SnB version 2.3: triplet sieve phasing for centrosymmetric structures.

A new version of the direct-methods program SnB has been developed. This version incorporates the triplet sieve method for phasing centrosymmetric structures in a way that is transparent to users. The triplet sieve procedure may decrease significantly the time required to achieve a solution for such structures.

Rapid and automated substructure solution by Shake-and-Bake.

Direct methods of phase determination have played an important role in determining heavy-atom substructures from difference amplitudes of native-derivative crystal pairs or crystals containing anomalously scattering atoms. The minimal principle-based Shake-and-Bake procedure is one of the most successful direct methods for heavy-atom substructure determination. The computer program SnB, which implements the Shake-and-Bake procedure and is part of the protein structure-determination package BnP, has recently been optimized for rapid and automated substructure determination. Specifically, SnB has been upgraded with (i) a newly developed statistical minimal function for higher success rates, (ii) an optimal FFT grid size for dramatic cost-effectiveness improvement, (iii) a dynamic figure of merit for automatic substructure-solution detection and (iv) a strategy of alternation of anomalous differences with isomorphous dispersive differences for virtually guaranteed substructure solution.

Optimizing statistical Shake-and-Bake for Se-atom substructure determination.

A novel statistical approach to the phase problem in X-ray crystallography was introduced in a recent paper [Xu & Hauptman (2004), Acta Cryst. A60, 153-157]. In this approach, a new minimal function based on the statistical distribution of structure-invariant values serves as the foundation of an optimization procedure called statistical Shake-and-Bake. Favorable application of this procedure to Se-atom substructure determination depends on the choice of the statistical interval over which the function is defined. The effects of interval variation have been studied for 19 Se-atom substructures ranging in size from five to 70 Se atoms in the asymmetric unit and the results have shown an overall improvement in success rate relative to traditional Shake-and-Bake. Statistical Shake-and-Bake is being incorporated as the default optimization procedure in newly distributed versions of the SnB and BnP computer programs.

Determining structure and function of steroid dehydrogenase enzymes by sequence analysis, homology modeling, and rational mutational analysis.

The short-chain oxidoreductase (SCOR) family of enzymes includes over 6,000 members identified in sequenced genomes. Of these enzymes, approximately 300 have been characterized functionally, and the three-dimensional crystal structures of approximately 40 have been reported. Since some SCOR enzymes are steroid dehydrogenases involved in hypertension, diabetes, breast cancer, and polycystic kidney disease, it is important to characterize the other members of the family for which the biological functions are currently unknown and to determine their three-dimensional structure and mechanism of action. Although the SCOR family appears to have only a single fully conserved residue, it was possible, using bioinformatics methods, to determine characteristic fingerprints composed of 30-40 residues that are conserved at the 70% or greater level in SCOR subgroups. These fingerprints permit reliable prediction of several important structure-function features including cofactor preference, catalytic residues, and substrate specificity. Human type 1 3beta-hydroxysteroid dehydrogenase isomerase (3beta-HSDI) has 30% sequence identity with a human UDP galactose 4-epimerase (UDPGE), a SCOR family enzyme for which an X-ray structure has been reported. Both UDPGE and 3-HSDI appear to trace their origins back to bacterial 3alpha,20beta-HSD. Combining three-dimensional structural information and sequence data on the 3alpha,20beta-HSD, UDPGE, and 3beta-HSDI subfamilies with mutational analysis, we were able to identify the residues critical to the dehydrogenase function of 3-HSDI. We also identified the residues most probably responsible for the isomerase activity of 3beta-HSDI. We test our predictions by specific mutations based on sequence analysis and our structure-based model.

Rational proteomics II: electrostatic nature of cofactor preference in the short-chain oxidoreductase (SCOR) enzyme family.

The dominant role of long-range electrostatic interatomic interactions in nicotinamide adenine dinucleotide/nicotinamide adenine dinucleotide phosphate (NAD/NADP) cofactor recognition has been shown for enzymes of the short-chain oxidoreductase (SCOR) family. An estimation of cofactor preference based only on the contribution of the electrostatic energy term to the total energy of enzyme-cofactor interaction has been tested for approximately 40 known three-dimensional (3D) crystal complexes and approximately 330 SCOR enzymes, with cofactor preference predicted by the presence of Asp or Arg recognition residues at specific 3D positions in the beta2alpha3 loop (Duax et al., Proteins 2003;53:931-943). The results obtained were found to be consistent with approximately 90% reliable cofactor assignments for those subsets. The procedure was then applied to approximately 170 SCOR enzymes with completely uncertain NAD/NADP dependence, due to the lack of Asp and Arg marker residues. The proposed 3D electrostatic approach for cofactor assignment ("3D_DeltaE(el)") has been implemented in an automatic screening procedure, and together with the use of marker residues proposed earlier (Duax et al., Proteins 2003;53:931-943), increases the level of reliable predictions for the putative SCORs from approximately 70% to approximately 90%. It is expected to be applicable for any NAD/NADP-dependent enzyme subset having at least 25-30% sequence identity, with at least one enzyme of known 3D crystal structure.

Rational proteomics I. Fingerprint identification and cofactor specificity in the short-chain oxidoreductase (SCOR) enzyme family.

The short-chain oxidoreductase (SCOR) family of enzymes includes over 2000 members identified in sequenced genomes. Of these enzymes, approximately 200 have been characterized functionally, and the three-dimensional crystal structures of approximately 40 have been reported. Since some SCOR enzymes are involved in hypertension, diabetes, breast cancer, and polycystic kidney disease, it is important to characterize the other members of the family for which the biological functions are currently unknown. Although the SCOR family appears to have only a single fully conserved residue, it was possible, using bioinformatics methods, to determine characteristic fingerprints composed of 30-40 residues that are conserved at the 70% or greater level in SCOR subgroups. These fingerprints permit reliable prediction of several important structure-function features including NAD/NADP cofactor preference. For example, the correlation of aspartate or arginine residues with NAD or NADP binding, respectively, predicts the cofactor preference of more than 70% of the SCOR proteins with unknown function. The analysis of conserved residues surrounding the cofactor has revealed the presence of previously undetected CH em leader O hydrogen bonds in the majority of the SCOR crystal structures, predicts the presence of similar hydrogen bonds in 90% of the SCOR proteins of unknown function, and suggests that these hydrogen bonds may play a critical role in the catalytic functions of these enzymes.

The crystal structure of a novel, inactive, lysine 49 PLA2 from Agkistrodon acutus venom: an ultrahigh resolution, AB initio structure determination.

The crystal structure of acutohaemolysin, a lysine 49 phospholipase A2 protein with 1010 non-hydrogen protein atoms and 232 water molecules, has been determined ab initio using the program SnB at an ultrahigh resolution of 0.8 A. The lack of catalytic activity appears to be related to the presence of Phe102, which prevents the access of substrate to the active site. The substitution of tryptophan for leucine at residue 10 interferes with dimer formation and may be responsible for the additional loss of hemolytic activity. The ultrahigh resolution of the experimental diffraction data permits alternative conformations to be modeled for disordered residues, many hydrogen atoms to be located, the protonation of the Nepsilon2 atom in the catalytic residue His48 to be observed experimentally, and the density of the bonding electrons to be analyzed in detail.

Sine-enhanced Shake-and-Bake: the theoretical basis and applications to Se-atom substructures.

Shake-and-Bake is a dual-space direct-methods procedure for crystal structure determination capable of providing ab initio solutions for structures containing as many as 1200 independent non-H atoms, as well as for heavy-atom substructures containing as many as 160 Se atoms in the asymmetric unit. In traditional Shake-and-Bake, phase refinement in reciprocal space utilizes the technique of parameter shift to reduce the value of a minimal function that considers only the mean-square differences between the current values of the cosine structure invariants and their expected values. A new type of minimal function, termed the sine-enhanced minimal function, considers both cosine and sine values of the structure invariants. Exhaustive tests on six Se-atom substructures, ranging in size from 12 to 160 Se atoms in the asymmetric unit, have shown that a two- to eightfold increase in the percentage of trials that converge to solution is attainable with the technique of sine-enhanced parameter shift. The corresponding sine-enhanced Shake-and-Bake, with suitable default parameter values, is being incorporated into a new distributed version of the SnB computer program.