Specificity and Kinetic Studies on the Cleavage of Various Prohormone Mono- and Paired-basic Residue Sites by Yeast Aspartic Protease 3*

The specificity and relative efficiency of cleavage of mono- and paired-basic residue processing sites by YAP3p was determined in vitro for a number of prohor- mone substrates: human ACTH 1–39 , bovine proinsulin, porcine cholecystokinin 33, cholecystokinin (CCK) 13– 33, dynorphin A(1–11), dynorphin B(1–13), and amidorphin. YAP3p generated ACTH 1–15 from ACTH 1–39 . It cleaved proinsulin at the paired-basic residue sites of the B-C junction as well as the C-A junction. Leu-en-kephalin-Arg and Leu-enkephalin-Arg-Arg were gener- ated from dynorphin A and dynorphin B, respectively. YAP3p generated Met-enkephalin-Lys-Lys from amidor- phin showing that cleavage by this enzyme can occur at a lone pair of Lys residues. CCK33 was cleaved at Lys 23 and Arg 9 , each containing an upstream Arg residue at the P6 and P5 position, respectively. K m values were between 10 (cid:50) 4 and 10 (cid:50) 5 M for the various substrates, with the highest affinity exhibited for

Yeast cells express an alternate enzyme, an aspartic protease, YAP3p, encoded by the YAP3 gene, which can process pro-␣-mating factor at paired-basic residues, when this prohormone is overexpressed in a mutant yeast strain lacking the KEX2p serine protease (1). While the substrate that YAP3p normally processes in yeast is not known, it most likely functions as a processing enzyme since it is localized in the secretory pathway and can process transfected anglerfish prosomatostatin correctly (2). We have previously overexpressed the YAP3 gene in the BJ3501 strain of Saccharomyces cerevisiae, purified the recombinant YAP3p enzymatic activity, and investigated its ability to process the prohormones, mouse pro-opiomelanocortin (POMC) 1 and its fragments, and anglerfish pro-somatostatin I and II (aPSS I and aPSS II) in vitro (3,4). Those studies revealed that YAP3p specifically cleaved the pairedbasic residues of POMC to yield adrenocorticotropin (ACTH), ␤-endorphin, and ␥ 3 -melanocyte stimulating hormone, and aPSS I to yield somatostatin 14. YAP3p cleaved aPSS II at a monobasic residue which has an upstream Arg in the P6 position, to yield somatostatin 28, but cleavage at the analogous monobasic residue cleavage site in aPSS I, which has the upstream Arg substituted by a histidine, was not observed (4). This latter finding indicated that in addition to recognizing the paired-basic residue motif, YAP3p also appeared to show specificity for a monobasic residue that had an Arg present upstream. Furthermore, YAP3p cleaved the monobasic residue site of aPSS II more rapidly than the paired-basic residue site of aPSS I, raising the possibility that YAP3p may have a greater preference for monobasic residue sites.
However, in vitro analysis of the specificity of YAP3p on POMC, aPSS I, and aPSS II alone is insufficient to completely define the specificity of YAP3p or to conclude that YAP3p would in general cleave the monobasic residue motif of prohormones more efficiently than the paired-basic residue motif. In this study, we have investigated extensively the specificity and kinetic parameters governing the processing of six different substrates of physiological importance; proinsulin, cholecystokinin 33, dynorphin A, dynorphin B, amidorphin, and adrenocorticotropic hormone, containing mono-, paired-, and tetrabasic residue sites, by YAP3p. This study also provides the first kinetic information on the efficiency of cleavage of paired versus monobasic residue sites by an aspartic protease processing enzyme. In addition, YAP3p has similar specificity to the proopiomelanocortin converting enzyme (EC 3.4.23.17), a mammalian prohormone processing aspartic protease functionally and immunologically related to YAP3p (3). Thus, kinetic studies on YAP3p as a model enzyme, which is obtainable in large quantities, will also likely provide a better understanding of the substrate parameters governing the specificity and efficiency of cleavage of prohormones by mammalian processing aspartic proteases.
Yeast Aspartic Protease 3-Transformed yeast, strain BJ3501 containing the YAP3 gene, was induced to express YAP3p activity as described previously (3). The YAP3p enzymatic activity that was secreted into the growth media was partially purified by concanavalin A affinity chromatography (4). For the kinetic experiments (Figs. 2 and 3), YAP3p was purified further by DEAE-Sepharose anion exchange chromatography, desalted, lyophilized, reconstituted with water, and stored at Ϫ20°C until use. One unit of YAP3p activity is expressed as the amount of enzyme that generates 0.18 g of ACTH 1-15 from 10 g of ACTH 1-39 in a 100-l 0.1 M sodium citrate buffer, pH 4.0, and 37°C for 30 min. In this preparation (1 unit of YAP3p Ӎ0.5 fmol), it was determined that no other proteolytic activity was present by the following criteria. 1) Growth media or concanavalin A-purified growth media from both untransformed and transformed/uninduced yeast cells were found to have no proteolytic activity when assayed by the sensitive 125 Ih ␤-Lipotropin assay. In this assay, activity was measured by the ability of a protease to generate trichloroacetic acid-soluble counts/min from 125 Ih ␤-Lipotropin. 2) Pepstatin A, which is a specific inhibitor of aspartic proteases, completely inhibited the secreted proteolytic activity from the galactose-induced transformed yeast. 3) Using ACTH 1-39 as substrate, the two products generated by YAP3p which have been identified previously by HPLC and amino acid sequencing (3), as ACTH 1-15 and CLIP 16 -39 , were stable over time. There were no anomalous cleavages in the presence of pepstatin A, indicating the absence of any other proteolytic activity including carboxypeptidase or aminopeptidase activity in the enzyme preparation.

Generation and Identification of Products Generated by YAP3p
Cholecystokinin 33-One hundred ng of CCK33 (26 pmol) was incubated with 0.07 pmol (140 units) of YAP3p in 100 l, 0.1 M sodium citrate, pH 4.0, for 1, 2, 5, and 8 h at 37°C in the presence and absence of 3.7 ϫ 10 Ϫ5 M pepstatin A (ICN Biomedicals Inc., Aurora, OH). Substrate alone and enzyme alone were incubated in parallel, as controls. The reaction was stopped by addition of 50 l of 0.1 M HCl and immediately frozen on dry ice until analysis. Products were separated from the substrate by DEAE-Sephadex A-25 ion exchange resin and then quantitated by radioimmunoassay (RIA) with an antibody that recognizes CCK8, CCK12, and CCK33 (5). The products were then further analyzed by Sephadex G-50 gel filtration chromatography, and each fraction was analyzed by RIA. The CCK8-sized peak of immunoreactivity from the Sephadex G-50 column was analyzed further by HPLC separation followed by RIA as described previously (5).
ACTH  , Dynorphin A, Dynorphin B/Rimorphin, Amidorphin, Proinsulin, Proinsulin B-C Junction Peptide, and Cholecystokinin 13-33-Ten g of each peptide (2-9 nmol) were incubated with enzyme (0.014 -0.7 pmol, i.e. 28 -1410 units) in 100 l, 0.1 M sodium citrate, pH 4.0, at 37°C. The reactions were stopped by the addition of 10 l of glacial acetic acid, and the products were separated on an LKB 2150 HPLC system using a Bio-Rad HiPore RP-318 column (5 ϫ 250 mm). Buffer A was 0.1% trifluoroacetic acid, and buffer B was 80% acetonitrile in 0.1% trifluoroacetic acid. The gradients for each substrate are indicated in their respective figure legends. Products were monitored by absorbance at 214 nm, and individual peaks were collected for identification by amino acid sequence analysis. The peptide products from several HPLC runs were pooled and prepared for N-terminal amino acid sequence analysis by microcentrifugation in a Pro-Spin cartridge (Applied Biosystems, Foster City, CA.). Amino acid sequence analysis was carried out by Edman degradation using an Applied Biosystems Model 470A Protein Sequencer with an on-line phenylthiohydantoin (PTH) analyzer. Control experiments for each substrate included incubations with pepstatin A, enzyme alone, and substrate alone.

Cleavage of Native and Denatured/Reduced Proinsulin
Twenty g of bovine proinsulin (2 mg/ml) was boiled for 5 min in 0.03 M sodium citrate, pH 4.0, and 2% (v/v) Triton X-100. After microcentrifugation to collect the condensate, dithiothreitol (20 mM final) and water was added to a final volume of 30 l. Two pmol of YAP3p was added and incubated for 1 h at 37°C after which 10 l of a 4 ϫ SDS sample buffer was added to the reaction mixture and then prepared for SDS-PAGE. The proteins were separated by SDS-PAGE on a 16% Tris/Tricine gel and stained by colloidal Coomassie Blue. An identical aliquot of proinsulin that had not been boiled was incubated with the same amount of YAP3p in the absence of dithiothreitol and detergent and analyzed in the same manner. Triton X-100 was added to the amount equivalent to that of the denatured/reduced sample prior to the SDS-PAGE in order to ensure similar electrophoretic patterns of the proinsulin products so that a more accurate comparison could be made by inspection of the intensity of the staining of the proteins in each reaction. An incubation of proinsulin alone was used as a negative control. The gel was scanned by a Scanmaker II scanner (Microtek International, Inc., Taiwan, Republic of China) using AdobePhotoshop TM LE software. The image was analyzed by NIH Image v1.57 to quantitate the amount of substrate remaining in comparison to the negative control.

Kinetic Analysis of Products Generated by YAP3p
Determination of K m and V max -The method of Lineweaver-Burk (6) was used to calculate K m and V max values for the generation of the major products from each substrate. Initially, a time course for the generation of each product was done by incubating YAP3p with the highest and lowest concentrations of each substrate that would be used in the kinetic assays. This was done to determine the fixed time frame for a linear response for the generation of the products. The amount of enzyme to be used for each substrate was determined individually such that Ͼ90% of the substrate remained after the reaction in order to minimize product inhibition. It was assumed that a linear response (i.e. initial rate, V 0 ), obtained at the two extreme substrate concentrations, would also generate linear responses at any substrate concentrations within these two limits. Four concentrations of each substrate, spanning approximate K m values obtained from preliminary experiments, were then incubated with the enzyme. The concentration of YAP3p used The kinetic assays were done in triplicate. Products were measured in centimeters of peak height when analyzed by reverse phase HPLC and detected by absorbance at 214 nm and converted to nanomoles of product generated using standard curves generated on the same HPLC column under identical gradient conditions. The product of the ACTH 1-39 reaction, ACTH 1-15 , was quantitated by comparison to a standard curve generated from ACTH 1-14 while the products of dynorphin A, dynorphin B, and amidorphin were quantitated by standard curves of Leu-enkephalin-Arg, Leu-enkephalin-Arg-Arg, and Met-enkephalin-Lys-Lys, respectively. The product of proinsulin, a proinsulin intermediate, was quantitated by a standard curve of proinsulin itself. Since there are no commercially available standards for the product of CCK13-33, it was quantitated based on the molar decrease of the substrate. The Lineweaver-Burk plots were generated by plotting 1/nmol of product generated/min versus 1/[substrate] using Computer Associates, Cricket Graph for Windows, v1.3.1. The corresponding kinetic parameters, K m and V max , were determined directly from these plots. Initial rate data for the hydrolysis of the proinsulin B-C junction peptide were obtained in two separate time course experiments using overlapping substrate concentrations and analyzed by Lineweaver-Burk plot (Fig. 3).
Determination of k cat -The absolute concentration of the active enzyme could not be determined since at present there are no irreversible active site titrants for YAP3p. Also, due to the relatively low affinity of YAP3p for pepstatin A (K i ϭ 3.3 ϫ 10 Ϫ7 M), 2 it was not possible to approximate the concentration of YAP3p by titration of the activity with this inhibitor. Enzyme concentration was therefore estimated based on protein concentration by comparison of the intensity of a Coomassie Blue-stained band of an aliquot of the YAP3p that was used in the kinetic reactions to a standard curve generated using purified YAP3p. Serial dilutions of purified YAP3p, quantitated by N-terminal amino acid analysis, were run on SDS-PAGE in parallel with an aliquot of the unknown YAP3p solution. The gel was stained by colloidal Coomassie Blue (Novex, San Diego, CA), dried, and scanned with a Digital CCD camera. The scanned image was processed using public shareware NIH Image, v1.57, and the areas of the stained proteins were quantitated. A plot of arbitrary units of area against the known YAP3p amounts loaded generated a standard curve with the equation, y ϭ 2.55 ϩ 0.51x, r 2 ϭ 0.987. YAP3p in the unknown sample, that fell within the line, was calculated from this standard curve. Using these data, k cat values, obtained from V max /[E] T ϭ kcat, where [E] T ϭ total enzyme, were calculated. Since the YAP3p protein concentration and not the active enzyme concentration was used to calculate k cat , the values obtained may be an underestimation. However, this will not affect the comparison of the relative cleavage efficiency (k cat /K m ) of YAP3p for the different substrates which have been assayed using the same enzyme preparation. Lys 59 -Arg 60 . Based on the average yields of PTH-derivatives, a 4:1 ratio is calculated between the two cleavages at the bond Lys 59 -Arg 60 and Arg 60 -Gly 61 , respectively. Peak C was identified as C peptide. Pepstatin A inhibited the generation of these products (Fig. 1B, lower panel). The V max for the processing of proinsulin at the B-C junction was determined to be 7.9 Ϯ 0.4 pmol/min ( Fig. 2A), and the K m , k cat , and k cat /K m values are shown in Table I. The cleavage specificity of proinsulin B-C junction peptide by YAP3p was shown to be identical with that of proinsulin itself, cleaving preferentially on the carboxyl side of the Arg-Arg pair and in the same ratio (data not shown). This peptide was processed by YAP3p with a V max of 214.9 Ϯ 26.4 pmol/min (Fig. 3) while the K m , k cat , and k cat /K m values for the generation of these products are reported in Table I.

Proteolytic Processing of Proinsulin by
Cleavage of Native and Denatured/Reduced Proinsulin-Both native and denatured/reduced proinsulin were cleaved by YAP3p (Fig. 4). However, only ϳ14% of the denatured/reduced proinsulin (lane 3) was cleaved compared to ϳ34% of the native proinsulin (lane 2) relative to the negative control (lane 1). An additional control experiment using ACTH 1-39 as a substrate in these denaturing/reducing conditions demonstrated that YAP3p activity was not inhibited when compared to its activity against ACTH  in the absence of these conditions (data not shown).
Proteolytic Processing of CCK33 by YAP3p-YAP3p generated CCK8-immunoreactive products from CCK33 in a timedependent manner (Fig. 5). Pepstatin A completely inhibited the generation of these products (Fig. 5). When products from the 5-h time point were separated on a Sephadex G-50 gel filtration column and assayed by RIA, two peaks of CCK8-immunoreactivity which co-eluted with CCK22 and CCK8 standards were observed (Fig. 6B). Further analysis of the CCK8-sized peak by HPLC showed an elution profile that was identical with oxidized CCK8 standards (7) (Fig. 6C), indicating that YAP3p cleaved on the carboxyl side of the monobasic Arg 9 of CCK33 to generate CCK8. To verify the cleavage by YAP3p at Lys 23 of CCK33 that would generate CCK22 (see Fig.  6A), CCK13-33 was incubated with YAP3p, and the products were analyzed by HPLC. Two homogenous products were generated (Fig. 7B, upper panel) and identified as CCK23-33 and CCK13-22, respectively, by amino acid sequence analysis, indicating cleavage by YAP3p on the carboxyl side of Lys 23 . The presence of pepstatin A completely inhibited the formation of these products (Fig. 7B, lower panel). No products were detected when substrate alone or enzyme alone were incubated (data not shown). The V max for the cleavage of CCK13-33 was determined to be 29.2 Ϯ 1.4 pmol/min (Fig. 2B), and the K m , k cat , and k cat /K m values are shown in Table I.
Proteolytic Processing of Dynorphin A, Dynorphin B, Amidor-   phin by YAP3p-YAP3p generated Leu-enkephalin-Arg and Leu-enkephalin-Arg-Arg from dynorphin A and dynorphin B, respectively (Figs. 8 and 9). Products generated from dynorphin A (Fig. 8B, upper panel) were identified as (R)IRPK, Leu-enkephalin-Arg-Arg (minor), and Leu-enkephalin-Arg (major), by amino acid analysis and comparison to standards that were run on the same HPLC system. In the peak identified as (R)IRPK, the ratio of this pentapeptide to its Arg truncated tetrapeptide (IRPK) was approximately 35:1 based on the yields of PTH-Arg and PTH-Ile at cycle 2 of the amino acid sequence analysis and comparison of peak heights between the peaks of Leu-enkephalin-Arg and Leu-enkephalin-Arg-Arg. This demonstrated that YAP3p preferentially cleaved between the Arg 6 -Arg 7 pair of dynorphin A to yield mainly Leu-enkephalin-Arg. Products generated from dynorphin B (Fig. 9B, upper panel) were identified as QFKVVT and Leu-enkephalin-Arg-Arg which demonstrated a processing pattern different from that of dynorphin A because the preferred site of cleavage was on the carboxyl side of the Arg 6 -Arg 7 pair of dynorphin B to yield Leu-enkephalin-Arg-Arg. Amidorphin was cleaved by YAP3p to generate primarily Met-enkephalin-Lys-Lys (Fig.  10B). The generation of all products by YAP3p was inhibited by pepstatin A (Figs. 8B, 9B, and 10B, lower panels), and no products were detected when substrate alone or enzyme alone were incubated (data not shown). The V max for the generation of Leu-enkephalin-Arg from dynorphin A was 118.5 Ϯ 4.4 pmol/ min, Leu-enkephalin-Arg-Arg from dynorphin B was 45.9 Ϯ 6.2 pmol/min, and Met-enkephalin-Lys-Lys from amidorphin was 30.1 Ϯ 3 pmol/min (Fig. 2, C, D, and E) and the K m , k cat , and k cat /K m values for the generation of these products are reported in Table I. Proteolytic Processing of ACTH 1-39 by YAP3p-Secreted YAP3p cleaved ACTH 1-39 to yield ACTH 1-15 and CLIP 16 -39 (data not shown), products identical with that previously reported for the cleavage of ACTH 1-39 by YAP3p purified from yeast cell extracts (3). The V max for the generation of ACTH 1-15 by YAP3p was determined to be 33.2 Ϯ 4.8 pmol/min (Fig. 2F), and the K m , k cat , and k cat /K m values are shown in Table I. FIG. 6. Proteolytic processing of cholecystokinin 33. A, primary amino acid sequence of CCK33. Arrowheads indicate the sites cleaved by YAP3p. R ϭ arginine, K ϭ lysine, Y* ϭ sulfated tyrosine. B, Sephadex G-50 gel filtration profile of CCK8-immunoreactive products generated by YAP3p. C, radioimmunoassay of fractions from reverse-phase HPLC separation of the CCK8-sized product generated from incubation of CCK33 with YAP3p. Products were purified by Sephadex G-50 filtration, and the CCK8-sized product was further analyzed by HPLC. Arrows in B and C indicate positions of standards. Numbering of the CCK peptides start from the C terminus.

DISCUSSION
Processing of prohormones to yield active hormones occurs most commonly at paired-basic residues and to a lesser extent at specific monobasic residue sites (8 -11). A number of prohormone processing enzymes capable of carrying out these specific mono/paired-basic residue cleavages have been described. They include the serine/subtilisin-like enzymes, furin and the proprotein convertases, (12)(13)(14)(15)(16)(17)(18)(19), and representatives from the thiol (20,21) and aspartic protease classes (1,(22)(23)(24)(25). YAP3p is a member of the aspartic protease family of prohormone processing enzymes. In this study, we have analyzed a set of substrates to further define the specificity of YAP3p and the mono/ paired-basic residue cleavage site motifs recognized by this enzyme. The catalytic efficiency (k cat /K m ) of YAP3p for the cleavage of these substrates was determined.
YAP3p cleaved bovine proinsulin at the B-C junction which contains the sequence PKARRE (Fig. 1A). Cleavage occurred preferentially on the carboxyl side of the Arg-Arg pair, putting the Lys at P4, while some cleavage was observed in between, putting the Lys at P3. This preference is similar to that of furin where an Arg at P4 enhances the cleavage rate of synthetic proalbumin peptides (26) while an Arg at P3 is deleterious (27). The synthetic substrate, Boc-RVRR-methylcoumarin amide was cleaved by YAP3p (3) exclusively in between the Arg-Arg pair, rendering the assignment of the upstream Arg in this substrate in the P3 position (Table II). The absence of any cleavage on the carboxyl side of the pair of Arg residues is presumably due to the steric hindrance of the 7-amino-4-methyl-coumarin moiety that would be in the S1Ј pocket of the active site. Anglerfish pro-somatostatin I (aPSS I), containing the sequence PRQRKA (Table II), was cleaved by YAP3p, similar to the proinsulin B-C junction, primarily on the carboxyl side of the Arg-Lys cleavage site, although some cleavage in between the pair was observed (4), indicating an overall degree of tolerance but not preference by YAP3p for a basic residue in the P3 position. In an effort to determine the role that structural conformation plays in the efficiency of this reaction, we compared the ability of YAP3p to cleave native versus denatured/ reduced proinsulin. YAP3p cleaved the native proinsulin better than the denatured/reduced proinsulin (Fig. 4), demonstrating that disruption of the conformation of the prohormone negatively affected the efficiency of its cleavage. When we analyzed the cleavage of the proinsulin B-C junction peptide, we found that the specificity was identical with that of the full-length prohormone itself, demonstrating that the primary sequence around the cleavage site was sufficient to dictate the specificity  8. Proteolytic processing of dynorphin A(1-11). A, primary amino acid sequence of dynorphin A (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11). Big and small arrows indicate major and minor sites cleaved by YAP3p. B, HPLC profile of products generated from dynorphin A incubated with YAP3p in the absence (upper panel) and presence (lower panel) of pepstatin A. Ten g of dynorphin A was incubated with 0.014 pmol of YAP3p for 2 h. Gradient was 10% B for 10 min and 10 -40% B in 50 min. but the K m was greatly increased.
The bovine proinsulin C-A junction contains the sequence, PPQKRG, i.e. a lone pair of basic residues. YAP3p cleaved the C-A junction of proinsulin preferentially in between the Lys-Arg pair indicating that an Arg residue in the P1Ј position is an acceptable position. Cleavage at this site was relatively slow compared to the B-C junction, perhaps due to the presence of the two prolines immediately upstream from the cleavage site (P3 and P4). Further support of this hypothesis is borne out by our studies with aPSS I and aPSS II (4). While aPSS I was cleaved by YAP3p at the Arg-Lys pair preferentially after the Lys to yield somatostatin 14 (Table II), the cleavage of aPSS II at the analogous Arg-Lys site was not detected, possibly due to the presence of the two prolines immediately upstream (at P5 and P6) from the expected cleavage site (Table II). The proline side chain is known to have less conformational freedom resulting in a more rigid structure around the peptide bond. This may prevent the substrate in the cleavage site region from assuming a structure that binds efficiently to the active site pocket of the aspartic protease which generally spans up to 10 amino acids (28,29). It is noteworthy that the specificity of YAP3p for the cleavage sites of bovine proinsulin is subtly different, especially at the C-A junction, to that observed in vivo where processing occurs exclusively on the carboxyl side of both sites.
YAP3p has been shown to cleave aPSS II in vivo and in vitro at a monobasic Arg residue within a motif containing an upstream Arg at P6 (2, 4). However, in these two studies, cleavage of aPSS I, where the upstream Arg at P6 is substituted by a histidine, was observed by one group (2) and not by the other (4). This discrepancy exemplifies the differences that can be observed between in vivo and in vitro studies where the regulation of enzyme to substrate ratio dictates the efficacy of the reaction. The finding that no cleavage by YAP3p was observed when the ␣-mating factor-leader-proinsulin fusion protein was mutated to delete the Arg at the Lys-Arg junction (1) ( Table II), suggests that YAP3p does not cleave monobasic sites without an additional basic residue upstream or downstream.
To determine if YAP3p will cleave a monobasic Arg, as well as a Lys, within a motif having an upstream basic residue, CCK33 was tested as a substrate. YAP3p cleaved sulfated CCK33 at two monobasic residue sites, each containing an upstream Arg at either the P6 or P5 position ( Fig. 6A and Table  II). Cleavage at Arg 9 generated CCK8, while cleavage at Lys 23 generated CCK22 (Fig. 6, B and C). YAP3p was also shown to cleave CCK13-33 to CCK13-22 and CCK23-33 at Lys 23 (Fig.  7B). These results indicate that YAP3p can recognize both a monobasic Lys or Arg apparently within a motif containing an upstream Arg. This cleavage specificity exhibited by YAP3p is similar to the CCK8 generating enzyme previously described from rat brain synaptosomes which is capable of both these cleavages (30). Based on the relative concentrations of the products generated at the 5-h time point (Fig. 6B), it appears that YAP3p cleaved preferentially at Lys 23 to generate CCK22 rather than at Arg 9 to generate CCK8. This result may be an indication that YAP3p prefers mono-Lys sites over mono-Arg sites, or simply that the upstream Arg in the P6 position is more favorable than in the P5 position. The presence of a negatively charged sulfated tyrosine close to Arg 9 may also render this site difficult to cleave. While the presence of an Arg in the P5 and P6 positions of the cleavage sites of CCK33, and that of Arg in the P6 position of the monobasic cleavage site of aPSS II appear to correlate with cleavage at these sites, the presence of other amino acids around the cleavage site may also be important. For instance, a bulky hydrophobic residue at the P2Ј position appears to also correlate with the cleavage at monobasic residues (Table II). The extent of the effect of these residues surrounding the cleavage site is currently being determined on one systematically varied substrate.
The cleavage of dynorphin A(1-11) (dynA) and dynorphin B(1-13) (dynB) was studied because both these substrates contain a potential mono-and paired-basic cleavage site. The results show that YAP3p had a preference for the paired-basic sites over the potential monobasic sites in both substrates, generating Leu-enkephalin-Arg and Leu-enkephalin-Arg-Arg from dynA and dynB, respectively (Figs. 8 and 9). It is interesting to note the presence of a basic residue in the P3Ј position (Arg in dynA and Lys in dynB) relative to the primary cleavage sites in both substrates. The preference for the paired basic residues exhibited by YAP3p for dynA and dynB demonstrates that a basic amino acid present downstream of the cleavage site (e.g. P3Ј) may also play a major role in determining the exact cleavage site of YAP3p. Cleavage studies of synthetic proalbumin peptides by YAP3p have verified the importance of downstream basic residues. 3 An almost 10-fold increase of relative cleavage activity of YAP3p over the wild type proalbumin sequence was shown for a peptide that substituted the Ala for an Arg in the P2Ј position. Since the aspartic proteases exhibit a high degree of symmetry in the active site, it is possible that a given substrate can bind in a manner such that the bond to be cleaved is regulated by basic residues in downstream as well as upstream positions.
The cleavage specificity of YAP3p for the substrates tested in this and previous studies (Table II) indicates that YAP3p recognizes the following motifs: a pair of basic residues or monobasic residues with an additional upstream basic residue within the P2-P6 position. However, cleavage of some monobasic sites without an upstream basic residue may be regulated by additional amino acids in the cleavage site. Cleavage at paired basic residues can occur either between or on the carboxyl side of the pair of basic residues, the preference being substrate-dependent and likely governed predominantly by the upstream and downstream basic residues surrounding the cleavage site. However, most of the substrates studied also contain nonpolar residues in the P3 position. The importance of such residues remains unclear, but the finding that a charged residue, Arg or Lys, as mentioned before, is tolerable in the P3 position suggests that S3 does not have a strict requirement for a given type of amino acid. Cleavage at a monobasic site occurs on the carboxyl side of the basic residue. The specificity of YAP3p shows some overlap with that of the mammalian proopiomelanocortin converting enzyme (3,31). Pro-opiomelanocortin converting enzyme has been shown to cleave at pairedbasic residue sites of POMC, proinsulin (22,32), and the monobasic residue motif at Lys 23 of CCK13-33. 4 However, pro-opiomelanocortin converting enzyme did not cleave ACTH   (22) or ␤-endorphin 1-31 (33).
Analysis of kinetic parameters (Fig. 2) for the cleavage of paired and monobasic residues shows that YAP3p cleaves prohormone substrates at these sites with k cat /K m values comparable to monkey cathepsin E for a variety of synthetic peptides mimicking the cleavage sites of some prohormone precursors (34). Cathepsin E may represent another member of the aspartic protease family of enzymes involved in intracellular precursor processing (35). A comparison of the catalytic efficiency (k cat /K m ) of cleavage of the motifs in this study suggests that YAP3p cleaves the motifs containing a paired-basic residue site, with or without a basic residue upstream or downstream, as in ACTH 1-39 , dynA, dynB, and amidorphin, 10 -100 times more efficiently than the monobasic residue motif as in CCK13-33. Moreover, having additional basic residues flanking the cleavage site between P2-P6 and P2Ј-P6Ј enhances the affinity of binding and catalytic efficiency. This is exemplified by the decrease in K m and increase in k cat /K m for ACTH  versus amidorphin, both of which are cleaved at a Lys-Lys pair, but ACTH 1-39 has 4 additional basic residues flanking the cleavage site. Tetrabasic residues may well be a highly efficient cleavage site for YAP3p. CCK13-33 and amidorphin were both cleaved on the carboxyl side of a Lys residue, but amidorphin was cleaved with a Ͼ10-fold higher efficiency than CCK13-33. This may be an indication of the preference by YAP3p for a basic residue in the P2 position, Lys in amidorphin, rather than P6, Arg in CCK13-33.
From the present studies it would appear that YAP3p can cleave all the motifs recognized by the subtilisin-like serine proteases, PC1/3, PC2, and furin. PC1/3 has been shown to cleave paired-basic residue sites (36 -38) and at a monobasic residue site with an upstream Arg in the P4 position (39), while PC2 cleaves only at paired-basic residues (15,19). Furin, on the other hand, prefers a paired-basic residue motif with an additional upstream Arg residue at the P4 position, although the basic residue at P2 appears not to be essential (12, 39 -42). Future kinetic studies on the cleavage of these substrates by PC1 and PC2 will be important in assessing whether these enzymes cleave the same substrates with different efficiencies and the role the structural conformation plays in dictating efficiency of cleavages, independent of the enzyme. Such kinetic information will also be important in determining the relative role the aspartic proteases and the subtilisin-like serine processing enzymes play in cleaving various prohormones, since both families of proteases can be found in the same endocrine cells, e.g. the presence of pro-opiomelanocortin converting enzyme, PC1, and PC2 in the pituitary intermediate lobe cells (22,43).