Nonuniform Sampling for NMR Spectroscopy

SAND: Automated Time-Domain Modeling of NMR Spectra Applied to Metabolite Quantification

Developments in untargeted nuclear magnetic resonance (NMR) metabolomics enable the profiling of thousands of biological samples. The exploitation of this rich source of information requires a detailed quantification of spectral features. However, the development of a consistent and automatic workflow has been challenging because of extensive signal overlap. To address this challenge, we introduce the software Spectral Automated NMR Decomposition (SAND). SAND follows on from the previous success of time-domain modeling and automatically quantifies entire spectra without manual interaction. The SAND approach uses hybrid optimization with Markov chain Monte Carlo methods, employing subsampling in both time and frequency domains. In particular, SAND randomly divides the time-domain data into training and validation sets to help avoid overfitting. We demonstrate the accuracy of SAND, which provides a correlation of ∼0.9 with ground truth on cases including highly overlapped simulated data sets, a two-compound mixture, and a urine sample spiked with different amounts of a four-compound mixture. We further demonstrate an automated annotation using correlation networks derived from SAND decomposed peaks, and on average, 74% of peaks for each compound can be recovered in single clusters. SAND is available in NMRbox, the cloud computing environment for NMR software hosted by the Network for Advanced NMR (NAN). Since the SAND method uses time-domain subsampling (i.e., random subset of time-domain points), it has the potential to be extended to a higher dimensionality and nonuniformly sampled data.

Biomolecular NMR spectroscopy in the era of artificial intelligence

Biomolecular nuclear magnetic resonance (NMR) spectroscopy and artificial intelligence (AI) have a burgeoning synergy. Deep learning-based structural predictors have forever changed structural biology, yet these tools currently face limitations in accurately characterizing protein dynamics, allostery, and conformational heterogeneity. We begin by highlighting the unique abilities of biomolecular NMR spectroscopy to complement AI-based structural predictions toward addressing these knowledge gaps. We then highlight the direct integration of deep learning approaches into biomolecular NMR methods. AI-based tools can dramatically improve the acquisition and analysis of NMR spectra, enhancing the accuracy and reliability of NMR measurements, thus streamlining experimental processes. Additionally, deep learning enables the development of novel types of NMR experiments that were previously unattainable, expanding the scope and potential of biomolecular NMR spectroscopy. Ultimately, a combination of AI and NMR promises to further revolutionize structural biology on several levels, advance our understanding of complex biomolecular systems, and accelerate drug discovery efforts.

Structural Fingerprinting of Antisense Oligonucleotide Therapeutics by Solution NMR Spectroscopy

PURPOSE

Antisense oligonucleotide (ASO) therapeutics are an emerging class of biopharmaceuticals to treat and prevent diseases, particularly those involving "undruggable" protein targets. Impurities generated throughout the ASO drug manufacturing and formulation pipeline can be detrimental to drug safety and efficacy. Therefore, analytical techniques are needed to rigorously characterize these molecules for quality assurance purposes.

METHODS

We demonstrate 1D and 2D nuclear magnetic resonance (NMR) spectroscopy methods that can generate high-resolution structural "fingerprints" of ASOs.

RESULTS AND CONCLUSIONS

1D 1H and 31P measurements are shown to provide rapid initial assessment of the ASO integrity. In particular, a well-resolved pair of 31P signals arising from the 5´-end of the phosphorodiamidate morpholino oligomer (PMO) are sensitive to complex formation and oligomerization state. 2D 1H-1H, 1H-13C, and 1H-15 N experiments, although less sensitive, are further shown to enable resonance assignment, which will allow the tracking of structural changes at high-resolution during the drug development and manufacturing processes. We further anticipate that the described NMR approaches will be broadly applicable to fully formulated ASO therapeutics, including modalities other than PMOs.

Structural Properties of CXCL4L1 and Its Recognition of the CXCR3 N-Terminus

Chemokine CXCL4L1, a homologue of CXCL4, is a more potent antiangiogenic ligand. Its structural property is correlated with the downstream receptor binding. The two chemokines execute their functions by binding the receptors of CXCR3A and CXCR3B. The receptors differ by an extra 51-residue extension in the CXCR3B N-terminus. To understand the binding specificity, a GB1 protein scaffold was used to carry different CXCR3 extracellular elements, and artificial CXCL4 and CXCL4L1 monomers were engineered for the binding assay. We first characterized the molten globule property of CXCL4L1. The structural property causes the CXCL4L1 tetramer to dissociate into monomers in low concentrations, but native CXCL4 adopts a stable tetramer structure in solution. In the titration experiments, the combination of the CXCR3A N-terminus and receptor extracellular loop 2 provided moderate and comparable binding affinities to CXCL4 and CXCL4L1, while sulfation on the CXCR3A N-terminal tyrosine residues provided binding specificity. However, the CXCR3B N-terminal extension did not show significant enhancement in the binding of CXCL4 or CXCL4L1. This result indicates that the tendency to form a chemokine monomer and the binding affinity together contribute the high antiangiogenic activity of CXCL4L1.

Pressure, motion, and conformational entropy in molecular recognition by proteins

The thermodynamics of molecular recognition by proteins is a central determinant of complex biochemistry. For over a half-century, detailed cryogenic structures have provided deep insight into the energetic contributions to ligand binding by proteins. More recently, a dynamical proxy based on NMR-relaxation methods has revealed an unexpected richness in the contributions of conformational entropy to the thermodynamics of ligand binding. Here, we report the pressure dependence of fast internal motion within the ribonuclease barnase and its complex with the protein barstar. In what we believe is a first example, we find that protein dynamics are conserved along the pressure-binding thermodynamic cycle. The femtomolar affinity of the barnase-barstar complex exists despite a penalty by -TΔS_conf of +11.7 kJ/mol at ambient pressure. At high pressure, however, the overall change in side-chain dynamics is zero, and binding occurs with no conformational entropy penalty, suggesting an important role of conformational dynamics in the adaptation of protein function to extreme environments. Distinctive clustering of the pressure sensitivity is observed in response to both pressure and binding, indicating the presence of conformational heterogeneity involving less efficiently packed alternative conformation(s). The structural segregation of dynamics observed in barnase is striking and shows how changes in both the magnitude and the sign of regional contributions of conformational entropy to the thermodynamics of protein function are possible.

Structural elucidation of the O-antigen polysaccharide from Escherichia coli O125ac and biosynthetic aspects thereof

Enteropathogenic Escherichia coli O125, the cause of infectious diarrheal disease, is comprised of two serogroups, viz., O125ab and O125ac, which display the aggregative adherence pattern with epithelial cells. Herein, the structure of the O-antigen polysaccharide from E. coli O125ac:H6 has been elucidated. Sugar analysis revealed the presence of fucose, mannose, galactose and N-acetyl-galactosamine as major components. Unassigned 1H and 13C NMR data from one- and two-dimensional NMR experiments of the O125ac O-antigen in conjunction with sugar components were used as input to the CASPER program, which can determine polysaccharide structure in a fully automated way, and resulted in the following branched pentasaccharide structure of the repeating unit: →4)[β-d-Galp-(1 → 3)]-β-d-GalpNAc-(1 → 2)-α-d-Manp-(1 → 3)-α-l-Fucp-(1 → 3)-α-d-GalpNAc-(1→, where the side chain is denoted by square brackets. The proposed O-antigen structure was confirmed by 1H and 13C NMR chemical shift assignments and determination of interresidue connectivities. Based on this structure, that of the O125ab O-antigen, which consists of hexasaccharide repeating units with an additional glucosyl group, was possible to establish in a semi-automated fashion by CASPER. The putative existence of gnu and gne in the gene clusters of the O125 serogroups is manifested by N-acetyl-d-galactosamine residues as the initial sugar residue of the biological repeating unit as well as within the repeating unit. The close similarity between O-antigen structures is consistent with the presence of two subgroups in the E. coli O125 serogroup.

Ligand-induced structural transitions combined with paramagnetic ions facilitate unambiguous NMR assignments of methyl groups in large proteins

NMR spectroscopy allows the study of biomolecules in close-to-native conditions. Structural information can be inferred from the NMR spectra when an assignment is available. Protein assignment is usually a time-consuming task, being specially challenging in the case of large, supramolecular systems. Here, we present an extension of existing state-of-the-art strategies for methyl group assignment that partially overcomes signal overlapping and other difficulties associated to isolated methyl groups. Our approach exploits the ability of proteins to populate two or more conformational states, allowing for unique NOE restraints in each protein conformer. The method is compatible with automated assignment algorithms, granting assignments beyond the limits of a single protein state. The approach also benefits from long-range structural restraints obtained from metal-induced pseudocontact shifts (PCS) and paramagnetic relaxation enhancements (PREs). We illustrate the method with the complete assignment of the 199 methyl groups of a MILproSVproSAT methyl-labeled sample of the UDP-glucose pyrophosphorylase enzyme from Leishmania major (LmUGP). Protozoan parasites of the genus Leishmania causes Leishmaniasis, a neglected disease affecting over 12 million people worldwide. LmUGP is responsible for the de novo biosynthesis of uridine diphosphate-glucose, a precursor in the biosynthesis of the dense surface glycocalyx involved in parasite survival and infectivity. NMR experiments with LmUGP and related enzymes have the potential to unravel new insights in the host resistance mechanisms used by Leishmania major. Our efforts will help in the development of selective and efficient drugs against Leishmania.

Structural Fingerprinting of Short Interfering RNA Therapeutics by Solution Nuclear Magnetic Resonance Spectroscopy

Nucleic acids are an increasingly popular platform for the development of biotherapeutics to treat a wide variety of illnesses, including diseases where traditional drug development efforts have failed. To date, there are 14 short oligonucleotide therapeutics and 2 messenger RNA (mRNA) vaccines approved by the U.S. Food and Drug Administration (FDA), which demonstrates the potential of nucleic acids as a platform for the development of safe and effective medicines and vaccines. Despite the increasing popularity of nucleic acid-based drugs, there has been a paucity of high-resolution structural techniques applied to rigorously characterize these molecules during drug development. Here, we present application of nuclear magnetic resonance (NMR) methods to structurally "fingerprint" short oligonucleotide therapeutics at natural isotope abundance under full formulation conditions. The NMR methods described herein leverage signals arising from the native structural features of nucleic acids, including imino, aromatic, and ribose resonances, in addition to non-native chemistries, such as 2'-fluoro (2'-F), 2'-O-methyl (2'-OMe), and phosphorothioate (PS) modifications, introduced during drug development. We demonstrate the utility of the NMR methods to structurally "fingerprint" a model short interfering RNA (siRNA) and a sample that simulated the drug product Givosiran. We anticipate broad applicability of the NMR methods to other nucleic acid-based therapeutics due to the generalized nature of the approach and ability to monitor many quality attributes simultaneously.

Lipid Nanodiscs for High-Resolution NMR Studies of Membrane Proteins

Membrane proteins (MPs) play essential roles in numerous cellular processes. Because around 70% of the currently marketed drugs target MPs, a detailed understanding of their structure, binding properties, and functional dynamics in a physiologically relevant environment is crucial for a more detailed understanding of this important protein class. We here summarize the benefits of using lipid nanodiscs for NMR structural investigations and provide a detailed overview of the currently used lipid nanodisc systems as well as their applications in solution-state NMR. Despite the increasing use of other structural methods for the structure determination of MPs in lipid nanodiscs, solution NMR turns out to be a versatile tool to probe a wide range of MP features, ranging from the structure determination of small to medium-sized MPs to probing ligand and partner protein binding as well as functionally relevant dynamical signatures in a lipid nanodisc setting. We will expand on these topics by discussing recent NMR studies with lipid nanodiscs and work out a key workflow for optimizing the nanodisc incorporation of an MP for subsequent NMR investigations. With this, we hope to provide a comprehensive background to enable an informed assessment of the applicability of lipid nanodiscs for NMR studies of a particular MP of interest.

Exclusively heteronuclear NMR experiments for the investigation of intrinsically disordered proteins: focusing on proline residues

NMR represents a key spectroscopic technique that contributes to the emerging field of highly flexible, intrinsically disordered proteins (IDPs) or protein regions (IDRs) that lack a stable three-dimensional structure. A set of exclusively heteronuclear NMR experiments tailored for proline residues, highly abundant in IDPs/IDRs, are presented here. They provide a valuable complement to the widely used approach based on amide proton detection, filling the gap introduced by the lack of amide protons in proline residues within polypeptide chains. The novel experiments have very interesting properties for the investigations of IDPs/IDRs of increasing complexity.

A simple approach for reconstruction of non-uniformly sampled pseudo-3D NMR data for accurate measurement of spin relaxation parameters

We explain how to conduct a pseudo-3D relaxation series NUS measurement so that it can be reconstructed by existing 3D NUS reconstruction methods to give accurate relaxation values. We demonstrate using reconstruction algorithms IST and SMILE that this 3D approach allows lower sampling densities than for independent 2D reconstructions. This is in keeping with the common finding that higher dimensionality increases signal sparsity, enabling lower sampling density. The approach treats the relaxation series as ordinary 3D time-domain data whose imaginary part in the pseudo-dimension is zero, and applies any suitably linear 3D NUS reconstruction method accordingly. Best results on measured and simulated data were achieved using acquisitions with 9 to 12 planes and exponential spacing in the pseudo-dimension out to ~ 2 times the inverse decay time. Given these criteria, in typical cases where 2D reconstructions require 50% sampling, the new 3D approach generates spectra reliably at sampling densities of 25%.

Local Deuteration Enables NMR Observation of Methyl Groups in Proteins from Eukaryotic and Cell-Free Expression Systems

Therapeutically relevant proteins such as GPCRs, antibodies and kinases face clear limitations in NMR studies due to the challenges in site-specific isotope labeling and deuteration in eukaryotic expression systems. Here we describe an efficient and simple method to observe the methyl groups of leucine residues in proteins expressed in bacterial, eukaryotic or cell-free expression systems without modification of the expression protocol. The method relies on simple stereo-selective 13 C-labeling and deuteration of leucine that alleviates the need for additional deuteration of the protein. The spectroscopic benefits of "local" deuteration are examined in detail through Forbidden Coherence Transfer (FCT) experiments and simulations. The utility of this labeling method is demonstrated in the cell-free synthesis of bacteriorhodopsin and in the insect-cell expression of the RRM2 domain of human RBM39.

FID-Net: A versatile deep neural network architecture for NMR spectral reconstruction and virtual decoupling

In recent years, the transformative potential of deep neural networks (DNNs) for analysing and interpreting NMR data has clearly been recognised. However, most applications of DNNs in NMR to date either struggle to outperform existing methodologies or are limited in scope to a narrow range of data that closely resemble the data that the network was trained on. These limitations have prevented a widescale uptake of DNNs in NMR. Addressing this, we introduce FID-Net, a deep neural network architecture inspired by WaveNet, for performing analyses on time domain NMR data. We first demonstrate the effectiveness of this architecture in reconstructing non-uniformly sampled (NUS) biomolecular NMR spectra. It is shown that a single network is able to reconstruct a diverse range of 2D NUS spectra that have been obtained with arbitrary sampling schedules, with a range of sweep widths, and a variety of other acquisition parameters. The performance of the trained FID-Net in this case exceeds or matches existing methods currently used for the reconstruction of NUS NMR spectra. Secondly, we present a network based on the FID-Net architecture that can efficiently virtually decouple ¹³C_α-¹³C_β couplings in HNCA protein NMR spectra in a single shot analysis, while at the same time leaving glycine residues unmodulated. The ability for these DNNs to work effectively in a wide range of scenarios, without retraining, paves the way for their widespread usage in analysing NMR data.

Structural basis of the dynamic human CEACAM1 monomer-dimer equilibrium

Human (h) carcinoembryonic antigen-related cell adhesion molecule 1 (CEACAM1) function depends upon IgV-mediated homodimerization or heterodimerization with host ligands, including hCEACAM5, hTIM-3, PD-1, and a variety of microbial pathogens. However, there is little structural information available on how hCEACAM1 transitions between monomeric and dimeric states which in the latter case is critical for initiating hCEACAM1 activities. We therefore mutated residues within the hCEACAM1 IgV GFCC' face including V39, I91, N97, and E99 and examined hCEACAM1 IgV monomer-homodimer exchange using differential scanning fluorimetry, multi-angle light scattering, X-ray crystallography and/or nuclear magnetic resonance. From these studies, we describe hCEACAM1 homodimeric, monomeric and transition states at atomic resolution and its conformational behavior in solution through NMR assignment of the wildtype (WT) hCEACAM1 IgV dimer and N97A mutant monomer. These studies reveal the flexibility of the GFCC' face and its important role in governing the formation of hCEACAM1 dimers and selective heterodimers.

Rapid nuclear magnetic resonance data acquisition with improved resolution and sensitivity for high-throughput metabolomic analysis

Nuclear magnetic resonance (NMR)-based metabolomics has witnessed rapid advancements in recent years with the continuous development of new methods to enhance the sensitivity, resolution, and speed of data acquisition. Some of the approaches were earlier used for peptide and protein resonance assignments and have now been adapted to metabolomics. At the same time, new NMR methods involving novel data acquisition techniques, suited particularly for high-throughput analysis in metabolomics, have been developed. In this review, we focus on the different sampling strategies or data acquisition methods that have been developed in our laboratory and other groups to acquire NMR spectra rapidly with high sensitivity and resolution for metabolomics. In particular, we focus on the use of multiple receivers, phase modulation NMR spectroscopy, and fast-pulsing methods for identification and assignments of metabolites.

Enabling NMR Studies of High Molecular Weight Systems Without the Need for Deuteration: The XL-ALSOFAST Experiment with Delayed Decoupling

Current biological research increasingly focusses on large human proteins and their complexes. Such proteins are difficult to study by NMR spectroscopy because they often can only be produced in higher eukaryotic expression systems, where deuteration is hardly feasible. Here, we present the XL-ALSOFAST-[13 C,1 H]-HMQC experiment with much improved sensitivity for fully protonated high molecular weight proteins. For the tested systems ranging from 100 to 240 kDa in size, 3-fold higher sensitivity was obtained on average for fast relaxing signals compared to current state-of-the-art experiments. In the XL-ALSOFAST approach, non-observed magnetisation is optimally exploited and transverse relaxation is minimized by the newly introduced concept of delayed decoupling. The combination of high sensitivity and superior artefact suppression makes it ideal for studying inherently unstable membrane proteins or for analysing therapeutic antibodies at natural 13 C abundance. The XL-ALSOFAST and delayed decoupling will therefore expand the range of biomolecular systems accessible to NMR spectroscopy.

Combining rapid 2D NMR experiments with novel pre-processing workflows and MIC quality measures for metabolomics

INTRODUCTION

The use of 2D NMR data sources (COSY in this paper) allows to reach general metabolomics results which are at least as good as the results obtained with 1D NMR data, and this with a less advanced and less complex level of pre-processing. But a major issue still exists and can largely slow down a generalized use of 2D data sources in metabolomics: the experiment duration.

OBJECTIVE

The goal of this paper is to overcome the experiment duration issue in our recently published MIC strategy by considering faster 2D COSY acquisition techniques: a conventional COSY with a reduced number of transients and the use of the Non-Uniform Sampling (NUS) method. These faster alternatives are all submitted to novel 2D pre-processing workflows and to Metabolomic Informative Content analyses. Eventually, results are compared to those obtained with conventional COSY spectra.

METHODS

To pre-process the 2D data sources, the Global Peak List (GPL) workflow and the Vectorization workflow are used. To compare this data sources and to detect the more informative one(s), MIC (Metabolomic Informative Content) indexes are used, based on clustering and inertia measures of quality.

RESULTS

Results are discussed according to a multi-factor experimental design (which is unsupervised and based on human urine samples). Descriptive PCA results and MIC indexes are shown, leading to the direct and objective comparison of the different data sets.

CONCLUSION

In conclusion, it is demonstrated that conventional COSY spectra recorded with only one transient per increment and COSY spectra recorded with 50% of non-uniform sampling provide very similar MIC results as the initial COSY recorded with four transients, but in a much shorter time. Consequently, using techniques like the reduction of the number of transients or NUS can really open the door to a potential high-throughput use of 2D COSY spectra in metabolomics.

2D J-correlated proton NMR experiments for structural fingerprinting of biotherapeutics

The higher order structure (HOS) of protein therapeutics is essential for drug safety and efficacy and can be evaluated by two-dimensional (2D) nuclear magnetic resonance (NMR) spectroscopy at atomic resolution. ¹Hn-¹⁵N amide correlated and ¹H-¹³C methyl correlated NMR spectroscopies at natural isotopic abundance have been demonstrated as feasible on protein therapeutics as large as monoclonal antibodies and show great promise for use in establishing drug substance structural consistency across manufacturing changes and in comparing a biosimilar to an originator reference product. Spectral fingerprints from ¹Hn-¹Hα correlations acquired using 2D homonuclear proton-proton J-correlated NMR experiments provide a complementary approach for high-resolution assessment of the HOS of lower molecular weight (<25 kDa) protein therapeutics. Here, we evaluate different pulse sequences (COSY, TOCSY and TACSY) used to generate proton-proton J-correlated NMR spectral fingerprints and appraise the performance of each method for application to protein therapeutic HOS assessment and comparability.

Taking Simultaneous Snapshots of Intrinsically Disordered Proteins in Action

Intrinsically disordered proteins (IDPs) as well as intrinsically disordered regions (IDRs) of complex protein machineries have recently been recognized as key players in many cellular functions. NMR represents a unique tool to access atomic resolution structural and dynamic information on highly flexible IDPs/IDRs. Improvements in instrumental sensitivity made heteronuclear direct detection possible for biomolecular NMR applications. The CON experiment has become one of the most useful NMR experiments to get a snapshot of an IDP/IDR in conditions approaching physiological ones. The availability of NMR spectrometers equipped with multiple receivers now enables the acquisition of several experiments simultaneously instead of one after the other. Here, we propose several variants of the CON experiment in which, during the recovery delay, a second two-dimensional experiment is acquired, either based on 1H detection (CON//HN) or on 15N detection (CON//btNH, CON//(H)CAN). The possibility to collect simultaneous snapshots of an IDP/IDR through different two-dimensional spectra provides a novel tool to follow chemical reactions, such as the occurrence of posttranslational modifications, as well as to study samples of limited lifetime such as cell lysates or whole cells.

Rapid NMR assignments of intrinsically disordered proteins using two-dimensional 13C-detection based experiments

An approach for rapid backbone resonance assignments in proteins using only two 2D NMR experiments is presented. The new method involves a combination of high-resolution 13Cα-detected NMR experiments and selective unlabeling of amino acid residues. The 13C detected 2D hNCA and 2D hNcoCA spectra of a uniformly labeled sample of the protein are analysed in concert with the 2D hNCA spectrum obtained for a selectively unlabeled sample. The combinatorial set of amino acid residues for selective unlabeling is chosen optimally to maximize the assignments. The method is useful for rapid assignment of proteins with low stability such as intrinsically disordered proteins and is applicable to deuterated proteins. This approach helped in assignments of 14.5 kDa human α-synuclein during the course of its aggregation.