An efficient copper-catalysed pyrrole synthesis

Copper-catalysed cyclisations of  -hydroxyhomopropargylic sulfonamides can be carried out using copper(II) acetate in hot toluene to provide generally excellent yields of the corresponding pyrroles.


Introduction
The development of efficient and environmentally acceptable syntheses of heteroaromatics continues to be of enormous importance because of the value of compounds containing such residues in most areas of 'effect' chemicals.Despite many years of effort, there still remain a number of deficiencies in the methods available for the elaboration of such compounds, especially in a general sense in a number of structural types.In the area of pyrroles, it is still not unusual to see quite poor yields reported for the synthesis of relatively simple pyrroles when these are used as precursors of more complex derivatives.These limitations in existing pyrrole synthesis 1 are evident from the large number of reactions types which have recently been applied to their preparation. 2Further, it is often the case that when an 'effect' chemical contains a pyrrole residue, its synthesis is quite inefficient and used simply to provide sufficient material; in some cases one wonders what would happen if a commercial scale-up actually became necessary.Our own recent contributions have been focussed on the use of iodonium and selenium electrophilic species to trigger 5-endo-dig cyclisations, illustrated by the conversion of the alkynyl sulfonamides 1 into iodopyrroles 2 (Scheme 1) which, while usually providing excellent yields of substituted iodopyrroles, does suffer from some limitations. 3cheme 1.An iodocyclization route to β-iodopyrroles.
Efficient though these are, with the presence of the iodine atom suggesting a number of opportunities for further elaboration, 4 the requirement for three equivalents of iodine rather detracts from this methodology, which would probably not be viable on a large scale on cost grounds alone.We therefore initiated a series of studies aimed at defining heterogeneous catalysts which are suitable for triggering such cyclisations of homopropargylic sulfonamides of the same type as 1, with a view to forming pyrroles (3) (Scheme 2).These could in particular offer the crucial advantages of defining a catalyst which could be both recoverable and reusable and thereby carry some significant 'green' credentials.Herein, we report a new cyclisation protocol for the very efficient generation of pyrroles from 3-alkynyl 2-hydroxysulfonamides.Scheme 2. A possible catalysed route pyrroles by 5-endo-dig cyclisation.
Fortunately, we had available a range of suitable precursors 6 from a separate study having the amine function protected by a toluenesulfonyl group.This was especially advantageous, as this group has provided very stable protection, and also apparently activation, in cyclisations previously studied by us, and can also be readily removed from pyrroles under, effectively, basic saponification conditions.These precursors were readily obtained from 1-alkynes 4 by sequential formylation 5 and condensation of the resulting ynals 5 with the tin(II) enolate of ethyl N-tosylglycinate, 6 as summarized in Scheme 3.This methodology, perhaps surprisingly, also works very well when applied to ketones; hence, we were also readily able to obtain more substituted precursors [e.g.8] (Scheme 4). 6heme 4. Synthesis of a more substituted precursor.
Ethyl N-tosylglycinate 9 was also used as the precursor to a pair of pseudosymmetrical precursors but by acting as an electrophile this time, in condensations with two equivalents of a lithiated alkyne, which gave excellent yields of the potential pyrrole precursors 10 (Scheme 5).Finally, two representative alkyl-substituted precursors, 13, were prepared from the aminoaldehyde 11, readily derived from 2-amino-1-butanol ( Scheme 6).The rather irritating exchange of protecting groups was necessary because the N-tosyl aldehydes corresponding to the N-Boc aldehydes 11 were highly unstable and failed to deliver acceptable yields of the desired aminoalcohols 13 in a direct fashion.
Our first success in this area, reported a while ago, 7 was the finding that various copper salts were capable of cyclising such substrates (e.g.6a) but to give, most often, a predominance of the hydroxy-dihydropyrroles 14a, accompanied by varying but generally much smaller returns of the corresponding pyrroles 15a (Scheme 7).Of a succession of copper salts tested, copper(II) acetate appeared to be an optimum choice, typically delivering overall yields of up to 90% and a ratio of products of ~9:1.Further, the best reaction conditions we found in these initial studies were somewhat inconvenient, consisting of heating the copper salt with a precursor such as ester 6a in a mixture of diethyl ether and pyridine held in a sealed tube at around 90 o C for a few hours.This somewhat bizarre solvent system was suggested by similar mixtures which are used in some of the classical acetylene coupling reactions; 8 our reasoning was simply that under such conditions, alkynes are probably activated by metal complexation, with copper ions at least, and hence that these might be effective in triggering the desired type of cyclisation.We were very surprised to isolate to intermediate hydroxy-dihydropyrroles such as 14a, as these have not been previously recorded and would be expected to very easily lose water to give the corresponding pyrroles (e.g.15a).This is certainly the case in related furan syntheses, when no traces of such dihydro intermediates were observed, despite a number of attempts. 3,9owever, such dihydrofuran species have been prepared and used in synthesis, for examples, as substrates in Claisen rearrangements. 10While these hydroxy-dihydropyrroles 14 will very likely turn out to be useful intermediates, their conversions into the corresponding pyrroles 15 were not, in our hands, completely efficient, despite trying many methods involving exposure to acid 7,11 or various derivatisations followed by base-induced elimination.Although the major diastereoisomer shown from the Kazmaier method 6 is not particularly well set up for such eliminations by reason of its syn-stereochemistry, we thought that the thermodynamic driving force of aromatization would be quite sufficient to trigger such transformations; evidently such eliminations were not so facile in these cases.We therefore wished to further develop this chemistry with the duel aims of rendering it a one-pot, two-step sequence of cyclisation and dehydration, thus avoiding a separate elimination step, and also of finding much simpler, greener and more convenient reaction conditions.Herein, we report a successful outcome of this study.
We first carried out a simple screen, using the representative substrate 6a and a higher boiling solvent, to determine how essential the use of ether was and if we could remove the requirement of using a sealed tube (Table 1).Only toluene proved to be useful, giving complete conversion although predominantly to the hydroxy-dihydropyrrole 14a along with much smaller amounts of the pyrrole 15a, in direct contrast of the failures associated with the other solvents tested.Ethyl acetate, ethanol and dimethylformamide were partly successful but were clearly not solvents of choice.Hence, at least, the need for a sealed tube was obviated as the use of a toluene-pyridine combination permitted the reaction mixture to be heated to a necessary temperature under a reflux condenser.We therefore initiated a more extensive systematic study of all features of the cyclisation, the results of which are presented in Table 2.
The original sealed tube conditions in ether (Entry 1) 7 were found to have too long a reaction time; one hour was sufficient (Entry 2), although the reaction was significantly slower in hot toluene (Entry 3) but at least the drawback of the sealed tube requirement had been obviated and the yield was very similar to those which had been obtained using ether; the cyclisation was also successful using the phenyl-substituted precursor 6b (Entry 4), which also gave significantly more pyrrole 15b, presumably because of greater conjugation.We were surprised to find that lowering the amount of pyridine to only two equivalents, based on the amount of precursor 6a, resulted in an equally efficient cyclisation in refluxing toluene, but gave a product containing much more pyrrole 15a (Entry 5).At the same time, we found that copper(II) acetate, which we presumed might be more stable with respect to its oxidation state, was equally effective in triggering the cyclisation and also gave a preponderance of pyrrole 15a (Entry 6).By contrast, copper(I) iodide gave no trace of cyclised products in two runs (Entries 7 and 8).We then questioned the need for including any pyridine at all and were delighted to find that it was indeed unnecessary (Entry 9) when we reverted back to using copper(II) acetate as the catalyst.A considerable bonus was that the product now consisted entirely of the pyrrole 15a, which could be isolated in a pure state in around 93% yield.The eight hour heating period was also unnecessary: simply refluxing the reaction mixture for an hour produced the same result (Entry 10).Finally, we checked that the copper acetate could indeed be used 'catalytically' as, up to this stage, we had routinely used one equivalent of this reagent.Happily, we found that only 10mol% was perfectly adequate to drive the combined cyclisation and dehydration steps to completion during one hour at reflux (Entry 11).That this temperature was required to achieve complete reaction in such a short time was confirmed when a reaction carried out at 70 o C overnight only resulted in 80% conversion to the pyrrole 15a (Entry 12).Having established what appeared to be an optimum set of conditions, we turned to an exploration of application of these to a more general series of cyclisation-dehydrations using a diversity of precursors, the results of which are set out in Table 3.The initial optimisation substrate 6a, when reacted on a preparative scale, gave the desired pyrrole 15a in a 98% isolated yield, without the need for any purification beyond a filtration through a short pad of silica gel (Entry 1).Similarly, the 5-phenyl pyrrole 15b was formed slightly more slowly but in comparable yield (Entry 2).Cyclisation of the enyne (entry 3) was strangely retarded, as the previous conjugated substrate 6b (Entry 2) also contained a conjugated alkyne and reacted more rapidly.The example with a terminal alkyne (entry 4) also required a much longer reaction time and one equivalent of catalyst to secure a good yield.Protection of a potentially competing 5endo-dig cyclisation of a hydroxyl group was not necessary (Entries 5 and 6) but was needed in the case of competing 5-exo-dig or 6-endo-dig cyclisations (Entry 7).Although the alternative products proved too sensitive to isolate and characterize, the attempt to cyclise the free alcohol derivative 6h gave only a 30% return of the expected pyrrole 15h (Entry 8).
2,4-Disubstituted pyrroles (entries 9 and 10) can also be prepared using this method; the necessary precursors to these 'pseudo-symmetrical' pyrroles were prepared by a double addition of the corresponding lithio-alkyne to N-tosyl ethyl glycinate.Unfortunately, the completely substituted precursors (entries 11 and 12) failed to cyclise despite prolonged heating and so it would seem that this method may be limited to the synthesis of products which do not contain adjacent substituents.Most of the entries contain an ester group, which is positioned at the 2position of the final pyrroles 4.This was due purely to synthetic expedience as the precursors 1 were readily prepared using the Kazmaier method. 6To demonstrate that the ester group is not necessary for cyclisation, the two alkyl-substituted precursors (entries 13 and 14) derived from commercial 2-amino-1-butanol were found to undergo comparably rapid and clean cyclisations when heated in toluene.This should mean that this methodology will find plenty of applications as the elaboration of such relatively simple pyrroles is often beset by poor yields.
Although many of the foregoing cyclisations were carried out using mixtures of stereoisomers, no difference in the rates of cyclisation of such pairs were observed under the reaction conditions used.At lower temperatures or perhaps by an earlier work up, it might be possible to observe such a rate difference.b One equivalent of catalyst was used; with 10 mol%, the yield was 50% after 12 h.
Similar cyclisations under the present conditions using carbamate derivatives in place of sulfonamides were unsuccessful in the case of both methoxycarbonyl and N-Boc derivatives 16 (Scheme 8); no traces of either the hydroxyl-dihydropyrroles 17 or the derived pyrroles 18 were observed by NMR analysis of the crude products, which were essentially unchanged starting materials.Although it is not clear why this is, it may be associated with the higher pKa values of such derivatives.To check that the catalyst was indeed necessary, the first two substrates 6a and 6b (entries 1 and 2) were separately heated in toluene alone when no cyclisation was observed.We are not certain of the exact nature of the catalyst: although high purity copper(II) acetate was used throughout, it is of course possible that small amounts of copper(I) are responsible for the catalytic effect.However, the greater electrophilicity of copper(II) ions suggest that these would be more effective catalyst species.Finally, we should point out that this type of cyclisation is not entirely new.In related work, a Japanese group have reported that copper(II) acetate in hot toluene is a most effective catalytic system for the cyclisation of sulfonamide derivatives of 2-alkynylanilines to give indoles. 12In similar fashion, they too found that the corresponding carbamate derivatives did not cyclise anything like as efficiently using this system.Although the copper catalyst was not recovered, it very likely could be and then reused.Hence, overall, this can be regarded as a very environmentally friendly procedure for pyrrole synthesis.Of course, as it usually the case, this only contributes one 'clean' step to the overall sequence required to reach the final pyrroles; as ever, there are necessary prior steps which produce the routine levels of waste, so there is still much to do!

Experimental Section
General.NMR spectra were recorded using a Bruker DPX spectrometer operating at 400 MHz for 1 H spectra and at 100.6 MHz for 13 C spectra respectively.Unless stated otherwise, NMR spectra were measured using dilute solutions in deuteriochloroform.All NMR measurements were carried out at 30 °C and chemical shifts are reported as ppm on the delta scale downfield from tetramethylsilane (TMS:  0.00) or relative to the resonances of CHCl3 (H 7.27 ppm in proton spectra and C 77.0 ppm for the central line of the triplet in carbon spectra respectively).Coupling constants ( 3 JHH) are reported in Hz.Infrared spectra were recorded as thin films on sodium chloride plates for liquids and as KBr disks for solids, using a Perkin-Elmer 1600 series FTIR spectrophotometer and sodium chloride plates.Low resolution mass spectra were obtained using a VG Platform II Quadrupole spectrometer operating in the electron impact (EI; 70 eV) or atmospheric pressure chemical ionization (ApcI) modes, as stated.High resolution mass spectrometric data was obtained from the EPSRC Mass Spectrometry Service, University College, Swansea, using the electrospray ionisation (ES) mode unless otherwise stated.Melting points were determined using a Kofler hot stage apparatus and are uncorrected.Elemental analyses were obtained using a Perkin Elmer 240C Elemental Microanalyser.All reactions were conducted in oven-dried apparatus under an atmosphere of dry nitrogen unless otherwise stated.All organic solutions from aqueous work-ups were dried by brief exposure to dried magnesium sulfate, followed by gravity filtration.The resulting dried solutions were evaporated using a Büchi rotary evaporator under water aspirator pressure and at ambient temperature unless otherwise stated.Column chromatography was carried out using Merck Silica Gel 60 (230-400 mesh).TLC analyses were carried out using Merck silica gel 60 F254 precoated, aluminium-backed plates, which were visualized using ultraviolet light or potassium permanganate or ammonium molybdenate sprays.Ether refers to diethyl ether and petrol to the fraction boiling 60-80C unless stated otherwise. 5o a stirred solution of a 1-alkyne (1.0 equiv.) in dry tetrahydrofuran (3 ml mmol -1 ) maintained at -78 o C was added butyl lithium (2.5 M in hexanes, 1.0 equiv.)and the resulting solution stirred for 0.5 h.Anhydrous dimethylformamide (2.0 equiv.) was then added dropwise, the cooling bath was removed and stirring continued for 0.5 h.The reaction mixture was then poured into a vigorously stirred, ice-cold 1:1 v/v biphasic mixture of 10% aqueous potassium dihydrogen phosphate (4.0 equiv.)and ether.The organic layer was separated and the aqueous layer extracted with an additional portion of ether.The combined ether solutions were dried, filtered and evaporated at ambient temperature.The residue was taken up in 10% ether-hexanes and filtered through a short pad of silica gel.Evaporation of the filtrate left the pure alkynal, which could be further purified by distillation.Hept-2-ynal (5a). 13By the general procedure, condensation between 1-hexyne (4.20 ml, 36.3 mmol) and dimethylformamide (5.65 ml, 73.2 mmol) in tetrahydrofuran (120 ml) followed by quenching with potassium dihydrogen phosphate (25.29 g, 145.2 mmol) in water (250 ml) gave the ynal 5a, obtained as a pale orange oil (2.80 g, 66%), b.p. 80 o C (kugelrohr oven temp.) at 0.6 mm Hg, H (400 MHz, CDCl3) 9.05 (app.s, 1H, CHO), 2.29 (t, 3 JHH 7.3 Hz, 2H, 4-CH2), 1.45 (quint., 3 JHH 7.3 Hz, 2H, 5-CH2), 1.35 (sext., 3 JHH 7.3 Hz, 2H, 6-CH2), 0.81 (t, 3 JHH 7.3 Hz, 3H, 7-Me). 13-Phenylprop-2-ynal (5b). 13By the general procedure, starting with phenylacetylene (3.23 ml, 29.0 mmol), the ynal 5b was obtained as an orange oil (2.81 g, 74%), b.p. 190 o C (oven temp.) at 0.6 mm Hg, H (400 MHz, CDCl3) 9.30 (app.s, 1H, CHO), 7.49-7.26(m, 5H).

N-Sulfonyl--hydroxy--amino ester synthesis using tin(II) enolates of glycinates and other
precursor syntheses: General procedure. 6o a stirred solution of lithium diisopropylamide (2.2 equiv.) in tetrahydrofuran (4-5 ml mmol -1 ) maintained at -78 o C was added a solution of ethyl N-tosylglycinate (1.00 equiv.) in tetrahydrofuran (1.0 ml mmol -1 ) followed by a solution of tin(II) chloride (2.50 equiv.) in tetrahydrofuran (0.50 ml mmol -1 ).The resulting mixture was stirred at the same temperature for 0.5h before the dropwise addition of an ynal or ynone (1.10 equiv.) in tetrahydrofuran (1.0 ml mmol -1 ).Stirring was then continued overnight without further addition of coolant.The resulting cloudy yellow solution was then quenched using phosphate buffer (pH 7), filtered through a pad of silica gel and the bulk of the tetrahydrofuran was evaporated.The aqueous residue was extracted with ether (3 × equal volume) and the combined organic solutions dried, filtered and evaporated.The

Table 1 .
An initial solvent study