a new variant with defective polymerization ... SDS-PAGE and reverse-phase HPLC analysis of purified ... Fibrinogen purification and chain separation.
British Journal of Haematology, 1998, 101, 24–31
Electrospray ionization mass spectrometry identification of fibrinogen Banks Peninsula (g280Tyr→Cys): a new variant with defective polymerization A N D R E W P. F E L L OW E S , S TE PHE N O. B R E N NA N , H AYL E Y J. R I D G WAY, DAV I D C. H E ATON * A ND P E T E R M. G E O RG E Molecular Pathology Laboratory and *Department of Haematology, Canterbury Health Laboratories, Christchurch, New Zealand Received 17 November 1997; accepted for publication 21 January 1998
Summary. Fibrinogen Banks Peninsula was identified in the mother of a patient referred for investigation following recurrent epistaxis. Coagulation tests revealed prolonged thrombin and reptilase times and a decreased functional fibrinogen level. Thrombin-catalysed release of fibrinopeptides A and B was normal, and no abnormalities were detected by DNA sequencing of the regions encoding the thrombin cleavage sites in the Aa and Bb genes. Reducing SDS-PAGE and reverse-phase HPLC analysis of purified fibrinogen chains were normal, as was electrospray ionization mass spectrometry (ESI-MS) analysis of isolated Aa and Bb chains. However ESI-MS revealed a mass of 48 345 D for the isolated g chains, 31 D less than the measured mass of control chains (48 376 D). Since normal and abnormal g chains were not resolved, this implies a 60–62 D mass
Fibrinogen plays a pivotal role in blood coagulation, since after cleavage by thrombin it polymerizes to form the fibrin clot (Henschen & McDonagh, 1986). The circulating 340 kD fibrinogen molecule is composed of six disulphide bonded chains ([Aa Bb g]2) which form a trinodular structure consisting of two outer D domains and a central E domain (Marchant et al, 1997). Polymerization is initiated by the thrombin-catalysed cleavage of fibrinopeptides A and B from the N-termini of the Aa and Bb chains of the E domain, resulting in the formation of fibrin ([a b g]2). Fibrin molecules interact initially via the new N-terminal a chain residues (Gly-Pro-Arg) which dock with C-terminal g chain residues in the D domain of adjacent fibrin molecules (Mosesson et al, 1995b). Affinity labelling with peptide analogues has established that the a chain Gly-Pro-Arg Correspondence: Dr Andrew P. Fellowes, Molecular Pathology Laboratory, Canterbury Health Laboratories, PO Box 151, Christchurch, New Zealand.
24
decrease in 50% of the molecules. A 60 D decrease was confirmed when DNA sequencing indicated heterozygosity for a mutation of Tyr→Cys at codon 280 of the g chain gene. Fibrin monomer polymerization revealed a delayed lag phase and reduced final turbidity and although factor XIIIa crosslinking of fibrinogen was normal, it is likely that this delay is due to impaired D:D self association. Recent crystallographic studies show residues g280 and g275 make contact across the D:D interface, suggesting a similar mechanism for the polymerization defects in fibrinogens Banks Peninsula and Tokyo II (g275Arg→Cys). Keywords: dysfibrinogenaemia, gamma chain, electrospray mass spectrometry, PCR, defective polymerization.
sequence docks to a site close to Tyr363 of the g chain (Yamazumi & Doolittle, 1992). Close association of the fibrin monomers through these D:E interactions allows stabilization of the growing polymer via end to end non-covalent D:D interactions between monomeric units. These interactions result in the formation of a half-staggered double-stranded fibrin polymer known as a protofibril. Subsequent stabilization of the protofibril is mediated by factor XIIIa covalent cross-linking of adjacent D domains (Henschen & McDonagh, 1986). Genetic mutation has provided a valuable means of probing the complex series of events involved in clot formation (Bithell, 1985). Since most variants are detected by clotting time assays, they usually have defective polymerization. Most cases are due to mutations at or near the thrombin cleavage site and result in impaired fibrinopeptide release (Matsuda et al, 1990; Ridgway et al, 1997b). Others, particularly those located in the C-terminal of the g chain, perturb the D domain, and can potentially affect q 1998 Blackwell Science Ltd
ESI-MS Identification of Fibrinogen Banks Peninsula either D:E or D:D interactions (Matsuda, 1996). Further polymerization aberrations result from mutation or deletion of the mobile a C-terminal (Ridgway et al, 1997a, 1996; Maekawa et al, 1991). Mutations characterized to date have been identified by either classic protein or DNA analytical techniques. However, both matrix-assisted laser desorption (MALDI) and electrospray ionization (ESI) mass spectrometry are being increasingly applied to the analysis of high molecular weight proteins (Karas, 1996; Smith et al, 1990), and we recently described the analysis of normal fibrinogen chains by ESI-MS (Brennan, 1997). Here, using this technique, we describe the detection of a new g280Tyr→Cys mutation in the g chain. The mutation, which is associated with mild bleeding, appears to impair polymerization by purturbing D:D interactions, since recent X-ray analysis has shown that this Tyr side chain hydrogen bonds across the D-D interface to the 275Arg side chain of a neighbouring monomer. MATERIALS AND METHODS Coagulation studies. Citrate anticoagulated blood was collected and standard coagulation assays were used to measure thrombin and reptilase times. Functional fibrinogen levels were measured by the Clauss method (Clauss, 1957). Fibrinogen purification and chain separation. Fibrinogen was precipitated from 0·8 ml of plasma by the addition of 0·225 ml of saturated ammonium sulphate (Brennan et al, 1995). The pellet was washed three times with 25% saturated ammonium sulfate, redissolved in 50 mM Tris, 50 mM NaCl, pH 7·4, and quantitated spectrophotometrically using an extinction coefficient (E1% 280) of 15·1. For chain separation, precipitates were dissolved in 0·4 ml of 8 M urea, 0·1 M Tris, 20 mM dithiothreitol, pH 8·0, and reduced for 4 h at 378C under N2. Up to 120 ml of the reduced protein was injected onto a Phenomenex (25 × 0·45 cm) C4 column. Chromatography was performed with the following solvent system; A, 0·05% trifluoracetic acid (TFA); B, 0·05% TFA in 60% acetonitrile. The column was equilibrated in 64% B and individual fibrinogen chains were eluted with a linear gradient from 64% to 85% B over 20 min at a flow rate of 1 ml/min. The effluent was monitored at 215 nm and peak crests spanning a volume of 200 ml were collected. Mass spectrometry. Peak crests were analysed directly by ESI-MS on a VG Platform quadrupole analyser operating in positive ion mode (Brennan, 1997). 10–30 ml of each peak was injected into the source at a flow rate of 10 ml/min. The probe was charged at þ3500 V and the source maintained at 608C. The mass range 700–1400 m/z was scanned every 2 s and a cone voltage ramp of 30–60 V was applied over this range. Up to 100 scans were averaged to form the raw data. Calibration, data processing and transformation were as described earlier (Brennan, 1997). For each chain analysed, an isolated control chain was analysed under identical conditions to give a figure for the normal mass. Fibrinopeptide release. The stoichiometry of fibrinopeptide release was determined using reverse-phase HPLC (Brennan et al, 1995). 50 ml of fresh plasma containing 2 ml of 2 mM q 1998 Blackwell Science Ltd, British Journal of Haematology 101: 24–31
25
phenanthroline was diluted in an equal volume of 50 mM Tris, 50 mM NaCl, pH 7·4, and induced to clot by the addition of 1 ml (1 U) of bovine thrombin (Parke-Davis) and 1 ml of 250 mM CaCl2. Reactions were terminated after 30 min by the addition of an equal volume of 49 mM phosphate buffer, pH 2·9. After boiling for 10 min the reaction mixture was microfuged and 50 ml of the supernatant was injected onto a 5 m Waters NovaPak (10 × 0·45 cm) C18 column. Chromatography was performed at a flow rate of 1 ml/min with a solvent system consisting of A, 49 mM phosphate buffer, pH 2·9; and B, a 1:1 mixture of A and acetonitrile. The gradient was from 25% to 45% B over 14 min and the effluent was monitored at 215 nm. DNA amplification and sequence analysis. All primer names refer to the position of the 50 base with respect to the gene sequence of Chung et al (1990). Genomic DNA was isolated from buffy coat by a standard procedure (Ciulla et al, 1988). A 611 bp fragment spanning exon 8 of the fibrinogen g chain was amplified by the polymerase chain reaction (Saiki et al, 1988) using primers Fn5515g (TATCTATTGCCTCCTGCCAG) and Fn6126g (ACTTGGTATTATCCACTTCC). Amplification was performed on a TC-1 thermal cycler (Perkin Elmer) for 30 cycles with denaturation at 948C for 30 s, annealing for 30 s at 608C, extension at 728C for 1 min, and a final extension at 728C for 7 min. Each 100 ml PCR contained 1 mmol/l of each primer, 200 mmol/l deoxynucleotide triphosphates, 50 mmol/l KCl, 10 mmol/l Tris, pH 8·3, 1·5 mmol/l MgCl2, 0·1% (w/v) gelatin, and 2 U of Taq DNA polymerase (Boehringer Mannheim). PCR products were purified using HiPure PCR purification cartridges (Boehringer Mannheim) and the DNA concentration estimated by comparison to HaeIII digested fX174 DNA on a 1% agarose gel (Gibco BRL) containing 0·5 mg/ml ethidium bromide. Cycle sequencing was performed with the internal primer Fn5607g (CCTACGAAAGAGGGAACTTC) using 33P radiolabelled terminators (Amersham) and ThermosequenaseTM (Amersham) according to the manufacturer’s instructions. Fibrin monomer repolymerization. Fibrin monomers were prepared by the incubation of a minimum of 0·1 mg of fibrinogen (1 mg/ml) in polymerization buffer (20 mM HEPES, 150 mM NaCl, pH 7·4) with 0·1 NIH units/ml of bovine thrombin for 3 h at 378C (Gorkun et al, 1994). The clot was wound onto a glass rod, washed five times in 150 mM NaCl and redissolved to a concentration of 3 mg/ml in 0·125% acetic acid at 48C. Solubilized monomer preparations were then repolymerized by diluting 1:10 in polymerization buffer. The polymers were again redissolved in 0·125% acetic acid and fibrin precipitation was repeated a further two times. 50 ml of fibrin monomer solution at a concentration of 1 mg/ml was then added to 450 ml of polymerization buffer and monomer aggregation recorded at 350 nm over a period of 30 min. Fibrinogen crosslinking. Purified human factor XIII (kindly provided by Dr K. Siebenlist) was activated to factor XIIIa as follows. 50 U/ml factor XIII was incubated with 1 U/ml of bovine thrombin (Sigma Biochemicals) at 378C for 1 h in a reaction containing 10 mM HEPES, pH 7·0, 50 mM NaCl and 50 mM DTT. Thrombin was inhibited by the addition of PPACK (Calbiochem) to 50 mM, followed by a further incubation at 378C for 10 min. Fibrinogen crosslinking was
26
Andrew P. Fellowes et al
performed by incubating fibrinogen (3 mg/ml), 10 mM HEPES, pH 7·0, 50 mM NaCl, 10 mM CaCl2, and 10 U/ml factor XIIIa at 378C. Reactions were terminated at various times by the addition of an equal volume of SDS sample buffer containing 10% b-mercaptoethanol. Crosslinking products were separated on a 7·5% SDS-PAGE gel and electroblotted to nitrocellulose membrane at 100 V for 1 h. Membranes were blocked with 5% casein, 25 mM Tris, pH 7·9, 75 mM NaCl, 0·5% Tween 20, and probed with the gchain specific monoclonal antibody, F3P (GTI) at a 1:500 dilution. After washing in 25 mM Tris, pH 7·9, 75 mM NaCl, 0·5% Tween 20, the secondary antibody (horse radish peroxidase conjugated anti-mouse IgG) was applied to the membranes for 1 h. Membranes were again extensively washed before the addition of ECL reagent (Amersham) and detection by exposure to Xomat AR film (Kodak). RESULTS Case report The patient studied in this report is the mother of the propositus originally mistakenly described as being heterozygous for the Bb 14 Arg→Cys mutation (Kaudewitz et al, 1986). Subsequent investigation revealed that the case histories of two Christchurch families with prolonged thrombin times had been inadvertently transposed and that the individual with the Bb 14 Arg→Cys mutation actually has a family history of thrombosis. The family described by Kaudewitz et al (1986) has now been clearly demonstrated to be homozygous for the normal Arg at position 14 of the Bb chain, and has a history of mild and variable bleeding. Coagulation studies The patient had extended thrombin and reptilase times of 46 and 34 s respectively (normal range 18–22). Similar results were obtained for her daughter (Kaudewitz et al, 1986). At 1·1 mg/ml, functional fibrinogen levels as determined by the Clauss method were exactly half that measured spectrophotometrically after quantitative purification. Fibrinopeptide release assays (not shown) were normal, with equimolar release of peptides A and B, indicating normal cleavage of the Aa and Bb chains by thrombin. PCR amplification and direct sequence analysis of exon 2 of the Aa and Bb genes confirmed a normal sequence at and surrounding the thrombin cleavage sites. Protein analysis Analysis of purified fibrinogen on 7·5% reducing SDS PAGE showed the expected pattern of bands corresponding to normal Aa, Bb and g chains (not shown). A normal pattern of peaks was also observed on reverse-phase HPLC separation (Fig 1). When HPLC peak crests were analysed by ESI-MS, the mass of the Aa chains from the patient (66 212 D) and the control (66 200 D) were in excellent agreement, and, as expected, higher than values predicted from the amino acid sequence alone (66 132 D) (not shown). This discrepancy results mainly from non-stoichiometric phosphorylation of the serine residues at positions 3 and 345 (Henschen &
Fig 1. Reverse-phase HPLC separation of reduced fibrinogen chains: (a) normal fibrinogen; (b) fibrinogen Banks Peninsula. Order of elution, Aa, Bb and g chain.
McDonagh, 1986). The mass of the dominant isoform of the Bb chain was 54 200 D. This was again in close agreement with a control value of 54 197 D and a predicted value of 54 213 D for a chain with a single sialic acid residue on its biantennary oligosaccharide side-chain. Examination of g chains, however, showed that the patient had a major isoform with a mass of 48 345 D. This was 31 D less than the control value of 48 376 D (Fig 2), the latter being in excellent agreement with the predicted value of 48 368 D for the mono-sialylated form of the g chain (Fig 2). DNA analysis We pursued the probability of a g chain mutation by DNA sequencing. Photoaffinity labelling of isolated D domains indicated that the Gly-Pro-Arg peptide interacts with residues in a 43 amino acid cyanogen bromide fragment from g337 to g379 (Shimizu et al, 1992). In addition, most previously described mutations in the g chain have been localized to the region spanning residues 268–375 (Mimuro et al, 1992). Since these residues are located in a region of the g chain encoded by exons 8 and 9 of the gene, PCR was used to examine these exons for mutations. Cycle sequencing of the exon 8 PCR product clearly showed a single base substitution at position 5744 where both G and A were present in the patient (Fig 3). This confirmed heterozygosity for a mutation in the g chain at position 280 where the normal tyrosine is replaced by a cysteine. q 1998 Blackwell Science Ltd, British Journal of Haematology 101: 24–31
ESI-MS Identification of Fibrinogen Banks Peninsula
27
Fig 2. Transformed electrospray ionization mass spectra of fibrinogen g chains: (a) normal; (b) g chains from the patient. The major signal corresponds to that of the monosialylated form of the chain and the higher mass signal to the fully sialylated form. Scans were made over a range 700–1400 m/z and transformed using MaxEnt software.
Fig 3. Cycle sequencing of the forward strand of exon 8 of the g chain gene. Left panel: normal DNA; right panel: DNA from the patient. Bands in both the G and A lanes can be seen in the patient’s sequence at position 5744 (arrowed) indicating heterozygosity for the mutation Tyr to Cys at position 280.
Fig 4. Fibrin repolymerization curves for the patient (O) and two normal control samples (S, A). The delayed lag phase seen in Banks Peninsula signifies a delay in the formation of the protofibril and the lower overall turbidity value indicates a thinning of fibres. q 1998 Blackwell Science Ltd, British Journal of Haematology 101: 24–31
Functional analysis Fibrin repolymerization studies (Fig 4) showed that the polymerization lag phase for Banks Peninsula fibrin was approximately twice that of the controls, reflecting a delay in the formation of the two-stranded protofibrils (Gorkun et al, 1994). However, the rate of polymerization following the lag phase was essentially normal, suggesting normal aggregation of protofibrils into fibres. Although protein concentrations in all experiments were identical, the final turbidity value for Banks Peninsula fibrin was only 80% of that for normal fibrin over the time course of the experiment. This reflects a reduced fibre diameter for the variant and suggests that in Banks Peninsula fibrin, thick fibre formation is somehow perturbed. Aggregation of monomeric fibrin units involves both D:E and D:D interactions. To determine which, if either, of these was affected by the substitution, factor XIIIa crosslinking studies were performed on the purified fibrinogen, and the rate of crosslinking compared to normal human fibrinogen.
28
Andrew P. Fellowes et al
Fig 5. Factor XIIIa catalysed crosslinking of fibrinogen g chains analysed by SDS-PAGE and immunoblotting: (a) fibrinogen Banks Peninsula; (b) normal fibrinogen. In both cases the ,95 kD g-dimer band increases in intensity in a time-dependent manner as assessed visually. The minor ,100 kD band present in panel a probably represents Aa·g heterdimers. Lane 1: 0 min; lane 2: 1 min; lane 3: 2 min; lane 4: 10 min; lane 5: 1 h; lane 6: 6 h; lane 7: 12 h; lane 8: 24 h.
When crosslinking profiles were probed with a g chain specific monoclonal antibody, two main bands were observed (Fig 5). The band at approximately 45 kD corresponds to monomeric g chain whereas the band at approximately 95 kD corresponds to covalently joined g dimers. This band showed a time-dependent increase in concentration, and the rate of dimer formation appeared identical in both Banks Peninsula and normal fibrinogens. This suggests that neither D:D self-association, nor g–g crosslinking is affected by the mutation. In addition to the expected bands, another minor , 100 kD band was sometimes observed. This was not related to factor XIIIa crosslinking since it was present before the addition of enzyme, and probably represents Aa·g heterodimers (Siebenlist & Mosesson, 1996). DISCUSSION Fibrinogen Banks Peninsula was initially detected through routine coagulation studies. An inherited defect was suggested since both the propositus and her mother exhibited decreased functional fibrinogen levels together with prolonged thrombin and reptilase times. The normal fibrinopeptide release and normal sequence flanking the thrombin cleavage sites suggested that the genetic lesion
might involve the functionally active regions located in the C-terminal of the g chain. This suspicion was strengthened by ESI-MS analysis which indicated a molecular mass of 48 345 D for the dominant mono-sialylated isoform of the g chain. This was 31 D less than the mass of the control (48 376 D). Since the normal and abnormal components are not resolved, what is measured is an average mass. This implies that in a 50:50 heterozygous mixture, half of the molecules are some 60–62 D lighter. This was confirmed by DNA sequence analysis which showed heterozygosity for a Tyr to Cys mutation at position 280 of the g chain. The mass decrease associated with this mutation (60 D) was in excellent agreement with the measured decrease. The presence of the new cysteine creates the possibility that it might form a disulphide bridge with a suitable free thiol such as the one at position 36 of serum albumin. This has been observed in fibrinogens IJmuiden (Bb14 Arg→Cys) and Nijmegan (Bb44 Arg→Cys) (Koopman et al, 1992). However, when Western blots of fibrinogen from the patient and a carrier of the Bb14 Arg→Cys mutation were probed with antisera to human albumin, only the Bb variant showed a cross-reactive band (not shown). Even at sensitivities capable of detecting 10¹4 mg, no albumin was observed binding to fibrinogen Banks Peninsula. This implies that the new cysteine is not available for reaction with other large macromolecules, a finding shared with fibrinogen Milano V (g275 Arg→Cys) (Steinmann et al, 1994). Whether it is adducted to other small sulphydral compounds, as occurs in fibrinogen Osaka II (g275 Arg→Cys) (Terukina et al, 1988), is not known since the chains were reduced in urea before mass measurement was performed. The prolonged thrombin and reptilase times associated with this substitution could be due to perturbation of any one the distinct functional sites in the g chain. These include sites for calcium binding, D:E interaction, and D:D interaction. Although calcium binds with high affinity to residues in the g C-terminal, it is now clear from calcium binding studies on several g chain variants (Furlan et al, 1996; Yoshida et al, 1992), together with recent structural studies (Yee et al, 1997) that residues around g280 are not involved in this process. Recent interest has centred on the role of g chain Cterminal residues in promoting D:D interactions. Based on other recent studies (Mosesson et al, 1995a; Niwa et al, 1996), it seems likely that the mutation at g280 perturbs the end to end association (D:D interaction) of adjacent fibrin monomers prior to factor XIIIa crosslinking, rather than having an effect on the initial polymerization event (D:E interaction). Evidence supporting this conclusion comes from recently published crystallographic data (Yee et al, 1997; Pratt et al, 1997; Spraggon et al, 1997). In these structures the polymerization pocket lies in a C-terminal subdomain of less ordered secondary structure extending from gGly287–Met379, whereas Tyr280 lies within the wellordered central domain which contains a five-stranded b sheet. Tyr280 is located in the last of the b sheet strands, and it is clear from the refined three-dimensional structure (Pratt et al, 1997) that it is well separated from the Gly-Pro-Arg binding pocket. This, together with the structural rigidity of q 1998 Blackwell Science Ltd, British Journal of Haematology 101: 24–31
ESI-MS Identification of Fibrinogen Banks Peninsula
29
Fig 6. A view across the D–D interface showing key residues involved in the normal dimerization process (black ball and stick). The contact made between Tyr 280 of molecule A and Arg 275 of molecule B can be clearly seen, as well as that between Arg 275 of molecule A and Ser 300 of molecule B (dotted lines). It can also be seen that Tyr 280 of molecule B does not make any inter-domain contacts. The position of Gly 268, which is substituted for Glu in fibrinogen Kurashiki I, is indicated in grey. GPRP peptides are shown as grey sticks. Hypothetical positions for residues 397–406 of molecule A (for which co-ordinates are not available) are shown as a dashed line, and the e-amino g-glutamyl crosslink is depicted as a solid line. The figure was prepared using RasWin from data deposited in the Brookhaven Protein Data Bank, accession no. 1FZB (Spraggon et al, 1997).
the central domain, makes it unlikely that substitutions affecting Tyr280 could affect polymerization by perturbing D:E interactions (see also Fig 6). It has been suggested that the ability of D domains to self associate can be measured in isolation from D:E interactions by examining the rate of factor XIIIa catalysed fibrinogen (as opposed to fibrin) crosslinking (Niwa et al, 1996). When factor XIIIa crosslinking was performed on fibrinogen Banks Peninsula, the rate of crosslink formation was indistinguishable from normal. Similar findings have been reported for the heterozygous g275 Arg→Cys variant, Tokyo II (Mosesson et al, 1995a). Here the authors concluded that both fibrin D:E interaction and factor XIIIa fibrinogen crosslinking occur normally. Factor XIIIa crosslinking experiments on the recently described homozygous variant fibrinogen Kurashiki I (g268 Gly→Glu) (Niwa et al, 1996) contrast with Tokyo II and Banks Peninsula. In Kurashiki I, factor XIIIa crosslinking of fibrinogen g chains was delayed. This is presumably because the mutation causes the substitution of a neutral glycine for an acidic glutamic acid, generating repulsive forces between D domains. q 1998 Blackwell Science Ltd, British Journal of Haematology 101: 24–31
When factor XIIIa crosslinked Tokyo II fibrin or fibrinogen was examined by electron microscopy, the polymers were not as well ordered as with normal fibrinogen, exhibiting numerous tapered terminating fibres. Ultrastructure studies show that Banks Peninsula clots also possess an abnormal fibre network reminiscent of Tokyo II (not shown). In addition, fibrin monomer repolymerization studies show a decreased ability to form fibres of normal thickness, and a delay in the formation of protofibrils. Since D:E interactions are unlikely to be affected by a substitution at g280, this most probably results from defective D:D interactions. A similar conclusion has been reached in the case of Tokyo II (g275Arg→Cys). Further supporting evidence for the role of these residues in promoting D:D interactions has recently been presented by Dempfle et al (1996) who showed that synthetic peptides from this region were able to impair fibrin polymerization. The recently solved crystal structure of the covalently joined D-dimer (Spraggon et al, 1997) indicates a relationship between fibrinogens Tokyo II and Banks Peninsula (Fig 6). On dimerization, the side-chain of g275Arg of molecule B
30
Andrew P. Fellowes et al
comes into close contact with the g280Tyr of molecule A, and a semi-reciprocal contact is formed between g300Ser of molecule B and g275Arg of molecule A. Thus in the normal dimer only one gTyr280 residue participates in an inter-domain contact, whereas both g275Arg residues are involved. It therefore seems likely that D:D association in fibrinogen Banks Peninsula would be less perturbed than in fibrinogen Tokyo II. Although we are unable to compare these variants directly, it is clear that both substitutions have a profound effect on fibrin polymerization, despite having normal factor XIIIa crosslinking. This suggests that the rate of factor XIIIa mediated fibrinogen crosslinking is largely independent of D:D self association, and may in fact be a poor measure of D domain interaction except in exceptional cases such as the homozygous, neutral to acidic substitution in fibrinogen Kurashiki I. This notion is supported by the peripheral location of the crosslinked segments (Fig 6) which would allow them great freedom of movement – indeed, neither of the crystal structures have precisely localized these residues, suggesting that they can adopt a number of different conformations. In summary, electrospray mass spectrometry has proved to be a powerful tool in the characterization of this new fibrinogen. The ability of ESI-MS to measure mass changes to within a few Daltons will enable us to detect mutations and modifications responsible for dysfibrinogenaemias, and this will lead to a better understanding of the functioning of this protein. ACKNOWLEDGMENT This investigation was supported by the Health Research Council of New Zealand, the Canterbury Medical Research Foundation and Lottery Health. REFERENCES Bithell, T.C. (1985) Hereditary dysfibrinogenemia. Clinical Chemistry, 31, 509–516. Brennan, S.O. (1997) Electrospray ionisation analysis of human fibrinogen. Thrombosis and Haemostasis, 78, 1055–1058. Brennan, S.O., Hammonds, B. & George, P.M. (1995) Aberrant hepatic processing causes removal of activation peptide and primary polymerisation site from fibrinogen Canterbury (Aa20 Val→Asp). Journal of Clinical Investigation, 96, 2854–2858. Chung, D.W., Harris, J.E. & Davie, E.W. (1990) Nucleotide sequences of the three genes coding for human fibrinogen. Fibrinogen, Thrombosis, Coagulation, and Fibrinolysis (ed. by C. Y. Liu and S. Chien), p. 39. Plenum Press, New York. Ciulla, T.A., Sklar, R.M. & Hauser, S.L. (1988) A simple method for DNA purification from peripheral blood. Analytical Biochemistry, 174, 485–488. Clauss, A. (1957) Gerinnungsphysiologische Schnellmethode zur Bestimmung des Fibrinogens. Acta Haematologica, 17, 237–246. Dempfle, C.E., Pfitzner, S.A., Lill, H. & Heene, D.L. (1996) Structural studies on fibrin polymerization sites using synthetic peptide analogues: characterisation of polymerization sites ‘A’ and ‘DDlong’. (Abstract). Fibrinolysis, 10, 3–4. Furlan, M., Stucki, B., Steinmann, C., Jungo, M. & Lammle, B. (1996) Normal binding of calcium to five fibrinogen variants with mutations in the carboxy terminal part of the g-chain. Thrombosis and Haemostasis, 76, 377–383.
Gorkun, O.V., Veklich, Y.I., Medved, L.V., Henschen, A.H. & Weisel, J.W. (1994) Role of the aC domains of fibrin in clot formation. Biochemistry, 33, 6986–6997. Henschen, A.H. & McDonagh, J. (1986) Fibrinogen, fibrin and factor XIII. Blood Coagulation (ed. by R. F. A. Zwaal and H. C. Hemker), p. 171. Elsevier Science Publishers B.V. (Biomedical Division), Amsterdam. Karas, M. (1996) Matrix-assisted laser desorption ionization MS: a progress report. Biochemical Society Transactions, 24, 897–900. Kaudewitz, H., Henschen, A.H., Soria, C., Soria, J., Bertrand, O. & Heaton, D. (1986) The molecular defect of the genetically abnormal fibrinogen Christchurch II. Fibrinogen and its Derivatives (ed. by G. Muller-Berghaus, V. Scheefers-Borchel, E. Sellmayr and A. Henschen), p. 31. Elsevier Science Publishers B.V. (Biomedical Division), Amsterdam. Koopman, J., Haverkate, F., Grimbergen, J., Engesser, L., Novakova, I., Kerst, A.F. & Lord, S.T. (1992) Abnormal fibrinogens IJmuiden (Bb Arg14→Cys) and Nijmegen (Bb Arg44→Cys) form disulfidelinked fibrinogen–albumin complexes. Proceedings of the National Academy of Sciences of the United States of America, 89, 3478– 3482. Maekawa, H., Yamazumi, K., Muramatsu, S., Kaneko, M., Hirata, H., Takahashi, N., de Bosch, N.B., Carvajal, Z., Ojeda, A. & ArochaPinango, C.L. (1991) An Aa Ser-434 to N-glycosylated Asn substitution in a dysfibrinogen, fibrinogen Caracas II, characterized by impaired fibrin gel formation. Journal of Biological Chemistry, 266, 11575–11581. Marchant, R.E., Barb, M.D., Shainoff, J.R., Eppell, S.J., Wilson, D.L. & Siedlecki, C.A. (1997) Three dimensional structure of human fibrinogen under aqueous conditions visualised by atomic force microscopy. Thrombosis and Haemostasis, 77, 1048–1051. Matsuda, M. (1996) The structure function relationship of hereditary dysfibrinogens. International Journal of Hematology, 64, 167–179. Matsuda, M., Yoshida, N., Terukina, S., Kensuke, Y. & Maekawa, H. (1990) Molecular abnormalities of fibrinogen: the present status of structure elucidation. Fibrinogen 4: Current Basic and Clinical Aspects (ed. by M. Matsuda, S. Iwanaga, A. Takada and A. Henschen), p. 139. Elsevier Science Publishers B.V. (Biomedical Division), Amsterdam. Mimuro, J., Muramatsu, S., Maekawa, H., Sakata, Y., Kaneko, M., Yoshitake, S., Okuma, M., Ito, Y., Takeda, Y. & Matsuda, M. (1992) Gene analyses of abnormal fibrinogens with a mutation in the g chain. International Journal of Haematology, 56, 129–134. Mosesson, M.W., Siebenlist, K.R., DiOrio, J.P., Matsuda, M., Hainfeld, J.F. & Wall, J.S. (1995a) The role of fibrinogen D domain intermolecular association sites in the polymerisation of fibrin and fibrinogen Tokyo II (g275 Arg to Cys). Journal of Clinical Investigation, 96, 1053–1058. Mosesson, M.W., Siebenlist, K.R., Hainfeld, J.F. & Wall, J.S. (1995b) The covalent structure of factor XIIIa crosslinked fibrinogen fibrils. Journal of Structural Biology, 115, 88–101. Niwa, K., Takebe, M., Sugo, T., Kawata, Y., Mimuro, J., Asakura, S., Sakata, Y., Mizushima, J., Maeda, A., Endo, H. & Matsuda, M. (1996) A gGly-268 to Glu substitution is responsible for impaired fibrin assembly in a homozygous dysfibrinogen Kurashiki I. Blood, 87, 4686–4694. Pratt, K.P., Cote, H.C.F., Chung, D.W., Stenkamp, R.E. & Davies, E.W. (1997) The primary fibrin polymerization pocket: three-dimensional structure of a 30-kDa C-terminal g chain fragment complexed with the peptide Gly-Pro-Arg-Pro. Proceedings of the National Academy of Sciences of the United States of America, 94, 7176–7181. Ridgway, H.J., Brennan, S.O., Faed, J.M. & George, P.M. (1997a) q 1998 Blackwell Science Ltd, British Journal of Haematology 101: 24–31
ESI-MS Identification of Fibrinogen Banks Peninsula Fibrinogen Otago: a major a chain truncation associated with severe hypofibrinogenaemia and recurrent miscarriage. British Journal of Haematology, 98, 632–639. Ridgway, H.J., Brennan, S.O., Fellowes, A.P. & George, P.M. (1997b) Rapid detection of the fibrinogen Aa16Arg→His mutation. Clinical Chemistry, 43, 2184–2186. Ridgway, H.J., Brennan, S.O., Gibbons, S. & George, P.M. (1996) Fibrinogen Lincoln: a new truncated a chain variant with delayed clotting. British Journal of Haematology, 93, 177–184. Saiki, R.K., Gelfand, D.H., Stoffel, S., Scharf, S.J., Higuchi, R., Horn, G.T., Mullis, K.B. & Erlich, H.A. (1988) Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science, 239, 487–491. Shimizu, A., Nagel, G.M. & Doolittle, R.F. (1992) Photoaffinity labeling of the primary fibrin polymerization site: isolation and characterization of a labeled cyanogen bromide fragment corresponding to g-chain residues 337–379. Proceedings of the National Academy of Sciences of the United States of America, 89, 2888– 2892. Siebenlist, K.R. & Mosesson, M.W. (1996) Evidence for intramolecular cross-linked Aa·g chain heterodimers in plasma fibrinogen. Biochemistry, 35, 5817–5821. Smith, R.D., Loo, J.A., Edmonds, C.G., Barinaga, C.J. & Udseth, H.R. (1990) New developments in biochemical mass spectrometry: electrospray ionization. Analytical Chemistry, 62, 882–899. Spraggon, G., Everse, S.J. & Doolittle, R.F. (1997) Crystal structures
q 1998 Blackwell Science Ltd, British Journal of Haematology 101: 24–31
31
of fragment D from human fibrinogen and its crosslinked counterpart from fibrin. Nature, 389, 455–462. Steinmann, C., Bo¨bgli, C., Jungo, M., Lammle, B., Heinemann, G., Wermuth, B., Redaelli, R., Baudo, F. & Furlan, M. (1994) Fibrinogen Milano V: a congenital dysfibrinogenaemia with a 275 Arg→Cys substitution. Blood Coagulation and Fibrinolysis, 5, 463–471. Terukina, S., Matsuda, M., Hirata, H., Takeda, Y., Miyata, T., Takao, T. & Shimonishi, Y. (1988) Substitution of gArg-275 by Cys in an abnormal fibrinogen, ‘fibrinogen Osaka II’: evidence for a unique solitary cystine structure at the mutation site. Journal of Biological Chemistry, 263, 13579–13587. Yamazumi, K. & Doolittle, R.F. (1992) Photoaffinity labeling of the primary fibrin polymerization site: localization of the label to gchain Tyr-363. Proceedings of the National Academy of Sciences of the United States of America, 89, 2893–2896. Yee, V.C., Pratt, K.P., Cote, H.C.F., Letrong, I., Chung, D.W., Davie, E.W., Stenkamp, R.E. & Teller, D.C. (1997) Crystal structure of a 30 kDa C-terminal fragment from the g chain of human fibrinogen. Structure, 5, 125–138. Yoshida, N., Hirata, H., Morigami, Y., Imaoka, S., Matsuda, M., Yamazumi, K. & Asakura, S. (1992) Characterization of an abnormal fibrinogen Osaka V with the replacement of g-arginine 375 by glycine: the lack of high affinity calcium binding to Ddomains and the lack of protective effect of calcium on fibrinolysis. Journal of Biological Chemistry, 267, 2753–2759.
