Heme-linked protonation of HCN, CO, NO and O2 complexes of reduced horseradish peroxidases.

Small reversible changes in the absorption spectra of HCN, CO, NO and O₂ complexes of ferrous diacetyldeuteroperoxidase A, not hitherto observed, were attributed to proton dissociation of a distal amino acid residue. From spectrophotometric titration data the pK_a was measured as 5.5 (HCN), 5.6 (ligand free), 6.0 (CO), 6.55 (NO) and 8.0 (O₂). The value of 8.0 for the pK_a of the O₂ complex was also obtained from a curve of pH dependence of proton uptake in the reaction of the ferrous enzyme with O₂. Absorption bands in the visible region were shifted to longer wavelengths in the order of CO to NO to O₂ which is the decreasing order of the energy of π¹ level of these diatomic ligands.The pK_a values for CO complexes of ferroperoxidases, isoenzymes A and (B+C) were varied with substituents at the 2 and 4 positions of deuterohemin IX, and the $\text{Δp}\text{K}{}_{\mathrm{<mtext>a</mtext>}}\text{Δp}\text{K}{}_3$ ratio was about 0.3 in both series of isoenzyme preparations, where pK₃ is a measure of basicity of pyrrole nitrogen.The present data support the previous conclusion (Yamada and Yamazaki (1974) Arch. Biochem. Biophys., 165, 728) that the pK_a for ferroperoxidases, measured from small reversible changes in the absorption spectra, represents a proton dissociation constant of a distal amino acid residue and that there is hydrogen bonding between the residue and a ligand atom directly bound to the iron atom.     

Axial ligation constants of [5,10,15,20-tetraphenylporphyrinato(2-)] cobalt

Axial ligation constants (log K^B) of bases for [5, 10,15,20-tetraphenylporphyrinato(2-)]cobalt ([Co^II (tpp)]) are reported. The log K^B values of pyridine derivatives except for 4-cyanopyridine show a good linear relationship in a plot of log K^B vs. pK_a of the axial ligand. 4-Cyanopyridine gives a larger log K^B than expected. The log K^B value of 1-methylimidazole for [Co^II(tpp)] is almost the same as that for tetrakis(p-methoxyphenyl)porphyrinatocobalt(II) ([Co((p-CH₃O)tpp)]), althoug the other log K^B values for [Co^II(tpp)] are always slightly larger than those for [Co((p-CH₃O)tpp)]. These results are discussed on the basis of the σ- and π-bonding abilities of the bases, and the solvent effects on the log K^B values. The lower base affinities of cobalt(II) capped porphyrins are also discussed.     

The effects of external potassium and long duration voltage conditioning on the amplitude of sodium currents in the giant axon of the squid, Loligo pealei

Giant axons were voltage-clamped in solutions of constant sodium concentration (230 mM) and variable potassium concentrations (from 0 to 210 mM). The values of the peak initial transient current, I_p, were measured as a function of conditioning prepulse duration over the range from less than 1 msec to over 3 min. Prepulse amplitudes were varied from E _m = -20 mv to E _m = -160 mv. The attenuation of the I_p values in high [K_o] was found to vary as a function of time when long duration conditioning potentials were applied. In both high and low [K_o], I_p values which had reached a quasi-steady—state level within a few milliseconds following a few milliseconds of hyperpolarization were found to increase following longer hyperpolarization. A second plateau was reached with a time constant of about 100–500 msec and a third with a time constant in the range of 30 to 200 sec. The intermediate quasi-steady—state level was absent in K-free ASW solutions. Sodium inactivation curves, normalized to I _pmax values obtained at either the first or second plateaus, were significantly different in different [K_o]. The inactivation curves, however, tended to superpose after about 1 min of hyperpolarizing conditioning. The time courses and magnitudes of the intermediate and very slow sodium conductance restorations induced by long hyperpolarizing pulses are in agreement with those predicted from the calculated rates and magnitudes of [K⁺] depletion in the space between the axolemma and the Schwann layer.     

Electrostatic control of enzyme reactions: The mechanism of inhibition of glucose oxidase by putrescine

The interaction of putrescine dihydrochloride with glucose oxidase is reported. At pH 7.65 glucose oxidase is strongly anionic (Z = ?80). The pK_a of an essential acidic group on the reduced form of the enzyme is extremely sensitive to ionic strength, as predicted by simple electrostatic theory [J. G. Voet, J. Coe, J. Epstein, V. Matossian, and T. Shipley (1981), Biochemistry, 20, 7182–7185]. Putrescine dihydrochloride was found to inhibit glucose oxidase at pH 7.65 at a constant ionic strength of 0.05. The kinetics do not obey simple competitive inhibition, however. The data can best be explained by a model in which change in the electrostatic potential of the enzyme on putrescine binding changes the observed pK_a of the essential acidic group. The pH dependence of putrescine inhibition supports this interpretation. At I = 0.05, 5 mM putrescine was found to change the pK_a of the essential acidic group from 7.6 to 7.1. The shift in the pK_a as a function of putrescine concentration at pH 7.7 and I = 0.05 also supports the model presented. The K_a for putrescine to the active form of the enzyme was calculated to be 4.2 mm.     

Capillary isoelectric focusing of a difficult-to-denature tetrameric enzyme using alkylurea–urea mixtures

Capillary isoelectric focusing (cIEF) is normally run under denaturing conditions using urea to expose any buried protein residues that may contribute to the overall charge. However, urea does not completely denature some proteins, such as the tetrameric enzyme Erwinia chrysanthemil-asparaginase (ErA), in which case electrophoresis-compatible alternative denaturants are required. Here, we show that alkylureas such as N-ethylurea provide increased denaturation during cIEF. The cIEF analysis of ErA in 8 M urea alone resulted in a cluster of ill-resolved peaks with isoelectric points (pI values) in the range 7.4 to 8.5. A combination of 2.0 to 2.2 M N-ethylurea and 8 M urea provided sufficient denaturation of ErA, resulting in a main peak with a pI of 7.35 and an acidic species minor peak at 7.0, both comparing well with predicted pI values based on the sum of protein residue pK_a values. Recombinant deamidated ErA mutants were also demonstrated to migrate to pI values consistent with predictions (pI 7.0 for one deamidation). The quantitation of ErA acidic species in samples from full-scale manufacturing (1.0–3.5% of total peak area) was found to be reproducible and linear. Use of alkylureas as denaturing agents in capillary electrophoresis and cIEF should be considered during biopharmaceutical assay development.     

Redox potentials of trinuclear μ-oxo ruthenium acetate clusters with N-heterocyclic ligands

The trinuclear clusters of general composition [Ru₃O(OOCCH₃)₆(N-Het)₃], where N-Het=pyridine and pyrazine derivatives, exhibit a series of reversible waves in the range of −1.8 to 2.4 V versus SHE, in acetonitrile, ascribed to the successive [cluster]^{−2/−1/0/+1/+2/+3} redox couples. The redox potentials decrease with the pK_a of the N-heterocyclic ligands according to the equations E°(+3/+2)= 2.24−0.023 pK_a; E°(+2/+1)=1.34−0.029 pK_a; E°(+1/0)=0.36−0.039 pK_a and E°(0/−1)=−0.68− 0.074 pK_a. The dependence is greater at lower oxidation states, reflecting the role of π-backbonding in the complexes.     

Role of Glu445 in the substrate binding of β-glucosidase

A previously uncharacterized gene in Neosartorya fischeri was cloned and expressed in Escherichia coli. It was found to encode a β-glucosidase (NfBGL1) distinguishable from other BGLs by its high turnover of p-nitrophenyl β-d-glucopyranoside (pNPG). Molecular determinants for the substrate recognition of NfBGL1 were studied through an initial screening of residues by sequence alignment, a second screening by homology modeling and subsequent site-directed mutagenesis to alter individual screened residues. A conserved amino acid, E445, in the substrate binding pocket of wild-type NfBGL1 was identified as an important residue affecting substrate affinity. Replacement of E445 with amino acids other than aspartate significantly decreased the catalytic efficiency (k_cat/K_m) of NfBGL1 towards pNPG, mainly through decreased binding affinity. This was likely due to the disruption of hydrogen bonding between the substrate and the carboxylate oxygen of the residue at position 445. Density functional theory (DFT) based studies suggested that an acidic amino acid at position 445 raises the substrate affinity of NfBGL1 through hydrogen bonding. The residue E445 is completely conserved indicating that this position can be considered as a crucial determinant for the substrate binding among GHs tested.     

Action on peptides by wheat carboxypeptidase

A kinetic analysis has been performed with purified wheat carboxypeptidase by the use of N-acyl dipeptides, Z-Gly-Pro-Leu-Gly (Z = benzyloxycarbonyl), angiotensin II and bradykinin. The values of k_cat were dramatically influenced by amino acid residues occupying the penultimate position from the carboxyl terminus of substrates. The structure of the substrate did not appreciably affect the K_m values.     

Purification and characterization of two β-N-acetylhexosaminidases from the tobacco hornworm,Manduca sexta (L.) (Lepidoptera: Sphingidae)

β-N-Acetylhexosaminidases were detected in 10 insects including species of Lepidoptera, Coleoptera, Hemiptera, and Orthoptera. Two enzymes were purified from the tobacco hornworm, Manduca sexta (L.). EI was detected in larval and pharate pupal molting fluid, integument, and pupal hemolymph while EII was found in larval and pupal hemolymphs. They are acidic hydrolases with similar molecular weights (6.1 × 10⁴), molar extinction coefficients at 280 nm (1.9 × 10⁵ liters mol^?1 cm^?1), and pH optima (pH 6). They differ in the number of polypeptide chains per molecule (EI is a single chain and EII consists of two polypeptide chains), amino acid composition, extent of glycosylation (EII is probably a glycoprotein), isoelectric point (pI_EI = 5.9 and pI_EII ～- 5.1), tissue distribution, and reactivities toward nitrophenylated N-acetylglucosamine (k_cat,I = 328 s^?1 and k_cat,II = 103 s^?1) and N,N′-diacetylchitobiose (k_cat,I = 307 s^?1 and k_cat,II = 3 s^?1). These results suggest that EI is a chitinase and that EII may function as a hexosaminidase in vivo.     

Voltage dependent potassium fluxes and the significance of action potentials in Acetabularia

Membrane potential, V_m, and K⁺(⁸⁶Rb⁺) fluxes have been measured simultaneously on individual cells of Acetabularia mediterranea. During resting state (resting potential approx. ?170 mV) the K⁺ influx amounts to 0.24–0.6 pmol · cm^?2 · s^?1 and the K⁺ efflux to 0.2–1.5 pmol · cm^?2 s^?1. According to the K⁺ concentrations inside and outside the cell (40 : 1) the voltage dependent K⁺ flux (zero at V_m = E_K = ?90 mV) is stimulated approx. 40-fold for V_m more positive than E_K.It is calculated that during one action potential (temporary depolarization to V_m more positive than E_K) a cell looses the same amount of K⁺, which leaks in during 10–20 min in the resting state (V_m = ?170 mV). Since action potentials occur spontaneously in Acetabularia, they are therefore suggested to have a significant function for the K⁺ balance of this alga.     

Kinetics of interaction of some α- and β-D-monosaccharides with concanavalin A

The rates of formation and dissociation of concanavalin A with some 4-methylumbelliferyl and p-nitrophenyl derivatives of α- and β-D-mannopyranosides and glucopyranosides were measured by fluorescence and spectral stopped-flow methods. All process examined were uniphasic. The second-order formation rate constants varied only from 6.8 · 10⁴ to 12.8 · 10⁴ M^?. s^?1, whereas the first-order dissociation rate constants ranged from 4.1. to 220 s^?1, all at ph 5.0, I = 0.3 M, and 25°C. Dissociation rates thus controlled the value of binding constant. The effect of temperature on these reactions was examined, from which enthalpies and entropies of activation and of reaction could be calculated. The effects of pH at 25°C on the reaction rates of 4-methylumbelliferyl α-D-mannopyranoside and 4-methylumbelliferyl α-D-glucopyranoside with concanavalin A were examined. The value of the binding constant K_ap (derived from the kinetics) at any pH could be related to the intrinsic binding constant K by the expression K_ap = K_aK(K_a + [H⁺])^?1. The values of K_a, the ionization constant of the protein segment responsive to sugar binding, were 3 · 10^?4 M and 1 · 10^?4 M for 4-methylumbelliferyl α-D-mannopyranoside and 4-methylumbelliferyl α-D-glucopyranoside, respectively. The binding constant of p-nitrophenyl α-D-mannopyranoside is surprisingly much less sensitive to a pH change from 5.0 to 2.7. Ionic strength had little effect on the binding characteristics of 4-methylumbelliferyl α-D-mannopyranoside to concanavalin A at pH 5.2 and 25°C.     

Purification and characterization of an intracellular N-terminal exopeptidase from Streptococcus durans

An intracellular N-terminal exopeptidase isolated from cell extracts of Streptococcus durans has been purified 470-fold to homogeneity (specific activity of 12.0 μmol/min per mg). In the absence of thiol compounds, the purified aminopeptidase undergoes a slow oxidation with a 70% loss of activity, which can be restored by the addition of 2 mM β-mercaptoethanol. The purified aminopeptidase (M_r 300 000) preferred L-peptide and arylamide substrates with small nonpolar or basic side chains. SDS electrophoresis yielded a single protein band corresponding to a molecular weight of 49 400, suggesting that the native enzyme is a hexameric protein. The enzyme-catalyzed hydrolysis of $\text{L}\text{-}\text{alanyl}\text{-p-}\text{nitroanilide}$ exhibited a bell-shaped pH dependence for log V_max/K_m(pK₁ = 6.35; pK₂ = 8.50) while the log V_max versus pH profile showed only an acid limb (pK = 6.35). Methylene blue-sensitized photooxidation of the enzyme resulted in the complete loss of activity, while L-leucine, a competitive inhibitor, partially protected against this inactivation. Amino acid analysis indicated that this photooxidative loss of activity corresponded to the modification of one histidine residue per enzyme monomer. N-Ethylmaleimide (100 mM) caused a 78% reduction in enzyme activity. Treatment of the enzyme with 1.0 mM hydrogen peroxide resulted in the oxidation of two cysteine residues per enzyme monomer and caused a 70% decrease in the catalytic activity.     

Heme-linked ionization and chloride binding in intestinal peroxidase and lactoperoxidase.

Hog intestinal peroxidase and bovine lactoperoxidase exhibited similar spectral shifts upon pH alteration. From spectrophotometric titrations, it was found that there are hemelinked ionizations of pK_a = 4.75 in intestinal peroxidase and pK_a = 3.5 in lactoperoxidase. The apparent pK_a (pK_a′) increased with the increase in chloride concentration. The pK_a′ vs log[Cl^?] plots showed that the chloride forms complex with the acid forms of these enzymes with a dissociation constant (pK = 2.7). Although the dissociation constant (K_d) of the peroxidase-cyanide complexes is nearly independent of pH, cyanide competed with chloride in the acidic pH region. The slopes of logK_d vs log[Cl^?] were 1.0 for intestinal peroxidase and 0.5 for lactoperoxidase. The reaction of hydrogen peroxide with these peroxidases was also affected by chloride, similarly as the reaction with cyanide was. The results were explained by assuming that protonation occurs at the distal base and destroys the hydrogen bond between the base and a water molecule at the sixth coordinate position of the heme iron.     

Kinetics of substitution of ferrocenyl-containing β-diketonato ligands by phenanthroline from β-diketonato-1,5-cyclooctadienerhodium(I) complexes

Second-order rate constants, k₂, for the substitution of the ferrocene-containing β-diketonato ligands FcCOCHCOR⁻ with R=CF₃ (ferrocenoyltrifluroacetonato, fctfa, pK_a 6.56), CCl₃ (ferrocenoyltrichloroacetonato, fctca, 7.13), CH₃ (ferrocenoylacatonato, fca, 10.01), Ph (anion of benzoylferrocenoylmethane, bfcm, 10.41) and Fc (anion of diferrocenoylmethane, dfcm, 13.1) (Ph=phenyl, Fc=ferrocenyl, values in brackets are the pK_a values of the free β-diketones) from the complexes [Rh(cod)(FcCOCHCOR)] with 1,10-phenanthroline (phen, cod=1,5-cyclooctadiene) at 25 °C were found to be 560 (R=CF₃), 1370 (CCl₃), 30 (Ph), 18 (CH₃) and 7.0 dm³ mol⁻¹ s⁻¹ (Fc), respectively. The temperature dependence of each reaction was determined and the large negative values obtained for activation, ΔS^#<−100 J K⁻¹ mol⁻¹ for all but R=CCl₃ (ΔS^#_CCl3=−81 J K⁻¹ mol⁻¹), suggests an associative substitution mechanism. The rate law of the reaction was found to be R={k_s+k₂[phen]}[Rh(cod)(FcCOCHCOR)]. Since the solvent-associated rate constant k_s≈0 for all R except Ph (k_s,RPh=0.06 s⁻¹) the solvent, methanol, plays a limited role in the reaction. Results are interpreted to imply that the rate-determining step during substitution is breaking of an RhO bond and not the formation of an RhN bond. The role of β-diketone pK_a and group electronegativity, χ, of each R group on the rate of substitution are also discussed.     

Purification and properties of neutral β-galactosidases from human liver

Two neutral β-galactosidase isozymes were purified from human liver. The initial step of purification was removal of the acidic β-galactosidases by adsorption on concanavalin A-Sepharose 4B conjugate. Subsequent purification steps included ammonium sulfate precipitation, diethylaminoethyl cellulose column chromatography, Sephadex G-100 gel filtration, and preparative polyacrylamide-gel isoelectric focusing. The final step of purification was affinity chromatography of the separated isoelectric forms on ?-aminocaproyl-β-d-galactosylamine-Sepharose 4B conjugate. The purified β-galactosidase isozymes had activity toward both β-d-galactoside and β-d-glucoside derivatives of 4-methylumbelliferone and p-nitrophenol with a pH optimum around 6.2. These enzyme forms were also found to possess lactosylceramidase II activity with a pH optimum in the range of 5.4 to 5.6, but not lactosylceramidase I activity and no activity toward galactosylceramide or GM₁-ganglioside. The molecular weight was found to be in the range of 37,500–39,500 for the two neutral isozymes and they had similar K_m and V values; the more acidic form (designated β-galactosidase N₁) was more heat stable than the other form (designated β-galactosidase N₂). Antibodies evoked against the N₁ and N₂ β-galactosidases gave identical precipitin lines retaining enzymatic activity. No cross-reactivity was observed between the neutral and the acidic isozymes when examined with the respective antisera.     

Metal ion-biomolecule interactions. II. Methylmercuration of the deprotonated amino groups in adenine,guanine, and cytosine derivatives,and its relationship to amino group acidity

Proton nmr spectroscopic evidence is presented for methylmercury(II) binding to the deprotonated amino groups in adenosine, 9-methyladenine, guanosine, 1-methylguanosine, and cytidine under basic conditions. Except for the guanosine case, ¹H nmr spectra of the products from aqueous or ethanolic 1:1 mixtures of substrate and MeHgOH are consistent with methylmercuration of the deprotonated amino groups. Guanosine undergoes initial binding of MeHg to N₁, and a second equivalent of MeHgOH is necessary to effect amino binding. The nmr spectra of the complexed adenine derivatives suggest that different geometrical isomers exist in (CD₃)₂SO solution, reflecting the partial double bond character of the C₆N bond in these systems. Using a correlation relating the magnitude of the ¹⁹⁹Hg-¹H coupling constant (J) for MeHg-ligand complexes with the ligand pK_a (J = ?3.88 pK_a + 248.5, extending over 13 pK units, based on a variety of N and O donor ligands), estimates (± 0.3 pK unit) of the pK_as of the amino groups of the above substrates have been made. In this way, pK_a values of 15.5 (cytidine), 17.0 (adenosine and 9-methyladenine), 15.1 (guanosine), and 14.9 (1-methylguanosine) are obtained. In the cases where comparisons with literature pK_a data can be made, good agreement is found.     

Hydrophobic binding sites of human liver alanine aminopeptidase

Fatty acids, alkyl amines, and amides of α-amino fatty acids inhibit human liver alanine aminopeptidase apparently by binding to residue binding site 1 of the active center, i.e., the N-terminal binding site. The pK_i values of the acids, amines, and amides increase until the overall chain length reaches eight carbons. The pK_i values are the same for members of the series with chain lengths longer than eight carbon atoms. Assuming an extended structure of the inhibitors, this site will accommodate amino acid side chains of not longer than 11.7 Å from the α-carbon to the end of the chain. Long chain amino acids inhibit by binding apparently at residue site 3. The pK_i values of dl-α-amino acids from α-aminobutyric acid to α-aminodecanoic acid increase with the addition of each methylene unit. Thus, site 3 will accommodate amino acid side chains which are at least 13.0 Å from the α-carbon to the end of the chain. Methanol and other organic solvents reversibly inhibit the binding of substrates at pH 6.9 without affecting the maximum rate of catalysis. At lower pH values, the maximum rate of catalysis is lowered. Sodium chloride also inhibits substrate hydrolysis at pH 6.9 but does not affect the maximum rate of catalysis. The pK_i values of fatty acids, alkyl amines, and amino acids are strongly decreased by methanol and slightly increased by sodium chloride. These data indicate that a major portion of the interactions of the enzyme with fatty acids, amines, and amino acids is of a hydrophobic nature.     

Selective modification of tyrosine-68 in β1-bungarotoxin from the venom ofBungarus multicinctus (Taiwan banded krait)

β₁-Bungarotoxin (β₁-Bgt) fromBungarus multicinctus (Taiwan banded krait) snake venom was subjected to tyrosine modification withp-nitrobenzenesulfonyl fluoride (NBSF) atpH 8.0 and the NBS derivatives were separated by high-performance liquid chromatography. The results of amino acid analysis revealed that only one Tyr residue out of 14 was modified, and the modified residue was identified to be Tyr-68 in the A chain of β₁-Bgt. Spectrophotometric titration indicated that the phenolic group of Tyr-68 has apK of 10.1. Modification of Tyr-68 in the A chain caused a selective loss in lethal toxicity, but had no effect on either enzymatic or antigenic activities. The Ca²⁺-induced difference spectra and fluorescence study indicated that β₁-Bgt possesses at least two different types of Ca²⁺-binding sites. However, modification of Tyr-68 in β₁-Bgt did not cause any change of the Ca²⁺-induced difference spectra and fluorescence spectra in native toxin and the two types of Ca²⁺-binding sites were retained. Moreover, the affinity of Tyr-68-modified β₁-Bgt for 8-anilinonaphthalene sulfonate was also unaffected in both the presence and absence of Ca²⁺. All of the results indicated that Tyr-68 is not involved in the Ca²⁺ and substrate bindings in the A chain of β₁-Bgt. It is concluded that lethal toxicity is not necessarily associated with enzymatic, antigenic, and Ca²⁺-binding activities in β₁-Bgt.     

Biochemical peculiarities of benzaldehyde lyase from Pseudomonas fluorescens Biovar I in the dependency on pH and cosolvent concentration

Benzaldehyde lyase from Pseudomonas fluorescens (BAL, EC 4.1.2.38) is a versatile catalyst for stereoselective carboligations. Nevertheless, rather inconsistent data about its biochemical properties are reported in literature. In this study, the dependency of BAL activity on ionic strength, pH, and concentration of DMSO was for the first time systematically investigated and interpreted. It was found that the activity of BAL strongly depends on all three parameters, and a correlation exists between the dependency on pH and DMSO concentration. This correlation could be explained by an interaction of DMSO with an ionic amino acid in the catalytic site. A model-based analysis indicated that the pK_a of this residue shifts to the alkaline milieu upon addition of DMSO. Consequently, the optimum pH also shifts to alkaline values when DMSO is present. Potentiometric experiments confirmed that the pK_a can most probably be attributed to Glu50 which governs the activity increase of BAL on the acidic limb of its pH-activity profile. With these findings, the apparently contradicting data from literature become comprehensible and optimal reaction conditions for synthesis can easily be deduced.     

ThepH dependence of the equilibrium constantK Hyd for the hydrolysis of the Lys15-Ala16 reactive-site peptide bond in bovine pancreatic trypsin inhibitor (aprotinin)

ThepH dependence of the equilibrium constant K_Hyd for the hydrolysis of the Lys¹⁵-Ala¹⁶ reactive-site peptide bond of the bovine pancreatic trypsin inhibitor (aprotinin) was investigated over thepH range 2.3–6.5. Solutions of aprotinin, modified aprotinin with the Lys¹⁵-Ala¹⁶ peptide bond cleaved and mixtures of both species were incubated with 10 mol% porcine β-trypsin. The state of equilibrium was determined by analytical cation-exchange HPLC. The K_Hyd values obtained did not exactly obey the simple equation of Dobry et al. (1952), which had to be used in an extended form with two additional parameters for a satisfactory fit. ThepH-independent equilibrium constant is 0.90 and thepK values of the Lys¹⁵ carboxyl group and of the Ala¹⁶ amino group are 3.10 and 8.22, respectively. ThepK of an additional group is apparently perturbed by the peptide-bond hydrolysis. It is 4.60 in the native and 4.40 in the modified aprotinin.