vitellogenin locus

Proc. Nati. Acad. Sci. USA Vol. 81, pp. 5777-5781, September 1984 Developmental Biology

Injection of partially purified estrogen receptor protein from Xenopus liver nuclei into oocytes activates the silent vitellogenin locus (steroid hormone/affinity chromatography/RNA blot hybridization/gene regulation)

JOHN KNOWLAND*t, IRENE THEULAZ*, CHRISTOPHER V. E. WRIGHTt, AND WALTER WAHLI* *Institut de Biologie animale, Universitt de Lausanne,

Batiment de Biologie, CH-1015 Lausanne, Switzerland

Communicated by J. B. Gurdon, May 21, 1984

ABSTRACT Injection of extracts from Xenopus liver nuclei that are enriched 2000 times in estradiol receptor into Xenopus oocytes induces transcription of the silent vitellogenin locus, which is activated in liver by estradiol, but not of the albumin locus, which is active in liver but suppressed by high levels of estradiol. Transcription initiates within the 5'-end region of the gene we have studied and probably continues into the 3' third. The activation seems to be very efficient, but most of the primary transcripts are probably rapidly and inaccurately processed. New proteins are also made and secreted by the oocytes.

It has long been suspected that steroid hormones activate genes by first binding to receptor proteins, forming complexes that then induce specific gene transcription. In Xenopus liver, estrogen produces the protein vitellogenin, which is the precursor of the egg-yolk proteins, by activating the viteilogenin locus (for reviews see refs. 1 and 2). The locus consists of four genes. There are two in group A (Al and A2), which have 95% base sequence homology, and two in group B (BJ and B2), which also have 95% sequence homology, While the genes in group A have about 80% homology with those in group B (1). There is circumstantial evidence that an estrogen receptor in the liver nuclei (3-5) is responsible for the activation of the locus, because its properties account for the main features of the activation by estradiol (5). As we have previously suggested (6), a possible test of the ability of the receptor to activate vitellogenin genes would be to inject it into Xenopus oocytes, in which the vitellogenin genes are inactive. Here we have injected a partially purified receptor preparation from Xenopus liver nuclei into oocyte cytoplasm, and we have tested for synthesis of new protein and for transcription of vitellogenin genes.

lulose filters that were treated according to Thomas (10). The yield was 40-50 pg from 10 oocytes; gel lanes typically contained RNA from 2.5 oocytes. Liver RNA was prepared according to Wahli et al. (11). Hybridization. Filters were incubated for 6-15 hr in 5 x SSPE (SSPE is 0.18 M NaCl/10 mM NaPO4/1 mM EDTA, pH 7.5), 50% (vol/vol) formamide, 1.1% NaDodSO4, denatured salmon sperm DNA at 200 pg/ml, 5 x Denhardt's solution and then hybridized with nick-translated probes for 65 hr at 420C with gentle agitation in a solution containing S x SSPE, 50% (vol/vol) formamide, 0.1% NaDodSO4, denatured DNA at 100-200 Ag/ml, and lx Denhardt's solution. Except where indicated, filters were washed for 2.5 hr at 420C in 2x NaCl/Cit (NaCl/Cit is 0.15 M NaCl/0.015 M trisodium citrate) and 0.1% NaDodSO4 containing denatured DNA at 10 ,.g/ml. Probes were removed before hybridization with a second probe by washing for 1-2 hr at 720C in 0.05% sodium pyrophosphate/0.2 mM EDTA/5 mM Tris-HCl, pH 8.0/0.1x Denhardt's solution, followed by rinsing in the same solution. Probes had specific activities of 0.7-1.5 x 108 cpm/,ug and were used at 1-2 x 106 cpm per ml. This procedure allows detection of picogram amounts of specific RNA (10).

RESULTS

from liver nuclei by salt extraction, ammonium sulfate precipitation, and affinity chromatography (5), giving approximately 2000-fold purification. Injection of Oocytes. Receptor preparations were thawed slowly, dialyzed for 10 min at 4°C against saline containing 5 nM estradiol (giving 90% saturation of receptor), and kept on ice. They were injected by standard procedures (7) into the cytoplasm of stage 6 oocytes, which were incubated at 19°C in saline containing 5 nM estradiol. For each sample, 40-50 oocytes were injected. Extraction and Fractionation of RNA. RNA extracted from oocytes (8) was treated with glyoxal for 10 min at 50°C, fractionated on 0.8% agarose gels (9), and transferred to nitrocel-

New Proteins Are Synthesized After Injection of Receptor Preparations into Oocytes. We injected oocytes in the cytoplasm with receptor preparations and incubated them in a physiological concentration of estradiol. Each oocyte received approximately 500,000 molecules of receptor, substantially more than one liver nucleus contains [100 in normal males and 1000 in estrogen-stimulated males (3)]. We found changes in protein synthesis, confined to later times and to the secreted proteins (Fig. 1), which should include any newly synthesized vitellogenin (12). The largest new protein runs on gels just ahead of authentic vitellogenin. In some tests, although not in all, antibodies against purified vitellogenin precipitated some of these proteins, the largest one included, raising the possibility that receptor preparations can activate the vitellogenin genes in oocytes. We therefore tested directly for RNA sequences derived from the vitellogenin genes. Estradiol Receptor Preparations Activate Transcription of Vitellogenin Sequences in Oocytes. We injected receptor preparations, extracted RNA, fractionated it, and probed for vitellogenin transcripts, using a mixture of two cDNA probes representing the Al and BJ genes (Fig. 2). Neither probe contained any plasmid sequences. We used a preparation of vitellogenin RNA from liver to indicate the position of full-length, mature mRNA (6.3 kb). Filters were washed at

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Abbreviation: kb, kilobase(s). tPermanent address: Department of Biochemistry, South Parks Road, Oxford OX1 3QU, UK.

MATERIALS AND METHODS Estradiol Receptor. Estradiol-receptor complex (3 x 1010 binding sites per ml; 8 ,ug of protein per ml) was prepared

5777

5778

Proc. NatL Acad Sci. USA 81

Developmental Biology: Knowland et al.

(1984)

1 2 3 4 5 6 7 8 9 10 11 121314 V

24 96 R2 R5 R2 R5

___~0

-

6.3

-

3 .2

- 1.8

s-* V

FIG. 3. Transcription of vitellogenin sequences in oocytes injected with receptor preparations. Oocytes were injected with saline and incubated for 0, 24, 48, 72, and 96 hr (lanes 1-5); with an active receptor preparation, and incubated for the same times (lanes 7-11); or with one that had been heated at 100'C for 10 min, and incubated for 72 or 96 hr (lanes 13 and 14). Lanes 6 and 12 wereempty. The probes were the EcoRI fragments of the Al and B) cDNA clones pXlvc 23 and pXlvc 10 (Fig. 2). V, vitellogenin mRNA. RNA lengths are given on the right in kb. Exposure time, 3 weeks.

FIG. 1. Newly synthesized proteins secreted from oocytes injected with receptor preparations. Fifteen oocytes were labeled in 50 lu containing 35 ,Ci (1 Ci = 37 GBq) of [35Slmethionine (>800 Ci/mmol; Amersham) for 16 hr starting at either 24 or 96 hr after injection. The labeled media were analyzed by using 7% gels (6) and fluorography; exposure time, 3 weeks. Two preparations of receptor (R2 and R5) were used, either in active form (+) or after inactivation by heating at 100°C for 10 min (-). The position of authentic vitellogenin (Vg) is shown.

relatively low stringency to allow detection of transcripts from any of the four vitellogenin genes (13). Fig. 3 shows that no signals are detected when oocytes are injected with saline (lanes 1-5). In oocytes injected with receptor preparations, vitellogenin sequences are not detectable either immekb

A group 2

''

mRNA 5 p

4

3

Xlvc 23 probes

Al

B group 2

mRNA 5

4

1

|

diately Ufter injection (lane 7) or at 24 hr (lane 8), but fragments approximately 3.2 and 1.8 kb long are detectable at 48 hr (lane 9), and the signals are stronger at 72 and 96 hr (lanes 10 and 11). Transcripts shorter than 1.8 kb were sometimes detected after the longer incubation times (lane 11). Boiled receptor preparations do not induce synthesis of the transcripts at either 72 hr (lane 13) or 96 hr (lane 14), showing that induction of vitellogenin sequences requires native receptor preparations. Treatment with alkali confirmed that the signals detected were due exclusively to RNA (data not shown). Transcription Induced by Receptor Preparations Produces Authentic Vitellogenin Sequences That Lack a Pbly(A) Terminus. We decided to focus on one of the four genes, b2, which has been characterized in detail (14). The 3.2- ahd 1.8-kb RNA species were consistently detected with the B2 probe from pXlvc 19 (Fig. 2), but neither was polyadenylylated (data tiot shown), anho full-length vitellogenin mRNA (6.3 kb) was not detected. This suggests that in oocytes the activation of the gehes and/or the processing of the primary transcripts is abnormal. We therefore tested for authentic vitellogenin sequences by examining the thermostability of the hybrids formed between the probe and the 3.2- and 1.8-kb RNA species. Fig. 4, lane 1, shows that the homoduplex formed between the probe and the DNA fragment used to make the probe is stable in 2 $ NaCl/Cit at 72°C, and lane 6 shows that even in 0.02 x NaCl/Cit at 72°C the homoduplex 'is stable, although inevitably melted to some extent. The probe does not hybridize to oocyte RNA (lanes 2 and 7) but does hybridize strongly to the 3.2-kb and 1.8-kb RNAs from oocytes that have been injected with receptor preparations, forming hybrids that are stable at 720C both in 2 x NaCl/Cit (lane J) and in 0.02 x NaCl/Cit (lane 8). In 0.02 x NaCl/Cit these hybrids

b 3 1

p XIvc

6 7 8 9 10 m

lo

4w

19 probe

4.3

.4.

'.

.4

.,.-

_

-

1.6

B2

B2

Restriction site maps of the cloned mRNAS used to make

probes. o, EcoRI sites; v, HindIlI sites;4, oligo(A) region of A group mRNA; and the poly(A) tail found in both groups; kb, kilobases. Adapted from ref. 13. ,

-

-A-i-

FIG. 4. Induced transcripts contain authentic vitellogenin seTwo identical sets of samples were fractionated and hybridized together to the B2 probe from pXlvc 19. One set (lanes 1-5), was washed for 2.5 hr at 72°C in 2x NaCl/Cit. The other set (lanes 6-10) was washed for 2.5 hr at 72°C in 0.02 x NaCl/Cit. Lanes 1 and 6, 5 pg of the DNA used to make the probe; lanes 2 and 7, oocyte RNA; lanes 3 and 8, RNA from oocytes injected with a receptor preparation and incubated for 96 hr; lanes 4 and 9, 5 pg of the EcoRI/HindIII fragment from the Al gene; lanes 5 and 10, 5 pg of the Bl gene fragment; m, DNA size markers, in kb. quences.

24 probe

_ B1

probe

FIG. 2.

2 3 4 5

Developmental Biology: Knowland et aL

Proc. NatL. Acad. Sci. USA 81 (1984)

display stability similar to that of homoduplex DNA, showing that they are closely homologous to the B2 gene. As shown in lanes 4 and 9 (position marked by arrow in lane 4), the B2 probe does not hybridize significantly to the Al gene [-80% homology with the B2 gene (13)], and the hybrid that it forms with the B1 gene (":95% homology), although stable at 720C in 2x NaCl/Cit (lane 5), is unstable in 0.02x NaCl/ Cit (lane 10). These comparisons show that the RNA sequences we detect in oocyteS injected with receptor preparations are more than 95% homologous to the B2 gene and are therefore most likely to be authentic B2 transcripts. Since a similar analysis has not been performed with the three other genes (Al, A2, and B1), it remains to be seen whether they are also activated. Transcription Induced by Receptor Preparations Initiates Within the 5'-End Region of the B2 Gene and Continues into the 3' Third. To determine which regions of the B2 gene are represented in the 3.2- and 1.8-kb fragments, we hybridized RNA to probes from different regions of the cloned locus (Fig. 5). With two genomic probes from the region flanking the 5' end of the gene (probes a and b), no transcripts were detected (lanes 1 and 4) even after a 5-week exposure. However, hybridization with probe c, which contains the 5' end of the gene (14), revealed the induced RNAs (lane 7), demonstrating that initiation of vitellogenin transcription in oocytes injected with receptor preparations takes place in the same region of the gene as in estrogen-stimulated hepatocytes. As B2 5'

5779

already shown above, the transcripts hybridize strongly to cDNA probes from the middle of the B2 gene, but the cDNA probe d, which represents the 3' third of the B2 mRNA, produced only very weak signals (arrowed in lane 12). We conclude that only a few sequences from the 3' region are present in the transcripts, agreeing with their lack of poly(A) termini. A control experiment (Fig. S E and F) shows that the lack of hybridization to probes a and b is not due simply to failure to bind vitellogenin transcripts to the filter. RNA from oocytes injected with receptor preparations and incubated for 0 (lane 17) or 48 hr (lane 18) does not hybridize to probe b, but hybridization of the same filter to an internal cDNA probe shows that vitellogenin transcripts are not present at time zero (lane 19) but are present after 48 hr (lane 20), confirming the results described above. We conclude that when oocytes are injected with extracts of liver nuclei containing estradiol-receptor complex, new transcription starts within the 5' end region of the B2 gene and continues to its 3' third. More precise mapping of the 5' end of the transcripts will be required to localize the initiation site. Our results on the sequence representation within the transcripts and on their length suggest that the primary transcripts are rapidly and inaccurately processed to form discrete products that accumulate in the oocyte. If any fulllength vitellogenin mRNA is made, the pool is too small and too dilute in total oocyte RNA to be detectable by using nick-translated probes.

GENE mRNA

3'

L-... i a

d

b

Probes

C

kb

?

0-1

0 1

U

Eco RI

10

5

A 1 2 3

B 4 56

C 7 8 9 10 11

D 12 13 14 15 16

I,

It

I X

e11

0* J;

Probe

a

b

..

..

X

c

F 19 20

M

i'.4s

.

-4

A..

E 17 18 w

.

P4:

I

d

b

pXlvc 10

FIG. 5. Hybridization of the sequences induced by receptor preparations to probes from different regions along the B2 locus. Oocytes were injected with receptor preparations or saline or left uninjected and incubated for 96 hr. Probes a, b, and c were genomic fragments sbbcloned in pBR322. Fragment c contains the 5' end of the gene. Probe d was the cDNA clone pXlvc 24, containing exons front the region indicated by the broken line. (A) Probe a hybridized to the following: lane 1, RNA from oocytes injected with a receptor preparation; lane 2, no RNA; lane 3, an EcoRI digest of a clone containing the probe. (h) Probe b hybridized to the following: lane 4, RNA from oocytes injected with a receptor preparation; lane 5, no RNA; lane 6, an EcoRI digest of a clone containing the probe. (C) Probe c hybridized to the following: lane 7, RNA from oocytes injected with a receptor preparation; lane 8, RNA from oocytes injected with saline; lane 9, RNA from uninjected oocytes; lane 10, no RNA; lane 11, an EcoRI digest of a clone containing the probe. (D) probe d hybridized to the following: lane 12, RNA from oocytes injected with a receptor preparation; lane 13, RNA from oocytes injected with saline; lane 14, RNA from uninjected oocytes; lane 15, no RNA; lane 16, a HindIII digest of the plasmid containing the probe. A 5-week exposure (data not shown) of the filters shown in A and B did not reveal any vitellogenin sequences. In E, RNA from oocytes injected with a receptor preparation and frozen immediately (lane 17) or after 48 hr (lane 18) was hybridized with probe b. The filter was washed again and hybridized with the internal cDNA probe from pXlvc 10. F shows that vitellogenin transcripts were not detected immediately after injecting receptor preparations (lane 19) but were detected at 48 hr (lane 20). A 3-week exposure revealed no other transcripts. M, DNA markers of 4.3 and 1.6 kb.

5780

Developmental Biology: Knowland et al. 3

2

1

5

4

6 7

m

A _ q'a

.'VP

-

4.3

-

1.6

B k

1.0 ,-.Am m

0

.s. 4p V:

-1.6 aP

0

FIG. 6. Test for albumin transcripts in oocytes injected with receptor preparations. RNA from male liver, uninjected oocytes, and oocytes injected with active or heat-inactivated receptor preparations was fractionated and hybridized (A) with an isolated 1.1-kb HindIII fragment from the Xenopus albumin plasmid pcXa 13 (16). Lanes 1 and 2, RNA from male liver to show the position of albumin transcripts on the filter (concentration in lane 1 = 16x that in lane 2); lane 3, oocyte RNA; lane 4, RNA from oocytes injected with a receptor preparation and incubated for 48 hr; lane 5, RNA from oocytes injected with a boiled preparation and incubated for 48 hr; lanes 6 and 7 (concentration in lane 7 = 5 x that in lane 6), male liver RNA; m, size markers, in kb. After the albumin probe had been removed, the filter was rehybridized with the B2 probe from pXlvc 19 (B), revealing vitellogenin transcripts in lane 4.

Injection of Receptor Preparations into Oocytes Does Not Activate Transcription of the Albumin Genes. To see whether the preparations we inject activate other liver-specific genes, we probed for transcripts of the albumin genes. In liver cells the maximal level of albumin mRNA is approximately half that of vitellogenin mRNA (15), so that if both genes are activated in oocytes and if their transcripts accumulate in the same proportion as in liver, transcripts from both should be detected. Fig. 6A, lane 4, shows that albumin transcripts were not detectable in the oocytes, although they were detectable in liver RNA (lanes 1, 6, and 7). Fig. 6B shows that

vitellogenin transcripts

were

detectable in lane 4 after hy-

bridization with a vitellogenin probe under the same conditions. We conclude that extracts of liver nuclei that contain estradiol receptor do not simply activate all liver-specific genes when they are injected into oucytes, although they do activate the vitellogenin locus.

DISCUSSION The results reported here strongly suggest that when extracts from Xenopus liver nuclei enriched 2000 times in estrogen receptor are injected into Xenopus oocytes, transcription of vitellogenin sequences is induced. The albumin genes are not activated, showing that the preparations do not nonspecifically activate all genes expressed in liver cells. Tests using the B2 gene show that transcription initiates within the 5' end region of the gene and continues into the 3' third, although the primary transcript seems to be rapidly and

probably inaccurately processed. The products of processing accumulate from zero with time, and none are detectable in oocytes injected with heat-inactivated preparations, showing that the extracts do not merely increase preexisting transcription or contain vitellogenin mRNA. They cannot be

Proc. NatL Acad. Sci. USA 81 (1984)

derived from transcription of any DNA that might survive receptor purification because DNA injected into oocyte cytoplasm is degraded (17). We therefore conclude that the transcripts derive from the vitellogenin sequences in the oocyte nuclei. This shows that activation of these sequences in oocytes, like the activation in liver, does not require DNA synthesis (18), which in turn suggests that changes in methylation are probably not required either (19, 20). We cannot test whether activation would occur in the absence of estradiol, because we have not succeeded in washing all the endogenous estradiol out of oocytes, which contain enormous amounts of the hormone (21). We do not know whether activation of the vitellogenin genes in oocytes could be achieved by injecting pure receptor because although the preparations we inject have been enriched 2000 times in receptor, substantial further purification will be needed to obtain homogeneous receptor-a difficult task compounded by the extremely small amounts available in liver [10 ng per g of estrogen-stimulated male liver (3)]. We have therefore injected the impure preparations into the cytoplasm of the oocytes rather than into the nuclei because it may be important to allow the oocyte nuclei, which have considerable powers of discrimination (reviewed in ref. 22), to select the factors that activate the vitellogenin sequences, which may include the estrogen receptor, from contaminants. We have attempted to estimate the extent of accumulation of the vitellogenin transcripts in oocytes by using hybridization of the 1.6-kb B2 probe to known amounts of cloned DNA as standards. The comparison between the strength of the signals obtained with these standards and with the induced RNAs suggests that the rate of transcription induced in oocytes must be of the same order of magnitude as that found in estrogen-treated liver cells, which synthesize between 2 and 20 molecules of vitellogenin mRNA per minute per cell in secondary stimulation (23, 24). Six independent preparations of receptor were capable of activating the vitellogenin genes, but only some oocytes responded to the preparations. The reasons for this are not clear, but not all oocytes can activate genes in somatic nuclei that have been injected into their own nuclei (25), and the same may apply. to activation of oocyte genes induced by injecting tissue-specific factors. Although we have not been able to demonstrate synthesis of normal vitellogenin gene products (full-length mRNA or authentic protein), the results show an effect of a nuclear extract enriched in estrogen receptor on gene expression. It remains to map the initiation site more precisely, to identify the type of RNA polymerase responsible, and to characterize in more detail the factors required for activation. We thank B. Westley and G. Ryffel for the albumin plasmid pcXa 13, P. Walker and J.-E. Germond for helpful discussions, and B. ten Heggeler and F. Caddick for photography. J.K. thanks the European Molecular Biology Organization for a long-term fellowship to visit Lausanne, and C.V.E.W. thanks the Science and Engineering Research Council for a Studentship for graduate training. This work was supported by grants from the Medical Research Council and the Cancer Research Campaign to J.K. and from the Swiss National Science Foundation and l'Etat de Vaud to W.W. 1. Wahli, W., Dawid, I. B., Ryffel, G. U. & Weber, R. (1981) Science 212, 298-304. 2. Shapiro, D. J. (1982) Crit. Rev. Biochem. 12, 187-203. 3. Westley, B. & Knowland, J. (1978) Cell 15, 367-374. 4. Hayward, M. A., Mitchell, T. A. & Shapiro, D. J. (1980) J. Biol. Chem. 255, 11308-11312. 5. Wright, C. V. E., Wright, S. C. & Knowland, J. (1983) EMBO J. 2, 973-977. 6. Wangh, L. J., Longthorne, R. F. & Knowland, J. (1976) in The Molecular Biology ofHormone Action: 34th Symposium of the Society for Developmental Biology, ed. Papaconstantinou, J. (Academic, New York), pp. 151-169.

Proc. Developmental Biology: Knowland et al. Acad. ScL USA 81 (1984) NaTL 7. Gurdon, J. B. (1977) Methods Cell Biol. 14, 125-139. 8. Probst, E., Kressman, A. & Birnstiel, M. L. (1979) J. Mol. Biol. 135, 709-732. 9. McMaster, G. K. & Carmichael, G. C. (1977) Proc. Natl. Acad. Sci. USA 74, 4835-4838. 10. Thomas, P. S. (1983) Methods Enzymol. lOOB, 255-266. 11. Wahli, W., Wyler, T., Weber, R. & Ryffel, G. U. (1976) Eur. J. Biochem. 66, 457-465. 12. Lane, C. D., Champion, J., Colman, A., James, T. C. & Applebaum, S. W. (1983) Eur. J. Biochem. 130, 529-535. 13. Wahli, W., Dawid, I. B., Wyler, T., Jaggi, R. B., Weber, R. & Ryffel, G. U. (1979) Cell 16, 535-549. 14. Germond, J.-E., ten Heggeler, B., Schubiger, J.-L., Walker, P., Westley, B. & Wahli, W. (1983) Nucleic Acids Res. 11, 2979-2997. 15. May, F. E. B., Ryffel, G. U., Weber, R. & Westley, B. R. (1982) J. Biol. Chem. 257, 13919-13923.

5781

16. Westley, B., Wyler, T., Ryffel, G. U. & Weber, R. (1981) Nucleic Acids Res. 9, 3557-3574. 17. Wyllie, A. H., Gurdon, J. B. & Price, J. (1977) Nature (London) 268, 150-152. 18. Green, C. D. & Tata, J. R. (1976) Cell 7, 131-139. 19. Folger, K., Anderson, J. N., Hayward, M. A. & Shapiro, D. J. (1983) J. Biol. Chem. 258, 8908-8914. 20. Gerber-Huber, S., May, F. E. B., Westley, B. R., Felber, B. K., Hosbach, H. A., Andres, A.-C. & Ryffel, G. U. (1983) Cell 33, 43-51. 21. Fortune, J. E. (1983) Dev. Biol. 99, 502-509. 22. De Robertis, E. (1983) Cell 32, 1021-1025. 23. Ryffel, G. U., Wahli, W. & Weber, R. (1977) Cell 11, 213-221. 24. Brock, M. L. & Shapiro, D. J. (1983) J. Biol. Chem. 258, 5449-5455. 25. Korn, L. J. & Gurdon, J. B. (1981) Nature (London) 289, 461465.