Skip to main content
NCBI home page
Proceedings of the National Academy of Sciences of the United States of America logoLink to Proceedings of the National Academy of Sciences of the United States of America
. 1997 Sep 2;94(18):9620–9625. doi: 10.1073/pnas.94.18.9620

Crystal structures of the copper and nickel complexes of RNase A: Metal-induced interprotein interactions and identification of a novel copper binding motif

R Balakrishnan *, N Ramasubbu †,‡, K I Varughese §, R Parthasarathy *
PMCID: PMC23236  PMID: 9275172

Abstract

We report the crystal structures of the copper and nickel complexes of RNase A. The overall topology of these two complexes is similar to that of other RNase A structures. However, there are significant differences in the mode of binding of copper and nickel. There are two copper ions per molecule of the protein, but there is only one nickel ion per molecule of the protein. Significant changes occur in the interprotein interactions as a result of differences in the coordinating groups at the common binding site around His-105. Consequently, the copper- and nickel-ion-bound dimers of RNase A act as nucleation sites for generating different crystal lattices for the two complexes. A second copper ion is present at an active site residue His-119 for which all the ligands are from one molecule of the protein. At this second site, His-119 adopts an inactive conformation (B) induced by the copper. We have identified a novel copper binding motif involving the α-amino group and the N-terminal residues.

Metal ions are responsible for a wide range of functions, including nucleophilic catalysis in enzymes such as carbonic anhydrase and carboxypeptidase, electron transfer in proteins such as rubredoxin and the cytochromes, and stabilization of protein structure in many proteins from zinc finger to alcohol dehydrogenase (1, 2) and in gene regulation (3, 4). Cu2+ and Zn2+ are essential trace elements present in many species (5, 6), whereas Ni2+ enters the cell by chronic exposure (7) and is a known carcinogen (8). Copper functions as a cofactor in copper–zinc superoxide dismutase and plays an important role in the defense against oxygen-derived free radicals that have deleterious effects on biological macromolecules (4). Because of the many biological roles played by these metal ions, efforts have been recently directed to engineer de novo metal binding sites to impart a desired property to a protein. For example, metal binding sites have been engineered to aid protein purification (9), to enhance protein stability (10, 11) and regulation of catalytic activity (12, 13), and to introduce metal ions into novel protein scaffolds (14, 15). In this regard structural studies of metalloproteins have greatly aided in engineering of novel metal binding sites into proteins.

However, presence of these transition metals in abnormal levels due to metabolic disorders or due to chronic exposure has been linked to DNA damage and eventually to cell death. For example, for Wilson and Menkes diseases, there is an aberrant copper metabolic pathway leading to copper accumulation (16). In Alzheimer disease, decreased levels of a metallothionein-like protein has been observed leading to an increase in the levels of zinc (17, 18). Inability of metalloproteins to perform normal functions has been directly attributed to the absence of metal ions leading to a conformational change, as shown by crystallographic, spectroscopic, and biochemical studies (19–23). But so far there has been limited evidence to elucidate how metals, upon binding, can force normal proteins to undergo conformational changes and thereby make them dysfunctional. It is necessary to develop a model to understand such unusual metal–protein interactions. In this regard we have used ribonuclease (RNase A) as a model nonmetalloprotein to determine (i) whether metal ion binding to RNase A induces a conformational change (local or global) and (ii) how do these changes affect the structural and functional properties of the protein. In particular, because it has been speculated that metal ions may cause the aggregation of the otherwise soluble proteins such as β amyloid protein of Alzheimer disease and the prion protein of prion diseases, we would like to determine whether metal binding could lead to oligomerization or aggregation of the protein in our model.

Reports (24) based on immobilized metal ion affinity chromatography (IMAC) studies showed that RNase A binds to copper and nickel. Moreover, from a historic perspective, although RNase A complexed with nickel (form I) was first obtained by Kunitz (25), attempts to reproduce the form I at the Protein Structure Project led by D. Harker were unsuccessful (25). However, the structure solution of a metal-free monoclinic form of RNase A (designated form II and crystallized at the Protein Structure Project) was reported first (26). Although, 14 more crystal forms of RNase A were crystallized, including a complex with copper (27), no structure of RNase A with a bound metal ion has ever been reported. Results from our structural investigation show that the binding of copper and nickel to RNase A causes local conformational changes and leads to different packing. We have also identified a new structural motif for binding of copper in proteins.

EXPERIMENTAL PROCEDURES

Crystals of RNase–Cu and RNase–Ni complexes were grown as described (28, 29). For the nickel complex, RNase A (Sigma, Lot 5503) was dialyzed against water to remove any salt, concentrated by using PEG (Mr 4000), and then lyophilized. For RNase–Cu complex crystals, the enzyme used was type XII-A from Sigma without further purification. Samples were left undisturbed in darkness for 3 months after which crystals appeared as thin plates. Crystals of copper complex are monoclinic, whereas the crystals of nickel complex are orthorhombic (Table 1).

Table 1.

Crystallographic parameters for RNase–Cu and RNase–Ni complexes

Structure detail Copper Nickel
Crystallization conditions
 Salt (0.1 M) CuCl2 NiCl2
 pH 5.5 5.5
 Precipitant MPD MPD
Data collection
 Space group C2 P212121
 a, Å 58.18 44.69
 b, Å 53.80 75.80
 c, Å 42.90 37.74
 α, degrees 90.00 90.00
 β, degrees 118.90 90.00
 γ, degrees 90.00 90.00
Volume, Å3 29,800 31,800
 Matthew’s parameter 2.18 2.32
 No. of unique reflections 6,930 4,205
 Rsvmm, % 4.1 4.7
Refinement
 No. of reflections >2σ 6610 4180
 Resolution range, Å 8.0–2.0 10.0–2.6
 Final R factor, % 20 17
 No. of protein atoms 1,196 1,196
 No. of solvent molecules 82 87
 No. of metal ions 2 1
Deviations from ideality
 Distances, Å 0.007 0.007
 Bond angles, degrees 1.4 1.4
 Torsion, degrees 26 25
Protein B factors, Å2
 Main chain 17.0 10.6
 Side chain 19.9 13.3
Open in a new tab

MPD, 2-methyl-2,4-pentanediol. 

X-ray diffraction data on the copper crystals were collected by using the Mar Image plate (to a resolution of 2.0 Å), whereas data on the nickel complex were obtained by using the multiwire area detector (to 2.6-Å resolution). X-ray source used to generate Cu–Kα radiation was a rotating anode, operating at 100 mA and 50 kV. Diffraction data were recorded as a series of discrete frames, although the individual frames were contiguous in that the beginning of each small oscillation range coincided with the end of the previous range. Data on the nickel complex were collected at room temperature, and data on the copper complex were collected at 4°C. The determination of crystal orientation, integration of reflection intensities, and data reduction were performed by using denzo and scalepack (30), as shown in Table 1.

The structures of both complexes were solved by molecular replacement method using the structure of phosphate-free ribonuclease (7RSA) as the search model (31). The program amore (32) was used to solve both the structures. Refinement (positional and B factor) was carried out by using x-plor (33). A resolution-dependent weighting scheme defined in x-plor was employed for both complexes. Difference density omit maps (Fo − Fc and 2Fo − Fc) maps were generated by using x-plor and the residues were fitted to the omit maps by using tom, a derivative of frodo (34). A complete set of omit maps in which 10 residues were successively left out over the entire course of the polypeptide chain were also examined, and further adjustments made as necessary. Difference electron density map did not reveal, as expected, any additional density corresponding to phosphate molecule. However, copper and nickel ions were observable at the 7σ level. Search for water molecules was initiated only after the refinement of the protein atoms converged to an R factor of 21 and 19% for copper and nickel complexes, respectively. Water molecules were selected from difference density maps generated by using x-plor and included in the refinement only when the following criteria were met: (i) peak height in the Fo − Fc map was greater than the 2σ level, (ii) the selected peak made sensible contact with the protein molecule, and (iii) the temperature factors were less than 60 Å2. The final crystallographic R factor was 20% and 17% for the RNase–Cu and RNase–Ni structures, respectively. The programs procheck (35) and x-plor were used to assess the correctness of the model. Visualization of electron density, model building, and refinement were carried out on a SGI work station.

RESULTS

The final refined models of the complexes consists of 124 amino acid residues, 2 Cu2+ ions, and 82 waters in RNase–Cu and 1 Ni2+ ion, and 87 waters in RNase–Ni structures. The relevant refinement parameters are given in Table 1. Both structures had very well-defined and continuous electron density at the 1α level for all main-chain atoms in the Fo − Fc omit maps. The overall topology observed in 7RSA (phosphate free) (31) and 5RSA‖ (phosphate bound) (36) have been preserved in these complexes also. The rms standard deviations as estimated from Luzzati plots are comparable for the copper and nickel complexes. However, the two complexes exhibit significant differences at the sites of coordination as described below.

Coordination of Metal Ion.

The major difference between the copper and nickel complexes lies in the number of binding sites (Fig. 1) and the nature of the ligands around these ions. The bond distances in the copper complex vary from 2.0 to 2.6 Å, and the distances in the nickel complex vary from 1.8 to 2.4 Å. In metalloproteins, the bond distances between the metal ion and its ligands usually fall within a very narrow range (37) but may vary from an average distance of 2.1 Å due to distorted geometry. For example, in ascorbate oxidase–copper complex (38), the distances between the metal ion and its ligands (O, N) range from 1.9 to 2.2 Å. In the nickel complex of urease (39), the distance between the ligands (O, N) and the metal ions varies from 2.0 to 2.3 Å. Also, no geometry restraints were applied for the metal coordination during refinement.

Figure 1.

Figure 1

Open in a new tab

Stereo view of the coordination spheres at the metal binding sites in the complexes. (A) Copper at the His-105 site. (B) Nickel at the His-105 site. (C) Copper at the His-119 site.

The copper ion binds to two sites, whereas the nickel ion binds to only one site; both copper and nickel have a common site near His-105 at which the coordination geometry is a distorted square pyramid. However, the coordinating groups (ligands) provided by two symmetry related molecules are different at this common site. Thus, the α-amino group of Lys-1, amide nitrogen and carbonyl oxygen of Glu-2 from one molecule, NE2 of His-105, and carbonyl oxygen of Tyr-76 from a symmetry-related molecule (1/2 − x, y − 1/2, −z + 1) ligand to Cu2+ at site I. In nickel, however, three water molecules coordinate in place of the α-amino group, carbonyl oxygen of Tyr-76, and amide nitrogen of Glu-2. In addition, carboxylate oxygen of Lys-1 from a symmetry-related molecule (x, y, z + 1) rather than Glu-2 coordinates to nickel at this site.

The second site for copper is at His-119, which is an active site residue. The ligands at this site are NE2 of His-119 and OE2 of Glu-111 from the same molecule of RNase A and three water molecules. Interestingly, the copper ion induces a conformational change in which His-119 has flipped its side-chain orientation significantly compared with 7RSA and nickel complex (Fig. 2). In this orientation, His-119 is out of the active site pocket. A histogram depicting the solvent-accessible surface area for all the His residues in RNase A structures shows that it is larger for His-119 compared with the corresponding His residues in either RNase–Ni or 7RSA structures. In contrast, this area is comparable for His-105 in all three structures. This suggests that in the flipped conformation, His-119 is more accessible and is better oriented to bind to metal ions. In the flipped conformation, the value of the angle χ1 of His-119 (−58.4°) is comparable to that found for His-119 in an inactive B conformation as observed in an inhibitor complex of RNase A, but it differs from the active conformation A of His-119 by as much as 143° (40). Because of this, the pKa of His-12 might be affected to an extent that RNase–Cu may become inactive.

Figure 2.

Figure 2

Open in a new tab

Superposition of structures of 7RSA, the RNase–Cu complex, and the RNase–Ni complex. Only 7RSA is shown as ribbon diagram. Side chains of residues near the metal binding sites are displayed in all three structures (copper, green; nickel, yellow; 7RSA, white). Note that (i) the flipping of the imidazole ring of His-105 by 180° and the consequent change in positions of copper (green) and nickel (yellow) ions and (ii) the orientation of the His-119 (green) in the RNase–Cu complex.

Metal-Induced Packing.

The coordination of copper and nickel shows that the binding around His-105 involves four residues from two molecules of RNase A in copper complex, whereas only two residues from two molecules of RNase A are involved in the nickel complex. As a result, the packing interactions in RNase–Cu are significantly different from those in the nickel complex. RNase–Cu complex exhibits a back to back packing whereas the RNase–Ni complex exhibits a linear array of a head to tail packing. Each molecule in RNase–Cu complex interacts with 7 symmetry-related molecules, giving rise to a total of 194 contacts. These contacts are distributed among 74 residues of which 36 are from the reference molecule. In the RNase–Ni complex, a total of 186 contacts occur between the reference molecule (44 residues) and 6 symmetry-related molecules (44 residues). In the copper complex, there are 16 H bonds and one salt bridge compared with only 10 H bonds and no salt bridge in the nickel complex. The lone salt bridge in the RNase–Cu structure is between OD2 (Asp-121) and NH1 (Arg-85) with a distance of 2.84 Å. Similar packing analysis of another RNase A (nonmetal) structure, 6RSA** (41), revealed the presence of 6 H bonds. Although the average number of contacts per molecule in the 6RSA structure is only 12, this number is significantly higher in copper (24 contacts) and nickel (27 contacts) complexes, implying a tight packing in the RNase–Cu and RNase–Ni structures. Although packing in a molecular crystal is a compromise among many different intermolecular and intramolecular forces, it is decided by the strongest forces corresponding to the shortest of these contacts. Thus, between the two complexes, although the nickel complex has more contacts <4.0 Å, the copper complex has more of the shorter meaningful contacts such as H bonds and salt bridges. Because of these shorter interactions, copper complex is tightly packed. Additional support for this inference comes from the comparison of the Matthews parameter (42). This parameter is small for copper (2.18) compared with nickel (2.32). Majority of the interactions occur around the region where the two metal ions are present. At these sites, the number of contacts observed in the copper complex is higher than in the nickel complex. Apparently, the differences in the packing are a direct consequence of the differences in the binding of copper and nickel because the orientation of His-105 and the coordinating ligands in the copper and nickel complexes are different. These differently dimerized RNase A molecules could act as nucleating centers for packing and may bring about the differences in the overall symmetry of the crystal.

DISCUSSION

Copper Binding Motif.

The comparison of the coordination of nickel and copper at site I reveals the inherent ability of copper to induce local conformational and electronic changes to a greater extent than that of nickel (Figs. 2 and 3). Clearly, the terminal Lys residue undergoes a significant conformational change to bind to copper ion. Because the crystals of copper complex were obtained from a solution whose pH was maintained at 5.5, the α-amino group is expected to be protonated. For this group to be able to coordinate to the copper ion, there must have been a major change in the local electronic environment to reduce the pKa from about 7.8. Crystal structures studies of a tripeptide, Gly-Gly-Gly, complexed with copper also revealed the involvement of the α-amino group in the formation of the complex (43). Similarly, in serum albumin, which is transporter protein for copper, the α-amino group and His residue are involved in the formation of a coordination complex with copper (44). However, the tripeptide crystallization and the serum albumin complex formation were carried out at an alkaline pH and, hence, one would expect the α-amino group to be neutral. Interestingly, a situation similar to RNase–Cu occurs in the crystal structure of thioredoxin (45), wherein copper binds at the N-terminal end using the α-amino group as a ligand. These authors report the pH of the mother liquor to be 3.8 but do not provide an explanation for the unusual binding of the α-amino group at this low pH. Clearly, Cu2+ induces local conformational changes at the N-terminal end to bind to the available ligands. We propose that this feature may be linked to a new structural motif for copper binding in proteins, and the motif is shown in Fig. 4. In this motif, the α-amino group is a constant electron donor to copper, although among O1, N2, and O2, the choice depends on the relative position and stereochemistry of the metal ion at the N terminus.

Figure 3.

Figure 3

Open in a new tab

Superposition of the N terminal residues (residues 1–5) of 7RSA (ribbon), RNase–Cu (thick line), and RNase–Ni (broken line). Note the approach of His-105 to copper and nickel ions and the relative disposition of other ligands. Lys-1 is at the bottom.

Figure 4.

Figure 4

Open in a new tab

Superposition of the atoms N1 and the copper metal ions in the copper binding motif involving the α-amino terminus and ligands from the first two residues in Gly-Gly-Gly (broken line), thioredoxin (light line), and RNase–Cu structures (thick line).

The coordination by the backbone electron donor atoms (N1, O1, N2, and O2) of the N-terminal end residues Lys and Glu to Cu2+ poses the question of the mechanism of such a binding. Interestingly, in a separate study, a step-by-step coordination induced by copper in canine serum albumin was proposed by using CD and NMR spectroscopy (45). Copper may be able to induce the RNase A groups into coordination in a similar step-by-step manner because, under similar circumstances, nickel generates a different complex. Thus, the complex formation by copper may be thermodynamically controlled for which we propose the following mechanism. (i) His-105 coordinates to copper. (ii) Oligomerization of RNase A molecules occurs as a result of binding to one of the N-terminal residues. (iii) A search for protein ligands for complex formation involving additional N-terminal residues may occur. In this process, the small degree of N-terminal flexibility observed in ribonuclease structures might also play a role. Nickel complex, on the other hand, may be more kinetically controlled. Thus, after the initial binding of His-105, coordination with water molecules is sufficient for the stabilization of the nickel complex. Comparison of the backbone torsion of the copper and nickel structures with the structure of phosphate-free RNase structure (7RSA) shows a partial unwinding of the N-terminal helix in these metal complexes (Fig. 3). A similar mechanism could explain the binding of copper in thioredoxin, wherein, at pH 3.8, which is closer to the pKa of Asp, copper could first bind to Asp (Asp-2 OD1 or Asp-10 OD2) and, upon oligomerization in the crystal lattice, could then bind to the α-amino group (Ser-1), amide nitrogen of Asp-2 N2, and water.

Although the carbonyl oxygen of Tyr-76 is a strong ligand to copper, it is conspicuously absent in the nickel structure. This arises in part due to a significant change in the value of χ2 angle for His-105 in the copper (86°) and nickel (−67°). Because of this conformational change (i.e., flipping of the imidazole ring by about 180°), the NE2 of His-105 is located near Tyr-76 (Fig. 2). The approach of the metal-bound His-105 of a symmetry-related molecule toward the N terminus in both complexes is different. Binding of copper to His-105 and hence the N terminus has reoriented the side chain of Lys-1. It is clear from our investigation that the formation and/or stability need not depend on the mere availability of ligands. The hybridization state of the metal ions may also play a role in the formation and stability of metal complexes as suggested by Irving and Williams (46).

Correlation with IMAC.

The crystal structures of RNase–Cu and RNase–Ni complexes provide a basis to interpret the results obtained from IMAC on RNase A (24). It has been reported that RNase A exhibits different binding strengths for immobilized copper and nickel. RNase A was retained on IDA-Cu(II) and IDA-Ni(II) columns at pH 7.0 and was eluted at pH 4.3 and 6.0, respectively. In IMAC the protein is retained on an IDA-M(II) column through its surface-exposed residues, such as His and Trp, and is released from the column by protonating the imidazole ring of the His residues (lowering the pH). The fact that RNase A could be displaced from IDA-Ni(II) at a much higher pH (6.0) than from IDA-Cu(II) implies that the protein–nickel complex is weak and hence could be displaced by a small decrease in pH from 7.0 to 6.0. From the present study, we attribute the stronger binding of RNase A to immobilized copper to the contribution from the two His residues (His-105 and -119) and a weaker one in nickel to the contribution from His-105 alone. In addition, the nickel ion is solvated at His-105 with three water molecules but copper has protein ligands contributing to the coordination sphere. Further, elution of RNase A from an IDA-Cu(II) column at pH 4.3 shows that the His pKa has decreased significantly to remain neutral and be bound to the metal ion.

Biological Implications of the Metal Ion Binding to RNase A.

It has been reported (12) that copper-bound recombinant trypsin becomes inactive when Arg-96 is replaced with His near the active-site His-57. Copper ion prevents the participation of the active-site His in the general base catalysis. X-ray crystal studies (47) of the copper complex of recombinant trypsin showed that the side chain of the catalytically essential His-57 is reoriented out of the active-site pocket to bind to copper. This relates very well to the RNase–Cu structure where a catalytically essential residue, His-119, has reoriented and formed a complex with copper ion (Fig. 2). However, unlike trypsin, where only one active-site residue has become unavailable for catalysis, in RNase–Cu structure, copper binding to His-119 could prevent both His-119 and His-12 from participating in catalysis. It has been previously documented (40) that in the presence of an inhibitor, the His-119 imidazole ring flips to an inactive B conformation. A similar effect has been brought upon by copper by flipping the His-119 side chain to an angle closer to that found in the inhibitor complex of RNase A. This strongly suggests that copper could behave like an inhibitor.

Cells contain an unexpectedly large number of distinct RNases that carry out the many specific reactions of RNA metabolism. The mechanisms by which these reactions are regulated by RNases is still not clear. These specific reactions are of the following two classes: (i) conversion of a RNA molecule to another, including separation of individual RNA species, maturation of 5′ and 3′ termini of individual RNA, and the removal of introns; and (ii) degradation of unwanted RNA or to generate nucleotides for subsequent RNA synthesis. Unwanted RNA could interfere with cell metabolism by competing with other functioning RNAs (48). In this regard, RNase–Cu structure provides some clue about a possible regulatory role of RNase because copper bound to an active-site residue (His-119) may render the enzyme inactive by a mechanism observed in trypsin. However, from these structural studies, how and why a particular RNase is selected for inactivation cannot be answered clearly without additional evidence on the regulation of RNase by metal ions. Another important observation in this investigation is the involvement of both copper and nickel in the oligomerization of RNase A in the crystal lattice. In a situation where the cells have high levels of trace metals, the metal ions can be sequestered by the protein via oligomerization, which can prevent metal toxicity and hence cell death.

Summary.

We have shown from the crystal structure of RNase–Cu and its comparison with other copper complexes of peptides and proteins that copper prefers the α-amino group as one of its ligands. This binding motif, which is unique to copper, involves a dimer of proteins (peptides). We also show that differences exist in the binding of copper and nickel. Finally, we suggest that the inherent ability of copper to induce local conformational changes at its binding sites may arise because of thermodynamic control in contrast to the binding of nickel.

Acknowledgments

This article is dedicated to the late Dr. D. Harker (former Director of Biophysics Department, Roswell Park Cancer Institute, Buffalo) who pioneered the structural investigation of RNase A, the first protein crystallography project in this country (25). We thank Dr. M. J. Levine for support, Dr. N. H. Xuong at University of California, San Diego, for access to the area detectors, and Ms. J. Mann for technical assistance. R.B. thanks the New York State Department of Health for a fellowship and Drs. S. T. Rao and E. Sulkowski for many useful discussions. This work was supported in part by U.S. Public Health Service Grants DE08240 (Research Center in Oral Biology) and HL55375 (K.I.V.).

ABBREVIATION

IMAC

immobilized metal ion affinity chromatography

Footnotes

Data deposition: The atomic coordinates and structure factors have been deposited in the Protein Data Bank, Chemistry Department, Brookhaven National Laboratory, Upton, NY 11973 (reference 1aqp, γ1aqprsf). This information is embargoed for 6 months (coordinates) and 6 months (structure factors) from the date of publication.

Space group, P21/a = 30.18 Å; b = 38.40 Å; c = 53.32 Å; α = 90.00°; β = 105.85°, γ = 90.00°.

‖

Space group, P21/a = 30.18 Å; b = 38.40 Å; c = 53.32 Å; α = 90.00°; β = 105.85°; γ = 90.00°.

**

Space group, P21/a = 30.3 Å; b = 38.35 Å; c = 53.70 Å; α = 90.00°; β = 106.4°; γ = 90.00°.

References

  • 1.Glusker J P. Adv Protein Chem. 1991;42:1–77. doi: 10.1016/s0065-3233(08)60534-3. [DOI] [PubMed] [Google Scholar]
  • 2.Regan L. Trends Biochem Sci. 1995;20:280–285. doi: 10.1016/s0968-0004(00)89044-1. [DOI] [PubMed] [Google Scholar]
  • 3.O’Halloran T V. Science. 1993;261:715–725. doi: 10.1126/science.8342038. [DOI] [PubMed] [Google Scholar]
  • 4.Zhou P, Thiele D J. Biofactors. 1993;4:105–115. [PubMed] [Google Scholar]
  • 5.Ettinger M J. In: Copper Proteins and Copper Enzymes. Lontie R, editor. Vol. 3. Boca Raton, FL: CRC; 1984. pp. 175–229. [Google Scholar]
  • 6.Bremner I, Beattie J H. Annu Rev Nutr. 1990;10:63–83. doi: 10.1146/annurev.nu.10.070190.000431. [DOI] [PubMed] [Google Scholar]
  • 7.Sunderman F W., Jr Food Cosmet Toxicol. 1971;9:105–120. doi: 10.1016/s0015-6264(71)80120-7. [DOI] [PubMed] [Google Scholar]
  • 8.Hartwig A, Kruger I, Beyersmann D. Toxicol Lett. 1994;72:353–358. doi: 10.1016/0378-4274(94)90048-5. [DOI] [PubMed] [Google Scholar]
  • 9.Arnold F H, Haymore B L. Science. 1991;252:1796–1797. doi: 10.1126/science.1648261. [DOI] [PubMed] [Google Scholar]
  • 10.Braxton S, Wells J A. Biochemistry. 1992;31:7796–7801. doi: 10.1021/bi00149a008. [DOI] [PubMed] [Google Scholar]
  • 11.Muheim A, Todd R J, Casimiro D R, Gray H B, Arnold F H. J Am Chem Soc. 1993;115:5312–5313. [Google Scholar]
  • 12.Higaki J N, Haymore B L, Chen S, Fletterick R J, Craik C S. Biochemistry. 1990;29:8582–8586. doi: 10.1021/bi00489a012. [DOI] [PubMed] [Google Scholar]
  • 13.Brinen L S, Scott Willet W, Craik C S, Fletterick R J. Biochemistry. 1996;35:5999–6009. doi: 10.1021/bi9530200. [DOI] [PubMed] [Google Scholar]
  • 14.Handel T M, Williams S A, DeGrado W F. Science. 1993;261:879–885. doi: 10.1126/science.8346440. [DOI] [PubMed] [Google Scholar]
  • 15.Pessi A, Binachi E, Crameri A, Venturini S, Tramontano A, Solazzo M. Nature (London) 1993;362:367–369. doi: 10.1038/362367a0. [DOI] [PubMed] [Google Scholar]
  • 16.Prohaska J R. Clin Physiol Biochem. 1986;4:87–93. [PubMed] [Google Scholar]
  • 17.Bush A I, Pettingell W H, Jr, Marc d. Paradis, Tanzi R E. J Biol Chem. 1994;269:12152–12158. [PubMed] [Google Scholar]
  • 18.Bush A I, Pettingell W H, Jr, Multhaup G, Marc d. Paradis, Vonsattel J-P, Gusella J F, Beyreuther K, Masters C L, Tanzi R E. Science. 1994;265:1464–1467. doi: 10.1126/science.8073293. [DOI] [PubMed] [Google Scholar]
  • 19.Grossmann J G, Neu M, Evans R W, Lindley P F, Appel H, Hasnain S S. J Mol Biol. 1993;229:585–590. doi: 10.1006/jmbi.1993.1063. [DOI] [PubMed] [Google Scholar]
  • 20.Bellsolell L, Prieto J, Serrano L, Coll M. J Mol Biol. 1994;238:489–495. doi: 10.1006/jmbi.1994.1308. [DOI] [PubMed] [Google Scholar]
  • 21.Coffer A I, Knowles P P. Biochim Biophys Acta. 1994;1209:279–285. doi: 10.1016/0167-4838(94)90197-x. [DOI] [PubMed] [Google Scholar]
  • 22.Gomis-Ruth F-X, Grams F, Yiallouros I, Nar H, Kusthardt U, Zwilling R, Bode W, Stocker W. J Biol Chem. 1994;269:17111–17117. [PubMed] [Google Scholar]
  • 23.Zhang F L, Fu H-W, Casey P J, Bishop W R. Biochemistry. 1996;35:8166–8171. doi: 10.1021/bi960574+. [DOI] [PubMed] [Google Scholar]
  • 24.Sulkowski E. Trends Biotech. 1985;3:1–7. [Google Scholar]
  • 25.Tulinsky A. Annu Rep Med Chem. 1996;31:357–366. [Google Scholar]
  • 26.Kartha G, Bello J, Harker D. Nature (London) 1967;213:281–291. doi: 10.1038/213862a0. [DOI] [PubMed] [Google Scholar]
  • 27.King M V, Bello J, Pignataro E H, Harker D. Acta Crystallogr. 1962;15:144–147. [Google Scholar]
  • 28.King M V, Magdoff B S, Adelman M B, Harker D. Acta Crystallogr. 1956;9:460–465. [Google Scholar]
  • 29.King M V. Biochim Biophys Acta. 1964;79:388–392. [PubMed] [Google Scholar]
  • 30.Otwinowski Z. In: Proceedings of the CCP4 Study Weekend. Sawyer L, Issacs N, Bailey S, editors. Daresbury Laboratory, Daresbury, U.K.: Science and Engineering Research Council; 1993. pp. 56–62. [Google Scholar]
  • 31.Wlodawer A, Svensson L A, Sjolin L, Gilliland G L. Biochemistry. 1988;27:2705–2717. doi: 10.1021/bi00408a010. [DOI] [PubMed] [Google Scholar]
  • 32.Navaza J. Acta Crystallogr A. 1994;50:157–163. [Google Scholar]
  • 33.Brunger, A. T. (1988) x-plor Manual (Yale Univ. Press, New Haven, CT).
  • 34.Jones T A. Methods Enzymol. 1985;115:157–171. doi: 10.1016/0076-6879(85)15014-7. [DOI] [PubMed] [Google Scholar]
  • 35.Laskowski R A, MacArthur M W, Moss D S, Thornton J M. J Appl Crystallogr. 1993;26:283–291. [Google Scholar]
  • 36.Wlodawer A, Sjolin L. Biochemistry. 1983;22:2720–2728. doi: 10.1021/bi00280a021. [DOI] [PubMed] [Google Scholar]
  • 37.Chakrabarti P. Protein Eng. 1990;4:49–56. doi: 10.1093/protein/4.1.49. [DOI] [PubMed] [Google Scholar]
  • 38.Messerschmidt A, Rossi A, Ladenstein R, Huber R, Bolognesi M, Gatti G, marchesini A, Petruzzelli R, Finazzi-Agro A. J Mol Biol. 1989;206:513–529. doi: 10.1016/0022-2836(89)90498-1. [DOI] [PubMed] [Google Scholar]
  • 39.Jabri E, Carr M B, Hausinger R P, Karplus P A. Science. 1995;268:998–1004. [PubMed] [Google Scholar]
  • 40.Zegers I, Maes D, Dao-Thi M-H, Poortmans F, Palmer R, Wyns L. Protein Sci. 1994;3:2322–2339. doi: 10.1002/pro.5560031217. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Borah B, Chen C-W, Egan W, Miller M, Wlodawer A, Cohen J S. Biochemistry. 1985;24:2058–2067. doi: 10.1021/bi00329a038. [DOI] [PubMed] [Google Scholar]
  • 42.Matthews B W. J Mol Biol. 1968;33:491–497. doi: 10.1016/0022-2836(68)90205-2. [DOI] [PubMed] [Google Scholar]
  • 43.Freeman H C, Robinson G, Schoone J C. Acta Crystallogr. 1964;17:719–730. [Google Scholar]
  • 44.Laussac J-P, Sarkar B. Biochemistry. 1984;23:2832–2838. doi: 10.1021/bi00307a046. [DOI] [PubMed] [Google Scholar]
  • 45.Katti S K, LeMater D M, Eklund H. J Mol Biol. 1990;212:167–184. doi: 10.1016/0022-2836(90)90313-B. [DOI] [PubMed] [Google Scholar]
  • 46.Irving H, Williams R J P. Nature (London) 1948;162:746–747. [Google Scholar]
  • 47.McGrath M E, Haymore B L, Summers N L, Craik C S, Fletterick R J. Biochemistry. 1993;32:1914–1919. doi: 10.1021/bi00059a005. [DOI] [PubMed] [Google Scholar]
  • 48.Deutscher M P. Trends Biochem. 1988;13:136–139. doi: 10.1016/0968-0004(88)90070-9. [DOI] [PubMed] [Google Scholar]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences

RESOURCES