High-Pressure NMR Experiments for Detecting Protein Low-Lying Conformational States

Summary

We provide a detailed description of the steps required to assemble a high-pressure cell, set up and record high-pressure NMR experiments, and finally analyze both peak intensity and chemical shift changes under pressure. These experiments can provide valuable insights into the folding pathways and structural stability of proteins.

Abstract

High-pressure is a well-known perturbation method that can be used to destabilize globular proteins and dissociate protein complexes in a reversible manner. Hydrostatic pressure drives thermodynamical equilibria toward the state(s) with the lower molar volume. Increasing pressure offers, therefore, the opportunities to finely tune the stability of globular proteins and the oligomerization equilibria of protein complexes. High-pressure NMR experiments allow a detailed characterization of the factors governing the stability of globular proteins, their folding mechanisms, and oligomerization mechanisms by combining the fine stability tuning ability of pressure perturbation and the site resolution offered by solution NMR spectroscopy. Here we present a protocol to probe the local folding stability of a protein via a set of 2D 1H-15N experiments recorded from 1 bar to 2.5 kbar. The steps required for the acquisition and analysis of such experiments are illustrated with data acquired on the RRM2 domain of hnRNPA1.

Introduction

It has long been recognized that higher-energy, sparsely populated conformational states of proteins and protein complexes play a key role in many biological pathways^1,2,3. Thanks to experiments based on Carr-Purcell-Meiboom-Gill (CPMG)⁴, Chemical Exchange Saturation Transfer (CEST)⁵, and dark-state exchange saturation transfer (DEST)⁶ pulse sequences (among others), solution NMR spectroscopy has emerged as a method of choice for characterizing transient conformational states⁷. Along with these experiments, perturbations such as temperature, pH, or chemical denaturants can be introduced to increase the relative population of higher energy conformational substates. Similarly, protein equilibria can also be perturbed by applying high hydrostatic pressure. Depending on the magnitude of the volume change associated with the corresponding conformational changes, an increase of pressure by a few hundred to a few thousand bars can significantly stabilize a higher energy state or cause a protein to completely unfold^8,9,10. Protein NMR spectra typically display two types of changes with hydrostatic pressure: (i) chemical shift changes and (ii) peak intensity changes. Chemical shift changes reflect changes at the protein surface-water interface and/or local compression of the protein structure on a fast time scale (relative to NMR time scale)¹¹. Crosspeaks exhibiting large non-linear chemical shifts pressure dependence can indicate the presence of higher energy conformational states^12,13. On the other hand, peak intensity changes point to major conformational transitions on a slow time scale, such as changes in folded/unfolded state populations. The presence of folding intermediates or higher energy states can be detected from large variations in the magnitude of the volume change upon unfolding measured for different residues of a given protein^14,15,16,17. Based on our experience, even small proteins that are typically classified as two-state folders exhibit non-uniform responses to pressure, which provides useful information about their local folding stability. Described here is a protocol for the acquisition and analysis of amide peak intensity and ¹H chemical shifts pressure dependence, using as a model protein the isolated RNA recognition motif 2 (RRM2) of the heterogeneous nuclear ribonucleoprotein A1 (hnRNPA1).

Protocol

NOTE: The protocol described here requires (i) a high-pressure pump and cell with a 2.5 kbar rated aluminum-toughened zirconia tube¹⁸, (ii) the software SPARKY¹⁹ for analysis of the NMR spectra, and (iii) a curve fitting software.

1. Sample preparation, assembly of the high-pressure cell, and setting up the experiments.

1. Choice of buffer: Use equal mixture of anionic and cationic buffers, such as phosphate and Tris^20,21.
   NOTE: The pKa of anionic buffers such as phosphate and MES is associated with a substantial reaction volume (i.e., the difference in the partial molal volumes of the acid and the ionized products). The pH of such buffers can therefore be significantly affected by a change of pressure (~0.25-0.5 pH unit/kbar).
2. Ensure that the required sample volume is similar to that of a standard 3 mm diameter NMR tube (~300μL).
3. Introduce the ¹⁵N-labeled sample with a glass pipette into the zirconia tube. Make sure the sample seats at the bottom of the tube. Complete with 200 μL of mineral oil to prevent the sample from mixing with the transmission liquid (e.g., water). Fill the rest of the tube with transmission liquid.
4. Put a single-use O-ring on top of the zirconia tube and slide the tube into the base (Figure 1A,B). Then, connect the tube to the high-pressure tether line and tighten the base to the cell first by hand. Then, apply 14.7 Nm of torque to prevent leaks at lower pressure (Figure 1C,D).
5. To check the integrity of the pressure cell assembly, pressurize the tube up to 300 bar outside the spectrometer using cell support and containment vessel. Wait for 15 min, reset the pressure to 1 bar and check for leaks with a clean lint-free wipe.
6. Insert the unpressurized tube into the spectrometer by carefully guiding the tether line. Slide the tube in the spectrometer until reaching the sample sitting position (Figure 1E).
7. Lock, shim, match, and tune the ¹H and ¹⁵N channels as usual.
   NOTE: Shims for high-pressure rated zirconia tubes are very different from standard NMR tubes. It is recommended to save the optimized shims for future use.
8. Set up a ¹H-¹⁵N-HSQC or TROSY-HSQC and record a reference experiment at atmospheric conditions (1bar).

2. Recording high-pressure NMR experiments

1. Gradually increase the pressure from 1 bar to 2.5 kbar with 500 bar increments to test the overall stability of the protein. Set the speed of the pressure pump by default at ~18 bar/s. If the precise folding/unfolding rates are not known, let the sample equilibrate 15-20 min after each 500-bar increment. Record a spectrum at 2.5 kbar.
2. Gradually decrease the pressure back to 1 bar with 500 bar steps to test the reversibility of the pressure perturbation. Record another spectrum at atmospheric conditions and compare the chemical shifts and peak intensities with that of the reference spectrum previously recorded in the same conditions.
   NOTE: If the native crosspeaks are more intense after the pressure run, it may indicate that small aggregates present in solution at atmospheric pressure may have dissociated and properly refolded. On the other hand, a loss in the intensity or significant chemical shift changes suggest that the protein may experience a non-reversible misfolding in high-pressure conditions.
3. Record a series of 2D experiments from 1 bar to 2.5 kbar every 500 bar. It is recommended to record additional experiments near the inflection point of the folding/unfolding transition to improve the precision of the fit.

3. Analyzing peak intensity changes

1. Process all the spectra and transfer the backbone assignment from the reference spectrum at 1 bar to the spectrum recorded at 500 bar, and then transfer the assignment from 500 bar to 1 kbar and so forth.
   NOTE: Because pressure induces a non-uniform shift of ¹H and ¹⁵N chemical shifts, do not simply copy the backbone assignment from one spectrum to the next. Adjust it manually.
2. In Sparky menu, click on Peak > Peak List (lt). In the Peak List window, click on Options and select the option to display both Frequencies (ppm) and Data Height. Save the list obtained for each spectrum.
3. In a curve fitting software, copy the crosspeak identity and peak intensity values to have the pressure values (in bar) as the X-axis variable and the intensity as the Y-axis variable.
4. If a complete or near complete (>80%) unfolding is observed, fit the individual peak intensity profiles to extract the free energy and volume change upon unfolding, respectively, using a simple two-state model:
                                        Eq. 1
   Where, "I" is the observed intensity of a crosspeak at a given pressure p and I_F is the intensity of the same crosspeak in a fully folded state. R is the gas constant, T is the absolute temperature, ΔG_U⁰ the standard Gibbs free energy difference between the unfolded and folded states at atmospheric pressure p₀ (1 bar), and ΔV_u the volume change upon unfolding. With the pressure p in bar and temperature T in kelvin, R = 1.987 cal/K, ΔG_U⁰ is in cal/mol, and ΔV_U is in cal/mol/bar. Multiply the ΔV_U values obtained from the fit by 41.84 to convert into mL/mol. ΔV_U values are typically in the range of -50 to -150 mL/mol for globular proteins²². Here, all the parameters are expressed with regard to the unfolding reaction but can easily be converted into folding reaction parameters (ΔG_U⁰ = -ΔG_F⁰ and ΔV_U = -ΔV_F).

4. Analyzing chemical shift changes

1. Arrange the columns in the software in order to have pressure points as variable and the ¹H chemical shifts extracted from Sparky lists as the Y-axis.
2. Fit the pressure dependence of the ¹H chemical shifts to a simple quadratic equation:
   δ(p) = δ₀(p₀) + B₁(p-p₀) + B₂ (p-p₀)²                    Eq. 2
   Where, δ(p) is the measured ¹H chemical shift of a crosspeak at a given pressure p and δ₀(p₀) the ¹H chemical shifts of the same crosspeak in the reference spectrum recorded at 1 bar. B₁ and B₂ represent the first and second-order parameters expressed in ppm/bar and ppm/bar², respectively.

Representative Results

The protocol described here was used to probe the pressure dependence of RRM2, the second RNA recognition motif of hnRNPA1 (residues 95-106), which is almost completely unfolded within the 2.5 kbar range (>90%). ¹H-¹⁵N spectra were collected at 1 bar, 500 bar, 750 bar, 1 kbar, 1.5 kbar, 2 kbar, and 2.5 kbar (Figure 2). Since none of the native crosspeaks were visible above the noise level at 2.5 kbar, all corresponding residues were attributed an intensity value of 0 at this pressure (Figure 3A). A total of 45 individual pressure intensity profiles were fitted according to Equation 1 (Eq. 1) to obtain the corresponding changes in standard Gibbs free-energy (ΔG_U⁰) and volume (ΔV_U) associated with the unfolding reaction (Figure 3A). The calculated ΔV_U values ranged from -41 to -88 mL/mol and ΔG_U⁰ from 1.8-3.3 kcal/mol. When mapped on the domain structure (pdb 1U1R), it appears that residues with the largest magnitude of volume change are found within the domain structural core (in red, in Figure 3B), while those with the smallest magnitude of volume change are mainly located in the loop connecting the β-strands 2 and 3, and the loop connecting β-strand 4 to the C-terminal helix (in green, Figure 3B). The pressure dependence of the native backbone ¹H chemical shifts was also analyzed to probe the degree of compressibility and conformational heterogeneity of the folded state ensemble. Individual profiles of ¹H chemical shifts as a function of pressure were fitted using Eq. 2 in order to extract site-specific linear (B₁) and non-linear (B₂) coefficients (Figure 3C). The residues with the largest non-linear coefficients were mostly found in the structural core of the domain (in yellow, Figure 3D), while residues with the smallest non-linear coefficients were mostly located within loops connecting the different structural motifs (in blue, Figure 3D).

Figure 1: High-pressure cell assembly and installation in the spectrometer. (A-D) The full high-pressure cell assembly requires a zirconia tube, a single-use O-ring, the cell base (A,B). The tube is connected to the tether line by tightening the base to the cell (C,D). 14.7 Nm of torque is required to prevent leaks at lower pressure. (E) The zirconia tube connected to the Xtreme-60 Syringe pump via the tether line is introduced into the spectrometer until reaching the normal sample sitting position. Please click here to view a larger version of this figure.

Figure 2: Unfolding of hnRNPA1 RRM2 domain under pressure. ¹H-¹⁵N HSQC spectra collected at (A) 1 bar, (B) 1,000 bar, (C) 1,500 bar, (D) and 2,500 bar show a complete unfolding of the RRM2 domain under pressure. Please click here to view a larger version of this figure.

Figure 3: Analysis of RRM2 peak intensity and ¹H chemical shifts pressure dependence. (A) Representative peak intensity profiles recorded for RRM2 between 1 and 2,000 bar. Solid lines represent the fit obtained using Eq. 1 to calculate the corresponding standard free-energy and volume changes associated with the unfolding reaction. (B) The top 25% of residues with the largest measured magnitude of ΔV_U are highlighted with red spheres on RRM2 reference structure (pdb 1U1R). Residues with the smallest magnitude of ΔV_U are shown in green. (C) Representative changes of ¹H chemical shifts measured for RRM2 between 1 bar and 2,000 bar. Solid lines represent the fit to Eq. 2 to calculate the linear (B₁) and non-linear (B₂) pressure coefficients. (D) The top 25% of residues with the largest B₂ coefficients are shown in yellow, while those with the smallest B₂ coefficients are shown in blue. Please click here to view a larger version of this figure.

Discussion

This study details a protocol implemented to probe protein structural and thermodynamics responses to pressure perturbation. The high-pressure experiments recorded here on RRM2 demonstrate that large variations in ΔV_U values, indicative of non-fully cooperative unfolding, can be found in a relatively small single domain protein. A similar picture emerges from the analysis of ¹H chemical shift changes under pressure. It should be noted Kalbitzer and coworkers have demonstrated that a more in-depth analysis of chemical shift changes can be performed linking the ratio between the non-linear and linear coefficients (B₂/B₁) with the ratio of the change in compressibility and volume (Δβ/ΔV)²³. These results highlight the ability of high hydrostatic pressure to destabilize individual structural motifs and subdomains based on the local thermodynamic stability and volume change associated with unfolding. Recording a simple set of pressure titration experiments can therefore provide valuable information regarding the folding mechanisms and subdomain architecture of a protein.

The whole pressure system (pump and controller) is relatively compact and can be installed on a standard utility cart for easy access to the spectrometer (Figure 1). All pulse sequences can be used under pressure without modification, including triple resonance experiments. The only limitation is that pressure-rated zirconia tubes reduce the sensitivity of NMR experiments by almost 50% compared to a standard 3 mm diameter NMR tube. This may present a challenge for protein samples that can't be concentrated to a high-enough level. Compared to other perturbation methods such as temperature or chemical denaturants, hydrostatic pressure presents the advantage of being, with very few exceptions, fully reversible. Because pressure favors the dissociation of protein complexes, there is, for example, very little risk of sample aggregation, a problem often encountered with high temperatures.

While hydrostatic pressure is a powerful tool for studying protein stability, it is important to keep in mind that there is no perfect method of perturbation. Each of them highlights the specificity of a protein free-energy landscape in a different way. For example, if two conformational states have the same molar volume, an increase in hydrostatic pressure will not significantly change their relative populations. Similarly, if two conformational states have a similar exposed surface area, adding chemical denaturants would not necessarily increase the relative population of one state compared to the other one. Ideally, different perturbation methods should be combined to obtain a complete description of a protein free-energy landscape^15,24. Finally, it should be noted that although the manuscript is focused on the pressure dependence of intensity and chemical shifts from series of 2D ¹H-¹⁵N experiments measured at equilibrium, a wide range of NMR experiments can be coupled to pressure perturbation to probe different aspects of the stability, the folding mechanism, and the conformational dynamics of proteins and proteins complexes^20,25.

Materials

Name	Company	Catalog Number	Comments
Bruker Nmr Cell 2.5 Kbar	Daedalus Innovations LLC	NMRCELL-B
Sparky3	University of California San Francisco, CA	N/A
Xtreme-60 Syringe pump	Daedalus Innovations LLC	XTREME-60