Current Opinion in Structural Biology

Published by Elsevier
Print ISSN: 0959-440X
Several databases of protein structural families now exist-organised according to both evolutionary relationships and common folding arrangements. Although these lag behind sequence databases in size, the prospect of structural genomics initiatives means that they may soon include representatives of many of the sequence families. To some extent, functional information can be derived from structural similarity. For some structural families, their function is highly conserved, whereas, for others, it can only be inherited or derived on the basis of additional information (e.g. sequence patterns, common residue clusters and characteristic surface properties).
When energy is a critical quantity, accurate biomolecular simulations rest in substantial part on accurate potential energy functions (force fields). Improvements in methodology for determining parameters--particularly, in the systematic use of computational data obtained from quantum chemical calculations--and enhancements in functional form are leading to better potential energy functions. New calculations have been developed for water (including calculations that incorporate electronic polarizability to take account of the degree to which a molecule can be polarized), proteins, nucleic acids, carbohydrates, lipids, and general organic molecules. Most notably, two new biomolecular force fields have recently been derived and significant redeterminations of the parameters of two existing biomolecular force fields have been carried out. Some progress has also been made in incorporating polarizability into potential energy functions for molecules in general and in improving the treatment of metal-ligand interactions in systems of biomolecular interest.
Recent successes in protein design have illustrated the promise of computational approaches. These methods rely on energy expressions to evaluate the quality of different amino acid sequences for target protein structures. The force fields optimized for design differ from those typically used in molecular mechanics and molecular dynamics calculations.
Different potential energy functions have predominated in protein dynamics simulations, protein design calculations, and protein structure prediction. Clearly, the same physics applies in all three cases. The differences in potential energy functions reflect differences in how the calculations are performed. With improvements in computer power and algorithms, the same potential energy function should be applicable to all three problems. In this review, we examine energy functions currently used for protein design, and look to the molecular mechanics field for advances that could be used in the next generation of design algorithms. In particular, we focus on improved models of the hydrophobic effect, polarization and hydrogen bonding.
Predicting protein sequences that fold into specific native three-dimensional structures is a problem of great potential complexity. Although the complete solution is ultimately rooted in understanding the physical chemistry underlying the complex interactions between amino acid residues that determine protein stability, recent work shows that empirical information about these first principles is embedded in the statistics of protein sequence and structure databases. This review focuses on the use of 'knowledge-based' potentials derived from these databases in designing proteins. In addition, the data suggest how the study of these empirical potentials might impact our fundamental understanding of the energetic principles of protein structure.
Beta-1,4-galactosyltransferase-1, a housekeeping enzyme that functions in the synthesis of glycoconjugates, has two flexible loops, one short and one long. Upon binding a metal ion and UDP-galactose, the loops change from an open to a closed conformation, repositioning residues to lock the ligands in place. Residues at the N-terminal region of the long loop form the metal-binding site and those at the C-terminal region form a helix, which becomes part of the binding site for the oligosaccharide acceptor; the remaining residues cover the bound sugar-nucleotide. After binding of the oligosaccharide acceptor and transfer of the galactose moiety, the product disaccharide unit is ejected and the enzyme returns to the open conformation, repeating the catalytic cycle.
PRD-containing proteins are bacterial transcriptional antiterminators and activators characterised by a duplicated phosphorylation domain involved in the regulation of catabolic operons. Recent genetic and biochemical studies have suggested how the activity of these regulators is positively or negatively controlled through the multiple phosphorylation of conserved histidyl residues. The regulation mode of these proteins has been examined in light of the recently determined first crystal structure of the phosphorylatable domain of the LicT antiterminator.
Of all the techniques that are currently available to measure binding, isothermal titration calorimetry is the only one capable of measuring not only the magnitude of the binding affinity but also the magnitude of the two thermodynamic terms that define the binding affinity: the enthalpy (AH) and entropy (AS) changes. Recent advances in instrumentation have facilitated the development of experimental designs that permit the direct measurement of arbitrarily high binding affinities, the coupling of binding to protonation/deprotonation processes and the analysis of binding thermodynamics in terms of structural parameters. Because isothermal titration calorimetry has the capability to measure different energetic contributions to the binding affinity, it provides a unique bridge between computational and experimental analysis. As such, it is increasingly becoming an essential tool in molecular design.
Fibroblast growth factors (FGFs) are among the best-studied heparin-binding proteins, and heparan sulfate proteoglycans regulate FGF signalling by direct molecular association with FGF and its tyrosine kinase receptor, FGFR. Two recently determined crystal structures of FGF-FGFR-heparin complexes have provided new structural information on how heparin binds to FGF and FGFR, and lead to different models for receptor dimerisation.
Pulsed electron paramagnetic resonance (EPR) distance measurement techniques target macromolecular structure elucidation at both the local and global level. Recent developments in pulse microwave technology and high-field EPR have led to the development of a variety of pulsed EPR distance measurement techniques. These methods have emerged as powerful tools for the determination of structure/function relationships in macromolecular systems. In this review article, we discuss recent applications of long-range and short-range EPR distance measurements.
Leucine-rich repeats (LRRs) are 20-29-residue sequence motifs present in a number of proteins with diverse functions. The primary function of these motifs appears to be to provide a versatile structural framework for the formation of protein-protein interactions. The past two years have seen an explosion of new structural information on proteins with LRRs. The new structures represent different LRR subfamilies and proteins with diverse functions, including GTPase-activating protein rna1p from the ribonuclease-inhibitor-like subfamily; spliceosomal protein U2A', Rab geranylgeranyltransferase, internalin B, dynein light chain 1 and nuclear export protein TAP from the SDS22-like subfamily; Skp2 from the cysteine-containing subfamily; and YopM from the bacterial subfamily. The new structural information has increased our understanding of the structural determinants of LRR proteins and our ability to model such proteins with unknown structures, and has shed new light on how these proteins participate in protein-protein interactions.
Single-subunit RNA polymerases are widespread throughout prokaryotic and eukaryotic organisms, and also viruses. T7 RNA polymerase is one of the simplest DNA-dependent enzymes, capable of transcribing a complete gene without the need for additional proteins. During the past two years, three illuminating crystal structures of T7 RNA polymerase complexed to either T7 lysozyme, which is a transcription inhibitor, an open promoter DNA fragment or a promoter DNA fragment being transcribed into RNA at initiation have been determined. For the first time, these structures describe in detail the intricate mechanism of transcription initiation by T7 RNA polymerase, which is likely to be a general model for other related RNA polymerases.
The alpha helices of transmembrane proteins interact to form higher order structures. These interactions are frequently mediated by packing motifs (such as GxxxG) and polar residues. Recent structural data have revealed that small sidechains are able to both stabilize helical membrane proteins and allow conformational changes in the structure. The strong interactions involving polar sidechains often contribute to protein misfolding or malfunction.
Beta-O-linked N-acetylglucosamine (O-GlcNAc) is an abundant modification of cytosolic and nuclear proteins that occurs in metazoans. O-GlcNAc is dynamically processed by a unique set of enzymes that actively add and remove the modification. Functionally, O-GlcNAc appears to regulate protein stability, subcellular localization and protein-protein interactions. The modification often acts in a reciprocal manner to O-phosphate modifications of proteins and together they can synergistically control the activity of many cellular processes. Recently, O-GlcNAc has been demonstrated to play a significant role in diseases such as diabetes, cancer and neurodegeneration. For example, the increased levels of O-GlcNAc that occur in diabetes are associated with decreased insulin responsiveness in adipocytes.
Ribonuclease III (RNase III) enzymes occur ubiquitously in biology and are responsible for processing RNA precursors into functional RNAs that participate in protein synthesis, RNA interference and a range of other cellular activities. Members of the RNase III enzyme family, including Escherichia coli RNase III, Rnt1, Dicer and Drosha, share the ability to recognize and cleave double-stranded RNA (dsRNA), typically at specific positions or sequences. Recent biochemical and structural data have shed new light on how RNase III enzymes catalyze dsRNA hydrolysis and how substrate specificity is achieved. A major theme emerging from these studies is that accessory domains present in different RNase III enzymes are the key determinants of substrate selectivity, which in turn dictates the specialized biological function of each type of RNase III protein.
Protein interface hot spots, as revealed by alanine scanning mutagenesis, continue to stimulate interest in the biophysical basis of molecular recognition. Although these regions apparently constitute fertile grounds for intermolecular interactions, no general algorithm has yet been developed that can predict hot spots based solely on their shape or composition. The discovery of structural plasticity in hot spot regions indicates that dynamic simulation techniques may be essential for achieving a predictive understanding of binding interface energetics. Future progress will depend as much on the application of new computational approaches for dissecting protein interfaces as on expanding our empirical databank of mutagenic substitutions and their effects. Despite our current theoretical shortcomings, recent methodological advances provide efficient experimental means of probing hot spots and enable immediate applications for hot spots in drug discovery.
Many bacterial pathogens manipulate the host cell cytoskeleton during infection. Such cytoskeletal modulation can occur at several points of contact between the pathogen and the host, and involves extracellular receptors, intracellular signal transduction and cytoskeletal proteins themselves. The field of bacterial pathogenesis has progressed dramatically over the past decade, such that structural knowledge is both timely and essential for a full appreciation of the biology at the pathogen-host interface. Several recent examples involving bacterial proteins that target actin, Rho family GTPases and extracellular receptors have contributed to a structural understanding of eukaryotic cytoskeletal modulation by pathogens.
Conserved RNA structures have traditionally been thought of as potential binding sites for protein factors and consequently are regarded as fulfilling relatively passive albeit important roles in cellular processes. With the discovery of riboswitches, RNA no longer takes a backseat to protein when it comes to affecting gene expression. Riboswitches bind directly to cellular metabolites with exceptional specificity and affinity, and exert control over gene expression through ligand-induced conformational changes in RNA structure. Riboswitches now represent a widespread mechanism by which cells monitor their metabolic state and facilely alter gene expression in response to changing conditions.
Several computational and experimental methods exist for identifying disordered residues within proteins. Computational algorithms can now identify these disordered sequences and predict their occurrence within genomes with relatively high accuracy. Recent advances in NMR and mass spectroscopy permit faster and more detailed studies of disordered states at atomic resolutions. Combining prediction, computation and experimentation is proposed to accelerate and enhance the characterization of intrinsically disordered protein.
It has long been appreciated that green fluorescent protein (GFP) autocatalytically forms its chromophore in a host-independent process; several of the initial steps in the reaction have recently been elucidated. Nevertheless, the end points of the process are unexpectedly diverse, as six chemically distinct chromophores, including two with three rings, have been identified. All fluorescent proteins continuously produce a low level of reactive oxygen species under illumination, which, in some cases, can lead to host cell death. In one extreme but useful example, the red fluorescent protein KillerRed can be used to selectively destroy cells upon brief illumination. Finally, when photophysical processes such as excited-state proton transfer, reversible photobleaching and photoactivation are understood, useful research tools, for example, real-time biosensors and optical highlighters, can result; however, side effects of their use may lead to significant artifacts in time-dependent microscopy experiments.
Sequence comparison is a major step in the prediction of protein structure from existing templates in the Protein Data Bank. The identification of potentially remote homologues to be used as templates for modeling target sequences of unknown structure and their accurate alignment remain challenges, despite many years of study. The most recent advances have been in combining as many sources of information as possible--including amino acid variation in the form of profiles or hidden Markov models for both the target and template families, known and predicted secondary structures of the template and target, respectively, the combination of structure alignment for distant homologues and sequence alignment for close homologues to build better profiles, and the anchoring of certain regions of the alignment based on existing biological data. Newer technologies have been applied to the problem, including the use of support vector machines to tackle the fold classification problem for a target sequence and the alignment of hidden Markov models. Finally, using the consensus of many fold recognition methods, whether based on profile-profile alignments, threading or other approaches, continues to be one of the most successful strategies for both recognition and alignment of remote homologues. Although there is still room for improvement in identification and alignment methods, additional progress may come from model building and refinement methods that can compensate for large structural changes between remotely related targets and templates, as well as for regions of misalignment.
RNA secondary structure is often predicted from sequence by free energy minimization. Over the past two years, advances have been made in the estimation of folding free energy change, the mapping of secondary structure and the implementation of computer programs for structure prediction. The trends in computer program development are: efficient use of experimental mapping of structures to constrain structure prediction; use of statistical mechanics to improve the fidelity of structure prediction; inclusion of pseudoknots in secondary structure prediction; and use of two or more homologous sequences to find a common structure.
The past several years have witnessed the emergence of a new world of nucleic-acid-based architectures with highly predictable and programmable self-assembly properties. For almost two decades, DNA has been the primary material for nucleic acid nanoconstruction. More recently, the dramatic increase in RNA structural information led to the development of RNA architectonics, the scientific study of the principles of RNA architecture with the aim of constructing RNA nanostructures of any arbitrary size and shape. The remarkable modularity and the distinct but complementary nature of RNA and DNA nanomaterials are revealed by the various self-assembly strategies that aim to achieve control of the arrangement of matter at a nanoscale level.
The alphabeta-tubulin dimer assembles into microtubules, essential polymers in all eukaryotic cells. Microtubules are highly dynamic, a property that derives from tubulin's GTPase activity. Both the bacterial homolog, FtsZ, and the recently discovered bacterial tubulins from Prosthecobacter self-assemble in a nucleotide-dependent manner into protofilaments similar to those that form the microtubule wall. A number of structural studies of alphabeta-tubulin, gamma-tubulin (the isoform involved in microtubule nucleation), FtsZ and bacterial tubulin, in a variety of nucleotide and polymerization states, have been reported in the past few years. These studies have revealed the similarities and differences between these structures and their possible functional implications. In particular, a two-state mechanism has been proposed for the recycling of alphabeta-tubulin during the microtubule disassembly-assembly cycle; this mechanism may be unique to eukaryotic dimeric tubulin and the microtubule structure.
Helical filaments were the first structures to be reconstructed in three dimensions from electron microscopic images, and continue to be extensively studied due to the large number of such helical polymers found in biology. In principle, a single image of a helical polymer provides all of the different projections needed to reconstruct the three-dimensional structure. Unfortunately, many helical filaments have been refractory to the application of traditional (Fourier-Bessel) methods due to variability, heterogeneity, and weak scattering. Over the past several years, many of these problems have been surmounted using single-particle type approaches that can do substantially better than Fourier-Bessel approaches. Applications of these new methods to viruses, actin filaments, pili and many other polymers show the great advantages of the new methods.
Depending on whether similar structures are found in the PDB library, the protein structure prediction can be categorized into template-based modeling and free modeling. Although threading is an efficient tool to detect the structural analogs, the advancements in methodology development have come to a steady state. Encouraging progress is observed in structure refinement which aims at drawing template structures closer to the native; this has been mainly driven by the use of multiple structure templates and the development of hybrid knowledge-based and physics-based force fields. For free modeling, exciting examples have been witnessed in folding small proteins to atomic resolutions. However, predicting structures for proteins larger than 150 residues still remains a challenge, with bottlenecks from both force field and conformational search.
Last year, atomic structures of the 50S ribosomal subunit from Haloarcula marismortui and of the 30S ribosomal subunit from Thermus thermophilus were published. A year before that, a 7.8 A resolution electron density map of the 70S ribosome from T. thermophilus appeared. This information is revolutionizing our understanding of protein synthesis.
Central issues concerning protein structure prediction have been highlighted by the recently published summary of the fourth community-wide protein structure prediction experiment (CASP4). Although sequence/structure alignment remains the bottleneck in comparative modeling, there has been substantial progress in fully automated remote homolog detection and in de novo structure prediction. Significant further progress will probably require improvements in high-resolution modeling.
Published rotamer libraries.
Rotamer libraries are widely used in protein structure prediction, protein design, and structure refinement. As the size of the structure data base has increased rapidly in recent years, it has become possible to derive well-refined rotamer libraries using strict criteria for data inclusion and for studying dependence of rotamer populations and dihedral angles on local structural features.
Only a tiny fraction of the many hundreds of known protein complexes are also of known three-dimensional structure. The experimental difficulties surrounding structure determination of complexes make methods that are able to predict structures paramount. The challenge of predicting complex structures is daunting and raises many issues that need to be addressed. To produce the best models, new prediction methods have to somehow combine partial structures with a mixed bag of experimental data, including interactions and low-resolution electron microscopy images.
Recent publications have expanded our knowledge of the major structural proteins of the human immunodeficiency virus as isolated proteins. The next challenge lies in understanding the changes in structure and the interactions of these components during assembly and maturation.
The past two years have seen the rapid development of new recognition methods for protein structure prediction. These algorithms 'thread' the sequence of one protein through the known structure of another, looking for an alignment that corresponds to an energetically favorable model structure. Because they are based on energy calculation, rather than evolutionary distance, these methods extend the possibility of structure prediction by comparative modeling to a larger class of new sequences, where similarity to known structures is recognizable by no other means. The strength of the evidence they offer should be judged by objective statistical tests, however, so as to rule out the possibility that favorable scores arise from chance factors such as similarity of length, composition, or the consideration of a large number of alternative alignments. Calculation of objective p-values by analytical means is not yet possible, but it would appear that approximate values may be obtained by simulation, as they are in gapped, global sequence alignment. We propose that the results of threading experiments should include Z-scores relative to the composition-corrected score distribution obtained for shuffled and optimally aligned sequences.
Since the discovery of the 26S proteasome, much progress has been made in determining the structure of this large dynamic protein complex. Until now, a vast amount of structural information of the proteasome has been obtained from all kinds of structure determination techniques, and the function of the protease core is well understood at atomic detail. Yet our understanding of the entire 26S proteasome structure, particularly its 19S regulatory complex, is still limited at a low-resolution blob-ology level. In this review, we highlight the recent progress made in understanding the mechanism of 20S gate opening by the proteasomal activators. We also emphasized the recent methodological advances, particularly in achieving the near atomic resolution by single particle electron cryomicroscopy, and the possible approaches that will enable more detailed structural analysis of the entire 26S proteasome.
During the past year, computational methods have been developed that use the rapidly accumulating genomic data to discover protein function. The methods rely on properties shared by functionally related proteins other than sequence or structural similarity. Instead, these 'nonhomology' methods analyze patterns such as domain fusion, conserved gene position and gene co-inheritance and coexpression to identify protein-protein relationships. The methods can identify functions for proteins that are without characterized homologs and have been applied to genome-wide predictions of protein function.
Zinc finger proteins are generally thought of as DNA-binding transcription factors; however, certain classes of zinc finger proteins, including the common C(2)H(2) zinc fingers, function as RNA-binding proteins. Recent structural studies of the C(2)H(2) zinc fingers of transcription factor IIIA (TFIIIA) and the CCCH zinc fingers of Tis11d in complex with their RNA targets have revealed new modes of zinc finger interaction with nucleic acid. The three C(2)H(2) zinc fingers of TFIIIA use two modes of RNA recognition that differ from the classical mode of DNA recognition, whereas the CCCH zinc fingers of Tis11d recognize specific AU-rich sequences through backbone atom interaction with the Watson-Crick edges of the adenine and uracil bases.
If protein structure prediction methods are to make any impact on the impending onerous task of analyzing the large numbers of unknown protein sequences generated by the ongoing genome-sequencing projects, it is vital that they make the difficult transition from computational 'gedankenexperiments' to practical software tools. This has already happened in the field of comparative modelling and is currently happening in the threading field. Unfortunately, there is little evidence of this transition happening in the field of ab initio tertiary-structure prediction.
A large number of zinc endopeptidases contain an HEXXHXXGXXH consensus motif in their catalytic site (single letter code; X is any amino acid residue). These enzymes can be grouped into four distinct families, the astacins, the adamalysins, the serralysins and the matrix metalloproteinases (matrixins). Despite a low degree of sequence similarity, their catalytic modules are topologically similar. A topology derived sequence alignment suggests that the four families form a superfamily, called the metzincins because of a perfectly superimposable methionine residue close to the zinc-binding active site. Topological similarity to the thermolysin-like enzymes indicates that these enzymes may have had a common ancestor.
In the past 18 months, two RING finger structures have been solved. They represent the first reported structures for this novel zinc-binding sequence motif. Both structures are significantly different from other zinc-binding domains, in terms of both their zinc-ligation scheme and their three-dimensional structures. The RING finger domain appears to be a convenient scaffold which can be altered to provide functional specificity in those proteins that contain the motif.
Structure-specific DNA nucleases play important roles in various DNA transactions such as DNA replication, repair and recombination. These enzymes recognize loops and branched DNA structures. Recent structural studies have provided detailed insights into the functions of these enzymes. Structures of Holliday junction resolvase revealed that nucleases are broadly diverged in the way in which they fold, however, are required to form homodimers with large basic patches of protein surfaces, which are complementary to DNA tertiary structures. Many nucleases maintain structure-specific recognition modes, which involve particular domain arrangements through conformal changes of flexible loops or have a separate DNA binding domain. Nucleases, such as FEN-1 and archaeal XPF, are bound to proliferating cell nuclear antigen through a common motif, and thereby actualize their inherent activities.
Accompanying recent advances in determining RNA secondary structure is the growing appreciation for the importance of relatively simple topological constraints, encoded at the secondary structure level, in defining the overall architecture, folding pathways, and dynamic adaptability of RNA. A new view is emerging in which tertiary interactions do not define RNA 3D structure, but rather, help select specific conformers from an already narrow, topologically pre-defined conformational distribution. Studies are providing fundamental insights into the nature of these topological constraints, how they are encoded by the RNA secondary structure, and how they interplay with other interactions, breathing new meaning to RNA secondary structure. New approaches have been developed that take advantage of topological constraints in determining RNA backbone conformation based on secondary structure, and a limited set of other, easily accessible constraints. Topological constraints are also providing a much-needed framework for rationalizing and describing RNA dynamics and structural adaptation. Finally, studies suggest that topological constraints may play important roles in steering RNA folding pathways. Here, we review recent advances in our understanding of topological constraints encoded by the RNA secondary structure.
Cell biology depends on the interactions of macromolecules, such as protein-DNA, protein-protein or protein-nucleotide interactions. GTP-binding proteins are no exception to the rule. They regulate cellular processes as diverse as protein biosynthesis and intracellular membrane trafficking. Recently, a large number of genes encoding GTP-binding proteins and the proteins that interact with these molecular switches have been cloned and expressed. The 3D structures of some of these have also been elucidated.
A great variety of protein systems have been investigated in the past year using structure-guided evolutionary strategies. On the basis of available 3D structural information, critical regions of proteins have been targeted for randomizing mutagenesis and active variants of the corresponding genes have been selected. These approaches help characterize structural and mechanistic features of proteins and have important implications for design.
Over the past decade, there has been a rapid rise in the use of three-dimensional (3D) animation to depict molecular and cellular processes. Much of the growth in molecular animation has been in the educational arena, but increasingly, 3D animation software is finding its way into research laboratories. In this review, I will discuss a number of ways in which 3d animation software can play a valuable role in visualizing and communicating macromolecular structures and dynamics. I will also consider the challenges of using animation tools within the research sphere. Copyright © 2015. Published by Elsevier Ltd.
The vast majority of membrane protein complexes of biological interest cannot be purified to homogeneity, or removed from a physiologically relevant context without loss of function. It is therefore not possible to easily determine the 3D structures of these protein complexes using X-ray crystallography or conventional cryo-electron microscopy. Newly emerging methods that combine cryo-electron tomography with 3D image classification and averaging are, however, beginning to provide unique opportunities for in situ determination of the structures of membrane protein assemblies in intact cells and nonsymmetric viruses. Here we review recent progress in this field and assess the potential of these methods to describe the conformation of membrane proteins in their native environment.
Over the past few years, evidence has accumulated that shows that circularly permuted proteins resulting from permutations in their coding genes can indeed occur naturally. In most instances, these circularly permuted amino acid sequences have been detected by sequence alignment of homologous proteins. Circular permutations may escape detection, however, when based on sequence comparisons alone, as recently illustrated by transaldolase, a member of the class I aldolase family.
Three-dimensional structures of NADH:ubiquinone oxidoreductase (or complex I) from the respiratory chain of mitochondria and bacteria have been recently studied by electron microscopy. The low-resolution structures all reveal a characteristic L shape for complex I; however, some of the differences among these structures may have important implications for the location of the functional elements of complex I, for example, the ubiquinone-binding site.
It is characteristic of eukaryotic transcription that a unique combination of multiple transcriptional regulatory proteins bound to promoter DNA specifically activate or repress downstream target genes; this is referred to as combinatorial gene regulation. Recently determined structures have revealed different modes of protein-protein interaction on the promoter DNA from near (e.g. the Runx1-CBFbeta-DNA, NFAT-Fos-Jun-DNA, GABPalpha-GABPbeta-DNA, Ets-1-Pax-5-DNA and PU.1-IRF-4-DNA complexes) and afar with DNA looping (e.g. the c-Myb-C/EBPbeta-DNA complex), and their regulatory mechanisms.
Top-cited authors
Bernard Henrissat
  • CNRS Marseille
Jeffery W Kelly
  • The Scripps Research Institute
Bostjan Kobe
  • The University of Queensland
Jane Dyson
  • The Scripps Research Institute
Ivet Bahar
  • University of Pittsburgh