P . JOHNSON AND W. E . P . NAISMITH 109 THE EFFECTS OF CERTAIN IONS AND NEUTRAL MOLECULES ON THE CONVERSION OF FIBRINOGEN TO FIBRIN BY SIDNEY SHULMAN School of Medicine, The University of Buffalo, Buffalo, New York Received 8th May, 1952 Certain reagents interfere strongly with the clotting of fibrinogen by thrombin. A large number of compounds have been studied under the same conditions of pH, ionic strength, and protein concentrations, and about 40 of these reagents were found to delay clotting for at least 24 h ; this has been termed inhibition. A like number of substances were found which, even at high concentrations, caused much less clotting delay ; this has been termed retardation. Inhibition is reversible and does not involve destruction of either protein. It can be achieved even if the reagent is added after the thrombin, at any time up to the moment of actual clotting.Correlations are made between the molec- ular structures of the reagents and their inhibitory effectiveness. Certain general struc- tural features may be required, though specific structural details are also significant. The clotting of fibrinogen by thrombin is markedly affected by the presence of certain reagents. Except for the very few substances that have been found to accelerate the reaction, these reagents may be classified into two main groups- retafders and inhibitors. Members of both groups will increase the time required for the first appearance of a clot after mixing the two protein solutions, and de- crease the final opacity of the resulting clot. The clotting time increase and the opacity decrease are progressively augmented with increasing concentration of reagent .Is 2 The retarders, even a t very high concentrations, increase the clotting time by a t most a factor of 10 or 20, compared to similar systems containing no reagent, while the more effective compounds, even at moderate concentrations,110 CONVERSION OF FIBRINOGEN TO FIBRIN delay clotting for at least 24 h, corresponding to a factor of about 120.This action has been defined as inhibition,3 since a 24-h delay in clotting generally means that the mixture will remain fluid indefinitely. This has generally been studied at pH 6.2 and ionic strength 0.45; the effects of altering these variables will be discussed below. Inhibition has been found to be a reversible phenomenon as shown by the fact that apparently normal clotting will result from dialysis of such mixtures against salt solution.Further evidence of the innocuous nature of such a reagent is pro- vided by mixing it with either of the proteins for 24 h. On removal of the reagent by dialysis, the protein solution remains clear and fluid and clots in a normal manner on mixing with the other protein. Some of the reagents have been ob- served to cause denaturation of fibrinogen, as evidenced by turbidity increase and formation of a fragile opague coagulum, but this occurs only at concentrations considerably higher than those required for inhibition. It was also found that inhibition can be produced by adding the reagent solu- tion after mixing the two protein solutions, provided the addition precedes the actual gelation.3 This has been termed post-thrombin inhibition, and is of im- portance in showing that the inhibitor must be influencing fibrinogen rather than thrombin, since the enzyme must have completed its action by the time of clotting, and any interference with thrombin should not then have any effect on the sub- sequent fibrinogen polymerization.The inhibitors can be conveniently classified into five groups, viz. neutral alcohols, neutral amides, dipolar ions, cations, and anions. Some representative examples 3 are listed in table 1, along with the minimum inhibiting concentration of each. Some reagents of similar structure that are only retarders 3 are shown in table 2. The inhibitors bromide and iodide were first studied by Edsall and Lever,4 who employed different conditions of protein concentrations and ionic strength, and a different criterion of clotting delay, so that comparison with the inhibitors as defined above is somewhat difficult. But they did find that guanidine and thiocyanate interfere markedly with clotting, as did also urea, iodide, and acetyltryptophanate.Under their conditions, bromide was not very effective. In a recent study,5 sodium bromide and iodide were examined in the same manner as in the earlier inhibitor investigation, with the results indicated in table 1. TABLE 1 .-INHIBITORS OF FIBRINOGEN CLOTTING Fibrinogen, 5.0 g/l. ; thrombin, 1-0 unit/ml ; pH 6-20 ; ionic strength, 0.45 neutral alcohols cations minimum inbibiting (molell.) reagent concentration reagent heptamethylene glycol 0.24-0.28 guanidine hexamethylene glycol 0-3 1-0-33 histamine*b pentamethylene glycol 0-43-0.48 imidazole tetramethylene glycol 0-90-1.01 n-butylamine *c tetrahydrofurfuryl alcohol 0.38-0.50 lithium bromide *d 1 : 4-cyclohexanediol 0.45-0'52 anions 1 : 2-cyclohexanediol 0.21 -0.43 minimum inhibiting concentration (mole/l.) 0.12-0.1 6 0.1 4-0.20 0.31-0.36 0.36-0.72 0.22-0'34 neui urea methylurea semicarbazide formamide sodium ascorbate *e 0.44-0.45 0.36-0.42 sodium tetrathionate 0.1 2-0.14 0.42-0.56 sodium thiocyanate 0.24-0'32 0'26-0.34 sodium sulphite *f 0.19-0.38 1.01-1.35 sodium bromide *g 0.34-0.45 sodium iodide *JJ 0.22-0.34 sodium benzenesulphonate 0.33-0-36 'ral amides neutral dipolar ion cys teine*a 0'34-0.42 * at a total ionic strength of : a, 0.45-0.56 ; b , 0 3 1-0.53 ; c, 0-63-0.8 1 ; d, 0.56-0-62 ; e, 0-89-0.90 ; f, 0.98-1.51 ; g, 0-62-0.68 ; h, 0-56-0-62.SIDNEY SHULMAN 111 The effectiveness of an inhibitor (excepting the inorganic ions) is markedly influenced by ionic strength.6 The minimum inhibiting concentration becomes smaller with an increase in ionic strength.This increased effectiveness can be accounted for in terms of an enhanced binding of inhibitor by fibrinogen, and in general, a decreased solubility of the inhibitor. It is essentially a salting-out effect. Before attempting to explain the action of the inhibitors, one may consider the influence of pH and ionic strength on clotting.7.4 At constant pH, say 6.2, in- creasing concentration of sodium chloride causes an increasing clotting time, but inhibition cannot be obtained by this means.At the same time the nature of the final clot shifts from opaque to transparent. Both of the criteria demonstrate a TABLE 2.-RETARDERS OF FIBRINOGEN CLOTTING Fibrinogen, 5.0 g/l. ; thrombin, 1.0 unit/ml ; pH 6.20 ; ionic strength, 0.45. reagent maximum relative concentration clotting time tested (mole/l.) increase t neutral alcohols trimethylene glycol 0.80 2.8 1 : 3-butanediol 0-43 11.8 glycerol 2.16 2.6 mannitol 0.44 1.8 ethanol 1 -09 2.0 ethylene glycol 3.14 4.3 cyclohexanol $ 0.19 11.2 biuret $ glycine taurine histidine *a lysine *b methylamine neutral amide 0.14 25.4 neutral dipolar ions 0.72 0-22 1.27 1.3 cat ions 0.41 1.5 0-76 21.8 0.36 2.7 anions sodium thiosulphate *c 0.35 5.8 sodium m-benzenedisulphonate *d 0.32 3.7 sodium isopropylsulphonate *e 0.57 5.1 sodium chloride*f 1.72 56.0 * at a total ionic strength of: a, 0.50; b, 0.83; c, 1.11 ; d, 0.77; e, 1-02; f, 1.72.$ concentration limited by reagent solubility. a factor of 5 corresponds to a clotting time of about 1 h. reduced tendency toward aggregation, and can be explained by the decreased activity coefficient of a single fibrinogen molecule relative to that of an aggregate caused by increasing salt concentration, as is also demonstrated by the effect of salt addition on the solubility of the protein. The effect of pH is a little more complicated. The isoelectric point of fibrinogen is about pH 5 and the protein has invariably been studied at pH values alkaline to this.At the lower values of pH, the clot is opaque, while at higher values it is transparent. This is illustrated in fig. 1. The actual pH of transition depends on the ionic strength. No com- pletely satisfactory explanation of this complicated behaviour has yet been offered, and we must presumably wait until there is some understanding of the distribution of ionizable groups on the fibrinogen molecule.112 CONVERSION OF FIBKINOGEN TO F l B R I N The effect of pH on clotting may now be considered. It has been found that a clotting time minimum occurs in the vicinity of pH 7, depending somewhat on FIG. 1 .-Clot opacity plotted against pH for 5.0 g/l. fibrinogen and 1.0 unit/ml thrombin. The figures denote values of ionic strength I 0.8 0.6 0.4 0.2 0 5 6 7 8 9 10 PH the ionic strength, and that clotting is progressively delayed at higher and lower values of pH, until inhibition is obtained at pH 5.3 and at pH 10.This is illus- 3 .O 2 .o In 3 c .- E + 0 .- c FIG. 2.-Logarithm of clotting g 1.0 time plotted against pH for 5.0 g/l. fibrinogen and 1.0 unit/ml thrombin. The figures denote values of ionic strength. - 0 5 6 7 8 9 10 PH trated in fig. 2. To account for this behaviour, two mechanisms were postulated.7 The clotting delay caused at high pH can be explained in terms of the increasing electrostatic repulsions between the more highly charged fibrinogen units. At lowSIDNEY SHULMAN 113 pH one can assume that some ionizable groups in one, or both, of the proteins is involved in the reaction.Considering the pH region in question, the only likely group is the irnidazole ring of histidine. According to this hypothesis, the neutral form of imidazole is reactive, while the addition of a proton produces an inactive component. Inhibition by acid was studied by Laki and co-workers, who were the first to establish the presence of a definite intermediate stage in the clotting of fibrinogen.gY9 By neutralizing acid-inhibited mixtures of fibrinogen and thrombin after various time intervals, it was shown that some transformation was occurring in the acid environment even though no clotting took place, for the neutralized mixture clotted faster, the longer it had first stood at low pH. The action of inhibitors other than acid or base can be interpreted in terms of the reagent being bound to protein, producing an equilibrium between free and 70 60 u) 5 0 L 1 0 r .c 40 E Q) .- c g 30 * 20 .- c c 0 - 10 0 10 20 30 4 0 5 0 Concentration of hexamethylene glycol in g./l.FIG. 3.Xlotting time plotted against concentration of hexamethylene glycol for 5.0 g/l. fibrinogen, 1.0 unit/ml thrombin, ionic strength 0.45, and pH 6-20. bound protein. The moderately high reagent concentration required shows that the binding is rather weak. At low concentrations of inhibitor, and at all retarder concentrations, clotting is delayed and opacity is reduced. This opacity result shows that the reagents interact primarily with fibrinogen. This conclusion is also supported by the phenomenon of post-thrombin inhibition, and by the results of physicochemical investigations 10-13 on certain inhibited clotting systems, re- vealing the presence of partially polymerized fibrinogen.The difference between inhibitors and retarders is not just one of degree of effectiveness. The basic difference is well illustrated by fig. 3 and 4, which show the clotting time as a function of concentration for a typical inhibitor, hexamethy- lene glycol, and a typical retarder, glycerol. It will be seen that the curves differ radically in shape. As already mentioned, a certain intermediate in the poly- merization process is found in inhibited systems. Preliminary studies on retarded clotting systems 14 seem to reveal a different intermediate, and this, if confirmed, would further demonstrate the fundamental difference in the two types of inter- ference with clotting.114 CONVERSION OF FIBRINOGEN TO FIBRIN An analysis of the structure of the inhibitor molecules reveals several general principles. It will be profitable to consider the organic and inorganic reagents separately.Among the organic inhibitors, the rather rigid structural requirements suggest that the inhibitor is bound to the protein at or near the reaction site, so that this site is blocked. The reagent structure requires a polar functional group, such as hydroxyl, sulphydryl, amino, or sulphonate, along with a minimum quan- tity of hydrocarbon residue. This is best illustrated among the alcohols. The terminal glycols containing 4-7 carbon atoms are inhibitors. If the quantity of hydrocarbon moiety is reduced, as in trimethylene or ethylene glycols, or if the number of hydroxyl groups is increased, as in mannitol, effectiveness is lost.Moving the hydroxyl groups in from the end of the cliain-compare 1 : 3-butanediol with tetramethylene glycol-removes effectiveness. This observation, along with the fact that the monohydric tetrahydrofurfuryl alcohol is an inhibitor, suggests that the glycols may, in interacting with the protein, assume a cyclic configuration with the two hydroxyl groups functioning essentially as a single unit. Confirmation of this concept would be sought in the behaviour of cyclohexanol; unfortunately, the concentration range is limited by its low solubility, but at the highest concen- tration studied it compared fairly well with hexamethylene glycol in delaying clot- ting.Inhibition was, however, readily obtained with 1 : 2-cyclohexanediol and 0 5 0 100 150 200 Concentration of Qlycerol in g./l. FIG. 4.-Clotting time plotted against concentration of glycerol for 5.0 g/l. fibrinogen, 1.0 unit/ml thrombin, ionic strength 0.45, and pH 6-20. 1 ~Ccyclohexanediol. Further support for this explanation comes from the results on the organic electrolytes, and only a brief discussion will be included here. Among the sulphonates, for example, several inhibitors were found, e.g. ben- zenesulphonate, p-toluenesulphonate, several xylenesulphonates, but neither iso- propylsulphonate, taurine, nor even rn-benzenedisulphonate are effective. In- hibition is clearly not due to the functional group alone, but depends also on the large hydrocarbon residue.This view is further strengthened by the finding of Mihalyi 15 that sodium taurocholate delays clotting by a factor of about 150 at 0-05 M, and that glycocholic acid has a similar action. Glycine and taurine, how- ever, are not inhibitors. The amides must function by a different mechanism. Urea and its derivatives do not require a hydrocarbon residue for effectiveness. It should be noted that of more than a dozen amides studied all were found to be real or potential inhibitors, the latter, e.g. biuret, being limited by solubility. It seems likely that their action is due to hydrogen bonding with the protein, causing very slight and reversible modifications that block or alter the reactive groups. The action of cysteine is probably similar to that of the alcohol reagents, with the sulphydryl group playing a role similar to that of a hydroxyl group.Its possible action as a reducing agent was ruled out ; 3 the same conclusions were reached for sodium ascorbate and sodium sulphite. The inorganic anions, bromide, iodide, thiocyanate, tetrathionate, as well as perhaps the other effective anions would seem to be bound to fibrinogen, increasing the negative net charge, just as does an increase in pH. The augmented repulsionSIDNEY SHULMAN 115 would be expected to diminish the aggregation of fibrils and would thus produce a transparent clot, as observed for low concentrations of these inhibitors at pH 6.2, or for high pH itself. The effect of these reagents on clotting is to delay it markedly, whereas an increase in pH from 6.2 to, say 7.2, markedly hastens clotting.This difference can be readily reconciled with the theory already mentioned, concerning the importance of the state of ionization of the imidazole group in this pH region. As Edsall and Lever 4 have pointed out, the binding of anions at pH 6.2 would still leave the imidazole groups largely in the unreactive acidic form, but the con- comitant increase in negative net charge would retard clotting. As for the action of some of the cations, a different mechanism must be pro- posed. Edsall and Lever,4 by studying the action of guanidine at pH 7.3 and ionic strength 0.135, found that it not only delayed clotting but produced a clot of higher turbidity than the control clot. This is the only reagent so far known to show this behaviour.They pointed out that by these two criteria the addition of guanidinium ions was exactly like the addition of hydrogen ions, that is, a shift to lower pH. This should not present any conflict with the theory of pH effect on clotting; one can postulate that a guanidinium ion can react with an imidazole residue just as does a proton. Independent evidence supporting this hypothesis is yet to be sought. Whether other cation inhibitors can be classed as similar to guanidine, for example, whether the opacity-increasing effect is duplicated, remains to be seen. One reagent that might show this behaviour, and if so, be explained by the same mechanism, is the lithium cation. Its strong interaction with fibrinogen and fibrin is demonstrated not only by the fact that lithium bromide inhibits clotting at a lower concentration than does sodium bromide, but also by the observation 5 that 3 M lithium chloride will dissolve fibrin formed from highly purified protein components, whereas 3 M sodium chloride is ineffective.With respect to histamine, imidazole, and n-butylamine, all of which demon- strate again the importance of a bulky hydrocarbon residue, the inhibitory action may possibly be due to a competition between the free reagent molecules and the corresponding imidazole or lysine side chains in fibrinogen. The result of putting a carboxyl group on the methylamine carbon atom is shown by glycine, which does not prolong but actually hastens clotting. This alteration is similar to the reduction in inhibitory effectiveness resulting from the analogous introduction of a carboxyl group in histamine to give histidine, as well as the loss of inhibition suffered by butylamine on the introduction of a carboxyl group to give lysine. Apparently these molecular alterations cause sufficient interference with the electrostatic attraction between the amino group and the negative protein site that binding is inadequate to cause inhibition. 1 Ferry and Morrison, J . Amer. Chem. SOC., 1947, 69, 388. 2 Ferry and Shulman, J. Amer. Chem. Soc., 1949, 71, 3198. 3 Shulman, Arch. Biochem., 1951, 30, 353. 4 Edsall and Lever, J . Biol. Chem., 1951, 191, 735. 5 Shulman and Katz, Fed. Proc., 1952, 11, 286. 6 Shulman, Herwig and Ferry, Arch. Biochem. Biophys., 1951, 32, 354. 7 Shulman and Ferry, J . Physic. Chem., 1950, 54, 66. 8 Laki and Mommaerts, Nafure, 1945, 156, 664. 9 Laki, Arch. Biochem. Biophys., 1951, 32, 317. 10 Shulman and Ferry, J . Physic. Chem., 1951, 55, 135. 11 Shulman, Ehrlich and Ferry, J. Amer. Chem. Soc., 1951, 73, 1388. 12 Foster, Samsa, Shulman and Ferry, Arch. Biochem. Biuphys., 1951,34,417. 1 3 Ehrlich, Shulman and Ferry, J. Amer. Chem. Soc., 1952, 74, 2258. 14 Ferry and Shulman, unpublished experiments. 15 MihAlyi, Hung. Acta Phusiol., 1948, 1, 179.