number 123 Contribution New Mitsunobu Reagents Tetsuto Tsunoda, Hiroto Kaku and Shô Itô Tokushima Bunri University, Pharmaceutical Sciences 1. Introduction Organic chemists have enjoyed these advantages of the Mitsunobu reaction in organic synthesis. However, the The Mitsunobu reaction is a well-established funda- reaction has a serious limitation (the so-called “the mental reaction and has been applied widely in organic restriction of pKa”); the acidic hydrogen in HA has to have synthesis. In the Mitsunobu reaction, a unique dehydration a pKa of less than 11 for the reaction to proceed satisfacto- occurs between alcohols and various Brønsted-Lowry rily. If HA has a pKa higher than 11, the yield of RA is acids (HA) utilizing a combination of diethyl considerably lower, and with HA having a pKa higher than azodicarboxylate (DEAD) - triphenylphosphine (TPP) 13, the desired reaction does not occur (for example: 1, 2, 4) (Scheme 1).1,2) Scheme 3). In order to overcome “the restriction of pKa”, we have developed new Mitsunobu reagents and applied them to organic synthesis.5) In this article, we would O O O O like to describe the results. EtO C NNCOEt EtO C N N C OEt H H 2 ROH + HA RA (i) pKa < 11 PPh O PPh O O 3 3 DEAD-TPP ex. 1 EtOH + HN Et N Nucleophiles (HA) O 91% O Oxygen Nitrogen Carbon Sulfur (pKa = 8.3) COOH O SH (ii) 11 < pKa < 13 H CN HN DEAD-TPP H CN TsNMe Me NTs O Me OH + H 51% Me OH (pK = 11.7) HN3 a (iii) pKa > 13 TsNMe H O O HO H OEt DEAD-TPP OEt TfNMe + OH H O O H OEt OEt O 0% O (pKa = 13.3) Scheme 1. Scheme 3. Without any prerequisite activation of the alcohol, this redox condensation reaction proceeds under mild conditions with complete Walden inversion of stereochem- istry (for example: Scheme 2),3) while DEAD is reduced to 2. Development of New Mitsunobu Reagents dihydro-DEAD (2) and TPP is oxidized totriphenylphosphine oxide (1) (Scheme 1). 2.1. New Azo-type Reagents To develop improved redox system, we considered the mechanism of the Mitsunobu reaction and its side OH TfNMe reaction. The desired Mitsunobu reaction proceeds TfNHMe probably through the generally accepted path a shown in OBz OBz DEAD-TPP 2) 87% Scheme 4. Scheme 2. number 123 O O EtO C N N C OEt + PPh3 O O desired reaction side reaction EtO N N OEt (path a) (path b) O O Ph3P 3 O O EtO N N OEt EtO N N OEt H ROH H 4 4 H A RO PPh3 H A RO PPh3 5 5 O O HA + EtO C N N C OEt + 2 RA + 12+ H R 6 Scheme 4. On the contrary, in the case of the reaction of less acidic of the reaction cyclized intramolecularly as shown in HA, the hydrazo anion 4 attacks the alkoxyphosphonium 5 Scheme 6.8) Thus, cyclic 1,6-dimethyl-1,5,7-hexahydro- directly to afford alkylated the hydrazine derivative 6 as a 1,4,6,7-tetrazocin-2,5-dione (DHTD)10) was also designed by-product (Scheme 4, path b),3) since the anion 4 is not to prevent the cyclization of acyclic azodicarboxamides to efficient in deprotonating the weakly acidic HA. In order to 9. overcome these drawbacks, “the restriction of pKa”, and expand the versatility of the original Mitsunobu reaction, O O R N new reagents which can be protonated by the less acidic O O R2N NR2 2 NR2 HA have been developed to replace the DEAD-TPP sys- R2N N N NR2 O N N O N N tem. PBu3 PBu3 PBu3 One way to improve the redox system would be to 8 O enhance the basicity of the anion 4 by replacement of the R2N NR2 +Bu3PO alkoxy group OEt in DEAD with a strong electron-donating NN group such as NR in a new anion 7 (Scheme 5).6,7) 2 9 Furthermore, it was also considered that the bulkiness of Scheme 6. the alkyl substituents on the NR2 group influenced the reactivity of new azo-type reagents. Thus, N,N,N',N'- tetraisopropylazodicarboxamide (TIPA),8) 1,1'- (azodicarbonyl)dipiperidine (ADDP),7) N,N,N',N'- 2.2. New Phosphorane-type Reagents tetramethylazodicarboxamide (TMAD) 8,9) have been developed as the new reagents. When the azo-type reagents were developed, it was found that maleic and fumaric acid derivatives, which were δ − O O O O identified as carbon analogs of DEAD and/or TMAD, R' mediated the condensation of benzyl alcohol with tosyl EtO NN OEt N NN NR'2 H R' δ+ H amide 10 (ex. Scheme 7).6) 4 7 O Me OO OO OO Me Me N N NCNNCN NCNNCN NCNNCN O O N N Me Me MeOC C OMe - PBu O Me H H 3 TIPA ADDP TMAD DHTD (1.5 equiv.) TsNMe Ph OH + TsNMe H 100 °C, 24 h Scheme 5. Ph 10 (pKa = 11.7) Ar, dry PhH in sealed tube (1.5 equiv.) 62% Those new azo compounds were combined with a more Scheme 7. nucleophilic phosphine than TPP, such as tributylphosphine (TBP), because of the lower reactivity of the azodicarboxamides as Michael acceptors, compared with Unfortunately, difficulty in reproducibility of the reac- DEAD. In the course of our study, we found that most of tion and in product isolation forced us to abandon this the new azo compounds and TBP used was consumed investigation. However, consideration of this reaction even in cases where no desired product was obtained. In mechanism revealed that the betaine 11 was formed such cases, a large amount of the oxadiazole 9 was instead of the Mitsunobu intermediate 3 (Scheme 8). 11 obtained probably through a new competitive side reaction might easily convert to 12, whose structure can be pathway, in which the betaine 8 produced in the first step generalized as 13. This is a phosphorus ylide, namely, number 123 phosphorane. After this discovery, it was suspected that The reaction was carried out usually under an anhydrous the Mitsunobu reagent, the combination of an azo com- argon atmosphere at 0 °C to room temperature. When the pound and a phosphine, could be replaced with an ylide. results were not satisfactory, the reaction shown in Scheme 6 may take place as a competitive reaction. In such cases, heating and/or usage of a large amount of the reagents O O O O O was often ineffective. R'O CH C OR' R'O C C OR' R'O C R" H H2 The desired products were purified by column PR3 PR3 PR 3 chromatography. In the traditional Mitsunobu reaction 11 13 12 using DEAD and TPP, one major problem was the X X CH CH laborious purification of the product from dihydro-DEAD and PR3 PR3 triphenylphosphine oxide because of their moderate Ylide Phosphorane polarity and half-crystalline nature. On the other hand, in the reaction of the new azo-type reagents, removal of the Scheme 8. hydrazo-compounds 14-17 (in place of dihydro-DEAD) and tributylphosphine oxide could be easily accomplished by SiO column chromatography because of their high If ylides could mediate the Mitsunobu-type reactions, 2 polarity. Furthermore, since the crystalline 14-17 were the reaction would proceed through the following reaction hardly soluble in many organic solvents, filtration of the pathway illustrated in Scheme 9, which does not take into reaction mixture after the addition of a solvent (such as consideration the problem of acid-base equilibrium. 1) The hexane, ether, and so on) was quite effective to remove alcohol is deprotonated by the ylide, then 2) the resulting them. As an alternative work-up for the reaction of TMAD alcoholate attacks the phosphonium part in the ylide to and DHTD, aqueous treatment of the reaction mixture was afford the alkoxy phosphonium. 3) The X-substituted also quite effective because of the good aqueous solubility methyl anion is protonated by the acidic HA, and then 4) of 16 and 17. The hydrazo-compounds 14-17 could be the resulting conjugate base A- reacts with the alkoxy recycled by reoxidation. phosphonium to give the desired A-R’ along with the phosphine oxide.5,6,11) O O O O XCH 3 NCNNCN NCNNCN AR' HH HH X PR3 A R' O PR3 + OPR3 AH H OR' 14 15 Scheme 9. O Me O O Me Me HN N N C N N C N H H HN N Me Me On the basis of the above working hypothesis, we O Me examined the reaction of several phosphoranes6) and found 16 17 that (cyanomethylene)tributylphosphorane (CMBP)6,11) and Scheme 11. less bulky (cyanomethylene)trimethylphosphorane (CMMP)6,12,13) had sufficient reactivity (Schemes 10). Especially, CMMP gave excellent results. In this article, 3.2. Phosphorane Reagents we describe the C-N, C-C, and C-O bond forming reactions as examples to reveal the features of the new Since CMBP and CMMP are very sensitive to air and reagents. moisture, all procedures for their purification should be carried out under a dry argon atmosphere, even for analy- sis by NMR, IR, and Mass spectra. CMBP was NC PMe3 NC PBu3 purified by distillation under reduced pressure. CMMP was recrystallized from benzene (toluene is unsuitable). When (cyanomethylene)tributylphosphorane (cyanomethylene)trimethylphosphorane CMBP is sealed in an ampule and CMMP is stored in a ( CMMP ) ( CMBP ) screw-top vial with a rubber septum, the reagents could be Scheme 10. kept for months at 10 °C under an argon atmosphere without decomposition.