Diverse N‐Heterocyclic Ring Systems via Aza‐Heck Cyclizations of N‐(Pentafluorobenzoyloxy)sulfonamides

Abstract Aza‐Heck cyclizations initiated by oxidative addition of Pd0‐catalysts into the N−O bond of N‐(pentafluoro‐benzoyloxy)sulfonamides are described. These studies, which encompass only the second class of aza‐Heck reaction developed to date, provide direct access to diverse N‐heterocyclic ring systems.

There has been ar esurgence of interest in the development of processes based on the Mizoroki-Heck reaction. [1] Notable contributions include boryl-Heck alkene functionalizations [2] and remote redox relay Heck CÀCb ond formations. [3] Our focus has been on the development of aza-variants of the Heck reaction, because of the importance of N-containing ring systems in drug discovery. [4][5][6][7] Within this context, the Narasaka process, [4] which involves the Pd-catalyzed cyclization of O-pentafluorobenzoyl ketoxime esters with alkenes,is unique in harnessing key steps that are analogous to the conventional Heck reaction:1)anunusual oxidative addition into the N À Ob ond of 1 to afford cationic imino-Pd intermediate 2; [7,8] 2) C À Nb ond forming alkene migratory insertion; [9] and 3) b-hydride elimination (Scheme 1A). Imino-Pd II intermediates 2 can also be exploited more widely in redox neutral processes,s uch as diverse alkene 1,2-carboaminations, [8] aryl CÀHa minations, [7a] alkene aziridinations, [10] alkene 1,2-iodoaminations, [11] aryne aminofunctionalizations, [12] and C À Cb ond activations. [13] Efforts to expand the range of redox active donors available for accessing aza-Pd II intermediates led us to consider whether activated hydroxysulfonamide derivatives might be viable (Scheme 1B). [14] In this approach, N-(pentafluorobenzoyloxy)sulfonamides 4a/b,w hich we have found easy to prepare on gram scale, [15] act as af ormal nitrene equivalent, but with key distinguishing aspects.F irst, as with nitrenes, 4a/b function as both anucleophile and electrophile, but, importantly,these features are decoupled, such that their unveiling can be orchestrated in acontrolled manner.Second, nucleophilic modification of 4a/b can be achieved under stereospecific Mitsunobu conditions and this allows readily available enantiopure secondary alcohols to be exploited in synthetic sequences. [16] Third, and most importantly, 5a/b do not function as an electrophile by direct reaction at nitrogen, with this reactivity facet instead controlled by the Pd-center of aza-Pd II species 6a/b.C onsequently,a lkylated derivatives 5a/b can, in principle,b ea dapted to asymmetric cyclizations [17] and cascade sequences, [18] as well as other processes typical of Pd-catalysis.H erein, we delineate preliminary studies towards this broad goal by reporting what is,t ot he best of our knowledge,o nly the second class of aza-Heck reaction developed to date (Scheme 1B,box). [19] Theprocess provides high versatility for the synthesis of complex N-heterocyclic ring systems [20] and can be integrated into cascade sequences to provide alkene 1,2-carboamination products.T his validates the broader N-heteroannulation strategy outlined in Scheme 1B. Initial studies focused on aza-Heck cyclization of monosubstituted alkene 7a,w hich was prepared in 70 %y ield by Mitsunobu alkylation of 4a with pent-4-enol (Scheme 2). [15] Under conditions related to those previously optimized for aza-Heck cyclizationso fo xime esters,w here P-(3,5-(CF 3 ) 2 C 6 H 3 ) 3 was identified as ap rivileged ligand, [5] ketone 8a' was isolated in 82 %y ield. 1 HNMR analysis of crude reaction mixtures indicated that 8a' ' forms via hydrolysis of initial aza-Heck product 8a.
Cyclization of 7a was considered relatively easy as both the N À Ob ond and alkene are sterically accessible.T o integrate the new process into synthetically attractive settings we sought substrates where b-hydride elimination to form hydrolytically sensitive enamides was not possible.A ccordingly we focused on cyclic alkene 7ba,which was expected to deliver bicyclic system 8b,d ue to the presumed mechanistic constraints of syn-amino palladation and syn-b-hydride elimination (Table 1). In the event, this system was challenging, with initial attempts generating 8b in only 34 %y ield as a3 :1 mixture with regioisomer iso-8b (entry 1);t his likely arises via Pd-hydride mediated isomerization of 8b.I nefficiencies were attributed to competing protodepalladation and b-hydride elimination at the stage of the aza-Pd II intermediate;this latter pathway led to the isolation of the corresponding aldehyde. [21] Optimization was undertaken focusing on activating group,s olvent, and ligand. O-Trifluoroacetyl acti-vated variant 7bc offered marginal efficiency gains (entry 3), whereas an O-Ms activated system 7bb was less effective. Less dissociating activating groups,s uch as O-Bz, were completely ineffective (see below). Fortunately,i tw as found that solvent effects were pronounced, with n-BuCN,M eCN, and THF all promoting cyclization of 7ba to target 8b in useful yield (entries 4,6,7). Them ost efficient method used am ixed-solvent system and sub-stoichiometric quantities of Et 3 N( see below; entry 5). Thep rocess is highly sensitive to the nature of the phosphine ligand, and, from an exhaustive screen of commercial variants,t he only other systems found to provide greater than 20 %y ield were PPh 3 ,d ppp,a nd P(4-(CF 3 )C 6 H 4 ) 3 .
Thescope of the aza-Heck process is outlined in Table 2, with fine tuning of reaction solvent required on acase-by-case basis.C yclization of 7c,w hich involves ac yclopentene, generated bicyclic system 8c in high yield and as as ingle diastereomer.E fficient cyclizations were observed for processes involving 1,2-disubstituted alkenes.F or example, 7d delivered 8d in 81 %yield and with complete selectivity over the corresponding enamide (cf. 7a to 8a). 1,1-Disubstituted alkenes are also tolerated, albeit with greater variation in efficiency.C yclization of 7f generated the challenging tetrasubstituted stereocenter of pyrrolidine 8f in 80 %y ield. More sterically demanding systems 7g and 7h were less effective,b ut still delivered targets 8g and 8h in workable yields.S ystems with substitution on the alkene tether can provide diastereoselective processes.F or example, 7k generated cis-2,5-disubstituted pyrrolidine 8k in 58 %y ield and more than 10:1 d.r;f or this process,a nN-tosyl protecting group was less effective. [15] Similar efficiencies were observed for 7j, 7l,a nd 7m,w ith the latter affording complex 2,2,5trisubstituted pyrrolidine 8m in high diastereoselectivity. Electron-deficient alkenes also participate:c yclization of acrylate 7n provided 8n in 78 %y ield, thereby validating anovel entry to versatile alkylidene pyrrolidines.
Thechemistry can be used to provide challenging bridged ring systems common to many alkaloid targets (Scheme 3). Forexample,cyclization of 7o,which involves acycloheptene constructed by RCM, [15] provided tropane 8o in 60 %yield;this is the core structure of multiple natural products including cocaine. [22] Alternatively,c yclization of 7p generated regioisomeric 6azabicyclo[3.2.1]octene scaffold 8p in 76 %y ield. [23] Thes tructures of 8oand 8pwere confirmed by X-ray diffraction. [15] Preliminary studies show that the chemistry will be of utility in other contexts.A ll aza-Heck processes described so far involve 5exo cyclization;h owever,e ven at the present level of development, 6exo cyclization is possible (Scheme 4A). Indeed, exposure of styrenyl system 7q to optimized conditions provided tetrahydroisoqui-Scheme 2. Afeasibility experiment. [a] In situ yield; 8b:iso-8b ratio is given in parentheses.

Angewandte Chemie
Communications noline 8q in 42 %yield. We have also assessed the possibility of alkene 1,2-carboamination processes by trapping the alkyl-Pd II intermediate generated after migratory insertion (Scheme 4B). Exposure of 7r to aza-Heck conditions afforded bicycle 8r in 86 %y ield, via Heck trapping of 7r' '.T he development of further alkene aza-functionalizations will be afocus of future work. Themechanism of the aza-Heck processes is likely akin to that of the Narasaka cyclization of O-pentafluorobenzoyl ketoxime esters (Scheme 5, 7d to 8d). [5,8] Pd 0 L n (L = P-(3,5-(CF 3 ) 2 C 6 H 3 ) 3 )g enerated in situ effects N-O oxidative addition of 7d to provide I;despite extensive efforts,wehave so far been unable to isolate aza-Pd II intermediates related to I.E fficient aza-Heck cyclization requires dissociation of pentafluorobenzoate from I to access cationic intermediate II. [8] This assertion is based on the observation that less dissociating leaving groups (for example,O -Bz) are ineffective,a nd chloride additives (for example, n-Bu 4 NCl) com- [a] Reaction solvent is specified in parentheses under each starting material. Full details are given in the Supporting Information.

Angewandte Chemie
Communications pletely suppress cyclization;inboth cases protodepalladation to the corresponding sulfonamide predominates.From II, synmigratory insertion of the alkene generates alkyl-Pd intermediate III.T he intermediacyo fIII is corroborated by the cyclization of 7r to 8r,while support for the feasibility of synstereospecific alkene migratory insertion is found in studies on aza-Wacker cyclizations. [24,25] From III, b-hydride elimination releases the product (8d)a nd Pd II -hydride IV,w hich undergoes base (Et 3 N) induced reductive elimination to close the catalytic cycle.T he equilibrium between neutral and cationic complexes I and II is shifted forward by triethylammonium mediated protodecarboxylation of the otherwise inhibitory pentafluorobenzoate leaving group.W eh ave previously shown that this process is rapid, [8] and 19 FNMR analysis of crude reaction mixtures has confirmed that it is operative in the current scenario.T his also accounts for the use of sub-stoichiometric (catalytic) quantities of Et 3 Nunder optimized conditions.
It is pertinent to comment on the synthetic scope of the prototype 5-exo aza-Heck processes outlined here versus complementary 5-exo aza-Wacker cyclizations of alkenyl NHsulfonamides,which require an external oxidant (for example, air or oxygen). [24] Despite extensive development, this latter method still has key limitations;f or example,c yclization of systems with large a-substituents (larger than methyl) have not been achieved (cf. 7j-m), hindered acyclicolefins do not participate (cf. 7h), and electron-deficient alkenes cannot be used due to competing conjugate addition (cf. 7n). Additionally,a za-Heck cyclization seems uniquely suited to demanding systems (Scheme 3) and cascade polycyclizations (Scheme 4B). Earlier work using oxime esters has also established N-O oxidative addition as aunified platform for the design of diverse redox-neutral alkene 1,2-carboamination processes that cannot be achieved using an aza-Wacker approach. [8] From ap ractical viewpoint, ap re-installed internal oxidant may be preferable for scale-up or redox sensitive substrates. Importantly,this unit can be brought in directly by Mitsunobu reaction of 4a/b,enabling atwo-step conversion of (enantiopure) alcohols to heterocyclic targets.A lkenyl NH-sulfonamides required for aza-Wacker cyclization are not usually prepared directly from the alcohol because the requisite primary sulfonamides do not engage efficiently in conventional Mitsunobu reactions. [26] Further potential advantages of the aza-Heck approach are that highly tunable phosphine ligands can be used (because oxidative conditions are avoided) and predictable syn-migratory insertion of the alkene can be expected. [24c] In summary,wereport aza-Heck cyclizations initiated by oxidative addition of Pd 0 -catalysts into the N À Ob ond of N-(pentafluorobenzoyloxy)sulfonamides.T hese studies provide direct access to N-heterocyclic ring systems that are not accessible using the Narasaka aza-Heck procedure. [20] The approach exploits stepwise unveiling of the nitrenoid character embedded within N-(pentafluorobenzoyloxy)sulfonamide reagents.S equential nucleophilic-electrophilic C À Nb ond forming strategies of this type,which involve the intermediacy of atunable aza-Pd II intermediate,should enable awide array of N-heteroannulation processes.Byanalogy to the utility of oxime ester derived imino-Pd intermediates (2), [4,5,[8][9][10][11][12][13] we also anticipate that the catalysis platform outlined here,which involves ar are example of oxidative addition of Pd 0 into an N À Ob ond, [7] should find broad applicability in the design of redox neutral C À Nb ond forming methods outside the immediate area of N-heterocyclic chemistry.