Membrane-bound enzymes: III. Protease activity in leucocytes in relation to erythrocyte membranes

Protease activity was detected in membranes of human and bovine erythrocytes prepared by the conventional procedures which include washing and removal of the “buffy layer”. The enzyme was extracted by 0.75 M KCNS or (NH₄)₂SO₄ and was activated by 0.4 to 0.5 M of the same salts. Colored, particulate hide powder-azure, membrane fractions and soluble proteins such as hemoglobin, casein or albumin were susceptible to hydrolysis by the membraneous protease.Partial purification of the enzyme was accomplished through disc-gel electrophoresis on polyacrylamide in the presence of 0.25% positively charged detergents like cetyltrimethylammonium bromide. An alkaline protease (pH 7.4) with properties similar to those of the erythrocyte enzyme was found in leucocytes. The similarity between the properties of the leucocytic and erythrocytic proteases and the correlation of the activity in erythrocyte membranes with the content of white cells in these preparations, suggest that enzymatic activities in the contaminating leucocytes are responsible for the activity of membraneous proteases in erythrocytes.     

An acid protease in human erythrocytes and its localization in the inner membrane.

The isolation of erythrocytes of high purity from human blood was achieved by a combination of the two well established methods cells in erythrocyte preparations of different purities was studied. The acid protease activity was recovered to a level comparable with the recovery of erythrocytes, while the neutral protease activity as detected by the release of acid-soluble peptides from hemoglobin or casein disappeared in proportion to the removal of white blood cells. An acid protease was solubilized from the membranes of the purified erythrocytes by the extraction with 1-butanol. The enzyme was active in a pH range from 2 to 4, and sensitive to pepstatin. It was named pH-3 protease after its pH optimum. Sealed ghosts with right-side-out membranes and inside-out vesicles with reverted membranes were prepared from the purified erythrocytes and compared with respect to pH-3 protease activity for its latency as well as its inactivation by tryptic digestion. The results obtained indicate that pH-3 protase is localized on the inner surface of erythrocyte membranes. The self-digestion experiments at pH 4 using the sealed ghosts showed higher availability to pH-3 protease of spectrin and IVa protein than the other membrane proteins, also suggesting the localization of an acid protease in the inner membranes of erythrocytes.     

Studies on extracellular proteases of Streptococcus sanguis. Purification and characterization of a human IgA1 specific protease.

Extracellular caseinolytic activity was found in the culture fluid of Streptococcus sanguis ATCC 10556 grown in a dialyzed culture medium. This activity was due to multiple proteases that differed in their elution from hydroxyapatite, sensitivity to enzyme inhibitors, specificity and optimum pH. IgA protease, which splits human immunoglobulin A1 into intact Fc and Fab could be effectively separated from these relatively non-specific proteases and purified to apparent homogeneity in 20% yield by a five-step procedure. Although the bulk of the dextran sucrase activity was separated from the IgA protease, a small amount of sucrase activity remained with the final IgA protease preparation. In polyacrylamide gel electrophoresis at pH 9.5 both activities were located in the single protein band detected in this preparation. A quantitative method for the assay of IgA protease was developed, based on radial immunodiffusion to quantitate the Fab produced. This was used to follow the specific activity and yield during purification, and to characterize some of the catalytic properties of the enzyme. At an enzyme/substrate ratio of 1: 400 (w/w) the protease could effect 50% proteolysis of IgA in overnight incubation at 37 degrees C. The optimum activity was at pH 8.0, and 50% inhibition was achieved at 4 . 10(-4) M o-phenanthroline or 8 . 10(-4) M ethylene diamine tetraacetate. Concentrations of diisopropyl phosphofluoridate, phenylmethyl-sulfonyl fluoride, iodoacetate and p-chloromercuribenzoate up to 10(-2) M were without effect on the IgA protease activity. Full reactivation of the chelator inhibited enzyme could be achieved by the addition of Mg2+, Mn2+ or Ca2+.     

Multiple proteases from Streptomyces moderatus. II. Physicochemical and enzymatic properties of the extracellular proteases

The physicochemical and enzymatic properties of five different extracellular proteases of Streptomyces moderatus were studied. The first protease was found to be a metal chelator sensitive protease with a Mr of 21,000 +/- 1000 a and a pI of 4.6. The second enzyme was an anionic trypsin-like protease (Mr 19,000 +/- 1000; pI 3.8) with a Km value of 4.76 X 10(-4) M on N-benzoyl-L-arginine-p-nitroanilide. A Km value of 1.52 X 10(-4) M was obtained when N-benzoyl-L-arginine ethyl ester was used as the substrate. The other three enzymes were found to be serine alkaline proteases with Mr's of 22,000, 29,000, and 23,000 +/- 1000 and with respective pI's of 7.8, 8.4, and 9.2. All the proteases showed optimum activity in the alkaline pH range. One of the three proteases was found to possess chymotrypsin and elastase-like properties. All five proteases were found to be unstable at temperatures above 60 degrees C. Except the trypsin-like protease, which was stable only in acidic pH, all other enzymes were found to be stable over a wide range of pH.     

Proteolytic enzymes in human fetal membranes and amniotic fluid. A comparison of normal and premature ruptured membranes

The amniotic and chorionic membranes obtained at term and term amniotic fluid contain a soluble protease activity which cleaves [14C]-labeled globin at acid pH. In contrast, a salt extract of the pellet fraction obtained from the fetal membranes displays only negligible protease activities at the pH range of 4-8. Specific activities of the proteases in the soluble and salt-extractable fractions of fetal membranes which were intact before onset of labor were not significantly different from the respective activities in cases of premature rupture of fetal membranes (PROM). However, the protease activity of the amniotic fluid was found to increase with advancing gestational age and to reach maximal activity at term. A heat-sensitive and nondializable protease inhibitory activity was found in term amniotic fluid. This inhibitory activity acted on the cytosolic protease of amniotic membranes from control and PROM cases, but not on the soluble protease of chorionic membranes, and had a similar potency in fluids from PROM cases or fluids collected at term. These results do not support a role for fetal membrane proteases, amniotic fluid proteases, or amniotic fluid protease inhibitory activities in the etiology of PROM. However, the observed changes in amniotic fluid protease activity with fetal age suggest a physiological role for the enzyme in normal fetal development.     

Phorbol diester treatment promotes enhanced adenylate cyclase activity in frog erythrocytes

Incubation of intact frog erythrocytes with 12-O-tetradecanoyl phorbol-13-acetate (TPA), a tumor-promoting phorbol diester which activates protein kinase C, results in an approximate two- to threefold increase in subsequently tested beta-adrenergic agonist-stimulated adenylate cyclase activity. This increase is due to an elevation in the Vmax of the enzyme rather than to a change in affinity for the agonist. TPA treatment of frog erythrocytes does not alter the affinity (KD) or the binding capacity (Bmax) for the beta-adrenergic antagonist [125I]cyanopindolol. In addition, agonist/[125I]cyanopindolol competition curves are not affected by TPA pretreatment nor is their sensitivity to guanine nucleotides. Incubation of frog erythrocyte membranes alone with TPA does not promote sensitization or activation of adenylate cyclase activity. Pretreatment of intact frog erythrocytes with TPA also produces approximately two- to threefold increases in basal, guanine nucleotide-, prostaglandin E1-, forskolin-, NaF-, and MnCl2-stimulated adenylate cyclase activities in frog erythrocyte membranes. This enhancement of adenylate cyclase activity by TPA is induced rapidly (t1/2 approximately equal to 5 min) and with an EC50 of about 10(-7) to 10(-6) M. Other tumor-promoting phorbol diesters or phorbol diester-like compounds including 4 beta-phorbol 12,13-dibutyrate, 4 beta-phorbol 12,13-didecanoate, and mezerein are effective in promoting enhanced adenylate cyclase activity. In contrast, phorbols such as 4 beta-phorbol, 4 alpha-phorbol 12,13-didecanoate, and 4-O-methylphorbol 12-myristate 13-acetate, which are inactive in tumor promotion and which do not activate protein kinase C, do not affect frog erythrocyte adenylate cyclase activity. These data are suggestive of a protein kinase C-mediated phosphorylation of one of the adenylate cyclase components that is distal to the receptor, i.e., the nucleotide regulatory and/or catalytic components.     

Regulatory differences between Ca(2+)-ATPase in plasma membranes from chicken (nucleated) and pig (anucleated) erythrocytes

Kinetic and regulatory properties of the plasma membrane Ca(2+)-ATPase activity from chicken (nucleated) erythrocytes were studied and compared to those from pig (anucleated) erythrocytes. In the absence of known activators: (1) Ca(2+) affinity for the Ca(2+)-ATPase activity from nucleated erythrocytes was 12-fold higher than that from pig erythrocytes, and thus the enzyme is sensitive to physiological Ca(2+) concentrations; (2) the enzyme from chicken erythrocytes showed two apparent Km values for ATP, as compared to one apparent Km value displayed by pig erythrocytes; (3) Ca(2+)-ATPase inserted in chicken erythrocyte membranes showed a low sensitivity to activation by phosphatidylinositol-4-phosphate; (4) when p-NPP was used as substrate, the activity of chicken erythrocytes was high, similar to that attained by pig erythrocytes, but barely sensitive to activation by dimethylsulfoxide and calmodulin. ATP hydrolysis was 10-fold lower than that displayed by pig erythrocytes and the maximal velocity was activated three-fold by calmodulin. The enzyme was insensitive to alkaline phosphatase treatment and showed a single phosphorylation band in electrophoresis, ruling out the possibility of previous modulation by endogenous kinases and/or by partial proteolysis. The differences may be attributed to some endogenous modulator, to distinct isoforms, or to a difference in the E(1)/E(2) states of the enzyme.     

Falstatin, a cysteine protease inhibitor of Plasmodium falciparum, facilitates erythrocyte invasion

Erythrocytic malaria parasites utilize proteases for a number of cellular processes, including hydrolysis of hemoglobin, rupture of erythrocytes by mature schizonts, and subsequent invasion of erythrocytes by free merozoites. However, mechanisms used by malaria parasites to control protease activity have not been established. We report here the identification of an endogenous cysteine protease inhibitor of Plasmodium falciparum, falstatin, based on modest homology with the Trypanosoma cruzi cysteine protease inhibitor chagasin. Falstatin, expressed in Escherichia coli, was a potent reversible inhibitor of the P. falciparum cysteine proteases falcipain-2 and falcipain-3, as well as other parasite- and nonparasite-derived cysteine proteases, but it was a relatively weak inhibitor of the P. falciparum cysteine proteases falcipain-1 and dipeptidyl aminopeptidase 1. Falstatin is present in schizonts, merozoites, and rings, but not in trophozoites, the stage at which the cysteine protease activity of P. falciparum is maximal. Falstatin localizes to the periphery of rings and early schizonts, is diffusely expressed in late schizonts and merozoites, and is released upon the rupture of mature schizonts. Treatment of late schizionts with antibodies that blocked the inhibitory activity of falstatin against native and recombinant falcipain-2 and falcipain-3 dose-dependently decreased the subsequent invasion of erythrocytes by merozoites. These results suggest that P. falciparum requires expression of falstatin to limit proteolysis by certain host or parasite cysteine proteases during erythrocyte invasion. This mechanism of regulation of proteolysis suggests new strategies for the development of antimalarial agents that specifically disrupt erythrocyte invasion.     

A serum factor stimulating cathepsin D-release from erythrocytes or ghosts

A serum factor preparation extensively purified from bovine serum stimulated cathepsin D-release from the rat blood cells in a concentration-dependent fashion within a range of physiological concentrations of the factor. Among the blood cells only the erythrocytes (or ghosts) were responsive to the factor, and the leucocytes and lymphocytes were unresponsive. The effects of Ca2+-concentrations, SH- blocking reagents, protease-inhibitors, calmodulin-inhibitors, calmodulin or EGTA-pretreatment of the ghosts on cathepsin D-release from the erythrocytes of ghosts in the presence or the absence of serum factor were investigated. The results suggested that the serum factor may first activate the Ca2+-calmodulin system via the mobilization of "Ca2+-pool" and then the calmodulin-dependent SH-protease in the erythrocyte plasma membranes. The activated protease in turn may break the linkage between cathepsin D and the plasma membranes, liberating cathepsin D activity into the incubation medium. The name of "calciferin" was proposed for the serum factor.     

Erythrocyte cytosolic free Ca2+ and plasma membrane Ca2+-ATPase activity in cystic fibrosis

The properties of the Ca2+, Mg2+-ATPase of erythrocyte membranes from patients with cystic fibrosis (CF) were extensively compared to that of healthy controls. Following removal of an endogenous membrane inhibitor of the ATPase, activation of the enzyme by Ca2+, calmodulin, limited tryptic digestion or oleic acid, as well as inhibition by trifluoperazine, were studied. The only properties found to be significantly different (CF cells vs controls) were calmodulin-stimulated peak activity (90 vs 101, P less than 0.02) and trypsin-activated peak activity (92 vs 102, P less than 0.02). No significant difference could be measured in the steady-state Ca2+-dependent phosphorylation of CF and control erythrocyte membranes indicating similar numbers of enzyme molecules per cell. The functional state of Ca2+ homeostasis in intact erythrocytes was investigated by measuring the resting cytosolic free Ca2+ levels using quin-2. Both CF and control erythrocytes maintained cytosolic free Ca2+ between 20 to 30 nM. Addition of 50 uM trifluoperazine resulted in an increase in erythrocyte cytosolic free Ca2+ to about 50 nM in both CF and control cells. Estimates of erythrocyte membrane permeability using the steady-state uptake of 45Ca into intact erythrocytes revealed no differences between CF and control cells. These results confirm that there is a small decrease in the calmodulin-stimulated activity of the erythrocyte Ca2+, Mg2+-ATPase in CF. However, this deficit is apparently not large enough to impair the ability of the CF erythrocyte to maintain normal resting levels of cytosolic free Ca2+.     

Activation of the Ca2+-ATPase of human erythrocyte membrane by an endogenous Ca2+-dependent neutral protease

Limited proteolysis of the plasma membrane calcium transport ATPase (Ca2+-ATPase) from human erythrocytes by trypsin produces a calmodulin-like activation of its ATP hydrolytic activity and abolishes its calmodulin sensitivity. We now demonstrate a similar kind of activation of the human erythrocyte membrane Ca2+-ATPase by calpain (calcium-dependent neutral protease) isolated from the human red cell cytosol. Upon incubation of red blood cell membranes with purified calpain in the presence of Ca2+ the membrane-bound Ca2+-ATPase activity was increased and its sensitivity to calmodulin was lost. In contrast to the action of other proteases tested, proteolysis by calpain favors activation over inactivation of the Ca2+-ATPase activity, except at calpain concentrations more than 2 orders of magnitude higher. Exogenous calmodulin protects the Ca2+-ATPase against calpain-mediated activation at concentrations which also activate the Ca2+-ATPase activity. Calcium-dependent proteolytic modification of the Ca2+-ATPase could provide a mechanism for the irreversible activation of the membrane-bound enzyme.     

AMP-deaminase activity of circulating leukocytes in the human

It is shown that the AMP-deaminase activity in leucocytes of the human peripheric blood in contrast with the enzyme from erythrocytes manifests its activity only if it is isolated in the presence of K+ or Na+ ions. Pi and GTP being inhibitors of the enzyme in different tissues including erythrocytes do not alter the AMP-deaminase activity in leucocytes. 3,3',5-triiodothyracetic acid markedly decreasing the AMP-deaminase activity of leucocytes does not affect the enzyme activity in the hemolyzate of erythrocytes. The results obtained have shown that the AMP-deaminase activity in leucocytes of the human peripheric blood possesses some regulatory properties differing from those of the enzyme in erythrocytes.     

Interaction of cysteine proteases with calciotropic hormone receptors

The effect of two cysteine proteases: papain and a cathepsin L-like enzyme purified from the oesophagus of Nephrops norvegicus (NCP) was studied on the specific binding of calcitonin (CT) and calcitonin gene related peptide (CGRP) to rat kidney and liver membranes, respectively. In addition, the response of adenylyl cyclase to increasing concentrations of these two enzymes was investigated. Each protease inhibited the initial CGRP and CT binding to rat liver and kidney membranes, respectively, in a manner not significantly different from that obtained in the presence of the unlabeled standard. The adenylyl cyclase activity in rat liver membranes was increased by the addition of each enzyme. The response was higher with papain that induced a fivefold increase of enzyme activity at a 4-microg/ml enzyme concentration. In rat kidney membranes, the magnitude of the response was identical with both enzymes. In contrast with NCP, papain induced a biphasic response. Leupeptin and E(64), two specific inhibitors of cysteine proteases, reversed the observed effects. Trypsin induced an inhibition of the liver membrane adenylyl cyclase activity and an activation in rat kidney membranes at low protease concentration. Thus, cysteine proteases are able to act, in vitro, at the receptor level in target organs specific for calciotropic hormones.     

Fusion of rat erythrocytes by membrane-mobility agent A2C depends on membrane proteolysis by a cytoplasmic calpain

The membrane-mobility agent 2-(2-methoxyethoxy)ethyl-cis-8-(2-octylcyclopropyl)octanoate (A2C) promotes fusion of rat, but not of human, erythrocytes. The difference in fusibility was shown to be correlated with membrane proteolysis, a process induced by Ca2+ in the rat erythrocytes or hemolysate-loaded ghosts, but not in the human cell. Membrane proteolysis is necessary but not sufficient for fusion. Fusion requires both Ca2+ and A2C [Kosower, N. S., Glaser, T. and Kosower, E. M. (1983) Proc. Natl Acad. Sci USA 80, 7542-7546]. Membrane proteolysis (Ca2+-dependent) and fusion (Ca2+ and A2C-dependent) requires a Ca2+-activated cytoplasmic thiol protease, as shown by the following observations. In intact rat erythrocytes, proteolysis and fusion are prevented by thiol alkylation and by inhibitors of Ca2+-dependent thiol proteases. Inhibitors to other proteases have no effect. Erythrocyte ghosts undergo proteolysis and fusion only when loaded with non-inhibited hemolysate, irrespective of membrane status (native or alkylated membrane). A partially purified cytosolic enzyme, identified as calpain I, promotes proteolysis in rat erythrocyte ghosts. A2C induces fusion only in such calpain-treated ghosts.     

Abundance of the Ca2+-pumping ATPase in pig erythrocyte membranes.

The Ca2+-pumping ATPase (Ca2+-ATPase) was purified from human and pig erythrocyte membranes by calmodulin affinity chromatography in the presence of phosphatidylcholine. The amount of enzyme present in pig erythrocytes is at least 7 times greater than that isolated from human erythrocyte ghosts. However, the properties of the enzyme from the two species are similar in many respects.     

An investigation of 2':3'-cyclic nucleotide 3'-phosphodiesterase (EC 3.1.4.37, CNP) in peripheral blood elements and CNS myelin

Activity of 2':3'-cyclic nucleotide 3'-phosphodiesterase (CNP) of human erythrocyte membranes was determined in the presence of various brain CNP inhibitory compounds. Also, the hydrolysis of 2':3'-cAMP and 2':3'-cCMP by CNP of human platelets and lymphocytes was confirmed by thin layer chromatography and CNP activity was measured in lymphocytes, platelets, erythrocytes and CNS myelin. Human erythrocyte CNP activity was reduced 75 percent by the organomercurial p-chloromercuriphenyl sulfonate (1 X 10(-4) M), 46 percent by thimerosal (1 X 10(-4) M) and 35 percent by cupric chloride (1 X 10(-3) M). The 2'-AMP or 2'-CMP isomer was produced, exclusively, by the hydrolysis of 2':3'-cAMP or 2':3'-cCMP, respectively, by CNP of human lymphocytes and platelets and indicates a CNP-like activity is not only present in erythrocytes and the central and peripheral nervous systems, but also platelets and lymphocytes. CNP activities of human erythrocytes, human human and rat lymphocytes and human platelets were less than 4 percent of the activity of human and bovine CNS myelin.     

Studies on adenosine 3′,5′-monophosphate phosphodiesterase of human erythrocyte membranes

In this work we show the existence of cyclic AMP phosphodiesterase (EC 3.1.4.17) in human erythrocyte membranes and have clarified some properties of the enzyme. In human erythrocytes, about 23% of the total cyclic AMP phosphodiesterase activity is in a membrane-bound form. Although it could be solubilized with Triton X-100 in 5 mM Tris-HCl buffer (pH 8.0), it was not solubilized by a low or high concentration of salt. The enzyme seems to be localized in the cytoplasmic surface, since it is detected in sealed inside-out vesicles of human erythrocyte membranes, but not in intact human erythrocytes. The optimum pH was found to lie between 7.4 and 8.0, and Mg²⁺ was found to be necessary for its activity. Ca²⁺ and calmodulin could not stimulate the activity of this enzyme. Theophylline was a strong inhibitor, but cyclic GMP could not inhibit the enzymic hydrolysis of cyclic [³²P]AMP and this membrane-bound enzyme therefore seems to be specific to cyclic AMP.     

Extracellular and membrane-bound proteases from Bacillus subtilis.

Bacillus subtilis YY88 synthesizes increased amounts of extracellular and membrane-bound proteases. More than 99% of the extracellular protease activity is accounted for by an alkaline serine protease and a neutral metalloprotease. An esterase having low protease activity accounts for less than 1% of the secreted protease. These enzymes were purified to homogeneity. Molecular weights of approximately 28,500 and 39,500 were determined for the alkaline and neutral proteases, respectively. The esterase had a molecular weight of approximately 35,000. Amino-terminal amino acid sequences were determined, and the actions of a number of inhibitors were examined. Membrane vesicles contained bound forms of alkaline and neutral proteases and a group of previously undetected proteases (M proteases). Membrane-bound proteases were extracted with Triton X-100. Membrane-bound alkaline and neutral proteases were indistinguishable from the extracellular enzymes by the criteria of molecular weight, immunoprecipitation, and sensitivity to inhibitors. The M protease fraction accounted for approximately 7% of the total activity in Triton X-100 extracts of membrane vesicles. The M protease fraction was partially fractionated into four species (M1 through M4) by ion-exchange chromatography. Immunoprecipitation and sensitivity to inhibitors distinguished membrane-bound alkaline and neutral proteases from M proteases. In contrast to alkaline and neutral proteases, proteases M2 and M3 exhibited exopeptidase activity.     

Enhancement of (Ca2+ + Mg2+)-ATPase activity of human erythrocyte membranes by hemolysis in isosmotic imidazole buffer. I. General properties of variously prepared membranes and the mechanism of the isosmotic imidazole effect.

1. Membranes prepared from human erythrocytes hemolyzed in isosmotic (310 imosM) imidazole buffer, pH 7.4, show enhanced and stabilized (Ca2+ + Mg2+)-ATPase activity compared with membranes prepared from erythrocytes hemolyzed in hypotonic (20 imosM) phosphate or imidazole buffer, pH 7.4. 2. Exposure of intact erythrocytes or well-washed erythrocyte membranes to isosmotic imidazole does not cause enhanced (Ca2+ + Mg2+)-ATPase activity. 3. Exposure of erythrocyte membranes, in the presence of isosmotic imidazole, to the supernatant of erythrocyte hemolysis or to a partially purified endogenous (Ca2+ + Mg2+)-ATPase activator, promotes enhanced (Ca2+ + Mg2+)-ATPase activity. Under appropriate conditions, NaCl can be shown to substitute for imidazole. The results demonstrate that imidazole does not act directly on the erythrocyte membrane but rather by promoting interaction between an endogenous (Ca2+ + Mg2+)-ATPase activator and the erythrocyte membrane.     

Novel alkaline- and heat-stable serine proteases from alkalophilic Bacillus sp. strain GX6638.

An alkalophilic Bacillus sp., strain GX6638 (ATCC 53278), was isolated from soil and shown to produce a minimum of three alkaline proteases. The proteases were purified by ion-exchange chromatography and were distinguishable by their isoelectric point, molecular weight, and electrophoretic mobility. Two of the proteases, AS and HS, which exhibited the greatest alkaline and thermal stability, were characterized further. Protease HS had an apparent molecular weight of 36,000 and an isoelectric point of approximately 4.2, whereas protease AS had a molecular weight of 27,500 and an isoelectric point of 5.2. Both enzymes had optimal proteolytic activities over a broad pH range (pH 8 to 12) and exhibited temperature optima of 65 degrees C. Proteases HS and AS were further distinguished by their proteolytic activities, esterolytic activities, sensitivity to inhibitors, and their alkaline and thermal stability properties. Protease AS was extremely alkali stable, retaining 88% of initial activity at pH 12 over a 24-h incubation period at 25 degrees C; protease HS exhibited similar alkaline stability properties to pH 11. In addition, protease HS had exceptional thermal stability properties. At pH 9.5 (0.1 M CAPS buffer, 5 mM EDTA), the enzyme had a half-life of more than 200 min at 50 degrees C and 25 min at 60 degrees C. At pH above 9.5, protease HS readily lost enzymatic activity even in the presence of exogenously supplied Ca2+. In contrast, protease AS was more stable at pH above 9.5, and Ca2+ addition extended the half-life of the enzyme 10-fold at 60 degrees C. In contrast, protease AS was more stable at pH above 9.5, and Ca2+ addition extended the half-life of the enzyme 10-fold at 60 degrees C. The data presented here clearly indicate that these two alkaline proteases from Bacillus sp. strain GX6638 represent novel proteases that differ fundamentally from the proteases previously described for members of the genus Bacillus.