Sparse (13)C labelling for solid-state NMR studies of P. pastoris expressed eukaryotic seven-transmembrane proteins

Solid-State NMR: Methods for Biological Solids

In the last two decades, solid-state nuclear magnetic resonance (ssNMR) spectroscopy has transformed from a spectroscopic technique investigating small molecules and industrial polymers to a potent tool decrypting structure and underlying dynamics of complex biological systems, such as membrane proteins, fibrils, and assemblies, in near-physiological environments and temperatures. This transformation can be ascribed to improvements in hardware design, sample preparation, pulsed methods, isotope labeling strategies, resolution, and sensitivity. The fundamental engagement between nuclear spins and radio-frequency pulses in the presence of a strong static magnetic field is identical between solution and ssNMR, but the experimental procedures vastly differ because of the absence of molecular tumbling in solids. This review discusses routinely employed state-of-the-art static and MAS pulsed NMR methods relevant for biological samples with rotational correlation times exceeding 100's of nanoseconds. Recent developments in signal filtering approaches, proton methodologies, and multiple acquisition techniques to boost sensitivity and speed up data acquisition at fast MAS are also discussed. Several examples of protein structures (globular, membrane, fibrils, and assemblies) solved with ssNMR spectroscopy have been considered. We also discuss integrated approaches to structurally characterize challenging biological systems and some newly emanating subdisciplines in ssNMR spectroscopy.

Towards a native environment: structure and function of membrane proteins in lipid bilayers by NMR

Solid-state NMR (ssNMR) is a versatile technique that can be used for the characterization of various materials, ranging from small molecules to biological samples, including membrane proteins. ssNMR can probe both the structure and dynamics of membrane proteins, revealing protein function in a near-native lipid bilayer environment. The main limitation of the method is spectral resolution and sensitivity, however recent developments in ssNMR hardware, including the commercialization of 28 T magnets (1.2 GHz proton frequency) and ultrafast MAS spinning (<100 kHz) promise to accelerate acquisition, while reducing sample requirement, both of which are critical to membrane protein studies. Here, we review recent advances in ssNMR methodology used for structure determination of membrane proteins in native and mimetic environments, as well as the study of protein functions such as protein dynamics, and interactions with ligands, lipids and cholesterol.

Improved Protocol for the Production of the Low-Expression Eukaryotic Membrane Protein Human Aquaporin 2 in Pichia pastoris for Solid-State NMR

Solid-state nuclear magnetic resonance (SSNMR) is a powerful biophysical technique for studies of membrane proteins; it requires the incorporation of isotopic labels into the sample. This is usually accomplished through over-expression of the protein of interest in a prokaryotic or eukaryotic host in minimal media, wherein all (or some) carbon and nitrogen sources are isotopically labeled. In order to obtain multi-dimensional NMR spectra with adequate signal-to-noise ratios suitable for in-depth analysis, one requires high yields of homogeneously structured protein. Some membrane proteins, such as human aquaporin 2 (hAQP2), exhibit poor expression, which can make producing a sample for SSNMR in an economic fashion extremely difficult, as growth in minimal media adds additional strain on expression hosts. We have developed an optimized growth protocol for eukaryotic membrane proteins in the methylotrophic yeast Pichia pastoris. Our new growth protocol uses the combination of sorbitol supplementation, higher cell density, and low temperature induction (LT-SEVIN), which increases the yield of full-length, isotopically labeled hAQP2 ten-fold. Combining mass spectrometry and SSNMR, we were able to determine the nature and the extent of post-translational modifications of the protein. The resultant protein can be functionally reconstituted into lipids and yields excellent resolution and spectral coverage when analyzed by two-dimensional SSNMR spectroscopy.

Codon Harmonization of a Kir3.1-KirBac1.3 Chimera for Structural Study Optimization

The expression of functional, folded, and isotopically enriched membrane proteins is an enduring bottleneck for nuclear magnetic resonance (NMR) studies. Indeed, historically, protein yield optimization has been insufficient to allow NMR analysis of many complex Eukaryotic membrane proteins. However, recent work has found that manipulation of plasmid codons improves the odds of successful NMR-friendly protein production. In the last decade, numerous studies showed that matching codon usage patterns in recombinant gene sequences to those in the native sequence is positively correlated with increased protein yield. This phenomenon, dubbed codon harmonization, may be a powerful tool in optimizing recombinant expression of difficult-to-produce membrane proteins for structural studies. Here, we apply this technique to an inward rectifier K⁺ Channel (Kir) 3.1-KirBac1.3 chimera. Kir3.1 falls within the G protein-coupled inward rectifier K⁺ (GIRK) channel family, thus NMR studies may inform on the nuances of GIRK gating action in the presence and absence of its G Protein, lipid, and small molecule ligands. In our hands, harmonized plasmids increase protein yield nearly two-fold compared to the traditional ‘fully codon optimized’ construct. We then employ a fluorescence-based functional assay and solid-state NMR correlation spectroscopy to show the final protein product is folded and functional.

Longitudinal Relaxation Optimization Enhances 1 H-Detected HSQC in Solid-State NMR Spectroscopy on Challenging Biological Systems

Solid-state (SS) NMR spectroscopy is a powerful technique for studying challenging biological systems, but it often suffers from low sensitivity. A longitudinal relaxation optimization scheme to enhance the signal sensitivity of HSQC experiments in SSNMR spectroscopy is reported. Under the proposed scheme, the ¹ H spins of ¹ H-X (¹⁵ N or ¹³ C) are selected for signal acquisition, whereas other vast ¹ H spins are flipped back to the axis of the static magnetic field to accelerate the spin recovery of the observed ¹ H spins, resulting in enhanced sensitivity. Three biological systems are used to evaluate this strategy, including a seven-transmembrane protein, an RNA, and a whole-cell sample. For all three samples, the proposed scheme largely shortens the effective ¹ H longitudinal relaxation time and results in a 1.3-2.5-fold gain in sensitivity. The selected systems are representative of challenging biological systems for observation by means of SSNMR spectroscopy; thus indicating the general applicability of this method, which is particularly important for biological samples with a short lifetime or with limited sample quantities.

Enhancing sensitivity of Double Electron-Electron Resonance (DEER) by using Relaxation-Optimized Acquisition Length Distribution (RELOAD) scheme

Over the past decades pulsed electron-electron double resonance (PELDOR), often called double electron-electron resonance (DEER), became one of the major spectroscopic tools for measurements of nanometer-scale distances and distance distributions in non-crystalline biological and chemical systems. The method is based on detecting the amplitude of the primary (3-pulse DEER) or refocused (4-pulse DEER) spin echo for the so-called "observer" spins when the other spins coupled to the former by a dipolar interaction are flipped by a "pump" pulse at another EPR frequency. While the timing of the pump pulse is varied in steps, the positions of the observer pulses are typically fixed. For such a detection scheme the total length of the observer pulse train and the electron spin memory time determine the amplitude of the detected echo signal. Usually, the distance range considerations in DEER experiments dictate the total length of the observer pulse train to exceed the phase memory time by a factor of few and this leads to a dramatic loss of the signal-to-noise ratio (SNR). While the acquisition of the DEER signal seems to be irrational under such conditions, it is currently the preferred way to conduct DEER because of an effective filtering out of all other unwanted interactions. Here we propose a novel albeit simple approach to improve DEER sensitivity and decrease data acquisition time by introducing the signal acquisition scheme based on RELaxation Optimized Acquisition (Length) Distribution (DEER-RELOAD). In DEER-RELOAD the dipolar phase evolution signal is acquired in multiple segments in which the observer pulses are fixed at the positions to optimize SNR just for that specific segment. The length of the segment is chosen to maximize the signal acquisition efficiency according the phase relaxation properties of the spin system. The total DEER trace is then obtained by "stitching" the multiple segments into a one continuous trace. The utility of the DEER-RELOAD acquisition scheme has been demonstrated on an example of the standard 4-pulse DEER sequence applied to two membrane protein complexes labeled with nitroxides. While theoretical gains from the DEER-RELOAD scheme increase with the number of stitched segments, in practice, even dividing the acquisition of the DEER trace into two segments may improve SNR by a factor of >3, as it has been demonstrated for one of these two membrane proteins.

Methyl-selective isotope labeling using α-ketoisovalerate for the yeast Pichia pastoris recombinant protein expression system

Methyl-detected NMR spectroscopy is a useful tool for investigating the structures and interactions of large macromolecules such as membrane proteins. The procedures for preparation of methyl-specific isotopically-labeled proteins were established for the Escherichia coli (E. coli) expression system, but typically it is not feasible to express eukaryotic proteins using E. coli. The Pichia pastoris (P. pastoris) expression system is the most common yeast expression system, and is known to be superior to the E. coli system for the expression of mammalian proteins, including secretory and membrane proteins. However, this system has not yet been optimized for methyl-specific isotope labeling, especially for Val/Leu-methyl specific isotope incorporation. To overcome this difficulty, we explored various culture conditions for the yeast cells to efficiently uptake Val/Leu precursors. Among the searched conditions, we found that the cultivation pH has a critical effect on Val/Leu precursor uptake. At an acidic cultivation pH, the uptake of the Val/Leu precursor was increased, and methyl groups of Val and Leu in the synthesized recombinant protein yielded intense ¹H-¹³C correlation signals. Based on these results, we present optimized protocols for the Val/Leu-methyl-selective ¹³C incorporation by the P. pastoris expression system.

On the use of Pichia pastoris for isotopic labeling of human GPCRs for NMR studies

NMR studies of human integral membrane proteins provide unique opportunities to probe structure and dynamics at specific locations and on multiple timescales, often with significant implications for disease mechanism and drug development. Since membrane proteins such as G protein-coupled receptors (GPCRs) are highly dynamic and regulated by ligands or other perturbations, NMR methods are potentially well suited to answer basic functional questions (such as addressing the biophysical basis of ligand efficacy) as well as guiding applications (such as novel ligand design). However, such studies on eukaryotic membrane proteins have often been limited by the inability to incorporate optimal isotopic labels for NMR methods developed for large protein/lipid complexes, including methyl TROSY. We review the different expression systems for production of isotopically labeled membrane proteins and highlight the use of the yeast Pichia pastoris to achieve perdeuteration and ¹³C methyl probe incorporation within isoleucine sidechains. We further illustrate the use of this method for labeling of several biomedically significant GPCRs.

Solid-state MAS NMR resonance assignment methods for proteins

The prerequisite to structural or functional studies of proteins by NMR is generally the assignment of resonances. Since the first assignment of proteins by solid-state MAS NMR was conducted almost two decades ago, a wide variety of different pulse sequences and methods have been proposed and continue to be developed. Traditionally, a variety of 2D and 3D ¹³C-detected experiments have been used for the assignment of backbone and side-chain ¹³C and ¹⁵N resonances. These methods have found widespread use across the field. But as the hardware has changed and higher spinning frequencies and magnetic fields are becoming available, the ability to use direct proton detection is opening up a new set of assignment methods based on triple-resonance experiments. This review describes solid-state MAS NMR assignment methods using carbon detection and proton detection at different deuteration levels. The use of different isotopic labelling schemes as an aid to assignment in difficult cases is discussed as well as the increasing number of software packages that support manual and automated resonance assignment.

Applications of solid-state NMR to membrane proteins

Membrane proteins mediate flow of molecules, signals, and energy between cells and intracellular compartments. Understanding membrane protein function requires a detailed understanding of the structural and dynamic properties involved. Lipid bilayers provide a native-like environment for structure-function investigations of membrane proteins. In this review we give a general discourse on the recent progress in the field of solid-state NMR of membrane proteins. Solid-state NMR is a variation of NMR spectroscopy that is applicable to molecular systems with restricted mobility, such as high molecular weight proteins and protein complexes, supramolecular assemblies, or membrane proteins in a phospholipid environment. We highlight recent advances in applications of solid-state NMR to membrane proteins, specifically focusing on the recent developments in the field of Dynamic Nuclear Polarization, proton detection, and solid-state NMR applications in situ (in cell membranes). This article is part of a Special Issue entitled: Biophysics in Canada, edited by Lewis Kay, John Baenziger, Albert Berghuis and Peter Tieleman.

Gd3+-chelated lipid accelerates solid-state NMR spectroscopy of seven-transmembrane proteins

Solid-state NMR (SSNMR) is an attractive technique for studying large membrane proteins in membrane-mimetic environments. However, SSNMR experiments often suffer from low efficiency, due to the inherent low sensitivity and the long recycle delays needed to recover the magnetization. Here we demonstrate that the incorporation of a small amount of a Gd³⁺-chelated lipid, Gd³⁺-DMPE-DTPA, into proteoliposomes greatly shortens the spin-lattice relaxation time (¹H-T ₁) of lipid-reconstituted membrane proteins and accelerates the data collection. This effect has been evaluated on a 30 kDa, seven-transmembrane protein, Leptosphaeria rhodopsin. With the Gd³⁺-chelated lipid, we can perform 2D SSNMR experiments 3 times faster than by diamagnetic control. By combining this paramagnetic relaxation-assisted data collection with non-uniform sampling, the 3D experimental times are reduced eightfold with respect to traditional 3D experiments on diamagnetic samples. A comparison between the paramagnetic relaxation enhancement (PRE) effects of Cu²⁺- and Gd³⁺-chelated lipids indicates the much higher relaxivity of the latter. Hence, a tenfold lower concentration is needed for Gd³⁺-chelated lipids to achieve comparable PRE effects to Cu²⁺-chelated lipids. In addition, Gd³⁺-chelated lipids neither alter the protein structures nor induce significant line-width broadening of the protein signals. This work is expected to be beneficial for structural and dynamic studies of large membrane proteins by SSNMR.

Nuclear magnetic resonance (NMR) applied to membrane–protein complexes

Increasing evidence suggests that most proteins occur and function in complexes rather than as isolated entities when embedded in cellular membranes. Nuclear magnetic resonance (NMR) provides increasing possibilities to study structure, dynamics and assembly of such systems. In our review, we discuss recent methodological progress to study membrane–protein complexes (MPCs) by NMR, starting with expression, isotope-labeling and reconstitution protocols. We review approaches to deal with spectral complexity and limited spectral spectroscopic sensitivity that are usually encountered in NMR-based studies of MPCs. We highlight NMR applications in various classes of MPCs, including G-protein-coupled receptors, ion channels and retinal proteins and extend our discussion to protein–protein complexes that span entire cellular compartments or orchestrate processes such as protein transport across or within membranes. These examples demonstrate the growing potential of NMR-based studies of MPCs to provide critical insight into the energetics of protein–ligand and protein–protein interactions that underlie essential biological functions in cellular membranes.