N- and C-capping preferences for all 20 amino acids in alpha-helical peptides.

We have determined the N- and C-capping preferences of all 20 amino acids by substituting residue X in the peptides NH2-XAKAAAAKAAAAKAAGY-CONH2 and in Ac-YGAAKAAAAKAAAAKAX-CO2H. Helix contents were measured by CD spectroscopy to obtain rank orders of capping preferences. The data were further analyzed by our modified Lifson-Roig helix-coil theory, which includes capping parameters (n and c), to find free energies of capping (-RT ln n and -RT ln c), relative to Ala. Results were obtained for charged and uncharged termini and for different charged states of titratable side chains. N-cap preferences varied from Asn (best) to Gln (worst). We find, as expected, that amino acids that can accept hydrogen bonds from otherwise free backbone NH groups, such as Asn, Asp, Ser, Thr, and Cys generally have the highest N-cap preference. Gly and acetyl group are favored, as are negative charges in side chains and at the N-terminus. Our N-cap preference scale agrees well with preferences in proteins. In contrast, we find little variation when changing the identity of the C-cap residue. We find no preference for Gly at the C-cap in contrast to the situation in proteins. Both N-cap and C-cap results for Tyr and Trp are inaccurate because their aromatic groups affect the CD spectrum. The data presented here are of value in rationalizing mutations at capping sites in proteins and in predicting the helix contents of peptides.     

Conformational behavior of ionic self-complementary peptides

Several de novo designed ionic peptides that are able to undergo conformational change under the influence of temperature and pH were studied. These peptides have two distinct surfaces with regular repeats of alternating hydrophilic and hydrophobic side chains. This permits extensive ionic and hydrophobic interactions resulting in the formation of stable beta-sheet assemblies. The other defining characteristic of this type of peptide is a cluster of negatively charged aspartic or glutamic acid residues located toward the N-terminus and positively charged arginine or lysine residues located toward the C-terminus. This arrangement of charge balances the alpha-helical dipole moment (C --> N), resulting in a strong tendency to form stable alpha-helices as well. Therefore, these peptides can form both stable alpha-helices and beta-sheets. They are also able to undergo abrupt structural transformations between these structures induced by temperature and pH changes. The amino acid sequence of these peptides permits both stable beta-sheet and alpha-helix formation, resulting in a balance between these two forms as governed by the environment. Some segments in proteins may also undergo conformational changes in response to environmental changes. Analyzing the plasticity and dynamics of this type of peptide may provide insight into amyloid formation. Since these peptides have dynamic secondary structure, they will serve to refine our general understanding of protein structure.     

NMR characterization of structure, backbone dynamics, and glutathione binding of the human macrophage migration inhibitory factor (MIF).

Human macrophage migration inhibitory factor is a 114 amino acid protein that belongs to the family of immunologic cytokines. Assignments of 1H, 15N, and 13C resonances have enabled the determination of the secondary structure of the protein, which consists of two alpha-helices (residues 18-31 and 89-72) and a central four-stranded beta-sheet. In the beta-sheet, two parallel beta-sheets are connected in an antiparallel sense. From the total of three cysteines present in the primary structure of MIF, none was found to form disulfide bridges. 1H-15N heteronuclear T1, T2, and steady-state NOE measurements indicate that the backbone of MIF exists in a rigid structure of limited conformational flexibility (on the nanosecond to picosecond time scale). Several residues located in the loop regions and at the N termini of two helices exhibit internal motions on the 1-3 ns time scale. The capacity to bind glutathione was investigated by titration of a uniform 15N-labeled sample and led us to conclude that MIF has, at best, very low affinity for glutathione.     

Conformational transitions and fibrillation mechanism of human calcitonin as studied by high-resolution solid-state 13C NMR

Conformational transitions of human calcitonin (hCT) during fibril formation in the acidic and neutral conditions were investigated by high-resolution solid-state 13C NMR spectroscopy. In aqueous acetic acid solution (pH 3.3), a local alpha-helical form is present around Gly10 whereas a random coil form is dominant as viewed from Phe22, Ala26, and Ala31 in the monomer form on the basis of the 13C chemical shifts. On the other hand, a local beta-sheet form as viewed from Gly10 and Phe22, and both beta-sheet and random coil as viewed from Ala26 and Ala31 were detected in the fibril at pH 3.3. The results indicate that conformational transitions from alpha-helix to beta-sheet, and from random coil to beta-sheet forms occurred in the central and C-terminus regions, respectively, during the fibril formation. The increased 13C resonance intensities of fibrils after a certain delay time suggests that the fibrillation can be explained by a two-step reaction mechanism in which the first step is a homogeneous association to form a nucleus, and the second step is an autocatalytic heterogeneous fibrillation. In contrast to the fibril at pH 3.3, the fibril at pH 7.5 formed a local beta-sheet conformation at the central region and exhibited a random coil at the C-terminus region. Not only a hydrophobic interaction among the amphiphilic alpha-helices, but also an electrostatic interaction between charged side chains can play an important role for the fibril formation at pH 7.5 and 3.3 acting as electrostatically favorable and unfavorable interactions, respectively. These results suggest that hCT fibrils are formed by stacking antiparallel beta-sheets at pH 7.5 and a mixture of antiparallel and parallel beta-sheets at pH 3.3.     

Molecular dynamics simulations of water within models of ion channels.

The transbilayer pores formed by ion channel proteins contain extended columns of water molecules. The dynamic properties of such waters have been suggested to differ from those of water in its bulk state. Molecular dynamics simulations of ion channel models solvated within and at the mouths of their pores are used to investigate the dynamics and structure of intra-pore water. Three classes of channel model are investigated: a) parallel bundles of hydrophobic (Ala20) alpha-helices; b) eight-stranded hydrophobic (Ala10) antiparallel beta-barrels; and c) parallel bundles of amphipathic alpha-helices (namely, delta-toxin, alamethicin, and nicotinic acetylcholine receptor M2 helix). The self-diffusion coefficients of water molecules within the pores are reduced significantly relative to bulk water in all of the models. Water rotational reorientation rates are also reduced within the pores, particularly in those pores formed by alpha-helix bundles. In the narrowest pore (that of the Ala20 pentameric helix bundle) self-diffusion coefficients and reorientation rates of intra-pore waters are reduced by approximately an order of magnitude relative to bulk solvent. In Ala20 helix bundles the water dipoles orient antiparallel to the helix dipoles. Such dipole/dipole interaction between water and pore may explain how water-filled ion channels may be formed by hydrophobic helices. In the bundles of amphipathic helices the orientation of water dipoles is modulated by the presence of charged side chains. No preferential orientation of water dipoles relative to the pore axis is observed in the hydrophobic beta-barrel models.     

Flexibility of beta-sheets: principal component analysis of database protein structures

Protein folds are built primarily from the packing together of two types of structures: alpha-helices and beta-sheets. Neither structure is rigid, and the flexibility of helices and sheets is often important in determining the final fold (e.g., coiled coils and beta-barrels). Recent work has quantified the flexibility of alpha-helices using a principal component analysis (PCA) of database helical structures (J. Mol. Bio. 2003, 327, pp. 229-237). Here, we extend the analysis to beta-sheet flexibility using PCA on a database of beta-sheet structures. For sheets of varying dimension and geometry, we find two dominant modes of flexibility: twist and bend. The distributions of amplitudes for these modes are found to be Gaussian and independent, suggesting that the PCA twist and bend modes can be identified as the soft elastic normal modes of sheets. We consider the scaling of mode eigenvalues with sheet size and find that parallel beta-sheets are more rigid than antiparallel sheets over the entire range studied. Finally, we discuss the application of our PCA results to modeling and design of beta-sheet proteins.     

A general model of the P2 protein of peripheral nervous system myelin based on secondary structure predictions, tertiary folding principles, and experimental observations

The amino acid sequence of the P2 protein of peripheral myelin was analyzed with regard to regions of probable alpha-helix, beta-structure, beta-turn, and unordered conformation by means of several algorithms commonly used to predict secondary structure in proteins. Because of the high beta-sheet content and virtual absence of alpha-helix shown by the circular dichroic spectra of the protein, a bias was introduced into the algorithms to favor the beta-structure over the alpha-helical conformation. In order to define those beta-sheet residues that could lie on the external hydrophilic surface of the protein and those that could lie in its hydrophobic interior, the predicted beta-strands were examined for charged and uncharged amino acids located at alternating positions in the sequence. The sequential beta-strands in the predicted secondary structure were then ordered into beta-sheets and aligned according to generally accepted tertiary folding principles and certain chemical properties peculiar to the P2 protein. The general model of the P2 protein that emerged was a "Greek key" beta-barrel, consisting of eight antiparallel beta-strands with a two-stranded ribbon of antiparallel beta-structure emerging from one end. The model has an uncharged, hydrophobic core and a highly hydrophilic surface. The two Cys residues, which form a disulfide, occur in a loop connecting two adjacent antiparallel strands. Two hydrophilic loops, each containing a cluster of acidic residues and a single Phe, protrude from one end of the molecule. The general model is consistent with many of the properties of the actual protein, including the relatively weak nature of its association with myelin lipids and the positions of amino acid substitutions. Alternative beta-strand orderings yield three specific models having different interstrand connections across the barrel ends.     

Rate of beta-structure formation in polypeptides

An explanation is suggested for why a marginally stable beta-structure folds extremely slowly; it is predicted that even a small increase in stability drastically accelerates beta-folding. According to the theory, this folding is a first-order phase transition, and the rate-limiting step is nucleation. The rate-determining "nucleus" (transition state) is the smallest beta-sheet that is sufficiently large to provide an overall free energy reduction during subsequent folding. If the stability of the beta-structure is low, the nucleus is large and possesses a high free energy due to having a large perimeter. When the net stability of the final beta-structure increases (due to either an increase of the beta-sheet stability or a decrease in stability of the competing structures, e.g., alpha-helices), the size and energy of a nucleus decrease and the rate of folding increases exponentially. This must result in a fast folding of polypeptides enriched by beta-forming residues (e.g., protein chains). The theory is developed for intramolecular beta-structure, but it can also explain the overall features of intermolecular beta-folding; it is applicable both to antiparallel and parallel beta-sheets. The difference in folding of beta-sheets, alpha-helices, and proteins is discussed.     

Porcine pancreatic phospholipase A2: sequence-specific 1H and 15N NMR assignments and secondary structure

The solution structure of porcine pancreatic phospholipase A2 (124 residues, 14 kDa) has been studied by two-dimensional homonuclear 1H and two- and three-dimensional heteronuclear 15N-1H nuclear magnetic resonance spectroscopy. Backbone assignments were made for 117 of the 124 amino acids. Short-range nuclear Overhauser effect (NOE) data show three alpha-helices from residues 1-13, 40-58, and 90-109, an antiparallel beta-sheet for residues 74-85, and a small antiparallel beta-sheet between residues 25-26 and 115-116. A 15N-1H heteronuclear multiple-quantum correlation experiment was used to monitor amide proton exchange over a period of 22 h. In total, 61 amide protons showed slow or intermediate exchange, 46 of which are located in the three large helices. Helix 90-109 was found to be considerably more stable than the other helices. For the beta-sheets, four hydrogen bonds could be identified. The secondary structure of porcine PLA in solution, as deduced from NMR, is basically the same as the structure of porcine PLA in the crystalline state. Differences were found in the following regions, however. Residues 1-6 in the first alpha-helix are less structured in solution than in the crystal structure. Whereas in the crystal structure residues 24-29 are involved both in a beta-sheet with residues 115-117 and in a hairpin turn, the expected hydrogen bonds between residues 24-117 and 25-29 do not show slow exchange behavior. This and the absence of several expected NOEs imply that this region has a less well defined structure in solution. Finally, the hydrogen bond between residues 78-81, which is part of a beta-sheet, does not show slow exchange behavior.     

On the significance of alternating patterns of polar and non-polar residues in beta-strands

A common assumption about protein sequences in beta-strands is that they have alternating patterns of polar and non-polar residues. It is thought that such patterns reflect the interior/exterior geometry of amino acid residue side-chains on a beta-sheet. Here we study the prevalence of simple hydrophobicity patterns in parallel and antiparallel beta-sheets in proteins of known structure and in the sequences of amyloidogenic proteins. The occurrence of 32 possible pentapeptide binary patterns (polar (P)/non-polar (N)) is computed in 1911 non-homologous protein structures. Despite their tendency to aggregate in experimentally designed proteins, the purely alternating hydrophobic/polar patterns (PNPNP and NPNPN) are most frequent in beta-sheets, typically occurring in antiparallel strands. The overall distribution of the pentapeptide binary patterns is significantly different in strands within parallel and antiparallel sheets. In both types of sheets, complementary patterns (where the hydrophobic and polar residues pair with one another) associate preferentially. We do not find alternating patterns to be common in amyloidogenic proteins or in short fragments involved directly in amyloid formation. However, we do note some similarities between patterns present in amyloidogenic sequences and those in parallel strands.     

Discovery of a significant,nontopological preference for antiparallel alignment of helices with parallel regions in sheets

To help elucidate the role of secondary structure packing preferences in protein folding, here we present an analysis of the packing geometry observed between alpha-helices and between alpha-helices and beta-sheets in 1316 diverse, nonredundant protein structures. Finite-length vectors were fit to the alpha-carbon atoms in each of the helices and strands, and the packing angle between the vectors, Omega, was determined at the closest point of approach within each helix-helix or helix-sheet pair. Helix-sheet interactions were found in 391 of the proteins, and the distributions of Omega values were calculated for all the helix-sheet and helix-helix interactions. The packing angle preferences for helix-helix interactions are similar to those previously observed. However, analysis of helix-strand packing preferences uncovered a remarkable tendency for helices to align antiparallel to parallel regions of beta-sheets, independent of the topological constraints or prevalence of beta-alpha-beta motifs in the proteins. This packing angle preference is significantly diminished in helix interactions involving mixed and antiparallel beta-sheets, suggesting a role for helix-sheet dipole alignment in guiding supersecondary structure formation in protein folding. This knowledge of preferred packing angles can be used to guide the engineering of stable protein modules.     

Structure of the lamprey yolk lipid-protein complex lipovitellin-phosvitin at 2.8 A resolution

The X-ray crystallographic structure of the lipid-protein complex lipovitellin-phosvitin has been determined with the multiple isomorphous replacement method using four heavy-atom derivatives. Lamprey yolk lipovitellin-phosvitin is a dimeric molecule of molecular weight 352,000. The monomer consists of three polypeptide chains. The smallest is known as phosvitin and has an extremely high phosphoserine content. The monomeric unit also contains about 16% (w/w) of non-covalently bound lipid, probably in a monolayer or bilayer-like configuration. Within each monomer is a "cavity" or region of low electron density. The cavity has a volume of about 68,000 A3 and is believed to contain the lipid in a presumably disordered state. The cavity is roughly conical in shape and is lined on two sides by seven and eight-stranded antiparallel beta-sheets. The base of the cavity opens away from the intersubunit interface, but appears partially closed off from solvent regions by additional antiparallel beta-sheet structure. The beta-sheets lining the sides of the cavity are surrounded by a shell of two curved layers of 16 interconnected helices. The helices in either layer of the shell are all roughly parallel to each other and antiparallel to all of the helices of the other layer. The connectivity of the helices resembles a "superhelix" and is different from the connectivities seen in proteins containing four-helix bundles. There are an estimated 1300 amino acids in lamprey lipovitellin-phosvitin and almost 1000 alanine residues have been modeled into electron density. The remaining residues are assumed to be disordered.     

Three-dimensional structure determined for a subunit of human tRNA splicing endonuclease (Sen15) reveals a novel dimeric fold

Splicing of eukaryal intron-containing tRNAs requires the action of the heterotetrameric splicing endonuclease, which is composed of two catalytic subunits, Sen34 and Sen2, and two structural subunits, Sen15 and Sen54. Here we report the solution structure of the human tRNA splicing endonuclease subunit HsSen15. To facilitate the structure determination, we removed the disordered 35 N-terminal and 14 C-terminal residues of the full-length protein to produce HsSen15(36-157). The structure of HsSen15(36-157), the first for a subunit of a eukaryal splicing endonuclease, revealed that the protein possesses a novel homodimeric fold. Each monomer consists of three alpha-helices and a mixed antiparallel/parallel beta-sheet, arranged in a topology similar to that of the C-terminal domain of Methanocaldococcus jannaschii endonuclease. The dimeric interface is dominated by a beta-barrel structure, formed by face-to-face packing of two, three-stranded beta-sheets. Each of the beta-sheets results from reciprocal parallel pairing of one beta-strand from one subunit with two other beta-strands from the symmetric subunit. The structural model provides insights into the functional assembly of the human tRNA splicing endonuclease.     

1H NMR assignment and secondary structure of the Ca2(+)-free form of the amino-terminal epidermal growth factor like domain in coagulation factor X

Blood coagulation factor X is composed of discrete domains, two of which are homologous to the epidermal growth factor (EGF). The N-terminal EGF like domain in factor X (fX-EGFN), residues 45-86 of the intact protein, contains a beta-hydroxylated aspartic acid and has one Ca2(+)-binding site. Using 2D NMR techniques, we have made a full assignment of the 500-MHz 1H NMR spectrum of Ca2(+)-free fX-EGFN. On the basis of this assignment and complementary NOESY experiments, we have also determined the secondary structure of Ca2(+)-free fX-EGFN in water solution. Residues 45-49 are comparatively mobile, whereas residues 50-56 are constrained by two disulfide bonds to one side of an antiparallel beta-sheet involving residues 59-64 and 67-72. Another antiparallel beta-sheet involves residues 76-77 and 83-84. A small, parallel beta-sheet connects residues 80-81 and 55-56 and thereby orients the two antiparallel beta-sheets relative to each other. Four beta-turns are identified, involving residues 50-53, 56-59, 64-67, and 73-76. Residues 78-82 adopt an extended bend structure. On the basis of secondary structure and the location of the three disulfide bonds, we find that Asp 46, Asp 48, and Hya 63 are sufficiently close to each other to form a Ca2(+)-binding site. However, the amino terminus of the Ca2(+)-free form of fX-EGFN is not part of a triple-stranded beta-sheet as in other EGF like peptides. Differences and similarities between fX-EFGN and murine EGF with respect to secondary structure and conformational shifts are discussed.     

In silico characterization of thermostable lipases

Thermostable lipases are of high priority for industrial applications as they are endowed with the capability of carrying out diversified reactions at elevated temperatures. Extremophiles are their potential source. Sequence and structure annotation of thermostable lipases can elucidate evolution of lipases from their mesophilic counterparts with enhanced thermostability hence better industrial potential. Sequence analysis highlighted the conserved residues in bacterial and fungal thermostable lipases. Higher frequency of AXXXA motif and poly Ala residues in lid domain of thermostable Bacillus lipases were distinguishing characteristics. Comparison of amino acid composition among thermostable and mesostable lipases brought into light the role of neutral, charged and aromatic amino acid residues in enhancement of thermostability. Structural annotation of thermostable lipases with that of mesostable lipases revealed some striking features which are increment of gamma turns in thermostable lipases; being first time reported in our paper, longer beta strands, lesser beta-branched residues in helices, increase in charged-neutral hydrogen bonding pair, hydrophobic-hydrophobic contact and differences in the N-cap and C-cap residues of the α helices. Conclusively, it can be stated that subtle changes in the arrangement of amino acid residues in the tertiary structure of lipases contributes to enhanced thermostability.     

Extended form of a retro-inverso peptide stabilized by beta-sheet unidirectional H-bonds: Crystallographic and NMR evidence.

The crystallographic investigation of the retro-inverso peptide Bz-S-gAla-R-mAla-NHPh reveals an extended backbone conformation where the NH groups of the gem-diamino alkyl moiety and the CO groups of the malonyl residue face side by side. This extended conformation, presenting all carbonyls on opposite sides of the NH groups, is stabilized by interstrand H-bonds running in a single direction of the parallel beta-sheets that characterize the crystal packing. These sheets differ from the beta-sheets formed by native amino acids only. (1)H-NMR nuclear Overhauser effect spectroscopy (NOESY) experiments suggest that a conformation similar to that found in the crystal also prevails in dimethylsulfoxide solution. Previous potential energy calculations of gem-diamino alkyl (g) and malonyl (m) Ala residues predicted that extended forms were less stable than the helical ones because of strong electrostatic repulsions between the parallel polar groups. Similar arguments were invoked to give more weight to helical forms of the retro-peptide units in the proposal of packing models of some nylons in their crystalline polar regions. The present findings show that both g and m Ala residues can experience the extended conformation in the beta-sheet aggregation. The energy increase occurring in one strand, due to the parallel orientation of consecutive peptide dipoles, is more than compensated by favorable cooperative interactions among head-to-tail aligned peptide dipoles of facing strands, resulting in the formation of two C==O...H==N H-bonds per residue.     

Mechanism of the reaction catalyzed by mandelate racemase. 2. Crystal structure of mandelate racemase at 2.5-A resolution: identification of the active site and possible catalytic residues

The crystal structure of mandelate racemase (MR) has been solved at 3.0-A resolution by multiple isomorphous replacement and subsequently refined against X-ray diffraction data to 2.5-A resolution by use of both molecular dynamics refinement (XPLOR) and restrained least-squares refinement (PROLSQ). The current crystallographic R-factor for this structure is 18.3%. MR is composed of two major structural domains and a third, smaller, C-terminal domain. The N-terminal domain has an alpha + beta topology consisting of a three-stranded antiparallel beta-sheet followed by an antiparallel four alpha-helix bundle. The central domain is a singly wound parallel alpha/beta-barrel composed of eight central strands of beta-sheet and seven alpha-helices. The C-terminal domain consists of an irregular L-shaped loop with several short sections of antiparallel beta-sheet and two short alpha-helices. This C-terminal domain partially covers the junction between the major domains and occupies a region of the central domain that is filled by an eight alpha-helix in all other known parallel alpha/beta-barrels except for the barrel domain in muconate lactonizing enzyme (MLE) [Goldman, A., Ollis, D. L., & Steitz, T. A. (1987) J. Mol. Biol. 194, 143] whose overall polypeptide fold and amino acid sequence are strikingly similar to those of MR [Neidhart, D. J., Kenyon, G. L., Gerlt, J. A., & Petsko, G. A. (1990) Nature 347, 692]. In addition, the crystal structure reveals that, like MLE, MR is tightly packed as an octamer of identical subunits. The active site of MR is located between the two major domains, at the C-terminal ends of the beta-strands in the alpha/beta-barrel domain. The catalytically essential divalent metal ion is ligated by three side-chain carboxyl groups contributed by residues of the central beta-sheet. A model of a productive substrate complex of MR has been constructed on the basis of difference Fourier analysis at 3.5-A resolution of a complex between MR and (R,S)-p-iodomandelate, permitting identification of residues that may participate in substrate binding and catalysis. The ionizable groups of both Lys 166 and His 297 are positioned to interact with the chiral center of substrate, suggesting that both of these residues may function as acid/base catalysts.(ABSTRACT TRUNCATED AT 400 WORDS)     

Structure and mechanism of activity of the cyclic phosphodiesterase of Appr>p, a product of the tRNA splicing reaction

The crystal structure of the cyclic phosphodiesterase (CPDase) from Arabidopsis thaliana, an enzyme involved in the tRNA splicing pathway, was determined at 2.5 A resolution. CPDase hydrolyzes ADP-ribose 1",2"-cyclic phosphate (Appr>p), a product of the tRNA splicing reaction, to the monoester ADP-ribose 1"-phosphate (Appr-1"p). The 181 amino acid protein shows a novel, bilobal arrangement of two alphabeta modules. Each lobe consists of two alpha-helices on the outer side of the molecule, framing a three- or four-stranded antiparallel beta-sheet in the core of the protein. The active site is formed at the interface of the two beta-sheets in a water-filled cavity involving residues from two H-X-T/S-X motifs. This previously noticed motif participates in coordination of a sulfate ion. A solvent-exposed surface loop (residues 100-115) is very likely to play a flap-like role, opening and closing the active site. Based on the crystal structure and on recent mutagenesis studies of a homologous CPDase from Saccharomyces cerevisiae, we propose an enzymatic mechanism that employs the nucleophilic attack of a water molecule activated by one of the active site histidines.     

The three-dimensional structure of bovine platelet factor 4 at 3.0-A resolution

Platelet factor 4 (PF4), which is released by platelets during coagulation, binds very tightly to negatively charged oligosaccharides such as heparin. To date, six other proteins are known that are homologous in sequence with PF4 but have quite different functions. The structure of a tetramer of bovine PF4 complexed with one Ni(CN)4(2-) molecule has been determined at 3.0 A resolution and refined to an R factor of 0.28. The current model contains residues 24-85, no solvent, and one overall temperature factor. Residues 1-13, which carried an oligosaccharide chain, were removed with elastase to induce crystallization; residues 14-23 and presumably 86-88 are disordered in the electron density map. Because no heavy atom derivative was isomorphous with the native crystals, the complex of PF4 with one Ni(CN)4(2-) molecule was solved using a single, highly isomorphous Pt(CN)4(2-) derivative and the iterative, single isomorphous replacement method. The secondary structure of the PF4 subunit, from amino- to carboxyl-terminal end, consists of an extended loop, three strands of antiparallel beta-sheet arranged in a Greek key, and one alpha-helix. The tetramer contains two extended, six-stranded beta-sheets, each formed by two subunits, which are arranged back-to-back to form a "beta-bilayer" structure with two buried salt bridges sandwiched in the middle. The carboxyl-terminal alpha-helices, which contain lysine residues that are thought to be intimately involved in binding heparin, are arranged as antiparallel pairs on the surface of each extended beta-sheet.