Comparison of human plasma fibrinogen subfractions and early plasmic fibrinogen derivatives.

Abstract Human plasma fibrinogen Subfractions I-8 and I-9 have previously been shown to consist of catabolic intermediates of fibrinogen characterized by the lack of COOH-terminal portions of Aα chains. The present study was undertaken to determine the structural relationships between these catabolites and core derivatives formed in vitro by the action of plasmin. First, these fractions were compared with plasma Subfraction I-4, which is representative of "native" fibrinogen, by polyacrylamide gel electrophoresis of unreduced samples in sodium dodecyl sulfate. Fraction I-4 resolved into two major bands, designated I and II (the higher the number the greater the anodal migration and the smaller the molecular size), and a trace component (Band IIIa). Fractions I-8 and I-9 exhibited four bands (IIIa, IV, V, and VI) of which Band IIIa was the major component and VI, the least abundant. Fractions I-8 and I-9 were also compared with subfractions isolated from stage 1 plasmic digests of Fraction I-4 (termed I-9Dn, where n = clottability of the digest from which the subfraction was obtained). An early plasmic subfraction, I-9D88, contained Bands IIIa (major component), V, and VI plus a minor amount of Band VII; a subfraction from a more advanced stage 1 digest (I-9D50) contained a faster migrating Band III (IIIb) and relatively increased amounts of Bands VI and VIII. Despite the fact that Fraction I-9D88 closely resembled its plasma counterpart (I-9) in clottability (96%), electrophoretic, chromatographic, and solubility behavior, several differences were apparent. Band IV was detected only in the plasma fractions; Band VII was present only in the plasmic derivatives. Each of the 10 core Aα chain remnants (viz., Aα/2, Aα/4, Aα/6 to Aα/13) identifiable by gel electrophoresis of reduced plasma fibrinogen subfractions can apparently be generated in vitro by plasmin digestion. However, whereas Aα/8 and Aα/9 were the most abundant of the small Aα/ remnants (mol wt g40,000) in Fractions I-8 and I-9, Aα/11 predominated in I-9D88. Moreover, Bβ chain degradation was somewhat more advanced in I-9D88 than in its plasma counterpart. The failure to demonstrate complete identity between I-9 and I-9D88 underscores the present uncertainty as to whether plasmin alone catalyzes the formation of circulating catabolites in vivo. Experiments involving gel slicing, reduction of the eluted protein, and re-examination by gel electrophoresis, plus related electrophoretic procedures and chromatographic analyses, yielded the following results. Formation of species migrating within Bands II, IIIa, IV, and V could be accounted for by the loss of COOH-terminal portions of the Aα chains; species migrating in Band IIIa or more anodally exhibited no intact Aα chains. Formation of Band IIIb (from species like those in Band IIIa) occurs by the loss of the NH2-terminal region of one Bβ chain; material within Bands VI and VII is associated not only with the removal of this region from the other Bβ chain but also with internal cleavages such as those eventually leading to the separation of Fragments D and E.


SUMMARY
Human plasma fibrinogen Subfractions I-8 and I-9 have previously been shown to consist of catabolic intermediates of fibrinogen characterized by the lack of COOH-terminal portions of Aa chains.
The present study was undertaken to determine the structural relationships between these catabolites and core derivatives formed in vitro by the action of plasmin.
First, these fractions were compared with plasma Subfraction I-4, which is representative of "native" fibrinogen, by polyacrylamide gel electrophoresis of unreduced samples in sodium dodecyl sulfate.
Fraction I-4 resolved into two major bands, designated I and II (the higher the number the greater the anodal migration and the smaller the molecular size), and a trace component (Band IIIa). Fractions I-8 and I-9 exhibited four bands (IIIa, IV, V, and VI) of which Band IIIa was the major component and VI, the least abundant.
Fractions I-8 and I-9 were also compared with subfractions isolated from stage 1 plasmic digests of Fraction I-4 (termed I-9Dn, where n = clottability of the digest from which the subfraction was obtained). An early plasmic subfraction, I-9Da8, contained Bands IIIa (major component), V, and VI plus a minor amount of Band VII; a subfraction from a more advanced stage 1 digest (I-9D50) contained a faster migrating Band III (IIIb) and relatively increased amounts of Bands VI and VII.
Despite the fact that Fraction I-9D88 closely resembled its plasma counterpart (I-9) in clottability (96%), electrophoretic, chromatographic, and solubility behavior, several differences were apparent. Band IV was detected only in the plasma fractions; Band VII was present only in the plasmic derivatives.
Each of the 10 core Aa chain remnants (viz., AcY/~, Aair, ALY/~ to Aa/,%) identifiable by gel electrophoresis of reduced plasma fibrinogen subfractions can apparently be generated in vitro by plasmin digestion.
However, whereas Ao!f8 and Aalg were the most abundant of the small ACY/ remnants (mol wt <40,000) in Fractions I-8 and I-9, Acu/ll predominated in I-9Ds8. Moreover, BP chain degradation was somewhat more advanced in I-9Ds8 than in its plasma counterpart.
The failure to demonstrate complete identity between I-9 and I-9Ds8 underscores the present uncertainty as to whether plasmin alone catalyzes the formation of circulating catabolites in Go.
* This work was supported by Grant HE-11409 from the National Heart and Lung Institute.

Experiments
involving gel slicing, reduction of the eluted protein, and re-examination by gel electrophoresis, plus related electrophoretic procedures and chromatographic analyses, yielded the following results.
Formation of species migrating within Bands II, IIIa, IV, and V could be accounted for by the loss of COOH-terminal portions of the Aa chains; species migrating in Band IIIa or more anodally exhibited no intact Aa! chains.
Formation of Band IIIb (from species like those in Band IIIa) occurs by the loss of the NHz-terminal region of one BP chain; material within Bands VI and VII is associated not only with the removal of this region from the other BP chain but also with internal cleavages such as those eventually leading to the separation of Fragments D and E.

Heterogeneity
of human plasma fibrinogen manifested by differences in solubility reflects the presence of early catabolic intermediates (e.g. those occurring in Fractions I-8 and I-9) that are more soluble, have a longer thrombin clotting time, and are of lower molecular weight than the parent material (occurring in Fraction I-4) from which they are formed (l-3). This molecular weight reduction occurs by proteolytic attack on Aa chains (3, 4) resulting in release of various portions of their COOH-terminal regions (2, 3). Ten cleavage sites along the Aa chain have been identified by gel electrophoresis of fibrinogen subfractions isolated from plasma; that is, 10 more or less discrete remnants, smaller than intact ACZ chain but retaining an intact NH2 terminus, were demonstrable (3).
The mechanism of fibrinogen catabolism is not known with certainty; it is possible that a number of circulating proteolytic enzymes play a role. Nevertheless, much of the catabolic activity resembles that of plasmin.
For example, it has been shown in vitro that the initial plasmin attack on fibrinogen occurs on Aa chains (3-8) and gives rise to derivative fractions (e.g. I-9Ds8) that are similar (3,9) to the plasma fibrinogen fractions of relatively high solubility (e.g. I-9). Furthermore, increased amounts of fibrinogen catabolites are found in the plasma of patients undergoing urokinase (9) or streptokinase (10) infusions, and similar fibrinogen derivatives can be produced in vitro by the addition of urokinase (2, 11) or streptokinase (2, 10) to plasma.
It has been shown (12) that plasminogen activated in plasma can retain a portion of its enzymic activity by virtue of the formation of a complex between plasmin and cu2-macroglobulin. This finding suggests at least one possible mechanism by which plasmin or plasm&like enzymes could express their effects in blood.
In a preceding study designed to educe the essential features of the fibrinogen molecule from the structures of the plasmic degradation products (13) attention was, of necessity, directed toward derivative species formed during the later stages of digestion (stages 2 and 3 in the terminology of Marder et al.,Ref. 14). It was clear, however, that the degradation undergone by fibrinogen at these stages was far beyond that seen in even the most soluble plasma fibrinogen fractions (3,13). In the present work, therefore, emphasis was placed on the detection and analysis of plasmic derivatives arising during digestive stage 1. Moreover, to determine the structural relationships between circulating catabolic fibrinogen intermediates and core derivatives formed in vitro by the action of plasmin, a detailed comparison was made between the plasma subfractions and those isolated from stage 1 plasmic digests.

TERMINOLOGY
The terminology applied in this study is the same as that developed previously (1,3,9,13) and conforms (insofar as guidelines are available) to the tentative recommendations (August 1972) of the Subcommittee on Nomenclature of the International Society on Thrombosis and Haemostatis.
A pertinent summary follows.
The intact subunit chains of fibrinogen are designated Aa, I@, and y. Cleavage at sites other than those attacked by thrombin is indicated by a solidus on the COOH-or NH2-terminal side. Thus Acr + ACY/ + /cy; BP +B@/ + l/3, and so forth. Chain fragments are further designated with a numerical subscript (3, 13, Table I); within a given set (e.g. ALY/ core remnants), a higher subscript indicates a lower molecular weight. MATERIALS AND METHODS was carried out at 6 ma per tube for 10 hours in G.2% gels. Bands IV and V, Gels 2 and S, are difficult to appreciate as distinct bands in the photograph, although they were evident upon visual inspection.

Preparation
(3), but these manipulations yielded I-SD which was more than 95% clottable only when the clottability of the digest itself was at least SS'$$. Similar observations were reported by Sherman et al. (9).
S-sulfo derivatives of fibrinogen or of plasmic subfractions were prepared according to the method of Pech&re et al. (17).
Electrophoretic and Related Procedures-SDS-electrophoresisl was performed essentially as described by Weber and Osborn (18). This technique was also used for t,he identification of chains retaining peptide A or B by treating with reptilase or thrombin prior to electrophoresis (3, 13). Throughout the investigation samples were examined after sulfitolysis or treatment with DTT or in unmodified form, and the concentrat,ion of acrylamide in the gels was varied, according to t,he requirement,s of a given experiment. In certain studies the gels were sliced (transversely) at l&mm intervals after electrophoresis of unmodified samples. Protein was eluted from the individual slices, reduced with DTT, and re-examined by SDS-electrophoresis.
This procedure, which for brevity is denoted simply "gel slicing", has previously been described in detail (13). In the present work it permitted assessment of Aor and Aa/t chains as well as of Acx/ remnants with molecular weight <40,000, but AOI/ remnants of intermediate size were obscured by intact Bp or y chains.
Chromatographic Procedures-DEAE-cellulose (Whatman DE23) gradient elution chromatography of fibrinogen or plasmic subfractions was carried out as reported elsewhere (19). CM-cellulose gradient elution chromatography in 8 M urea (Whatman CM23) for separation of S-sulfo derivative chains was carried out as recently described (3).

Characterization of Plasma Fibrinogen Fractions by SDS-Alectrophoresis-
Solubility may be regarded as an index of the average molecular size of fibrinogen catabolites present in plasma. fibrinogen fractions.
SDS-electrophoresis of unreduced samples ( Fig. 1) provided a means for determining the distribution of molecular sizes within each fraction.
Unmodified Fract'ion I-4 exhibited two major bands (I and II); Fractions I-8 and I-9 ' The abbreviations used are: SDS-electrophoresis, polyacrylamide gel electrophoresis in sodium dodecyl sulfate; DTT, dithiothreitol. exhibited f'our other bands (IIIa to VI).
Bands I and II were unique to I-4, although there was some overlap of the anodal margin of Band 11 with material present in I-8 or I-9. Band IIIa, which was a trace component of I-4, constituted the major band of Fractions I-8 and I-9. Bands IV and V were absent from Fraction I-4 but were present as minor components of both I-8 and I-9. Band VI, containing the fastest migrating material in T-8 and I-9, appeared as a minor, diffuse zone. The relative distribution of derivative bands in I-8 and I-9 was consistent with the average molecular weights of these fractions (3): Hand IIla, the highest molecular weight component, was more abundant in I-8; Bands IV and V (lower molecular weight) were more abundant in I-9.
The degree of intactness of Aa chains has been shone (3) to account for differcnccs among the various plasma fibrinogen fractions (6. Fig. 2, Gels 1 and .Z). To correlate these findings with the size heterogeneity of core species observed within a given plasma fraction (Fig. l), preparative amounts2 of unmodified samples were subjected to SDS-clectrophoresis, followed by gel slicing as described under "JIaterials and ;\lethocls." After DTT reduction, all protein slices from gels of Fraction I-4 displayed hcu and Acx/~ chains, whereas these chains aere virtually absent from the slices of Fraction I-9 (not shown).
Thus, fibrinogen species migrating in Bands I and II are charactcrizcd by the presence of ACY and AcY/~. Successively anodal slices from I-4 exhibited a progressively decreasing ratio of Aa to Aa/*. This observation, however, does not imply that the difference between Bands I and II is due solely to the content of remnant Aoclz; additional differences are probably attributable to smaller Aoc/ remnants present in Fraction I-4 (3). In fact, no slice from I-4 contained exclusively hollz or Aa. The latter finding suggested that even species migrating within Band I had uuder-2 The load conditions required for preparative electrophoresis (13) preclude the resolution obtainable in analytical experiments. A gel slice (or group of slices) does not necessarily correspond to a single band.  Ao/rr relative to ALY/II or As/g, although the latter two were always the most abundant. Bands IIIa, IV, and V were thus SDS-electrophoresis of the "y" peak of S-sulfo I-SDS8 demoncharacterized by high levels of small Aa/ remnants, but con-strated, in addition to intact y chains, a band known (13) (Fig. 3) exhibited this shoulder as well as all of the similarities between Acr/ chains in plasmic derivatives and those major characteristics of the profile of I-9. (Consistent with in plasma subfractions were explored by comparing 1-9D% with j the fact that separation on DEAE-cellulose occurs primarily on Fractions I-4 and I-9. SDS-electrophoresis of reduced I-9D** the basis of charge rather than molecular size, SDS-electro- (Fig. 2 Furthermore, SDS-electrophoresis of pooled column fractions continuous with the anodal margin of the y chain (Fig. 2, Gel 3); showed that each chromatographic peak of S-sulfo I-9D@ con-its presence in either I-9 or I-9Ds8 could be demonstrated by tained the type of chains found in the corresponding peak of treating the sample with reptilase or thrombin (not shown). S' S-sulfo I-9. That is, the first major peak ("7") contained mostly mce ACY/~ is detectable in I-4 only as a minor component of y chains; the second ("B/3"), mostly BP chains; and so forth isolated S-sulfo ACY chain preparations (3), its identification in (Fig. 4). Material eluted in the "A& peak is considered in a a The /83 derivative was initially characterized as a y chain in later section; that eluted in the "7" and the "BP" regions is Ref. 3 because of its electrophoretic migration and its failure to described in the following paragraph. react with thrombin. Its subsequent identification as a Bg chain remnant has been developed in Ref. 13. reptilase (R) or thrombin (T) before electrophoresis.
whole samples of I-SD@ (as well as in somewhat more advanced derivative fractions) indicated that it had been formed as a result of plasmic hydrolysis.
Electrophoretic patterns of S-sulfo Aa chain preparations (Fig. 4) suggested that remnant Aa/d was somewhat more abundant in Fraction I-9D@ than in the parent material, I-4. This notion was confirmed by densitometry, which showed that Aa/r amounted to 9.4% (range for three analyses, 8.2 to 10.6%) of the I-4 Aa! chainpopulationand 13.3% (rangefor four analyses, 13.0 to 13.6%) of that of I-9D@. These values implied that plasmin had catalyzed the formation of Aa/r in vitro. Despite the fact that these Aa/ remnants can be produced by plasmic hydrolysis in vitro, early plasmic derivatives (viz. I-9DS8) differ from Fraction I-9 in the distribution of the various Aor/ remnant chains.
In I-9, the major Atu/ remnants of molecular weight <40,000 are Aa!s and AcY/~; little or no Acu/ll is evident ( Fig. 2 and Ref. 3). In sharp contrast, AcY/,~ is the most abundant Acu/ remnant in I-9Dw (Figs. 2 and 4). Its distribution in the core species comprising 1-9D= was further investigated by analysis of gel slices obtained by preparative electrophoresis of unmodified I-9Da. Although the Acu/,, remnant was present in all protein-containing slices, the [ALY/~ + Acx/~]:[ALY/~ to Acu/io] ratio, as assessed either visually or bydensitometry (Table  II), increased with the anodal position of the gel slice. No gel slice of I-9Ds8 exhibited an [Am/ii + Acu/iJ :[Aa/, to Aar/ro] ratio as low as that of Fraction I-9 (which had almost no detectable Aa/,, band under these load conditions).
Further Characterization of Stage 1 Core Derivatives by SDS-Electrophoresis-Comparison of the starting material for plasmin digestion, viz. Fraction I-4 (Fig. 1, Gel 1), with theunfractionated 88% clottable digest (Gel 4) showed the persistence of Band I in the latter.
Moreover, the digest contained increased amounts of material overlapping the positions of Bands II and III, indicating that material migrating as Bands II and III had been formed from Band I during this early digestive phase. Electrophoresis of unmodified I-9Dm (Fig. 1, Gel 5) yielded a band pattern similar in several respects to that of plasma Fractions I-8 and I-9. The main band corresponded to Band IIIa (Fig.  1, Gels 2 and S), although it did overlap somewhat the position of Band IIIb (Gels B and 7). In addition, Fraction I-9Da exhibited three faster bands, the first two of which migrated approximately in the positions of Bands V and VI.
(There was no evidence of a derivative migrating as Band IV.) The fastest Band (VII) had no identifiable counterpart in the plasma fractions and appeared in minor amounts at this phase. Densitometric ratio of material in the [Am/u + As/J region to that in the [Aa/ to Aa/l~] region.
Unmodified Fraction I-9D88 was subjected to SDS-electrophoresis on a 5y0 gel, the gel was sliced (1.5-mm slices, see Footnote 2), and the protein was eluted, reduced, and subjected to SDS-electrophoresis in 9% gels under standard conditions (see legend to Fig. 2). These gels were stained with Coomassie brilliant blue and analyzed by densitometry. Slices were numbered consecutively from the cathodal end of the gel; those containing no protein (1 to 4 and beyond 8) were omitted from the table. Since AU/~~ did not always separate from L/n, the two remnants were considered as a single region. Experiments described above showed that the feature distinguishing derivative Band IIIa of Fraction I-9 from Bands I and II of Fraction I-4 was the degree of ACX chain fragmentation. This distinction was valid for Band IIIa of Fraction I-9Ds6 as well since, like I-9, it had undergone little attack on chains other than ACX (Fig. 2, cf. Refs. 3 and 13). However, plasmic derivative fractions showing appreciable degradation of other chains required additional investigation to establish the structural features of their derivative bands.
For example, Fraction I-9D50 ( Fig. 2, Gel 4)  The //I3 remnanta (migrating in the cathodal region of the y band) was present in relatively large amounts.
An additional remnant (mol wt -9000) of uncertain origin (Fig. 2), unreactive with reptilase or thrombin, was also evident in I-9D50. In fractions like this, the composition of derivative bands separated by SDS-electrophoresis of unmodified sample ( Fig. 1) was investigated by electrophoresis of reduced protein obtained from gel slicing experiments.  Fig. 1, Gel 7). The gel slices (1.5 mm thick) from which the samples were eluted are numbered consecutively beginning at the cathodal end of the gel.
In contrast to that of I-9 or 1-9Da, most of the Band III material from I-SDS0 or I-9Dz0 was distinctly more anodal (i.e. Band IIIb, Fig. 1). Band V was no longer evident in either of the latter preparations, but there were increased amounts of Bands VI and VII.
The subunit chain structure of these derivative bands, assessed as described above, indicated that the B/3 chain was enriched in gel slices taken from the cathodal region (i.e. from Band IIIb) but was virtually absent from the anodal slices (Fig. 5). Conversely, the [/-yi, /p4, B/3/d region was relatively abundant in anodal slices and depleted in cathodal slices. The finding that cathodal samples contained the //It remnant as well as the B/3 chain suggested that core species migrating as Band IIIb, in addition to having undergone extensive degradation of both Acu chains (a feature also characteristic of Band IIIa), had released the NHz-terminal region of one B/3 chain as a BP/ remnant (cf. Fig. 6). The absence of B/3 chains in anodal gel slices permitted the additional conclusion that species migrating as Bands VI and VII had released the NH%-terminal regions of both B/3 chains. However, the migration of Bands VI and VII (Fig. 1) was so rapid relative to Band IIIb that the mere release of the NH&erminal region of the second B/l chain appeared insufficient to account for the difference.
The cathodal to anodal distribution of the [/ri, /B4, B/3/,] band suggested that internal plasmin attack related to the ultimat'e separation of Fragments D and E might account for the formation of Bands VI and VII.
This possibility is considered again below and depicted schematically in Fig. 6.

Consideration of Heterogeneity Shown in Previous Polyacrylamide Gel Electrophoretic
Studies-Heterogeneity of circulating fibrinogen as a function of size was first suggested by studies of its solubility (1) and proved by molecular weight determination on plasma Subfractions I-4 and I-8 (2). Disc electrophoresis (20) carried out at that time showed that material from I-8 migrated farther than did that from I-4 (2). Since I-4 and I-8 did not differ significantly with respect to charge, as assessed by chromatography on DEAE-cellulose (2) or by electrophoresis FIG. 6. Schematic diagram of core species migrating as Bands I to VII.
The general chain structure of core species migrating within the various bands and of chain remnants released from the core during their formation are depicted by the drawings.
The interchain disulfide bridges (-) which join the Acu ( (m), and y (0) chains are shown forming two hexagons linking all six chains in two regions (13). No attempt has been made to indicate the exact location of intrachain disulfide bridges (13) or the precise spatial orientation of the chains. Cleavage sites are on cellulose acetate strips (l), the differences in migration in polyacrylamide gels appeared to be due primarily to the molecular sieving properties of this medium. Electrophoretic heterogeneity was also found within each plasma subfraction (2) as well as within early plasmic derivatives (9), although the significance of these observations was not fully appreciated. Mills and Karpatkin (21) identified six core components by SDS-electrophoresis of solubilized clots prepared from plasma fractions differing in solubility; these components appear to correspond to Bands I to VI described above. Their results and the present work strongly imply that the electrophoretic heterogeneity observed within each subfraction in earlier studies (2, 9) can indeed be attributed to size differences.
Considerations of Core Species Identijied by SDS-Electrophoresis of Unreduced Samples (Fig.   6)-The present study has shown that plasma fibrinogen includes a number of (multichain) derivative species indentifiable by their migration rates in SDSelectrophoresis.
These derivatives, which were found in the electrophoretic bands designated II, IIIa, IV, and V, were shown (by experiments involving separation of the subunit chains) to arise via the release of certain portions of the COOH-terminal region of Acr chains. Species migrating in Bands II, IIIa, and indicated by numbers corresponding to terminology summarized in the text and in Table I. Heavy lines show conversion pathways. Thin solid lines show remnants released during these conversions. Thin broken lines show regions from which remnants originate. The release of specific internal portions of Aa, BP, or y chains as an essential feature in the formation of core species migrating within Bands VI and VII is somewhat speculat,ive (see text); this is indicated by appropriately placed queslion marks.
V are also formed during early phases of the plasmic hydrolysis of "native" fibrinogen (Fraction I-4) in vitro (Fig. 1). Since subfractions comprised of these plasmic derivatives (e.g. I-9Dss) contain almost no intact Aa! chains but manifest only minor attack on chains other than Acu, it may be concluded that they too are formed by the release of COOH-terminal peptides from both Aa! chains. Band IIIb, occurring in appreciable amounts at a somewhat later phase of plasmic degradation (i.e. I-SDS0 and I-9D2'J, Fig. l), is formed from species like those in IIIa by the release of material from the NHt-terminal region of one BP chain.
Bands VI and VII become the major core derivatives during the late phases of stage I. Subsequently, these derivatives can undergo cleavage resultin g in the separation of Fragments D and E or, prior to separation of Fragments D and E, they may release a large remnant containing the COOH-terminal region of one BP chain (i.e. /BG) to form derivatives migrating as Band VIII (13). Electrophoresis of reduced samples prepared from gel slices (Fig. 5 and Ref. 13) has indicated that virtually all available /a! and BP/ remnants are released during the formation of Bands VI and VII.
However, the electrophoretic positions of these bands relative to other core derivatives strongly suggest that the release of /a: and BP/ remnants is insufficient to account for t,he formation of species migrating as Bands VI and VII.
This apparently requires additional attack on irlternal sites, such as those eventually cleaved to separate Fragments D and E. The following considerations of this attack are based upon information obtained by carrying plasmic digest,ion beyond stage 1 (13) and on the amino acid sequence of the NHs-terminal portion of fibrinogen (22). Since remnants /&, and /pl both contain the COOII-terminal portion of the Is/3 chain (Ref. 13 and Table I) and each is produced by cleavage at a different site, cleavage at both of these sites should give rise to a derivative chain (/p/) of mom lecular weight -9,000.
However, such a remnant has not yet been identified. Remnant /o(/u (mol wt 6,700,Ref. 13), comprises part of Fragment D and appears to account for a portion of the Aa: chain between Aa/yj (Table I), which is formed during stages 1 and 2, and Aa/1r (Table I), which has been itlerltified in early forms of Fragment E (13). The fate of the r(Jmainder of this Aa chain segment is uncertain. It is not known whether such a /a/ remnant or the /p/ remnant alluded to above is released or whether it remains covalently linked to the core. The NHa-terminal sequence of t,he y chain (22) indicates that scission at appropriate combinations of plasmin cleavage sites (i.e. positions 35, 53, 58, and 62) should release part of the y chain as one or more /y/ remnants (Fig. 6).
In their studies of the plasmic digestion of fibrinogen, 1Iills and Karpatkin (5) identified two early core components by SDS-electrophoresis.
These products, which they designated "I" and "II", correspond to our plasmic derivative Hands IIIb and [VI + VII], respectively.
They attributed the formation of product "I" to the release of the COOH-terminal region from one Aa chain and that of product "II" to the release of the corresponding portion from the other. In addition, these in vestigators indicated that other degradative events, such as the loss of the NHS-terminal fragment from the l%fl chain (i.e. l%fi/), were associated with the formation of derivative species withill product "II" (5). Our present analyses strongly indicate that products "I" and "II'! could not have been formed in the manner proposed.
Neither L& chain of product "I" (i.e. derivative Band IIIb) is intact, and appreciable degradation of the l!@ chain has occurred as well. Even the release of all available /a and I!@/ remnants does not result in the formation of product "II" (i.e. Bands VI + VII). Pizzo et al. (8) have also examined unreduced plasmic digests by SDS-electrophoresis. They concluded that there are several major species of "Fragmcnt X" formed although they did not provide data bearing on the specific structural features of the various derivatives. Discrepancies between certain of their conclusions (e.g. the site of initial 13p chain attack) and ours relating to the formation of early core derivatives are attributable to several of their underlying assumptions, which are not consistent with presently available information (3, 13). Comparison of Circulating Catabolifes and Plasmic Derivatives-Previous studies (9), in which Fraction I-8 and a plasmic derivative Fraction I-8D, (obtained from a digest of 85y0 clottability) were compared, demonstrated that the latter closely resembled its plasma "solubility counterpart" in chromatographic behavior on DEAE-cellulose, electrophoretic behavior (in agarose, polyacrylamide gels, or cellulose acetate strips), NHz-terminal groups, fibrinopeptide content, molecular size, clottability, and rate of aggregation of fibrin monomers prepared from these fractions.
More recent observations (3) included a preliminary comparison of the subunit structures of I-8, I-9, I-811, and I-9D (the latter two fractions from a digest of 88% clottability) by SDS-electrophoresis; these results and our present studies permit a fairly detailed summary of the similarities and differences among these fractions.
The close similarity between I-9 and I-9D@ parallels that previously shown for Fractions I-8 and I-8D (3, 9). Fraction I-9 and its plasmic counterpart are virtually indistinguishable by chromatography on DEAE-cellulose (Fig. 3) or by chromatography of their S-sulfo derivatives on C11-cellulose (Fig. 4, present xork and Fig. 2, Ref. 3). Analyses4 of unchromatographed preparations (Fig. 2) or of material obtained from the Acr chain peak of CLI-cellulose chromatograms (Fig. 4) indicate that all .Zol/ remnant chains (i.e. Acu/z to AcY/~~) identifiable in circulating fibrinogen can also be formed in vifro by plasmin action. This conclusion differs from that implied by Mills and Karpatkin; they concluded that chains corresponding in size to Acull and AcY/~~ orcurrcd only in plasmic digests (5). ljy contrast, our results suggest that AU/ 1L and Acz/~~ arc formed in plasmic digests and in vivo, although they have a much lower frequency in the latter instance.
SDS-electrophorcsis of samples from plasmic digests showed that core derivatives corresponding in size to those found in I-4 (i.e. lland II), I-8, and I-9 (i.e. IIIa and V) were formed as the conscqucncc of Aol chain attack.
These results differ sharply and inexplicably from those of Smith and Frank (23), who reported that they were unable to produce plasmic derivatives with molecular weights similar to those of plasma catabolites.
I>iffcrences between I-9 and L91Y8 include the absence of derivative Hand IV from the latter and the paucity of remnants Acz/~~ and holllz in the former.
In addition, l{p chain degradation is somewhat more advanced ill plasmic Fraction I-9D8* than in its plasma counterpart (I-9). Remnants /& and /pa can be demonstrated in the chromatographic fractions of S-sulfo I-9T>88 (Fig. 4) as well as in gels of cross-linkctl fibrin from this preparation (24). The /pa chain is also identifiable in I-9 (3, 24), but /@z has not been detected.
A band corresponding to [/rl, /p4, I%@/5] was revealed by electrophorrsis of the "7" peak of S-sulfo I-91Y8 (Fig. 4); in retrospect, such a band (albeit very small) could also be identified in S-sulfo y chain preparat,ions of plasma fibrinogen fractions (3), although there was no indication of the presence of the Up/, remnant.
Thus, even as early a plasmic subfraction as I-9DR8 manifests somewhat more dcgradation than the clottable fibrinogen catabolites isolated from plasma.
Tnasmuch as Fragment X (14) results from considerably more advanced digestion, the normal concentration of this fragment in circulating plasma must be very low." The failure to demonstrate complete identity between I-9 and I-9Dx8 underscores the uncertainty as to whether plasmin alone catalyzes the formation of circulating catabolites in tivo. Nevertheless, available evidence indicates that plasmin or a plasmin-like enzyme(s) catalyze(s) the formation of circulating catabolites and permits the speculation that some or all of the *In agreement with the observation of Mills and Karpatkin (4) we have found an increase in the amount of ACY/Z during early phases of plasmic hydrolysis.
By using relat,ively heavy loads for SDS-electrophoresis we have demonstrated small amounts of ALY/,~ in the "ACZ" peak of S-sulfo I-9D8* (cf. Fig. 4), whereas under the same conditions this remnant was not seen in the "Aol" peak of S-sulfo I-4. 5 Band VII was previously shown to correspond to Fragment X (13). This band was not demonstrable in plasma subfractions, hence Fragment X, if present in plasma at all, must be represented by Band VI of Fraction I-9 (Fig. 1). observed features of these circulating derivatives may be related 9. to the modifying effects of the plasma environment (12).

10.
Acknowledgments-We are grateful to David Amrani and 11.
Ruth Umfleet for their assistance in carrying out these studies, 12. to Lynne Goldstein for her secretarial services, and to the Medical Illustrations/Photography section at State University of New 13. York for their cooperation and technical excellence.