Using Integrative Modeling Platform to compute,validate, and archive a model of a protein complex structure

Biology is advanced by producing structural models of biological systems, such as protein complexes. Some systems are recalcitrant to traditional structure determination methods. In such cases, it may still be possible to produce useful models by integrative structure determination that depends on simultaneous use of multiple types of data. An ensemble of models that are sufficiently consistent with the data is produced by a structural sampling method guided by a data‐dependent scoring function. The variation in the ensemble of models quantified the uncertainty of the structure, generally resulting from the uncertainty in the input information and actual structural heterogeneity in the samples used to produce the data. Here, we describe how to generate, assess, and interpret ensembles of integrative structural models using our open source Integrative Modeling Platform program ( https://integrativemodeling.org ).     

Modeling conformational states of proteins with AlphaFold

Many proteins exert their function by switching among different structures. Knowing the conformational ensembles affiliated with these states is critical to elucidate key mechanistic aspects that govern protein function. While experimental determination efforts are still bottlenecked by cost, time, and technical challenges, the machine-learning technology AlphaFold showed near experimental accuracy in predicting the three-dimensional structure of monomeric proteins. However, an AlphaFold ensemble of models usually represents a single conformational state with minimal structural heterogeneity. Consequently, several pipelines have been proposed to either expand the structural breadth of an ensemble or bias the prediction toward a desired conformational state. Here, we analyze how those pipelines work, what they can and cannot predict, and future directions.     

Accurate structural correlations from maximum likelihood superpositions

The cores of globular proteins are densely packed, resulting in complicated networks of structural interactions. These interactions in turn give rise to dynamic structural correlations over a wide range of time scales. Accurate analysis of these complex correlations is crucial for understanding biomolecular mechanisms and for relating structure to function. Here we report a highly accurate technique for inferring the major modes of structural correlation in macromolecules using likelihood-based statistical analysis of sets of structures. This method is generally applicable to any ensemble of related molecules, including families of nuclear magnetic resonance (NMR) models, different crystal forms of a protein, and structural alignments of homologous proteins, as well as molecular dynamics trajectories. Dominant modes of structural correlation are determined using principal components analysis (PCA) of the maximum likelihood estimate of the correlation matrix. The correlations we identify are inherently independent of the statistical uncertainty and dynamic heterogeneity associated with the structural coordinates. We additionally present an easily interpretable method (“PCA plots”) for displaying these positional correlations by color-coding them onto a macromolecular structure. Maximum likelihood PCA of structural superpositions, and the structural PCA plots that illustrate the results, will facilitate the accurate determination of dynamic structural correlations analyzed in diverse fields of structural biology.     

A method for studying the structure of uniaxially aligned biopolymers using solid state 15N-nmr: Application to Bombyx mori silk fibroin fibers

Recent advances in the application of solid state nmr spectroscopy to uniformly aligned biopolymers have opened a window through which to view the detailed structure of biological macromolecules that are unable to be seen with standard techniques for structure determination such as x-ray diffraction. Atomic resolution structural details are obtained from solid state nmr data in the form of bond orientations, which yield the relative positions of specific atoms within the molecule. For static aligned systems such as fibers, in which rapid reorientation about the axis of alignment does not occur, it has generally been necessary to perform trial and error line-shape simulations to extract structural details from nmr spectra arising from a single type of nuclear spin interaction. In the present work, a new method is developed in which solid state ¹⁵N-nmr spectra obtained from uniaxially aligned molecules placed with the axis of alignment both parallel and perpendicular to the magnetic field are analyzed to yield the orientations of specific molecular bonds. Analytical expressions are derived that utilize spectral features read from 15N chemical shift anisotropy line shapes to calculate a discrete number of possible orientations for a specific site. The ¹⁵N-¹H dipolar interaction is employed to further narrow the number of unique orientations possible for a given site. With this method, a neighborhood of possible orientations is quickly determined, and full line-shape simulations within this region of allowed space can be performed to refine the limits of orientation. This technique demonstrates the use of a single type of isotopic label to determine the orientation of a specific molecular group such as a peptide plane within a protein. Results from the application of this method to the Bombyx mori silk fibroin protein provide structural detail that is consistent with currently accepted structural models based on fiber diffraction studies. © 1993 John Wiley & Sons, Inc.     

The Ligand-Free State of the TPP Riboswitch: A Partially Folded RNA Structure

Riboswitches are elements of mRNA that regulate gene expression by undergoing structural changes upon binding of small ligands. Although the structures of several riboswitches have been solved with their ligands bound, the ligand-free states of only a few riboswitches have been characterized. The ligand-free state is as important for the functionality of the riboswitch as the ligand-bound form, but the ligand-free state is often a partially folded structure of the RNA, with conformational heterogeneity that makes it particularly challenging to study. Here, we present models of the ligand-free state of a thiamine pyrophosphate riboswitch that are derived from a combination of complementary experimental and computational modeling approaches. We obtain a global picture of the molecule using small-angle X-ray scattering data and use an RNA structure modeling software, MC-Sym, to fit local structural details to these data on an atomic scale. We have used two different approaches to obtaining these models. Our first approach develops a model of the RNA from the structures of its constituent junction fragments in isolation. The second approach treats the RNA as a single entity, without bias from the structure of its individual constituents. We find that both approaches give similar models for the ligand-free form, but the ligand-bound models differ for the two approaches, and only the models from the second approach agree with the ligand-bound structure known previously from X-ray crystallography. Our models provide a picture of the conformational changes that may occur in the riboswitch upon binding of its ligand. Our results also demonstrate the power of combining experimental small-angle X-ray scattering data with theoretical structure prediction tools in the determination of RNA structures beyond riboswitches.     

Structure of the coat protein in Pf1 bacteriophage determined by solid-state NMR spectroscopy

The atomic resolution structure of Pf1 coat protein determined by solid-state NMR spectroscopy of magnetically aligned filamentous bacteriophage particles in solution is compared to the structures previously determined by X-ray fiber and neutron diffraction, the structure of its membrane-bound form, and the structure of fd coat protein. These structural comparisons provide insights into several biological properties, differences between class I and class II filamentous bacteriophages, and the assembly process. The six N-terminal amino acid residues adopt an unusual "double hook" conformation on the outside of the bacteriophage particle. The solid-state NMR results indicate that at 30 degrees C, some of the coat protein subunits assume a single, fully structured conformation, and some have a few mobile residues that provide a break between two helical segments, in agreement with structural models from X-ray fiber and neutron diffraction, respectively. The atomic resolution structure determined by solid-state NMR for residues 7-14 and 18-46, which excludes the N-terminal double hook and the break between the helical segments, but encompasses more than 80% of the backbone including the distinct kink at residue 29, agrees with that determined by X-ray fiber diffraction with an RMSD value of 2.0 A. The symmetry and distance constraints determined by X-ray fiber and neutron diffraction enable the construction of an accurate model of the bacteriophage particle from the coordinates of the coat protein monomers.     

X-ray diffraction and time-resolved fluorescence analyses of Aequorea green fluorescent protein crystals

The energy transfer protein, green fluorescent protein, from the hydromedusan jellyfish Aequorea victoria has been crystallized in two morphologies suitable for x-ray diffraction analysis. Hexagonal plates have been obtained in the P6122 or P6522 space group with a = b = 77.5, c = 370 A, and no more than three molecules per asymmetric unit. Monoclinic parallel-epipeds have been obtained in the C2 space group with a = 93.3, b = 66.5, c = 45.5 A, beta = 108 degrees, and one molecule per asymmetric unit. The monoclinic form is better suited for use in a structure determination, and a data set was collected from the native crystal. Time-resolved fluorescence measurements of large single crystals are possible due to the unique, covalently bound chromophore present in this molecule. Fluorescence emission spectra of Aequorea green fluorescent protein in solution and from either the hexagonal or monoclinic single crystal show similar profiles suggesting that the conformations of protein in solution and in the crystal are similar. Multifrequency phase fluorimetric data obtained from a single crystal were best fit by a single fluorescence lifetime very close to that exhibited by the protein in solution. The complementary structural data obtained from fluorescence spectroscopy and x-ray diffraction crystallography will aid in the elucidation of this novel protein's structure-function relationship.     

B-DNA structure is intrinsically polymorphic: even at the level of base pair positions

Increasingly exact measurement of single crystal X-ray diffraction data offers detailed characterization of DNA conformation, hydration and electrostatics. However, instead of providing a more clear and unambiguous image of DNA, highly accurate diffraction data reveal polymorphism of the DNA atomic positions and conformation and hydration. Here we describe an accurate X-ray structure of B-DNA, painstakingly fit to a multistate model that contains multiple competing positions of most of the backbone and of entire base pairs. Two of ten base-pairs of CCAGGCCTGG are in multiple states distinguished primarily by differences in slide. Similarly, all the surrounding ions are seen to fractionally occupy discrete competing and overlapping sites. And finally, the vast majority of water molecules show strong evidence of multiple competing sites. Conventional resolution appears to give a false sense of homogeneity in conformation and interactions of DNA. In addition, conventional resolution yields an average structure that is not accurate, in that it is different from any of the multiple discrete structures observed at high resolution. Because base pair positional heterogeneity has not always been incorporated into model-building, even some high and ultrahigh-resolution structures of DNA do not indicate the full extent of conformational polymorphism.     

Protein structure determination in solution by NMR spectroscopy

The introduction of nuclear magnetic resonance (NMR) spectroscopy as a second method for protein structure determination at atomic resolution, in addition to x-ray diffraction in single crystals, has already led to a significant increase in the number of known protein structures. The NMR method provides data that are in many ways complementary to those obtained from x-ray crystallography and thus promises to widen our view of protein molecules, giving a clearer insight into the relation between structure and function.     

Structural studies of chromatin remodeling factors

Changes of chromatin structure require participation of chromatin remodeling factors (CRFs), which are ATP-dependent multisubunit complexes that change the structure of the nucleosome without covalently modifying its components. CRFs act together with other protein factors to regulate the extent of chromatin condensation. Four CRF families are currently distinguished based on their structural and biochemical characteristics: SWI/SNF, ISWI, Mi-2/CHD, and SWR/INO80. X-ray diffraction analysis and electron microscopy are the main methods to obtain structural information about macromolecules. CRFs are difficult to obtain in crystal because of their large sizes and structural heterogeneity, and transmission electron microscopy (TEM) is mostly employed in their structural studies. The review considers all structures obtained for CRFs by TEM and discusses several models of CRF–nucleosome interactions.     

IPET and FETR: experimental approach for studying molecular structure dynamics by cryo-electron tomography of a single-molecule structure

The dynamic personalities and structural heterogeneity of proteins are essential for proper functioning. Structural determination of dynamic/heterogeneous proteins is limited by conventional approaches of X-ray and electron microscopy (EM) of single-particle reconstruction that require an average from thousands to millions different molecules. Cryo-electron tomography (cryoET) is an approach to determine three-dimensional (3D) reconstruction of a single and unique biological object such as bacteria and cells, by imaging the object from a series of tilting angles. However, cconventional reconstruction methods use large-size whole-micrographs that are limited by reconstruction resolution (lower than 20 Å), especially for small and low-symmetric molecule (<400 kDa). In this study, we demonstrated the adverse effects from image distortion and the measuring tilt-errors (including tilt-axis and tilt-angle errors) both play a major role in limiting the reconstruction resolution. Therefore, we developed a “focused electron tomography reconstruction” (FETR) algorithm to improve the resolution by decreasing the reconstructing image size so that it contains only a single-instance protein. FETR can tolerate certain levels of image-distortion and measuring tilt-errors, and can also precisely determine the translational parameters via an iterative refinement process that contains a series of automatically generated dynamic filters and masks. To describe this method, a set of simulated cryoET images was employed; to validate this approach, the real experimental images from negative-staining and cryoET were used. Since this approach can obtain the structure of a single-instance molecule/particle, we named it individual-particle electron tomography (IPET) as a new robust strategy/approach that does not require a pre-given initial model, class averaging of multiple molecules or an extended ordered lattice, but can tolerate small tilt-errors for high-resolution single “snapshot” molecule structure determination. Thus, FETR/IPET provides a completely new opportunity for a single-molecule structure determination, and could be used to study the dynamic character and equilibrium fluctuation of macromolecules.     

Solvent structure in crystals of trypsin determined by X-ray and neutron diffraction.

The solvent structure in orthorhombic crystals of bovine trypsin has been independently determined by X-ray diffraction to 1.35 A resolution and by neutron diffraction to 2.1 A resolution. A consensus model of the water molecule positions was obtained using oxygen positions identified in the electron density map determined by X-ray diffraction, which were verified by comparison to D2O-H2O difference neutron scattering density. Six of 184 water molecules in the X-ray structure, all with B-factors greater than 50 A2, were found to be spurious after comparison with neutron results. Roughly two-thirds of the water of hydration expected from thermodynamic data for proteins was localized by neutron diffraction; approximately one-half of the water of hydration was located by X-ray diffraction. Polar regions of the protein are well hydrated, and significant D2O-H2O difference density is seen for a small number of water molecules in a second shell of hydration. Hydrogen bond lengths and angles calculated from unconstrained refinement of water positions are distributed about values typically seen in small molecule structures. Solvent models found in seven other bovine trypsin and trypsinogen and rat trypsin structures determined by X-ray diffraction were compared. Internal water molecules are well conserved in all trypsin structures including anionic rat trypsin, which is 65% homologous to bovine trypsin. Of the 22 conserved waters in trypsin, 19 were also found in trypsinogen, suggesting that they are located in regions of the apoprotein that are structurally conserved in the transition to the mature protein. Seven waters were displaced upon activation of trypsinogen. Water structure at crystal contacts is not generally conserved in different crystal forms. Three groups of integral structural water molecules are highly conserved in all solvent structures, including a spline of water molecules inserted between two beta-strands, which may resemble an intermediate in the formation of beta sheets during the folding of a protein.     

Radial distributions of water—water distances in protein crystals

An accurate protein crystallographic structure determination requires a knowledge of the solvent contribution to the diffraction pattern. As resolutions improve, research groups are reporting coordinates of large numbers of water molecules. We examine the accuracy of these coordinates by presenting radial distributions of water—water distances from refinements at different stages and interpreting them in terms of preferred hydrogen-bonding distances and problems in solvent electron density map interpretation. Marked differences between the distributions suggest that wide variations exist in the water molecule selection and refinement criteria employed by different research groups which mask possible real differences in solvent structure.     

Structural variants of human T200 glycoprotein (leukocyte-common antigen).

Structural variation in the primary structure of human T200 glycoprotein has been detected. Three cDNA variants have been characterized each of which encode T200 molecules that differ in size as a result of sequence differences in their amino-terminal regions. The largest form of the molecule is distinguished from the smallest by an insert of 161 amino acids, after the first eight amino-terminal residues. The other variant has an insert at the same location of 47 amino acids identical to residues 75-121 in the larger insert. Both extra domains are rich in serine and threonine residues and are likely to display multiple O-linked oligosaccharides. These structural variants which probably arise by cell-type-specific alternative splicing provide a molecular basis for the previously observed structural and antigenic heterogeneity of T200 glycoprotein. In addition to the variable amino-terminal region, the external domain of human T200 glycoprotein consists of a second cysteine-rich region of about 400 amino acids, a single transmembrane-spanning region and a large cytoplasmic domain of 707 amino acids shared by all of the structural variants and highly conserved between species. The gene encoding human T200 is located on the long arm of chromosome 1.     

Detection and separation of heterogeneity in molecular complexes by statistical analysis of their two-dimensional projections

Progress in molecular structure determination by cryo electron microscopy and single particle analysis has led to improvements in the resolution achievable. However, in many cases the limiting factor is structural heterogeneity of the sample. To address this problem, we have developed a method based on statistical analysis of the two-dimensional images to detect and sort localised structural variations caused, for example, by variable occupancy of a ligand. Images are sorted by two consecutive stages of multivariate statistical analysis (MSA) to dissect out the two main sources of variation, namely out of plane orientation and local structural changes. Heterogeneity caused by local changes is detected by MSA that reveals significant peaks in the higher order eigenimages. The eigenimages revealing local peaks are used for automated classification. Evaluation of differences between classes allows discrimination of molecular images with and without ligand. This method is very rapid, independent of any initial three-dimensional model, and can detect even minor subpopulations in an image ensemble. A strategy for using this technique was developed on model data sets. Here, we demonstrate the successful application of this method to both model and real EM data on chaperonin-substrate and ribosome-ligand complexes.     

Protein structure refinement based on paramagnetic NMR shifts: applications to wild-type and mutant forms of cytochrome c.

A new approach to NMR solution structure refinement is introduced that uses paramagnetic effects on nuclear chemical shifts as constraints in energy minimization or molecular dynamics calculations. Chemical shift differences between oxidized and reduced forms of horse cytochrome c for more than 300 protons were used as constraints to refine the structure of the wild-type protein in solution and to define the structural changes induced by a Leu 94 to Val mutation. A single round of constrained minimization, using the crystal structure as the starting point, converged to a low-energy structure with an RMS deviation between calculated and observed pseudo-contact shifts of 0.045 ppm, 7.5-fold lower than the starting structure. At the same time, the procedure provided stereospecific assignments for more than 45 pairs of methylene protons and methyl groups. Structural changes caused by the mutation were determined to a precision of better than 0.3 A. Structure determination based on dipolar paramagnetic (pseudocontact) shifts is applicable to molecules containing anisotropic paramagnetic centers with short electronic relaxation times, including numerous naturally occurring metalloproteins, as well as proteins or nucleic acids to which a paramagnetic metal ion or ligand may be attached. The long range of paramagnetic shift effects (up to 20 A from the iron in the case of cytochrome c) provides global structural constraints, which, in conjunction with conventional NMR distance and dihedral angle constraints, will enhance the precision of NMR solution structure determination.     

Bicelles: A natural ‘molecular goniometer’ for structural,dynamical and topological studies of molecules in membranes

Major biological processes occur at the biological membrane. One of the great challenges is to understand the function of chemical or biological molecules inside the membrane; as well of those involved in membrane trafficking. This requires obtaining a complete picture of the in situ structure and dynamics as well as the topology and orientation of these molecules in the membrane lipid bilayer. These led to the creation of several innovative models of biological membranes in order to investigate the structure and dynamics of amphiphilic molecules, as well as integral membrane proteins having single or multiple transmembrane segments. Because the determination of the structure, dynamics and topology of molecules in membranes requires a macroscopic alignment of the system, a new membrane model called ‘bicelles’ that represents a crossover between lipid vesicles and classical micelles has become very popular due to its property of spontaneous self-orientation in magnetic fields. In addition, crucial factors involved in mimicking natural membranes, such as sample hydration, pH and salinity limits, are easy to control in bicelle systems. Bicelles are composed of mixtures of long chain (14–18 carbons) and short chain phospholipids (6–8 carbons) hydrated up to 98% with buffers and may adopt various morphologies depending on lipid composition, temperature and hydration. We have been developing bicelle systems under the form of nano-discs made of lipids with saturated or biphenyl-containing fatty acyl chains. Depending on the lipid nature, these membranous nano-discs may be macroscopically oriented with their normal perpendicular or parallel to the magnetic field, providing a natural ‘molecular goniometer’ for structural and topological studies, especially in the field of NMR. Bicelles can also be spun at the magic angle and lead to the 3D structural determination of molecules in membranes.