Which strategy for a protein crystallization project?

The three-dimensional, atomic-resolution protein structures produced by X-ray crystallography over the past 50+ years have led to tremendous chemical understanding of fundamental biochemical processes. The pace of discovery in protein crystallography has increased greatly with advances in molecular biology, crystallization techniques, cryocrystallography, area detectors, synchrotrons and computing. While the methods used to produce single, well-ordered crystals have also evolved over the years in response to increased understanding and advancing technology, crystallization strategies continue to be rooted in trial-and-error approaches. This review summarizes the current approaches in protein crystallization and surveys the first results to emerge from the structural genomics efforts.

Molecular tuning of an EF-hand-like calcium binding loop. Contributions of the coordinating side chain at loop position 3.

Calcium binding and signaling orchestrate a wide variety of essential cellular functions, many of which employ the EF-hand Ca2+ binding motif. The ion binding parameters of this motif are controlled, in part, by the structure of its Ca2+ binding loop, termed the EF-loop. The EF-loops of different proteins are carefully specialized, or fine-tuned, to yield optimized Ca2+ binding parameters for their unique cellular roles. The present study uses a structurally homologous Ca2+ binding loop, that of the Escherichia coli galactose binding protein, as a model for the EF-loop in studies examining the contribution of the third loop position to intramolecular tuning. 10 different side chains are compared at the third position of the model EF-loop with respect to their effects on protein stability, sugar binding, and metal binding equilibria and kinetics. Substitution of an acidic Asp side chain for the native Asn is found to generate a 6,000-fold increase in the ion selectivity for trivalent over divalent cations, providing strong support for the electrostatic repulsion model of divalent cation charge selectivity. Replacement of Asn by neutral side chains differing in size and shape each alter the ionic size selectivity in a similar manner, supporting a model in which large-ion size selectivity is controlled by complex interactions between multiple side chains rather than by the dimensions of a single coordinating side chain. Finally, the pattern of perturbations generated by side chain substitutions helps to explain the prevalence of Asn and Asp at the third position of natural EF-loops and provides further evidence supporting the unique kinetic tuning role of the gateway side chain at the ninth EF-loop position.

Structures of a hemoglobin-based blood substitute: insights into the function of allosteric proteins.

BACKGROUND: . Potential blood substitutes can be based on hemoglobin. Two problems must be overcome with acellular hemoglobin-based blood substitutes, however: the oxygen affinity of purified human hemoglobin is too high for it to deliver oxygen to tissues, and hemoglobin tetramers dissociate into alphabeta dimers that can cause kidney damage. A modified form of hemoglobin, rHb 1.1, has reduced oxygen affinity as the result of an Asnbeta 108-->Lys mutation, and dimerization is prevented by the insertion of a glycine residue between the sequences of the normal alpha chains to produce one covalently continuous di-alpha-chain. Determination of the structure of rHb 1.1 would provide structure-based explanations for the altered properties of rHb 1.1. RESULTS: . We determined the structures of the deoxy form of rHb 1.1 at 2.0 resolution and of cyanomet-rHb 1.1 at 2.6 resolution. Deoxy-rHb 1.1 adopts the classic 'T state' quaternary structure, but cyanomet-rHb 1.1 adopts a novel quanternary structure, the B state. The most striking feature of the tertiary structures is a charged hydrogen bond involving Lysbeta 108 that is broken in the T-->B state transition. The glycine bridge within the di-alpha-chain is well defined in both structures and appears to cause adoption of the B state instead of the previously observed ligand-bound quaternary structures R or Y/R2. CONCLUSIONS: . A charged hydrogen bond between Lysbeta 108 and Tyrbeta35 is broken in the transition between the deoxy and ligand-bound forms of rHb 1.1. This structural change reduces the oxygen affinity of rHb 1.1 by changing the relative stability of deoxy and ligand-bound states. Furthermore, our observations highlight the importance of small conformational changes in allosteric proteins, even in their most rigid domains. Three ligand-bound quaternary structures of hemoglobin (R, Y/R2 and B) have now been described. In contrast, only one quaternary structure has been observed for deoxyhemoglobin (T). The structural degeneracy of the high oxygen affinity form of hemoglobin is an important reminder that allosteric proteins may have multiple quaternary structures that are functionally very similar. This degeneracy of quaternary structures has important implications for the regulation of allosteric proteins, because different quaternary structures may be stabilized by different allosteric effectors.

Preliminary X-ray diffraction studies of an RNA pseudoknot that inhibits HIV-1 reverse transcriptase.

Crystals of a 26-nucleotide pseudoknot RNA, PK26, have been grown. The RNA was produced using phosphoramidite chemistry and was purified by denaturing polyacrylamide electrophoresis. The crystallization was robust with respect to changes in the number of nucleotides and to the salt used as precipitant. The crystals belong to space group P4(1)22 or P4(3)22 with unit-cell dimensions a = b = 61.6, c = 98.9 A. The best crystals diffract X-rays to 2.9 A. Three different sequences incorporating a single 5-bromo-deoxyuridine or 5-bromo-uridine nucleotide were also crystallized. Two of these derivatives are being used to determine the structure by multiple isomorphous replacement.

Crystal structure of a group I ribozyme domain: principles of RNA packing.

Group I self-splicing introns catalyze their own excision from precursor RNAs by way of a two-step transesterification reaction. The catalytic core of these ribozymes is formed by two structural domains. The 2.8-angstrom crystal structure of one of these, the P4-P6 domain of the Tetrahymena thermophila intron, is described. In the 160-nucleotide domain, a sharp bend allows stacked helices of the conserved core to pack alongside helices of an adjacent region. Two specific long-range interactions clamp the two halves of the domain together: a two-Mg2+-coordinated adenosine-rich corkscrew plugs into the minor groove of a helix, and a GAAA hairpin loop binds to a conserved 11-nucleotide internal loop. Metal- and ribose-mediated backbone contacts further stabilize the close side-by-side helical packing. The structure indicates the extent of RNA packing required for the function of large ribozymes, the spliceosome, and the ribosome.

RNA tertiary structure mediation by adenosine platforms.

The crystal structure of a group I intron domain reveals an unexpected motif that mediates both intra- and intermolecular interactions. At three separate locations in the 160-nucleotide domain, adjacent adenosines in the sequence lie side-by-side and form a pseudo-base pair within a helix. This adenosine platform opens the minor groove for base stacking or base pairing with nucleotides from a noncontiguous RNA strand. The platform motif has a distinctive chemical modification signature that may enable its detection in other structured RNAs. The ability of this motif to facilitate higher order folding provides one explanation for the abundance of adenosine residues in internal loops of many RNAs.

The structure of an RNA dodecamer shows how tandem U-U base pairs increase the range of stable RNA structures and the diversity of recognition sites.

BACKGROUND: Non-canonical base pairs are fundamental building blocks of RNA structures. They can adopt geometries quite different from those of canonical base pairs and are common in RNA molecules that do not transfer sequence information. Tandem U-U base pairs occur frequently, and can stabilize duplex formation despite the fact that a single U-U base pair is destabilizing. RESULTS: We determined the crystal structure of the RNA dodecamer GGCGCUUGCGUC at 2.4 A resolution. The molecule forms a duplex containing tandem U-U base pairs, which introduce an overall bend of 11-12 degrees in the duplex resulting from conformational changes at each interface between the tandem U-U base pairs and a flanking duplex sequence. The formation of the U-U base pairs cause small changes in several backbone torsion angles; base stacking is preserved and two hydrogen bonds are formed per base pair, explaining the stability of the structure. CONCLUSIONS: Tandem U-U base pairs can produce stable structures not accessible to normal A-form RNA, which may allow the formation of specific interfaces for RNA-RNA or RNA-protein recognition. These base-pairs show an unusual pattern of hydrogen-bond donors and acceptors in the major and minor grooves, which could also act as a recognition site.

Data reduction from twinned RNA crystals.

Methods were developed to process diffraction data from epitaxically twinned crystals. Four programs for data reduction and two display programs were developed to augment the data-reduction program XDS [Kabsch (1988). J. Appl. Cryst. 21, 916-924]. The programs can be generalized for use with other data-reduction software that provides the user with a list of the reflections used to determine lattice constants and crystal orientation. LATTICE_VIEW generates a PDB file containing 'water molecules' at the reciprocal-space coordinates of the strong spots found in the initial data frames. The PDB file is visualized to identify spots that belong to the same lattice, obtain unit-cell dimensions for a lattice, and assess data quality. VECTOR_MATCH is used to find additional spots belonging to a lattice. ACCOUNT4 determines which spots have been processed by XDS. COMFORT discards reflections that are too close to a reflection in another lattice. The display programs provide useful visual information on the quality of the crystal orientations used. Data with an R(merge) of 7.1% at 2.4 A resolution were obtained from epitaxically twinned crystals of an RNA dodecamer. The data were of sufficient quality to solve the structure with a combination of molecular replacement and single isomorphous replacement methods.

Crystallization and structure determination of RNA.

Progress in the synthesis, purification and crystallization of RNA has resulted in the determination of several X-ray crystal structures of RNA molecules over the past few years. Methods proven and under development will lead to future structure determinations and shed light on the structural basis for RNA's many functions.

Crystallization of ribozymes and small RNA motifs by a sparse matrix approach.

The three-dimensional structures of RNA enzymes form catalytic centers that include specific substrate binding sites. High-resolution determination of these and other RNA structures is essential for a detailed understanding of the function of RNA in biological systems. The crystal structures of only a few RNA molecules are currently known. These include tRNAs, which were produced in vivo and contained modified bases, and short oligonucleotide duplexes lacking tertiary interactions. Here we report that a number of different RNA molecules of 4-50 kDa, all synthesized in vitro, have been crystallized. A highly successful method for the growth of RNA crystals based on previously reported conditions for tRNA crystallization is presented. This method is rapid and economical, typically requiring 1.1 mg of RNA to set up an experiment and 2 weeks to complete the observations. Using this technique, we have obtained crystals of 8 of 10 different RNA molecules tested, ranging in size from a dodecamer duplex to a 208-nucleotide catalytic intron. Several of these crystal forms diffract to high resolution; in one case, we have collected a 2.8-A native data set for a 160-nucleotide domain of the group I self-splicing intron from Tetrahymena thermophila. The solution of these RNA structures should reveal aspects of tertiary structure that relate to RNA function and catalytic mechanisms.

Influence of a mutation in the putative nucleotide binding site of the nitrogen regulatory protein NTRC on its positive control function.

A mutation, serine 170 to alanine, in the proposed ATP binding site of the activator protein NTRC prevents transcriptional activation at sigma 54-dependent promoters both in vivo and in vitro. The rate of phosphorylation of the mutant protein by NTRB and the stability of mutant NTRC-phosphate were similar to those of wild-type NTRC. The phosphorylated mutant protein shows only a slight decrease in affinity (around 2-fold) for tandem NTRC binding sites in the Klebsiella pneumoniae nifL promoter suggesting that the mutation primarily influences the positive control function of NTRC. Moreover the mutant protein is trans dominant to the wild-type protein with respect to transcriptional activation at both the glnAp2 and nifL promoters. In vitro footprinting experiments reveal that the mutant protein is unable to catalyse isomerisation of closed promoter complexes between sigma 54-RNA polymerase and the nifL promoter to open promoter complexes. However, the mutant protein retains the ability to increase the occupancy of the -24, -12 region by sigma 54-RNA polymerase, forming closed complexes at the nifL promoter, which are not detectable in the absence of NTRC. These data support a model in which the activator influences the formation of closed complexes at the nifL promoter in addition to its role in catalysing open complex formation.

Designing an allosterically locked phosphofructokinase.

Six site-directed mutants of Escherichia coli phosphofructokinase (PFK) were made in an attempt to produce an enzyme "locked" in the inactive or "T"-state. The kinetic properties of the mutants were examined as a function of the substrates fructose 6-phosphate (Fru6P) and ATP, the positive effector GDP, and the negative effector phosphoenolpyruvate (PEP). All mutants exhibited lower activity than wild-type PFK. Three mutants (RS63, LV153, and VT246) had apparent dissociation constants for substrates and effectors similar to those of wild type. One mutant, HN160, had a 10-fold reduced affinity for Fru6P and reduced apparent affinity for the effectors. Two mutants, SN159 and T(GS)156, exhibited hyperbolic kinetics consistent with a "locked" T-state protein. Surprisingly, T(GS)156 showed hyperbolic activation in response to the physiological inhibitor PEP. The mutant PFK properties are discussed in terms of the PFK structure. These results suggest that the kinetic properties of PFK are sensitive to interactions in the homotropic interface; residues 156-160 in particular are critical in mediating the interactions between effector and active sites and in the T to R quaternary transition.

Effect of hydrostatic pressure on the solvent in crystals of hen egg-white lysozyme.

The mass density of protein crystals can be measured in Ficoll gradients as a function of hydrostatic pressure. Carbon tetrachloride-toluene mixtures provide convenient density markers, and the compressibility of these standards is reported. Measurements on tetragonal crystals of hen egg-white lysozyme yielded densities at room temperature of 1.2367(+/- 0.0010) g cm-3 at 1 atm and 1.2586(+/- 0.0017) g cm-3 at 1000 atm (1 atm = 101,325 Pa). When combined with the unit cell dimensions at these two pressures these values lead to an estimated compression (fractional change in volume) of the crystal solvent at 1000 atm of 0.0369(+/- 0.0054). This value is comparable to that of a 0.7 M solution of NaCl. From an approximate estimate of the Donnan effect for the crystal in the 1.4 M-NaCl mother liquor, the crystal solvent contains 0.8 M-Na+ and 2.5 M-Cl-. It is concluded that the compressibility of solvent in lysozyme crystals is, within experimental error, the same as bulk solvent and does not exhibit the dramatically altered compressibility expected of an ice or glass-like solid. The crystallographically observable water sites, 151 at 1 atm and 163 at 1000 atm, showed a tendency to increase the number of hydrogen bonds made to other water sites at the expense of hydrogen bonds made to protein. The explanation for this phenomenon is presently unknown. Water sites that occur in both structures tend to have comparable temperature factors and show some tendency to follow the pressure-induced changes in protein atom positions. The compression expected for the water molecules themselves is too small to be observable at the resolution of the X-ray data collected in this study.

Identification of structural motifs from protein coordinate data: secondary structure and first-level supersecondary structure.

A computer program is described that produces a description of the secondary structure and supersecondary structure of a polypeptide chain using the list of alpha carbon coordinates as input. Restricting the term "secondary structure" to the conformation of contiguous segments of the chain, the program determines the initial and final residues in helices, extended strands, sharp turns, and omega loops. This is accomplished through the use of difference distance matrices. The distances in idealized models of the segments are compared with the actual structure, and the differences are evaluated for agreement within preset limits. The program assigns 90-95% of the residues in most proteins to at least one type of secondary element. In a second step the now-defined helices and strands are idealized as straight line segments, and the axial directions and locations are compiled from the input C alpha coordinate list. These data are used to check for moderate curvature in strands and helices, and the secondary structure list is corrected where necessary. The geometric relations between these line segments are then calculated and output as the first level of supersecondary structure. A maximum of six parameters are required for a complete description of the relations between each pair. Frequently a less complete description will suffice, for example just the interaxial separation and angle. Both the secondary structure and one aspect of the supersecondary structure can be displayed in a character matrix analogous to the distance matrix format. This allows a quite accurate two-dimensional display of the three-dimensional structure, and several examples are presented. A procedure for searching for arbitrary substructures in proteins using distance matrices is also described. A search for the DNA binding helix-turn-helix motif in the Protein Data Bank serves as an example. A further abstraction of the above data can be made in the form of a metamatrix where each diagonal element represents an entire secondary segment rather than a single atom, and the off-diagonal elements contain all the parameters describing their interrelations. Such matrices can be used in a straightforward search for higher levels of supersecondary structure or used in toto as a representation of the entire tertiary structure of the polypeptide chain.

Crystal structure of hen egg-white lysozyme at a hydrostatic pressure of 1000 atmospheres.

The crystal structure of tetragonal hen egg-white lysozyme at a hydrostatic pressure of 1000 atmospheres has been determined by X-ray diffraction to a nominal resolution of 2 A. The crystals, originally grown in 0.83 M-NaCl, had to be transferred to 1.4 M-NaCl to prevent crystal cracking at 300 to 400 atm. The a and b axes of the unit cell contracted by 0.6%, whilst the c axis increased by 0.1%. The unit cell volume contracted by 1.1%. Both the 1 atm and the 1000 atm structures were refined by restrained least-squares to yield final R factors of 14.9% in each case. Since the data were collected by an accurate difference protocol, the change in structure is considered to be more accurate than the absolute structure. The probable accuracy of the atomic shifts is shown to be +/- 0.06 A. The estimated volume decrease of the whole molecule corresponded to an isothermal compressibility of 4.7 X 10(-3) kbar-1. The contraction was non-uniformly distributed. Domain 2 (residues 40 to 88) was essentially incompressible, whilst domain 1 (residues 1 to 39, 89 to 129) had a compressibility of 5.7 X 10(-3) kbar-1. The interdomain region was also compressible. The average B factor decreased about 1 A2 at 1000 atm, but there was a wide range of decreases and increases in individual values. The pressure-induced deformation was analyzed with difference distance matrices. The beta-sheet (residues 42 to 60) and helix 2 (residues 24 to 36) were deformed the least under pressure. The other helices were more deformed and one loop region (residues 61 to 87) actually appeared to expand. The main-chain atoms of the beta-sheet and helix 2 were used to perform a least-squares superposition of the 1 atm and 1000 atm models. The root-mean-square pressure-induced shift for all atoms was 0.2 A, with a few atoms moving more than 1 A. There was no evidence for co-ordinated movement about the hinge axis defined by alpha carbon atoms 38 and 97. The 1 atm and 1000 atm refined structures included 151 and 163 ordered water molecules, respectively. The changes in these ordered water molecules and the mean compressibility of all of the solvent in the crystal will be described elsewhere.

Crystallographic observations of the metal ion triple in the active site region of alkaline phosphatase.

Diffraction analysis reveals three metal ion binding sites, M1, M2 and M3, in each of two symmetric active centers 32 A apart in alkaline phosphatase from Escherichia coli with intermediate distances within the center of 4, 5 and 7 A for M1-M2, M2-M3 and M1-M3, respectively. A fourth site, M4, has been reported 25 A away. Arsenate, a product analog, binds adjacent to M1 and M2. The active serine residue, 102, which is phosphorylated during normal enzymatic turnover, is also adjacent to M1 and M2 and arginine 166 is adjacent to the arsenate. The implication with respect to the mechanism is that M1, M2 and Arg 166 neutralize and redistribute charges within the phosphate group, activate the serine hydroxyl, and stabilize transition states during bond formation and breakage. Three sites, A, B and C, have been deduced from solution studies and defined specifically on the basis of nuclear magnetic resonance data, binding studies and activity data. The evidence suggests correspondence of A to M1, B to M2, and C to M3. Strong antagonism between binding at M1 and M2 is evidenced crystallographically by a pseudo-saturation, which is relieved by phosphate binding. Local destabilization of the protein, particularly residues 323 through 333, is produced by removal of metals from the crystal.

Sendai virus-mediated lysis of liposomes requires cholesterol.

Vesicles were constituted with glycophorin, the Sendai virus receptor of human erythrocytes, and loaded with calcein, a polar derivative of fluorescein, at self-quenching concentrations. On exposure to Sendai virus and mild hypo-osmotic stress, vesicles of the appropriate composition released a significant portion of their internal contents, as indicated by an increase in calcein fluorescence. Susceptible liposomes were not induced to leak by heat-inactivated virus or by trypsin-treated virus. The response of the vesicles to virus attachment is thus analogous to virus-induced hemolysis and presumably involves fusion of the vesicle and virus membranes. In addition to glycophorin and phosphatidylcholine, cholesterol was absolutely required for the lytic response to the virus. The need for cholesterol was not attributable to inactivation of the virus by liposomes without cholesterol. The presence of gangliosides increased the encapsulated volume of the liposomes, but gangliosides did not effectively substitute for glycophorin. Thin-layer chromatography of lipid extracted from incubated virus and liposomes containing a small amount of a fluorescent phosphatidylcholine indicated that phosphatidylcholine in the vesicle is not chemically altered by functional interaction with the virus.