THE JOURNAL 0 1988 by The American Society for Biochemistry

OF BIOLOGICAL CHEMISTRY and Molecular Biology, Inc

Vol. 263. , No. 6, Issue of February 25, pp. 2625-2631, 1988 Printed in U . S . A .

Estrogen-induced Destabilization of Yolk Precursor Protein mRNAs in Avian Liver*

(Received for publication, September 30, 1987)

David A. Gordon$, Gregory S. ShelnessO, Monica Nicosia, and David L. Williams From the Department of Pharmacological Sciences, Health Sciences Center, State Uniuersity of New York at Stony Brook, Stony Brook, New York 1 I794

In addition to regulating gene expression at the level of transcription, estrogen is generally considered to selectively stabilize induced mRNAs against degrada- tion. As a result of mRNA stabilization, estrogen-in- duced mRNAs accumulate to much higher levels in target cells, and the encoded proteins are made at much greater rates than would occur on the basis of tran- scriptional activation alone. The present study exam- ined the effect of estrogen on the stabilities of avian liver mRNAs that code for the yolk precursor proteins apolipoprotein (apo) I1 and vitellogenin (VTG) 11. The results show that the degradation rates of apoII and VTG I1 mRNAs during hormone withdrawal are dra- matically altered by the duration of prior estrogen treatment. During the 2 days required for the hor- monal inductions of these mRNAs to new steady states, the turnover rates of both mRNAs were the same in the presence and absence of estrogen (tlh = 13 h). This result indicates that mRNA stabilization does not con- tribute to the extensive accumulation of apoII and VTG I1 mRNAs. When the duration of estrogen treatment was extended beyond 3 days, however, hormone with- drawal led to the rapid (& = 1.5 h) and selective destabilization of these mRNAs. This result suggests that estrogen induced a destabilization activity that was only functional following hormone withdrawal. Thus, the point at which estrogen alters mRNA stabil- ity is at the level of mRNA degradation. An absence of detectable apoII mRNA degradation intermediates during either the slow or rapid mode of mRNA decay suggests that the rate-limiting step for apoII mRNA turnover is an estrogen-sensitive targeting event that marks the mRNA for rapid degradation.

In addition to regulating gene expression at the level of transcription, estrogens are believed to selectively stabilize induced mRNAs against degradation (Darnell, 1982; Darnell et al., 1986; Shapiro and Brock, 1985; Shapiro et al., 1987). As a result of mRNA stabilization, estrogen-induced mRNAs accumulate to much higher levels in target cells, and the encoded proteins are made at much greater rates than would occur on the basis of transcriptional activation alone. This

* This work was supported by Grant DK 18171 from the National Institutes of Health. Partial support was provided by an award from Innovative Research of America (to D. A. G.) . The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ Trainee in Pharmacological Sciences (Grant GM 07518). Recip- ient of an Advanced Predoctoral Fellowship from the Pharmaceutical Manufacturer’s Association Foundation.

$ Trainee in Pharmacological Sciences (Grant GM 07518).

conclusion is based on observations that estrogen-induced mRNAs decay faster during hormone withdrawal than in the presence of estrogen. This was first noted with the secondary stimulated chick oviduct in which hormone withdrawal caused translatable ovalbumin mRNA to decay 5-10-fold faster than predicted from the degradation half-life ( thJ measured in the presence of estrogen (Palmiter, 1973; Palmiter and Carey, 1974; Harris et al., 1975). Subsequent studies with hybridiza- tion probes confirmed this result and extended it to the estrogen-induced mRNAs for conalbumin, ovomucoid, and lysozyme (Cox, 1977; Hynes et al., 1979; Shepherd et al., 1980). Similar experiments with chick liver showed that the mRNAs for the estrogen-induced yolk proteins, apolipoprotein (apo) 11 and vitellogenin (VTG),’ were 7-8-fold more stable in the presence of estrogen than during hormone withdrawal (Wis- kocil et al., 1980). Similarly, VTG mRNA in Xenopus liver is as much as 30-fold more stable in the presence of estrogen (Brock and Shapiro, 1983). Little is known about the mech- anism through which estrogen regulates mRNA stability, but the process likely involves factors which interact with mRNA sequence or structural elements to account for the specificity and reversibility of mRNA stabilization (Brock and Shapiro, 1983; Shapiro and Brock, 1985; Shapiro et al., 1987; Brawer- man, 1987).

In the present study, we have examined the stability of apoII and VTG I1 mRNAs in chick liver in the presence of estrogen and during hormone withdrawal after various periods of hormone treatment. The results show that the degradation rates of both mRNAs during hormone withdrawal are altered dramatically by the duration of prior estrogen treatment. The degradation rates were the same in the presence and absence of estrogen for the initial 3 days during which time the fully induced steady state levels of apoII and VTG I1 mRNAs were attained. This result indicates that selective mRNA stabili- zation does not play a role in the extensive cellular accumu- lation of these mRNAs. Further hormone stimulation, how- ever, led to the accumulation of an activity that rapidly and selectively destabilized apoII and VTG I1 mRNAs upon estro- gen withdrawal. The destabilization activity was inactive in the presence of estrogen. Thus, the point of regulation by estrogen is at the level of mRNA degradation. The rapid destabilization of apoII mRNA upon hormone withdrawal was not accompanied by a change in the polysome distribution of the mRNA or an accumulation of degradation intermediates. These observations suggest that the rate-limiting step in apoII mRNA decay is an estrogen-sensitive targeting event that marks the mRNA for rapid degradation.

The abbreviations used are: VTG, vitellogenin; TAM, tamoxifen citrate.

2625

2626 mRNA Destabilization

EXPERIMENTAL PROCEDURES

Animals and Hormone Treatments-White Leghorn cockerels (2- 4 weeks old) were purchased from SPAFAS, Norwich, CT. They received food and water ad libitum. Estrogen was administered by subcutaneous implantation of constant release pellets (Innovative Research of America) containing 100 mg of 17P-estradiol. Tamoxifen citrate (TAM) was administered by intramuscular injection (100 mg/ kg) as described (Capony and Williams, 1981). Animals were killed by cervical dislocation. Liver samples were immediately frozen in liquid N, and stored at -80 "C.

RNA Preparation, Northern Blot Analysis, Dot Blot Analysis, and Polyribosome Isolation-RNA was prepared by the method of Cox (1968) with modifications (Protter et al., 1982). Since the animals were not fasted, RNA preparations contained significant amounts of glycogen which interfered with the determination of RNA concentra- tion by ultraviolet absorbance. An accurate estimate of RNA concen- tration was obtained by adjusting an expendable sample of RNA (100 pg) to 1.25% trichloroacetic acid on ice for 10 min. RNA was collected by centrifugation, the supernatant containing the glycogen was dis- carded, the pellet was washed twice with 100% ethanol, dried by lyophilization, and redissolved in 50 mM Tris, pH 8.3, and the absorbance at 260 nm was determined. For Northern blot analysis, total cellular RNA was electrophoresed on 1.2% agarose gels contain- ing 2.2 M formaldehyde, transferred to nitrocellulose, and hybridized (Thomas, 1980) with a nick-translated apoII cDNA probe (Protter et Q L , 1982). Dot blot analysis was carried out by dotting increasing amounts (up to 6 pg) of RNA onto nitrocellulose as described (White and Bancroft, 1982) and hybridizing with nick-translated probes for apoII, VTG I1 (Protter et al., 1982), or chicken serum albumin mRNAs (Elbrecht et al., 1984). After autoradiographic exposure, the dots were cut from the nitrocellulose, and radioactivity was determined by scintillation spectrometry.

Polysome profiles were obtained by a variation of the method of Palmiter (1974). Frozen liver (0.5 g) was homogenized in 4.5 ml of 25 mM Tris-HC1, pH 7.5, 25 mM NaCl, 5 mM MgCl,, 2% Triton X-100, 1 mg/ml sodium heparin in a 10-ml Dounce homogenizer. The ho- mogenate was centrifuged for 5 min at 17,000 X g, and 5 AZ60 nm,l Cm

units of supernatant were loaded on a linear 0.5-1.5 M sucrose gradient containing 25 mM Tris-HC1, pH 7.5, 25 mM NaCl, 0.2 mg/ ml sodium heparin, 2.5 mM MgCl,. Gradients were centrifuged at 4 "C for 2.5 h at 39,000 rpm in an SW-41 rotor (Beckman). The absorbance profile at 260 nm was determined with an ISCO UA-5 absorbance detector. For analysis of apoII mRNA, 18 fractions of 0.6 ml were collected and assayed directly by dot blot analysis (White and Ban- croft, 1982)

DNA-excess Solution Hybridization AssQ~s-A~oII mRNA was measured with a DNA-excess solution hybridization assay using a single-stranded 240-nucleotide cDNA fragment as described (El- brecht et Q L , 1984, Williams et al., 1986). A similar assay was con- structed to measure apoII mRNA nuclear precursors by employing a single-stranded hybridization probe derived from the first intron of the apoII gene. The probe was synthesized from subclone mp9apoIIABP (Shelness and Williams, 1984) by priming DNA syn- thesis with the M13 universal primer followed by digestion at a unique BstEII site located just 3' to the exon I-intron A boundary. The probe corresponds to nucleotides 101 to 235 according to the numbering system of van het Schip et al. (1983). The single-stranded intron probe was purified by gel electrophoresis and adsorption to hydroxylapatite as described (Williams et al., 1986). Hybridization assays were exactly as described (Newman et al., 1985, Williams et al., 1986). A DNA-excess solution hybridization assay for chicken apoB mRNA was carried out in the same fashion using a 233- nucleotide single-stranded apoB cDNA probe (Kirchgessner et al., 1987).

Nuckar Run-on Transcription Assays-Groups of four chickens were treated with estrogen as described above for 1 day or 14 days after which liver nuclei were prepared (Evans et al., 1981). Nuclei equivalent to 100 p g of DNA were incubated exactly as described (Evans et Q L , 1981) to allow nascent RNA chains to extend in the presence of [a-32P]UTP, and RNA was isolated as described by McKnight and Palmiter (1979). Radiolabeled RNA was hybridized (McKnight and Palmiter (1979) to nitrocellulose filters containing 1 pg of pBR322 DNA or a 500-base-pair fragment of apoII cDNA (corresponding to nucleotides -10 to 543 of apoII mRNA) (Wieringa et al., 1981) derived from an SP6 clone kindly supplied by Dr. John Burch (Institute for Cancer Research, Philadelphia, PA). The same clone was used to synthesize an apoII mRNA hybridization standard

radiolabeled with L3H]UTP (Melton et Q L , 1984) and to synthesize large quantities of unlabeled apoII mRNA to be used as a competitor in the hybridization assays. Filters containing apoII DNA or pBR322 DNA were incubated with [32P]RNA in the presence of [3H]apoII mRNA which served as an internal standard for hybridization effi- ciency. After hybridization, ribonuclease treatment, and washing (McKnight and Palmiter, 1979), filters were counted by liquid scin- tillation spectrometry. Relative transcription rates were expressed as parts per million of total [32P]UTP incorporation. Raw data were corrected by subtraction of radioactivity on the filters containing pBR322 DNA, normalization to the size of the apoII gene (2.9 kilobase pairs) compared to the cDNA probe (0.5 kilobase pair), and correction for the hybridization efficiency (25%) determined with the [3H]apoII mRNA internal standard. ParalIel hybridizations which also con- tained unlabeled SP6-derived apoII mRNA (1 pg) as competitor were used to confirm the specificity of the hybridization signal.

RESULTS

Half-times for mRNA Decay Determined from the Approach to Steady State-In order to determine the mRNA decay constants from approach to steady state data, chickens were treated with estrogen by subcutaneous implantation of con- stant release pellets containing 17P-estradiol. At various times after implantation, liver RNA was prepared and assayed for apoII mRNA by DNA-excess solution hybridization and for VTG 11 mRNA by dot blot analysis. Fig. 1A shows that apoII mRNA increased from a basal level of 0.5 mRNA copies/cell to a steady state level of approximately 15,000 mRNA copies/ cell. The steady state level was attained at about 2 days of hormone treatment and was maintained for up to 15 days. The accumulation of VTG I1 mRNA (Fig. 1B) showed very similar kinetics. The tl,* values for the degradation of these mRNAs were estimated from the approach to steady state by the method of Berlin and Schimke (1965) as elaborated by Watson et al. (1981). For this purpose, the steady state was taken as the mean of the mRNA values between 2 and 14 days of hormone treatment. Fig. 2 shows that the accumula- tions of apoII and VTG I1 mRNAs are described by the same line which yields a first order decay constant of 0.055 hr" and a tl,, of 12.6 h.

Half-times for mRNA Decay upon Hormone Withdrawal- Messenger RNA turnover following estrogen withdrawal was initiated by removal of hormone pellets and administration of the anti-estrogen, TAM, by intramuscular injection. TAM appears to exert its antiestrogenic effects by competitively inhibiting the binding of estrogen to its receptor (Capony and Rochefort, 1978; Katzenellenbogen, 1980) and should serve to rapidly terminate the activity of the estrogen-receptor com- plex and eliminate the effects of residual hormone in the animal. TAM is usually characterized as a partial agonist or a mixed agonist-antagonist which displays some degree of estrogenic activity depending on the animal species. In the chicken oviduct (Sutherland et al., 1977) and liver (Capony and Williams, 1980, 1981), TAM has been characterized as a pure antagonist although further studies showed it to have very limited agonist activity of less than 0.1% compared to 17P-estradiol (Blue and Williams, 1981). Previous studies (Shepherd et al., 1980; Capony and Williams, 1981) and additional experiments described below substantiate this point and show that for practical purposes TAM can be used as an antagonist to rapidly interrupt the action of estrogen.

Fig. 3 shows the decay of apoII and VTG I1 mRNAs following hormone withdrawal and TAM administration after 14 days of estrogen treatment. Both messages decayed rapidly with no apparent lag period and with t, values of approxi- mately 1.5 h. In contrast to this result, Fig. 4 shows that apoI1 and VTG I1 mRNAs decayed very slowly when hormone was withdrawn after 1 day of estrogen treatment. In this case, the t3,> values for both mRNAs were approximately 13 h. When

mRNA Destabilization 2627

TIME AFTER ESTROGEN IMM cows)

6

w.w- 0 2 4 8 a I O 12 14 16

TIME F~FTER ESTROGEN IMPLFNT (wys) FIG. 1. Approach to steady state of apoII and VTG I1

mRNAs. Chickens were treated with estrogen by subcutaneous im- plantation of constant release pellets containing 100 mg of 17p- estradiol. At the indicated times, livers were removed, and total RNA was prepared. ApoII mRNA ( A ) and VTG I1 mRNA ( B ) were meas- ured as described under “Experimental Procedures.” Each data point is the mean f S.E. of values from 4-12 animals.

these data (Fig. 4) were displayed as the log of the mRNA remaining uersus time, a linear least squares fit yielded a correlation coefficient of 0.99, indicating that mRNA decay approximates a first order process (not shown). This was also seen with the data in Fig. 3 in which case the correlation coefficient was 0.90.

The effects of hormone withdrawal and TAM administra- tion on nuclear events was tested by monitoring the levels of apoII mRNA precursors. For this purpose, a DNA-excess solution hybridization assay was constructed with a single- stranded probe derived from intron A of the apoII gene. As shown in Fig. 5, intron A-containing precursors were present at about 75 molecules/cell or approximately 0.6% of the level of mature apoII mRNA after 1 day of estrogen treatment. After pellet removal and TAM administration, apoII pre- mRNA decreased rapidly with a tlh of about 15 min and was near background levels by 90 min. This result indicates that pellet removal and TAM administration quickly and effec- tively terminated the production of apoII mRNA precursors.

* RPO I 1 mRNA + VTG. I I mRNR * RPO I 1 mRNA + VTG. I I mRNR

0.00 rt\ 0

L (L \

-0.25 - L2I

C

-0.50 -

J O 2 4 6 e IO 12 14

TIME AFTER ESTROGEN IMPLANT (HRS)

FIG. 2. Determination of rate constants for apoII and VTG I1 mRNA degradation in the presence of estrogen. Data from the ascending portion of the approach to steady state curves in Fig. 1 were plotted according to the method of Watson et al. (1981) to obtain the degradation rate constants for apoII and VTG I1 mRNAs.

A APO I I mRNR o VTG I 1 mRNR

0-4 ’ I 0 6 12 18 24

TIME AFTER ESTROGEN WITHORAWAL (HRS)

FIG. 3. Decay of apoII and VTG I1 mRNAs after 14 days of estrogen pretreatment. Chickens were treated for 14 days with estrogen by subcutaneous implantation of constant release pellets as described in Fig. 1. Estrogen with withdrawn by pellet removal and TAM administration (100 mg/kg), and liver RNA was prepared at the indicated times. ApoII (A) and VTG I1 (0) mRNAs were measured by solution hybridization and dot blot analysis, respectively. Data are expressed as the percent of control f S.E.

Thus, the slow decay of mature apoII mRNA following hor- mone withdrawal after 1 day of estrogen treatment (Fig. 4) was not due to continued mRNA production.

The results shown in Figs. 3 and 4 suggest that apoII and VTG I1 mRNAs decay in a slow mode or in a fast mode depending upon the duration of prior estrogen treatment. In order to examine the transition from the slow to the fast decay mode, animals were treated with estrogen for various times prior to hormone withdrawal. ApoII and VTG I1 mRNAs were measured at 3 h after withdrawal, a time interval that readily distinguishes the two modes of decay. As shown in Fig. 6, mRNA decay was unchanged at 1 and 3 days of estrogen treatment but increased thereafter until it was max- imal by 7 days. No further change was seen between 7 and 14 days.

These results suggest that estrogen induced a new or en-

2628 mRNA Destabilization

‘OWL A APO I 1 mRNA

0 0 6 12 18 24

TIME RFTER ESTROGEN WITHDRAWAL (HRS)

FIG. 4. Decay of apoII and VTG I1 mRNAs after estrogen pretreatment for 1 day. Chickens were treated for 1 day with estrogen prior to pellet removal and TAM administration. Other details as in Fig. 3.

Y U L

R a I \ I

0 r j0l 4 0 b 0 30 60 go

TIME RFTER ESTROGEN WITHDRAWAL (MIN)

FIG. 5. Decay of apoII pre-mRNA during hormone with- drawal. Chickens were treated with estrogen for 1 day prior to pellet removal and TAM administration as in Fig. 3. ApoII pre-mRNA was measured by solution hybridization with an intron A probe as de- scribed under “Experimental Procedures.” Data points are the mean -+ S.E. of values from four animals.

hanced capacity to destabilize apoII and VTG I1 mRNAs. The induced destabilization activity appeared to be inactive in the presence of the hormone, however, since the steady state mRNA levels were maintained (Fig. 1) throughout the period when the destabilization activity accumulated (Fig. 6). An alternate explanation is that transcription increased concom- itantly with the destabilization activity to maintain the steady state. In order to distinguish between these possibilities, nu- clear run-on transcription assays were carried out after l day and 14 days of estrogen treatment. Previous studies have shown that transcription of the apoII gene is completely dependent upon estrogen treatment (Strijker et al., 1986). The results in Table I show that the relative apoII transcription rate was the same at 1 day and 14 days of estrogen treatment. This result indicates that the induced destabilization activity was quiescent until estrogen was withdrawn.

Specificity of the Destabilization Actiuity-The specificity of the estrogen-induced destabilization activity was tested by monitoring apoB and serum albumin mRNAs. Hepatic apoB

0 RPO I I mRNA El VTG I I rnRNA

TIME OF ESTROGEN TREATMENT (DAYS)

FIG. 6. Accumulation of apoII mRNA and VTG I1 mRNA destabilization activity. Chickens (eight/group) were treated with estrogen by pellet implantation for the indicated times after which pellets were removed and TAM administered to half the chickens in each group. Animals were killed 3 h later, liver RNA was prepared, and apoII and VTG I1 mRNAs were assayed by solution hybridization and dot blot analysis, respectively. Bars represent the mean (f S.E.) percent of each mRNA remaining compared to control animals not withdrawn from hormone.

TABLE I Relative transcription rate of the apoIZ gene

Relative transcription rates are expressed as parts per million of total [32P]UTP incorporation into RNA in the nuclear run-on assay as described under “Experimental Procedures.” Each value is the mean f S.E. of assays from four chickens.

Treatment -Competitor +Cornpetitof Net

ppm * S.E. Estrogen (1 day) 870 f 192 252 + 12 618 Estrogen (14 days) 846 f 54 258+ 108 588

“These values were obtained in parallel hybridizations which contained 1 pg of SP6-derived apoII mRNA as competitor in addition to the [32P]RNA.

synthesis exhibits both estrogen-independent and estrogen- stimulated components (Capony and Williams, 1980, 1981). ApoB synthesis and apoB mRNA (Kirchgessner et aL, 1987) increase rapidly after hormone treatment and reach a maxi- mum between 12 and 24 h. TAM completely inhibits this increase and rapidly reverses (tLh = 3 h) estrogen-stimulated apoB synthesis when administered subsequent to estrogen (Capony and Williams, 1981). The effect of estrogen treat- ment on apoB mRNA turnover was tested after 1 day and 14 days of estrogen treatment by measuring apoB mRNA at 2 h after hormone withdrawal and TAM administration. The amount of apoB mRNA remaining at 2 h after withdrawal was not substantially different after 14 days (69 k 5% of control) or 1 day (60 f 9% of control) of prior estrogen treatment. This result indicates that apoB mRNA decay was not markedly enhanced by estrogen treatment as occurred with apoII and VTG I1 mRNAs.

Serum albumin mRNA is not acutely regulated by estrogen in the chicken although a slow decrease in serum albumin mRNA accompanies prolonged hormone treatment (Williams et al., 1978). In order to test the effects of hormone withdrawal on albumin mRNA, two groups of animals were treated with hormone implants for 2 days or 15 days. Within each group, one set of animals was withdrawn from hormone by implant removal and TAM administration 1 day prior to death. Al- bumin mRNA was measured by dot blot analysis. As shown

mRNA Destabilization 2629

in Table 11, hormone withdrawal had no effect on the levels of albumin mRNA in either group.

Characteristics of ApoII mRNA Decay-ApoII mRNA was examined by Northern blot analysis a t steady state and at various times after hormone withdrawal. Fig. 7 shows that apoII mRNA was present as a single hybridizing band after 14 days of estradiol treatment (lunes A-C) or after 14 days of estradiol followed by withdrawal for 2 (lunes D-F) or 24 h (lunes G and H ) . Smaller apoII mRNA fragments were not detected even with prolonged autoradiographic exposures. Similar results were obtained with RNA samples from animals withdrawn from hormone after 1 day of estrogen treatment (data not shown). Thus, with the sensitivity provided by Northern blot analysis, an accumulation of mRNA degrada- tion intermediates was not seen whether apoII mRNA decayed rapidly or slowly.

TABLE I1 Effect of hormone withdrawal on serum albumin mRNA

Groups of four animals were treated with estrogen implants for the indicated times. Animals were withdrawn from hormone by implant removal and TAM treatment 1 day prior to death. Albumin mRNA was measured by dot blot analysis of 3 Kg of total liver RNA as described under “Experimental Procedures.”

Treatment Albumin mRNA

cpm & S.E. Estrogen (2 days) 4640 f 210 Estrogen (1 day/withdrawn 1 day) 4900 f 290 Estrogen (15 days) 3400 f 550 Estrogen (14 days/withdrawn 1 day) 3330 f 290

A B C D E F G H

1 “.

1 ” v-

FIG. 7. Northern blot analysis of apoII mRNA after 14 days of estrogen treatment and following hormone withdrawal. The gel was loaded with total RNA from animals treated with estrogen for 14 days (lanes A-C) or treated with estrogen for 14 days followed by pellet removal and TAM treatment for 2 h (lunes D-F) or 24 h (lanes G and H ) . The amount of RNA run in each lane was adjusted to give approximately the same amount of apoII mRNA except in lane H which contained twice as much apoII mRNA.

The distribution of apoII mRNA in polysomes was exam- ined by dot blot analysis of sucrose gradient fractions after sedimentation of liver cytosol extracts. Fig. 8 shows that apoII mRNA was found almost exclusively in polysomes after 1 day or 14 days of estrogen treatment. A similar distribution of apoII mRNA was seen at 2 h and 24 h after hormone with- drawal in animals that received either 1 day or 14 days of estrogen treatment (Fig. 8). These data indicate that there is no substantial shift of apoII mRNA from polysomes to the free mRNP fraction when hormone withdrawal initiates either the slow mode or rapid mode of mRNA decay. Minor differences in the gradient distribution of apoII mRNA were not reproducible and presumably reflect procedural variation. Variable amounts of apoII mRNA (5-896 of the total) seen in the top gradient fractions (less than 80 S ) were degraded as judged by Northern analysis (data not shown). Since degraded apoII mRNA was not detected in RNA purified directly from liver by guanidine HCl extraction (Fig. 7), degraded apoII mRNA in the gradient profiles most likely reflects procedural degradation and not in vivo degradation intermediates.

DISCUSSION

The results of this study show that the degradation rate of apoII mRNA is altered by the duration of estrogen treatment. When hormone was withdrawn after short periods of estrogen treatment, the mRNA decayed slowly with a tM = 13 h. This tl/A is the same as determined by monitoring the approach of apoII mRNA to the new steady state during continuous estro- gen treatment. When estrogen was administered for 14 days, however, hormone withdrawal resulted in a much more rapid mRNA decay characterized by a t,,* = 1.5 h. Measurements of mRNA decay after intermediate periods of estrogen treatment showed a gradual shift from the slow mode to the rapid mode of decay. These data indicate that estrogen induced an activity responsible for the rapid destabilization of apoII mRNA. The same pattern of induced destabilization was seen with VTG I1 mRNA, another estrogen-induced yolk protein mRNA. In contract, hormone withdrawal had little effect on serum al- bumin mRNA after either short or long term estrogen treat- ment. The duration of estrogen treatment also had no major effect on the turnover of apoB mRNA, an mRNA that is made in the absence of hormone and at higher levels after estrogen treatment (Capony and Williams, 1980,1981; Kirch- gessner et al., 1987). Thus, the induced destabilization activity appears to be specific for apoII and VTG I1 mRNAs.

The induced destabilization activity appears to be quiescent in the presence of estrogen and is only activated following hormone withdrawal. This conclusion follows from the obser- vations that the steady state level of apoII mRNA was con- stant over a 14-day period, and the relative transcription rate of the apoII gene was the same after 1 and 14 days of estrogen treatment. If the induced destabilization activity were func- tional in the presence of estrogen, a 6-10-fold increase in the transcription rate would be needed to maintain the steady state of mature apoII mRNA. As judged by nuclear run-on transcription measurements, this is not the case.

Estrogen is generally considered to selectively stabilize induced mRNAs against degradation (Darnell, 1982; Shapiro and Brock, 1985; Darnell et al., 1986; Shapiro et al., 1987). As a result of mRNA stabilization, estrogen-induced mRNAs accumulate to much higher levels in the cell, and the encoded proteins are made a t much greater rates than would occur on the basis of transcriptional activation alone. This conclusion has been reached on the basis of observations that estrogen- induced mRNAs decay faster after hormone withdrawal than in the presence of estrogen (Palmiter and Carey, 1974; Cox,

2630

A mRNA Destabilization

B

24 Hr E - O . D . 260 na

a

w

4 6 8 10 12 1 4

0

N U

I u a - U m

t- 0

w

35 24 Hr E - 2 Hr Tam

*PO I1 n F w A - O . D . 260 nm

24 Hr E - 24 Hr Tam - O . D . 260 na

u, I I I

2 4 6 E 10 12 14 Bc t ion No.

m U 0 I-

m

2 Wks E 3) I - A D O I1 -A - 0 . D . 260 nm

2 Wks E - 2 Hr Tam 3) 1 C A P O -*

- 0 . 0 . 260 nm

4

U m

I- O

m

/ 8 4

0 2 4 6 'e i O j 2 i 4 F r a c t i o n No.

2 Wks E - 24 Hrs Tam - O.D. 260 nm

~~

a i \

P-

u-

10-

S-

O -; 1 1

F r a c t i o n No. 0 2 4 6 'e i o i 2 ' 4

FIG. 8. Distribution of apoII mRNA in polyribosomes during estrogen treatment and after hormone withdrawal. Polyribosomes prepared from liver after 1 day (column A ) or 14 days (column B ) of estrogen treatment alone (top profiles) or after pellet removal and TAM administration for 2 h (middle profiles) or 24 h (bottom profiles) were analyzed by sucrose gradient sedimentation (left to right) as described under "Experimental Procedures." The scale of the absorbance profile at 260 nm (solid line) is 1 absorbance unit from the bottom to the top of each graph. The apoII mRNA distribution was determined by dot blot analysis of gradient fractions. Each data point represents the percent of total apoII mRNA radioactivity summed over the entire profile.

mRNA Destabilization 2631

1977; Hynes et al., 1979; Shepherd et al., 1980; Wiskocil et al., 1980, Brock and Shapiro, 1983). In some cases, it was noted that the data also were consistent with the possibility that estrogen might regulate a degradative capacity (Palmiter and Carey, 1974; Shepherd et al., 1980). The results of the present study show that the turnover rates for apoII and VTG I1 mRNAs during hormone withdrawal are altered by the dura- tion of prior estrogen exposure. This result has two important consequences. First, mRNA turnover rates were the same in the presence and absence of estrogen for the first 3 days of estrogen treatment during which time the steady state levels of apoII and VTG I1 mRNAs were attained. This result indicates that selective mRNA stabilization does not play a role in the extensive cellular accumulation of apoII and VTG I1 mRNAs. The extensive accumulation of these mRNAs appears to be due solely to enhanced transcription and the moderate stability (tXh = 13 h) of apoII and VTG I1 mRNAs. Second, estrogen regulates a destabilization activity that is specific for apoII and VTG I1 mRNAs, but not for all estrogen- regulated mRNAs. This is conceptually and mechanistically different than the idea that estrogen stabilizes these mRNAs against a degradation mechanism that is common to many mRNAs.

The extent to which the present results can be generalized to other estrogen-regulated systems and mRNAs is not known. Previous results obtained for egg white mRNAs in chick oviduct (op. cit .) are consistent with the idea that estrogen regulates mRNA destabilization as occurs in chick liver. In Xenopus liver, VTG mRNA is degraded about 30 times faster during hormone withdrawal than in the presence of estrogen (Brock and Shapiro, 1983). These data are also consistent with the current findings in chick liver.

A possible functional role for this estrogen-induced mRNA destabilization activity is to rapidly terminate the synthesis of the apoII and VTG I1 proteins when the estrogen level drops at the end of a laying cycle in the hen. Since as much as 25% of liver protein synthesis is devoted to apoII and VTG I1 (Chan et al., 1976 Jost et al., 1978, Williams et al., 1978), the rapid termination of estrogen-induced protein synthesis could remove an unnecessary demand on the liver thereby increasing its capacity to carry out other essential functions. Consistent with this possibility is that estrogen-induced apoB synthesis, which accounts for an additional 10-15% of liver protein synthesis, rapidly (t!,? = 3 h) returns to its uninduced level following hormone withdrawal (Capony and Williams, 1981).

Little is known about the mechanisms of mRNA degrada- tion or how specific mRNAs can be selectively targeted for decay. Irrespective of whether apoII mRNA was subject to the slow or rapid mode of decay, an accumulation of degra- dation intermediates was not detected by Northern analysis following hormone withdrawal. Similarly, apoII mRNA was present almost exclusively in polysomes in the presence of estrogen or after hormone withdrawal. These observations suggest that the rate-limiting step in apoII mRNA degrada- tion is a targeting event that marks the mRNA for degrada- tion. Once targeted for degradation, apoII mRNA decays very rapidly without an increase in the steady state level of deg- radation intermediates. The targeting event may be the step in apoII mRNA degradation that is regulated by estrogen. We have recently identified hypersensitive sites for apoII mRNA

degradation in polyribosomes in. uitro (Shelness et al., 1987). A set of cleavages occurs in the 3‘ noncoding region at 5’- AAU-3’ trinucleotides. These cleavage sites map to single- stranded bulge and interior loops in a secondary structure model of apoII mRNA (Shelness and Williams, 1985). S1 nuclease protection assays show that apoII mRNA molecules cleaved selectively at these sites are also found at low levels in RNA purified directly from liver by guanidine HC1 extrac- tion.2 Studies are in progress to determine whether endonu- cleolytic cleavage at these sites targets apoII mRNA for rapid degradation and whether estrogen regulates this event.

Acknowledgements-Excellent technical assistance was provided by Penelope Strockbine. We thank Dr. John Burch for SP6 subclone containing apoII cDNA sequences.

REFERENCES Berlin, C. M., and Schimke, R. T. (1965) Mol. Phrmacol. 1 , 149-156 Blue, M. L., and Williams, D. L. (1981) Biochem. Biophys. Res. Commun. 9 8 ,

Brawerman, G. (1987) Cell 48,5-6 Brock, M. L., and Shapiro, D. J. (1983) Cell 3 4 , 207-214 Capony, F., and Rochefort, H. (1978) Mol. Cell. Endocrinol. 11 , 181-198 Capony, F., and Williams, D. L. (1980) Biochemistry 19,2219-2225 Capony, F., and Williams, D. L. (1981) Endocrinology 108 , 1862-1868 Chan, L., Jackson, R. L., OMalley, B. W., and Means, A. R. (1976) J . Clin.

Cox, R. A. (1968) Methods Enzymol. 12B, 120-129 Cox. R. F. (19771 Biochemistrv 16.3433-3443

785-791

Znuest. 5 8 , 368-379

Darhell, J. E:, Jr. (1982) Natire 297, 365-371 Darnell, J., Lodish, H., and Baltimore, D. (1986) Molecular Cell Biology, p.490,

Elbrecht, A,, Lazier, C. B., Protter, A. A,, and Williams, D. L. (1984) Science

Evans, M. I., Hager, L. J., and McKnight, G. S. (1981) Cell 2 5 , 187-193 Harris, S. E., Rosen, J. M., Means, A. R., and O’Malley, B. W. (1975) Biochem-

Scientific American Books, New York

225,639-641

Hynes, N. E., Groner, B., Sippel, A. E., Jeep, S., Wurtz, T., Nguyen-Huu, M.

Jost. J.-P.. Ohno. T.. Panyim. S.. and Schuerch. A. R. (1978) Eur. J . Biochern.

istry 14,2072-2081

C., Giesecke, K., and Schutz, G. (1979) Biochemistry 18, 616-624 . . - .

Katzenellenbogen, B. S. (1980) Annu. Reu. Physiol. 4 2 , 17-35 Kirschgessner, T. G., Heinzmann, C., Svenson, K. L., Gordon, D. A,, Nicosia,

McKnight, G. S., and Palmiter, R. D. (1979) J. Biol. Chem. 254,9050-9058 Melton, D. A,, Krieg, P. A,, Rebagliati, M. R., Maniatis, T., Zinn, K., and

Newman, T. C., Dawson, P. A,, Rudel, L. L., and Williams, D. L. (1985) J. Biol.

Palmiter, R. D. (1973) J. Biol. Chem. 248,8260-8270 Palmiter, R. D. (1974) Biochemistry 13,3606-3615 Palmiter, R. D., and Carey, N. H. (1974) Proc. Natl. Acad. Sci. U. S. A. 71,

Protter, A. A,, Wang, S.-Y., Shelness, G. S., Ostapchuk, P., and Williams, D.

Shapiro, D. J., and Brock, M. L. (1985) in Biochemical Actions of Hormones

Shapiro, D. J., Blume, J. E., and Nielsen, D. A. (1987) BioEssays 6 , 221-226

Shelness, G. S., and Williams, D. L. (1985) J. Biol. Chem. 2 6 0 , 8637-8646 Shelness, G. S., and Williams, D. L. (1984) J. Biol. Chem. 2 5 9 , 9929-9935

Shelness, G. S., Binder, R., Hwang, S.-P. L., MacDonald, C., Gordon, D. A,, and Williams, D. L. (1987) in Molecular Biology of RNA: New Perspectives (Inouye, M., and Dudock, B., eds) Academic Press, Orlando, FL, in press

Sheuherd. J. H.. Mulvlhill. E. R.. Thomas. P. S.. and Palmiter. R. D. (1980) J.

84,3551361

M., Lusis, A. J., and Williams, D. L. (1987) Gene (Amst.), in press

Green, M. R. (1984) Nucleic Acids Res. 12,7035-7056

Chem. 260,2452-2457

2357-2361

L. (1982) Nucleic Acids Res. 10, 4935-4950

(Litwak, G., ed) Vol. 12, pp. 139-172, Academic Press, Orlando, FL

, , Ckll B i d 87,’142-153

Gene (Amst.) 45,27-35

, I . .

Strijker, R., Vanassen, G. B., Dikkesch, B. D., Gruber, M., and AB, G. (1986)

Sutherland, R., Mester, J., and Baulieu, E.-E. (1977) Nature 267,434-435 Thomas, P. S. (1980) Proc. Natl. Acad. Sci. U. S. A. 77,5201-5205 van het Schip, A. D., Meijlink, F. C. P., Strijker, R., Gruber, M., van Vliet, A.

J., van de Klundert, J. A. M., and AB, G. (1983) Nucleic Acids Res. 11,2529- 2540

Watson. G.. Davev. R. A.. Labarca. C.. and Paieen. K. (1981) J. Biol. Chern. 266,3005-301f

. . - .

White, B. A,, and Bancroft, F. C. (1982) J. Biol. Chem. 257,8569-8573 Wieringa, B., AB, G., and Gruber, M. (1981) Nucleic Acids Res. 9,489-499 Williams. D. L.. Wane S.-Y.. and Klett. H. (1978) Proc. Natl. Acad. Sci. U. S.

A. 75,5974-5978 ” ’

Methods Enzymol. 129, 671-689

Deeley, R. G. (1980) Proc. Natl. Acad. Sci. U. S. A. 77,4474-4478

. . .

Williams, D. L., Newman, T. C., Shelness, G. S., and Gordon, D. A. (1986)

Wiskocil, R., Bensky, P., Dower, W., Goldberger, R. F., Gordon, J. I., and

R. Binder and D. L. Williams, unpublished results.