Synthesis and evaluation of 4-(α-arylvinyl)pyridines as activated linker groups for application in antibody–drug conjugates

Abstract

With the aim of increasing the cysteine-selective bioconjugation reactivity of 4-vinylpyridines, a series of 4-(α-arylvinyl)pyridines were synthesised by Suzuki–Miyaura cross-coupling of 4-(α-bromovinyl)pyridine. The rate of thia-Michael addition of cysteine-containing glutathione correlated with σ_x of the aryl substituent (ρ = +0.94). Introduction of electron-donating and electron-withdrawing substituents at the 3-position of 2,6-dimethyl-4-vinylpyridine all resulted in a reduction in the rate of glutathione bioconjugation, and a shift from the optimum pK_a value of approximately 6–7 with the pK_a values determined by pH gradient NMR titration. Potential ‘dual-armed’ 2,6-disubstituted-4-vinylpyridine antibody–drug linker groups developed previously were modified by the α-vinyl arylation methodology. This resulted in an increase in the rate of glutathione bioconjugation of approximately an order of magnitude for the p-nitrophenyl derivatives, providing opportunities for the development of new antibody–drug conjugates.

---

Introduction

Antibody–drug conjugates (ADCs) are therapeutic agents combining the target-specific binding of an antibody with a potent cytotoxic payload.¹ Both components are covalently attached to a linker group, and for the antibody this typically exploits chemoselective bioconjugation of a nucleophilic cysteine-derived thiol with an electrophilic Michael acceptor.² Maleimide is used as the latter in several clinically approved ADCs, but these are prone to disassembly by retro-Michael addition.³ Selective cysteine-based conjugation of proteins has also been achieved by exploiting the β-vinyl carbon electrophilicity of 2- and 4-vinylpyridines,⁴ and ester functionalised 4-vinylpyridine 1 (Fig. 1) is an ADC linker significantly less prone to reversible thiol addition.^5,6 This intrinsic stability is often a desirable feature for the design of ADC therapeutics, as deconjugation of the potent cytotoxic drug during plasma circulation can increase systemic toxicity and reduce on-target exposure leading to a narrower therapeutic index.^7,8 Related vinylheterocycles investigated for bioconjugation include vinylpyrimidines,⁹ divinylpyrimidines¹⁰ and N-methylated 2-vinylpyridinium salts.¹¹

Fig. 1 4-Vinylpyridine based ADC linker groups 1 ^5,6 and 2.¹²

With the objective of delivering high-DAR (drug–antibody ratio) conjugates, we recently reported a series of difunctionalised ‘dual-armed’ derivatives of 4-vinylpyridine, capable of linking two cytotoxic payloads per conjugation moiety.¹² Model thia-Michael bioconjugation reactions using glutathione, a cysteine-containing tripeptide, revealed that of the many 2,6-disubstituted derivatives synthesised, only diester 2 approached the reactivity of 1.

With a view to increasing the rate of bioconjugation of ‘dual-armed’ 4-vinylpyridine-based linkers, we reasoned that the latent polarity of the vinylpyridine core (A – Fig. 2) pointed to the introduction of an electron-withdrawing group (EWG) at either the α- (B) or 3-position (C) as a means of increasing the electrophilicity of the β-vinyl carbon. This paper describes our work on the synthesis and model reactivity studies of such derivatives, leading to the identification of an optimised ‘dual-armed’ and activated 4-vinylpyridine ADC linker.

Fig. 2 The latent polarity of a 2,6-disubstituted 4-vinylpyridine (A) and positioning of an electron-withdrawing group (EWG) for greater β-carbon electrophilicity (B and C).

Results and discussion

Alkenes containing electron-withdrawing substituents are prone to polymerisation,¹³ and at the start of this programme it was conspicuous that 4-(α-nitrovinyl)pyridine was not known, and there was only a single, characterisation free, report of 4-(α-trifluoromethylvinyl)pyridine 5.¹⁴ We obtained a low yield of 5 by palladium catalysed cross-coupling of 4-pyridineboronic acid pinacol ester 3 and 2-bromo-3,3,3-trifluoropropene 4 (Scheme 1),¹⁵ followed by pyridinium salt precipitation rather than chromatography as a means of purification. Following liberation of 5 with aqueous sodium hydrogen carbonate, it was found to be highly prone to decomposition on storage, rendering analysis difficult and further testing impractical.

Scheme 1 Synthesis of 4-(α-trifluoromethylvinyl)pyridine 3. (i) PdCl₂(dppf) (10 mol%), Cs₂CO₃ (3 eq.), DME/H₂O, 80 °C, 8 h. (ii) 2 M HCl/Et₂O – filtration. (iii) Aq. NaHCO₃ – extraction with CH₂Cl₂.

As an alternative to the direct attachment of a strong electron-withdrawing substituent, we reasoned that introduction of an α-aryl group would enable tuning of the reactivity profile of the 4-vinylpyridine component by variation of one or more aryl substituents. Although various methodologies for the synthesis of 4-(α-aryllvinyl)pyridines have been reported,¹⁶ palladium-catalysed cross-coupling appeared to offer the simplest approach. To this end we first synthesised 4-(α-bromovinyl)pyridine 9 in 73% yield overall as previously reported (Scheme 2).¹⁷ As found for 5, pyridine 9 also proved to be highly reactive, decomposing rapidly as evidenced by the formation of a black solid when left to stand either in solution or neat at −20 °C in the dark. To minimise this 9 was used in a subsequent reaction immediately following generation.

Scheme 2 Synthesis of 4-(α-bromovinyl)pyridine 9.¹⁷ (i) TfOH (1.1 eq.), Et₂O, 0 °C, 1 h. (quant.). (ii) Br₂ (2 eq.), CHCl₃, 0 °C to RT, 2 h. (86%). (iii) NEt₃ (3 eq.), MeCN, 0 °C to RT, 3 h. (85%).

Two methods of palladium catalysis were used: starting either with Pd(OAc)₂ and employing the water soluble triarylphosphine ligand TPPTS 11 [conditions (i) – Scheme 3], or with Pd(PPh₃)₄ [conditions (ii)]. The former aided the isolation of the potentially polymerisation-sensitive cross-coupled product without the use of column-chromatography, although all of 10a–10o, containing both electron-withdrawing and electron-donating substituents, were sufficiently stable to enable full-characterisation.

Scheme 3 Synthesis of 4-(α-arylvinyl)pyridines 10a–10o. (i) Pd(OAc)₂ (3–5 mol%), 11 (6–10 mol%), K₂CO₃ (3 eq.), 1 : 1 dioxane/H₂O, 90 °C, 16 h. (ii) Pd(PPh₃)₄ (10 mol%), K₂CO₃ (3 eq.), 5 : 1 dioxane/H₂O, 90 °C, 16 h.

With these compounds in hand we then determined the rate of glutathione (GSH) addition to 10a and all of the meta and para mono-substituted aryl derivatives (excepting 10m – Scheme 4). As previously developed,¹² this model bioconjugation reaction¹⁸ utilised one equivalent of the cysteine-containing tripeptide as a solution in D₂O (0.22 M), added to the vinylpyridine in DMSO-d₆ (0.073 M). Pyrazine was included as an internal standard for integration with the reaction monitored using ¹H NMR spectroscopy. The rates of this second order addition reaction^10,11 are given in Table 1.

Scheme 4 Reaction of 4-(α-arylvinyl)pyridines with glutathione (GSH) for the determination of rate as a function of the position and identity of the substituent X.

Table 1 Rate of reaction of selected 4-(α-arylvinyl)pyridines 10 with glutathione

Entry	Substrate (X)	kobs/M^−1 min^−1	Entry	Substrate (X)	kobs/M^−1 min^−1
1	10a (H)	0.088	2	10d (p-NO2)	0.480
3	10e (m-NO2)	0.245	4	10h (m-CN)	0.230
5	10i (p-CN)	0.577	6	10j (p-OMe)	0.033
7	10k (m-OMe)	0.080	8	10l (p-F)	0.127
9	10n (p-Me)	0.065	10	10o (m-Me)	0.086

---

For comparison, the rate of addition of GSH to 4-vinylpyridine was determined as 0.14 M⁻¹ min⁻¹. Thus the introduction of an α-phenyl substituent is deactivating (entry 1, 10a), but this can be more than compensated for by the introduction of a para-nitro (10d, entry 2) or para-cyano (10i, entry 5) substituent, with the latter giving an approximately four-fold increase in rate over 4-vinylpyridine. A plot of log(k_x/k_H) vs. Hammett substituent constant (σ_x) further confirmed the influence of a substituent's identity and position on reactivity (Fig. 3), and the value of the reaction constant (ρ = +0.94) is consistent with increased electron-density in the transition state of rate-determining S–C bond forming conjugate addition. Furthermore, the relative electrophilicity of a substrate, as measured by β-carbon ¹³C NMR chemical shift,¹⁹ correlates with σ_x (Fig. 4) and by extension with the rate of glutathione addition.²⁰

Fig. 3 Plot of log(k_x/k_H) vs. Hammett substituent constant (σ_x) for glutathione addition to 4-(α-arylvinyl)pyridines 10 (ρ = 0.94, R² = 0.89).

Fig. 4 Plot of β-carbon ¹³C chemical shift vs. Hammett substituent constant (σ_x) 4-(α-arylvinyl)pyridines 10 (R² = 0.97).

We next examined the influence of a substituent at the 3-position, and for this study it was convenient to start with 4-bromo-2,6-dimethylpyridine 11a due to the simplicity of mono-nitration to give 11b (Scheme 5). Subsequent reduction to 11c was followed by diazotisation/fluorination, methylation or acetylation to give 11d, 11e and 11f respectively. Subsequent Suzuki–Miyaura cross-coupling with potassium vinyltrifluoroborate¹² gave vinylpyridines 12a–12f. The rate of GSH addition to each of these was determined as before (Table 2).

Scheme 5 Synthesis of 4-vinyl-2,6-dimethylpyridine 12a and 3-substituted derivatives 12b–f. (i) HNO₃ (3 eq.), H₂SO₄, 100 °C, 16 h. (40%). (ii) Fe (12 eq.), EtOH, HCl_(aq), Δ, 2 h. (99%). (iii) HBF₄ (2 eq.), NaNO₂ (1 eq.), H₂O, 0 °C, PhMe, Δ, 16 h. (34%). (iv) 37% CH₂O (2.5 eq.), HCO₂H (6.5 eq.), Δ, 16 h (46%). (v) Ac₂O (1.2 eq.), AcCl (2.6 eq.), NEt₃ (2.2 eq.), CH₂Cl₂, RT, 2 d. (59%). (vi) Potassium vinyltrifluoroborate (2 eq.), NEt₃ (3 eq.), PdCl₂(dppf) (5 mol%), 2 : 1 PhMe/i-PrOH, 90 °C.

Table 2 Rate of GSH addition, ¹³C NMR and pK_a analysis of vinylpyridines 12a–f

Entry	Substrate (X)	kobs/M^−1 min^−1	β-Carbon^a ^13C (ppm)	pKa
a CDCl3.b Not determined due to instability.c Not determined due to poor aqueous solubility (see SI).
1	12a (H)	1.2	118.2	7.03 ± 0.08
2	12b (NO2)	0.073	122.7	2.83 ± 0.04
3	12c (NH2)	0.49	119.1	7.44 ± 0.07
4	12d (F)	0.28	120.5	ND^b
5	12e (NMe2)	0.3	117.3	7.30 ± 0.07
6	12f (NAc2)	0.2	121.8	ND^c

---

The value of k_obs (M⁻¹ min⁻¹) obtained for 12a (entry 1, X = H) is similar to the values reported previously for 2,6-disubstituted vinylpyridines 1 (0.80) and 2 (0.75). None of the 3-substituted derivatives improved the rate, and the reactivity of 12b containing the strongly electron-withdrawing nitro-substituent was much reduced. As for the 4-(α-arylvinyl)pyridines 10 discussed above, the β-carbon ¹³C chemical shifts (Table 2) increase in value with the increasing electron-withdrawing nature of the substituent X.²¹ However, in contrast to the α-(4-nitrophenyl) derivative 10d, the introduction of a nitro substituent at the 3-position to give 12b decreases rather than increases reactivity.

To examine this further we used our previously developed pH gradient NMR titration methodology for pK_a determination (Table 2).¹² In this earlier study on 2,6-disubstituted-4-vinylpyridines a linear relationship was observed between the rate of GSH addition and pK_a within the range of 1 (low reactivity) to 6–7 (high reactivity) [for 1 (pK_a = 6.18) and 2 (pK_a = 6.59)]. With pK_a > 7 the reactivity decreased. The hypothesis that pyridine basicity principally controls reactivity was rationalised by participation in rate-determining conjugate addition of the corresponding pyridinium species, with the reactivity of this protonated intermediate attenuated by electron-donating substituents (where pK_a > 7).¹² This mechanistic proposal is supported by the similar outcomes observed for the 3-substituted derivatives 12, and in particular the decreased reactivity of 12b as a result of the decrease in pK_a to 2.83 (Table 2, entry 2). This analysis also explains the higher reactivity of 2,6-dimethyl-4-vinylpyridine 12a (rate = 1.2) compared to less basic 4-vinylpyridine (rate = 0.14) for which a pK_a of 5.6 has been determined.²²

The above results revealed that improved bioconjugation reactivity in a ‘dual-armed’ vinylpyridine-based linker was most likely to result from the introduction of a para-nitrophenyl or para-cyanophenyl moiety at the α-position. In addition to 2, we also applied this adaptation to 2,6-diester 13 (pK_a = 5.36, rate = 0.17 M⁻¹ min⁻¹).¹² The methodology described above for the synthesis of 10 was found to be applicable to both compounds (Scheme 6). In each case the α-bromo derivative generated by elimination was not isolated but instead was used immediately in a Suzuki–Miyaura cross-coupling to give α-aryl derivatives 14–17.

Scheme 6 Synthesis of activated functionalised ‘dual-armed’ 4-vinylpyridine derivatives 14–17. (i) TfOH (1.1 eq.), CHCl₃, 0 °C, 30 min. (ii) Br₂ (2 eq.), CHCl₃, RT. (iii) NEt₃ (3 eq.), MeCN, RT. (iv) p-XC₆H₄B(OH)₂ (1.5 eq.), PdCl₂(PPh₃)₂ (5 mol%), K₂CO₃ (3 eq.), 5 : 1 dioxane/H₂O, 90 °C.

The determination of rates of GSH addition as before (Table 3) revealed an approximately order of magnitude increase in reactivity for the para-nitrophenyl containing derivatives 14 and 16 (entries 1 and 3) relative to their corresponding parents (2 and 13 respectively). Significantly, the similarity of the pK_a values of 14 (6.21, entry 1) and 2 (6.59)¹² points to the 8-fold increase in the rate of GSH addition to the former being the result of increased β-carbon electrophilicity without the para-nitrophenyl substituent significantly attenuating pyridine basicity. This is similarly the case for 16 (pK_a 5.37, entry 3) compared to 13 (5.36).¹² The rate increases for the para-cyanophenyl derivatives 15 and 17 are smaller (entries 2 and 4). These differences between the influence of the p-nitro and p-cyano substituents are reflected in the higher β-carbon ¹³C chemical shift and higher pK_a value of the compounds containing the former.

Table 3 Rate of GSH addition, ¹³C NMR and pK_a analysis of activated functionalised 4-vinylpyridine derivatives 14–17 and diaryl derivative 18

Entry	Substrate (X)	kobs (M^−1 min^−1)	Rate increase	β-Carbon^a ^13C (ppm)	pKa ^b	pKa change
a CDCl3.b Of the pyridine Moiety. The pKa of the amine component of 15 was determined as 2.69 ± 0.05 (3.12 for 2 ^12).c Compared to 2 (0.75).d Compared to 13 (0.17).e Compared to 2 (6.59).f Compared to 13 (5.36).
1	14 (NO2)	5.8	8^c	119.3	6.21 ± 0.03	−0.38^e
2	15 (CN)	2.7	4^c	118.8	6.11 ± 0.013	−0.48^e
3	16 (NO2)	2.3	14^d	119.5	5.37 ± 0.12	0.01^f
4	17 (CN)	0.4	3^d	118.8	4.79 ± 0.05	−0.57^f
5	18	2.6	NA	120.2	NA	NA

---

Finally, as the Hammett substituent constant (σ) for the 4-pyridyl group (0.94)²³ is similar to σ_p for the nitro group (0.78) we reasoned that 1,1-di(para-nitrophenyl)ethene 19 ²⁴ may also display reactivity towards GSH addition. This is the case (entry 5), although with a rate significantly less than that of 14. Thus the presence of strong electron-withdrawing groups in both aryl substituents of a 1,1-diarylethene results in the thia-Michael reaction of the glutathione cysteine component, a reaction accelerated further by a sufficiently basic pyridine moiety.

Conclusions

The rate of thia-Michael addition of cysteine-containing glutathione to 4-(α-arylvinyl)pyridines, a model bioconjugation reaction, increases with the electron-withdrawing capacity of the substituted aryl group. The highest rates, which are approximately 3–4 times faster than 4-vinylpyridine, were obtained with a p-cyanophenyl or p-nitrophenyl aryl moiety. In contrast, introduction of a 3-nitro substituent to 2,6-dimethyl-4-vinylpyridine results in a significant decrease in rate. In both cases the introduction of an electron-withdrawing group increases the electrophilicity of the β-vinyl carbon of the Michael acceptor as determined by ¹³C NMR chemical shift. However, in the latter case the introduced 3-nitro group also significantly decreases pyridine basicity, and thus the percentage of the corresponding activated pyridinium ion generated by acid catalysis. Application of the methodology developed for the synthesis of the 4-(α-arylvinyl)pyridines enabled the direct modification of previously developed ‘dual-armed’ 2,6-disubstituted-4-vinylpyridine ADC linkers.¹² The rate of model bioconjugation to the new 4-(α-arylvinyl) derivatives increases by up to an order of magnitude with a p-nitrophenyl group. These results provide opportunities for protein bioconjugation validation as the next step in the development of new ADCs.

Author contributions

TES, GRS, MW and CJR conceived the project. JA and JMS prepared the compounds and performed the kinetic analysis under the supervision of JT, PLKH, GRS, CJR and TES. MW designed and performed the pH gradient NMR titration experiments. CJR wrote the manuscript, with suggestions from all other authors.

Conflicts of interest

There are no conflicts to declare.

Data availability

Additional data supporting this article are available in the supplementary information (SI). Supplementary information: synthetic details, other experimental procedures, NMR spectra. See DOI: https://doi.org/10.1039/d6ob00614k.

Acknowledgements

Iksuda Therapeutics Ltd and the MRC are acknowledged for funding (JA, JMS). MW thanks UKRI for a Future Leaders Fellowship (MR/Z000114/1) and the Royal Commission for the Exhibition of 1851 for a Research Fellowship.

References

1. (a) S. J. Walsh, J. D. Bargh, F. M. Dannheim, A. R. Hanby, H. Seki, A. J. Counsell, X. Ou, E. Fowler, N. Ashman, Y. Takada, A. Isidro-Llobet, J. S. Parker, J. S. Carroll and D. R. Spring, Chem. Soc. Rev., 2021, 50, 1305–1353; (b) C. Dumontet, J. M. Reichert, P. D. Senter, J. M. Lambert and A. Beck, Nat. Rev. Drug Discovery, 2023, 22, 641–661.
2. P. Ochtrop and C. P. R. Hackenberger, Curr. Opin. Chem. Biol., 2020, 58, 28–36.
3. S. C. Alley, D. R. Benjamin, S. C. Jeffrey, N. M. Okeley, D. L. Meyer, R. J. Sanderson and P. D. Senter, Bioconjugate Chem., 2008, 19, 759–765.
4. R. Sebastiano, A. Citterio, M. Lapadula and P. G. Righetti, Rapid Commun. Mass Spectrom., 2003, 17, 2380–2386.
5. A. G. Watts, T. Kantner and A. B. Mackenzie, WO2010/070300A2, 2010.
6. A. D. Baxter, C. M. Birchall, D. J. Mansell, J. H. Mysliwy and J. Thirlway, WO2016/067021, 2016.
7. R. P. Lyon, J. R. Setter, T. D. Bovee, S. O. Doronina, J. H. Hunter, M. E. Anderson, C. L. Balasubramanian, S. M. Duniho, C. I. Leiske, F. Li and P. D. Senter, Nat. Biotechnol., 2014, 32, 1059–1062.
8. L. N. Tumey, M. Charati, T. He, E. Sousa, D. Ma, X. Han, T. Clark, J. Casavant, F. Loganzo, F. Barletta, J. Lucas and E. I. Graziani, Bioconjugate Chem., 2014, 25, 1871–1880.
9. H. Seki, S. J. Walsh, J. D. Bargh, J. S. Parker, J. Carroll and D. R. Spring, Chem. Sci., 2021, 12, 9060–9068.
10. S. J. Walsh, S. Omarjee, W. R. J. D. Galloway, T. T.-L. Kwan, H. F. Sore, J. S. Parker, M. Hyvönen, J. S. Carroll and D. R. Spring, Chem. Sci., 2019, 10, 694–700.
11. M. J. Matos, C. D. Navo, T. Hakala, X. Ferhati, A. Guerreiro, D. Hartmann, B. Bernardim, K. L. Saar, I. Compañón, F. Corzana, T. P. J. Knowles, G. Jiménez-Osés and G. J. L. Bernardes, Angew. Chem. Int. Ed., 2019, 58, 6640–6644.
12. M. Wallace, J. M. Sharpe, K. Baj, M. Ngwube, J. Thirlway, P. L. Kerigan Higgs, G. R. Stephenson, J. A. Iggo, T. E. Storr and C. J. Richards, Analyst, 2025, 150, 2827–2879.
13. G. Raffyson, S. de Freitas and C. L. Firme, J. Mol. Model, 2013, 19, 5267–5276.
14. H. Mizuno, WO2009/066786A1, 2009.
15. G. P. Lahm, D. Cordova, J. D. Barry, T. F. Pahutski, B. K. Smith, J. K. Long, E. A. Benner, C. W. Holyoke, K. Joraski, M. Xu, M. E. Schroeder, T. Wagerle, M. J. Mahaffey, R. M. Smith and M.-H. Tong, Bioorg. Med. Chem. Lett., 2013, 23, 3001–3006.
16. (a) S. Zhu, J. Qin, F. Wang, H. Li and L. Chu, Nat. Commun., 2019, 10, 1–7; (b) J. Liu, H. Yi, X. Zhang, C. Liu, R. Liu, G. Zhang and A. Lei, Chem. Commun., 2014, 50, 7636–7638.
17. J. J. Concepcion, D. K. Zhong, D. J. Szalda, J. T. Muckerman and E. Fujita, Chem. Commun., 2015, 51, 4105–4108.
18. Although it differs in steric and other local environmental factors that can influence cysteine–thiol reactivity in proteins, non-volatile and crystalline glutathione (GSH) is a convenient substrate for reactivity studies. It has been used previously as a model substrate for electrophilic covalent inhibitors of cysteine-containing proteins. See: (a) V. J. Cee, L. P. Volak, Y. Chen, M. D. Bartberger, C. Tegley, T. Arvedson, J. McCarter, A. S. Tasker and C. Fotsch, J. Med. Chem., 2015, 58, 9171–9178; (b) A. Birkholz, D. J. Kopecky, L. P. Volak, M. D. Bartberger, Y. Chen, C. M. Tegley, T. Arvedson, J. D. McCarter, C. Fotsch and V. J. Cee, J. Med. Chem., 2020, 63, 11602–11614.
19. H. Neuvonen and K. Neuvonen, J. Chem. Soc., Perkin Trans. 2, 1999, 1497–1502.
20. For β-carbon ¹³C NMR chemical shift vs. rate gives R² = 0.93 excluding 10d and 10i which are slower than this correlation predicts. See SI.
21. For 12a–f a plot of β-carbon ¹³C chemical shift vs. σ_m gives R² = 0.88 and vs. σ_p gives R² = 0.74 (see SI). These data highlight the relationship between the electronic properties of the 3-substituent and the β-vinyl carbon ¹³C chemical shift, but being ortho steric influences may also be significant.
22. A. Pietrzyk, R. Wiley and D. McDaniel, J. Org. Chem., 1957, 22, 83–84.
23. A σ value of 2.57 was also determined for the 4-pyridinium group. See; J. H. Blanch, J. Chem. Soc. B, 1966, 937–939.
24. R. Ramage, A. J. Blake, M. R. Florence, T. Gray, G. Raphy and P. L. Roach, Tetrahedron, 1991, 47, 8001–8024.

---