Jan 29, 1976 - of vitellogenin mRNA. This mRNA was of the same size as that found in the liver of the rooster actively synthesizing vitello- genin in response to ...
Proc. Natl. Acad. Sc. USA
Vol. 73, No. 5, pp. 1442-1446, May 1976 Biochemistry
Induction of vitellogenin synthesis by estrogen in avian liver: Relationship between level of vitellogenin mRNA and vitellogenin synthesis (estrogen induction/translational control/cell-free protein synthesis)
KATHLEEN P. MULLINIX, WALDEMAR WETEKAM, ROGER G. DEELEY, JEFFREY I. GORDON, MARILYN MEYERS, KENNETH A. KENT, AND ROBERT F. GOLDBERGER Laboratory of Biochemistry, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 20014
Communicated by Lawrence Bogorad, January 29, 1976
pears in the serum after about 4 hr, reaches a level of approximately 3 mg/ml, and declines, reaching an undetectable level within 1 week (1); stimulation of the rooster with a second injection of estrogen 10-14 days after the first one also results in the appearance of vitellogenin in the serum in about 4 hr but causes the concentration of vitellogenin to increase more rapidly, to reach a higher level, and to take longer to disappear (10 days) than does the primary stimulation; tertiary stimulation is attended by a response even more extensive and long-lasting than the response to secondary stimulation (7-9). It is known that with a primary injection of estrogen de novo RNA synthesis is required for induction of vitellogenin synthesis (4, 6, 10). On the basis of these findings one could speculate that the ratelimiting step in vitellogenin synthesis is transcription of the vitellogenin gene, but other possibilities must also be considered. This paper reports the results of experiments designed to determine whether the presence and amount of vitellogenin messenger RNA in rooster liver is correlated with the amount of vitellogenin found in the serum following primary, secondary, and tertiary injections of estrogen.
We have investigated the estrogen-mediated ABSTRACT induction of vitellogenin synthesis in rooster liver. We compared the concentrations of vitellogenin messenger RNA (mRNA) in the liver with the concentrations of vitellogenin in the sera of roosters that had received various treatments with estrogen. We found no vitellogenin mRNA in the livers of the unstimulated roosters. An initial injection of estrogen was attended by de novo synthesis of vitellogenin mRNA in the liver and accumulation of vitellogenin in the serum. When vitellogenin was no longer present in the serum or liver (the "post-estrogen-serum-negative" state), the liver was found to contain appreciable amounts of vitellogenin mRNA. This mRNA was of the same size as that found in the liver of the rooster actively synthesizing vitellogenin in response to estrogen. Whereas vitellogenin mRNA was in large polysomes in the livers of the roosters actively synthesizing vitellogenin, the vitellogenin mRNA in the liver of the post-estrogen-serum-negative rooster was not associated with polysomes. The possible relevance of these findings to the fact that the rooster responds differently to a primary stimulation with estrogen than to subsequent stimulations is discussed.
Vitellogenin is the precursor of the chicken egg yolk phosphoproteins phosvitin and lipovitellin (ref. 1; J. L. Christmann, M. G. Grayson, and R. C. C. Huang, personal communication). This protein is-normally synthesized in the liver of the laying hen and is transported through the blood to the oviduct, where it is deposited in the developing oocyte (2). Vitellogenin is not normally synthesized by immature chickens or by roosters, but its synthesis can be induced in these animals by estrogen (3-5). Finding no structure into which it can be deposited, vitellogenin accumulates in the plasma of the estrogen-stimulated rooster, reaching a concentration as high as 5 mg/ml (1). Study of the induction of vitellogenin synthesis in the rooster provides an opportunity to understand the mechanisms by which hormones regulate specific gene expression in the liver, an organ of diverse synthetic potential. While the vitellogenin gene is not transcribed at all in normal rooster liver, transcription of this gene following stimulation of the rooster with estrogen results in the synthesis and accumulation of so much vitellogenin mRNA that its presence is detectable in crude polysomal RNA (6). In addition, vitellogenin, the protein specified by this messenger RNA, is of unusual chemical structure and is thus easily identifiable (1). An intriguing aspect of vitellogenin induction is that, after the rooster has responded to an initial injection of estrogen and the serum vitellogenin level has returned to zero, the animal responds to subsequent injections of estrogen with a faster rate of accumulation of vitellogenin in the serum and an increased maximal level of serum vitellogenin (7, 8). Prior to estrogen stimulation, vitellogenin is not detectable in the plasma of the rooster; following a primary stimulation, vitellogenin first ap-
MATERIALS AND METHODS Estrogen Treatment of Roosters and Antibody Preparation. Estrogen treatment of roosters and preparation of purified antivitellogenin and antilipovitellin antibodies were as described (1, 6). Identification of Vitellogenin in Serum and Liver of Roosters. Vitellogenin in serum was determined by a colorimetric analysis of protein-bound phosphate, by Ouchterlony double immunodiffusion analysis with highly purified monospecific antibody (1), and by immunoprecipitation of 3P-labeled vitellogenin from serum of roosters that had been injected intraperitoneally with H?33PO4 (2 mCi/kg) 1 hr before exsanguination. In the immunoprecipitation experiments, immunoprecipitates were dissociated and denatured by heating at 1000 for 5 min in 0.1 M Tris base containing dithiothreitol (0.1 M) and sodium dodecyl sulfate (1%) (11). These denatured immunoprecipitates were subjected to electrophoresis on 5% polyacrylamide gels containing 0.1% sodium dodecyl sulfate (12). After electrophoresis the gels were sliced and the radioactivity in each slice was determined as described by Deeley et al. (1). To determine the amount of vitellogenin in the liver, extracts were made of livers from roosters labeled with H3asPO4 as described above. Livers were homogenized in 0.1 M sodium phosphate, pH 7.2, containing Triton X-100 (0.6%) sodium L-deoxycholate (1.2%), and a-toluenesulfonyl fluoride (50 ,ug/ml). After low-speed centrifugation, the supernatant frac1442
Biochemistry: Mullinix et al. were centrifuged at 235,000 X ga, for 1 hr. The supernatant fractions were subjected to indirect immunoprecipitation
tions
described by Wetekam et al. (6) and the radioactivities of the precipitates were determined in toluene-based scintillation fluid, using the Beckman LS-355 spectrometer. Isolation of Polysomes and Polysomal RNA. Polysomes were prepared from rooster liver by the method of Palmiter (13). These polysomes were centrifuged on 10-34% isokinetic sucrose gradients under the conditions described by Noll (14). The gradients were fractionated and the absorbances were monitored with an Isco model 640 fraction collector. The fractions were pooled, as indicated in the figures, and extracted with phenol. The RNA was precipitated with ethanol, centrifuged, and dissolved in water to yield concentrations between 1 mg/ml and 3 mg/ml. Twenty microliters of these solutions were tested in the in vitro system for the capacity to direct the cell-free synthesis of vitellogenin (see below). Polysomal RNA was prepared by deproteinization of polysomes with phenol and chloroform as described by Palmiter (13). Immediately after the last extraction with chloroform the aqueous phase, containing the RNA, was heated to 600 for 3 min to denature the RNA. This material was then applied directly to linear gradients of 5-20% sucrose in 0.01 M N-2hydroxyethylpiperazine-N'-2-ethanesulfonic acid (Hepes), pH 7.5, containing EDTA (0.001 M), as described by Palmiter (15). The gradients were centrifuged at 40 at 67,000 X g., for 16 hr in a SW 27 rotor in a Beckman L-5 65 centrifuge, after which they were fractionated as described above. The RNA obtained by ethanol precipitation of sucrose gradient fractions of polysomal RNA was dissolved in 0.1-0.5 ml of water to yield concentrations between 1 mg/ml and 5 mg/ml. Twenty microliters of these solutions were used in each assay (total volume 0.1 ml) for vitellogenin mRNA activity. Cell-Free Protein Synthesis. Cell-free protein synthesis was carried out utilizing a rabbit reticulocyte lysate as described in detail previously (6). The products of the cell-free protein synthesis reactions were analyzed immunologically as described previously (6).
Proc. Natl. Acad. Sci. USA 73 (1976)
Table 1. Vitellogenin-specific mRNA in rooster liver
as
RESULTS As has been previously shown in several other laboratories (7, 8, 16), we found that the magnitude of the response to estrogen with respect to vitellogenin synthesis increases upon secondary and tertiary injections of estrogen. Serum levels of vitellogenin reach higher levels and are maintained for longer periods after
secondary and tertiary injections of estrogen than after a primary
injection.
Polysomal RNA was extracted from the livers of roosters at various times after the injection of estrogen and the ability of the various RNA preparations to direct the synthesis of vitellogenin in a rabbit reticulocyte cell-free protein synthesis system was determined. Table 1 shows the results of some of these experiments. As we have reported previously (6), polysomal RNA from unstimulated roosters did not contain vitellogenin mRNA, whereas this mRNA was present in estrogen-stimulated roosters that were actively synthesizing vitellogenin. In addition, we found vitellogenin mRNA in livers of roosters that had been injected with estrogen but no longer had vitellogenin in their plasma ("post-estrogen-serum-negative" roosters). The same result was obtained each of the fifteen times this experiment was repeated. Our data indicated that estrogen-treated roosters were serum-negative (down to the basal level found in unstimulated roosters) 6 days after a primary, 11 days after a secondary, and 15 days after a tertiary injection of estrogen, as
1443
Source of RNA
In vitro assaya
Activity
(cpm)b
(%)
Unstimulated roosters 0 Estrogen-induced; secondaryC 560 100 PESNd (14 days after primary) 155 28 775 Estrogen-induced; tertiary 100 PESNd 250 32 (20 days after secondary) a Cell-free protein synthesis was carried out in a rabbit reticulocyte lysate for 90 min as described previously (6), using between 50 and 100 ,g of RNA per assay (0.1 ml total volume). The labeled amino acid used was ['4C]leucine (0.5 ACi of [14C]leucine per 0.1 ml incubation, specific activity 280 mCi/mmol). The newly synthesized protein was precipitated by the indirect immunoprecipitation reaction described previously (6), with antilipovitellin antibody, and the radioactivity was corrected for the nonspecific precipitation with preimmune serum. b The results shown are averages of assays carried out in triplicate expressed in terms of the activity of 50 fg of RNA. The standard deviation of the assay was A 2%. For the various RNA samples tested, the nonspecific precipitation obtained with preimmune serum ranged between 250 and 360 cpm. c For each preparation of polysomal RNA from the liver of a postestrogen-serum-negative rooster, a comparable preparation of RNA was made simultaneously from the liver of a vitellogeninsynthesizing rooster of the same age that differed only in its exposure to estrogen. d Post-estrogen-serum-negative.
determined by a colorimetric determination of protein-bound phosphate in the serum. In order to interpret the meaning of the finding that vitellogenin mRNA was present in livers of post-estrogen-serumnegative roosters, it was necessary to first determine whether or not any vitellogenin was present in the sera or in the livers of these animals. By Ouchterlony double immunodiffusion analyses no vitellogenin could be found in either sera or liver extracts from post-estrogen-serum-negative roosters under conditions in which 2-5% of the fully induced level of vitellogenin could be detected. Immunoprecipitation of sera from roosters that had been injected with Ha1PO4 1 hr before being sacrificed was carried out using purified antilipovitellin antibody, which has been shown to react with vitellogenin (1). The immunoprecipitates were subjected to electrophoresis in polyacrylamide gels containing sodium dodecyl sulfate and the patterns of radioactivities in the fractionated gels were determined. While vitellogenin was present in immunoprecipitates from the sera of roosters actively synthesizing vitellogenin, there was none detectable in the sera of either post-estrogen-serumnegative or unstimulated roosters. In addition, the amount of 33P present in these sera as protein-bound phosphate was analyzed after the delipidation procedure of Goldstein and Hasty (16). One milliliter of serum from an unstimulated rooster contained 230 cpm of aP; 1 ml of serum from a stimulated rooster (2 days after a primary estrogen stimulation) gave 5270 cpm of MP; and 1 ml of serum from a post-estrogen-serumnegative rooster (12 days after a primary estrogen stimulation) gave 217 cpm of 33P. Since 100 cpm above background are readily detectable, we conclude that if the serum of the postestrogen-serum-negative rooster contains any vitellogenin at all, it must be less than 2% of that found in the serum of the stimulated rooster. Finally, in order to rule out the possibility that the absence of vitellogenin from the serum was due to a
1444
Biochemistry: Mullinix et al.
defect in transport from the liver, we determined the amounts of vitellogenin in extracts of livers of roosters that had been injected with 3a33P04, as described above. These extracts were subjected to highly sensitive and specific indirect immunoprecipitation analyses as described previously (6). Whereas vitellogenin was readily detectable in liver extracts of roosters actively synthesizing vitellogenin, none was detectable in the extracts from post-estrogen-serum-negative or unstimulated roosters. By these criteria, we have determined that vitellogenin is not present in the sera or in the livers of unstimulated roosters or post-estrogen-serum-negative roosters. The products of cell-free protein synthesis reactions, programmed with polysomal RNA from post-estrogen-serumnegative roosters, were characterized by reactivity with monospecific antibody and by the competition between the protein synthesized in titro and authentic vitellogenin in the indirect immunoprecipitation reaction. With results analogous to those reported by Wetekam et al. (6), who studied products of cell-free protein synthesis directed by polysomal RNA from estrogen-treated, vitellogenin-synthesizing roosters, we found that purified vitellogenin competed completely with the product of the cell-free synthesis directed by polysomal RNA from post-estrogen-serum-negative roosters. Sucrose gradient centrifugation of polysomal RNA was carried out in order to determine the size of the vitellogenin mRNA from roosters actively synthesizing vitellogenin and to compare its size with that of the vitellogenin mRNA found in post-estrogen-serum-negative roosters. Fig. lA-C shows absorbance traces and vitellogenin mRNA activity distributions of liver polysomal RNA preparations from unstimulated, estrogen-treated (vitellogenin-synthesizing), and post-estrogenserum-negative roosters. After heat denaturation the preparations were centrifuged on low-salt, 5-20% linear sucrose gradients. Under these conditions, any breaks in the RNA not otherwise apparent would be easily identified as causing the RNA to sediment more slowly than expected for the whole vitellogenin mRNA. Indeed, when proper precautions were not taken to prevent degradation, such a change in sedimentation pattern was readily observed after denaturation. The gradients were fractionated and the RNA in each fraction was tested in the rabbit reticulocyte cell-free system. There was no vitellogenin mRNA in any of the fractions of RNA from the liver of control (unstimulated ) roosters. When polysomal RNA from viiellogenin-synthesizing roosters was tested the major portion of vitellogenin mRNA applied to the gradient was found in the fraction containing the largest RNA, with a size larger than 28 S. Fig. 1B shows the results of an experiment in which 63% of the vitellogenin mRNA was found in this fraction. This is the approximate size expected for a messenger RNA that specifies a protein the size of vitellogenin, which has a subunit molecular weight of 240,000 (1). Analyses of fractions of sucrose gradients in which denatured polysomal RNA from post-estrogenserum-negative roosters had been centrifuged showed that this vitellogenin mRNA was of the same size as that from estrogen-treated, vitellogenin-synthesizing roosters. Fig. 1C shows results of an experiment in which 67% of the vitellogenin mRNA was in that fraction of the sucrose gradient that contained RNA larger than 28 S. In the many experiments performed the recoveries of vitellogenin mRNA activity after sucrose gradient centrifugation were variable; however, the distribution of this activity consistently gave the results illustrated in Fig. 1B and C. Fig. ID-F shows the results of some of the experiments in which we determined the size distribution of polysomes in the livers of untreated, estrogen-treated, and post-estrogen-
Proc. Natl. Acad. Sci. USA 73 (1976)
25
Is
i 75 en
