Heterogeneity of bovine fibrinogen and fibrin.

Abstract We have attempted to relate the heterogeneity found when native bovine fibrinogen, which is known to have the subunit structure [Aα Bβ γ]2, was chromatographed on DEAE-Sephadex A-50, to electrophoretic differences among the constituent polypeptide chains and differences in the content of charged groups. Bovine fibrinogen, purified from the blood of individual animals, eluted from DEAE-Sephadex A-50 as three major peaks, each of which appeared heterogeneous. Two γ chains were distinguished by polyacrylamide gel electrophoresis of the reduced, S-carboxymethylated derivative of unchromatographed fibrinogen in 8 m urea at pH 8.6. Fibrin from the first major chromatographic peak contained only the more cationic γ chain. Fibrin from the second peak contained equal amounts of the two γ chains as judged by densitometry of stained gels. Fibrin from the third peak contained predominantly the more anionic γ chain. These findings indicate that separation into the three major peaks is due to the three possible combinations of two different γ chains in native fibrinogen. Since a Ferguson plot of the relative mobilities of the two γ chains versus gel concentration in 8 m urea at pH 9.5 yielded parallel lines, and since the two γ chains were not separated by polyacrylamide gel electrophoresis in 0.1% sodium dodecyl sulfate at pH 7.0, it is likely that the two γ chains differ in charge but not in size. Analyses of fractions within each of the major peaks suggest that heterogeneity within the peaks is due to differences in phosphate content, which varied from 0.6 to 3.8 moles of phosphate per mole of fibrinogen. These observations indicate that within an individual animal there are as many as 36 different fibrinogen molecules, differing either in the composition of γ chains or in the content of bound phosphate.

Bovine fibrinogen, purified from the blood of individual animals, eluted from DEAE-Sephadex A-50 as three major peaks, each of which appeared heterogeneous. Two y chains were distinguished by polyacrylamide gel electrophoresis of the reduced, S-carboxymethylated derivative of unchromatographed fibrinogen in 8 M urea at pH 8.6. Fibrin from the first major chromatographic peak contained only the more cationic y chain.
Fibrin from the second peak contained equal amounts of the two y chains as judged by densitometry of stained gels. Fibrin from the third peak contained predominantly the more anionic y chain. These findings indicate that separation into the three major peaks is due to the three possible combinations of two different y chains in native fibrinogen.
Since a Ferguson plot of the relative mobilities of the two y chains uersus gel concentration in 8 M urea at pH 9.5 yielded parallel lines, and since the two y chains were not separated by polyacrylamide gel electrophoresis in 0.1% sodium dodecyl sulfate at pH 7.0, it is likely that the two y chains differ in charge but not in size. Analyses of fractions within each of the major peaks suggest that heterogeneity within the peaks is due to differences in phosphate content, which varied from 0.6 to 3.8 moles of phosphate per mole of fibrinogen.
These observations indicate that within an individual animal there are as many as 36  terminus, and y with an NH-in-terminal tyrosine (I). Upon proteolysis by thrombin, the A and B peptides are released from the ALY and Bb chains, leaving the (Y, 6, and y chains of fibrin. The three polypeptide chains of fibrinogen are covalently linked by disulfide bonds (2). Recent work suggests that there are also disulfide bonds holding ALY and y chains together to give a dimeric structure with 2-fold symmetry (3).
A variety of techniques has been used to demonstrate heterogeneity in fibrinogen preparations.
In human fibrinogen, BlombSick el al. (16) showed by sequence analysis that fibrinopeptide A was heterogeneous. The Ser.3 was partially phosphorylated, and two NHz-terminal amino acid residues were found.
It has been pointed out (17) that if the phosphorylated and nonphosphorylated Aol chains had a random distribution in the fibrinogen dimer, one could formulate a number of symmetrical and asymmetrical variants.
The same proposition can be made in respect to heterogeneity of y chains and to other sites of phosphorylation, and thus, a large number of variants are possible.
We have devised a chromatographic procedure which separates native bovine fibrinogen into three major peaks, each of which itself appears heterogeneous, and have attempted to relate the heterogeneity noted on chromatography to heterogeneity at the subunit level and to differences in charged groups.
In this paper we present evidence that separation into the three major peaks is due to the presence of two different y chains which can be distinguished electrophoretically, whereas heterogeneity within the major peaks is due to differences in phosphate content. The distribution of protein among chromatographic fractions is similar to the distribution expected from random combination of these two types of determinants. further purified using the method of Glover and Shaw (18). U.S. standard thrombin, Lot B3, was used as a standard of activity (19). Chromatography columns were constructed from polymethylacrylate tubes in which the resin beds were supported by porous polyethylene discs.
Puri&ation of Fibrinogen-Fibrinogen from the plasma of in dividual cows was prepared using a modification of the method of Capet-Antonini (20). Eight parts of blood, spurting from the great vessels of a freshly killed animal, were mixed with 1 part of 0.1 M sodium-EDTA, pH 7.0. Cells were removed by two centrifugations at 5000 X g for 20 min. Barium sulfate, 100 g per liter, and c-ACA,~ 6.5 g per liter, were added, the mixture stirred for 1 hour, and the supernatant collected by centrifugation.
To the supernatant solid ammonium sulfate, 129 g per liter, was added at 4" with constant stirring.
The precipitate was collected by centrifugation, washed several times with 25% saturated ammonium sulfate, and dissolved in 0.05 M sodium phosphate, 0.01 M sodium-EDTA, pH 6.6. A second, identical ammonium sulfate precipitation yielded a tenacious, white precipitate which was dissolved in 0.01 M Tris, 0.1 M sodium chloride, pH 7.4, and dialyzed at 4" against 5 volumes of the same buffer. A small amount of precipitate was removed by centrifugation.
The solution was cooled to O&l", and e-ACA and ethanol were added to give 0.1 M and 7 y0 solutions, respectively. The precipitate was collected by centrifugation and dissolved in 0.01 M Tris, 0.01 M sodium-EDTA, 0.1 M sodium chloride, pH 7.4. The fibrinogen solution was frozen and stored at -30".
Centrifugation was done in a Sorvall RC-2B refrigerated cen-clottable proteins, fibrin was formed and eluted in lo-or 20.ml plastic syringes. A polyethylene disc cut to the internal diameter of the syringe was placed in the bottom of the barrel, and polyethylene tubing was fitted over the hub so that flow could be controlled with a clamp.
The syringe was placed upright and partially filled with 1 KIM Tris, 0.14 M sodium chloride, pH 6.9. The buffer was allowed to flow out until the level was just above the disc. The syringe was then filled with fibrinogen solution in Tris-sodium chloride buffer, and the fibrinogen was clotted with thrombin, 1 U.S. unit per ml. After 4 hours, Trissodium chloride buffer was layered on top of the clot, the top of the syringe was closed with a rubber stopper equipped with a plastic fitting and inflow tubing, and the clot was eluted with buffer from a reservoir until the thrombin concentration of the effluent fell to lop3 to lop4 U.S. units per ml. At room temperature and fibrin concentrations above 2 mg per ml, the bed of fibrin remained intact while buffer was passed through it at flow rates of 0.1 to 0.5 ml per min and pressure heads of 10 to 20 cm of water.
Syneresis of the eluted clot was induced by disturbing the clot with a spatula, the fibrin was dissolved in 1 M sodium bromide, 0.07 M sodium acetate, pH 5.0, and the soluble fibrin was dialyzed for 72 hours against at least six changes of 200 to 1000 volumes of 2 mM acetic acid. The clear fibrin solution was lyophilized. Fibrin was reduced and carboxymethylated using the method of Murano et al. (13). trifuge using GSA or GS3 rotors.
Protein precipitates were collected by centrifugation at 5000 x g for 20 min at 4". Protein solutions were handled in polycarbonate or polyethylene containers.
Dialysis tubing was prepared by heating the tubing successively at 80" in 0.1 M sodium carbonate, 0.1 M sodium-EDTA, 0.1 M sodium chloride, and distilled water. pH was measured at 25" using a Radiometer pHm 25 pH-Meter.
The pH values quoted for the various buffers are not corrected to the temperature at which the buffers were used.
Chromalography-Fibrinogen to be chromatographed was thawed and dialyzed at 4" against 0.05 M sodium phosphate, 0.01 M sodium-EDTA, pH 7.4. The solution was cleared of a small amount of precipitate by centrifugation and charged onto DEAE-Sephadex A-50 equilibrated with phosphate-EDTA buffer. The protein was eluted with a linear 0 to 0.20 M sodium chloride gradient made up in phosphate-EDTA buffer. Chromatography of large amounts of protein was done in a jacketed column, 5 x 80 cm, cooled to 3" with a Lo-Temptrol 154 constant temperature bath (Precision Scientific Co.). Smaller samples were chromatographed in columns, 1.2 x 25 cm, constructed so that four columns could be developed in parallel from the same gradient reservoirs.
The tops of these columns opened into a common chamber with a volume of approximately 5 ml, and the chamber was connected to the gradient reservoirs.
It was found that if the columns were packed simultaneously to the same height using the same batch of resin, the flow rates and the conductivities of the effluents during gradient elution were identical, and duplicate samples of purified fibrinogen from single animals gave identical elution profiles.
The A280 was measured with a Beckman DB spectrophotometer and conductivity with a Serfass conductivity bridge, model RCM 15Bl.
Fractions were pooled and concentrated by precipitating the protein with 33% saturated ammonium sulfate at 4".

Polyacrylamide
Gel Electrophoresis-Polyacrylamide gel electrophoresis in 0.1% sodium dodecyl sulfate at pH 7.0 was performed in 5% gels using the method of Weber and Osborn (21). Prior to electrophoresis, samples were incubated for 10 min at 100" in 4 M urea, 1 To sodium dodecyl sulfate, and 1 y0 P-mercaptoethanol.
Electrophoresis in 8 AI urea was performed in gels of the following composition: 0.25 mg per ml of ammonium persulfate, 8 M urea, and 0.40 M Tris-chloride, pH 8.6. Samples were dissolved in 10 M urea, 0.01 M sodiumEDTA, 0.1 M glycine, pH 9.6 (22), and subjected to electrophoresis for 135 min at 4 ma per gel. The reservoir buffer contained Tris, 0.6 g per liter, and glycine, 2.88 g per liter.
To determine relative mobility as a function of gel concentration, multiphasic Buffer System A of Rodbard and Chrambach (23) was used, the upper and lower gels being made 8 M in urea. Gels were stained with Coomassie brilliant bIue and destained by diffusion (21). Densitometer tracings were made on a PhotoVolt scanner.
Analytical Procedures-Fibrinogen concentration was determined spectrophotometrically, using an absorption coefficient at 280 nm of At.:; = 1.523 (24), or by the method of Lowry et al. (25), using unchromatographed fibrinogen as a standard. Clottable protein was estimated using the method of Laki (26). Phosphate and sialic acid assays were performed on protein solutions which had been dialyzed for 72 hours against at least six changes, 200 to 1000 volumes each, of 2 mM acetic acid. Phosphate was analyzed by the method of Bartlett (27). Samples were read at 830 nm with a Beckman DU spectrophotometer, using the near-infrared filter slide. Sialic acid was determined by the method of Warren (28) after hydrolysis in 0.1 N sulfuric acid at 80" for 1 hour, using N-acetylneuraminic acid as a standard.

Curbozymethy&ion
of Fib&-To remove thrombin and non-Ten liters of blood, obtained from a single cow, yielded 8 to 10 2 The abbreviation used is: E-ACA, c-aminocaproic acid.
Polyacrylamide gel electro-phoresis in sodium dodecyl sulfate (Fig. 1A) showed that 99% of the stain was in the triplet characteristic of the Aa, BP, and y chains of fibrinogen (29). In addition, there were two minor bands of lesser mobility (Components I and II) which were not completely eluted from the fibrin clot, and a doublet of very low mobility (Component III) which appeared to be the major nonclottable component.
On prolonged storage at -30" a small amount of material of greater mobility than the fibrinogen triplet (Component IV) appeared, indicating degradation.
The molecular weights of the four minor components were determined on a plot of electrophoretic mobility (relative to bromphenol blue) versu.s the logarithm of the known molecular weights of reduced albumin, immunoglobulin light chain, and immunoglobulin heavy chain. The molecular weights were: I, 108,000; II, 155,000; III, 280,000; and IV, 35,000. Because molecular weight standards larger than 68,000 were not used, the estimated molecular weights of Components I, II, and III are subject to more than the usual 10% uncertainty of this method (21). Polyacrylamide gel electrophoresis of reduced and carboxymethylated subunits in 8 M urea, pH 8.6, revealed that the ACY and BP chains moved as diffuse bands, whereas the y chain appeared as a doublet (Fig. 1B). In the fibrinogen preparation on which the chromatographic studies reported in this paper were performed, the proportion of the more anionic to the more cationic y chain (hereafter called yr and 72, respectively, as suggested by the IUPAC-IUB Commission of Biochemical Nomenclature (30)) was estimated by densitometry to be 0.36: 0.64. In this fibrinogen preparation there were 2.6 moles of phosphate and 7.2 moles of sialic acid per 340,000 g of protein.
The elution profile for this single animal fibrinogen preparation on DEAE-Sephadex A-50 is shown in Fig. 2. Three main peaks, each with several shoulders, were seen. The effluent was pooled and concentrated as indicated.
Of the protein, 38.6% was in Fractions 1 to 8, 41.3% was in Fractions 9 to 16, and 14.1% was in Fractions 17 to 21. Upon rechromatography, protein from each of the main peaks eluted at the same position in the gradient as originally (Fig. 3A). Adjacent fractions from the same peak did not elute coincidentally (Fig. 3B), suggesting that the shoulders seen within the main peaks represent heterogeneity within the peaks. During these studies, pH, ionic strength, and temperature were monitored closely. It was found that temperature variations of 3" caused dramatic artifacts in the elution patterns when shallow gradients were run. The scatter of the data in Fig. 3, A and B, corresponds to the temperature cycle of the cold room in which chromatography was done.
Polyacrylamide gel electrophoresis in sodium dodecyl sulfate of the reduced fractions from DEAE-Sephadex chromatography indicated that in all fractions intact Aa, BP, and y chains predominated (Fig. 4A). The small amount of degraded material (Component IV) eluted in Fractions 1 to 4. The high molecular weight, nonclottable contaminant (Component III) eluted with the fibrinogen in Fractions 16 to 21. Component III was estimated by densitometry to be 1.1, 5.5, 10.6, and 18.8% of the protein in Fractions 16, 18, 20, and 21, respectively.
The two minor components (I and II) which could not be eluted from fibrin were seen throughout.
Because the nonclottable component constituted up to 18.8% of the protein in the Peak III fractions, as judged by densitometer tracings of sodium dodecyl sulfate gels, we chose to study fibrin rather than fibrniogen when comparing the major peaks to one another.
Fractions from the three major peaks were pooled as in Fig. 2 and clotted. Elution of the fibrin clots com-B FIG. 1. Polyacrylamide gel electrophoresis of purified fibrinogen, fibrin, and clot liquor. A, sodium dodecyl sulfate gels: 1, purified fibrinogen, 20 pg; 2, fibrin solubilized after eluting the clot with buffer, 13 rg; 9, the initial effluent from the clot, concentrated by precipitating with 5y0 trichloroacetic acid, 1% acetic acid, 12.5 pg. The bands corresponding to the Aa, 01, BP, p, and y chains are assigned according to the data of McKee, et al. (29). The minor clottable components are labeled Z and ZZ, the major nonclottable component is labeled ZZZ, and the band of greater mobility than the y chain (very faint in these gels) is labeled IV. B, alkaline urea gels of carboxymethylated purified fibrinogen, 15 pg. The positions of the Aa and BP chains are assigned according to the data of Takagi and Iwanaga (22). The assignment of the y chains was proven by subjecting purified carboxymethylated y chains isolated on carboxymethylcellulose to electrophoresis using the method of Murano, et al. (13). The anode is towards the right; the direction of electrophoresis was from cathode to anode. The densitometer tracing is reduced to the dimension of the gel.
pletely removed the nonclottable component as assayed by sodium dodecyl sulfate gel electrophoresis.
The fibrin was reduced, carboxymethylated, and analyzed by gel electrophoresis in 8 M urea (Fig. 4B). Protein from the first major peak contained a single y chain, y2, protein from the second peak contained two y chains, yl and y2, and protein from the third peak contained predominantly yl. A Ferguson plot of relative mobility versus gel concentration (23) gave a set of parallel lines for the two y chains (Fig. 5).
Fractions from Peak I and Peak III were mixed and rechromatographed to see whether the two different y chains could hybridize to form Peak II material (Fig. 3A). After incubation at 37" for 20 min, there was no increase in the amount of protein eluting in the area of Peak II. Tubes were pooled into 23 fractions as indicated and concentrated for sodium dodecyl sulfate gel elec-To account for the heterogeneity seen within each of the major peaks, fractions within the peaks were analyzed for phosphate and sialic acid. In these experiments fibrinogen was not converted to fibrin. Therefore, the high molecular weight protein (Component III in the sodium dodecyl sulfate gels) contaminated fibrinogen in fractions from Peak III.
The sialic acid content of the various fractions ranged between 6.7 and 9.0 males/340,000 g of protein and did not vary in a systematic manner from fraction to fraction (Table I). In contrast, the phosphate content varied from 0.6 to 3.8 males/340,000 g, and fibrinogen from the ascending portions of the major peaks contained less phosphate than that from the descending portions (Table I).
Elution profiles similar to that shown in Fig. 2 have been found for fibrinogen preparations purified from five different cows and also for a preparation purified from commercially obtained bovine Fraction I using Laki's method (26). The phosphate content and y chain composition of fractions of two of the single animal preparations have been determined.
It is the data from one of these single animal preparations that are presented in Fig. 4B and Table I. The data from the other single animal preparation are similar. DISCUSSION Our studies indicate that within an individual cow there are many different types of fibrinogen molecules which can be partially separated chromatographically, and that these different molecules are a result of two types of heterogeneity, heterogeneity of y chains and heterogeneity of phosphorylation. A model is presented in Fig. 6 relating the elution profile to y chain heterogeneity and phosphate content.
It is based on the assumptions: (a) that there are two sites for phosphorylation per set of three polypeptide chains, or four sites per fibrinogen molecule; (b) that these sites are phosphorylated independently of one another; (c) that phosphorylation at the first site cannot be distinguished chromatographically from phosphorylation at the second; (d) that the two types of y chains randomly combine in trophoresis, sialic acid determination, and phosphate determination. For conversion to fibrin and subsequent carboxymethylation, protein was furt.her pooled into Fractions I, ZZ, and ZZZ, representing each of the major peaks. The increase in sodium chloride concentration at the first arrow represents 0.2 M sodium chloride applied when the gradient was 90% completed.
The column was stripped with 0.5 M sodium hydroxide, which broke through at the second arrow. The yield of protein was 95.4y0.
forming the dimeric structure; (e) that the chromatographic difference between the two types of y chains is approximately 5 times greater than the difference between a phosphorylated and nonphosphorylated molecule; and (f) that the separation effected by an incremental increase in chloride ion concentration is not constant throughout the linear gradient. The values for y1:y2 (0.36:0.64) and phosphate content (2.6 males/340,000 g) were determined analytically.
In this model there are eight different sets of three polypeptide chains of six different chromatographic valences. These combine to give 36 different kinds of fibrinogen molecules of 15 different chromatographic valences. The model agrees well wit#h the elution profile in Fig. 2, suggesting that heterogeneity of y chains and of phosphorylation account for all of the chromatographic heterogeneity and that these two types of determinants randomly combine. However, more complex models, in which there are more than two sites for phosphorylation, in which chromatographic determinants do not randomly combine, or in which there are other, as yet undetected, sources of heterogeneity (31), can be formulated. To decide among such models, additional information is needed. One must know: (a) how the two y chains differ from one another; (b) at what point in the lifetime of the fibrinogen molecules examined, synthesis, secretion, circulation, or purification, the combination of determinants occurs; and (c) if one component is selectively purified reIative to another during the isolation procedure.
Fibrinogen preparations from a number of animals should be examined.
Several reports in the literature support the conclusions of this investigation and suggest that our results, obtained with bovine fibrinogen, may also apply to humans and other species. Brada In contrast, Finlayson and Mosesson (9) reported that bovine fibrinogen eluted from DEAEcellulose as a single peak; however, the peak they observed was asymmetrical.
The relationship between y chain heterogeneity reported from other laboratories and that presented in the present paper is unclear. Mosesson  in 8 M urea for S-sulfa y chains from both their first and second peaks. They suggested that the chromatographic heterogeneity reported by Gerbeck et al. (10) for bovine fibrinogen was homologous to this second type of heterogeneity and not to the type of heterogeneity which accounted for the charge difference between Peaks 1 and 2 of human fibrinogen. We believe the type of heterogeneity of y chains that we have found is homologous to the y,y' heterogeneity of Mosesson et al (11) which accounted for the difference between Peaks 1 and 2 of human fibrinogen.
We do not know how our findings are related to those of Gerbeck et aE. (10).
The finding that the relative mobilities of the two y chains A discontinuous buffer system was used as described in the "Experimental Procedures." The concentration of the stacking gel was 2.5% (w/v), and the ratio of monomer to cross-linker was 20: 1. The + pH of the separating gel was 9.5. The data plotted are from an experiment using S-carboxymethyl fibrin from Peak II. Similar results were obtained using S-carboxymethyl fibrinogen derived from unchromatographed protein. l and 0 represent yl and 72, respectively. n represents the midpoint of the broad band containing 01 and p chains, which were not separated well in this system. Phosphate and sialic acid content of the pooled and concentrated fractions were determined as described under "Experimental Procedures" and are expressed as moles per 340,000 g of protein.

6901
Phosphate assays were performed in duplicate, + and sialic acid assays in quadruplicate. The average range of the sialic acid analyses was 2.3 males/340,000 g of protein. gel electrophoresis of chromatographically separated fibrinogen and fibrin. A, sodium dodecyl sulfate gels: the fraction (see Fig. 2) is indicated below each gel; 5 to 25 pg of protein were electrophoresed; the band assignments are the same as in Fig. 1A. B, alkaline urea gels: fibrin from Peak I, 15 pg; fibrin from Peak II, 17 rg; and fibrin from Peak III, 15 pg. The anode is towards the right; the direction of electrophoresis was from cathode to anode.
The densitometer tracings are reduced to the dimensions of the gels. Band assignments are the same as in Fig. 1B. are parallel over a range of polyacrylamide gel concentrations in 8 M urea, pH 9.5, indicates that the two chains differ in charge but do not differ grossly in size (23). The structural basis for the charge differences is not known. Gerbeck et al. ( Fig. 2. Band centers are placed so as to optimize agreement with peaks seen on chromatography (see Fig. 2) and with phosphate content (see Table I).
The molecules within each peak are listed above the peak and the phosphate content is noted in parentheses. We also considered the possibility that purified fibrinogen could those (0.46:0.44:0.10) which can be calculated assuming random dissociate into subunits of molecular weight 180,000 as described combination of nonphosphorylated and phosphorylated A pepby Capet-Antonini and Guinand (32), and that if these subunits tides, if the former accounts for 14s2 and the latter 352 of the were nonidentical, they could randomly reassociate to give sevtotal A peptides. era1 species upon rechromatography.
However, when we mixed In bovine fibrinogen the phosphate is confined to the parts fractions from Peaks I and III that should have been able to of the molecule which remain after thrombin proteolysis (36,37). give rise to Peak II material by such hybridization, only the Krajewski and Cierniewski (38) reported that the phosphorus original fractions were observed (Fig. 3A). This is in accord was located on the Acu and R/3 chains and occurred as phosphowith recent reports (33,34) in which subunits of molecular weight serine.
Their findings support our assumption that there are 180,000 were not found and with the report that the Aor and y two sites of phosphorylation per set of Aoc, I@, and y chains, chains are held together by disulfide bonds (3). and that these sites are incompletely phosphorylated. UlombSck el al. (16) demonstrated that a serine in human The functional and physiological significance of the chemical heterogeneity described liere isnot known.
We have seen the ably differs from species to species. As we have discussed, Mosesson and Finlayson (11) observed a ratio of 0.07 :0.93 for human fibrinogen, whereas we have observed a ratio of 0.36:0.64 for bovine fibrinogen. The finding of phosphate in secretory proteins probably reflects post-translational modification of these molecules (39). Several clinical rep0rt.s suggest that the amount of phosphate in human fibrinogen may vary depending on physiological and pathophysiological circumstances. Witt and Miiller (40) found that fetal fibrinogen purified from human cord blood contained significantly more phosphate than that purified from adults.
A detailed study of fibrinogen purified from a patient with a congenital dysfibrinogenemia (fibrinogen Baltimore) revealed that the A peptide from the defective fibrinogen was more completely phosphorylated than the A peptide from normal fibrinogen (41). Shainoff et al. (42) found a high content of phosphorylated A peptide in fibrinogen from a patient with hemangioma and low blood levels of fibrinogen. The same authors also reported a decrease in phosphorylated A peptide and an increase in nonphosphorylated A peptide during incubation of blood and plasma and suggested that fibrinogen may be fully phosphorylated when secreted and progressively dephosphorylated in the circulation. Our studies, by describing a method by which fibrinogen may be separated into several fractions and relating the heterogeneity noted in native fibrinogen to the occurrence of two different y chains and to differences in phosphate content, should facilitate further investigations into the physiological and pathophysiological importance of fibrinogen heterogeneity.