Highly Efficient Camphor-Derived Oxaziridines for the Asymmetric Oxidation of Sulfides to Chiral Sulfoxides Vassilios Meladinis, Uwe Verfürth, and Rudolf Herrmann* Organisch-Chemisches Institut der Technischen Universität München, Lichtenbergstraße 4, D-8046 Garching, Bundesrepublik Deutschland Dedicated to Prof. Dr. Ivar Ugi on the occasion o f his 60th birthday Z. Naturforsch. 45b, 1689- 1694 (1990); received May 18, 1990 Asymmetric Oxidation, N-Sulfonyl-oxaziridines, Chiral Sulfoxides, Camphorsulfonic Acid

Chiral N-sulfonyl-oxaziridines derived from 8 -camphorsulfonic acid and fenchone have been evaluated as asymmetric oxidizing agents for the conversion of sulfides to chiral sulf­ oxides. There is no correlation between the redox potentials nor the lvO NMR chemical shifts of the oxaziridines and their relative oxidation rates, nor with the enantiomeric excesses achieved, indicating that steric effects are responsible for their behaviour. The results are con­ sistent with an attack of one sulfur lone pair at the oxaziridine oxygen in such a way that both sulfur lone pairs lie in the plane of the oxaziridine ring. The most efficient oxaziridines, the camphorlactone-sulfonyloxaziridine [(4aS,9aR)-10,10-dimethyl-6,7-dihydro-4H-4a,7-meth- ano-oxazirino[3,2-j]oxepino[3,4-c]isothiazol-9(5 H)-one 3,3-dioxide] and the 3-endo-bromo- camphorsulfonyloxaziridine [(4aS,8 S ,8 aR)-8-bromo-9,9-dimethyl-5,6,7,8-tetrahydro-4 H- 4a,7-methano-oxazirino-2,l-benzisothiazole 3,3-dioxide] allow the preparation of chiral sul­ foxides with up to 85% enantiomeric excess.

Introduction crowded oxaziridines activated by an electron- Chiral sulfoxides play a prominent role among withdrawing sulfonyl group at nitrogen give the the chiral auxiliaries used for the synthesis of enan- best results. This concept is ideally realized in the tiomerically pure compounds, such as natural bicyclic system of camphor, which is easily deriva- products [1], and remarkable success has been tized to form many useful chiral auxiliaries [12]. achieved [2], particularly for terpenes. The classi­ The main advantage is that most reactions at the cal Andersen synthesis of such chiral sulfoxides camphor skeleton are highly stereoselective, and from diastereomerically pure sulfinates with Gri- only few procedures involve the separation of dia- gnard or lithium reagents [3] can be applied only to stereoisomers. We have prepared some new oxa­ a limited number of sulfoxides, and thus, the oxi­ ziridines derived from 8-camphorsulfonic acid dation of sulfides to chiral sulfoxides, as the most which are modified in the position 3 of the cam­ direct access, remains a challenging task. Among phor molecule, and wish to report here on the the methods available, the microbiological proce­ results in the oxidation of sulfides to chiral sulf­ dures [4] allow the preparation of almost enan- oxides. tiomerically pure sulfoxides of special types. The more general chemical methods include catalytic Results and Discussion oxidations, using bovine serum albumin or fla­ vines [5], and Kagan’s modification of the Sharp- We have reported on the synthesis of the cam­ less oxidation (titanium/tartrate catalyst) [6] seems phor-derived oxaziridines 2 [9], 3 [10], 4, 5, 6, 7, to be well suited for most purposes. Non-catalytic and 8 [13], included in this study. The preparation processes are of interest only where they apply ex­ of the fenchone-derived oxaziridine 9 will be re­ ceedingly mild reaction conditions. This is the case ported elsewhere. The oxidations were performed with chiral oxaziridines [7], A variety of such oxi­ in CC14 at standard concentrations (0.05 mol/1 for dants has already been prepared for this purpose both oxaziridine and sulfide), as this is the solvent [8-11], and it was found that rigid and sterically which gives the highest enantiomeric excesses in almost all cases [9, 11]. The oxaziridines are not * Reprint requests to Dr. R. Herrmann. always completely soluble, which means that long­ Verlag der Zeitschrift für Naturforschung, D-7400 Tübingen er reaction times are required, especially in the case 0932 - 0776/90/1200 -1689/$ 01.00/0 of the camphorlactone-sulfonyloxaziridine 3. 1690 V. M eladinis et al. ■ Highly Efficient Camphor-Derived Oxaziridines

parameters, to see if it might be possible to predict the reactivity and selectivity of new oxaziridines. We therefore measured the redox potentials and the pO NMR spectra of some oxaziridines. These data are listed in Table I. As no l70 NMR data have been reported for oxaziridines up to now, we have also included the racemic oxaziridines 10, 11, and 12, for comparison. It is interesting to note that all oxaziridines give signals between -18 and -20 ppm, a very narrow range compared with the large differences ob­ served between compounds of the same type, e.g. methanol (-37 ppm) and ethanol ( + 6 ppm). It seems that the geometry of the three-membered ring mainly governs the chemical shifts, and not the influence of the electron donor or acceptor ca­ pacity of the substituents. It is very striking that the strongly electron-withdrawing sulfonyl substi­ tuents at nitrogen do not have significant effects, nor does the structure of the carbon substituents. The linewidths (50- 160 Hz) are normal for the ob­ servation temperature (37 C). This narrow range of the chemical shifts excludes reliable correlations with reaction rates or enantioselectivity. The l70 chemical shifts of other oxygens in the molecules lie in the range expected for this functional groups [20 ]. As judged from the redox potentials, the elec- First, we were interested in correlations of reac- tron acceptor capacity of the oxaziridines increas- tion rates and enantioselectivities of the oxidation es in the direction 3<6< 1<2<9<7<8<4. of sulfides to chiral sulfoxides with other physical The reduction waves of all oxaziridines are irrever-

Table I. Redox properties 3 and l70 NMR data of some oxaziridines.

Compound Eped (V ) ö (l70 ) (in acetone-d6)b Remarks C-N-O others

1 -2.30 -18.8 (50 Hz) 175.3, 178.3 (S02) [8] 2 -2 .2 6 -- [9] 3 -2.40 -18.1 (160 Hz) 158.8 (-0-), 180.9, 183.0 (SO,), [ 1 0 ] 538.8 (C = 0 ) 4 -1 .9 7 - - [13] 6 -2 .3 6 - - [13] 7 -2 .1 7 -- [13] 8 -1 .9 9 —— [13]; very slow oxidation of sulfides 9 -2 .2 5 —— very slow oxidation of sulfides 10 - -1 8 .4 (6 0 Hz) 138.2, 144.0 (SO,) [14]; racemic 11 - -19.9(120 Hz) NO^ overlaps with acetone [15]; racemic 12 - -19.5 (130 Hz) NO: overlaps with acetone [15]; racemic

a Peak potential of the irreversible cathodic wave, in 0.1 M NBu4BF4 in tetrahydrofuran, relative to internal FcH/ FcH+; b chemical shifts relative to internal (Me-d 3)-,CpO (S = 569 ppm); linewidth in brackets. V. M eladinis et al. ■ Highly Efficient Camphor-Derived Oxaziridines 1691 sible, which points to a chemical reaction occur­ oxaziridine has to be tested individually for its effi­ ring after the initial one-electron reduction. It is ciency as chiral oxidation reagent, as no predic­ not obvious what this reaction might be. The time tions can be made from other data. It seems that necessary for complete conversion of the sulfide this efficiency is mainly determined by steric (Table II) is used as a rough measure for the reac­ effects. tion rates; the low solubility of some oxaziridines The differing steric requirements of the substi­ does not allow to determine exact rates. For the tuents in the position 3 of the camphor skeleton oxidation of methyl phenyl sulfide, the reactivity lead to a different behaviour of the oxaziridines in increases in the direction 7 — 6<3<4<5<§2. 8 the oxidation of sulfides. This is illustrated for the and 9 react very slowly with sulfides, and only very oxidation of methyl phenyl sulfide in Schemes low yields of sulfoxide were obtained with these 2 -4 . oxaziridines. They are therefore not included in For the oxidation of sulfides by oxaziridines, it Table II. is generally agreed that the approach of the sulfur As for the enantioselectivities of the same oxida­ to be oxidized to the oxygen of the oxaziridine tion, we have found an increasing efficiency in the occurs in such a way that both lone pairs of sulfur direction 1 — 5< 6< 7< 2 — 4<3. No correla­ lie in the plane of the three-membered ring. It then tion with the reaction rates nor the redox poten­ depends on the space available on the left or right tials could be found, which means that any new side of the oxaziridine group how the larger or

Table II. Oxidation of sulfides R '-S -R 2 to chiral sulfoxides with oxaziridines11.

Entry Oxaziridine R 1 R2 Time [h] Yield [%] Mb2b e.e. Configuration Ref. [%]

1 2 Ph Me 1 80 - 8 6 59 S [9, 16] 2 3 Ph Me 6 8 85 -1 1 5 79 S 3 4 Ph Me 48 55 -8 4 57 S 4 5 Ph Me 24 65 + 4.0 3 R 5 6 Ph Me 72 60 + 14 9 R 6 7 Ph Me 72 63 -3 6 24 S 7 2 4-M e-P h Me 2 0 79 -7 9 56 S [9, 17] 8 3 4-M e-P h Me 70 90 - 1 0 2 72 s 9 2 4-M e-P h Et 3 8 6 -1 1 7 61 s [9, 17] 10 3 4-M e-P h Et 72 95 -1 4 5 76 s 11 4 4-M e-P h Et 48 49 -1 6 2 85 s 12 2 4-Me-Ph Ph 24 8 8 -7 .0 27 s [9, 17] 13 3 4-M e-P h Ph 72 85 -9 .0 35 s 14 3 2-M e-P h Ph 72 8 8 -8 0 ?° 9 - 15 2 cyclohexyl Ph 3 79 - 9 7 d 49 9 [9, 18] 16 3 cyclohexyl Ph 72 95 - 132d 67 9 17 2 P h -C H , Et 2 0 79 + 17 32 R [9, 19] 18 3 Ph-CH^ Et 70 80 + 24 45 R 19 4 P h -C H ' Et 48 51 + 25 47 R 2 0 2 Ph-CH^ Ph 3 6 8 — 9 1 e 36 s [4, 9] 21 4 P h -C H i Ph 48 45 94° 37 s ?c 2 2 2 2,4,6-Me,Ph Ph 48 65 197 9 [9] 23 3 2,4,6-Me,Ph Ph 160 70 -1 9 9 ?c 9 24 2 C H ,=C H Ph 4 75 — 12 0 e 39 s [6 . 9] 25 3 c h ;= c h Ph 6 6 65 - 142e 46 s 26 2 c h ;= c h c h , Ph 3 78 -107 61 s [4. 9] 27 3 c h > c h c h 2 Ph 62 90 - 1 2 2 70 s 28 3 »-octyl Pr" 72 90 + 9.0 9c 9 - 29 4 «-octyl Pr« 48 43 + 15 9c ? 9 30 2 EtO.CCH, Bu" 2 0 71 - 8 . 2 39 [9] 31 3 EtOXCH; Bu" 1 2 0 99 -1 6 78 9

a Standard conditions: concentration 0.05 mol/1 of oxaziridine and sulfide in CC1 4 at r.t.; b c 1, ethanol, if not other­ wise specified;c e.e. could not be determined with all chiral shift reagents tested; d c 1, acetonitrile;e c 1, acetone. 1692 V. M eladinis et al. • Highly Efficient Camphor-Derived Oxaziridines V v

• \vs Scheme 2. Selectivity in the oxida­ tion of methyl phenyl sulfide by oxa­ ziridines 1, 4, and 5.

Scheme 3. Selectivity in the oxi­ dation of methyl phenyl sulfide by oxaziridines 6 and 7. V

w w s s Scheme 4. Selectivity in the oxi­ " i dation of methyl phenyl sulfide " i by oxaziridines 2 and 3. smaller group of the sulfide is orientated. The en- reacts with the oxygen [21]. However, it should be antioselectivity of the oxidation is due to non- kept in mind that there is a strong solvent depend- bonded interactions which have to be minimized in ence in the selectivity of such oxidations,, and even the transition state, and therefore it is the “lower” the direction of the enantioselectivity may be re­ electron pair (as depicted in Schemes 2-4) which versed [9, 21], which suggests the participation of V. M eladinis et al. • Highly Efficient Camphor-Derived Oxaziridines 1693

solvent molecules in the transition state. Restrict­ gen; from an X-ray structure of 3 [10], this distance ing the discussion to solvents with low polarity, is calculated to be only 2.86 Ä (Scheme 4). Thus, such as CC14 in our case, will probably render such the space available on the left side of the three- a simple picture more valid. membered ring in 3 is dramatically reduced, and Let us consider first oxaziridine 4 (Scheme 2). high enantiomeric excesses of (S)-sulfoxides can be The very voluminous bromine in the endo -position expected. Indeed, 3 gives the highest e.e. of methyl reduces dramatically the space available on the left phenyl sulfoxide (79%, entry 2). In some cases, the side of the oxaziridine, and an efficient orientation results with the tWo-bromo-oxaziridine 4 are of the large group (phenyl) to the right side can be slightly better (entries 11, 19, and 29 vs. 10, 18, and expected, leading to a high e.e. of (S)-methyl phe­ 28), which means that both oxaziridines can be nyl sulfoxide (entry 3 in Table II). On the other considered as valuable new oxidizing agents for hand, an e.ro-orientation of the bromine, as in 5, sulfides, and to obtain optimum results, both should have almost no effect on the space available should be compared. The highest e.e. was observed in the vicinity of the oxaziridine, and 5 should in the case of the ethyl /?-tolyl sulfoxide with 4 therefore behave in the same way as the unsubsti­ (85%, entry 11). As for the preference of (R)-ben- tuted 1. This is indeed true; both oxaziridines give zyl ethyl sulfoxide (entries 17-19), one can imag­ low enantioselectivities, with a slight preference ine that the introduction of flexibility in the sulfide for the (R)-sulfoxide (entry 4). The low e.e. indi­ by the CH2 group between phenyl and sulfur will cates that the difference in the space available on allow the planar phenyl group to turn away from both sides is quite small. The oxaziridine 6 with the the camphor skeleton, and thus the ethyl group rigid and almost planar dioxolane ring has quite a may appear bulkier to the oxaziridine and stay on similar geometry as 1, and only the oxygen lone the right side, which leads to the (R)-sulfoxide pairs occupy somewhat more space than the hy­ after oxidation. drogens of 1 (Scheme 3). The hydrogens of the The new oxaziridines are compatible with many methylene bridge of the dioxolane point down­ other functional groups, e.g. carbon-carbon dou­ wards and do therefore not interfere with the ap­ ble bonds (entries 24-27) and ester groups (entries proach of sulfides. The result (a small preference 30-31). Growing applications of these and other for the (R)-sulfoxide, entry 5) indicates that the chiral oxaziridines in the synthesis of sulfoxides as space available is still greater on the left side of the intermediates for naturally occurring products can oxaziridine ring. The situation is completely differ­ therefore be expected. We will report on further ent in the case of oxaziridine 7 containing the di- camphor- and fenchone-derived oxaziridines in oxane ring. It has a high conformational flexibili­ due course. ty, and it has been found by measuring the nuclear Overhauser enhancements that the preferred con­ Experimental formation of this ring is chair-like [13], As shown NMR spectra were obtained on a Bruker AM in Scheme 2, this implies that the trimethylene 360 instrument at 48.82 MHz for nO at 37 °C, in bridge is bent forward and thus is able to block the d6-acetone which was also used as internal refer­ left side of the oxaziridine ring to a high extent. ence ((5 = 569 ppm). Redox potentials were deter­ A considerable preference for the (S)-sulfoxide mined by cyclovoltammetry with a EG & PAR 173 results (entry 6). potentiostat/galvanostat and a EG & PAR 175 The exchange of the sp3 carbon in the position 3 universal programmer, in a solution of 0.1 mol/1 of oxaziridine 1 with a sp2 carbon to form 2 brings NB u 4BF4 in tetrahydrofuran, relative to the inter­ the carbonyl oxygen in a position where its lone nal ferrocene/ferrocenium redox couple (Ep/2 = pairs reduce the space on the left side of the oxazi­ 0 V). Optical rotations were measured with a Roussel Jouan Digital 71 polarimeter. The ridine ring; it is therefore not surprising that the syntheses and properties of the chiral oxaziridines (S)-sulfoxide is formed with good enantioselectivi- derived from camphorsulfonic acid have been de­ ty (entry 1). The enlargement of the five-mem- scribed earlier [8-10, 13], as well as the racemic bered ring which contains the carbonyl group by oxaziridines 10, 11, and 12 [14, 15], The synthesis an additional oxygen (oxaziridine 3) brings the of the fenchone-derived oxaziridine 9 will be carbonyl oxygen very close to the oxaziridine oxy­ reported elsewhere. 1694 V. M eladinis et al. ■ Highly Efficient Camphor-Derived Oxaziridines

Asymmetrie oxidation o f sulfides to chiral tiomeric excesses of the sulfoxides are determined sulfoxides by polarimetry. In the cases marked with ? in To a solution of the sulfide (5 mmol) in CC14 Table II, the optical rotation of enantiomerically (100 ml) is added the oxaziridine (5.5 mmol), and pure sulfoxides is not known, and could not be the mixture is stirred at room temperature for the determined by NMR with lanthanide shift re­ time indicated in Table II, with exclusion of mois­ agents. ture. With oxaziridine 3, it is sufficient in almost all cases to filter off the precipitated imine and We wish to thank Prof. Dr. Ivar Ugi, TU evaporate the solvent to obtain pure sulfoxides München, for supporting this work. Prof. Dr. Fer­ (less than 2% impurities detected by 360 MHz 'H nanda N. N. Carvalho, Lisboa, Portugal, for her NM R), with only few exceptions (entries 18, 23, help in the electrochemical measurements, and 25, and 31 in Table II). In these cases and with all Dipl.-Chem. Manuela Klein, TU München, for other oxaziridines, the crude sulfoxide is purified helpful collaboration. Financial support from by preparative thin layer chromatography (com­ Deutsche Forschungsgemeinschaft is gratefully mercial plates (Merck), diethyl ether). The enan­ acknowledged.

[1] G. Solladie, Chem. Scr. 25, 149 (1985); G. H. Billmers, J. Org. Chem. 48, 2672 (1983); F. A. Posner, Acc. Chem. Res. 20, 72 (1987); K. Krohn, Davis. J. P. McCauley (Jr.). and M. E. Harakal, J. Nachr. Chem. Tech. Lab. 35, 22 (1987); G. Solladie. Org. Chem. 49, 1465(1984). Pure Appl. Chem. 60, 1699 (1988); G. H. Posner. in [9] G. Glahsl and R. Herrmann, J. Chem. Soc. Perkin S. Patai, Z. Rappoport, and C. Stirling (eds): The Trans. 1 1988, 1753. Chemistry of Sulphones and Sulphoxides, p. 233, [10] V. Meladinis, R. Herrmann, O. Steigelmann, and G. John Wiley & Sons, Chichester (1988). Müller, Z. Naturforsch. 44b, 1453 0989). [2] C. Iwata, M. Fujita, Y. Moritani, K. Hattori, and T. [11] F. A. Davis, M. C. Weismiller, G. S. Lai, B. C. Imanishi, Tetrahedron Lett. 28, 3135 (1987); S. G. Chen, and R. Przeslawski, Tetrahedron Lett. 30, Pyne, Tetrahedron Lett. 28, 4737 (1987); T. Satoh. 1613 (1989); F. A. Davis, R. Thimma Reddy, and T. Oohara, Y. Ueda, and K. Yamakawa, Tetra­ M. C. Weismiller, J. Am. Chem. Soc. Ill, 5964 hedron Lett. 29, 313 (1988); R. A. Holton, R. M. (1989). Kennedy, H.-B. Kim, and M. E. Krafft, J. Am. [12] W. Oppolzer, Tetrahedron 43, 1969 (1987). Chem. Soc. 109, 1597 (1987); P. Bravo, G. Resnati, [13] U. Verfürth and R. Herrmann. J. Chem. Soc. Perkin F. Viani, and A. Arnone, Tetrahedron 43, 4635 Trans. I, in press. (1987). [14] M. Klein, Dissertation, Technische Universität [3] K. K. Andersen, Tetrahedron Lett. 3, 9? (1962); J. München (1990). Drabowicz and M. Mikolajczik, J. Ore. Chem. 47, [15] J. Bjorgo, D. R. Boyd, R. M. Campbell, N. J. 3325 (1982). Thompson, and W. B. Jennings, J. Chem. Soc. [4] D. R. Light, D. J. Waxman, and C. Walsh. Bio­ Perkin Trans. II 1976, 606. chemistry 21, 2490 (1982); D. J. Waxman, D. R. [16] U. Folli. D. Iarossi, F. Montanari, and G. Torre, J. Light, and C. Walsh, Biochemistry 21, 2499 (1982); Chem. Soc. C 1968, 1317. B. J. Auret, D. R. Boyd, H. B. Henbest, and S. [17] K. Mislow, M. M. Green, P. Laur, J. T. Melillo, T. Ross, J. Chem. Soc. C 1968, 2371; H. Ohta, Y. Oka- Simmons, and A. L. Ternay (Jr.), J. Am. Chem. Soc. moto, and G.-I. Tsuchihashi, Chem. Lett. 1984, 205; 87, 1958 (1965); A. L. Cope and E. A. Caress, J. H. L. Holland. H. Popperl, R. W. Ninnis, and P. C. Am. Chem. Soc. 8 8 , 1711 (1966). Chenchaiah, Can. J. Chem. 63, 1118 (1985); B. J. [18] T. Komori and T. Nonaka. J. Am. Chem. Soc. 106, Auret, D. R. Boyd. R. Dunlop, and A. F. Drake, J. 2656(1984). Chem. Soc. Perkin Trans. 1 1988, 2827. [19] K. Mislow. M. M. Green, and M. Raban, J. Am. [5] T. Sugimoto, T. Kokubo, J. Miyazaki, S. Tanimoto, Chem. Soc. 87, 2761 (1965). and M. Okano, J. Chem. Soc. Chem. Commun. [20] W. McFarlane and H. C. E. McFarlane, in J. Mason 1979, 402; S. Shinkai, T. Yamaguchi, O. Manabe. (ed.): Multinuelear N M R , p. 403, Plenum Press, and F. Toda. J. Chem. Soc. Chem. Commun. 1988. New York (1987); E. Block. A. A. Bazzi. J. B. Lam­ 1399. bert. S. M. Wharry, K. K. Andersen. D. C. Dittmer. [6 ] P. Pitchen and H. B. Kagan. Tetrahedron Lett. 25, B. H. Patwardhan, and D. J. H. Smith. J. Org. 1049 (1984); E. Dunach and H. B. Kagan. Nouv. J. Chem. 45, 4810 (1980);.!. C. Dyer, D. L. Harris, and Chim. 9, 1 (1985); S. H. Zhao. O. Samuel, and H. B. S. A. Evans (Jr.), J. Org. Chem. 47, 3660 (1982). Kagan. C.R. Acad. Sei. (Paris) 304, 273 (1987). [21] F. A. Davis, J. P. McCauley (Jr.), S. Chattopa- [7] F. A. Davis and A. C. Sheppard. Tetrahedron 45, dhyay, M. E. Harakal, J. C. Towson. W. H. Wat­ 5703(1989). son. and I. Tavanaiepour, J. Am. Chem. Soc. 109, [8 ] F. A. Davis, R. H. Jenkins (Jr.). S. B. Awad. O. D. 3370 (1987); F. A. Davis, J. C. Towson, M. C. Stringer. W. H. Watson, and J. Galloy, J. Am. Weismiller. S. Lai. and P. J. Carrol. J. Am. Chem. Chem. Soc. 104, 5412 (1982); F. A. Davis and J. M. Soc. 110, 8477(1988).