Stereospecific synthesis of allylic and homoallylic alcohols from functionalized propargylic alcohols

1-Substituted 4,4,5,5-tetraethoxy-2-pentyn-1-ols undergo stereospecific reduction to allylic and homoallylic alcohols under the right conditions. Hydrogenation over Lindlar’s catalyst gave the corresponding ( Z )-allylic alcohols in excellent yield provided potassium carbonate was added . Reduction was also achieved with lithium aluminum hydride, but the product appeared to be solvent and temperature dependent . In THF at -15 o C the corresponding ( E )-allylic alcohols were formed, in better than 70% yield from secondary propargylic alcohols, but below 60% from tertiary ones; in refluxing diethyl ether the products were the corresponding 1-substituted derivatives of homoallylic alcohol ( E )-4,5,5-triethoxypent-3-en-1-ol, obtained in 93% yield in the best case.


Introduction
Recently we reported a high-yield synthesis of 3,3,4,4-tetraethoxybut-1-yne (1), denoted TEB. 1-3 In spite of the compound's polar and bulky 1,1,2,2-tetraethoxyethyl group (TEE) the corresponding propargylic alcohols (2) are obtained when TEB acetylide is generated with base in the usual way and subsequently treated with an aldehyde or a ketone.Generally the alcohols are formed in good to excellent yields, but there are, as expected, 4-6 some exceptions (see experimental).
Alcohols 2 are somewhat special due to the TEE moiety which is sterically demanding, has a large number of oxygen atoms that can engage in complex formation, and contains ethoxy groups that can act as a leaving group under the right reaction conditions.Consequently, it is quite conceivable that conversion of 2 to the corresponding (Z)-and (E)-allylic alcohols, by standard hydrogenation over Lindlar's catalyst and by treatment with lithium aluminum hydride, respectively, can be hampered by various effects caused by TEE.As reported here, that appeared indeed to be the case.
Ethyl acetate; K 2 CO 3 ; rt H 2 ; Lindlar's catalyst On the basis of these observations the way the hydrogen was added to the suspension was studied.When the gas was introduced into the flask through a glass tube just above the stirred suspension (method 1), days were required to achieve complete conversion even of the most reactive alcohols.However, when hydrogen was bubbled directly into the suspension just above the stirring bar (method 2), the reaction was much faster and even hindered propargylic alcohols gave the corresponding allylic alcohols.
Based on these findings eight propargylic alcohols were hydrogenated using the two methods outlined above (for details, see Experimental).The results, compiled in Table 1, show that all the alcohols were converted to the corresponding (Z)-allylic alcohol in almost quantitative yield within a fairly short time (8 hours) when method 2 was applied.However, when hydrogenation was performed according to the less efficient method 1, reactivity differences became apparent.Thus, whereas 5,5,6,6-tetraethoxyhex-3-yn-2-ol (2b) was reduced quantitatively in 24 hours, the tertiary analogues (2g -2i) reacted more reluctantly and required up to three days of hydrogenation to achieve a good yield of the propargylic alcohol (Table 1).Another interesting observation was that the only primary propargylic alcohol investigated, 4,4,5,5-tetraethoxypent-2-yn-1-ol (2a), deviated from this trend and turned out to require about the same reaction time as the secondary analogues.
The Z configuration of the allylic alcohols was substantiated by inspection of the vicinal coupling constant ( 3 J HH ) across the C=C bond in their 1 H-NMR spectra.In every case this coupling constant was 14.1 Hz or smaller for the allylic alcohols obtained by hydrogenation of 2 over Lindlar's catalyst, and at the same time more than 1.8 Hz smaller than the corresponding 3 J in the proton spectra of the corresponding (E)-allylic alcohols obtained by LAH reduction of 2 (vide infra).Based on the general and well documented observation that 3 J trans > 3 J cis for 3 J HH across C=C bonds 14 it was concluded that the allylic alcohols afforded by hydrogenation over Lindlar's catalyst exhibit Z stereochemistry.The observation that 2 was not hydrogenated in the absence of potassium carbonate indicates that the reactivity is strongly influenced by the hydroxyl group's ability to engage in hydrogen bonding.In order to test this hypothesis 4,4,5,5-tetraethoxypent-2-yn-1-ol (2a) was converted to the corresponding acetate, 4,4,5,5-tetraethoxypent-2-ynyl acetate (4), which was exposed to hydrogen in the presence of Lindlar's catalyst, but in the absence of potassium carbonate (Scheme 3), i.e. conditions under which 2a does not react.To our satisfaction ester 4 was consumed relatively quickly (complete consumption after 15 h) and afforded (Z)-4,4,5,5tetraethoxypent-2-enyl acetate ((Z)-5) in 95% yield, and this clearly indicates that our line of reasoning is correct.This conclusion was further supported by hydrogenation of a 1:1 mixture of 2a and 4 over Lindlar's catalyst in the presence of potassium carbonate.When the progress of the reaction was monitored by thin-layer chromatography, the acetate (4) appeared to react significantly faster than the alcohol, and furthermore, when 4 had been completely consumed and the reaction was quenched and worked up, the corresponding alkenes (Z)-3a and (Z)-5 were isolated in a 1:5 ratio (see Experimental).

Scheme 3
Primary alcohol 2a is significantly less sterically crowded around the OH group then the secondary and tertiary analogues (2b -2f and 2g -2j, respectively), and the hydrogen-bonding effects should therefore be significantly more pronounced for 2a than for the other alcohols.If the hydrogen-bonding effects contribute more to retard the hydrogenation than moderate steric effects, the rate of hydrogenation for 2a should be similar to or smaller than for the secondary analogues.The results for compounds 2b -2f in Table 1 are in accordance with this line of thinking, but to test the hypothesis, 5,5-diethoxy-4-hydroxypent-2-ynyl acetate (7) and the corresponding "double propargylic" diol, 5,5-diethoxypent-2-yne-1,4-diol (9), were prepared as outlined in Scheme 4 and hydrogenated in the presence of potassium carbonate (method 2; see Experimental).Both 7 and 9 gave the corresponding (Z)-allylic alcohol, (Z)-5,5-diethoxy-4hydroxypent-2-enyl acetate ((Z)-10) and (Z)-5,5-diethoxypent-2-ene-1,4-diol ((Z)-11), respectively, under these conditions, but the reaction appeared to be significantly slower for 9 than for 7.In order to compare the relative reactivity more accurately a 1:1 mixture of 7 and 9 was hydrogenated over Lindlar's catalyst in accordance with method 2. The progress of the reaction was monitored by thin-layer chromatography, and when the more reactive compound, 7, had been completely consumed and the reaction was quenched and worked up, alkenes (Z)-10 and (Z)-11 were isolated in a 2:1 ratio (see Experimental) (Scheme 4).
The influence of the sterically demanding TEE group on the hydrogenation is difficult to assess from the results presented above, but an impression is obtained by comparing the rate of hydrogenation for acetates 4 and 6.Both compounds give the corresponding Z alkene, (Z)-5 and (Z)-5,5-diethoxy-4-oxopent-2-enyl acetate ((Z)-12) (Scheme 5), respectively, in high yield when hydrogenated in the absence of potassium carbonate, but the reaction appeared to be significantly slower for 4 than for 6.In order to compare the relative reactivity more accurately a 1:1 mixture of 4 and 6 was hydrogenated over Lindlar's catalyst in accordance with method 1.The progress was monitored by thin-layer chromatography, and when the more reactive compound, 6, had been completely consumed and the reaction was quenched and worked up, alkenes (Z)-5 and (Z)-12 were isolated in a 1:8 ratio (see Experimental).Thus, the ketal moiety seems to slow down the reactivity of the triple bond a lot, although some of the difference is probably due to electronic influence of the carbonyl group, which is conjugated to the triple bond in 6.

Scheme 5 (E)-Allylic alcohols
In order to convert 2 to the corresponding (E)-allylic alcohols 3 a standard procedure with literature precedence, viz.treatment with lithium aluminum hydride (LAH) in diethyl ether at room temperature, 15 was applied first.When 4,4,5,5-tetraethoxypent-2-yn-1-ol (1a) was reacted following this procedure, (E)-4,4,5,5-tetraethoxypent-2-en-1-ol ((E)-3a) was obtained, but in addition a significant amount of a by-product, homoallylic alcohol (E)-4,5,5-triethoxypent-3-en-1-ol (13a), was formed (Scheme 6).Studies of the literature revealed that formation of (E)-allylic alcohol has been favoured by lowering the reaction temperature, 16,17 and sure enough, when 2a was reacted with LAH at -15 o C, the yield of (E)-3a increased whereas the by-product formation dropped considerably.Even better results were obtained when the reaction at low temperature was repeated with THF as solvent; 16 under these conditions 2a gave (E)-3a only in almost quantitative yield (Table 3).Five other propargylic alcohols were reduced with LAH under the same conditions and in all cases the corresponding (E)-allylic alcohol was the only major product obtained.The secondary alcohols ((E)-3b, (E)-3e, and (E)-3f) were obtained in good to excellent yields, whereas the tertiary analogs ((E)-3g and (E)-3h) were obtained in 54% only ( As pointed out previously the E configuration of the alcohols could in all cases be substantiated by the 1 H-NMR spectroscopy and inspection of the vicinal coupling constant ( 3 J HH ) across the C=C bond because this 3 J HH was both large (> 15.6 Hz) and at the same time significantly larger than the corresponding coupling constant for the Z analogues (in the case of 3f the difference was as much as 4.1 Hz; see Table 2). 14 The concurrent formation of allylic alcohols (E)-3 and homoallylic alcohols (E)-13 can be accounted for by the mechanism outlined in Scheme 7, using propargylic alcohol 2a as an example.Hydride attack of the triple bond to give carbanion 14 has solid literature support. 18 This intermediate can undergo two transformations, either ring closure to form metallocycle 15, as described by Borden and others, 18 or expulsion of one of the ethoxy groups next to the anion to give allene 16 in accordance with reports by Cowie, Claesson and co-workers. 17,19Based on literature precedence 20 16 will undergo hydride transfer and ultimately form metallocycle 17 under the prevailing reaction conditions.Neither 15 nor 17 are stable under aqueous conditions; 18,20 both are therefore cleaved upon hydrolysis and form (E)-3a and (E)-13a, respectively.

Scheme 7
Despite numerous reports of allene formation from propargylic alcohols with a leaving group attached to C-4, 21 none of our experiments gave any indications of formation of such byproducts.NMR and IR spectroscopic techniques were used extensively to detect signals due to the allene moiety, 22 but all attempts were unsuccessful.

Homoallylic alcohols
As reported above LAH reduction of 4,4,5,5-tetraethoxypent-2-yn-1-ol (2a) in diethyl ether at room temperature gave two products, mainly (E)-4,4,5,5-tetraethoxypent-2-en-1-ol ((E)-3a), but also some of homoallylic alcohol (E)-4,5,5-triethoxypent-3-en-1-ol ((E)-13a).The amount of the latter product decreased when the reaction temperature was lowered, and this, combined with the fact that (E)-13a formation is envisaged to go via allene intermediate 16 (Scheme 7) which reacts more effectively at elevated temperature, 20 triggered us to believe that homoallylic-alcohol formation could be rendered more favourable by running the LAH reduction in refluxing ether or at higher temperature in a different solvent.Small-scale experiments were performed with 2a to test this prediction, and to our satisfaction 13a turned out to be the only major product when the reduction was carried out in refluxing diethyl ether (bp 34 o C).Five propargylic alcohols were then reacted on a larger scale under the same conditions.All the compounds reacted similar to 2a; the reaction was over in less than 20 minutes, and as seen from Table 3, the corresponding homoallylic alcohols were obtained in good to excellent yield after isolation by flash chromatography.
A noteworthy feature of the transformation is its stereospecificity, which in all cases led to formation of one isomer only, the E stereoisomer.The E configuration was substantiated by NOESY experiments, which gave spectra with significant cross peaks between the methine proton in the diethoxymethyl moiety and the protons in the methylene group on the other side of the C=C bond, but no cross peaks between the same methine proton and the olefinic proton.A few words regarding the workup of the reaction mixtures are in place since the products contain acid-labile groups, i.e. an enol-ether moiety as well as an acetal function.If hydrolysis was carried out with an aqueous acid instead of water so that the hydrolysate remains acidic at the end, partial decomposition of the product takes place.Several products were detected, including dihydropyran 18 (Scheme 8), the formation of which is expected on the basis of the work of Newman. 9Similar reactions took place when 13 was purified by column chromatography with certain types of silica gel as the stationary phase (conceivably due to variable density and nature of acidic sites).These transformations are currently under investigation.

Scheme 8
Experimental Section General Procedures.IR spectra were recorded on a Nicolet Impact 410 infrared spectrophotometer.NMR spectra were run on a Bruker Spectrospin AC 200 F or a Bruker Spectrospin DMX 400.Chemical shifts are reported downfield from TMS and coupling constants are given in Hz.GC analyses were performed on a HP 5890 Gas Chromatograph with a flame ionization detector and a HP Ultra 1 column (100% dimethyl-polysiloxane, 25 m, 0.2 mm i.d., 0.33 µm).Flash chromatography was performed with Silica gel (230-400 mesh) as the stationary phase and mixtures of hexane and ethyl acetate as the mobile phase.Thin-layer charomatographic (TLC) analyses of the reaction mixtures were carried out with Silica gel (60 F 254 ) on aluminium sheets with mixtures of hexane and ethyl acetate as the mobile phase.Mass spectra were obtained on a VG 7070 Micromass spectrometer and an Autospec Ultima instrument, a three-sector instrument with EBE geometry from Micromass, both operated in the EI mode at 70 eV.
Chemicals.3,3,4,4-Tetraethoxybut-1-yne (TEB) was synthesized from ethyl vinyl ether as described in the literature. 2 Ethylmagnesium bromide was prepared according to the literature, and the reagent's concentration was determined by titration following literature procedure. 23All other chemicals and reagents, the syntheses of which are not described or referred to in this section, were commercially available from Aldrich and Fluka (now SigmaAldrich) and were, with the exception of some solvents, used without further purification.Ethanol, hexane, ethyl acetate, tetrahydrofuran (THF), and diethyl ether were dried and purified as described in the literature. 24  Synthesis of propargylic alcohols.General procedures Method A. The reactions carried out with NaNH 2 generated in situ were performed following the procedure described by Furniss et al. 25 TEB (1.0 g, 4.3 mmol) was reacted with sodium amide prepared from Na (0.20 g, 8.7 mmol in all cases except one, the preparation of 2f) and NH 3 (50 mL), the resulting acetylide was treated with 5.2 mmol of aldehyde or ketone, and the reaction was quenched by adding NH 4 Cl (0.9 g, 0.17 mmol).The product was worked up by flash chromatography as specified for each alcohol.Method B. The reactions performed with commercially available butyllithium as base were carried out following a procedure published by Brandsma. 26The amounts of the reactants and butyllithium are specified for each alcohol.Method C. The reactions with ethylmagnesium bromide as base were performed as follows: An ether solution of ethylmagnesium bromide (ca 1.2 equiv.) was added dropwise to a stirred solution of 3,3,4,4-tetraethoxybut-1-yne (TEB) (1.0 equiv.) in THF.When the addition was complete, the reaction mixture was stirred at reflux for 80 min and cooled to rt before aldehyde or ketone (ca 1.2 equiv.) was added dropwise.The resulting mixture was stirred at 35 o C for 3 h before it was cooled to rt and quenched by adding a saturated aqueous solution of NH 4 Cl (approximately 0.5 mL/mmol TEB).The mixture was extracted with dichloromethane (3 x 10 mL), and the combined organic phases were dried (MgSO 4 ), filtered, and concentrated in vacuo on a rotary evaporator.Finally, the product was isolated by flash chromatography.
Reduction of a 1:1 mixture of 4 and 6.Ketal 4 (0.151 g, 0.5 mmol) and ketone 6 (0.114 g, 0.5 mmol) were dissolved in ethyl acetate (2 mL) together with Lindlar catalysts (0.005 g) and K 2 CO 3 (0.0003 g).H 2 -gass was bubbled through the reaction mixture for 30 min.The reaction was monitored by TLC and stopped when all of the more reactive alkyne, conjugated ynone 6, had been consumed.The Lindlar catalysts and K 2 CO 3 were then filtered from the reaction mixture and ethyl acetate was evaporated on a rotary evaporator.NMR analysis of the crude mixture showed that it contained the hydrogenated products (Z)-5 and 12 in a 1:8 ratio.The product mixture was subjected to flash chromatography (90:10; hexane-ethyl acetate) and pure 12 (0.103 g, 91%) as a clear liquid and (Z)-5 (0.020 g, 12%) as a clear liquid were obtained.

Reduction of propargylic alcohols (2) to homoallylic alcohols (13). General procedure
The propargylic alcohol was added dropwise to a mixture of LAH (3.0 equiv.) in dry, refluxing (35 o C) diethyl ether.The reaction mixture was left stirring at reflux until the reaction was complete (20 min).Most of the solvent was evaporated on a rotary evaporator and the residue was mixed with water (30 mL), and then extracted with dichloromethane (3 x 30 mL).The combined organic phases were dried (MgSO 4 ), filtered, and concentrated in vacuo on a rotary evaporator.The crude product was purified by flash chromatography (80:20, hexane-ethyl acetate).TLC analyses indicated that the products were pure, but the 1 H-and 13 C-NMR spectra showed that the compounds contained one or several minor contaminants.

Table 2 )
, conceivably reflecting the steric congestion around the hydroxyl group.

Table 2 .
Conversion of propargylic alcohols 2 to the corresponding (E)-allylic alcohols ((E)-3) by treatment with LAH in THF at -15 o C

Table 3 .
Conversion of propargylic alcohols 2 to the corresponding homoallylic alcohols 13 by LAH in refluxing diethyl ether