Nov 1, 1989 - PETER DAVIES*. * Tenovus Institute for Cancer Research, University of Wales. College of Medicine, Heath Park, Cardiff CF4 4XX, U.K. and ?
BIOCHEMICAL SOCIETY TRANSACTIONS
560
AMP plays a significant role in thyroid gene activation with TSH and Graves' immunoglobulin, thyroid mRNA activation may also be modulated by other cellular messengers. 1. Dumont, J. E. & Lamy, F. (1980) The Thyroid Gland (De
Visscher, M., ed.), pp. 153- 167, Raven Press, New York 2. Nunez, J. & Pommier, J. (1982) Vifam.Horm. 39,175-229 3. Kung, A. W. C., Collison, K., Banga, J. P. & McGregor, A. M. (1988) FEBS Len 232, 12- 16
4. Collison, K. S., Banga, J. P., Barnett, P. S., Kung, A. W. C. & McGregor, A. M. ( 1 989) J. Mol. Endocrinol. 3, 1-5 5. Macchia, E., Concetti, G., Borgoni, F., Fenzi, G. F. & Pinchera, A. (1988) Clin. Endocrinol. 28, 147- 156
Received 2 1 November 1989
Intronic steroid response elements in prostate binding protein genes NEIL K. RUSHMERE,* FRANK CLAESSENS,? BEN PEETERS,t WILFRIED ROMBAUTS? and PETER DAVIES* * Tenovus Institute for Cancer Research, University of Wales College of Medicine, Heath Park, Cardiff CF4 4XX, U.K. and ?Biochemistry Division, K . U. Leuven, 3000 Leuven, Belgium Androgen receptors are assumed to influence gene expression through interaction with a nucleotide response element similar to that recognized by glucocorticoid, mineralocorticoid and progesterone receptors [ 11. This assumption arises from the efficiency in heterologous systems by which androgens act through the defined glucocorticoid response elements (GRE) of the mouse mammary tumour virus Abbreviations used: GRE, glucocorticoid response element;
ARE, androgen response element; MMTV, mouse mammary tumour virus; LTR, long terminal repeat.
(MMTV) long terminal repeat (LTR) [2-41 or the tyrosine aminotransferase gene [ 51. However, androgen response elements (ARE) in androgen-responsive genes have been less easy to delimit. High-affinity androgen-receptor-binding regions have been detected in the promoter and first intron of the rat ventral prostate binding protein C3( 1 ) [6-81 and other constituent polypeptide genes [ 91. Affinity for androgen receptors has been tracked to a 450 b p fragment of the C3(1) intron containing two putative GRE-like elements: 5 '-TAGCCAAGTTGTTCT-3' in the non-coding strand and, centred 52 bp downstream in the coding strand, 5'-AGTACGTGATGTTCT-3' (see Fig. 1). Transfection analysis [lo] has attributed A R E status to the second element and none to the first. However, mobility shift analysis of androgen-receptor binding to restriction fragments labelled at either end (Fig. 1) indicates significant binding affinity in the first, non-coding strand, element. The abolition of binding to the second element can be attributed to its proximity
(a1
(bl
1 2
3
L 5 6 7 8 9 10 11 12 nnnnnn 8- P B - F 0-Sc B-Sa B - D B - H . ( I )
1
2
3
L 5
6 7
8 9 1 0 1 1
12
nnnnnn
P-B
P-H
P-D
P-Sa
P-Sc
P-F
a**immmm-
450 b 450 395 303 257 202
343 b 248 b 193 b 1L7 b
107
55
3'
5' 0
I
B
107 I
202 I
257
303
I
I
H
D
Sa
I
I
I
395
L50
I I
F ~ G A A C A A C T T G G C TAGTACGTGATGT ~ TCT'
Fig. 1. The 450 bp Bglll-Pvull (B-17 fragment of the C3(1) gene first intron was (a) 5' (7'4 polytiudeide kinuse) or (h) 3' (K1enow)-labelled with -"I' and digested with one of Hinfl(lf), Dral(D), Sacl(Sa), Scal(Sc) or Fokl(F) Electrophoretically purified, electroeluted, specific radiolabelled DNA fragments were re-electrophorcsed, alone (oddnumbered lanes), or after incubation with androgen-receptor complexes and poly(dl-dC) (even-numbered lanes), in polyacrylamide gels 12). Coding strand positions and sequences of putative response elements are shown. 1990
633rd MEETING, LONDON to the Fokl cutting site. For further analysis of the ScaI-FokI and Fok I-Pvull fragments, pairs of complementary oligonucleotides representing these regions were synthesized, respectively, A R E 1:
5 ’-CTATAA ATATCACAGCTCGAATGTTAGCCAAGTTGTTCTCTCTTA ACTCT-3' and ARE2:
5 ’-TATAGGATGTTTGA ACATAGTACGTGATGTTCTCAAGATAGTAATGAAAT-3’ in each of which the putative 15-nucleotide hormone response element, comprising the two ‘arms’ of the imperfect 6-base palindrome and the 3-base spacer [ 11, is underlined. In DNA cellulose competition assays, both A R E l (50% competition at 0.05 g equivalents) and ARE2 (73% competition at 0.05 g equivalents) effectively reduced binding of androgen receptor to immobilized calf-thymus DNA. Mutation of the right arm to TTTTCT [ l o ]abolished competitive ability in both cases. Alteration of the left arm of A R E l to TTGCAA and that of ARE2 to ATTAAG reduced competitive effectiveness at 0.05 g equivalents to 24% and 41‘%, respectively. Deletion of the three bases immediately 5’ to the TGTTCT motif, i.e. those prospectively responsible for the critical spacing of dyadic response elements [ l l ] , abolished the effectiveness of ARE2 and reduced that of ARE 1 t o 28% at 0.05 g equivalents. In the latter case, however, removal of the ‘spacer’ creates TGTTAGCCATGTTCT, which has greater similarity to ARE2 than does the natural sequence. In terms of binding of androgen receptor, therefore, both putative intronic AREs behave as response elements, with the right arm of the imperfect palindrome placing more stringency on binding than the left arm. ARE2 has the greater similarity to the response elements of the MMTV LTR and, most remarkably, to the most active synthetic ele-
56 1 ment GGTACANNNTGTTCT devised by Ham el af. 131. This is in line with the relative performances of A R E l and ARE2 in transfection assays [ 101, but their role in vivo, and the requirement for the co-operativity of other enhancer factors, including possible A R E elements in the promoter region, remains open to critical appraisal. I . Beato, M., Chalepakis, G., Schauer, M. & Slater, E. P. ( 1989)J. Steroid Riochem. 32,737-748 2. Cato, A. C. B., Henderson, D. & Ponta, H. ( 1 987) E M H O J . 6 , 363-368 3. Ham, J., Thomson, A,, Needham, M., Wehh, P. & Parker, M. (1988)Nucleic Acids Res. 16,5263-5276 4. Otten, A. D., Sanders, M. M. & McKnight, G. S. (1088) Mol. Endocrinol. 2, 143- 147 5. Denison, S. H., Sands, A. & Tindall, D. J. (1989) Endocrinolohy 124, 1091-1093 6. Perry, S. T., Viskochil, D. H., Ho, K.-C., Fong, K., Stafford, D. W., Wilson, E. M. & French, F. S. ( 1 985) in Hegulution 01 Androgen Action (Bruchovsky, N., Chapdelaine, A. & Neumann, F., eds.), pp. 167- 174, Congressdruck R. Bruckner, Berlin 7. Rushmere, N. K., Parker, M. G. & Davies, P. (1987) Mol. Cell. Endocrinol. 51,259-265 8. Van Dijck, Winderickx, J., Heyns, W. & Verhoeven, G. ( I 989) Mol. Cell. Endocrinol. 64, 195-204 9. Claessens, F., Rushmere, N. K., Peeters, B., Davies, P. & Rombauts, W. ( 1 989)Arch. Int. I’hysiol. Hiochem. 97, B I 5 10. Claessens, F., Celis, L., Peeters, B., Heyns, W., Verhoevcn, G. & Romhauts, W. ( 1 989) Biochem. Biophys. Rex Cbmmun. 164, 833-840 1 1 . Umesono, K. & Evans, R. M. ( 1989) (211 (Cumhridgc., Mus.~.) 57,1139-1 146 12. Fried, M. & Crothers, D. M. (1981) Nucfeic Acids Hex 9, 6 505-6525
Received 22 November 1989
Functional characterization of an androgen response element F. CLAESSENS? N. RUSHMERE,t L. CELIS,* B. PEETERS,* P. DAVlESt and W. ROMBAUTS* *AJdelirigBiochemie, c‘umpi4s Gasthuisberg, K . U. Leuven, .WO Leirven. Rel~iirmarid t Terioviu Institute tor Cancer Keseurch, Utiive&y of Wules, College of Medicine, Curd# CF4 4xx. U.K . The synthesis of prostatic binding protein (PBP) in rat ventral prostate is strongly dependent on the presence of androgens. This dependency can at least in part be explained by transcriptional regulation of the genes coding for the components C I , C2 and C 3 of PBP. Wc have screened the complete C I , C2A and C3( I ) genes and their flanking sequences for regions displaying affinity in vim) f o r the androgen receptor in a DNA-cellulose competition assay. Two affinity regions were detected in each gene: one in the promoter region and one in the first intron. These results are in agreement with earlier reports on the 5’ part o f theC3( I ) g e n e [ l ] . While attempts t o delimit androgen response elements (AREs) in the promoter region of C3(1) have been unsuccessful 121, we now have demonstrated that the 450 bp intronic fragment o f C3( 1 ) confers androgen responsiveness Ahhreviations used: PBP, prostatic binding protein: ARE, androgen response clement: GRE. glucocorticoid response element: PRE, progesterone response element; tk-CAI; thymidine kinase promoter in front of a chloramphenicol acetyltransferase: MMTV LIK. mouse mammary tumour virus long terminal repeat: 5a-DH7; 5 *-dihydrotestosterone.
VOl. 1 x
to a heterologous promoter in T-47D cells 131. A 204 bp subfragment, containing two elements: core I and core 11, which resemble the consensus for the glucocorticoid response element (GRE) [4] displays identical properties (Table I). The inductive effect of 100 nM-Sa-DHT was abolished by addition of 100 ,LAMof the anti-androgenic compound nilutamide (RU 23908; obtained from J. P. Raynaud, Roussel Uclaf), and is therefore mediated by the androgen receptor. In each core element we have substituted the G in 5‘TGTTCT-3’ by T. Functional analysis has revealed that only core 11 has the ability to act as an ARE (Table 1 ).This clement (S’-AGTACGTGATGTTCT-3’) shows a high degree of similarity to the GRE consensus (S’-GGTACAnnnTGTTCT-3‘), which can also act as an ARE [51.The substitution in core I has no effect on the androgen response. Mutation of core I1 in the 204 bp intronic fragment results in a total loss o f affinity ;TI vitro o f the fragment for the androgen receptor in DNA-cellulose cornpetition analysis, while mutation o f core I results only in a minor decrease. Furthermore, core II, cloned as an oligonucleotide in the polylinker site of the pBLCAT2 vector 161, clearly confers affinity in v i m to a vector fragment. Surprisingly, this oligonucleotide construct is not responsivc to androgens when transfccted in T-47D cells (Table 1 ). Because of the resemblance of the C 3 ARE with the GRE/PRE consensus, we have tried t o induce the androgen responsivc tk-CAT constructs with progcsteronc and dcxamethasone and found that they are inducible with thcsc steroids. Moreover, the substitution in the ARE (core 11)
