Insight into Residues Critical for Antithrombin Function from Analysis of an Expanded Database of Sequences That Includes Frog, Turtle, and Ostrich Antithrombins Marija Backovic and Peter G. W. Gettins* Department of Biochemistry and Molecular Biology, College of Medicine, University of Illinois at Chicago, Chicago, Illinois 60612-4316 Received March 4, 2002
Abstract: Complete sequences were determined for frog, turtle, and ostrich antithrombins. Protein sequence comparisons with the other 10 known antithrombin sequences and with sequences of other serpins have provided striking evidence for the conservation of the heparin activation mechanism and new insight into those residues important for heparin binding, for heparin activation, and for reactive center loop function, as well as an indication of which glycosylation sites might be needed for function. Importantly, an understanding of, as yet, poorly understood antithrombin-protein interactions will be greatly aided by this expanded database and comparative analysis. Keywords: antithrombin • human • ostrich • frog • turtle • protein sequence • heparin activation mechanism • sequence comparison • factor Xa inhibition
Introduction Antithrombin is the principal inhibitor of blood coagulation proteinases, requires heparin for activation, has potent antiangiogenic activity in certain conformations, and is a member of the serpin superfamily of proteins.2 Characterization of spontaneous human antithrombin mutations that are correlated with thrombosis has been a useful method for probing the role of specific residues in various antithrombin functions. One way to obtain such information that does not depend on identifying natural mutations correlated with a specific binding defect is through extensive primary structure comparisons for antithrombins from different species. However, only 10 antithrombin sequences have been available, from a limited range of species. To make sequence comparisons more useful for identification of critical residues or regions, we have determined three new antithrombin sequences, from frog, turtle, and ostrich, and carried out an analysis of all 13 antithombin sequences in terms of the roles of particular residues and regions.
Results and Discussion Sequences of Frog, Turtle, and Ostrich Antithrombins. Three new antithrombin sequences, from ostrich, turtle, and * To whom correspondence should be addressed. Phone: +1 312 996 5534. Fax: +1 312 413 0364. E-mail:
[email protected]. 10.1021/pr025515z CCC: $22.00
2002 American Chemical Society
frog, were determined, as described in the Experimental Procedures. These are shown in Figure 1, aligned with the 10 available antithrombin sequences. Whereas mature human antithrombin contains 432 residues,3 mature ostrich and turtle antithrombins each contain 436 residues and frog antithrombin contains 439 residues. Most of the differences result from a variable insertion between residues 36 and 37. The molecular weights of the polypeptide portion of ostrich, turtle, and frog antithrombins are 49.5, 49.6, and 49.9 kDa, respectively, compared to 49.0 kDa for human antithrombin. The amino acid compositions are similar for both charged and uncharged residues (Table 1) and give calculated PI values of 6.26, 7.15, and 6.65, respectively, not taking into account any contribution from carbohydrate. All four glycosylation sites present in human antithrombin are present in frog antithrombin, while turtle and ostrich are each missing the site at asparagine 96. Overall, human antithrombin shares 72% and 71% identity, respectively, with ostrich and turtle antithrombin, corresponding to about 120 differences between human and ostrich or turtle antithrombin. Frog antithrombin bears even less identity to human antithrombin, with around 140 amino acids in frog that are substituted by different amino acids in human antithrombin. Addition of three novel antithrombin sequences represents a doubling of the number of sequences that are less than 80% identical to human antithrombin (Table 2) and decreases the average identity score from 44% to 39%. This corresponds to 170 amino acids that are intolerant to any changes and share 100% identity among all the species, 123 that are more than 70% identical, and 140 that are less than 70% identical. A phylogenetic tree (Figure 2) shows that the largest differences from human antithrombin is for the avian, reptile, amphibian, and fish antithrombins. Comparison of turtle and frog antithrombins with those of the most closely related available species, ostrich and salmon, respectively, show only 80% and 61% identity. Comparison of Conserved Residues in Antithrombin and Other Serpins. A recent analysis of 433 serpin sequences identified 51 residues that are g70% conserved.4 Forty-two of these are buried and are likely to be important for the formation and stability of the metastable serpin fold and perhaps also for the ability to undergo facile conformational change during proteinase inhibition, which involves large-scale movement of strands of β-sheet A over the underlying residues present in several of the helices. These 51 residues are highly conserved in all 13 antithrombins. Only seven residues are different, but Journal of Proteome Research 2002, 1, 367-373
367
Published on Web 06/14/2002
Antithrombin Sequence Comparisons
technical notes
Figure 1. Aligned sequences of 13 antithrombins. Mature human antithrombin numbering is used. The reported sequence of chicken antithrombin probably lacks the first 12 or so residues of the mature protein. Secondary structural elements are based on the structure of native antithrombin and are shown above the sequences using solid arrows to indicate β-sheet and coils to represent R-helices. Orange triangles indicate basic residues known to be involved in binding heparin. Blue stars indicate glycosylation sites in human antithrombin. The three cysteine pairs are identified by green numbers (1, 2, and 3). Positions showing 100% conservation are dark blue, those with conservation between 70% and 100% are light blue, and those with less than 70% conservation are uncolored. These percentages are based on strict conservation. Accessibility, indicated by acc below the sequences, is shown as white for buried residues, cyan for partially buried residues, and blue for exposed residues on the basis of the structure of native antithrombin. 368
Journal of Proteome Research • Vol. 1, No. 4, 2002
technical notes
Backovic and Gettins
Table 1. Amino Acid Composition of Human, Ostrich, Turtle, and Frog Antithrombins amino acid
human
ostrich
turtle
frog
arginine lysine histidine aspartate glutamate serine threonine asparagine glutamine cysteine glycine alanine valine leucine isoleucine methionine tyrosine phenylalanine tryptophan proline
22 35 5 23 38 32 23 23 13 6 18 31 28 40 22 12 10 26 4 21
25 31 4 21 37 30 23 23 12 6 23 28 31 39 27 9 13 27 5 22
21 34 6 20 35 37 24 27 14 6 20 24 23 37 30 14 13 23 5 23
21 34 5 21 35 29 27 28 17 6 16 30 29 44 24 9 12 26 4 22
Table 2. Pairwise Percentage Sequence Relationships between 13 Mature Antithrombins species
1 2 3 4 5 6 7 8 9 10 11 12 13
1
2
3
4
5
6
7
8
9
10
11
12
13
human - 88 89 90 90 86 83 67 62 59 72 71 67 bovine - 96 91 86 83 81 66 62 60 72 70 65 sheep - 91 87 85 82 67 62 61 73 71 67 pig - 87 85 82 68 63 60 72 72 67 mouse - 83 82 66 63 61 70 70 66 rabbit - 79 66 62 60 70 70 67 guinea pig - 64 60 60 68 67 64 chicken - 61 57 85 72 63 salmon - 71 64 62 61 fugu fish - 59 59 59 ostrich - 80 67 turtle - 68 frog -
these are substituted by amino acids with similar physicochemical properties. There are, however, almost 250 additional amino acids that are more than 70% identical among antithrombins from all 13 species. Conservation of amino acids that are apparently not essential for the serpin fold and the common mechanism of inhibition strongly suggests that some may be important specifically for unique aspects of antithrombin structure and function. Comparison of the conservation patterns of secondary structure elements common to all the serpins with the equivalent regions of antithrombin shows that, with the exception of strands s6B, s4C, and helix I, all these secondary structure elements have a higher degree of conservation within antithrombins from various species (Table 3). Even more striking, all of the residues from helices B, P, D, and I1 and β-strands s3A, s5A, s2B, s3B, and s5B are g70% identical among antithrombins from all the species (Table 2 and Figure 3). These regions are therefore good candidates for binding sites or for regions that may be important for the mechanism of heparin activation and proteinase inhibition. Unlike these highly conserved regions, other secondary elements show much higher variability. A general trend is that residues that are more buried are also more conserved (Figure 3), consistent with the importance of protein core packing for the stability and function of the whole protein. Finally, all six cysteines, and by implication all three disulfides, are conserved among all 13 antithrombins. These are known to be required in human
Figure 2. Phylogenetic tree of antithrombins from the 13 species derived from protein sequence analysis. Shown is the single most parsimonious tree found, depicted as an unrooted tree, and obtained by analysis with the PHYLIP package.35 Table 3. Conservation of Sequence within Secondary Structural and Other Elements for Antithrombins and Serpins Generally element
no. of residues
no. >70% identical in all ATs
no. >70% identical in all serpins
hAa s6B hB hC hP hD s2A hE s1A hF hF-to-s3A s3A hF1 s4C s3C s1B s2B s3B hG hH s2C s6A hI hI1 hI1-to-s5A s5A RCL s1C s4B s5B
23 3 12 10 5 17 11 10 3 19 19 10 3 7 14 5 6 7 7 9 10 8 7 3 21 10 23 3 7 8
17 2 12 7 5 16 9 8 2 10 8 10 1 1 9 2 6 7 3 5 4 7 3 3 13 10 15 1 5 8
1 2 4 2 0 0 0 1 0 6 2 6 0 3 3 0 1 2 0 0 0 2 3 1 2 2 2 0 2 3
a Abbreviations of secondary structure elements correspond to those used in Figure 1, with h used to indicate helix and s to indicate strand of β sheet. Two loops linking secondary structure elements and located at the bottom of β-sheet A are also included.
antithrombin for retention of the ability to bind heparin with high affinity and to inhibit proteinases.5,6 Journal of Proteome Research • Vol. 1, No. 4, 2002 369
Antithrombin Sequence Comparisons
technical notes
Figure 3. Stereoview of native heparin-activated antithrombin, showing the pattern of conservation of all residues, using the same color code as in Figure 1. The figure was drawn using MOLMOL,36 using the coordinate file 1AZX.
Identification of Residues Critical for Heparin Binding. Antithrombin is unusual among serpins in having a relatively low rate of reaction with target proteinases in the absence of the activating glycosaminoglycan heparin.2 Even prior to an X-ray crystal structure of antithrombin bound to the highaffinity pentasaccharide, there had been proposals for which residues were critical for high-affinity heparin binding as well as for transmission of the conformational change to the interaction site with factor Xa, based on biochemical studies and on natural antithrombin variants with altered inhibitory or cofactor properties.7,8 However, the X-ray structure of the heparin-human antithrombin complex provided the best insight into what the contact residues are and what the nature of the conformational changes are.9 This structure showed changes in the heparin-binding region involving extension of helix D at its C terminus and formation of a new one-turn helix P, oriented orthogonally to helix D, at its N terminus. Also, as proposed from solution studies, these local conformational changes result in the expulsion of the partially constrained RCL from β-sheet A. Although the crystal structure of the pentasaccharide-antithrombin complex revealed contact residues for this short heparin (Figure 4), it neither proved the relative importance of individual contact residues nor established what additional residues are involved in binding more physiologically relevant longer chain heparins. We now consider in parallel the findings (i) from biochemical studies of antithrombin variants, (ii) from an analysis of the conservation of these residues among antithrombins from different species, and (iii) from the interactions observed for these residues in the crystal structure of the complex. Residues K11 and R13 are part of the flexible N-terminal loop, which becomes ordered upon pentasaccharide binding. In the X-ray structure of the antithrombin-pentasaccharide complex,9 both K11 and R13 interact with the pentasaccharide. Whereas the side chain of K11 makes a salt bridge with the carboxylate group on sugar ring E, R13 forms a main-chain hydrogen bond with 370
Journal of Proteome Research • Vol. 1, No. 4, 2002
a sulfate group on sugar ring F. That the basic lysine side chain at position 11 is required is supported by the absolute conservation of this residue among all 13 antithrombins. Although only the backbone of residue 13 is involved in heparin binding in the crystal structure, the residue is arginine in 11 species and lysine in the other two, suggesting that there may also be an important function for the basic side chain of R13. The X-ray structure shows that helix A residues R46 and R47 form a network of contacts with pentasaccharide units.9 However, heparin affinity and kinetic studies on variants at these positions have shown that, while the R47H variant binds pentasaccharide and full length heparin 20-30-fold weaker than wild-type antithrombin, an R46A variant shows only a 3-4-fold reduction in affinity.10 Moreover, mutations at position 47, but not at position 46, have been associated with thrombosis.11 Our sequence data are consistent with a minor role for R46 and a major role for R47 in heparin binding; thus, R46 is poorly conserved, being proline in six out of 13 species, whereas R47 is strictly conserved in all species, with the exception of chicken, where it is also proline. It has been suggested12 that K45 in chicken antithrombin, which is asparagine in all other species, might have taken over the role of R47, making the number of positive charges in helix A the same in chicken as in human antithrombin. The paramount importance of helix D in heparin binding is clearly reflected in its very high conservation, with all of its residues being g70% identical (Figure 1), whereas none of them is conserved among the other serpins (Table 3). Both K125 and R129 are involved in salt bridge interactions with either heparin carboxylate (K125) or sulfate (R129) moieties.9 An earlier chemical modification study had implicated K125 in heparin binding,13 while later binding studies on a K125M variant showed 17-27% reduction in heparin affinity.14 R129, however, is the most important helix D residue, since its substitution by histidine leads to a 28-35% loss of heparin binding energy.15 Strict conservation of these two residues is consistent with the
technical notes
Backovic and Gettins
Figure 4. Stereo representation of the heparin binding site of antithrombin, showing the high-affinity heparin pentasaccharide and contact residues from antithrombin on the basis of the crystal structure of the heparin pentasaccharide-antithrombin complex.9 The designation of heparin sugar rings (D, E, F, G, and H) is as used in the text and other publications. The figure was prepared using Molscript37 and the coordinate file 1AZX.
mutagenesis studies. Moreover, residues that surround K125 and R129 in both unactivated and activated antithrombins are also well conserved (I7 forms hydrogen bond with K125, whereas N45 and E414 form bonds with R129 in unactivated and activated states, respectively). There is also a striking conservation of most other residues, many of which are hydrophobic, which may be required for correct packing of helix D against the body of the protein. Indeed, a H120C variant of human antithrombin was unstable and had lowered heparin affinity, even though the imidazole side chain does not directly interact with heparin in the wild-type protein.16 Whereas R132, K133, K136, and K139 have all been proposed to form an extended binding site for longer chain heparin,17 the crystal structure shows that only R132 and K133 become part of the helix D extension upon binding of heparin pentasaccharide, with K136 as the helix-capping residue.9 Consistent with the importance of R132 and K133, as well as of K136, for heparin binding, all three residues are well conserved, with the only allowed changes being arginine to lysine or vice versa. This is borne out experimentally by a number of studies.17-19 Position 139 shows high variability, with a basic residue, lysine, being present only in human and rabbit antithrombins, and hydrophobic or even negatively charged residues being found in the other 11 species. This suggests no involvement of residue 139 in either heparin binding or activation, which is in agreement with a recent mutagenesis study19 but differs from an earlier study claiming major involvement of K139.17 Residue K114 becomes part of the new helix P (residues 113118) upon heparin binding. This residue was proposed to be critically involved in heparin binding through multiple salt bridge interactions with charged groups on heparin, including the critical 3-O-sulfate of sugar residue F9. This residue contributes more than any other single residue to heparin binding in large part through stabilization of the conformationally altered state of antithrombin.20 Its critical role is consistent with the high sequence conservation, in which the
positive charge is completely conserved and only a single change to arginine is found. The rest of helix P is also well conserved, with all of its residues being g70% identical. Heparin Activation of Antithrombin. It has been suggested, based on a comparison between the crystal structures of the pentasaccharide-antithrombin complex, of a P14C-fluorescein antithrombin and of native antithrombin, that the RCL residue E381 (P13) is important in transmission of the heparin-induced conformational change from the heparin binding site to the RCL.21 P13 forms a network of hydrogen bonding and salt bridge interactions in the heparin complex, and the “activated” P14C-fluorescein species, that differs from the network in native antithrombin, suggesting a linkage between heparin binding, the switch in salt-bridge interactions and activation of antithrombin. Strict conservation of residue P13 in all species suports the importance of this residue and contrasts with significant variability at P13 among other serpins.4,22 Of the seven residues implicated in the salt-bridge and hydrogenbonding network, five are completely conserved (E195, Y220, K222, R324, and E374) and may be important for antithrombin activation, whereas the other two (K139 and R197 in human antithrombin) show sufficient variation to be unlikely candidates for involvement in the mechanism. The hinge region residues P14 and P15 are also very important for antithrombin, since they are partially inserted in β-sheet A in the native state and their expulsion is either essential for, or coincident with, heparin-induced activation. P15 is glycine in >70% of serpins, suggesting that it is important for general serpin function. P14, however, is threonine in 60% of serpins, and serine in relatively few. In antithrombin, however, serine is strictly conserved. P14 is surrounded by residues 223 and 225 from s3A and 375 from s5A. Whereas these three residues from s3A and s5A are g50% conserved among serpins, they are 100% conserved in the 13 antithrombins. Thus, serine at P14, and appropriate contact residues in the flanking Journal of Proteome Research • Vol. 1, No. 4, 2002 371
technical notes
Antithrombin Sequence Comparisons
strands of β-sheet A, may be important for maintenance of the low activity loop-inserted state of native antithrombin. Conservation of Other RCL Residues. The remainder of the RCL1 can be divided into two parts; residues P2, P1, and P1′, which are the primary determinants of antithrombin specificity, and residues P12-P3. The former, with the sequence GRS, are absolutely conserved in all species. Whereas the conservation of the P1 arginine is understandable in terms of the proteinase inhibitory specificity of antithrombin, the conservation of a P2 glycine is initially surprising, given the preference of thrombin for proline at the P2 position of peptide substrates. However, an attempt to enhance the rate of reaction of antithrombin with thrombin by changing the P2 residue to proline gave only a 2-fold rate increase in the presence of heparin, but at the expense of a 10-fold reduction in the efficiency of complex formation,23 suggesting that P2 glycine is an optimum residue that represents a compromise between rate of interaction with thrombin and ability to efficiently form inhibited complex. The role of P1′ was established from a naturally occurring P1′ S f L variant that reacted more slowly than wild-type, but with similar efficiency of inhibition. The defect was identified as destabilization of the transition state leading from the initial noncovalent encounter complex to the acyl enzyme intermediate.24 Within the region P12-P3, only P3-P7 and P9 show less than 70% identity among the species, although all are small hydrophobic amino acids. In contrast, P8 threonine and P10 alanine, with their side chains pointing into solution in the native state and buried in cleaved or covalently complexed states with proteinase, are strictly conserved. This high conservation is likely to result from the constraints imposed during burial of the side chain when antithrombin forms covalent complexes with target proteinases. Finally, a residue that, although not within the RCL, has been suggested to play an important role in stabilizing the low activity state of antithrombin25 is E255, which in the crystal structure of native antithrombin forms a salt bridge with the P1 arginine. Recent analysis of P1W26 and fluorescently labeled P1C27 antithrombin variants, however, suggests that this interaction is of negligible energy and consequently not mechanistically important. Given this, it is noteworthy that E255 is completely conserved, suggesting that it may play a role in antithrombin function, albeit not the one originally suggested. Here, perhaps, its proximity to the P1 residue indicates a role in exosite interaction with factor Xa. Conservation of Glycosylation Sites. Human antithrombin has four glycosylation sites, at positions 96, 135, 155, and 192. It circulates as a mixture of two glycoforms: the dominant (∼90%) fully glycosylated R isoform and the minor β isoform that lacks a carbohydrate at position 135.28 The β isoform has higher heparin affinity29 and may be physiologically the more important form.30 Decreased heparin affinity also resulted, in recombinant BHK-derived antithrombin, from a core fucose residue in the carbohydrate moiety attached to Asn155.31 Both findings suggest that glycosylation affects antithrombin activity, which may be especially important for regulation of antithrombin activity in vivo. (1) Abbreviations used: R1PI, R1-proteinase inhibitor; P1, P2, etc., designation of residues in the reactive center loop, using the nomenclature of Schechter and Berger1 in which the scissile bond is between residues P1 and P1′; residues N-terminal to this are designated P2, P3, etc. and those C-terminal P2′, P3′, etc.; RCL, reactive center loop.
372
Journal of Proteome Research • Vol. 1, No. 4, 2002
Despite the effects of carbohydrate at positions 135 and 155 in human antithrombin, N192 is the only glycosylation site that is conserved. The N96 glycosylation site is present in nine species, the N135 site in 11 species, and the N155 in 10 species. Fugu fish has only two potential glycosylation sites, at N96 and N192, while other species have either three or four potential glycosylation sites. Concerning the potential physiological importance of R- and β-glycoforms of antithrombin in humans, it may be significant that the 11 species that retain the N135 glycosylation site all have the recognition sequence NKS rather than NKT, whereas almost all of the other glycosylation sites in these antithombins have threonine in the third position. Since serine rather than threonine lowers the efficiency of protein glycosylation, this sequence may be a requirement to permit generation of both R- and β-type isoforms in human antithrombin.32,33 If this is so, it is possible that retention of the NKS sequence in other antithrombins is to ensure production of both isoforms, with the implication that this is biologically important. The absence of this glycosylation site from salmon and fugu fish might then reflect differences in physiology relative to other species. Conservation of Heparin Activation Mechanism and Other Properties of Antithrombin. While antithrombin from closely related species might be expected to have similar properties and serve similar roles, as has been shown from comparison of human and bovine antithrombins, it is a more open question whether all aspects of human antithrombin function are present in widely divergent species. Our database provides insight into this. Thus, the almost complete conservation of helix D residues and of residues that form helix P, despite poor conservation of this region in other serpins, argues strongly both for high affinity heparin binding in all antithrombins, as well as for an equivalent mechanism of activation against target proteinases. Similarly, within the RCL, the retention of the P14 and P13 residues, as well as flanking contact residues found in the loop-inserted native state of human antithombin, argue for a similar loop-expulsion during heparin-induced activation, while the conserved pattern of surface charged residues also argues for common involvement of a factor Xa exosite in the activation mechanism. The absolute conservation of GRS at P2-P1′ also argues for identical targets and the same considerations of specificity and efficiency of proteinase inhibition as for human antithrombin. In the area of anti-angiogenic activity of antithrombin and the clearance of antithrombinproteinase complexes by LRP, whether these mechanisms are conserved across species must await better definition of what the epitopes are in human antithrombin.
Experimental Procedures Cloning of Antithrombin cDNAs. Total RNA was isolated from frozen livers according to the RNeasy Mini Kit (Quiagen) protocol. The initial PCR was performed on AP-primed cDNAs, which were obtained using 3′ RACE Kit (Gibco BRL) according to the manufacturer’s instructions. Pairs of primers used for the initial PCR were degenerate and designed based on the available antithrombin sequences and were: 7fd and 4rd, 3fd and 4rd, and 3fd and 8rd for ostrich, 3fd and 4rd, and 7fd and 4rd for turtle, and 7fd and 4rd for frog antithrombin (Table 4). The PCR was run on a PerkinElmer GeneAmp 2400 cycler using the following conditions: initial denaturation temperature 94 °C with 40 cycles of amplification (30 s at 94 °C, 30 s at 55 °C, and 60 s at 72 °C). Amplified fragments were cloned into pCR2.1-TOPO using the TopoTA cloning Kit (Invitrogen) ac-
technical notes
Backovic and Gettins
Table 4. DNA Oligonucleotide Primers Used for PCR Amplification of Antithrombin Sequences primer name
primer sequence (5′-3′)a
3fd 4rd 7fd 8rd 1fo 2ro 4ro 6ro 2rf 4rf 5ff 7ff 9ff 2rt 4rt
GAY CAR RTS CAY TTY TTC TTY GC CRC TGC CYT CYT CRT TYA CCT C GGT BCT IGT YAA CAC MAT YTA YTT YAA GGG GYA RCT YSA KSA CCA TGG TGA TGT CIW CHC C CGG CTT TAC AAG AAA GCC AAC AAA TCC TCG G CCG AGG ATT TGT TGG CTT TCT TGT AAA GCC G GCT CCA TAA ACT ATT TCG CTA ATG TTC TGG CCT TCT GGG ATC ACT TCT GTA ATG CG CCT GCT CCA CTG TTG TTA GCG GAG TCT CTT GCG CCT TAA CAC TGA AGG AAT CTT CCA CCC GGA ATC G GCA GGA GGA AGG ACA GAC TTG TAT GTG TCC G CTT TGC AAA GCC CAA CTG CCG GGT ATC CAA TAA AAC AGA GAA GCG CCA TGG TGA TGT CAT CCC CTT TGT ATG GAA GC GCA GCT ATA CCT GGC AGC TTG GCA CTT TCA GGG
a The IUB codes for mixed bases are as follows: R ) G/A; Y ) C/T; M ) A/C; K ) G/T; S ) G/C; W ) A/T; H ) A/C/T; B ) G/T/C; V ) G/C/A; D ) G/A/T; N ) A/G/C/T.
cording to the manufacturer’s protocol. Multiple clones were tested for the presence of an insert of the correct size by PCR, using M13 forward and reverse primers provided in the kit. The inserts were sequenced in both directions, and gene-specific primers were designed based on the sequences obtained. These primers were used for reading 3′ and 5′ ends of antithrombin cDNA using 3′RACE and 5′RACE kits (Gibco BRL). 3′RACE was carried out according to the manufacturer’s protocol: cDNAs were obtained using AP primer, and the gene specific primers used for the PCR were: 1fo for ostrich, 7fd for turtle and 5ff, 7ff or 9ff for frog antithrombin. The reverse primer used for all PCR reactions was the AUAP primer provided by the manufacturer. 5′RACE was done on cDNAs primed with the 6ro primer for ostrich, 4rt for turtle and 4rf for frog antithrombin. Subsequent PCR was done using the AAA primer supplied with the kit and primers 2ro or 4ro for ostrich, 2rt for turtle, and 2rf for frog. All amplification products of correct size were cloned into pCR2.1-TOPO, and DNA from colonies containing inserts of correct size were sequenced in both directions. Multiple independent clones were sequenced for each species antithrombin. Sequence Alignment and Data Analysis. Complete sequences were assembled using Sequencher 2.0.6 software, and the alignment was carried out using ClustalW.34 Identity scores were obtained using JellyFish (www.biowire.com). Antithrombin protein sequences were assembled into a maximum parsimony tree using PROTPARS from the PHYLIP analysis package.35 Sequences have been deposited with GenBank with accession nos. of AF411691 for ostrich, AF411692 for turtle, and AF411693 for frog antithrombins.
Acknowledgment. We thank Sergey Popov, University of Illinois at Chicago, Brad Andersen, Fish R Us Inc., Savanna, IL, and Freedom Sausage Inc., Earlville, IL, for providing us with frog, turtle, and ostrich liver, respectively. We thank Francis Peterson for technical assistance and Steven Olson for helpful comments on the manuscript. This work was supported by Grant No. R37 HL49234 from the National Institutes of Health.
Supporting Information Available: Nucleotide sequences. This material is available free of charge via the Internet at http://pubs.acs.org. References (1) Schechter, I.; Berger, A. Biochem. Biophys. Res. Commun. 1967, 27, 157-162. (2) Gettins, P. G. W.; Patston, P. A.; Olson, S. T. Serpins: Structure, function and biology; R. G. Landes Co.: Austin, 1996. (3) Chandra, T.; Stackhouse, R.; Kidd, V. J.; Woo, S. L. C. Proc. Natl. Acad. Sci. U.S.A. 1983, 80, 1845-1848. (4) Irving, J. A.; Pike, R. N.; Lesk, A. M.; Whisstock, J. C. Genome Res. 2000, 10, 1845-1864. (5) Longas, M. O.; Ferguson, W. S.; Finlay, T. H. J. Biol. Chem. 1980, 255, 3436-3441. (6) Ferguson, W. S.; Finlay, T. H. Arch. Biochem. Biophys. 1983, 221, 304-307. (7) Huber, R.; Carrell, R. W. Biochemistry 1989, 28, 8951-8966. (8) Van Boeckel, C. A. A.; Grootenhuis, P. D. J.; Visser, A. Nature Struct. Biol. 1994, 1, 423-425. (9) Jin, L.; Abrahams, J. P.; Skinner, R.; Petitou, M.; Pike, R. N.; Carrell, R. W. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 14683-14688. (10) Arocas, V.; Bock, S. C.; Olson, S. T.; Bjo¨rk, I. Biochemistry 1999, 38, 10196-10204. (11) Stein, P. E.; Carrell, R. W. Nature Struct. Biol. 1995, 2, 96-113. (12) Tejada, M. L.; Deeley, R. G. Thromb. Haemost. 1995, 73, 654661. (13) Peterson, C. B.; Noyes, C. M.; Pecon, J. M.; Church, F. C.; Blackburn, M. N. J. Biol. Chem. 1987, 262, 8061-8065. (14) Fan, B.; Turko, I. V.; Gettins, P. G. W. Biochemistry 1994, 33, 14156-14161. (15) Desai, U. R.; Swanson, R.; Bock, S. C.; Bjo¨rk, I.; Olson, S. T. J. Biol. Chem. 2000, 275, 18976-18894. (16) Fan, B.; Turko, I. V.; Gettins, P. G. W. FEBS Lett. 1994, 354, 8488. (17) Kridel, S. J.; Chan, W. W.; Knauer, D. J. J. Biol. Chem. 1996, 271, 20935-20941. (18) Meagher, J. L.; Huntington, J. A.; Fan, B.; Gettins, P. G. W. J. Biol. Chem. 1996, 271, 29353-29358. (19) Arocas, V.; Turk, B.; Bock, S. C.; Olson, S. T.; Bjo¨rk, I. Biochemistry 2000, 39, 8512-8518. (20) Arocas, V.; Bock, S. C.; Raja, S. M.; Olson, S. T.; Bjo¨rk, I. J. Biol. Chem. 2001, 276, 43809-43817. (21) Huntington, J. A.; Mccoy, A.; Pei, X.; Gettins, P. G. W.; Carrell, R. W. J. Biol. Chem. 1999, 275, 15377-15383. (22) Hopkins, P. C. R.; Whisstock, J. Science 1994, 265, 1893-1894. (23) Chuang, Y.-J.; Gettins, P. G. W.; Olson, S. T. J. Biol. Chem. 1999, 274, 28142-18149. (24) Olson, S. T.; Stephens, A. W.; Hirs, C. H. W.; Bock, P. E.; Bjo¨rk, I. J. Biol. Chem. 1995, 270, 9717-9724. (25) Pike, R. N.; Potempa, J.; Skinner, R.; Fitton, H. L.; Mcgraw, W. T.; Travis, J.; Owen, M.; Jin, L.; Carrell, R. W. J. Biol. Chem. 1997, 272, 19652-19655. (26) Chuang, Y.-J.; Swanson, R.; Raja, S. M.; Bock, S. C.; Olson, S. T. Biochemistry 2001, 40, 6670-6679. (27) Futamura, A.; Beechem, J. M.; Gettins, P. G. W. Biochemistry 2001, 40, 6680-6687. (28) Peterson, C. B.; Blackburn, M. N. J. Biol. Chem. 1985, 260, 610615. (29) Turko, I. V.; Fan, B.; Gettins, P. G. W. FEBS Lett. 1993, 335, 9-12. (30) Frebelius, S.; Isaksson, S.; Swedenborg, J. Arterioscler. Thromb. Vasc. Biol. 1996, 16, 1292-1297. (31) Garone, L.; Edmunds, T.; Hanson, E.; Bernasconi, R.; Huntington, J. A.; Meagher, J. L.; Fan, B.; Gettins, P. G. W. Biochemistry 1996, 35, 8881-8889. (32) Kasturi, L.; Eshleman, J. R.; Wunner, W. H.; Shakin-Eshleman, S. H. J. Biol. Chem. 1995, 270, 14756-14761. (33) Kasturi, L.; Chen, H. G.; Shakin-Eshleman, S. H. Biochem. J. 1997, 323, 415-419. (34) Thompson, J. D.; Higgins, D. G.; Gibson, T., J. Nucleic Acids Res. 1994, 22, (35) Felsenstein, J. Cladistics 1989, 5, 164-166. (36) Koradi, R.; Billeter, M.; Wu ¨ thrich, K. J. Mol. Graphics 1996, 14, 52-55. (37) Kraulis, P. J. J. Appl. Crystallogr. 1991, 24, 946-950.
PR025515Z
Journal of Proteome Research • Vol. 1, No. 4, 2002 373
