Towards a Catalytic Atroposelective Synthesis of Diaryl Ethers via C(sp²)-H Alkylation Using Nitroalkanes

Abstract

Herein we report studies towards a small molecule catalytic approach to access atropisomeric diaryl ethers that proceeds via a C(sp²)-H alkylation using nitroalkanes as the alkyl source. A quaternary ammonium salt derived from quinine containing a sterically hindered urea at the C-9 position was found to effect atroposelective C(sp²)-H alkylation with moderate to good enantioselectivities across several naphthoquinone-containing diaryl ethers. Products can then be isolated in greater than 95:5 er after one round of trituration. For several substrates that were evaluated we observed a ‘nitroethylated’ product in similar yields and selectivities.

Graphical Abstract

Atropisomerism is a type of chirality that occurs when there is hindered rotation about a bond, typically between two sp² atoms, wherein the rotational isomers are enantiomeric.^1–5 While molecules that don’t possess adequate barriers to rotation aren’t considered atropisomeric,⁶ they possess the potential to be atropisomeric when engaging a chiral receptor (i.e. protein binding or interacting with a chiral catalyst). ^7–12

Previous work from our group and others have demonstrated that atropisomer confirmation can be leveraged to improve various properties of biologically active small molecules.^13–16 This work in part serves as call to action for the development of new synthetic methods towards atropisomeric compounds. While there have been several excellent examples of atroposelective catalysis over the past few decades,^17–20 the vast majority of examples have been on biaryl scaffolds. Examples of atroposelective catalysis on non-biaryl atropisomers have largely focused on benzamides and anilides.^2,8,9

Diaryl ethers are a type of atropisomer that have been largely overlooked by the enantioselective catalysis community despite their prevalence in natural products, exemplified by the macrocyclic diaryl ethers in vancomycin. Furthermore, atropisomerically unstable diaryl ethers are common motifs in drug discovery (Figure 1). Thus far, the literature involving diaryl ethers is highlighted by some nice diastereoselective examples en route to vancomycin²¹ and several excellent studies by Clayden and coworkers^22–26 wherein they characterized the stereochemical stabilities of differently substituted diaryl ethers, and developed diastereoselective routes to atropisomeric diaryl ethers. Notably, in collaboration with Turner, they disclosed the only catalytic atroposelective route towards diaryl ethers currently in the literature, employing oxidase or reductase enzymes to desymmetrize prochiral diaryl ethers (Scheme 1A).

Figure 1:

Illustrative examples of diaryl ethers

Scheme 1:

A) Previous catalytic atroposelective route towards diaryl ethers. B) Proposed enantioselective alkylation of diaryl ethers.

Other than these groundbreaking examples, work on the catalytic atroposelective synthesis of diaryl ethers has been scarce. Perhaps the major reason for this is that diaryl ethers possess two axes (Figure 1) that can complicate reaction development, analysis and in some cases, result in racemization at a lower-than-expected energy via a concerted gearing mechanism.^23,27 When defining the chirality of diaryl ethers, Clayden and others have made an analogy to atropisomeric biaryl systems, defining chirality based on the orientation of the substituents across both aryl planes when looking down one aryl-ether axis (see Figure 1, priority assigned in example according to R group’s number).

Recent work from our group has focused on the atroposelective synthesis of naphthoquinone-based biaryls via the rigidification of a rapidly interconverting axis via a 1,4-nucleophilic addition into the quinone.²⁸ As there is a lack of enantioselective routes towards atropisomeric diaryl ethers, we decided to test whether this approach could be extended to this scaffold (Scheme 1B). We chose to evaluate naphthoquinones such as 1a where the aryl groups possessed a large tert-butyl group and a 2^nd smaller substituent ortho-to the ether axis, as Clayden has shown that having one large quaternary substituent is often a prerequisite to obtain stereochemically stable diaryl ethers.

While evaluating nucleophiles for the addition into 1a, we observed that the use of nitromethane in the presence of excess Cs₂CO₃ and tetrabutylammonium bromide (TBAB) resulted in the isolation of a C(sp²)-H methylated product 2a (Table 1, entry 1), in line with seminal work reported by Mukherjee.^29,30 Quinine derived quaternary amines with hydroxyl substitution at the cinchona alkaloid C-9 position gave nearly no observable selectivities (see SI for details). On the other hand, catalyst C1, which has a C-9 stereochemically inverted Boc-protected amine (Table 1, entry 2), yielded preliminary levels of enantioselectivity, albeit with low conversions to 2a. The selectivity could be further improved to an enantiomeric ratio (er) of 85:15 when tert-butyl urea-containing catalyst C2 was used.

Table 1:

Optimization of C(sp²)-H Methylation 1a

Entry	Solvent	Catalyst	Base	Sieves	Yield (%)	er[b]
1	PhMe	TBAB	Cs2CO3	-	18	---
2	PhMe	C1	Cs2CO3	-	18	62:38
3	PhMe	C2	Cs2CO3	-	17	85:15
4	PhMe/H2O	C2	Cs2CO3	-	40	50:50
5	PhMe	C2	Na2CO3	-	20	60:40
6	PhMe	C2	K2CO3	-	15	65:35
7	PhMe	C2	K3PO4	-	60	80:20
8	PhMe	C2	K3PO4	4Å	51	75:25
9	PhMe	C1	K3PO4	4Å	36	64:36
10	PhMe	C2	K3PO4	3Å	68	81:19
11	CH2Cl2	C2	K3PO4	3Å	38	68:32
12	MTBE	C2	K3PO4	3Å	71	73:27

^[a]Reactions were performed on a 0.028 mmol scale with 10 equiv of both base and nitromethane in 0.1M solvent.

^[b]er was calculated by HPLC and results are reported as an average from at least two trials. See the Supporting Information (SI) for more details.

The addition of water, or the use of other carbonate bases, resulted in a loss in selectivity with minimal increase in yield (Table 1, entries 4–6), however the use of tribasic potassium phosphate (Table 1, entry 7) resulted in an increase in yield to 60% with only a slight drop in selectivity (er of 80:20) when compared to Cs₂CO₃. Finally, in line with Mukherjee’s work, we found that the addition of 3Å molecular sieves resulted in yields of 2a approaching 70% with a small increase in selectivity to 81:19 er (Table 1, entry 10). While these selectivities may be considered moderate, 1a can be further enantioenriched via trituration out of isopropanol to allow access of 2a in greater than 97:3 er. 2a proved to be moderately stable, with an experimentally determined barrier to rotation at 65^◦C of 26.6 kcal/mol, likely resulting in what LaPlante calls a ‘Type-II atropisomer” that would be expected to display significant racemization at room temperature over the course of a few weeks.

Extensive experimentation failed to yield further increases in er. Nonetheless, we decided to evaluate the conditions in Table 1, entry 10 across several differentially substituted substrate analogs (Figure 2).^31–33 Substrate 1b, where the aryl group possesses a 6-methyl substitution, resulted in a decrease in yield and enantioselectivity in the isolated product 2b to 54% yield with an er of 71:29. Substrate 1c, which possesses a 4-phenyl and 6-methyl substitution yielded 2c with similar selectivities. These results could be explained by an increase in electron density on the ether, reducing the electrophilicity of the quinone. Diaryl ether 1d, which possesses an electron deficient para-aryl ring, reinforces this hypothesis as the reaction proceeded with a large increase in yield of 2d while retaining the moderate enantioselectivity (er of 71:29).

Figure 2:

[a] Reactions were performed on a 0.028 mmol scale with 10 equiv of both base and nitromethane in 0.1M PhMe. ^[b] The formation of both products was observed for substrates a-i. The major product is shown above with its respective yield and er. More information about product ratios can be found in the Supporting Information. ^[c] er was determined via HPLC through an average from at least two trials. ^[d]2a was triturated with HPLC grade isopropanol. ^[e] A mixture of both methylated and ‘nitroethylated’ product was detected via mass spectroscopy; only the methylated product 2e was isolated and characterized. ^[f]3f was triturated with HPLC grade hexanes. See SI for more details.

Substrate 1e, which possesses 4-phenyl and 2-chloro substitution on the aryl ring also yielded moderate selectivities and yields of 2e (47% yield, 72:28 er), however, we also observed small amounts of ‘nitroethylated’ 3e (vide infra). Substrate 1f, which possesses 6-Br substitution, yielded a 1:3 mixture of methylated 2f (10%) and ‘nitroethylated’ 3f (30%) yield. Interestingly, while 2f was obtained in 78:22 er, 3f was obtained in only 60:40 er. Surprisingly, one round of trituration with hexanes allowed for isolation of 3f in 99:1 er, albeit in low overall yield. Substrate 1g, which possesses 4,6-diphenyl substitution yielded exclusively ‘nitroethylated’ 3g with a 75:25 er. We observed similar results with 1h and 1i which gave ‘nitroethylated’ 3h and 3i with similar yields and selectivities.

Referring back to Mukherjee’s hypothesized mechanism of C(sp²)-H alkylation using nitroalkanes, we postulate that our diaryl ether substrates proceed via a quinone methide intermediate.²⁹ If this intermediate is long lived enough, then another equivalent of nitronate anion can add 1,4 into the quinone-methide to yield the ‘nitroethylated’ product (Scheme 2). While the exact mechanism of quinone-methide stabilization is unknown, we suspect it is due to a subtle electronic effect as ‘nitroethylation’ is only observed with substrates that possess electron neutral or donating substitution para- to the ether.

Scheme 2:

Proposed mechanism for nucleophilic methylation. The key intermediate involves a quinone-methide which is influenced heavily by electronic effects off of the aryl ring. Stabilization of the intermediate can induce subsequent addition of nitromethane, forming the ‘nitroethylated’ product. We hypothesize that the urea catalyst interacts with both the diaryl ether and carbonyl oxygen of the substrate, thereby locking it into the (S_a)-exo- confirmation. Subsequent tautomerization or 2^nd addition of nitronate followed by oxidation provide diaryl ether products 2a and 3a, respectively.

We next sought to define the stereochemical induction of this reaction. As we were unable to obtain suitable crystals, we compared experimental and computational circular dichroism (CD) spectra, a method that is gaining acceptance in the stereochemical community.^34–36 We obtained CD spectra of highly enantioenriched (er>97:3) 2a and compared it to computational CD spectra of 2a of all possible stable conformations about both ether axes (Scheme 3). This analysis suggests that the major product is in the S_a configuration with the proximal quinone carbonyl endo- to the aryl ring (S_a-endo).

Scheme 3:

A dihedral angle contour diagram, generated through a force field molecular model, conveys an energetically favorable pathway towards racemization through a concerted gearing mechanism (See the SI for more details). Substrate-catalyst interaction can induce the high energy exo confirmation, followed by immediate “gearing” to the respective endo- enantiomer.

As we generated these conformations for the computational CD studies we observed striking differences in their predicted energies (Scheme 3). For example, the S_a-endo conformation of 2a is predicted to be significantly more stable than the S_a-exo conformation. Interestingly, we observed the opposite trend for the starting material 1a, with the exo- conformation being more stable (see SI for more details).

To further investigate this, we generated contour energy maps of the rotational landscape about both axes (Scheme 3) of the diaryl ether. Consistent with results from Clayden,²³ these maps demonstrate that there is a low energy pathway for interconversion between the exo- and endo- conformations of a given enantiomer that proceeds via a concerted gearing mechanism. Thus, it is likely that the exo- conformation of the starting material is more stable and will likely be the conformation that interacts with the catalyst, however, the addition of the methyl group will lead to an immediate conformational gear shift to the endo- enantiomers.

Our working model for stereoinduction is shown in the proposed transition state in Scheme 2 in which we propose that the urea moiety of the catalyst to hydrogen bond with both the ether oxygen and one of the quinone carbonyls of the lower energy exo- diaryl ether conformation. From here, the diaryl ether is preorganized into the S_a-atropisomer to avoid steric interactions between the tert-butyl group and the quinuclidinium that would be present in the R_a atropisomer/catalyst complex. We postulate that the hydrogen bonding will activate the diaryl ether for nucleophilic attack by nitronate anion which will then undergo HNO₂ eliminatiom followed by tautomerization to give alkylated quinone, or subsequent attack of nitronate followed by oxidation to give the ‘nitroethylated’ byproduct. At this point, it is likely that both products will rapidly relax to the endo- conformation, perhaps providing a release mechanism from the catalyst.

In conclusion we have disclosed to the best of our knowledge the first example of a small molecule catalytic synthesis of diaryl ethers. While our selectivities were moderate-to-good, highly enantioenriched ethers can be accessed via trituration. We also discussed several mechanistic aspects of this work. We hope that these studies will serve as a starting place for future efforts towards the enantioselective syntheses of diaryl ethers and related atropisomers.