Mild olefin formation via bio-inspired vitamin B12 photocatalysis

Dehydrohalogenation, or elimination of hydrogen-halide equivalents, remains one of the simplest methods for the installation of the biologically-important olefin functionality. However, this transformation often requires harsh, strongly-basic conditions, rare noble metals, or both, limiting its applicability in the synthesis of complex molecules. Nature has pursued a complementary approach in the novel vitamin B12-dependent photoreceptor CarH, where photolysis of a cobalt–carbon bond leads to selective olefin formation under mild, physiologically-relevant conditions. Herein we report a light-driven B12-based catalytic system that leverages this reactivity to convert alkyl electrophiles to olefins under incredibly mild conditions using only earth abundant elements. Further, this process exhibits a high level of regioselectivity, producing terminal olefins in moderate to excellent yield and exceptional selectivity. Finally, we are able to access a hitherto-unknown transformation, remote elimination, using two cobalt catalysts in tandem to produce subterminal olefins with excellent regioselectivity. Together, we show vitamin B12 to be a powerful platform for developing mild olefin-forming reactions.


General Experimental Section
The commercially available reagents were purchased at highest commercial quality and directly used without further purification. All the air or moisture-sensitive reactions were set-up using oven-dried glassware under an inert atmosphere. The reactions were monitored by thin-layer chromatography (TLC) carried out on Merck silica gel pre-coated (60 F254) glass plates (0.25mm). Purification of reaction products were carried out by flash chromatography using silica gel 60 (230-400 mesh). The NMR yields of the compounds were calculated using 1,3,5trimethoxybenzene (proton signal at δ 6.08 ppm) as an internal standard. 1 H and 13 C NMR spectra were recorded on Bruker 600 (600 MHz). The spectra are referenced relative to residual CD3CN or CDCl3 proton signals at δ 1.94 ppm and δ 7.26 ppm, respectively. The chemical shifts are reported in part per million (ppm) from high to low frequency and referred to the residual solvent resonance peak. Coupling constant (J) are reported in Hz. The multiplicity of 1 H signals are indicated as: s = singlet, d = doublet, t = triplet, p = pentet, m = multiplet. IR data were recorded with Bruker Alfa Platinum ATR single reflector spectrometer by applying the compounds as a thin film directly on the ATR unit and the data are presented as most characteristic absorption frequencies in cm -1 . High resolution mass spectra (HRMS) were recorded in Agilent UHPLC TOF mass spectrometer using electrospray ionization (ESI-positive) method.

3a-Bromocholest-5-ene(7)
3a-Bromocholest-5-ene was prepared following the procedure reported in the literature. (8) A mixture of cholesterol (193 mg, 0.50 mmol) and PPh3 (197 mg, 0.75 mmol) was dissolved in dichloromethane (10 mL) at ambient temperature. Then CBr3 (182 mg, mmol) was added to the reaction mixture in portion and stirred overnight at room temperature. The solvent was removed by vacuo and the residue was purified via column chromatography (silica gel; hexane) as white solid (0.185 g, 82%

2, 3-Dihydro-1H-inden-2-yl 4-methylbenzenesulfonate(9)
The title compound was prepared following the procedure reported in the literature. CH2Cl2 (15 mL), and cooled to 0 °C. Finally, tosyl chloride (1.05 g, 5.50 mmol) was slowly added to the cold reaction mixture and was allowed to warm to room temperature. After 17 h, saturated aqueous solution of NaHCO3 (10 mL) was added and the aqueous layer was extracted with CH2Cl2 (1 x 20 mL). The combined organic phase was dried over MgSO4, filtered, and concentrated under reduced pressure. The resulting crude residue was subjected to flash chromatography (silica gel, hexane/ EtOAc = 10:1, v/v) The title compound was obtained as a colorless solid in (1.12 g, 78%

Optimization of VB12-Photocatalyzed Olefination of Alkyl halide
In order to optimize the reaction condition of VB12-photocatalytic dehydrohalogenation reaction, 1-chlorooctane was used as the model substrate. All the control reactions (Table S1, entries 1-14) were set-up at 0.1 mmol scale and in 0.1 molar concentration (except Table S1, entry 14) at room temperature. Terminal alkene product was obtained in all the control reactions, the isomerized (semiterminal) product was notably absent in all the control reactions where the yields are specified. The catalytic reaction did not proceed forward in the absence of light, reductant and base (Table S1, entries 2-5). No reaction was observed when reductants such as phenylsilane, B2Pin2, zinc (Zn) and manganese (Mn) were used. Further, no reaction occurred when a weak inorganic base NaHCO3 was substituted by equal amount of triethylamine (Table S1, entry 6). Interestingly, when the amount of NaBH4 was lowered from 4.0 equivalents to 1.5 equivalent, the efficiency of the reaction was significantly reduced, only producing 56% of the desired product (Table S1, entry 7). Various cobalt pre-catalysts were investigated, it was found that AdoCbl was unsuccessful to catalyze the reaction (Table S1, entries 8), whereas cobaloxime pyridine chloride (COPC) produced some desired product 3a in 34% yield (Table S1, entrie 9). Among the various solvents screened, acetonitrile was found to the best that gave desired terminal alkene selectively. Low selectivity was observed upon using dimethylforamide (DMF), with the yield of desired terminal alkene 3a decreasing to 35% and remainder of the mass balance corresponding to side product 3b (Table S1 entry 10). Similarly, the catalytic action of our method was greatly reduced by the polar solvents dimethyl sulfoxide (DMSO) and acetone, favoring the formation of reduced side product 3b, with olefinic product 3a being formed in 17% and 21% respectively (Table S1, entries [11][12]. Furthermore, rate of the reaction was slower when the catalyst loading of C2 was reduced from 5 Ph OMs mol% to 2.5 mol% (Table S1, entry 13) and with the decrease in concentration of the reaction (Table S1, entry 14). Finally, we investigated the effect of some common mild bases under our optimized reaction condition, and upon using Na2CO3, only 36% of desired product was formed with only 50% conversion (Table S1, entry 15). Whereas the mild base K2CO3 was able to drive the reaction to completion producing low yield of the desire terminal alkene (Table S1, entry 16).  VB12 C2 (6.80 mg, 0.005 mmol, 5.0 mol%), NaBH4 (15.1 mg, 0.399 mmol, 4.0 equiv.), NaHCO3 (12.6 mg, 0.149 mmol, 1.2 equiv.), internal standard 1,3,5-trimethoxybenzene (2.0 -5.0 mg) and a stir bar were added to an oven-dried 8-mL glass vial. A rubber cap was fitted to the vial, which was then evacuated and backfilled with nitrogen (3 cycles). The vial was detached from the nitrogen line, and 1mL of acetonitrile was added via syringe. Then the nitrogen gas was bubbled through the reaction mixture for about 10 min and alkyl halide substrate (0.10 mmol) was added and nitrogen was bubbled again for 5 min. The reaction mixture was placed under blue light after sealing the punctured holes of the vial cap with vacuum grease and electric tape. The reaction mixture was stirred for 16 h, and then about 0.1ml of the reaction mixture was filtered through silica (using short pipet column) to remove solid residue. The silica was washed with CDCl3 (0.1 mL). The filtrate was then washed with DI water and the lower CDCl3 layer was taken for 1 H NMR spectroscopy after drying with MgSO4.

General Method of Remote Elimination
VB12 C2 (6.80 mg, 0.005 mmol, 5.0 mol%), cobalt salen complex C4 (3.00 mg, 0.005 mmol, 5.0 mol%) NaBH4 (15.1 mg, 0.399 mmol, 4.0 equiv.), NaHCO3 (12.6 mg, 0.149 mmol, 1.2 equiv.), internal standard 1,3,5-trimethoxybenzene (2.0 -5.0 mg) and a stir bar were added to an ovendried 8-mL glass vial. A rubber cap was fitted to the vial, which was then evacuated and backfilled with nitrogen (3 cycles). The vial was detached from the nitrogen line, and 1mL of acetonitrile was added via syringe. Then the nitrogen gas was bubbled through the reaction mixture for about 10 min and alkyl sulfonate substrate (0.10 mmol) was added and nitrogen was bubbled again for 5 min. The reaction mixture was placed under blue light after sealing the punctured holes of the vial cap with vacuum grease and electric tape. The reaction mixture was stirred for 16 h, and then about 0.1ml of the reaction mixture was filtered through silica (using short pipet column) to remove solid residue. The silica was washed with CDCl3 (0.1 mL). The filtrate was then washed with DI water and the lower CDCl3 layer was taken for 1 H NMR spectroscopy after drying with MgSO4.

Note:
In the dehydrohalogenation and remote elimination reactions, some corresponding alkane as side product was absorbed which made the isolation of the desired product difficult because of their similar Rf. Therefore, the yield of the target compounds have been reported based on the mmol of internal standard, i.e. 1,3,5, trimethoxybenzene (all the target compounds are known and previously characterized).

3-Butenyloxytrimethylsilane (Table 2, entry 2)
The title compound was synthesized using procedure described in general method 4. The NMR yield of the product was calculated using 1, 3, 5-trimethoxybenzene as an internal standard and was found to be 60 % yield. The compound was identical to the compound reported in the literature by proton NMR.(13)

4-(4-methoxy phenyl)-1-butene (Table 2, entry 7)
The title compound was synthesized using procedure described in general method 4. The NMR yield of the product was calculated using 1, 3, 5-trimethoxybenzene as an internal standard and was found to be 71% yield. The compound was identical to the compound reported in the literature by proton NMR. (14) Hex-5-en-1-yl 2-(3-(trifluoromethyl)phenyl)acetate ( The title compound was synthesized using procedure described in general method 4. The NMR yield of the product was calculated using 1, 3, 5-trimethoxybenzene as an internal standard and was found to be 57% yield. The compound was identical to the compound reported in the literature by proton NMR.(4)

Hex-5-en-1-yl furan-2-caboxylate (Table 2, entry 9)
The title compound was synthesized procedure described in general method 4. The NMR yield of the product was calculated using 1,3,5-trimethoxybenzene as an internal standard and was found to be 61% yield. The compound was identical to the compound reported in the literature by proton NMR. (15) Hex-5-en-1-yl thiophene-2-caboxylate ( Table 2, entry 10) The title compound was synthesized using procedure described in general method 4. The NMR yield of the product was calculated using 1,3,5-trimethoxybenzene as an internal standard and was found to be 66% yield. The compound was identical to the compound reported in the literature by proton NMR. (15) Hexyl-5-en-1-yl quinoline-3-carboxylate (Table 2, entry 11) The title compound was synthesized using procedure described in general method 4. The NMR yield of the product was calculated using 1,3,5-trimethoxybenzene as an internal standard and was found to be 53% yield. The compound was identical to the compound reported in the literature by proton NMR.(4) (Table 2, entry 12) The title compound was synthesized using procedure described in general method 4. The NMR yield of the product was calculated using 1,3,5-trimethoxybenzene as an internal standard and was found to be 69% yield. The compound was identical to the compound reported in the literature by proton NMR.(16) (Table 2, entry 13)

1-(Hex-5-en-1-yloxy)-4-fluorobenzene
The title compound was synthesized using procedure described in general method 4. The NMR yield of the product was calculated using 1,3,5-trimethoxybenzene as an internal standard and was found to be 55% yield. The compound was identical to the compound reported in the literature by proton NMR.(17) (Table 2, entry 14)

3,5-Cholestadiene
The title compound was synthesized using procedure described in general method 4. The NMR yield of the product was calculated using 1,3,5-trimethoxybenzene as an internal standard and was found to be 52% yield. The compound was identical to the compound reported in the literature by proton NMR. (18) 1H-Indene (Table 3, entry 1) The title compound was synthesized using procedure described in general method 4. The NMR yield of the product was calculated using 1,3,5-trimethoxybenzene as an internal standard and was found to be 87% yield. The compound was identical to the compound reported in the literature by proton NMR.
The title compound was synthesized using procedure described in general method 4. The NMR yield of the product was calculated using 1,3,5-trimethoxybenzene as an internal standard and was found to be 54% yield. The compound was identical to the compound reported in the literature by proton NMR.(20) (Table 3, entry 3)

1-Dodecene
The title compound was synthesized using procedure described in general method 4. The NMR yield of the product was calculated using 1,3,5-trimethoxybenzene as an internal standard and was found to be 56% yield. The compound was identical to the compound reported in the literature by proton NMR.(4) (Table 3, entry 4)

1-Dodecene
The title compound was synthesized using procedure described in general method 4. The NMR yield of the product was calculated using 1,3,5-trimethoxybenzene as an internal standard and was found to be 56% yield. The compound was identical to the compound reported in the literature by proton NMR.(4) (Table 3, entry 5)

1-(But-3-ene-1-yl)-4-methoxybenzene
The title compound was synthesized using procedure described in general method 4. The NMR yield of the product was calculated using 1,3,5-trimethoxybenzene as an internal standard and was found to be 75% yield. The compound was identical to the compound reported in the literature by proton NMR. (14) Ph Me(H 3 C) 8

H 3 CO
Ph Pent-4-en-1-ylbenzene (Table 3, entry 6) The title compound was synthesized using procedure described in general method 4. The NMR yield of the product was calculated using 1,3,5-trimethoxybenzene as an internal standard and was found to be 88% yield. The compound was identical to the compound reported in the literature by proton NMR. (20) 1-Dodecene (Table 3, entry 7) The title compound was synthesized using procedure described in general method 4. The NMR yield of the product was calculated using 1,3,5-trimethoxybenzene as an internal standard and was found to be 53% yield. The compound was identical to the compound reported in the literature by proton NMR.(4) (Table 4, entry 6)

1-(4-Methoxyphenyl)but-2-ene
The title compound was synthesized using procedure described in general method 5. The NMR yield of the product was calculated using 1,3,5-trimethoxybenzene as an internal standard and was found to be 41% yield. The compound was identical to the compound reported in the literature by proton NMR.(21)

3-Penten-1-ylbenzene (Table 4, entry 7)
The title compound was synthesized using procedure described in general method 5. The NMR yield of the product was calculated using 1,3,5-trimethoxybenzene as an internal standard and was found to be 52% yield. The compound was identical to the compound reported in the literature by proton NMR. (22) 2-Dodecene (

Me(H 2 C) 6
The title compound was synthesized using procedure described in general method 5. The NMR yield of the product was calculated using 1,3,5-trimethoxybenzene as an internal standard and was found to be 66% yield. The compound was identical to the compound reported in the literature. (23)

A. Inhibition Experiments i) TEMPO Inhibition Experiments
In An aliquot of the reaction was drawn using syringe and passed through a short silica plug and diluted with ACN-d3. and the reaction was monitored via NMR. The yields were calculated using 1, 3, 5-trimethoxybenzene as an internal standard. The 1-octene A product was identical to authentic material (Aldrich) by 1 H NMR spectroscopy.

ii) 1,1-Diphexylethylene Inhibition Experiment
The reaction was set up following the standard method as described in the TEMPO inhibition experiment in dry acetonitrile using 1-bromooctane substrate (0.03 ml, 0.20 mmol) and 1,1-Diphenylhexylethylene (18.0 mg, 0.10 mmol). An aliquot of the reaction was drawn using syringe and passed through a short silica plug and diluted with CDCl3 and NMR was taken after aqueous workup and drying with Na2SO4. The yield was calculated using 1, 3, 5-trimethoxybenzene as an internal standard. The 1-octene A product was identical to authentic material (Aldrich) by 1 H NMR spectroscopy.

iii) BHT Inhibition Experiment
The reaction was set up following the standard method as described in the TEMPO inhibition experiment using 1-bromooctane substrate (0.03 ml, 0.20 mmol) and 2,6-Di-tert-butyl-4methylphenol (22.0 mg, 0.10 mmol). An aliquot of the reaction was drawn using syringe and passed through a short silica plug and diluted with CDCl3 and NMR was taken after aqueous workup and drying with Na2SO4. The yield was calculated using 1, 3, 5-trimethoxybenzene as an internal standard. The 1-octene A product was identical to authentic material (Aldrich) by 1 H NMR spectroscopy.

B. Radical Clock Experiment
The reaction was set-up following the procedure described in the general method 3 using 6-bromo-1-hexene (16.3 mg, 0.100 mmol) as substrate. The reaction went to completion after 20 h to form 5-hexadiene A (52%) along with reduced side product B (32%) as shown in the Scheme S1. In this transformation, no trace of cyclic product was observed supporting our proposed reaction mechanism showing evidence of the rapid Co(II)-perpetrated HAT step in the catalytic cycle. The NMR yields were calculated using 1,3,5-trimethoxybenzene as an internal standard.
The product A is identical to the literature value by 1 H NMR spectroscopy. (24) Scheme S3: The mechanistic study showing the formation of hexadiene product over cyclic product.

C. Isotope Labeling Experiment
The isotope labeling experiments were conducted following the procedure described in general method 3 using 1-bromo-3-phenylpropane (19.9 mg, 0.10 mmol) as substrate and 4 equiv. of NaBD4 (16.7 mg, 0.40 mmol). After 22h, from 1 H NMR analysis; 48% of starting material was left unreacted and 3-phenylpropene product formed was calculated to be 48% (Scheme S2a). Whereas, when the same substrate was subjected to the reaction procedure described in general method 4 (remote elimination) both the subterminal alkene and terminal alkenes in 29% each along with 40% unreacted starting material (Scheme S2b). The NMR yield were obtained using 1,3,5trimethoxybenzene as internal standard. No deuterium incorporation in the final products was observed. The trans-1-Phenyl-1-propene product and allylbenzene products were identical to authentic material (Aldrich) by 1 H NMR spectroscopy.     Table 2