Role of tyrosine and tryptophan residues in the structure-activity relationships of a cardiotoxin from Naja nigricollis venom

This paper is an attempt to localize the critical area determining toxicity in a snake cardiotoxin. Toxin gamma is a single-chain polypeptide of 60 amino acids, which has been isolated from the venom of the African spitting cobra, Naja nigricollis. Three aromatic residues, namely, Trp-11, Tyr-22, and Tyr-51, have been individually modified by chemical means. The structure of the native toxin and of each derivative has been carefully investigated by circular dichroism, fluorescence, proton magnetic resonance spectroscopy, and two specific monoclonal antibodies. None of the chemical modifications alters the overall structure of the toxin, which in all cases remains folded into three adjacent loops (I, II, and III) rich in beta-pleated sheet emerging from a small globular region containing four disulfide bridges. A number of subtle changes, however, have been detected in the structure of each derivative compared with that of the native toxin. In particular, nitration of Tyr-51 provoked a structural perturbation in the globular region. Nitration of Tyr-22 induces a more substantial change in the beta-sheet area of the molecule. Thus, the strong inter-ring NOE that is observed in the native toxin between Tyr-22 and Tyr-51 vanishes in the Tyr-22 derivative, and significant changes are observed in the globular region. In contrast, no alteration of the beta-sheet structure of loops II and III has been detected after modification of Trp-11. All changes observed for this derivative remain located in the vicinity of the indole side chain of Trp-11 in loop I. The biological consequences of the modifications were measured: the lethal potency in vivo in mice and the cytotoxic activities in vitro on FL-cells. Lethal activities correlate with cytotoxicity: Tyr-51 modified toxin is equally potent as native toxin, whereas Tyr-22 and Trp-11 derivatized toxins are characterized by substantially lesser activities, the Trp-11 derivatized toxin being the least potent. We conclude that (1) Tyr-51 is not involved in the functional site of the toxin, although it is in interaction with the core of the molecule, (2) Tyr-22 may play a dual structural and functional role, and (3) Trp-11 is in, or in close proximity to, the functional site of the toxin. These data indicate the importance of loop I in determining toxicity of the cardiotoxin.     

The status of tyrosyl residues in a Formosan cobra cardiotoxin.

Spectrophotometric titration of Formosan cobra cardiotoxin showed that two of the three tyrosyl residues were titrated freely with a normal apparent pKa of 9.6 whereas the remaining one ionized at pH above 11.0. Nitration of cardiotoxin in Tris . HCl buffer with tetranitromethane resulted in the selective nitration of tyrosine 11 and tyrosine 22. It also revealed that tyrosine 51 was the abnormal one in the spectrophotometric titration. Complete nitration occurred in the presence of 6.0 M guanidine hydrochloride. Compared with the conformation of native cardiotoxin, the peptide conformation of the partially nitrated cardiotoxin did not change significantly but the conformation of the completely nitrated cardiotoxin changed remarkably. The biological activity of cardiotoxin was indeed affected by nitration, but the immunological activity was nearly intact even when all the tyrosine residues were nitrated.     

Delineation of the functional site of a snake venom cardiotoxin: preparation, structure, and function of monoacetylated derivatives

Toxin gamma, a cardiotoxin from the venom of the cobra Naja nigricollis, was modified with acetic anhydride, and the derivatives were separated by cation-exchange and reverse-phase chromatography. Nine monoacetylated derivatives were obtained, and those modified at positions 1, 2, 12, 23, and 35 were readily identified by automated sequencing. The overall structure of toxin gamma, composed of three adjacent loops (I, II, and III) rich in beta-sheet, was not affected by monoacetylation as revealed by circular dichroic analysis. Trp-11, Tyr-22, and Tyr-51 fluorescence intensities were not affected by modifications at Lys-12 and Lys-35, whereas Trp-11 fluorescence intensity slightly increased when Lys-1 and Lys-23 were modified. The cytotoxic activity of toxin gamma to FL cells in culture was unchanged after modification at positions 1 and 2, whereas it was 3-fold lower after modification at Lys-23 and Lys-35. The derivative modified at Lys-12 was 10-fold less active than native toxin. Using two isotoxins, we found that substitutions at positions 28, 30, 31, and 57 did not change the cytotoxic potency of toxin gamma. A good correlation between cytotoxicity, lethality, and, to some extent, depolarizing activity on cultured skeletal muscle cells was found. In particular, the derivative modified at Lys-12 always had the lowest potency. Our data show that the site responsible for cytotoxicity, lethality, and depolarizing activity is not diffuse but is well localized on loop I and perhaps at the base of loop II. This site is topographically different from the AcChoR binding site of the structurally similar snake neurotoxins.     

Circular dichroism study of the unfolding-refolding of a cardiotoxin from Taiwan cobra (Naja naja atra) venom

Circular dichroism spectroscopy has been used to study the unfolding-refolding process of a cardiotoxin from Taiwan cobra (Naja naja atra) venom upon addition of fluoroalcohols or sodium dodecyl sulfate (SDS) to its aqueous solution. In these experiments, the disulfide bridges remained intact. The unfolding process has been found to be reversible both for fluoroalcohols and for SDS unfolding. The reversibility of the unfolding-refolding process of cardiotoxin in aqueous mixtures of fluoroalcohols was dependent on the volume per volume ratio of alcohol to water. SDS did not unfold the secondary structures of cardiotoxin whereas its tertiary structure was affected. If the SDS concentration in aqueous solution exceeded the critical micelle concentration value of SDS, a quasi-refolded state of cardiotoxin was observed. The mechanism of unfolding-refolding is discussed in terms of molecular interactions which might govern the protein conformation in solution.     

Effects of Ca2+ and membrane surface charge on the direct lytic activity of cobra cardiotoxin - a membrane spin assay.

Direct membrane lytic activity of cobra (Naja Naja atra) cardiotoxin was studied by measuring the release of trapped spin label (2,2,6,6-tetramethyl-piperidinyloxyl)-choline chloride from liposomes. Results suggest that (a) presence of membrane negative charge and Ca²⁺ are required for optimal direct lytic activity of cardiotoxin, (b) direct Ca²⁺ binding to cardiotoxin may affect the equilibrium between its membrane active and inactive form.     

The 1H nuclear-magnetic-resonance spectra of Neurotoxin I and cardiotoxin Vii4 from Naja mossambica mossambica.

Two toxins from the venom of Naja mossambica mossambica, neurotoxin I and cardiotoxin VII4, were investigated in aqueous solution by high-resolution 1H nuclear magnetic resonance (NMR) techniques at 360 MHz. The spectral characterization of the proteins included determination of the number of slowly exchanging amide protons which can be observed in 2H2O solution, measurement of the amide proton chemical shifts and exchange rates, characterization of the aromatic spin systems and the internal mobilities of aromatic rings, and studies of the pH dependence of the NMR spectra. For numerous resonances of labile and non-labile protons quite outstanding pH titration shifts were observed. It is suggested that these NMR parameters provide a useful basis for comparative structural studies of different proteins in the large group of homologous snake toxins. As a first application the NMR data presently available in the literature on neurotoxin II from Naja naja oxiana, toxin alpha from Naja nigricollis and erabutoxin a and b from Laticauda semifasciata have been used to compare these three proteins with neurotoxin I from Naja mossambica mossambica. This preliminary comparative study provides evidence that the same type of spatial structure prevails for these four homologous neurotoxins and that the folding of the backbone corresponds quite closely to that observed in the crystal structure of erabutoxin b. A second application is the comparison of cardiotoxin VII4 from Naja mossambica mossambica with the neurotoxins. The experimental data indicate that the folding of the polypeptide backbone is closely similar, but that the cardiotoxin molecule is markedly more flexible than the neurotoxins.     

Far-u.v. CD-spectroscopy and immunological properties of synthetic sequential peptides derived from cardiotoxin VII1 of Naja nivea venom: an amphipathic alpha-helix formed by sequence 15-25 of a beta-protein

1. Immunological properties of cardiotoxin V(II)1 of Naja nivea were investigated. 2. Polyvalent antiserum raised against the cardiotoxin was tested for its interaction with synthetic peptides of overlapping sequence in order to locate possible sequential epitopes. 3. The conformation of each synthetic peptide in various solvents was determined by circular dichroism spectroscopy for relating immunological to structural properties. 4. It was found that sequential epitopes are absent in this cardiotoxin, but that region 15-25, although part of a beta-structured region, could be a possible T-cell epitope through the formation of an amphipathic helix.     

Snake venom cardiotoxins and bee venom melittin activate phospholipase C activity in primary cultures of skeletal muscle

The effects of cardiotoxin fractions from Naja naja kaouthia and Naja naja atra snake venoms and synthetic melittin peptide were examined on lipolytic activity in red blood cells and primary skeletal muscle cultures. Both native cardiotoxin fractions caused considerable production of free fatty acids in red blood cells. This production was abolished when the fractions were first treated with p-bromophenacyl bromide to reduce the venom phospholipase A2 activity contamination. In equine and human primary cultures of skeletal muscle, the N. n. kaouthia cardiotoxin (10 microM) and melittin (2 microM) caused a breakdown of phospholipids and production of free fatty acids and diacylglycerol in the absence of lysophospholipid formation. Additionally, melittin at higher concentrations (10 microM) caused triglyceride breakdown. These studies do not support the suggestion that snake venom cardiotoxins and melittin selectively activate endogenous phospholipase A2 activity. Instead, the toxins primarily activate endogenous phospholipase C activity and, in the case of melittin at high concentrations, triglyceride lipase activity.     

Chemical modification of cytochrome P-450 LM4. Identification of functionally linked tyrosine residues

Cytochrome P-450 LM4 (RH, reduced flavoprotein:oxygen oxidoreductase (RH-hydroxylating), EC 1.14.14.1) from rabbit liver microsomes was chemically modified with tetranitromethane. Nitration of two tyrosine residues inhibits the p-nitrophenetole O-deethylase activity of the enzyme by about 80%. Sequencing the 3-nitrotyrosine-containing peptides after HPLC tryptic peptide mapping reveals that mainly Tyr-243 and Tyr-271 are nitrated, whereas Tyr-71, Tyr-188 and Tyr-365 are modified to a lower extent. Nitration of tyrosine residues affects the complex formation with p-nitrophenetole, alpha-naphthoflavone and metyrapone as indicated by an increased affinity towards p-nitrophenetole and by a decreased affinity for the latter compounds. Furthermore, nitration interferes with the electron transfer from NADPH-cytochrome P-450-reductase to cytochrome P-450 LM4 resulting in a slowed down reduction reaction. The results suggest that Tyr-243 and Tyr-271 of cytochrome P-450 LM4 are functionally involved in the interaction with NADPH-cytochrome P-450 reductase.     

The oxidation of methionine and its effect of the properties of cardiotoxin VII1 from Naja melanoleuca venom.

Methionine residues 24 and 26 of cardiotoxin VII1 from Naja melanoleuca were oxidised to sulphoxides using N-chlorosuccinimide at pH 8.5. The number of equivalents of oxidant required for complete oxidation suggested that the methionine side-chains existed in a relatively "exposed" conformational state in cardiotoxin. The oxidised cardiotoxin was devoid of lethality. It was also non-haemolytic, both on its own and in the presence of phospholipase A2. However, it was still able to precipitate with anti-cardiotoxin antibody. CD studies indicated that the polypeptide backbone conformation was intact in the oxidised cardiotoxin but some perturbation of tyrosine residues was evident. The possibility of a direct or indirect involvement of the methionine residues in the biological activity of the cardiotoxin is discussed.     

Transition of environmental polarity of tyrosine-76 of Trimeresurus flavoviridis phospholipase A2 upon active site ligand binding

Modification of Trimeresurus flavoviridis phospholipase A2 with a 5-fold molar excess of tetranitromethane produced 40% active mononitrotyrosyl phospholipase A2 in which Tyr-76 was specifically nitrated. This is in contrast to the case of mammalian pancreatic phospholipases A2 where Tyr-70 but not Tyr-76 was nitrated. When Ca2+ was bound to T. flavoviridis mononitrotyrosyl phospholipase A2, nitrated tyrosine (Tyr(NO2))-76 moved from a less polar site to a polar site with the decrease of the pKa value of its hydroxyl group. Nitration of Tyr-76 did not influence the binding affinity to Ca2+. Addition of laurylphosphorylcholine to mononitrotyrosyl phospholipase A2 in the presence of Ca2+ caused the movement of Tyr(NO2)-76 from a polar environment to a less polar environment with the rise in the pKa value. Tyrosine-76 is located in the site whose environmental polarity is affected by the binding of the ligands to the active site. As Tyr-76 is located in the site not proximal to the active site, it could be assumed that the conformational change induced by the binding of the ligands extends to the region remote from the active site in T. flavoviridis phospholipase A2. This might provide evidence of long-range diffusional coupling between remote sites in the noncooperative globular protein.     

Photochemically induced nuclear polarization study of exposed tyrosines, tryptophans, and histidines in postsynaptic neurotoxins and in membranotoxins of elapid and hydrophid snake venoms

The accessibility of surface tyrosines, histidines, and tryptophans in snake venom neurotoxins (short and long) and in membranotoxins to excited triplet 10-(carboxyethyl)-flavin was studied by photochemically induced dynamic nuclear polarization at 270 MHz. Trp-29 is accessible in the short neurotoxins--erabutoxins a, b, and c and cobrotoxin--and also in the long neurotoxins--alpha-cobratoxin and alpha-bungarotoxin. Tyr-25 is practically inaccessible in all neurotoxins. Tyr-39 in cobrotoxin and Tyr-55 in alpha-bungarotoxin are accessible. His-6 (revised sequence) is inaccessible in the erabutoxins while His-26 is only very weakly accessible. His-22 of alpha-cobratoxin is inaccessible as are His-4 and -68 in alpha-bungarotoxin and His-4 of cobrotoxin. His-33 of cobrotoxin is accessible. The rigidity order alpha-bungarotoxin greater than or equal to alpha-cobratoxin greater than or equal to erabutoxins, with respect to the unfolding effect of 7 M urea, was deduced in this study. In the membranotoxins studied (cardiotoxin and its analogues I, II, and IV as well as cytotoxin I and II), the two tyrosines Tyr-25 and Tyr-58 are only weakly accessible. Tyr-14 is completely accessible and so is in all probability Tyr-29. These studies allow deductions to be made about the accessibilities in analogous systems. Thus, the accessibility of His-33 and the inaccessibility of His-4 in cobrotoxin can be used to deduce the conformations of these residues in a large group of neurotoxins including the alpha-toxin of Naja nigricollis, neurotoxin II of Naja naja oxiana, and neurotoxins I and III of Naja mossambica mossambica.(ABSTRACT TRUNCATED AT 250 WORDS)     

Sequence characterization of venom toxins from Thailand cobra

Several toxins with distinct pharmacological properties were isolated from the venom of Thailand cobra (Naja naja siamensis) by cation-exchange chromatography. Two neurotoxins and one basic toxin with cardiotoxic activity were further purified and sequenced. The neurotoxins characterized were closely similar to the previously reported long- and short-chain neutrotoxins. The complete sequences of one minor neurotoxin and one cardiotoxin analogue were determined with the automatic protein sequencer in non-stop single runs of Edman degradation coupled with C-terminal sequence determination with carboxypeptidase digestion. The minor neurotoxin consists of 62 amino-acid residues with 8 cysteine residues and is found to be almost identical to cobrotoxin, a major toxic component of Formosa cobra (Naja naja atra). The sequence comparison of the 60-residue cardiotoxin with other reported cytotoxins of snake venoms indicates that 8 cysteine residues at the positions 3, 14, 21, 38, 42, 53, 54, and 59 are invariant among all sequences, with only two conservative changes at other positions along the sequence. The upshot of this report exemplified the facile sequence analysis of venom toxins by the application of pulsed-liquid phase protein sequencer and also revealed new analogues of a minor neurotoxin and one major cardiotoxin reported previously on the same species of Thailand cobra.     

The role of Galleria mellonella hemocytes in melanin formation

The strong paralytic and lethal action of the venom of the cobra Naja mossambica mossambica on locusts is mainly due to its cardiotoxic components. When cardiotoxin was applied to the locust extensor tibiae nerve-muscle preparation, it caused a gradual and irreversible decrease in the membrane resting potential of the muscle fibers. The time course of this potential drop was dose dependent. The electrical responses of the muscle fiber, caused by stimulation of the motor nerve, progressively decreased due to the depolarization of the membrane. The membrane-depolarizing action of cardiotoxin could be prevented by high calcium (10 mm) and by lanthanum (1 mm) in the bathing solution. In nerve-muscle preparations obtained from cardiotoxin-paralyzed locusts, normal resting and action potentials were recorded for at least 2 hr after cardiotoxin was injected into the animals.     

Studies on the status of tyrosyl residues in phospholipases A2 fromNaja naja atra andNaja nigricollis snake venoms

Two phospholipases A₂ (PLA₂) fromNaja naja atra andNaja nigricollis snake venoms were subjected to tyrosine modification withp-nitrobenzenesulfonyl fluoride (NBSF) atpH 8.0. Three major NBS derivatives from each PLA₂ were separated by high-performance liquid chromatography. The results of amino acid analysis showed that only two Tyr residues out of nine were modified, and the modified residues were identified to be Tyr-3 and Tyr-63 (or Tyr-62) in the sequence. Spectrophotometric titration indicated that the phenolic group of Tyr-3 and Tyr-63 (or Tyr-62) had apK of 10.1 and 11.0, respectively. The reactivity of Tyr-3 toward NBSF was not affected in the presence or absence of Ca ²⁺; however, the reactivity of Tyr-63 (or Tyr-62) toward NBSF was greatly enhanced by Ca²⁺. Modification of Tyr-63 (or Tyr-62) resulted in a marked decrease in both lethality and enzymatic activity. Conversely, modification of Tyr-3 inN. naja atra PLA₂ could cause more than a sixfold increase in lethal potency, in sharp contrast to the loss of enzymatic activity.Tyrosine-63-modifiedN. naja atra PLA₂ exhibited the same Ca²⁺-induced difference spectra as that of native PLA₂, indicating that the Ca²⁺-binding ability of Tyr-63-modifiedN. naja atra PLA₂ was not impaired. However, Tyr-3-modified PLA₂ and all Tyr-modifiedN. nigricollis CMS-9 were not perturbed by Ca²⁺, revealing that the Ca²⁺-binding ability have been lost after tyrosine modification. These results suggest that Tyr-62 inN. nigricollis CMS-9 and Tyr-3 in both enzymes are involved in Ca²⁺ binding. AtpH 8.0, both native PLA₂ enzymes enhance the emission intensity of 8-anilinonaphthalene sulfonate (ANS) dramatically, while all of the Tyr-modified derivatives did not enhance the emission intensity at all either in the presence or absence of Ca²⁺, suggesting that the hydrophobic pocket that interacts with ANS might be the substrate binding site, in which Tyr-3 and Tyr-63 (or Tyr-62) are involved.     

Lipid-protein interactions. A comparative study of the binding of cardiotoxins and neurotoxins to phospholipid vesicles

It has recently been shown that cardiotoxin II from Naja mossambica mossambica specifically interacts with negatively charged phospholipids (Dufourcq, J. and Faucon, J.F. (1978) Biochemistry 17, 1170–1176). In order to investigate whether or not short neurotoxins give rise to similar interactions, four techniques have been used, namely intrinsic fluorescence, fluorescence polarization of 1,6-diphenylhexatriene, turbidity measurements and release of 6-carboxyfluorescein trapped inside single shelled vesicles.Neurotoxin III from Naja mossambica mossambica and neurotoxin I from the venom of the scorpion Androctonus australis Hector, specifically interact with negatively charged phospholipids leading to changes in tryptophan fluorescence and to a decrease of the fluidity of the bilayer. Cardiotoxin II from the same snake venom gives similar results. On the other hand, it seems that either a very weak or no interaction at all occurs in the case of neurotoxin I from the same Naja venom.There are important differences in the behaviour of cardiotoxin and neurotoxins: (i) neurotoxins lead to only weak release of 6-carboxyfluorescein from lipid vesicles, whereas cardiotoxin II induces fast and quantitative escape of the dye and then a general breakdown of the vesicular structure; (ii) binding of neurotoxins can be easily reversed by 100–200 mM NaCl or less than 1 mM Ca²⁺ and so it is essentially electrostatic, whereas binding of cardiotoxin II seems to involve some hydrophobic contribution.The short neurotoxins and cardiotoxins from snake venom having a great homology in sequence, their differences on binding properties are discussed in terms of changes in a particular area of the sequence.     

Kinetics of erythrocyte lysis by snake venom cardiotoxins

Hemolysis of guinea pig erythrocytes by snake venom cardiotoxins was investigated with a semi-automatic method based on light-scattering changes of erythrocyte suspensions at 700 nm which are directly related to hemoglobin release. Small amounts of phospholipase-free cardiotoxin (<100 μg) could be conveniently and rapidly assayed with the high reproducibility in a recording spectrophotometer, and reliable kinetic data were accumulated.Cardiotoxins from two different genera (Hemachatus haemachates and Naja mossambica mossambica) displayed virtually identical hemolytic properties. Hemolysis increased linearly with time, in contrast with a sigmoidal pattern when phospholipase was present as an impurity. Low concentrations of Ca²⁺ (<1 mM) stimulated cardiotoxin action. A limiting plateau rate of hemolysis reached during concentration dependence experiments in which the level of either cardiotoxin or of erythrocytes was varied, suggested that the interaction of cardiotoxin with erythrocyte membranes is a saturation phenomenon only at a high ratio of cardiotoxin: erythrocytes. No hemolysis was observed with an homologous neurotoxin of S-methylated cardiotoxin, providing evidence for specificity. The linear Arrhenius plots obtained for the temperature dependence of cardiotoxin-induced hemolysis strengthened the conclusion that its action involves more than a detergent-like effect on membrane phospholipids.     

Comparison of phospholipase activity with direct and indirect lytic effects of animal venoms upon human red cells.

1. The venoms of 22 species of arthropods, saurians, elapids and crotalids were studied concerning the phospholipase activity and the presence of a direct and an indirect lytic effect upon human red cells. 2. The venoms from the spiders Latrodectus and "tarantula", and the venoms from the scorpions of the genus Centruroides are not haemolytic and do not have phospholipase activity. 3. Only the venoms of Apis mellifera and Naja naja siamensis have shown direct lytic effect. 4. All other venoms studied have an indirect haemolytic effect associated to a phospholipase activity, but there is indication that other agents might be implicated in the haemolytic processes.     

Studies on the status of tyrosyl residues in phospholipases A2 fromNaja naja atra andNaja nigricollis snake venoms

Two phospholipases A₂ (PLA₂) fromNaja naja atra andNaja nigricollis snake venoms were subjected to tyrosine modification withp-nitrobenzenesulfonyl fluoride (NBSF) atpH 8.0. Three major NBS derivatives from each PLA₂ were separated by high-performance liquid chromatography. The results of amino acid analysis showed that only two Tyr residues out of nine were modified, and the modified residues were identified to be Tyr-3 and Tyr-63 (or Tyr-62) in the sequence. Spectrophotometric titration indicated that the phenolic group of Tyr-3 and Tyr-63 (or Tyr-62) had apK of 10.1 and 11.0, respectively. The reactivity of Tyr-3 toward NBSF was not affected in the presence or absence of Ca ²⁺; however, the reactivity of Tyr-63 (or Tyr-62) toward NBSF was greatly enhanced by Ca²⁺. Modification of Tyr-63 (or Tyr-62) resulted in a marked decrease in both lethality and enzymatic activity. Conversely, modification of Tyr-3 inN. naja atra PLA₂ could cause more than a sixfold increase in lethal potency, in sharp contrast to the loss of enzymatic activity. Tyrosine-63-modifiedN. naja atra PLA₂ exhibited the same Ca²⁺-induced difference spectra as that of native PLA₂, indicating that the Ca²⁺-binding ability of Tyr-63-modifiedN. naja atra PLA₂ was not impaired. However, Tyr-3-modified PLA₂ and all Tyr-modifiedN. nigricollis CMS-9 were not perturbed by Ca²⁺, revealing that the Ca²⁺-binding ability have been lost after tyrosine modification. These results suggest that Tyr-62 inN. nigricollis CMS-9 and Tyr-3 in both enzymes are involved in Ca²⁺ binding. AtpH 8.0, both native PLA₂ enzymes enhance the emission intensity of 8-anilinonaphthalene sulfonate (ANS) dramatically, while all of the Tyr-modified derivatives did not enhance the emission intensity at all either in the presence or absence of Ca²⁺, suggesting that the hydrophobic pocket that interacts with ANS might be the substrate binding site, in which Tyr-3 and Tyr-63 (or Tyr-62) are involved.     

Chemical modification of human alpha 1-proteinase inhibitor by tetranitromethane. Structure-function relationship.

Nitration of tyrosine residues of alpha 1-proteinase inhibitor (alpha 1-PI) by tetranitromethane yielded a product that maintained its inhibitory activity against trypsin but lost most of its inhibitory activity against elastase. Chemical analysis of the product showed that four out of the six tyrosine residues in alpha 1-PI had been nitrated to various degrees: Tyr-38 and Tyr-297 were not nitrated, whereas Tyr-138, Tyr-160, Tyr-187 and Tyr-244 were nitrated to extents in the range 40-80%. We interpreted these data to mean that modification of these tyrosine residues decreased the association constant between alpha 1-PI and the proteinases and that the decrease differs from one proteinase to the other. When either alpha 1-PI-trypsin or alpha 1-PI-elastase complex was nitrated, nitration took place only to a very slight extent at these latter four tyrosine residues. On the other hand, Tyr-38 and Tyr-297 underwent nitration to about 20%. We concluded that Tyr-138, Tyr-160, Tyr-187 and Tyr-244 were located on the surface of alpha 1-PI that interacts with either trypsin or elastase in the formation of complexes, and were therefore protected from nitration.