The molecular basis for partial androgen insensitivity associated with adult onset spinal/bulbar muscular atrophy was investigated by transient transfection of human androgen receptor (AR) expression vectors containing increasing CAG repeat lengths in the first exon. An inverse relationship was observed between CAG repeat length and AR mRNA and protein levels. Trinucleotide repeat lengths of 43 and 65 associated with spinal/bulbar muscular atrophy decreased AR mRNA and protein levels but did not alter equilibrium binding affinity for [3H]R1881 or inherent transcriptional activity of AR, expressed as androgen-dependent fold induction of a mouse mammary tumor virus promoter-luciferase reporter vector. The findings indicate that glutamine expansion up to 66 residues in the NH2-terminal domain of AR does not alter AR functional activity. Rather, CAG repeat expansion in the region of the first exon reduces AR mRNA and protein expression. The study reveals a previously unrecognized effect of CAG repeat length on AR mRNA expression and a novel molecular mechanism for androgen resistance.

No abstract

A composite mouse androgen receptor DNA sequence was obtained by amplifying genomic DNA or cDNA using the polymerase chain reaction. The open reading frame was 2,697 basepairs, encoding a polypeptide of 899 amino acids (98,204 mol wt). Amino acid sequence comparisons indicated that the mouse androgen receptor (AR) is 97% homologous with rat AR and 83% with human AR. The amino acid sequences of the three receptors are identical within the DNA- and steroid-binding domains. Northern blot analysis revealed the predominant mouse AR mRNA to be 10 kilobases (kb). A 1.7-kb mRNA species was detected in mouse kidney using a cDNA probe containing only 5' untranslated AR sequence. Lack of hybridization with AR-coding sequence probes suggested that the 1.7-kb mRNA was not a truncated form of AR mRNA. Sequencing of genomic DNA isolated from testicular feminized (Tfm) mice revealed a single base deletion in the N-terminal domain, resulting in a frameshift mutation. Cycloheximide treatment caused a dramatic increase in AR mRNA in kidneys of Tfm mice, but not wild-type mice, suggesting that the Tfm mutation results in an unstable AR mRNA.

Full-length wild type and deletion mutant human androgen receptors (AR) were transiently expressed in monkey kidney COS cells to identify the phosphorylated amino acid residues. Phosphoamino acid analysis indicated serine (Ser) and threonine (Thr) residues as the major sites of phosphorylation. Both NH2- and carboxyl-terminal fragments containing the DNA-binding domain were highly phosphorylated, suggesting the presence of phosphorylation sites throughout the protein. Site-directed mutagenesis of wild type and deletion mutant AR at proline-directed consensus phosphorylation sites replaced Ser or Thr residues with Ala; wild type and mutant ARs were expressed in the presence of [32P]orthophosphate and isolated by immunoprecipitation using AR-specific antipeptide antibodies. Three proline-directed phosphorylation sites were identified: Ser 81 and 94 in the NH2-terminal region and Ser 650 in the hinge region. Expression of a series of NH2-terminal AR fragments provided evidence for additional sites in the NH2-terminal region. The effect of loss of each phosphorylation site on receptor function was determined by introducing the Ser to Ala mutations into full-length AR. Substituting Ser 81 and 94 with Ala had little effect on transcriptional activity when assayed by transient cotransfection. Substituting Ser 650 with Ala in the hinge region reduced transcriptional activity up to 30%. The results suggest at least three proline-directed phosphorylation sites in AR, one of which, serine 650, contributes to optimal gene activation by AR.

The molecular basis for the different physiological effects of testosterone (T) and dihydrotestosterone (DHT) was investigated using recombinantly expressed wild-type and mutant androgen receptor (AR). Rates of androgen dissociation from nuclear and cytoplasmic AR were compared with hormone- and concentration-dependent receptor degradation rates. T dissociates from AR 3 times faster than DHT or methyltrienolone (R1881) and is less effective in stabilizing the receptor. Analysis of AR deletion mutants and AR/glucocorticoid receptor chimeras indicates that the AR NH2-terminal domain has a specific role in stabilizing the receptor by slowing the rate of ligand dissociation and AR degradation. Amino acid mutations that abolish receptor dimerization, nuclear localization, or DNA-binding activity have no significant effect on androgen dissociation or AR degradation. A naturally occurring steroid-binding domain mutation (Val889 to Met) that causes androgen insensitivity, but does not alter equilibrium androgen binding affinity, lowered the androgen-binding capacity as a result of increased rates of androgen dissociation and AR degradation. Thus, AR stabilization and function require prolonged receptor occupancy with androgen, with a similar extent of stabilization observed at higher concentrations of faster dissociating androgens and lower concentrations of slower dissociating androgens. Retention of receptor-bound androgen is enhanced by an interaction between the AR NH2-terminal and steroid-binding domains. The ligand specificity and concentration dependence of receptor stabilization provide an explanation for physiological differences in the actions of T and DHT.