A protein sequencer from samples electrophoretically transferred from gels to Immobilon membranes by the procedure of Matsudaira (19). 93-

Results

93-

Characterization of the proenzyme Purification and mol wt. The procedure described previously (8), i.e. ammonium sulfate precipitation, DEAEcellulose chromatography, and gel exclusion chromatography, resulted in a 100-fold purification of the proenzyme. One additional step, ion exchange chromatography on a Mono Q 5/5 column, resulted in a 30- to 40-fold additional purification, with a 40-55% recovery. In a typical preparation, 1.56 mg enzyme, with a specific activity of 13.5 U//ug, were loaded onto the column; 31 Hg enzyme, with a specific activity of 374, were eluted in a 55% yield. A 3800-fold overall purification resulted. A polyacrylamide gel and a Western blot of the purified enzyme are shown in Fig. 1. Three bands appear in the gel (Fig. 1A), at 77,000, 32,000, and 29,000 Da. An immunoblot (Fig. IB) with rabbit antihuman prothrombin antibody showed that only the 77,000 Da band was related to prothrombin. Reduction of the samples with mercaptoethanol did not alter the results (data not shown). Attempts to separate the 77,000-Da band (the proenzyme) from the impurities by alteration of the gradient on the Mono Q column resulted in losses of up to 90% of the proenzyme. Based on these results (and others given below), we conclude that the 77,000-Da band is the uterine proenzyme, that it is immunologically related to prothrombin, and that the lower mol wt bands are not immunologically related to prothrombin. Stability of proenzyme dependence of activation

and enzyme and calcium

The highly purified proenzyme was not stable to storage, presumably because it was in very dilute solution. The proenzyme was completely activated after a 15-min incubation at 37 C with the uterine membrane activator. It retained 86% of its activity after 1 h at 37 and 67% of its activity after 3 days at room temperature in the presence of the activator and CaCl2. A calcium concentration curve of activation was constructed with CaCl2-EGTA buffers, calculated with the algorithm of Goldstein (20). No activation occurred at micromolar concentrations of calcium; maximum activation was at 2 mM calcium and above (data not shown).

554267- : 3642-

30-

v frty

30-

FIG. 1. Gel electrophoresis and immunoblotting of the uterine proenzyme after purification. A, A SDS-gel stained with Coomassie blue. Lane 1, Proenzyme (825 U) after the Mono Q step; lane 2, 2057 U proenzyme after the Superose step. B, Immunoblots of a SDS-gel. Lane 1, Mono Q-purified proenzyme (792U); lane 2, 594 U Superose enzyme. The primary antibody was rabbit antihuman prothrombin. No bands were visible if the primary antibody was omitted or if nonimmune rabbit IgG was used. The mol wt of the standard proteins are given on the left. The arrows indicate the proenzyme bands.

Similarity of the proenzyme and prothrombin Both human and bovine prothrombin have been extensively studied (21). Their amino acid sequences and mechanisms of proteolytic activation to thrombin are well known. Although rat prothrombin has been purified (22), it has been studied in much less detail. We compared the rat uterine proenzyme's properties with the previously reported properties of other prothrombins and directly with rat prothrombin in the following gel experiments. Amino terminal sequence. The N-terminal sequence of the proenzyme was determined through the first 15 amino acids by the procedure of Matsudaira (19). The proenzyme was electrophoresed on a SDS-gel; the proteins were transferred to Immobilon, stained briefly with Coomassie blue, and destained, and the 77,000-Da band was cut out of the membrane. The fragment of membrane was placed directly into the chamber of the protein sequencer and analyzed by Dr. F. Carl Haase of the Biotechnology Division of Rohm and Haas Corp (Spring-

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 November 2015. at 15:56 For personal use only. No other uses without permission. . All rights reserved.

ESTROGEN INCREASES UTERINE PROTHROMBIN ACTIVITY

170

house, PA). The sequence is presented in Table 1, along with the published partial sequences of human, bovine, and rat prothrombin. The proenzyme partial sequence matches the available rat N-terminal sequence and agrees well with the 15 N-terminal amino acids of human and bovine prothrombins. Our determination of a single sequence indicated that the 77,000-Da band was a single protein. yCarboxylglutamate (Gla) content of the protein. Prothrombin, like other proteins that contain Gla residues, is precipitated by BaSO4. In an experiment designed to show that the proenzyme had a similar property, 904.2 U proenzyme were incubated with 6.7 mg BaSO4 for 30 min at 4 C. The precipitate was collected by centrifugation, washed with 10 mM Tris, and eluted with 100 /A 0.02 M citrate, pH 7.8, for 2 h at 4 C. The eluate was activated and assayed for activity in the TAME assay. Of the initial 904 U, 27 U (3%) were found in the BaSO4 supernatant and 301 U (33%) in the citrate eluate. Warfarin, a dicoumarol analog of vitamin K, is an anticoagulant widely used as rodent poison. In the intact animal it prevents the posttranslational, vitamin Independent modification of specific Glu residues to Gla in several coagulation proteins, including prothrombin (23). The biosynthesis of inactive forms of those proteins results in death from a failure of coagulation. To determine the effect of warfarin on the proenzyme, 20- to 22day-old rats were treated with sublethal levels of warfarin for 48 h, then injected with E2 and killed 3 h later. Proenzyme was prepared from the uterine cytosols through the ammonium sulfate step, activated, and assayed for TAME activity. Control groups were given no warfarin or E2. The results are shown in Table 2. While warfarin had no effect on the E2-induced increase in wet weight of the uterus or on the increase in total soluble TABLE 1. Amino-terminal sequences of prothrombins Bovine (21) Human (21) Rat (22) Proenzyme (this study)

KGNLE KGNLE

ANTGFLEEVR ANT-FLEEVR ANSGFLEEL ANSGFLEELR

KGNLE

Endo • 1990 Vol 126 • No 1

protein, proenzyme activity was undetectable in the warfarin-treated nonestrogenized animals, and in the estrogenized animals it was only 16% of the values in controls without warfarin. We conclude from these experiments that the proenzyme contained Gla residues. Immunological identity. Immunological identity of the proenzyme and rat prothrombin (22) was shown by double diffusion (Fig. 2). Antibody to rat prothrombin was placed in the center well, and alternating wells contained rat prothrombin or uterine proenzyme. The continuous precipitin line, without spurs or tails, indicated that the proenzyme preparation contained a protein that was immunologically identical to rat prothrombin. Activation of proenzyme and mol wt of the active species

Prothrombin can be activated by prothrombinase, a membrane-bound complex of coagulation factors Xa and Va, by factor Xa alone, or by venom from the snake Echis carinatus. Factor Xa or prothrombinase activation results in the formation of several intermediate mol wt forms (21, 22, 24). These forms and their mol wt include prethrombin 1 (55,000) and fragment 1 (25,000), generated by proteolysis of prothrombin by the newly generated thrombin, and prethrombin 2 (38,000) and fragment 1-2 (35,000), generated by factor Xa. Fragment 1-2 is also degraded by thrombin to fragments 1 and 2 (12,000); thrombin itself (37,000) is formed from prethrombin 2 after hydrolysis of a short peptide. Immunoblots of activated proenzyme and prothrombin. We activated the uterine proenzyme and rat prothrombin to determine whether the products of activation were the same. Immunoblots were used to show that the active forms of the enzyme were immunologically related to prothrombin. The activators were highly purified human coagulation factor Xa or uterine membrane-bound procoagulant. Proenzyme and rat prothrombin each were activated, run on SDS-gels, blotted to nitrocellulose, and probed with antibodies to human and rat prothrombin. The results are shown in Fig. 3. The proenzyme and prothrombin both gave bands at 77,000 Da. The pro-

TABLE 2. Effect of warfarin on uterine proenzyme activity

Treatment

Uterine wet wt (mg)

Protein in ammonium sulfate fraction (ng)

Proenzyme activity in ammonium sulfate fraction (U)

Control-control Control-E2 Warfarin-control Warfarin-E2

25.7 32.8 26.9 31.2

305 310 266 278

0.70 2.85 ND 0.42

Twenty- to 22-day-old rats were injected with 9 ng warfarin, then given 8 mg/liter in their drinking water. After 48 h they were injected with E2 and killed 3 h later. Uteri were collected, weighed, and homogenized. Ammonium sulfate fractions (0-55%) of cytosols were prepared, dialyzed, and assayed for proenzyme. There were 53-60 rats in each group; all data are given per rat. ND, not detectable.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 November 2015. at 15:56 For personal use only. No other uses without permission. . All rights reserved.

ESTROGEN INCREASES UTERINE PROTHROMBIN ACTIVITY

171

FlG. 2. Double immunodiffusion of uterine proenzyme and rat prothrombin. Five microliters of antiserum, rabbit antirat prothrombin on the left and nonimmune rabbit serum on the right, were placed in the center well. Wells 1, 3, and 5 contained 165 TAME U purified rat prothrombin, while wells 2, 4, and 6 contained 55 U proenzyme at the Superose stage of purification. Samples were allowed to diffuse for 24 h at room temperature, then the slide was washed, dried, and stained with Coomassie blue.

FlG. 3. Immunoblots of proenzyme and prothrombin before and after activation. Proenzyme after Mono Q purification was used for each experiment. A, Activation of 1362 U proenzyme and 1100 U prothrombin by human factor Xa. The Xa was generated from factor X by incubation with Russell's viper venom; 1.75 ng factor Xa were used per activation in 2 mM CaCl2 at 37 C for 45 min. The primary antibody was antihuman prothrombin. Lane 1, Proenzyme; lane 2, prothrombin; lane 3, activated proenzyme; lane 4, activated prothrombin; lane 5, activator. B, Activation of 880 U proenzyme and 825 U prothrombin by 1.6 jug uterine activator in 2 mM CaCl2 for 30 min at 37 C. The primary antibody was antirat prothrombin. Lane 1, Rat prothrombin; lane 2, activated rat prothrombin; lane 3, activator; lane 4, proenzyme; lane 5, activated proenzyme. Positions of the standard proteins are indicated by the numbers at the left of each photograph. Arrows indicate the bands generated by activation.

A

thrombin lane also had a band at 55,000 Da, which was an indication of mild degradation of prothrombin after prolonged storage at —20 C. After activation with human factor Xa (Fig. 3A), the proenzyme gave a band at 55,000 Da and faint bands near the 36,000-Da marker. Activated prothrombin gave similar faint bands near the 36,000Da marker. The lowest band, at 25,000 Da, was present in the activator and both experimental lanes, but was darker in the activated proenzyme channel. Activation by the uterine procoagulant (Fig. 3B) gave the same bands from prothrombin and proenzyme. Again, prothrombin gave bands at 55,000 and 25,000 Da before activation. In each lane prominent bands at 55,000, 38,000, 37,000, and 25,000 Da were seen. Similar results

B

were obtained with E. carinatus venom as the activator (data not shown). Inhibition by PPACK and Western blots. The effect of PP ACK on the appearance of bands in the immunoblots of activated enzyme was investigated. Since PPACK inhibits thrombin and not factor Xa, its addition after a brief time of activation should result in the retention of bands of protein that would otherwise disappear through thrombin autoproteolysis. The activation of 1557U uterine proenzyme was initiated, and PPACK was added to a final concentration of 0.1 ixM after 1, 5, 10, and 30 min of activation. The hydrolysis of TAME was completely inhibited in all samples. The samples were all probed with antihuman prothrombin after electrophoresis and

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 November 2015. at 15:56 For personal use only. No other uses without permission. . All rights reserved.

172

ESTROGEN INCREASES UTERINE PROTHROMBIN ACTIVITY

Endo • 1990 Vol 126-No 1

prestained standards. Yellow bands appeared in the overtransfer to nitrocellulose (see Fig. 4). In all channels, lay in areas that corresponded to 55,000 and 36,000bands appeared for prothrombin, prethrombin 1, pre42,000 Da within 12-15 min at room temperature. No thrombin 2, and thrombin. When the inhibitor was added after 1 min, an additional band for fragment 1 • 2 at 35,000 other yellow spots appeared within 30 min. In another experiment with 454 U activated enzyme, the SDS was Da was visible. This band did not appear in the other displaced by Triton and bands were sliced from the gel channels, but a lower mol wt band at 25,000 Da correby reference to the prestained standards, macerated, sponding to fragment 1 was present. extracted with 10 mM Tris buffer, pH 7.7, for 4 h, and Mol wt of the active species. Of the intermediates gener- assayed for activity against both the chromogenic subated by activation of prothrombin, several, including strate and TAME. The total recovery of TAME activity thrombin, are active against low mol wt substrates (25), was 46%. Of the recovered activity, 2% was between although only thrombin is active in coagulation. Thus, it 67,000-93,000 Da, 64% was at 55,000 Da, and 34% was was of interest to determine which activation products between 36,000-42,000 Da in both assays. Similar results of the uterine proenzyme were active against the small were obtained when an entire channel from the gel was substrates used here. Proenzyme was activated with utersliced into 2-mm slices, and each was macerated, exine procoagulant and electrophoresed in a 10% gel with tracted with buffer, and assayed. No activity was found SDS. The gel was soaked in 2.5% Triton X-100 to in regions of the gel where the lower mol wt protein displace the SDS, and activity was measured in the gel. impurities were found with the Coomassie blue stain. In two experiments with 3850 and 2030 U enzyme, a A similar experiment was performed with 2200 U rat piece of filter paper soaked with 1 mM D-Pheprothrombin, activated by either E. carinatus venom or PipecolylArg-pNA and 0.17 M CAPS, pH 9.5, was placed uterine procoagulant. Again, the yellow bands in the over the gel and observed until yellow zones appeared. overlay with the chromogenic substrate appeared first at The mol wt of the active species was inferred from the 55,000 Da and then between 36,000-42,000 Da only in relative positions of the yellow zones of activity and the the channels that contained activated enzyme. We conclude that the 77,000-Da proenzyme upon activation generates proteins of the same mol wt and enzymatic activity as authentic rat prothrombin. Enzymatic similarity of the enzyme and thrombin Several specific substrates and inhibitors of thrombin are known. The activated enzyme was compared to thrombin with respect to inhibitor specificity and substrate reactivity. Inhibitors. Thrombin is specifically inhibited by hirudin (a peptide anticoagulant derived from leeches) and by PPACK (26, 27). The hirudin inhibition of rat thrombin and uterine enzyme were studied in parallel experiments. E. carinatus venom was used to activate 4.1 TAME U rat prothrombin; 5.5 U uterine proenzyme were activated by uterine procoagulant. Hirudin was added to each, and the enzymes were assayed after a 10-min delay. The concentration of hirudin that resulted in 50% inhibition was 0.027 U/ml for thrombin and 0.02 U/ml for enzyme. In similar experiments with PPACK the concentrations required for 50% inhibition were 7.4 nM for 4.1 U thrombin and 5.0 nM for 3.7 U enzyme. FIG. 4. PPACK inhibition and the products of activation of proenzyme by the uterine procoagulant. PPACK was added to 0.1 /uM at timed intervals after the initiation of activation. Activation was stopped by adding sample buffer and boiling the samples. Lane 1, PPACK added after 1 min of activation; lane 2, 5 min; lane 3, 10 min; lane 4, 30 min. The mol wt of the standards are given on the right. The numbers on the left give the mol wt of the proenzyme and the bands generated from it.

Substrates. In addition to the hydrolysis of TAME, a relatively nonspecific substrate, the uterine enzyme hydrolyzed a variety of tri- and tetra-peptide-pNAs. The relative rates of hydrolysis of 14 of these substrates were determined; the 3 most rapidly hydrolyzed were D-IleProArg-pNA, tosyl-GlyProArg-pNA, and D-Phe-

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 November 2015. at 15:56 For personal use only. No other uses without permission. . All rights reserved.

ESTROGEN INCREASES UTERINE PROTHROMBIN ACTIVITY

173

and fragment 1 result from hydrolysis of prothrombin at Arg155 (human prothrombin numbering) by thrombin; prethrombin 2 and fragment 1.2 result from hydrolysis by factor Xa at Arg273. Catalytically active thrombin, a two-chain molecule held together by a disulfide bond, is generated by hydrolysis at Arg322 by factor Xa. An additional small peptide (1000 Da) may be hydrolyzed off prethrombin 2 to generate a-thrombin. Prethrombins 1 and 2 which have been hydrolyzed at Arg322 are catalytically active against small molecules, but do not clot fibrinogen (24, 25). Thrombin also digests fragment 1.2 rapidly to fragment 1 and fragment 2. To prove that the uterine proenzyme and enzyme are prothrombin and thrombin, it was necessary to demonstrate the presence of these mol wt species in gels as well as to show that the proteins are the same immunologically and enzymatically. The proenzyme has been partly purified. The preparations, while not pure, had only one band at 77,000 Da, which cross-reacted with antiprothrombin antibodies. A comparison of the proenzyme's properties with those of rat, human, and bovine prothrombin showed that it was prothrombin. The enzymatic properties of the proenzyme used for comparison included its mol wt, its aminoterminal sequence, and its Gla content (demonstrated by binding to BaSO4 and reduction in activity by administration of warfarin to the rats). The enzymatic properties of the activated enzyme that were like thrombin included inhibition by PPACK and hirudin, reactivity with certain tripeptide-pNAs, digestion of actin at defined points in the molecule, and formation of clots from fibrinogen. The immunoblots demonstrated all of the bands that would be expected from prothrombin activation. Activation by the uterine membrane preparation gave prominent bands at 38,000 and 37,000 Da, which corresponded to prethrombin 2 and thrombin. The two additional bands at 55,000 and 25,000 Da corresponded to prethrombin 1 and fragment 1, which were formed by the newly generated thrombin. The results were the same for authentic rat prothrombin and uterine proenzyme. Activation by highly purified human factor Xa generated faint bands in the area where thrombin and prethrombin 2 were expected. Enough thrombin was generated to activate the prothrombin, which resulted in the formation of prethrombin 1 and fragment 1. Only fragment 1.2 and fragment 2 did not appear in these gels. Fragment

PipecolylArg-pNA. The other 11 substrates were hydrolyzed only 1-10% as fast as these 3. Kinetic constants were determined for the 3 and are compared to the constants for thrombin in Table 3. The Km values are all very similar. The maximum velocity (Vmax) values are in the same order as the kcat values for thrombin. Values for kcat, which is defined as Vmax/enzyme concentration, cannot be calculated for the uterine enzyme since its concentration was unknown. The uterine enzyme digested actin in a 2-h incubation at 37 C. When the products of the reaction were separated on a SDS-gel (Fig. 5), the 42,000-Da band of actin was reduced in intensity, and new bands appeared at 40,000, 31,000, 15,000, and 14,000 Da. Residues at the sites of proteolysis of 42 fig actin by 7.7 U enzyme were identified after gel electrophoresis and transfer to an Immobilon membrane as described by Matsudaira (19). The 40,000-Da fragment's amino-terminal sequence was Ala-Val-Phe-Pro-Ser-Ile-Gly, which corresponded to positions 31-38 of actin (28). The 30,000-Da fragment's sequence was Ala-Asn-Arg-Glu-Lys-Met-Thr-Gln-IleMet, which was the sequence of residues 115-124 of actin. Thus, the uterine enzyme hydrolyzed actin after ProArg30 and Pro-Lysn4, but did not hydrolyze the protein after the Arg-Pro-Arg40. These products of enzymatic digestion were the same as those described by Muszbek et al. (29) for bovine thrombin. The major physiological role of thrombin is the clotting of fibrinogen. The capacity of the uterine enzyme to clot fibrinogen was shown by the following experiment. The purified uterine proenzyme after activation by a uterine procoagulant was added to a fibrometer cup which contained fibrinogen; the time required for clot formation was recorded. The results are given in Table 4. The activated enzyme catalyzed the formation of a fibrin clot. Increased amounts of enzyme resulted in more rapid clot formation. Purified rat thrombin was included as a positive control. Hirudin added at 1 NIH U to 92.1 TAME U enzyme completely inhibited coagulation. Uterine activator alone did not clot fibrinogen. Discussion In the activation of prothrombin various polypeptides are generated by the action of both the added activator and newly generated thrombin (21). Thus, prethombin 1 TABLE 3. Kinetic constants for uterine enzyme and bovine thrombin Thrombin, pH7.8 (32) Substrate PhePipecolylArg-pNA IleProArg-pNA TosGlyProArg-pNA

Km

( M M)

1.50 1.20 3.61

kcat

Uterine enzyme, pH 9.5 1

(sec" )

98.4 74.0 102

Km (MM)

Vmax (nmol/Vg-min)

1.4 1.4 1.5

0.89 ± 0.06 0.62 ± 0.025 1.08 ± 0.08

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 November 2015. at 15:56 For personal use only. No other uses without permission. . All rights reserved.

ESTROGEN INCREASES UTERINE PROTHROMBIN ACTIVITY

174

3

4

7

8

FIG. 5. Digestion of actin by the uterine enzyme. Proenzyme (261 U) purified through the Superose step was activated and incubated with 50 fig actin at 37 C for various time intervals. Samples were boiled with SDS, run on a 12.5% gel, and stained with Coomassie blue. Lane 1, Standards; lane 2, 2-h incubation with 2.9 U; lane 3, 1.5 h; lane 4, 1 h; lane 5, 30 min; lane 6, 20 min; lane 7, 10 min; lane 8, actin alone. The faint high mol wt bands are due to activator and enzyme. No digestion occurred with proenzyme or activator alone. TABLE 4. Clot formation catalyzed by purified uterine enzyme Enzyme (TAME U)

Clotting time (sec)

None (activator alone) 46 92 138 92 + hirudin 247 U thrombin

>1000 521 ± 131 211 ± 0.3 127.1 ± 10.7 >1000 70.9 ± 3.9

Proenzyme was activated by the uterine membrane activator and added to a fibrometer cup which contained 500 jug fibrinogen dissolved in 0.1 M Tris, pH 7.7-0.18 M NaCl. Each result is the mean of at least three determinations.

1.2 was rapidly broken down by thrombin to fragment 1, which was readily visible, and fragment 2, which never appeared in our gels. It may have lost the immunogenic epitope and, therefore, should be nonreactive in the immunostaining. Inhibition of the activation process after a short period of time with PPACK, a specific thrombin inhibitor, showed the rapidity of the activation process and the generation of the various activation products. The addition of PPACK after 1 min of activation allowed bands for fragment 1.2 as well as thrombin and prethrombin 2 to be visualized. After 5 min activation was virtually complete, the amount of thrombin generated was sufficient to completely degrade fragment

E n d o • 1990 Vol 126 • No 1

1.2, and a band for fragment 1 was visible. The results after 10 and 30 min of activation were similar. Enzymatically active forms of the enzyme after activation had mol wt of 55,000 and 36,000-42,000. Similar results were found for both the chromogenic substrates and [3H]TAME. Because of rapid diffusion of the chromogenic substrates we could not determine which of the immunologically detected bands in the 36,000-42,000 mol wt range had the activity. The active form at 55,000 Da was prethrombin 1. Protein contaminants of the proenzyme preparations with mol wt of 32,000 and 29,000 had neither immunological cross-reactivity nor enzymatic activity against various thrombin substrates. We conclude that the proenzyme found in the rat uterus after estrogen stimulation is prothrombin. The products formed after activation in vitro are consistent with activation by factor Xa, either free or as part of a prothrombinase complex (24). Generation of Xa by a plasma membrane estrogen-responsive procoagulant with distinct similarities to tissue factor, an initiator of the coagulation cascade, is described in the following paper (30). The changes in permeability of the uterus after estrogen administration allow a well documented transudation of plasma proteins into the uterus, which provides a ready source of prothrombin. Although little attention has been given to the possibility that these plasma proteins may have a local uterine function, the availability of prothrombin as a consequence of estrogen stimulation raises the possibility that thrombin generation is a component of the pleiotropic response of the immature rat uterus to estrogens. The local generation of thrombin, a known mitogen for fibroblasts and other cell lines (10-13, 31), may provide a paracrine factor that is active in the development of a mature, reproductively competent uterus as a response to estrogen stimulation.

Acknowledgments We thank Drs. S. A. Kumar, Michael Fasco, James Dias, and Brian Pentecost for helpful discussions.

References 1. Katzenellenbogen BS, Gorski J 1975 Estrogen actions on syntheses of macromolecules in target cells. In: Litwack G (ed) Biochemical Actions of Hormones. Academic Press, New York, vol 3:187 2. Kumar SA, O'Connor DL, Seeger JI, Beach TA, Dickerman HW 1983 Purification and characterization of creatine kinase, an estrogen-induced uterine protein (IP) from immature rats. Biochem Biophys Res Commun 111:156 3. Clark JH, Peck Jr EJ 1979 Female Sex Steroids: Receptors and Function. Springer-Verlag, Berlin p 79 4. Finlay TH, Katz J, Rasums A, Seiler S, Levitz M 1981 Estrogenstimulated uptake of al-protease inhibitor and other plasma proteins by the mouse uterus. Endocrinology 108:2129 5. Finlay TH, Katz J, Kirsch L, Levitz M, Nathoo SA, Seiler S 1983 Estrogen-stimulated uptake of plasminogen by the mouse uterus. Endocrinology 112:856 6. Peterson RP, Spaziani E 1969 Cycloheximide and cortisol inhibi-

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 November 2015. at 15:56 For personal use only. No other uses without permission. . All rights reserved.

ESTROGEN INCREASES UTERINE PROTHROMBIN ACTIVITY

7. 8.

9. 10. 11. 12. 13.

14. 15. 16. 17. 18.

tion of estradiol-stimulated uterine uptake and distribution of homologus serum albumin and alpha globulin in the rat. Endocrinology 85:932 Henrikson KP, Dickerman HW 1983 An estrogen stimulated, calcium-dependent tosyl arginine methyl ester (TAME) hydrolase in immature rat uterus. Mol Cell Endocrinol 32:143 Henrikson KP, Jazin EE, Dickerman HW 1987 Separation and identification of two components of an estrogen-responsive, calcium-dependent arginine esteropeptidase. J Steroid Biochem 26:189 Jazin EE, Dickerman HW, Henrikson KP 1988 Estrogen stimulates a uterine plasma membrane protease activator. Endocrinology 122:500 Medrano EE, Cafferata EGA, Larcher F 1987 Role of thrombin in the proliferative response of T-47D mammary tumor cells. Exp Cell Res 172:354 Sue-A-Quan AI, Simard G, Connolly JA 1988 Thrombin modulates muscle cell growth and differentiation. J Cell Biol 107:56a (Abstract) Gordon EA, Carney DH 1988 Role of protein kinase C in thrombinstimulated cell proliferation. J Cell Biol 107:257a Perdue JF, Lubenskyi W, Kivity E, Sonder SA, Fenton IIJW 1981 Protease mitogenic response of chick embryo fibroblasts and receptor binding/processing of human alpha thrombin. J Biol Chem 256:2767 Imanari T, Wilcox GM, Pisano JJ 1976 Sensitive radiochemical esterolytic assays for urokinase. Clin Chim Acta 71:267 Izquierdo C, Burguillo FJ, Bardsley WG 1987 The non-Michaelian action of thrombin on peptdie-p-nitroanilide substrates. Biochem J 243:329 Laemmli UK 1970 Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680 Bio-Rad Corp 1988 Catalog. Bio-Rad, Richmond, CA, p 102 Bradford MM 1976 A rapid and sensitive method for the quantitation of microgram quantities of protein using the principle of protein-dye binding. Anal Biochem 72:848

175

19. Matsudaira P 1987 Sequence from picomole quantities of proteins electroblotted onto polyvinylidene difluoride membranes. J Biol Chem 262:10035 20. Goldstein DA 1979 Calculation of the concentrations of free cations and cation-ligand complexes in solutions containing multiple divalent cations and ligands. Biophys J 26:235 21. Shapiro SS, McCord S 1978 Prothrombin. In: Spaet TH (ed) Progress in hemostasis and thrombosis. Grune and Stratton, New York, vol 4:177 22. Grant GA, Suttie JW 1976 Rat prothrombin: characterization and activation. Arch Biochem Biophys 176:650 23. Swanson JC, Suttie JW 1985 Prothrombin biosynthesis: characterization of processing events in rat liver microsomes. Biochemistry 24:3890 24. Krishnaswamy S, Church WR, Nesheim ME, Mann KG 1987 Activation of human prothrombin by human prothrombinase. J Biol Chem 262:3291 25. Rosing J, Zwaal RFA, Tans G 1986 Formation of meizothrombin as intermediate in factor Xa-catalyzed prothrombin activation. J Biol Chem 261:4224 26. Walsmann P, Markwardt F 1981 Biochemische und pharmakologische aspekte des thrombininhibitors hirudin. Die Pharmazie 10:653 27. Kettner C, Shaw E 1979 Phenylalanylprolylarginine chloromethyl ketone: a selective affinity label for thrombin. Thrombosis Res 14:969 28. Collins JH, Elzinga M, Jackman N 1975 The primary structure of actin from rabbit skeletal muscle. J Biol Chem 250:5915 29. Muszbek L, Gladner JA, Laki K 1975 The fragmentation of actin by thrombin. Isolation and characterization of the split products. Arch Biochem Biophys 167:99 30. Jazin EE, Dickerman HW, Henrikson KP 1990 Estrogen regulation of a tissue factor-like procoagulant in the immature rat uterus. Endocrinology 126:176 31. Carney DH, Cunningham DD 1978 Role of specific cell-surface receptors in thrombin-stimulated cell division. Cell 15:1341

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 November 2015. at 15:56 For personal use only. No other uses without permission. . All rights reserved.