Forcing the reversibility of a mechanochemical reaction

Mechanical force modifies the free-energy surface of chemical reactions, often enabling thermodynamically unfavoured reaction pathways. Most of our molecular understanding of force-induced reactivity is restricted to the irreversible homolytic scission of covalent bonds and ring-opening in polymer mechanophores. Whether mechanical force can by-pass thermodynamically locked reactivity in heterolytic bimolecular reactions and how this impacts the reaction reversibility remains poorly understood. Using single-molecule force-clamp spectroscopy, here we show that mechanical force promotes the thermodynamically disfavored SN2 cleavage of an individual protein disulfide bond by poor nucleophilic organic thiols. Upon force removal, the transition from the resulting high-energy unstable mixed disulfide product back to the initial, low-energy disulfide bond reactant becomes suddenly spontaneous, rendering the reaction fully reversible. By rationally varying the nucleophilicity of a series of small thiols, we demonstrate how force-regulated chemical kinetics can be finely coupled with thermodynamics to predict and modulate the reversibility of bimolecular mechanochemical reactions.

for fragmentation peak values. (c) CID of +14 Da peak at 11076.5 Da also identifies fragment ion series corresponding to the I27E24C-k55C construct. b-ions corresponding to the N-terminal region of the sequence localise the +14 Da mass addition to the His-tag region (the protein purification tag). Blue separators indicate identified b-and y-ions in the fragmentation spectra. The oxidised cysteine residues are highlighted in grey and the possible methylation event (+14 Da) is indicated in green. See supplementary Table 5 for fragmentation peak values. The methylation of I2725/55 is present in the first 7 residues -very likely to be in the second Arg from the N-terminal, which is in the protein purification histidine tag-does not affect the percentage of disulfide bond reduction. Empty circles represent value from individual experiment (n=3 for each compound; error bars: s.d).

Supplementary
Supplementary figure 20. The percentage of disulfide bond reduction by penicillamine depends on the protein:nucleophile ratio. (a) The percentage of reduced I2724/55 containing a single penicillamine modification depends on the concentration of deprotonated penicillamine. The percentage is obtained from the raw intensity, with no background subtraction, and therefore is likely to be a large overestimation of the actual percentage. The intensity of the 11266 Da peak (blue arrow) in the deconvoluted mass spectrum decreases with decreasing protein:penicillamine ratio. Empty circles represent value from individual experiment (n=3 for each condition; error bars: s.d) (b) 1:54 (c) 1:16, and (d) 1:6.

Supplementary figure 21. The kinetics of disulfide bond reduction is largely dominated by the concentration of deprotonated thiols. (a)
We conducted the Ellman's assay at two pH values using the same total concentration (0.02 mM) of NAC, thereby modulating the concentration of the deprotonated species. At pH = 7.5 (dark green), approximately 0.62% of NAC is in the deprotonated form and the reduction of all DTNB molecules is almost instantaneous. By contrast, using the same assay performed at pH = 2.5, where 0.000006% of NAC molecules are deprotonated, we do not measure any change in absorbance over 10 mins. (b) A two-force protocol first unfolds the protein structure (grey) and then measures the kinetics of disulfide bond reduction in the presence of NAC at different pH values (green coloured). All trajectories are acquired at the same total concentration of NAC (32mM), however in each case the concentration of the deprotonated species is modulated by adjusting the pH. The fastest kinetics (dark green) corresponds to pH = 7.5, whereas the slowest (light green) corresponds to pH = 5.8. (c) Investigating the reformation percentage at tq = 1s, tq = 8s, and tq = 16s for four different compounds, spanning a wide range of pKa values (8.4 -9.8), demonstrates that while a kinetic barrier is present, the tq = 8s quench time used throughout the manuscript is enough to overcome the kinetic barrier, hence probing the thermodynamics of the reverse reaction (From 1s quench to 16s nNAC-ME=34, 24 and 19; n1mercapto-2propanol=19, 39 and 19; all error bars: s.d).

Supplementary Figure 24. The correlation between the energetics of the mechanochemical reaction and the reversibility is recapitulated in the homologous protein I2732/75. (a)
Disulfide bond rupture by penicillamine in the structurally homologous protein I2732/75, which contains a single disulfide bond between residues 32 and 75, is hallmarked by steps of ∼15 nm. (b) The relationship between the energetics of the disulfide rupture and the subsequent reversibility observed in I2724/55 (circles) is observed in I2732/75 (triangles), albeit with a lower disulfide yield. (c) The same trend is observed when considering the pKa of the attacking thiol (From left to right bond reformation events of I2732/75 n= 20, 20 and 13; all error bars: s.d).

Supplementary Figure 25. The percentage of proteins that successfully refold in the absence of a disulfide bond is well predicted by DFT calculations. (a)
The 5-pulse mechanochemical assay used in this study can readily distinguish whether a protein has refolded with or without a disulfide bond in the structure. While disulfide reformation is identified in the test pulse by steps of ∼15 nm followed by ∼10 nm, when a protein folds void of a disulfide is hallmarked by ∼25 nm steps (inset: red asterisks). This scenario occurs when a free thiolate in solution attacks the mixed disulfide, removing the modification from the protein and forming a homodimer species. As such, the protein harbours two reduced cysteines and is able to effectively refold into its native structure, harbouring a mechanical stability that is intermediate between a misfolded conformation harbouring a mixed disulfide bond, and a stiff protein form that hallmarks a successfully refolded and oxidised protein. The occurrence of the attack of the mixed disulfide by solution thiolates can be rationalized by the results of DFT free energy calculations using either the (b) M062X, or (c) B3LYP functionals (from left to right n= 31, 18, 30, 38, 49, 21, 30, 31 and 16; all error bars: s.d).

Supplementary Figure 26. Cysteine-methyl-ester displays time dependent dimerization. (a)
Mass spectrometry measurements immediately after the preparation of cysteine-methylester (0.2 mM) reveal that the majority of the compound is in monomeric form (m/z = 136 Da). (b) Mass spectrometry measurements taken 25 hours after solution preparation reveal that cysteine-methyl-ester has completely dimerized (m/z = 269 Da). (c) This dimerization is readily observed in the single molecule experiments. Disulfide rupture traces obtained immediately after cysteine-methyl-ester preparation (green), consistently display all 8 rupture events within the first 10 s (red arrows). However, traces obtained after 2.5 hours (grey), reveal that while 6 protein unfolding events are observed (inset: 15 nm steps, black asterisks), only 2 disulfide bond rupture events subsequently occur (red arrows). This dramatic slowing of the rate of disulfide cleavage signifies a reduction in the concentration of attacking species. For this reason, our single molecule exeriments were limited to ∼3 hours. There is a dramatic reduction in the rate of disulfide cleavage between the start of the experiment (green) and after 13 hours (grey). For this reason, our single molecule exeriments were limited to ∼3 hours.

Supplementary Tables
Supplementary Table 1