Multiplex Trifluoromethyl and Hydroxyl Radical Chemistry Enables High-Resolution Protein Footprinting

Abstract

Hydroxyl radical-based protein footprinting (HRPF) coupled with mass spectrometry is a valuable medium-resolution technique in structural biology, facilitating the assessment of protein structure and molecular-level interactions in solution conditions. In HRPF with X-rays (XFP), hydroxyl radicals generated by water radiolysis covalently label multiple amino acid (AA) side chains. However, HRPF technologies face challenges in achieving their full potential due to the broad (>10³) dynamic range of AA reactivity with ^•OH and difficulty in detecting slightly modified residues, most notably in peptides with highly reactive residues like methionine, or where all residues have low ^•OH reactivities. To overcome this limitation, we developed a multiplex labeling chemistry that utilizes both CF₃ radicals (^•CF₃) produced from a trifluoromethylation (TFM) reagent and OH radicals (^•OH), under controlled and optimized radiolysis doses generated by X-rays. We optimized the dual ^•CF₃/^•OH chemistry using model peptides and proteins, thereby extending the existing ^•OH labeling platform to incorporate simultaneous ^•CF₃ labeling. We labeled >50% of the protein sequence and >80% of protein solvent-accessible AAs via multiplex TFM labeling resulting in high-resolution footprinting, primarily by enhancing the labeling of AAs with low ^•OH reactivity via the ^•CF₃ channel, while labeling moderate and highly ^•OH-reactive AAs in both ^•CF₃ and ^•OH channels. Moreover, the low reactivity of methionine with ^•CF₃ enabled the detection and quantification of additional AAs labeled by ^•CF₃ within methionine-containing peptides. Finally, we found that the solvent accessibility of protein AAs directly correlated with ^•CF₃ labeling, demonstrating that multiplex TFM labeling enables a high-resolution assessment of molecular interactions for enhanced HRPF.

Figure 1.

Proposed pathway for TFM-based multiplex labeling of proteins in •CF₃/•OH channels. Water radiolysis generates •OH, and CF₃SO₂Na is used as the TFM reagent. For simplicity, direct attack of radicals on proteins is not shown.

Figure 2.

Optimized multiplex TFM labeling for six model peptides (x-axis) as a function of CF₃SO₂Na concentrations (0.75 mM-30 mM) and a constant absorbed •OH dose (32.2 × 10¹⁵, 76 μm Al). The Log10 of averaged modified fraction derived from (A) •CF₃ (solid bars) and (B) •OH (patterned bars) labeled peptides was plotted on y-axis. Individual peptides were irradiated for 20 ms and their titration curves are shown in SI Figure S1. Experiments were conducted in duplicate, and error bars represent the standard error.

Figure 3.

Optimized multiplex TFM labeling for six model peptides (x-axis) as a function of the absorbed •OH dose and a constant 2.5 mM CF₃SO₂Na concentration. The Log₁₀ of modified fraction (y-axis) for (A) •CF₃ (solid bars) and (B) •OH (patterned bars) labeled peptides was plotted against absorbed •OH doses ranging from 3.9 × 10¹⁵ to 53.9 × 10¹⁵ •OH. Individual peptides were irradiated for 20 ms and titration curves at 2.5 mM and 7.5 mM CF₃SO₂Na are shown in SI Fig S2.

Figure 4.

Amyl-B labeling in multiplex TFM experiment. Sample was exposed for 20 ms at 32.2 × 10¹⁵ •OH dose and 2.5 mM CF₃SO₂Na. (A) The extracted ion chromatogram displays •CF₃ (orange) and •OH (green and pink) adducts and the unmodified Amyl-B (blue). (B) The unmodified fraction (y-axis) of •OH-labeled and (C) •CF₃-labeled AAs in the Amyl-B were plotted against exposure time (x-axis) to generate dose-response curves. Rate constants for modified AAs are shown in SI Table S2.

Figure 5.

The Venn diagrams of (A) lysozyme and (B) CaM AAs labeled in optimized multiplex TFM labeling. The crystal structures display labeled AAs for (C) lysozyme (PDB 6LYZ) and (D) CaM (PDB 1CLL). The AAs modified during multiplex TFM labeling are categorized based on their detected adducts into three groups: purple (•CF₃-modifications only), blue (both •CF₃- and •OH-modifications), and green (•OH-modifications only). The lysozyme and CaM were exposed with X-rays for 20 ms on the 17-BM beamline.

Figure 6.

The frequencies of labeled AAs for (A) six model peptides (solid bars) and for (B) proteins (lysozyme and CaM) (patterned bars) were compared between optimized multiplex TFM (•CF3 and •OH) and optimized HRPF (•OH) labeling. The AAs were categorized according to their •OH reactivity.