Binding activity of Streptococcus canis for albumin and other plasma proteins

All 24 cultures of Streptococcus canis examined bound 125I-labelled human albumin, IgG and fibrinogen; but neither IgA nor haptoglobin. Binding of human albumin was time-dependent, saturable and reversible by the addition of unlabelled albumin. The binding of 125I-labelled human albumin could be inhibited completely by unlabelled albumin preparations from humans, mice and dogs, and partly by bovine albumin. In contrast, binding of 125I-labelled human albumin was not inhibited by unlabelled rabbit albumin, human IgG or human fibrinogen. Data from competition experiments of two S. canis cultures with high 125I-labelled albumin-binding activities yielded KD values of 10 and 15 nmol l-1, respectively. The estimated number of binding sites per bacterial cell ranged from 30,000 to 57,000. The binding factor for albumin could be isolated from S. canis by boiling the bacteria at pH 2, and it was purified by affinity chromatography on human albumin-Sepharose. The isolated albumin-binding proteins had a molecular mass of approximately 51 kDa and inhibited binding of 125I-labelled albumin to S. canis. They formed complexes with human albumin that altered its electrophoretic mobility.     

Binding of plasma proteins to Candida species in vitro

The ability of purified human albumin, fibrinogen and transferrin to bind to Candida species was measured by immunofluorescence. The proteins all bound with high avidity to germ-tubes formed by Candida albicans, but did not bind to blastospores of C. albicans or other pathogenic Candida species, not even to parent blastospores bearing germ-tubes. The extent of binding of the proteins to C. albicans germ-tubes varied between growth media and from germ-tube to germ-tube. Strains of C. albicans that did not form germ-tubes were incapable of binding any of the proteins. There was evidence that purified fibrinogen bound to germ-tubes with higher avidity than albumin and transferrin. When germ-tubes were treated with whole human plasma or serum, indirect immunofluorescence revealed that proteins were bound all over the surface of C. albicans blastospore-germ-tube units, indicating behaviour different from that seen with the purified proteins tested alone or in mixtures. C. albicans cells grown in the presence of azole antifungal agents bound purified plasma proteins in the same way as cells untreated with the drugs. The results of this study suggest that binding of host proteins to the surface of C. albicans may not be a property related directly to virulence of the fungus in vivo.     

Sequence content of alpha-fetoprotein, albumin and fibrinogen polypeptide mRNAs in different organs, developing tissues and in liver during carcinogenesis in rats

To investigate the variable gene activities of alpha-fetoprotein, albumin and fibrinogen polypeptides as markers of 'liver specific proteins' in different developing organs or tissues, we have used specific complementary DNA probes to detect and to quantitate alpha-fetoprotein, albumin and fibrinogen polypeptide mRNA, respectively, in RNA fractions, prepared from various tissues of rats at different stages of fetal and postnatal development and from hepatomas induced by diethylnitrosamine. The results indicate that there is no consistent relationship between sequence content of alpha-fetoprotein, albumin and fibrinogen polypeptide mRNA in different developing tissues. Intestines which are like the liver also of endodermal origin do not contain alpha-fetoprotein, albumin and fibrinogen polypeptide mRNAs, while kidneys which are mesodermal in origin were found to be alpha-fetoprotein, albumin and fibrinogen polypeptide mRNA producers in neonatal life. In yolk sac, only alpha-fetoprotein and fibrinogen polypeptide mRNA could be detected. In the liver, the increased level of albumin and fibrinogen polypeptide mRNA during fetal and neonatal development is accompanied with a diminished amount of alpha-fetoprotein mRNA. The neosynthesis of alpha-fetoprotein mRNA in the liver during carcinogenesis occurred without a decreased content of albumin and fibrinogen polypeptide mRNAs. These findings suggest that complex mechanisms of gene regulation are involved in variable gene activities of alpha-fetoprotein, albumin and fibrinogen polypeptides in cells of different organs or tissues developed from a single cell.     

A fibrinolytic metalloprotease from the fruiting bodies of an edible mushroom, Armillariella mellea

A fibrinolytic metalloprotease has been purified from the fruiting bodies of the edible honey mushroom (Armillariella mellea). The enzyme has a molecular weight of 18538.1508, as measured by MALDI-TOF mass spectrometry and includes Zn2+ ion as found by ICP/MS. The N-terminal amino acid sequence, XXYNGXTXSRQTTLV, do not match any known protein or open reading frame. It hydrolyzes fibrinogen as well as fibrin, but does not show any proteolytic activity for other blood proteins such as thrombin, human albumin, bovine albumin, human IgG, hemoglobin, or urokinase. This protease hydrolyzes both A alpha and B beta subunits of human fibrinogen with equal efficiency. The enzyme activity was strongly inhibited by EDTA and 1,10-phenanthroline, indicating that the enzyme is a metalloprotease. No inhibition was found with PMSF, E-64, pepstatin, and 2-mercaptoethanol. The activity of the purified enzyme was slightly increased by Mg2+, Zn2+, and Co2+, but the enzyme was totally inhibited by Hg2+. It has broad substrate specificity for synthetic peptides, and a pH optimum at 7, suggested that the purified enzyme was a neutral protease. It was thermally stable up to 60 degrees C and the maximum fibrinolytic activity was at 55 degrees C.     

Purification and characterization of Atlantic salmon (Salmo salar) fibrinogen

This study describes a purification protocol of salmon fibrinogen that gives a consumable and highly clottable fibrinogen. Some characteristics of salmon and human fibrinogen are compared. Fibrinogen was purified from barium sulphate adsorbed plasma of Atlantic salmon, using two steps of 25% ammonium sulphate precipitation followed by ultrafiltration. The clottability of the purified salmon fibrinogen was 91%. The Aalpha chains of salmon fibrinogen were heterogeneous with a molecular mass of 90-110 kDa, compared to approximately 67 kDa of human fibrinogen Aalpha chains. The Bbeta and gamma chains of salmon and human fibrinogen had molecular masses of approximately 55 and 50 kDa, respectively. Western blotting revealed that polyclonal rabbit anti-human fibrinogen antibodies had affinity for the gamma chains of salmon fibrinogen, making it possible to study factor XIII activity in purified salmon fibrinogen. Cross-linking of either gamma-gamma or gamma-alpha chains was not detected upon incubation of the purified fibrinogen with thrombin and calcium alone, but was detected when clotting was performed in plasma indicating absence of factor XIII activity in the purified product.     

Sequence content of α-fetoprotein, albumin and fibrinogen polypeptide mRNAs in different organs, developing tissues and in liver during carcinogenesis in rats

To investigate the variable gene activities of α-fetoprotein, albumin and fibrinogen polypeptides as markers of ‘liver specific proteins’ in different developing organs or tissues, we have used specific complementary DNA probes to detect and to quantitate α-fetoprotein, albumin and fibrinogen polypeptide mRNA, respectively, in RNA fractions, prepared from various tissues of rats at different stages of fetal and postnatal development and from hepatomas induced by diethylnitrosamine. The results indicate that there is no consistent relationship between sequence content of α-fetoprotein, albumin and fibrinogen polypeptide mRNA in different developing tissues. Intestines which are like the liver also of endodermal origin do not contain α-fetoprotein, albumin and fibrinogen polypeptide mRNAs, while kidneys which are mesodermal in origin were found to be α-fetoprotein, albumin and fibrinogen polypeptide mRNA producers in neonatal life. In yolk sac, only α-fetoprotein and fibrinogen polypeptide mRNA could be detected. In the liver, the increased level of albumin and fibrinogen polypeptide mRNA during fetal and neonatal development is accompanied with a diminished amount of α-fetoprotein mRNA. The neosynthesis of α-fetoprotein mRNA in the liver during carcinogenesis occurred without a decreased content of albumin and fibrinogen polypeptide mRNAs. These findings suggest that complex mechanisms of gene regulation are involved in variable gene activities of α-fetoprotein, albumin and fibrinogen polypeptides in cells of different organs or tissues developed from a single cell.     

Characterization of a metalloenzyme from a wild mushroom, Tricholoma saponaceum

Two kinds of metalloendopeptidases from the fruiting bodies of Tricholoma saponaceum (TSMEP1 and TSMEP2) have been purified, and TSMEP1 has been characterized based on their fibrinolytic activity. The enzymes have the same N-terminal amino acid sequence, Ala-Leu-Tyr-Val-Gly-X-Ser-Pro-X-Gln-Gln-Ser-Leu-Leu-Val, but slightly different molecular weights of 18,147 and 17,947, as measured by matrix assisted laser desorption ionization time of flight (MALDI-TOF) mass spectrometry. The N-terminal sequence do not match with any known protein or open reading frame. TSMEP1 hydrolyzes fibrinogen as well as fibrin, but does not show any proteolytic activity for other blood proteins such as thrombin, human albumin, human IgG, hemoglobin, or urokinase. The enzyme hydrolyzes both A alpha and B beta subunits of human fibrinogen with equal efficiency but didn't show any reactivity for the gamma form of human fibrinogen. The enzymatic activity is strongly inhibited by EDTA and 1,10-phenanthroline, indicating that the enzymes are metalloproteases. No inhibition was found with phenylmethylsulfonyl fluoride (PMSF), L-trans-epoxysuccinyl leucylamido-(4-guanidino)-butane (E-64), pepstatin and 2-mercaptoethanol. The activity of the purified enzyme was increased by Mg2+, Fe2+, Zn2+, and Co2+, and slightly decreased by Ca2+, but the enzyme activity was dramatically decreased by Cu2+, and totally inhibited by Hg2+. It has broad substrate specificity for synthetic peptides, and keep the high activity from pH 7.5 to 9, suggesting that the purified enzyme was a basic protease. The enzyme was stable up to 30 degrees C and the maximum fibrinolytic activity was at 55 degrees C.     

Human serum albumin as a new interacting partner of prolactin inducible protein in human seminal plasma

Prolactin inducible protein (PIP) is a 17 kDa glycoprotein. It binds to many proteins including fibrinogen, actin, keratin, myosin, immunoglobulin G, CD4, and human zinc-alpha-2 glycoprotein. Its ability to bind a large array of proteins indicates its multifaceted role in various biological processes, such as fertility, immunoregulation, antimicrobial activity, apoptosis, and tumor progression. Here, we present the first report of native human serum albumin (HSA)-PIP complex formation in seminal plasma. The complex was purified by chromatographic separation techniques, analyzed by gel electrophoresis, identified by MALDI-TOF mass spectrometry and validated by co-immunoprecipitation coupled with western blotting experiments. Moreover, the behavior of complex in solution was analyzed by dynamic light scattering and interacting residues were identified by in silico protein-protein docking. The purified protein complex shows two bands (67 kDa and 17 kDa) on SDS-PAGE gel and a single band (~85 kDa) on native PAGE gel. The predicted complex structure has 13 intermolecular hydrogen bonds, which may contribute to the overall stability of the complex. As HSA has been known to preserve the motility of sperm, native HSA-PIP complex formation may point towards an important role of PIP, which can directly be correlated with male fertility/infertility.     

Synthesis and secretion of fibrinogen and albumin by isolated rat hepatocytes

In order to investigate the regulation of synthesis of some of the plasma proteins, especially fibrinogen, at the cellular level, we have chosen to work with suspensions of hepatocytes isolated by the perfusion of rat liver with crude bacterial collagenase. By adding soybean trypsin inhibitor to the collagenase and by avoiding mechanical damage, we have prepared cell suspensions that synthesize and secrete fibrinogen and albumin and that survive for longer than twenty hours. The fibrinogen secreted is clottable and shows the same pattern in acrylamide gel electrophoresis as fibrinogen purified from rat plasma. After a three hour lag, the rate of synthesis of fibrinogen, as measured by a solid-phase radioimmunoassay, continually accelerates, so that rates several fold greater than the $\text{in vivo}$ rate have been observed after twenty hours incubation. Cycloheximide (0.1 mM) completely abolished the appearance of fibrinogen in the medium; whereas colchicine (0.3 mM) reduced the rate by 85%. Insulin and cortisol succinate enhanced fibrinogen synthesis and secretion. The albumin secretion profile differs in several respects from that of fibrinogen, reflecting differences in intracellular pool levels and probably distinct regulatory mechanisms.     

Fibrin assembly after fibrinopeptide A release in model systems and human plasma studied with magnetic birefringence.

Magnetically induced birefringence was used to monitor fibrin polymerization after the release of the small negatively charged A fibrinopeptides from human fibrinogen by the action of the snake-venom-derived enzymes reptilase and ancrod. A range of conditions was investigated. Fibrin polymerization in solutions of purified fibrinogen shows a distinct break near the gelation point. On addition of Ca2+ or albumin the lag period is shortened, fibre thickness is increased and the break in assembly almost vanishes, probably because both of these additives promote lateral aggregation. There are minor differences in the kinetics, depending on the venom enzyme used. The kinetics of fibrin assembly in model systems containing either Ca2+ or albumin and in human plasma with a largely dormant coagulation cascade are very similar. Therefore in the latter condition there is no significant alteration in the assembly process due to interaction between fibrin or the venom enzymes and any of the plasma proteins. When the cascade is activated, the polymerization progress curves have a character that resembles a combination of the reactions observed when the venom enzymes and endogenously generated thrombin separately induce coagulation, except for a region near gelation where, paradoxically, polymerization appears to be slower on activation. The low-angle neutron-diffraction patterns from oriented gels made with thrombin or reptilase are identical. Therefore at low resolution the packing of the monomers within fibres is the same when fibrinopeptide A only or both fibrinopeptides A and B are removed.     

Complex formation of platelet membrane glycoproteins IIb and IIIa with the fibrinogen D domain

Glycoprotein IIb (GPIIb) and glycoprotein IIIa (GPIIIa) form a macromolecular complex on the activated platelet surface which contains the fibrinogen-binding site necessary for normal platelet aggregation. To identify the specific region of the fibrinogen molecule responsible for its interaction with the GPIIb-GPIIIa complex, purified fragment D1 (Mr = 100,000) and fragment E (Mr = 50,000) were prepared from plasmin digests of purified human fibrinogen. In addition, the polypeptide chain subunits A alpha, B beta, and gamma of fibrinogen were prepared. Using an enzyme-linked immunosorbent assay we have demonstrated that isolated fragment D1 in a solid phase system forms a complex with a mixture of GPIIb and GPIIIa. The binding of the GPIIb-GPIIIa mixture to fragment D1-coated plates reached saturation at 8 nM and to fibrinogen-coated plates at 24 nM. Isolated A alpha, B beta, and gamma chains were not reactive with added glycoproteins. Fragment E coated directly on plastic plates or immobilized on antibody-coated plastic plates did not form a complex with GPIIb-GPIIIa. Only fluid phase fibrinogen and fragment D1 but not fragment E were inhibitory toward formation of a complex between solid phase fibrinogen and GPIIb-GPIIIa. Isolated A alpha, B beta, and gamma chains at concentrations equivalent to fluid phase fibrinogen were inactive. Binding of fragment D1 but not fragment E to the GPIIb-GPIIIa complex was also demonstrated by rocket immunoelectrophoresis of the membrane glycoprotein mixture through a gel containing the individual fragments and subsequent autoradiography of the complex following exposure to 125I-anti-fibrinogen. These observations with isolated platelet membrane glycoproteins provide strong evidence that each of the D domains of the fibrinogen molecule interacts directly with the GPIIb-GPIIIa complex on the activated platelet surface, thus allowing formation of a tertiary molecular "bridge" across the surface of two adjacent activated platelets.     

Albumin indirectly modulates fibrin and protofibrin ultrastructure

Albumin modulation of fibrin and protofibrin coagulation parameters was studied. Cation-depleted, fatty acid free, human and bovine albumins decrease fibrin clot turbidity in a concentration-dependent manner. Albumin also inhibits the formation of protofibrin gels, induced by addition of 25 microM Zn(II) to protofibrils, though it does not bind to (proto)fibrin. In order to verify that competition for cations underlies the influence of albumin, fibrinogen was dialyzed against cation-depleted albumin. Elemental analysis indicates a redistribution of Zn(II) from the fibrinogen to the albumin compartment, and the resultant fibrin clots are less turbid. Apparently, cation-depleted albumin acts as a competitor for divalent cations. The ability of albumin to compete for available Zn(II) was also expressed in gels formed by pH-jump experiments, in which fibrin monomer, maintained soluble at pH 4.9, is induced to change phase by addition of NaOH to pH 7.4. While turbidimetric evidence indicates that individual fibrin fibers simply become thinner with albumin, scanning electron micrographs (SEM) reveal a more complex effect on ultrastructure. Though albumin does not bind to the gels, fibrin gels produced with albumin show major changes in fiber ultrastructure, particularly evident in gels formed in the presence of Zn(II). These structural modifications are discussed within the context of the "excluded volume" effect, in which "crowding" by albumin alters (proto)fibrin reactivity and ultrastructure.     

Bindings of plasma proteins to streptococci of serological group L with special reference to their immunoglobulin G Fc-receptor activity

Of 33 streptococcal cultures belonging to serological group L, all bound human immunoglobulin (Ig) G, fibrinogen, and fibronectin; 32 bound bovine IgG; 31 bound alpha 2-macroglobulin; 5 bound albumin; and none bound either haptoglobin or IgA. The binding sites for IgG could be isolated from the L streptococci by trypsinization and purified by affinity chromatography on human IgG-Sepharose. The purified Fc receptors reacted with IgG subclasses 1, 2, 3, 4 of humans, 1 and 2 of bovines, ovines, and caprines as well as a, b, c, and T of equines. They had a molecular mass of approximately 49,000 Da. Thus, the Fc receptors from L streptococci corresponded to type III Fc receptors of Streptococcus dysgalactiae.     

Effect of GPIIb-IIIa complex ligands on calcium ion movement and cytoskeleton organization in activated platelets

We studied the influence of the occupancy of the fibrinogen receptor (GP IIb-IIIa complex) on two early aspects of agonist induced platelet activation: the increase of the intracellular Ca2+ concentration and the cytoskeleton reorganization. A monoclonal antibody, a peptide containing the RGD sequence and fibrinogen purified from human plasma were used as GP IIb-IIIa ligands. The obtained results demonstrated that fibrinogen receptor occupancy inhibits Ca2+ movement and cytoskeleton reorganization caused by low thrombin concentration and ADP.     

Identification of N-homocysteinylation sites in plasma proteins

A protocol for the identification of N-homocysteinylation sites in plasma proteins is described. Human plasma or purified fibrinogen is subjected to trypsin digestion and analysis of N-Hcy-peptides by liquid chromatography/mass spectroscopy (LC/MS). Human fibrinogen is isolated from the plasma by the glycine precipitation method. Identification of N-Hcy-Lys-peptides in tryptic digests of in vivo-derived samples is facilitated by the use of N-Hcy-albumin and N-Hcy-fibrinogen synthesized in vitro from commercially available human proteins. This protocol allows identification of N-homocysteinylation sites at Lys4, Lys12, Lys137, and Lys525 in albumin directly in trypsin-digested human serum samples. N-Hcy-Lys562, N-Hcy-Lys344, and N-Hcy-Lys385 were identified in human fibrinogen from patients with cystathionine β-synthase deficiency. The protocol can be completed in 4 days.     

On the mechanism of fibrin-specific plasminogen activation by staphylokinase

The mechanism of plasminogen activation by recombinant staphylokinase was studied both in the absence and in the presence of fibrin, in purified systems, and in human plasma. Staphylokinase, like streptokinase, forms a stoichiometric complex with plasminogen that activates plasminogen following Michaelis-Menten kinetics with Km = 7.0 microM and k2 = 1.5 s-1. In purified systems, alpha 2-antiplasmin inhibits the plasminogen-staphylokinase complex with k1(app) = 2.7 +/- 0.30 x 10(6) M-1 s-1 (mean +/- S.D., n = 12), but not the plasminogen-streptokinase complex. Addition of 6-aminohexanoic acid induces a concentration-dependent reduction of k1(app) to 2.0 +/- 0.17 x 10(4) M-1 s-1 (mean +/- S.D., n = 5) at concentrations greater than or equal to 30 mM, with a 50% reduction at a 6-aminohexanoic acid concentration of 60 microM. Staphylokinase does not bind to fibrin, and fibrin stimulates the initial rate of plasminogen activation by staphylokinase only 4-fold. Staphylokinase induces a dose-dependent lysis of a 0.12-ml 125I-fibrin-labeled human plasma clot submersed in 0.5 ml of citrated human plasma; 50% lysis in 2 h is obtained with 17 nM staphylokinase and is associated with only 5% plasma fibrinogen degradation. Corresponding values for streptokinase are 68 nM and more than 90% fibrinogen degradation. In the absence of a fibrin clot, 50% fibrinogen degradation in human plasma in 2 h requires 790 nM staphylokinase, but only 4.4 nM streptokinase. These results suggest the following mechanism for relatively fibrin-specific clot lysis with staphylokinase in a plasma milieu. In plasma in the absence of fibrin, the plasminogen-staphylokinase complex is rapidly neutralized by alpha 2-antiplasmin, thus preventing systemic plasminogen activation. In the presence of fibrin, the lysine-binding sites of the plasminogen-staphylokinase complex are occupied and inhibition by alpha 2-antiplasmin is retarded, thus allowing preferential plasminogen activation at the fibrin surface.     

Molecular weight of human fibrinogen derived from phosphorus determinations

1. Fibrin clots obtained from diluted human plasma with bovine thrombin often contain amounts of phospholipids that cannot be diminished by further plasma dilution. 2. The ;cold insoluble residue' obtained during fibrinogen preparation has a higher phosphorus content than the purified fibrinogen. 3. Evidence showed that adsorption of phospholipids or phosphorus-containing fibrinopeptides on purified fibrinogen or fibrin was unlikely. 4. O-Phosphorylserine was detected in acid hydrolysates of human fibrin. 5. On the basis of phosphorus determinations the average molecular weight of human fibrinogen cannot be less than 342000 (304000-383000) for a group of ten donors, and 265000 for two other persons, assuming 1 phosphorus atom/molecule and incomplete splitting of the phosphorus-containing fibrinopeptide. Complete splitting of the phosphopeptide would require molecular weights twice as high. 6. Fibrinolysis was a possible cause of lower phosphorus contents found in isolated fibrinogen and fibrin from a donor who showed apprehension during blood collection and in a fibrinogen preparation that had been submitted to prolonged dialysis.     

Analysis of fibrinogen genes in patients with congenital afibrinogenemia

Several cDNA clones coding for A alpha, B beta and gamma chains of fibrinogen have been isolated from a human liver cDNA library. They were selected by differential hybridization with probes raised against fractionated liver mRNA (positive probes) and muscle and albumin mRNA (negative probes), then firmly identified by positive hybridization selection. Three of these clones, encoding A alpha, B beta and gamma fibrinogen chain sequences, were further characterized by restriction mapping and used as probes to characterize fibrinogen mRNAs from adult and fetal liver and fibrinogen genes in normal individuals and two afibrinogenemic patients. The results indicate that there is a single copy of the fibrinogen genes which are present and grossly intact in afibrinogenemic DNA.     

Ristocetin-dependent reconstitution of binding of von Willebrand factor to purified human platelet membrane glycoprotein Ib-IX complex

Whether the human platelet membrane glycoprotein (GP) Ib-IX complex is the receptor for ristocetin-dependent binding of von Willebrand factor (vWF) has been examined by reconstitution with the purified components using a solid-phase bead assay. Purified GP Ib-IX complex was bound and orientated on the beads via a monoclonal antibody, FMC 25, directed against the membrane-associated region of the complex. Specific binding of 125I-labeled vWF to the GP Ib-IX complex coated beads was strictly ristocetin dependent with maximal binding occurring at ristocetin concentrations greater than or equal to 1 mg/mL. Ristocetin-dependent specific binding of 125I-labeled vWF was saturable. The observed binding was specific to the interaction between vWF and the GP Ib-IX complex since there was no ristocetin-dependent specific binding of vWF if the physicochemically related platelet membrane glycoprotein, GP IIb, was substituted for the GP Ib-IX complex in a corresponding bead assay. Further, neither bovine serum albumin nor other adhesive glycoproteins, such as fibrinogen or fibronectin, specifically bound to the GP Ib-IX complex in the presence of ristocetin. Ristocetin-dependent binding of vWF to platelets and to GP Ib-IX complex coated beads was inhibited by monoclonal antibodies against a 45,000 molecular weight N-terminal region of GP Ib but not by monoclonal antibodies directed against other regions of the GP Ib-IX complex. Similar correspondence between platelets and purified GP Ib-IX complex with respect to the ristocetin-dependent binding of vWF was obtained with anti-vWF monoclonal antibodies.(ABSTRACT TRUNCATED AT 250 WORDS)     

Anticoagulant proteases from western diamondback rattlesnake (Crotalus atrox) venom

Crotalus atrox venom contains agents that render human fibrinogen and plasma incoagulable by thrombin. To elucidate the mechanism of alteration of fibrinogen clotting function by the venom, four immunochemically different proteases, I, II, III, and IV, were purified from the venom by anion-exchange chromatography and column gel filtration. All four proteases had anticoagulant activity rendering purified fibrinogen incoagulable. Proteases I and IV do not affect fibrinogen in plasma but in purified fibrinogen cleave the A alpha chain first and then the B beta and gamma chains. Both enzymes are metalloproteases containing a single polypeptide chain with 1 mol of zinc, are inhibited by (ethylenedinitrilo)tetraacetate and human alpha 2-macroglobulin, and have an optimal temperature of 37 degrees C and an optimal pH of 7. Protease I has a molecular weight (Mr) of 20 000 and is the most cationic. Protease IV has an Mr of 46 000 and is the most anionic glycoprotein with one free sulfhydryl group. Proteases II and III degrade both purified fibrinogen and fibrinogen in plasma, cleaving only the B beta chain and leaving the A alpha and gamma chains intact. Both enzymes are alkaline serine proteases, cleave chromogenic substrates at the COOH terminal of arginine or lysine, are inhibited by diisopropyl fluorophosphate and phenylmethanesulfonyl fluoride, and have an optimal temperature of 50-65 degrees C. Protease II is a single polypeptide chain glycoprotein with an Mr of 31 000. Protease III is a two polypeptide chain protein with an Mr of 24 000, each of the two chains having an Mr of 13 000; its activity is not affected by major protease inhibitors of human plasma. Proteases II and III are enzymes with unique and limited substrate specificity by cleaving only the B beta chain, releasing a peptide of Mr 5000 and generating a fibrinogen derivative of Mr 325 000, with intact A alpha and gamma chains and poor coagulability. Since the two enzymes are active in human plasma and serum, it is postulated that proteases II and III can mediate anticoagulant effects in vivo after envenomation.