Lewis acid-catalyzed Diels-Alder reaction of 2-cyclopentenones with Danishefsky’s diene: double bond isomerization of tetrahydro-1 H - indene-1,5(7a H )-diones, and attempts on an asymmetric catalysis

Details on the isolation of ketals 14


NMR-spectroscopic analysis of the inseparable double bond isomers 10, 12
The NMR-spectroscopic analysis of the isolated mixture of double bond isomers 10 and 12 is discussed in the following: Initially we expected the two compounds to be the two regioisomers 10 and 11 as shown below.

Scheme S1
If indeed the regioisomers 10 and 11 were formed, we should have found a multiple degenerated doublet in the range of 5 and 7 ppm in the 1 H-NMR spectrum for each of the vinylic protons 6-H and 7-H (regioisomer 10) and 4-H und 5-H (regioisomer 11) respectively.As shown in the enlargement of the olefinic part of the 1 H-NMR spectrum between 5.80 and 6.35 ppm in Figure S1, the NMR experiment provided a doublet of doublets (J = 10.1 Hz, J = 2.6 Hz) at 5.84 ppm and a doublet of doublets of doublets (J = 10.1 Hz, J = 4.2 Hz, J = 0.7 Hz) at 6.24 ppm.The third multiplet at 6.33 ppm was identified to be a triplet of doublets of doublets (J = 3.7 Hz, J = 3.2 Hz, J = 0.7 Hz).The integration of the triplet of doublets of doublets showed a value of one whereas the doublet of doublets of doublets and the doublet of doublets contained an integration of 0.75 each.This and the fact that both had an identical coupling constant (10.1 Hz) led us to the assumption that these two protons are located in the same molecule.The triplet of doublets of doublets was then inevitably assigned to the second compound.On closer examination of the integration we found a ratio of 0.75:1.75 or 0.43:0.57 of the two isomers that is attended by GC analysis (cf.Table 1).Conflicting our first theory of the two regioisomers 10 and 11 was the fact that we found only three signals for the four expected vinylic protons and the presence of nearly equimolar mixtures made us suspicious of the supposed structure of the two isolated products.In order to explore the real structures of the two compounds the HSQC spectrum of the mixture was examined carefully.A sector of the HSQC spectrum is shown in Figure S2.One of the carbon atoms (139.3 ppm) in the vinylic part of the spectrum shows no cross peak to any of the protons.Therefore it was assigned as a quaternary carbon.This assumption was confirmed in a DEPT-spectrum that is not shown here.ddd (J = 10.1 Hz, J = 4.2 Hz, J = 0.7 Hz) 6.24 ppm 0.75 H tdd (J = 3.7 Hz, J = 3.2 Hz, J = 0.7 Hz) 6.33 ppm 1 H dd (J = 10.1 Hz, J = 2.6 Hz) 5.84 ppm 0.75 H

Figure S2
HSQC spectrum of the inseparable mixture of double bond isomers 10, 12 At this point it was clear that one of the compounds had a quaternary olefinic carbon atom and owing to the two olefinic protons the other compound had to be one of the two initially expected ones.Altogether there were four possible isomers (Scheme S2).

Schema S2
The next issue was the position of the double bond.In order to solve that problem, the HMBC spectrum was investigated more closely.Possible correlations between the olefinic protons 6-H and 7-H (isomers 10, 10') and 4-H, 5-H (isomers 11, 11') respectively, and the carbonyl carbon atoms C-1, C-5 (isomers 10, 10') and C-1, C-6 (isomers 11, 11') respectively, should be observed in that spectrum and so the problem of the position of the double bond should be resolved unambiguously.In the case of isomer 10 there should be a cross peak of the protons 6-H and 7-H to both of the carbonyl C-atoms C-1 and C-5 whereas in the case of isomer 11 the olefinic protons 4-H and 5-H should correlate only to one of the carbonyl C-atoms, namely C-6.The same applies for the isomers 10' and 11'.A correlation of the olefinic proton 7-H to both of the carbonyl carbons should be detected in the case of isomer 10' and only one cross peak should be seen in the case of isomer 11'.A part of the HMBC spectrum is shown in Figure S3.

Figure S3
HMBC spectrum of the inseparable mixture of double bond isomers 10, 12 The olefinic protons at 6.24 ppm (ddd) and 6.33 ppm (tdd) are unambiguously correlated to two C-atoms.This verifies the structure of the two double bond isomers, that is about the two compounds 10 and 10' (= 12) (Scheme S3).It is not uncommon that there is no correlation in the HMBC-spectrum between the dd at 5.84 ppm and the carbonyl C-atoms, since such two-bondcouplings are not always visible in these HMBC spectra.

Scheme S3
This result also explains why there was obtained only one product in the selective reduction of the enone moiety.

Details on the isolation of ketals 14, 15
In the following the isolation of the pure monoketal 14 and diketal 15 and their NMR spectroscopic analysis is specified: As described in the main part the obtained monoketal 14 and the bisketal 15 could not be separated by chromatography (Scheme S4).But initially the structure of the two isolated products (one with m/z 194, the other with m/z 226) was not confirmed -at least four different monoketals 14 -14d and two different diketals 15, 15b were possible (Scheme S4).

Scheme S4
We felt confident that after reduction of the enone moiety of the monoketal to the corresponding alcohol, the remaining diketal could be separated from the resulting alcohol.The mixture of the monoketal and diketal was reduced with 4 equiv.NaBH 4 in the presence of 0.5 equiv.LiClO 4 in CH 3 CN. 2 Actually we obtained after reduction a mixture of diketal and diasteriomeric allylic alcohols (Scheme S5).The desired diketal could then be separated via preparative HPLC and analyzed in NMR experiments.The NMR sample, dissolved in CDCl 3 , was subjected to another GC-MS analysis and we found out that the compound was not the same anymore than before the NMR experiment: Under the slightly acid conditions of CDCl 3 one of the ketal protecting groups was cleaved off.This resulting monoketal showed exactly the same retention time as the monoketal gained during the first ketalization and so we assumed that the two monoketals had the same structure.There is shown the olefinic section of the resulting 1 H-NMR of the obtained monoketal in Figure S4.position of the ketal was determined by comparison of the chemical shifts of the carbonyl carbon in the 13 C-NMR spectrum with the 13 C-NMR spectrum of the double bond isomer 10.It is well known that carbonyl carbons in 5-membered rings are detected at lower field (approx.210-220 ppm) whereas conjugated carbonyl carbons are detected at higher field (approx.190-200 ppm). 3he double bond isomer 10 shows the carbonyl C-atom in the 5-membered ring at 211.5 ppm whereas the conjugated carbonyl carbon in the 6-membered ring shows up at 194.4 ppm.The carbonyl C-atom in the described monoketal shows one signal at 195.9 ppm.We concluded that the carbonyl carbon must be the one conjugated in the 6-membered ring (Scheme S5).

Scheme S5
The olefinic part of the 1 H-NMR spectrum of the diketal, dissolved in benzene-d 6 is shown in Figure S5.Again a doublet of doublets (5.67 ppm, J = 10.2Hz, J = 3.1 Hz) and a doublet of doublets of doublets (5.61 ppm, J = 10.2Hz, J = 1.6 Hz, J = 0.7 Hz) was observed.Since the difference of the chemical shifts (v 1 -v 2 ) ≈ J 1,2 the two signals are distorted.Again the similarity of the spectrum to the one of double bond isomer 10 is evident.Thus the structures of the resulting monoketal 14 and the diketal 15 could be established.L-Proline (5.00 g, 43.2 mmol) was dissolved in 86 mL of methanol.Potassium carbonate (5.97 g, 43.2 mmol) was added and the mixture was cooled to 0 °C.Then 9 mL (10.3 g, 95.0 mmol) ethyl chloroformate were added slowly.The resulting solution was allowed to warm up to room temperature and was stirred for further 12 h.The resulting colorless solid was filtered off, the solvent was reduced in vacuo and the remaining colorless oil was transferred to 20 mL of H 2 O. Then 25 mL of CH 2 Cl 2 were added and the layers were separated.The aqueous layer was extracted with 3 x 10 mL CH 2 Cl 2 and the combined dd (J = 10.2Hz, J = 3.1) 5.67 ppm 1 H ddd (J = 10.2Hz, J = 1.6 Hz, J = 0.7 Hz) 5.61 ppm 1 H organic layers were dried over MgSO 4. Concentration under reduced pressure afforded the product as colorless oil (7.62 g, 38 mmol, 87%).
A mixture of rotamers was obtained.In an oven dried Schlenk flask 7.36 g (0.30 mol) magnesium turnings were covered with 75 mL of THF.A small amount of bromobenzene (2 mL, 0.02 mol) was added to the mixture and the reaction was started using a heat gun.Then a mixture of 13.8 mL (20.7 g, 0.13 mol) bromobenzene and 25 mL of THF was added dropwise and after complete addition the grey solution was refluxed for further 30 min.Under an inert gas atmosphere (S)-N-(ethoxcarbonyl)proline methyl ester (27) (7.62 g, 37.8 mmol) was dissolved in 75 mL THF and cooled to 0 °C.The Grignard-solution was added slowly over a period of one hour and the resulting solution was stirred for further 10 h at room temperature.The reaction was hydrolyzed using 50 mL sat.NaHCO 3 solution.After the addition of 100 mL of dichloromethane the layers were separated and the aqueous layer was extracted with 3 x 15 mL CH 2 Cl 2 .The combined organic layers were dried over MgSO 4 and concentrated in vacuo.Purification by flash chromatography (2:1 hexanes/ethyl acetate) afforded the product as a colorless solid (3.60 g, 11.1 mmol, 30%).

29
(S)-N-Ethoxy-α,α-diphenyl-2-pyrrolidinemethanol (28) (3.60 g, 11.1 mmol) was dissolved in 46 mL of methanol.Potassium hydroxide (7.60 g, 0.13 mol) was added and the resulting solution was refluxed for 60 h.After concentration the resulting yellow oil was transferred to 30 ml H 2 O and 30 mL of dichloromethane were added.The layers were separated and the aqueous layer was extracted with 4 x 40 mL CH 2 Cl 2 .The combined organic layers were washed with brine, dried over MgSO 4 and concentrated in vacuo.After recrystallization from 3 mL of heptane 2.60 g (0.010 mol, 86%) of a colorless solid were obtained.Spectroscopic data were consistent with previously reported data for this compound. 4

(S)-Ethyl 2-(hydroxydinaphthalen-2-ylmethyl)pyrrolidine-1-carboxylate (30), (S)-1,1-Di(naphthalen-2-yl)tetrahydropyrrolo[1,2-c]oxazol-3(1H)-one (31)
In an oven dried Schlenk flask 0.60 g (24.7 mmol) magnesium turnings were covered with 20 mL of THF.A small amount of 2-bromonaphthalene (0.50 g, 2.40 mmol) in 2 mL of THF was added to the reaction mixture.The reaction was started using one crystal of iodine.Then a mixture of 1.99 g (9.6 mmol) 2-bromonaphthalene and 8 mL of THF was added dropwise and after complete addition the grey solution was refluxed for further three hours.Then 0.61 g (3.03 mmol) (S)-N-(ethoxycarbonyl)proline methyl ester (27) were dissolved in 30 mL of THF and cooled to 0 °C.Since the Grignard-solution was poorly soluble in cold THF, it was dissolved in hot benzene.The addition of the Grignard-solution was completed after one hour and the mixture was stirred for further 10 h at room temperature.The reaction was hydrolyzed using 10 mL sat.NaHCO 3 solution.After the addition of 20 mL of dichloromethane the layers were separated and the aqueous layer was extracted with 3 x 15 mL CH 2 Cl 2 .The combined organic layers were dried over MgSO 4 and concentrated in vacuo.Purification by flash chromatography (2:1 hexanes/ethyl acetate) afforded 0.50 g (42%) of a mixture (1:1) of desired product 30 and the cyclic carbamate 31.The mixture of (S)-1,1-di(naphthalen-2-yl)tetrahydropyrrolo[1,2-c]oxazol-3(1H)-one 31 and (S)-ethyl 2-(hydroxydinaphthalen-2-ylmethyl)pyrrolidine-1-carboxylate 30 (0.38 g, 1.24 mmol) was dissolved in 10 mL of methanol and an excess of potassium hydroxide (2.40 g, 0.04 mol) was added.The mixture was stirred for 15 h at room temperature, methanol was distilled off and the residue was dissolved in 5 mL of dichloromethane.Water was added (5 mL), the layers were separated and the organic layer was dried over MgSO 4 .After concentration in vacuo 0.27 g (0.76 mmol, 80%) of desired product was isolated as a colorless solid.In an oven dried Schlenk flask 4.3 g (0.18 mol) magnesium turnings were covered with 30 mL of THF.A small amount of 1-bromonaphthalene (0.5 g, 2.41 mmol) in 3 mL of THF was added to the reaction mixture.The reaction was started using one crystal of iodine.Then a mixture of 17.8 g (86.0 mmol) 1-bromonaphthalene and 20 mL of THF was added drop wise and after complete addition the grey solution was refluxed for further three hours.Then 4.43 g (22.0 mmol) (S)-N-(ethoxycarbonyl)proline methyl ester 27 were dissolved in 30 mL of THF and cooled to 0 °C.Since the obtained Grignard-solution was poorly soluble in cold THF, it was dissolved in hot benzene.The addition of the Grignard-solution was completed after one hour and the mixture was stirred for further 14 h at room temperature.The reaction was hydrolyzed using 100 mL sat.NaHCO 3 solution.After the addition of 100 mL of dichloromethane the layers were separated and the aqueous layer was extracted with 3 x 25 mL CH 2 Cl 2 .The combined organic layers were dried over MgSO 4 and concentrated in vacuo.Purification by flash chromatography (1:1 hexanes/ethyl acetate) afforded 1.45 g (22%) of monoarylated product 33.
The reaction was hydrolyzed using 10 mL sat.NaHCO 3 solution.After the addition of 10 mL of dichloromethane the layers were separated and the aqueous layer was extracted with 3 x 5 mL CH 2 Cl 2 .The combined organic layers were dried over MgSO 4 and concentrated in vacuo.
Purification by flash chromatography (2:1 hexanes/ethyl acetate) afforded the product as a colorless solid (0.10 g, 0.30 mmol, 33%).R f = 0.6 (2:1 hexanes/ethyl acetate, anisaldehyde)  21 mmol) was dissolved in 3 mL of methanol and an excess of potassium hydroxide (0.50 g, 9.00 mmol) was added.The mixture was refluxed for 15 h, methanol was distilled off and the residue was dissolved in 3 mL of dichloromethane.Water was added (3 mL), the layers were separated and the organic layer was dried over MgSO 4 .After concentration in vacuo 0.06 g (0.20 mmol, 93%) of desired product was isolated as a yellow oil.In an oven dried Schlenk flask 1.94 g (0.08 mol) magnesium turnings were covered with 10 mL of diethyl ether.A small amount of methyl iodide (0.31 mL, 0.71 g, 5.00 mmol) was added to the mixture and the reaction was started using a heat gun.In the following a mixture of 2.18 mL (4.97 g, 35.0 mol) methyl iodide and 15 mL of Et 2 O was added drop wise and after complete addition the colorless solution was refluxed for further 30 minutes.Under an inert gas atmosphere (S)-N-(ethoxcarbonyl)proline methyl ester 27 (1.00 g, 4.95 mmol) was dissolved in 20 mL Et 2 O and cooled to 0 °C.The obtained Grignard-solution was added slowly over a period of one hour and the resulting solution was stirred for further 10 h at room temperature.The reaction was hydrolyzed using 20 mL sat.NaHCO 3 solution.After the addition of 20 mL of dichloromethane the layers were separated and the aqueous layer was extracted with 3 x 10 mL CH 2 Cl 2 .After drying over MgSO 4 concentration afforded 1.00 g (4.95 mmol, 100%) of the desired product as yellow oil.

37
(S)-N-Ethoxy-α,α-dimethyl-2-pyrrolidinemethanol 36 (0.50 g, 2.48 mmol) was dissolved in 30 mL of methanol and an excess of potassium hydroxide (4.48 g, 0.08 mol) was added.The mixture was refluxed for 15 h, methanol was distilled off and the residue was dissolved in 5 mL of dichloromethane.Water was added (5 mL), the layers were separated and the organic layer was dried over MgSO 4 .After concentration in vacuo 0.10 g (0.77 mmol, 32%) of desired product was isolated as a yellow oil.Spectroscopic data were consistent with previously reported data for this compound. 9

Tri(o-tolyl)boroxin (38)
o-Tolylboronic acid (0.92 g, 6.80 mmol) was dissolved in 20 mL of toluene, heated to 150 °C and was directly distilled off via a claisen bridge.This procedure was repeated three times.The favored product was obtained as a colorless solid (0.65 g, 1.84 mmol, 81%).

Figure S1 1 H
Figure S1 1 H-NMR spectrum of the inseparable mixture of double bond isomers 10, 12

Figure S4 1 H
Figure S4 1 H-NMR spectrum of the obtained monoketal 14