Determination of protein conformation and orientation at buried solid/liquid interfaces

Abstract

Protein structures at solid/liquid interfaces mediate interfacial protein functions, which are important for many applications. It is difficult to probe interfacial protein structures at buried solid/liquid interfaces in situ at the molecular level. Here, a systematic methodology to determine protein molecular structures (orientation and conformation) at buried solid/liquid interfaces in situ was successfully developed with a combined approach using a nonlinear optical spectroscopic technique - sum frequency generation (SFG) vibrational spectroscopy, isotope labeling, spectra calculation, and computer simulation. With this approach, molecular structures of protein GB1 and its mutant (with two amino acids mutated) were investigated at the polymer/solution interface. Markedly different orientations and similar (but not identical) conformations of the wild-type protein GB1 and its mutant at the interface were detected, due to the varied molecular interfacial interactions. This systematic strategy is general and can be widely used to elucidate protein structures at buried interfaces in situ.

Fig. 1. (a) Crystal structure of WT GB1 (PDB ID: 3gb1). (b) Crystal structure of WT GB1 with Q32 and N35 plotted in magenta sticks. (c) Schematic of the SFG prism geometry used in this study to collect SFG spectra from the PS/protein solution interfaces.

Fig. 2. SFG ssp spectra collected from proteins (a) WT NL, (b) WT Leu, (c) MT NL, (d) MT Leu, (e) MT Val, (f) MT Phe, (g) MT Lys and (h) MT Ile adsorbed at the PS/protein solution interfaces. Black dots are experimental data points and red lines are fitted spectra.

Fig. 3. SFG ppp spectra collected from proteins (a) WT NL, (b) WT Leu, (c) MT NL, (d) MT Leu, (e) MT Val, (f) MT Phe, (g) MT Lys and (h) MT Ile absorbed at the PS/protein solution interfaces. Black dots are experimental data points and red lines are fitted spectra.

Fig. 4. (a) Final score map of the spectral matching between the reconstructed experimentally collected WT GB1 SFG spectra (after deconvoluting the non-resonant contribution in ssp and ppp spectra and PS contribution in the ppp spectra) and the calculated WT GB1 SFG spectra as a function of protein orientation based on the simulated WT GB1 structure at 781 ns. The orientations at (30°, 50°) and (150°, 230°) in the map possess the highest matching scores (0.66). The spectral comparisons (b) between the reconstructed WT NL SFG spectra and the calculated WT NL spectra (based on the simulated structure at 781 ns) and (c) between the reconstructed WT Leu SFG spectra and the calculated WT Leu spectra (based on the simulated structure at 781 ns) at an orientation of (30°, 50°) (or (150°, 230°)) which have the highest matching scores. The orientation visualizations of (d) WT GB1 (with the 781 ns simulation structure) at (30°, 50°) and (e) WT GB1 (with the 781 ns simulation structure) at (0°, 0°). The (0°, 0°) orientation is the protein orientation obtained from the MD simulation result without further rotating the protein.

Fig. 5. (a) Final score map of spectral matching between the reconstructed experimentally collected MT GB1 SFG spectra (after deconvoluting the non-resonant contribution in ssp and ppp spectra and PS contribution in the ppp spectra) and the calculated MT GB1 SFG spectra as a function of orientation based on the simulated MT GB1 structure at 972 ns. The orientations at (30°, 100°) and (150°, 280°) shown in the map possess the highest matching score (0.19). The spectral comparisons between the reconstructed experimental spectra and calculated spectra using the simulated MT structure at 972 ns with an orientation of (30°, 100°) (or (150°, 280°)) for (b) MT NL, (c) MT Leu, (d) MT Val, (e) MT Phe, (f) MT Lys and (g) MT Ile. The orientation visualizations of (h) MT GB1 with a simulated structure at 972 ns with the most likely orientation at (30°, 100°) and (i) MT GB1 with a simulated structure at 972 ns without rotation at (0°, 0°).

Fig. 6. Orientation comparisons of GB1 at the PS/protein solution interfaces: (a) and (b) the best matched orientations of WT GB1 based on the SFG data analysis using the GB1 crystal structure, (c) the best matched WT GB1 orientation after rotating all the simulated structures (based on the simulated structure at 781 ns) – replotted from Fig. 4(d). (d) and (e) The best matched orientations of MT GB1 based on the SFG data analysis using the GB1 crystal structure, (f) the best matched MT GB1 orientation after rotating all the simulated structures (based on the simulated structure at 972 ns) – replotted from Fig. 5(h). (g) The best matched WT GB1 orientation from the simulated structure without rotation (based on the simulated structure at 821 ns), and (h) the best matched MT GB1 orientation from the simulated structure without rotation (based on the simulated structure at 887 ns).

Fig. 7. Orientation visualizations of (a) the simulated WT GB1 structure at 781 ns of (30°, 50°) and (b) the simulated MT GB1 structure at 972 ns of (30°, 100°). Residues Q32 and N35 are shown in magenta sticks. Residues F30, Y33 and N37 are shown in red sticks.