AlphaFold2 as a replacement for solution NMR structure determination of small proteins: Not so fast!

The determination of a protein's structure is often a first step towards the development of a mechanistic understanding of its function. Considerable advances in computational protein structure prediction have been made in recent years, with AlphaFold2 (AF2) emerging as the primary tool used by researchers for this purpose. While AF2 generally predicts accurate structures of folded proteins, we present here a case where AF2 incorrectly predicts the structure of a small, folded and compact protein with high confidence. This protein, pro-interleukin-18 (pro-IL-18), is the precursor of the cytokine IL-18. Interestingly, the structure of pro-IL-18 predicted by AF2 matches that of the mature cytokine, and not the corresponding experimentally determined structure of the pro-form of the protein. Thus, while computational structure prediction holds immense promise for addressing problems in protein biophysics, there is still a need for experimental structure determination, even in the context of small well-folded, globular proteins.

Zinc finger structure determination by NMR: Why zinc fingers can be a handful.

Zinc fingers can be loosely defined as protein domains containing one or more tetrahedrally-co-ordinated zinc ions whose role is to stabilise the structure rather than to be involved in enzymatic chemistry; such zinc ions are often referred to as "structural zincs". Although structural zincs can occur in proteins of any size, they assume particular significance for very small protein domains, where they are often essential for maintaining a folded state. Such small structures, that sometimes have only marginal stability, can present particular difficulties in terms of sample preparation, handling and structure determination, and early on they gained a reputation for being resistant to crystallisation. As a result, NMR has played a more prominent role in structural studies of zinc finger proteins than it has for many other types of proteins. This review will present an overview of the particular issues that arise for structure determination of zinc fingers by NMR, and ways in which these may be addressed.

NMR Structure Determinations of Small Proteins Using only One Fractionally 20% 13C- and Uniformly 100% 15N-Labeled Sample.

Uniformly 13C- and 15N-labeled samples ensure fast and reliable nuclear magnetic resonance (NMR) assignments of proteins and are commonly used for structure elucidation by NMR. However, the preparation of uniformly labeled samples is a labor-intensive and expensive step. Reducing the portion of 13C-labeled glucose by a factor of five using a fractional 20% 13C- and 100% 15N-labeling scheme could lower the total chemical costs, yet retaining sufficient structural information of uniformly [13C, 15N]-labeled sample as a result of the improved sensitivity of NMR instruments. Moreover, fractional 13C-labeling can facilitate reliable resonance assignments of sidechains because of the biosynthetic pathways of each amino-acid. Preparation of only one [20% 13C, 100% 15N]-labeled sample for small proteins (<15 kDa) could also eliminate redundant sample preparations of 100% 15N-labeled and uniformly 100% [13C, 15N]-labeled samples of proteins. We determined the NMR structures of a small alpha-helical protein, the C domain of IgG-binding protein A from Staphylococcus aureus (SpaC), and a small beta-sheet protein, CBM64 module using [20% 13C, 100% 15N]-labeled sample and compared with the crystal structures and the NMR structures derived from the 100% [13C, 15N]-labeled sample. Our results suggest that one [20% 13C, 100% 15N]-labeled sample of small proteins could be routinely used as an alternative to conventional 100% [13C, 15N]-labeling for backbone resonance assignments, NMR structure determination, 15N-relaxation analysis, and ligand-protein interaction.

High-Resolution Protein 3D Structure Determination in Living Eukaryotic Cells.

Proteins in living cells interact specifically or nonspecifically with an enormous number of biomolecules. To understand the behavior of proteins under intracellular crowding conditions, it is indispensable to observe their three-dimensional (3D) structures at the atomic level in a physiologically natural environment. We demonstrate the first de novo protein structure determinations in eukaryotes with the sf9 cell/baculovirus system using NMR data from living cells exclusively. The method was applied to five proteins, rat calmodulin, human HRas, human ubiquitin, T. thermophilus HB8 TTHA1718, and Streptococcus protein G B1 domain. In all cases, we could obtain structural information from well-resolved in-cell 3D nuclear Overhauser effect spectroscopy (NOESY) data, suggesting that our method can be a standard tool for protein structure determinations in living eukaryotic cells. For three proteins, we achieved well-converged 3D structures. Among these, the in-cell structure of protein G B1 domain was most accurately determined, demonstrating that a helix-loop region is tilted away from a ß-sheet compared to the conformation in diluted solution.

Direct Activation of Human MLKL by a Select Repertoire of Inositol Phosphate Metabolites.

Necroptosis is an inflammatory form of programmed cell death executed through plasma membrane rupture by the pseudokinase mixed lineage kinase domain-like (MLKL). We previously showed that MLKL activation requires metabolites of the inositol phosphate (IP) pathway. Here we reveal that I(1,3,4,6)P4, I(1,3,4,5,6)P5, and IP6 promote membrane permeabilization by MLKL through directly binding the N-terminal executioner domain (NED) and dissociating its auto-inhibitory region. We show that IP6 and inositol pentakisphosphate 2-kinase (IPPK) are required for necroptosis as IPPK deletion ablated IP6 production and inhibited necroptosis. The NED auto-inhibitory region is more extensive than originally described and single amino acid substitutions along this region induce spontaneous necroptosis by MLKL. Activating IPs bind three sites with affinity of 100-600 µM to destabilize contacts between the auto-inhibitory region and NED, thereby promoting MLKL activation. We therefore uncover MLKL's activating switch in NED triggered by a select repertoire of IP metabolites.

CS-ROSETTA.

Chemical Shift-Rosetta (CS-Rosetta) is an automated method that employs NMR chemical shifts to model protein structures de novo. In this chapter, we introduce the terminology and central concepts of CS-Rosetta. We describe the architecture and functionality of automatic NOESY assignment (AutoNOE) and structure determination protocols (Abrelax and RASREC) within the CS-Rosetta framework. We further demonstrate how CS-Rosetta can discriminate near-native structures against a large conformational search space using restraints obtained from NMR data, and/or sequence and structure homology. We highlight how CS-Rosetta can be combined with alternative automated approaches to (i) model oligomeric systems and (ii) create NMR-based structure determination pipeline. To show its practical applicability, we emphasize on the computational requirements and performance of CS-Rosetta for protein targets of varying molecular weight and complexity. Finally, we discuss the current Python interface, which enables easy execution of protocols for rapid and accurate high-resolution structure determination.

Chemical shift-based methods in NMR structure determination.

Chemical shifts are highly sensitive probes harnessed by NMR spectroscopists and structural biologists as conformational parameters to characterize a range of biological molecules. Traditionally, assignment of chemical shifts has been a labor-intensive process requiring numerous samples and a suite of multidimensional experiments. Over the past two decades, the development of complementary computational approaches has bolstered the analysis, interpretation and utilization of chemical shifts for elucidation of high resolution protein and nucleic acid structures. Here, we review the development and application of chemical shift-based methods for structure determination with a focus on ab initio fragment assembly, comparative modeling, oligomeric systems, and automated assignment methods. Throughout our discussion, we point out practical uses, as well as advantages and caveats, of using chemical shifts in structure modeling. We additionally highlight (i) hybrid methods that employ chemical shifts with other types of NMR restraints (residual dipolar couplings, paramagnetic relaxation enhancements and pseudocontact shifts) that allow for improved accuracy and resolution of generated 3D structures, (ii) the utilization of chemical shifts to model the structures of sparsely populated excited states, and (iii) modeling of sidechain conformations. Finally, we briefly discuss the advantages of contemporary methods that employ sparse NMR data recorded using site-specific isotope labeling schemes for chemical shift-driven structure determination of larger molecules. With this review, we aim to emphasize the accessibility and versatility of chemical shifts for structure determination of challenging biological systems, and to point out emerging areas of development that lead us towards the next generation of tools.

Analysis of Prion Protein Structure Using Nuclear Magnetic Resonance Spectroscopy.

Nuclear magnetic resonance (NMR) spectroscopy is a powerful experimental tool for obtaining information on three-dimensional (3D) structures of proteins at atomic resolution. In inherited forms of prion diseases, misfolding of cellular prion protein, PrPC, into its pathological form, PrPSc, is caused by mutations in the human prion protein gene (PRNP). Understanding of the earliest stages of the conformational changes leading to spontaneous generation of prions in inherited forms of prion diseases may benefit from detailed structural analysis of different human (Hu) PrP variants. Here, we describe the protocol for structure determination of HuPrP variants by NMR spectroscopy in solution that consists of preparation of NMR samples, acquisition of NMR data, NMR resonance assignments, and structure calculation.

NMR solution structure determination of large RNA-protein complexes.

Structure determination of RNA-protein complexes is essential for our understanding of the multiple layers of RNA-mediated posttranscriptional regulation of gene expression. Over the past 20years, NMR spectroscopy became a key tool for structural studies of RNA-protein interactions. Here, we review the progress being made in NMR structure determination of large ribonucleoprotein assemblies. We discuss approaches for the design of RNA-protein complexes for NMR structural studies, established and emerging isotope and segmental labeling schemes suitable for large RNPs and how to gain distance restraints from NOEs, PREs and EPR and orientational information from RDCs and SAXS/SANS in such systems. The new combination of NMR measurements with MD simulations and its potential will also be discussed. Application and combination of these various methods for structure determination of large RNPs will be illustrated with three large RNA-protein complexes (>40kDa) and other interesting complexes determined in the past six and a half years.

Proteins of well-defined structures can be designed without backbone readjustment by a statistical model.

We report that using mainly a statistical energy model, protein sequence design for designable backbones can be carried out with high confidence without considering backbone relaxation. A recently-developed statistical energy function for backbone-based protein sequence design has been rationally revised to improve its accuracy. As a demonstrative example, this revised model is applied to design a de novo protein for a target backbone for which the previous model had relied on after-design directed evolution to produce a well-folded protein. The actual backbone structure of the newly designed protein agrees excellently with the corresponding target. Besides presenting a new protein design protocol with experimentally verifications on different backbone types, our study implies that with an energy model of an appropriate resolution, proteins of well-defined structures instead of molten globules can be designed without the explicit consideration of backbone variations due to side chain changes, even if the side chain changes correspond to complete sequence redesigns.

Dynamic Local Polymorphisms in the Gbx1 Homeodomain Induced by DNA Binding.

The Gastrulation Brain Homeobox 1 (Gbx1) gene encodes the Gbx1 homeodomain that targets TAATTA motifs in double-stranded DNA (dsDNA). Residues Glu17 and Arg52 in Gbx1 form a salt bridge, which is preserved in crystal structures and molecular dynamics simulations of homologous homeodomain-DNA complexes. In contrast, our nuclear magnetic resonance (NMR) studies show that DNA binding to Gbx1 induces dynamic local polymorphisms, which include breaking of the Glu17-Arg52 salt bridge. To study this interaction, we produced a variant with Glu17Arg and Arg52Glu mutations, which exhibited the same fold as the wild-type protein, but a 2-fold reduction in affinity for dsDNA. Analysis of the NMR structures of the Gbx1 homeodomain in the free form, the Gbx1[E17R,R52E] variant, and a Gbx1 homeodomain-DNA complex showed that stabilizing interactions of the Arg52 side chain with the DNA backbone are facilitated by transient breakage of the Glu17-Arg52 salt bridge in the DNA-bound Gbx1.

The acidic domain is a unique structural feature of the splicing factor SYNCRIP.

The splicing factor SYNCRIP (hnRNP Q) is involved in viral replication, neural morphogenesis, modulation of circadian oscillation, and the regulation of the cytidine deaminase APOBEC1. It consists of three globular RNA-recognition motifs (RRM) domains flanked by an N-terminal acid-rich acidic sequence segment domain (AcD12-97 ) and a C-terminal domain containing an arginine-glycine-rich sequence motif (RGG/RXG box), which are located near to the N- and C-terminals, respectively. The acid-rich sequence segment is unique to SYNCRIP and the closely related protein hnRNP R, and is involved in interactions with APOBEC1. Here, we show that while AcD12-97 does not form a globular domain, structure-based annotation identified a self-folding globular domain with an all α-helix architecture, AcD24-107 . The NMR structure of AcD24-107 is fundamentally different from previously reported AcD molecular models. In addition to negatively charged surface areas, it contains a large hydrophobic cavity and a positively charged surface area as potential epitopes for intermolecular interactions.

Contact-assisted protein structure modeling by global optimization in CASP11.

We have applied the conformational space annealing method to the contact-assisted protein structure modeling in CASP11. For Tp targets, where predicted residue-residue contact information was provided, the contact energy term in the form of the Lorentzian function was implemented together with the physical energy terms used in our template-free modeling of proteins. Although we observed some structural improvement of Tp models over the models predicted without the Tp information, the improvement was not substantial on average. This is partly due to the inaccuracy of the provided contact information, where only about 18% of it was correct. For Ts targets, where the information of ambiguous NOE (Nuclear Overhauser Effect) restraints was provided, we formulated the modeling in terms of the two-tier optimization problem, which covers: (1) the assignment of NOE peaks and (2) the three-dimensional (3D) model generation based on the assigned NOEs. Although solving the problem in a direct manner appears to be intractable at first glance, we demonstrate through CASP11 that remarkably accurate protein 3D modeling is possible by brute force optimization of a relevant energy function. For 19 Ts targets of the average size of 224 residues, generated protein models were of about 3.6 Å Cα atom accuracy. Even greater structural improvement was observed when additional Tc contact information was provided. For 20 out of the total 24 Tc targets, we were able to generate protein structures which were better than the best model from the rest of the CASP11 groups in terms of GDT-TS. Proteins 2016; 84(Suppl 1):189-199. © 2015 Wiley Periodicals, Inc.

NMR structures of α-proteobacterial ATPase-regulating ζ-subunits.

NMR structures of ζ-subunits, which are recently discovered α-proteobacterial F1F0-ATPase-regulatory proteins representing a Pfam protein family of 246 sequences from 219 species (PF07345), exhibit a four-helix bundle, which is different from all other known F1F0-ATPase inhibitors. Chemical shift mapping reveals a conserved ADP/ATP binding site in ζ-subunit, which mediates long-range conformational changes related to function, as revealed by the structure of the Paracoccus denitrificans ζ-subunit in complex with ADP. These structural data suggest a new mechanism of F1F0-ATPase regulation in α-proteobacteria.

NMR structural studies of the first catalytic half-domain of ubiquitin activating enzyme.

We report a high resolution NMR structure and (15)N relaxation studies of the first catalytic cysteine half-domain (FCCH) of the mouse ubiquitin-activating enzyme E1, together with interaction studies of FCCH and the other catalytic E1 subdomain - SCCH (second catalytic cysteine half-domain). In solution, mouse FCCH forms a well-defined six-stranded antiparallel ß-barrel structure, a common fold for many proteins with a variety of cellular functions. (15)N relaxation data reveal FCCH complex backbone dynamics and indicate which residues experience slow intramolecular motions. Some of these residues make contacts with the polar face of ubiquitin in the co-crystal structure of yeast E1 and ubiquitin. However, the titration of FCCH with ubiquitin does not show any visible chemical shift changes in the 2D (1)H/(15)N HSQC spectra of the FCCH. The 2D (1)H/(15)N HSQC experiments performed both for each catalytic half-domain individually and for their equimolar mixture in the milimolar concentration range display no detectable chemical shift perturbation, suggesting a lack of interaction between the two subdomains unless they are covalently linked via the adenylation domain.

NMR-profiles of protein solutions.

NMR-Profiles are quantitative one-dimensional (1D) presentations of 2D [¹5N, ¹H]-correlation spectra used to monitor the quality of protein solutions prior to and during NMR structure determinations and functional studies. In our current use in structural genomics projects, an NMR-Profile is recorded at the outset of a structure determination, using a uniformly ¹5N-labeled microscale sample of the protein. We thus assess the extent to which polypeptide backbone resonance assignments can be achieved with given NMR techniques, for example, conventional triple resonance experiments or APSY-NMR. With the availability of sequence-specific polypeptide backbone resonance assignments in the course of the structure determination, an "Assigned NMR-Profile" is generated, which visualizes the variation of the ¹5N - ¹H correlation cross peak intensities along the sequence and thus maps the sequence locations of polypeptide segments for which the NMR line shapes are affected by conformational exchange or other processes. The Assigned NMR-Profile provides a guiding reference during later stages of the structure determination, and is of special interest for monitoring the protein during functional studies, where dynamic features may be modulated during physiological processes.