Integrating Molecular Models Into CryoEM Heterogeneity Analysis Using Scalable High-resolution Deep Gaussian Mixture Models

Structural flexibility and/or dynamic interactions with other molecules is a critical aspect of protein function. CryoEM provides direct visualization of individual macromolecules sampling different conformational and compositional states. While numerous methods are available for computational classification of discrete states, characterization of continuous conformational changes or large numbers of discrete state without human supervision remains challenging. Here we present e2gmm, a machine learning algorithm to determine a conformational landscape for proteins or complexes using a 3-D Gaussian mixture model mapped onto 2-D particle images in known orientations. Using a deep neural network architecture, e2gmm can automatically resolve the structural heterogeneity within the protein complex and map particles onto a small latent space describing conformational and compositional changes. This system presents a more intuitive and flexible representation than other manifold methods currently in use. We demonstrate this method on both simulated data as well as three biological systems, to explore compositional and conformational changes at a range of scales. The software is distributed as part of EMAN2.

In the last decade, various coarse-grained elastic network models have been developed to study the large-scale motions of proteins and protein complexes where computer simulations using detailed all-atom models are not feasible. Among these models, the Gaussian Network Model (GNM) and Anisotropic Network Model (ANM) have been widely used. Both models have strengths and limitations. GNM can predict the relative magnitudes of the fluctuations well, but due to its isotropic assumption, it can not be applied to predict the directions of the fluctuations. In contrast, ANM adds the ability to do the latter, but it loses a significant amount of precision in the prediction of the magnitudes. In this article, we develop a generalized spring tensor model (STeM) that is able to predict well both the magnitudes and the directions of the fluctuations. STeM also outperforms ANM in explaining protein conformation changes. All of these are accomplished without sacrificing the essential features that have made ANM and GNM attractive.

We investigated both conformational changes in the junctional foot protein (JFP) and Ca2+ release from sarcoplasmic reticulum (SR) in parallel after stimulation of triadic vesicles by the JFP-specific ligand, polylysine. To monitor protein conformational change, the JFP was labeled in a site-directed fashion with the fluorescent conformational probe methylcoumarin acetate (MCA) (Kang, J. J., Tarcsafalvi, A., Carlos, A. D., Fujimoto, E., Shahrokh, Z., Thevenin, B. J.-M., Shohet, S. B., and Ikemoto, N. (1992) Biochemistry 31, 3288-3293). The induction of SR Ca2+ release by polylysine produced a rapid increase in the fluorescence intensity of the JFP-bound MCA. The polylysine concentration dependence of the fluorescence change was essentially the same as that of Ca2+ release, suggesting that the two events are tightly coupled. However, the rate constant of MCA fluorescence change was much larger than that of Ca2+ release; i.e. the conformational change preceded Ca2+ release. Prevention of protein conformational change by lysine (0.2 M) inhibited Ca2+ release from SR. Inhibition of Ca2+ release by Mg2+ (5 mM), however, had little effect on the conformational change. These results suggest that binding of polylysine to the JFP produces conformational changes in the protein, which in turn activates the Ca2+ channel, leading to Ca2+ release from the SR.

Many large protein complexes undergo extensive conformational changes as part of their functionality. Tracing these changes is important for understanding the way these proteins function. It is not always possible to obtain a high resolution structure for very large complexes. Electron cryo-microscopy (Cryo-EM) enables the representation of large macromolecular structures at a medium resolution level (4–8A). Traditional conformational search methods cannot be applied to medium resolution data where structural information may be partial or missing. Additionally, simulating large scale conformational changes in proteins require a massive amount of computational efforts. We apply a search method from robotics to structural information obtained from medium resolution Cryo-EM maps, modeled to approximate backbone trace level. The pathways obtained by this method can be useful in understanding protein motions, providing reliable results for the medium resolution data. To provide a baseline validation for our method, we tested it on Adenylate Kinase and Cyanovirin. To test the data on actual cryo-EM determined structures, we simulated the conformational opening of the GroEL single ring complex. We show that we can produce low energy conformational pathways which correspond to known structural data. The method presented here is a promising step towards exploring the conformational motion of even larger complexes.

It was found that the variety of function-related conformational changes ("movements") in proteins is beyond the earlier simple classifications. Here we offer biochemists a more comprehensive, transparent, and easy-to-use approach allowing a detailed and accurate interpretation of such conformational changes. It makes possible a more multifaceted characterization of protein flexibility via identification of rigidly and nonrigidly repositioned fragments, stable and nonstable fragments, and domain and nondomain repositioning. "Coordinate uncertainty thresholds" derived from computed differences between independently determined coordinates of the same molecules are used as the criteria for conformational identity. "Identical" rigid substructures are localized in the distance difference matrices (DDMs). A sequence of simple transformations determines whether a structural change occurs by rigid-body movements of fragments or largely through non-rigid-body deformations. We estimate the stability of protein fragments and compare stable and rigidly moving fragments. The motions computed with the coarse-grained elastic networks are also compared to those of their DDM analogues. We study and suggest a classification for 17 structural pairs, differing in their functional states. For five of the 17 proteins, conformational change cannot be accomplished by rigid-body transformations and requires significant non-rigid-body deformations. Stable fragments rarely coincide with rigidly moving fragments and often disagree with the CATH identifications of domains. Almost all monomeric apo chains, containing stable fragments and/or domains, indicate instability of the entire molecule, suggesting the importance of fragments and domains motions prior to stabilization by substrate binding or crystallization. Notably, kinases exhibit the greatest extent of nonrigidity among the proteins investigated.

Analysis of ligand-induced structural changes in proteins is challenging due to the lack of experimental methods suited for detection and characterisation of both ligand binding and induced structural changes. We have explored biosensors with different detection principles to study interactions between ligands and acetylcholine binding proteins (AChBPs), soluble homologues of Cys-loop ligand gated ion channels (LGICs) that undergo similar structural changes as LGICs upon ligand binding. X-ray crystallography was used to identify binding sites and establish if the detected conformational changes involved small changes in loop C or major structural changes in the pentamer associated with ion channel opening. Experiments were initially focused on ligands exhibiting complex surface plasmon resonance (SPR) biosensor sensorgrams or detected by second harmonic generation (SHG) biosensor analysis. Surface acoustic wave (SAW) and SHG biosensors confirmed that complexities in SPR data were indeed due to ligand-induced conformational changes. Grating coupled interferometry (GCI) biosensor sensorgrams were less complex, despite similar detection principles. switchSENSE biosensor analysis revealed that ligands resulted in either a compaction or expansion of the protein structure. X-ray crystallography of the protein-ligand complexes was only successful for 7 out of 12 ligands, despite nM-μM affinities. Crystals were not obtained for the two compounds shown by SHG analysis to induce large structural changes, while electron densities were not seen in the structures for some ligands. The work presented herein shows that several biosensor technologies have a unique capability to detect and discriminate binding and ligand induced conformational changes in proteins, also when interactions are rapid, weak and structural changes are small. However, they are complementary and provide different information.

A global exploration of the complex interplay between protein conformational changes and phosphorylation events in cell-cell interactions is crucial for understanding the dynamic nature of protein modifications. This understanding is essential for developing novel targeted therapies for pancreatic cancer, particularly in the context of interactions between pancreatic cancer cells (PCCs) and pancreatic stellate cells (PSCs). However, protein conformational changes that occur during PCCs-PSCs interactions remain poorly studied, and the relationship between these changes and phosphorylation is not well understood. Here, we present a comprehensive mass spectrometry-based study investigating the interplay between protein conformational alterations and phosphorylation in a coculture system of pancreatic ductal adenocarcinoma cells (PANC-1) and PSCs. Our results demonstrate that 435 proteins exhibiting conformational changes were detected during the coculture of PANC-1 with PSCs, primarily involving proteins associated with the tricarboxylic acid (TCA) cycle, glycolysis/gluconeogenesis, and carbon metabolism. Our findings also highlight a potential association between phosphorylation and protein conformational changes. Moreover, we identified five potential conformational targets, including ACLY, ACO1, ACO2, IDH1, and OGDH, which may provide valuable insights into the molecular pathways underlying gemcitabine resistance in pancreatic cancer. Overall, these results offer insights into protein conformational changes and their potential link to phosphorylation in the context of cancer-stromal cell interactions, paving the way for structure-based, targeted therapeutic strategies for pancreatic cancer treatment.

Changes in the bioactivity of a protein after being adsorbed on a material surface may result from conformational changes of the protein. Unfortunately, however, direct evidence of such conformational changes of proteins adsorbed on a flat material surface is sparse so far. This is because probing the conformation of an adsorbed protein on material surfaces, especially flat ones, remains a challenge due to considerable experimental difficulties. In this study, the surface-enhanced Raman scattering (SERS) technique is used to characterize the conformational changes of a protein (lysozyme) adsorbed on tailored flat gold substrates with different chemistries. Two such substrates are formed by self-assembly of octadecanethiol and thiolated PEG on gold chips (Au-C18 and Au-PEG). Preliminary results reveal that, compared to the hydrophobic Au-C18 surface, the hydrophilic Au-PEG surface has much smaller effect on the conformation of lysozyme in aqueous solution, which thereby keeps its high bioactivity. The conformational changes of lysozyme adsorbed on material surfaces with different chemistries are well correlated with changes in its bioactivity.

Characterizing conformation changes in proteins through the torsional elastic response

Understanding the dynamics of large-scale conformational changes in proteins still poses a challenge for molecular simulations. We employ transition path sampling of explicit solvent molecular dynamics trajectories to obtain atomistic insight in the reaction network of the millisecond timescale partial unfolding transition in the photocycle of the bacterial sensor photoactive yellow protein. Likelihood maximization analysis predicts the best model for the reaction coordinates of each substep as well as tentative transition states, without further simulation. We find that the unfolding of the alpha-helical region 43-51 is followed by sequential solvent exposure of both Glu46 and the chromophore. Which of these two residues is exposed first is correlated with the presence of a salt bridge that is part of the N-terminal domain. Additional molecular dynamics simulations indicate that the exposure of the chromophore does not result in a productive pathway. We discuss several possibilities for experimental validation of these predictions. Our results open the way for studying millisecond conformational changes in other medium-sized (signaling) proteins.

Environmental contextSingle-walled carbon nanotubes can have adverse physiological effects by interacting with proteins. Using serum bovine albumin as a model protein, we investigate the conformational changes in proteins at the tertiary structure level upon interaction with carbon nanotubes. This specific study of a model protein helps our understanding of the general binding mechanisms involved, and allows us to predict the potential adverse effects of carbon nanotubes interacting with other proteins. AbstractConsidering the large-scale production of diversified nanomaterials, it is of paramount importance to unravel the structural details of interactions between nanoparticles and biological systems at the molecular level, with the aim to reveal the potential adverse biological impacts. Herein, with single-walled carbon nanotubes (SWCNTs) acting as model nanomaterials and bovine serum albumin (BSA) acting as a model protein, a combination of spectroscopy experiments and molecular modelling was applied to help us unravel some important issues on the mechanism of protein interactions with SWCNTs. As a result, SWCNTs were first proved to bind at subdomain IB of BSA based on fluorescence experiments and molecular dynamics modelling. In addition, hydrophobic interactions were recognised as the driving force governing the binding behaviour between the SWCNTs and BSA. As a consequence, SWCNT binding led to a conformational change both at the secondary and tertiary structure levels. Insight into the binding details between BSA and SWCNT can help understand the recognition mechanism between SWCNTs and proteins, thus be helpful to predict the potential adverse effects of SWCNTs.

Effect of biogenic polyamine spermine on the structure and function of photosystem I

Kinetic analysis of conformational changes of proteins at interfaces is crucial for understanding many biological processes at membrane surfaces. In this study, we demonstrate that surface-selective sum frequency generation (SFG) spectroscopy can be used to investigate kinetics of conformational changes of proteins at interfaces. We focus on an intrinsically disordered protein, human islet amyloid polypeptide (hIAPP) that is known to misfold into the beta-sheet structure upon interaction with membranes. Using the ssp polarization setting (s-polarized SFG, s-polarized visible, and p-polarized infrared), we observe changes in the amide I spectra of hIAPP at the air/water interface after addition of dipalmitoylphosphoglycerol (DPPG) that correspond to the lipid-induced changes in secondary structures. We also used the chiral-sensitive psp polarization setting to obtain amide I spectra and observed a gradual buildup of the chiral structures that display the vibrational characteristics of parallel beta-sheets. We speculate that the second-order chiral-optical response at the antisymmetric stretch frequency of parallel beta-sheet at 1622 cm(-1) could be a highly characteristic optical property of the beta-sheet aggregates not only for hIAPP, but possibly also for other amyloid proteins. Analyzing the achiral and chiral amide I spectra, we conclude that DPPG induces the misfolding of hIAPP from alpha-helical and random-coil structures to the parallel beta-sheet structure at the air/water interface. We propose that SFG could complement existing techniques in obtaining kinetic and structural information for probing structures and functions of proteins at interfaces.

Chapter 25 - Conformational changes of biomolecules in ion-exchange chromatography

Semi-automated screen for global protein conformational changes in solution by ion mobility spectrometry–massspectrometry combined with size-exclusion chromatography and differential hydrogen–deuterium exchange