Mutation of Highly Conserved Arginine Residues Disrupts the Structure and Function of Annexin V

Archives of Medical Research 30 (1999) 360–367

ORIGINAL ARTICLE

Mutation of Highly Conserved Arginine Residues Disrupts the Structure and Function of Annexin V Begoña Campos,* Songtao Wang,** Gregory S. Retzinger,*** Marcia A. Kaetzel,* Barbara A. Seaton,**** Norman J. Karin,***** J. David Johnson** and John R. Dedman* *Department of Molecular and Cellular Physiology, University of Cincinnati, College of Medicine, Cincinnati, OH, USA **Department of Pathology and Laboratory Medicine, University of Cincinnati, College of Medicine, Cincinnati, OH, USA ***Ohio State University, Department of Medical Biochemistry, Columbus, OH, USA ****Structural Biology Group, Department of Physiology, Boston University School of Medicine, Boston, MA, USA *****Department of Integrative Biology, University of Texas Medical School, Houston, TX, USA Received for publication January 29, 1999; accepted June 9, 1999 (99/010).

Background. Annexins are a family of structurally related proteins that bind to phospholipid membranes in a Ca21-dependent manner. Annexins are characterized by highly conserved canonical domains of approximately 70 amino acids. Annexin V contains four such domains. Each of these domains has a highly conserved arginine (R). Methods. To evaluate the role of the conserved arginines in the molecular structure of annexin V, negatively charged amino acids were substituted for arginines at positions R43, R115, R199, and R274 using site-directed mutagenesis. Results. Mutants R199D and R274E were rapidly degraded when expressed in bacteria, and were not further characterized. R43E exhibited an electrophoretic mobility similar to the wild-type protein, while R115E migrated significantly in a slower fashion, suggesting a less compact conformation. R43E and R115E exhibited much greater susceptibility to proteolytic digestion than the wild type. While Ca21-dependence for phospholipid binding was similar in both mutants (half-maximal 50–80 mM Ca21), R43E and R115E exhibited a 6- and 2-fold decrease in phospholipid affinity, respectively. Consistent with the different phospholipid affinities of the annexins, a phospholipid-dependent clotting reaction, the activated partial thromboplastin time (aPTT), was significantly prolonged by the wild-type protein and mutants R115E and R115A. The aPTT was unaffected by R43E. Conclusions. Our data suggest that mutation of these highly conserved arginine residues in each of the four canonical domains of annexin have differential effects on the phospholipid binding, tertiary structure, and proteolytic susceptibility of annexin V. The site I mutation, R43E, produced a large decrease in phospholipid affinity associated with an increase in proteolytic susceptibility. The site II mutation, R115E, produced a small change in phospholipid binding but a significant modification of electrophoretic mobility. Our data suggest that highly conserved arginine residues are required to stabilize the tertiary structure of annexin V by establishing hydrogen bonds and ionic bridges. © 1999 IMSS. Published by Elsevier Science Inc. Key Words: Annexins, Calcium-binding proteins, Mutagenesis, Structure.

Introduction Annexins are a family of structurally related proteins that bind to phospholipid membranes in a Ca21-dependent manner. Binding to phospholipid surfaces is a central feature of Address reprint requests to: John R. Dedman, M.D., Department of Molecular and Cellular Physiology, University of Cincinnati College of Medicine, Cincinnati, OH 45267-0576 USA. Tel.: (1513) 558-4145; FAX: (1513) 558-5738; E-mail: [email protected].

all proposed biological functions of annexins (for review, see 1–3). Members of this protein family have been shown to bind and bundle actin (4), modulate ion channels (5–7), form Ca21 channels (8), promote fusion of membranes (9) and budding of coated pits (10), mediate mitogenic signal transduction (11), and inhibit phospholipase A2 and blood coagulation (12,13). Annexins do not contain the EF-hand Ca21-binding motif as found in calmodulin, nor do they possess sequence

0188-4409/99 $–see front matter. Copyright © 1999 IMSS. Published by Elsevier Science Inc. PII S0188-0128(99)00 0 4 0 - 8

Campos et al. / Archives of Medical Research 30 (1999) 360–367

similarity with other Ca21 and phospholipid-binding proteins, such as protein kinase C (14). A comparison of the amino acid sequence between annexins reveals a common core structure consisting of four repeated domains. Each domain is composed of approximately 70 amino acids. This core structure is preceded by a highly variable amino terminal region. Huber et al. (15) and Concha et al. (16) found that the four-domain structure of annexin V is arranged in a planar, circular array. Each domain is composed of five alpha helices of 7–15 amino acid residues separated by loops of various lengths. The helices are designated A, B, C, D, and E. Two helix-loop-helix units are formed by helices A and B, and by D and E, respectively. Ca21 ions are atomically coordinated within these loops. These structures interact directly with membrane phospholipid head groups. The C helix lies perpendicular to, and stabilizes, the overall structure by capping the other four helices. Recently, Swairjo et al. (17) presented evidence for a calcium-bridging mechanism, in that the surface amino acid side-chains and the phospholipid head groups are both involved in Ca21 coordination. All annexins from vertebrates contain an invariant arginine (R) in each domain. Huber et al. (15) noted that these highly conserved charged amino acids (R45, R117, R201, and R276) are buried within the interior of human annexin V. Each of these arginines is positioned near the C-terminus of the B helix in domains I, II, III, and IV. Of the five helices in each domain, the B helix tends to be the most hydrophobic, constituting much of the domain core. The arginine side-chains are oriented inward toward the protein interior rather than outward toward the solvent (Figure 1). In rat annexin V, R43 and R115, located in domains I and II, respectively, hydrogen-bond with the preceding residue in the B helix, A42, and S114, respectively. These hydrogenbonded pairs then form a salt bridge with an acidic residue in the adjacent A helix, D18, and D90, respectively (17). In domain III, R199 coordination differs from R43 and R115 in that its guanidinium group lies almost parallel to the B helix (Figure 1) rather than perpendicular to it. In this orientation, R199 is not positioned to form a hydrogen bond with the D173 in the preceding A helix; D173 hydrogen bonds with H203, linking the A and C helices non-covalently. These functional groups are also highly conserved. In domain IV, the R274 interactions are unique. No acidic residues are present at the required position in the preceding helix. R274 hydrogen bonds instead with the acidic D278 in the neighboring C helix. This interaction reinforces the connection between the AB helix-loop-helix unit and the perpendicular capping C helix. The location of R274 in the central pore of the protein also allows this positively charged residue to participate in ionic interactions between domains such as that with E119. R274 and R115 are in close proximity in the central channel, a region containing many charged residues. In the present study, the unique role of these highly con-

361

served arginines was investigated using site-directed mutagenesis of rat annexin V. The physical properties and the phospholipid binding of the bacterially expressed wild-type and mutant annexin V were compared. These data suggest that the conserved arginines play a role in maintaining the tertiary structure of the protein.

Materials and Methods Mutagenesis procedure. The 1.5 kb XbaI/SacI fragment of the annexin V cDNA (18) cloned in the NotI site of pBSSKII was subcloned into the corresponding site of M13mp19, after which single-stranded template DNA was prepared (19). Single-primer mutagenesis was then performed on the M13 template with mutagenic oligonucleotides, as described by Kunkel et al. (20). Clones containing the desired mutation were identified by DNA sequencing using Sequenase Version 2.0 (Amersham Pharmacia Biotech, Piscataway, NJ, USA) following the manufacturer’s protocol. The M13mp19AV mutant replicative form was then prepared from clones containing the desired mutation(s), digested with NcoI/EcoRI, and ligated into NcoI-cut expression vector pKK233-2 (Amersham Pharmacia Biotech), which contains the hybrid trc promoter and the gene for ampicillin resistance. Following ligation of the NcoI end of the insert to the vector, the remaining free ends were blunted with Klenow and ligated (21). Presence of the desired mutation in the final plasmid construct was verified by double-strand DNA sequence analysis. Bacterial expression of annexin V. Cultures of Escherichia coli strain JM105 containing the desired plasmid were grown overnight at 378C in L-broth 1 100 mg/mL ampicillin and then diluted 1:10 in L-broth containing 100 mg/mL ampicillin. This starting culture (OD600 z 0.05) was grown at 378C until the OD600 reached 0.3–0.5. Isopropyl-b-D-thiogalactopyranoside (IPTG) was then added to a final concentration of 3 mM, after which growth was maintained for 4 h. Bacteria were harvested by centrifugation, washed once with 100 mL PBS containing 10 mM ethylenediaminetetraacetic acid disodium salt (EDTA) and stored at 2808C. Electrophoretic analysis of wild-type and mutant annexin V. The IPTG-induced bacterial cultures (3 mL) of E. coli containing wild-type or mutated recombinant annexin V were centrifuged. The pellets were dissolved in 1 mL of SDS sample buffer containing 0.5 M Tris-HCl (pH 6.8), 10% glycerol, 15% 2-mercaptoethanol, 6% sodium dodecyl sulfate (SDS), and 10 mM EDTA, and heated to 958C for 5 min. DEAE pre-purification of wild-type and mutant annexin V. E. coli–expressing recombinant wild-type or mutant annexin V were disrupted by sonication three times for 30 sec each on ice in imidazole buffer (20 mM imidazole, 1 mM EDTA,

362

Campos et al./ Archives of Medical Research 30 (1999) 360–367

Figure 1. Location of arginine targets of site-directed mutagenesis in wild-type rat annexin V. Domains are identified by Roman numerals. Helix-loop-helix units in each domain (except domain II) are oriented similarly: A helix, left; B helix, right; Ca21 is shown in the connecting loop. Domain II is reoriented to show its interface with domain IV. Hydrophobic residues that contribute to the domain core are shown explicitly, although they are also present in other domains.

pH 7.4, 1 mM phenylmethylsulfonyl fluoride (PMSF), 1 mM pepstatin, 1 mM leupeptin). The sonicated suspensions were centrifuged (15 min at 25,000 3 g) and the supernatants applied to DE-52 (Whatman) columns previously equilibrated

with imidazole buffer. The columns were then washed with 20 column volumes of the same buffer. The proteins were eluted sequentially with 100 mM, 150 mM, and 200 mM NaCl. The unbound material and washes were analyzed by

Campos et al. / Archives of Medical Research 30 (1999) 360–367

sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) using the system described by Bio-Rad Laboratories (Hercules, CA, USA). Proteins were transferred to nitrocellulose by the procedure of Towbin et al. (22), and immunoblot analysis was performed by incubating the blot with an antiserum directed against rat kidney annexin V (23) and visually detected using horseradish peroxidase-conjugated secondary antibody (Cappel) and 4-chloro-1-naphthol. Chymotryptic digestion of wild-type and mutant annexin V. Approximately 10 mg of each DEAE-purified annexin was incubated at 378C in 20 mM Tris-HCl, 200 mM NaCl, 5 mM CaCl2, 1 mM 1,4 dithiothreitol (DTT) (pH 7.4), with 1 mg of chymotrypsin. At 0, 1, 5, and 60 min of incubation, an aliquot was removed and added to SDS sample buffer [0.5 M Tris-HCl (pH 6.8), 10% glycerol, 2% SDS, and 15% 2-mercaptoethanol] at 958C to arrest proteolysis. Samples were analyzed by immunoblot, as previously described. Isolation of wild-type and mutant annexin V. IPTG–stimulated E. coli–expressing recombinant wild-type or mutated annexin V were sonicated for 2–3 min on ice in sonication buffer [PBS, 10 mM EDTA, 6 M urea, 1% Triton X-100, 0.5 mg/mL leupeptin, 0.5 mg/mL pepstatin, and 0.2-mM phenylmethanesulfonyl fluoride (PMSF)]. Both suspensions were centrifuged (15 min at 25,000 3 g) and the supernatants recovered and dialyzed three times at 48C against 1 L of dialysis buffer (50 mM TRIS-HCl, pH 8.0, 100 mM NaCl, 3 mM NaN3, and 1 mM EDTA). Each dialyzed sample was centrifuged (20 min at 15,000 3 g), and the resulting supernatant filtered through a 0.22-mm nitrocellulose membrane. The annexin V was then purified as described previously by Kaetzel et al. (23). In brief, each extract was brought to 200 mM in free Ca21 using an Orion standard (0.1000 M CaCl2) and was then applied to a phenyl-sepharose column conditioned with brain phospholipids (Sigma-Aldrich, St. Louis, MO, USA) and equilibrated with column buffer containing 200 mM-free CaCl2. The phenyl-sepharose column was sequentially washed with three column volumes of column buffer containing 0.2 M NaCl, no NaCl, and 0.5 M NaCl. Annexin V was eluted from the phenyl-sepharose by Ca21 chelation in a buffer containing 1 M NaCl and 1 mM EDTA without added Ca21. Fractions were analyzed by immunoblot, as described previously. Ca21-dependent phospholipid binding assay. Dansyl phosphatidylserine (DPS; Molecular Probes, Eugene, OR, USA) was incorporated into small unilamellar vesicles (liposomes) by tip sonication of a mixture of DPS (20%) and phosphatidylcholine (80%). DPS-liposomes were separated from free DPS on a Sephadex G-50 column equilibrated with 10 mM MOPS, 90 mM KCl at pH 7.0. Blood coagulation assay. Activated partial thromboplastin times (aPTTs) were measured using normal pooled plasma

363

chelated with citrate (0.013 M), actin FSL (Dade actin FSL activated PTT reagent, Dade Bohring, Inc., Deerfield, IL, USA) and an automated photometer (MLA 900, Medical Laboratory Automation, Plainville, NY, USA) according to a modification of Fritsma (24). In brief, 50 mL of PBS containing a precise quantity of annexin was mixed with an equivalent volume of aPTT reagents. After incubating this mixture at 378C for 30 min, 50% of the mixture was then used to determine the aPTT of normal pooled plasma.

Results Mutation and expression of annexin V. In order to explore the functional role of the conserved arginines in positions 43, 115, 199, and 274, they were substituted by a negatively charged amino acid using site-directed mutagenesis. R115 was also changed to the neutral amino acid, alanine (A). The following mutations were observed: in domain I, codon GCC (Arg43) was changed to CTC (R43E); in domain II, TCC (Arg115) was changed to CTC (R115E) or to CGC (R115A); in domain III, GCG (Arg199) was changed to CTG (R199D), and in domain IV, TCC (Arg274) was changed to CTC (R274E). These mutations were confirmed by double-stranded DNA sequencing. R199D and R274E proteins proved extremely unstable when expressed in bacteria, even in the protease-deficient bacterial strain Y1089 (Gibco BRL Life Technologies, Gaithersburg, MD, USA) and were not further characterized. Electrophoretic mobility. Purified wild-type rat recombinant annexin V was first compared with annexin V purified from rat kidney. Both forms of annexin V migrated identically in SDS-PAGE and had similar chromatographic properties (data not shown). However, during initial characterization of the recombinant annexin V mutants, differences in SDS-PAGE migration and DEAE anion-exchange chromatography were observed. Bacteria expressing the individual mutants were solubilized, denatured, and subjected to SDS-PAGE (Figure 2). A significant decrease (z3 kDa) in the relative electrophoretic mobility of annexin V occurred when the arginine substitution was in R115 of site II. A partially unfolded protein would be less compact and would cause the protein to migrate more slowly during SDSPAGE. The fact that changes in position 43 caused no mobility changes demonstrates that conserved arginines in domains I and II may provide distinct structural roles. Resistance to proteolysis. In order to determine whether annexin V mutants contain partially unfolded tertiary structures, wild-type and mutant proteins were treated with chymotrypsin. Under the conditions used, recombinant wild-type annexin V demonstrated limited digestion by chymotrypsin during the 60-min incubation period. In sharp contrast the R43E and R115E mutants, representing the first and second

364

Campos et al./ Archives of Medical Research 30 (1999) 360–367

Figure 2. Electrophoretic analysis of annexin V and its mutants. Cultures of E. coli containing the wild-type or mutated recombinant annexin V R43E and R115E were centrifuged. The pellets were dissolved in 1 mL of sample buffer and heated as described in Methods. Each sample was subjected to electrophoresis in a 12%-acrylamide-SDS gel. Immunoblot detection was performed using antiserum produced against rat kidney annexin V.

domain repeats, respectively, were rapidly digested within the first 5 min of incubation (Figure 3). These results strongly suggest that domain I and II mutations cause changes in tertiary structure, producing less compact polypeptides than the wild-type protein. Ca21-dependent phospholipid binding. To analyze the Ca21dependent phospholipid affinity of wild-type annexin V and its mutants, binding to fluorescent phospholipid, dansylphosphatidylserine (DPS) was assessed. Figure 4 shows that recombinant wild-type annexin V bound most tightly with a Kapp of approximately 13 nM. The R43E mutation in site I produced a z6-fold reduction in its affinity for DPS (Kapp

Figure 3. Immunoblot analysis of chymotrypsin digestion of annexin V wild-type and mutants. Annexin V wild-type and R43E and R115E were incubated at 378C in digestion buffer with 1 mg of chymotrypsin. At indicated times, an aliquot was removed and added to 958C sample buffer to arrest proteolysis. Immunoblots are performed as described in the Materials and Methods section.

z73 nM). R115E and R115A mutants produced a z2-fold decrease in their affinity for DPS (Kapp z23 or 25 nM, respectively). DPS:PC liposomes were also titrated with annexin and its mutants. Wild-type annexin V bound to these liposomes with a Kapp of z4 nM, while the R115E bound with z2-fold lower affinity (Kapp z8.3 nM) and the R43E mutant bound with z4-fold lower affinity (Kapp z17 nM). Thus, similar decreases in affinity were observed for both pure DPS and DPS incorporated into liposomes. In all cases, DPS binding was strictly dependent on Ca21 and could be reversed by EGTA. The Kd of annexin V binding to DPS was determined by titrating 25, 100, 250, and 500 nM DPS with annexin V and plotting the Kapp as a function of annexin concentration. The plot was fit to a linear regression (r 5 0.996), which intercepted the Y axis [zero(DPS)] at 0.2 6 0.35 nM. Therefore, the Kd of annexin V for DPS is z0.2 nM.

Figure 4. Binding of wild-type annexin V and its mutants to dansyl-phosphatidylserine (DPS). Increasing concentrations of protein were added to 0.5 mM DPS in a buffer of 10 mM MOPS, 90 mM KCl, 1 mM CaCl2 at pH 7.0. Each point is the average of three titrations 6 SE. The increase in DPS fluorescence is shown as a function of annexin concentration. A 100%-fluorescence increase corresponds to a 1.5- to 1.7-fold increase over basal fluorescence. DPS fluorescence was monitored at 500 nm with excitation at 340 nm on a Perkin-Elmer LS-5 spectrofluorometer (Perkin-Elmer Co., Norwalk, CT, USA) at 208C.

Campos et al. / Archives of Medical Research 30 (1999) 360–367

Figure 5. Calcium dependence of annexin V and its mutants R43E and R115E binding to DPS:PC liposomes. Increasing concentrations of calcium were added to 0.5 mM DPS in DPS:PC liposomes (20:80, by weight) in the presence of 100 nM wild-type annexin (d), or R43E (r) or R115E (m) in 10 mM MOPS, 90 mM KCl at pH 5 7.0. Calcium-dependent increases in DPS fluorescence were followed at 500 nm with excitation at 340 nm at 208C. Each point is the average of three titrations 6 SE. The 100%-fluorescence increase corresponds to a 1.2-fold increase over basal (no added calcium) DPS fluorescence.

Ca21-dependent phospholipid-binding mutant protein properties were compared with wild-type protein properties by following those binding to DPS:PC-containing liposomes (Figure 5). Wild-type annexin V, R43E, and R115E half-maximally bound at z53 mM, 64 mM, and 84 mM Ca21, respectively. Therefore, mutations in domains I and II do not significantly affect Ca21-dependent binding to DPS:PC liposomes. Inhibition of blood coagulation. By virtue of its high affinity for phospholipid surfaces, annexin V is a potent inhibitor of coagulation (25). It was of interest, therefore, to compare the anticoagulant activity of the wild-type protein to that of its mutants. For this purpose, a modification of activated partial thromboplastin time (aPTT) was used. As a preliminary screen, high concentrations (2.28 mM) of the various annexins were tested to determine whether the aPTT of normal pooled plasma was prolonged. At this high concentration, the wild-type protein and mutants R115E and R115A did prolong the aPTT, while mutant R43E had no effect. Dose-response profiles of the wild-type protein and mutants R115E and R115A were determined and found to be similar (Figure 6). These data indicate that the anticoagulant activity of annexin V is sensitive to the mutation of R43, which prolonged the aPTT 15%, but was indifferent to mutation R115. The results are consistent with the observation that high affinity for anionic phospholipid surfaces, a property shared by the wild-type protein and mutants R115E and R115A, contributes to the anticoagulant activity of annexin V. All mutants studied retained Ca21-dependent phospho-

365

Figure 6. Anticoagulant activity of annexin V and its mutants. APTTs were performed using normal pooled plasma to which wild-type annexin V (d), mutant R43E (r), mutant R115E (m), or mutant R115A (j) had been added. The shaded zone indicates the aPTT of normal pooled plasma in the absence of annexin. The asterisk (*) indicates that no clot was detected in plasma samples containing a concentration of R115A in excess of 0.60 mM.

lipid binding and were sensitive to proteolytic digestion. Mutant R43E displayed reduced affinity for phospholipids, and mutants R115E and R115A displayed a reduced rate of electrophoretic mobility.

Discussion Site-directed mutagenesis is a valuable approach to evaluate the functional importance of highly conserved motifs, which have been identified by sequence comparisons and 3-D structural data. Annexins contain an invariant arginine in each of four 70 amino acid-repeat domains. The unique structural role of arginines arises from the fact that the guanidinium side-chain is longer and has greater potential for hydrogenbonding and ionic interactions when compared with lysine or histidine (26). This property enables arginine to interconnect discontinuous stretches of the polypeptide chain to stabilize higher order conformations. In fact, the frequency of arginine is greater in proteins isolated from thermophilic organisms that live at extremely high temperatures when compared with their mesophilic counterparts (27). The highly conserved arginines evaluated in this study play a major role in the stability of annexin V. The primary intent of using SDS-PAGE is to denature and separate proteins according to relative amino-acid length. Many proteins, however, are resistant to heat and detergent denaturation and the migration rate reflects the state of tertiary structure. Calmodulin, for example, retains the ability to bind Ca21 following boiling in the presence of re-

366

Campos et al./ Archives of Medical Research 30 (1999) 360–367

ducing agents and SDS. The Ca21-bound form has a greater a-helical content and migrates faster on SDS-PAGE than the Ca21-free form (28). An initial observation in this study was that R115E migrated more slowly than the wild-type annexin V. DNA sequence data of each annexin V mutant studied herein confirmed that the predicted amino-acid length of the wild-type and mutant proteins were identical—319 residues. Others (29,30) have observed changes in electrophoretic mobility as a consequence of a single amino-acid substitution. In fact, Donato et al. (31) identified two isoforms of annexin V in bovine brain that migrate at 33 and 37 kDa, respectively, which were demonstrated by Learmonth et al. (32) to be the result of a naturally occurring single amino-acid substitution of glutamate to lysine at position 125. Our results show that annexin V binds phospholipids with 0.2 nM affinity. This is similar to placental annexin V binding to 20% PS/80% PC liposomes with a Kd of 0.1 nM (33). Further, we demonstrate that 53 mM Ca21 will produce half-maximal binding of annexin V to DPS:PC liposomes. This result is in good agreement with the results of Schlaepfer et al. (34), who demonstrated that 53 mM Ca21 will produce half-maximal binding of annexin V to PS liposomes. Local destabilization of the tertiary structure would also lead to partial unfolding of core residues and produce greater susceptibility to proteolysis. Localized unfolding would cause R43E and R115E mutants to exhibit increased sensitivity to chymotryptic digestion while retaining the ability to bind Ca21. Changes in the tertiary structure of mutant proteins do not dramatically affect the Ca21 dependence of phospholipid binding. R43E mutation does, however, produce a dramatic decrease in affinity for phospholipids. In contrast, mutations R115E and R115A produce little change in phospholipid affinity. These results also demonstrate that changing positively charged for negatively charged amino acid had the same effect as substitution of conserved arginine for a non-charged group. Such an interpretation is supported by the observation that wild-type annexin V and the R115E and R115A mutants significantly prolong the aPTT, while R43E does not affect aPTT. Tait and Smith (35) produced four individual mutants of annexin V by substituting alanine for arginine 200, histidine 204, arginine 206, or lysine 207. They found that recombinant proteins bound phospholipids in a Ca21-dependent manner. These residues are located on or near the concave surface of the molecule, far from the Ca21membrane-binding surface. It is, therefore, not surprising that Ca21-dependent binding to phospholipids in their mutants is unaffected. AB and DE helix connectors are short loops that contain Ca21-binding sites. We postulate that structural changes in the mutant R43E annexin V molecule completely disrupted the Ca21-bridge in domain I and decreased phospholipid affinity, supporting the idea that structural complementarity

between annexin and phospholipids is required for phospholipid binding (17). It is possible that some Ca21-bridging regions remain intact in R115E and R115A, allowing phospholipid binding in domain II. These data suggest that Ca21binding sites may function independently because mutant proteins maintained the capacity to bind phospholipids in a Ca21-dependent manner. The net effect of polar interactions involving R43 and R115 appears to increase the stability of helix-loop-helix Ca21-binding units in domains I and II, respectively. In all annexin V domains, the interface between helices A and B contains many hydrophobic core residues (16). While hydrophobic interactions stabilize overall polypeptide folding, polar residues are positioned to further align helices into proper tertiary structure. Mutation of either R43 or R115 to an acidic residue appears to disrupt this important hydrogenbonding network, and destabilizes the native configuration. In summary, wild-type annexin V, despite the absence of disulfide bonds, is an exceptionally stable protein. Each domain possesses a hydrophobic core, locked into place by highly conserved electrostatic interactions (15,16). Each of the four conserved arginines evaluated in this study appears to be involved in these important structural interactions. Mutation of these invariant-to-acidic-residue argines disrupts the conformational integrity of the protein. Our studies have shown that each domain may be disrupted independently, consistent with the four-domain structure of annexin V, where each domain forms an independent folding unit. Mutational analysis of annexins can allow evaluation of the pathophysiologic roles of these proteins, such as the role of annexin V as a thromboregulatory protein in the Antiphospholipid Antibody Syndrome (34) or the role of annexin II in leukemia (35) and other possible annexinopathies (36) that have not been discovered.

Acknowledgments We would like to thank Glenn Doerman for the preparation of the graphics. This study was supported by the following grants: National Institutes of Health #GM-44554 (BAS), #DK46433 (JRD and BC), and #DK33727 (JDJ); the American Heart Association #9930171(BC), and American Heart Association Ohio Valley Affiliate #9806236 (BC).

References 1. Raynal P, Pollard HB. Annexins: the problem of assessing the biological role for a gene family of multifunctional calcium- and phospholipid-binding proteins. Biochim Biophys Acta 1994;1197;63. 2. Swairjo MA, Seaton BA. Annexin structure and membrane interactions: a molecular perspective. Annu Rev Biophys Biomol Struct 1994;23:193. 3. Seaton BA, editor. Annexins: molecular structure to cellular function. Austin: R.G. Landes Co.1996. 4. Jones PG, Moore GJ, Waisman DM. A nonapeptide to the putative F-actine binding site of annexin-II tetramer inhibits its calcium-dependent activation of actin filament bundling. J Biol Chem 1992;267:13993.

Campos et al. / Archives of Medical Research 30 (1999) 360–367 5. Díaz-Muñoz M, Hamilton SL, Kaetzel MA, Hazarika P, Dedman JR. Modulation of Ca21 release channel activity from sarcoplasmic reticulum by annexin VI (67-kDa calcimedin). J Biol Chem 1990;265:15894. 6. Kaetzel MA, Chan HC, Dubinsky WP, Dedman JR, Nelson DJ. A role for annexin IV in epithelial cell function. Inhibition of calcium-activated chloride conductances. J Biol Chem 1994;269:5297. 7. Naciff JM, Behbehan MM, Kaetzel M, Dedman, JR. Annexin VI modulates Ca21 and K1 conductances of spinal cord and dorsal root ganglion neurons. Am J Physiol Cell Physiol 1996;271(Cell Physiol 40):C2004. 8. Rojas E, Pollard HB, Haigler HT, Parra C, Burns AL. Calcium-activated endonexin II forms calcium channels across acidic phospholipid bilayer membranes. J Biol Chem 1990;265:21207. 9. Creutz C. The annexins and exocytosis. Science 1990;258:924. 10. Lin HC, Sudhof TC, Anderson RGW. Annexin VI is required for budding of clathrin-coated pits. Cell 1992;70:283. 11. Haigler HT, Schlaepfer DD, Burgess WH. Characterization of lipocortin I and an immunologically unrelated 33-kDa protein as epidermal growth factor receptor/kinase substrates and phospholipase A2 inhibitors. J Biol Chem 1987;262:6921. 12. Tait JF, Sakata M, McMullen BA, Miao CH, Funakoshi T, Hendrickson LE, Fujikawa K. Placental anticoagulant proteins: isolation and comparative characterization four members of the lipocortin family. Biochemistry 1988;27:6268. 13. Andree HAM, Stuart MCA, Hermens WT, Reutelingsperger CPM, Hemker HC, Frederik PM, Willems GM. Clustering of lipid-bound annexin V may explain its anticoagulant effect. J Biol Chem 1992; 267:17907. 14. Smith VL, Dedman JR. Stimulus response coupling: the role of intracellular calcium-binding proteins. Boca Raton, FL: CRC Press;1990. 15. Huber R, Berendes R, Burger A, Schneider M, Karshikov A, Luecke H, Romisch J, Paques E. Crystal and molecular structure of human annexin V after refinement. Implications for structure, membrane binding and ion channel formation of the annexin family of proteins. J Mol Biol 1992;223:683. 16. Concha NO, Head JF, Kaetzel MA, Dedman JR, Seaton BA. Rat annexin V crystal structure: Ca(2+)-induced conformational changes. Science 1993;261:1321. 17. Swairjo MA, Concha NO, Kaetzel MA, Dedman JR, Seaton BA. Ca(21)-bridging mechanism and phospholipid head group recognition in the membrane-binding protein annexin V. Nat Struct Biol 1995;2:968. 18. Wallner BP, Mattaliano RJ, Hession C, Cate RL, Tizard R, Sinclair LK, Foeller C, Chow EP, Browning JL, Ramachandran KL, Pepinsky RB. Cloning and expression of human lipocortin, a phospholipase A2 inhibitor with potential anti-inflammatory activity. Nature 1986;320:77. 19. Sambrook J, Fritsch EF, Maniatis T. Molecular cloning. A laboratory manual. 2nd ed. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press;1989. 20. Kunkel TA, Roberts JD, Zabour RA. Rapid and efficient site-specific mutagenesis without phenotype selection. Methods Enzymol 1987;154:367.

367

21. Amann E, Brosius J, Ptashne M. Vectors bearing a hybrid trp-lac promoter useful for regulated expression of cloned genes in Escherichia coli. Gene 1983;25:167. 22. Towbin HT, Staehlin T, Gordon J. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc Natl Acad Sci USA 1979;76:4350. 23. Kaetzel MA, Hazarika P, Dedman JR. Differential tissue expression of three 35-kDa annexin calcium-dependent phospholipid-binding proteins. J Biol Chem 1989;264:14463. 24. Fritsma GA. The role of intracellular calcium-binding proteins in stimulus-response coupling. In: Corriveau DM, Fritsma, GA, editors. Hemostasis and thrombosis in the clinical laboratory. Philadelphia: Lippencott Co.;1989. p. 92. 25. Sammaritano RL, Gharavi AE, Soberano C, Levy RA, Lockshin MD. Phospholipid binding of antiphospholipid antibodies and placental anticoagulant protein. J Clin Immunol 1992;12:2735. 26. Mrabet NT, Van den Broeck A, Van den Brande I, Stanssens P, Laroche Y, Lambeir AM, Matthijssens G, Jenkins J, Chiadmi M, van Tilbeurgh H, Rey F, Janin J, Quax W, Lasters I, De Maeyer M, Wodak SJ. Arginine residues as stabilizing elements in proteins. Biochemistry 1992;31:2239. 27. Argos P, Rossmann MG, Grau UM, Zuber H, Frank G, Tratschin JD. Thermal stability and protein structure. Biochemistry 1979;18:5698. 28. Dedman JR, Kaetzel MA. Calmodulin purification and fluorescent labeling. Methods Enzymol 1983;102:1. 29. Fasano O, Aldrich T, Tamanoi F, Taparowsky E, Purth M, Wigles M. Analysis of the transforming potential of the human H-ras gene by random mutagenesis. Proc Natl Acad Sci USA 1984;81:4008. 30. Wittinghofer A, Pai EF. The structure of Ras protein: a model for a universal molecular switch. Trends Biochem Sci 1991;16:382. 31. Donato R, Giambanco I, Aisa MC, Ceccarelli P, Di Geronimo G. Identification of two novel acidic, calcium-binding proteins in bovine brain. Cell Biol Int 1988;12:565. 32. Learmonth MP, Howell SA, Harris ACM, Amess B, Patel Y, Giambanco J, Bianchi R, Pula G, Ceccarelli P, Donato R, Green BN, Artken A. Novel isoforms of CaBP 33/37 (annexin V) from mammalian brain: structural and phosphorylation differences that suggest distinct biological roles. Biochim Biophys Acta 1192;1160:76. 33. Tait JF, Gibson D, Fujikawa K. Phospholipid binding properties of human placental anticoagulant protein-I, a member of the lipocortin family. J Biol Chem 1989;264:7944. 34. Rand JH. Antiphospholipid antibody syndrome: new insights on thrombogenic mechanism. Am J Med Sci 1998;316:142. 35. Menell JS, Cesarman GM, Jacovina AT, McLaughlin MA, Lev EA, Hajjar KA. Annexin II and bleeding in acute promyelocytic leukemia. N Engl J Med 1999;340:994. 36. Rand JH. Annexinopathies: a new class of diseases. N Engl J Med 1999;340:1035.