High-performance solvent suppression for proton detected solid-state NMR

Abstract

High-sensitivity proton detected experiments in solid-state NMR have been recently demonstrated in proton diluted proteins as well as fully protonated samples under fast magic-angle spinning. One key element for performing successful proton detection is effective solvent suppression achieved by pulsed field gradients (PFG) and/or saturation pulses. Here we report a high-performance solvent suppression method that attenuates multiple solvent signals simultaneously by more than a factor of 10,000, achieved by an optimized combination of homospoil gradients and supercycled saturation pulses. This method, which we call Multiple Intense Solvent Suppression Intended for Sensitive Spectroscopic Investigation of Protonated Proteins, Instantly (MISSISSIPPI), can be applied without a PFG probe. It opens up new opportunities for two-dimensional heteronuclear correlation spectroscopy of hydrated proteins at natural abundance as well as high-sensitivity and multi-dimensional experimental investigation of protein-solvent interactions.

Figure 1

MISSISSIPPI-based solid-state ¹⁵N-¹H heteronuclear correlation pulse sequence employing homospoil gradients and saturation pulses for efficient suppression of multiple solvent signals. Optionally, RFDR (the π pulse train and bracketing π/2 pulses on ¹H channel) or spin-diffusion (omitting the π pulses) can be used to study protein-solvent interactions as well as proton-proton distance constraints within the protein [12, 20]. For 2D acquisition, States-TPPI phase increment was applied to the π/2 pulse immediately following t₁ [23, 24]. XY-16 phases [25] are used for RFDR pulses; other phases are labeled above the pulses. Narrow and wide filled rectangles represent π/2 and π pulses, respectively.

Figure 2

Proton spectra demonstrating the solvent suppression efficiency of MISSISSIPPI. (a) Spectrum excited by a single π/2 pulse. (b)-(e) Spectra with MISSISSIPPI suppression and ¹⁵N-editing. Two transients for (a), (b) and (d), and 16 transients for (c) and (e) were acquired. The 10 ms free induction decays in (b) and (c) were truncated to 3.4 ms for (d) and (e), respectively. (f) For comparison, a 256-transient spectrum was acquired using a previous solvent suppression method [13], also truncated to 3.4 ms. Spectr (a)-(f) were acquired on a 750 MHz spectrometer with 36 kHz sample spinning rate, for 0.9 μmol natural abundance nanocrystalline protein GB1 precipitated with methyl-2,4-pentanediol and isopropanol. (g) For another comparison, a 4-transient spectrum was acquired using MISSISSIPPI on a GB1 sample uniformly ²H,¹³C,¹⁵N-enriched, back-exchanged with H₂O; free induction decay evolved to 14 ms. No apodization was applied to any spectra and intensity per transient is plotted for all spectra.

Figure 3

The investigation of protein-solvent interactions through high-sensitivity proton detection. ¹⁵N-edited spectra using either (a) spin-diffusion or (b) RFDR ¹H-¹H mixing to establish protein solvent correlations; the spectra in grey are the corresponding control spectra collected with ¹⁵N channel blanked off. The spectra were acquired with four transients on a 750 MHz spectrometer with 39 kHz sample spinning rate, for 0.9 μmol protein GB1 that was uniformly ²H,¹³C,¹⁵N-enriched, back-exchanged with H₂O, and precipitated with methyl-2,4-pentanediol (MPD) and isopropanol (IPA). Relatively weak π pulses with half the nutation frequency of the π/2 ¹H pulses are used during RFDR. Recycle delay of 1.5 was used in (a) but 3 s in (b) to reduce duty cycle; 4 transients were acquired for each spectrum. Free induction decays were truncated to 14 ms before Fourier transform without apodization.

Figure 4

Trajectories for protein and solvent signals with (a) spin diffusion and (b) RFDR ¹H-¹H mixing. The build-up curves (open symbols, right vertical axis) are plotted in a different scale from the decay curves (filled symbols, left axis). The curves for amide protons (integration over 11 −6 ppm) and all protons (11 −0.9 ppm) were fitted with a single exponential decay model; the curves for water (5.3 −4.6 ppm), alcohol (4.3 −3.9 ppm) and methyl (1.4 −0.9 ppm) signals were fitted with a single exponential growth model in (a) and a dual growth-decay exponential model in (b). Fit parameters are listed in Table 1. Data with a 1-ms time step (τ_M) were acquired and fitted for the curves but only every third point is shown for clarity.