Effect of Neurophysin on Enzymatic Maturation of Oxytocin from Its Precursor*

We examined the extent to which rates of enzymatic conversion of the oxytocin biosynthetic precursor to mature peptide are modulated by intramolecular and intermolecular assembly of precursor and polypeptide intermediates. The biosynthesized precursor contains hormone and neurophysin sequences linked by a Gly-Lys-Arg sequence and undergoes enzymatic processing reactions which include endoproteolytic cleavage at the Lys-Arg dibasic sequence, carboxypeptidase B-like exoproteolytic cleavage, and enzymatic amidation. We evaluated the effect of neurophysin on such processing reactions using semisynthetic precursors of oxytocin1 bovine neurophysin I and synthetic oxytocinyl precur- sor intermediates as substrates. Neurophysin I at high concentration (0.7 mM) reduced the rates of carboxy- peptidase B-like conversion of oxytocinyl-Gly-Lys-Arg to oxytocinyl-Gly and the enzymatic amidation of oxytocinyl-Gly to mature (C-terminal amidated) oxytocin. The dependence of rate suppression on the con- centrations of peptide substrate and neurophysin I suggested that suppression is due to intermolecular for- mation of hormone-neurophysin complexes which are aggregated at least to dimers. An analogous intramo- lecular neurophysin effect was found for endoproteolytic processing of semisynthetic The purified filtration Sephadex RP-HPLC on ODS an elution two consecutive linear from 0.1% and from latter to 0.1% in 70% at 40 The main peak were collected and lyophilized. The yield of final product calculated from Boc-Arg(tosy1)-phenylacetamidomethylresin was about 50%. Amino acid composition analysis after hydrolysis HCl 110 for 24 in uacw Asp Glu Pro Gly %Cys Ile Leu Tyr (1.09), and Arg S-protein substituted for Site-specific Cleavage Semisynthetic IN"'- were in 4.2 p1 mM MES, and incubated at 37 with 4.2 p1 of endoproteinase Lys-C (Boehringer dissolved (at 38.5 ng/pl) in the same buffer. The final concentration of precursors was 700 p ~ ; the ratio of precursor to enzyme was about 4001 by weight. The amount of endoproteinase Lys-C was chosen to slow the reactions enough to follow their time courses. After incubation for 15,30,45, and 60 min, 1.5 p1 of each sample were withdrawn, mixed with 30 pl of 10% acetic acid, and frozen until analysis. Samples were analyzed by RP-HPLC on Zorbax CN (0.46 X 25 cm) using an elution by consecutive linear gradients of 90% TEAP, 10% acetonitrile at zero time to 75% TEAP, 25% acetonitrile at 20 min, to 72% TEAP, 28% acetonitrile at 50 min, and then to 40% TEAP, 60% acetonitrile at 55 min. The elution profiles were monitored by UV absorbance at 210 nm. Limited proteolysis reactions also were carried out using diluted precursors (5.9 nmol of proteins dissolved in 420 p1 of 10 mM Na-H2P04/NaOH buffer, pH 5.7), which were incubated with 4.2 pl of endoproteinase Lys-C solution (38.5 nglpl). The final precursor concentration was 14 p ~ of Semisynthetic Precursors by Trypsin-Semisyn-thetic Pro"-OT/BNPI and [Na'-Ac,AlaZ]pro"-OT/BNPI (4.9 nmol each) were dissolved in 58 pl of 150 mM MES, pH 5.5, and incubated at 37 "C with 5.3 pl(0.53 pg in water) of L-1-tosylamido-2-phenylethyl chloromethyl ketone-treated trypsin (Millipore Corp.). The final concentration of precursor was 77 p ~ , and the ratio of precursors to trypsin was about 1001 by weight. After 5 and 30 min and 1,2, and 4 h, a 9-pl aliquot of each reaction was withdrawn, mixed with 30 pl of 10% acetic acid, and stored frozen until analysis. Samples were analyzed by RP-HPLC on Zorbax CN (0.46 X 25 cm) with elution of consecutive linear gradients of 90% TEAP, 10% acetonitrile at zero time to 75% TEAP, 25% acetonitrile at 20 min, to 72% TEAP, 28% acetonitrile at 50 min, and to 40% TEAP, 60% acetonitrile at 55 min. The elution profiles were monitored by UV absorbance at 210 nm.

We examined the extent to which rates of enzymatic conversion of the oxytocin biosynthetic precursor to mature peptide are modulated by intramolecular and intermolecular assembly of precursor and polypeptide intermediates. The biosynthesized precursor contains hormone and neurophysin sequences linked by a Gly-Lys-Arg sequence and undergoes enzymatic processing reactions which include endoproteolytic cleavage at the Lys-Arg dibasic sequence, carboxypeptidase B-like exoproteolytic cleavage, and enzymatic amidation. We evaluated the effect of neurophysin on such processing reactions using semisynthetic precursors of oxytocin1 bovine neurophysin I and synthetic oxytocinyl precursor intermediates as substrates. Neurophysin I at high concentration (0.7 mM) reduced the rates of carboxypeptidase B-like conversion of oxytocinyl-Gly-Lys-Arg to oxytocinyl-Gly and the enzymatic amidation of oxytocinyl-Gly to mature (C-terminal amidated) oxytocin. The dependence of rate suppression on the concentrations of peptide substrate and neurophysin I suggested that suppression is due to intermolecular formation of hormone-neurophysin complexes which are aggregated at least to dimers. An analogous intramolecular neurophysin effect was found for endoproteolytic processing of semisynthetic precursors. Endoproteinase Lys-C cleaved the Lys"-Arg12 peptide bond in a native-like semisynthetic precursor at a significantly slower rate than it did an assembly-deficient precursor analogue. The difference in semisynthetic precursor endoproteolysis rates is most substantial at the high concentrations at which the native-like precursor would form dimers but the assembly-deficient analogue would not. The native-like semisynthetic precursor was more stable than the assembly-deficient precursor analogue to tryptic digestion. The concentration-dependent effects of neurophysin, both intramolecularly as a precursor domain and intermolecularly as an interacting protein, are likely to occur in the secretory granules in which the biosynthetic precursors are packaged. The molecular organization of both hormone/neurophysin precursors and the noncovalently complexed hormone-neurophysin intermediates can be expected to play a role in modulating enzymatic processing reactions that lead to mature neurohypophysial hormones.
The neurohypophysial nonapeptide hormones oxytocin and vasopressin are synthesized as parts of common precursor * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. proteins with neurophysins I and 11, respectively. The complete amino acid sequences of the precursors, as deduced by sequencing cloned complementary DNA (1,2), have the general form, hormone-Gly-Lys-Arg-neurophysin-carboxyl-terminal extension. Production of biologically active oxytocin from its precursor occurs in secretory granules during axonal transport by post-translational enzymatic processing reactions that include endoproteolytic cleavage at the Lys-Arg sequence between the hormone and neurophysin sequence, exoproteolytic removal of the terminal basic residues (Arg and Lys) of hormone intermediates by a carboxypeptidase Blike enzyme and perhaps an aminopeptidase, and amidation at sites marked by a C-terminal Gly. Processing enzymes which account for all of these conversion reactions have been detected in secretory granules. tion of a dibasic endoprotease which cleaves pro-opiomelanocortin, proinsulin, and proarginine vasopressin/neurophysin 11. A secretory granule-associated carboxyeptidase B-like enzyme has been reported (5-8) and recently purified and cloned (9). And, Bradbury et al. (10) and Eipper et al. (11) have detected and purified an amidating enzyme which converts peptides terminating in X-Gly-COOH sequences to X-amides. Characterization of the reactions and specificities of the range of processing enzymes identified in vitro has been accomplished mainly with small molecular weight synthetic peptide substrates. As a result, current understanding of the molecular mechanisms underlying neuropeptide precursor processing as it would occur in secretory granules, including the impact of precursor folding and multimolecular assembly, is still rudimentary.

769
Neurophysin Effect on Enzymatic Maturation of Oxytocin neurophysin domains interact intramolecularly and the entire molecule dimerizes (12,13). Mature hormones and neurophysins are present in granules at concentrations of 60-70 mM a n d t h u s are likely t o be present predominantly as dimers or larger aggregates (14). Similarly, precursors most likely assemble into self-associated forms upon signal peptide cleavage and intragranular packaging (13). Thus, enzymatic processing of precursors occurs on hormone substrates that almost certainly interact with neurophysins in self-associated complexes. The impact of such molecular interactions on the enzymatic processing reactions which form oxytocin and vasopressin has not been evaluated in depth.
In this study, we have examined the effects of the hormoneneurophysin interaction on enzymatic processing using both semisynthetic oxytocinyl precursors and synthetic oxytocin intermediates as substrates. We compared endoproteolysis rates with trypsin and endoproteinase Lys-C for semisynthetic precursors in which the neurophysin and hormone domains interact to different extents. Carboxypeptidase B and amidation rates were compared for oxytocinyl peptides alone and in the presence of neurophysin.

MATERIALS AND METHODS
Synthesis of Oxytocinyl-Gly-Lys-Arg (OT-GKR)-OT-GKR was synthesized, analogously as before (la), by conventional solid-phase peptide synthesis using Boc-Arg (tosy1)-phenylacetamidomethyl resin (Vega Biotechnologies, Inc., Tucson, AZ). After stepwise coupling of Boc-amino acids using the DCC method and, for Asn and Gln, the dicyclohexylcarbodiimide-hydroxybenzotriazole active ester method, the protected peptide resin was treated with hydrogen fluoride at 0 "C for 1 h. Formation of the disulfide bridge between 1/zCys residues 1 and 6 was carried out with potassium ferricyanide. The crude product was purified by gel filtration on Sephadex G-25 (1.7 X 140 cm) in 10% acetic acid and then by RP-HPLC on Zorbax ODS (0.94 X 25 cm) with an elution by two consecutive linear gradients, from 0.1% trifluoroacetic acid in 20% acetonitrile at zero time to 0.1% trifluoroacetic acid in 40% acetonitrile at 30 min and then from the latter to 0.1% trifluoroacetic acid in 70% acetonitrile at 40 min. The main peak fractions were collected and lyophilized. The yield of final product calculated from Boc-Arg(tosy1)-phenylacetamidomethylresin was about 50%. Amino acid composition from analysis after hydrolysis with constant boiling HCl at 110 "C for 24 h in uacw was: Asp (LOO), Glu (0.92), Pro (1.05), Gly (2.21), %Cys (1.74), Ile (0.84), Leu (1.09), Tyr (0.86), Lys (1.09), and Arg (1.10).
Preparation of Oxytocinyl-Gly (OT-GI-OT-G was prepared from OT-GKR by treatment with porcine pancreatic carboxypeptidase B (Sigma). OT-GKR (0.5 mg by weight) was dissolved in 180 pl of 0.2 M sodium hydrogen carbonate, pH 8.3, and incubated with 20 pl of carboxypeptidase B solution (5.6 mg of protein/ml in 0.1 M NaC1) at 37 "C for 30 min. The reaction was monitored, and the OT-G produced was purified by RP-HPLC on Zorbax CN (0.46 X 25 cm) with elution by three consecutive linear gradients, from 95% TEAP (67 mM phosphoric acid/triethanolamine buffer, 3.0). 5% acetonitrile at zero time to 85% TEAP, 15% acetonitrile at 15 min, from the latter to 82% TEAP, 18% acetonitrile at 36 min, and then from the latter to 30% TEAP, 70% acetonitrile at 50 min.
The fractions containing OT-G were collected and lyophilized. OT-G then was desalted on the same column with an elution of consecutive linear gradients, from 0.1% trifluoroacetic acid in 5% acetonitrile at zero time to 0.1% trifluoroacetic acid in 25% acetonitrile at 10 min and then from the latter to 0.1% trifluoroacetic acid in 30% acetonitrile at 40 min. The yield was about 40%. Amino acid composition from analysis after acid hydrolysis (as above) was: Asp (0. Boc groups were removed with dry trifluoroacetic acid. Yields of deprotected products were 10-16% based on starting amount of acetimidated bovine neurophysin I. Amino acid analyses after acid hydrolyses (as above) and amino-terminal sequence analysis confirmed that the synthetic peptides were coupled to the a-amino group of c-amino acetimidated neurophysin I in a molar ratio of 1:l. Analytical Affinity Chromatography of Oxytocinyl Peptide Intermediates on Neurophysin Matrix-The binding affinities of OT-GKR, OT-GK, and OT-G for neurophysin were measured by analytical HPLAC using bovine neurophysin I1 immobilized on highly crosslinked agarose (15). The matrix (272 nmol of BNPII/ml-bed volume) was packed in an Omni column (Pierce Chemical Co., 0.66 X 7.5 cm). Peptides were dissolved at 1 mg/ml in 0.4 M ammonium acetate, pH 5.7. Different amounts (10,7.5,5,2.5, and 1 pl) of each peptide sample were eluted on the affinity matrix with the same buffer at a flow rate of 1 ml/min. Elution profiles were monitored by UV absorbance at 226 nm.
Isoiation of Neurosecretory Granules-Fresh bovine neurointermediate pituitaries were obtained from Treuth and Sons (Catonsville, MD). Isolation of neurosecretory granules was carried out, at 0 "C on ice, unless stated otherwise, following the method of Russell (16). The pituitaries (12 pieces) were minced and homogenized in 20 ml of 0.25 sucrose containing 10 mM HEPES, pH 7.0. After continuous centrifugation (5 "C) at 3,000 X g for 2 min and at 4,000 X g for 15 min (Sorvall RCLB, SS-34 rotor), the supernatant was centrifuged at 26,000 X g for 15 min (Sorvall RCPB, SS-34 rotor) to obtain the crude secretory granule pellet. This pellet was suspended in 1 ml of 30% Percoll in 0.25 M sucrose (d 1.065), placed on 10 ml of 30% Percoll, and centrifuged at 50,000 X g for 45 min (Beckman L5-65 ultracentrifuge, Ti5O-rotor). Six fractions of 50 drops (2 ml) each were collected by puncturing the bottom of the tube, and 8 ml of 0.25 M centrifuged at 105,000 X g for 45 min (Beckman LS-65, SW 41 Ti sucrose solution were added to each fraction. All fractions then were rotor) to remove Percoll. The liquid layer (0.5 ml) of secretory granules (above the semisolid Percoll layer) was withdrawn using a Pasteur pipette. After sonication, fractions 1-4 (highest density fractions) were found to have carboxypeptidase B-like enzyme activity specifically activated by cobalt ion. The lysates of these fractions were stored at -70 "C until used for carboxypeptidase B assays.
Purification of Amidation Enzyme-The purification of peptidylglycine a-amidating monooxygenase from frozen bovine neurointermediate pituitary was as reported before (17,18). Assays utilized purified peptidyl-glycine a-amidating monooxygenase B pooled after final gel filtration on Sephadex G-75 in 0.1 M NHIHCO~ and stored at -20 "C in the presence of 50% ethylene glycol and 1.0 mg/ml bovine serum albumin. The particular preparation of peptidyl-glycine a-amidating monooxygenase B utilized consisted of approximately equal amounts of 38-and 43-kilodalton proteins having the same amino-terminal amino acid sequence (18). Sample analysis was by RP-HPLC as described above for preparation of OT-G. Elution profiles were monitored by UV absorbance at 220 nm. The degree of conversion from OT-GKR to OT-GK and OT-G was calculated from peak area (height X width at half-height). Controls were performed in which ribonuclease S-protein instead of BNPI was added to oxytocinyl peptides in enzyme reaction mixtures. Amidation of OT-C-OT-G (4.2 nmol, 4.5 pg) and bovine neurophysin I (4.2 nmol, 39.5 pg; or 42 nmol, 395 pg), in water, were transferred to small centrifuge tubes and lyophilized. The lyophilized samples were dissolved in 45 pl of 150 mM MES, pH 5.5, or 150 mM TES, pH 8.0,8.5, and 9.0, titrated with NaOH and incubated with 11 ng of purified peptidyl-glycine a-amidating monooxygenase B in a final volume of 60 pl containing 3 p~ CuSO,, 2 mM ascorbate, and 0.1 mg/ml catalase. The molar ratio of peptidyl-glycine a-amidating monooxygenase B to OT-G, based on a 40-kilodalton mass for the former, was approximately 1:15,000. After incubation at 37 "C for 8 h, 40-p1 aliquots were withdrawn and stored frozen until analysis. Samples were analyzed by RP-HPLC on Zorbax CN (0.46 X 25 cm) using the same elution gradient scheme as for assay of carboxypeptidase reactions. Elution profiles were monitored by UV absorbance at 215 nm. The ratio of conversion from OT-G to amidated oxytocin Enzymatic Maturation of Oxytocin 771 was calculated from peak areas. In controls, S-protein was substituted for BNPI in OT-G reaction mixtures.

Site-specific Cleavage of Semisynthetic Precursors by Endoprotei-
nase Lys-C-Semisynthetic precursors Pro"-OT/BNPI and IN"'-Ac,AlaZ]pro"-OT/BNPI (5.9 nmol) were dissolved in 4.2 p1 of 150 mM MES, pH 5.5, and incubated at 37 "C with 4.2 p1 of endoproteinase Lys-C (Boehringer Mannheim) dissolved (at 38.5 ng/pl) in the same buffer. The final concentration of precursors was 700 p~; the ratio of precursor to enzyme was about 4001 by weight. The amount of endoproteinase Lys-C was chosen to slow the reactions enough to follow their time courses. After incubation for 15,30,45, and 60 min, 1.5 p1 of each sample were withdrawn, mixed with 30 pl of 10% acetic acid, and frozen until analysis. Samples were analyzed by RP-HPLC on Zorbax CN (0.46 X 25 cm) using an elution by consecutive linear gradients of 90% TEAP, 10% acetonitrile at zero time to 75% TEAP, 25% acetonitrile a t 20 min, to 72% TEAP, 28% acetonitrile at 50 min, and then to 40% TEAP, 60% acetonitrile a t 55 min. The elution profiles were monitored by UV absorbance at 210 nm. Limited proteolysis reactions also were carried out using diluted precursors (5.9 nmol of proteins dissolved in 420 p1 of 10 mM Na-H2P04/NaOH buffer, pH 5.7), which were incubated with 4.2 pl of endoproteinase Lys-C solution (38.5 nglpl). The final precursor concentration was 14 p~.

Binding of Synthetic Oxytocinyl Precursor Intermediates to
Neurophysin-An initial hypothesis in this study was that the synthetic oxytocinyl precursor intermediates OT-GKR, OT-GK, and OT-G are likely to bind to neurophysin noncovalently during processing and that these interactions could affect processing reactions by carboxypeptidase B-like and amidation enzymes. In order to confirm the expectation of neurophysin binding by oxytocinyl substrates, extents of neurophysin interaction were measured for OT-GKR and OT-G by analytical HPLAC on BNPII-highly cross-linked agarose as used previously (15)  neurophysin noncovalently in secretory granules during enzymatic processing.
Effect of Neurophysin on Carboxypeptidase B-like Processing of OT-GKR-We previously reported the action of a carboxypeptidase B-like activity in bovine posterior pituitary secretory granule lysate on OT-GKR (8). We found that rates of conversion of OT-GKR to OT-GK and OT-G could be determined by monitoring product and substrate chromatographically on Zorbax CN (Fig. 2). Using this method of analysis, we measured rates of conversion of 0.71 mM OT-GKR incubated with lysate in the presence of 0, 0.18, 0.35, 0.71, and 1.42 mM BNPI. The elution profiles obtained in the absence of neurophysin are shown in Fig. 2, and the data for all neurophysin concentrations are shown in Fig. 3. A suppression of reaction rate was observed a t all neurophysin concentrations except the lowest, 0.18 mM neurophysin I, a t which a slight but reproducible acceleration was observed. A slight acceleration also was observed with 71 PM OT-GKR and 71 PM BNPI uersus 71 PM OT-GKR alone (data not shown). The acceleration of carboxypeptidase B-like activity also was produced with S-protein, a polypeptide of almost the same molecular weight as BNPI but with no known binding site for oxytocin, and thus was concluded to be nonspecific. However, rate suppression was not detected at any concentration of Sprotein. These results argue strongly that the BNPI suppression of carboxypeptidase B-like conversion of OT-GKR is specific and depends on neurophysin binding to substrate. Available affinity data show that the K d values for oxytocin-BNPI binding and dimerization of oxytocin-bound bovine neurophysin are and M, respectively. Thus, it is most likely that, at the concentrations at which BNPI affects rate suppression, BNPI and oxytocin-containing peptides exist as noncovalent complexes that are predominantly dimerized. Thus, rate suppression appears to be correlated with formation of substrate-neurophysin dimers or higher aggregates.

Carboxypeptidase-B-like enzyme
FIG. 2. Enzymatic conversion of OT-GKR by secretory granule carboxypeptidase B-like activity. OT-GKR (71 nmol) was dissolved in 90 pl of 0.2 M sodium acetate, pH 5.5, containing 1 mM CoC12 and incubated with 10 pl of granule lysate at 37 "C. At 1, 3, and 9 h, 15 pl of sample were withdrawn, mixed with 35 p1 of TEAP, and stored at -70 "C until analyzed. In the right-most elution profile, the chromatographed sample was a mixture of a 7-pl aliquot of the 6-h reaction and 4.3 nmol of authentic oxytocin. Sample analyses were by elution on Zorbax CN as described under "Materials and Methods." Effect of Neurophysin on Arnidation of OT-G-Given the above, the effect of neurophysin on enzymatic conversion of OT-G to mature, COOH-terminally amidated oxytocin was evaluated. OT-G and mature oxytocin are distinguishable chromatographically as shown in Fig. 2. Oxytocinoic acid, which contains a free a-carboxyl instead of an amide group at Gly', elutes at the position of OT-G in the RP-HPLC system. When OT-G was incubated with purified peptidyl-glycine a-amidating rnonooxygenase B, significant conversion to oxytocin was detected only in alkaline buffers, with an optimal pH of about 8.5 (Fig. 4). No amidation conversion of OT-G to oxytocin was observed at the reported internal pH of 5.7 for mature secretory granules (19, no), although peptidylglycine a-amidating monooxygenase B is active at that pH based on the observed conversion of D-Tyr-Glu-Gly (18). When 70 PM OT-G was incubated with purified peptidylglycine a-amidating monooxygenase B (molar ratio of enzyme to substrate was 1:15,000) in 150 m M TES buffer, pH 8.5, for 8 h, 8% conversion was observed; no other degradation products were detected. In these amidation reactions with OT-G, conversion levels were kept relatively low by using low enzyme amounts. This condition was set up purposely in order to keep the reaction rates in the linear range. Higher levels of conversion could be observed by adding more enzyme. Repeat experiments were performed routinely to confirm the pattern of enzyme activity variation observed with the low conversion levels used.
T o evaluate the effect of neurophysin on amidation, the amidation assays were performed in the presence of 70 and 700 PM BNPI at pH 8.0, 8.5, and 9.0. Addition of 700 PM BNPI suppressed the amidation rate; but, again, at lower concentrations of BNPI (70 PM), rate enhancement was observed (Fig. 4). When the same reactions were done in the presence of 70 and 700 PM S-protein instead of BNPI, rate suppression was not seen at either concentration and only rate enhancement was observed, of a magnitude similar to that found with BNPI. Again, results argue that suppression of amidation of OT-G by high concentration of BNPI is specific, whereas the low concentration enhancement is not. The concentration dependence of rate suppression suggests a correlation of this rate effect with formation of self-associated substrate-neurophysin complexes. Noncovalent hormoneneurophysin binding and, to a lesser extent, neurophysin selfassociation are weakened as the pH is raised above 6 (21). However, both interactions do occur at pH values at least as high as 8 and also can be expected to potentiate each other by positive cooperativity.
Site-specific Cleavage of Semisynthetic Precursors by Endoproteinase Lys-C-The neurophysin effect on enzymatic processing of peptide intermediates leads to the postulate that endoproteolytic processing of oxytocinyl and vasopressinyl precursors themselves might be affected by their assembly properties. To test this idea, the susceptibility of pro"-OT/ BNPI and [Nu'-Ac,Ala']pro"-OT/BNPI to endoproteinase Lys-C was evaluated. Pro"-OT/BNPI folds with oxytocin and neurophysin domains interacting intramolecularly and, as a consequence, self-associates with relatively high affinity into precursor dimers and possibly higher order aggregates (13). In contrast, in the a-acetyl-Ala' variant, the acetylated, Ala' hormone domain does not interact with the neurophysin domain, and, thus, precursor derivatives self-associate with only low affinity (13).
Endoproteinase Lys-C cleaves only on the carbonyl side of lysyl residues. Since the 2 lysyl residues in the neurophysin domains of semisynthetic precursors are acetimidated, the enzyme can cleave only at the lysine in the linker region between hormone and neurophysin domains. Both pro*-OT/ BNPI and [Ne'-Ac,Ala']pro"-OT/BNPI were treated with endoproteinase Lys-C. Enzymatic degradation was carried out at pH 5.5, close to the pH of mature secretory granules (19, 20), in order to maintain the interaction between the hormone and neurophysin domains in pro"-OT /BNPI (22, 23). When either precursor (700 JLM) was incubated with enzyme, peaks corresponding to OT-GK, [N"'-Ac,Ala']OT-GK, and arginyl-[diAcetJBNPI were detected by 15 min (Fig. 5 ) . Under these conditions, the degradation of pro"-OT/BNPI was much slower than that of [N"'-Ac,Ala2]pro"-OT/BNPI. When lower concentrations (14 p M ) of each precursor were incubated with enzyme, the degradation rate of pro"-OT/BNPI was much greater than that of [Ne'-Ac,Ala2]pro"-OT/BNPI. Thus, the endoproteolysis rate of pro"-OT/BNPI, but not of [N"'-Ac,Ala2]pro"-OT/BNPI, was dependent on substrate concentration. At both concentrations used here (14 and 700 JLM), the hormone domain of pro"-OT/BNPI is expected to bind to the neurophysin domain in folded precursor, as judged by analytical affinity chromatography and CD measurement (12, 13). Thus, precursor folding does not interfere with endoproteolysis between the two domains and conceivably could enhance this degradation. But, when the concentration of precursor is sufficiently high, cleavage of pro'-OT/BNPI is suppressed, apparently due to self-association of the precursor to dimer or higher aggregated forms. The differential degradation rates of pro'-OT/BNPI and the N"'-Ac,Ala2 species at low concentration likely reflect the effects of intra-and intermolecular conformation (see "Discussion"). &action products were assayed by RP-HPLC as described under "Materials and Methods." After a 5-min reaction at 37 "C, peptide bond hydrolysis between Arg" and Ala13 was complete as judged by disappearance of intact substrate. The amount of [diAcetIBNPI, the first product formed and subsequently hydrolyzed to des-l-8-[diAcet] BNPI, was determined with time of reaction. The peak positions of both [diAcet]BNPI and the des-1-8 derivative were determined using standard [dAcetJBNPI (12,13) and the tryptic product formed from it analogously as before (24). The relative amount of [diAcetIBNPI remaining was determined from peak areas as the percentage ratio tween Lys"-Arg" was not detected in either precursor form tested here. Interestingly, Parish et al. (4) have reported that the endoproteolytic cleavage of provasopressin by an endoprotease purified from neural lobe occurs mainly at the carbonyl side of Arg.
The results of trypsin reaction provided a useful insight into the stabilization of neurophysin in intermediate complexes by following the time course of degradation of [diAcet] BNPI by trypsin. This degradation is expected to be due mainly to cleavage at the Ar$-Glng bond of [diAcet]BNPI to produce des-l--8-[diAcet]BNPI since this cleavage is the first step of degradation of unprotected neurophysin by trypsin (24). Only small amounts of other degradation products were detected in 4 h. The degradation of [diAcet]BNPI to des-l-8-[diAcet]BNPI proceeded faster in the mixture with [Nul-Ac,Ala2]0T-GKR than in that with OT-GKR (Fig. 6). These results suggest that formation of the complex of [diAcet] BNPI and OT-GKR protects the noncovalently bound hormone and neurophysin components from further endoproteolysis. [N"'-Ac,Ala2]OT-GKR does not bind to [diAcetIBNPI.

DISCUSSION
In this study, we have made several observations showing that enzymatic processing of oxytocin/neurophysin precursor and precursor intermediates is modulated by folding and multimolecular assembly. As shown in Figs. 3 and 4, both carboxypeptidase B-like and amidation enzyme reactions were suppressed by relatively high concentrations of neurophysin. Since OT-GKR binds to neurophysin with an affinity constant of K, 2.6 X lo4 M" (this study) and the complex of OT-GKR and neurophysin dimerizes with a constant of K, 1.7 X lo7 "' (12), the OT-GKR-neurophysin complex can be expected to dimerize readily when formed. At 0.71 mM OT-GKR and 0.18, 0.35, 0.71, and 1.42 mM neurophysin I, 25, 49, 99, and loo%, respectively, OT-GKR can be estimated to be present as dimer. The suppression effect on carboxypeptidase B-like and amidation reactions can be correlated with the formation of aggregates of monomer complex, presumably mostly dimer, but perhaps tetramer or higher order aggregation states as well. Tetramers of hormone-neurophysin com-plexes have been suggested from NMR (25), analytical ultracentrifuge (26), and x-ray diffraction (27) analyses.
Corresponding to these results with noncovalent hormoneneurophysin intermediate complexes, rate suppression by neurophysin domain was observed for precursor endoproteolysis. This can be seen in the variation of endoproteinase Lys-C cleavage rates at different precursor concentrations. When semisynthetic precursors at the relatively low concentration of 14 p M were treated with enzyme, cleavage at Lys" proceeded faster for pro"-OT/BNPI than for [Nu'-Ac,Ala*]pro"-OT/BNPI. Since the hormone domain of pro"-OT/BNPI can bind to the neurophysin domain at concentrations a t least as low as 1 p~ as judged by CD and affinity chromatographic analysis (12,13), folding of the close-to-native precursor per se, including the intramolecular domain-domain interaction, does not interfere with endoproteolytic cleavage a t Lys". Monomer folding in pro'-OT/BNPI even appears to cause a rate increase. Whether this latter effect might be related to the acceleration of carboxypeptidase B-like and amidation reactions in the presence of low concentrations of neurophysin I is not known at present.
In contrast, endoproteolysis of pro"-OT/BNPI proceeded significantly more slowly than that of [N"'-Ac,Ala2]pro"-OT/ BNPI at higher concentration, namely 700 p~. The cleavage rate for [N"'-Ac,Alaz]pro"-OT/BNPI is relatively unchanged at the higher concentration. It is very likely that reduction in rate for pro"-OT/BNPI is due to aggregation (self-association) of pro"-OT/BNPI. Association of pro'-OT/BNPI to liganded neurophysin has been measured to have K, lo7 "I. Thus, the precursor can be assumed to be effectively all dimer (or, again, possibly larger aggregates) at 700 pM.
Currently available data argue rather convincingly that oxytocinyl precursor and intermediate complexes exist mainly as dimers or higher order aggregates at the very high concentrations that exist in granules and thus that these self-assembled forms, and not monomers, are the predominant intragranular forms that exist during enzymatic processing (Fig.  7). The data presented here strongly suggest that enzymatic conversion of precursors to mature peptides is slowed substantially by this multimolecular assembly. However, two somewhat different possibilities for how molecular organization slows processing rates cannot be distinguished a t present. The fully aggregated precursors and intermediates may be low rate, but nonetheless the major substrates of enzymatic conversion in granules. Alternatively, fully assembled precursors may be converted at only extremely low rates or not at all, with degradation occurring predominantly with a relatively small pool of precursors and peptides which are dissociated of Oxytocin 775 to monomers. The latter seems less likely since concentration conditions which should lead to virtually full (>95%) selfassociation lead to observed rate decreases in the range of only 50-70%. The above arguments notwithstanding, the current results confirm the view that the rates of enzymatic conversion of oxytocinyl precursor to mature hormone likely are controlled by well-ordered folding and assembly events which occur in secretory granules. Judging from ongoing semisynthesis experiments, the vasopressinyl precursor also appears to fold and to self-associate? Thus, the neurophysin effect observed in this work for oxytocinyl precursor and intermediates is likely to be a general factor in modulating the rates of maturation of both neurohypophysial peptide hormones. In addition, multimolecular assembly of the precursors and intermediate complexes is likely to protect hormones and neurophysins from excessive enzymatic degradation, as seen in the precursor tryptic digestion experiment in this study. The actual biological role of processing rate modulation by the molecular organization of precursors, intermediates, and products remains to be clarified.