Xenopus fibrinogen. Characterization of subunits and hormonal regulation of biosynthesis.

In this paper we describe the purification and characterization of Xenopus plasma fibrinogen and the hormonal factors which regulate synthesis and secretion of fibrinogen in liver parenchymal cells in primary culture. As in other vertebrate species, Xenopus fibrinogen is composed of three nonidentical polypeptide chains, A alpha, B beta, and gamma. In contrast to mammalian fibrinogens, the B beta chain of Xenopus fibrinogen has a higher apparent molecular weight than the A alpha chain. The gamma chain has the lowest molecular weight in the frog protein, as in that of other species. The relatively large size of the frog B beta chain results from the unusually large size of the NH2-terminal B fibrinopeptide, which is released by thrombin cleavage of fibrinogen. Hormonal regulation of fibrinogen biosynthesis was examined using a primary cell culture system. Purified Xenopus liver parenchymal cells, maintained for several weeks in a defined culture medium, gradually decrease the synthesis and secretion of fibrinogen. Sustained production of this protein is dependent upon the addition of a glucocorticoid, dexamethasone, to the culture medium. Fibrinogen production is suppressed if an estrogen, estradiol-17 beta, is added to the culture medium together with dexamethasone and triiodothyronine. The Xenopus system provides new insight into the structure of fibrinogen, the evolution of this protein, and the hormonal factors which regulate its synthesis.

Primary cultures ofXenopus liver tissue and purified parenchymal cells can be maintained for several weeks in defined media (14-16). Under these conditions, the cells cease the production of many major secreted proteins, but they remain responsive to hormonal stimulation. We have shown that estradiol-17b added to the culture medium induces the synthesis and secretion of vitellogenin, the precursor of the major egg yolk proteins (14). Thyroid hormones and glucocorticoids function as co-hormones to enhance estrogen induction of secreted vitellogenin (15, 16). When the glucocorticoid dexamethasone is added to the culture medium alone, it acts to sustain or induce the synthesis and secretion of several proteins which are normally secreted by freshly excised liver (14, 16).
In all vertebrate species which have been studied, fibrinogen is a multimeric molecule with a total molecular weight of about 340,000 (for reviews see Refs. 17 and 18). Each molecule is composed of two sets of three nonidentical subunits designated Aa, BB, and y. In mammalian species, the relative sizes of the subunits, from largest to smallest, is A&, BP, and y.
During the coagulation process, the proteolytic enzyme thrombin cleaves two small polar peptides called the A and B fibrinopeptides from the NHp-tennini of the A& and BP fibrinogen chains, respectively. The resulting molecules of fibrin polymerize to form the matrix of the blood clot.
In this communication we describe the isolation and characterization of fibrinogen from Xenopus plasma. Our characterization of frog fibrinogen has allowed us to identify the fibrinogen polypeptides synthesized and secreted by liver cells in culture and to analyze the hormonal factors which regulate fibrinogen biosynthesis in this system.

EXPERIMENTAL PROCEDURES
Preparation of Xenopus Fibrinogen-Adult female frogs, 20 animals, were anesthetized by immersion in 0.2 g/100 ml of ethyl maminobenzoate, methanesulfonic acid salt (Aldrichl. The animals were dissected abdominally exposing the heart, taking care to avoid the central vein. Heparin (Sigma type H7005), 0.1 ml of a 1500-unit/ 4600 Xenopus Fibrinogen accomplished by the procedure of Bergstrom and Wallen (21). The resulting pellet was redissolved as a concentrated solution. Fibrinogen was then further purified by precipitation with glycine (20). The final protein pellet was dissolved and dialyzed at 4 "C in order to remove the cryoprecipitate (22). The purified protein was stored at -20 "C in 0.3 M NaCI, 10 m~ Tris-HCI, pH 7.3. The yield of purified fibrinogen was 35 mg/100 ml of plasma. Upon addition of thrombin, 95% of the purified protein formed a fibrin clot. Protein concentration was measured by the method of Lowry et al. (23). Bovine fibrinogen was purified as described (24). Human fibrinogen was a gift from Dr. M. Mosesson (Department of Medicine, Mount Sinai Medical Center, Milwaukee, WI).
Selective Enzymatic Digestion of Fibrinogen-For experiments in which unfractionated Xenopus plasma was digested with thrombin, the plasma was collected using 0.6 x 0.15 M NaCI, 0.015 M sodium citrate without other protease inhibitors. Bovine thrombin was purified according to the procedure of De Vreker et al. (25). The preparation started with 20,000 NIH units of thrombin and the final purified material was dissolved in 0.3 M NaCI, 0.03 M Na phosphate, pH 7.4, to give a n A~H O " , , , of 0.15. Batroxobin (Pentapharm, Basel, Switzerland) a t 500 BU/mg was prepared as a stock solution a t 1 mg/ml in 0.15 M NaCI. Copperhead Venzyme was obtained from Dr. J. Shainoff (Thrombosis Research Section, The Cleveland Clinic Foundation, Cleveland, OH), as a stock solution at 15 a-N-tosyl-L-arginine methyl ester units/& (26). Digestion reactions contained 7.5 pg of fibrinogen in 20-50 pl with final concentrations of 0.015 M Tris-HCI, pH 7.5, and 0.14 M NaCI. Digestion of radioactively labeled secreted fibrinogen was accomplished by including up to one-half volume of tissue culture medium in appropriate reactions. Amounts of enzymes used and incubation conditions are given in the figure legends. During the course of these investigations we observed that the ability of Venzyme to selectively digest only the BP polypeptide of Xenopus fibrinogen was enhanced by addition of the solution used in an in vitro translation system (27). For this reason the enzyme digestion with Venzyme also contained one-tenth volume of: 25 p~ of each amino acid (except methionine), 1 mM ATP, 0.2 m~ GTP, 8 mM creatine phosphate, 24 mM Hepesl, pH 7.6, 600 PM spermidine, 2 mM dithiothreitol, 120 mM potassium acetate, and 2 mM magnesium acetate. Preliminary experiments suggest that this enhanced specificity is due to the inhibition of a secondary protease activity present in the preparation of Copperhead Venzyme which digests the Aa fibrinogen polypeptide of Xenopus but does not digest the Aa polypeptide of either human or bovine fibrinogen.
Liver Cell Cultures-Liver parenchymal cells were prepared and purified as described previously (14) with the following modifications. Enzymatic perfusion utilized 25 ml of Type 1 collagenase (Sigma) at 200 units/ml without inclusion of bovine serum albumin and the disaggregation was completed in vitro with an additional 50 ml of the enzyme solution. No tissue pulp remained after this treatment. The final cell washes were carried out using 60% modified Coon's Medium as described (15). Cells were plated a t 3 X 10" cells/ml into 24-or 96well plastic dishes (Costar), in 1.5 or 0.15 ml of 60% modified Coon's Medium with 12.5 X IO"' units/ml insulin (Sigma) without added steroid or thyroid hormones. Each well contained a glass coverslip which had been previously coated with 200 or 20 pI of human fibrinectin (10 pg/ml). One-half the volume of the tissue culture medium in each well was changed every other day. Steroid hormones were dissolved as 25x stocks and were added to cultures as indicated in the text. Radioactive labeling of secreted proteins with ['%]methionine was carried out as described (15) except that heparin (1500 units/ml) was included in the labeling medium to prevent the degradation of fibrinogen as described by Grieninger et al. (13). For some experiments, cells were taken from a normal adult male animal while in other experiments cells were prepared from a male injected with 2 mg of estradiol-17P either two weeks or more than three months prior to sacrifice, as indicated in the figure legends.
Immunoprecipitation of Fibrinogen-A rabbit antiserum was prepared using the preparation of purified fibrinogen. The protein at 3.9 mg/ml in 0.3 M NaCI, 10 mM Tris-HCI, pH 7.3, was diluted with an equal volume of Freund's complete adjuvant (Difco) and 780 pg of fibrinogen was injected intradermally. Five weeks later, the animal was injected with 400 pg of fibrinogen in Freund's incomplete adjuvant and was bled seven days later. The resulting antiserum was used ' The abbreviations used are: Hepes, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid; SDS, sodium dodecyl sulfate. 63K, 59K, 55K, etc., M , = 63,000, 59,000, 55,000, etc. without further purification except that heparin and phenylmethanesulfonylfluoride, final concentrations 15 and 150 pg/ml, respectively, were added before use to inhibit rabbit thrombin activity. Immunoprecipitations were carried out using Staphylococcus aureus to precipitate the antigen-antibody complex according to the method of Kessler (28).
Polyacrylamide Gel Electrophoresis a n d Autoradiography-SDS-polyacrylamide gel electrophoresis and autoradiography were carried out as described previously (15). Gels were poured as exponential gradients using acrylamide (Eastman) stock solutions in the proportions of 2 volumes of 9 g/100 ml:l volume of 15 g/100 ml. Samples were dissolved in a modified sample buffer containing: 62.5 5% (v/v) 2-mercaptoethanol, 300 mM 2-mercaptoethane sulfonic acid, 0.025 g/100 ml bromphenol blue. Protein molecular weights were determined by Comparison with molecular weight standards purchased from Bethesda Research Laboratories.

Fibrinogen Structure
Resolution of X e n o p u s Plasma Fibrinogen Subunits-Fibrinogen purified from X e n o p u s plasma is resolved by electrophoresis into four polypeptide bands ( Fig. 1, lune C). The apparent molecular weights of these polypeptides are 63K, 59K, 55K, and 52K. The same protein bands are observed in whole plasma (Fig. 1, lane B), but the relative amount of the 59K polypeptide decreases and that of the 55K polypeptide increases during the process of fibrinogen purification. These observations suggest that the 59K polypeptide is converted to the 55K polypeptide and that this process continues during the protein isolation procedure. This hypothesis is supported Identification of Plasma Fibrinogen Subunits-In order to relate the four observed polypeptides to the three subunits of fibrinogen, we investigated which polypeptide bands are sensitive to digestion with specific proteolytic enzymes known to selectively digest particular subunits of fibrinogen. Thrombin cleaves the A and B fibrinopeptides from the A a and BP subunits of fibrinogen, giving rise to the a and P chains of fibrin. The y subunit is not digested by this enzyme. Addition of thrombin to either total Xenopus plasma or purified fibrinogen results in cleavage of the three largest fibrinogen polypeptides ( Fig. 1, lanes A and D). The cleavage products of the 63K and 59K polypeptides migrate at 56K-57K. Two polypeptides, at 56.5K and 57K, can be resolved when the gels are run for an extended time. The 55K polypeptide is cleaved to a product of molecular weight 53K, which in some cases is resolved from the uncleaved 52K polypeptide. Thus, the pattern of thrombin-dependent polypeptide cleavages indicates that the three largest polypeptides are forms of the Aa and BP subunits, while the 52K polypeptide is the y subunit.
Specific identification of the frog Aa and BP subunits was accomplished by the technique of selective enzymatic digestion of subunits using snake venom proteases. The enzymes batroxobin and Venzyme selectively digest the Aa and BP subunits of fibrinogen, respectively (26,30,31). Human, bovine, and frog fibrinogens were digested with these enzymes and the subunits of each fibrinogen were separated electrophoretically and examined for the disappearance of particular bands as well as the appearance of cleavage products. In our gel system, undigested human fibrinogen is resolved into five bands: a triplet of bands for the Aa subunit, one BP band, and one y band (Fig. 2, lane -). Intact bovine fibrinogen is resolved into three bands, one Aa, one BP, and one y band (Fig. 2, lane -), plus a minor band above y which may be the y' form of this subunit (29). The subunits from all three species migrate in the same molecular weight range.
Digestion of both human and bovine fibrinogens with batroxobin, results in cleavage of the largest polypeptides only, and in the appearance of new bands of slightly lower molecular weight (Fig. 2, lane B ) . This control result c o n f m s that, under our experimental conditions, batroxobin digestion is selective for the A a subunit and that the slow-migrating polypeptides are the Aa chains in these two mammalian species. In contrast, digestion of Xenopus fibrinogen with batroxobin results in cleavage of the two middle bands, but not of the largest polypeptide. Thus, both of the middle two polypeptides are forms of the Aa subunit of Xenopus fibrinogen.
Batroxobin fails to digest the largest polypeptide of Xenopus fibrinogen. Since this chain is digested by thrombin, it must be the BP subunit. Selective digestion of fibrinogen with Venzyme was used to prove this deduction. Control digestions using human and bovine fibrinogens demonstrate that for each of these species only the middle band is digested (Fig. 2,  lane V). Although Venzyme does not digest these polypeptides to completion under our conditions, it is specific for BP subunits. In contrast to mammalian fibrinogen, treatment of Xenopus fibrinogen with Venzyme results in digestion of the slowest migrating band. This finding directly confms that the largest polypeptide of Xenopus fibrinogen is the BP subunit.
In summary, by the criterion of selective sensitivity to three complementary enzymes, we conclude that the 63K polypeptide is the BP subunit of Xenopus fibrinogen, while the 59K and 55K polypeptides are alternate forms of the A a subunit. The 52K polypeptide, which is not cleaved by any of the proteolytic enzymes employed, is the y subunit. The Aa subunit of other species is known to be susceptible to proteolytic degradation at the COOH-terminal end (37), but the observed reversal in the relative sizes of the Xenopus BP and A a subunits is not due to partial proteolytic degradation of Aa chains. This conclusion is based on the fact that the Aa chain of newly secreted fibrinogen molecules comigrates with the larger form of the plasma Aa subunit, but is smaller than the secreted BP subunit (see below). Two forms of the Xenopus plasma A a subunit are observed. Upon cleavage of the NHn-terminal fibrinopeptide with thrombin, both forms are decreased in molecular weight by the same amount. Thus, we conclude that the smaller Aa form arises from the primary secreted polypeptide by removal of a 4K dalton COOH-terminal peptide.

Xenopus Fibrinogen
the P chain is about 2000 daltons smaller than the corresponding BP subunit. In contrast, thrombin cleavage of the Xenopus BP subunit gives rise to a P chain which is 6500 daltons smaller (63K to 56.5K). Thus, the apparent molecular weight of the frog B fibrinopeptide is several times greater than the size of mammalian B fibrinopeptides.
Although the Xenopus BP subunit differs from that of mammals, available evidence indicates that it is similar to that of other nonmammalian vertebrates. In the cases of chicken fibrinogen (39), and salmon fibrinogen (40), the subunits have been resolved and the BP has been shown to be larger than the Aa subunit. In another species of frog, Rana esculanta, the fibrinopeptide B has been reported to contain 43 amino acids, including tyrosine-O-S04 (41).2 This modified amino acid is characteristic of the B fibrinopeptides of many mammals and has not been found in any A fibrinopeptides. The total molecular weight calculated for these amino acids is 4387. Thus, an increased number of amino acids could partially account for the apparent size of the Xenopus B fibrinopeptide. Post-translational addition of carbohydrate residues at a site within the B fibrinopeptide accounts for the additional increase in molecular weight." In the lamprey, one of the most primitive extant vertebrates, both the size and modification of the B fibrinopeptide are similar to the structure we have observed for the frog. It comprises 36 amino acids, including tyrosine-O-S04, and it is glycosylated (42, 43). The structure of fibrinogen in the lamprey has proven difficult to analyze, but two alternate interpretations of the data suggest that the primary form of the Aa subunit is smaller than the BP subunit (44,45).
The A fibrinopeptide of frogs is similar in size to that of higher vertebrate species and distinctly different from that of the lamprey. Each of the two forms of the Xenopus plasma Aa subunit gives rise to an a chain which is 2000 daltons smaller (59K to 57K and 55K to 53K) (Fig. 2). This result is consistent with the Rana A fibrinopeptide which contains 15-17 amino acids with a total molecular weight of about 1700 (41).2 The A fibrinopeptides of at least 40 mammals, and of the chicken, and a lizard, have been sequenced. In all cases they contain 13-19 amino acids (32, 38). In contrast, the lamprey A fibrinopeptide is only six amino acids long (43).
These comparisons of nonmammalian and mammalian fibrinogen polypeptides and fibrinopeptides suggest that major alterations have occurred in the length and modification of the fibrinopeptides during the course of evolution. Such changes are distinct from the frequent amino acid substitutions which characterize the fibrinopeptides of mammals (32). We hypothesize that the A fibrinopeptide may have increased in length while the B fibrinopeptide of nonmammalian vertebrates may have arisen as an extended, glycosylated peptide and this structure persisted until emergence of the mammals. If this view is correct, fibrinogen molecules with A and B fibrinopeptides of approximately equal length may be restricted to the mammals. Changes in the length and modification of fibrinopeptides can be expected to have dramatic effects on the solubility of fibrinogen molecules. These concepts can be tested by characterization of the fibrinogens from other nonmammalian vertebrate species.

Fibrinogen Synthesis and Hormonal Regulation
Liver Cell Secreted Fibrinogen-Characterization of fibrinogen from Xenopus plasma enabled us to identify fibrinogen secreted by liver cells in culture. We have previously described methods for the purification of Xenopus liver parenchymal cells and the establishment of long-term primary cultures (15). The cells in these cultures synthesize and secrete into the culture medium proteins which can be detected by radioactive labeling with [:'%]methionine followed by gel electrophoresis and autoradiography. The pattern of radioactive secreted proteins includes three bands that co-migrated with the polypeptides of purified fibrinogen (Fig. 3, lanes A and D). In the region of the two plasma Aa chains there is only one radioactive band which co-migrates with the larger form of this subunit. These radioactive liver secreted proteins were conf m e d to be the subunits of fibrinogen by immunoprecipitation of secreted polypeptides with a rabbit antiserum prepared against plasma fibrinogen. Autoradiography of immune precipitates of radioactively labeled secreted proteins demonstrated that the anti-fibrinogen serum selectively precipitates the radioactive fibrinogen polypeptides (Fig. 3, lane B ) . These polypeptides are not precipitated by nonimmune serum (Fig.   3, lane 0. We used the method of selective enzymatic digestion with several proteases to identify the individual subunit polypeptides of secreted fibrinogen. The results of this analysis are shown in Fig. 4. As in the case of plasma fibrinogen, Venzyme digests the 63K radioactive band generating the 56.5K polypeptide (Fig. 4, lane V). Hence, these are the BP subunit and the P chain, respectively. Batroxobin cleaves the 59K band, identifying it as the Aa subunit, and generates the a chain at  (Fig. 4, lane B ) . Thrombin digests both the 63K and the 59K secreted protein bands (Fig. 4, lane T). The 52K radioactive band is not cleaved by any of these enzymes and hence is the y subunit. All of the cleavage products co-migrate with the subunit chains of plasma fibrin.
These results demonstrate that even in newly synthesized molecules the BP subunit of Xenopus fibrinogen is larger than the Aa subunit. Thus, the unusual relative sizes of the subunits does not result from proteolytic cleavage in the plasma. The secreted Aa subunit exists in a single 59K form. This fact supports our conclusion that the 55K plasma form is derived from the 59K polypeptide. Hormonal Regulation of Fibrinogen Synthesis-Xenopus provides an attractive experimental system in which to investigate the complex factors which regulate liver protein synthesis (14-16). In vivo experiments demonstrate that fibrinogen biosynthesis can be both positively and negatively regulated. The Xenopus liver primary cell culture system that we have developed has the experimental advantage that either purified parenchymal cells or pieces of tissue can be maintained for several weeks in a fully defined culture medium free of steroid and thyroid hormones. It is therefore possible to eliminate the action of endogenous hormones present in the cells at the start of the culture period.
Glucocorticoid Stimulation of Fibrinogen Synthesis-When liver parenchymal cells are maintained in a defined culture medium for an extended period in the absence of added steroid hormones, they cease the synthesis of almost the entire set of normal secreted proteins. Synthesis and secretion of this set of proteins is sustained if dexamethasone is present. Addition of dexamethasone several weeks after the start of the culture period reinduces the synthesis and secretion of these proteins. The polypeptides of fibrinogen are among those secreted proteins whose synthesis is most rapidly responsive to the absence or presence of dexamethasone.
The results shown in Fig. 5 demonstrate these phenomena. Three sets of duplicate cultures were established in a defined medium lacking steroid hormones. Two days later one set of duplicates was given dexamethasone and the added hormone was maintained at each change of the culture medium. All cultures were maintained for 21 days, a t which time another set of cultures received dexamethasone. Twenty-four hours later, secreted protein synthesis was monitored in all six cultures. The autoradiogram of the separated proteins secreted into the culture medium demonstrates that, in cultures which did not receive dexamethasone, the level of most secreted proteins had decreased significantly (Fig. 5, lanes -DX).
Cells cultured in the continuous presence of dexamethasone (Fig. 5, lanes +DX) continued to produce high levels of the entire set of secreted proteins. Cultures receiving the glucocorticoid 24 h before being assayed (Fig. 5,

lanes -+ + DX)
were reinduced to synthesize fibrinogen and some other secreted proteins. All three subunits of fibrinogen appear coordinately, consistent with the concept that these liver cells secrete intact fibrinogen molecules. Our findings complement and extend studies which have been carried out using liver from other animals (9, 11, 13).
The dose of dexamethasone used here was either lo-" or M. Even lower doses of dexamethasone, M, are also effective if thyroid hormones are added as well (16). When liver cells are cultured for extended periods in the absence of added glucocorticoids, the addition of thyroid hormones alone has no effect on the pattern of secreted proteins. Thus, thyroid hormones appear to enhance the action of low levels of endogenous glucocorticoids or exogenously added dexamethasone, but thyroid hormones do not act alone as inducers of specific secreted proteins (15). These observations may ex- plain reports indicating a direct effect of thyroid hormones on fibrinogen synthesis in relatively short-term primary cultures of chick liver cells (33). Insulin has also been reported to induce increases in liver synthesis of fibrinogen and other secreted proteins in shortterm primary cultures of chick liver cells (34). Insulin at 12.5 X units/ml is present in all our culture media. Omission of insulin results in an overall decrease in the incorporation of the radioactive precursor into liver secreted proteins and in a generally less healthy appearance of these cells. Insulin alone does not induce or sustain the synthesis of specific secreted proteins.
Estrogen Suppression of Fibrinogen Synthesis-Estrogens regulate frog liver production of vitellogenin, the precursor of major egg yolk proteins (35). Extensive hormone treatment induces very high levels of vitellogenin in the plasma of male animals, as well as the disappearance of virtually all other plasma proteins (36). These dramatic alterations in plasma protein composition reflect biosynthetic changes in the liver. Livers from males given chronic estrogen treatment synthesize and secrete vitellogenin, but not the normal set of secreted Estrogen suppression of fibrinogen production. A set of triplicate cultures was prepared from a normal male frog and was incubated for 22 days in the absence of steroid and thyroid hormones. On day 23, cultures were labeled with ["S]methionine to monitor secreted protein synthesis (1anesA). On day 25, cultures received lo-" M estradiol-17/3, M dexamethasone, and lo-' M triiodothyronine. The cultures were labeled again on day 26 (lanes B ) and day 31 (lanes C ) of the culture period. In each case labeling was carried out for 16 h and the radioactive media were collected and replaced with fresh media and appropriate hormones. 25 pl of each culture medium was mixed with approximately 1 pg of carrier vitellogenin and was then analyzed by SDS-polyacrylamide gel electrophoresis followed by autoradiography. Addition of carrier vitellogenin is necessary for sharp resolution of radioactive vitellogenin bands. proteins (16). In an effort to understand the hormonal and cellular basis for this suppression, we have previously shown that cultured pieces of tissue from an estrogen suppressed animal are dependent on dexamethasone for reinduction of fibrinogen and other normal proteins (16). Reinduction failed to occur, however, in the presence of estradiol-17/3, triiodothyronine, and dexamethasone, the combination of hormones which induces the highest level of vitellogenin production.
The results presented in Fig. 6 demonstrate that estrogen suppression can be achieved in cultures of isolated liver cells from a normal male animal. If cells are maintained for several weeks in the absence of steroid hormones, they show decreased secretion of most normal proteins, particularly fibrinogen (Fig. 6, lanes A ) . If a mixture of estradiol-17/3, dexamethasone, and triiodothyronine is then added, within 24 h the estrogen induces the synthesis of vitellogenin while the glucocorticoid induces the synthesis of fibrinogen and several other proteins (Fig. 6, lanes B ) . T h e thyroid hormone enhances the actions of both steroids (15, 16). When the mixed hormone treatment is extended for one week, vitellogenin production is sustained at a high level while the synthesis and secretion of fibrinogen in particular is suppressed (Fig. 6, lanes 0. These results demonstrate that estrogen suppression occurs by a mechanism operative within isolated normal liver cells, despite the fact that these cells are exposed to, and initially respond to, inducing levels of dexamethasone.
Suppression could arise by a failure of liver cells either to transcribe or to translate the messenger RNAs for particular proteins. We have examined the polypeptides translated in vitro from mRNA prepared from the livers of both estrogeninduced and glucocorticoid-induced or normal animals. Estrogen-induced RNA is depleted of translatable messages for serum albumin and the fibrinogen polypeptides. This observation, together with the fact that liver cells from estrogensuppressed animals can be reinduced by dexamethasone to synthesize these proteins, lead us to hypothesize that both estrogen suppression and glucocorticoid induction occur by mechanisms involving gene transcription. Experiments are currently underway to test this hypothesis.