Room-temperature electrochemical hydroxyalkylation of pyridines

Abstract

Direct and selective functionalization of ubiquitous pyridine scaffold is of paramount importance across numerous fields, yet remains challenging due to the inherent inertness of pyridines and the presence of competing reactive sites. Here, using an undivided cell equipped with a zinc cathode, a graphite anode, and a TEMPO mediator, we report a room-temperature electrochemical strategy that enables the direct hydroxyalkylation of diverse pyridines with readily available carbonyl compounds. This method features a broad substrate scope, operational simplicity, metal-free conditions, and high step/atom economy. It overcomes the limitations of conventional approaches requiring pre-functionalized substrates or stoichiometric activators, establishing a practical platform for direct access to C4- or C2-hydroxyalkyl pyridines, depending on the substitution pattern of the pyridine substrates. Mechanistic studies reveal that acetic acid activates both reactants, and the products are formed via cross-coupling of cathodic reduction-induced pyridyl radicals and hydroxy α-radicals followed by protonation and anodic oxidative aromatization.

Direct and selective functionalization of pyridine scaffold is of importance across numerous fields, yet remains challenging due to the inherent inertness of pyridines and the presence of competing reactive sites. Here, the authors report a room-temperature electrochemical strategy that enables the direct hydroxyalkylation of diverse pyridines with carbonyl compounds.

Introduction

Pyridine represents the most prevalent N-heterocyclic scaffold in FDA-approved pharmaceuticals¹ and serves as a key structural motif in numerous agrochemicals, functional materials, ligands, and natural products^2–6. Among its derivatives, 4-hydroxyalkylated pyridines constitute a important subclass due to their distinctive properties and broad applications^7–14. As illustrated in Fig. 1A, representative examples include pyridoxine (1, a natural form of vitamin B6)¹¹, a CCR6 antagonist (2) for autoimmune diseases⁷, and LY2874455 (3), a commercially available second-generation FGFR inhibitor used in cancer therapy⁸. Additionally, compounds 4 and 5 function as a surface-grafting agent for MOF-based polymers and a precursor for polymer-supported catalysts, respectively^12,13. Despite their significance, synthetic routes to 4-hydroxyalkyl pyridines remain notably scarce. Conventional syntheses typically rely on either initial preparation of C4-halopyridines followed by lithiation, nucleophilic addition, and protonation or Grignard addition to preinstalled 4-pyridyl ketones and subsequent hydrolysis (Fig. 1B)^15,16. These methods often require harsh conditions, exhibit poor functional group tolerance, suffer from low step and atom economy, and generate substantial metal wastes. Thus, the development of efficient and direct strategies for diverse synthesis of 4-hydroxyalkyl pyridines is desirable.

Fig. 1. Approaches for the functionalization of pyridines.

A Representative examples of 4-hydroxyalkylated pyridines. B Multistep synthesis of C4-hydroxyalkyl pyridines. C Strategies for C4-functionalization of pyridines: state-of-the-art. D This work: direct room temperature C4-functionalization of pyridines by paired electrolysis.

Despite the challenges posed by pyridine’s inherent thermodynamic stability, kinetic inertness, and multiple competitive reaction sites, several elegant strategies have been established for direct functionalization at the C2^17–22 and C3^23–32 positions. In contrast, functionalization at the C4 position typically relies on multistep sequences employing pre-activated substrates such as pyridinium salts, N-oxides, or pyridyl salts (Figs. 1C-1, 1C-2 and 1C-3)^33–47. Although Minisci-type radical addition enables direct functionalization of heteroarenes^17–22, it generally yields C2-substituted products alongside minor amounts of C4 regioisomers (Fig. 1C-4). In this context, cross-coupling between pyridyl radicals generated via single-electron reduction (SER) and polarity-matched radicals represents a promising strategy for achieving direct C4 functionalization, particularly for C2-substituted pyridines. Nonetheless, this objective remains a yet underdeveloped subject in synthetic chemistry.

In recent years, electrochemistry has emerged as an appealing tool in synthetic chemistry owing to the facile tunability of key reaction parameters (e.g., electrolyte, electrode, current, and potential)^48–72. Capitalizing on these advantages, we previously reported a direct method for constructing N-heterocyclic vicinal amino alcohols via electroreductive cross-coupling of the quinolyl C2-site with carbonyl compounds⁷³. Inspired by this work, we envisioned a practical electrochemical cross-coupling strategy for direct C4-functionalization of C2-substituted pyridines (A) with carbonyls (B). As shown in Fig. 1D, our proposed strategy involves the simultaneous cathodic reduction of pyridine A and carbonyl agent B, generating radicals A-1 and B-1. Subsequent radical-radical cross-coupling, followed by protonation and anodic oxidative aromatization of the resulting intermediate C-1, affords the desired product C. Notably, being significantly different from the reactive quinoline reactions employing a sacrificial zinc anode⁷³, the electrons released from the anodic oxidation of the coupling adduct (C-1) in this approach are subsequently consumed in the cathodic reduction process. Such an electrochemical strategy^74–76 substantially improves the synthetic efficiency while avoiding the need for sacrificial oxidants or reductants, thereby providing a greener and more environmentally benign route to functionalize pyridines.

However, realizing the above synthetic purpose presents several critical challenges: (1) Rate matching issue: The high reduction potentials (E_red < −2.2 V vs. SCE) of pyridines need to be lowered to match those of carbonyl compounds. This adjustment is essential for compatible reaction kinetics and to prevent overreduction of the more reactive carbonyl agents (B) into alcohols (B”). (2) Selectivity control: Diverse radical intermediates generated during pyridine reduction require precise reaction modulation for high C4-regioselectivity (C rather than C’). Additionally, radicals derived from carbonyl reduction should exhibit sufficient persistence to timely trap the pyridyl radicals in situ. This persistence is crucial for achieving high chemoselectivity by suppressing homocoupling processes (A’ and B’). (3) Requiring effective redox neutral electrolysis: Coupling adduct (C-1) must rapidly diffuse away from the cathode’s electrical double layer to undergo efficient anodic oxidative aromatization, thereby preventing its decomposition to achieve efficient transformation.

Here, by integrating acetic acid activation of both reactants with TEMPO-mediated oxidative aromatization, we report a room-temperature electrochemical strategy for direct C4- or C2-hydroxyalkylation of pyridines with readily available carbonyls. This method circumvents the limitations of conventional pyridyl functionalization that require pre-functionalized steps or stoichiometric activators. Furthermore, it achieves air/moisture tolerance, broad substrate scope, operational simplicity, metal-free conditions, high step and atom economy, and diverse synthesis.

Results

Reaction condition optimizations

To evaluate the feasibility of the envisioned synthetic protocol, we chose the coupling reaction between 2-phenylpyridine (A₁) and 4’-methoxyacetophenone (B₁) as a model system for parameter screening. The reaction was conducted in an undivided cell equipped with a graphite anode and a zinc cathode, using N,N-dimethylformamide (DMF) as the solvent and tetra-n-butylammonium tetrafluoroborate (n-Bu₄NBF₄, 0.1 M) as the electrolyte. The reaction was applied with constant current of 25 mA for 9 h in the presence of TEMPO (2,2,6,6-tetramethylpiperidin-1-oxyl) and acetic acid (AcOH). As shown in Table 1, this reaction afforded the C4-hydroxyalkyl product C₁ with an optimal isolated yield of 81% (entry 1). Excellent regioselectivity was confirmed by GC-MS and HPLC analyses (C₁:C₁’ > 20:1). Control experiments confirmed that both electricity (entry 2) and AcOH (entry 3) were indispensable for the reaction. The absence of TEMPO significantly decreased the product yield, along with low conversion of B₁ (entry 4), indicating that TEMPO acts as a reaction mediator. Current variations proved critical: no reaction occurred at 15 mA (entry 5) with substrates fully recovered, while increasing the current to 30 mA reduced the yield (entry 6). Moderate yields were obtained when adjusted the stoichiometric ratio of TEMPO/AcOH or replaced AcOH with HCOOH (entries 7–9). Neither ferrocene nor n-Bu₄NI enabled the formation of C1 (entries 10–11). This is attributed to the insufficient oxidizing capability of ferrocenium ions and elemental iodine, which are unable to effectively promote the oxidation of the intermediate—a key step in product formation. Additionally, these mediators may undergo competitive side reactions, further reducing the overall reaction efficiency. Solvent screening demonstrated that DMF outperformed NMP and acetone (entry 12). Cathode substitution (graphite or nickel) diminished the yield (entry 13), while platinum or zinc anodes performed the reaction poorly (entry 14). Notably, identical yields were obtained under argon and air atmospheres (entry 15), whereas the addition of H₂O (10 equiv.) led to a slight decrease in yield (entry 16). This indicates that the reaction exhibits good compatibility with water and oxygen. Moreover, both extended and shortened reaction times failed to improve the outcomes (Supplementary Table S1).

Table 1.

Optimization of reaction conditions^a

Entry	Variation	C1, yieldb (%)
1	none	81 (C1/C1’ > 20:1)
2	w/o electricity	N.R.
3	w/o AcOH	N.R.
4	w/o TEMPO	37
5	15 mA, 15 h	Trace
6	30 mA, 7.5 h	68
7	AcOH (3 equiv.)	55
8	TEMPO (0.25 equiv.)	64
9	HCOOH instead of AcOH	45
10	Fc instead of TEMPO	Trace
11	n-Bu4NI instead of n-Bu4NBF4	N.R.
12	NMP or Acetone instead of DMF	10, 21
13	GF (-) or Ni (-) instead of Zn (-)	53c, 71
14	Zn (+) or Pt (+) instead of GF (+)	15, 19
15	argon instead of air	81
16	H2O (10 equiv.)	74

^aConditions: Unless otherwise stated, all the reactions were performed in undivided cell with DMF (10 mL), A1 (0.2 mmol), B1 (0.6 mmol), TEMPO (0.5 equiv.), AcOH (5 equiv.), n-Bu₄NBF₄ (0.1 M, 1 mmol), current (25 mA), at room temperature for 9 h by using graphite plate (GF) as anode, and zinc plate (Zn) as cathode. ^bIsolated yield, C1/C1’ ratio determined by GC-MS. ^c4.5 h. AcOH Acetic acid, TEMPO 2,2,6,6-tetramethyl-1-piperinedinyloxy, NMP N-methyl-2-pyrrolidone; Fc Ferrocene; Ni nickel plate, Pt platinum plate, N.R. no reaction, PMP 4-methoxyphenyl.

Substrates scope investigation

With the optimal conditions established, we set out to evaluate the substrate scope of the developed synthetic protocol. As shown in Fig. 2, diverse ketones and aldehydes (B₁–B₂₈, B₄₀; see Supplementary Fig. S1 for their structures) were employed to couple with pyridines. Ketones lacking redox-sensitive groups (B₁–B₂₃) gave C4-hydroxyalkyl pyridines (C₁–C₂₃) in moderate to good isolated yields. Electron-donating groups and halogens were well tolerated (C₁–C₉), and 1-(4-chlorophenyl)ethan-1-one (B₆) required shorter reaction time to prevent hydrodehalogenation of product C₆. After stoichiometric adjustment, trifluoromethyl-containing aryl ketone B₁₁ was able to afford product C₁₁ in moderate yield. 3-Acetylthiophene provided product C₁₂ in acceptable yield. Cyclic and different alkyl-substituted ketones reacted smoothly (C₁₃–C₁₉). Trifluoromethyl (B₂₀–B₂₂) and vinyl ketones (B₂₃) also generated the desired products in moderate yields. Noteworthy, 3-penten-2-one yielded C_23a along with the regioisomer C_23b, the latter suggests a spin center shift (SCS) of the hydroxy α-radical. Gratifyingly, aryl aldehydes (B₂₄–B₂₈) also underwent efficient transformation, delivering products C₂₄–C₂₈ in reasonable yields. Notably, cyclohexanone (B₄₀) failed to react productively, and pyridine homocoupling product was obtained instead, which stems from the excessively high reduction potential of cyclohexanone.

Fig. 2. Variation of ketones and aldehydes.

Isolated yield of major regioisomeric products is given; the regioisomeric ratio was determined by GC-MS. Standard conditon: A (0.2 mmol), B (3 equiv.), 9 h (unless otherwise mentioned). ^aUsing graphite plate as the cathode and 35 mA. ^bWith heteroarene (3 equiv.) instead. ^cWith 5 equiv. of B.

Next, we investigated the reaction of diverse N-heteroarenes (A₁−A₂₈, Supplementary Fig. S1). Given the obvious variation in reduction potentials among these substrates, adjustments to the standard reaction conditions were necessary (Fig. 3). A range of functionalities (e.g., −OMe, −Cl, −i-Pr, −CO₂Et, −Me, −TMS, −F) proved compatible, affording the desired products (C₂₉−C₃₈) in moderate to good yields. Notably, electron-rich pyridines required either a graphitic cathode or extended reaction time to improve the yields. Thiophene-substituted pyridines yielded the desired products (C₃₉ and C₄₀) within shorter times. Electron-withdrawing groups (−CO₂Et, −CF₃) and unsaturated substituents (vinyl, alkynyl) were well tolerated (C₄₁−C₄₄). It is noteworthy that 2,2’-bipyridine underwent 4,4’-bisfunctionalization to afford compound C₄₅, demonstrating the potential of the present chemistry in developing valuable bidentate N-ligands. Pyridine (A₁₉) and 2-picoline (A₂₀) also underwent hydroxyalkylation efficiently, affording the corresponding C4-functionalized products (C₄₆ and C₄₇) in 40% and 24% yields, respectively. Interestingly, when equimolar amounts of aldehyde and amine were added, the in-situ-generated imines (B₃₅−B₃₆) could also be used in the coupling reaction with A₁, affording the C4-aminoalkyl pyridines (C₄₈, C₄₉) in reasonable yields. This example further demonstrates the broader potential of the present synthetic method. In addition to the C4-functionalization of C2-substituted pyridines, the method is also applicable to C3- and C4-substituted pyridines to afford predominantly C2-functionalized products. Interestingly, extending the reaction time for the coupling of 4-phenylpyridine (A₂₁) and B₁ from 4 h to 9 h shifted the product from the C2-hydroxyalkyl species (C₅₁’, 41%) to the 2,6-dihydroxyalkyl product (C₅₀’). 4-(Trifluoromethyl)pyridine (A₂₂) also reacted smoothly with 4-methoxyacetophenone, affording the desired product (C₅₂’) in 29% yield. Both 3-phenylpyridine (A₂₃) and 3-(trifluoromethyl)pyridine (A₂₄) underwent smooth transformation to form their respective C2-functionalized products C₅₃’ and C₅₄’, with this selectivity presumably governed by steric effects. In contrast, the disubstituted pyridines (A₂₅, A₂₆) afforded products C₅₅ and C₅₆ with exclusive C4-selectivity. Fused pyridines were also suitable substrates. For instance, thieno[3,2-b]pyridine (A₂₇) and furo[3,2-b]pyridine (A₂₈) afforded mixtures of major C4- and minor C2-regioisomers (C₅₇/C₅₇’, C₅₈/C₅₈’). Conversely, quinolines (A₂₉, A₃₀) and 4,7-phenanthroline (A₃₁) selectively provided C2-functionalized products (C₅₉’−C₆₁’), a class of potential hemilabile bidentate ligands. These examples show that the present synthetic protocol is also applicable for direct functionalization of fused pyridines, and the regioselectivity differences are attributed to the combined effects of radical stability and the steric influence.

Fig. 3. Substrate variation.

Standard condition: A (0.2 mmol), B (3 equiv.), 9 h. Isolated yield of major regioisomeric products is given; the regioisomeric ratio was determined by GC-MS. ^aGraphite plate as cathode. ^bWith 35 mA, graphite plate as cathode, A (3 equiv.), B (0.2 mmol), TEMPO (1 equiv.). ^cWith 15 mA, pyridine (3 equiv.) and H₂O (5 equiv.) were used instead of AcOH. ^dMixture of diastereomers.

Investigations of the reaction mechanism

To gain mechanistic insights into the established reaction, we initially conducted various control experiments. As shown in Fig. 4A, separate treatment of 2-phenylpyridine A₁ and 4-methoxyacetophenone B₁ under the standard conditions yielded the homocoupling products D₁ and D₂, respectively (Figs. 4A-I and 4A-II). Introducing the radical scavenger TEMPO (2.5 equiv.) into the model reaction completely suppressed the formation of product C₁; instead, a TEMPO-trapping pyridine adduct TEMPO-B₁ was observed (Fig. 4A-III). These results indicate a radical-involved pathway. Subsequently, a radical clock experiment using cyclopropyl ketone B₃₀ and A₁ afforded a ring-opening product (C₆₂), demonstrating that the ketone also acts as a radical precursor (Fig. 4B). Finally, a kinetic isotope effect (KIE) experiment with equimolar A₁ and A₁-d₆ afforded the desired product in 34% yield with a k_H/k_D value of 1.94 (Fig. 4C). This result suggests that C–H bond cleavage at the pyridyl C4 position is likely involved in the rate-determining step, though the presence of other rate-limiting processes cannot be excluded⁷⁷.

Fig. 4. Mechanism research.

A Control experiments. B Radical clock experiment. C Kinetic isotope effect experiment. D Cyclic voltammetry experiments. E Plausible mechanism.

Next, we performed cyclic voltammetry (CV) and open-circuit voltage (OCV) experiments to probe the electrochemical behavior of the model reaction (Fig. 4D and Supplementary Fig. S9). The CV of 2-phenylpyridine (A₁) displayed a reversible redox process with a reduction peak at −2.42 V (vs. SCE in DMF), while 4′-methoxyacetophenone (B₁) exhibited a reduction potential of −2.28 V. Upon addition of AcOH, both A₁ and B₁ showed two reduction peaks (A₁: −2.26 V and −2.79 V; B₁: −2.13 V and −2.55 V). The positive shift in reduction potentials and the enhanced catalytic current show that AcOH activates both reactants. Their similar onset potentials (−1.68 V) and peak potentials provide compelling evidence for concurrent reduction. This conclusion was further supported by OCV experiments (Supplementary Fig. S9), which revealed a stable cathodic plateau at −1.95 V (vs. SCE). Notably, the oxidation peak intensities of A₁ and B₁ decreased in the presence of AcOH, suggesting that AcOH stabilizes their radical intermediates generated by cathodic reduction.

Based on the above findings and the transformations employing redox neutral electrolysis^78,79, we propose a plausible reaction pathway for the reaction of A₁ and B₁ to form product C₁ (Fig. 4E). Initially, cathodic reduction of proton-activated 4’-methoxyacetophenone (B₁-H) generates a persistent benzyl radical (or hydroxy α-radical B₁’), while electroreduction of protonated 2-phenylpyridine (A₁-H) yields the pyridyl radical INT-1. Simultaneously, anodic single-electron oxidation of TEMPO (I) forms oxoammonium species II. Subsequent cross-coupling of radicals INT-1 and B₁’ affords the coupling adduct INT-2. Species II then mediates oxidative aromatization of INT-2, delivering product C₁ along with release of TEMPOH (III). Finally, anodic oxidation of III regenerates TEMPO (I), completing the catalytic cycle. Calculation of the Faradaic efficiency (FE) for this transformation revealed a relatively low value under standard conditions (3.86%; Supplementary Table S1). Given that hydrogen evolution is not significant for most substrates, we attribute this low FE primarily to an unproductive cycle, wherein radicals generated upon single-electron reduction of the substrate are re-oxidized by TEMPO prior to cross-coupling. As product formation necessitates radical cross-coupling following simultaneous single-electron reduction of substrates with high reduction potentials, the observed low FE remains acceptable for a reaction conducted in an undivided cell with non-sacrificial anode. Comparable FE values have been reported in similar electrochemical systems involving multiple single-electron transfers of the substrates with high reduction potential⁸⁰.

Further, DFT calculations were applied to analyze the proposed reaction pathway. All geometries were fully optimized, and the parameters are given in Supplementary Table S3 in the SI. As shown in Fig. 5, A₁-H is initially generated from the protonation of A₁ due to the presence of acetic acid in the reaction system. This protonation step will facilitate a subsequent single electron transfer process and generate radical INT-1. Spin density analysis of INT-1 reveals that the electron spin is predominantly located on the C4 and C6 positions, with little spin density at the C2 position. Mulliken spin population calculation further supports that the electron population on the C4 position (0.367) is significantly higher than on the C2 position (0.0896). These results suggest that the C4 position will be more reactive towards radical coupling. To study the chemoselectivity of the radical coupling step, the reaction energy barrier between radical B₁’ (Generated from cathodic reduction of ketone B₁) and INT-1 is calculated for C2 and C4 positions. The radical combination energy barrier at the C4 position is 9.7 kcal/mol, which is 2.3 kcal/mol lower than that at the C2 position. Thus, the radical coupling process will preferentially occur at the C4 position and afford INT-2 rather than INT-3. Subsequently, given the nucleophilic positions in INT-2 and potential hydrogen donor characteristics, the subsequent reaction between TEMPO⁺ (Generated from the anodic oxidation of TEMPO) may have multiple pathways. When the C3 carbon attacks the nitrogen in the N = O double bond of TEMPO⁺, the reaction energy barrier is 21.1 kcal/mol, which is relatively high. In addition, this step is also thermodynamically unfavorable (ΔG = 17.7 kcal/mol). Thus, the following cope elimination from INT-5 to form the aromatized product INT-4 is unlikely to occur. Attempts to locate a stable intermediate via nucleophilic attack from the nitrogen atom of INT-2 were unsuccessful, possibly due to the steric hindrance between the tera-methyl groups in TEMPO⁺ and the phenyl group in INT-2. Alternatively, a hydride transfer mechanism which is favorable both kinetically and thermodynamically was identified^81–83, as 1,4-dihyropyridines are well-documented hydrogen donors^84–86. The hydrogen transfer between INT-2 and TEMPO⁺ has a low energy barrier (15.9 kcal/mol) and favored thermodynamics (ΔG = −37.1 kcal/mol), which will lead to the formation of protonated product INT-4 and TEMPOH. Final deprotonation mediated by acetate affords the desired product C₁. This deprotonation step may also be assisted by substrate A₁.

Fig. 5. DFT calculations for the reaction pathways.

All energies are calculated at M06-2X-D3/def2-TZVPP/SMD(DMF)/M06-2X-D3/def2-SVP/PCM(DMF) level of theory. Spin density and spin population of INT-1 are shown.

Derivatizations of the products

Finally, we explored the synthetic utility of the developed methodology. Even with reduced loadings of TEMPO/AcOH (Fig. 6A), the reaction of A₁ (3 mmol) with B₁ (9 mmol) afforded the product C₁ in good yield (62%, 0.57 g). This yield represents a significant improvement over the multistep procedure reported previously¹⁴ and demonstrates the scalability potential of the protocol. As shown in Fig. 6B, treating C₁ to HCl in toluene (12 h) followed by neutralization afforded the 1,1-diarylethylene D₃ in 85% yield. Hydroboration-oxidation of D₃ gave the secondary alcohol D₄ in 88% yield, while epoxidation with m-CPBA followed by ring-opening yielded the vicinal diol D₅ in 50% yield. Enabled by the pyridyl directing group, palladium-catalyzed ortho-C–H acylation of the phenyl moiety in C₁ provided D₆ in 39% yield. Direct esterification of C₁ with ibuprofen furnished D₇ in 51% yield. We next applied the synthetic method to the functionalization of bioactive substrates (Fig. 6C). Reaction of Isophorone (B₃₁) with A₁ gave the desired product C_63a in merely 16% yield, while the 1,3-rearrangement product C_63b predominated with a 56% yield. Tonalide (B₃₂), Epoxyprogesterone (B₃₃) and L-menthol-derived ketone (B₃₄) reacted efficiently with A₁ to exclusively afford C4-functionalized pyridines (C₆₄–C₆₆). Furthermore, the reactions of Myrtenol-derived pyridine A₃₂ and Abametapir A₃₃ with B₁ proceeded smoothly, affording C₆₇ and C₆₈ in moderate to good yields.

Fig. 6. synthetic applications.

A Amplified synthesis. B Product derivatization. C Post-functionalization of biomedical molecules. ^aProceeds under standard conditions. ^bProceeds under standard conditions but using graphite plate as cathode, 5 h. ^cC₁ (0.2 mmol), HCl (3 equiv.), toluene (2 mL), r.t., 20 h. then add Na₂CO₃ (5 equiv.). ^dD₃ (0.2 mmol), BH₃·THF (5 equiv.), THF (2 mL), r.t. 20 h. then add KOH (3 equiv.), H₂O₂ (5 equiv.). ^eD₃ (0.2 mmol), m-CPBA (2.5 equiv.), THF (2 mL), r.t., 20 h. then add Na₂SO₃ (3 equiv.). ^fC₁ (0.2 mmol), PdCl₂ (5 mol%), ^tBuOOH (3.0 equiv.), benzyl alcohol (3 equiv.), PhCl (0.5 mL), air, 140 ^oC, 24 h. ^gC₁ (0.2 mmol), DL-ibuprofen (1.2 equiv.), DCC (2 equiv.), DMAP (0.2 equiv.), DCM (3 mL).

Discussion

In summary, we have developed an electrochemical method that enables the direct hydroxyalkylation of diverse pyridine derivatives with readily available carbonyl compounds at room temperature. Governed primarily by electronic and steric effects, unsubstituted and C2-substituted pyridines mainly yield C4-functionalized products, whereas C3- and C4-substituted pyridines, along with fused systems, generally afford C2-functionalized products. By employing a zinc cathode, graphite anode, and TEMPO mediator, this strategy offers broad substrate scope, operational simplicity, metal-free conditions, and high step- and atom-economy. The present work overcomes the limitations of conventional pyridyl functionalization that require pre-functionalized substrates or stoichiometric activators, establishing a practical platform for direct access to hydroxyalkyl pyridines. In addition, preliminary studies indicate that fused pyridines and imines are also amenable to this transformation, demonstrating broad potentials of the present chemistry. Mechanistic studies reveal that acetic acid activates both reactants, while cathodic reduction generates key pyridyl and hydroxy α-radicals. Product formation proceeds via electroreduction-induced radical-radical cross-coupling, followed by anodic oxidative aromatization. In this work, the strategy of performing electroreduction-induced dearomatization and subsequent intermediate transformation in a single operation is expected to advance the development of efficient methodologies for (hetero)aryl functionalization. Furthermore, it paves the way for constructing valuable functional ligands and for functionalizing bioactive molecules to promising derivatives.

Method

General procedure for the hydroxyalkylation of pyridines in DMF

An undivided electrolytic cell equipped with a magnetic stir bar was subsequently charged with TEMPO (15.6 mg, 0.1 mmol, 0.5 equiv.), A (0.2 mmol, 1 equiv.), carbonyl substrate B (0.6 mmol, 3 equiv.), n-Bu₄NBF₄ (330 mg, 0.1 M), DMF (10 mL) and AcOH (58 uL, 1 mmol, 5 equiv.) under air condition. After fully stirring and dissolution, a graphite plate anode and a zinc plate cathode were installed. Then the cell was electrolyzed at room temperature under a constant current of 25 mA for 9 h (If not otherwise mentioned). Upon the completion of electrolysis, the resulting mixture was poured into saturated NH₄Cl solution (50 mL) and the electrodes and cell were rinsed with EtOAc, then extracted the mixture with EtOAc (35 mL × 3). The combined organic layers were dried over anhydrous Na₂SO₄, filtered, and concentrated under reduced pressure. The resultant residue was purified by preparative TLC on silica to afford hydroxyalkylated pyridines C.

(Note: Both the dimensions of graphite and zinc electrodes are approximately 5.25 cm ×  0.8 cm × 0.2 cm. The effective immersion areas of graphite anode and zinc cathode are approximately 2.56 cm²).

Author contributions

M.Z. conceived the idea, analyzed the data, directed the project, and wrote the manuscript. C.-Q.Z. and M.-R.W. carried out all the catalytic experiments. C.-Q.Z. drew the structures of all the obtained compounds, analyzed the single crystal structures, synthesized the raw materials and carried out NMR tests. M.-R. W. performed the DFT calculations. P.H.D. discussed the mechanistic aspects and revised the manuscript. All the authors have read the manuscript and agree with its content. All the authors contributed to the discussion during the whole project.

Peer review

Peer review information

Nature Communications thanks Youai Qiu and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

All NMR spectra, mass spectrometry data, and experimental procedures generated in this study are provided in the Supplementary Information. The computational coordinate data generated in this study are provided in the Source Data, The X-ray crystallographic coordinates for structures reported in this study have been deposited at the Cambridge Crystallographic Data Centre (CCDC), under deposition numbers of CCDC 2416007 (C₂₁). These data can be obtained free of charge from The Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif. All data are available from the corresponding author upon request. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-026-71858-2.