Diastereoselective synthesis of trisubstituted ole ﬁ ns using a silicon-tether ring-closing metathesis strategy

The diastereoselective synthesis of trisubstituted ole ﬁ ns with concomitant C – C bond formation is still a di ﬃ cult challenge, and ole ﬁ n metathesis reactions for the formation of such alkenes are usually not high yielding or/and diastereoselective. Herein we report an e ﬃ cient and diastereoselective synthesis of trisubstituted ole ﬁ ns ﬂ anked by an allylic alcohol, by a silicon-tether ring-closing metathesis strategy. Both E - and Z -trisubstituted alkenes were synthesised, depending on the method employed to cleave the silicon tether. Furthermore, this methodology features a novel Peterson ole ﬁ nation for the synthesis of allyldimethylsilanes. These versatile intermediates were also converted into the corresponding allylchlorodimethylsilanes, which are not easily accessible in high yields by other methods.


Introduction
The trisubstituted E olefin motif with a methyl substituent is present in numerous polyketide natural products. Among these, callipeltoside A 1 and dolabelide C 2 possess another common feature: an allylic alkoxy group on the lone substituent of the trisusbstituted olefin (highlighted in red in Scheme 1). During our previous approaches to the synthesis of dolabelide C, 3 we had envisaged appending the side chain by a cross-metathesis (CM) reaction between an allylic alcohol and a gem-disubstituted olefin to form the C24-C25 trisubstituted olefin. Unfortunately, all attempts to perform CM between diverse allylic alcohols ( protected or not) and the required gem-disubstituted olefin for dolabelide C led to dimerisation of the allylic alcohol partner. 4 Silicon-tether ring-closing metathesis (RCM) reactions with O-Si-O linkages have been widely used for the formation of alkenes, when the corresponding CM reactions are not high yielding or/and diastereoselective. 5 Previous works by Kobayashi et al. 6 and Evans et al. 7 illustrate the use of such linkage for the formation of E and Z trisubstituted olefins, respectively, via eight-membered rings (Scheme 2).
However, there are much fewer examples of RCM reactions with O-Si-C tethers. The reported formations of trisubstituted olefins from allyl silanes lead to bicyclic products, 8 and the others focus on vinyl silanes. 9 We propose a strategy for an efficient and diastereoselective synthesis of trisubstituted olefins flanked by an allylic alcohol, which involves a RCM reaction with a O-Si-C tether. In our preliminary communication, 10 the metathesis precursors C were prepared by dehydrogenative coupling of allylic alcohols A and allyldimethylsilanes B, and the metathesis products D were converted into the E olefin (Scheme 2). Herein, we report our preliminary attempts to synthesise compounds C and a synthesis of versatile allylchorodimethylsilanes from silanes B. Furthermore, the Z olefin isomer can also be accessed by choosing the appropriate method to cleave the tether (Scheme 2).

Results and discussion
The preparation of the RCM substrates C we first envisioned is shown in Scheme 3: it would entail silyl ether formation between allylic alcohols A and allylchlorodimethylsilanes B′.
The first issue to address was the preparation of compounds B′. Allylchlorodimethylsilanes can be prepared by reaction of the corresponding allyl Grignard reagents with dichlorodimethylsilane, but the yields are usually low. For example, allylchlorodimethylsilane was synthesised in 44% yield 11 using this method, and the methallyl derivative in 39% yield. 12 Organoindium derivatives have been reported to favour monoaddition of allyl nucleophiles, thus leading to better yields of allylchlorosilanes. 13 As a preliminary trial, allylbromide, when treated with indium and dichlorodimethylsilane in 1,3dimethyl-2-imidazolidinone (DMI), led to allylchlorodimethylsilane, which was not isolated but directly reacted with alcohol 1 in the presence of triethylamine and DMAP to furnish the desired silyl ether 2 in 56% yield (Scheme 4). RCM reaction of compound 2 with Grubbs 2 catalyst proceeded smoothly 14 to furnish cyclic silyl ether 3 in 74% yield, thus validating our strategy.
We then prepared several allylic bromides with a gem-disubstituted olefin, which are models or racemic versions of the side chain of dolabelide C, by a known method. 15 Esters 4a-c 10 were subjected to trimethylsilylmethylmagnesium chloride in the presence of anhydrous cerium trichloride, and the resulting β-hydroxysilanes were stirred with silica gel in dichloromethane to furnish trimethylsilanes 5a-c in 61-91% yield (Scheme 5). Bromination of the Peterson olefination products 5a-c was performed with pyrrolidone hydrotribromide 6 in THF to give the allylic bromides 7a-c in 88-91% yield.
The allylchlorodimethylsilane formed in situ from 7a and dichlorodimethylsilane in the presence of indium, was treated with alcohol 1 to furnish silyl ether 8a in 58% yield (Scheme 6). Unfortunately, this reaction was not reproducible, and variable yields were obtained after further attempts. Moreover, compounds 8b and 8d could not be prepared by this method. These poor results might be due to the additional steric hindrance present in 7b and 7d 16 (R = OTES and OPMB, respectively), which renders the indium derivatives less reactive. 17 When magnesium was employed instead of indium with compound 7b, silyl ether 8b was produced in only 4% yield, along with several side-products resulting from the cleavage of the triethylsilyl ether. The same reaction was tried on the PMB derivative 7d; however, in this case, only the Würtz coupling product was obtained. This Barbier-type reaction was also attempted with zinc, 16,18 but the desired silyl ethers 8a,b, d were never observed.
Inspired by a report from Maruoka and coworkers who described the preparation of bromosilanes from the corresponding silanes, 19 we converted allylic bromides 7a-c into silanes 9a-c in 81-86% yield, by reaction of the Grignard reagents derived from 7a-c with chlorodimethylsilane (Scheme 7). It was found that slow addition of a solution of the bromide compound to a solution of 3.5 equivalents of magnesium and 1 equivalent of chlorodimethylsilane was necessary to reduce the formation of the Würtz coupling products. Maruoka et al. used bromine for the conversion of the silanes to the bromosilanes, but these conditions are not compatible with the alkene present in 9a-c. Several chlorination conditions were employed -CuCl 2 , catalytic CuI; 20 FeCl 3 , acetyl chloride 21without success. Finally, the use of hexachloroethane and catalytic palladium dichloride furnished the chlorosilane intermediates, which were treated with allylic alcohol 1 in the presence of triethylamine and DMAP to give the desired chlorosilanes 8a-c in 47-77% yield. 22 At this point, we were satisfied to have found a reliable method for the synthesis of RCM precursors 8a-c via the chlorosilanes, but the overall sequence was too long to be practical, so we sought to shorten the route in two different ways. Silyl ethers 8a-c could be prepared directly from silanes 9a-c by a copper-catalysed dehydrogenative coupling, 23 bypassing one step and the need to deal with sensitive chlorosilanes. Furthermore, we envisioned a novel Peterson olefination for the synthesis of allyldimethylsilanes 9a-c from esters 4a-c, which would proceed in an analogous fashion as the for-mation of allyltrimethylsilanes 5a-c (Scheme 5). As previously reported in our communication, 10 both steps proceeded uneventfully (Scheme 8). It is worth noting that this new olefination reaction provides an easy access to allyldimethylsilanes, but also to the corresponding chlorosilanes (see Scheme 7).
RCM reactions of silyl ethers 8a-c in the presence of the Grubbs 2 catalyst in refluxing dichloromethane gave the cyclic products 10a-c in 63-94% yield. These cyclic silyl ethers are unstable on silica gel, and the yield of 10a could be improved from 63% (ref. 10) to 83% simply by performing the purification by a quick filtration through a silica plug instead of a column chromatography. As previously reported, direct cleavage was attempted on 10b with diverse fluoride sources and Lewis acids, but without success, leading to isomerisation product 13b′ or elimination products 14b and 14b′ (see Table 1 for structures of these compounds). 10 We thus resorted to a two-step process, starting with cleavage of the silicon-oxygen bond with methyllithium to give allyltrimethylsilanes 11a-c (Scheme 9).
Removal of the trimethylsilyl group was then optimised on compound 11b (Table 1). Employing Lewis acids such as BF 3 ·OEt 2 or Sc(OTf ) 3 (entries 1 and 2) only led to elimination product 14b′, with the TES group still in place. When 11b was treated with TBAF under mild conditions (entry 3), the starting compound was recovered in full. Product 12b, resulting from direct protodesilylation of the TMS group was obtained under more forcing conditions (entry 4), along with isomerisation product 13b and elimination product 14b; in all those compounds, the TES group had been removed. In order to reduce the amount of elimination product and favor the reprotonation, water was added to the reaction mixture (entry 5); in this case, the desired olefin 12b was produced in 35% yield, along with 35% of the isomerised olefin 13b. Although these conditions were not ideal, they were applied to 11a and 11c, for which the TMS removal would not be disrupted by the TES group cleavage. The desired products were obtained in 68% and 64% yields, respectively, along with 28% of the corresponding isomerisation product in each case (Scheme 9).

Organic & Biomolecular Chemistry Paper
Trimethylsilyl group removal was further optimised on compound 11a by varying the fluoride source (Scheme 10). When using solid TBAF hydrate, the reaction was complete after 30 min at low temperature, and the amount of isomerised product 13a was reduced. With TBAF(t-BuOH) 4 , the amount of isomerised product was further decreased, but the yield of the desired product 12a was only moderate.
As previously reported in our communication, 10 this new synthetic route was applied to the synthesis of the C16-C30 fragment of dolabelide C.
We then reasoned that this methodology could also be employed for the construction of Z-trisubstituted olefins, if the silyl substituent was used to extend the chain (Scheme 11). Alcohol 1 was reacted with known methallyldimethyl-chlorosilane 12 in the presence of imidazole and DMAP to furnish silyl ether 15 in quantitative yield. 24 Ring-closing metathesis of 15 with the Grubbs 2 catalyst gave cyclic silyl ether 16, which was subjected without purification to Tamao-Fleming oxidation, and diol 17 was obtained in excellent yield for the two steps. In this manner, a terminal alkene (in red in 1) can be converted into a Z-trisubstituted olefin (in red in 17) in three high yielding steps, and the primary alcohol is a convenient handle for subsequent elongation of the chain.

Conclusions
In conclusion, we have developed a new methodology for the construction of E-trisubstituted olefins flanked by an allylic alcohol, featuring a silicon-tether ring-closing metathesis reaction. In the course of synthesising the metathesis substrates, we designed a novel Peterson olefination reaction that enabled direct access to allyldimethylsilanes. These allysilanes are versatile intermediates that can be converted into the corresponding chlorosilanes, which are otherwise difficult to obtain in good yields. This methodology was applied to the synthesis of the C16-C30 fragment of dolabelide C, showcasing its practical aspect. Furthermore, Z-trisubstituted olefins could also be prepared when Tamao-Fleming oxidation was used to cleave the silicon tether.   1 H NMR spectra were recorded with CDCl 3 as the solvent using residual CHCl 3 (δ = 7.26) as internal standard, and for 13 C NMR spectra the chemical shifts are reported relative to the central resonance of CDCl 3 (δ = 77.16). Signals in NMR spectra are described as singlet (s), doublet (d), triplet (t), quartet (q), quintet (quint), septet (sept), multiplet (m), broad (br) or combination of these, which refers to the spinspin coupling pattern observed. Spin-spin coupling constants reported are uncorrected. Two-dimensional (COSY, HSQC, HMBC, NOESY) NMR spectroscopy was used where appropriate to assist the assignment of signals in the 1 H and 13 C NMR spectra. IR spectra were obtained employing a Shimadzu FTIR-8400 instrument with a Golden Gate™ attachment that uses a type IIa diamond as a single reflection element so that the IR spectrum of the compound (solid or liquid) could be detected directly (thin layer). High resolution mass spectra were recorded under ESI and CI conditions by the analytical services at the University of Glasgow. Flash chromatography was executed under forced flow conditions, using the indicated solvent system and EMD Guduran silica gel 60 as solid support. Thin layer chromatography (TLC) was carried out on Merck silica gel 60 covered aluminum sheets, and monitored by UV-light or by staining with a solution of anisaldehyde or potassium permanganate.
Allyl bromide (0.70 g, 5.8 mmol, 1.8 equiv.) was added to a stirred solution of indium (0.44 g, 3.8 mmol, 1.2 equiv.) in DMI (7.5 mL) at room temperature under argon. The mixture was stirred for 2 h at room temperature under argon. Dichlorodimethylsilane (0.48 g, 3.8 mmol, 1.2 equiv.) was added to the mixture and the reaction was stirred for 3 h at 70°C under argon. The mixture was allowed to cool down and was added dropwise to a stirred solution of 1 (0.50 g, 3.1 mmol), triethylamine (0.78 g, 7.7 mmol, 2.5 equiv.) and DMAP (0.19 g, 1.5 mmol, 0.50 equiv.) in THF (10 mL) at 0°C under argon. The reaction mixture was stirred overnight at room temperature under argon. The reaction mixture was then quenched with a saturated aqueous solution of sodium hydrogen carbonate (10 mL) and extracted with ether (3 × 10 mL

Trimethyl(2-methyleneheptyl)silane 5a
Cerium(III) chloride heptahydrate (27.8 g, 74.5 mmol, 3.50 equiv.) was added to a 500 mL three-necked round-bottomed flask and dried under vacuum at 120°C for 2 h, 140°C for 2 h and then at 160°C for 2 h. The flask was allowed to cool down and was purged with argon for 10 min, and dry THF (200 mL) was added. The mixture was stirred for 20 h at room temperature under argon to give the cerium(III) chloride-THF complex as a white precipitate. A trimethylsilylmethylmagnesium chloride solution (1.30 M in THF, 57.0 mL, 74.5 mmol, 3.50 equiv.) was added dropwise at −78°C under argon. The gray suspension was stirred for 30 min, and then ethyl caproate (3.07 g, 21.3 mmol) in dry THF (15 mL) was added at −78°C. The reaction was stirred at −78°C for 2 h, the flask was then removed from the cold bath, and the mixture was stirred at room temperature for 15 h. The reaction mixture was cooled to 0°C, a saturated aqueous solution of ammonium chloride (200 mL) was added at 0°C, and the mixture was stirred for 20 min. Water was added, and the mixture was extracted with ether (3 × 150 mL). The combined ether extracts were washed with water and brine (2 × 300 mL), then dried (Na 2 SO 4 ), filtered and concentrated in vacuo to give a yellow oil. The oil was then stirred overnight with silica in dichloromethane. Filtration, followed by concentration in vacuo gave the allylic silane 5a  13  Triethyl(2-((trimethylsilyl)methyl)hept-1-en-4-yloxy)silane 5b The crude silane 5b was obtained from the corresponding ester 4b (4.94 g, 18.0 mmol) according to the procedure described above for silane 5a. The crude oil was purified by column chromatography on silica gel ( petroleum ether-ether 98/2) and the resulting oil was stirred overnight with silica in dichloromethane. The mixture was then filtered and concentrated in vacuo to give the allylic silane 5b (4.81 g, 85%) as a clear oil. 1

(4-(Benzyloxy)-2-methyleneheptyl)trimethylsilane 5c
Cerium(III) chloride heptahydrate (21.3 g, 57.1 mmol, 3.50 equiv.) was dried under vacuum at 120°C for 2 h, 140°C for 2 h and then at 160°C for 2 h. The flask was allowed to cool down to room temperature and was purged with argon for 10 min, and dry THF (190 mL) was added. The mixture was stirred for 20 h at room temperature under argon to give the cerium(III) chloride-THF complex as a white precipitate. To a stirred solution of magnesium (1.4 g, 57 mmol, 3.5 equiv.) and 1,2-dibromoethane (3 drops) in THF (48 mL) at 65°C was added slowly a solution of chloromethyltrimethylsilane (8.0 mL, 58 mmol, 3.5 equiv.) in THF (12 mL). The solution was then stirred for 2 h at room temperature. The above solution was then added dropwise (over 1 h) at −78°C to the cerium. The gray suspension was stirred for 30 min, and then ester 4c (4.09 g, 16.3 mmol) in dry THF (20 mL) was added at −78°C. The reaction was stirred at −78°C for 2 h, the flask was then removed from the cold bath, and the mixture was stirred at room temperature for 15 h. A saturated aqueous solution of ammonium chloride (200 mL) was added at 0°C, and the mixture was stirred for 20 min. Water was added, and the mixture was extracted with diethyl ether (3 × 200 mL). The combined ether extracts were washed with water and brine (2 × 300 mL), then dried (Na 2 SO 4 ), filtered and concentrated in vacuo to give a yellow oil. The oil was stirred one day with silica gel in dichloromethane. The mixture was then filtered and concentrated in vacuo and the resulting crude oil was purified by column chromatography on silica gel ( petroleum etherether 98/2) to give the allylic silane 5c (2.90 g, 61%) as a clear oil. 1

2-(Bromomethyl)hept-1-ene 7a 25
Pyrrolidone hydrotribromide (4.1 g, 7.8 mmol, 1.0 equiv.) was added to a stirred solution of the allylic silane 5a (1.44 g, 7.82 mmol) and pyridine (4.1 mL, 51 mmol, 6.5 equiv.) in dry THF (322 mL) at −10°C under argon. The mixture was stirred for 2 h and allowed to warm to room temperature during this period. A saturated aqueous solution of sodium thiosulfate (250 mL) was added and the resulting mixture was stirred for 5 min. The mixture was extracted with ether (3 × 200 mL). The ether extracts were combined and washed with a saturated aqueous solution of copper sulfate (2 × 300 mL) and brine (2 × 300 mL), then dried (Na 2 SO 4 ), filtered and concentrated in vacuo to deliver a yellow oil. Purification by column chromatography on silica gel ( petroleum ether-ether 99/1) gave the bromide 7a (1.36 g, 91%) as a clear oil. Obtained data in accordance with previously reported data. 25

Paper
Organic & Biomolecular Chemistry