Presenilin-1 and -2 are molecular targets for gamma-secretase inhibitors.

Presenilins are integral membrane protein involved in the production of amyloid beta-protein. Mutations of the presenilin-1 and -2 gene are associated with familial Alzheimer's disease and are thought to alter gamma-secretase cleavage of the beta-amyloid precursor protein, leading to increased production of longer and more amyloidogenic forms of A beta, the 4-kDa beta-peptide. Here, we show that radiolabeled gamma-secretase inhibitors bind to mammalian cell membranes, and a benzophenone analog specifically photocross-links three major membrane polypeptides. A positive correlation is observed among these compounds for inhibition of cellular A beta formation, inhibition of membrane binding and cross-linking. Immunological techniques establish N- and C-terminal fragments of presenilin-1 as specifically cross-linked polypeptides. Furthermore, binding of gamma-secretase inhibitors to embryonic membranes derived from presenilin-1 knockout embryos is reduced in a gene dose-dependent manner. In addition, C-terminal fragments of presenilin-2 are specifically cross-linked. Taken together, these results indicate that potent and selective gamma-secretase inhibitors block A beta formation by binding to presenilin-1 and -2.

A combinatorial approach to the identification of dipeptide aldehyde inhibitors of beta-amyloid production.

In an effort to rapidly identify potent inhibitors of Abeta production and to probe the amino acid sequence specificity of the protease(s) responsible for the production of this peptide, a large number of dipeptide aldehydes were combinatorially synthesized and manually evaluated for their inhibitory properties. The starting point for this study was the dipeptide aldehyde carbobenzoxyl-valinyl-phenylalanal previously shown to inhibit the production of Abeta in CHO cells stably transfected with the cDNA encoding betaAPP695. Pools of related dipeptide aldehydes were combinatorially synthesized, and the most active pool was deconvoluted, resulting in the identification of the most active inhibitor of this pool. Systematic optimization of this inhibitor resulted in a series of dipeptide aldehydes with enhanced potencies relative to carbobenzoxyl-valinyl-phenylalanal. The most active dipeptide aldehydes were those that possessed hydrophobic amino acids at both the P1 and P2 positions. The most potent compound identified in this study was 3, 5-dimethoxycinnamamide-isoleucinyl-leucinal with an IC(50) of 9.6 microM, approximately 10-fold more active than carbobenzoxyl-valinyl-phenylalanal. In immunoprecipitation experiments using antibodies directed toward either Abeta1-40 or Abeta1-42, 3,5-dimethoxycinnamamide-isoleucinyl-leucinal, like carbobenzoxyl-valinyl-phenylalanal, preferentially inhibited the shorter 1-40 form of Abeta, whereas the longer 1-42 form was not as strongly inhibited. These results suggest that dipeptide aldehydes related to carbobenzoxyl-valinyl-phenylalanal inhibit Abeta through similar mechanisms and demonstrate the utility of a combinatorial synthesis approach to rapidly identify potent inhibitors of Abeta production.

Engineered metalloregulation in enzymes.

Protein engineering of metal-dependent enzyme activity is now possible due to the wealth of information available about metalloproteins. The results emerging from these studies provide insight into our understanding of the chemistry of metals in macromolecular environments as well as the biology of metal-protein interactions.

Crystallographic analysis of trypsin-G226A. A specificity pocket mutant of rat trypsin with altered binding and catalysis.

The crystal structure of trypsin-G226A has been determined, in the presence of benzamidine, to a resolution of 1.75 A with an R-factor of 14.6%. The mutation was designed to alter substrate specificity by disrupting arginine binding, but was previously found to disrupt catalysis to a greater extent than binding. The arginine analog, benzamidine, has rotated 40 degrees and 49 degrees and translated 1.1 A in the specificity pocket, relative to the position in wild-type trypsin. The salt-bridge between the amidinium group of benzamidine and the carboxylate of D189 as well as four other hydrogen bonds have been replaced by a set of six new hydrogen bonds. Based on these interactions, computer modeling of an arginine substrate demonstrates that arginine terminal nitrogen atoms can occupy the new benzamidine nitrogen positions with torsion angle adjustments and without short contacts. In the secondary orientation, arginine substrates appear to be forced out of alignment with the active site. This may account for the larger drop in kcat with arginine relative to lysine substrates. A second possible cause of the altered activity is a change of the enzyme structure with concomitant loss of activity. No evidence of such a change is seen in the co-ordinates or temperature factors of the trypsin-G226A-benzamidine complex. A226 disrupts mainly the co-ordinates of amino acids with which it has direct contacts such that the effects of the mutation are absorbed locally.

Crystal structure of rat trypsin-S195C at -150 degrees C. Analysis of low activity of recombinant and semisynthetic thiol proteases.

The X-ray crystal structure of trypsin-S195C, a rat anionic trypsin mutant in which the active site serine has been replaced by cysteine, was determined at -150 degrees C and room temperature to 1.6 A resolution, R = 15.4% and 1.8 A resolution, R = 15.0%, respectively. Cryo-crystallography was employed to improve the quality of the diffraction data and the resulting structure by eliminating radiation damage and decreasing atomic thermal motion. The average temperature factor decreased by 10 A2 relative to that of the room temperature structure. No radiation-induced decay of the data was detected. The side-chains of the catalytic cysteine and histidine of trypsin-S195C are found with 25% occupancy in secondary orientations rotated 104 degrees and 90 degrees out of the active site, respectively. These alterations, as well as more subtle changes in the active site may be caused by the oxidation of the catalytic sulfur to sulfenic acid. The position of the carbonyl carbon of the tetrahedral intermediate analog, p-amidinophenylpyruvic acid, modeled into trypsin-S195C, is 1.1 A from the catalytic sulfur. The large size and altered approach of the catalytic sulfur to substrates could account for the observed low catalytic activity relative to wild-type trypsin. In addition to the benzamidine in the specificity pocket, two additional binding sites for benzamidine are characterized. One of these mediates an intermolecular contact that appears to maintain the crystal lattice.

Regulation of serine protease activity by an engineered metal switch.

A recombinant trypsin was designed whose catalytic activity can be regulated by varying the concentration of Cu2+ in solution. Substitution of Arg-96 with a His in rat trypsin (trypsin R96H) places a new imidazole group on the surface of the enzyme near the essential active-site His-57. The unique spatial orientation of these His side chains results in the formation of a stable, metal-binding site that chelates divalent first-row transition-metal ions. Occupancy of this site by a metal ion prevents the imidazole group of His-57 from participating as a general base in catalysis. As a consequence, the primary effect of the transition metal ion is to inhibit the esterase and amidase activities of trypsin R96H. The apparent Ki for this inhibition is in the micromolar range for copper, nickel, and zinc, the tightest binding being to Cu2+ at 21 microM. Trypsin R96H activity can be fully restored by removing the bound Cu2+ ion with EDTA. Multiple cycles of inhibition by Cu2+ ions and reactivation by EDTA demonstrate that reversible regulatory control has been introduced into the enzyme. These results describe a novel mode of inhibition of serine protease activity that may also prove applicable to other proteins.

Introduction of a cysteine protease active site into trypsin.

Active site serine 195 of rat anionic trypsin was replaced with a cysteine by site-specific mutagenesis in order to determine if a thiol group could function as the catalytic nucleophile in serine protease active site environment. Two genetically modified rat thiol trypsins were generated; the first variant contained a single substitution of Ser195 with Cys (trypsin S195C) while the second variant contained the Ser195 to Cys as well as an Asp102 to Asn substitution (trypsin D102N,S195C) that more fully mimics the putative catalytic triad of papain. Both variants were expressed as his J signal peptide-trypsin fusion proteins to high levels under the control of the tac promoter. The mature forms of both variants were secreted into the periplasmic space of Escherichia coli. Trypsin S195C shows a low level of activity toward the activated ester substrate Z-Lys-pNP, while both trypsin S195C and trypsin D102N,S195C were active toward the fluorogenic tripeptide substrate Z-GPR-AMC. Esterase and peptidase activities of both thiol trypsin variants were inhibited by known Cys protease inhibitors as well as by specific trypsin inhibitors. The kcat of trypsin S195C was reduced by a factor of 6.4 x 10(5) relative to that of trypsin while the kcat of trypsin D102N,S195C was lowered by a factor of 3.4 x 10(7) with Z-GPR-AMC as substrate. Km values were unaffected. The loss of activity of trypsin D102N,S195C was partially attributed to an inappropriate Asn102-His57 interaction that precludes the formation of the catalytically competent His57-Cys195 ion pair although loss of the negative charge of D102 at the active site probably contributes to diminished activity.(ABSTRACT TRUNCATED AT 250 WORDS)

Crystal structures of two engineered thiol trypsins.

We have determined the three-dimensional structures of engineered rat trypsins which mimic the active sites of two classes of cysteine proteases. The catalytic serine was replaced with cysteine (S195C) to test the ability of sulfur to function as a nucleophile in a serine protease environment. This variant mimics the cysteine trypsin class of thiol proteases. An additional mutation of the active site aspartate to an asparagine (D102N) created the catalytic triad of the papain-type cysteine proteases. Rat trypsins S195C and D102N,S195C were solved to 2.5 and 2.0 A, respectively. The refined structures were analyzed to determine the structural basis for the 10(6)-fold loss of activity of trypsin S195C and the 10(8)-fold loss of activity of trypsin D102N,S195C, relative to rat trypsin. The active site thiols were found in a reduced state in contrast to the oxidized thiols found in previous thiol protease structures. These are the first reported structures of serine proteases with the catalytic centers of sulfhydryl proteases. Structure analysis revealed only subtle global changes in enzyme conformation. The substrate binding pocket is unaltered, and active site amino acid 102 forms hydrogen bonds to H57 and S214 as well as to the backbone amides of A56 and H57. In trypsin S195C, D102 is a hydrogen-bond acceptor for H57 which allows the other imidazole nitrogen to function as a base during catalysis. In trypsin D102N,S195C, the asparagine at position 102 is a hydrogen-bond donor to H57 which places a proton on the imidazole nitrogen proximal to the nucleophile. This tautomer of H57 is unable to act as a base in catalysis.(ABSTRACT TRUNCATED AT 250 WORDS)

An expression system for trypsin.

The eukaryotic serine protease, rat anionic trypsin, and various mutants created by site-directed mutagenesis have been heterologously expressed in Escherichia coli. The bacterial alkaline phosphatase (phoA) promoter was used to control the expression of the enzymes in an induced or constitutive fashion. The DNA coding for the eukaryotic signal peptide of pretrypsinogen was replaced with DNA coding for the phoA signal peptide. The phoA signal peptide successfully directs the secretion of the mammalian trypsinogen to the periplasmic space of E. coli. Active trypsin was expressed in the periplasm of E. coli by deleting the DNA coding for the activation hexapeptide of the zymogen. The activity of trypsin in the periplasm suggests that the enzyme is correctly activated and has folded such that the 12 cysteine residues involved in the six disulfide bonds of rat anionic trypsin have paired correctly. A transcription terminator increased the level of expression by a factor of two. However, increasing the copy number of the plasmid decreased the levels of expression. Localization of the active enzyme in the periplasm allows rapid screening of modified trypsin activities and facilitates the purification of protein to homogeneity and subsequently to crystallinity.

Detection of intermediate species in the refolding of bovine trypsinogen.

The mixed disulfide of bovine trypsinogen and glutathione was refolded at pH 8.6 and 4 degrees C with a mixture of 3 mM cysteine and 1 mM cystine catalyzing disulfide interchange. The folding process was monitored by analysis of quenched samples with isoelectric focusing and size-exclusion chromatography. Isoelectric focusing showed a progressive change from a pI of 5.2 for the mixed disulfide derivative to a pI of 9.3 for native trypsinogen. A number of principal intermediates were detected as a function of the refolding time. These intermediates were also separated and further characterized by size-exclusion chromatography on columns of TSK G2000 SW operated in the high-performance liquid chromatographic mode. Rechromatography of a series of sequential fractions taken from the parental peak was necessary to resolve and characterize the principal intermediates. The loss of glutathione moieties produced a partly folded structure with an apparent hydrodynamic volume (Stokes radius, Rs) of 33.9 A. These structures became compact with time, and more intermediates were detected between 33.9 and 29.2 A. Finally, a change in conformation, resembling a two-state transition, changed the molecules of Rs 29.2 to the compact structure of native trypsinogen (22.4 A). The rate of formation of the native structure was determined from the progress curves derived from isoelectric focusing and size-exclusion chromatography.(ABSTRACT TRUNCATED AT 250 WORDS)

Independent refolding of domains in the pancreatic serine proteinases.

Ser-neotrypsinogen and Val-neotrypsinogen are two-chain modifications of bovine trypsinogen produced on limited proteolysis with trypsin. Ser-neotrypsinogen has Lys131-Ser132 cleaved in the connecting peptide (the autolysis loop) linking the amino- and carboxyl-terminal domains. Val-neotrypsinogen has Arg105-Val106 cleaved which is located within the amino-terminal domain. The mixed disulfide derivative of Ser-neotrypsinogen was successfully refolded. A functional molecule was regenerated from the polypeptide fragments with the correct molecular weight of neotrypsinogen in an overall yield of 7%. Val-Neotrypsinogen could not be refolded. The first-order rate constants for the regeneration of Ser-neotrypsinogen were determined from the formation of active enzyme molecules as a function of time and from the regain of the correct molecular weight. Both kinetic values were the same indicating that refolding of the polypeptide chains first forms globular domain structures. The two domains then associate and the disulfide bonds between the domains and the correct geometry of the active site residues are formed last. The same kinetic results were also found in refolding Thr-neochymotrypsinogen (Duda, C. T., and Light, A. (1982) J. Biol. Chem. 257, 9866-9871) where peptide bond cleavage also occurred in the connecting peptide. These observations support the hypothesis that the pathway of folding of serine proteinases proceeds with the independent refolding of domains.

The identification of neotrypsinogens in samples of bovine trypsinogen.

Isoelectric focusing of commercial samples of bovine trypsinogen detected a component with a lower isoelectric pH than that of trypsinogen. The isoelectric pH was 8.75 compared to 9.3 for trypsinogen, and the amount of the component varied from 16 to 41% of the total protein. The protein (24,000 Da) was converted to fragments of 13,800 and 10,500 Da on reduction with dithioerythritol, showing that the component was a modified form of trypsinogen containing a cleaved peptide bond. The cleavage site was established from the study of four polypeptide fragments which were isolated from the fully reduced and S-carboxymethylated trypsinogen. The molecular weights, amino acid compositions, and amino-terminal sequences of these fragments identified a cleavage of Lys 131-Ser 132, namely from a Ser-neotrypsinogen, or at Arg 105-Val 106, from a Val-neotrypsinogen. Val-neotrypsinogen was the more abundant of the two and was approximately 71% of the total neotrypsinogen in the trypsinogen sample. Both neotrypsinogens were converted to active trypsin molecules in high yields, showing that the zymogens closely resembled the conformation of intact trypsinogen. Presumably, the neotrypsinogens were produced during the isolation of the zymogen when pancreatic tissue was partly autolyzed and active trypsin was present.