Polymer-Supported Reagents

The use of polymer-supported reagents and scavengers provides an attractive and practical method for the clean and efficient preparation of novel chemical entities. These methods can be extended in a multistep fashion to provide access to more complex structures, including biologically active natural products. In our major review we covered all known supported reagents, catalysts and scavenging agents as useful directory to assist with future synthesis planning in the chemical community. Our group has a long track record in the development and application of polymer-supported reagents in organic synthesis and here we provide a few examples of our work in this area.


DEVELOPMENT OF POLYMER-SUPPORTED REAGENTS

We have developed many new polymer-supported reagents in the course of our research, too many to list in detail. Very often, we also found new applications by using, or modifying or re-inventing, existing supported reagents as well. These are a small selection of specifically developed reagents we introduced to tackle particular synthesis challenges.

Polymer-supported iridium catalyst
We prepared a polymer-supported iridium catalyst and used in the isomerisation of the double bonds in aryl allylic derivatives with excellent trans selectivity and without the need for conventional work-up procedures.

Polymer-supported thiolating agent
We developed a thiolating reagent for the conversion of carbonyls to thiocarbonyls and demonstrated its use on a range of amides. Secondary or tertiary amides were converted cleanly and efficiently through to the corresponding thioamides and primary amides were converted to the corresponding nitriles. While reactions could be aided by conventional heating, we found that if microwave heating was used, in the presence of an ionic liquid, enhanced reaction rates are achieved.

Polymer supported perruthenate
Steve Ley was the inventor of the widely used TPAP oxidation agent, and of course we had to prepare a polymer supported perruthenate reagent. We used it in the conversion of primary and secondary alcohols to aldehydes and ketones, respectively, which afforded pure products without the need for conventional work-up procedures.


THE POLYMER-SUPPORTED REAGENT APPROACH TO NATURAL PRODUCT SYNTHESIS

The use of polymer-supported reagents in our group is extensive, particularly as we now incorporate them into our flow chemistry projects. What we offer are a few discussions on a limited number of natural products that we have synthesised using polymer-supported reagents. At their time, these syntheses were often “world-firsts” and were major landmarks for us early on. These are a few examples excerpted from one of our book chapter reviews.

Oxomaritidine and Epimaritidine
Our first serious application of supported reagents for natural product synthesis was published in 1999 [1]. We reported concise routes to two amaryllidaceae alkaloids; oxomaritidine (1) and epimaritidine (2) in just five and six steps, respectively. Supported reagents, featured in all of the steps, were used in a sequential fashion and led to pure products following a simple filtration to remove spent reagents.

The route began with the quantitative oxidation of 3,4-dimethoxybenzyl alcohol (3) to the corresponding aldehyde 4 using our own polymer-supported perruthenate (PSP), a catalytic oxidant [2]. Next, the reductive amination of aldehyde 4 was achieved firstly by coupling with the phenolic amine 5 and followed by reduction using a polymer-supported borohydride reagent to furnish the corresponding secondary amine 6 under previously optimised conditions [3].  Amine 6 was protected as trifluoroacetate 7 by treatment with trifluoroacetic anhydride and immobilized dimethylaminopyridine (PS-DMAP) as the catalyst and base. The resulting product 7 underwent smooth oxidative coupling to the spirodienone 8 using polymer-supported hypervalent iodine diacetate, a reagent developed specifically for the task [4]. In order to guarantee good conversion in this process, it was essential to use trifluoroethanol as the solvent. Finally, a wet carbonate ion-exchange resin simultaneously deprotected and initiated cyclisation (via conjugate addition) to give the first of the natural products, oxomaritidine (1), in essentially quantitative yield after filtration and solvent evaporation. Subsequent stereoselective reduction of oxomaritidine (1), using an immobilized copper boride (or nickel boride) equivalent, delivered the second natural product epimaritidine (2) in an excellent 50% overall yield over the six step sequence. The synthesis was scaleable to deliver gram quantities.

Epibatidine
The potent analgesic alkaloid epibatidine (18) [5] isolated from the Equadorian poison frog Epipedobates tricolour, was a more challenging compound to synthesize using immobilization techniques. This synthesis involved the orchestrated employment of 10 polymer-supported reagents and scavengers to give epibatidine (18) in greater than 90% purity without any chromatographic steps.

The synthesis began by transforming the commercially available acid chloride 19 to aldehyde 20. This was achieved in a two-step process by reduction of the acid chloride to the intermediate alcohol with polymer-supported borohydride, then partial re-oxidation by the PSP reagent to deliver aldehyde 20. Alternative immobilized oxidants such as supported permanganate [6] and diacetoxyiodobenzene were equally efficient. Oxidants such as Magtrieve (magnetised CrO2 and MnO2), although performing the reaction, were less suitable as they required considerably longer reaction times. Aldehyde 20 then underwent straightforward conversion to the nitrostyrene 21. A Henry reaction promoted by the basic Amberlite resin (IRA 420 OH form) and followed by elimination of an intermediate trifluoroacetate using polymer-supported diethylamine as a base gave 21. By NMR analysis at each stage, products were shown to be better than 95% pure. In an important modification to the above sequence of reactions, it was found that by the incorporation of supported reagents contained in sealed porous polymer pouches the conversion of chloride 19 to nitrostyrene 21 was possible in a one-pot operation. Thus, when an individual reaction was deemed to be complete the pouch was simply removed, washed with solvent, and the next reagent pouch added to the flask. This process obviated the need for filtration between individual steps.

In the next phase of the synthesis, a regioselective Diels-Alder reaction of nitrostyrene 21 with 2-tert-butyldimethylsilyloxybutadiene in a sealed tube at 120 oC and work-up with a volatile acid (TFA) to hydrolyse the intermediate enol ether gave ketone 22 exclusively as the trans-substituted product.  Stereoselective reduction of the carbonyl and corresponding mesylate formation gave 23, again using an immobilized suite of reagents.  Selective reduction of the nitro 23 to amine 24 in the presence of other sensitive functionalities was next achieved using polymer-supported nickel boride. Treatment of amine 24 with PS-BEMP initiated an intramolecular displacement of the mesylate which, following a scavenging step with a polystyrene aminomethyl resin to remove excess mesylate, yielded the natural product precursor 25. Epimerisation to epibatidine (18) was readily achieved by microwave heating in the presence of potassium tert-butoxide. The natural product was ultimately isolated by a catch-and-release technique.

Plicamine
Much of the above constitutes the preparation of relatively simple structures. However, these syntheses actually helped prepare the ground for more challenging targets such as the synthesis of the alkaloid (+)-plicamine (27) [7]. Extensive use of parallel optimisation methods and focussed microwave techniques, to achieve fast reaction times, were used to complete the total synthesis in just six weeks without rehearsal of any reactions using conventional solution-phase methods or separation techniques. In this respect this synthesis stands out. Please refer to the full details of the work [8] which describes how the route (in multi-gram quantities) can be modified to afford analogues or generate related structural scaffolds for further chemical decoration. Also of note is that since this synthesis relied on a single asymmetric centre in the starting material to control all the others, by use of the opposite enantiomer, the unnatural (–)-plicamine enantiomer can also be prepared and examined for biological activity.

Polymer-supported hypervalent iodine reagent performed well to convert phenol 28 to spirodienone 29. Nafion-H (fluorosulfonic acid resin) catalysed the final cyclisation of 29 to form the tricyclic core of the natural product in virtually quantitative yield. After stereo- and regioselective reduction of 30 using supported borohydride, the very hindered intermediate alcohol was methylated by treatment with trimethylsilyl diazomethane and supported sulfonic acid resin to give 31. This process is very mild and is to be recommended in difficult situations. The remaining steps to (+)-plicamine from 31 were relatively straightforward. However, the final oxidation of amine 32 to (+)-plicamine required significant development. This was eventually achieved using CrO3 and 3,5-dimethylpyrazole followed by scavenging with Amberlyst 15 resin. The chromium salts were efficiently removed by filtration through a mixed bed containing Varian Chem Elut CE 1005 and Montmorillonite K10 clay to give (+)-plicamine (27).

The route to plicamine was also be diverted at an earlier stage to afford two further natural products, plicane (33) and obliquine (34) [9]. The ability to divert material in this way is attractive and was facilitated by using immobilized reagent methods. These more advanced syntheses clearly illustrated the opportunities created by embracing these methods for multi-step transformations.

Carpanone
The deceptively complex natural product carpanone (41) was made from commercially available sesamol in a relatively simple set of reactions using immobilized reagents.

After allylation of sesamol using allyl bromide and PS-BEMP, the product underwent an extremely clean Claisen rearrangement using a combination of toluene and an ionic liquid to absorb the energy from an external microwave source. These binary conditions were simple to operate, since after heating then cooling, the mixture was separated using a simple liquid handler to remove the product-containing toluene layer. The ionic liquid can also be recovered and re-used in further experiments. The toluene fraction containing 42 was then used directly in the double bond isomerisation reaction to yield conjugated compound 43. This was achieved stereoselectively (11:1 trans:cis) by use of a new immobilised iridium catalyst 44 [10] developed specifically for this project. However, it has also been shown to be general for other double bond isomerisations at room temperature [11].  Lastly following original work by Chapman using other oxidants, the phenolic styrene was converted to carpanone (41) by application of a modified Jacobsen Salen cobalt complex under catalytic conditions with molecular oxygen. After scavenging with a carbonate resin to remove unreacted phenols and a trisamine resin to remove an unwanted aldehyde by-product, the crystalline natural product was obtained in excellent yield and purity. Mechanistically, carpanone was finally formed by oxidative dimerisation through carbon coupling of the phenol (43) followed by a highly stereoselective intramolecular Diels-Alder reaction.

Epothilones
The epothilones have generated wide interest in the broad scientific community owing to their ability to inhibit tumour cell proliferation by inducing mitotic arrest through microtubules stabilization. Owing to this activity, they have also become principal targets for many synthesis groups. Indeed they provide an excellent platform to explore the full armoury of supported reagents and scavengers for complex molecule synthesis. The epothilones lend themselves well to convergent synthetic approaches as they can be easily disconnected to more manageable fragments for analogue development; a crucial aspect of pharmaceutical drug programmes. For the synthesis of epothilone C (66) and epothilone A (67) by epoxidation, a convergent synthesis plan was devised that would require the coupling of three major fragments 68, 69 and 70.  An important criterion for this route was conducting the work in an efficient and clean manner using only immobilized reagents, scavengers and catch-and-release techniques. The overall aim was to avoid chromatography, crystallisation, distillation and water washes which are common in conventional approaches to these molecules.  Consequently, using only immobilized reagents, scavengers and catch-and-release techniques can achieve this goal [12].

In the full paper on this work, a number of alternative routes to the various fragments were reported [13]. Three routes to fragment A (68) were investigated, but the one shown in below was the shortest, the most efficient, and also proved to be the most easily scaled.  The key feature was the formation of the C2-C3 bond with concomitant introduction of the desired C3 stereocentre by application of an asymmetric Mukaiyama aldol reaction. In this synthesis, the aldol reaction progressed in excellent yield to give the alcohol 71 with an enantiomeric excess of 92%. This was achieved using a complex of borane with N-tosylphenylalanine as the chiral ligand. Work-up of this reaction required the addition of a minimum amount of water and a boron selective scavenger Amberlite IRA-743 to quench the reaction and remove the contaminating boric acid. Filtration and solvent removal produced a suspension of amino acid and aldol product 71.  However, the insolubility of the N-protected amino acid in non-polar solvents allowed dissolution of the desired aldol product. Subsequent filtration enabled the amino acid to be recovered and recycled whilst concentration of the filtrate gave the purified alcohol 71. After protection as its tert-butyldimethylsilylenol ether 72 and reaction with (trismethylsilylmethyl)lithium followed by a scavenger quench using a carboxylic acid resin, direct filtration and solvent removal yielded the ketone 73.  α-Methylation of ketone 73 via its lithium enolate and work-up again with a polymer-supported carboxylic acid gave fragment A (68) in just six steps from commercially available starting material.

The preparation of the second key coupling partner fragment B (69) was achieved in just five steps from the commercially available bromide (–)-(R)-3-bromo-2-methyl-1-propanol (74). Protection of the alcohol as its THP-ether using a polymeric sulfonic acid followed by Finkelstein halide exchange gave iodide 75. Homologation of alkyl iodide 75 with a cuprate derived from 3-butenylmagnesium bromide produced the corresponding alkene 76. Addition of a carboxylic acid resin and the trisamine resin quenched the reaction and scavenged dissolved copper salts. Finally after deprotection of the THP acetal, the resulting alcohol was oxidised to the fragment B aldehyde 69. Several oxidants were investigated for this process, but the most expedient proved to be pyridinium chlorochromate PCC on basic alumina.

The final fragment C (70) was constructed  in a convergent fashion from (S)-α-hydroxyl-g-lactone (77) and the chloromethyl triazole hydrochloride (78). Lactone 77 was therefore elaborated to ketone 79. Ketone 79 was coupled with a phosphonate derived from thiazole 78 via a highly stereoselective Horner-Wadsworth-Emmons reaction. A polymer-supported aldehyde was used to scavenge any excess phosphonate from this reaction to yield the bis-TBS protected adduct 80. TBS-protected diol 80 was taken through to fragment C by previously described steps. Ultimately iodide 70 was captured onto a polymer-supported triphenylphosphine to produce Wittig salt 81 as the coupling precursor.

With all the fragments now in hand, the final fusion of the components began. Previously, the stereoselective aldol coupling of fragments A and B to form the C6-C7 bond was shown to be highly sensitive to proximal and remote functionality in both fragments. In this work, although bearing close similarity to previous studies, the coupling of ketone 68 with the aldehyde 69 was a novel combination. Using LDA as a base this aldol coupling proceeded in quantitative yield with better than 13:1 stereoselectivity for the desired product. An acetic acid quench was followed by treatment with a diamine functionalised polymer. This resin served two purposes; to remove the excess acid and to sequester a small amount of unreacted aldehyde.

The elaborated adduct 82 was then silyl-protected and the double bond cleaved by ozonolysis. Work-up of the intermediate ozonide was achieved by application of immobilized triphenylphosphine. This is an excellent procedure as the triphenylphosphine oxide produced in normal solution work-ups and in Wittig chemistry causes practical difficulties that often require multiple chromatographic separations to obtain pure material.  Here simple filtration was sufficient. The resin bound phosphonium salt 81 derived from fragment C was treated with an excess of sodium hexamethyldisilazide (NaHMDS) followed by washing with anhydrous THF to give an isolable salt-free ylide. This was coupled stereoselectively to give the cis-olefin which required application of a dilute methanolic solution of camphorsulfonic acid to effect selective removal of the primary TBS-protecting group to give the free alcohol 84. This process required a separate scavenging step using an immobilized carbonate resin to remove the acid with the volatile MeOTBS by-product being removed under reduced pressure.

Catalytic tetra-N-propylammonium perruthenate (TPAP) oxidation of alcohol 84 followed by filtration through a pad of silica gel to remove morpholine and ruthenium by-products gave an intermediate aldehyde which was immediately oxidised to the corresponding acid using a previously developed modified Pinnick procedure [14]. Finally, selective desilylation prior to macrocyclisation was finally achieved found using tetrabutyl ammonium fluoride (TBAF) solution to afford intermediate 85. Immobilized versions of fluoride or other acidic resins led to complex mixtures. As a consequence of using the TBAF procedure, an aqueous extraction was necessary, and this gave pure deprotected alcohol 85. This was the first and only water wash used in the whole synthesis to this point. The final steps to epothilone A (67) simply required application of the Yamaguchi macrolactonisation procedure using PS-DMAP and a catch-and-release purification to produce epothilone C. Upon epoxidation with DMDO, epothilone A was obtained.

Without a doubt, this synthesis constitutes a triumph for the utility of supported reagents and scavenging techniques for multi-step complex molecule assembly. The high stereoselectivity and overall yield in this synthesis of epothilone A compares well with the best of all of the previous and conventional, routes. By the discussed route, the target molecule was delivered in 29 steps, with the longest linear sequence being only 17 steps from readily available materials.


1. Synthesis of the alkaloids (±)-oxomaritidine and (±)-epimaritidine using an orchestrated multi-step sequence of polymer supported reagents S.V. Ley, O. Schucht, A.W. Thomas, P.J. Murray J. Chem. Soc., Perkin Trans. 11999, 1251-1252

2. (a) Polymer supported perruthenate (PSP): a new oxidant for clean organic synthesis B. Hinzen and S.V. Ley J. Chem. Soc., Perkin Trans. 1 1997, 1907-1908 (b) Polymer supported perruthenate (PSP): clean oxidation of primary alcohols to carbonyl compounds using oxygen as cooxidant B. Hinzen, R. Lenz, S.V. Ley Synthesis 1998, 977-979

5. Synthesis of the potent analgesic compound (±)-epibatidine using an orchestrated multi-step sequence of polymer supported reagents J. Habermann, S.V. Ley, J.S. Scott J. Chem. Soc. Perkin Trans. 11999, 1253-1256

6. Clean five-step synthesis of an array of 1,2,3,4,-tetra-substitued pyrroles using polymer supported reagents M. Caldarelli, J. Habermann, S. V. Ley J. Chem. Soc. Perkin Trans. 1 1999, 107-110

8.  Total synthesis of the amaryllidacea alkaloid (+)-plicamine using solid-supported reagents I.R. Baxendale, S.V. Ley, C. Piutti, M. Nesi, Tetrahedron200258, 6285.

10. A polymer-supported iridium catalyst for the stereoselective isomerisation of double bonds I.R. Baxendale, A-L Lee, S.V. Ley Synlett2002, 516

12. A total synthesis of epothilones using solid-supported reagents and scavengers R.I. Storer, T. Takemoto, P.S. Jackson, S.V. Ley Angew. Chem. Int. Edn., 200342, 2321

13. Multi-step application of immobilized reagents and scavengers: a total synthesis of epothilone C R.I. Storer, T. Takemoto, P.S. Jackson, D.S. Brown, I.R. Baxendale, S.V. Ley Chem. Eur. J200410, 2529-2547

14. Solid-supported reagents for the oxidation of aldehydes to carboxylic acids T. Takemoto, K. Yasuda, S.V. Ley Synlett 2001, 1555