ted telescoped cross metathesis-ring closing aza-Michael reaction sequence : step-economical access to nicotine – lobeline hybrid analogues †

Nicotinic acetylcholine receptors (nAChRs) belong to the family of pentameric ligand-gated channels. As they play a signicant role in cognitive and sensory gating processes, nAChRs comprise potentially therapeutic targets in manifold brain disorders. One of the major challenges in drug discovery targeting nAChRs, is to develop compounds that can selectively bind one receptor subtype. If fragment-based approach to nd out selective receptor ligands is nowadays widely established in drug discovery and chemical biology, the application of this concept on nAChRs is delayed due to the lack of readily available structural information and sensitive biophysical screening methods. Only one fundamental example of a fragment merging optimization of ligand has been described for the acetylcholine binding protein, a model protein for the extracellular a7 nAChR-subtype domain. Therefore, the concept of molecular hybridization continues to be an alternative answer to nd out new nAChRs ligands. More potent and selective ligands have been developed by combining the pharmacophores of natural alkaloids that exhibit a pronounced nAChR pharmacological activity such as nicotine (1), epibatidine (2), anatoxine (3) and lobeline (4) (Fig. 1). A part of our research program aims at shaping new stepeconomical synthetic processes of new Lobelia alkaloids


Introduction
Nicotinic acetylcholine receptors (nAChRs) belong to the family of pentameric ligand-gated channels.As they play a signicant role in cognitive and sensory gating processes, nAChRs comprise potentially therapeutic targets in manifold brain disorders. 1One of the major challenges in drug discovery targeting nAChRs, is to develop compounds that can selectively bind one receptor subtype.If fragment-based approach to nd out selective receptor ligands is nowadays widely established in drug discovery and chemical biology, the application of this concept on nAChRs is delayed due to the lack of readily available structural information and sensitive biophysical screening methods.Only one fundamental example of a fragment merging optimization of ligand has been described for the acetylcholine binding protein, a model protein for the extracellular a7 nAChR-subtype domain. 2 Therefore, the concept of molecular hybridization continues to be an alternative answer to nd out new nAChRs ligands.More potent and selective ligands have been developed by combining the pharmacophores of natural alkaloids that exhibit a pronounced nAChR pharmacological activity such as nicotine (1), epibatidine (2), anatoxine (3) and lobeline (4) (Fig. 1). 3 A part of our research program aims at shaping new stepeconomical synthetic processes of new Lobelia alkaloids analogues as ligands of nAChR-subtypes. 4In the context of targeting a4b2-subtype, the synthesis of hybrid molecules appeared to us as an interesting challenge and we decided to investigate the preparation of original lobeline-natural nAChR ligands chimeric analogues 5 by connecting relevant pharmacophores (Fig. 1).
A key advantage of the concept of molecular hybridization is its capacity to create highly chemically diverse molecules with a high degree of congenital resemblance, an essential criterion for relevant structure-activity relationship studies.We thus made an effort at designing a diastereoselective synthetic pathway that could rapidly reach structural and functional diversity.The shortness and the exibility of this synthetic strategy will be insured by a cross metathesis-ring closing aza-Michael (CM-RCAM) sequence (Scheme 1).
One of the challenges of our approach lies in the crossmetathesis of tethering N-heteroaromatic-containing olenic substrates 6 with sensitive electron-poor olenic coupling partners 7. Retrosynthetic analysis suggests that the transformation of the hydroxyl group of 8 into an amine function could easily lead to the formation of the key CM-RCAM precursors 6.It was further expected that the synthesis of the alcohol 8 would be secured by the condensation and 3-butenylmagnesium bromide with 2-substituted-5-carboxaldehydes 10.
We present herein a short and efficient access to nicotinenatural nAChR ligand hybrids for which each synthesis step offers both diversity and exibility with three distinct sites of modulation (R 1 , R 2 and R 3 ) and involves simple and commercially available precursors (7, 9 and 10).

Results & discussion
Our synthetic efforts are depicted in Scheme 2 and started with the efficient preparation of the pent-4-en-1-ol 8a and 8b.The olenic Grignard reagent was easily prepared starting from 4bromo-1-butene in the presence of magnesium before reacting with the commercially available pyridine-3-carboxaldehyde 10a or 6-chloropyridine-3-carboxaldehyde 10b. 5 The hydroxyl function of the pyridinic pentenols 8 was then efficiently transformed in a leaving group by mesylation under standard conditions.The mesylates 11a and 11b were isolated without further purication in 96% and 92% yields, respectively.Our initial attempts to synthesize the required amine moiety were envisaged via a two-step sequence including the introduction of an azido group followed by its reduction into primary amine.Treatment of the mesylated derivative 11a with sodium azide led to the desired azide 12a in a 80% yield.Unfortunately, the Staudinger reduction of 12a under usual conditions failed, giving a complex mixture of products.In the same way, other methods for converting an azido group into an amine function using propane-1,3-dithiol, 6 or indium metal and ammonium chloride 7 gave a mixture of inseparable polar products.
As the transformation of the azide function appeared to be quite problematic, an alternative synthetic pathway was investigated.In this way, we evidenced that the mesylate 11a easily underwent nucleophilic substitution (S N 2) in the presence of aqueous methylamine in DMF at 60 C, affording the pentenamine 14a in a satisfying 60% yield.The yield was increased to 75% by using a 2 M THF solution of methylamine rendering the purication step simpler as well.These amination conditions were also successfully applied to 11b delivering 14b in a good 70% yield.Nonetheless, a longer reaction time was necessary to reach full conversion: four days were needed for 14b instead of twelve hours in the case of 14a suggesting a deactivating effect of the chlorine atom on the nucleophilic substitution.Interestingly, it must be pointed out that these N-methyl-pent-4-en-1amines 14 did not require any further purication, before being engaged in the following step, highlighting the synthetic process cleanliness.At this stage of our synthesis, the protection of the methylamine group of 14 by an electron-withdrawing group was envisaged to avoid potential deactivating coordination between the amine function by the ruthenium-based catalyst during the olen cross-metathesis (CM).N-Methylamines 14 were thus treated with Boc 2 O providing the advanced key intermediates 6a and 6b in high yields (90% and 75%, respectively).
Only few examples of tandem CM/aza-Michael process have already been reported in the literature.4c,8 The bibliography becomes particularly poor for substrates bearing a strong Lewis base such as a free amine or a pyridine that may dramatically deactivate the catalyst. 9To the best of our knowledge, the reactivity of these potentially metal-coordinating substrates has never been described in the presence of electronically biased olens, such as a,b-unsaturated ketones.Following our recent results in this eld, 4c we decided to screen the efficiency of ve commercially available Ru complexes (Fig. 2).These precatalysts were selected based on their catalytic features: three standard metathesis complexes such as the Grubbs 2nd generation I (G-II), 10a the Grubbs-Hoveyda 2nd generation II (GH-II) 10b and the indenylidene-based complex III (M2); We rst evaluated the feasibility of this reaction by submitting the pyridinic olenic derivatives 14 and 6 with methyl acrylate 7a or ethyl acrylate 7b.As earlier described by the group of Cossy on pyridinic homoallylic alcohols, 9a the CM involving the pyridinic olenic amine 14a in the presence of the Grubbs catalysts I or II (10 mol%) in reuxing dichloromethane for 24 h only permitted the recovery of the starting material.These results were in coherence with the known poisonous character of both the amine and the pyridine substituents which deactivate the ruthenium catalyst by coordinating the metal centre.9b The study was thus pursued using the pyridines 6a and 6b and our results are summarized in Table 1.In this way, starting from the N-Boc protected derivative 6a, the desired cross-coupled product 15aa was formed in an encouraging 55% yield in the presence of the 2nd generation Grubbs catalyst I under the same aforementioned reaction conditions (Table 1, entry 1).The use of GH-II catalyst (II) provided the CM product 15aa in an enhanced 60% yield even with a lower 5 mol% catalyst loading (Table 1, entries 2 and 3).Moreover, the reaction time was dramatically reduced from 24 h to 45 min or 60 min when microwave heating was employed (Table 1, entries 1-3).Interestingly, yields were improved up to 70% by using more hindered catalysts such as IV or V with ethyl acrylate 7b as olenic partner (Table 1, entries 4-6).
Encouraged by these successful results with alkyl acrylates, we next examined the less studied cross-metathesis coupling with several vinyl ketones.Starting from methyl vinyl ketone 7c, the desired enone 15ac was isolated in a good 70% yield by combining the GH-II catalyst II with microwave irradiation (Table 1, entry 7).
Phenyl vinyl ketone 12 7d was next submitted to crossmetathesis with the terminal alkene 6a (Table 1, entries 8-11).The reaction was particularly reluctant under conventional heating conditions and did not work whatever the catalyst used.Contrastingly, the coupling product 15ad was easily obtained in a modest 50% yield, in only 3 hours combining microwave irradiation with GH-II catalyst II (Table 1, entry 8).Remarkably, following fully optimized conditions (i.e.7.5 mol% of the precatalyst IV, 1 h of microwave irradiation), the expected product 15ad was produced in a 70% yield (Table 1, entry 10).This protocol was nally extended to the more reactive chloropyridine scaffold 6b, and very satisfyingly, the highly functionalized Michael acceptor 15bd was isolated with a complete E selectivity in a 90% yield (Table 1, entry 11).This result was in accordance with the demonstrated favourable effect of a chlorine electron-withdrawing C-2 substituent that reduces the Lewis basicity of the pyridinic nitrogen atom.9a With the precursors 15 in hands, we focused on the cyclising aza-Michael step.Even though amino-enone 15ab underwent Boc-deprotection in the presence of a catalytic amount of HCl in i-PrOH at 60 C, the RCAM did not occur.Pleasingly, microwave heating (100 C, 200 W) efficiently reached the whole cascade in only 45 min, yielding the targeted pyrrolidines 5ab and 5ac in 50% to 65% yields (Scheme 3, eqn (1)).In addition, the best conditions for the tandem deprotection-cyclization sequence applied to 15ad were reached under ultrasound exposure for 3 hours and to 15bd by using conventional heating at 80 C for 2 hours.
In our continuing interest for developing economically favorable synthetic procedure, we surmised that the sequential CM/RCAM process could be telescoped.Indeed, Fustero and coworkers reported diastereoselective domino cross-metathesis/ aza-Michael reaction catalyzed by ruthenium complexes with Lewis acid as co-catalyst for the synthesis of piperidine, pyrrolidine and lactam derivatives.8c,e,f Directly applied to our pyridinic olens, the cascade process using either Ti(OiPr) 4 or  Extension of these optimized conditions provided the isolation of the original pyrrolidines 5ab, 5ac, 5ad and 5bd in 50-70% yields over three steps (Scheme 3, eqn ( 2)).More in details, a mixture of both diastereomers was obtained.If the 2,5-cis diastereoisomer was kinetically favoured, it rapidly equilibrated in solution toward the thermodynamically more stable 2,5-trans epimer through a well-known auto-catalyzed retro-aza-Michael/ aza-Michael cyclization process.4d The relative conguration of the major diastereoisomer was established by nOesy experiments and revealed a trans conguration between the both hydrogen atoms of the pyrrolidine C-2 and C-5 atoms (Scheme 3, eqn (2)).

Conclusions
We shaped a synthetic pathway allowing a short and efficient access to a series of lobeline-nicotine, pelletierine-nicotine and lobeline-nicotine-epibatidine hybrids.More generally, we demonstrated that this approach based on a challenging monotope sequential microwave-mediated cross-metathesis/ cyclizing aza-Michael reaction is expandable to the diastereoselective preparation of 2,5-trans disubstituted pyrrolidines.This synthetic strategy has the advantages of (i) involving simple starting reagents and available Ru-precatalysts, (ii) producing several purication-free synthetic intermediates and (iii) facilitating the introduction of molecular diversity.Moreover, the developed methodology enabled, under microwave irradiation, the telescoping of the cross-metathesis and the intramolecular aza-Michael reaction into a single efficient process, readily amenable to scale up.New applications of this strategy for the enantioselective synthesis of pyrrolidines as well as the nAChR-subtypes binding affinity and activity of selected nicotine-lobeline hybrid analogues are currently in progress and will be reported in due course.
Scheme 1 Retrosynthetic strategy for the preparation of hybrids 5.

BF 3 $
OEt 2 as Lewis acid co-catalyst failed, providing complete degradation of the starting materials.To our delight, under microwave irradiation, sequential addition of a catalytic amount of concentrated hydrochloric acid aer the crossmetathesis completion (controlled by TLC), initiated the N-Boc cleavage and activated the subsequent aza-Michael induced ring closure.

Table 1
Optimization of the cross-metathesis of pyridine derivatives with electron-poor olefinic partners a