A high-field cellular DNP-supported solid-state NMR approach to study proteins with sub-cellular specificity

Abstract

Studying the structural aspects of proteins within sub-cellular compartments is of growing interest. Dynamic nuclear polarization supported solid-state NMR (DNP-ssNMR) is uniquely suited to provide such information, but critically lacks the desired sensitivity and resolution. Here we utilize SNAPol-1, a novel biradical, to conduct DNP-ssNMR at high-magnetic fields (800 MHz/527 GHz) inside HeLa cells and isolated cell nuclei electroporated with [¹³C,¹⁵N] labeled ubiquitin. We report that SNAPol-1 passively diffuses and homogenously distributes within whole cells and cell nuclei providing ubiquitin spectra of high sensitivity and remarkably improved spectral resolution. For cell nuclei, physical enrichment facilitates a further 4-fold decrease in measurement time and provides an exclusive structural view of the nuclear ubiquitin pool. Taken together, these advancements enable atomic interrogation of protein conformational plasticity at atomic resolution and with sub-cellular specificity.

Fig. 1. (A): Confocal z-stack of FITC-SNAPol-1 (green) treated HeLa cells counterstained with Hoechst (blue) along with an illustrative FITC-SNAPol-1 intensity profile taken along the arrow depicted in the green panel. (B/C): SNAPol-1 DNP enhancement on 8 million HeLa cells electroporated with 1.2 mM [¹³C,¹⁵N] Ub treated with 30 mM SNAPol-1 embedded in “DNP Juice” (6 : 4 d8, ¹²C₃-enriched glycerol : D₂O with 1 × Hank's buffered salt solution HBSS) as previously described in ref. . In (C), polarization buildup (T_B) on the carbonyl signal was best fit by a biexponential with error bars denoting the standard deviation (s.d.) of the signal to noise ratio (S/N) and Sy.x the s.d. residual of the fit. See ESI for more details on error analysis.

Fig. 2. (A) Overlay of 2D SQSQ planes extracted from ¹³C–¹³C–¹³C DQSQSQ 3D experiments recorded on 8 million HeLa cells electroporated with 1.2 mM [¹³C,¹⁵N] Ub and doped with 30 mM AMUPol, at 400 MHz (orange) and with 30 mM SNAPol-1 at 800 MHz (blue). 1D projections detail linewidth improvement (full width at half maximum (FWHM)). (B) Ubiquitin sequence showing the increase in assignments between Ub in-cell samples measured at 400 (25 residues) and 800 MHz (41 residues) (assignment required 2 or more resolved correlations).

Fig. 3. (A) Confocal z-stack of cell nuclei stained with Hoechst (blue) treated FITC-SNAPol-1 (green). (B) Confocal z-stack of cell nuclei isolated from HeLa cells electroporated with 1.2 mM of TAMRA-Ub (red) stained with Hoechst (blue). Panel (C/D) SNAPol-1 DNP enhancement (judged on Ub carbonyl signal 170 to 185 ppm) on cell nuclei isolated from 16 million HeLa cells electroporated with 1.2 mM [¹³C,¹⁵N] Ub and treated with 30 mM SNAPol-1 dissolved in “DNP Juice” (6 : 4 d8, ¹²C₃-enriched glycerol : D₂O with 1 × HBSS) as previously described in ref. . Polarization buildup (T_B) on the carbonyl signal was best fit by a mono-exponential with Sy.x denoting the residual s.d. of the fit. Error bars are not visible due to high S/N ratio. See also Table S1.

Fig. 4. Panel (A) overlay of aliphatic 2D projections and 1D slices extracted from ¹³C–¹³C–¹³C DQSQSQ 3D experiment conducted at 800 MHz on 8 million HeLa cells (blue) electroporated with 1.2 mM [¹³C,¹⁵N]-Ub and on 1.6 × 10⁷ isolated cell nuclei (red), both doped with 30 mM SNAPol-1. Black crosses denote literature assignments (BMRB ID: 15410 and 7111). Panel (B) sequence showing the assigned Ub residues at 800 MHz in isolated cell nuclei (49 residues) versus whole cells (41 resides). Panel (C) comparison of experimental 2D projections for L50 and T66 Cα/β overlayed with literature secondary chemical-shift values (with crosses denoting the average secondary structure chemical-shift) taken from ref. .