Site-specific and Degree-controlled Alkyl Deuteration via Cu-catalyzed Redox-neutral Deacylation

Abstract

Deuterated organic compounds become increasingly important in many areas; however, it remains challenging to install deuterium site-selectively to unactivated aliphatic positions with control of degree of deuteration. Here we report a Cu-catalyzed degree-controlled deacylative deuteration of diverse alkyl groups with methylketone (acetyl) moiety as a traceless activating group. The use of N-methylpicolino-hydrazonamide (MPHA) promotes efficient aromatization-driven C–C cleavage. Mono-, di- and tri-deuteration at specific sites can be selectively achieved. The reaction is redox neutral with broad functional group tolerance. The utility of this method has been demonstrated in forming a complete set of deuterated ethyl groups, merging with the Diels–Alder reaction, a net devinylative deuteration, and the synthesis of the d₂-analogue of Austedo.

Graphical Abstract

Deuterated organic molecules have found important and broad applications in many fields. For example, incorporation of deuterium is frequently used to generate more metabolically stable and less toxic therapeutics (Scheme 1A),¹ e.g., deutetrabenazine (Austedo).² Owing to the kinetic isotope effect, deuterium-labeled compounds serve as routine tools for mechanistic studies in both organic and medicinal chemistry research.³ In addition, their significance to the development of NMR spectroscopy measurements cannot be ignored. Moreover, they have been employed in quantitative mass spectrometry to determine protein conformations.⁴ Further, deuterated organic materials have received increasing attentions due to improved properties over their undeuterated counterparts.⁵

Scheme 1.

Alkyl Deuteration and Hydrodeacylation of Ketones

As such, D-labelling strategies that introduce deuterium efficiently and site-specifically would greatly benefit the aforementioned applications.⁶ While hydrogen isotope exchange has been a powerful method to directly replace ¹H with deuterium,⁷ it remains a substantial challenge to install deuterium site-selectively at unactivated aliphatic positions (not acidic or adjacent to radical stabilizing groups) without using directing groups (Scheme 1B).⁸ Other D-labelling approaches, such as reductive deuteration, are also widely useful, though the degree of deuteration (i.e., the number of deuteriums introduced at a specific position) is generally dictated by the functional group precursor.⁹ For instance, it is still difficult to selectively install one, two and three deuteriums at an unactivated terminal position without using different substrates.^6,7c In addition, reductive conditions often require more expensive deuterium sources such as D₂ or deuteride equivalents.⁹ To allow efficient and site-specific deuteration at unactivated C(sp³) positions¹⁰ with control of degree-of-deuteration, here we describe a complementary and deconstructive strategy that utilizes readily available ketones (i.e., acetyl group) as a traceless activating group (Scheme 1C).

Recently, we developed a new deacylation reagent, N-methylpicolino-hydrazonamide (MPHA), that promotes efficient aromatization-driven Cu-mediated C–C cleavage of ketones to yield olefins.¹¹ The reaction starts with first forming a pre-aromatic intermediate (PAI) via the condensation between ketone and MPHA, which is subsequently oxidized by Cu(OAc)₂ to form a delocalized radical (R1). Driven by forming an aromatic triazole byproduct, selective C–C cleavage then occurs at the more substituted side to produce a carbon-center radical species (R2),¹² which is further oxidized by another equivalent of Cu(OAc)₂ to give the olefin product. We hypothesized that, instead of the previous oxidative pathway mediated by stoichiometric Cu(II), a redox-neutral hydrodeacylation could be realized with catalytic copper by utilizing the N–H hydrogen of PAIs as the H-atom source to quench the R2 radical (Scheme 1C). Considering that the relatively acidic ketone α hydrogens and the N–H hydrogen of PAIs can be easily deuterated at different stages via H/D exchange with D₂O,¹³ various degrees of deuteration could be achieved, in principle, via different combinations of operations from the same ketone precursor. For example, if only the N–H hydrogen of PAIs is deuterated, a mono-deuteration process is anticipated. On the other hand, if only the ketone α positions are deuterated, di-deuteration products should be afforded after ¹H-quenched deacylation. Clearly, tri-deuteration is expected if both ketone α positions and the N–H hydrogen of PAIs are deuterated.

To examine the hypothesis, the pristine reaction, namely the Cu-catalyzed hydrodeacylation, was explored first, and methyl ketone 1a was selected as a model substrate (Table 1). With 10 mol% 1-AdCO₂H as the acid catalyst and Al₂O₃ as the dehydrating agent,^11,14 the condensation between MPHA and 1a proceeded smoothly with almost full conversion. After careful examination of various copper salts, solvents, temperature, and other parameters, the desired deacylation product (3a) was obtained in 87% yield using CuOAc as the catalyst in dried DMA (0.5 M) at 140 ^oC for 3 hours (entry 1). Other hydrazonamides were also tested, which were less efficient than MPHA (see Supporting Information Table S1). The use of excess MPHA greatly reduced the yield (entry 2), likely due to that the residual free MPHA can strongly coordinate with Cu thus inhibiting its activity. The yield was slightly compromised in the absence of 1-AdCO₂H or using TsOH instead (entries 3 and 4). Comparable results were obtained without acidic alumina or using neutral alumina (entries 5 and 6). The C–C cleavage step can also occur at much lower temperatures. Nearly the same yields were observed at 100 and 90 ^oC with a longer reaction time, although the reaction became somewhat less efficient at 80 ^oC (entries 7-9). The control experiment indicated that CuOAc was essential to this transformation (entry 10).¹⁵ The yield was slightly dropped when using 20 mol% CuOAc (entry 11). Replacing CuOAc with the corresponding Cu(II) salt also worked albeit in lower yield (entry 12).

Table 1.

Selected Reaction Optimization

entry	variation from ‘the above condition’	yield of 3aa	entry	variation from ‘the above condition’	yield of 3aa
1	none	87%	7	100 °C and 20 h @ step 2	86%
2	1.2 equiv MPHA	47%	8	90 °C and 20 h @ step 2	88(87b)%
3	w/o 1-AdCO2H	73%	9	80 °C and 20 h @ step 2	74%
4	TsOH instead of 1-AdCO2H	74%	10	w/o CuOAc	<5%
5	w/o Al2O3 acidic	83%	11	20 mol% CuOAc instead	82%
6	Al2O3 neutral instead of Al2O3 acidic	83%	12	Cu(OAc)2 instead of CuOAc	65%

^aThe yield is based on ¹H NMR analysis of the crude reaction mixture.

^bIsolated yield.

The scope of the hydrodeacylation was next examined under the milder conditions described in entry 8, Table 1. A wide range of methyl ketones smoothly underwent the site-selective C–C cleavage (Table 2A).¹⁶ Due to the relatively mild and redox-neutral conditions, diverse functional groups, including acetal (3a), trimethylsilyl (3b), nitro (3c), Bpin (3d), phenol (3g), bromide (3h), esters (3k-3l), ethers (3q, 3n) and N-Boc groups (3m, 3o), were all tolerated. Notably, in the presence of an aryl methyl ketone (3f), the condensation with MPHA and the C–C fragmentation occurred selectively at the dialkyl ketone moiety. In addition, substrates containing heterocycles, such as quinoline (3j), N-methylindole (3v), protected piperidines (3t, 3x) and dibenzothiophene (3u), also worked well. Besides forming primary carbons, α-substituted ketones are also competent substrates to generate methylene groups (3w-3aa) (Table 2B).¹⁷ A terminal mono-fluorinated alkyl product (3w) was obtained, and the moderate yield was due to the labile alkyl–F bond under the reaction conditions. A strained cyclobutane (3z) and a phenanthrene (3aa) survived in this reaction. Gratifyingly, substrates derived from various natural products uneventfully delivered the desired deacylation products, despite the existence of labile or reactive functional groups, e.g., epoxide (3ab), ketal (3ac), xanthine (3ad), polyene (3ae) and cyclopentanone (3af) (Table 2C). The addition of 20 mol% 2,4,6-triisopropylthiolphenol as a H-atom transfer catalyst was found beneficial for more challenging substrates (3ab and 3af).¹⁸ Thus, this method shows promise for late-stage modifications of complex bioactive compounds.

Table 2.

Substrate Scope of the Pristine Reaction^a

^aUnless otherwise noted, all reactions were run on a 0.2 mmol scale.

^bThe reaction was run at 140 °C, 3 h in Step 2.

^c20 mol% 2,4,6-triisopropylthiolphenol was added.

^d0.4 mmol scale.

^e0.5 mmol scale.

^f0.1 mmol scale.

Next, the proposed deacylative degree-controlled deuteration was systematically explored with the objective of using inexpensive D₂O as the deuterium source (Table 3). After carefully optimization, three sets of reaction protocols were ultimately developed to realize selective mono-, di- and tri-deuteration, respectively (see Supporting Information Tables S2–7 for details). For the mono-deuteration, the key was to ensure that the N–H hydrogen in PAIs remains fully deuterated in the C–C-cleavage step. Upon condensation between the ketone and MPHA, treatment with a mild base (i.e., MeNH₂) and D₂O gave almost full deuteriation at the N–H of the PAI intermediate. To avoid undesired D/H exchange on the N–D in the C–C cleavage step (for a control experiment, see the Supporting Information), additional D₂O and deuterated 2,4,6-triisopropylthiolphenol were used. NaDCO₃ was employed as a base buffer to avoid undesired acid-catalyzed α-deuteration,¹⁹ which could lead to over-deuterated products. As a result, this protocol (condition a) generally gave high overall mono-deuteration for a range of substrates (Table 3) along with high isotope purity observed from the ¹³C NMR and HRMS (see Supporting Information). To realize di-deuteration, α-deuteration of the ketone substrates was first carried out via H/D exchange with D₂O using pyrrolidine as the catalyst,¹³ and then the key was to avoid deuterium loss during the PAI formation. The optimal conditions (condition b) were found to use fully deuterated MPHA (prepared via H/D exchange of MPHA with D₂O) to condense with the α-deuterated ketone; the resulting PAI can easily lose the D on the nitrogen via H/D exchange with H₂O in the C–C cleavage step. With condition b in hand, generally 1.84–1.99D can be introduced by this method. As an exception, the somewhat deuteriation loss with 3l was due to the intramolecular 1,5-hydron-atom-transfer (HAT). It is more straightforward to develop tri-deuteration conditions. Starting from the α-deuterated ketone, the desired 3D (or 2D when forming a methylene group) products were obtained in high efficiency under condition c, in which all the proton sources were replaced with the corresponding deuterated ones. In general, degree-controlled deuterations can be achieved at benzylic, α-to-oxygen and unactivated terminal or internal positions with good overall yields, as shown in the representative examples in Table 3. Note that di- and tri-deuterated products can be obtained in one pot directly from the undeuterated ketone substrates.

Table 3.

Examples of the Degree-controlled Deuteration^a

^aUnless otherwise noted, all reactions were conducted on a 0.1 mmol scale.

^bThe product was further treated with DBU (1 equiv) in MeOH (0.2 M) at 60 °C for 7 h to remove deuterium at the methine position.

^cPyrrolidine was replaced by Et₃N (10 mol%).

^dThe product was further treated with Ag₂CO₃ (5 mol%) and JohnPhos (5 mol%) in MeOH (0.2 M) at 80 °C for 24 h to remove deuterium on the imidazole moiety.

^eThe product was further treated with NaOH in DCE/1,4-dioxane/H₂O at 70 °C for 3 hours to remove deuterium at the cyclopentanone α-position.

^f0.2 mol scale.

Applications of the deacylative deuteration strategy was further investigated. First, site-selective multi-deuterations can be realized using ketone as a traceless handle (Scheme 2A). The carbonyl moiety can assist deuteration not only at the α-position, but also at the β-position. For example, one or two-D can be selectively installed at the ketone β-position via Pd-catalyzed deuteration of the enone or ynone (followed by H/D exchange at the α-position), respectively. Capitalized on this deconstructive strategy, now a complete set of deuterated ethyl moieties can be generated through combinations of different β-deuterated ketones with different deacylative deuteration conditions. In other words, all possible scenarios of deuterated ethyl groups can be prepared in a selective manner, which, to our knowledge, has not been realized by other methods.

Scheme 2.

Synthetic Applications

In addition, the intermediates generated from the Diels–Alder reactions of 2-butenone and 1,3-dienes can lead to interesting products that are hard to synthesize otherwise (Scheme 2B). For example, an unusual [3.2.1] nitrogen-containing bicycle (7) was synthesized efficiently from the Diels–Alder²⁰/deacylation sequence. Interestingly, a rare formal 1,2-radical-rearrangement was observed in this scaffold, which is expected to be mediated by the alkene moiety.²¹ The deuterated version of the reaction under condition c provided the product with two deuterium atoms installed on vicinal carbons. In addition, the Diels–Alder intermediate from myrcene (4) underwent the deacylation smoothly, which introduced two deuteriums into the cyclohexene methylene that is the conventionally hard-to-do-so position. Moreover, through merging with the Wacker oxidation, a net devinylative deuteration was illustrated using an oleic acid derivative (Scheme 2C), which is complementary to Kwon’s hydrodealkenylation method.²²

Finally, to show the utility of this method, synthesis of the d₂-analogue of Austedo was accomplished in four steps (Scheme 2D). While Austedo (d₆-tetrabenazine) represents the first FDA-approved deuterated drug,² preparation of the corresponding unsymmetrical and less deuterated analogues, e.g.,17, was not reported. Our approach started from commercially available homoveratrylamine (11) and employed Bischler–Napieralski cyclization and hydrolysis to afford compound 12 in good yield.²³ The following Dieckmann cyclization and phenol alkylation generated diketone 16. The less bulky methyl ketone reacted chemoselectively and served as a surrogate for a deuterated methyl group. Using condition b (as an example), 1.94 D incorporation was realized at the upper methoxy group. While the α-positions of the cyclic ketone were partially deuterated during the process, subsequent treatment with pyrrolidine and water restored ¹Hs at these acidic sites and furnished the synthesis of d₂-tetrabenazine with good diastereoselectivity. This synthesis suggests that controlled partial deuteration could be achieved and used for late-stage modifications.

In summary, we have developed a unique deconstructive approach to allow alkyl deuteration in a site-specific and degree-controlled manner using readily available ketones as traceless handles. The reaction is redox neutral without using stoichiometric metals, and D₂O is employed as the deuterium source. It also shows a broad scope and high functional group tolerance. This strategy is anticipated to be valuable for preparing complex deuterated or potentially tritiated molecules, particularly when fine tuning of their properties is needed.