Adenylyl cyclase (AC) activity in membrane particles of corpora lutea (CL) from humans and cynomolgus monkeys was examined at various stages of the menstrual cycle and pregnancy. AC activity was monitored by the conversion of [alpha-32P]ATP into [32P]cAMP under basal conditions and in the presence of several activators: NaF (10 mmol/L) plus forskolin (100 mumol/L); hCG (10 micrograms/mL); guanyl 5'-yl-imidodiphosphate [GMP-P(NH)P; 100 mumol/L]; and hCG (10 micrograms/ml) plus GMP-P(NH)P (100 mumol/L). The groups of human CL were midluteal (n = 10), late luteal (n = 4), following cycle (old CL; n = 5), and early pregnancy (6-11 weeks; n = 10). The groups of monkey CL were early luteal (n = 4), midluteal (n = 5), and pregnancy at term (n = 3). Luteal AC activity changed significantly during the menstrual cycle. In newly (less than 48 h after ovulation) formed CL, the enzyme was unresponsive to hCG, and total AC activity, as determined by NaF plus forskolin, averaged 86.5 +/- 28.9 (+/- SE) pmol cAMP/min.mg protein. As the CL developed, AC activity increased. Thus, in the midluteal phase, maximal hCG responsiveness in the presence of guanine nucleotide was 125 +/- 27 and 232 +/- 15 pmol/min.mg in human and monkey CL, respectively. No hCG responsiveness was detected in the late luteal phase or in the old CL. Maximal AC activity was also high in the midluteal phase (382 +/- 56 and 256 +/- 28 pmol/min.mg in human and monkey CL, respectively); the activity remained fairly high during the late luteal phase and then declined to less than 100 pmol/min.mg in the follicular phase of the next cycle. During early pregnancy, luteal AC was unresponsive to hCG stimulation, yet basal levels, maximal activity, and the characteristics of stimulation by nonhormonal activators were similar, if not identical, to those at the midluteal phase of the menstrual cycle. At term pregnancy, the enzyme remained unresponsive to hCG. However, basal activity and stimulation by NaF and forskolin were remarkably elevated, being between 2- and 7-fold higher than corresponding stimulations in the midluteal phase. We conclude that 1) AC activity in human luteal membranes is highly dependent on hormonal changes and functional state of the ovary, 2) the activity of luteal AC is similar in the CL of humans and cynomolgus monkeys, and 3) the AC system in the primate CL is functionally active during and at the end of pregnancy.(ABSTRACT TRUNCATED AT 400 WORDS)