Quaternization of Vinyl/Alkynyl Pyridine Enables Ultrafast Cysteine‐Selective Protein Modification and Charge Modulation

Abstract Quaternized vinyl‐ and alkynyl‐pyridine reagents were shown to react in an ultrafast and selective manner with several cysteine‐tagged proteins at near‐stoichiometric quantities. We have demonstrated that this method can effectively create a homogenous antibody–drug conjugate that features a precise drug‐to‐antibody ratio of 2, which was stable in human plasma and retained its specificity towards Her2+ cells. Finally, the developed warhead introduces a +1 charge to the overall net charge of the protein, which enabled us to show that the electrophoretic mobility of the protein may be tuned through the simple attachment of a quaternized vinyl pyridinium reagent at the cysteine residues. We anticipate the generalized use of quaternized vinyl‐ and alkynyl‐pyridine reagents not only for bioconjugation, but also as warheads for covalent inhibition and as tools to profile cysteine reactivity.

Electrophile concentration was 3.0 mM in all cases. A 1 H NMR spectrum was recorded every 85 s (number of scans: 16). Around 5 min were needed to record the first spectrum after mixing the reagents. The observed second-order rate constants kobs (i.e. k2) were derived from the slope of a linearly-fitted plot of the inverse of the electrophile concentration (1/[E]) versus time. Figure S1. a) Monitoring of reaction between 1 (blue) and PrSH (green) in sodium phosphate buffer in D2O (pH 7.6, 100 mM) by 1 H NMR (400 MHz) at 298 K. The corresponding adduct is shown in orange. b) Estimation of the second-order reaction constant (kobs) using the linearlyfitted regions of the 1/ [1]) versus time plot. c) Overlay of 1 H NMR spectra at different reaction times. Blue, green and orange arrows point to signals of compounds 1, PrSH and the corresponding adduct, respectively. d) Relative ratio (%) of compounds 1, PrSH and the reaction adduct at different reaction times. e) Estimation of the intrinsic nucleophilic rate constant ( "#$% & ) from the observed kinetic rate constant derived from the initial measurements (kobs) (in blue), thiol acidity constant (Ka) and buffer acidity ([H + ]).   Figure S8. a) Monitoring of reaction between 2 (blue) and Ac-Cys-NH2 (green) in sodium phosphate buffer in D2O (pH 7.6, 100 mM) by 1 H NMR (400 MHz) at 298 K. The corresponding adduct is shown in orange. b) Estimation of the second-order reaction constant (kobs) using the linearly-fitted region of the 1/ [2]) versus time plot. c) Overlay of 1 H NMR spectra at different reaction times. Blue, green and orange arrows point to signals of compounds 2, Ac-Cys-NH2 and the corresponding adduct, respectively. d) Relative ratio (%) of compounds 2, Ac-Cys-NH2 and the reaction adduct at different reaction times. e) Estimation of the intrinsic nucleophilic rate constant ( "#$% & ) from the observed kinetic rate constant derived from the two initial measurements (kobs) (in blue), cysteine thiol acidity constant (Ka) and buffer acidity ([H + ]). Figure S9. a) Monitoring of reaction between NEM (blue) and Ac-Cys-NH2 (green) in sodium phosphate buffer in D2O (pH 7.6, 100 mM) by 1 H NMR (400 MHz) at 298 K. The corresponding adduct is shown in orange. b) Estimation of the second-order reaction constant (kobs) using the linearly-fitted region of the 1/[NEM]) versus time plot. c) Overlay of 1 H NMR spectra at different reaction times. Blue, green and orange arrows point to signals of compounds NEM, Ac-Cys-NH2 and the corresponding adduct, respectively. d) Relative ratio (%) of compounds NEM, Ac-Cys-NH2 and the reaction adduct at different reaction times. e) Estimation of the intrinsic nucleophilic rate constant ( "#$% & ) from the observed kinetic rate constant derived from the two initial measurements (kobs) (in blue), cysteine thiol acidity constant (Ka) and buffer acidity ([H + ]). Figure S10. a) Monitoring of reaction between 2 (blue) and Ac-Lys-NH2 (green) in sodium phosphate buffer in D2O (pH 7.6, 100 mM) by 1 H NMR (400 MHz) at 298 K. The corresponding adduct is shown in orange. b) Estimation of the second-order reaction constant (kobs) using the linearly-fitted region of the 1/ [2]) versus time plot. c) Overlay of 1 H NMR spectra at different reaction times. Blue, green and orange arrows point to signals of compounds 2, Ac-Lys-NH2 and the corresponding adduct, respectively. d) Relative ratio (%) of compounds 2, Ac-Lys-NH2 and the reaction adduct at different reaction times. e) Estimation of the intrinsic nucleophilic rate constant ( "#$'( ) ) from the observed kinetic rate constant (kobs) (in blue), lysine ammonium acidity constant (Ka) and buffer acidity ([H + ]).

pH 5.7
The second-order reaction constants of the reactions of electrophiles 1-4 with small-molecule model N-acetylcysteine amide (Ac-Cys-NH2) were determined by 1 H NMR (400 MHz) at 298 K in sodium phosphate buffer in D2O (pH 5.7, 100 mM). Electrophile concentration was 3.0 mM in all cases. A 1 H NMR spectrum was recorded every 85 s (number of scans: 16). Around 5 min were needed to record the first spectrum after mixing the reagents. The observed second-order rate constants kobs (i.e. k2) were derived from the slope of a linearly-fitted plot of the inverse of the electrophile concentration (1/[E]) versus time. Figure S11. a) Monitoring of reaction between 1 (blue) and Ac-Cys-NH2 (green) in sodium phosphate buffer in D2O (pH 5.7, 100 mM) by 1 H NMR (400 MHz) at 298 K. The corresponding adduct is shown in orange. b) Estimation of the second-order reaction constant (kobs) using the linearly-fitted regions of the 1/[1] versus time plot. c) Overlay of 1 H NMR spectra at different reaction times. Blue, green and orange arrows point to signals of compounds 1, Ac-Cys-NH2 and the corresponding adduct, respectively. d) Relative ratio (%) of compounds 1, Ac-Cys-NH2 and the reaction adduct at different reaction times. e) Estimation of the intrinsic nucleophilic rate constant ( "#$% & ) from the observed kinetic rate constant derived from all the measurements (kobs) (in blue), cysteine thiol acidity constant (Ka) and buffer acidity ([H + ]). Figure S12. a) Monitoring of reaction between 2 (blue) and Ac-Cys-NH2 (green) in sodium phosphate buffer in D2O (pH 5.7, 100 mM) by 1 H NMR (400 MHz) at 298 K. The corresponding adduct is shown in orange. b) Estimation of the second-order reaction constant (kobs) using the linearly-fitted regions of the 1/[2] versus time plot. c) Overlay of 1 H NMR spectra at different reaction times. Blue, green and orange arrows point to signals of compounds 2, Ac-Cys-NH2 and the corresponding adduct, respectively. d) Relative ratio (%) of compounds 2, Ac-Cys-NH2 and the reaction adduct at different reaction times. e) Estimation of the intrinsic nucleophilic rate constant ( "#$% & ) from the observed kinetic rate constant derived from the initial measurements (kobs) (in blue), cysteine thiol acidity constant (Ka) and buffer acidity ([H + ]). Figure S13. a) Monitoring of reaction between 3 (blue) and Ac-Cys-NH2 (green) in sodium phosphate buffer in D2O (pH 5.7, 100 mM) by 1 H NMR (400 MHz) at 298 K. The corresponding adduct is shown in orange. b) Estimation of the second-order reaction constant (kobs) using the linearly-fitted regions of the 1/ [3] versus time plot. c) Overlay of 1 H NMR spectra at different reaction times. Blue, green and orange arrows point to signals of compounds 3, Ac-Cys-NH2 and the corresponding adduct, respectively. d) Relative ratio (%) of compounds 3, Ac-Cys-NH2 and the reaction adduct at different reaction times. e) Estimation of the intrinsic nucleophilic rate constant ( "#$% & ) from the observed kinetic rate constant derived from all the measurements (kobs) (in blue), cysteine thiol acidity constant (Ka) and buffer acidity ([H + ]). Figure S14. a) Monitoring of reaction between 4 (blue) and Ac-Cys-NH2 (green) in sodium phosphate buffer in D2O (pH 5.7, 100 mM) by 1 H NMR (400 MHz) at 298 K. The corresponding adduct is shown in orange. b) Estimation of the second-order reaction constant (kobs) using the linearly-fitted regions of the 1/ [4] versus time plot. c) Overlay of 1 H NMR spectra at different reaction times. Blue, green and orange arrows point to signals of compounds 4, Ac-Cys-NH2 and the corresponding adduct, respectively. d) Relative ratio (%) of compounds 4, Ac-Cys-NH2 and the reaction adduct at different reaction times. e) Estimation of the intrinsic nucleophilic rate constant ( "#$% & ) from the observed kinetic rate constant derived from the initial measurements (kobs) (in blue), cysteine thiol acidity constant (Ka) and buffer acidity ([H + ]).

The effect of pH on reactivity
Pyridine protonation might, in principle, accelerate the reaction due to nitrogen quaternization; however, thiol protonation would slow it down due to the decreased concentration of nucleophilic thiolate in solution. Therefore, as pH becomes more acidic, the concentration of the more reactive pyridinium electrophile increases while the concentration of the more reactive thiolate nucleophile decreases. Hence, the reaction outcome in terms of kinetics would strongly depend on the relative pKa's of both reagents, providing an opportunity to fine-tune reactivity and, potentially, chemo-and site-selectivity. Figure S15. Predominant electrophilic (pyridine/pyridinium) and nucleophilic (thiol/thiolate) species depending on their protonation state as a function of the environmental pH.
To test this concept, additional calculations on the nucleophilic addition of 1-propanethiolate anion to protonated versions of small-molecule models 1 and 3 (labelled as 1' and 3', respectively) were first performed. Of note, PrSwas able to spontaneously deprotonate alkynyl pyridinium 3', giving a largely thermodynamically stable neutral complex (3'_preTS_SMe). The activation barrier from this complex to a zwitterionic -and probably unrealistic-transition state (3'_TS_SMe) was calculated to be slightly higher than that calculated for the unprotonated derivative (ΔG ‡ = 26.1 kcal mol -1 for 3' vs. ΔG ‡ = 24.3 kcal mol -1 for 3).
These computational predictions suggest that alkynyl pyridines should not be very reactive at slightly acidic pH likely due to their higher acidity, while vinyl pyridines should be quite reactive do to (partial) quaternisation.
We verified these predictions by performing the reactions between electrophiles 1-4 and cysteine model Ac-Cys-NH2 at pH 5.7 (see section 1. Reaction kinetics pH 5.7). In agreement with the calculations, vinyl pyridine 1 was significantly more reactive at a slightly acidic than at neutral pH, achieving a 61% yield of the addition product within 1 h (10% at pH 7.6). On the other hand, alkynyl pyridine 3 was much less reactive (0-1% yield of addition product at both pH 5.7 and 7.6) than the vinyl pyridine analogue. S18

Quantum Mechanical Calculations
Computational Details. Full geometry optimizations were carried out with Gaussian 16 [1] using the M06-2X hybrid functional [2] and 6-31+G(d,p) basis set in combination with ultrafine integration grids. Bulk solvent effects in water were considered implicitly through the IEF-PCM polarizable continuum model. [3] The possibility of different conformations was taken into account. Frequency analyses were carried out at the same level used in the geometry optimizations, and the nature of the stationary points was determined in each case according to the appropriate number of negative eigenvalues of the Hessian matrix. The quasiharmonic approximation reported by Trular et al. was used to replace the harmonic oscillator approximation for the calculation of the vibrational contribution to enthalpy and entropy. [4] Scaled frequencies were not considered. Mass-weighted intrinsic reaction coordinate (IRC) calculations were carried out by using the Gonzalez and Schlegel scheme [5] in order to ensure that the TSs indeed connected the appropriate reactants and products. Gibbs free energies (ΔG) were used for the discussion on the relative stabilities of the considered structures. Free energies calculated using the gas phase   ionization. The mobile phases are 95% aqueous acetonitrile with 0.05% formic acid and 10 mM ammonium acetate with 0.1% formic acid. The separation technology is based on a 50x4.6 mm C18 column (currently a Phenomenex Kinetix solid core column). There are several methods available enabling the user to produce mass spectra for compounds up to 2kDa in positive and negative modes of ionization. In some cases, a Waters LCT Premier combined with an Agilent 1100 autosampler was also used. The system runs using 50% aqueous acetonitrile with 0.25% formic acid as mobile phase and can measure accurate masses from 150 Da to 1500 Da. UV spectra for recording absorption curves and kinetics of the reactions were recorded using a Cary 300 UV spectrometer. CD measurements were made in a 1 mm cuvette using Applied Photophysics' Chirascan CD spectrometer equipped with a Quantum TC125 temperature control unit 25 ºC.

Synthesis of compound 5
Scheme S1. Schematics of the steps towards the synthesis of 5.

Synthesis of H-VC-PAB-MMAE.
This procedure is a modification of a literature protocol. [7]

General procedure for protein and antibody conjugation with quaternized vinyl pyridine derivatives
To an eppendorf tube with NaPi (50 mM or 20 mM, pH 8.0) and DMF (10% of total volume), an aliquot of a stock solution of protein (final concentration 10 μM) was added. Afterwards, a solution of the quaternized pyridine derivative (1 to 10 equiv.) in DMF was added and the resulting mixture was vortexed for 10 seconds. The reaction was mixed for 1 or 2 h at 37 ºC. A 10 μL aliquot of each reaction time was analysed by LC-MS and conversion to the expected product was observed.

LC-MS method for analysis of protein conjugation
LC-MS was performed on a Xevo G2-S TOF mass spectrometer coupled to an Acquity UPLC system using an Acquity Q6 UPLC BEH300 C4 column (1.7 mm, 2.1 × 50 mm). Solvents A, a water with 0.1% formic acid and B, 71% acetonitrile, 29% water and 0.075% formic acid were used as the mobile phase at a flow rate of 0.2 mL min -1 . The gradient was programmed as follows: 72% A to 100% B after 25 min then 100% B for 2 min and after that 72% A for 18 min. The electrospray source was operated with a capillary voltage of 2.0 kV and a cone voltage of 40 V. Nitrogen was used as the desolvation gas at a total flow of 850 L h -1 . Total mass spectra were reconstructed from the ion series using the MaxEnt algorithm preinstalled on MassLynx software (v. 4.1 from Waters) according to the manufacturer's instructions. To obtain the ion series described, the major peak(s) of the chromatogram were selected for integration and further analysis.

Stability of bioconjugates in human plasma
A 20 μL aliquot of the bioconjugate (10 μM) in NaPi buffer (20 mM, pH 8.0) was thawed. 1 μL of reconstituted human plasma was added at room temperature and the resulting mixture vortexed for 10 seconds. The resulting reaction mixture was then mixed at 37 ºC. After 1 and 48 h, a 10 μL aliquot of each reaction mixture was analysed by LC-MS.

Annexin V-315C Modification and Characterization
Annexin

Analysis of Secondary Structural Content by CD
Circular dichroism (CD) spectroscopy was used to analyse protein secondary structure in solution. Samples were concentrated to 100 nM in NaPi buffer (50 mM, pH 8.0). CD measurements were made using Applied Photophysics' Chirascan CD spectrometer equipped with a Quantum TC125 temperature control unit 25 ºC. The data was acquired in a 1 mm cuvette path length with a response time of 1 s, a per-point acquisition delay of 5 ms and a pre and postscan delay of 50 ms. Spectra were averaged over three scans, in a wavelength range from 190 nm to 260 nm, and the spectrum from a blank sample containing only buffer was subtracted from the averaged data.

Determination of FcRn Binding by Surface Plasmon Resonance (SPR)
A BIAcore 3000 instrument (GE Healthcare) was used with CM5 sensor chips coupled with mFcRn (cynomolgus monkey FcRn) or hFcRn (1,000 resonance units) using amine coupling chemistry as described by the manufacturer. Langmuir model (BIAcore AB). Data was reference cell adjusted and zero-adjusted.

Microfluidic Determination of Electrophoretic Mobility
Device fabrication.
The electrophoresis device was fabricated using standard soft lithography techniques [11] of poly(dimethylsiloxane) (PDMS, Dow Corning) using SU-8 on silicon wafer master. To decrease the autofluorescence of the PDMS carbon nanopowder (13 nm, Plasmachem GMBH) was added to the PDMS mixture prior to curing. In addition, the channels were sealed with quartz microscope glass (Alfa Aesar, 76.2x25.4x1.0mm). This was done by plasma treated both PDMS and quartz surfaces with oxygen plasma for 30 s (Electronic Diener Femto, 40% power for 30 seconds) and heating at 95 ºC for a 1 min. An additional hydrophobic treatment was performed right before the experiments with oxygen plasma (500 s with 80 % power).

Microfluidic electrophoresis experiments.
The experiments were conducted with microfluidic free flow electrophoresis device [12] withdrawing at 400 rate from the fluid outlets while pushing the electrode solution at 65 from the electrode inlets. Buffer and the analyte sample were loaded to the device using gel and an electric field is simply from the measured voltage in the channel and device diameter: The voltage in the channel was calculated by measuring the voltage efficiency of each device. This was done by comparing the measured current of the buffer and highly conductive