Selective bromochlorination of a homoallylic alcohol for the total synthesis of (−)-anverene

The scope of a recently reported method for the catalytic enantioselective bromochlorination of allylic alcohols is expanded to include a specific homoallylic alcohol. Critical factors for optimization of this reaction are highlighted. The utility of the product bromochloride is demonstrated by the first total synthesis of an antibacterial polyhalogenated monoterpene, (−)-anverene.


General information
All reactions were conducted in oven-or flame-dried glassware under an atmosphere of nitrogen or argon unless otherwise noted. Commercial reagents and solvents were used as received unless otherwise noted with the exception of the following: hexanes (ACS grade, 4.2% various methylpentanes), toluene, tetrahydrofuran, acetonitrile, methanol, benzene, and dichloromethane were dried by passing through a bed of activated alumina in a JC Meyer Solvent System. Flash column chromatography was performed using F60 silica gel (40-63 μm, 230-400 mesh, 60 Å) purchased from Silicycle. Analytical thin-layer chromatography (TLC) was carried out on 250 μm 60-F 254 silica gel plates purchased from EMD Millipore, and visualization was effected by observation of fluorescence-quenching with ultraviolet light and staining with either p-anisaldehyde. Proton nuclear magnetic resonance ( 1 H NMR) and carbon nuclear magnetic resonance ( 13 C NMR) spectra were recorded on Varian Inova 600 or Varian Inova 500 spectrometers operating at 600 and 500 MHz respectively for 1 H, and at 150 and 125 MHz for 13 C. Chemical shifts are reported in parts per million (ppm) with respect to residual protonated solvent for Uncorrected melting point data were collected using a Thomas Hoover Uni-Melt apparatus. Table S1: Scope of bromochlorination for selected substrates: a Low estimate of ee due to moderate peak overlap with dibromide in HPLC Note: an asterisk (*) indicates that the absolute configuration has been assigned by analogy to other substrates. See the following text for details.

S4
Discussion on the scope and mnemonic of the bromochlorination reaction: Despite the clear utility of the bromochlorination reaction, it is not general to all allylic or homoallylic alcohols. The most capricious aspect of the bromochlorination reaction tends to be the regioselectivity, but there are also substrate classes for which no enantioselectivity is seen at all. These limitations are summarized here and in Table S1 above. The bromochlorination of prenol S1a is included for reference (Table S1, entry 1) [1].
Alkyl-trans-1,2-disubstituted allylic alcohols s2a and parent allyl alcohol S3a were bromochlorinated with high enantioselectivity, but with poor regioselectivity, rendering the resultant motifs completely inaccessible in pure form (Table S1, entries 2 and 3). Surprisingly, the major regioisomeric bromochloride arose from 6-endo chloride delivery to the distal olefinic carbon producing S2b and S3b in high enantioselectivity. However, the minor 5-exo regioisomeric bromochlorides S2c and S3c were formed with low enantioselectivity, favoring the opposite pseudoenantiomeric dihalide. This was proven by subjecting the inseparable mix of regioisomeric bromochlorides S2b and S2c to radical debromination conditions. The resulting known alkyl chlorides were chromatographically separable, permitting purification and assignment by optical rotation [2]. These results indicate that both regioisomeric bromochlorides arise from the same intermediate bromonium for this substrate, and the disparity in enantioselectivity among them suggests that most background reactivity is funneled into the 5-exo regioisomeric bromochloride S2c.
Homoprenol 6 (Table S1, entry 4) is not the only homoallylic alcohol that can be bromochlorinated with high enantioselectivity. Cis-or trans-1,2-disubstituted substrates S4a and S5a, as well as terminal olefin S6a were bromochlorinated with apparent high regio-and enantioselectivity (Table S1, entries 5-7). Unfortunately, the chemoselectivity for entries 5-7 was poorer than that seen for homoprenol (5); bromochlorides S4b, S5b and S6b were each contaminated with a significant amount of dibromide. De-dibromination of these dihalides with NaI was unsuccessful; the dibromide proved to be too robust, and no reactivity was seen after 3 hours. Further, low isolated yields were obtained for these bromochlorides due to their volatility. The bromochlorination of these substrates was not optimized any further, but the high regioselectivity seen for S4b and S6b indicates a potentially attractive alternative route to the inaccessible dihalide moieties in S2c and S3c, respectively.
Further examination into the scope of the bromochlorination reaction revealed some interesting trends with regard to the substitution pattern for both allylic and homoallylic alcohols. For example, the allylic alcohol S7 is unique among all other allylic alcohols tried: it was bromochlorinated in under 10% ee. With regards to homoallylic alcohols, substitution at the proximal carbon, as seen in S8 and S9, also resulted in no enantioselectivity. Homocinnamyl alcohol (S10), an electronically biased substrate, was bromochlorinated with high regioselectivity to yield the 7-endo bromochloride corresponding to substrate-controlled Markovnikov regioselectivity, but in only 20% ee, and with very poor chemoselectivity. Bromochlorination was attempted on one bishomoallylic alcohol, S11, but no enantioselectivity was seen for this substrate.
Dry hexanes (400 mL) were added to a dried 500 mL round-bottomed flask containing a 37 mm rod-shaped stirbar under N 2 atmosphere. Solid ClTi(OiPr) 3 (2.63 mL, 11 mmol, 1.1 equiv) was melted at 60 °C and added as a neat, viscous liquid by syringe using a long, thick (20 gauge) metal needle (commercial solutions of ClTi(OiPr) 3 in hexanes deliver identical results). Ti(OiPr) 4 (0.59 mL, 2 mmol, 0.2 equiv) was added as a neat liquid. Homoprenol (5, 1.00 g, 10.0 mmol, 1 equiv) was added as a solution in 3 mL hexanes. With vigorous stirring, (S,R)-4 (430 mg, 1.0 mmol, 0.1 equiv) was added dropwise over 1 minute as a solution in 10 mL hexanes. The pale yellow homogeneous solution was brought to −15 °C in a cryocool. The septum and N 2 balloon were removed, briefly exposing the reaction to ambient air. N-Bromosuccinimide (2.14 g, 12 mmol, 1.2 equiv) was quickly added, and the reaction was capped with a polyethylene yellow stopper. The reaction was stirred at 1500 rpm, maintaining a strong vortex for 17 hours, after which TLC analysis of the pale yellow cloudy suspension indicated reaction completion. The reaction was quenched with 50 mL 1 M aq. Na 2 SO 3 , and allowed to warm to room temperature.
Then, 100 mL ether and 20 mL conc. HCl were added, and the mixture was stirred vigorously for 15 minutes. The now clear layers were separated, the aqueous layer was washed 2 times with ether, and the combined organic layers were washed with 50 mL saturated aq. NaHCO 3 . The combined organic layers were dried over Na 2 SO 4 and solvent was removed by rotary evaporation yielding the crude bromochloride 6 as an orange oil, 2.51 g. 1 H NMR analysis indicated an 8.2:1.0:2.6 mixture of desired 6:constitutional isomer 7:dibromide 8. HPLC analysis indicated the desired product 6 was formed in 89% ee.
To selectively de-dibrominate the undesired dibromide sideproduct, the crude oil was dissolved in 10 mL acetone in a scintillation vial, and NaI (600 mg, 4 mmol, roughly 2 equiv relative to dibromide) was added, yielding a fully homogeneous orange solution. The vial was capped tightly and brought to 65 °C in a sand bath with vigorous stirring for 3 hours. The dark brown opaque mixture was cooled to room temperature. TLC analysis indicated a mixture of 6 and 5. The reaction was quenched with 10 mL 1 M aq. Na 2 SO 3 , and partitioned between 50 mL ether and 50 mL water. The layers were separated and the organic layer was washed with brine. The organic layer was dried over MgSO 4 , filtered, and the solvent was removed by rotary evaporation. Most residual homoprenol was removed by rotary evaporation at room temperature and at a vacuum of 2-5 torr for 30 minutes.
The dark brown crude oil was purified by column chromatography (silica gel, 25% ethyl acetate/hexanes) to yield the product 6 (1.38 g, 6.4 mmol, 64% yield) as an 8.2:1 mixture of 6:7 without dibromide contamination. HPLC analysis indicated the desired product remained at 89% ee.  according to the following Dess-Martin oxidation procedure on 3 mmol scale. A 100 mL round-bottomed flask was charged with substrate alcohol 6 (643.7 mg, 2.986 mmol, 1 equiv), NaHCO 3 (1.26 g, 5 equiv), and DCM (30 mL, not rigorously dried). The pale yellow homogeneous solution was cooled to 0 °C, and Dess-Martin periodinane (1.53 g, 1.2 equiv) was added all at once. The reaction was allowed to warm slowly to rt. After 90 minutes, TLC analysis indicated full conversion. The reaction was quenched by adding 10 mL saturated aq. Na 2 S 2 O 3 and 10 mL saturated aq. NaHCO 3 , and was stirred vigorously for 15 minutes. The reaction was partitioned between 30 mL DCM and 30 mL water, and the aqueous layer was extracted with DCM 3 times. The combined organic layers were dried over Na 2 SO 4 , and the solvent was removed by rotary evaporation at room temperature at a vacuum no stronger than 200 torr. Residual DCM was removed by azeotroping with pentane, again at a vacuum no stronger than 200 torr. This yielded the crude product aldehyde 3 as a yellow oil (605.7 mg, 2.837 mmol, 95% yield). 1 H NMR analysis indicated >97% purity by weight with respect to residual solvent. The crude aldehyde 3 was taken onto the next step without further purification.
Physical properties: yellow oil. Somewhat volatile; b.p. not measured.   were removed, briefly exposing the reaction to ambient air. N-Bromosuccinimide (300 mg, 1.7 mmol, 1.2 equiv) was quickly added, and the reaction was capped with a polyethylene yellow stopper. The reaction was stirred at 1500 rpm, maintaining a strong vortex for 42 hours. TLC analysis at this point indicated incomplete conversion.
The reaction was quenched with 10 mL 1 M aq. Na 2 SO 3 and allowed to warm to room temperature. Then, 10 mL ether and 10 mL conc HCl were added, and the mixture was stirred vigorously for 15 minutes. The now clear layers were separated, the aqueous layer was washed 2 times with ether, and the combined organic layers were washed with 20 mL saturated aqueous NaHCO 3 . The combined organic layers were dried over Na 2 SO 4 and solvent was removed by rotary evaporation yielding the crude tetrahalide 11 as a grainy yellow oil. 1

S12
(−)-(2R,3S,5S)-3,6-Dibromo-2,5-dichloro-2,6-dimethylheptanal (S13): The aldehyde S13 was made from the corresponding alcohol 11 according to the following Dess-Martin oxidation procedure on 1 mmol scale. A 100 mL round-bottomed flask was charged with substrate alcohol 11 (395.4 mg, 1.066 mmol, 1 equiv), NaHCO 3 (450 mg, 5.3 mmol, 5 equiv) and DCM (10 mL, not rigorously dried). The pale yellow homogeneous solution was cooled to 0 °C, and Dess-Martin periodinane (551 mg, 1.3 mmol, 1.2 equiv) was added all at once. The reaction was allowed to warm slowly to room temperature. After 90 minutes, TLC analysis indicated full conversion. The reaction was quenched by adding 5 mL saturated aq. Na 2 S 2 O 3 and 5 mL saturated aq. NaHCO 3 , and was stirred vigorously for 15 minutes. The reaction was partitioned between 30 mL DCM and 30 mL water, and the aqueous layer was extracted with DCM 3 times. The combined organic layers were dried over Na 2 SO 4 , and the solvent was removed by rotary evaporation. The crude pale yellow oil S13 (392.1 mg, 1.063 mmol, 99% yield) was taken onto the next step without further purification.