Formalizing the mechanism of the allylic substitution reaction (S N ’): application to the highly enantio-and diastereo-selective syntheses of 1-phenyl-2-vinylcyclopentanes

discovered


Introduction
Several years ago we described a high yielding synthesis of cyclopentane derivatives 4a bearing a benzylic quaternary center and substituted on the adjacent carbon by a methyl group, from benzylselenide 1a bearing a CC double bond four carbons away from the benzylic carbon. 1,2It involved the intermediate formation of the related benzyllithium, its exo-trig addition 3 across the built-in CC double bond, and protonation of the resulting cyclopentylmethyllithiums 3a (Scheme 1, entries a,b).
Interestingly, we found that this carbocyclization reaction occurs with high stereocontrol that depends upon the solvent and the temperature used. 1 It leads to the stereoisomer rac-4a' in which the phenyl group and methyl group on the adjacent carbons are cis when the reaction is carried out in pentane or diethyl etherhexanes between -20 and 20 °C and to rac-4a" in which the same substituents are trans when the reaction is instead carried out in THF-hexanes at -78 °C. 1 We later observed similar stereochemical features from benzylselenides 1b possessing either a Z-or Eallylic ether moiety (Scheme 1, entries c,d), 2,4 but were unable to determine whether the process that involves the intermediate formation of benzyllithiums (Scheme 1, entries c,d) occurs through a stepwise (Scheme 2, entry a) or a concerted allylic substitution reaction (Scheme 2, entry b).

Scheme 1 Scheme 2 Scheme 3
Since this reaction involves an intramolecular process, It was expected to favor the allylic substitution reaction (S N ') at the expense of the competing substitution reaction (S N ), that is beneficial for the interpretation of the experiments and for the synthetic value of the process.
The S N ' reaction has been the subject of very much experimental and theoretical work over the last 60 years, [5][6][7][8][9][10] that attests the significance attached to the phenomenon. 8This reaction can deliver up to four stereoisomeric products possessing either a Z-CC double bond (Figure 1, S Z , A Z ) or E-CC double bond (Figure 1, S E , A E ) from a single stereoisomer of the allylic electrophile, with a net preference for the formation of the latter products possessing E-CC double bonds.Those products formally result from the attack, on each of the two remarkable conformers of the starting material, of a nucleophile entering from the same face of the CC double bond than the departing group (syn-mode; Figure 1, S Z , S E ) or from its opposite face (anti-mode; Figure 1, A Z , A E ).

Figure 1
Surprisingly, although a great many experiments have been carried out on a wide variety of starting materials [5][6][7][8][9] involving charged and uncharged nucleophiles, different families of leaving groups and experimental conditions that include metal-catalyzed processes, 11,12 over a long period, only rare systematic studies have been carried out and up to now it has been impossible to predict with confidence the stereochemical outcome of any S N ' reaction.
The transcript 13 of a part of an interview, still of highly topical interest, with the late Prof. Gilbert Stork, who provided seminal contributions [5][6][7][8][9]11,[14][15][16][17][18] to the mechanism of the SN' reaction, sheds a proper light on this process: "That was good and bad. It turne out to be sort of the wrong kind of reaction to get involved with.It was intriguing at the time.It turned out to be (i) enormously more complicated than anyone knows; even today no one understands it, and (ii) not important.It became a known piece of work because there were not that many qualitative mechanistic studies at that time".
We selected as starting materials unsaturated benzylselenides 1c bearing a benzyloxy group at their allylic carbon atoms possessing all the fixed 8(S) stereochemistry and either Z-(1c Z ) or E-(1c E ) Δ [6][7] CC double bonds.We decided for convenience, to carry out the reactions on mixtures of the two epimeric selenides possessing the [2(R) and 2(S)] configuration at their benzylic carbons, expecting that it would circumvent inherent synthetic difficulties and will not interfere with the cyclization reaction due to the well-known ease by which benzyllithium intermediates interconvert. 19,20e expected that the process would initiate a series of asymmetric inductions from the allylic (S)-C-8 carbon that would allow to control the stereochemistry at each newly created asymmetric centers depending upon the nature of the solvent (Ether or THF) and therefore to produce from each of the two couples [2(R) and 2(S)] of stereoisomers [1c Z or 1c E ] a major product different from the others.
The strategy developed for the synthesis of each of the two epimeric mixtures of starting materials 1c E and 1c Z disclosed in Figure 2 is commented upon here.Their synthesis along with the related experimental part is presented below.The strategy involves the selection of: 1. commercially available 21 scalemic (S)-3-butyn-2-ol (5) able to deliver to 1c a four carbon unit (C-6 to C-9) with an hydroxyl group on a (S)-(C-8) carbon atom and carrying a terminal CC triple bond possessing the aptitude to be (i) easily metallated and alkylated at its sp carbon, after protection of its hydroxyl group and (ii) stereoselectively reduced to a disubstituted CC double bond (Δ 6-7 CC double bond) possessing either the (Z) or the (E) stereochemistry, using respectively either the Lindlar catalyst 22 that finally leads to 1c Z (Figure 2, entry a), or Red-Al 23 that takes advantage of its hydroxyl directed hydroalumination, leading finally to 1c E (Figure 2, entry b), 2. 2,2-bis(methylseleno)ethylbenzene, readily available 2,24 from acetophenone and methylselenol, is the precursor of the 1-lithio-1-ethyl-1-methylseleno-benzene 2,25 expected to bring the two-carbon unit of the chain (C-1 and C2) and the benzylic carbon (C-2) flanked with the methylseleno moiety, potential precursor of the corresponding benzyllithium 2c (Figure 2).Both benzyllithiums are prone to be alkylated by alkyl halides (by substitution) or CC double bonds (by addition), 2 3. complementary compounds bearing three carbon straight chain (C-3,C-5) that possess different leaving groups at each of their termini to allow selective sequential alkylations with the functionalized acetylide and then with the 1-lithio-1-ethyl-1-(methylseleno)benzene.

Experimental results
We have systematically carried out the reactions between epimeric mixtures of benzyl selenides 1c Z and 1c E and butyllithiums at -78 °C using diethyl ether-hexanes (Scheme 4, entries a,c) or THF-hexanes (Scheme 4, entries b,d) as solvents.As previously observed, the cleavage of their CSe bond is efficiently achieved by n-BuLi in THF-hexanes, whereas it requires the more reactive t-BuLi if ether-hexanes is instead used (Scheme 4, entries a,c).Each of those reactions led in reasonably high yields (65-88 %) to a stereoisomeric mixture of 1-phenyl-2propenylcyclopentane derivatives 4c in which the one possessing the E-CC double bond largely prevails (100-93 %, Scheme 4).
(i) The stereochemistry of the major products (4c E1 -4c E4 , Scheme 4) was found to be dependent both the nature of the solvent used and on the stereochemistry of the starting materials 1c.In accordance with our previous results (Schemes 1), those (4c E1 and 4c E3 , Scheme 4, entries a,c) resulting from the reactions carried out in ether-hexanes possess a cis-relationship between the phenyl and the propenyl groups and are produced through the syn-E-mode (Scheme 4, entry a) whereas those (4c E2 and 4c E4 ) generated in THFhexane possess a trans-relationship between the same groups and their formation instead involve the anti-E-mode (Scheme 4, entries b,d).
Each pairs of products 4c E1 and 4c E3 (Scheme 4, compare entry a with c) and 4c E2 and 4c E4 (Scheme 4, compare entry b with d), generated from different starting materials but in the same solvents, are enantiomers.Whereas each pair of products 4c E1 and 4c E2 (Scheme 4, compare entry a with b) and 4c E3 and 4c E4 (Scheme 4, compare entry c with d) generated from the same starting material but in different solvents are diastereoisomers with cyclopentane rings on which the carbons bearing the phenyl ring possess the same stereochemistry, and consequently the ones to which is attached the propenyl side chain bears an inverted stereochemistry.(ii) The stereochemistry of the minor products (4c E3 , 4c Z1 , and 4c E1 , Scheme 4, entries b-d) that are formed in less that 7% besides almost all the major products (4c E2 -4c E4 , Scheme 4, entries b-d, except Scheme 4, entry a) also depends on the stereochemistry of the starting material and the solvent.They all nevertheless exhibit a cis-relationship on the cyclopentane ring between the phenyl and propenyl groups and possess all a benzylic carbon that is epimeric to that of the related major stereoisomer.The minor stereoisomer 4c Z1 (Scheme 4, entry c) is the only product that possesses a Z-CC double bond.Interestingly its formation as the one of the major stereoisomer 4c E3 involves the syn-mode (although it is the syn-Zmode instead of the syn-E-mode; Scheme 4, entry c).

Scheme 4
We have also observed in the tandem Se/Li exchange-carbocyclisation reactions carried at 0 °C instead of -78 °C that the amount of the minor isomer always increases at the expanse of the major product (Schemes 5).This is particularly the case of reactions performed in THF-hexanes (Schemes 5, entries b,d).It still affects the reaction of 1c Z1 that delivers in ether-hexanes compound 4c Z1 , possessing a Z-propenyl side chain in quite high yield (30 %, Scheme 5B, entry c) but does not affect the outcome of the reaction involving its E-stereoisomer 1c E1 , performed in the same mixture of solvents (Scheme 5A, entry a).

Interpretation of the results
We show in Schemes 5, for each product formed, even the minor ones, the conformation of the related "transition state" and the "mode" (syn-mode or anti-mode) implied in each of their formations.It leads us, by including also the results shown in Scheme 4, to propose the following observations to rationalize the stereochemistry of the products obtained from those reactions.
(i) Reactions carried out in ether-hexane, involve the syn-mode, suggesting that a compact transition state is favored, in which the lithium cation is tightly linked to the benzylic carbon and coordinated by a lone pair of the alkoxy group, as well as to the π bond of the CC double bond of the allyl ether, 2,4 and those of the aromatic ring 26,27 (Schemes 4, 5, entries a,c).(ii) Reactions carried out in THF-hexanes involve the anti-mode, suggesting that the intramolecular interactions discussed above no longer exist due to the selective complexation of the "lithium cation" by the lone pairs of the oxygen atoms of the more basic THF.This favors an "extended conformation" in which the complexed benzyllithium could initiate the S N ' reaction via a back-side attack, avoiding as much as possible the unfavorable steric interactions (Scheme 4,5, entries b,d).(iii) The formation in high yields of the major stereoisomers (4c

Structure determinations
We have not been able to separate effectively the major stereoisomers of the cyclized products 4c in all experiments involving 1c Z and 1c E shown in Schemes 4 and 5 (entries a-d) and therefore we have not been able to determine directly their ratios and consequently their structures.We have nevertheless been able to do so by combining different techniques, taking into account: (1) that each of the four stereoisomeric (2-methyl-2-phenylcyclopentyl)methanols 11 (Scheme 7, Table 1, entry b) readily accessible, in a single pot process by sequential ozonolysis of the crude mixtures of 4c followed by in situ reduction of the resulting ozonides with sodium borohydride, 28 has been easily separated by HPLC using a "chiral column" allowing the determination of their relative ratio in each experiment, 29 and (2) that each of the related crystalline camphenoates 13 (Scheme 7, Table 1, entries c), readily prepared by reaction of commercially available (-)camphanic acid chloride 12 with compounds 11, 30 has been isolated by column chromatography on SiO 2 and its structure unambiguously determined by X-ray crystallography 31 (Scheme 7, Table 1, entry d).
Finally, the stereochemistry of the CC double bond of 4c has been assessed 32 by 1 H NMR spectroscopy of the crude mixtures of each experiment, taking into account the chemical shifts and value of the coupling constant of their hydrogens linked to the two adjacent vinylic carbons.

Scheme 7
Table 1. 31 Stereochemistry of compounds 4c E , 11 and 13 depicted in Scheme 7 a

Synthetic significance of the results
We have reported above the synthesis of each of the four stereoisomers of the cyclopentane derivatives 4c E bearing a phenyl-substituted quaternary carbon next to a tertiary carbon bearing a E-propenyl side chain, by cyclization that produces a new bond between those two carbon atoms with unusually high stereocontrol.
Those are versatile precursors of: (a) The whole series of scalemic 1-phenyl-1-methyl-2-propenyl-cyclopentanes 4c Z possessing instead Zpropenyl moiety that cannot be obtained by carbocyclization of 1c (Scheme 4).It would involve their sequential ozonolysis to the corresponding aldehydes 14 using ozone/dimethyl sulfide 33 followed by Zstereoselective Wittig reaction using the Schlosser conditions involving ethylidenetriphenylphosphorane in DMSO 34 (Scheme 8, entry a).(b) 1-methyl-1-phenyl-2-vinyl-cyclopentanes 4b, previously available as racemates from 2-phenyl-2selenomethyl-7-octene 1b (Scheme 1, entries c,d), that can be generated by a similar method as reported in the previous paragraph (Scheme 8, entry b) but instead involving methylene triphenylphosphorane. 35 Scheme 8 (c) Scalemic 1,2-dimethyl-1-phenyl-cyclopentanes 4a available as racemates from 2-phenyl-2-selenomethyl-7-octene 1b (Scheme 1, entries a,b) 1,2 that can be readily synthesized starting from the related cyclopentylmethanols 11 as outlined in Scheme 9 by reductive ozonolysis 28 of 4c E followed by sulfanylation of their hydroxyl group and reduction of the resulting sulfonates by lithium triethylborohydride (Scheme 9). 36This set of reactions has been carried out at an early stage of our research on a racemic mixture of 4a E1 +4a E3 as well as on a racemic mixture of 4a E2 +4a E4 obtained from rac-1c Z in ether-hexanes and THF-hexanes to determine their relative stereochemistry.

Scheme 9
Finally, each of the enantiomers of 1-methyl-1-methyl-2-vinyl cyclopentanes 4c E whose structures are disclosed in Scheme 4 can be produced on reaction of butyllithiums either in ether-hexanes or THF-hexanes from the different pairs of stereoisomeric benzyl selenides 1c whose structures are shown in Scheme 10, and possessing the following characteristics:  the same (R)-stereochemistry at the allylic 8-positions and either a Z-or a E-CC double bond (Scheme 10, upper left),  the same (S)-stereochemistry at the 8-allylic positions and a double bond possessing either a Z or a Estereochemistry (Scheme 10, upper right),  the same (Z)-stereochemistry of their CC double bonds and either an (R)-or (S)-stereochemistry at the 8allylic position (Scheme 10, lower left)  the same (E)-stereochemistry of the CC double bonds and either an (R)-or (S)-stereochemistry at the 8allylic position (Scheme 10, lower right).

Contextualization of the results
Although the allylic substitution reaction (S N ') has been the subject of extensive work 2,4,[6][7][8][9][10][11][12][13][14][15][16][17][18] since the seminal discoveries of Winstein 37 and Stork, 14,15 it still lacks proper models to predict with confidence the outcome of any reaction belonging to that field or to suggest conditions that could allow the synthesis of any specific stereoisomer of a given substance through an S N ' reaction. 13The intramolecular version the S CN ', 18 to which this work belongs, offers the advantage to avoid competing direct substitution reactions (S N ) that are usually observed.It leads to cyclic compounds, including alkenyl substituted five-membered heterocycles (Scheme 11) 11,17 and carbocycles (Scheme 12, 13) 6,18,[38][39][40] whose stereochemistry at the carbon on the cycle bearing the alkenyl group as well as of CC double bond offer precious indications about the mechanism of the reaction.
We first provide a brief historical background to the S N ' reaction that will allow inclusion of our work into a wider perspective.
Winstein 37 and Stork 14,15 very early recognized that the S N ' reactions could take place stereoselectively with the incoming nucleophile and the departing group lying on the same side (syn-mode) or the opposite side (anti-mode) (Figure 1).
Stork described the first syn-S N ' reaction 14,15 and twenty-four years later the first anti-S N ' reaction. 16Since the original paper from Stork, there has been considerable discussion as to whether concerted S N ' reactions ever occur 9 and this concept has been even described in early times as "unreasonable" or "abhorrent". 9,17,41][7][8][9][10] Those assessments proved to be incorrect, after the experimental results reported later by Stork. 16,17ost of the reactions so far described did indeed involve the syn-mode [5][6][7][8][9][10] and generate compounds bearing usually E-CC double bonds, [5][6][7][8][9][10] unless it is part of a medium ring.Although products possessing the Zstereochemistry have been from time to time described, often as side products, 6,12,16,17,38 they usually proceed through the syn-mode.
It has been reported that (i) "Experience suggests that soft nucleophiles give syn-stereochemistry and hard nucleophiles anti", 8,9,17 (ii) "Theory suggests 9 that the syn-mode is involved for neutral nucleophiles while anionic nucleophiles approach from the anti-direction" 8,9 and (iii) "Evidence is meager and contradictory….and …. small variations produce strikingly variable results". 8,9ur experimental results clearly contradict those statements.We agree with the view of Overton 42 who wrote "It becomes apparent that, contrary to the long-held view that S N ' reactions proceed with synstereochemistry, the whole spectrum spanned by the syn-and anti-extremes is to be expected depending, in any particular case, on the nature of the displacing and displaced groups, counter ions, and solvent"; we propose to add "temperature".In fact, the difficulties encountered in rationalizing the results published are due to the large number of parameters that play a crucial role in the process and the widespread differences between the examples that have to be compared.6][7][8][9] They all involve as starting materials unsaturated straight-chain organometallics that produce five-membered rings 6,11,17,18,[38][39][40] via intramolecular allylic substitution reactions leading to the departure of a benzoate (Scheme 11, 11 Scheme 12, 17 Scheme 14 18 ) or an alkoxide located on the allylic site (Scheme 13, 38 Scheme 15, 39 Scheme 16 6 ).As general trends, the reactions reported in Schemes 11-16 produce compounds in which the fivemembered ring 6,11,17,18,38,39 is substituted by a E-CC double bond and only rarely by a Z-CC double bond (Scheme 12, 7 Scheme 13, 38 Scheme 16, entry b 6 ).

ARKAT USA, Inc
Most of the cyclization reactions take place through the syn-mode except those disclosed in Scheme 11, 17 Scheme 12 7 and Scheme 14 18 that involve instead the anti-mode.
One striking difference between our work and that already published is the narrow window on which we have made systematic variations (stereoisomers, solvents of same kind, temperatures) as compared to the widespread data on which correlations have been made previously (very different types of starting material, especially nucleophilic entities and leaving groups, have been studied, using a wide variety of solvents and temperatures).
As general trends, S N ' reactions that involve a metal cation [5][6][7][8] or eventually a proton, 43 favor, as in our case, highly structured transition states involving chelation by atoms bearing lone pairs and π-bonds leading to the syn-mode.However this organization can be prevented when the reactions are performed in polar solvents, at high temperature or in cases of unfavorable steric interactions [5][6][7][8] favoring thus the anti-mode.
We did not find experimental proofs confirming the assessment of Stille 38 that the reactions disclosed in Scheme 13 proceed through an anti-mode, and we rationalize the anti-mode involved in the reactions implying metal thiolates 17 (19a, Scheme 12) or a metal malonate 18 (24, Scheme 14) by poorer chelation of (i) the soft thiolate 17 to the hard counter-cation (19, Scheme 12) and (ii) the delocalized enolate in case of the sodio-malonate 18 (24, Scheme 14) that disfavor the chelated preorganization leading to the syn-mode.
The case of α-alkoxyalkenyl lithiums 30a E and 30b E (Scheme 16) 6 attracted our attention since it shares some similarity with that of α-phenyl alkenyllithiums 2 E we have disclosed in Schemes 4 and 5 (entry c).They all bear a CC double bonds possessing the E-stereochemistry and a quaternary carbanionic center to which are attached groups (alkoxy and phenyl respectively) able to coordinate the lithium cation.They both deliver, through a syn-mode, cyclopentane derivatives in which those groups are cis to the pending CC double bonds (31a E , 31b z , 4c E3 , 4c Z1 ), and last but not least whereas one of the organolithium epimers delivers cyclopentane derivatives possessing E-CC double bonds through the syn-E-mode (31a E , Scheme 16, entry a; 4c E3 , Scheme 6, entry a) the other unusually produce mainly its stereoisomers possessing Z-CC double bonds (31b z , Scheme 16, entry b; 4c Z1 , Scheme 6, entry b) through the syn-Z-mode.
There are however striking differences, since (i) α-alkoxyalkenyl lithiums 30a E and 30b E are expected to be configurationally stable, 6,44 whereas benzyl lithiums such 2c E7 and 2c E3 have been found to interconvert already at -78 °C. 19,20This did not prove to be a problem because the latter have been found to adapt to the experimental conditions; (ii) although correlations about the outcome of the two types of reaction disclosed above fit very well (compare Schemes 4,5; entries c with Scheme 16, syn-mode in each case), they have been carried out in different solvents (THF for α-alkoxyalkenyl lithiums 30 E and ether for α-phenylalkenyl lithiums 2 E ) that have been found, at least for α-phenylalkenyl lithiums (compare Schemes 4,5; entries c with entries d), to lead to very different stereochemical outcome: syn-mode in ether, anti-mode in THF!We assume therefore that THF does not affect the ability of the alkoxy-group attached to the carbanionic center of 30 E to complex the "lithium cation" that leads to the compact transition state required for the syn-mode, whereas it destroys the weaker complexation of the same cation by the electron cloud of the phenyl ring 1,2,4,26,27 in α-phenylalkenyl lithiums 2 E that is only observed when the less basic ether is instead used.

Synthesis of the starting materials
The multistep syntheses of two isomeric Z-and E-[(8S)-8-(benzyloxy)-2-phenylnon-6-en-2-yl](methyl)selane 1c Z and 1c E , reported in Scheme 17 and Scheme 18 respectively, follows the retrosynthetic analysis shown in Scheme 3.Each of them was carried out from the commercially available (S)-3-but-3-yn-2-ol 5 21 and involve in each case protection of its hydroxyl group that allow the stepwise alkylation of their terminal acetylenic carbon and at the last stage the introduction of a benzylseleno moiety through the corresponding αselenobenzyl lithium 32. 1,2,25he synthesis of the Z-stereoisomer 1c Z involves the shortest of the two routes (Scheme 17) that uses the benzyl protected propargylic alcohol 33 its metalation by n-butyllithium in THF-hexanes at -78 °C and subsequent alkylation of the resulting acetylide by 1-bromo-3-chloropropane that takes place selectively on the carbon bearing the bromine atom, finally leading to 6 in 75 % yield.Catalytic dihydrogenation of 6 generates 7 Z in 85 % yield possessing a Z-CC double bond was achieved using Lindlar catalyst 22 in the presence of quinoleine to avoid over reduction and the formation of the target compound 1c Z has been achieved in 90 % yield by reacting the chloride 7 Z with 1-phenyl-1-methylseleno-ethyllithium 32. 1,2,25

Scheme 17
The synthesis of the E-stereoisomer 1c E (Scheme 18) is lengthier due to the exchange of the original tertbutyl dimethylsilyl protecting group that is required to allow the metalation/alkylation process leading to 35 but needs to be removed to offer one hand to carry the stepwise introduction of the two hydrogens in a transrelationship on the CC double bond using RedAl 23 followed by the hydrolysis of the resulting aluminum alcoholate leading to 9 E .An orthogonal deprotection/protection was required to avoid inadequate deprotection of the THP group and to promote the requested benzylation of the allyl alcohol moiety leading to 36 E .Selective deprotection of the THP group leaving untouched the benzyloxy ether was achieved by acid catalyzed methanolysis leading to the alcohol 37 E that on reaction with carbon tetrabromide/triphenylphosphine reagent 46 leads to the Eunsaturated bromide 10 E pendant of Z-unsaturated chloride 7 E whose reaction with 1-phenyl-1-(methylseleno)ethyl lithium 32 1,2,25 provides stereoselectively 1c E .

Experimental Section
General.(i) Thin layer chromatography (TLC) was conducted on pre-coated aluminum sheets with 0.20 mm Machevery-Nagel Alugram SIL G/UV254 with fluorescent indicator UV254.Column chromatography was carried out using Merck Gerduran silica gel 60 (particle size 63-200 m) (ii) Melting points (M.p.) were measured on a Büchi Melting Point B-545 in open capillary tubes and have not been corrected.(iii) Nuclear magnetic resonance (NMR) 1 H, 13 C and 19 F spectra were obtained on a 400 MHz NMR (Jeol JNM EX-400) Chemical shifts were reported in ppm according to tetramethylsilane using the solvent residual signal as an internal reference (CDCl 3 :  H = 7.26 ppm,  C = 77.16ppm; DMSO-d 6 :  H = 2.50 ppm,  C = 39.52 ppm).Coupling constants (J) were given in Hz (J 1 : ortho, J 2 : meta, J 3 : para).Resonance multiplicity was described as s (singlet), d (doublet), t (triplet), dd (doublet of doublets), dt (doublet of triplets), q (quartet), m (multiplet) and br (broad signal).Carbon spectra were acquired with a complete decoupling for the proton.(iv) Infrared spectra (IR) were recorded on a Perkin-Elmer Spectrum II FT-IR System single-reflection ATR mounted with a diamond mono-crystal (v) Chiral high-performance liquid chromatography (HPLC) analysis were performed through a chiral analytical column CHIRALCEL OJ-H (Daicel Chemical Industries, Ltd.) (0.25 m × 4.6 mm) coated with tris-(4-methylbenzoate) cellulose on 5 m silica-gel substrate).Column type: Eluent: n-Hexane/i-Propanol 99/1; Flow rate: 2 ml/min, Injection: 10 µl of a 5 mg/ml solution, Detection: UV (220 nm), Peaks at: 14.4 min (11 1 ), 21.2 min (11 2 ), 24.5 min (11 3 ), 33.3 min (11 4 ) using a Merck-Hitachi 655A equipment using a UV detector (vi) Xray analyses have been carried out by the "Laboratoire de Chimie Moléculaire Structurale", UNamur using NOMIUS CAD-4 diffractometer and the K α ray of copper (: 1.54178 nm).Product's structures have been resolved with the program SIR92 and refined with the program SHELXL97.CCDC quotation refers to the crystal structures related to 13 1 -13 4 have been deposited at the Cambridge Crystallographic Data Centre and allocated the deposition numbers: 13 1 (CCDC 923101), 13 2 (CCDC 923100), 13 3 (CCDC 923103), 13 4 (CCDC 923102) (vii) Chemicals were purchased from Sigma Aldrich, Acros Organics, TCI and ABCR and were used as received.Solvents were purchased from Sigma Aldrich, while deuterated solvents from Eurisotop.Diethyl ether and THF were distilled from sodium-benzophenone-cetyl, toluene was refluxed over calcium hydride and dichloromethane (CH 2 Cl 2 ) was refluxed over phosphorus pentoxide.Anhydrous DMF was purchased from Acros Organics.Hydrochloric acid (HCl 32%) was purchased from Fischer Scientific.MeOH and CHCl 3 were purchased as reagent-grade and used without further purification (viii) Low temperature baths were prepared using different solvent mixtures depending on the desired temperature: -78°C with acetone/dry ice, -40 °C with CH 3 CN/liquid N 2 , -10 °C with ice-H 2 O/NaCl, and 0 °C with ice/H 2 O. Anhydrous conditions were achieved by drying 2-neck flasks by flaming with a heat gun under vacuum and then purging with argon.The inert atmosphere was maintained using argon-filled balloons equipped with a syringe and needle that was used to penetrate the silicon stoppers used to close the flasks' necks.Additions of liquid reagents were performed using dried plastic or glass syringes.

E1, 4c E2, 4c E3 and 4c E4 )
in reactions carried out at low temperature reported in Scheme 4, implies that each pair of epimeric unsaturated benzyllithiums

4c E1, 4c E2, 4c E3 and 4c E4 listed
in Scheme 4 and the remainder cyclizes to the minor stereoisomers 4c E3 , 4c Z1 , and 4c E1 .The routes shown in Scheme 6 exemplify the role of the temperature on the equilibrium involved for example when the epimeric mixture of intermediates

2c ES and 2c ER are generated from 1c Z (1c ZS +1c ZR )
and s-butyllithium in ether-hexane at -78 °C and 0 °C, delivering various amounts of