Structural Genomics Initiatives Research Articles

Enzyme Function Prediction, Discovery, and Characterization in Undergraduate Biochemistry Teaching and Research Labs

The Structural Genomics Initiative was an effort by consortiums to solve as many unique protein structures as possible; the Protein Data Bank contains a number of enzymes whose structures have been solved, but for which no enzymatic activity has been determined. The Enzyme Function Initiative is an effort to determine as many unique enzyme functions as possible. There are a number of putative NUDIX Hydrolase superfamily members whose structures have been solved, but for which no enzymatic activity has been determined. We have catalogued the structurally-determined enzymes within the NUDIX Hydrolase superfamily using BLAST, Dali, SCOP, and PDB. We then began to characterize these enzymes in the biochemistry teaching laboratory and are finishing their characterization in the research lab. In the biochemistry lab course, the students have expressed his-tagged Nudix Hydrolases (PDB entries 2AZW, 2PQV, 3QSJ, and 3R03) from plasmids obtained from DNASU, purified the enzymes using nickel affinity chromatography, and then did enzyme assays to determine the substrates. From the assays, the students discovered Nudix Hydrolase activities for these enzymes, which we have finished characterizing in the research lab.

The SGC beyond structural genomics: redefining the role of 3D structures by coupling genomic stratification with fragment-based discovery.

The ongoing explosion in genomics data has long since outpaced the capacity of conventional biochemical methodology to verify the large number of hypotheses that emerge from the analysis of such data. In contrast, it is still a gold-standard for early phenotypic validation towards small-molecule drug discovery to use probe molecules (or tool compounds), notwithstanding the difficulty and cost of generating them. Rational structure-based approaches to ligand discovery have long promised the efficiencies needed to close this divergence; in practice, however, this promise remains largely unfulfilled, for a host of well-rehearsed reasons and despite the huge technical advances spearheaded by the structural genomics initiatives of the noughties. Therefore the current, fourth funding phase of the Structural Genomics Consortium (SGC), building on its extensive experience in structural biology of novel targets and design of protein inhibitors, seeks to redefine what it means to do structural biology for drug discovery. We developed the concept of a Target Enabling Package (TEP) that provides, through reagents, assays and data, the missing link between genetic disease linkage and the development of usefully potent compounds. There are multiple prongs to the ambition: rigorously assessing targets’ genetic disease linkages through crowdsourcing to a network of collaborating experts; establishing a systematic approach to generate the protocols and data that comprise each target’s TEP; developing new, X-ray-based fragment technologies for generating high quality chemical matter quickly and cheaply; and exploiting a stringently open access model to build multidisciplinary partnerships throughout academia and industry. By learning how to scale these approaches, the SGC aims to make structures finally serve genomics, as originally intended, and demonstrate how 3D structures systematically allow new modes of druggability to be discovered for whole classes of targets.

On the dynamical incompleteness of the Protein Data Bank.

Major scientific challenges that are beyond the capability of individuals need to be addressed by multi-disciplinary and multi-institutional consortia. Examples of these endeavours include the Human Genome Project, and more recently, the Structural Genomics (SG) initiative. The SG initiative pursues the expansion of structural coverage to include at least one structural representative for each protein family to derive the remaining structures using homology modelling. However, biological function is inherently connected with protein dynamics that can be studied by knowing different structures of the same protein. This ensemble of structures provides snapshots of protein conformational diversity under native conditions. Thus, sequence redundancy in the Protein Data Bank (PDB) (i.e. crystallization of the same protein under different conditions) is therefore an essential input contributing to experimentally based studies of protein dynamics and providing insights into protein function. In this work, we show that sequence redundancy, a key concept for exploring protein dynamics, is highly biased and fundamentally incomplete in the PDB. Additionally, our results show that dynamical behaviour of proteins cannot be inferred using homologous proteins. Minor to moderate changes in sequence can produce great differences in dynamical behaviour. Nonetheless, the structural and dynamical incompleteness of the PDB is apparently unrelated concepts in SG. While the first could be reversed by promoting the extension of the structural coverage, we would like to emphasize that further focused efforts will be needed to amend the incompleteness of the PDB in terms of dynamical information content, essential to fully understand protein function.

Structural basis of the substrate specificity and instability in solution of a glycosidase from Lactobacillus plantarum

Statistics from structural genomics initiatives reveal that around 50–55% of the expressed, non-membrane proteins cannot be purified and therefore structurally characterized due to solubility problems, which emphasized protein solubility as one of the most serious concerns in structural biology projects. Lactobacillus plantarum CECT 748T produces an aggregation-prone glycosidase (LpBgl) that we crystallized previously. However, this result could not be reproduced due to protein instability and therefore further high-resolution structural analyses of LpBgl were impeded. The obtained crystals of LpBgl diffracted up to 2.48Å resolution and permitted to solve the structure of the enzyme. Analysis of the active site revealed a pocket for phosphate-binding with an uncommon architecture, where a phosphate molecule is tightly bound suggesting the recognition of 6-phosphoryl sugars. In agreement with this observation, we showed that LpBgl exhibited 6-phospho-β-glucosidase activity. Combination of structural and mass spectrometry results revealed the formation of dimethyl arsenic adducts on the solvent exposed cysteine residues Cys211 and Cys292. Remarkably, the double mutant Cys211Ser/Cys292Ser resulted stable in solution at high concentrations indicating that the marginal solubility of LpBgl can be ascribed specifically to these two cysteine residues. The 2.30Å crystal structure of this double mutant showed no disorder around the newly incorporated serine residues and also loop rearrangements within the phosphate-binding site. Notably, LpBgl could be prepared at high yield by proteolytic digestion of the fusion protein LSLt-LpBgl, which raises important questions about potential hysteretic processes upon its initial production as an enzyme fused to a solubility enhancer.

Complex Reconstitution and Characterization by Combining Co-expression Techniques in Escherichia coli with High-Throughput.

Single protein expression technologies have strongly benefited from the Structural Genomics initiatives that have introduced parallelization at the laboratory level. Specifically, the developments made in the wake of these initiatives have revitalized the use of Escherichia coli as major host for heterologous protein expression. In parallel to these improvements for single expression, technologies for complex reconstitution by co-expression in E. coli have been developed. Assessments of these co-expression technologies have highlighted the need for combinatorial experiments requiring automated protocols. These requirements can be fulfilled by adapting the high-throughput approaches that have been developed for single expression to the co-expression technologies. Yet, challenges are laying ahead that further need to be addressed and that are only starting to be taken into account in the case of single expression. These notably include the biophysical characterization of the samples at the small-scale level. Specifically, these approaches aim at discriminating the samples at an early stage of their production based on various biophysical criteria leading to cost-effectiveness and time-saving. This chapter addresses these various issues to provide the reader with a broad and comprehensive overview of complex reconstitution and characterization by co-expression in E. coli.

Application of in situ diffraction in high-throughput structure determination platforms.

Macromolecular crystallography (MX) is the most powerful technique available to structural biologists to visualize in atomic detail the macromolecular machinery of the cell. Since the emergence of structural genomics initiatives, significant advances have been made in all key steps of the structure determination process. In particular, third-generation synchrotron sources and the application of highly automated approaches to data acquisition and analysis at these facilities have been the major factors in the rate of increase of macromolecular structures determined annually. A plethora of tools are now available to users of synchrotron beamlines to enable rapid and efficient evaluation of samples, collection of the best data, and in favorable cases structure solution in near real time. Here, we provide a short overview of the emerging use of collecting X-ray diffraction data directly from the crystallization experiment. These in situ experiments are now routinely available to users at a number of synchrotron MX beamlines. A practical guide to the use of the method on the MX suite of beamlines at Diamond Light Source is given.

ASSIST: a fast versatile local structural comparison tool

Structural genomics initiatives are increasingly leading to the determination of the 3D structure of target proteins whose catalytic function is not known. The aim of this work was that of developing a novel versatile tool for searching structural similarity, which allows to predict the catalytic function, if any, of these proteins. The algorithm implemented by the tool is based on local structural comparison to find the largest subset of similar residues between an input protein and known functional sites. The method uses a geometric hashing approach where information related to residue pairs from the input structures is stored in a hash table and then is quickly retrieved during the comparison step. Tests on proteins belonging to different functional classes, done using the Catalytic Site Atlas entries as targets, indicate that the algorithm is able to identify the correct functional class of the input protein in the vast majority of the cases. The application was developed in Java SE 6, with a Java Swing Graphic User Interface (GUI). The system can be run locally on any operating system (OS) equipped with a suitable Java Virtual Machine, and is available at the following URL: http://www.computationalbiology.it/software/ASSISTv1.zip.

Squeezing the most from every crystal: the fine details of data collection.

Modern synchrotron beamlines offer instrumentation of unprecedented quality, which in turn encourages increasingly marginal experiments, and for these, as much as ever, the ultimate success of data collection depends on the experience, but especially the care, of the experimenter. A representative set of difficult cases has been encountered at the Structural Genomics Consortium, a worldwide structural genomics initiative of which the Oxford site currently deposits three novel human structures per month. Achieving this target relies heavily on frequent visits to the Diamond Light Source, and the variety of crystal systems still demand customized data collection, diligent checks and careful planning of each experiment. Here, an overview is presented of the techniques and procedures that have been refined over the years and that are considered synchrotron best practice.

Microcrystal manipulation with laser tweezers

X-ray crystallography is the method of choice to deduce atomic resolution structural information from macromolecules. In recent years, significant investments in structural genomics initiatives have been undertaken to automate all steps in X-ray crystallography from protein expression to structure solution. Robotic systems are widely used to prepare crystallization screens and change samples on synchrotron beamlines for macromolecular crystallography. The only remaining manual handling step is the transfer of the crystal from the mother liquor onto the crystal holder. Manual mounting is relatively straightforward for crystals with dimensions of >25 µm; however, this step is nontrivial for smaller crystals. The mounting of microcrystals is becoming increasingly important as advances in microfocus synchrotron beamlines now allow data collection from crystals with dimensions of only a few micrometres. To make optimal usage of these beamlines, new approaches have to be taken to facilitate and automate this last manual handling step. Optical tweezers, which are routinely used for the manipulation of micrometre-sized objects, have successfully been applied to sort and mount macromolecular crystals on newly designed crystal holders. Diffraction data from CPV type 1 polyhedrin microcrystals mounted with laser tweezers are presented.

Development and Testing of a Systematic Approach for Computational Enzyme Function Determination

With the success of the Structural Genomics Initiative and other programs, the number of uncharacterized but crystallized proteins is on the rise. To date, there are over 3500 structures in the Protein Data Bank tagged as unknown function. The use of computational tools to speed up the analysis of these structures has been widely adopted, with varying measures of success. ProMOL, a plugin for PyMOL, was developed in a collaboration between RIT and Dowling College. ProMOL functions by searching for conserved active site “motif templates” in a queried protein structure. A combination of ProMOL, NCBI's BLAST algorithm, and University of Helsinki's DALI alignment server were used to attempt to assign functions to 4 previously uncharacterized proteins: PDBIDs 3DS8, 3FLE, 3LP5, and 2L57. The results from the computational analysis were then confirmed in vitro via functional assays. The computational approach successfully assigned the function of 3DS8, 3FLE (both lipases), and 2L57 (a glutaredoxin). It appears that a combination of BLAST alignment, DALI alignment, and ProMOL alignment is sufficient to computationally predict the function of uncharacterized proteins. This approach may be suitable for classroom settings as well, due to the low cost and ease of accessibility of these tools. The research is supported by 2R15GM078077–02 and 3R15GM078077–02S1.

Bridging the Gap between Sequence and Function

Structural genomics initiatives have generated a massive quantity of high resolution structures and sequenced genomes from archaea, bacteria, viruses and eukaryotes. These numbers are expected to grow rapidly within the coming years. The available data constitute an invaluable resource for the prediction of protein structure and function and will allow us to obtain a more comprehensive structure-based understanding of biological function. The acquisition of new biochemical functionality in the course of evolution does not necessarily involve the transfer of whole genes, but may be limited to the transfer of functional domains. In fact, most of a protein's amino acids serve structural roles and may exhibit a low degree of sequence conservation even among closely related genomes. Therefore, it is critical to identify and compare structurally related domains across a wide spectrum of organisms to reveal unique metabolic functionalities.This presentation will outline examples that have combined sequence alignments, homology modeling and biophysical approaches to predict function. Integration of models with existing knowledge about genomic context, biochemical pathways and sparse experimental data, such as small-angle X-ray scattering and spectroscopic data, enables us to accurately identify domains of unknown function. We also show how minor changes in otherwise highly conserved active sites can significantly affect functionality. There is a growing need for intelligent prediction-based strategies that can tap into our enormous genomic and structural databases and help bridge the gap between sequence and function.

Fragment and Conquer: From Structure to Complexes to Function

In this issue of Structure, Shumilin and colleagues show the power of metabolite screening by means of X-ray crystallography to link an orphan protein domain with an orphan biochemical function. This result paves the way for large-scale functional annotation and sets new objectives for structural genomics.

Publishing structural genomics results continued: the SSGCID Special Issue

Over the past decade, structural genomics (SG) initiatives have provided a large number of new macromolecular structures. A conservative estimate based on data from the Structural Biology Knowledgebase (http://www.sbkb.org) is that about 15% of all structures in the Protein Data Bank (PDB) were deposited by one of the worldwide SG centers. This remarkable record of productivity continues. One of the founding mandates of Acta Crystallographica Section F was to give SG organizations a platform to disseminate their results to the wider scientific community. A fresh approach to fulfilling this mandate appeared in our December 2009 issue: a special section featuring structures produced by the RIKEN–UK SG collaboration. The logical extension of this approach appeared as the October 2010 issue: the first full issue of Acta Cryst. F devoted specifically to the output of a single SG center. This first special issue featured 35 papers from the Joint Center of Structural Genomics (JCSG) at the Scripps Institute (La Jolla, USA) and all articles were open access (see Weiss & Einspahr, 2010 ▶). Now we happily acknowledge that its sequel has been published as the September 2011 issue. This issue exclusively featured papers from the Seattle Structural Genomics Center for Infectious Diseases (SSGCID) and again all the papers are open access. In an effort headed and coordinated by Darren Begley (Emerald BioStructures, Bainbridge Island, Washington, USA), 21 structure papers were assembled giving an overview of the SSGCID efforts. These papers were accompanied by eight laboratory communications, describing technical aspects of the work within the SSGCID. We would like to congratulate the SSGCID for this fine assembly of papers and we would like to thank all the referees who devoted their time in order to make this issue possible. While assembling these special issues has proven to be a lot of work not only for the authors but also for the reviewers and editors, we hope that others will take advantage of the opportunities afforded by special issues or sections. To our knowledge, these have been well received but we also welcome feedback on this approach from you, our readers.

Multi-class protein fold classification using a new ensemble machine learning approach.

Protein structure classification represents an important process in understanding the associations between sequence and structure as well as possible functional and evolutionary relationships. Recent structural genomics initiatives and other high-throughput experiments have populated the biological databases at a rapid pace. The amount of structural data has made traditional methods such as manual inspection of the protein structure become impossible. Machine learning has been widely applied to bioinformatics and has gained a lot of success in this research area. This work proposes a novel ensemble machine learning method that improves the coverage of the classifiers under the multi-class imbalanced sample sets by integrating knowledge induced from different base classifiers, and we illustrate this idea in classifying multi-class SCOP protein fold data. We have compared our approach with PART and show that our method improves the sensitivity of the classifier in protein fold classification. Furthermore, we have extended this method to learning over multiple data types, preserving the independence of their corresponding data sources, and show that our new approach performs at least as well as the traditional technique over a single joined data source. These experimental results are encouraging, and can be applied to other bioinformatics problems similarly characterised by multi-class imbalanced data sets held in multiple data sources.

Hidden conformations in protein structures

Prediction of interactions between protein residues (contact map prediction) can facilitate various aspects of 3D structure modeling. However, the accuracy of ab initio contact prediction is still limited. As structural genomics initiatives move ahead, solved structures of homologous proteins can be used as multiple templates to improve contact prediction of the major conformation of an unsolved target protein. Furthermore, multiple templates may provide a wider view of the protein's conformational space. However, successful usage of multiple structural templates is not straightforward, due to their variable relevance to the target protein, and because of data redundancy issues. We present here an algorithm that addresses these two limitations in the use of multiple structure templates. First, the algorithm unites contact maps extracted from templates sharing high sequence similarity with each other in a fashion that acknowledges the possibility of multiple conformations. Next, it weights the resulting united maps in inverse proportion to their evolutionary distance from the target protein. Testing this algorithm against CASP8 targets resulted in high precision contact maps. Remarkably, based solely on structural data of remote homologues, our algorithm identified residue-residue interactions that account for all the known conformations of calmodulin, a multifaceted protein. Therefore, employing multiple templates, which improves prediction of contact maps, can also be used to reveal novel conformations. As multiple templates will soon be available for most proteins, our scheme suggests an effective procedure for their optimal consideration. A Perl script implementing the WMC algorithm described in this article is freely available for academic use at http://tau.ac.il/~haimash/WMC.

Protein domain assignment from the recurrence of locally similar structures

Domains are basic units of protein structure and essential for exploring protein fold space and structure evolution. With the structural genomics initiative, the number of protein structures in the Protein Databank (PDB) is increasing dramatically and domain assignments need to be done automatically. Most existing structural domain assignment programs define domains using the compactness of the domains and/or the number and strength of intra-domain versus inter-domain contacts. Here we present a different approach based on the recurrence of locally similar structural pieces (LSSPs) found by one-against-all structure comparisons with a dataset of 6373 protein chains from the PDB. Residues of the query protein are clustered using LSSPs via three different procedures to define domains. This approach gives results that are comparable to several existing programs that use geometrical and other structural information explicitly. Remarkably, most of the proteins that contribute the LSSPs defining a domain do not themselves contain the domain of interest. This study shows that domains can be defined by a collection of relatively small locally similar structural pieces containing, on average, four secondary structure elements. In addition, it indicates that domains are indeed made of recurrent small structural pieces that are used to build protein structures of many different folds as suggested by recent studies.

INTEGRATING COMPUTATIONAL PROTEIN FUNCTION PREDICTION INTO DRUG DISCOVERY INITIATIVES.

Pharmaceutical researchers must evaluate vast numbers of protein sequences and formulate innovative strategies for identifying valid targets and discovering leads against them as a way of accelerating drug discovery. The ever increasing number and diversity of novel protein sequences identified by genomic sequencing projects and the success of worldwide structural genomics initiatives have spurred great interest and impetus in the development of methods for accurate, computationally empowered protein function prediction and active site identification. Previously, in the absence of direct experimental evidence, homology-based protein function annotation remained the gold-standard for in silico analysis and prediction of protein function. However, with the continued exponential expansion of sequence databases, this approach is not always applicable, as fewer query protein sequences demonstrate significant homology to protein gene products of known function. As a result, several non-homology based methods for protein function prediction that are based on sequence features, structure, evolution, biochemical and genetic knowledge have emerged. Herein, we review current bioinformatic programs and approaches for protein function prediction/annotation and discuss their integration into drug discovery initiatives. The development of such methods to annotate protein functional sites and their application to large protein functional families is crucial to successfully utilizing the vast amounts of genomic sequence information available to drug discovery and development processes.

Publishing structural genomics results: the JCSG Special Issue

In the last decade we have witnessed the advent of structural genomics initiatives around the world. While traditional hypothesis-driven research continues to produce a wealth of new macromolecular structures, the contribution of structural genomics structures to the Protein Data Bank (PDB) has become ever more important. For example, the USA-based Protein Structure Initiative (PSI) alone has deposited more than 5000 structures with the PDB during phases I and II of the project. While the various structural genomics initiatives have been immensely successful in producing structures, they have found publishing their results to be yet another bottleneck to overcome. Presently, many thousands of structures in the PDB have no publication associated with them at all. When Acta Crystallographica Section F was founded, one of its mandates was to provide an efficient and effective medium for presenting structural genomics results. We have sought new ways to ease the publication bottleneck and, with the help and advice of structural genomics leaders, now offer a platform to showcase these structures, and the experiments upon which they are based, to the wider structural and scientific communities. The key feature is the collection of the results produced by an individual structural genomics effort into a special section, allowing the structural genomics group to organize and annotate their structures and put groups of structures into context. The first example was published in December 2009 and contained eight papers from the RIKEN–UK structural genomics consortium in a special section. We have now completed the ultimate expression of this strategem, our first structural genomics special issue of Acta Crystallographica Section F, published online and entirely open access at http://journals.iucr.org/f/issues/2010/10/00/issconts.html. The Joint Center for Structural Genomics (JCSG), headed by Ian Wilson at the Scripps Institute (La Jolla, USA), contributed 35 fine papers to this issue and the center deserves our thanks for this concerted effort and perseverance. The October 2010 publication of the JCSG Special Issue defines the successful conclusion of an evolving process that started about two years ago during the 2008 IUCr Congress in Osaka, and it marks an important milestone for our journal – the first issue to focus exclusively on structures from an individual structural genomics group We have so far received a lot of positive feedback about this special issue from readers and from the structural genomics community as a whole. However, it is no more than a start, albeit clearly a successful one. We anticipate more special issues and special sections in the future and we already have new requests and proposals under consideration. Finally, we would like to thank our reviewers for their extraordinary generosity of time and energy in refereeing the papers in this special issue.

Ligands in crystal structures that aid in functional characterization.

Liganded structures can be instrumental in assigning function to uncharacterized proteins by revealing active sites, conserved residues, binding motifs, and substrate specificity. This introduction provides an overview and commentary on the value of liganded structures emerging from the JCSG structural genomics initiative.

Viewing the human microbiome through three-dimensional glasses: integrating structural and functional studies to better define the properties of myriad carbohydrate-active enzymes.

Recent studies have provided an unprecedented view of the trillions of microbes associated with the human body. The human microbiome harbors tremendous diversity at multiple levels: the species that colonize each individual and each body habitat; the genes that are found in each organism's genome; the expression of these genes and the interactions and activities of their protein products. The sources of this diversity are wide-ranging and reflect both environmental and host factors. A major challenge moving forward is defining the precise functions of members of various families of proteins represented in our microbiomes, including the highly diverse carbohydrate-active enzymes (CAZymes) involved in numerous biologically important chemical transformations, such as the degradation of complex dietary polysaccharides. Coupling metagenomic analyses to structural genomics initiatives and to biochemical and other functional assays of CAZymes will be essential for determining how these as well as other microbiome-encoded proteins operate to shape the properties of microbial communities and their human hosts.