<strong>Palladium</strong>-<strong>Catalyzed</strong> <strong>Cyclization</strong> <strong>Reactions</strong> <strong>of</strong> <strong>Acetylene</strong>-<strong>Containing</strong> Amino AcidsFULL PAPERSthe literature, several other successful examples <strong>of</strong> this type <strong>of</strong>reaction involving oxygen nucleophiles are known. [6]Analogously, these types <strong>of</strong> Pd-catalyzed coupling/cyclizationreactions can also be carried out with nitrogen nucleophiles.[7] For example, tosylated carbamates, [8] tosylatedamides [9] and sulfonamides [10] are known to react with arylhalides and vinyl triflates to give the corresponding crosscoupledtosylated oxazolidinones, tosylated lactams andsulfonylated piperidines, respectively, all containing an (E)-configured double bond.Inversely, in the group <strong>of</strong> Hiemstra acetylene-substitutedlactams and oxazolidinones (viz. 7) were cyclized under similarconditions to afford the corresponding bicyclic (Z)-substitutedenamides (Scheme 4). [11] A plausible reaction mechanism thatexplains these contradictary results involves the intermediate8, in which palladium(II) coordinates to the carbamate nitrogenatom before the nitrogen-carbon bond is formed. Thisintramolecular activation <strong>of</strong> the triple bond is then followed bynucleophilic attack <strong>of</strong> the nitrogen from within the coordinationsphere <strong>of</strong> palladium to initially give the bicyclic species 9.Reductive elimination <strong>of</strong> Pd(0) finally gives rise to the (Z)-configured olefin 10.OOScheme 4.OONHNPdL Ph8The precedence <strong>of</strong> these literature examples, combined withthe relatively facile biocatalytic access to enantiomericallypure acetylene-containing amino acids developed in our groupin collaboration with DSM (Geleen, The Netherlands) [12]inspired us to use these trifunctional amino acids [13] in similarPd-catalyzed reactions. Logically, these amino acids ± containingboth an oxygen and a nitrogen nucleophile ± might uponsuitable protection give access to a variety <strong>of</strong> enantiomericallypure substituted lactones, as well as to the correspondingnitrogen heterocycles. [14]Results and Discussion7Pd(PPh 3 ) 4 ,TBAC, PhIK 2 CO 3 , MeCNreflux, 3 hIn order to obtain the cyclization precursors, the enantiomericallypure amino acids 11 and 12 [12] were protected withdifferent protecting groups. In case <strong>of</strong> the oxypalladationprecursors, the nitrogen atom was protected according toliterature procedures [15] involving either tosylation or acylationin an aqueous medium or irradiation in a microwave in thepresence <strong>of</strong> phthalic anhydride. [16] In all cases, the N-protectedamino acids 13 ± 16 were obtained in good yields without loss <strong>of</strong>OOOON−Pd(0)Ph10 (60%)NPdL L9PhHHP 1NP( ) nCO 2 H(R)-13: (83%)n = 1, P = Ts, P 1 = H(S)-14: (73%)n = 2, P = Ts, P 1 = H(R)-15: (83%)n = 1, P = Boc, P 1 =H(S)-16: (91%)n = 1, P = P 1 = Phthrac-17: (91%)n = 2, P = P 1 = Phtha or b or cH 2 N( ) nCO 2 H(S/R)-11: n = 1(S/R)-12: n = 2d, eHNPCO 2 Meenantiopurity according to chiral HPLC analysis (Scheme 5).Additionally, rac-12 was converted to racemic phthalimide 17.Conversely, the enantiomerically pure amidopalladationprecursors 18 ± 20 were obtained by esterification using thionylchloride (2 equiv) in MeOH at reflux temperature andsubjection <strong>of</strong> the crude residue to Et 3 N (5 equiv) and p-toluenesulfonyl chloride (TsCl, 2 equiv) or p-nitrobenzenesulfonylchloride (NsCl, 2 equiv) in CH 2 Cl 2 . Alternatively, thecrude esterified residue was dissolved in pyridine and treatedwith TsCl or NsCl to give the same products in similar yields.The palladium-catalyzed cyclizations with the carboxylicacid as the nucleophile are shown in Table 1. Subjection <strong>of</strong> theN-protected amino acids 13 ± 17 to a mixture <strong>of</strong> 10 mol % <strong>of</strong> aPd(II) catalyst and 15 mol % <strong>of</strong> Et 3 N in various solvents atdifferent temperatures led to five- and six-membered lactones.The tosylamide (R)-13 was cyclized in a satisfactory yield <strong>of</strong>66% at 70 8C (entry 1). Changing the catalyst to Pd(OAc) 2 ledto a much faster reaction at room temperature in the same yield(66%, entry 2). A single attempt to use an inorganic base(K 2 CO 3 ) appeared fatal to the reaction (entry 3). Subjection <strong>of</strong>( ) n(S)-18: (94%)n = 1, P = Ts(R)-19: (72%)n = 2, P = Ts(R)-20: (59%)n = 2, P = NsScheme 5. Reagents and conditions: 13, 14 ! 15, 16 (a) TsCl(1.4 equiv), 1 M NaOH (1.1 equiv), rt, 16 h; 13 ! 17 (b) di-tertbutyldicarbonate (1.1 equiv), NaHCO 3 (2 equiv), dioxane/H 2 O,reflux, 4 h; 13, 14 ! 18, 19 (c) phthalic anhydride (1 equiv), H 2 O,microwave (600 W, 5 min); (d) SOCl 2 (2 equiv), MeOH, reflux;(e) TsCl or NsCl (2 equiv), Et 3 N (5 equiv), CH 2 Cl 2 , rt, 16 h.Table 1.( ) n 10% <strong>of</strong> catalyst( ) nP 1 P 1ON CO 2 H base, solvent , TNPP O13-17 21-25T timeentry substrate catalyst base solvent product(°C) (h)1 (R)-13 Pd(MeCN) 2 Cl 2 Et 3 N THF 70 16 (R)-212 (R)-13 Pd(OAc) 2 Et 3 N THF rt 1 (R)-213 (R)-13 Pd(MeCN) 2 Cl 2 K 2 CO 3 DMF 80 64 (R)-15 Pd(OAc) 2 Et 3 N THF rt 2 (R)-225 (S)-16 Pd(OAc) 2 Et 3 N THF rt 1.5 (S)-236 rac-14 Pd(MeCN) 2 Cl 2 Et 3 N MeCN 60 16 rac-247 rac-17 Pd(OAc) 2 Et 3 N THF 60 5 rac-25[a] Isolated yields after column chromatography.[b] ee >97% according to 1 H NMR using Eu(hfc) 3 as the shift reagent in CDCl 3 .yield(%) [a]6666 [b]06342 [b]3224Adv. Synth. Catal. 2002, 344, 70±83 71
FULL PAPERSLarissa B. Wolf et al.the Boc-protected acid 15 to similar conditions resulted atroom temperature in 22 in 63% (entry 4). Likewise, thephthalimide-protected amino acid 16 cyclized in a lower yield.In contrast with the five-membered rings, the six-memberedlactones 24 and 25 were obtained in somewhat disappointingyields (entries 6 and 7), which is probably due to the lowertendency <strong>of</strong> Pd to react in a 6-exo-fashion. [4d,6f] The ee <strong>of</strong> 21 and23 was determined via 1 H NMR experiments using a chiralshift reagent [Eu(hfc) 3 in CDCl 3 ] showing an ee higher than97%, which led us to conclude that virtually no racemizationoccurred in these cyclization reactions.In addition to these reactions, other cyclization reactionswere carried out in the presence <strong>of</strong> various aryl halides or avinyl triflate to introduce an organic substituent in thecyclization step, which are summarized in Table 2. The generalreaction conditions involved 10 mol % <strong>of</strong> a Pd(0) catalyst,Et 3 N (5 equiv), tetrabutylammonium chloride (TBAC,2 equiv) and the aryl halide (2 equiv) in MeCN at 608C.When tosylamide (R)-13 was reacted under these conditionsneither cyclization nor introduction <strong>of</strong> the aryl group wasobserved. Although the starting material had disappeared,amidopalladation could be excluded on the basis <strong>of</strong> NMRanalysis. More gratifyingly, the Boc-protected amino acid (R)-15 reacted under similar conditions to the cross-coupledlactone (R)-27 in a moderate yield <strong>of</strong> 44%, without detectableracemization. The phthalimide-protected amino acid (S)-16was treated with iodobenzene and p-iodoanisole to afford thelactams 28 and 29, respectively, in similar yields (entries 3 ± 6).Unfortunately, in all these cases complete racemizationoccurred at the temperature required for cyclization. Apparently,the phthalimide protecting group renders the a-proton<strong>of</strong> the amino acid sufficiently acidic to be prone to racemizationunder the slightly basic reaction conditions. No cyclizationoccurred with p-nitroiodobenzene (entry 7), which was probablydue to the decreased reactivity <strong>of</strong> the organopalladiumintermediate. Treatment <strong>of</strong> (S)-16 with vinyl triflate 26 [17]afforded 30, albeit in only 23% (entry 8). Furthermore, inanalogy with the aforementioned oxypalladation reactions,Table 2.R10% <strong>of</strong> catalyst( ) nRX, Et 3 N( )nP 1P ON CO 2 HTBAC1NMeCN, 60 °CPP O13, 15-17 27-31entry123456789substrate(R)-13(R)-15(S)-16(S)-16(S)-16(S)-16(S)-16(S)-16rac-17catalystRXtime (h)Pd(PPh 3 ) 4 PhI 2Pd(OAc) 2 /2PPh 3 PhI 5Pd(OAc) 2 /2PPh 3 PhI 6Pd(PPh 3 ) 4 PhI 5Pd(OAc) 2 /2PPh 3 PhI 5Pd(PPh 3 ) 2 Cl 2 PhI 5Pd(PPh 3 ) 4 p-MeOC 6 H 4 I 24Pd(PPh 3 ) 4 p-NO 2 C 6 H 4 I 16Pd(PPh 3 ) 4 t Bu OTf1626product(R)-2728282829044 [b]43411841023[a] Isolated yields after column chromatography.[b] The geometry <strong>of</strong> the (E)-double bond was proven via 1 H NMR NOE experiments.3031yield(%) [a]31formation <strong>of</strong> a six-membered lactone appeared difficult asindicated by the relatively lowyield <strong>of</strong> 31 (31%, entry 9).In summary, the yields <strong>of</strong> the cross-coupling reactions turnedout to be somewhat disappointing; this could either be due to alower reactivity <strong>of</strong> the amino acids towards oxypalladationreactions compared to −regular× carboxylic acids, but on theother hand also to the relative instability <strong>of</strong> the enol esters thatare formed. The fact that analysis by TLC always showedcomplete conversion and no starting material was recovered inthese reactions seems to support the latter hypothesis.The Pd-catalyzed cyclizations with the enantiopure propargylglycine-derivedprecursor (S)-18 are shown in Table 3.The anticipated five-membered endocyclic enamide (S)-32was formed in 76% yield using 10 mol % <strong>of</strong> Pd(PPh 3 ) 4 and5 equiv <strong>of</strong> K 2 CO 3 in DMF at 80 8C (entry 1). Unfortunately,partial racemization occurred under these conditions leadingto an enantiopurity <strong>of</strong> the product <strong>of</strong> only 33%. To circumventthe undesired racemization, catalyst, solvent, reaction temperatureand the base were varied. By changing the solvent fromDMF to THF and the catalyst from Pd(PPh 3 ) 4 to Pd(OAc) 2 /2PPh 3 , the ee increased from 91% at 60 8C (entry 2) to >99% at40 8C (entry 3), albeit that the yield dropped to 48%. Adding asmaller amount <strong>of</strong> base, i.e., one equiv instead <strong>of</strong> five equiv,resulted in a dramatic lowering <strong>of</strong> the yield (18%), while stillpartial racemization was observed (entry 5). Changing the baseto the less basic salt NaHCO 3 did not improve the yield either(entry 6). When using the organic base Et 3 N, no cyclizationproduct was observed at all (entry 7). In addition, the influence<strong>of</strong> different ligands was studied. Several combinations <strong>of</strong>catalysts and ligands with different bite angles [18] werescreened, but none <strong>of</strong> these combinations had a dramaticinfluence on the yield or the reaction time. The chloridesources were also varied to study their effect on the yield andthe reaction rate at 40 8C. The use <strong>of</strong> TBAC led to anTable 3.HN CO 2 MeTs(S)-18entry123456789101112131415catalyst10% <strong>of</strong> catalystbase (5 equiv)solvent, TPd(PPh 3 ) 4Pd(OAc) 2 /2PPh 3Pd(OAc) 2 /2PPh 3Pd(PPh 3 ) 4Pd(OAc) 2 /2PPh 3Pd(OAc) 2 /2PPh 3Pd(OAc) 2 /2PPh 3Pd(OAc) 2 /xantphosPd 2 (dba) 3 /xantphosPd(OAc) 2 /dppePd(OAc) 2 /dppbPd(OAc) 2 /2PPh 3 /TBACPd(OAc) 2 /2PPh 3 /LiClPd(OAc) 2baseK 2 CO 3K 2 CO 3K 2 CO 3K 2 CO 3K 2 CO 3cNaHCO 3Et 3 NK 2 CO 3K 2 CO 3K 2 CO 3K 2 CO 3K 2 CO 3K 2 CO 3K 2 CO 3K 2 CO 3N CO 2 MeTs(S)-32solventDMFTHFTHFTHFTHFTHFTHFTHFTHFTHFTHFTHFTHFDMFTHFT(°C)806040406060606060606040406080time(h)3.544848242424325421244855.5OPPh 2 PPh 2xantphosyield(%) [a]76654836181806664444064261225ee(%) [b]3391>99[a] Isolated yields after column chromatography.[b] The ee was determined by chiral HPLC (Chiralpak OD; eluent: 20% IPA/heptane).[c] One instead <strong>of</strong> five equiv <strong>of</strong> base was added.84618972 Adv. Synth. Catal. 2002, 344, 70±83