Fibrinogen and Fibrin Structure and Fibrin Formation Measured by Using Magnetic Orientation

Proc. Nati Acad. Sci. USA Vol. 80, pp. 1616-1620, March 1983 Biophysics

Fibrinogen and fibrin structure and fibrin formation measured by using magnetic orientation (Cotton-Mouton effect/magnetic birefringence/secondary structure/fiber structure/ldnetics of polymerization) J.-M. FREYSSINET*, J. TORBETt, G. HUDRY-CLERGEON*, AND G. MARETt *Laboratoire d'H6matologie, Unite 217 Institut National de la Sante et de la Recherche Mdicale, D6partement de Recherche Fondamentale, Centre d'Etudes Nuclkaires, 85 X, F-38041 Grenoble Cedex, France; and tHochifeld-Magnetlabor des Max-Planck-Institutes fir Festkorperforschung, 166 X, F-38042 Grenoble Cedex, France Communicated by Manfred Eigen, December 21, 1982 ABSTRACT Accurate birefringence measurements show that packwith three-dimensionalorder, probably in atetragonal unit fibrinogen orients to a small degree in high magnetic fields. This cell with a = b = 185 A and c = 446 A and containing eight mol- effect can be explained as due to the molecule having about 30% ecules (14). The mechanism of assembly of fibrin has been ex- (by weight) a-helix oriented relatively parallel to the long axis. Bi- tensively studied (2, 15-18) but due to its complexity and the refringence measurements on fully oriented fibrin suggest that limitations of the experimental investigations many important aligned a-helical content is less than that estimated for fibrinogen. points remain to be elucidated. It is important to have as precise But because of limitations in the analysis this difference must be knowledge as possible about the fibrin polymerization process; viewed with caution. Highly oriented fibrin results when poly- not only is it of interest in itself but also it may be an excellent merization takes place slowlyin a strong magneticfield. Low-angle model for other aggregation processes of biological macromol- neutron diffraction patterns from oriented fibrin made in the ecules (19). presence of EDTA, made in the presence of calcium, or stabilized the orien- with factor XIIIa are very similar, showing that the packing of the The present study shows that measuring gradual molecules within the fibers is the same or very similar in these dif- tation of fibrin in a high magnetic field reveals kinetic features ferent preparations. The induced magnetic birefringence was used of the polymerization reaction, and it also demonstrates that to follow fibrin formation under conditions in which thrombin was structural details of the secondary structure of the fibrinogen rate limiting. The fiber network formed by approximately the ge- molecule or the fibrin monomer can be assessed by studying lation point constitutes a kind of matrix or frame that is largely their behavior in the field. built upon during the remaining ==85% of the reaction. After gelation the reaction is pseudo-first order. MATERIALS AND METHODS The arrest of blood loss from an injured vessel, hemostasis, re- Protein Preparation. Purified bovine fibrinogen (>98% quires the participation of several plasma proteins and also clottable protein) was obtained as described in ref. 20. Unless platelets, cells that form occlusive aggregates at the site of the specified otherwise the experiments were performed in 0.05 M rupture. The last stage of the blood clotting process is the en- Tris'HC1 buffer containing 0.1 M NaCl, 0.5 mM EDTA, and zyme-catalyzed activation of a soluble plasma protein, fibrin- 0.01% (wt/vol) NaN3 (pH7.5). Thrombin and reptilase (Both- ogen, which then undergoes polymerization to form an insol- rops atrox serine proteinase) were purchased from the Institut uble fibrin gel, thus mechanically reinforcing the platelet plug. de Serotherapie Hematopoietique (Paris) and Laboratoire Stago The limited cleavage of fibrinogen by thrombin, a serine pro- (Asnieres, France), respectively. All measurements were made teinase, is the result of a series of steps involving many other at 200C. clotting factors; much is known about this sequence of highly Samples for neutron diffraction experiments were made by regulated events (for a recent and exhaustive review see ref. 1). forming fibrin (polymerization time 1 hr) in a magnetic field Thrombin also converts factor XIII into factor XIIIa, the plasma that had an average value of about 15 teslas (1 tesla, T, = 104 transglutaminase which, in the presence of calcium, crosslinks gauss). The orientation was performed in lH20 buffer, which adjacent fibrin monomers of a fiber by forming E-(y-gluta- was subsequently replaced as required with 2H20 buffer by dif- myl)lysyl pseudo peptide bonds (2). fusion. The sample thickness was 0.1 or 0.2 cm. The trinodular elongated (450-A-long) structure for the fi- Measurements of the Magnetically Induced Birefringence. brinogen molecule proposed by Hall and Slayter (3) is the most The samples were contained in quartz cells that had an optical widely accepted model, and it has obtained additional support path length of3, 1, or 0.1 cm. These were placed in a temper- from recentwork on native fibrinogen (4-8) or slightly modified ature-stabilized (±0.1C) sample holder within a Bitter type fibrinogen (9-11). Fibrin monomers are produced by thrombin, magnet (maximal field 13.5 T) that had a small radial optical which releases the small negatively charged fibrinopeptides A bore. The magnetic birefringence An was sensitively measured and B. The monomers associate in a longitudinal half-staggered (resolution An 10-10, A = 6,328 A) by using a combined pho- arrangement to generate the two-stranded fibrin protofibril (12, toelastic modulation and compensation technique (21). Polar- 13), then these protofibrils associate laterally to form the thicker izer and analyzer were crossed and at 450 with respect to the fibrin fibers (12). In a recent study, we have shown that when field direction. A 50-kHz modulation of the birefringence was polymerization of fibrin takes place slowly in a high magnetic produced by a photoelastic modulator, resulting in a 100-kHz field one ends up with a highly oriented gel on which neutron intensity modulation of the photodiode output. Any superim- low-angle diffraction studies demonstrate that the protofibrils posed steady-state (magnetic) birefringence produced an additional 50-kHz photodiode output, which was phase-sensi- The publication costs of this article were defrayed in partby page charge tively detected, converted to dc, and used (as error signal in a payment. This article must therefore be hereby marked "advertise- feedback loop) to compensate the steady-state birefringence by ment" in accordance with 18 U. S. C. §1734 solely to indicate this fact. means of a Pockels cell. Hence in the compensated case the 1616 Downloaded by guest on September 25, 2021 Biophysics: Freyssinet et al. Proc. NatL Acad. Sci. USA 80 (1983) 1617 voltage across the Pockels cell was a direct measure of the mag- Lorentz-Lorenz formula netic birefringence An. In order to measure the large birefringence given by fibrin polymerization the Pockels cell was re- 2ir (n2 + 2)2 CNA An, = -a. [3] placed by an electrically driven Babinet-Soleil compensator. 9 no Mr Neutron Diffraction Measurements. The neutron diffraction patterns were obtained on the small-angle scattering cam- The Lorentz-Lorenz formula does not have afirm theoretical era D 11 (22) at the Institut Laue-Langevin (Grenoble). The basis for solutions or gels of anisotropic particles, but because scattered neutrons were detected on a two-dimensional (64 X the analysis is based on comparison with measurements of known 64) BF3 multidetector. The wavelength used was 10 A and AA/ structures this is not a handicap. A was 8% (full width, half maximum). The specimen-to-detector Now consider that polymerization and orientation occur si- distance was 2.55 m. Awater spectrum, which is isotropic under multaneously and that the An from the oriented polymer con- these conditions, was used to correct for detector response. centration, c2(t), is large compared to that of the unpolymerized material concentration cl(t). Both concentrations are dependent Theory on time, t, and add up to the total concentration c = cl(t) + c2(t). At the beginning (t = 0), c1(O) = c, c2(0) = 0, and An(O) 0, We are interested in a dilute solution of elongated molecules and at the end of polymerization (t -m oo), cl(oo) = 0, c2(oo) =c, that after limited proteolysis polymerize to form large fibers. and An(oo) = An,. Consider the molecules to be rotationally symmetric about their At intermediate times long axis. The following quantities are defined: c = concentration; Mr = molecular weight; NA = Avogadro's number (cNA/ A'~t' - 21r(n- + 2)2 C2(t)NA Mr = number of molecules per unit volume); k = Boltzmann 9n. MrA, [4] constant; T = absolute temperature; A = wavelength of light; no = refractive index, which at lowconcentration is equal to that which, when combined with Eq. 3, gives of water (1.33); H = magnetic field strength; and Aa = all - a1, the optical anisotropy, and Ax = All - Xl, the diamagnetic An(t) = - c2(t) and Ans - An(t) = -s cl(t). [5] anistropy-i.e., the difference in their values parallel and per- C C pendicular to the axis of symmetry (XI and X1 are always negative). The induced birefringence, An, is used to monitor the Thus An(t) is directly proportional to the polymer concen- orientation, An = nil - no, the difference in the refractive in- tration. This is valid even if the alignment is not complete, pro- dices of light linearly polarized parallel and perpendicular to vided the degree of orientation is the same throughout and, as the applied magnetic field direction. An, = birefringence at full above, the signal from the unpolymerized material is relatively orientation; in this study the symmetry axis orients parallel to weak, in which case Ans is not given by Eq. 3 but is equal to the the field direction. birefringence at the end of the reaction. For a first-order or In a magnetic field diamagnetically anisotropic particles ex- pseudo-first-order reaction c1(t) = cekt, in which k is the ap- perience a torque that acts to bring the axis of smallest absolute parent rate constant. Hence from Eq. 5 diamagnetic susceptibility nearer to the field direction. This An, - An(t) = Ane-t. [6] means that if Axis positive, I,,ii < Ix±,I the Xm axis moves nearer to the field direction, whereas if AXis negative, IxllI > Ix-I, the Comparative measurements All axis moves away from the field direction. The magnetic orientation is opposed by the randomizing effect of Brownian mo- The ratio of the birefringences of two aligned samples (sub- tion. In general the magnetic orienting energy AVH2/2 < kT, scripts 1 and 2) at the same concentration is calculated by using so only weak orientation is attained. However, because dia- Eq. 3: An,1/An,2 = Mr2Aa1/Mr1Aa2. If both sorts of particles magnetic anisotropy is an additive property, if N particles are have the same structure and flexibility (or more generally the aligned as in a polymer with their symmetry axes parallel, then same Aa per unit mass) and differ only in length, then this ratio their total anisotropy is NAX(for a stiffpolymer) and ifN is large is equal to 1. However, if particle 1 has a fractionf that has the enough the orienting energy will be significantly larger than kT. same Aa per unit mass as particle 2 and its remaining fraction For better than 90% orientation parallel to the field (positive (1 - J) has zero optical anisotropy, then AX) NAXH2/2 > 16kT. For weakly orienting samples (AXH2 < 0.6kT) the induced An,, birefringence is given by (An << Ans). Ans2= [7] 21r (n 2 + 2)2 H2 CNA An = - - AXAa. [1] Similarly, making use of Eq. 2, the ratio of the Cotton- 135 no kT Mr Mouton constants is given by CM1/CM2 = Mr2AalAX1/ Mnj Aa2AX2. For particles that differ only in length (or gen- When An is plotted against H2 a straight line results; from this erally if they have the same AaAX/M2) this simplifies to CM1/ line the Cotton-Mouton constant, CM (CM = An/AH2), is ob- CM2 = MrI/Mr2. Again, if particle 1 has a fraction fwith the tained same structure as particle 2 and the diamagnetic and optical an- of the remainder are then 21T (n2+ 2)2 1 CNA isotropies 0, CM = -AXAa. [2] CM1 135 no AkT Mr -: Mrlf2 [8] CM2 Mr2 Thus CM is proportional to the product AaAX. Ifit is positive the direction of maximal susceptibility is one ofminimal polar- Above it is assumed that the Aa values and Ax values are izability, and if it is negative the directions ofmaximal suscep- additive. This is justifiable for the latter, but it is not so reliable tibility and polarizability coincide. for Aa because it may have both a shape, Aa,, and intrinsic, The birefringence ofa fully aligned sample is according to the AaI, component and to a first approximation Aa = Aa, + Aa,. Downloaded by guest on September 25, 2021 1618 Biophysics: Freyssinet et al. Proc. Nad Acad. Sci. USA 80 (1983) RESULTS AND DISCUSSION lowing assumptions implicit in the analysis. First, that the ar- Structure of Fibrinogen in Solution. The magnetically in- omatic residues in the a-helices contribute little, as in the phages, duced birefringence of fibrinogen in solution (Fig. 1) varies lin- to the optical and magnetic anisotropies and, second, that the early (Eq. 2) with H2 up to the highest fields used. No saturation contribution from the rest of the molecule is small. The latter was observed, indicating that only partial orientation takes place. assumption is reasonable because these regions are expected to The birefringence is positive, as is that of oriented fibrin (see be rather globular and therefore unlikely to be very anisotropic. later), so it is reasonable to conclude that the long axis, which Structure of Oriented Fibrin Gels. When fibrin is formed is the axis of rotational symmetry of the fibrinogen molecule, in a strong magnetic field it can be aligned parallel to the field tends to orient in the direction of the magnetic field, as do the (14). At near complete orientation as checked by neutron dif- monomers in fibrin (14). This is thus the axis of smallest nu- fraction, Ans/c is 1.24 (± 0.07) x 10-' ml mg-'. The corre- merical diamagnetic susceptibility and Ax is positive, as is Aa sponding values for Pf1 and fd are, respectively, 6.27 x 10-5 from Eq. 2. and 6.05 x 10-5 ml mg-' (23). Thus from Eq. 7,f = 0.20 when The specific Cotton-Mouton constant (i.e., CM divided by fibrin is compared with these two viruses. This suggests that the concentration) of fibrinogen is 1.58 (± 0.1) X 10-7 T-2 cm2 about 20% of fibrin is a-helical with an axis relatively parallel mg'1. It is independent of the protein concentration in the range to the fiber axis as in Pf1 and fd. The assumptions are the same 10 to 80 mg/ml and of the salt concentration between 0.05 and as those invoked for the interpretation ofthe CM offibrinogen. 0.45 M NaCl. These results demonstrate that over these ranges Because in both cases the analysis is not rigorous one must be of conditions there is little intermolecular interaction or sig- cautious about attributing significance to the difference be- nificant change in conformation. tween the estimates of the aligned a-helical contents offibrin- It is useful to compare the CM of fibrinogen with that of two ogen and fibrin. filamentous bacteriophages (23), Pf1 and fd, whose structures The low-angle neutron diffraction pattern given by magnet- are well characterized (24-26). The ratio of the specific CMs is ically oriented fibrin made in the absence of calcium has been 8.0 x 10-4 (fibrinogen/Pfl) and 16.2 X 10-4 (fibrinogen/fd), reported (14). In Fig. 2 the low-angle diffraction patterns can whereas the ratios of the molecular weights are, respectively, be compared for fibrin made without calcium, fibrin with cal- 9.07 x 10-3 and 2 x 10-2. From Eq. 8, f = 0.30 or 0.28, de- cium, and fibrin covalently linked by factor XIIIa. The positions pending on whether the comparison is with Pfl or fd. In order of the maxima are identical, therefore the packing or unit cell to assess the significance of this result it is necessary to refer to the structures of the phages. Both viruses are long (Pf1 22m, A B fd 0.9 /Lm) with a diameter of about 60 A. They contain cir- cular single-stranded DNA (5% by weight in Pf1, 14% in fd) en- capsulated in a helical arrangement of coat protein monomers, which account for about all of the phage protein. In Pf14% and in fd 12% of the amino acids are aromatic. The coat protein is almost 100% a-helical; the helix lies at an average angle ofabout 200 to the phage long axis. Although the a-helices, aromatic amino acid residues, and DNA bases are all potential contributors, it turns out that the Cotton-Mouton effect comes largely from the a-helices, because the average orientation of the other groups relative to the phage long axis is such that they have little in- C H .D fluence (23). Thus as far as the magnetic birefringence measurements are concerned fibrinogen behaves as though there is about 30% (fl- 0.30, above) a-helix oriented relatively parallel to the molecule long axis. This value is similar to that from other _(E>A0050 sources (27-30) but is nevertheless circumscribed by the fol- C7~~~ A-1 .

2.0 C, FIG. 2. Neutron diffraction pattern from fibrin (10 mg/ml) ori- r_4 1.5 ented by polymerization in a magnetic field,H, of about 15 T. The con- tour levels are the same in A, B, and C. The samples were formed in 1H20 buffer, which was replaced by 2H20 buffer inA, B, and C, but not D. The sample thickness was 0.2 cm in A, B, and C and 0.1 cm in D. - 10 The variation in the buffer conditions and the units of thrombin used to initiate polymerization were as follows. A, 0.5 mM EDTA, 0.05 Na- tional Institutes of Health (NIH) unit/ml; B, 2 mM CaCl2, 0.018 NIH 0.5 unit/ml; C, 2 mM CaCl2 and 3% (vol/vol) bovine plasma as a source of factor XII, 0.018 NIH unit/ml; D, 0.5 mM EDTA, 0.05 NIH unit/ml. Exposure time ofpatternD was 3 times that of patternsA, B, or C. The positions ofthe peaks are the same inthefourpatterns. The meridional reflections have a 223 ± 5 Aperiodicity andthe sharp equatorial peaks 50 100 150 are at 184 ± 5 A (the very strong equatorial peaks nearest tothe center are due to leakage of the direct beam round the cadmium beam catcher). H2, T2 The fibrin fibers were not exactly perpendicular to the neutron beam and for this reason the meridional reflections are not symmetric in in- FIG. 1. Continuously measured magnetically induced birefrin- tensity about the center. This effect is particularly obvious inD, where gence, An, of a solution of fibrinogen (c = 20 mg/ml) at 220C. It took there are no meridional peaks in the lower half of the pattern. See ref. about 60 sec to sweep up the field and down again. 14 for a more detailed discussion of the diffraction results. Downloaded by guest on September 25, 2021 Biophysics: Freyssinet et al. Proc. Natl. Acad. Sci. USA 80 (1983) 1619 (14) is the same in all samples. Fig. 2 also shows that the re- placement of '1H20 buffer with 21H20 buffer has no effect on the peak positions but does greatly enhance the signal relative to the noise. Formation of Fibrin. As shown previously (14), when fibrin formation is initiated by adding arate-limitingamount ofthrom- bin to a solution offibrinogen in a strong magnetic field the vari- ~o4 ation of the induced birefringence is sigmoidal (e.g., Fig. 3). In X EDTA the early stages the signal is weak because fibrinogen molecules 2 and fibrin monomers and oligomers have a small diamagnetic anisotropy and consequently their magnetic orientation energy is small compared to kT. However, when large ordered aggre- ~. B gates are formed the magnetic orienting energy becomes suf- Thrombin ficiently large, due to the additive property of diamagnetic an- Reptilase isotropy, to give rise to significant orientation and, when there 4 are enough of them, to a large induced birefringence. The degree of orientation when polymerization is complete depends, 2 in otherwise fixed conditions (see below), on the magnetic field strength and the rate ofpolymerization. The final birefringence varies linearly with the square of the field strength under con- 10 20 30 40 50 ditions ofincomplete alignment. Reducing the rate ofpolymer- Time, min ization increases the final value of the birefringence for a fixed field strength until a limiting value is reached; this value can be FIG. 4. Variation with time of the specific birefringence as polymerization proceeds in a constant.field of 1 T. (A) For the upper curve improved with difficulty either by increasing the field strength the buffer contained 2 mM CaCl2 and polymerization was initiated at or by further reducing the rate of polymerization. Then the fi- t =0 with thrombin at0.05 NIH unit/ml. Forthe lowercurve the buffer brin fibers are fully aligned parallel to the field direction (14). contained 0.5 mM EDTA and thrombin was added at 0.2 NIH unit/ml. No systematic attempt was made to determine the minimal field The concentration offibrinogen was the same in bothsamples, 4.8 mg/ that would give rise to full orientation, although we find that it ml. Under these conditions the reaction proceeds too rapidly for full is 7 T or less. The dependence on the rate of polymerization alignmentto beattained. (B) The fibrinogen concentration was4.8 mg/ ml and 0.5 mM EDTA was present in the buffer. The polymerization shows that for good alignment the fibers must orient highly be- was initiated att = Oby thrombin (0.1 NIH unit/ml) orreptilase (amount, fore there is any substantial development of interlinking con- adjusted to give similar clotting time). Reptilase cleaves only the A fi- nections. If the connections form before fiber orientation the brinopeptides (see ref. 15). network is fixed in position at incomplete alignment. Increasing the ionic strength of the fibrinogen solution greatly thermore, as demonstrated above, better orientation is achieved reduces the magnetically induced birefringence and therefore when the rate of polymerization is reduced. Thus the fibers must the degree of orientation when fibrin is formed. Using 0.15 M orient before significant interlinkinghas developed (i.e., before NaCl instead of 0.10 M NaCl in the buffer results in almost no gelation). Combining the foregoing results, we conclude that birefringence and a transparent gel. Conversely; the addition of before gelation has occurred many protofibrils have already co- calcium markedly increases the induced birefringence in com- alesced to form fibers that are poorly interlinked. parison with fibrin formed with no calcium, at the same rate of Polymerization brought about by a rate-limiting amount of polymerization, and under conditions that do not give full align- reptilase gives rise to about the same induced birefringence as ment (Fig. 4A). Thus, because high salt concentration is known is obtained by using thrombin when the reaction rates are about to decrease the fiber thickness (16) whereas the presence of cal- equal (Fig. 4B). With the observations above this implies that cium is known to increase it (31), the simplest interpretation of reptilase action gives rise to fibrin fibers of appreciable thick- these results is that the fibrin fibers must be relatively thick for ness; as Hantgan and Hermans (16) found. useful magnetic orientation. The thicker the fiber the larger will In Fig. 3 (broken line) the data are plotted in accordance with be its diamagnetic anisotropy and the stiffer it will be. Fur- Eq. 6. After approximately the time of gelation (as estimated by eye in a control sample outside the magnet) a straight line re- *10 'o sults, which indicates that beyond this point the reaction follows AD pseudo-first-order kinetics. At gelation about 15% (i.e., c2/c = o. x 0.15, c = 4.8 in -5.0 - mg/ml, Eq.- 5) of the protein' in the initial so- x lution is involved in network formation. C-. bD Fig. 5 shows how the developmentofthe birefringence curve is influenced by cutting the field at different stages of the re- 0.OA action. Below a critical point, which is again near the time of C! *1.0 C. gelation, a marked reduction in the induced birefringence results, whereas when the field is cut a little later there is, sur- 0 prisingly, little effect. These observations can be explained as O + 10 20 30 40 5C follows. In the former circumstances once the field is cut new Field on Time, min fibers continue, to be formed with little preferred directionality. However, about the time of gelation an aligned network has de- FIG. 3. Change with time of the specific birefringence (left-hand veloped throughout and the orientation of this matrix or frame scale, -) as polymerization takes place in a constant field of 11 T. A logarithmic plot of these data is also given (right-hand scale, ---) is subsequently followed. If this were not so then, contrary to in accordance with Eq. & The buffer contained 0.5 mM EDTA and observation, on cutting the field after gelation the birefringence thrombin at 0.05 NIH unit/ml was added to the fibrinogen solution (c at the end of the reaction would be much less than that obtained = 4.8 mg/ml) at t = 0. when the field is maintained throughout. There are -probably Downloaded by guest on September 25, 2021 1620 Biophysics: Freyssinet et al. Proc. Nad Acad. Sci. USA 80 (1983) occasionally formed by very long protofibrils making lateral at- tachments to more than one fiber. From about the time of gelation the overall process follows pseudo-first-order kinetics. Our study also demonstrates that in the case of orientable polymers (biological ornot) measuring the magnetically induced

ba birefringence, which is proportional to the concentration of 6 ~ ...... monomers that have undergone polymerization, can be an excellent technique with which to follow the assembly process. J.T. thanks the Deutsche Forschungsgemeinschaft for support. We 10 20 30 40 50 are indebted to the Service National des Champs Intenses (Centre Na- Field on Time, min tional de la Recherche Scientifique) and to the Institut Laue-Langevin, both in Grenoble.

FIG. 5. Variation with time of the specific birefringence as poly- 1. Jackson, C. M. & Nemerson, Y. (1980) Annu. Rev. Biochem. 49, merization proceeds in a constant field of 11 T, which was cut at dif- 765-811. ferent times during the reaction. , Field on during whole course of 2. Doolittle, R. F. (1981) in Haemostasis and Thrombosis, eds. Bloom, the reaction; . , field switched off at time t3 (when the field was A. L. & Thomas, D. P. (Churchill Livingstone, London), pp. 163- switched off at any time above t2 a similar curve was obtained); ----, 191. field switched offjust below t2. When the field was switched off at time 3. Hall, C. E. & Slayter, H. S. (1959)J. Biophys. Biochem. Cytot 5, ti the signal decreased somewhat and then stabilized. It took about 30 11-15. sec to reduce the field from 11 to 0 T. The fibrinogen concentration was 4. Fowler, W. E. & Erickson, H. P. (1979)J. Mol Biol 134, 241-249. 4.8 mg/ml, and 0.5 mM EDTA and thrombin at 0.085 NIH unit/ml were 5. Telford, J. N., Nagy, J. A., Hatcher, P. A. & Scheraga, H. A. (1980) added at t = 0. Proc. Natl. Acad. Sci. USA 77, 2372-2376. 6. Estis, L. F. & Haschemeyer, R. H. (1980) Proc. Nati Acad. Sci. USA 77, 3139-3143. relatively few free ends in the gel, so the principal process of 7. Price, T. M., Strong, D. D., Rudee, M. L. & Doolittle, R. F. (1981) growth is predominantly due to the lateral attachment of mono- Proc. Nati Acad. Sci. USA 78, 200-204. mers, dimers, or oligomers to existing fibers in a way resem- 8. Williams, R. C. (1981) J. Mol. Biol 150, 399-408. 9. Tooney, N. M. & Cohen, C. (1977) J. Mol Biol 110, 363-385. bling crystallization. If this is so then few new fibers can form 10. Weisel, J. W., Warren, J. S. & Cohen, C. (1978)J. Mol Biol 126, after gelation. However, if very long protofibrils continue to be 159-183. produced some of them may attach laterally to more than one 11. Hewat, E. A., Tranqui, L. & Wade, R. H. (1982)J. Mol Biol 161, fiber, giving rise to new interlinks or fibers while nevertheless 459-477. largely conforming to the orientation of the existing matrix. 12. Ferry, J. D. (1952) Proc. Natl Acad. Sci. USA 38, 566-569. Thickerfibers will grow more quickly because they have a larger 13. Fowler, W. E., Hantgan, R. R., Hermans, J. & Erickson, H. P. (1981) Proc. NatL Acad. Sci. USA 78, 4872-4876. surface to which additions can be made. So the final form of the 14. Torbet, J., Freyssinet, J.-M. & Hudry-Clergeon, G. (1981) Na- gel depends on steps preceding gelation. This could be one of ture (London) 289, 91-93. the factors controlling the size of a clot in vivo. 15. Blomback, B., Hessel, B., Hogg, D. & Therkildsen, L. (1978) Because above the critical point the polymerization and ori- Nature (London) 275, 501-505. entation of fibers follow the same kinetics with or without the 16. Hantgan, R. R. & Hermans, J. (1979) J. Biol Chem. 254, 11272- field and the coagulation times and the turbidities are very sim- 11281. 17. Collen, D., Vandereycken, G. & De Maeyer, L. (1970) Nature ilar, we conclude that the action of thrombin on fibrinogen mol- (London) 228, 669-671. ecules is not modified to a significant extent by the high mag- 18. Smith, G. F. (1980) Biochem. J. 185, 1-11. netic field. 19. Qosawa, F. & Kasai, M. (1962)J. Mol. Biol 4, 10-21. Hantgan and Hermans (16) have used stopped-flow light 20. Kekwick, R. A., Mackay, M. E., Nance, M. M. & Record, B. H. scattering to follow the events during polymerization in con- (1955) Biochem. J. 60, 671-683. ditions when thrombin was not rate limiting. When compari- 21. Maret, G. & Dransfeld, K. (1977) Physica B 86-88, 1077-1083. 22. Ibel, K. (1976) J. Appl Crystallogr. 9, 630-643. sons are made it is important to remember that in our experi- 23. Torbet, J. & Maret, G. (1981) Biopolymers 20, 2657-2669. ments thrombin concentration, on the contrary, was rate limiting. 24. Day, L. A. & Wiseman, R. L. (1978) in The Single-Stranded DNA Their measurements are informative about the early events but Phages, Cold Spring Harbor Laboratory Monograph Series, eds. reveal little about changes that occur later, whereas in our ex- Denhardt, D. T., Dressler, D. & Ray, D. S. (Cold Spring Harbor periments the converse is true. The scheme proposed by Hant- Laboratory, Cold Spring Harbor, NY), Vol. 8, pp. 605-625. gan and Hermans (16) is as follows: monomers polymerize to 25. Makowski, L., Caspar, D. L. & Marvin, D. A. (1980)J. Mol Biol 140, 149-181. form protofibrils; long protofibrils associate laterally to form fi- 26. Nave, C., Brown, R. S., Fowler, A. G., Ladner, J. E., Marvin, bers; occasionally a long protofibril associates with two different D. A., Provencher, S. W, Tsugita, A., Armstrong, J. & Perham, fibers to form an interfiber connection; finally, a network of in- R. N. (1981)J. Mol Biol 149, 675-707. terlinked fibers develops. Now, incorporating the results re- 27. Cohen, C. & Szent-Gyorgyi, A. G. (1957) J. Am. Chem. Soc. 79, ported here, we propose: monomers polymerize to form pro- 248. tofibrils; long protofibrils associate laterally to produce fibers, 28. Mihalyi, E. (1965) Biochim. Biophys. Acta 102, 487-499. 29. A. Z. Acta 663-671. interlinks are rare; occasionally a fiber or protofibril joins two Budzynski, (1971) Biochim. Biophys. 229, 30. Doolittle, R. F., Goldbaum, D. M. & Doolittle, L. R. (1978) J. different fibers and a fully interlinked network develops (ge- Mol, Biol 120, 311-324. lation occurs); and the matrix existing at gelation is built up by 31. Hantgan, R. R., Fowler, W. E., Erickson, H. P. & Hermans, J. the lateral growth of existing fibers, and possibly new fibers are (1980) Thromb. Haemostasis 44, 119-124. Downloaded by guest on September 25, 2021