Racemic or enantioselective osmium-catalyzed dihydroxylation of olefins under near-neutral conditions

K 3 Fe(CN) 6 and NaIO 4 serve as catalytic co-oxidants for osmium-catalyzed dihydroxylations that are performed under near-neutral conditions with K 2 S 2 O 8 as the stoichiometric oxidant and Na 2 HPO 4 as the base. By using either quinuclidine or hydroquinidine 1,4-phthalazinediyl ether [(DHQD) 2 Phal], good yields of racemic or enantioenriched diols are obtained. This simple, biphasic procedure offers advantages over other neutral dihydroxylation protocols that use N -methylmorpholine oxide as the stoichiometric oxidant


Introduction
During the course of our unpublished investigations directed toward the synthesis of cortistatin A (3) from 2 (common atoms in 2 and 3 highlighted in blue), we were confronted by what seemed to be a routine osmiumcatalyzed dihydroxylation of the allyl furan 1 to give the dihydroxylated product 2. To our surprise, when 1 was subjected to standard Sharpless asymmetric dihydroxylation conditions (AD) using AD-mix-β, [1][2][3] the reaction required four to five days and provided the diol 2 in only 20-40% yield (Scheme 1, Condition A) together with a mixture of unidentified side products.We suspected that prolonged reaction times under the basic reaction conditions might be leading to deleterious aldol reactions involving the cyclopentanone moiety, but efforts to increase the rate of reaction using known hydrolysis aides such as MeSO2NH2 4,5 and PhB(OH)2 6,7 failed to provide any improvement.After some experimentation, we found that supplementing the standard AD-mix with K2S2O8 significantly increased the rate of the reaction and improved the yield to 55% (Condition B). 8 Further increasing the quantity of K2S2O8 and using K3Fe(CN)6 as the catalytic co-oxidant dramatically accelerated the reaction and enabled the isolation of 2 in 76% yield in less than 16 h without employing ligands or hydrolysis aides (Scheme 1, Condition C).

Scheme 1
Osmium-catalyzed dihydroxylations can generally be accelerated by either increasing the rate of olefin oxidation through the addition of ligands or by enhancing the rate of hydrolysis of the intermediate osmate ester, thereby returning osmium to the catalytic cycle.The latter can be facilitated by the addition of hydrolysis aides, such as MeSO2NH2, to the mixture or by maintaining the pH around 12 because an AD reaction typically starts at a pH of about 12.2 but drops to 9.9 over the course of the reaction. 9Some oxidants, such as NaClO2 10 and NaOCl, 11 appear to accelerate the hydrolysis of the osmate ester by releasing hydroxide ions, but hydroxide ions are not produced when K2S2O8 is used as the terminal oxidant.Indeed, when K2S2O8 is used, we observed that the pH of the reaction was about 11.3, a full pH unit less than the standard AD conditions.The capability of K2S2O8 to serve as an oxidant under less basic conditions led us to wonder if the pH could be further lowered, so that the asymmetric osmium-catalyzed dihydroxylations could be applied to base-sensitive substrates.Such a modification would offer several advantages over other currently available protocols that use NaHCO3 to buffer the reaction to pH 10.3, 12 but AD reactions do not turn over if NaHCO3 is replaced with K2CO3. 1 Use of N-methylmorpholine oxide (NMO) as the terminal oxidant can allow for the dihydroxylation of olefins under neutral, [13][14][15] or even acidic conditions, 16 but these reactions often suffer from slower rates and inferior enantioselectivities when compared to the normal biphasic Sharpless AD conditions.The reduced enantioselectivities have been attributed to a secondary catalytic cycle, which occurs when the oxidant is in the same phase as the osmium catalyst. 1We thus explored the possibility of developing a Sharpless-style AD that could be performed at near-neutral pH, so it could be applied to base-sensitive substrates without sacrificing enantioselectivity.

Results and Discussion
The first step toward modifying Condition C to convert 1 into 2 involved screening different bases, and in initial studies we found that replacing K2CO3 with either NaHCO3 or Na2HPO4 gave ~20% of the desired diol 2, while buffering the reaction medium to pH 9.8 and 8.6, respectively.Having established that hydroxylation did occur at lower pH, we set to the task of optimizing the more challenging dihydroxylation of methyl cinnamate (4) as the model substrate using Na2HPO4 as the base.We hypothesized that the reaction was not proceeding to completion because of slow catalytic turnover resulting from the absence of ligands or hydrolysis aides coupled with the lower pH. 9To remedy this problem, we investigated a variety of additives, and we discovered that adding quinuclidine 17 as a ligand and MeSO2NH2 4,5 to facilitate hydrolysis led to a complete reaction and provided the diol 5 in 66% yield (Table 1, entry 1).When K3Fe(CN)6 was omitted as the co-oxidant from the reaction, no 5 was isolated (Table 1, entry 2), while increasing the stoichiometry of K3Fe(CN)6 led to only a marginal increase in yield (Table 1, entry 3).On the other hand, increasing the stoichiometry of the base Na2HPO4 from three to four equivalents (equiv), furnished the 5 in 91% yield (Table 1, entry 4).Inasmuch as K3Fe(CN)6 functions only as a co-oxidant, we were curious whether any other cooxidant might be used in lieu of K3Fe(CN)6.
Several co-oxidants were examined as possible replacements for K3Fe(CN)6, and although use of NaIO4, KBrO3, and NaClO2 afforded 5, NaIO4 emerged as the best co-oxidant giving 5 in 63% yield (Table 1, entry 5).Increasing the stoichiometry of the co-oxidant provided the diol 5 in 87% yield (Table 1, entry 6).Interestingly, whereas the K3Fe(CN)6 may be used as the co-oxidant under more basic conditions, the rate of dihydroxylation of 4 using NaIO4 decreases at higher pH.In fact, use of K2CO3 as the base with NaIO4 as co-oxidant gave only small quantities of 5 (Table 1, entry 7).That NaIO4 may be employed as the catalytic co-oxidant for olefin dihydroxylations under near neutral conditions is perhaps at first surprising and thus warrants brief comment.Although there are reports of a RuO4-catalyzed dihydroxylation that uses NaIO4 as the stoichiometric oxidant, 18,19 use of NaIO4 with OsO4 usually results in oxidative cleavage of the olefin, a reaction widely-known as the Johnson-Lemieux oxidation.Indeed, when methyl cinnamate (4) was treated with a stoichiometric quantity of NaIO4 in the presence of a catalytic amount of OsO4, the olefin suffered the expected oxidative cleavage to form benzaldehyde; no 5 was isolated.Notably, the pH of that reaction was 5.7, much lower than the pH of 8.6 that was used for the dihydroxylation of 4 using NaIO4.Indeed, it has been reported that the Johnson-Lemieux reaction occurs only slowly at neutral pH or in the presence of K2CO3, 20 and we found in an exploratory experiment that reducing the pH of the NaIO4 co-catalyzed dihydroxylation using phosphoric acid resulted in the formation of some benzaldehyde.
Having established optimized conditions for the racemic dihydroxylation of methyl cinnamate (4) using either NaIO4 or K3Fe(CN)6 as the catalytic co-oxidant, we explored the substrate scope with several substituted styrenes as standard substrates (Table 2, entries 1-4).The yields using either NaIO4 or K3Fe(CN)6 were comparable, and di-and trisubstituted alkenes are suitable substrates, although yields for the latter are somewhat lower.To demonstrate the potential utility of these conditions for the dihydroxylation of base sensitive substrates, 9-decenal ( 12) was chosen as a test compound.Dihydroxylation of 12 using AD-mix-β, with and without added NaHCO3 as a buffer, provided diol 13 in 33% and 51% yields, respectively; the absolute stereochemistry of 13 is tentatively assigned based upon literature precedent for 1-decene. 21On the other hand, dihydroxylation of 12 under near-neutral conditions with K3Fe(CN)6 as the co-catalyst provided the diol 13 in 82% yield, whereas use of NaIO4 as the co-oxidant gave 13 in only 45% yield (Table 2, entry 5).Because reactions with K3Fe(CN)6 in the presence of Na2HPO4 are generally faster that those using NaIO4, we attribute this discrepancy to the instability of 13 to prolonged exposure under the reaction conditions.
We then applied our modified procedure for olefin dihydroxylation to the enantioselective variant.Because reactions using dihydroquinidine 1,4-phthalazinediyl diether [(DHQD)2Phal] are significantly faster than those using quinuclidine as a ligand, reduced quantities of (DHQD)2Phal are required.Generally, the enantioselective dihydroxylations with (DHQD)2Phal in the presence of NaIO4 are higher yielding than those using K3Fe(CN)6 as the co-oxidant, but the yields in the enantioselective processes are lower than those with quinuclidine as the ligand (Table 2, entries 1-5).The enantioselectivities were comparable to the those reported in the literature, 4 and they did not depend on whether NaIO4 or K3Fe(CN)6 served as the co-oxidant.Although the yield of racemic 13 was better when K3Fe(CN)6 was used as the co-oxidant, NaIO4 gives superior yields under the conditions for enantioselective dihydroxylation of 12 (Table 2, entry 5). a Experimental details given under the general procedure for racemic dihydroxylation of olefins, using 3 equiv of Na2HPO4unless otherwise indicated.b ) Experimental details given under the general procedure for enantioselective dihydroxylation of olefins, using 3 equiv of Na2HPO4 unless otherwise indicated.c ) Enantiomeric excess (ee) was determined according to procedures reported by Sharpless. 21d ) 4 equiv of Na2HPO4 were used.e ) 0.1 equiv of co-oxidant was used.f ) MeSO2NH2 was not used.g) The 1 H-NMR spectrum of 13 suggested it existed as a mixture of aldehyde hemiacetals, so it was characterized as 1,2,10-decanetriol (see Experimental Section for details).h) ND (not determined).

Conclusions
Experiments to identify mild conditions that effect the dihydroxylation of the base-sensitive allyl furan 1 to give the diol 2 revealed that using K2S2O8 as the stoichiometric oxidant and K3Fe(CN)6 as a catalytic co-oxidant leads to faster reactions at a pH that is lower than other traditional osmium-catalyzed methods.Further optimization of these conditions led to a protocol for performing racemic and enantioselective dihydroxylations of alkenes that proceed under near-neutral conditions.In this modified procedure, Na2HPO4 serves as the base, MeSO2NH2 facilitates hydrolysis of the intermediate osmate ester, and either quinuclidine or (DHQD)2Phal are used as the ligand.NaIO4 was found to function as a catalytic co-oxidant, and in some cases of the enantioselective variant it outperformed K3Fe(CN)6.The utility of this procedure was demonstrated by obtaining good yields of diol even in the presence of a base-sensitive aliphatic aldehyde.We believe these conditions will expand the scope of the osmium-catalyzed dihydroxylation reaction to include a wider range of substrates, especially those that are sensitive to basic conditions.

Experimental Section
General procedure for racemic dihydroxylation of olefins Water (5 mL/mmol substrate) was added to a solid mixture of K2S2O8 (1.5 equiv), Na2HPO4 (3 or 4 equiv), K3Fe(CN)6 (0.2 equiv) or NaIO4 (0.2 equiv), MeSO2NH2 (1 equiv) and K2OsO2(OH)4 (0.05 equiv) at room temperature, and the mixture was stirred for 5 min.Quinuclidine (0.3 equiv), tert-BuOH (5 mL/mmol substrate), and the olefin (1 equiv) were then added sequentially, and the reaction was stirred at room temperature until the olefin was consumed as judged by TLC.A solution of saturated aqueous Na2S2O3 was added, and the mixture was extracted with CH2Cl2 (3 x 5 mL/mmol substrate).The combined organic extracts were dried (Na2SO4), filtered, and concentrated under reduced pressure, and the crude product was purified by flash chromatography to provide the pure diol.

Synthesis of 9-decenal (12)
Pyridinium chlorochromate (PCC) (830 mg, 3.85 mmol) was added to a slurry of basic alumina (3.84 g) in CH2Cl2 (6 mL) and stirred at room temperature for 2 h.Dec-9-en-1-ol (300 mg, 355 µL, 1.92 mmol) was added, and the mixture was stirred at room temperature for 3 h.The reaction mixture was filtered through a silica plug and eluted with a mixture of EtOAc/hexanes (1:3, 50 mL).The combined filtrate and washings were removed under reduced pressure to provide 288 mg (97%) of 12 that was used in the dihydroxylation reaction without further purification.Characterization of diol 13 as its reduction product 1,2,10-decanetriol NaBH4 (78 mg, 2.06 mmol) was added to a solution of 13 (78 mg, 0.414 mmol) in MeOH (5 mL) at room temperature.The reaction was stirred for 4 h, whereupon AcOH (~0.1 mL) was added, and the mixture was concentrated under reduced pressure.The crude product was purified by flash chromatography eluting with a gradient of acetone/hexanes (1:2 → 3:4) to provide 54 mg (69%) of the triol as a white solid: mp 53-55 C;