Geometric Recognition as a Tool for Predicting Structures of Molecular Complexes

Summary The structures of two binary complexes (the TEM-1/BLIP complex and the trypsin/amiloride complex) were predicted prior to their experimental determination and compared to the corresponding experimental structures when these became available. In both predictions the rigid-body geometric docking algorithm ranked the correct solution among the top ones. Additional information concerning the structure and chemical character of the binding site of one of the molecules in the complex was used to single out the correct solution. The results indicate that the combination of geometric surface matching with biochemical information produces a useful tool for structure prediction.     

Ligand-induced structural changes to maltodextrin-binding protein as studied by solution NMR spectroscopy

Solution NMR studies on the physiologically relevant ligand-free and maltotriose-bound states of maltodextrin-binding protein (MBP) are presented. Together with existing data on MBP in complex with beta-cyclodextrin (non-physiological, inactive ligand), these new results provide valuable information on changes in local structure, dynamics and global fold that occur upon ligand binding to this two-domain protein. By measuring a large number of different one-bond residual dipolar couplings, the domain conformations, critical for biological function, were investigated for all three states of MBP. Structural models of the solution conformation of MBP in a number of different forms were generated from the experimental dipolar coupling data and X-ray crystal structures using a quasi-rigid-body domain orientation algorithm implemented in the structure calculation program CNS. Excellent agreement between relative domain orientations in ligand-free and maltotriose-bound solution conformations and the corresponding crystal structures is observed. These results are in contrast to those obtained for the MBP/beta-cyclodextrin complex where the solution state is found to be approximately 10 degrees more closed than the crystalline state. The present study highlights the utility of residual dipolar couplings for orienting protein domains or macromolecules with respect to each other.     

The solution structure and dynamics of the DH‐PH module of PDZRhoGEF in isolation and in complex with nucleotide‐free RhoA

The DH‐PH domain tandems of Dbl‐homology guanine nucleotide exchange factors catalyze the exchange of GTP for GDP in Rho‐family GTPases, and thus initiate a wide variety of cellular signaling cascades. Although several crystal structures of complexes of DH‐PH tandems with cognate, nucleotide free Rho GTPases are known, they provide limited information about the dynamics of the complex and it is not clear how accurately they represent the structures in solution. We used a complementary combination of nuclear magnetic resonance (NMR), small‐angle X‐ray scattering (SAXS), and hydrogen‐deuterium exchange mass spectrometry (DXMS) to study the solution structure and dynamics of the DH‐PH tandem of RhoA‐specific exchange factor PDZRhoGEF, both in isolation and in complex with nucleotide free RhoA. We show that in solution the DH‐PH tandem behaves as a rigid entity and that the mutual disposition of the DH and PH domains remains identical within experimental error to that seen in the crystal structure of the complex, thus validating the latter as an accurate model of the complex in vivo. We also show that the nucleotide‐free RhoA exhibits elevated dynamics when in complex with DH‐PH, a phenomenon not observed in the crystal structure, presumably due to the restraining effects of crystal contacts. The complex is readily and rapidly dissociated in the presence of both GDP and GTP nucleotides, with no evidence of intermediate ternary complexes.     

Analytical ultracentrifugation combined with X-ray and neutron scattering: Experiment and modelling

Analytical ultracentrifugation and solution scattering provide different multi-parameter structural and compositional information on proteins. The joint application of the two methods supplements high resolution structural studies by crystallography and NMR. We summarise the procedures required to obtain equivalent ultracentrifugation and X-ray and neutron scattering data. The constrained modelling of ultracentrifugation and scattering data is important to confirm the experimental data analysis and yields families of best-fit molecular models for comparison with crystallography and NMR structures. This modelling of ultracentrifugation and scattering data is described in terms of starting models, their conformational randomisation in trial-and-error fits, and the identification of the final best-fit models. Seven applications of these methods are described to illustrate the current state-of-the-art. These include the determination of antibody solution structures (the human IgG4 subclass, and oligomeric forms of human IgA and its secretory component), the solution structures of the complement proteins of innate immunity (Factor H and C3/C3u) and their interactions with macromolecular ligands (C-reactive protein), and anionic polysaccharides (heparin). Complementary features of joint ultracentrifugation and scattering experiments facilitate an improved understanding of crystal structures (illustrated for C3/C3u, C-reactive protein and heparin). If a large protein or its complex cannot be crystallised, the joint ultracentrifugation-scattering approach provides a means to obtain an overall macromolecular structure.     

Ligand-induced domain rearrangement of fatty acid beta-oxidation multienzyme complex

The quaternary structure of a fatty acid beta-oxidation multienzyme complex, catalyzing three sequential reactions, was investigated by X-ray crystallographic and small-angle X-ray solution scattering analyses. X-ray crystallography revealed an intermediate structure of the complex among the previously reported structures. However, the theoretical scattering curves calculated from the crystal structures remarkably disagree with the experimental profiles. Instead, an ensemble of the atomic models, which were all calculated by rigid-body optimization, reasonably explained the experimental data. These structures significantly differ from those in the crystals, but they maintain the substrate binding pocket at the domain boundary. Comparisons among these structures indicated that binding of 3-hydroxyhexadecanoyl-CoA or nicotinamide adenine dinucleotide induces domain rearrangements in the complex. The conformational changes suggest the structural events occurring during the chain reaction catalyzed by the multienzyme complex.     

Shaker pore structure as predicted by annealed atomic simulation using symmetry and novel geometric restraints

Recent studies making use of channel-blocking peptides as molecular calipers have revealed the architecture of the pore-forming region of Shaker-type potassium channels. Here we show that the low-resolution, experimentally derived geometric information can be incorporated as restraints within the context of an annealed molecular dynamics simulation to predict an atomic structure for the channel pore which, by virtue of restraints, conforms to the experimental evidence. The simulation is reminiscent of the computational method employed by nuclear magnetic resonance (NMR) spectroscopists to resolve solution structures of biological macromolecules, but in lieu of restraints conventionally derived from NMR spectra, novel restraints are developed that include side-chain orientation of amino acid residues and assumed symmetry of protein subunits. The method presented here offers the possibility of expanding cooperation between simulation and experiment in developing structural models, especially for systems such as ion channels whose three-dimensional structures may not be amenable to determination by direct methods at the present time.     

Crystal structure of uronate dehydrogenase from Agrobacterium tumefaciens

Uronate dehydrogenase from Agrobacterium tumefaciens (AtUdh) belongs to the short-chain dehydrogenase/reductase superfamily and catalyzes the oxidation of D-galacturonic acid and D-glucuronic acid with NAD(+) as a cofactor. We have determined the crystal structures of an apo-form of AtUdh, a ternary form in complex with NADH and product (substrate-soaked structure), and an inactive Y136A mutant in complex with NAD(+). The crystal structures suggest AtUdh to be a homohexamer, which has also been observed to be the major form in solution. The monomer contains a Rossmann fold, essential for nucleotide binding and a common feature of the short-chain dehydrogenase/reductase family enzymes. The ternary complex structure reveals a product, D-galactaro-1,5-lactone, which is bound above the nicotinamide ring. This product rearranges in solution to D-galactaro-1,4-lactone as verified by mass spectrometry analysis, which agrees with our previous NMR study. The crystal structure of the mutant with the catalytic residue Tyr-136 substituted with alanine shows changes in the position of Ile-74 and Ser-75. This probably altered the binding of the nicotinamide end of NAD(+), which was not visible in the electron density map. The structures presented provide novel insights into cofactor and substrate binding and the reaction mechanism of AtUdh. This information can be applied to the design of efficient microbial conversion of D-galacturonic acid-based waste materials.     

On the stability of different experimental dimeric structures of the SL1 sequence from the genomic RNA of HIV-1 in solution: a molecular dynamics simulation and electrophoresis study

SL1 is a stem-loop RNA sequence from the genome of HIV-1 thought to be the initiation site for the dimerization of the retroviral genomic RNA. The aim of this study is to check the stability in solution of different experimental dimeric structures available in the literature. Two kinds of dimer have been evidenced: an extended duplex looking like a double helix with two internal bulges and a kissing complex in which the monomers with a stem/loop conformation are linked by intermolecular loop-loop interactions. Two divergent experimental structures of the kissing complex from the Lai isolate are reported in the literature, one obtained from NMR (Mujeeb et al., Nature Structural Biology, 1998, Vol. 5, pp. 432-436) and the other one from x-ray crystallography (Ennifar et al., Nature Structural Biology, 2001, Vol. 8, pp. 1064-1068). A crystallographic structure of the Mal isolate was also reported (Ennifar et al., Nature Structure Biology, 2001, Vol. 8, pp. 1064-1068). Concerning the extended duplex, a NMR structure is available for Lai (Girard et al., Journal of Biomolecular Structure and Dynamics, 1999, Vol. 16, pp. 1145-1157) and a crystallographic structure for Mal (Ennifar et al., Structure, 1999, Vol. 7, pp. 1439-1449). Using a molecular dynamics technique, all these experimental structures have been simulated in solution with explicit water and counterions. We show that both extended duplex structures are stable. On the contrary, the crystallographic structures of the Lai and Mal kissing complexes are rapidly destabilized in aqueous environment. Finally, the NMR structure of the Lai loop-loop kissing complex remains globally stable over a 20 ns MD simulation, although large rearrangements occur at the level of the stem/loop junctions that are flexible, as shown from free energy calculations. These results are compared to electrophoresis experiments on dimer formation.     

Three-dimensional reconstruction of the actin cytoskeleton from stereo images

In eucaryotic cells, actin filaments are abundant components in the cytoskeleton where they form a complex three dimensional (3D) structural network that provides the cell with its shape and mechanical properties. However, understanding the structural and mechanical properties of actin filaments composing the cell cytoskeleton is often hampered by the inability to faithfully reconstruct the three-dimensional geometric relationships. This paper presents a vision-based reconstruction approach that automatically reconstitutes the three-dimensional structures of cytoskeletal polymers from stereo image pairs taken at the different tilt angles. The approach finds corresponding points between two images and recovers the depth information about the structures. The computational process consists of three major procedures: feature representation, stereo matching, and disparity refinement, implemented in a multi-resolution manner based on a coarse-to-fine strategy. The reconstruction depicts the three-dimensional structure of cytoskeletal polymers and their geometric relationships. New and useful information becomes available and allows quantitative analysis of the structure. Measurement of the cytoskeleton geometrical properties and the filament concentration in a defined volume are obtained by direct calculation.     

Monitoring DNA/poly-L-lysine polyplex formation with time-resolved multiangle laser light scattering

Nonviral DNA complexes show promise as alternative and attractive gene delivery vectors for treating genetic diseases. Nonviral DNA complexes are typically formed by combining DNA with various condensing/complexing agents such as lipids, polyelectrolytes, polymers, polypeptides, and surfactants in solution. DNA/poly-L-lysine polyplex formation kinetics are probed by time-resolved multiangle laser light scattering (TR-MALLS), which yields the time evolution of the supramolecular complex mass and geometric size. Primary polyplexes whose geometric size is smaller than individual DNA molecules in solution are formed very rapidly upon mixing DNA and poly-L-lysine. Over time, these primary polyplexes aggregate into larger structures whose ultimate size is determined primarily by the relative concentrations of DNA and poly-L-lysine. This final polyplex size varies with the DNA/poly-L-lysine mass ratio in a non-monotonic fashion, with the maximum polyplex size occurring at a DNA/poly-L-lysine mass ratio of approximately two to three (charge ratio near unity). The utility of TR-MALLS for monitoring the temporal evolution of DNA loading and supramolecular complex size growth (mean square radius and molar mass) throughout the DNA/poly-L-lysine polyplex formation process is demonstrated. The polyplex DNA loading and size, both geometric and molar mass, are key to understanding the transfection process and for developing optimal gene therapy vectors.     

A geometric arrangement algorithm for structure determination of symmetric protein homo-oligomers from NOEs and RDCs

Nuclear magnetic resonance (NMR) spectroscopy is a primary tool to perform structural studies of proteins in physiologically-relevant solution conditions. Restraints on distances between pairs of nuclei in the protein, derived from the nuclear Overhauser effect (NOE), provide information about the structure of the protein in its folded state. NMR studies of symmetric protein homo-oligomers present a unique challenge. Using X-filtered NOESY experiments, it is possible to determine whether an NOE restrains a pair of protons across different subunits or within a single subunit, but current experimental techniques are unable to determine in which subunits the restrained protons lie. Consequently, it is difficult to assign NOEs to particular pairs of subunits with certainty, thus hindering the structural analysis of the oligomeric state. Computational approaches are needed to address this subunit ambiguity, but traditional solutions often rely on stochastic search coupled with simulated annealing and simulations of simplified molecular dynamics, which have many tunable parameters that must be chosen carefully and can also fail to report structures consistent with the experimental restraints. In addition, these traditional approaches rarely provide guarantees on running time or solution quality. We reduce the structure determination of homo-oligomers with cyclic symmetry to computing geometric arrangements of unions of annuli in a plane. Our algorithm, disco, runs in expected O(n2) time, where n is the number of distance restraints, potentially assigned ambiguously. disco is guaranteed to report the exact set of oligomer structures consistent with the distance restraints and also with orientational restraints from residual dipolar couplings (RDCs). We demonstrate our method using two symmetric protein complexes: the trimeric E. coli diacylglycerol kinase (DAGK) and a dimeric mutant of the immunoglobulin-binding domain B1 of streptococcal protein G (GB1). In both cases, disco computes oligomer structures with high precision and also finds distance restraints that are either mutually inconsistent or inconsistent with the RDCs. The entire protocol DISCO has been completely automated in a software package that is freely available and open-source at www.cs.duke.edu/donaldlab/software.php.     

Get5 carboxyl-terminal domain is a novel dimerization motif that tethers an extended Get4/Get5 complex

Tail-anchored trans-membrane proteins are targeted to membranes post-translationally. The proteins Get4 and Get5 form an obligate complex that catalyzes the transfer of tail-anchored proteins destined to the endoplasmic reticulum from Sgt2 to the cytosolic targeting factor Get3. Get5 forms a homodimer mediated by its carboxyl domain. We show here that a conserved motif exists within the carboxyl domain. A high resolution crystal structure and solution NMR structures of this motif reveal a novel and stable helical dimerization domain. We additionally determined a solution NMR structure of a divergent fungal homolog, and comparison of these structures allows annotation of specific stabilizing interactions. Using solution x-ray scattering and the structures of all folded domains, we present a model of the full-length Get4/Get5 complex.     

Assembly of a polytopic membrane protein structure from the solution structures of overlapping peptide fragments of bacteriorhodopsin.

Three-dimensional structures of only a handful of membrane proteins have been solved, in contrast to the thousands of structures of water-soluble proteins. Difficulties in crystallization have inhibited the determination of the three-dimensional structure of membrane proteins by x-ray crystallography and have spotlighted the critical need for alternative approaches to membrane protein structure. A new approach to the three-dimensional structure of membrane proteins has been developed and tested on the integral membrane protein, bacteriorhodopsin, the crystal structure of which had previously been determined. An overlapping series of 13 peptides, spanning the entire sequence of bacteriorhodopsin, was synthesized, and the structures of these peptides were determined by NMR in dimethylsulfoxide solution. These structures were assembled into a three-dimensional construct by superimposing the overlapping sequences at the ends of each peptide. Onto this construct were written all the distance and angle constraints obtained from the individual solution structures along with a limited number of experimental inter-helical distance constraints, and the construct was subjected to simulated annealing. A three-dimensional structure, determined exclusively by the experimental constraints, emerged that was similar to the crystal structure of this protein. This result suggests an alternative approach to the acquisition of structural information for membrane proteins consisting of helical bundles.     

Morphology of the procarboxypeptidase A-S6 complex

Bovine pancreatic procarboxypeptidase A is secreted as a non-covalent association of three different proteins (pro CPA-S6). The free native subunits can be obtained by dissociation of the complex by dimethylmaleylation. Moreover, two specific binary complexes resulting from the high affinity of procarboxypeptidase A (subunit I) for its other two partners (subunits II and III) can also be obtained.In order to better understand the function of the association, an investigation of the morphology of the ternary complex by solution X-ray scattering has been carried out. The radii of gyration of all the molecular species have been obtained and the experimental results have been interpreted in terms of compact objects of simple shape. The various components correspond to globular particles as shown by the value of the ratio Rg/M^1/3. This is confirmed by the moderate anisotropy of the simple geometric shapes determined using an assumed value of 0.3 g H₂O/g protein for the hydration. The distances between the centres of gravity of pairs of species strongly suggest that the components are in the closest distance configuration or close to it. However, the binary complex I–III appears to be more open than the complex I–II. Finally, a model of the interaction between carboxpeptidase A and its activation peptide has been constructed by comparing the hypothetical geometric model of subunit I to the crystallographically determined structure of carboxypeptidase A.Abbreviations pro CPA procarboxypeptidase A - pro CPA-S6 (or T.C.) ternary complex with a sedimentation coefficient of 6S - CPA carboxypeptidase A     

Molecular dynamics and binding specificity analysis of the bovine immunodeficiency virus BIV Tat-TAR complex

We have performed molecular dynamics (MD) simulations, with particle-mesh Ewald, explicit waters, and counterions, and binding specificity analyses using combined molecular mechanics and continuum solvent (MM-PBSA) on the bovine immunodeficiency virus (BIV) Tat peptide-TAR RNA complex. The solution structure for the complex was solved independently by Patel and co-workers and Puglisi and co-workers. We investigated the differences in both structures and trajectories, particularly in the formation of the U-A-U base triple, the dynamic flexibility of the Tat peptide, and the interactions at the binding interface. We observed a decrease in RMSD in comparing the final average RNA structures and initial RNA structures of both trajectories, which suggests the convergence of the RNA structures to a MD equilibrated RNA structure. We also calculated the relative binding of different Tat peptide mutants to TAR RNA and found qualitative agreement with experimental studies.     

Progress in computation and amide hydrogen exchange for prediction of protein-protein complexes

The macromolecular docking problem that must be solved for experimental biologists is prediction of the structures of complexes for which the components are known or reliably modeled in the unbound state, but the structure of the complex is unknown. The current state of the art in macromolecular docking is such that solving this problem usually requires supplementary experimental chemical and/or biological information to evaluate computational predictions. Amide (1)H/(2)H exchange measured by mass spectroscopy is a promising approach for obtaining such information, because it can reveal interfacial regions of each member of the complex and identify regions of conformational flexibility in the structure. In a previous article (Anand et al., Proc Natl Acad Sci USA 2003;100:13264-13269), we used (1)H/(2)H exchange data to predict the structure of a complex between regulatory and catalytic subunits of protein kinase A. Comparison of the prediction with a recent crystal structure determination (Kim et al., Science 2005;307:690-696) showed large conformational change in the regulatory subunit on formation of the complex. Analysis of the prediction, previous CAPRI results, novel data processing methods for the (1)H/(2)H exchange data, and new fragment docking computations give grounds for cautious optimism that this method can be useful even in cases of substantial conformational change.     

Pictures in cell biology. Structures of nuclear-transport components.

The past three years have seen the solution of several nuclear transport component structures and recently of the structure of a regulator bound to part of a nuclear pore complex (NPC) protein. These structures have provided a wealth of valuable information about the proteins involved and suggested strategies for further investigation of their properties. We do not have space here to go into detail about this information, so instead we are illustrating the structures and providing primary references enabling interested readers to find further information. On this page, we are concentrating on the GTPase Ran and proteins that modulate its activity, and on the facing page are the other transport factors, some of which also interact directly with Ran. Notably absent at the moment are the nuclear pore complex component s, apart from one domain of RanBP2. Only when theses are characterized fully will we really be able to understand how transport substrates move across the nuclear envelope.