High-throughput three-dimensional protein structure determination

GPU-powered tools boost molecular visualization

Recent advances in experimental structure determination provide a wealth of structural data on huge macromolecular assemblies such as the ribosome or viral capsids, available in public databases. Further structural models arise from reconstructions using symmetry orders or fitting crystal structures into low-resolution maps obtained by electron-microscopy or small angle X-ray scattering experiments. Visual inspection of these huge structures remains an important way of unravelling some of their secrets. However, such visualization cannot conveniently be carried out using conventional rendering approaches, either due to performance limitations or due to lack of realism. Recent developments, in particular drawing benefit from the capabilities of Graphics Processing Units (GPUs), herald the next generation of molecular visualization solutions addressing these issues. In this article, we present advances in computer science and visualization that help biologists visualize, understand and manipulate large and complex molecular systems, introducing concepts that remain little-known in the bioinformatics field. Furthermore, we compile currently available software and methods enhancing the shape perception of such macromolecular assemblies, for example based on surface simplification or lighting ameliorations.

Exploring protein structural dissimilarity to facilitate structure classification

BACKGROUND

Classification of newly resolved protein structures is important in understanding their architectural, evolutionary and functional relatedness to known protein structures. Among various efforts to improve the database of Structural Classification of Proteins (SCOP), automation has received particular attention. Herein, we predict the deepest SCOP structural level that an unclassified protein shares with classified proteins with an equal number of secondary structure elements (SSEs).

RESULTS

We compute a coefficient of dissimilarity (Omega) between proteins, based on structural and sequence-based descriptors characterising the respective constituent SSEs. For a set of 1,661 pairs of proteins with sequence identity up to 35%, the performance of Omega in predicting shared Class, Fold and Super-family levels is comparable to that of DaliLite Z score and shows a greater than four-fold increase in the true positive rate (TPR) for proteins sharing the Family level. On a larger set of 600 domains representing 200 families, the performance of Z score improves in predicting a shared Family, but still only achieves about half of the TPR of Omega. The TPR for structures sharing a Super-family is lower than in the first dataset, but Omega performs slightly better than Z score. Overall, the sensitivity of Omega in predicting common Fold level is higher than that of the DaliLite Z score.

CONCLUSION

Classification to a deeper level in the hierarchy is specific and difficult. So the efficiency of Omega may be attractive to the curators and the end-users of SCOP. We suggest Omega may be a better measure for structure classification than the DaliLite Z score, with the caveat that currently we are restricted to comparing structures with equal number of SSEs.

Scoring functions for de novo protein structure prediction revisited

De novo protein structure prediction methods attempt to predict tertiary structures from sequences based on general principles that govern protein folding energetics and/or statistical tendencies of conformational features that native structures acquire, without the use of explicit templates. A general paradigm for de novo prediction involves sampling the conformational space, guided by scoring functions and other sequence-dependent biases, such that a large set of candidate ("decoy") structures are generated, and then selecting native-like conformations from those decoys using scoring functions as well as conformer clustering. High-resolution refinement is sometimes used as a final step to fine-tune native-like structures. There are two major classes of scoring functions. Physics-based functions are based on mathematical models describing aspects of the known physics of molecular interaction. Knowledge-based functions are formed with statistical models capturing aspects of the properties of native protein conformations. We discuss the implementation and use of some of the scoring functions from these two classes for de novo structure prediction in this chapter.

Stereospecific assignments of protein NMR resonances based on the tertiary structure and 2D/3D NOE data

In many cases of protein structure determination by NMR a high-quality structure is required. An important contribution to structural precision is stereospecific assignment of magnetically nonequivalent prochiral methylene and methyl groups, eliminating the need for introducing pseudoatoms and pseudoatom corrections in distance restraint lists. Here, we introduce the stereospecific assignment program that uses the resonance assignment, a preliminary 3D structure and 2D and/or 3D nuclear Overhauser effect spectroscopy peak lists for stereospecific assignment. For each prochiral group the algorithm automatically calculates a score for the two different stereospecific assignment possibilities, taking into account the presence and intensity of the nuclear Overhauser effect (NOE) peaks that are expected from the local environment of each prochiral group (i.e., the close neighbors). The performance of the algorithm has been tested and used on NMR data of alpha-helical and beta-sheet proteins using homology models and/or X-ray structures. The program produced no erroneous stereospecific assignments provided the NOEs were carefully picked and the 3D model was sufficiently accurate. The set of NOE distance restraints produced by nmr2st using the results of the SSA module was superior in generating good-quality ensembles of NMR structures (low deviations from upper limits in conjunction with low root-mean-square-deviation values) in the first round of structure calculations. The program uses a novel approach that employs the entire 3D structure of the protein to obtain stereospecific assignment; it can be used to speed up the NMR structure refinement and to increase the quality of the final NMR ensemble even when no scalar or residual dipolar coupling information is available.

Effect of pH, urea, peptide length, and neighboring amino acids on alanine α-proton random coil chemical shifts

Accurate random coil alpha-proton chemical shift values are essential for precise protein structure analysis using chemical shift index (CSI) calculations. The current study determines the chemical shift effects of pH, urea, peptide length and neighboring amino acids on the alpha-proton of Ala using model peptides of the general sequence GnXaaAYaaGn, where Xaa and Yaa are Leu, Val, Phe, Tyr, His, Trp or Pro, and n = 1-3. Changes in pH (2-6), urea (0-1M), and peptide length (n = 1-3) had no effect on Ala alpha-proton chemical shifts. Denaturing concentrations of urea (8M) caused significant downfield shifts (0.10 +/- 0.01 ppm) relative to an external DSS reference. Neighboring aliphatic residues (Leu, Val) had no effect, whereas aromatic amino acids (Phe, Tyr, His and Trp) and Pro caused significant shifts in the alanine alpha-proton, with the extent of the shifts dependent on the nature and position of the amino acid. Smaller aromatic residues (Phe, Tyr, His) caused larger shift effects when present in the C-terminal position (approximately 0.10 vs. 0.05 ppm N-terminal), and the larger aromatic tryptophan caused greater effects in the N-terminal position (0.15 ppm vs. 0.10 C-terminal). Proline affected both significant upfield (0.06 ppm, N-terminal) and downfield (0.25 ppm, C-terminal) chemical shifts. These new Ala correction factors detail the magnitude and range of variation in environmental chemical shift effects, in addition to providing insight into the molecular level interactions that govern protein folding.

A protein folding degree measure and its dependence on crystal packing, protein size, secondary structure, and domain structural class

Comparing two or more protein structures with respect to their degree of folding is common practice in structural biology despite the fact that there is no scale for a folding degree. Here we introduce a formal definition of a folding degree, capable of quantitative characterization. This enables ordering among protein chains based on their degree of folding. The folding degree of a data set of 152 representative nonhomologous proteins is then studied. We demonstrate that the variation in the folding degree seen for this data set is not due to crystallization artifacts or experimental conditions, such as resolution, refinement protocol, pH, or temperature. A good linear relationship is observed between the folding degree and the percentages of secondary structures in the protein. The folding degree is able to account for the small changes produced in the structure due to crystal packing and temperature. Automating the classification of proteins into their respective structural domain classes, namely mainly-alpha, mainly-beta, and alpha-beta, is also possible.

Identification, expression, and purification of a unique stable domain from human HSPC144 protein

HSPC144 is a newly identified gene in human CD34(+) hematopoietic stem/progenitor cells. In this work, we have expressed and purified the 225-residue protein from Escherichia coli BL21 (DE3) and identified a stable fragment HSPC144-P (residues 44-225) by limited proteolysis method. The HSPC144-P fragment exhibits high stability with a little increase of secondary structure percentage as compared with the full-length protein. We anticipated that the N-terminally truncated protein possesses a more compact structure. By sequence analysis, the proteolytic fragment shares a great similarity with DUF589 domain, a previously identified domain with unknown function. This novel domain is highly conserved in Thy28 proteins and is worthy of structural and functional studies. We have subcloned this homologous domain from HSPC144 protein and purified to homogeneity for structure analysis. The (15)N and (15)N/(13)C-labeled DUF589 domain samples have been prepared successfully and determination of the NMR structure is in progress.

Rapid purification and crystal structure analysis of a small protein carrying two terminal affinity tags

Small peptide tags are often fused to proteins to allow their affinity purification in high-throughput structure analysis schemes. To assess the compatibility of small peptide tags with protein crystallization and to examine if the tags alter the three-dimensional structure, the N-terminus of the chicken alpha-spectrin SH3 domain was labeled with a His6 tag and the C-terminus with a StrepII tag. The resulting protein, His6-SH3-StrepII, consists of 83 amino-acid residues, 23 of which originate from the tags. His6-SH3-StrepII is readily purified by dual affinity chromatography, has very similar biophysical characteristics as the untagged protein domain and crystallizes readily from a number of sparse-matrix screen conditions. The crystal structure analysis at 2.3 A resolution proves native-like structure of His6-SH3-StrepII and shows the entire His6 tag and part of the StrepII tag to be disordered in the crystal. Obviously, the fused affinity tags did not interfere with crystallization and structure analysis and did not change the protein structure. From the extreme case of His6-SH3-StrepII, where affinity tags represent 27% of the total fusion protein mass, we extrapolate that protein constructs with N- and C-terminal peptide tags may lend themselves to biophysical and structural investigations in high-throughput regimes.

Rainbow tags: a visual tag system for recombinant protein expression and purification

Visualization systems for tracking proteins are standard experimental tools in most areas of biological research apart from protein purification. Here, we have sought to plug this gap by producing red and yellow visual tags using the heme-binding domain of mosquito cytochrome b5 and the flavin mononucleotide (FMN)-binding domain of human P450 reductase. Tests with colorless glutathione-S-transferase (GST) show them to be simple and effective tools for visually identifying correctly folded protein and tracking protein molecules through protein expression and purification. Furthermore, the characteristic absorbance signatures of the colored tags can be used to quantify protein concentrations directly, which allows purification to be linked to colorimetric detection. This technology, which we call Rainbow Tagging, facilitates expression and downstream processing of recombinant proteins, paving the way for the development of automated high-throughput protein expression systems.

High-throughput measurement of protein stability in microtiter plates

The direct determination of protein stability at high throughput has applications in proteomics, directed evolution, and formulation. Each application places different requirements on the accuracy of stability or transition midpoint determination. The measurement of protein stability by chemical denaturation has been previously performed at medium throughput and high accuracy using autotitrating fluorometers, after removal of proteins from the 96-well plate format in which they were expressed and purified. Herein we present a higher-throughput method for measuring and indexing the stability of proteins maintained within the 96-well format using a fluorescence microplate reader. Protein unfolding transitions were monitored by tryptophan fluorescence at 340 nm and assessed using bovine and equine cytochrome c (cyt c), as well as bovine serum albumin (BSA) stabilized with various amounts of palmitic acid. Two different approaches for generating unfolding curves in microtiter plates have been evaluated for their accuracy and applicability. Unfolding curves generated by the serial addition of denaturant into single wells allowed high-throughput stability screens capable of identifying protein variants with unfolding midpoint differences of 0.15 M denaturant concentration or larger. Such a method would be suitable for screening large numbers of proteins, as typically generated for directed evolution. Unfolding curves generated using one well per denaturant concentration allowed for medium-throughput stability screening and generated more accurate and precise stability values (C(1/2) +/- 0.05 M, m(G), and DeltaG(H2O)) for cyt c that are similar to values reported in literature. This method is suitable for screening the smaller numbers of proteins generated in proteomic research programmes. By using BSA stabilized with various palmitate concentrations and simple numerical indexing, it was shown that both experimental methods can successfully rank the order of protein stability.

Laboratory scale structural genomics

At Lawrence Livermore National Laboratory, the development of the TB structural genomics consortium crystallization facility has paralleled several local proteomics research efforts that have grown out of gene expression microarray and comparative genomics studies. Collective experience gathered from TB consortium labs and other centers involved in the NIH-NIGMS protein structure initiative allows us to explore the possibilities and challenges of pursuing structural genomics on an academic laboratory scale. We discuss our procedures and protocols for genomic targeting approaches, primer design, cloning, small scale expression screening, scale-up and purification, through to automated crystallization screening and data collection. The procedures are carried out by a small group using a combination of traditional approaches, innovative molecular biochemistry approaches, software automation, and a modest investment in robotic equipment.

Reference structure tomography

Reference structure tomography (RST) uses multidimensional modulations to encode mappings between radiating objects and measurements. RST may be used to image source-density distributions, estimate source parameters, or classify sources. The RST paradigm permits scan-free multidimensional imaging, data-efficient and computation-efficient source analysis, and direct abstraction of physical features. We introduce the basic concepts of RST and illustrate the use of RST for multidimensional imaging based on a geometric radiation model.

In vivo uniform 15 N-isotope labelling of plants: Using the greenhouse for structural proteomics

Isotope labelling of proteins is important for progress in the field of structural proteomics. It enables the utilisation of the power of nuclear magnetic resonance spectroscopy (NMR) for the characterisation of the three-dimensional structures and corresponding dynamical features of proteins. The usual approach to obtain isotopically labelled protein molecules is by expressing the corresponding gene in bacterial or yeast host organisms, which grow on isotope-enriched media. This method has several drawbacks. Here, we demonstrate that it is possible to fully label a plant with (15)N-isotopes. The advantage of in vivo labelling of higher organisms is that all constituting proteins are labelled and become available as functional, post-translationally modified, correctly folded proteins. A hydroponics set-up was used to create the first example of a uniformly (15)N-labelled (> 98%) plant species, the potato plant (Solanum tuberosum L., cv. Elkana). Two plants were grown at low costs using potassium-[(15)N]-nitrate as the sole nitrogen source. At harvest time, a total of 3.6 kg of potato tubers and 1.6 kg of foliage, stolons and roots were collected, all of which were fully (15)N-labelled. Gram quantities of soluble (15)N-labelled proteins (composed mainly of the glycoprotein patatin and Kunitz-type protease inhibitors) were isolated from the tubers. NMR results on the complete proteome of potato sap and on an isolated protease inhibitor illustrate the success of the labelling procedure. The presented method of isotope labelling is easily modified to label other plants. Its envisioned impact in the field of structural proteomics of plants is discussed.

High-throughput crystallography to enhance drug discovery

Protein crystallography has traditionally been regarded as a resource-intensive, time-consuming technique that, with some notable exceptions, has not made a significant impact on drug discovery. However, inspired by successes in the genome-sequencing initiatives, recent years have seen major changes in X-ray crystallography methodologies and the concept of high-throughput crystallography has emerged. Advances have been made in all phases of the process, including improved molecular biology, protein expression, crystallization and structure determination. This transformation has allowed X-ray crystallography to impact more broadly in the drug-discovery process, extending its utility from structure-based lead optimisation to novel fragment-based lead generation approaches.

Facilities and methods for the high-throughput crystal structural analysis of human proteins

Facilities and methods for the high-throughput crystal structure analysis of human proteins are described as recently established in the Protein Structure Factory, a Berlin-area structural genomics project. Genes encoding human proteins are expressed in either recombinant Escherichia coli or yeast (Saccharomyces cerevisiae or Pichia pastoris). To facilitate and standardize protein purification, the target proteins are produced with various tags for affinity chromatography. For high-throughput crystallization, a robotic station is being set up that has the capacity to handle 960 000 experiments simultaneously. The resulting protein crystals will be subjected to X-ray diffraction experiments at the third-generation synchrotron storage ring BESSY where protein crystallography beamlines are currently under construction. The Protein Structure Factory's strategy for high-throughput production and structure analysis of human proteins is evaluated based on first results.

Micromixer-based time-resolved NMR: applications to ubiquitin protein conformation

Time-resolved NMR spectroscopy is used to studychanges in protein conformation based on the elapsed time after a change in the solvent composition of a protein solution. The use of a micromixer and a continuous-flow method is described where the contents of two capillary flows are mixed rapidly, and then the NMR spectra of the combined flow are recorded at precise time points. The distance after mixing the two fluids and flow rates define the solvent-protein interaction time; this method allows the measurement of NMR spectra at precise mixing time points independent of spectral acquisition time. Integration of a micromixer and a microcoil NMR probe enables low-microliter volumes to be used without losing significant sensitivity in the NMR measurement. Ubiquitin, the model compound, changes its conformation from native to A-state at low pH and in 40% or higher methanol/water solvents. Proton NMR resonances of the His-68 and the Tyr-59 of ubiquitin are used to probe the conformational changes. Mixing ubiquitin and methanol solutions under low pH at microliter per minute flow rates yields both native and A-states. As the flow rate decreases, yielding longer reaction times, the population of the A-state increases. The micromixer-NMR system can probe reaction kinetics on a time scale of seconds.

Specific radiation damage can be used to solve macromolecular crystal structures

The use of third generation synchrotron sources has led to renewed concern about the effect of ionizing radiation on crystalline biological samples. In general, the problem is seen as one to be avoided. However, in this paper, it is shown that, far from being a hindrance to successful structure determination, radiation damage provides an opportunity for phasing macromolecular structures. This is successfully demonstrated for both a protein and an oligonucleotide, by way of which complete models were built automatically. The possibility that, through the exploitation of radiation damage, the phase problem could become less of a barrier to macromolecular crystal structure determination is discussed.

Reconstructing the binding site of factor Xa in trypsin reveals ligand-induced structural plasticity

In order to investigate issues of selectivity and specificity in protein-ligand interactions, we have undertaken the reconstruction of the binding pocket of human factor Xa in the structurally related rat trypsin by site-directed mutagenesis. Three sequential regions (the "99"-, the "175"- and the "190"- loops) were selected as representing the major structural differences between the ligand binding sites of the two enzymes. Wild-type rat trypsin and variants X99rT and X(99/175/190)rT were expressed in yeast, and analysed for their interaction with factor Xa and trypsin inhibitors. For most of the inhibitors studied, progressive loop replacement at the trypsin surface resulted in inhibitory profiles akin to factor Xa. Crystals of the variants were obtained in the presence of benzamidine (3), and could be soaked with the highly specific factor Xa inhibitor (1). Binding of the latter to X99rT results in a series of structural adaptations to the ligand, including the establishment of an "aromatic box" characteristic of factor Xa. In X(99/175/190)rT, introduction of the 175-loop results in a surprising re-orientation of the "intermediate helix", otherwise common to trypsin and factor Xa. The re-orientation is accompanied by an isomerisation of the Cys168-Cys182 disulphide bond, and burial of the critical Phe174 side-chain. In the presence of (1), a major re-organisation of the binding site takes place to yield a geometry identical to that of factor Xa. In all, binding of (1) to trypsin and its variants results in significant structural rearrangements, inducing a binding surface strongly reminiscent of factor Xa, against which the inhibitor was optimised. The structural data reveal a plasticity of the intermediate helix, which has been implicated in the functional cofactor dependency of many trypsin-like serine proteinases. This approach of grafting loops onto scaffolds of known related structures may serve to bridge the gap between structural genomics and drug design.

Trypsin mutants for structure-based drug design: expression, refolding and crystallisation

New techniques in drug discovery are essential for the fast and efficient development of novel innovative drugs to deal with the challenges of the future. Structure determinations of various members of serine proteinases have provided a basis for computer-based drug design within this class of enzymes. In many proteins of interest, however, this course is blocked through a lack of suitable crystals. As a strategy for circumventing such problems, we have investigated the use of surrogate proteins for studying protein-ligand interactions. To test the feasibility of this approach, we have chosen bovine trypsin as a scaffold to reconstruct the ligand binding site of factor Xa. The simple modular design of trypsin, its readiness to crystallise and straightforward handling lends itself to such drug design by proxy. The expression, folding, purification, crystallographic and kinetic characterisation of bovine trypsin forms with factor Xa phenotype are presented.

Towards structural genomics of RNA: rapid NMR resonance assignment and simultaneous RNA tertiary structure determination using residual dipolar couplings

We report a new residual dipolar couplings (RDCs) based NMR procedure for rapidly determining RNA tertiary structure demonstrated on a uniformly (15)N/(13)C-labeled 27 nt variant of the trans-activation response element (TAR) RNA from HIV-I. In this procedure, the time-consuming nuclear Overhauser enhancement (NOE)-based sequential assignment step is replaced by a fully automated RDC-based assignment strategy. This approach involves examination of all allowed sequence-specific resonance assignment permutations for best-fit agreement between measured RDCs and coordinates for sub-structures in a target RNA. Using idealized A-form geometries to model Watson-Crick helices and coordinates from a previous X-ray structure to model a hairpin loop in TAR, the best-fit RDC assignment solutions are determined very rapidly (<five minutes of computational time) and are in complete agreement with corresponding NOE-based assignments. Orientational constraints derived from RDCs are used simultaneously to assemble sub-structures into an RNA tertiary conformation. Through enhanced speeds of application and reduced reliance on chemical shift dispersion, this RDC-based approach lays the foundation for rapidly determining RNA conformations in a structural genomics context, and may increase the size limit of RNAs that can be examined by NMR.

Structure-based screening of low-affinity compounds

Conventional bioassay-based screening remains a mainstream approach for lead discovery. However, its limitations have meant that other, more biophysical methods, such as X-ray crystallography and NMR, are now being developed as lead discovery tools. These methods are particularly effective at detecting the binding of low affinity, low molecular weight compounds and transforming them into novel potent leads using structure-guided chemistry. Here, we describe some of the technologies and approaches that are being developed in structure-based screening using X-ray crystallography, which promise to have a major impact on lead discovery.

High-throughput crystallography for lead discovery in drug design

Knowledge of the three-dimensional structures of protein targets now emerging from genomic data has the potential to accelerate drug discovery greatly. X-ray crystallography is the most widely used technique for protein structure determination, but technical challenges and time constraints have traditionally limited its use primarily to lead optimization. Here, we describe how significant advances in process automation and informatics have aided the development of high-throughput X-ray crystallography, and discuss the use of this technique for structure-based lead discovery.