T S Golding

Estrogen receptor and its ligand, estradiol, have long been thought to be essential for survival, fertility, and female sexual differentiation and development. Consistent with this proposed crucial role, no human estrogen receptor gene mutations are known, unlike the androgen receptor, where many loss of function mutations have been found. We have generated mutant mice lacking responsiveness to estradiol by disrupting the estrogen receptor gene by gene targeting. Both male and female animals survive to adulthood with normal gross external phenotypes. Females are infertile; males have a decreased fertility. Females have hypoplastic uteri and hyperemic ovaries with no detectable corpora lutea. In adult wild-type and heterozygous females, 3-day estradiol treatment at 40 micrograms/kg stimulates a 3- to 4-fold increase in uterine wet weight and alters vaginal cornification, but the uteri and vagina do not respond in the animals with the estrogen receptor gene disruption. Prenatal male and female reproductive tract development can therefore occur in the absence of estradiol receptor-mediated responsiveness.

We employed homologous recombination in mouse embryonic stem cells to disrupt the estrogen receptor (ER) gene. Subsequently generated mice that are homozygous for the gene disruption, termed ERKO, possess no demonstrable wild-type ER by Western blot analysis. However, the presence of residual high affinity binding, as detected by [3H]estradiol binding assays and sucrose gradients in uterine extracts from ERKO females prompted further investigation of transcription and translation products from the disrupted ER gene. Analysis of ERKO uterine messenger RNA (mRNA) by reverse transcriptase-polymerase chain reaction demonstrated that although no full-length wild-type ER mRNA was present, two smaller transcripts, labeled E1 and E2, were identified and partially sequenced. Both ERKO transcripts are splicing variants that result in the disrupting NEO sequence being partially or completely removed from the mRNA. In the ERKO-E2 variant, this results in a frame shift and the creation of at least two stop codons downstream. In the ERKO-E1 variant, the ER reading frame is preserved and encodes for a smaller mutant ER that could be the source of the residual estradiol binding. When this mutant form is overexpressed and characterized in vitro, it results in a smaller protein of the predicted size that possesses significantly reduced estrogen-dependent transcriptional activity compared with that of the wild-type ER. Despite residual amounts of an impaired ER variant, estrogen insensitivity in the female ERKOs was confirmed by the failure of estrogen treatment to induce known uterine markers of estrogen action, such as increased DNA synthesis, and transcription of the progesterone receptor, lactoferrin, and glucose-6-phosphate dehydrogenase genes. Furthermore, serum levels of estradiol in the ERKO female are more than 10-fold higher than those in the wild type, consistent with a syndrome of hormone insensitivity.

Increased proteolytic degradation of the estrogen receptor (ER) was detected in uterine cytosol of estradiol-treated ovariectomized mice compared to saline controls. Estradiol had no direct effect on the proteinase activity or susceptibility of the ER to the enzyme. The proteolytic activity gradually increased after a single injection of estradiol with early increases at 2 and 8 h followed by a progressive increase which reached a maximum at 36 h. The proteinase(s) activity resulted in cleavage of the native ER form of 65,000 (65 K ER) to a product of limited proteolysis having an apparent molecular weight of 54,000 (54 K ER). The pH optimum for this proteinase activity was 6.0. The proteinase was inhibited by 2.5 mM p-chloromercuribenzoic acid and 2.5 mM p-chloromercuriphenylsulfonate and partially inhibited by 2.5 mM iodoacetamide but not by 1 mg/ml leupeptin, 0.1 mg/ml antipain, 0.1 mg/ml chymostatin, 0.1 mg/ml pepstatin, 0.1 mg/ml E-64, 2.5 mg/ml soybean trypsin inhibitor, 2.5 mM phenylmethylsulfonylfluoride, 2.5 mM diisopropylfluorophosphate, and 10 mM EGTA. The results suggested that the proteinase(s) had a thiol group essential for its activity. Estrogen receptor in the mouse uterine cytosol fraction appears to be degraded sequentially in two steps in which 65 K ER is cleaved to a 54 K ER which upon longer incubation is further degraded to a 37 K form. The second step was inhibited by leupeptin, antipain, chymostatin, E-64, and p-chloromercuribenzoic acid. A possible function of the 54 K ER under physiological conditions is discussed since the 54 K ER was also found in nuclear samples. This form of the ER still retains the ability to bind estradiol and DNA.

Mouse uterine estrogen receptor (ER) was purified about 11,000-fold from normal mouse uteri by affinity chromatography. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of the purified ER demonstrated a major component of 65,000 mol wt with minor fragments of the 54,000- and 37,000-dalton species, as judged by affinity labeling with [3H]tamoxifen aziridine and immunodetection with an ER monoclonal antibody (H-222). The minor fragments were not detected with additional monoclonals (H-226 or D-547), which recognize different domains on the ER molecule. Two-dimensional gel electrophoresis revealed that the major component with a mol wt of 65,000 had a pI of about 6.5. The 54,000- and 37,000-dalton components had similar pI values. Saturation binding and Scatchard plot analysis of purified ER yielded one class of binding sites with an apparent dissociation constant of about 1.4 nM. Changes in the hormonal status resulted in changes in the size of the ER even in the presence of molybdate and leupeptin.

Holomeric estrogen receptor (ER) prepared from ovariectomized mouse uteri displays heterogeneous electrophoretic mobility when analyzed by NaDodSO4/PAGE. ER derived from nuclei (ERn) appears as a closely spaced doublet having apparent molecular masses of 66.4 and 65 kDa, while ER from the cytosolic compartment (ERc) has a single band of 65 kDa. Both partially purified ERc and the 8S form of unactivated ERc show only the 65-kDa band. The appearance of the ERn doublet is hormonally inducible, and the relative proportions of the two doublet bands are influenced by the type of hormone treatment, with weakly estrogenic compounds yielding the lower band as predominant while potent estrogens increase the proportion of the upper band. Steroid binding of the ERn doublet was determined by [3H]tamoxifen aziridine affinity labeling of both the 66.4- and the 65-kDa peptides; binding to the 65-kDa peptide was predominant. The ERn doublet displays a time dependency after estrogen administration with maximal amounts occurring in a bimodal fashion at 1 and 8 hr.