Synthesis and Characterization of the First Potent Inhibitor of Yapsin 1 IMPLICATIONS FOR THE STUDY OF YAPSIN-LIKE ENZYMES* □ S

The potent peptidic inhibitor, Y1, of the basic residue-specific yeast aspartyl protease, yapsin 1, was synthesized and characterized. The inhibitor was based on the peptide sequence of a cholecystokinin 13–33 analog that yapsin 1 cleaved with an efficiency of 5.2 (cid:1) M s (Olsen, C., Blundell, T. L., and Loh, Y. P. (1998) Biochemistry 37, 2768–2777). The apparent K i of Y1 for the inhibition of yapsin 1 was determined to be 64.5 n M , and the mechanism is competitive. Y2 was also developed as an analog of Y1 for coupling to agarose beads. The resulting inhib-itor-coupled agarose beads were successfully used to purify yapsin 1 to apparent homogeneity from conditioned medium of a yeast expression system. Utilization of this new reagent greatly facilitates the purification of yapsin 1 and should also enable the identification of new yapsin-like enzymes from mammalian and nonmamma-lian sources. In this regard, Y1 also efficiently inhibited Sap9p, a secreted aspartyl (cid:2) aliquots removed at and and fluo- was determined (excitation, A Lineweaver-Burk plot of the (1/ V o versus 1/[S]) generated a straight line from which the K m of 24.2 (cid:2) M and V max of 751 (cid:2) g/min was calculated. , yapsin 1 was pre-equilibrated for in M 4.0, with dilutions of Y1 . Enzymatic activity of yapsin 1 was then determined with the substrate. A Dixon plot of 1/ V o versus [I] generated a straight line from which K i of 64.5 n M was calculated using the derivation of the Lin- eweaver-Burk equation for competitive inhibition and the K m and V max parameters determined in the previous experiment. Both experiments described in A and B were performed three times.

Yeast aspartyl protease 3 was the third aspartyl protease to be cloned from Saccharomyces cerevisiae (1) after saccharopepsin and barrierpepsin. It was cloned based on its ability to partially suppress the pro-␣-mating factor processing defect in yeast mutants that lacked the normal processing enzyme, kex2p, a subtilisin-like serine protease (2). The ability to correctly process the pro-␣-mating factor demonstrated that this new aspartyl protease could cleave a prohormone at paired basic residue cleavage sites. This identified the new enzyme as a unique member of the aspartyl protease family of endoproteases whose specificity for basic residues was in direct contrast to all other aspartyl pro-teases whose classical specificity for hydrophobic residues was well documented (3). Since its discovery, yeast aspartyl protease 3 has been renamed yapsin 1 (4) to represent it as the first cloned member of this novel subclass of aspartyl proteases. This subclass currently includes five members, yapsin 1, yapsin 2 (5), and yapsin 3 (6) from S. cerevisiae, Sap9p from Candida albicans (7), and the previously described pro-opiomelanocortin converting enzyme (now named yapsin A) (8) from bovine pituitary secretory granules (9).
Since the discovery of yapsin 1, a large body of work has emerged describing its nature with respect to its physical and chemical properties and has been comprehensively reviewed in the Handbook of Proteolytic Enzymes (4). To further investigate the mechanism of action of the enzyme, we have attempted to crystallize yapsin 1 but have been unsuccessful. Although purification of yapsin 1 by classical procedures has been accomplished (10,11), the presence of excess N-linked glycosylation on yapsin 1 interfered with the crystallization process. In addition, the deglycosylated yapsin 1 was unstable. For more traditional aspartyl proteases, inhibitor columns are often used to facilitate purification, and inhibitors are used to stabilize the protease during crystallization. However, pepstatin A, an aspartyl protease-specific inhibitor, is the only compound reported to inhibit yapsin 1 and shows only modest inhibition (K i APP ϭ 0.4 M (12). We therefore undertook to synthesize a more potent inhibitor that could be used to purify yapsin 1 and subsequently stabilize the deglycosylated form of yapsin 1 during crystallization. The development of new, more potent inhibitors for members of the yapsin subclass of aspartyl proteases is also relevant for the treatment of human disease. In this respect, the discovery that one of the secreted aspartyl proteases, Sap9p, of the opportunistic yeast, C. albicans, is a member of the yapsin family (7) renders the use of a specific inhibitor of Sap9p potentially important in the control of its pathogenesis.

Synthesis of Inhibitor
Materials were obtained from commercial suppliers and employed without further purification unless otherwise stated. Dihydropyranyl resin was purchased from NovaBiochem (San Diego, CA). Anhydrous N,N-dimethylformamide (DMF) 1 and N-methylpyrrolidinone (NMP) were purchased from Aldrich. The following solvents were distilled * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. under N 2 from specified drying agents: tetrahydrofuran and diethyl ether from sodium/benzophenone ketyl and methylene chloride (CH 2 Cl 2 ) and dichloroethane and pyridine from calcium hydride. For the general solid phase work-up procedure, excess reagent and the reaction solution were filtered away from support-bound material using polypropylene cartridges with 70-PE frits (Speed Accessories) attached to Teflon stopcocks. Cartridges and stopcocks were purchased from Applied Separations (Allentown, PA). Immobilized diaminodipropylamine gel and the polystyrene columns (1-5-ml bed volumes) for affinity column chromatography were purchased from Pierce. For a detailed description of the chemical reactions and compound characterization see the supplemental data. Compound 1 shown in the scheme in Fig. 1 is referred to as the inhibitor Y1. Compound 2, an analog of Y1 used for coupling to agarose beads, is shown in the scheme in Fig. 2 and is referred to as the inhibitor Y2.

K i Determination of Inhibitor Y1 for Yapsin 1
We utilized an internally quenched fluorescent substrate, Abz-Arg-Val-Lys-Arg-Gly-Leu-Ala-Tyr(NO 2 )-Asp-OH (Bachem, Torrance, CA) to determine the K i of Y1 for yapsin 1. This substrate was previously developed (13) as a substrate for furin, a subtilisin-like proprotein convertase (14); however, the Lys-Arg motif within this substrate is a well characterized yapsin 1 cleavage site (15). We first determined the K m and V max of yapsin 1 for this substrate as follows. Yapsin 1 was purified from a yeast expression system as described previously (11). Purified yapsin 1 (ϳ7 nM; ϳ37 ng) was incubated with 5-40 M of the substrate at 37°C in 100 l of 0.1 M ammonium acetate, pH 4.0. 15-l aliquots were removed at 5, 10, 20, 30, 40, and 60 min and added to 1 ml of 50 mM Tris-HCl, pH 7.5, to stop the enzymatic reaction. The samples were kept on ice until assayed for fluorescence in a Perkin-Elmer LS-5 luminescence spectrometer (emission, 320 nm; excitation, 425 nm). A nonlinear regression fit to each time course allowed the calculation of the initial rate of enzymatic activity at each substrate concentration and a Lineweaver-Burk plot of 1/V o versus 1/[S] was generated to determine K m and V max . These experiments were performed three times.
Preliminary experiments that included "recovery of activity" (16) and Lineweaver-Burk plots (data not shown) indicated that Y1 was a reversible competitive inhibitor of yapsin 1. To determine the K i , purified yapsin 1 (ϳ7 nM) was incubated with dilutions of Y1 (ranging from 250 to 15.6 nM) in 98 l 0.1 M ammonium acetate, pH 4.0, for 20 min at room temperature. 2 l of 0.5 mM substrate (final substrate concentration, 10 M) were then added, and the reactions were allowed to continue for 10 min at 37°C. 15 l were then assayed for fluorescence as described above. This experiment was performed three times. A Dixon plot of 1/V o versus [I] was generated from which K i was calculated using the derivation of the Lineweaver-Burk equation for competitive inhibition and the K m and V max parameters determined in the previous experiment.

Preparation of Affinity Column
Preparation of Inhibitor Y2 for Attachment to Affinity Column-Inhibitor Y2 was prepared from support-bound tosylate 11 ( Fig. 1) (100 mg, 48.8 mol, 0.488 mmol/g) according to the solid phase synthesis methods to prepare inhibitor Y1 up through coupling Fmoc-Val (see the supplemental data). At this stage Fmoc-Arg (2,2,5,7,8-pentamethylchroman-6-sulfonyl) was coupled using 0.3 M Fmoc-Arg (2,2,5,7,8-pentamethylchroman-6-sulfonyl), 0.3 M PyBOP, 0.3 M HOAt, and 0.9 M iPr 2 EtN in NMP with shaking overnight. The resin was washed with NMP (four times), tetrahydrofuran (twice), CH 2 Cl 2 (three times), and ether (three times), and then the resin was treated with 20% piperidine in NMP solution for 30 min. The resin was then treated with 0.3 M 4-carboxybenzaldehyde, 0.3 M PyBOP, 0.3 M HOAt, and 0.9 M iPr 2 EtN in NMP with shaking overnight. Inhibitor Y2 was cleaved from the solid support using 95:5 trifluoroacetic acid/H 2 O for 30 min followed by rinses with 95:5 trifluoroacetic acid/H 2 O (once) and CH 2 Cl 2 (twice) and then concentration under reduced pressure. The mixture was purified by reverse phase HPLC as described for inhibitor Y1 (see the supplemental data), and the mass was confirmed by mass spectrometry. HPLC purification provided 17.6 mg (16.1 mol) of inhibitor Y2 (Fig. 2  Mixing was then carried out by rotating the suspension at room temperature for 12 h. The gel was then washed with DMF (three times). The loading level of the gel was determined by Fmoc quantitation (gel total, 33.4 mol, 0.0309 mequiv/g, 84% loading efficiency based on a 20 equiv/ml loading level of diaminodipropylamine gel). The gel was treated with 20% piperidine in NMP for 3 h and then was washed with DMF (three times). To the resulting gel 16 (Fig. 2) was added inhibitor Y2 (

Purification of Yapsin 1 by Affinity Chromatography
The inhibitor-coupled agarose beads (Fig. 2, compound 3) were equilibrated with equilibration buffer (0.1 M ammonium acetate, pH 4.0). A C-terminally truncated form of yapsin 1, lacking its GPI membrane anchor, described previously (17), was overexpressed and secreted into the culture medium from a yeast strain engineered to express this form of the enzyme. The culture supernatant was adjusted to pH 7.0 with 1 M Tris/Cl and concentrated by batch application to DEAE-Sepharose beads. After washing, the bound protein was eluted with 0.2 M NaCl and aliquoted. A 0.9-ml aliquot of the concentrated medium enriched in yapsin 1 was adjusted to 0.1 M ammonium acetate, pH 4.0, and applied to the inhibitor column. After collecting the flow through, the column was washed with 6 ϫ 0.9 ml of equilibration buffer followed by 6 ϫ 0.9 ml of equilibration buffer containing 0.5 M NaCl. The yapsin 1 was eluted with 9 ϫ 0.9 ml of 20 mM Tris/Cl, pH 8.5, containing 2 M LiBr.
Yapsin 1 immunoreactivity was assayed by Western blot analysis. 10 l of each sample were fractionated by denaturing polyacrylamide gel electrophoresis on 4 -12% NuPAGE gels under reducing conditions using MES as running buffer (Invitrogen). After transferring the protein to nitrocellulose, the blots were probed with yapsin 1 antiserum MW283 (17) at 1:10,000, and the signal was visualized by enhanced chemiluminescence. Yapsin 1 enzymatic activity was assayed by incubating 1 l of each fraction from the purification procedure with 25 g of rat proenkephalin (pro-Enk) (5 mg/ml), purified as described previously (18), in 0.1 M ammonium acetate, pH 4.0, for 12 h. The reaction volume was 100 l. Yapsin 1 cleaves rat pro-Enk efficiently to release its N terminus, syn-enkephalin-Lys (pro-Enk 1-74 ). 2 20 l of each reaction was analyzed by colloidal Coomassie Blue staining after denaturing polyacrylamide gel electrophoresis on 4 -12% NuPAGE gels under reducing conditions using MES as running buffer.

Analysis of Affinity-purified Yapsin 1
Fractions containing pro-Enk cleaving activity and yapsin 1 immunoreactivity were pooled and concentrated by centrifugation through 30-kDa Filtron Nanosep membranes. An aliquot of the final sample (ϳ200 ng) was deglycosylated by endoglycosidase H as described previously (17) and analyzed by Western blot analysis and silver staining. 10 g of the starting material was also analyzed by Coomassie Blue staining. Specific activity measurements were carried out on the starting material and on the final purified yapsin 1 using the quenched fluorescent substrate. Briefly, yapsin 1 enzymatic activity was assayed by incubating 2 l from both samples with 10 M of the quenched fluorescent substrate in 100 l of 0.1 M ammonium acetate, pH 4.0, for 20min at 37°C. 900 l of 50 mM Tris-HCl, pH 7.5, were added to the reaction, and the fluorescence was determined as described above. The units of fluorescence were then corrected to units/g/min.

Inhibition of Sap9p Cleavage of Proenkephalin by Inhibitor Y1
A C-terminally truncated form of secreted aspartyl protease 9 (Sap9p), lacking its GPI membrane anchor, was expressed in a Pichia pastoris expression system. Sap9p was expressed and secreted into the culture medium after induction, and the culture supernatant was concentrated by centrifugation filtration through 30-kDa Filtron Macrosep membranes. The presence of Sap9p in the induced preparation was confirmed by Coomassie Blue staining of an induced protein of the expected size after SDS-PAGE and by the ability of the induced preparation alone to cleave adrenocorticotropin hormone  with identical specificity to that of yapsin 1. 2 To test the inhibitor against Sap9p, we used purified rat pro-Enk as a substrate. Aliquots of the conditioned medium containing Sap9p were preincubated with Y1 (1-100 M) in 0.1 M ammonium acetate, pH 4.0. for 20 min at room temperature, after which 25 g of purified rat pro-Enk was added, and the reaction continued for 12 h at 37°C. The reaction volume was 100 l. The controls included omission of enzyme, omission of inhibitor, or addition of 2.5 mM pepstatin A, a specific inhibitor of aspartyl proteases. After incubation, 20 l of each reaction was analyzed by colloidal Coomassie Blue staining after denaturing polyacrylamide gel electrophoresis on 4 -12% NuPAGE gels under reducing conditions using MES as running buffer (Invitrogen). Compound 2, an analog of the inhibitor Y1 described in Fig. 1, was coupled to agarose beads according to the scheme outlined here and described under "Materials and Methods" to generate the affinity matrix. a, DMF, pH 4.7 buffer; b, piperidine, NMP; c, DMF pH 4.7 buffer.
Purification of Sap9p by Affinity Chromatography 10 ml of conditioned medium containing Sap9p was adjusted to pH 4.0 with 1 M ammonium acetate, pH 4.0. The sample was centrifuged at 13,000 ϫ g for 5 min to remove particulate matter, and the supernatant was applied to the inhibitor column (Fig. 2, compound 3). The flow through fraction was reapplied twice before continuing with the purification procedure exactly as described for the purification of yapsin 1. 1-l aliquots from each fraction were assayed for pro-Enk processing activity. Fractions E2 and E3 were pooled and concentrated by centrifugation filtration through 30-kDa Nanosep filters, and an aliquot of the final sample was analyzed by Coomassie Blue. In addition, 5 g of purified Sap9p were run on a denaturing 4 -12% NuPAGE gel, transferred to polyvinylidene difluoride, and subjected to Edman degradation using an Applied Biosystems model 470A protein sequencer with an on-line phenylthiohydantoin analyzer.

RESULTS
Synthesis of Inhibitor-The inhibitor Y1 (Fig. 1, compound 1) was identified upon the preparation and screening of a focused library of 45 compounds based upon a known yapsin 1 substrate, RVSMIKNR. The key design feature was to replace the Lys-Asn dipeptide in the peptide substrate with the stable hydroxyethylamine peptide isostere incorporating a lysine side chain at the P1 position. The hydroxyethylamine isostere has been extensively used for the inhibition of the HIV-1 protease with several marketed AIDS drugs based upon this template. However, these isosteres incorporate hydrophobic side chains at the P1 position, and to our knowledge, the incorporation of basic side chains at the P1 position has not previously been reported.
The preparation of Y1 is shown in Fig. 1 as compound 1. Asymmetric Sharpless epoxidation of allylic alcohol 4 (19) provides epoxy alcohol 5 in 73% yield with Ͼ95% enantiomeric purity as measured by 1 H NMR of the corresponding Mosher ester. Benzoyloxy diol 6 was then prepared by treating epoxy alcohol 5 with (i-PrO) 2 Ti(OBz) 2 . Acetalization with 2,2-dimethoxy propane and p-toluenesulfonic acid was followed by exchange of the phthalimide group for the Boc group under standard conditions to provide alcohol 7. Treatment of alcohol 7 with toluenesulfonyl chloride in pyridine provided the tosylate 8. Azide displacement was followed by removal of the isopropylidene group to provide diol 9. Selective tosylation of the primary alcohol with toluenesulfonyl choride in pyridine provided intermediate 10 for loading onto the solid support. Intermediate 10 was then attached to the support using commercially available dihydropyranyl substituted resin (20) to provide support-bound inhibitor 11. The synthesis of Y1 was then performed according to previously developed solid phase synthesis methods (21). Displacement of the primary tosylate with butylamine was followed by acylation of the amine with ␣-N-Boc-Arg to provide  , y ϭ 0.0014x ϩ 0.1274, r 2 ϭ 0.99 (Fig. 3B). From the derivation of the Dixon plot, competitive inhibition predicts that y ϭ 1/V max (1 ϩ K m /[S]) ϭ 0.123, when x ϭ 0. Our data therefore fits well with that of a competitive inhibitor because y ϭ 0.1274 when x ϭ 0. Also from the equation for competitive inhibition, when y ϭ 0, x ϭ ϪK i (1 ϩ [S]/K m ), from which K i calculates as 64.5 nM.
Preparation of the Inhibitor-coupled Agarose Beads-To prepare the affinity column for yapsin purification, the diaminodipropylamine (diaminodipropylamine) form of Sephadex was coupled with active ester 15. Removal of the Fmoc group then provides the alkoxylamine derivatized agarose beads 16, which was then covalently linked to inhibitor Y2 by oxime bond formation (Fig. 2).
Purification of Yapsin 1 by Affinity Chromatography-Yapsin 1 was expressed and secreted into the culture medium of a yeast expression system, which has been described previously (11,17). After concentration of the medium by anion exchange chromatography using DEAE-Sepharose beads, the sample was applied to the inhibitor column. All of the yapsin 1 that was present in the starting material (Fig. 4A, lane S) bound to the column as evidenced by the absence of any immunoreactive yapsin 1 in the flow through fraction (Fig. 4A, lane FT) or any of the wash fractions. The yapsin 1 remained bound even during the stringent second wash containing 0.5 M NaCl (Fig. 4A,  2nd wash). The yapsin 1 was eluted from the inhibitor column by changing the pH from 4.0 to 8.5 and by including 2 M LiBr in the eluant (Fig. 4A, elution 1-9). Consistent with the elution profile of immunoreactive yapsin 1, yapsin 1 enzymatic activity was demonstrated in the eluate as seen by the generation of syn-Enk from pro-Enk (Fig. 4B, elution 1-9). Note that in the absence of immunoreactive yapsin 1 in the flow through and first three wash fractions, pro-Enk was presumably degraded by other proteases present in the culture supernatant that did not bind to the column.
Analysis of Affinity-purified Yapsin 1-Fractions 2-9 of the eluate were combined and concentrated. By Western blot analysis, the purified yapsin 1 gave two bands at ϳ90 and 120 -150 kDa, corresponding to glycosylated yapsin 1 (17) (Fig. 5B, lane Ϫ). Upon deglycosylation by endoglycosidase H treatment, a single immunoreactive anti-yapsin 1 band is seen by Western blot (Fig. 5B, lane ϩ), and only one band is apparent by silver stain analysis (Fig. 5C, lane ϩ), indicating that the purity of the enzyme approached homogeneity. An aliquot of the starting material (ϳ10 g) was analyzed by Coomassie Blue staining in which yapsin 1 can be seen in addition to other proteins (Fig.  5A, lane S). Using the quenched fluorescent substrate, specific activity assays for the starting material and the purified yapsin 1 yielded 95 and 752 units/g/min, respectively, indicating a fold purification of ϳ8.
Inhibition of Sap9p Cleavage of Proenkephalin by Y1-Sap9p was expressed and secreted into the culture medium of a P. pastoris expression system. Coomassie Blue staining (not shown) confirmed the presence of Sap9p in the induced medium and its absence in the uninduced medium, whereas activity assays using adrenocorticotropin hormone 1-39 as a substrate confirmed that Sap9p can cleave at basic residue cleavage sites. 2 Sap9p was shown here to process the prohormone, pro-Enk, to generate syn-Enk (Fig. 6, lane 1). This activity was inhibited by pepstatin A (Fig. 6, lane 5) as well as by Y1 in a dose-dependent manner. At 100 M of Y1, almost complete inhibition of the Sap9p activity was observed (Fig. 6,  lane 2), whereas at 10 M, ϳ50% inhibition was observed (Fig.   6, lane 3), and at 1 M almost no inhibition was observed (Fig.  6, lane 4).
Purification of Sap9p by Affinity Chromatography-Sap9p was purified from the conditioned medium of a P. pastoris expression system. The enzyme specifically eluted from the column in fractions E2 and E3, as measured by the disappearance of pro-Enk substrate and the generation of syn-Enk product in the assay of these fractions (Fig. 7). The smaller band of syn-Enk observed in E2 is consistent with syn- Enk (1-74), i.e. syn-Enk-Lys, and the slightly bigger product band in E3 is consistent with a Met-enkephalin extended form of syn-Enk, syn- Enk (1-82), i.e. syn-Enk-Lys-Lys-Tyr-Gly-Gly-Phe-Met-Lys-Arg 82 . Products with similar mobility on SDS-PAGE were generated by purified yapsin 1 and identified by mass spectros- were chemically coupled to an agarose support resin, described in the legend to Fig. 2, and the resulting ϳ1-ml affinity matrix was used for the purification of yapsin 1 by affinity chromatography. Yapsin 1 was expressed and secreted from a yeast expression vector described previously (17), and a concentrated preparation of the conditioned culture medium was used as the starting material. Aliquots from each sample were analyzed by Western blot for yapsin 1 immunoreactivity, and the enzymatic activity of yapsin 1 was assayed using purified rat pro-Enk as substrate. Yapsin 1 cleaves pro-Enk at basic residue cleavage sites and generates syn-Enk-Lys (pro-Enk 1-74 ). The doublet of pro-Enk is due to differential glycosylation. Note that yapsin 1 immunoreactivity is only observed in the eluate (E2-E9) and that syn-Enk-Lys is efficiently generated by yapsin 1 in these fractions. copy as syn-Enk (1-74) and syn-Enk (1-82). 2 Coomassie stain analysis of the concentrated eluted Sap9p showed two major bands (Fig. 8): one band at an apparent molecular mass of ϳ60 kDa and one band at an apparent molecular mass of ϳ5 kDa. N-terminal amino acid sequence analysis of the upper band gave the following sequence beginning with Leu 154 , L-F-G-F-G-X-I-Y. The lower band gave the following sequence starting with Asp 51 , D-G-S-L-D-M-T-L-T-N-K-Q-T-F-Y. This demonstrated that Sap9p exists as a heterodimer and confirms the predicted pro-region processing site as Lys 49 -Arg 50 . The site where cleavage occurs to generate the two subunits in Sap9p is equivalent to the position where yapsin 1 is cleaved to generate its two subunits. However, because of the apparent size of the smaller band (ϳ5 kDa), it is likely that additional processing has occurred on its carboxyl end to generate a protein of this size instead of one of the expected size, ϳ10 kDa (amino acids 51-153). We propose to call the Sap9p subunits ␣ and ␤, in keeping with the similar nomenclature used for yapsin 1.

DISCUSSION
Yapsin 1 is the model enzyme for the new subclass of aspartyl proteases that cleave proproteins at specific basic residue cleavage sites. Extensive analyses of its specificity (15,(22)(23)(24) have revealed that substrates containing multiple basic residues surrounding the scissile bond are cleaved with high efficiencies. The nature of this specificity is not completely understood even though molecular modeling (Protein Data Bank code 1YPS) (24) has revealed that yapsin 1 has a more open active site compared with other aspartyl proteases and that many of its subsites are electronegative, thus allowing for the accom-modation of positively charged residues within and surrounding the cleavage site (24). We have attempted to crystallize yapsin 1 to obtain x-ray crystal structure data but have been unsuccessful. The main problem resided in the high degree of glycosylation added to yapsin 1 in the yeast, ranging from 50 to 300% of its calculated molecular mass. Removal of this excess sugar can be achieved readily by endoglycosidase H or Nglycanase treatment; however, under these conditions, the deglycosylated yapsin 1 is not sufficiently stable for crystals to form.
To solve this problem we have synthesized an inhibitor of yapsin 1 that could be used in co-crystallization studies with deglycosylated yapsin 1. The design of the inhibitor Y1 (Fig. 1,  compound 1) was based on the sequence of a cholecystokinin 13-33 analog that yapsin 1 was shown to cleave with relatively high efficiency (5.2 ϫ 10 5 M Ϫ1 s Ϫ1 ) (24). The sequence of this peptide surrounding the scissile bond was RVSMIKNR, where cleavage occurred on the carboxyl side of the single lysine residue. Evaluation of peptide analogs demonstrated that for good cleavage efficiency, basic residues should be present at the P1, P6, and especially P2Ј positions (24).
FIG. 6. Inhibition of Sap9p by the new inhibitor. Sap9p was expressed in a P. pastoris expression system (described elsewhere (34)). The conditioned media containing Sap9p was pre-equilibrated with various concentrations of Y1 (0 -100 M) for 10 min at room temperature. 25 g of purified pro-Enk was added, and the samples were incubated for 12 h at 37°C. The reactions were analyzed by Coomassie Blue staining after SDS-PAGE. Note the disappearance of pro-Enk and the appearance of syn-Enk in the absence of Y1 (lane 1), indicating a yapsin 1-like processing profile. 2.5 mM pepstatin A completely inhibited the activity demonstrating that the proteolytic activity was due to the aspartyl protease Sap9p (lane 5). Sap9p was inhibited in a dose-dependent manner by the Y1 (lanes 2-4). B, no enzyme added.
FIG. 7. Purification of Sap9p from conditioned media by affinity chromatography. Sap9p expressed and secreted from a P. pastoris expression system was used as the starting material for purification by the inhibitor column. Aliquots from each sample were analyzed for enzymatic activity using purified rat proenkephalin as substrate. The reactions were run on denaturing/reducing gels for analysis by Coomassie Blue staining. Sap9p was eluted primarily in E2 and E3 as evidenced by the absence of pro-Enk and generation of syn-Enk-Lys in E2 and an extended form of syn-Enk in E3. B, no enzyme; S, starting material; FT, flow through.  Fig. 7 were pooled and concentrated, and an aliquot was analyzed by Coomassie Blue under denaturing/reducing conditions. Two bands stained strongly, both of which were identified as Sap9p related by N-terminal amino acid sequencing. The upper band described here as the ␤-subunit started with Leu 154 of Sap9p and has an apparent molecular mass of 62 kDa. The lower band described as the ␣-subunit started with Asp 51 of Sap9p and has an apparent molecular mass of 5 kDa. Therefore, the key inhibitor design feature was to replace the scissile Lys-Asn dipeptide in the peptide substrate with the stable hydroxyethylamine peptide isostere that incorporated a lysine side chain at the P1 position (Fig. 1). Notably, the hydroxyethylamine isostere serves as a very effective pharmacophore for inhibiting aspartyl proteases as demonstrated by several approved HIV-1 protease inhibitors that incorporate this isostere. To further take advantage of the known octamer substrate structure activity relationships, a small collection of potential inhibitors, was prepared by parallel synthesis with basic functionality introduced at the P2Ј position as introduced by the R1 and R2 diversity elements. Screening the compound collection for inhibition of yapsin 1 resulted in the identification of inhibitor Y1 (Fig. 1).
Using the quenched fluorescent substrate, Y1 was shown to inhibit purified yapsin 1 with an apparent K i of 64.5 nM (Fig. 3), and its mechanism of inhibition is characteristic of competitive inhibition. Its inhibitory property was maintained after coupling to agarose beads as seen by the ability of the inhibitor column to bind and elute yapsin 1, demonstrating that enzyme-inhibitor binding was reversible (Figs. 4 and 5). We estimate that with 10 mol of inhibitor coupled to the beads, its capacity, even at 10% (i.e. 1mole availability), would be sufficient to purify Ͼ10 mg of yapsin 1. As expected from conditioned medium that is enriched in yapsin 1 (Fig. 5A, lane S), the purification procedure resulted in only an ϳ8-fold increase in specific activity as measured by the quenched fluorescent substrate.
In the yeast S. cerevisiae, there are three members of the yapsin family that not only share sequence identity but also cleavage specificity. However, whereas each enzyme can cleave at basic residues, they exhibit some differences. For example, yapsin 1 cleaves adrenocorticotropin hormone 1-39 at its tetra-basic residue cleavage site extremely well, whereas yapsin 3 does not appear to cleave it at all (6). It is because of this selective specificity that a broad range of heterologously expressed proteins in S. cerevisiae are cleaved in a limited way by these enzymes (25,26). To circumvent this problem, researchers have utilized yeast mutants that lack either yapsin 1 (27)(28)(29) or both yapsin 1 and 2 (25,30) to express their proteins. However, although recovery of intact protein increases in this double mutant, the overall yield may be limited because this mutant exhibits impaired growth at 37°C (5). Because the yapsins are GPI membrane-anchored proteases located on the extracellular side of the cell membrane (5,17,31), addition of specific and potent yapsin inhibitors to the culture medium may be an alternative approach to minimize proteolysis while maximizing the expression potential from wild type yeast.
This new inhibitor described here also inhibited Sap9p (Fig.  6), a secreted aspartyl protease from C. albicans. In addition, Sap9p was purified from conditioned medium by affinity chromatography using the inhibitor column ( Figs. 7 and 8). The ability to specifically inhibit Sap9p may prove worthwhile in the prevention or impairment of the virulence of C. albicans, an opportunistic yeast prevalent in immuno-compromised patients. This is because all the SAP enzymes described so far have a paired basic cleavage site in their pro-region (7) that is predicted to be cleaved by an appropriate enzyme to generate the mature SAPs that are subsequently involved in virulence (32). Although this paired basic amino acid motif is described as a Kex2p cleavage site, it is not known with certainty what enzyme actually performs this function in vivo. In Kex2p-null mutants of C. albicans, the activation of Sap2p still occurs, albeit at reduced levels (33), demonstrating that another enzyme is involved in this activation, spec-ulated previously to be Sap9p (7). If Sap9p, which exhibits specificity for this type of motif, is involved in the maturation of some or all of the SAPs, then inhibition of Sap9p may represent a pivotal step in the prevention of virulence of this pathogen.
Our sequencing results of the purified Sap9p showed a distinct similarity to yapsin 1 with respect to its processing into two subunits. Processing of Sap9p into its two subunits occurs after a single Lys residue, Lys 153 , that aligns almost exactly with the subunit cleavage site of yapsin 1 and occurs within a loop insertion which, when compared with other aspartyl proteases, seems to be unique to yapsin 1, yapsin 2, and Sap9p. The mature N-terminal amino acid of the ␣ subunit was determined to be Asp 51 , confirming the prediction that the proregion cleavage site occurred after the pair of basic residues, Lys 49 -Arg 50 . It is not known whether a pseudo-Sap9p is generated during activation similar to yapsin 1 (11).
We have provided here the first report of a potent inhibitor of the novel class of basic residue-specific aspartyl proteases. The inhibitor Y1 should serve as a powerful chemical tool for the study of the increasing number of yapsin-like enzymes in the same way that pepstatin has been essential for the study of more traditional aspartyl proteases. In particular, we plan to obtain co-crystals of Y1 with yapsin 1 to enable x-ray structure determination. In addition, inhibitor-coupled agarose beads are effective for rapid purification of yapsin-like enzymes and may prove useful for the identification and purification of additional yapsin-like enzymes from other eukaryotic species. Finally, further optimization of Y1 could provide useful compounds for the inhibition of potentially critical proteases involved in pathogenic mechanisms.