Human Fibrinogen Heterogeneities STRUCTURAL AND RELATED STUDIES OF PLASMA

The heterogeneity of human plasma fibrinogen manifested as differences in solubility reflects the presence of early catabolic intermediates, which are more soluble, have a longer thrombin-clotting time, and are of a lower molecular weight than the parent material from which they are formed. In this study the S-s&o subunits from fibrinogen fractions of low (I-4) and high (I-8 and I-9) solubility were compared. Separation of AC+ BP, and y chains was achieved by gradient elution chromatography on CM-cellulose. The tryptic peptide maps were characteristic for each type of chain. The only observable differences were in the maps of Aa chains; in those of I-8 and I-9 a considerable number of spots were absent or reduced relative to I-4. NH2-terminal analyses of isolated chains before and after thrombin treatment showed the characteristic residues of the Aa, BP, and y chains, respectively. Thus the NHa-terminal portions of all chains of I-4, I-8, and I-9 were intact. Molecular weight estimation of subunit chains was made by polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate under conditions in which differences in mobility due to the presence or absence of A or B peptides were distinguishable. Electrophoresis in this medium before and after reptilase or thrombin treatment served for identification of the chains. The y chains moved as a single band in this system (molecular weight, 50,700) and were indistinguishable from one another; the same was true of BP chains (molecular weight, 60,400). However, electrophoresis of Acu chains revealed the presence of 13 more-or-less distinct bands (numbered in order of increasing anodal mobility), of which all but two (Band 3, a BP contaminant and Band 5, a y chain contaminant) were shown to be intact Aar chain or COOH-terminally degraded ACY remnants. The molecular weights ranged from that of intact Aa: chain (70,900) to 15,400

intermediates, which are more soluble, have a longer thrombin-clotting time, and are of a lower molecular weight than the parent material from which they are formed. In this study the S-s&o subunits from fibrinogen fractions of low (I-4) and high (I-8 and I-9) solubility were compared. Separation of AC+ BP, and y chains was achieved by gradient elution chromatography on CM-cellulose. The tryptic peptide maps were characteristic for each type of chain. The only observable differences were in the maps of Aa chains; in those of I-8 and I-9 a considerable number of spots were absent or reduced relative to I-4.
NH2-terminal analyses of isolated chains before and after thrombin treatment showed the characteristic residues of the Aa, BP, and y chains, respectively.
Thus the NHa-terminal portions of all chains of I-4, I-8, and I-9 were intact. Molecular weight estimation of subunit chains was made by polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate under conditions in which differences in mobility due to the presence or absence of A or B peptides were distinguishable.
Electrophoresis in this medium before and after reptilase or thrombin treatment served for identification of the chains. The y chains moved as a single band in this system (molecular weight, 50,700) and were indistinguishable from one another; the same was true of BP chains (molecular weight, 60,400). However, electrophoresis of Acu chains revealed the presence of 13 more-or-less distinct bands (numbered in order of increasing anodal mobility), of which all but two (Band 3, a BP contaminant and Band 5, a y chain contaminant) were shown to be intact Aar chain or COOH-terminally degraded ACY remnants. The molecular weights ranged from that of intact Aa: chain (70,900)  (Band 13). The highest proportion of intact ACV chains was in I-4; the lowest was in I-9, which had virtually no detectable intact Aoc chain.
Conversely, I-9 contained the highest proportion of Aoc remnant chains; I-4, the lowest. These data suggested that cleavage at any or all of at least 10 sites along the Aa chain resulted in the formation of early catabolic intermediates.
Comparisons with highly clottable derivative fractions I-SD and I-9D produced by plasmin treatment of I-4 in vitro showed that they were quite similar to their respective plasma counterparts (I-S and I-9) in that the y and BP chains appeared to be intact, but few or no intact Aa chains were present.
These observations thus strengthen the notion that it is plasmin which catalyzes the formation of early intermediates in vivo.
Examination of Fragment 'X,' another fibrinogen derivative produced by plasmin digestion, revealed considerable depletion of intact B/3 as well as of Aa: chains, indicating that Fragment 'X' was more extensively degraded than I-8D or I-9D.
Human plasma fibrinogen, as well as that of bovine and rabbit origin, is heterogeneous from the standpoint of solubility (3-i). This heterogeneity is a reflection of the relatively lower moleculnr weight of material with higher solubility (7-9). Two of the most highly soluble plasma subfractions have been named I-8 and I-9 (3). Because Fraction I-8 was obtained in a more highly clottable form, many of the biophysical and biochemical studies (8, 10) were carried out on it, although I-9 exhibited the higher solubility.
Comparisons of human plasma Fraction I-4 with I-8 (3, 8) and a remarkably similar fraction' (I-8D) obtained after limited plasmin degradation of Fraction I-4 (II), led to the hypothesis that the higher solubility material was a catabolic 1 Sherman et al. (11) termed the plasmin-degraded material "I-8D5" to indicate that it was a derivative (D) and that there were five subfractionation (D,) steps involved in its isolation and repurification to high clottability. In the present paper, that fraction is referred to simply as I-8D; by the same token, therefore, plasmin-degraded material having solubility equivalent to that of plasma fraction I-9 is termed I-9D. 5210 itltermediate derived from higher molecular weight species. Conclusive support for the catabolite hypothesis was provided by turnover studies of radioactively labeled rabbit fibrinogen; :I classical precursor-product relationship (12) was shown between the low and higher solubility forms of clottable material (7). The fact that in both I-8 and I-8D the transformation from low to higher solubility material occurred by release of peptide material from the COOH-terminal portion of the parent species (8, 11) suggested that the formation of the in viva derivatives w-as catalyzed by plasmin.
More recent studies of the sequence of plasmin degradation of fibrinogen (13, 14) have shown that the ;\a chain2 is the first to be degraded; related studies of fibrin sltbutlits formed from human fibrinogen of high and low solubility indicated that the high solubility material lacked intact her chains (13), an observation which further supported the notion that formation of these metabolites in viva is catalyzed by plasmin.
The studies to be reported here were concerned with the detailed subunit structure of fibrinogen fractions of low (I-4) and high solubility (I-8 and I-9). These were compared to ascertain the exact features which account for previously shown differences it1 solubility and thrombin-clotting times (3). Also included ivere comparisons with certain plasmin-mediated in vitro derivatives, viz. Fractions I-8D and I-9D and Fragment 'X' (15), to determine their structural relationships to one another and to the intermediates formed in viva.

P'ibrinogen
Subfractions and Derivatives-Fractions I-4, I-8, and I-9 were prepared from outdated human ACD plasma as described by Mosesson and Sherry (3). The clottability of I-9, usually about 90yp when prepared by the original method, was increased to more than 96Cyc (four preparations) by the following modifications.
After the usual procedure of precipitating Fraction I-8 at 2.1 M glycine concentration, the residual supernatant solution was treated at room temperature with PEG3 (average molecular wieght 6000; final concentration 4%) and centrifuged at 4000 to 5000 x g. The precipitate, Fraction I-9, was dissolved in 0.27 M NaCl-0.01 M sodium phosphate buffer, pH 6.4, to a protein concentration of 1 to 2c/,; reprecipitation by adding solid @-alanine to a final concentration of 2.4 I\T (250 g of /%alanine per liter of solution) was carried out at room temperature, followed by cooling to l-2" for 30 min or more and centrifugation at 10,000 to 12,000 X g. The resulting Fraction I-9 was more than 967; coagulable with thrombin (rallge for four preparations, 96 to 990/c). Limited plasmin degradation of Fraction I-4 (original clottability >98%) was carried out as previously described (11). Clottability of the digest prior to fractionation was 887,. Fractions I-8D and I-9D isolated from this digest by ethanol and glycine subfractionation (3, 11) were processed to a more highly clottable form ( >96'$$ by reprecipitation procedures similar to those described for purification of I-9. Fragment 'X' (15) 2 The nomenclature used for the subunit chains of fibrinogen, that is, Ao, Bp, and y, conforms to the recent recommendations of the Committee on Nomenclature of the International Society on Thrombosis and Haemostasis (Oslo, Norway, 1971). Formerly, the chains had been designated or(A), p(B), and y.
S-Sulfofibrinogen was preljared according to the method of Pechere et al. (16).
NH,-terminal Analyses-NH?-terminal analysis was performed by an adaptation of the DNS method (17), previously described in detail (18). The relative amounts of NHz-terminal residues were estimated visually.
Thrombin-treated chains for NHSterminal analysis were prepared as follows.
Freeze-dried subunit chains (about I-to 1.5.mg samples) were suspended in 0.2 ml of 0.2 M N-ethylmorpholine, pH 8.4 + 0.1, and treated with two successive additions of 10 ~1 of thrombin.
The suspension was incubated overnight at room temperature, then dissolved by addition of 0.4 ml of freshly deionized 10 M urea, and carried through the DNS-labeling procedure.
Human thrombin (lot I-I-I) was provided by Dr. 1). L. Aronson, Division of Biologics Standards, National Institutes of Health; it was reconstituted with water to a concentration of approximately 100 U.S. units per ml and st,ored at -20".
Tryptic Peptide il4apping-Mapping of the S-sulfo chains was done by a modification (19) of the method of Katz et al. (20). Digestion was carried out with trypsin-L-l-tosylamido-2-phenylethyl chloromethyl ketone (Worthington Biochemicals Corp.) ; maps were stained with either a collidine-ninhydrin or a cadmium acetate-ninhpdrin stain. Electrophoretic and Related Procedures-Gel electrophoresis was carried out at 20" (57; polyacrylamide gel slabs, 3 mm thick) in Tris-EDTA-borate buffer (21) containing freshly deionized urea (8 M; buffer pH 8.6) or no urea (buffer pH 8.4). The former conditions were very similar to those described by Takagi and Iwanaga (22). A vert.ical electrophoresis assembly (E-C Apparatus Corp., Philadelphia, Pa.) and a pulsed power supply (Ortec, Oak Ridge, Tenn.) were employed.
Polyacrylamide gel electrophoresis in acetic acid-urea solutions (23) was carried out in a disc gel apparatus with a regulated power supply (Buchler Instruments) in tubes, 5 x 75 mm, at a gel concentration of lO(;/; (pH 2.7, 2 M urea). SDS-polyacrylamide gel electrophoresis was performed essentially as described by Weber and Osborn (24) in 9"/c gels. For disulfide bond reduction, when desired, DTT (14 mM final concentration) was employed. Samples for application were made by mixing with an equal volume of 500/c (v/v) glycerol-O.1 M sodium phosphate, pH 7.0.
The precision in estimating the molecular weight of unknown bands from these markers was hl.5 to 3.5%.
Isolated X-sulfofibrinogen chains were reacted with reptilase (prepared from the venom of Bothrops atros; a gift from Pentapharm Ltd. Basel, Switzerland) or thrombin in the following way. Freeze-dried samples were solubilized in deionized 10 M urea to a concentraCion of about 10 mg per ml and then dialyzed overnight against 300 to 500 volumes of 0.2 M sodium phosphate, pH 7.0. After dialysis samples contained a uniform fine granular suspension.
To 50 ~1 of each sample (estimated final protein concentration, 4.5 Ing per ml) was added two ~-PI portions of reptilase a,t an interval of 30 to 60 min, or 2.5 ~1 of thrombin. The suspension was allowed to incubate with stirring for 3 hours at room temperature and overnight at 5". The amount of enzyme added caused unmodified fibrinogen (Fraction I-4, 2 mg per ml, which had been carried through the same dialysis procedure) to clot 30 s after the first reptilase addition and in less than 10 s after treatment with thrombin. A portion of the enzyme-treated S-sulfa sample was mixed with an equal volume of 9 M urea, 2% SDS solution prior to application as described above. Standard markers in the same solvent were run in separate gels.
Gels were stained with Amido schwarz IOB or Coomassie brilliant blue. Amido schwarz-stained gels could readily be counterstained with Coomassie blue (26) which in turn could be displaced by restaining with Amido schwarz; this provided a useful system for examining the same bands stained with a more or less sensitive stain.
Densitometric scans of the gels were carried out with an Aminco filter fluorometer equipped with a horizontal thin layer scanning attachment; a Wratten 23A filter was employed. To eliminate variable light refraction encountered during scanning of cylindrical gels laid directly on glass plates, gels were submerged in water in a clear plastic trough specially constructed for that purpose.
Gel slabs could be placed directly on glass plates over the light source.
Agarose electrophoresis (proteins stained with light green) and immunoelectrophoresis (27) were performed at room temperature in 17, agarose in barbital buffer (pH 8.6) on flexible plastic film (Cronar, 70-mm movie film, Du Pont), 90 to 100 mm in length, for 1 to 2 hours at 60 to 80 volts.
Immunodiffusion was carried out at room temperature for 24 to 48 hours with rabbit anti-human fibrinogen serum which had been made specific by absorption with human serum.

Chromatographic
Procedures-DEAE-cellulose (Whatman DE 23) analytical column chromatography of unmodified fibrinogen was carried out at 2" on columns, 0.9 x 30 cm, employing a combined pH and phosphate gradient (18). Fractions corresponding to I%, of the total gradient were collected.
CM-cellulose gradient chromatography (Whatman CM 23) for the separation of S-sulfofibrinogen chains was carried out on columns, 0.9 X 30 or 1..5 X 30 cm, at room temperature in 8 M urea. A sodium acetate gradient from 0.005 to 0.15 M with respect to Na+ was formed with a nine-chamber gradient device (Buchler Instruments) containing 50 or 75 ml per chamber. The Na+ molarity in each chamber was, consecutively: 0.005, 0.02 x 4, 0.04 x 3, 0.15. Dilution of the stock buffer solution (0.2 M sodium acetate, 8 M urea, pII 5.3 & 0.1) with 8 M urea to prepare the individual buffers resulted in a fall in pH; e.g. the pH in Chamber 1, 0.005 Na+, was 4.7. Urea solutions were prepared from reagent grade material and deionized on a mixed bed resin (Biorex AC 501-X8) immediately prior to use. Fractions corresponding to lcc of the total gradient were collected. The void volume of the system, was assumed from previous estimates (18) to be 69; of the total gradient.
Columns were regenerated by flushing with stock 0.2 M sodium acetate buffer followed by starting buffer. Pooled fractions were estensively dialyzed against water and freeze-dried.

Chromatographic,
Blectrophoretic, and Immunochemical Charac- The ordinates for the theoretical phcbsph&e and pH gradients (---) are on the right.
The point at which the gradient was begun is indicated by the vertical arrow. Each chromatogram was obtained separately under identical conditions.
were carried out to extend the results of previous studies of I-4 and I-8 (8) by comparing them with I-9, which has an even higher solubility and longer thrombin-clotting time than I-8 (3). Electrophoresis of these unmodified fractions in 59;, polyacrylamide gels at pH 8.4 showed that they migrated as single bands; Fractions I-8 and I-9 were indistinguishable from one another although both had a slightly greater anodal mobility4 than I-4. Agarose electrophoretic studies indicated a marginally greater anodal mobility of I-8 and I-9 compared with I-4, but this difference could not be shown by the positions of the fibrinogen precipitin arcs in concomitant immunoelectrophoretic experiments.
Immunodiffusion experiments comparing I-4, I-8, and I-9 resulted in fibrinogen precipitin lines of complete identity. Thus, by these several criteria, I-9 could not be distinguished from I-8.
The chromatographic elution pattern of I-8 on DE.4E-cellulose was very similar to that of I-4 ( Fig. 1) except for the previously observed (8) tendency for slightly earlier elution of the first major peak of I-8. I-9 displayed a shoulder on the ascending limb of Peak 1 which was not present in either I-4 or I-8 (Fig. l), a relatively minor but consistent (three preparations examined) distinguishing feature of this material. Fragment 'X' also migrated as a single band but had a significantly greater mobility than any of these fractions. 6 In all such chromatographic experiments, the frontal fraction varied from 3 to 250/, of the total eluted protein depending upon the particular preparation and upon the amount of material loaded on the column.
The nature of the frontal fraction is not relevant to these studies but is considered in Paper III of this series (28) which relates primarily to-, chain heterogeneity.
The over-all elution profiles of I-4, I-8, and I-9 were quite similar except that whereas the Rcu peak of I-4 or I-8 had a distinct shoulder on the descending limb, the corresponding peak of I-9 was more symmetrical. Trypfic Peptide Xapping-The tryptic peptide map was characteristic for each chromatographically isolated chain. The maps of y chains from I-4, I-8, and I-9, were indistinguishable from one another.
The same was true of the respective B/3 chain maps." These findings were in sharp contrast to that for hoc chains, which differed both quantitatively and qualitatively (Fig. 3). All peptide spots in the I-8 and I-9 ACY maps were also present in the pattern of I-4. However, in I-8 and I-9 a considerable number of spots were absent (solid arrows) or clearly reduced (dashed arrows) relative to those of I-4, and the reduction appeared more severe in I-9. !\TH2-terminal Analysis-These analyses not only served for identification of the isolated chains and assessing the presence or absence of thrombin-susceptible peptide bonds, but were also useful for estimating the degree of contamination of one preparation of chains with others.
The Aoc chains (two preparations each from I-4, I-8, and I-9) had the terminal alanine and aspartic acid residues characteristic of human A peptides (29) ; these were replaced by glycine after thrombin treatment.
Prior to thrombin treatment, small amounts of NHz-terminal glycine were detected in I-4 and I-8 Aar chains but not in those from I-9. The presence of a y chain contaminant, shown by other techniques to be present in all Aa preparations (see below), was suggested in a single I-8 ha! preparation by the presence of detectable terminal tyrosine in thrombin-t,reated or untreated samples. Recause the NHz-terminal pyroglutamyl of the B@ chain is unreactive with DNS chloride, BP contamination of Aol chains could not be evaluated by this analysis, but other data have indicat,ed the consistent presence of small amounts of B/3 chains in Aa: preparations (see below). Owing to the unreactive NHB-terminal residue, B/3 chains (two preparations of each fraction) revealed only trace amounts of terminal amino acids. The most evident of these was tyrosine (all preparations), indicatin g the presence of contaminating y chains. After thrombin treatment to remove the B peptide the expected KHz-terminal glycine appeared. The y chains (two preparations of each fraction) had only t'yrosine detectable before or after thrombin treatment. Electrophoretic Analysis (Figs. 4 to 9)-Urea-polyacrylamide gel electrophoresis at pH 8.6 resulted in clear separation of I-4 S-sulfa subunit chains (Fig. 4). The relative mobilities of the chains were consistent with their elution behavior during CMcellulose chromatography.
In addition to the major band corresponding to each S-sulfo chain, small amounts of other bands were usually seen.
The y chains of I-4, I-8, and I-9 were indistinguishable from 6 The tryptic peptide maps of the Bp and y chains from I-4, I-8, and I-9 (in addition to the Aol chain shown here) are available as JBC Document Number 72M-81, in the form of 1 microfiche. one another in this analyt,ical system, as were the respective BP chains.
In contrast to that of I-4, the Aa bands of I-8 and I-9 were more diffuse, somewhat more anodal than that of I-4 and tended to overlap the B/I position.
Additionally, in one of t\\-o preparations of I-8 (Fig. 4) there was a second major band which migrated more slowly than any Aa band.7  The foregoing resulk suggested that the differences between Fractions I-4, I-8, and I-9 were related to differences in the Aa! chain population.
This notion was investigated, and ultimately substantiated, by the following series of experiments in which SDS-polyacrylamide gel electrophoresis was the principal technique.
It had already been shown (see above) that the NH2terminal portions of the Acr and B/3 chains were intact and could be cleaved by thrombin.
It was also established at the outset, through analyses of isolated Aoc, BP, and y chains, that the individual chains, regardless of their molecular size, could be differentiated by comparing their gel patterns before and after reaction with reptilase or thrombin* (Figs. 5 and 7). This was possible because the electrophoretic technique was capable of distinguishing molecular weight differences in chains resulting from the loss of an A or B peptide (molecular weight 1500 to 1600) ; actual computed reductions were between 700 and 2800, a result consistent with the precision of the method.
The B/3 chains of I-4, I-8, and I-9 moved as single bands and were indistinguishable from one another; the same was true of y chains (Fig. 5). Upon treatment with thrombin, the position of BP chains shifted toward the anode whereas that of the y chains did not.8 After reaction with reptilase, most of the BP band underwent no change in mobility; however, two additional minor bands (arrows, Fig. 5) were detected.
The band of higher molecular weight was barely appreciated as a faint band continuous with the anodal portion of the B/? band in I-4 and I-9 or as a faint but discrete band in I-8. Its appearance was con- The lower band, molecular weight 56,000, differed from that of the intact Bfi chain (60,400) by 4,400, an amount suggesting that it had been derived from a BP chain by cleavage at a site more internal than the thrombin-susceptible site at position 14 (perhaps at position 42). Thus, in spite of the fact that reptilase does not act exclusively on the Acx chain (35), the reaction was sufficiently specific to enable Acu and Pi3 chains (and by inference, y chains as well) to be distinguished from one another. ln contrast to B/? and y chains, a complex band pat,tern xvas obtained with chromat'ographically isolated ACY chains. There were 11 to 13 more-or-less distinct bands, which were numbered consecutively in order of increasing anodal mobility (Figs. 6 and 7). For clarity, bands were further classified as Group I, II, or III if their relative migration was close to that of intact   Fig. 6). I-4 I-9 Aa, BP, or y chains, respectively; Group IV bands were those which migrated considerably faster than any of the intact chains. There were usually two Group I bands (occasionally this group was resolved into a triplet) both of which were enzymically identifiable as ACY chains. Of these bands, Number 1 was assumed to correspond to intact Aar chain whereas Number 2 was regarded as an ACY remnant of slightly reduced molecular size (Table I). Of the Group II bands one proved to be a B/3 chain (Band 3) because it reacted only with thrombin, while the other (Band 4) met the enzymic criteria for an Aor chain. Band 5 (Group III), on the basis of its characteristic molecular weight and its nonreactivity with reptilase or thrombin, is most likely a y chain contaminant.
Band 6 is an AOL remnant which can also be observed in unchromatographed fibrinogen chain preparations as a discrete band or as a minor band continuous with the y chain at its anodal border (Fig. 9) ; it becomes clearly evident after treatment with reptilase or thrombin.
Of the Group IV chains, Bands 12 and 13 were the faintest which could be enzymically identified as Aa! remnants, whereas Bands 8 and 9 were t,he predominant species (e.g. Fig. 7). The relative proportion of the total Aoc chain population represented by Group IV varied directly (Table I;  as Group I, ZZ, or ZZZ if their relative migration is close to that of intact Aol, BP, or -Y chains, respectively, and as Group IV if their migration is considerably faster than that of any intact chain (see Table I). The I-9 and I-4 gels were those shown in Fig. 6; the I-8 gel scanned was obtained by running a 45-pg sample in the same experiment as the 112.pg sample of I-8 shown in Fig. 7. From the foregoing data it was evident that the differences between I-4, I-8, and I-9 were in the degree of intactness of their ACY chains. Densitometric scans9 (Fig. 8)  Furthermore, it could be estimated visually that Fraction I-9 had virtually no intact (i.e. Band 1) Aoc chains whereas about 50(yC of the Group I bands in I-8 and most of t.hoae in I-4 were of this variety.
I-nderstandably, I-9 had the highest proportion of the smaller .\a remnants.
Comparisons with MD, I-9D, and Fragment 'P-Comparisons of whole I-8 and I-9 with the plasmin-mediated Fractions I-8D and I-9D and Fragment 'X' were carried out by SDS-gel electrophoresis to assess and compare the type and degree of degradation of their component chains (Fig.  9). I-8D and I-9D were quite similar to their respective counterparts, I-8 and I-9, in that the B/3 and T chains appeared to be intact and few Group I ha chains were detectable?0 The I-81) and I-91) bands which were visible and identifiable as .\a! remnants by enzymic analysis corresponded in size to the ;\a! remnants 2, 8, 9, 11, and 12 which had been identified in whole preparations of I-8 or I-9 or their isolated chains.
In contrast to I-8 and I-9 there was little indication of the presence of Bands 6 and 10 in MD or I-9D. Furthermore, in I-81) and I-9D there were considerable quantities of the Aa remnant corresponding in size to Band 11, whose tollcentration was too low to be detected in equivalent amounts of whole I-8 and I-9 (Fig. 9) from Band 1 to Band 2 has been reported to occur during plasmin hydrolysis (13) and since the ratio of Band 2 to Band 1 is greater in I-8 than in I-4 and almost infinite in I-9 (Fig. 7) with Blomback Fractions I-l and I-3 (4) from fresh-frozen plasma, a doublet has been noted in the hoi position.
Thus, all plasma fibrinogen fractions appear to contain at least some degraded Aac chains. As a result of the present studies at least ten cleavage sites on the Aoc chain (presumably lysyl-X or arginyl-X bonds; see Fig. 10) have been recognized by the demonstration of at least ten Aa remnants of discrete molecular weight.
Although conversion to Bands 2 and 9, for example, occurs with relatively high frequency, the es:& degradation sequence (or sequences) cannot be deduced from these data.
It is therefore convenient to regard cleavage of any or all of these bands as "Step I" in the formation of the early catabolic intermediates (Fig. 10). Plasmin degradation of fibrinogen in vitro is currently viewed as occurring in three relatively distinct stages (15), in the first of which (Stage I) clottable derivatives are present; the second and third stages are characterized in part by the absence of clott,able derivatives.
Presently available information permits a more complete picture of the composition of Stage I components to be drawn and a comparison of the well characterized early clottable derivatives (11,15,43) to be made. The present data, as well as t,hose of others (13, 14) clearly indicate that it is the Acu chain which is first degraded (Step 1, Fig. 10) and that this occurs by COOH-terminal attack without loss of clottability (3, 11). The formation of the plasmin-mediated Fractions I-8D and I-9D (11) occurs in this manner.
Following this there is hydrolysis of the B/3 chain (13, 14) which is not associated (at least initially) with the loss of clottability (13, 44). Shainoff and co-workers (44, 45) have reported that prior to appreciable loss of clottability the BP chain loses a fragment containing the II peptide indicating that the early fragmentation of the BP which are recognizable in the acid gel system will prove to be the same ACZ remnants differentiat,ed in the SDS system, although not necessarily in the same relative positions. I3 Unpublished results.
chains occurs in the NH-terminal region (Step 2, Fig. 10) ; how ever, evidence relating to the presence or complete absence of the B peptide in the remaining core molecule has not as yet been reported.r4 Fletcher et al. (43) have defined "first derivative" as that product retaining clottability.
This definition therefore includes derivatives in which degradation of Ati chain or of both Aa and IQ3 chains has occurred.
The Fragment 'X' (15) preparations (two) which we studied had extensively degraded BSp and ;\a chains ( Fig. 9). Since it is not highly clottable, other degradative steps must have occurred.
This material is therefore a more advanced Stage I derivative than those characterized b? Sherman et al. (11) or defined by Fletcher et al. (43).
Recent studies (6, 9) of bovine plasma fibrinogen of high solubility have suggested that the catabolic pathway outlined above (cf. Fig. 10) may not obtain in this species. Previous investiga tions of early intermediate human fibrinogens (3,8,11) had shown that the NHn-terminal region of the molecule cont,aining the A and B peptides remained intact and that the rate of release of these peptides by thrombin was the same as for "native" Fraction I-4 (8). The lengthening of the thrombin time was therefore attributable solely to delayed aggregation as was shown by direct experimentation (8). Although as in the corresponding human fractions, the molecular weight of bovine I-9 is lower than that of I-4 (9) and intact Aoc chain is lacking in the former, the NH?terminal residues (approximately 6 moles of tyrosine per 350,000 g) are considerably different from those found for Fraction I-4 (6). More importantly, there is no change in P;iHs-terminal residues after reaction of bovine I-9 with thrombin, although the fraction itself is highly clottable.
If this rnaterial proves to be an intermediate metabo1it.e analogous to human material, a different catabolic pathway than that for human fibrinogen must exist? and site (or sites) of thrombin action different from those generally recognized must be postulated for this species of fibrinogen.
The additional possibility that bovine I-9 could be a "fetal fibrinogen" has been raised (9) and must also be considered.
Acknowledgments-We are grateful to Dr. Michael Potter, National Cancer Institute, for the use of his laboratory facilities, and to Miss B. Lynne Srmstrong for skilled technical assistance.
Addendum-Since t,his manuscript was submitted, Pizzo et al. (49) have also confirmed that degradation of the Ao( chain precedes that of the B/3 chain (13, 14) and have presented their deductions of t'he sequence of fibrinogen hydrolysis by plasmin in vitro. Since these investigators assumed that their fraction lacking intact B/3 chains was the same as fraction I-8D (see 14  Our demon-23, atration of the intactness of the B/? chains in I-SD and I-9D, as well as our subsequent studiesi in which the B/3 remnant 24. corresponding to their /3' chain has been shown to be unreactive with thrombiu, strongly indicates that initial cleavage of the 25. U/3 chain takes place in the NHz-terminal region, as originally reported by Shainoff and co-workers (44, 45).