The rotational correlation time ( τ_c ) is a one of the fundamental parameters for characterizing biomacromolecules in solution. The [15N, 1H]-TRACT (TROSY for Rotational Correlation Times) NMR experiment is a widely utilized method for determining τ_c by measuring the differential 15N−1H transverse relaxation rates of α and β states. However, extracting accurate relaxation rates from these experiments often requires researchers to rely on ad hoc scripts or tedious workflows across multiple software packages, which can introduce systematic errors during peak integration and baseline correction. To address this bottleneck, we introduce pyTRACTnmr, an open-source, modular Python application featuring a modern graphical user interface designed specifically for streamlined analysis of [15N, 1H]-TRACT data. Built on the PySide6 framework and leveraging the nmrglue library for robust data ingestion, the software provides an interactive environment for visual baseline correction, precise integration window definition, and non-linear least-squares fitting of relaxation decays. By replacing rigid automated processing with a user-friendly, visually guided workflow, pyTRACTnmr minimizes noise amplification and ensures highly reproducible analysis, and significantly simplifies the workflow.
Relaxation phenomena greatly influence magnetic resonance experiments, from setting repetition rates and measurement times through to measuring internal motion of molecules. RelCalc provides a convenient and efficient means for symbolic and numerical evaluation of relaxation rates using Bloch, Wagness and Redfield (BWR) theory. Users supply interactions within a spin system arranged appropriately in the molecular frame. The effects of optional fluctuations around a symmetry axis, and then under either isotropic or anisotropic global tumbling are then computed. Relaxation rates linking pairs of Liouvillian basis operators are then returned symbolically, formatted as a sum of weighted reduced spectra density functions. Parameters are then supplied to replace symbolic quantities for values for rapid computation of numerical rates. We demonstrate the utility of the software by exploring relaxation interference (TROSY) effects and ’long-lived states’ in X_2 and AX_n systems, which, for n=1,2, and 3, describe NH, NH _2 and CH _3 groups in proteins, and small molecules with n=4,6,8,12 and 20 arranged as platonic solids. The implementation is highly efficient with 65,536 relaxation rates required to perform complete Liouvillian simulation of AX_3 methyl groups rotating about a symmetric axis in the presence of adjacent static ‘external’ spins being symbolically calculated in a few seconds on a single processor on a 2021 laptop. From this result, a complete set of numerical rates are then computed in under one second. RelCalc is implemented in python and is freely available.
β-sandwich domains of the intermediate immunoglobulin fold (IgI) are key building blocks of many large, multidomain proteins – such as titin. The uniformity of the basic fold of these domains does not seem to be an impediment to their adaptation to a wide variety of functions. IgI domains in different regions of titin have to fulfil distinct functions which is seen in subtle differences of sequence, structure and stability. This is particularly true for the N2A region whose IgI domains are subtly distinct from other parts of I-band titin. We have already shown the unusual structure and properties of IgI domain I83 which is able to bind to calcium and is part of a binding site for F-actin and p94. To continue our exploration of the titin IgI domains of the N2A region we describe here the solution NMR structure of domain I82 of murine titin. The structure of the murine domain is virtually identical to the human homologue. However, the NMR investigation reveals the existence of distinct conformers around a highly conserved glycine, part of the tyrosine corner motif at the junction of the EF-loop with the F-strand. 15N relaxation data show substantial line broadening for residues around this glycine, confirming conformational exchange on the fast-intermediate time scale. The unusual dynamics could for the first time explain the high level of conservation of the tyrosine corner glycine via a function in the folding of the domain.
Monoclonal antibodies (mAbs) are essential therapeutic agents whose efficacy depends critically on the integrity of their higher-order structure (HOS). Subtle perturbations in HOS, arising from post-translational modifications, chemical degradation or formulation conditions, can impair antigen binding or induce immunogenicity. Nuclear Magnetic Resonance (NMR) spectroscopy, and particularly methyl NMR, offers a powerful means to probe the structure and dynamics of large proteins, such as antibodies, at atomic scale resolution. While 2D 1H-13C methyl correlation spectra at natural abundance provide valuable structural fingerprints, their limited resolution and sensitivity hamper sophisticated NMR studies. Isotopic labeling of methyl groups with 13CH₃, combined with tailored deuteration, greatly enhances spectral quality. However, such labeling strategies remain underdeveloped in mammalian systems like CHO cells, which are the standard hosts for therapeutic antibody production. Here, we report an optimized labeling strategy enabling selective incorporation of the ¹³CH₃-labeled methyl groups from the six methyl-bearing amino acids into antibodies expressed in CHO cells. This includes enzymatic synthesis of regio- and stereoselectively labeled isoleucine and valine with partial deuteration and an optimized protocol for their incorporation into CHO-produced proteins. When applied to an anti-LAMP1 therapeutic antibody, this strategy yields highly resolved methyl NMR spectra, enabling comprehensive HOS assessment. It allows the transfer of previously assigned Fab and Fc methyl resonances, resulting in approximately 84
Over the last 20 years, the number of large proteins accessible to protein-NMR analysis has increased significantly due to the development of selective [1H,13C]-methyl labelling techniques in combination with methyl-TROSY based NMR experiments. Structure-based strategies for the assignment of [¹H,¹³C]-methyl groups rely on comparing spatial constraints derived from methyl–methyl NOEs to a known three-dimensional structure of the protein of interest. Cross peaks in methyl-TROSY spectra are assigned to specific methyl groups by matching methyl–methyl NOEs, as observed for example in 4D HMQC-NOESY-HMQC spectra, with distances derived from a structural model. This process is commonly referred to as a “methyl walk”. Here, we present AMIGO (Automated Methyl assignment via Iterative Graph Optimization), a novel assignment algorithm that formalises the intuitive methyl walk procedure by constructing graphs with nodes representing specific methyl groups and edges reflecting methyl-methyl NOEs or short methyl-methyl distances in a model. “Building blocks” consisting of nodes and edges are then generated to reconcile structure-based and NOE-based graphs. Assignments are achieved through permutation and concatenation of individual “building blocks” in a modular fashion, enabling efficient computation even for large proteins. Additional experimental restraints, such as paramagnetic relaxation enhancements (PREs) or pseudocontact shifts (PCSs), can be integrated to validate and extend the assignments. The performance of AMIGO was validated using 11 proteins that had previously been assigned and 32 NOE networks that had been generated synthetically.
Solid-state NMR spectroscopy is often limited by low spectral resolution, a problem typically addressed using fast magic-angle spinning (MAS) and ¹H detection, which require costly specialized hardware. Here, we demonstrate that the super-resolution method—previously applied in solution-state NMR—can be successfully implemented in solid-state NMR to enhance resolution. Applying dynamic number of scans (DNS) sampling to 2D ¹³C-¹³C DARR experiments on the AP205 capsid protein yielded an effective doubling of resolution, halving peak widths from 180 Hz to 87 Hz. Furthermore, DNS acquisition provides a significant advantage over post-acquisition apodization of conventional data with a 20
Peak information extraction is an important step for all types of protein NMR experiments, from structure determination to titration experiments. In most cases, extraction of peak information, or peak-picking, is done semi-automatically or manually, relying on the spectroscopist’s expertise to guard against errors in the algorithm. Optimus Peak is designed for fully automatic peak-picking in multi-dimensional protein NMR spectra for both solid-state and solution NMR. The program shows a low false positive rate, due to its more robust evolutionary algorithm-based fitting. The program is also designed to fit together with other programs by being implemented as a terminal based command line program. Optimus Peak was tested using both solid-state and solution NMR experiments, for both 2D and 3D experiments, using 67 spectra from 12 different proteins, consisting of 26 different types of NMR experiments.
Isotopic enrichment of pharmacologically relevant protein targets is crucial for structural studies by nuclear magnetic resonance (NMR) and plays a key role in advancing structure-guided drug discovery. Many clinically important drug targets require expression in eukaryotic systems-such as mammalian, yeast, or insect cells-rather than prokaryotic hosts. This requirement limits the feasibility of high-throughput isotopic labeling and poses challenges for obtaining uniformly isotope-labeled proteins suitable for NMR analysis. While several enrichment strategies have been developed, no broadly applicable enrichment platform has emerged for eukaryotic expression systems. In this study, we introduce Cupriavidus necator as an alternative biological source for 15N and 13C isotopic enrichment to support protein production in eukaryotic systems. To evaluate this approach, we selected the kinase domain of EPHA2, a receptor tyrosine kinase implicated in colorectal cancer progression and an important target for therapeutic inhibitor development. Isotopic incorporation was quantified using liquid chromatography-mass spectrometry (LC-MS), revealing enrichment levels of 79% for 15N and 69% for 13C. These results demonstrate that Cupriavidus necator can serve as a robust and flexible platform for generating isotopically enriched biomolecules compatible with eukaryotic protein expression, thereby enabling NMR investigations of disease-relevant protein targets.
The R3H domain of the human protein Sµbp-2 was produced with 5-fluoro-L-isoleucine (FIle) and 5,5-difluoro-L-isoleucine (diFIle) as probes for detection by 19F-NMR spectroscopy. The fluorinated protein, produced by cell-free protein synthesis, was obtained more easily with diFIle than FIle as FIle readily hydrolysed at pH 7.5 with the release of fluoride. The 19F-NMR spectra showed large chemical shift ranges but were heterogeneous. The heterogeneities arose from difficulties to fully exclude canonical isoleucine, the presence of multiple conformations and limited stability of the proteins, with the sample made with diFIle being particularly prone to precipitation. 19F resonance assignments were obtained by comparison of the chemical shifts of γ1-protons with those observed in the wild-type protein. Non-uniform cross-peak intensities observed in short-delay 1H,19F correlation experiments suggest incomplete averaging of 3JHF couplings and therefore preferential rotamer populations of the CH2F and CHF2 groups.
Band-selective excitation short-transient (BEST) sequences are widely used for protein NMR experiments that start with amide proton magnetization, such as 1H-15N HSQC, 1H-15N TROSY, and multidimensional backbone assignment experiments, because the optimization of amide proton longitudinal relaxation afforded by the BEST methodology allows for much greater sensitivity when using short scan times. Here we show that the BEST methodology can be easily incorporated in sequences for measuring proton transverse relaxation rates (1H R2), which are typically used to determine paramagnetic relaxation enhancements (PREs). The resulting BEST-HSQC-PRE and BEST-TROSY-PRE experiments afford similar or better sensitivity for measuring PREs compared to previous methods, provide equally accurate measurements of transverse relaxation rates (and therefore PREs), and allow shorter scan times to be used.
In this communication we describe a new scheme to process the data from stimulated echo protein diffusion experiments. For a series of gradient-encoded proton spectra f_k(ω ) considered over the selected spectral region (ω _left,ω _right) , we build a model to approximate the unique (protein-dependent) shape of the spectrum. Taking a cue from the optimal filtration theory, f_model(ω ) is constructed as the intensity-weighted combination of f_k(ω ) . The so obtained f_model(ω ) is then used to fit the individual spectra f_k(ω ) , thus providing highly accurate estimates for the integral signal intensities that are subsequently used for Stejskal-Tanner-type analyses. This algorithm has been implemented as a part of a new web server, named DDfit ( https://ddfit.org , mirror at https://ddfit.bio-nmr.spbu.ru/ ). The server accepts spectrometer data from the standard stimulated and double-stimulated echo experiments by Bruker, as well as custom-designed experiments. The server is easy to use, with data processing taking no more than several seconds. Our tests using simulated as well as experimental data found that DDfit determines protein diffusion coefficients with both accuracy and precision, offering several-fold improvement in precision compared to other processing schemes.
We demonstrate here a pulse sequence based on rotational-echo double resonance (REDOR) that can help distinguish resonances based on the ψ -torsion angle, which allows it to distinguish α -helical and β -sheet regions in solid proteins under magic-angle-spinning. The method relies on conformation-dependent differences in distances between an amide ^1 H and amide ^15 N nuclei not covalently attached to it. Dephasing from this remote ^15 N nucleus is obtained in presence of the much stronger one-bond dipole-dipole coupling by using the θ -REDOR sequence. Experiments are demonstrated on perdeuterated (and 70 ^13 C, ^15 N labeled model protein GB1 at the magic-angle spinning frequency of 41.67 kHz. This method will be useful in simplifying chemical-shift assignments in proteins where the structure is already known, and we anticipate a direct application in determining secondary structures without relying on ^13 C chemical-shifts.
Under inflammatory conditions, the ubiquitin-like modifier FAT10 serves as a tag for protein degradation by the 26S proteasome. FAT10 is degraded along with its substrates and this process is independent of the segregase VCP/p97, which, in the regular ubiquitin pathway of degradation, is required if a substrate lacks a disordered initiation region. FAT10 itself is loosely folded and its tendency to aggregate has complicated investigations of its structure, interaction, and function. Recently, hydrogen-deuterium exchange in combination with mass spectrometry has suggested that, in preparation of degradation by the proteasome, the adapter protein NUB1 traps FAT10 in a mostly unfolded state by capturing a β-strand. β-strand capture was subsequently confirmed by magic-angle spinning (MAS) NMR spectroscopy of a stabilized variant of the N-domain of FAT10 in complex with NUB1L, the longer splice variant of NUB1. MAS NMR, in addition, revealed that the N-domain of FAT10 and NUB1L form a fuzzy complex and that the N-terminus of FAT10 is positioned for initiation of degradation by specific non-covalent interaction with NUB1L. Here, we report the investigation of the wild-type N-domain of FAT10 by MAS NMR. Co-sedimentation with NUB1L yields high-quality spectra, which enable sequential assignment of resonances. Based on MAS NMR data, the complexes of the wild-type and stabilized N-domain of FAT10 with NUB1L appear identical. The N-terminal residue of FAT10 again shows up prominently in the spectra, even though it is this time an Ala, not a Gly. Our experiences suggest that co-sedimentation in combination with MAS NMR is generally helpful in the exploration of conditional folds of intrinsically disordered proteins.
Monoclonal antibodies (mAbs) are leading therapeutic agents due to their high specificity and limited side effects. Ensuring their structural integrity under stress and maintaining batch consistency require robust quality control. Methyl 2D NMR has emerged as a powerful tool to probe mAb structure at natural isotopic abundance, enabling spectral fingerprint comparisons across production batches to detect subtle structural changes. However, extracting atomic-level structural information requires assignment of methyl resonances to their amino acids. While such assignments are available for several antigen-binding fragments (Fabs), no comprehensive assignment has been reported for the crystallisable fragment (Fc). In this study, we present the methyl group assignment of the 50-kDa Fc fragment of an immunoglobulin G1 (IgG1) antibody. Using cell-free expression, strategic isotopic labelling, and high-quality 2D and 3D NMR experiments, we successfully assigned 94% of methyl resonances of a non-glycosylated Fc. Given that therapeutic mAbs are typically produced in Chinese Hamster Ovary (CHO) cells, we transferred this assignment to the methyl spectrum of a glycosylated Fc fragment obtained by the enzymatic cleavage of a CHO-produced mAb at natural abundance, achieving 83% assignment coverage. This assignment was then used to investigate the impact of methionine oxidation on Fc structure at atomic resolution using NMR. The methyl group assignment transforms 2D methyl NMR fingerprinting into a powerful tool for quality control. It enables the direct comparison of spectra acquired on mAbs produced at natural abundance, allowing the detection and localisation of chemical modifications and structural changes without the need for isotopic labelling. This approach offers a robust solution for monitoring the structural integrity of therapeutic antibodies throughout development and manufacturing.
M9 minimal media and its enhanced variants (M9 + and M9++) are widely used for recombinant protein expression in Escherichia coli, particularly for isotopic labeling required in structural biology techniques such as NMR spectroscopy. This study investigates how different compositions of M9-based media (M9, M9+, and M9++) influence bacterial growth, metabolic stress, and central carbon metabolism during recombinant expression of the protein. Using 1D ¹H NMR spectroscopy and multivariate statistical analysis, we observed distinct media-dependent metabolic shifts. Standard M9 exhibited limited bacterial growth and heightened stress-related fermentation, indicated by high ethanol and acetate levels. In contrast, M9 + significantly increased biomass but promoted pronounced overflow metabolism. M9 + + presented intermediate biomass levels and markedly reduced overflow metabolites, favoring biosynthesis pathways, notably increasing valine, acetoin, and formate concentrations. These findings suggest that further optimization of glucose concentration, nitrogen sources, and phosphate buffering could significantly improve the metabolic balance of M9++, creating an enhanced medium tailored for efficient, high-quality recombinant protein expression and isotopic labeling in E. coli.
Owing to their high specificity and therapeutic effectiveness, monoclonal antibodies (mAbs) have rapidly become one of the leading classes of biologic drugs used to treat critical illnesses. The antigen-binding fragment (Fab) of mAbs plays a key role in the antigen recognition, so its structural characterization is essential, as even a slight change to its Higher Order Structure (HOS) can impact the antibody’s potency. Recently, 2D methyl NMR has been introduced as a powerful method to assess both the structure and integrity of therapeutic Fab fragments. However, the identification of methyl group resonances in NMR spectra remains rare since Fabs are large heterodimers of 50 kDa. Here, we present the methyl group assignment of an IgG1 Fab produced in a cell-free system with an optimal isotope labelling. We first assigned 99
The efficient computer optimization of magnetic resonance pulses and pulse sequences involves the calculation of a problem-adapted cost function as well as its gradients with respect to all controls applied. The gradients generally can be calculated as a finite difference approximation, as a GRAPE approximation, or as an exact function, e.g. by the use of the augmented matrix exponentiation, where the exact gradient should lead to best optimization convergence. The calculation of the exact gradient, on the other hand, is computationally demanding and is one of the determining factors for the overall time needed for optimization. Its improved calculation, e.g. by faster analytical solutions, is therefore highly desired. The majority of pulse optimizations involve a single spin 1/2, for which propagation is either represented by 3D-rotations or quaternions. Here, highly efficient analytical solutions for gradients for both cases are derived with respect to various possible controls. Controls are either x and y pulses, but also z-controls, as well as gradients with respect to amplitude and phase of a pulse shape. In addition, analytical solutions with respect to pseudo controls, involving holonomic constraints to maximum rf-amplitudes, maximum rf-power, or maximum rf-energy, are introduced. Using the hyperbolic tangent function, maximum values are imposed in a fully continuous and differentiable way. The obtained analytical gradients allow the calculation two orders of magnitude faster than the augmented matrix exponential approach, automatically speeding up any optimization in any optimization program by the same factor for the corresponding set of controls. The use of exact gradients for different controls is finally demonstrated in a number of optimizations involving broadband pulses of potential use in biomolecular applications for ^15 N, ^13 C, and ^19 F.
We present a novel approach for side-chain-selective deuteration of proteins to improve 1Hα spectral resolution and to simplify side-chain signals in 1H-detected protein solid-state NMR (SSNMR) with a simple bio-expression method using E. coli BL21 (DE3). 1H-detected SSNMR using ultra-fast magic-angle spinning (MAS) at a spinning rate of 60 kHz or higher is attracting attention as a powerful method of protein structure determination. However, even with ultra-fast MAS at 100 kHz, the 1H line broadening due to 1H-1H dipolar interactions cannot be eliminated, posing an obstacle to signal assignment and structure determination. To improve resolution for SSNMR-based protein structural analysis, we developed a method to selectively deuterate side-chains at a high deuteration level while maintaining the protons at the α-position. This selective labeling method is based on the transamination reaction in the amino-acid biosynthesis pathway and switching a medium from an unlabeled H2O medium containing D-glucose (glucose), ammonium chloride, and amino acid mixture for rapid cell growth to a labeled H2O medium containing [2H, 13C]-glucose, 15N-labeled ammonium chloride, and a [2H, 13C, 15N]-labeled amino-acid mixture just before the induction. With [2H, 13C]-labeled glucose and a [2H, 13C, 15N]-labeled amino-acid mixture as the carbon sources, this medium-switching method provides a simple and efficient means to express a selectively deuterated protein GB1 domain (GB1) sample, which is achieved by promoting efficient back-protonation at the α-position via the transamination reaction while retaining side-chain deuterons to a large extent. The yield of the GB1 protein was found to be enhanced by a factor of ca. 1.5 with the medium-switching method, compared with that for the expression with a traditional M9 minimal medium in H2O without medium-switching. For the selectively deuterated GB1 sample, the resultant 1H resolution for resolved 1Hα peaks in 1H-detected 2D 1H/13C correlation SSNMR at a MAS rate of 70 kHz was improved by a factor of 1.21 on average, compared with the corresponding resolution for a fully protonated, uniformly 13C- and 15N-labeled GB1 sample. Furthermore, side-chain signal assignment is facilitated by utilizing residual protons of the side chains. Our results also suggest that the side-chain deuteration level can be altered by adjusting the level of the deuterated amino-acid mixture in the expression system.
The use of 1H detection, made possible by very fast magic-angle spinning (MAS), has revolutionized the field of biomolecular solid-state NMR. In the past, 1H detection was often paired with deuteration schemes to achieve the highest possible resolution needed for protein structural characterization. However, with modern probes capable of MAS rates over 100 kHz, deuteration is no longer required, resulting in a need to measure long-range distances in fully protonated systems. In this study, we evaluate the potential of two 3D pulse sequences, (H)NCOH and (H)NCAH, to measure long-range C-H correlations in a fully protonated protein sample at a MAS rate of 105 kHz. Our results show that the (H)NCOH spectrum contains multiple sequential and structurally relevant long-range CO-H contacts for each residue, capturing HN contacts up to 6 Å despite transfers to side chain protons. Conversely, the (H)NCAH spectrum yields fewer Cα-HN correlations, with those present mostly from intraresidue aliphatic proton contacts. Therefore, in protonated proteins, the extensive 1H network leads to dipolar truncation in the Cα-H experiment, while the CO-H correlations observed are comparable to those in deuterated samples. These findings highlight the feasibility of conducting distance measurements based on long-range cross polarization, on more accessible and affordable samples, expanding the scope of proton detection for systems where deuteration and back-exchange are not possible.
Cellular functions require biomolecules to transition among various conformational sub-states in the energy landscape. A mechanistic understanding of cellular functions requires quantitative knowledge of the kinetics, thermodynamics, and structural features of the biomolecules experiencing exchange between several states. High-power Relaxation Dispersion (RD) NMR experiments have proven very effective for such measurements if the exchange occurs in timescales ranging from microseconds to milliseconds. However, scanning the significantly larger kinetic window within the time limit of instrumental availability and sample stability requires careful optimization of experiments. Understanding biomolecular functions at a mechanistic level depends on fitting such experimental data to theoretical models. However, the reliability of the fit parameters depends on the measurement schemes and is sensitive to experimental noise. Here, we benchmark different measurement schemes along with theoretical models for sub-millisecond timescale exchange and determine the robustness of these models in providing information when the measurements contain noise. Our results show that kinetics can be measured reliably from such experiments. The structural features of the exchanging sub-states, encoded in the chemical shift differences between the states, can be fitted, albeit with significant uncertainties. Information about the minor states is difficult to obtain exclusively from the RD data due to large uncertainties and sensitivity to noise.
