15N NMR spectroscopy was used to examine the active-site histidyl residue of alpha-lytic protease in peptide boronic acid inhibitor complexes. Two distinct types of complexes were observed: (1) Boronic acids that are analogues of substrates form complexes in which the active-site imidazole ring is protonated and both imidazole N-H protons are strongly hydrogen bonded. With the better inhibitors of the class this arrangement is stable over the pH range 4.0-10.5. The results are consistent with a putative tetrahedral intermediate like complex involving a negatively charged, tetrahedral boron atom covalently bonded to O gamma of the active-site serine. (2) Boronic acids that are not substrate analogues form complexes in which N epsilon 2 of the active-site histidine is covalently bonded to the boron atom of the inhibitor. The proton bound to N delta 1 of the histidine in these histidine-boronate adducts remains strongly hydrogen bonded, presumably to the active-site aspartate. Benzeneboronic acid, which falls in this category, forms an adduct with histidine. In both types of complexes the N-H protons of His-57 exchange unusually slowly as evidenced by the room temperature visibility of the low-field 1H resonances and the 15N-H spin couplings. These results, coupled with the kinetic data of the preceding paper [Kettner, C. A., Bone, R., Agard, D. A., & Bachovchin, W. W. (1988) Biochemistry (preceding paper in this issue)], indicate that occupancy of the specificity subsites may be required to fully form the transition-state binding site. The significance of these findings for understanding inhibitor binding and the catalytic mechanism of serine proteases is discussed.

Two affinity-purified polyclonal antibodies have been generated that differentially recognize two mol wt (Mr) variants of prostaglandin H synthase (PGS) in the rat ovary: antibody-2 recognized PGS of 72,000 Mr (PGS72), and antibody-3 recognized PGS of 69,000 Mr (PGS69). Immunoblot analyses showed that PGS72 was rapidly induced by LH in granulosa cells of preovulatory (PO) follicles and was associated with the increased production of prostaglandins (PGs) obligatory for ovulation. PGS72 was low (negligible) in other ovarian tissues, including PO follicles, corpora lutea, and interstitium. In contrast, PGS69 was constitutively present in small antral and PO follicles (primarily in thecal cells), was unaffected by LH, and was found at higher levels in corpora lutea throughout pregnancy and in the ovarian interstitium. PGS69 (but not PGS72) was also detected by immunoblots in rat adrenal glands, heart, uterus, and kidney. Immunofluorescent localization of PGS72 and PGS69 to ovarian tissue sections confirmed the cell-specific distribution of PGS observed by immunoblot analyses of cell extracts. Immunofluorescent detection of PGS72 required methanol fixation, whereas PGS69 was also observed with paraformaldehyde fixation and Triton X-100 permeabilization, further suggesting biochemical differences in these molecules. Immunoreactive PGS69 in PO follicles, thecal cells, and granulosa cells was associated with low amounts of indomethacin-sensitive production of PGs by these tissues in vitro, which was unaffected by inhibitors of transcription or translation. In contrast, stimulation of PGs in PO follicles by LH in vitro correlated with the marked induction of PGS72, but not PGS69, and was sensitive to both transcriptional and translational inhibitors. Collectively, these studies provide the first evidence that the rat ovary contains two immunologically distinct forms and Mr variants of PGS, each of which is selectively regulated by hormones, localized to specific cell types, differentially sensitive to inhibitors of transcription/translation, and differentially solubilized for immunocytochemical localization.

The individual tautomeric forms of histidine have been directly observed in 15N NMR spectra at -55 O C in ethanol/water solution. The chemical shifts of the nitrogen atoms in the tautomers agree well with values previously estimated for them from 15N NMR spectra of N" and Ne2 methyl histidines and of histidine in the dry powder state. A comparison of the chemical shift position of N6l in the amphion and anion forms of histidine shows that N*1 of the NeLH tautomer forms at most only a very weak hydrogen bond with the a-amino group. The existence of this hydrogen bond has been proposed previously to explain the effect ionization of the amino group has on the tautomeric equilibrium constant. The pH dependence of the 15N chemical shifts of histidine at 25 O C in the ethanol/water cosolvent closely matches that observed in water alone, showing that the ethanol cosolvent solution used in the low-temperature work has little or no effect on the intrinsic l5N shifts of the individual tautomeric forms or on the acid-base and tautomeric equilibria of the imidazole ring. The significance of these results for determining tautomeric equilibrium constants of imidazoles and for interpreting l5N chemical shift data from histidyl residues in proteins is discussed.Recognizing the frequency with which histidyl residues play key functional roles in proteins, early biochemical applications of NMR spectroscopy devoted considerable effort to resolving, assigning, the studying the behavior of NMR signals from these residues in onedimensional spectra.I Observation of the imidazole ring Hel proton2 was the most widely used a p p r~a c h ,~~ in part because it could be appiied to proteins as they are found in nature, Le., incorporation of isotopic labels was not necessary, and in part because of the sensitivity advantage of IH NMR spectroscopy.However, the information about histidyl residues extractable from the chemical shift behavior of this proton is somewhat limited. It reveals whether or not the imidazole ring titrates. If it titrates, the pH dependenke of the lHfL signal yields a microscopic pK, value, and if it does not titrate the chemical shift of this signal may reveal whether the ring is protonated or neutral, although for nontitrating imidazole rings the IHfl chemical shift is often ambiguous on this point. One-dimensional I3C NMR spectroscopy of I3CeI histidyl labeled proteins, developed at about the same time as the Hel proton NMR approach,a11 offers some Autbr to whom oorrwpondence should be addressed. t Present a l d w :

Two distinct isoforms of prostaglandin (PG) endoperoxide synthase (PGS) have been identified in rat ovarian tissues: rPGSi (mol wt, 70,000-72,000) is induced by FSH and LH in preovulatory follicles, whereas the other isoform (mol wt, 69,000) is not. Induction of rPGSi is associated with LH-stimulated increases in PG biosynthesis obligatory for ovulation. Because GnRH, like LH, can also stimulate the synthesis of PGs and ovulation in the rat, this study was undertaken to determine which isoform of PGS might be induced by GnRH, in what cell type, and by what intracellular pathways. Results show that GnRH at relatively low concentrations (10(-8)-10(-7) M) induced the same isoform of PGS (rPGSi) in the same cell type (preovulatory granulosa cells) and within the same 5- to 7-h time course as did LH. Unlike LH and FSH, GnRH did not cause a major increase in cAMP, nor did GnRH induce luteinization. The effects of GnRH on rPGSi in preovulatory follicles were not mimicked by known activators of protein kinase-C (phorbol myristate acetate, bryostatin, diacyglycerol, and (+/-)ionomycin). Epidermal growth factor (but not basic fibroblast growth factor or platelet-derived growth factor), which activates a receptor-associated tyrosine kinase, caused a small increase in rPGSi. Genistein, a selective inhibitor of tyrosine kinases, blocked GnRH and LH induction of rPGSi. Taken together these results suggest that the mechanisms by which GnRH and LH selectively induce rPGSi in granulosa cells of preovulatory follicles before ovulation may converge at some step within a cellular tyrosine kinase cascade. Furthermore, the mechanisms responsible for inducing rPGSi are distinct from those required for cellular luteinization.

Prostaglandin endoperoxide synthase (PGS) catalyzes the rate-limiting step in the synthesis of prostaglandins E and F2 alpha that are obligatory for ovulation. To understand the molecular mechanisms by which LH regulates the induction of PGS in rat preovulatory (PO) follicles, we established an in vitro system which mimics in vivo induction of the enzyme. We show that the rapid increase in PGS enzyme: 1) is stimulated by LH, FSH, and forskolin (cAMP) in a time- and dose-dependent manner that is distinct from changes in steroidogenic enzymes analyzed in the same follicles; 2) is unaltered by end products (PGE and PGF2 alpha) of the reaction or inhibitors (indomethacin) of enzyme activity; 3) is blocked by inhibitors of transcription (alpha-amanitin) and translation (cycloheximide) at a step distal to production of cAMP and activation of A-kinase. Analyses of PGS mRNA by Northern blots using a mouse PGS cDNA probe revealed a PGS transcript of 2.8 kilobases that was present but in low abundance in PO follicles and decreased rapidly (1-4 h) as a consequence of LH/human CG stimulation. This hormone-induced decrease in PGS mRNA appeared to be transient because PGS transcripts were present in corpora lutea at a level similar to that in PO follicles. These results raise the possibility that the marked increase in PGS enzyme in granulosa cells of PO follicles that occurs as a consequence of the LH/human CG surge may not involve a cAMP regulated increase in transcription of the PGS gene itself. Or, if increased transcription of PGS mRNA does occur, it is rapid and coupled with cotranslational degradation of the mRNA. Alternatively, transcriptional (alpha-amanitin-sensitive) regulation of a separate gene may be obligatory for increased translation of PGS mRNA or posttranslational modification (stabilization?) of the enzyme. In summary, the LH-stimulated appearance of PGS in granulosa cells of PO follicles before ovulation is mediated by cAMP in a complex manner involving transcriptional regulation (PGS gene?) and translational control of PGS mRNA. The transient appearance of the PGS enzyme represents a unique pattern of response by granulosa cells of PO follicles to LH/cAMP and thereby may involve novel intracellular factors and regulatory processes.