Studies on the Lithiation, Borylation, and 1,2‐Metalate Rearrangement of O‐Cycloalkyl 2,4,6‐Triisopropylbenzoates

Abstract A broad range of acyclic primary and secondary 2,4,6‐triisopropylbenzoate (TIB) esters have been used in lithiation‐borylation reactions, but cyclic TIB esters have not. We have studied the use of cyclic TIB esters in lithiation‐borylation reactions and looked at the effect of ring size (3‐ → 6‐membered rings) on the three key steps of the lithiation‐borylation protocol: deprotonation, borylation and 1,2‐metalate rearrangement. Although all rings sizes could be deprotonated, the cyclohexyl case was impractically slow, and the cyclopentyl example underwent α‐elimination faster than deprotonation at −78 °C and so could not be used. Both cyclobutyl and cyclopropyl cases underwent rapid borylation, but only the cyclobutyl substrate underwent 1,2‐metalate rearrangement. Thus, the cyclobutyl TIB ester occupies a “Goldilocks zone,” being small enough for deprotonation and large enough to enable 1,2‐migration. The generality of the reaction was explored with a broad range of boronic esters.


Introduction
Boronic ester homologation represents ap owerful method in asymmetric synthesis. [1] Originally done through substrate control using ac hiral diol on the boronic ester backbone [2] it was later shown that reagent control offered greater flexibility and when used in an iterative manner was also more effective and more efficient. Both substituted chlorosulfoxides [3] and hindered carbamates or benzoates [4] have been investigated for reagent-controlled homologation, the latter showing broader substrate scope.T he hindered carbamates or benzoate esters (for example, O-alkyl 2,4,6triisopropylbenzoate esters (TIB esters) [5] )a re deprotonated by strong base in the presence of diamine ligands and trapped by boronic esters,l eading to intermediate boronate complexes,which upon warming, undergo 1,2-metalate rearrangement, forming homologated boronic esters (Scheme 1A). [6] Thec arbamate (or TIB ester) serves not only as ad irecting and stabilizing group for the lithiation step,b ut also as al eaving group in the 1,2-metalate rearrangement step.T he reaction is stereospecific,e nabling homologated boronic esters to be produced with high enantioselectivity from enantioenriched lithiated carbamates/TIB esters.T his lithiation-borylation protocol has been used extensively in the synthesis of natural [1,7] and unnatural [8] products.B oth pri-Scheme 1. A) Current lithiation-borylation conditionswith acyclic alkyl substrates.B )F actors affecting deprotonationa nd 1,2-metalaterearrangement of cyclic TIB esters. C) This work:the effect of ring size on lithiation-borylation processes. TMEDA = N,N,N',N'-tetramethylethylenediamine, TIB = 2,4,6-i-Pr 3 C 6 H 2 C(=O), TP = 2,4,6-i-Pr 3 C 6 H 2 . mary and secondary carbamate/TIB esters can be employed, but in the case of secondary substrates which do not have additional anion-stabilizing groups (for example,a ryl, [9] allyl, [10] or propargyl [11] )m uch more forcing conditions are required for deprotonation. [12] In these cases TIB esters were found to facilitate deprotonation better than the corresponding carbamates.T he lithiation-borylation methodology has been explored extensively on acyclic substrates but little work has been done on cyclic substrates that do not have additional anion-stabilizing groups.S ince carbocycles are ubiquitous in nature we were interested in exploring the scope of this methodology to cyclic variants bearing different ring sizes.
In traversing the series of cyclic alkanol TIB esters, deprotonation becomes progressively easier with decreasing ring size due to the increasing s-character of the C À Hb ond (Scheme 1B). [13] In fact, there is only one report on the deprotonation of cyclic TIB esters,t hat of the cyclopropyl TIB ester,t he most acidic in the series. [14] Indeed, the high acidity of cyclopropanes enabled even bromocyclopropane to be deprotonated with lithium tetramethylpiperidine (LiTMP), and when (1-bromocyclopropyl)lithium was generated in the presence of boronic esters,b orylation and 1,2metalate rearrangement ensued. [15] Numerous reports detail the lithiation of benzo fused 5-and 6-membered systems however these benefit from stabilization by the adjacent phenyl ring. [7b,16] For1,2-migration of the boronate complexes, the opposite trend is expected:i tb ecomes progressively easier with increasing ring size due to the reduction in ring strain in the transition state (TS). In the case of the reaction of (1-bromocyclopropyl)lithium with boronic esters,1 ,2-migration is enabled by having av ery good leaving group.T he competing trends in ease of lithiation with decreasing ring size and ease of 1,2-migration with increasing ring size warranted afull investigation of lithiation-borylation of cyclic substrates. We now report that of the 3-6-membered rings investigated, only 4-membered rings can be successfully employed. They occupy a" Goldilocks zone", where small rings are required for deprotonation and large rings are required for 1,2migration (Scheme 1b), although this study revealed afurther important factor in determining success,t he stability of the intermediate lithiated TIB ester towards a-elimination. This work is not only of fundamental interest, but it is also of practical utility since the success with 4-membered rings leads to cyclobutane products which are useful and highly soughtafter, particularly in medicinal chemistry as they provide ad efined spatial arrangement of groups due to their rigid scaffolds. [17]

Results and Discussion
Our investigation began with the synthesis of the cyclic TIB esters 1-4.A fter some experimentation we found that the cyclopropyl ester could be made by S N 2d isplacement of the corresponding bromide with 2,4,6-triisopropylbenzoic acid (TIB acid) (Scheme 2), whilst the 4-6-membered ring TIB esters were best made by Mitsunobu reaction of the corresponding alcohols.C yclopentanol and cyclohexanol proceeded smoothly (see Supporting Information for details), however in the case of the cyclobutanol, asmall amount (5 %) of cyclopropylmethyl TIB ester 5 was obtained. This was presumably formed from rearrangement of ac yclobutyl carbocation, generated by an S N 1p athway,t oacyclopropylmethyl carbocation which was then trapped by the carboxylic acid (Scheme 2). [18] This side product was not separable by chromatography,b ut fortunately the minor component did not interfere with the subsequent chemistry.
There are three steps associated with lithiation-borylation reactions:1 )d eprotonation to form the organolithium;2 ) borylation to form the boronate complex;and 3) 1,2-metalate rearrangement. In situ IR spectroscopy was used to optimise the first two steps,l ithiation and borylation, as it allows determination of reaction times and desired stoichiometry in asingle experiment, where the signal intensity of the carbonyl group is followed for each intermediate. [19] In all examples, as olution of TIB ester 1-4 (0.3 m,E t 2 O) and N,N,N',N'tetramethylethylenediamine (TMEDA, 1.2 equiv) was cooled to À78 8 8Ca nd then s-BuLi was added (1.3 m in cyclohexane, 1.2 equiv). Studying the most acidic cyclopropylv ariant first (Scheme 3A), upon addition of base ar apid decrease in the intensity of the signal attributed to the starting TIB ester ( % 1730 cm À1 )w as observed. At the same time,asignal at al ower wavenumber 1648 cm À1 grew in intensity,w hich we attributed to the lithiated species 1-Li,w hich then plateaued and remained horizontal over 10 minutes showing that the lithiated cyclopropyl TIB ester was chemically stable under the reaction conditions.T he lithiation was essentially instantaneous,being complete by the end of the dropwise addition of the base.Asolution of phenethylboronic acid pinacol ester (1.0 m,E t 2 O) was then added over 2minutes and another peak appeared (1682 cm À1 ), again instantaneously,w hich is indicative of the boronate complex 1-B.E ven though the organolithium is tertiary (albeit bearing as mall cyclopropyl group) rapid borylation ensued.
Thecyclobutyl TIB ester 2 was then subjected to the same sequence (Scheme 3B). In this case deprotonation was no longer instantaneous but it was still rapid, taking 10 minutes to reach aplateau. In contrast to the cyclopropyl example,the lithiated cyclobutyl substrate 2-Li was not chemically stable under the reaction conditions and decayed slowly over time. Ab roader second peak appeared at lower wavenumber 1581 cm À1 ,w hich is attributed to the carboxylate salt 6,a nd Scheme 2. Synthesis of small ring benzoatesand side-products observed.

Angewandte Chemie
Forschungsartikel was observed to grow after % 50 minutes indicating slow decomposition by a-elimination. Addition of phenethyl boronic acid pinacol ester then led to instantaneous borylation, like the cyclopropyl example,a ffording the cyclobutyl boronate complex 2-B.Asthe deprotonation of the cyclobutyl TIB ester was especially facile,w ea lso attempted the lithiation-borylation of bromo-and chlorocyclobutane,i n as imilar way to that employed for bromocyclopropane. However,weonly observed trace product, presumably due to the enhanced rate of a-elimination from having ab etter leaving group (see Supporting Information for details).
Increasing the ring size further to the cyclopentyl TIB ester 3 led to unexpected results.W hile the starting material was converted more slowly than the smaller ring systems (as expected), there was no observable lithiated species 3-Li. Instead, only the broad peak of the carboxylate 6 was observed (Scheme 3C). This indicates that the lithiated species undergoes decomposition (presumably by a-elimina-tion) at afaster rate than its formation. This was confirmed by addition of d 4 -methanol (5 equiv) 4h after addition of the base (when the formation of carboxylate had stopped by in situ IR spectroscopy), which led to al ow recovery (27 %) of non-deuterated (0 %d euterium incorporation) starting material. Attempts to stabilize the cyclopentyl lithiated species using the diisopropyl carbamate (a better stabilizing group and worse leaving group) were unsuccessful and no lithiated species was observed by in situ IR spectroscopy (see Supporting Information for details). We were surprised at the high instability of the lithiated cyclopentyl TIB ester and carbamate.F inally,t he cyclohexyl TIB ester 4 showed much slower deprotonation still. After 22 ha tÀ78 8 8Ct he reaction reached ap lateau, with the in situ IR spectroscopy trace showing low conversion (Scheme 3D). Quenching the reaction at this time with d 4 -methanol gave 93 %r ecovery with only 26 %deuteration which is clearly impractical. Being able to observe the lithiated benzoate 4-Li by in situ IR Scheme 3. In situ IR spectroscopy traces for 3-6-membered cycloalkyl TIB esters. A) Lithiation and borylation of the cyclopropyl TIB ester.The trace shows that the deprotonation is very rapid, the lithiated species is stable at À78 8 8C, and that borylation is rapid. B) Lithiation and borylation of the cyclobutyl TIB ester.T he trace shows that deprotonationi sr apid, the lithiated species slowly decomposes at À78 8 8C, and that borylation is rapid. C) Lithiationo fthe cyclopentyl TIB ester.The trace shows that the lithiated species is not stable and decomposes to the carboxylate. D) Partial lithiation and deuteration of the cyclohexylT IB ester.T he trace shows that the lithiation is slow and inefficient.

Angewandte Chemie
Forschungsartikel spectroscopy and the fact that we isolated 4-D showed that the lithiated benzoate 4-Li was much more stable than the 5membered ring analogue 3-Li.A tÀ60 8 8Ct he deprotonation plateaued after 4h and quenching with d 4 -methanol gave similar levels of recovery (95 %) and deuterium incorporation (21 %) as at À78 8 8C. We have previously shown that s-butyl TIB ester gave 35 %yield in alithiation-borylation process [12] at À60 8 8Cafter a2hlithiation time,indicating that cyclohexyl TIB ester is deprotonated even more slowly than acyclic substrates.H aving established that both the cyclopropyl and cyclobutyl TIB esters could be lithiated and borylated, our attention turned to the 1,2-metalate rearrangement. Beginning with the cyclopropyl boronate complex 1-B we attempted to promote 1,2-metalate rearrangement but even under av ariety of conditions (e.g.M gBr 2 ,s olvent swap to CHCl 3 ), all reactions either returned starting material or resulted in decomposition (see Supporting Information for details). This clearly indicated that the barrier to 1,2-migration for the 3membered ring was higher than alternative decomposition pathways or reversion to the starting components.I nterestingly,c yclopropyl boronate complexes bearing ab romide leaving group do undergo 1,2-metalate rearrangement, [15] highlighting the difference the nature of the leaving group can make.T urning to the cyclobutyl boronate complex 2-B, we found that this time the 1,2-metalate rearrangement began to occur at room temperature,but the reaction was slow.Even with heating in Et 2 O, boronate complex remained (Table 1, entries 1and 2). Use of MgBr 2 was not effective at promoting the 1,2-metalate rearrangement (entry 3), but we found that as olvent switch to CHCl 3 followed by heating to 60 8 8C enabled complete 1,2-metalate rearrangement to occur in just 3h,f urnishing the cyclobutylboronic ester 7 in 67 %i solated yield. Solvent exchange to an on-coordinating solvent, like CHCl 3 ,h as previously been found to promote 1,2-migration of recalcitrant boronate complexes. [20] As mall amount of Omigration of the pinacol group was also observed for all entries (8: < 10 %). Having developed asuccessful lithiationborylation protocol for the cyclobutyl TIB ester (entry 4), we explored the scope of this process with different boronic esters (Scheme 4).
Ther eaction proceeded well for ad iverse collection of primary boronic esters including those bearing nitrile (10), ester (11)and azide (12)functional groups.Alower yield was observed for the azide 12 presumably due to competing nucleophilic addition of the organolithium 2-Li to the azide in the starting boronic ester.R eaction with ac omplex lithocholic acid derivative 13 also proceeded in good yield. Secondary boronic esters also worked well, with examples including cyclohexyl 14,cyclopropyl 17, N-Boc-pyrrolidine 18 and piperidines 19 and 21.A lthough a-amino substrates are poor migrating groups, [20a] they nevertheless proceeded in moderate yields (18 and 21). Furthermore,u sing chiral and non-racemic boronic esters the 1,2-metalate rearrangement was found to be completely stereospecific (15 and 16). To demonstrate scalability, N-Boc-piperidine 19 was prepared on gram scale.I nt he case of the menthyl derivative 20,l ittle product was formed but switching to the less hindered neopentyl glycol ester resulted in an increased 60 %y ield. Unusually,t he neopentyl glycol boronic ester product was stable to silica gel chromatography.T his hindered secondary boronic ester turned out to be the limit of reactivity with secondary TIB esters.Noboronate complex was observed by in situ IR spectroscopy with t-Bu pinacol boronic ester but boronate was observed using the neopentyl glycol ester (see Supporting Information for details). However,d espite formation of the neopentyl glycol boronate complex, no product was obtained after attempted 1,2-metalate rearrangement. Presumably,t he hindered boronate complex reversed to starting materials upon heating. Similar observations were observed with acyclicsecondary TIB esters,indicating that it is apparently too demanding for this methodology to place two quaternary centres next to each other using boronic esters,a lthough this problem could be overcome using boranes. [21] Ar ange of sp 2 boronic esters were also explored. Both electron poor and electron rich aromatics 22-24,a sw ell as heteroaromatics,such as benzofuran 26 and indole 27 worked well giving the products in good yield. Finally,alkenyl boronic esters performed well, providing tertiary allylic boronic esters 28-30 in good yields.
To further illustrate the utility of the cyclobutyl boronic ester products,wetransformed the boronic ester functionality present in substrate 19 into ar ange of functional groups (Scheme 5). Zweifel olefination with propenyllithium gave the olefin 31 in excellent yield, [22] and alkynylation with vinyl carbamate gave the alkyne 34 in high yield. [23] Thet ertiary boronic ester underwent aM atteson homologation to give ap rimary boronic ester product 32. [24] Theb oronic ester was also converted into the tertiary amine in moderate yield, which was protected as the carbamate 33. [25]

Conclusion
We have studied the lithiation-borylation of as eries of cyclic TIB esters and have shown that the success of the process is governed by adelicate balance of factors involving [a] The ratio of O-migrated side-product (d % 50 ppm) to C-migrated product (d % 32 ppm) to boronatec omplex (d % 8ppm), ratio determined by 11  ease of lithiation, stability of the organolithium, and ease of 1,2-migration. Of the ring sizes studied, deprotonation became progressively slower going from 3-! 6-membered rings,w ith the cyclohexyl substrate being too slow to be practical. Theo rganolithium intermediate was prone to aelimination and while the lithiated 3-and 4-membered rings were stable,the lithiated cyclopentyl TIB ester underwent aelimination faster than deprotonation. The3 -a nd 4-membered rings both underwent rapid deprotonation and trapping with boronic esters but the 3-membered ring did not undergo 1,2-metalate rearrangement, presumably because of the high strain in the TS of the migration. Thec yclobutyl ring did undergo 1,2-metalate rearrangement. Thus,t he cyclobutyl ring occupies a" Goldilocks zone", where the ring is small enough to promote deprotonation, but large and flexible enough to allow the 1,2-metalate rearrangement to occur, and the organolithium is sufficiently stable.T he process shows broad substrate scope,a nd the applicability of these boron substituted cyclobutanes has been demonstrated by transforming the boronic ester into ar ange of functional groups.