Desymmetrization Approach to the Synthesis of Optically Active P-Stereogenic Phosphin-2-en-4-ones

Two synthetic protocols for the conversion of 1-phenylphosphinan-4-ones to novel P-stereogenic 1-phenylphosphin-2-en-4-ones by enantioselective deprotonation followed by oxidation and by asymmetric organocatalytic halogenation accompanied by elimination have been developed. These two-step one-pot transformations provide convenient access to optically active 1-phenylphosphin-2-en-4-one 1-sulfide and 1-phenylphosphin-2-en-4-one 1-oxide of 96 and 55% enantiomeric purities, respectively.


■ INTRODUCTION
Cyclic nonracemic phosphines constitute an important group of organophosphorus compounds that are sought for their advantageous performance as organocatalysts and as ligands in various asymmetric processes. 1 Numerous chiral five-membered (phospholane) 2 and four-membered (phosphetane) ligands 3 have been developed to meet the demand. In contrast, the corresponding chiral six-membered carbon-phosphorus heterocycles (phosphinanes) have received relatively little attention 4,5 due, most probably, to scarcity of convenient methods enabling their synthesis in suitably functionalized and nonracemic forms. 6 For illustration, all the optically active phosphines and phosphine oxides containing phosphorus embedded in the sixmembered ring, which have been synthesized to date, are collected in Figure 1A,B.
There has recently been considerable interest in preparation of P-stereogenic phosphorus compounds by desymmetrization reactions starting from P-prochiral precursors. 7 Synthesis of cyclic phosphine derivatives by this route can start either from an acyclic, 4e4i or from a cyclic 8,9 precursor. In the latter case, the reported precedents included phosphol-3-ene oxide 8 and its epoxide, 9 phosphetane sulfide, 3d and phospholane sulfide 2f as well as phospholane borane and phosphinane boranes. 4a In this paper, we wish to report our results on evaluation of enantioselective desymmetrization of 1-phenylphosphinan-4one (1) by employing its carbonyl function in two independent two-step processes designed to lead to the formation of optically active 1-phenylphosphin-2-en-4-one derivatives 4 ( Figure 1C). The target phosphin-2-en-4-one, equipped with a versatile enone functionality, represents a novel phosphinane scaffold potentially amenable to rich chemistry further downstream.

■ RESULTS AND DISCUSSION
Of the known synthetic methods used frequently for desymmetrization of prochiral ketones, 10 enantioselective deprotonation, 10a10b and enantioselective α-halogenation, 10f10g seemed to be most suitable for accomplishing our goal. Accordingly, the two alternative paths that we have designed to lead to optically active 4 are based on these two desymmetrization processes (Scheme 1).
The desymmetrization by path A involves asymmetric deprotonation of 1-phenylphosphinan-4-one (1) by a chiral base and conversion of the resulting lithium enolate to the silyl enol ether 2 by quenching with TMSCl. 10a10b The desymmetrization by path B entails transformation of phosphinanone 1 into a chiral α-halogenated derivative 3, which could be achieved by organocatalytic asymmetric α-halogenation. 10f10g Both synthetic procedures make use of the ketone functionality of phosphinanone 1, and both result in the overall asymmetric transformation of the remote prochiral phosphorus center in ketone 1 into a P-stereogenic one in enone 4 via intermediate 2 or 3.
Since the time the enantioselective deprotonation of cyclic ketones by a chiral lithium amide was first demonstrated in 1986, 10a10b the method has been widely utilized in asymmetric synthesis for generating chirality centers in cyclic ketones by desymmetrization. 11 Although efficient desymmetrizations of a number of oxa-, aza-, and thia-heterocyclic ketones by chiral lithium amides have been already demonstrated, 10f1012 the corresponding P-heterocyclic analogs have not been investigated before. Thus, we started with checking the viability of enantioselective deprotonations of 1-phenylphosphinan-4-one 1-oxide (1a), 1-borane (1b), and 1-sulfide (1c) using lithium amide derived from amine (S,S)-5 as the model base premixed with an excess of TMSCl before addition of a ketone (ISQ -in situ quench) 13 (Table 1).
As shown in Table 1 (entries 1−3), phosphinanone oxide 1a failed to provide silyl enol ether 2a, whereas borane 1b and sulfide 1c gave the expected enol ethers 2b and 2c, respectively, albeit in low yields and with very low ee. Subsequent testing of phosphinanone sulfide 1c revealed that addition of 0.5 equiv. of LiCl allowed increasing the yield and ee of silyl enol ether 2c to more acceptable levels (entry 5) and that allowing lithium amide to react with phosphinanone sulfide 1c before TMSCl was introduced (EQ -external quench) 14 gave slightly better results than the ISQ alternative (entries 5 and 6). As checked under these conditions again, the amount of 0.5 equiv. of LiCl was sufficient; increasing its loading to 1 equiv. did not bring about improvement of ee. The Journal of Organic Chemistry pubs.acs.org/joc Article The details of further optimization of these reaction conditions, which included variations of molarity, stoichiometry, and temperature, are presented in Table 2.
As shown in Table 2, lowering of concentration led to improvement of enantioselectivity, but unfortunately, it led to a substantial decrease in yield (entries 1−3). The concentration of 0.025 M was deemed a practical compromise and was then used in subsequent trials. A substantial increase of enantioselectivity to 74% ee at 85% conversion was observed when 3 equiv. of amine 5 was used instead of 1.5 equiv. (entries 4 and 5). In addition, lowering of the reaction temperature to −90°C resulted in further enhancement of enantioselectivity up to 87% ee at 81% conversion (entry 10). Finally, checking the concentration factor once again confirmed that its lowering resulted in a substantial decrease in yield, but this time, it was not even accompanied by an increase of enantioselectivity observed before (cf. entries 10 and 12).
Once the optimization of the reaction conditions was completed, also other amine catalysts were tested for their efficiency in desymmetrization of 1-phenylphosphinan-4-ones 1c and 1b. The results obtained with chiral monoamines 5−13, 15, and 16 and diamines 14 and 17−19 are displayed in Table 3.
Inspection of the results collected in Table 3 reveals that the best enantioselectivities in desymmetrization of phosphinanone sulfide 1c were achieved with C 2 -symmetric lithium bis(αarylethyl)amides derived from amines (S,S)-5 and (S,S)-6, i.e., 87 and 76% ee, respectively. The C 1 -symmetric α-phenylethylamine derived bases 7−12 and 16 were also effective in desymmetrizing phosphinanone sulfide 1c and gave silyl enol ether 2c in good yield and with enantioselectivity reaching 59% ee. Diamines 14 and 17−19 gave slightly lower enantioselectivities than the monoamines. In turn, desymmetrization of phosphinanone borane 1b carried out with lithiated 5−19 under the same conditions gave silyl enol ether 2b in generally better yields but with much lower enantioselectivities than sulfide 2c. For borane 2b, the best ee's were again achieved with lithium amides derived from (S,S)-5 and (S,S)-6, i.e., 61% ee at 95% conversion and 52% ee at 68% conversion, respectively.
Next, we turned our attention to the oxidation of silyl enol ethers 2b,c required for their conversion into phosphinenones 4b,c. Our initial attempts involved use of the well-known procedures utilizing Pd(OAc) 2 in acetonitrile, 15 DDQ in benzene, and trityl tetrafluoroborate in dichlorometane 16 as the oxidizing agents, but with these reagents, phosphinenones 4 were produced in very low yields (Table 4, entries 1−3). Subsequent treatment of silyl enol ethers 2c and 2b with ceric ammonium nitrate (CAN) in DMF 17 led to the formation of phosphinenones 4c and 4a in 69 and 74% yields, respectively (entries 4 and 5). It should be noted, however, that under these conditions, phosphinenone borane 4b could not be obtained due to concurrent oxidation of the P center during the reaction course. Finally, using Nicolaou et al.'s procedure for the oxidation of silyl enol ethers to α,β-unsaturated carbonyl compounds utilizing the IBX·MPO complex as the oxidant, 18 phosphinenones 4c and 4a (from 2b) were obtained in high yields, 80 and 73%, respectively (entries 8 and 9).
Encouraged by the latter's promising results and taking into account the fact that silyl enol ethers 2c and 2b proved to be    highly susceptible to hydrolysis during chromatographic purification, we decided to combine the best desymmetrization and oxidation protocols found for phosphinanone 1c in a onepot process to avoid substantial loss of the intermediate silyl enol ether during isolation (Scheme 2).   Determined for the crude reaction mixture by GC−MS and 31 P NMR analysis. b Identified as oxides 4a and 1a due to P-oxidation occurring under the reaction conditions. c Isolated yield of enone 4. d rt = 18−22°C.

g (5 mmol) Scale
The Journal of Organic Chemistry pubs.acs.org/joc Article    the oxidation step. Importantly, however, recrystallization of the isolated sulfide 4c of 73% ee from hexane/i-PrOH allowed its enantiomeric purity to increase to 96% ee.
In the second part of our study, we turned our attention to another organocatalytic strategy expected to be suitable to achieve our goal. In 2005, Jørgensen et al. 10g described the first enantioselective α-bromination of ketones utilizing N-bromosuccinimide (NBS) and 4,4-dibromo-2,6-di-tert-butyl-cyclohexa-2,5-dienone (20) as the brominating agents and (S)proline and a C 2 -symmetric imidazolidine as the chiral catalysts. These reagents enabled the formation of stereogenic C−Br centers with up to 94% ee in high yields. 10g We decided then to check the viability of this protocol in the asymmetric αbromination of phosphinanones 1a,c, which, when followed by elimination of HBr, could lead to the target optically active phosphinenones 4a,c.
We started our investigations with a brief screening of solvents and additives in α-bromination of oxide 1a and sulfide 1c, using NBS (or 20) as the brominating agent and (S)-proline as the model chiral catalyst. At the outset, we were pleased to find that elimination of HBr started to occur already under the bromination conditions and that practically quantitative elimination of HBr could be achieved by simply raising the temperature at the end of the reaction to 60°C for half an hour. We included this maneuver to the screening conditions to make the planned synthesis of phosphinenones 4a,c a one-pot process ( Table 5).
As can be seen from the collected data, a change of solvent as well as an added acid 19 can strongly influence the outcome of the reaction (Table 5, entries 7−13). With added benzoic acid, the enantiomerically enriched 4a was obtained with 34% ee and in 47% yield, what constituted a significant improvement over the reaction run without this additive in the same solvent (DCM) (cf. entries 7 and 12). In turn, changing the solvent to THF or DMF resulted in a marked increase of the conversion, but the observed enantioselectivity was significantly lowered. Thus, the conditions utilizing DCM and added benzoic acid (entry 12), which best compromised the conversion and induction levels, were selected for screening of a number of other chiral amine catalysts in the next optimization step. The results of this screening are summarized in Table 6. The reactions of all tested amines were performed with and without benzoic acid, but only the better result of these two runs has been listed for clarity.
Also listed in Table 6 are the results of desymmetrization of phosphinanone sulfide 1c carried out with compound 20 as the brominating agent under otherwise the same conditions. These reactions proceeded relatively well and afforded enone 4c in good yields (56−84%) but with only moderate enantiomeric enrichment (8−38% ee). Possibly the best match of yield and enantiomeric purity of 4c was achieved with DACH derivative (R,R)-26 (66% and 38% ee, respectively) and with imidazolidine (S,S)-25 (84% and 33% ee, respectively).
Looking for further improvement, we also decided to briefly check the efficiency of analogous enantioselective α-chlorinations, which have been recently demonstrated to be highly efficient in the case of six-membered-ring ketones. 10f The results of screening experiments involving chlorination of phosphinanones 1a,c by NCS and PhICl 2 in the presence of (S)-proline and other amine catalysts, followed by DBU-assisted elimination of HCl from intermediate α-chloro ketone 3-Cl to give enone 4a,c, are collected in Table 7.
The collected data reveal that PhICl 2 as the chlorine source gave better conversions than NCS and that addition of benzoic  4). Similarly, chlorinations of phosphinanone 1a with amines 17, 21, and 24 as the catalysts also led to the formation of racemic enone 4a, although in these cases with remarkably high conversions of 91, 87, and 88%, respectively (entries 8−10). In turn, imidazolidine (S,S)-25, the reported excellent catalyst for the asymmetric α-chlorination of sixmembered-ring ketones, 10f afforded enantioenriched enone 4a of only 30 (with PhICl 2 ) or 21% ee (with NCS) in moderate 34% and good 74% yields, respectively (entries 5 and 6). It is important to note, however, that in these two cases, as determined by comparison of the pertinent CSP-HPLC chromatograms, the use of (S,S)-25 as the catalyst led to the formation of enone 4a enriched in the enantiomer opposite to that found in predominance in 4a obtained by the αbromination procedure utilizing the same (S,S)-25 as the catalyst. Interestingly, an attempted reaction of sulfide 1c under exactly the same conditions failed completely (entry 7). At this point, considering that the prospect of getting high enantioselectivity in desymmetrizations of phosphinanones 1a,c by αchlorination did not look promising, further optimization of this process was discontinued. Nonetheless, despite the fact that the chlorination procedure did not provide the expected improvement of enantioselectivity in the studied syntheses of optically active phosphinenones 4, the developed one-pot chlorinationelimination procedure is likely to find use as an effective method for synthesis of racemic phosphinenone oxide 4a (cf. entries 8− 10). All in all, it is tempting to conclude that enantioselective αhalogenation of phosphinanone 1, a six-membered-ring ketone possessing a phosphorus function in the γ position, is considerably more challenging than the parent cyclohexanone and related six-membered-ring ketones., 10f10g Moreover, a poor result of our attempted organocatalytic desymmetrization of phosphinanone oxide 1a via enamine oxidation under recently reported optimized conditions 20 shown to be effective in converting a whole variety of mono and doubly 4-substituted cyclohexanones to the corresponding cyclohexenones of very high enantiomeric purity corroborates this notion further (Scheme 3).

■ CONCLUSIONS
Even though the asymmetric deprotonation and asymmetric halogenation of phosphinanone 4 have turned out to be less efficient than those of carbocyclic ketones, the developed onepot enolization-oxidation and halogenation-elimination procedures have for the first time provided access to the new Pstereogenic phosphin-2-en-4-one derivatives in nonracemic forms. A good level of asymmetric induction (87% ee at 81% conversion) can be achieved by enantioselective deprotonation of phosphinanone sulfide 1c at −90°C using 3 equiv. of lithium amide derived from commercially available amine S,S-5. Subsequent in situ oxidation of the formed enantiomerically enriched silyl enol ether 2c by IBX·MPO converts it to optically active phosphinenone 4c, the enantiopurity of which can be upgraded to 96% ee by recrystallization. Desymmetrization of phosphinanone oxide 1a can be best achieved by asymmetric αbromination using (S)-proline amide 24 as the catalyst to provide enriched 3-bromophosphinanone 3, which, in turn, undergoes in situ elimination of HBr to afford phosphinenone 4a of 55% ee in 54% yield. The analogous asymmetric αchlorination-elimination procedure offers very low or even no enantioselectivity in desymmetrization of phosphinanone 1a. Nevertheless, it allows obtaining phosphinenone oxide 4a in very high yields (cf. Table 7, entries 8−10) and may thus constitute a useful route to rac-4a.

■ EXPERIMENTAL SECTION
General Information. All reactions were performed under an argon atmosphere using Schlenk techniques or in a 10 mL glass reaction tubes with a screw cap. Only dry solvents were used, and the glassware was heated under vacuum prior to use. THF was dried over sodium/ benzophenone ketyl. LiCl was dried in a Schlenk tube under vacuum at 150°C for 5 h. TMSCl, NBS, MPO, DMSO, chiral amines 5, 6, 13, 14, 21, and (S)-proline (19) were purchased from commercial sources and used without further purification. Solvents for chromatography and extraction were commercially available and used as received without further purification. Solvents for crystallization and Et 3 N were distilled once before use. Room temperature (rt) means a range of temperatures from 18 to 22°C.
The NMR spectra were recorded with a Bruker Ascend (500 MHz) spectrometer in CDCl 3 as a solvent at room temperature unless otherwise noted. Chemical shifts (δ) are given in ppm relative to tetramethylsilane ( 1 H), residual CHCl 3 ( 13 C), or external 85% H 3 PO 4 ( 31 P) as a reference. The following abbreviations are used in reporting NMR data: s (singlet), d (doublet), t (triplet), q (quartet), m (multiplet), br (broad). Coupling constants (J) are in Hz. Highresolution mass spectrometry analyses were obtained on a Shimadzu LCMS IT -TOF spectrometer. Elementary analyses were performed on a PerkinElmer CHN 2400 analyzer. Melting points were determined on a Buchi Melting Point M -560 in a capillary tube and are uncorrected. Mass spectra were recorded with a GC−MS spectrometer working in electron ionization (EI) mode. Chiral HPLC analysis was performed on a Shimadzu HPLC using Chiralcel columns. Optical rotations were measured on a PerkinElmer 341LC spectrometer using a 1 mL cell with a 10 mm path length and are reported as follows: [α] D 20 (c g/100 mL, solvent). Thin layer chromatography (TLC) was performed with precoated silica gel plates and visualized by potassium permanganate (KMnO 4 ) staining or exposing to iodine vapor. The reaction mixtures were purified by column chromatography over silica gel (60−240 mesh). The chiral amines 7−12, 21 15−16, 22 17, 23 18, 24 25, 25 26, 26 27, 27 and 28−30 28 were prepared according to the literature procedures. Analytical data for those amines are in accordance with those previously reported. The reagents IBX 29 and 4,4-dibromo-2,6-ditert-butyl-cyclohexa-2,5-dienone (20) 30 were synthesized according to reported procedures, and their properties matched those previously reported.
General Experimental Procedure for the Desymmetrization of 1-Phenylphosphinan-4-ones by Enantioselective Enolization. 14 The synthesis of silyl enol ethers 2b and 2c (the external quench procedure (EQ)) is as follows. In a flame-dried Schlenk tube (20 mL) equipped with a magnetic stirrer and inert gas inlet, the lithium amide base was formed by addition of n-BuLi (0.19 mL, 1.6 mol/dm 3 solution in hexanes; 0.49 mL, 0.15 mmol) to a solution of the chiral secondary amine (0.3 mmol) and LiCl (2.1 mg, 0.05 mmol) in THF (4 mL) under nitrogen at −78°C (dry ice/acetone bath). After 5 min, the solution was allowed to warm to room temperature and then recooled to −90°C (methanol/liquid nitrogen bath) before addition of a solution of 1phenylphosphinan-4-one 1-sulfide (1c) or 1-phenylphosphinan-4-one 1-borane (1b) (0.1 mmol) in THF (1 mL). After 30 min, Me 3 SiCl (0.063 mL, 0.5 mmol) was added to the reaction mixture, which was then stirred at −90°C for further 45 min. After that time, the solution was allowed to warm to room temperature and the solvent was evaporated. The residue was quickly purified on a silica gel column (hexane/THF = 6:1) to give silyl enol ether 2b or 2c as a colorless oils. 2b and 2c are highly susceptible to hydrolysis under extraction and column chromatography conditions, and the reported yields and enantiomeric excesses refer to those determined for crude products. Enantiomeric excess of 2b and 2c was determined by HPLC analysis on a Chiralcel AS-H column using hexane/i-PrOH (90/10).
General Procedure for the Organocatalytic α-Halogenation of 1-Phenylphosphinan-4-ones. In a flame-dried Schlenk tube (10 mL) equipped with a magnetic stirrer, the halogenating agent ((NBS, 20, NCS, or PhICl 2 ) (0.15 mmol)) was added to a mixture of phosphinanone 1a or 1c (21 or 22 mg, respectively, 0.1 mmol), PhCOOH (2.4 mg, 0.02 mmol), and organocatalyst (0.02 mmol) in DCM (2 mL) at 0°C (ice/water bath), and the reaction mixture was allowed to warm to room temperature and stirred for further 16 or 24 h at that temperature. Then, in chlorination reactions, DBU (22.8 mg, 0.15 mmol) was added to effect elimination of HCl from the intermediate chloro ketone 3-Cl, and the reaction mixture was stirred at room temperature for 1 h. In bromination reactions, the reaction mixture was warmed up to 60°C (heating mantle) for 30 min to complete quantitative elimination of HBr from the intermediate bromo ketone 3. Then, evaporation of the reaction mixture gave crude enone 4a or 4c. The crude products could be purified on a silica gel column using either DCM/THF = 10:1 for enone 4a or hexane/THF = 8:1 for enone 4c to give the pure products as colorless oils. Yields of 4a and 4c were determined by GC−MS analysis and confirmed by 31 P NMR spectroscopy. Enantiomeric excess was determined by HPLC analysis using CSP.
One-Pot Procedure for Direct Synthesis of Phosphin-2-en-4one 4c from Phosphinanone 1c. In a flame-dried Schlenk tube (400 mL) equipped with a magnetic stirrer and inert gas inlet, the lithium amide base was formed by addition of n-BuLi (1.6 mol/dm 3 solution in hexanes; 4.6 mL, 7.37 mmol) to a solution of (−)-bis[(S)-1phenylethyl]amine (S,S-5) (3.38 mL, 14.73 mmol, 3 equiv.) and LiCl (104 mg, 2.46 mmol, 0.5 equiv.) in THF (200 mL) under nitrogen at −78°C (dry ice/acetone bath). After 5 min, the solution was allowed to warm to room temperature and then recooled to −90°C before addition of a solution of 1c (1.1 g, 5 mmol) in THF (20 mL). After 30 min, Me 3 SiCl (3.1 mL, 24.5 mmol, 5 equiv.) was added to the reaction mixture, which was then stirred at −90°C (methanol/liquid nitrogen bath) for further 45 min. After this time, the solution was allowed to warm to room temperature and the solvent was evaporated (during the evaporation, the temperature of the solution should be kept below 25°C ) to give crude silyl enol ether 2c. The silyl enol ether 2c was obtained in 82% yield (determined by 31 P NMR spectroscopy) and with an enantiomeric excess of 76% (determined by chiral HPLC analysis using a Chiralcel AS-H column). Then, following the published oxidation protocol, 18 equimolar amounts of IBX and MPO (2.06 g of IBX and 0.92 g of MPO, 1.5 equiv.) dissolved in DMSO (5 mL) were added in one portion at room temperature to the crude vacuum-dried silyl enol ether 1c dissolved in 3 mL of DMSO. The solution was stirred vigorously for 2 h at room temperature. After this time, the reaction mixture was diluted with aqueous HCl (5%) and extracted with DCM (five times). The combined organic phase was dried (MgSO 4 ), and the solvent was removed in vacuum to afford the crude product, which was further purified by silica gel column chromatography (hexane/THF = 6:1) to give enone 4c as a light yellow oil in 48% overall yield (two steps) (0.52 g, 2.4 mmol) and with 73% ee (determined by HPLC analysis using a Chiralcel OJ-H column). Repeated recrystallizations (three times) of (−)-4c (73% ee) from a hexane/i-PrOH mixture allowed to increase its enantiopurity of the levorotatory enantiomer of 4c left in the mother liquor up to 96% ee.
Catalytic Desymmetrizing Dehydrogenation of Phenylphosphin-2-en-4-ones through Enamine Oxidation. Reactions were performed according to the literature procedure 20 at room temperature. To a 10 mL flask were added phenylphosphinan-4-one 1a−c (0.041−0.045 g, 0.2 mmol), catalyst (20 mol %, 0.04 mmol), pentanedioic acid (7.3 mg, 30 mol %, 0.06 mmol), and diethyl ether (0.1 mL). The reaction system was gently stirred for half an hour. Then IBX (56 mg, 0.2 mmol) was added followed by 0.1 mL of diethyl ether. After 48 h, the reaction system was diluted with ether and immediately passed through a thin layer of silica gel. The remaining organic phase was concentrated in vacuum. Yield was determined by 31 P NMR analysis, and enantiomeric excess was determined by CSP-HPLC analysis.