Pentanidium-Catalyzed Direct Assembly of Vicinal All-Carbon 1 Quaternary Stereocenters through C(sp 3 )-C(sp 3 ) Bond Formation

12 The stereoselective construction of vicinal all-carbon quaternary stereocenters has long been a formidable 13 synthetic challenge. Direct asymmetric coupling of a tertiary carbon nucleophile with a tertiary carbon 14 electrophile is the most straightforward approach but it is sterically and energetically disfavored. Herein, 15 we described a catalytic asymmetric substitution, where racemic tertiary bromides directly couple with 16 racemic secondary or tertiary carbanion, creating a series of congested carbon (sp 3 )-carbon(sp 3 ) bonds, 17 including isolated all-carbon quaternary stereocenters, vicinal tertiary/all-carbon quaternary stereocenters 18 and vicinal all-carbon quaternary stereocenters. This double stereoconvergent process, using pentanidium 19 as catalyst, affords substituted products in good enantioselectivities and diastereoselectivities.


Introduction 21
The use of high-throughput synthetic practices, in tandem with extensive use of Pd-coupling chemistry in 22 medicinal chemistry laboratories world-wide, has led to a propensity of achiral, aromatic compounds in 23 screening libraries. 1 Many secondary metabolites with interesting pharmacological activities contain all-24 carbon quaternary stereocenters. 2,3,4 Introducing all-carbon quaternary stereocenters into molecules will 25 improve structural diversities in screening libraries. However, stereoselective construction of all-carbon 26 quaternary stereocenters remains a significant challenge in synthetic chemistry. 5,6 Amongst the limited 27 2 number of strategies, for the formation of this highly congested moiety, double Heck coupling, 7,8 double 28 Aldol reaction, 9 and double allylation 10 have been reported to be useful ( Figure 1a). Separately, the use of 29 multi-substituted alkenes in [3+2] annulation, 11,12 Diels-Alder 13-15 and other cycloadditions 16,17 is another 30 common approach (Figure 1a). Recent advances include dearomatization addition of b-naphthols on 3-31 bromooxindoles, 18 Claisen rearrangement of g,d-unsaturated carbonyl compounds, 19 dialkylation of 32 bisoxindoles, 20 phosphine-catalyzed cyclization of allenes 21 and a nucleophilic substitution at a quaternary 33 carbon center with concomitant opening of a cyclopropane ring. 22,23 On the other hand, direct radical 34 coupling of two C(sp 3 ) centers is a promising possibility as it can overcome steric hindrance; but currently 35 it is limited to a narrow substrates scope such as bisoxindoles and chiral auxiliaries need to be deployed if 36 prepare vicinal all-carbon quaternary stereocenters through a catalytic asymmetric coupling of two 38 tertiary C(sp 3 ) centers. This will be the most direct and convenient and yet, conceivably, the most 39 sterically challenging approach. Nucleophilic substitution at a quaternary carbon center is difficult and 40 can be made improbable if the nucleophile is also a bulky tertiary carbanion. 41 We have been developing chiral cationic salts such as pentanidium and bisguanidinium as phase transfer 42 and ion-pair catalysts. 28 Using these catalysts, we recently reported an enantioconvergent halogenophilic 43 nucleophilic substitution (SN2X) to generate enantioenriched quaternary stereocenters using thiols and 44 azides. 29-31 In a conventional SN2 substitution, the nucleophile displaces a carbon-bound leaving group X, 45 often a halogen, by attacking the carbon face opposite the C-X bond; while in the SN2X reaction, the 46 nucleophile approach a carbon-bound leaving group X from the front, making it an ideal sterically-47 immune synthetic approach. Shortly thereafter, a more in-depth investigation of the azide-substitution 48 with tertiary bromide, revealed that it is a dynamic kinetic resolution, modulated by base present in the 49 reaction. 32 Herein, we report our recent progress into the use of nucleophilic substitutions to construct 50 vicinal all-carbon quaternary stereocenters, using insights from our previous works, through direct 51 3 coupling of racemic tertiary electrophiles with racemic tertiary nucleophiles using chiral cations as 52 catalysts (Figure 1c).

57
We began our investigation by extending our previous work on enantioconvergent SN2X substitution. 58 Instead of thiols and azides, we wanted to demonstrate that carbon nucleophiles can add to racemic 59 tertiary bromides. Firstly, methyl 2-bromo-2-cyanoacetate 1a was chosen as the model and various carbon 60 pronucleophiles activated by an electron-withdrawing group such as acetophenone, isobutyronitrile and 2-61 nitropropane, were examined under basic condition (Scheme 1a). We found that only protonated product 62 1a-H was obtained via a base-mediated SN2X debromination process. Further exploration revealed that 63 carbon pronucleophiles with two electron withdrawing groups, such as malononitrile and dialkyl 64 malonate, afforded the desired substituted products (Scheme 1b).

67
Subsequently, we found that in the presence of pentanidium PN1-3 or bisguanidinium BG1-3 as catalyst, 68 substituted product 2a was obtained with moderate yields and ee values ( Table 1, entries 1-6). 69 Bisguanidinium BG1, bearing 3,5-bis(trifluoromethyl)benzyl groups, provided the most promising results 70 (entry 4). Further optimization by investigating various bases (entries 7-8), solvents (entries 9-11) and 71 temperature (entries 12-13) revealed that the ideal condition was using BG1 as catalyst, 4M aq. KOH (1.5 72 equiv.) as base in toluene at -30 o C. Lowering the reaction temperature further to -40 o C led to a 73 significant decreased in yield, due to an increase formation of protonated product 1a-H (entry 13). When 74 methyl ester 1a is change to ethyl ester 1b, ee value of adduct 2b is improved to 84% (entry 14). Further 75 increase in steric bulk of the tertiary bromides led to iso-propyl ester 2c and tert-butyl ester 2d with even 76 higher ee values (entries 15-16). However, changing dimethyl malonate to diethyl malonate or 77 diisopropyl malonate only led to an increased formation of 1a-H. 78 79

83
Under the ideal set of conditions developed above, various tertiary bromides 3d-16d were further 84 evaluated ( Figure 2). Both electron-withdrawing group and electron-donating groups of the benzyl-85 substituted substrates were tolerated (2e-2i). Replacing the phenyl group with a naphthyl group, thiophene 86 or pyridine also resulted in good yields and ee values of the adducts 2j and 2k, 2l respectively. Tertiary 87 bromides with alkyl groups can afford the desired substituted adducts in good yields and ee (2m-2n). The 88 reaction was also effective for tertiary bromides bearing allylic or alkene substituents (2o-2q).

96
Following the success of generating enantioenriched quaternary carbon centers through the addition of 97 dimethyl malonate to racemic tertiary bromides, we wonder if significant diastereoselectivity will be 98 observed when esters groups on malonates were different. Thus, tertiary bromide 1a was treated with 99 ethyl methyl malonate 18a (Table 2, entry 1); it was found, after screening our catalyst library, that PN1 100 can provide adduct 19a with moderate enantioselectivity and some diastereoselectivity. By introducing 101 iPr (18b), Bn (18c) or tBu (18d) groups to monomethyl malonates to increase steric discrimination, we 102 found that diastereoselectivities increased correspondingly (entries 2-4). However, using ethyl iso-propyl 103 malonate 18e did not further improve the diastereoselectivity observed and the yield decreased 104 dramatically (entry 5). When ethyl tert-butyl malonate 18f was used, mostly protonated product 1a-H was 7 obtained (entry 6). Thiolate 18g produced the corresponding adduct but ee and dr values obtained were 106 moderate (entry 7). Amide 18h was also examined but no desired adduct was observed (entry 8). Further 107 investigations were conducted with methyl tert-butyl malonate 18d (entries 9-11) by varying different 108 bromides and found 1d gave 19k the best results with 90% ee and 49:1 dr (entry 11). 109 110

114
With this optimized reaction conditions in hand, various tertiary bromides were studied (Figure 3). 115 Tertiary bromides with benzylic substitutions, heterocycles, alkyl and allylic substituents that were 116 investigated, afforded their corresponding adducts 19l-19s in good yields and stereoselectivity.

139
Subsequently, we identified cyclic b-ketone ester 21a as a suitable model to study this reaction (Table 3). 140 It allowed the coupling with tertiary bromide 1a to proceed (entry 1). From our previous studies, we 141 concluded that steric effect played a crucial role in enantioselectivity and diastereoselectivity. When 142 tertiary bromide 1b was investigated, we found that it led to an increased yield of protonated product and 143 with tertiary bromide 1c, no desired product was obtained. On the other hand, changing cyclic b-ketone 144 ester 21 led to more interesting results. When tert-butyl ester 21c was used, both ee and dr values of the 145 corresponding adduct were increased (Table 3, entries 1-3). We hypothesized that the protonated product 146 could be suppressed if we removal of water from the reaction condition. Thus, in order to improve the 147 yield of adduct 22, we need to choose a more suitable base. We investigated a series of bases ranging 148 from powdered hydroxides salts to carbonates (Table 3, entries 4-10). We found that carbonate salts gave 149 reproducible results with high yields and stereoselectivities; in particular, Cs2CO3 proved to be the more 150 reliable (entry 9). With Cs2CO3 and at a lower reaction temperature, the ideal reaction condition was 151 found (entry 11).

158
With the goldilocks zone identified, we expanded investigation of the scope of the tertiary bromides that 159 we can used. We report successful cases, which the reaction proceeded smoothly with good yields and 160 stereoselectivities (Figure 4, 22d-w). For benzyl substitutions in bromides, both electron-withdrawing and 161 electron-donating groups were tolerated (22d-k). Heterocycle such as thiophene was well tolerated (22l). 162 Simple alkyl groups also produced good results (22m-p). Olefin containing alkyl chains were transformed 163 into the desired product with good yields and stereoselectivities (22q). Substitution on cyclic b-ketone 164 ester 21e-f was also well tolerated (22r-w). Attempts to expand to other tertiary carbon nucleophiles such 165 as tert-butyl 1-oxo-1,2,3,4-tetrahydronaphthalene-2-carboxylate and tert-butyl 2-oxocyclopentane-1-166 carboxylate were not successful. We continue to explore other potential tertiary carbon nucleophiles. 167

12
In order to gain a better understanding of the mechanism, control experiments were designed accordingly. 171 Firstly, a carbanion-exchange experiment was conducted between tertiary bromide 1a and cyclic 172 b-ketone ester 21c. The reaction temperature was lowered from -20 o C to -40 o C and the reaction was 173 quenched using sat. NH4Cl after 8 hours. The transfer of Br atom from 1a to 21c was evident through the 174 significant production bromide 23 (Scheme 3a). However, both protonated product 1a-H and bromide 23 175 were obtained as racemic mixtures. Separately, a carbanion-trapping experiment using acrylonitrile, 176 further substantiate the presence of a carbanion intermediate, generated from tertiary bromide 1d (Scheme 177 3b). The conjugated addition product 24 was obtained with moderate enantioselectivity, pointing to close 178 ion-pair interaction of the carbanion with bisguanidinium BG1. Next, we prepared the enantioenriched 179 tertiary bromide 23 by using preparative high-performance liquid chromatography and subjected them to 180 our conditions separately (Scheme 3c). We found that both enantioenriched tertiary bromides 23 were 181 transformed to the same stereoisomer 22c. Lastly, a base mediated racemization was observed when 182 treating enantioenriched bromides 23 with Cs2CO3, this indicated a Cs2CO3 induced dynamic kinetic 183 resolution prior the C-C bond coupling, which contributes to the high stereoselectivity (Scheme 3d). 184 Based on previous investigations (Scheme 4a) and our preliminary studies, we proposed that cyclic b-185 ketone ester 21c and tertiary bromide underwent carbanion-exchange through SN2X (Scheme 4b). Cyclic 186 b-ketone ester bromide 23 that is generated in this step can undergo further racemization through SN2X 187 that is modulated by base. Finally, SN2 substitution occurred between the PN1 paired carbanion generated 188 from tertiary bromide A and cyclic b-ketone ester bromide 23 to install the vicinal all-carbon quaternary 189 stereocenters through the coupling of two C(sp 3 ) centers. 190

191
In conclusion, we have successfully developed the pentanidium-catalyzed direct coupling of tertiary 192 carbon nucleophiles and tertiary carbon electrophiles through C(sp 3 )-C(sp 3 ) bond formation. These 193 reactions allowed the direct construction of the challenging vicinal all-carbon quaternary stereocenters in 194