Superactivation of thermolysin by acylation with amino acid N-hydroxysuccinimide esters.

Synthesis of a series of active N-hydroxysuccinimide esters of aliphatic and aromatic amino acids has yielded a new class of reagents for the covalent modification of proteolytic enzymes such as thermolysin. The activities of aliphatic acyl amino acid thermolysins are from 1.7 to 3.6 times greater than that of the native enzyme when hydrolyzing durylacryloyl-Gly-Leu-NH2, the substrate employed most widely. By comparison, the aromatic acylamino acid derivatives are "superactive," their activities being as much as 70-fold greater. Apparently, the aromatic character of the amino acid introduced is a critical variable in the determination of the functional response. The increased activity is completely restored to that of the native enzyme by deacylation with nucleophiles, such as hydroxylamine, and the rate of restoration of native activity is a function of the particular acyl group incorporated. Preliminary evidence regarding the chemical properties of the modified enzyme suggests that tyrosine, rather than lysine, histidine, or arginine, may be the residue modified. The functional consequences of successive modification with different reagents, moreover, indicate that each of them reacts with the same protein residue. The competitive inhibitors beta-phenyl-propionyl-Phe and Zn-2+ do not prevent modification with these active esters. Hence, the site(s) of their inhibitory action differ(s) from that at which modification occurs. The structure of the substrate is also a significant variable which determines the rate at which each acyl amino acid thermolysin hydrolyzes peptides. Depending on the particular substrate, the activity of aromatic derivatives can be as much as 400-fold greater than that of the native enzyme, and the resultant activity patterns can be ordered in a series characteristic for each enzyme derivative.     

Chemical modification of D-amino acid oxidase. Amino acid sequence of the tryptic peptides containing tyrosine and lysine residues modified by fluorodinitrobenzene

D-Amino acid oxidase can be inactivated by covalent modification of predominantly tyrosine residue(s) at pH 7.4 by a low molar excess of fluorodinitrobenzene, which appears to act as an active site-directed reagent (Nishino, T., Massey, V., and Williams, C. H., Jr. (1980) J. Biol. Chem. 255, 3610-3616). Peptide mapping by high performance liquid chromatography of tryptic digests of protein modified with radiolabeled reagent revealed two major radioactive fractions with substantially different retention times which were not observed in protein modified in the presence of benzoate, a potent competitive inhibitor. Isolation and sequence analysis of the major radiolabeled peptides, as well as other direct chemical analyses, are presented which unambiguously demonstrate that these fractions represent modification of two different regions of the protein. The majority of the radiolabel was found within a 61-amino acid residue peptide containing an O-DNP-tyrosine residue exclusively at position 17. The substantial sequence surrounding this tyrosine residue indicates that it is different from that shown to react with N-chloro-D-leucine (Ronchi, S., Galliano, M., Minchiotti, L., Curti, B., Rudie, N. G., Porter, D. J. T., and Bright, H. J. (1980) J. Biol. Chem. 255, 6044-6046). The second fraction consisted of a 12-residue peptide containing an epsilon-DNP-lysine residue at position 5. Together, these two modified amino acids represented 0.89 mol of DNP incorporated/protein monomer. Both modifications must contribute to inactivation to account for the 90% decrease in enzymatic activity. Evidence is presented which suggests that both groups are within the active center of the enzyme and are modified in a mutually exclusive manner.     

The structure of tyrosine aminotransferase. Evidence for domains involved in catalysis and enzyme turnover

The primary structure of tyrosine aminotransferase, as deduced from the nucleotide sequence of complementary DNA, was confirmed by fast atom bombardment mass spectrometry of tryptic peptides derived from the purified protein. Limited digestion of the native enzyme with trypsin released an acetylated, amino-terminal peptide; the new amino terminus in the modified enzyme was Val65. Endogenous proteases generated a chromatographically separable form of tyrosine aminotransferase that began at Lys35. Neither trypsin nor the other proteases altered the catalytic activity of tyrosine aminotransferase. Reduction of the holoenzyme with sodium borohydride yielded a major tryptic peptide containing phosphopyridoxamine bound to lysine 280, which probably functions in transamination. The carboxyl terminus of tyrosine aminotransferase contains features that typify proteins with short half-lives; it includes two negatively charged, hydrophilic segments that are enriched for glutamyl residues and are similar to a PEST region in ornithine decarboxylase (Rogers, S., Wells, R., and Rechsteiner, M. (1986) Science 234, 364-368). Tyrosine aminotransferase belongs to a superfamily of enzymes which includes aspartate aminotransferase and can be aligned so that many invariant, functional residues coincide. Like the isoenzymes of aspartate aminotransferase, tyrosine aminotransferase may contain two domains, with a central, catalytic core, and a small domain made up of both amino- and carboxyl-terminal components. We speculate that the exposed small domain may confer the unusually rapid degradative rate that characterizes this enzyme.     

The use of [3H]aniline to identify the essential carboxyl group in the bovine mitochondrial F1-ATPase that reacts with 1-(ethoxycarbonyl)-2-ethoxy-1,2-dihydroquinoline

The inactivation of the bovine heart mitochondrial F1-ATPase with 1-(ethoxycarbonyl)-2-ethoxy-1,2-dihydroquinoline (EEDQ) in the presence of [3H]aniline at pH 7.0 led to the covalent incorporation of 3H into the enzyme. When the ATPase was inactivated by 94% with 0.9 mM EEDQ in the presence of 3.6 mM [3H]aniline in a large-scale experiment in which the protein concentration was 21 mg/ml, 4.2 mol [3H]anilide were formed per mol enzyme, of which 0.35 mol was incorporated per mol of the alpha subunit and 1.0 mol was incorporated per mol of the beta subunit. Examination of a tryptic digest of the isolated alpha subunit revealed that the majority of the 3H was contained in a single tryptic peptide, which, when purified, was shown to contain the [3H]anilide of a glutamic acid residue which corresponds to alpha-Glu-402 of the Escherichia coli F1-ATPase. This residue was labeled to the extent of about 1.0 mol/mol enzyme. Analysis of tryptic peptides purified from the isolated beta subunit showed that 0.8 and 1.5 mol, respectively, of the [3H]anilides of beta-Glu-341 and beta-Glu-199 were formed per mol MF1 during the inactivation of the enzyme at 21 mg/ml. When the ATPase was inactivated by 90% at a protein concentration of 1.7 mg/ml by 0.9 mM EEDQ in the presence of 1.7 mM [3H]aniline, 3.1 mol [3H]anilide were formed per mol enzyme. From the analysis of the radioactive peptides purified from a tryptic digest of the labeled ATPase from this experiment it was estimated that 0.7 mol of the [3H]anilide of alpha-Glu-402, 0.3 mol of the [3H]anilide of beta-Glu-341, and 1.5 mol of the [3H]anilide of beta-Glu-199 were formed per mol F1-ATPase. Since beta-Glu-199 is labeled to the same extent in the two experiments while alpha-Glu-402 and beta-Glu-341 were not, suggests that the modification of beta-Glu-199 is responsible for inactivation of the enzyme by EEDQ.     

Antigenic determinants of acylphosphatase from porcine skeletal muscle

Analysis of the quantitative precipitin reaction of acylphosphatase from porcine skeletal muscle with rabbit antiserum indicated the presence of at least two antigenic determinants on the porcine enzyme molecule. Immunological cross-reactivities of acylphosphatases from equine and rabbit skeletal muscles were examined. In double immunodiffusion with the antiserum, the precipitin lines of the porcine and equine enzymes completely fused, while the rabbit enzyme gave no precipitin line. The reaction between the 125I-labeled porcine enzyme and its antibody was inhibited to the same extent by the porcine and equine enzymes, but not by the rabbit enzyme. The three enzymes were similar in net charge and molecular weight on polyacrylamide gel electrophoreses. No conformational difference among the three enzymes was observed in their circular dichroism spectra. The amino acid composition of the rabbit enzyme differed from those of the porcine and equine enzymes in the contents of Glu, Gly, Lys, and Arg. Differences in the sequence of the rabbit enzyme from that of the porcine enzyme were investigated by comparison of the peptide maps of the tryptic peptides of the two enzymes. Four peptides of the rabbit enzyme were located at different positions from those of the porcine enzyme. Three of the four peptides from both enzymes were sequenced and all the tryptic peptides of both enzymes were characterized by amino acid analysis. The tryptic peptides of rabbit enzyme were tentatively aligned on the basis of their amino acid compositions and sequence homologies, compared with the corresponding peptides of the porcine enzyme. Among five amino acid residues of the porcine enzyme, Arg-4, Asp-28, Arg-31, Glu-56, and Ile-68, which are replaced in the rabbit enzyme, Arg-4 and Asp-28 are considered to be included in the antigenic determinants.     

Histidine and not tyrosine is required for the peroxide-induced formation of haem to protein cross-linked myoglobin

Peroxide-induced oxidative modifications of haem proteins such as myoglobin and haemoglobin can lead to the formation of a covalent bond between the haem and globin. These haem to protein cross-linked forms of myoglobin and haemoglobin are cytotoxic and have been identified in pathological conditions in vivo. An understanding of the mechanism of haem to protein cross-link formation could provide important information on the mechanisms of the oxidative processes that lead to pathological complications associated with the formation of these altered myoglobins and haemoglobins. We have re-examined the mechanism of the formation of haem to protein cross-link to test the previously reported hypothesis that the haem forms a covalent bond to the protein via the tyrosine 103 residue (Catalano, C. E., Choe, Y. S., Ortiz de Montellano, P. R., J. Biol. Chem. 1989, 10534 - 10541). Comparison of native horse myoglobin, recombinant sperm whale myoglobin and Tyr(103) --> Phe sperm whale mutant shows that, contrary to the previously proposed mechanism of haem to protein cross-link formation, the absence of tyrosine 103 has no impact on the formation of haem to protein cross-links. In contrast, we have found that engineered myoglobins that lack the distal histidine residue either cannot generate haem to protein cross-links or show greatly suppressed levels of modified protein. Moreover, addition of a distal histidine to myoglobin from Aplysia limacina, that naturally lacks this histidine, restores the haem protein's capacity to generate haem to protein cross-links. The distal histidine is, therefore, vital for the formation of haem to protein cross-link and we explore this outcome.     

Characterization of the catalase-peroxidase KatG from Burkholderia pseudomallei by mass spectrometry

The electron density maps of the catalase-peroxidase from Burkholderia pseudomallei (BpKatG) presented two unusual covalent modifications. A covalent structure linked the active site Trp111 with Tyr238 and Tyr238 with Met264, and the heme was modified, likely by a perhydroxy group added to the vinyl group on ring I. Mass spectrometry analysis of tryptic digests of BpKatG revealed a cluster of ions at m/z 6585, consistent with the fusion of three peptides through Trp111, Tyr238, and Met264, and a cluster at m/z approximately 4525, consistent with the fusion of two peptides linked through Trp111 and Tyr238. MS/MS analysis of the major ions at m/z 4524 and 4540 confirmed the expected sequence and suggested that the multiple ions in the cluster were the result of multiple oxidation events and transfer of CH3-S to the tyrosine. Neither cluster of ions at m/z 4525 or 6585 was present in the spectrum of a tryptic digest of the W111F variant of BpKatG. The spectrum of the tryptic digest of native BpKatG also contained a major ion for a peptide in which Met264 had been converted to homoserine, consistent with the covalent bond between Tyr238 and Met264 being susceptible to hydrolysis, including the loss of the CH3-S from the methionine. Analysis of the tryptic digests of hydroperoxidase I (KatG) from Escherichia coli provided direct evidence for the covalent linkage between Trp105 and Tyr226 and indirect evidence for a covalent linkage between Tyr226 and Met252. Tryptic peptide analysis and N-terminal sequencing revealed that the N-terminal residue of BpKatG is Ser22.     

Phaseolus vulgaris cytoplasmic leucyl-tRNA synthetase. Purification and comparison of its catalytic, structural, and immunological properties with those of the chloroplastic enzyme

The cytoplasmic leucyl-tRNA synthetase was purified from bean (Phaseolus vulgaris) leaves. After ammonium sulfate fractionation and chromatography on Sephadex G-50, DEAE-cellulose, hydroxylapatite, and phosphocellulose, complete purification was achieved by blue Sepharose CL-6B chromatography using specific elution with pure yeast tRNALeu1. The enzyme was purified 1050-fold and had a specific activity of 940 nmol of leucyl-tRNA formed/min/mg of protein. Polyacrylamide gel electrophoresis of the native enzyme showed one band, but the denatured enzyme showed two bands. These two protein bands are structurally related. The smallest protein appears to be a cleavage product from the largest one, suggesting the presence of a sensitive cleavage site in the cytoplasmic leucyl-tRNA synthetase. The cytoplasmic enzyme is a monomer (Mr = 130,000), larger than its chloroplastic counterpart (Mr = 120,000). The two enzymes differ in their substrate (tRNA) specificity, tryptic peptide map, and amino acid composition. Antibodies were raised against the cytoplasmic enzyme and against the chloroplastic enzyme and no cross-immunological reaction was detected, showing that the two enzymes do not share any antigenic determinant. Taken together, these results suggest that P. vulgaris cytoplasmic and chloroplastic leucyl-tRNA synthetases are coded for by different genes.     

Identification of essential tyrosine and lysine residues in angiotensin converting enzyme: evidence for a single active site

Inactivation of rabbit lung angiotensin converting enzyme (ACE) by 1-fluoro-2,4-dinitrobenzene (Dnp-F) has been shown to be due primarily to the modification of a tyrosine residue [Bünning, P., Kleeman, S.G., & Riordan, J.F. (1990) Biochemistry (preceding paper in this issue)]. Rabbit testicular ACE is also inactivated by Dnp-F. The specific site of modification has been identified by peptide mapping of tryptic digests of the Dnp-modified protein. Two principal 340-nm-absorbing peaks, not observed with protein modified in the presence of inhibitor, have been characterized. Amino acid and sequence analyses show that these peptides contain two distinct residues that have been selectively modified. The sequence of the major (greater than 90% of the total) modified peptide is YVEFTNK with the Dnp group on tyrosine. The sequence of the second, minor peptide is KVQDLQR with the Dnp group on lysine. Identical peptides were obtained from Dnp-modified rabbit lung ACE. These modified amino acids correspond to residues 200 and 118, respectively, in testicular ACE (human enzyme numbering). Both peptides are present only in the carboxy-terminal half-domain of lung ACE, corresponding to residues 776 and 694, respectively. These results indicate that the Dnp-F sensitive, catalytically functional active site is located in the "testicular" half of lung ACE.     

Synthetic peptides as inactivators of multimeric enzymes: inhibition of Plasmodium falciparum triosephosphate isomerase by interface peptides.

Synthetic peptides corresponding to two distinct segments of the subunit interface of the homodimeric enzyme triosephosphate isomerase (residues 9-18, ANWKCNGTLE, peptide I; residues 68-79, KFGNGSYTGEVS, peptide II) from Plasmodium falciparum (PfTIM) have been investigated for their ability to act as inhibitors by interfering with the quaternary structure of the enzyme. An analog of peptide II containing cysteine at the site corresponding to position 74 and tyrosine at position 69 in the protein sequence KYGNGSCTGEVS (peptide III) was also investigated. A substantial fall in enzyme activity was observed following incubation of the enzyme with peptide II, whereas peptide I did not show any appreciable inhibition. The inhibitory effect was more pronounced on two mutants of PfTIM (Y74C and Y74G), both of which have reduced stability compared to the wild-type protein due to an interface cavity. The IC50 value determined for peptide II is in the range of 0.6-0.8 microM. This study suggests that interface peptides of oligomeric enzymes can be used to inhibit dimeric enzymes by disrupting their native multimeric states and may provide lead structures for potential inhibitor design.     

Soybean nodule sucrose synthase (nodulin-100): further analysis of its phosphorylation using recombinant and authentic root-nodule enzymes.

Sucrose synthase (SS) is a known phosphoserine-containing enzyme in legume root nodules and various other plant "sink" tissues. In order to begin to investigate the possible physiological significance of this posttranslational modification, we have cloned a full-length soybean nodule SS (nodulin-100) cDNA and overexpressed it in Escherichia coli. Authentic nodule SS and recombinant wild-type and mutant forms of the enzyme were purified and characterized. We document that a conserved serine near the N-terminus (Ser(11)) is the primary phosphorylation site for a nodule Ca(2+)-dependent protein kinase (CDPK) in vitro. Related tryptic digestion and mass spectral analyses indicated that this target residue was also phosphorylated in planta in authentic nodulin-100. In addition, a secondary phosphorylation site(s) in recombinant nodule SS was implicated given that all active mutant enzyme forms (S11A, S11D, S11C, and N-terminal truncation between Ala(2) and Arg(13)) were phosphorylated, albeit weakly, by the CDPK. This secondary site(s) likely resides between Glu(14) and Met(193) as evidenced by CNBr cleavage and phosphopeptide mapping. Phosphorylation of the recombinant and authentic nodule Ser(11) enzymes in vitro by the nodule CDPK had no major effect on the sucrose-cleavage activity and/or kinetic properties. However, phosphorylation decreased the apparent surface hydrophobicity of the recombinant wild-type enzyme, suggesting that this covalent modification could potentially play some role in the documented partitioning of nodulin-100 between the nodule symbiosome/plasma membranes and cytosol in planta.     

Chorismate mutase/prephenate dehydrogenase from Escherichia coli K12: purification, characterization, and identification of a reactive cysteine

The bifunctional enzyme involved in tyrosine biosynthesis, chorismate mutase/prephenate dehydrogenase, has been isolated from extracts of a regulatory mutant of Escherichia coli K12. The pure enzyme is a homodimer of total molecular weight 78 000 and displays Michaelis-Menten kinetics for both activities. Fingerprinting and amino acid sequencing of tryptic and thermolytic peptides of the S-[14C]carboxymethylated enzyme allowed the identification of three unique cysteine-containing sequences per subunit. Chemical modification of the native enzyme with 5,5'-dithiobis(2-nitrobenzoate) or iodoacetamide showed that one sulfhydryl group per subunit was particularly reactive, and the integrity of this group was essential for both enzymic activities. This work supports previous proposals for a close spatial relationship between the active sites.     

Sulfation, the up-and-coming post-translational modification: its role and mechanism in protein-protein interaction

Tyrosine sulfation is a post-translational modification entailing covalent attachment of sulfate to tyrosine residues. It takes place in the trans-Golgi, is necessary for the bioactivity of some proteins, and improves their ability to interact with other proteins. In the present work, we show that a protein containing a sulfated tyrosine with a delocalized negative charge forms a salt bridge with another protein if it has two or more adjacent arginine residues containing positive delocalized charges. These noncovalent complexes are so stable that, when submitted to collision induced dissociation, the peptides forming the complex dissociate. Just one covalent bond fragments, the covalent bond between the tyrosine oxygen and the SO3 sulfur, and is represented by the appearance of a new peak (basic peptide + SO3), suggesting that in some instances covalent bonds will break down before the noncovalent bonds between the arginine guanidinium and SO3 dissociate. The data implies that the dissociation pathway is preferred; however, fragmentation between tyrosine and the sulfate residue is a major pathway.     

Multiple covalent modifications of Trg, a sensory transducer of Escherichia coli

The sensory transducers of Escherichia coli are integral membrane proteins that mediate the tactic response of cells to chemical stimuli. Adaptation to environmental stimuli is correlated with methylation of the transducer proteins. Two transducer genes, tsr and tar, exhibit extensive homologies while no homology has been detected between a third transducer, trg, and those genes. The Tsr and Tar proteins have been shown to contain multiple sites for methylation as well as two sites for another modification that requires an active cheB gene product and is designated the CheB-dependent modification. In this study, covalent modifications of the Trg protein were characterized by analysis of tryptic peptides. We found that methylation occurred at several sites on the Trg protein and that the protein contained at least three sites for CheB-dependent modification, two of which were located on a tryptic peptide that contains both methionine and lysine. This tryptic peptide is analogous to the methionine- and lysine-containing methyl-accepting peptides isolated from the Tsr and Tar proteins and like those peptides may contain several methyl-accepting sites. We estimated the pKa of the group created by the CheB-dependent modification on the methionine- and lysine-containing peptide of Trg to be between pH 2.2 and 5.8. This result supports the idea that the CheB-dependent modification is an enzymatic deamidation of glutamine to glutamic acid.     

Role of the N-terminus of rat pheochromocytoma tyrosine hydroxylase in the regulation of the enzyme's activity

Activation of rat pheochromocytoma tyrosine hydroxylase by limited tryptic proteolysis was investigated. The modifications produced upon the enzyme's structure were analyzed with the use of sodium dodecyl sulfate/polyacrylamide gel electrophoresis and tyrosine hydroxylase activity was measured all through the digestion. During the proteolysis the activity of tyrosine hydroxylase was elevated threefold at the same time as a 56-kDa tryptic fragment was formed. When the enzyme was phosphorylated, at its N-terminal region, by a kinase copurified with tyrosine hydroxylase, the major 56-kDa species did not appear to be phosphorylated on the autoradiograph, suggesting that it was derived from the native subunit by cleavage of the N-terminal of the protein. The reactivity of the 2/40/15 anti-(tyrosine hydroxylase) monoclonal antibody with the N-terminal of tyrosine hydroxylase was also investigated, using the Western-blot technique. This antibody reacted with the 62-kDa hydroxylase subunit but not with the 60-kDa tryptic fragment; the amino acid sequences of these two species showed that the 60-kDa fragment lacked the first 16 N-terminal amino acids of the native molecule. These results suggest that the N-terminal region of tyrosine hydroxylase is apparently responsible for an inhibition of the hydroxylase activity and that the first N-terminal amino acids of the hydroxylase are necessary for the recognition of the enzyme by its antibody.     

Formation of a Schiff base intermediate is not required for the adenine glycosylase activity of Escherichia coli MutY

The mutY gene product of Escherichia coli is a 39-kDa protein that catalyzes the removal of adenine bases mispaired with 2'-deoxyguanosine and 7,8-dihydro-8-oxo-2'-deoxyguanosine (OG) in DNA. Although adenine removal proceeds via monofunctional glycosylase activity, MutY is able to form covalent adducts with substrate DNA in the presence of borohydride, a trait otherwise known to be associated only with enzymes having bifunctional glycosylase/AP lyase activity. To help identify active site residues involved in the formation of MutY-DNA adducts in the presence of borohydride, a series of site-directed mutant forms of MutY were generated. Our data show that Lys 142 is the primary residue involved in cross-link formation. The absence of Lys 142 results in near elimination of the enzyme-DNA adducts formed relative to wild-type, suggesting that this residue is the primary one involved in forming covalent associations with DNA during MutY catalysis. Importantly, the enzymatic activity and DNA binding of the K142A enzyme is nearly identical to the WT enzyme. This shows that formation of the covalent intermediate is not required for adenine removal by MutY. Furthermore, this suggests that the covalent intermediate is formed by reaction of Lys 142 with the OG/G:(AP site) product, and this may be a consequence of MutY's unusually high affinity for the product of its glycosylase action.     

Preganglionic stimulation increases the phosphorylation of tyrosine hydroxylase in the superior cervical ganglion by both cAMP-dependent and Ca2+-dependent protein kinases

Electrical stimulation of the preganglionic cervical sympathetic trunk increases the phosphorylation of tyrosine hydroxylase in the superior cervical ganglion of the rat by a nicotinic mechanism and by a noncholinergic mechanism. We have measured the incorporation of [32P]Pi into specific tryptic phosphopeptides in tyrosine hydroxylase in order to identify the protein kinases that phosphorylate this enzyme in electrically stimulated ganglia. 32P-labeled tyrosine hydroxylase was isolated from the ganglion by immunoprecipitation and polyacrylamide gel electrophoresis and was subjected to tryptic hydrolysis. Seven tryptic peptides were resolved from these hydrolysates by two-dimensional thin-layer electrophoresis and chromatography. Preganglionic stimulation (20 Hz, 5 min) increased the incorporation of 32P into four of these peptides. In the presence of cholinergic antagonists, however, electrical stimulation increased the labeling of only one phosphopeptide. From a comparison of the effects of preganglionic stimulation with the effects of agonists that activate specific protein kinases, we conclude that electrical stimulation increases the phosphorylation of tyrosine hydroxylase by both a cAMP-dependent protein kinase and a Ca2+/calmodulin-dependent protein kinase. The nicotinic component of preganglionic stimulation appears to be mediated by a Ca2+/calmodulin-dependent protein kinase, while the noncholinergic component appears to be mediated by cAMP-dependent protein kinase. Although protein kinase C can phosphorylate tyrosine hydroxylase, this kinase does not appear to participate in the stimulation-induced phosphorylation of tyrosine hydroxylase in the superior cervical ganglion.     

Reductive alkylation of lysine residues in subtilisin DY

Only lysine epsilon-amino groups (and the N-terminal alpha-amino group) in native subtilisin DY were reductively alkylated by glyceraldehyde in the presence of sodium cyanoborohydride. The modified protein molecule was cleaved by TosPheCH2Cl-trypsin or cyanogen bromide and the two sets of peptides obtained were fractionated and purified by gel filtration and HPLC. For determination of the degree of modification of each lysine residue, selected peptides were subjected to sequence analysis combined with quantitative estimation of the containing PTH-Lys and PTH-epsilon-DHP-Lys. The data obtained showed that the lysine residues in positions 12, 15, 27, 43, 136, 141, 265 were entirely modified, those in positions 170, 184, 237 were partially modified, and Lys22 and Lys94 were unaccessible for the reagent. The caseinolytic activity decreased by 23% when the maximum number of lysine residues (8.6 of the total 12 residues) in subtilisin DY were modified. The CD-spectra of native and modified enzyme showed only slight differences. Both these experiments suggest that the lysine residues do not take part directly in the catalytic reaction but are responsible for maintaining the native three-dimensional enzyme structure. The data obtained for the accessibility of the different lysine residues in subtilisin DY correlated very well with the positions of these residues in a video model of the structure of subtilisin Carlsberg, thus suggesting that the spatial structures of these two enzymes are very similar.