Mono-endo -6- N,N -diethylcarbamoyl and Bis-endo , endo -6,12-N , N - diethylcarbamoyl derivatives of Tröger’s Base. Synthesis and exo-endo isomerization study

An efficient synthetic route to the mono-endo -6- N,N -diethylcarbamoyl and bis-endo , endo -6 , 12-N,N - diethylcarbamoyl derivatives of Tröger’s base (TB), endo - 7 and endo - 8 , is reported. Studies of reaction time, proton source


Introduction
First synthesized in 1887, 1 Tröger's base, 2,8-dimethyl-6H,12H-5,11-methanodibenzo[b,f] [1,5]diazocine, (TB, 1, (S,S) in Fig. 1), is a unique V-shaped C 2 -symmetric molecule with chirality arising from pyramidal nitrogen atoms structurally incapable of undergoing inversion.Many derivatives of TB have been developed for diverse applications.4][15][16][17] Both TB itself and certain derivatives have elicited interest as ligands in catalysis. 18,19ecent reviews detail the numerous possible applications of this important molecule and its congeners. 20 Due to the versatility of 1 and its derivatives, many research groups 24 including ours, [25][26][27][28][29][30] have developed procedures to modify 1 to prepare new derivatives and analogues.Having functional groups pointing towards the inside of the chiral pocket of 1 might allow the installation of new binding sites on the concave face of TB.To date, the only method available to synthesize endo-functionalized derivatives of 1 has been by oxidation of one or both of the benzylic methylene groups, which results in a twisted amide or diamide, respectively. 26,31uch amides lack the conjugation of conventional amides and are consequently more reactive, often behaving like ketones.Thus, a Wittig reaction and subsequent catalytic hydrogenation of the resulting alkenes and ester reduction selectively affords the endo-substituted compounds 4 and 6 shown in Scheme 1. 26 The major drawbacks of this method are the oxidation step used to prepare both monolactam 2 (10%) and dilactam 5 (28%) and the difficulties removing the phosphine oxide by-products resulting from the Wittig reaction, leading to the derivatives 4 and 6 after hydrogenation.
As part of our long-standing interest in lithiation chemistry of TB, 18,25,[32][33] we considered carbanionic chemistry to functionalize one or both of the endo positions of TB.We have previously reported 25 on the LDAmediated synthesis of mono-exo-6-and bis-exo,exo-6,12-N,N-dialkylcarbamoyl derivatives of TB, exo-7 and exo-8, respectively (Figure 2).Herein, we disclose a comprehensive and efficient synthesis of endo-7 and endo-8 through an exo,endo-isomerization study to achieve optimization of conditions for the formation of both exo and endo isomers of the mono-and di-carbamoyl derivatives.
Thus, we report a thorough investigation of factors of temperature, time, amount and type of base and chelating additive, and type of electrophile quench for the establishment of optimized conditions to obtain preparative amounts of 7 and 8, respectively.Additionally, we describe the reactions of a weak base (MeONa) and HCl with both proton and deuterium quench of TB and the mono-carbamoyl derivative 7. Of interest is the fact that isomerization of TB under acidic conditions occurs, as predicted from racemization studies, 13 via methylene bridge opening-reclosure and therefore, as expected, in the case of both 1 and derivative 7, does not lead to proton-deuterium exchange.

Results and Discussion
To begin, variation of the reaction temperature for exo-7 to endo-7 isomerization was studied and the results are displayed in Table 1.Thus, exo-7 and DIPA (diisopropylamine) were combined in a minimum amount of THF and the mixture was cooled to the desired temperature.Slow dropwise addition of n-BuLi was followed by stirring for 2 h and then protic quench with excess methanol.After workup, the crude product was dissolved in chloroform-d and the endo:exo ratio was established by comparison of the relative integrals of the peaks at δ5.40 and δ4.91 ppm in the 1 H NMR, which correspond to the proton alpha (H-6, Fig 2 ) to the carbonyl group in endo-7 and exo-7, respectively.Correspondingly for endo-8 and exo-8, the peaks at δ4.91 and δ4.89 ppm, were used.As seen in Table 1, the temperature that favors endo-7 product formation appears to be -20 °C (entry 5).At -10 °C, on the other hand, decarbamoylated product 1 was obtained (entry 6) by a Haller-Bauer process 34 in addition to a minor ketone product resulting from n-BuLi addition to the amide.The structural assignments were based on the 1 H NMR spectra (see the Supporting Information).We next tested a set of selected n-BuLi-amine base combinations for the isomerization reaction (Table 2).Not surprisingly, at -20 °C, n-BuLi alone (entry 1), with TMEDA (entry 2), or with DIPEA (Hünig's base) (entry 3) afforded the decarbamoylated product 1, the result of a Haller-Bauer reaction 34 LDA, whether generated In situ (entry 4) or as a commercial stock solution (entry 5), afforded a high ratio of endo-7:exo-7.A stock solution of LiHMDS, a weaker base (pK a ~26 of the corresponding acid) compared to LDA (pK a ~36 of the corresponding acid), gave a lower ratio of endo-7:exo-7 (entry 6).Two other lithium dialkylamine bases generated in situ from diethylamine (DEA) and dibutylamine (DBA) were tested.The latter (LiDBA) gave the best result (compare entries 7 and 8, Table 2).
Optimization of reaction time for the formation of endo-7 product was next studied (Table 3).Using the same conditions as those described for temperature optimization (Table 1) but maintaining the temperature at -20 °C, the reaction time was varied.The optimum time for the formation of endo-7 was 3 h (entry 4) after which time the product distribution remained constant (entries 4 and 5).This suggests that deprotonation is not rapid, as might be expected, since a bulky base must effect a proton abstraction on the concave face of endo-7.

Entry
Base (equiv)/Amine (equiv) Endo-7:Exo-7 Ratio a In view of the favorable isomerization result using LiDBA generated in situ, we tested this combination in the reaction as a function of time (Table 4).Rewardingly, the maximum endo-7:exo-7 ratio was observed in shorter reaction times (entry 3) than that obtained using LDA as the base (Table 2, entry 4) and this ratio underwent essentially no variation with increasing time periods (Table 4, entries 4-6).Variation of the n-BuLi-DBA stoichiometry from 1.1, to 2 or 3 equivalents under the conditions given in Table 4 showed only a minor effect on the endo-7:exo-7 ratio.This is not surprising, as we expect the endo:exo ratio to be determined at the acidic quench stage and the quench to be kinetic in nature.A preparative scale synthesis of endo-7 from exo-7 was carried out based on the optimum conditions for the isomerization, affording endo-7 in 86% yield of isolated product.A preparative synthesis of endo-8 was similarly carried out using the optimum conditions, as developed for the isomerization for exo-7, to afford the endo-8 in 54% yield of isolated product in addition to recovered starting material, exo-8 (25%) (Scheme 2).The structure of endo-8 was confirmed by single crystal X-ray analysis (Figure 3).Scheme 2. Conversion of exo-8 to endo-8 under LiDBA conditions.All of the preceding experiments involved quenching under kinetic conditions, i.e, the stereochemistry of the product is determined in the quenching of the intermediate enolate.We also wished to investigate the isomerization process under conditions of thermodynamic control.Thus, exo-7 and endo-7, (see Table 5, entries 1 and 2) were separately treated with a catalytic amount of sodium methoxide (0.05 equiv) in methanol for a period of 2 weeks at room temperature.We observed that, under these conditions, endo-7 underwent complete isomerization to exo-7 (Table 5, entry 2).Conversely, exo-7 did not undergo isomerization (Table 5, entry 1).Repeating the isomerization experiments of exo-7 and endo-7 in methanol-d 4 , we observed the exchange of the C-6 α-proton in the both cases for deuterium (Table 5, entries 3 and 4).These observations suggest that the isomerization of endo-7 and exo-7 takes place via a carbanion intermediate, most likely an aminoenolate, and that exo-7 is the more stable of the two isomers.As enantiomerically pure 1 is known to racemize under weakly acidic conditions, 35 treating either exo-7 or endo-7 with an acid should also result in isomer equilibration.This was indeed the case; treating both isomers independently with a catalytic amount of aqueous HCl resulted in the isolation of exo-7 in both cases, indicating that this isomer is the more stable of the two.Performing the experiment in methanol-d 4 using a catalytic amount of DCl (Table 5, entries 7 and 8), showed no exchange of any hydrogens in exo-7.This is consistent with the isomerization mechanism involving opening of the methylene bridge to form an iminium ion, 35 and not an enolization mechanism and raises the interesting question of whether any enantiomerically pure TB derivative with an exo substituent placed at a benzylic methylene group would be stereochemically stable in acid.We are interested in examining this question.
7][38] The amide was truncated to the corresponding dimethyl species (9) in order to simplify the calculations (Scheme 3).Scheme 3. In silico mechanism for the deprotonation of 9 and diisopropylamide.
The results showed exo-9 to be 2.0 kcal/mol lower in energy than endo-9, supporting the idea that exo-9 is the thermodynamically favored product.This is consistent with the experimental results obtained for both the acid and base-mediated equilibration of exo-and endo-7 (Table 5).
In order to assist in rationalizing the results of exo to endo isomerization using kinetic bases, experimental findings were compared to a computational examination of the reaction mechanism for the isomerization using an experimentally relevant deprotonated DIPA reaction initiator.These calculations examined the energies for each isomer to undergo complexation with a DIPA anion in the absence of lithium cations, followed by a kinetically limiting proton transfer to form a stable enolate-DIPA complex, which can then dissociate in THF to produce the free enolate intermediate.Examination of the computational mechanism (Scheme 3, Figure 4) indicates that both stereoisomers undergo deprotonation by the diisopropylamide anion, with a thermodynamic driving force of more than 13 kcal/mol and a kinetic barrier less than 20 kcal/mol, suggesting relatively rapid formation of the enolate species, although deprotonation is expected to be slower

AUTHOR(S)
for exo-9 vis-à-vis endo-9, as deprotonation of the former results from removal of an endo hydrogen residing on the concave face of the molecule.Initial complex formation between 9 and the diispropylamide anion leads to a hydrogen-bonded complex 10.This proceeds through transition state 11 to give an amide enolate complexed to diisopropylamine (12).Dissociation leads to the free enolate 13.Conversely, the quenching of the enolate intermediate in the presence of DIPA, may be treated as the protonation of the enolate complex by DIPA.Under these conditions, the thermodynamically favored exo-9 product is predicted to form via a transition state that is 5.0 kcal/mol higher in energy than the TS for the endo-9.This elevated transition state energy appears to stem from the sterically hindered approach of the DIPA proton source on the endo face of the aminoenolate.Instead, the open exo-face should more easily enable the approach of a proton source.Taken together, these computational results help confirm that quenching of the enolate should produce exo-9 as a thermodynamically favored product, while endo-9 will be kinetically favored (see Table S2, Supporting Information).

Conclusions
We have investigated the synthesis of endo-7 and endo-8 from their corresponding exo isomers (Figure 2).The reaction conditions were systematically optimized so that the maximum yield of the desired isomer was obtained.Furthermore, we established preparative routes to the endo-isomers from Tröger's base (1) (Scheme 4).This new approach to endo-substituted Tröger's base derivatives improves the overall yield over the previously reported method.The new method exhibits an overall yield of 60% for amide endo-7 and 52% for amide endo-8 over 2 steps as compared to 7% and 16% over 3 steps for the previous method, which was used to prepare esters 4 and 6, respectively (Scheme 1).8b Finally, our study paves the way for the use of other functional groups as side chains that are able to induce enolization, to yield endo substituents on the Tröger's base skeleton.Further studies regarding the chemistry of endo-functionalized TB derivatives will be reported in due course.

Experimental Section
General.Exo-7 and exo-8 were synthesized according to the protocol reported in our previous publication. 25ll chemicals were used as received from commercial sources without further purification.DIPA, DBA, TMEDA, DIPEA, DEA were dried with KOH pellets overnight and distilled under house vacuum prior to use.THF was dried with sodium and distilled under nitrogen.PE refers to petroleum ether (bp 40-60 °C).Precoated Merck silica gel 60 F 254 plates were used for TLC analysis.Column chromatography was performed on silica gel (Davisil 35-70 µm). 1 H and 13 C NMR spectra were recorded on a 400 NMR spectrometer.Chemical shifts (δ) are reported relative to a shift-scale calibrated with residual NMR solvent peak CDCl 3 (7.26ppm for 1 H NMR and 77.23 for 13 C NMR).Melting points were recorded on a Fischer-Jones apparatus, using plate technology.IR spectra were recorded on a FTIR spectrophotometer.HRMS were obtained on a Q-TOF microinstrument.Elemental analyses were performed by Mikroanalytisches Laboratorium KOLBE (Mülheim an der Ruhr, Germany).

Isomerization studies of exo-7 and endo-7
With MeOH/MeONa.Exo-7 or endo-7 (50.0 mg, 0.143 mmol) was dissolved in methanol (2 mL).Sodium (0.50 mg, 0.022 mmol) was added.The mixture was stirred for 24 h and neutralized by slow addition of HCl in methanol-d 4 (1.0 M).The solvent was removed in vacuo, the resulting residue was dissolved in CDCl 3 , and a 1 H NMR spectrum was recorded on the resulting solution to determine the product ratio.
With CD 3 OD/CD 3 ONa.Exo-7 or endo-7 (50.0 mg, 0.143 mmol) was dissolved in methanol-d 4 (2 mL).Sodium (0.50 mg, 0.022 mmol) was added and the resultant mixture was stirred for 24 h.The mixture was made neutral by slow addition of DCl in methanol-d 4 (1.0 M).The solvent was removed in vacuo, the residue was dissolved in CDCl 3 , and a 1 H-NMR spectrum was recorded of the resulting solution.With HCl/MeOH/H 2 O. Exo-7 or endo-7 (50.0 mg, 0.143 mmol) was dissolved in MeOH (5 mL).Aqueous HCl (1 M, 1 mL) was added and the mixture was stirred over 8-12 h at rt.The mixture was made alkaline (pH 10) by adding aqueous NaOH (1.0 M).Dichloromethane (5 mL) was added, the phases were separated, and the aqueous phase was extracted with additional dichloromethane (2×5 mL).The combined organic phases were collected and dried over Na 2 SO 4 , subjected to filtration, and the solvent was removed in vacuo.The residue was dissolved in CDCl 3 and the 1 H-NMR spectrum was recorded on the resulting solution.With DCl/CD 3 OD.Exo-7 or endo-7 (50.0 mg, 0.143 mmol) was dissolved in DCl in methanol-d 4 (1.0 M, 1 mL)).The mixture was stirred for 24 h and neutralized by slow addition of NaOMe in methanol-d 4 (1.0 M).The solvent was removed in vacuo.The residue was dissolved in methanol-d 4 and a 1 H NMR spectrum was recorded of the resulting solution.
Optimimized isomerization protocols for exo-7 and endo-7 Endo-7 (0.050 g, 0.14 mmol), was dissolved in freshly distilled THF (2 mL) under N 2 .Amine (1.1 equiv) was added and the solution was cooled to the desired temperature.A solution of n-BuLi (1.1 equiv) was added dropwise over 30 min.The reaction mixture was left for the required time and then quenched with the desired quenching medium.Water (5 mL) and dichloromethane (5 mL) were sequentially added and the phases separated.The aqueous phase was washed with additional dichloromethane (2×5 mL).The combined organic phases were washed with brine (5 mL) and dried (Na 2 SO 4 ) and subjected to filtration.The solvent was removed in vacuo to dryness.The residue was dissolved in deuterated chloroform and a 1 H NMR spectrum was recorded of the resulting solution.

Scheme 1 .
Scheme 1. Synthesis of twisted amides and TB derivatives.

Figure 2 .
Figure 2. Derivatives of Tröger's base of the present study.

a
Derived from1 H NMR spectra of the worked-up reaction mixtures (see Experimental Section) by comparison with authentic pure materials.Reaction run until no change in composition of reaction mixture.exo-7 and endo-7 were the only products unless otherwise stated.b Compound 1 was the only isolated material.

Figure 4 .
Figure 4. Free energy surface for the reaction mechanism outlined in Scheme 3.

Table 1 .
Isomerization of exo-7 to endo-7.Optimization of the Reaction Temperature

Table 3 .
Isomerization of exo-7 to endo-7.Optimization of the Reaction Time using DIPA as Base a Derived from 1 H NMR spectra of the worked-up reaction mixtures (see Experimental Section) by comparison with authentic pure materials.Reaction run until no change in composition of reaction mixture.exo-7 and endo-7 were the only products.b Variation of the quenching reagent (MeOH, water, NaHCO 3 aq sat'd solution, and i-PrOH/THF showed minor variation in endo-7:exo-7 ratios (79:21 -82:18).

Table 4 .
Isomerization of exo-7 to endo-7.Optimization of reaction conditions using n-BuLi/DBA a Derived from 1 H NMR spectra of the worked-up reaction mixtures (see Experimental Section) by comparison with authentic pure materials.Reaction run until no change in composition of reaction mixture.exo-7 and endo-7 were the only products.