Regio- and Stereoselective Electrochemical Alkylation of Morita-Baylis-Hillman Adducts.

Electrosynthesis is effectively employed in a general regio- and stereoselective alkylation of Morita-Baylis-Hillman compounds. The exposition of N-acyloxyphthalimides (redox-active esters) to galvanostatic electroreductive conditions, following the sacrificial-anode strategy, is proved an efficient and practical method to access densely functionalized cinnamate and oxindole derivatives. High yields (up to 80%) and wide functional group tolerance characterized the methodology. A tentative mechanistic sketch is proposed based on dedicated control experiments.

Morita–Baylis–Hillman (MBH) adducts (1) are arguably referred to as “privileged scaffolds” within the synthetic community, due to their ease of preparation,[1] ready diversification and faceted reactivity. In recent years, SN2 or SN2′ substitution reactions on MBH compounds has faced high interest with a variety of ionic-based nucleophiles,[2] as well as their transformation in reactive dipolar species.[3] These led to the development of a large number of uncatalyzed or Lewis-based catalyzed protocols, targeting, among others, functionalized methacrylates, cinnamates, azine heterocycles, and indole derivatives.[4] On the contrary, the rapidly expanding radical chemistry scenario has scarcely permeated the functionalization of MBH adducts. Here, some elegant examples have recently been disclosed under metal- or photocatalytic regimes with a focus on specific classes of stabilized alkyl radicals (Scheme a).[5,6] Protocols displaying a wide scope of simple and functionalized radicals are much more narrow in number, with notable shortcomings related to the employment of stoichiometric organic reductants[7] or poor stereochemical outcomes.[8] Proceeding under mild conditions, displaying exquisite functional group tolerance and complementary selectivity, electrochemical synthesis “eChem” has rapidly paralleled other enabling techniques, opening unforeseen opportunities in organic synthesis.

Scheme 1

Previous Methodologies for the Radical Alkylation of MBH Acetates and Present eChem Protocol

NPhth: phthalimide.

Our current research interest toward the implementation of site-selective radical-based transformations[9] prompted us to envision eChem as a valuable direct toolbox for the functionalization of MBH adducts 1 with “green electrons”.[10]

To realize such a strategy, we deemed a general, cheap, easy to prepare, and broadly applicable class of radical precursors to be necessarily employed. N-Acyloxyphthalimides (redox-active esters, RAE, 2), prepared in one step from inexpensive N-hydroxyphthalimide and ubiquitous carboxylic acids, represent a class of benchmark radical precursors.[11] RAEs have been extensively employed in metal- or photoassisted[12] and (to a lower extent) electrochemical generation of radicals under reductive conditions.[13] Among others, C–H functionalizations and cross-coupling processes were primary targeted,[14] whereas the use of RAEs on eChem derivatizations of olefins have faced less attention.[15] In particular, the condensation of RAEs 2 and MBH acetates 1 is unprecedented (Scheme b), posing some important questions toward the overall chemo- and regioselectivity of the process. In this report, we disclose a highly stereoselective (the E-isomers were exclusively isolated), electroreductive strategy for the regioselective SN2′ radical alkylation of MBH adducts 1 with RAEs 2. The protocol targets the formation of α-substituted α,β-unsaturated esters or ketones 3, under mild and organic reductant-free conditions. It is worth mentioning that alternative synthetic methodologies for the installation of alkyl (i.e., methyl) groups at the β-postion of the MBH acceptors requires the utilization of harmful organometallic reagents such as trialkyl-Al and trialkyl-In compounds.[16]

Our investigation started by subjecting MBH acetate 1a and RAE 2a (1 equiv) to a constant current electrolysis of 10 mA.[17] A graphite (C) cathode and a Zn sacrificial anode were used as electrodes with tetraethylammonium tetrafluoroborate (TEABF4, 1 equiv) as the supporting electrolyte in DMF (Table , entry 1). Encouragingly, the desired product 3aa was isolated in 25% yield as a single E isomer, along with 5% of the reduction–deacetoxylation product 1a′. Exclusive SN2′ radical trapping was recorded, and no over-reduced product 3aa′ was detected. The high E selectivity of cinnamates 3 means the present eChem approach is a desirable and complementary alternative to the poorly stereoselective photocatalytic strategy, aiming at similar tasks (vide infra).[8] Moreover, the use of a nontoxic and inexpensive sacrificial anode (Zn) circumvents the use of organic reductants whose cost and separation from the reaction mixture might be burdensome.

Table 1

Optimization of the Reaction Conditions.a

entry	electrolyte (equiv)	anode (+)||cathode (−)	I (mA)	yieldb (%)
1c	TEABF4 (1)	Zn(+)||C(−)	10	25
2c	TEABF4 (2)	Zn(+)||C(−)	10	38
3	TEABF4 (2)	Zn(+)||C(−)	10	59
4	TEABF4 (2)	Zn(+)||C(−)	4	67
5	TEABF4 (2)	Zn(+)||C(−)	2	79
6	TBAPF6 (2)	Zn(+)||C(−)	4	65
7	LiBF4 (2)	Zn(+)||C(−)	4	66
8	TEABF4 (2)	Zn(+)||RVC(−)	4	65
9	TEABF4 (2)	Zn(+)||Nid(−)	4	54
10	TEABF4 (2)	Mg(+)||C(−)	4	54
11	TEABF4 (2)	Ni(+)||C(−)	4	50
12e	TEABF4 (2)	C(+)||C(−)	4	34

Reaction conditions, unless otherwise noted: 1a (0.15 mmol), 2a (0.30 mmol), electrolyte (0.15 or 0.30 mmol), dry DMF (3 mL), CCE (10, 4, or 2 mA; 2 F/mol), rt. E/Z ratios were determined via 1H NMR spectroscopy on the reaction crude mixtures and were always found to be >20:1.

Isolated yields after flash chromatography.

2a (0.15 mmol).

Ni foam.

4 (0.30 mmol) added.

In order to optimize the reaction conditions, the equivalents of both the electrolyte (2 equiv, entry 2, 38% yield) and of 2a (2 equiv, entry 3, 59% yield) were beneficially increased. Next, we reasoned that a slower generation of the radical species, as the result of a lower current, could be beneficial in avoiding radical–radical homocoupling and other undesired side reactions. Interestingly, by lowering the current value from 10 to 4 mA (entry 4, 67% yield) and further to 2 mA (entry 5, 79% yield) a significant increase in reaction efficiency was recorded (with 2 mA current, the formation of 1a′ was completely suppressed). On the other hand, salts such as tetrabutylammonium hexafluorophosphate (TBAPF6, entry 6) and LiBF4 (entry 7) behaved similarly to TEABF4. Cathodes different from graphite, such as RVC (reticulated vitreous carbon, entry 8) and Ni foam (entry 9), and sacrificial anodes such as Mg and Ni (entries 10 and 11), delivered 3aa in comparable or lower yields with respect to the optimal set (entry 5). To demonstrate the superiority of the strategy based on the sacrificial anode, an attempt using Hantzsch ester 4 as the terminal reductant and two graphite (C) electrodes was performed (entry 12).[15a] Here, product 3a was isolated in low yield (34%) along with 22% of reduced byproduct 1a’.

Having established the optimal reaction conditions (Table , entry 5), the generality of the electrochemical alkylation was tested (Scheme ) on different MBH acetates (or carbonates) 1 (or 5). Cinnamates 3, having both electron-withdrawing (3ba–3da, 3ga) and electron-donating (3ea, 3fa) substituents at the para or ortho position of the benzene ring, were productively formed (61–77% yield). Naphthalene (1h) and heteroarenes (thiophene 1i and quinoline 1j) on the MBH acceptor were also well tolerated, although in the case of 1j a complete reduction of the C–C double bond occurred, resulting in saturated compound 3ja′ as the sole product (64%). Pleasingly, MBH acetate 1k, derived from an aliphatic aldehyde (i.e., hydrocinnamaldehyde) was proved to be productive (3ka, 62% yield) and product 3la, having a conjugated diene moiety, could also be formed in synthetically useful 57% yield. It is worth mentioning that radical-sensitive moieties such as benzylic (1m) and propargylic esters (1n) were adequately tolerated in the present eChem alkylation process (yield: 76% and 74%, respectively). Finally, the protocol was not limited to ester-like MBH adducts; as a matter of fact, when methyl ketone 1o was subjected to the optimal conditions the desired α-alkylated enone 3oa was isolated in synthetically useful amounts (76% yield).

Scheme 2

Scope of the Present eChem Methodology

Reaction conditions: 1 (0.15 mmol), 2 (0.3 mmol), TEABF4 (0.30 mmol), dry DMF (3 mL), CCE (2 mA; 4 F/mol1), Zn(+) C(−), rt. Isolated yields after flash chromatography; E/Z ratios were determined via 1H NMR spectroscopy on the reaction crude mixtures.

Reaction performed on 1.0 mmol of 1a at 3.0 mA, see the Supporting Information for details.

Additionally, the electrochemical alkylation protocol was extended to MBH carbonates derived from N-methylisatins (5a,b) in the presence of RAEs 2a and 2c. Delightfully, the corresponding oxindoles 6 were isolated with high diastereoselectivity (>20:1) and useful yields (51–66%) as a result of an alkylation–reduction sequence. To the best of our knowledge, this protocol represents the first example of SN2′-type radical alkylations of oxindole-based MBH acceptors. The exclusive isolation of the fully reduced compounds 6 and 3ja′ might be rationalized in term of lower reduction potential of α,β-unsaturated esters featuring conjugated highly electron-deficient moieties. Over-reductive SET processes and hydrogen abstraction of the resulting radical anions could lead to “zinc-enolate” intermediates that will smoothly undergo protonation during the aqueous reaction quenching.

Next, we examined the adaptability of the disclosed strategy to different radical precursors 2b–j. Methyl (2b), primary (2c), secondary (2d and 2f), and tertiary (2e) alkyl radicals were all installed regio- and stereoselectively at MBH acetate 1a, highlighting the generality of the methodology (3ab–af, 49–80% yield). As a rule of thumb, the higher the substitution (hence, the stability) of the radical the higher the yield observed. Tolerance toward an unprotected indole group was also ascertained (3ag, 61% yield). Pleasingly, amino acid derived RAEs 2h (Boc-Ala) and 2i (Boc-Pro) as well as dehydrocholic acid derived 2j generated competent alkyl radicals for the alkylation of 1a (3ah-aj, 66–75% yield). This demonstrates the possibility to employ naturally occurring acids as precursors of functionalized radicals as well as developing bioconjugation protocols via a late-stage eChem approach. The millimole scale reaction on the model substrates 1a and 2a was also effectively carried out, delivering the desired 3aa in 83% isolated yield.

To gain mechanistic insights into the formation of byproduct 1a′, voltametric experiments on 1a were then carried out (see the Supporting Information). Unfortunately, no clear evidence for a reductive event was recorded, suggesting that 1a′ could result from successive reactive steps. However, a control experiment run on 1a in the absence of 2a (optimal reaction conditions) furnished 1a′ in 18% yield, along with poor recovery of unreacted 1a (34%, decomposition to unidentified byproducts, Scheme a). This result rules out an exclusive RAE-mediated formation of 1′ from 1 and proves MBH acceptors 1 are not inert toward reductive conditions, rendering a judicious choice of the reaction parameters pivotal for their productive employment in electroreductive processes.

Scheme 3

Control Experiments

Finally, our attention was caught by the stereochemical discrepancy between our protocol (E/Z > 20:1) and previously reported photochemical radical alkylations of MBH derivatives (E/Z ca. 1:1).[8,18] To prove a possible photomediated isomerization of the final cinnamyl C–C double bond, we subjected (E)-3aa (E/Z > 20:1) to visible-light irradiation in the presence of a common triplet-emitter Ir-based photosensitizer (Ir[dF(CF3)ppy]2(dtbppy)PF6, (1 mol %, Scheme b). Interestingly, 3aa was recovered in 95% yield as a 1:1.7 E/Z mixture, showing the intrinsic unsuitability of some photocatalytic methodologies aiming at the stereoselective formation of α-substituted cinnamates.[19]

Mechanistically, the eChem cycle depicted in Scheme is tentatively proposed. This starts with the cathodic fragmentation of 2a into CO2, phthalimide anion, and the desired cyclohexyl radical. This event is in accordance with cyclovoltammetry experiments run on 2a, showing a clear irreversible cathodic event at −1.57 V vs ferrocene (−1.15 V vs Ag/AgCl).[15a] The possible involvement of the MBH adduct into cathodic reduction was judged unlikely due to the absence of reductive signals on the collected voltametric spectra in the same reductive windows of 2a (see the Supporting Information). Addition of the cyclohexyl radical onto the electrophilic double bond of 1a gives the key intermediate A, which can evolve following two alternative pathways. Radical fragmentation (path a) would directly render the observed product 3aa and the acetoxy radical, undergoing reduction to the acetate anion. On the other hand, intermediate A can first be reduced and then deliver 3aa through E1cB elimination. As depicted, cathodic reductions are coupled with anodic formation of Zn2+ ions, following the sacrificial anode strategy.

Scheme 4

Proposed Mechanistic Profile

In conclusion, we have developed a novel electrochemical alkylation of MBH adducts with redox-active esters as radical precursors. Under galvanostatic conditions and employing a sacrificial anode, a wide range of α-substituted α,β-unsaturated esters or ketones were formed chemo- and stereoselectively (E/Z ratio >20:1). Extension of the protocol to isatin-derived MBH carbonates resulted in the formation of substituted oxindoles following an alkylation–reduction sequence. Control experiments and mechanistic proposals completed the present investigation.