BERT W. O’MALLEY, SOPHIA Y. TSAI, MILAN BAGCHI, NANCY L. WEIGEL, WILLIAM T. SCHRADER and MING-JER TSAI,     Department of Cell Biology, Baylor College of Medicine, Houston, Texas 77030

Publisher Summary

This chapter presents experimental approaches to elucidate the precise molecular mechanism by which steroid receptors regulate the initiation of target gene transcription and the role of the hormonal ligand in this process. For determining the direct actions of receptor on DNA transcription, a cell-free—reconstituted—transcription system is required. In such a system, the concentration of receptor, ligand, general transcription factors, and target genes could be manipulated. In the initial experiments, the effect of purified cPR (50–80% pure) or a purified cPR derivative expressed in Escherichia coli are measured on transcription of templates that lacked or contained progesterone response elements (PREs). To examine the mechanism by which progesterone receptor interacts with general transcription factors to stimulate RNA synthesis in vitro, it was investigated whether the receptor participates in the formation of a stable preinitiation—rapid start—complex of transcription factors. In this investigation described in the chapter, in an attempt to investigate the mechanism by which regulatory proteins such as steroid hormone receptors interact with core promoters to enhance transcription, the cell-free transcription system was used.

I Background Review

In the early 1960s, predominant theories existed that steroid hormones acted at the level of the cell membrane to facilitate ion or substrate transport or to catalyze energy exchange. When labeled estradiol was synthesized by Jensen’s laboratory, a new series of experiments was initiated that led to the observation that estradiol bound with specificity and high affinity to an intracellular protein in target cells, termed a receptor (Jensen et al., 1966, 1968; O’Malley et al., 1979; Gorski and Cannon, 1976; Gorski et al., 1968; Yamamoto, 1985). A flurry of experiments extended this concept to progestins, androgens, glucocorticoids, and other steroid hormones. These receptors were noted to be capable of binding to DNA and were thought to be concentrated in the nucleus following hormone administration. In separate experiments during this period, steroid hormones were shown also to increase the incorporation of radiolabeled nucleotides and amino acids into precipitable macromolecules. Heretical voices suggested that steroid hormones may activate genes via the concurrent theories of Jacob and Monod generated from bacterial studies. For the most part, such investigators were correct in their interpretations, but critics were entitled to point out that their results could emanate from increased transport of labeled precursors into cellular pools rather than from a net increase in macromolecular synthesis.

The next important observations were qualitative and involved experiments which showed that steroid hormones (1) caused appearance and accumulation of new species of nuclear (hybridizable) RNAs, (2) caused stimulation of de novo synthesis of new specific proteins, (3) caused a corresponding increase in the net cellular level of specific mRNAs, and (4) stimulated the rate of transcription of certain nuclear genes (O′Malley et al., 1969, 1979; O’Malley and Means, 1974). These observations argued for a nuclear role for steroid hormone and were accompanied by studies which revealed that receptors had an inherent affinity for DNA (O′Malley, 1984; Yamamoto, 1985). At this point in the early 1970s, the circle was closed and the primary pathway for steroid hormone action was defined as follows: steroid → (steroid—receptor) → (steroid-receptor-DNA) → mRNA → protein → functional response (Fig. 1).

Steroid receptors were purified then to single species and characterized as to size, charge, etc. Antibodies were developed and structural domains were postulated by proteolytic analyses. After an initial description by Jensen et al. (1966, 1968), assays of sex steroid receptors became commonplace in the diagnosis and assignment of therapy for breast cancer (Jensen and DeSombre, 1973). Investigators isolated specific target genes for steroid hormones, defined their structure, and proved that cis-acting regulatory sequences were located near such genes, usually in the 5’ flanking sequences (Payvar et al., 1983; Renkawitz et al., 1984; Yamamoto, 1985;). When such sequences, termed steroid response elements (SREs), were occupied by receptors these genes came under the control of their respective hormones.

The molecular biological community became increasingly interested in these receptors for steroid hormones as they came to realize that they were the most highly studied and purified trans-activation factors for control of eukaryotic transcription. Also, they were the specific activators of an emerging and fascinating genetic cis element, the enhancer. In the past 5 years, a great deal more has been learned of the structure–function relationships of steroid receptors and the mechanisms by which they interact with DNA. Biochemical studies in the late 1970s suggested that steroid receptors, thyroid receptors, and receptors for vitamins such as vitamin D3 were likely to belong to a family of gene regulatory proteins. Furthermore, it was known that these functional proteins were organized into domains which contained the functions of (1) specific and high-affinity ligand binding, (2) specific DNA binding, and (3) “transcriptional modulation.”

Molecular cloning of the steroid/thyroid and vitamin receptors proved this “superfamily” concept (Giguere et al., 1987; Evans, 1988; Beato, 1989; O’Malley, 1990). A surprising observation has been that certain oncogenes such as v-erbA are members of this receptor gene family. The avian erythroblastosis virus appears to have captured the cellular gene coding for thyroid receptor and this retrovirus uses this mutated molecule for its own oncogenic purposes. Further sequence analyses not only substantiated the existence of functional domains within receptors but showed that they could be rearranged as independent cassettes within their own molecules or as hybrid molecules with other regulatory peptides (Green and Chambon, 1987). Perhaps the most intensively studied domain has been that responsible for DNA binding. This domain has been shown to be a cysteine-rich region which is capable of binding zinc in a manner which creates two peptide projections referred to as “zinc fingers” (Evans, 1988). These zinc fingers promote the interactions of receptors with target enhancers and clearly mark each as a member of this evolutionarily conserved family. A broad region of the C-terminal domain was shown to comprise the ligand-binding site and a transcriptional activation region; the heterogeneous N-terminal domains of receptors were revealed to contain an additional transcriptional modulation domain (for review, see Beato, 1989).

The precise sequences of the steroid response elements (SREs) which are regulated by steroid hormones have been described for all members of this family to date. In general they are 15-base pair (bp) core sequences, composed of two half-sites of 5 or 6 bp arranged in a dyad axis of symmetry (inverted repeats); the half-sites are split by a few central base pairs of random composition (Strahle et al., 1989). The SREs for various receptors share similarities in sequence and, in fact, the identical sequence allows activation by glucocorticoid, progesterone, and androgen receptors. One copy of such an SRE is sufficient usually to bring a promoter under moderate hormonal control and two copies often provide a synergistic response to the cognate hormone.

The precise mechanism of interaction of receptors with their target SREs has come under close scrutiny of late. After cytoplasmic synthesis, many steroid receptors form complexes with heat shock proteins, such as hsp90 and hsp70 (Schuh et al., 1985; Catelli et al., 1985; Pratt et al., 1988). Extracts of receptors for glucocorticoid, progesterone, and estrogen exist in aggregate form with these heat shock proteins when prepared from cells; thyroid and vitamin D3 receptors appear not to be extracted as such complexes. In cells, this interaction may promote proper folding and stability of the molecule; in complex with hsp90, receptors cannot bind to DNA. When a receptor is complexed with heat shock proteins in vitro and in vivo, binding of hormone causes dissociation of the complex (Denis et al., 1988; Kost et al., 1989). The exact physiologic meaning of this association of certain receptors with heat shock proteins is unclear at present but it is thought that heat shock proteins aid posttranslational folding of the molecules, prevent denaturation,...