ON THE NATURE OF FORCES OPERATING IN BLOOD CLOTTING : I. THE PARTICIPATION OF ELECTROSTATIC ATTRACTION

The results of the study of the inhibiting effect of neutral salts upon the clotting tendency of fibrinogen by thrombin may be summarised as follows: Salts like NaCl and KCl inhibit only weakly. Salts of the same cation (K^•) with monovalent anions of different ionic radius are the more active the larger the anion (Cl'',Br'',I''). Salts of the same cation with anions of different valency are the more active the higher the charge of the anion (1–1 <1–2 <1–3 <1–4). Salts with the same anion with cations of different valency show stronger inhibition in the case of cations of higher charge (K^•,Na^• < Mg^••, Ca^••, Sr^••, Ba^••). Salts with the same anion and cations of the same charge, but of different radius, are the more active the larger the cation (but with an inversion between Mg^•• and Ca^•• in the series of the alkali earths, which is not infrequent in biocolloids). These results show that the clotting of fibrinogen with thrombin is, at least partly, caused by a coacervation process, due to electrostatic attraction between positive and negative groups. Its nature and localisation will be dealt with in the next paper of this series.     

Mechanisms involved in fibrin formation

That the role of thrombin in the conversion of fibrinogen to fibrin is essentially enzymatic, is established not only by the minute amounts of thrombin which are effective but also by the complete independence of fibrin yields and thrombin concentrations over a very wide range of thrombin dilutions and clotting times. The thrombin-fibrinogen reaction, in the phase beyond the "latent period" at least, seems fundamentally "first order." Technical requirements of the experiments leading to these conclusions include: (1) a highly purified (e.g. 97 per cent "clottable") fibrinogen, (2) absence of traces of thrombic impurities in the fibrinogen, (3) absence of fibrinolytic protease contaminant of the thrombin and the fibrinogen, and (4) sufficient stability of the thrombin even at very high dilutions. Four conditions affecting thrombin stability have been investigated. Fibrin yields are not significantly modified by numerous experimental circumstances that influence the clotting time, such as (1) temperature, (2) pH, (3) non-specific salt action due to electrical (ionic) charges, which alter the Coulomb forces involved in the fibrillar aggregation, (4) specific ion effects, whether clot-accelerating (e.g. Ca⁺⁺) or clot-inhibitory (e.g. Fe(CN)₆'), (5) occluding (adsorptive) colloids, which have a "fibrinoplastic" action, e.g. (a) acacia and probably (b) fibrinogen which has been mildly "denatured" by salt-heating, acidification, etc. The data with which several European workers have attempted to substantiate the idea of a two-stage thrombin-fibrinogen reaction with an intermediary "profibrin" (allegedly partly "denatured") have been reanalyzed with controls which lead us to very different conclusions, viz. (1) denaturation and fibrin formation are independent; (2) partial denaturation is "fibrinoplastic" (see above); and (3) conditions of strong salinity and acid pH (5.1) usually do not completely prevent the thrombin-fibrinogen reaction but merely prolong the "latent" phase and lessen the time required for completion of essentially the same reaction (fibrin polymerization) when more favorable clotting conditions are restored. Thus, our experiments advance the modern concepts concerning the coagulation mechanisms along lines that, for the most part, agree with those of the Harvard physical chemists, and we oppose the European views concerning a two-stage reaction, "profibrin," and "the denaturase theory" of clotting.     

Clotting of fibrinogen. 2. Calorimetry of the reversal of the effect of calcium on clotting with thrombin and with ancrod

When clotting is effected by thrombin in the presence of calcium, the endotherm for the D nodules of fibrinogen broadens significantly and then becomes narrow again, while increasing in size. Clotting effected by the snake venom enzyme Ancrod, which releases only the A fibrinopeptides from the E nodule, shows only the broadening of the D endotherm. Accordingly, significant interactions of the D nodules of fibrinogen become possible only when the B fibrinopeptides of the E nodule are released on clotting. When calcium present during clotting is removed from the fibrin clot with ethylenediaminetetraacetic acid, the endotherm for the D nodules of fibrin shows nearly complete reversal if clotting was effected with Ancrod but appears to be divided into two endotherms if clotting was effected with thrombin. At neutral pH, new endotherms were observed for fibrinogen in the temperature range 105-140 degrees C.     

Changes in pH associated with clotting of fibrinogen. Kinetic studies of the pH shift and correlation of the pH change with the release of fibrinopeptides and the ensuing polymerization

The effect of the initial pH and the concentrations of thrombin, fibrinogen, and Ca2+ upon the rate of pH change associated with clotting of bovine fibrinogen by human thrombin was investigated at pH 6.80, 7.80, and 8.80, 0.3 ionic strength, 25 degrees C, and 19.5 mg/mL final fibrinogen concentration. At pH 6.80 and 7.80, the reaction was first order, with rate constant k1. At pH 8.80, a first-order reaction of the release of H+ (k1) was followed by a partial rebinding of these in a reaction consecutive to the first one (k2). At each of the above pH values, k1 was proportional to thrombin concentration in the 0.05-3.0 min-1 range investigated. The k1 constants were 0.111 +/- 0.001, 0.250 +/- 0.005, and 0.190 +/- 0.002 min-1 (NIH thrombin units)-1 mL-1 at pH 6.80, 7.80, and 8.80, respectively. Plots of log rate vs log thrombin concentration of these data were linear with slopes close to 1 at all three pH values. The rate of the second reaction (k2) was independent of both the thrombin and the initial fibrinogen concentration. The pH dependence of k1 exhibited a bell-shaped curve that could be resolved into the effect of one group with a pK of 7.27 that increased the rate and another with a pK of 9.22 that decreased the rate. With constant thrombin concentration but varying fibrinogen concentration, plots of 1/k1 vs [fibrinogen] were linear, but the lines did not pass through the origin. From the slope and intercept, kcat and KM of the Michaelis-Menten equation could be calculated. The same parameters were obtained also from initial velocity vs [fibrinogen] plots. Values of kcat were consistent and accurate; those of KM were more scattered. KM was (22.4-34.2) X 10(-6) M at pH 6.80 and approximately 7 X 10(-6) M in the pH 7.26-8.80 range. The latter value, pertaining to the release of H+ ions, is in agreement with values in the literature for KM of the release of fibrinopeptide A by thrombin in the 7.4-8.0 pH range. The value of kcat s-1 (unit of thrombin)-1 mL-1 increases from 1.2 X 10(-10) s-1 unit of thrombin-1 mL-1 at pH 6.80 to 2.46 X 10(-10) at pH 7.80 and then decreases to 2.01 X 10(-10) 10(-1) (units of thrombin)-1 mL-1 at pH 8.80. The kcat values are significantly lower than those in the literature for the release of fibrinopeptide A.(ABSTRACT TRUNCATED AT 400 WORDS)     

Mutation of the H-helix in antithrombin decreases heparin stimulation of protease inhibition

Blood clotting proceeds through the sequential proteolytic activation of a series of serine proteases, culminating in thrombin cleaving fibrinogen into fibrin. The serine protease inhibitors (serpins) antithrombin (AT) and protein C inhibitor (PCI) both inhibit thrombin in a heparin-accelerated reaction. Heparin binds to the positively charged D-helix of AT and H-helix of PCI. The H-helix of AT is negatively charged, and it was mutated to contain neutral or positively charged residues to see if they contributed to heparin stimulation or protease specificity in AT. To assess the impact of the H-helix mutations on heparin stimulation in the absence of the known heparin-binding site, negative charges were also introduced in the D-helix of AT. AT with both positively charged H- and D-helices showed decreases in heparin stimulation of thrombin and factor Xa inhibition by 10- and 5-fold respectively, a decrease in affinity for heparin sepharose, and a shift in the heparin template curve. In the absence of a positively charged D-helix, changing the H-helix from neutral to positively charged increased heparin stimulation of thrombin inhibition 21-fold, increased heparin affinity and restored a normal maximal heparin concentration for inhibition.     

The conversion of fibrinogen to fibrin. XIII. Dissolution of fibrin and inhibition of clotting by various neutral salts

1. Fibrin clots prepared in the absence of calcium can be dissolved in solutions of lithium chloride and bromide and sodium bromide and iodide, as well as of guanidine hydrochloride and urea. These salts do not denature fibrinogen under the same conditions of concentration, temperature, and time. Sedimentation experiments on the fibrin solutions show in each case a single sharp peak with a sedimentation constant close to that of fibrinogen. 2. At lower concentrations, these salts inhibit the clotting of fibrinogen by thrombin, but in the case of lithium bromide and sodium iodide, at least, allow an intermediate polymer to accumulate whose sedimentation constant is close to that of the polymer observed in systems inhibited by hexamethylene glycol or urea.     

THE ORIGIN OF THE ELECTRICAL CHARGES OF COLLOIDAL PARTICLES AND OF LIVING TISSUES

1. When a solution of a salt of gelatin or crystalline egg albumin is separated by a collodion membrane from a watery solution (free from protein) a potential difference is set up across the membrane in which the protein is positively charged in the case of protein-acid salts and in which the protein is negatively charged in the case of metal proteinates. The turning point is the isoelectric point of the protein. 2. Measurements of the pH of the (inside) protein solution and of the outside watery solution show that when equilibrium is established the value pH inside minus pH outside is positive in the case of protein-acid salts and negative in the case of metal proteinates. This is to be expected when the P.D. is caused by the establishment of a Donnan equilibrium, since in that case the pH should be lower outside than inside in the case of a protein-acid salt and should be higher outside than inside in the case of a metal proteinate. 3. At the isoelectric point where the electrical charge is zero the value of pH inside minus pH outside becomes also zero. 4. It is shown that a P.D. is established between suspended particles of powdered gelatin and the surrounding watery solution and that the sign of charge of the particles is positive when they contain gelatin-acid salts, while it is negative when the powdered particles contain metal gelatinate. At the isoelectric point the charge is zero. 5. Measurements of the pH inside the powdered particles and of the pH in the outside watery solution show that when equilibrium is established the value pH inside minus pH outside is positive when the powdered particles contain a gelatin-acid salt, while the value pH inside minus pH outside is negative when the powdered particles contain Na gelatinate. At the isoelectric point the value pH inside minus pH outside is zero. 6. The addition of neutral salts depresses the electrical charge of the powdered particles of protein-acid salts. It is shown that the addition of salts to a suspension of powdered particles of gelatin chloride also diminishes the value of pH inside minus pH outside. 7. The agreement between the values 58 (pH inside minus pH outside) and the P. D. observed by the Compton electrometer is not only qualitative but quantitative. This proves that the difference in the concentration of acid (or alkali, as the case may be) in the two phases is the only cause for the observed P.D. 8. The Donnan theory demands that the P.D. of a gelatin chloride solution should be 1½ times as great as the P.D. of a gelatin sulfate solution of the same pH and the same concentration (1 per cent) of originally isoelectric gelatin. This is found to be correct and it is also shown that the values of pH inside minus pH outside for the two solutions possess the ratio of 3:2. 9. All these measurements prove that the electrical charges of suspended particles of protein are determined exclusively by the Donnan equilibrium.     

Peroxynitrite-mediated oxidation of fibrinogen inhibits clot formation

The clotting activity of human fibrinogen was fully inhibited in vitro by peroxynitrite. The decrease of activity followed an exponential function and the concentration of peroxynitrite needed to inhibit 50% of fibrinogen clotting was 22 microM at 25 degrees C. The oxidative modification(s) induced by the peroxynitrite system (i.e. ONOO-, ONOOH and ONOOH*) appeared specifically to affect fibrin clot formation (through the inhibition of fibrinogen polymerization) since the interaction of peroxynitrite-modified fibrinogen with thrombin appeared to be unaffected. The addition of NaHCO3 decreased the peroxynitrite effect on fibrinogen clotting, suggesting that the reactive species formed by the reaction of CO2 with peroxynitrite are less efficient oxidants of peroxynitrite itself. Similar effects were observed after addition of bilirubin, which also exerted a significant protection against peroxynitrite-mediated modification of fibrinogen.     

The binding of calcium to fibrinogen: influence on the clotting process.

The influence of Ca2+ on the basic reaction between thrombin and fibrinogen was investigated. The results demonstrate that: (a) A Ca2+-dependent dimeric intermediate is formed during the early step of the clotting process. This dimeric intermeidate is shown to be formed by the association of an intact fibrinogen molecule and a fibrin monomer devoid in only the peptide A, (b) Ca2+ enhances the proteolytic step as illustrated by the measure of the kinetics of H+ release at pH 8.6. On the basis of these observations it is proposed that Ca2+ catalyses the proteolysis of fibrinogen by thrombin through the formation of a Ca2+-dependent dimer.     

Inhibition of thrombin by synthetic hirudin peptides

To investigate the role of different regions of hirudin in the interaction with the proteinase thrombin, segments of hirudin containing 15-51 residues were synthesized. The C-terminal segment 40-65 inhibited the fibrinogen clotting activity of thrombin but not amidolysis of tosyl-Gly-Pro-Arg-p-nitroanilide. Central peptide 15–42 was insoluble at pH 7, but peptide 15-65 inhibited fibrinogen clotting and amidolysis to an equal extent. The N-terminal loop peptide 1-15 had no inhibitory activity and did not affect the potency of peptide 15-65. These data suggest that the central region inhibits catalysis.     

Review of some unusual effects of calcium binding to fibrinogen

Calcium binding curves of human and bovine fibrinogen were obtained by using a calcium sensitive electrode. The two were identical and showed 2 high, 2-3 medium and more than 15 low affinity sites. Differential scanning calorimetry at neutral pH demonstrated the presence of the D and E domains of fibrinogen; however, at pH 3.5 the D-domain was split into two. The presence of the subdomains was demonstrated also by digestion by pepsin at this pH. Combination of digestion of fibrinogen and of its fragments with different enzymes and temperatures identified up to 12 subdomains in the original molecule. Clotting of fibrinogen by thrombin at pH 7.0 was investigated also by differential scanning calorimetry. In the absence of Ca2+ clotting elicited a 40% increase in the enthalpy of thermal denaturation of the D domain of fibrinogen, but the position of the peak increased only by 0.4 degrees C. However, with clotting in the presence of 10(-3) M calcium the former increased by 70-75% and the latter by 11.0 degrees C, while these parameters of the E-domain remained unchanged. Changes of bound calcium during clotting were also measured with the calcium sensitive electrode. These had to be corrected, because the drop in free calcium was partly compensated by release of some calcium that was already bound to fibrinogen. Log of the half time of calcium uptake plotted against log thrombin concentration indicated a first order process with respect to thrombin concentration, moreover, the rate determined corresponded to that of the conformation change measured by calorimetry. The calcium uptake was correlated with release of the fibrinopeptides. Release of fibrinopeptide B follows parallel to binding of calcium and that of fibrinopeptide A is about fourfold faster. Polymerization and formation of thick bundles of fibrin is connected with release of fibrinopeptide A. Clotting with Ancrod, an enzyme that releases only fibrinopeptide A, showed only minimal binding of calcium. The polymerization inhibiting tetrapeptide Gly-Pro-Arg-Pro also depressed binding of calcium. These data suggest that a calcium-binding site must be in the proximity of the site of release of fibrinopeptide B and of a polymerization site.     

Antithrombin activity of a peptide corresponding to residues 54-75 of heparin cofactor II

Heparin cofactor II (HCII) is a highly specific serine proteinase inhibitor, which complexes covalently with thrombin in a reaction catalyzed by heparin and other polyanions. The molecular basis for the thrombin specificity may be explained by the identification here of a segment of HCII including residues 54-75 that binds to thrombin. A synthetic peptide, HCII(54-75), based on this segment of HCII, Gly-Glu-Glu-Asp-Asp-Asp-Tyr-Leu-Asp-Leu-Glu- Lys-Ile-Phe-Ala-Glu-Asp-Asp-Asp-Tyr-Ile-Asp inhibited thrombin's cleavage of fibrinogen. Clotting activity of thrombin was inhibited 50% at a concentration of 28 microM. Polyacrylamide gel electrophoresis showed that HCII(54-75) inhibited thrombin's cleavage of both the A alpha and B beta polypeptides in fibrinogen. However, the peptide did not block thrombin's active site, as hydrolysis of chromogenic substrates was not inhibited. HCII(54-75) probably binds to the same site on thrombin as do carboxyl-terminal residues of hirudins, thrombin inhibitors of leeches. HCII(54-75) inhibited binding of thrombin to a synthetic peptide corresponding to residues 54-66 of hirudin PA, but the hirudin peptide was about 30-fold more potent in binding and clotting assays. Both synthetic peptides, as a result of their polyanionic character, might be expected to stimulate the reaction of HCII with thrombin. However, the hirudin-related peptide inhibited this reaction, suggesting that it blocked a site on thrombin required for interaction with HCII. HCII(54-75) had a net stimulatory effect on the thrombin-HCII reaction as a consequence of its lower affinity for thrombin and greater negative charge relative to the hirudin-related peptide. These studies suggest that residues 54-75 of HCII interact with a noncatalytic binding site on thrombin and that this interaction contributes to efficient inhibition of thrombin by HCII.     

Gel formation by fibrin oligomers without addition of monomers

Soluble fibrin oligomers were formed by reacting fibrinogen with thrombin under fine clotting conditions where the action of thrombin is the rate-determining step for polymerization, and by inhibiting the reaction shortly before gelation. Oligomeric fibrin was separated from unreacted fibrinogen and small oligomers by gel permeation chromatography. Electron microscopy revealed that the largest soluble fibrin oligomers resemble the protofibrils present in fine clots, but are somewhat shorter and entirely lack the twisted, trifunctional junctions that contribute to the elastic properties of fine clots. When thrombin was added to the soluble fibrin oligomers, polymerization resumed and clots were formed at a more rapid rate than from fibrinogen at the same concentration and resulted in a less-opaque clot under coarse clotting conditions. The results confirm a prediction of a theory for the polymerization of fibrin and provide additional evidence that the final state of a coarse fibrin clot depends on the mobility of protofibrils during its formation.     

Inhibition of the enzymatic activity of thrombin by concanavalin A

Concanavalin A, a carbohydrate lectin derived from the jack bean, prolongs the thrombin clotting time of human plasma or purified fibrinogen. Prolongation is due to delay in peptide release from fibrinogen. The rate of fibrin monomer polymerization is not affected. Hydrolysis of protamine sulfate by thrombin is inhibited by concanavalin A. All inhibitory effects are prevented by α-methyl-D-mannoside. Concanavalin A does not delay clotting of fibrinogen by reptilase (releases fibrinopeptide A only) or by Ancistrodon contortrix contortrix (releases fibrinopeptide B initially followed by a small amount of A). It is concluded that concanavalin A binds to a carbohydrate on the thrombin molecule thus inhibiting its enzymatic activity.     

Purification and characterization of a fibrinolytic enzyme and identification of fibrinogen clotting enzyme in a marine green alga, Codium divaricatum

A fibrinolytic enzyme was isolated from a marine green alga, Codium divaricatum, and designated C. divaricatum protease (CDP). This protease effectively hydrolyzed fibrinogen A alpha chain, while it had very low hydrolyzing efficiency for B beta and gamma chains. This property was similar to that of alpha-fibrinogenase isolated from snake venom. Protease activity peaked at pH 9, and was completely inhibited by diisopropyl fluorophosphate (DFP) and phenylmethylsulfonyl fluoride (PMSF), identifying it as a serine protease. Its molecular form was single polypeptide structure and molecular weight was estimated as 31,000 by SDS-PAGE. Fibrinogen clotting enzyme was also identified in a fraction by ion-exchange chromatography. Analysis of clots formed by the enzyme and by thrombin by SDS-PAGE showed that the fibrinogen clotting enzyme would act like thrombin and have high substrate specificity.     

Characterization of the fibrin polymer structure that accelerates thrombin cleavage of plasma factor XIII

The effect of plasmin-derived fibrin(ogen) degradation products on alpha-thrombin cleavage of plasma Factor XIII was studied to identify the fibrin polymer structure that promotes Factor XIIIa formation. Fibrin polymers derived from fibrinogen and Fragment X enhanced the rate of thrombin cleavage of plasma Factor XIII in plasma or buffered solutions. The concentrations of fibrinogen and Fragment X that promoted half-maximal rates of Factor XIIIa formation were 5 and 40 micrograms/ml, respectively. Fragments Y, D, E, D-dimer, and photooxidized fibrinogen did not enhance thrombin cleavage of Factor XIII. Although purified Fragment D1 inhibited fibrin gelation, the soluble protofibrils promoted thrombin activation of Factor XIII. Noncrosslinked fibrin fibers failed to enhance thrombin cleavage of Factor XIII. In conclusion, soluble fibrin oligomers function to promote thrombin cleavage of plasma Factor XIII during blood clotting.     

Bacitracin and gramicidin S as biospecific ligands for affinity chromatography of human thrombin

It was shown that the cyclic polypeptide antibiotic bacitracin is a competitive inhibitor of fibrinogen clotting by thrombin. Biospecific adsorbents for isolation of thrombin by gramicidin S and bacitracin attachment to silochrome S-80 modified by gamma-glycido-oxypropyl groups were synthesized. The thrombin yield at pH 7.2 and 8.0 was 76.5-96%, purification--6.2-11.6-fold, specific coagulating activity--940-1750 NIH u./mg protein. At pH 6.1 the enzyme does not practically bind to the adsorbents. In all probability, the differences in thrombin binding are due to conformational changes in the enzyme molecule, when pH changes from 6.1 to 7.2. Possible application of the synthesized adsorbents for obtaining laboratory and commercial preparations of thrombin and their perspective use for purification of other blood plasma serine proteinases possessing a narrow specificity are discussed.     

Electrostatic influence of local cysteine environments on disulfide exchange kinetics

The ionic strength dependence of the bimolecular rate constant for reaction of the negative disulfide 5,5'-dithiobis (2-nitrobenzoic acid) with cysteines in fragments of naturally occurring proteins was determined by stopped-flow spectroscopy. The Debye-Hückel relationship was applied to determine the effective charge at the cysteine and thereby determine the extent to which nearby neighbors in the primary sequence influence the kinetics. Corrections for the secondary salt effect on cysteine pKs were determined by direct spectrometric pH titration of sulfhydryl groups or by observation of the ionic strength dependence of kinetics of cysteine reaction with the neutral disulfide 2,2'-dithiodipyridine. Quantitative expressions was verified by model studies with N-acetyl-cystein. At ionic strengths equal to or greater than 20 mM, the net charge at the polypeptide cysteine site is the sum of the single negative charge of the thiolate anion and the charges of the amino acids immediately preceding and following the cysteine in the primary sequence. At lower ionic strengths, more distant residues influence kinetics. At pH 7.0, 23 degree C, and an ionic strength of 20 mM, rate constants for reaction of the negative disulfide with a cysteine having two positive neighbors, one positive and one neutral neighbor, or two neutral neighbors are 132000, 3350, and 367 s-1 M-1, respectively. This corresponds to a contribution to the activation energy of 0.65- 1.1 kcal/mol per ion pair involved in collision between the cysteine and disulfide regions. The results permit the estimation that cysteine local environments may provide a means of achieving a 10(6)-fold range in rate constants in disulfide exchange reactions in random-coil proteins. This range may prove useful in developing strategies for directing disulfide pairing in synthetic proteins.     

THE FLOCCULATION MAXIMUM (pH) OF FIBRINOGEN AND SOME OTHER BLOOD-CLOTTING REAGENTS. (RELATIVE TURBIDIMETRY WITH THE EVELYN PHOTOELECTRIC COLORIMETER)

By means of a novel adaptation of the Evelyn photoelectric colorimeter to the measurement of relative turbidities, the question of the flocculation maximum (F.M.) in acetate buffer solutions of varying pH and salt content has been studied on (a) an exceptionally stable prothrombin-free fibrinogen and its solutions after incipient thermal denaturation and incomplete tryptic proteolysis, (b) plasma, similarly treated, (c) prothrombin, thrombin, and (brain) thromboplastin solutions. All the fibrinogens show a remarkable uniformity of the precipitation pattern, viz. F.M. =4.7 (±0.2) pH in salt-containing buffer solutions and pH = 5.3 (±0.2) in salt-poor buffer (N/100 acetate). The latter approximates the isoelectric point (5.4) obtained by cataphoresis (14). There is no evidence that denaturation or digestion can produce any "second maximum." The data support the view that fibrin formation (under the specific influence of thrombin) is intrinsically unrelated to denaturation and digestion phenomena, although all three can proceed simultaneously in crude materials. A criticism is offered, therefore, of Wöhlisch''s blood clotting theory. Further applications of the photoelectric colorimeter to coagulation problems are suggested, including kinetic study of fibrin formation and the assay of fibrinogen, with a possible sensitivity of 7.5 mg. protein in 100 cc. solution.     

PLATELET DISINTEGRATION DURING CLOTTING

Sequential alterations in the ultrastructure of blood platelets observed during the clotting of human platelet-rich plasma are described, with emphasis on disintegration of the platelet. As the clotting reaction proceeds, aliquots of citrated and recalcified citrated plasma are fixed by adding buffered OsO₄. After recalcification a lag period of about 10 minutes is followed by an interval of rapidly occurring changes which include reorientation of cytoplasmic contents, progressive central degeneration, and disruption of the platelet limiting membrane. Shortly thereafter vesicular platelet remains are seen at the peripheries of loosely arranged and widely spaced masses of granular material. As clot retraction proceeds, these masses gradually come closer together, become more and more compact, and finally disappear. At the same time the vesicles undergo progressive disintegration until at the end of the experiment, 2 hours after recalcification, only a few are found randomly distributed in a dense clot. The significance of progressive disintegration and the origin of the vesicles observed in the later stages of the experiment are discussed in relation to clot formation and clot retraction.