Inactivation of active thrombin-activable fibrinolysis inhibitor takes place by a process that involves conformational instability rather than proteolytic cleavage.

Thrombin-activable fibrinolysis inhibitor (TAFI) is present in the circulation as an inactive zymogen. Thrombin converts TAFI to a carboxypeptidase B-like enzyme (TAFIa) by cleaving at Arg(92) in a process accelerated by the cofactor, thrombomodulin. TAFIa attenuates fibrinolysis. TAFIa can be inactivated by both proteolysis by thrombin and spontaneous temperature-dependent loss of activity. The identity of the thrombin cleavage site responsible for loss of TAFIa activity was suggested to be Arg(330), but site-directed mutagenesis of this residue did not prevent inactivation of TAFIa by thrombin. In this study we followed TAFI activation and TAFIa inactivation by thrombin/thrombomodulin in time and characterized the cleavage pattern of TAFI using matrix-assisted laser desorption ionization mass spectrometry. Mass matching of the fragments revealed that TAFIa was cleaved at Arg(302). Studies of a mutant R302Q-TAFI confirmed identification of this thrombin cleavage site and, furthermore, suggested that inactivation of TAFIa is based on its conformational instability rather than proteolytic cleavage at Arg(302).


Thrombin-activable fibrinolysis inhibitor (TAFI) is present in the circulation as an inactive zymogen.
Thrombin converts TAFI to a carboxypeptidase B-like enzyme (TAFIa) by cleaving at Arg 92 in a process accelerated by the cofactor, thrombomodulin. TAFIa attenuates fibrinolysis. TAFIa can be inactivated by both proteolysis by thrombin and spontaneous temperaturedependent loss of activity. The identity of the thrombin cleavage site responsible for loss of TAFIa activity was suggested to be Arg 330 , but site-directed mutagenesis of this residue did not prevent inactivation of TAFIa by thrombin. In this study we followed TAFI activation and TAFIa inactivation by thrombin/thrombomodulin in time and characterized the cleavage pattern of TAFI using matrix-assisted laser desorption ionization mass spectrometry. Mass matching of the fragments revealed that TAFIa was cleaved at Arg 302 . Studies of a mutant R302Q-TAFI confirmed identification of this thrombin cleavage site and, furthermore, suggested that inactivation of TAFIa is based on its conformational instability rather than proteolytic cleavage at Arg 302 .
Thrombin-activable fibrinolysis inhibitor (TAFI) 1 (1), also known as plasma procarboxypeptidase B (2) or procarboxypeptidase U (3,4), provides an important link between coagulation and fibrinolysis (5,6). TAFI is released into the circulation by the liver (2, 7) as a proenzyme of a carboxypeptidase B and can be activated (TAFIa) by thrombin, a process in which thrombomodulin acts as a cofactor (8). TAFIa down-regulates fibrinolysis presumably by removing C-terminal lysines from fibrin that is already partially degraded by plasmin. Those lysines act as ligands for the lysine-binding sites of plasminogen and tis-sue-type plasminogen activator (9). Removal of the lysines attenuates the fibrin cofactor function of tissue-type plasminogen activator mediated plasminogen activation, resulting in prevention of accelerated plasmin formation and subsequent down-regulation of fibrinolysis (10).
Although the mechanism of TAFI regulation has been studied extensively, its mechanism of inactivation is still poorly understood. TAFI shows ϳ40% sequence homology to tissue procarboxypeptidases, that are activated as a result of a singlesite cleavage by trypsin-like enzymes that release the activation peptides (11)(12)(13)(14)(15). Upon incubation of TAFI with trypsin, TAFIa activity first increased and then decreased (2), in contrast to tissue carboxypeptidases which are not inactivated by trypsin. The trypsin-cleavage sites in TAFI were identified by SDS-PAGE and sequencing (2). TAFI was not only cleaved at Arg 92 but also at Arg 330 . The latter cleavage yielded a 25-kDa inactive fragment and a C-terminal polypeptide of 71 amino acids. Similar cleavage patterns were found upon incubation of TAFI with either plasmin or thrombin (2).
When 125 I-TAFI was incubated with thrombin, the generation of a 35-kDa fragment became apparent. During the inactivation, this fragment was further proteolyzed into 25-and 14-kDa fragments (1). The N-terminal residue of the 25-kDa band was identified as Ala 93 .
A TAFI mutant was generated in which Arg 330 was changed into Gln and compared with both wild type recombinant TAFI and plasma-derived TAFI (16). Upon incubation with thrombin/thrombomodulin, the R330Q-TAFI mutant was activated and inactivated at similar rates as the two wild type forms.
Besides by proteolysis by thrombin, TAFIa is also inactivated by spontaneous, temperature-dependent inactivation (3,8,17,18). The intrinsic fluorescence of TAFIa was quenched during incubation at 37°C, suggesting that the decrease in activity is due to conformational changes of TAFIa (17).
We studied the mechanism of proteolytic activation of TAFI and inactivation of TAFIa by thrombin/thrombomodulin using HPLC in combination with matrix-assisted laser desorption ionization mass spectrometry (MALDI-MS). Furthermore, we analyzed a mutant form of TAFI in which the putative thrombin cleavage site at Arg 302 was altered.
Cloning and Expression of Recombinant TAFI (rTAFI) and R302Q-TAFI-The TAFI cDNA I.M.A.G.E. Consortium Clone ID 194171 (19) was amplified by polymerase chain reaction using the forward primer TTTCTCGAGTCTAGAGCCACCATGAAGCTTTGCAGCCTTG (Kozak sequence underlined, XhoI and XbaI sites in italic) and the reverse primer CGGGGTACCTTAAACATTCCTAATGAC (KpnI site in italic). Polymerase chain reactions (100 l) contained approximately 50 ng of each primer, 200 M dNTPs, 2 ng of template DNA, and 5 units of Pwo polymerase. The polymerase chain reaction cycling conditions were as follows: 1 min 95°C, 1 min 55°C and 2 min 72°C for 25 cycles. The polymerase chain reaction product was incubated with 2 units of Taq polymerase and 200 M dATPs at 72°C for 30 min to generate 3Ј A overhangs for cloning it into the PCR R II-TOPO vector according to the manufacturers recommendations. TAFI was excised from this vector using XhoI and KpnI and ligated into the eukaryotic expression vector pcDNA3.1 (Invitrogen, San Diego, CA). The Arg 302 to Gln mutation (R302Q-TAFI) was generated according to the QuikChange TM Sitedirected mutagenesis kit protocol (Stratagene, La Jolla, CA), using the forward primer GTGTTTCCATATTCGTATACACAAAGTAAAAGCAA-AGACC and the reverse primer GGTCTTTGCTTTTACTTTGTGTATA-CGAATATGGAAACAC (mutated base underlined). Both constructs were confirmed by sequencing, and were found to contain the Thr 147 isoform of TAFI (20). Baby hamster kidney cells were grown in Dulbecco's modified Eagle's medium/F-12 nutrients mixture supplemented with 5% heat-inactivated fetal calf serum and penicillin/streptomycin, at 37°C in 95% air, 5% CO 2 . Transfections were performed as described previously (21). One day after transfection, G418 (1 mg/ml) was added as selecting agent. Ten days later, surviving clones were picked and TAFI expression was detected by an enzyme-linked immunosorbent assay for TAFI (22). Stably expressing clones were cultured in 175-cm 2 flasks. Once cells had grown to confluency, the medium was changed to UltraCHO (30 ml/flask, BioWhittaker, MD). The medium was refreshed twice a week for 3 weeks. The media were pooled, Pefabloc SC was added to 0.2 mM and stored at Ϫ20°C until use.
Purification of TAFI-Fresh frozen citrated human plasma was obtained from the local blood bank. One unit of plasma (ϳ300 ml) was supplemented with 5 mM benzamidine-HCl, 3 mM EDTA, 0.05 mg/ml Polybrene, and 0.1 M ⑀-aminocaproic acid (⑀-ACA) and applied to a PD-10 Sephadex G-25M column linked to a CNBr-activated Sepharose column to which a monoclonal antibody directed against TAFI (MoAb Nik-9H10 (22), 1 mg/ml, 5 ml) was coupled. Both columns were equilibrated in TBS (50 mM Tris, 150 mM NaCl, pH 7.4) supplemented with the protease inhibitors mentioned above. Unbound proteins were washed away with 10 bed volumes of 0.5 M NaCl in TBS with protease inhibitors. The protease inhibitors were removed by washing the column with 50 mM Tris, pH 7.4. Bound protein was eluted with 0.1 M glycine, pH 2.7. The eluate was collected in 1/10 (v/v) 1 M Tris, pH 9.0. This procedure was repeated and elution fractions of two units of plasma with an A 280 Ͼ 0.1 were pooled, diluted 3 times in H 2 O, and applied to a protein G-Sepharose column to remove traces of IgG. The nonadherent material was applied to a Q-Sepharose column equilibrated in 50 mM Tris, pH 7.4. Unbound protein was washed away with equilibration buffer until the A 280 Ͻ 0.01. Bound protein was eluted from the Q-Sepharose column with a linear gradient (0 -250 mM NaCl in 50 mM Tris, pH 7.4). Fractions containing TAFI, as determined with a TAFI enzyme-linked immunosorbent assay (22), were pooled and stored at Ϫ70°C.
Purification of rTAFI and R302Q-TAFI-rTAFI and R302Q-TAFI were purified by applying the pooled cell culture media to a PD-10 Sephadex G-25M column linked to the MoAb Nik-9H10 Sepharose column. Both columns were equilibrated in TBS and washed with 0.5 M NaCl in TBS. Bound protein was eluted with 0.1 M glycine, pH 2.7, and collected in 1/10 (v/v) 1 M Tris, pH 9.0. TAFI concentrations were determined with a TAFI enzyme-linked immunosorbent assay (22) and with a BCA kit (Pierce) using bovine serum albumin as a standard.
Activation of TAFI-A mixture of thrombin (20 nM), thrombomodulin (5 nM), CaCl 2 (5 mM), and purified TAFI (2.8 or 5.4 M) was incubated at 37°C and aliquots were removed at various time points and added to PPACK (30 M) to inhibit thrombin activity. Experiments were performed in the presence and absence of 5 mM ⑀-ACA. To distinguish between proteolytic inactivation by thrombin and spontaneous inactivation of TAFIa, thrombin activity was inhibited after 10 min by the addition of 30 M PPACK. At several time points aliquots were taken for measurement of TAFIa activity, SDS-PAGE, or mass spectrometral analysis.
TAFIa Activity Assay-The TAFIa activity assay was performed essentially as described (22). Briefly, aliquots of the activation mixtures were diluted to a final concentration of 0.2 M TAFI (0.4 M R302Q-TAFI) in a mixture of PPACK (30 M) and hippuryl-Arg (4 mM) in TBS, 0.1% bovine serum albumin. After thorough mixing, substrate conversion was allowed for 10 min at room temperature, stopped by 1 M HCl (20 l), and an equivalent volume of 1 M NaOH was added. Then 25 l of 1 M sodium phosphate (pH 7.4) and 30 l of 6% cyanuric chloride dissolved in 1,4-dioxane were added. After the color was developed under extensive vortexing, samples were centrifuged (2 min, 14,000 rpm) to remove excess cyanuric chloride and denatured protein. Onehundred l of the supernatant was transferred to a 96-well microtiter plate. Absorbance was measured at 405 nm in a multiscan photometer (V max reader, Molecular Devices, Menlo Park, CA).
Fragment Separation by Reverse Phase HPLC-TAFI (100 l, 5.4 M) was activated for 0, 10, 30, 60, or 120 min with thrombin, thrombomodulin, and CaCl 2 as described above and applied to a C4 column (5 m, 0.46 ϫ 15 cm) connected to a Hewlett-Packard HPLC 1050 system equilibrated in H 2 O, 0.1% trifluoroacetic acid. A linear gradient from 0 to 60% acetonitrile in H 2 O, 0.1% trifluoroacetic acid (1 ml/min, 2%/min) was used to elute bound protein. Eluting peaks were collected separately and lyophilized. Lyophilized HPLC samples were dissolved in 5 l of 45% acetonitrile in H 2 O, 0.1% trifluoroacetic acid, extensively vortexed, and subjected to mass spectral analysis.
Mass Spectrometry-MALDI-MS was performed on a Dynamo DY-100 mass spectrometer (Thermo BioAnalysis) using 3,5-dimethoxy-4hydroxycinnamic acid as a matrix. Electrospray ionization mass spectrometry (ESI-MS) was performed on an API-III triple quadrupole electrospray mass spectrometer (PE-Sciex). From the experimental mass to charge (m/z) ratios from all the observed protonation states of the protein, the mass was calculated using MacSpec software (Sciex). The mass found for TAFIa (35,808 Da fragment) was used to calibrate MALDI-MS spectra. Theoretical masses of proteins and protein fragments were calculated using MacProMass software (Beckman Research Institute, Duarte, CA) and Protein Analysis Worksheet Software, PAWS 6.0b2.
SDS-PAGE and N-terminal Sequencing-Aliquots of the activation mixtures were analyzed by SDS-PAGE (15%) under reducing conditions. For N-terminal sequencing the proteins were transferred to polyvinylidene difluoride membranes and sequenced according to the Edman degradation method on a Perkin-Elmer/Applied Biosystems type 476A. Sequencing was performed by the Utrecht Sequence center.

RESULTS
Activation of TAFI and Inactivation of TAFIa-To investigate the regulation of TAFIa activity, TAFI was incubated with thrombin, thrombomodulin, and CaCl 2 at 37°C. At various time points aliquots were taken from the activation mixture, thrombin activity was inhibited by the addition of PPACK and TAFIa activity toward the substrate hippuryl-Arg was measured. Under these conditions, maximal TAFIa activity was reached within 5 min, after which the activity decreased (Fig. 1).
Analysis by HPLC and Mass Spectrometry-To identify the proteolytic fragments generated upon activation and inactivation by thrombin/thrombomodulin, we analyzed samples of the activation mixtures after 0, 10, 30, 60, and 120 min by HPLC ( Fig. 2A) and MALDI-MS (Fig. 2B).
At 0 min, one major peak on the HPLC spectrum was observed, with a molecular mass of 55.6 kDa on the MALDI spectrum. This corresponds with the molecular mass of TAFI of Ϯ60 kDa as measured on SDS-PAGE (2). The 55.6-kDa peak was broad, reflecting the heterogeneity of TAFI glycosylation.
After 10 min activation, the HPLC spectrum revealed two additional peaks (peak 2 and 3) compared with the starting material. Peak 3 contained a fragment of 19.4 kDa. This fragment, like the 55.6-kDa protein, appeared as a broad peak on the MALDI spectrum, reflecting a heterogeneous glycosylated protein. TAFI has four potential N-linked glycosylation sites located in its activation peptide, Asp 22 , Asp 51 , Asp 63 , and Asp 86 (2). Therefore, the 19.4-kDa fragment is most likely the Nterminal activation peptide (amino acid 1-92). Peak 2 yielded a mass of 35.8 kDa and a small amount of a 24.5-kDa fragment on MALDI-MS. The peaks on the MALDI spectrum were sharp, indicating that these fragments were homogenous and not glycosylated.
The mass of the fragment in HPLC peak 2 was analyzed by ESI-MS and revealed a molecular mass of 35,808 Ϯ 9 Da (Fig.  3), which corresponds to the mass of TAFIa. This mass was used to calibrate the masses found by MALDI-MS, whenever the 35.8-kDa fragment was present in the sample.
After 30 min, TAFIa activity was no longer detectable (Fig.  1). The HPLC spectrum showed the same peaks as after 10 min with one additional peak, peak 4, which appeared as a shoulder of peak 2. Peaks 2 and 4, analyzed together by MALDI-MS, contained 2 fragments of 35.8 and 24.5 kDa. The amount of the 24.5-kDa fragment had increased considerably compared with 10-min TAFI activation, whereas the 55.6-kDa TAFI peak had decreased. After 60 min peaks 2 and 4 containing the 24.5-and 35.8-kDa forms were the major peaks on HPLC and MALDI.
After a 120-min incubation, the major fragment found was the 24.5-kDa fragment whereas minor amounts of the starting material and the 35.8-kDa form were still detected. Peak 5 was also analyzed, it contained a 11.1-kDa polypeptide.
Mass Matching-To identify the proteolytic fragments within the TAFI sequence, the experimentally obtained masses were matched to theoretical masses. In addition, the N-terminal sequences of the fragments were determined. The masses of the 55.6-and 19.4-kDa proteins could not be calculated due to the presence of glycosylated residues. The observed 55.6-kDa protein corresponded to the reported mass of TAFI (2) and started with TAFI's N-terminal sequence Phe-Gln-Ser-Gly-Gln, thus confirming its identity. The 19.4-kDa fragment most likely represented the activation peptide (Phe 1 -Arg 92 ). The observed 35,808 Ϯ 9 Da fragment (ESI-MS) matched to the calculated mass of TAFIa (Ala 93 -Val 401 , 35,813 Da), starting with the sequence Ala-Ser-Ala-Ser-Tyr-Tyr-Glu. The N terminus of the 24.5-kDa fragment was identical to that of the 35.8-Da fragment and its experimental mass (average mass of 4 measurements 24,494 Da) corresponded to the calculated mass of a fragment spanning from Ala 93 to Thr 301 (24,526 Da), rather than from Ala 93 to Arg 302 (24,682 Da). The observed 11.1-kDa fragment was matched to the fragment spanning from Ser 303 to Val 401 (theoretical mass 11,149 Da) and its N-terminal sequence, Ser 303 -Lys-Ala-Lys-Asp-His, was in agreement with this. The results suggest that thrombin proteolyzed TAFIa at Arg 302 , after which Arg 302 was cleaved off by the carboxypep-  Table I.
Stability of TAFIa-Comparison of the appearance and disappearance of the activity with the generation of fragments indicated that the presence of TAFIa did not correlate very well with the presence of the 35.8-kDa form. The 35.8-kDa fragment was present for at least 60 min, whereas the activity was no longer detectable after 30 min. Analysis of the reaction mixture under reducing conditions on SDS-PAGE yielded similar molecular weight fragments as obtained under nonreducing conditions with MALDI-MS. This suggests that the 35.8-kDa form of TAFIa was inactivated spontaneously, without detectable proteolytic cleavage. To study the discrepancy between the presence of the 35.8-kDa fragment and TAFIa activity, TAFI was activated for 10 min by thrombin/thrombomodulin whereafter thrombin activity was inhibited by PPACK. In this way, proteolytic inactivation of TAFIa by thrombin was prevented, whereas spontaneous inactivation was allowed. TAFIa activity and fragment formation were analyzed in time by the activity assay (Fig. 4A) and SDS-PAGE (Fig. 4B), respectively. In the absence of PPACK, proteolysis continued whereas in the presence of PPACK, no further proteolysis was observed (Fig. 4B). Despite the absence of proteolytic fragmentation, the rate of TAFIa inactivation in the presence of PPACK was identical to the rate of TAFIa inactivation in the absence of PPACK (Fig.  4A). Addition of ⑀-ACA, which was shown before to prevent spontaneous inactivation of TAFIa (11,17,18), resulted in reduced rates of inactivation both in the presence and absence of PPACK (Fig. 4A). Presence of ⑀-ACA seemed to slow down proteolysis in the absence of PPACK (Fig. 4B). These experiments suggest, in agreement with the results obtained with the MALDI-MS, that inactivation of TAFIa takes place in a process involving spontaneous inactivation rather then proteolysis.
Analysis of R302Q-TAFI-The role of proteolytic inactivation of TAFIa by thrombin was investigated by analysis of a mutant form of TAFI in which the putative cleavage site Arg 302 was changed to Gln (R302Q-TAFI). R302Q-TAFI exhibited activity upon incubation with thrombin/thrombomodulin toward the substrate hippuryl-Arg (Fig. 5A), although its specific activity was approximately half that of either rTAFI or plasmaderived TAFI. SDS-PAGE analysis of the activation mixtures of R302Q-TAFI and rTAFI at the indicated time points (Fig. 5B), showed that the 35.8-kDa fragment of R302Q-TAFIa was not proteolyzed by thrombin, even when 10 times higher concentrations of thrombin/thrombomodulin were used. This suggested that Arg 302 is indeed the major cleavage site involved in proteolytic degradation of TAFIa by thrombin/thrombomodulin. Furthermore, this suggested that cleavage at Arg 302 is not involved in TAFIa inactivation. As was observed for plasmaderived TAFI, the presence of ⑀-ACA stabilized R302Q-TAFIa and rTAFIa and resulted in the generation of more TAFIa activity. DISCUSSION TAFI is activated by thrombin in a process that is stimulated by thrombomodulin. Inactivation of TAFIa was first reported to take place by proteolytic cleavage at Arg 330 by thrombin, plasmin, and trypsin. However, a mutant form of TAFIa in which Arg 330 was replaced by Gln was still inactivated by thrombin/ thrombomodulin, indicating that cleavage at Arg 330 was not  responsible for inactivation (16). Inactivation of TAFIa was also reported to take place in a spontaneous temperature-dependent reaction, that did not involve proteolysis (3,8,17,18).
To study the process of inactivation of TAFIa we incubated TAFI with thrombin/thrombomodulin and correlated the formation of proteolytic fragments with the activation of TAFI and inactivation of TAFIa. The activation of TAFI was shown to result from cleavage at Arg 92 , releasing a 19.4-kDa polypeptide from the 55.6-kDa proenzyme. Activation of tissue procarboxypeptidases, which show homology to TAFI, is also achieved by trypsin cleavage of an activation peptide (11)(12)(13)(14)(15). Both the 55.6-kDa TAFI and 19.4-kDa activation peptide appeared heterogeneous on MALDI-MS spectra, in contrast to the homogeneous 35,808 Da fragment that was identified as TAFIa by ESI-MS and N-terminal amino acid analysis. This is in agreement with the presence of four potential Asn-linked glycosylation sites in the activation peptide, whereas no glycosylation sites are present in the 35.8-kDa TAFIa form.
The 35.8-kDa form of TAFIa was further proteolyzed by thrombin into 24.5-kDa and 11.1-kDa polypeptides. The N terminus of the 24.5-kDa fragment was Ala 93 and the N terminus of the 11.1-kDa fragment was Ser 303 , indicating that cleavage had taken place at Arg 302 . The 24.5-kDa fragment, however, matched the theoretical mass of a fragment spanning amino acids Ala 93 to Thr 301 , rather than Ala 93 to Arg 302 . This apparent discrepancy is most likely a result of cleavage of the Cterminal Arg 302 by the carboxypeptidase activity of TAFIa itself.
Our results thus suggest that the major thrombin/thrombomodulin cleavage site in TAFIa, is Arg 302 . This was supported by a recombinant TAFI mutant in which Arg 302 was replaced by Gln. R302Q-TAFIa was not proteolyzed by thrombin/thrombomodulin into the 24.5-and 11.1-kDa fragments indicating that Arg 302 is indeed the major cleavage site of thrombin/ thrombomodulin in TAFIa. This is in contrast to previous reports that identified Arg 330 as the major cleavage site (2). However, trypsin and plasmin were used in a part of those studies and it is possible that those enzymes cleaved TAFIa at different sites then Arg 302 . In addition, small differences in molecular weight might have been undetectable on SDS-PAGE, when plasmin, trypsin, and thrombin/thrombomodulin cleavage patterns were compared. It should be noted that thrombin/thrombomodulin contains both free thrombin and thrombin-thrombomodulin complex.
There was a discrepancy between the time course for loss of TAFIa activity and the persistent presence of the 35.8-kDa polypeptide that is considered to be the activated form of TAFI. Notably, the fragment was still present at times that TAFIa activity was no longer detectable. This was previously reported (8) and explained by intrinsic instability of TAFIa, suggesting that inactivation of TAFIa can take place by a process that does not involve proteolysis. In agreement with this, we found that R302Q-TAFIa, which cannot be cleaved at Arg 302 , was still inactivated. This strongly suggests that inactivation of TAFIa does not require any proteolytic cleavage but can be entirely caused by conformational instability of TAFIa. This hypothesis was further supported by experiments in which thrombin activity was inhibited by PPACK after activation of TAFI. Inactivation of TAFIa continued at the same rate as in the absence of PPACK, despite the fact that proteolysis by thrombin was prevented, suggesting that cleavage at Arg 302 was not essential for inactivation. Furthermore, ⑀-ACA was shown to stabilize TAFIa, whereas proteolytic cleavage still continued at a slower rate.
Our results indicate that inactivation of TAFIa is the result of the conformational instability of TAFIa and not a direct result of proteolysis of TAFIa. It is possible that the conformational changes that take place as a result of the conformational instability make the inactivated 35.8-kDa form more susceptible to proteolysis. The mechanism by which ⑀-ACA stabilizes TAFIa is unknown, but is likely to involve lysine-binding sites. C-terminal lysine residues of partially degraded fibrin may have a similar stabilizing effect on TAFIa. If so, TAFIa would be most effective in situations where partially degraded fibrin is present because fibrin-bound TAFIa would presumably be resistant to spontaneous loss of TAFIa activity.