Pyrrole Chemistry. Part XII. The Mechanism of the Reaction Between the Pyrrole Grignard Reagent and Acylating Agents1 C. E. LOADER AND HUGH J. ANDERSON Department of Chemistry, Memorial University, St. Jolzn's, Newfoundlarzd Received October 5, 1970 Deuterium labeling experiments have shown that, in the reaction between the pyrrole Grignard reagent and methyl 1-pyrrolecarboxylate to form 1,2'-dipyrrolyl ketone, direct acylation occurs at the 2-position of the pyrrole Grignard reagent, rather than initial substitution onto the I-position, followed by rearrangement to the 2-position. At an intermediate stage in the formation of the ketone, ring interchange takes place between the 1-substituted pyrrole rings in the intermediate and unused Grignard reagent, and as a result both pyrrole rings in the final product may originate from the Grignard reagent. Des experiences de marquage par le deuterium ont montrc que, dans la reaction entre le reactif de Grignard du pyrrole et du pyrrolecarboxylate-1 de methyle pour former la dipyrrolyle-1,2' cctone, l'acylation directe se fait en position 2 sur le reactif de Grignard, plutbt qu'une substitution initiale en position 1, suivie d'un rearrangement en position 2. Au cours d'une Ctape intermediaire de la formation de la cctone, il se produit un Bchange de cycles entre le reactif de Grignard substituc en position 1 du pyrrole intermediaire et le reactif de Grignard non utilisk; il en rcsulte donc que les deux cycles pyrroles du produit final peuvent provenir du reactif de Grignard. Canadian Journal of Chemistry, 49, 1064 (1971) In most of its reactions with acylating agents, the pyrrole Grignard reagent 1 gives products resulting from substitution at the 2-position of the pyrrole ring (1). It was noted in the previous paper in this series (2) that the pyrrole Grignard reagent reacted with alkyl carbonates to give alkyl 1- pyrrolecarboxylates rather than the alkyl 2- pyrrolecarboxylates that might have been expected from the reaction. No reason for the selectivity of the acylation in this reaction is known. In this paper some work on the mechanism of the acylation reaction of the pyrrole Grignard reagent with esters is reported. The intermediate in the reaction of methyl carbonate with the pyrrole Grignard reagent is probably a salt similar to 2. The salt is hydrolyzed in the work-up of the reaction to give methyl 1-pyrrolecarboxylate (3), the main reaction product (Scheme 1). Although the same kind of intermediate seems likely from the reaction of ethyl acetate with pyrrole Grignard reagent the main product isolated from that reaction is not 1-acetylpyrrole but 2-acetylpyrrole (4) (Scheme 2) (3). A possible reaction mechanism for the formation of 2-substituted pyrroles in the Grignard reaction is that substitution onto the 1-position of the pyrrole ring occurs first and is followed by rearrangement to give 2- or 3-substituted 'For Part XI see ref. 2. products. Some evidence against this mechanism in the alkylation of the pyrrole Grignard reagent has already been reported (4). If this mechanism does, however, prevail, the formation of 1-substituted products may simply be explained as being cases in which the rearrangement to the 2-position of the pyrrole ring does not occur. Alternatively, initial substitution might take place onto the 2-position with migration onto the 1-position of the pyrrole ring being possible in some cases. This latter possibility seems unlikely as a carbon-carbon bond must be broken if 1-substituted products are to be formed. Finally, direct substitution onto both the 1- and the 2-positions may occur. A convenient method of studying the mechanism of acylation of the pyrrole Grignard reagent is offered in the case of the acylation of the reagent with methyl 1-pyrrole carboxylate. Here, if the initial attack is at the 1-position of the pyrrole ring, the intermediate 5 may be formed which then rearranges to finally give 1,2'-dipyrrolyl ketone (6). In the intermediate 5 the two pyrrole rings are indistinguishable, one of the rings being from the Grignard reagent and the other from the ester. Thus, if the intermediate 5 exists, both of the pyrrole rings should appear as the 2-substituted ring in the final product, 1,2'-dipyrrolyl ketone, with equal frequency. If the intermediate 5 does not exist and direct substitution onto the 2-position of the pyrrole ring in the Grignard reagent occurs, the pyrrole
LOADER AND ANDERSON: PYRROLE CHEMISTRY. XI1 rings do not become indistinguishable. Thus, the 2-substituted pyrrole ring in the final product will originate from the Grignard reagent and the I-substituted from the ester. In the latter case the probable intermediate involved in the reaction is the salt 7. 1 1 Meo-zN;OMgBr UJJ pared (4) and converted to the pyrrole-d, Grignard reagent. The d, Grignard reagent was allowed to react with the 1-pyrrolecarboxylate in the usual way and the 1,2'-dipyrrolyl ketone was isolated by careful fractional distillation of the reaction products. Also isolated in the distillations were pyrrole, methyl I-pyrrolecarboxylate and traces of 2,2'-dipyrrolyl ketone. The recovered pyrrole was a mixture of deuterated pyrroles as shown by mass spectrometry, but the recovered methyl 1-pyrrolecarboxylate was almost completely free of deuteration. The 1,2'-dipyrrolyl ketone was found to be a mixture of the d,, d, and undeuterated ketones (8, 9, and 6 respectively), the d, product predominating. The mass spectrum (Table 1) of the mixture D)mJD D I D, I lfih 7 D$H - In order to investigate the path of the re- - action of the pyrrole Grignard reagent with D D D methyl 1-pyrrolecarboxylate, pyrrole-d, was pre- 8 9
1066 CANADIAN JOURNAL OF CHEMISTRY. VOL. 49, 1971 TABLE 1. Mass spectra of deuterium labeled 1,2'-dipyrrolyl ketone* Ketone mle A B C D 'A, 1,2'-Dipyrrolyl ketone; B, from methyl I-pyrrole-d4-carboxylate and the pyrrole Grignard reagent; C, 1,2'-dipyrrolyl- 1'-d ketone D from methyl I-pyrrolecarboxylate and pgrrole-d4 Grignard reagent. of ketones showed that about 12% of the undeuterated product 6 was formed and about 19 % of the d7 compound 9 (any N-deuterium on the Zsubstituted ring having been equilibrated to natural abundance with water in the work-up). The remaining 70% of the product was the d3 compound 8. Clearly some form of 'ring interchange' takes place during the reaction. Because no tri- or tetradeuterio methyl 1- pyrrolecarboxylate is formed in the reaction, the possibility of ring interchange between the pyrrole rings of the ester and the Grignard reagent can be excluded. That the isomerization takes place in the mass spectrometer can be excluded for two reasons. The proton resonances of 1,2'-dipyrrolyl ketone could clearly be seen in the p.m.r. spectrum of the reaction product and it is difficult to devise a mechanism for the formation of the d7 ketone 9 in the mass spectrometer by rearrangement or by pyrolysis in the inlet system. In the latter case the most likely product is a d, ketone, which is not found, or the undeuterated product 6, which is present, but it is unlikely that any is formed in this way (see below). The results discussed above suggest that the exchange between the pyrrole rings takes place at some intermediate stage in the chemical reaction. The possibility that the ring interchange may involve the 1,2'-dipyrrolyl ketone was considered. However, it was found that interchange neither took place between the ketone and the pyrrole-d, Grignard reagent nor was there any between the ketone and pyrrole-d4 under the reaction conditions used. The d4 ketone peak (mle 164) is very small compared with the d, ketone peak (mle 163) and can be attributed almost entirely to natural abundance of isotopes in the d3 product. The presence of the d7 ketone 6 and the undeuterated ketone 9 can be readily explained on the basis of the intermediate 7, their formation arising from ring interchange between the molecules of intermediate and the pyrrole Grignard reagent. If the intermediate 7 is important, since little if any d, ketone is formed in the reaction, we must conclude that the deuterated ring does not interchange with the undeuterated ring although it might interchange with the Grignard reagent. A more likely possibility is that ring interchange occurs between the undeuterated ring and the Grignard reagent or with free d, pyrrole (present due to the incomplete formation of the Grignard reagent). In this way the formation of the d, ketone and undeuterated pyrrole can be explained. The ring interchange will, at the same time as giving the d7 intermediate (and thus the d7 ketone 9), give undeuterated pyrrole Grignard reagent which can react with unused methyl 1- pyrrolecarboxylate to give the deuterium free ketone 6 in the product. Because 1,2'-dipyrrolyld4 ketone is not formed in the reaction in an amount equal to that of the 1,2'-dipyrrolyl-d, ketone, for the reasons given above the most likely intermediate in the reaction is similar to 7. Thus, direct attack at the 2-position of the pyrrole Grignard reagent by the methyl l-pyrrolecarboxylate predominates over attack at the 1- position followed by migration to the 2-position. For further confirmation of this mechanism a sample of methyl 1-pyrrole-(2,3,4,5-d4)-carboxylate was prepared and allowed to react with the pyrrole Grignard reagent to give a
LOADER AND ANDERSON: PYRROLE CHEMISTRY. XI1 deuterated 1,2'-dipyrrolyl ketone with the deuteration having been introduced via the ester rather than via the pyrrole Grignard reagent as in the earlier case. The mass sdectrum of the ketone showed deuterated parent ions at mle 164, 163, and 160. The most abundant of these was the ion at mle 164 (31 %) and is clearly due to the rl, ketone 10 and the product to be expected if no ring interchange takes place. However, the ion mle 160 must be due to the undeuterated ketone 6 and indicates that ring interchange does occur. (The ion mle 163 is almost exclusively due to the presence of about 30% trideuterated methyl 1-pyrrolecarboxylate in the I -pyrrole-d4-carboxylate used to prepare the ketone.) In this case the formation of d7 ketone in the product requires not only ring interchange but reaction of the exchanged pyrrole Grignard reagent to give the d7 product. The abundance of the ions due to the ketone formed by this mechanism showed wide variations from reaction to reaction as might be expected. In contrast, the abundance of the ions due to the ketone resulting from ring interchange only, remained almost constant. There was no evidence to indicate that ring interchange involving the 2-substituted pyrrole ring in the intermediate took place. Once formed the carbon-carbon bond between the 2-substituted pyrrole ring and the rest of the intermediate is not broken. CO CO d H d D the to give the ions mle 94 and 66. The ion mle 94 is the base peak and is only a little more abundant than the ion mle 67. The fragmentation of the monodeuterio ketone 11 is dominated by the ions mle 95, 68, and 67. Since these ions are the same as those from the undeuterated compound + 1 we can assume that the fragments still contain the deuterium atom and thus that the fragmentation is dominated by the cleavage of the bond between the nitrogen of the l-substituted pyrrole ring and the carbonyl group. The main fragmentation paths for 1,2'-dipyrrolyl ketone are shown in Schemes 3 and 4. The deuterated 1,2'-dipyrrolyl ketone from the Grignard reaction of methyl l-pyrrolecarboxylate with the d4 pyrrole Grignard reagent showed parent ions at mle 160, 163, 164, and 167. The most abundant ion was at mle 97 and corresponds to the ion at mle 94 in the nondeuterated compound. It is a deuterated ion arising from the d3 and d7 ketones 8 and 9. Accurate mass measurements of the ions mle 160, 163, and 167 showed them to correspond to the deuterated ketones of formulae CgH,N20, CgH,D3N20, and C,HD7N20 respectively and confirms the nature of the reaction products, as shown in Table 2. The cyclic mechanism (Scheme 4) for the TABLE 2. High resolution mass measurement data for deuterated 1,2'-dipyrrolyl ketone from methyl I-pyrrolecarboxylate and pyrrole-d4 Grignard reagent 10 11 Ion mle Found mle Calculated mle Formula Mass Spectra (Table 1) 160 160.0640 160.06366 C9H8Nz0 The mass spectrum of 1,2'-dipyrrolyl ketone 6 163 163.0826 163.0825 C9H,D,N20 167.1076 C9HD7Nz0 shows that the parent ion (mle 160) fragments 167 167.1080
1068 CANADIAN JOURNAL OF CHEMISTRY. VOL. 49, 1971 formation of the mle 67 ion in the mass spectrum of 1,2'-dipyrrolyl ketone is supported by the mass spectra of the deuterated ketones. The dl ketone 11 shows an ion at mle 68 which replaces the mle 67 ion in the mass spectrum of the undeuterated ketone and indicates that the 1'-proton (deuteron) is present in these ions. Thus, the abundant mle 71 ion in the mass spectrum of the mixture of ketones from methyl 1-pyrrole-d4-carboxylate and the pyrrole Grignard reagent - arises from the 1'-substituted pyrrole-d4 ring in the ketone mixture along with the 1-proton from the 2-substituted pyrrole ring in the same molecule. By applying this mechanism to the fragmentation of the deuterated ketones it is possible to explain the presence in the mass spectra of all the ions corresponding to the mle 67 ion found in the mass spectrum of 1,2'-dipyrrolyl ketone. This work has shown that the acylation of the pyrrole Grignard reagent with methyl 1- pyrrolecarboxylate proceeds by direct attack at the 2-position of the pyrrole ring in the Grignard reagent rather than by initial attack at the 1- position followed by rearrangement to the 2- position. This is also likely to be true for the acylation of the pyrrole Grignard reagent by other esters excluding the alkyl carbonates and a few other acylating agents (5) where a different mechanism or combination of mechanisms must operate to allow for the formation of the l-substituted pyrroles found as the main reaction products. We are continuing our study of the pyrrole Grignard reagent with particular interest in the ring interchange reaction and its mechanism. Experimental High resolution mass spectra were recorded on a C. E. C. 21-llOB spectrometer using a direct inlet system. Gas chromatography was carried out on a Beckman GC-2A chromatograph equipped with a 13-112 in. column packed with Apiezon L on firebrick. The remainder of the equipment used has already been described (2). I,2'-Dipyrrolyl Ketone Pyrrole Grignard reagent uas prepared under nitrogen from magnesium (0.6 g), bromoethane (3.0 g), and pyrrole (1.67 g) in ether (20 ml) (2). The solution was cooled to room temperature and methyl 1-pyrrolecarboxylate (3.15 g) was added in ether (10 ml) with mechanical stirring. The stirred mixture was then heated under a reflux condenser for 2 h. It was then cooled ana dilute HCI (1 M) was added until all of the precipitated solid dissolved. The aqueous layer was separated from the organic layer and extracted several times with ether. The organic layer and the ether extracts were combined, washed with water, sodium bicarbonate solution, water, and finally dried (MgSO,). The ether was removed and the residue was fractionally distilled under vacuum. The distillate consisted of pyrrole, methyl l-pyrrolecarboxylate (1.2 g) and 1,2'-dipyrrolyl ketone (1.1 g, 27%). The dipyrrolyl ketone was recrystallized from petroleum, m.p. 61-62", lit. (6) m.p. 58.5-60". Full details of the physical properties of these products have already been reported (6). Reactiorl of' Methyl 1-Pyrrolecarboxylate with Pyrrole-d4 Grignard Reagent The pyrrole-d4 Grignard reagent was prepared from pyrrole-d, (7) as in the preparation of pyrrole Grignard reagent above. The reagent was then used in the reaction with methyl 1-pyrrolecarboxylate as in the previous experiment. The deuterated 1,2'-dipyrrolyl ketone was repeatedly recrystallized from petroleum, m.p. 60-61". The low boiling fractions contained pyrrole and methyl 1-pyrrolecarboxylate and were purified by gas chromatography at a column temperature of 103". The mass spectrum of the recovered methyl 1-pyrrolecarboxylate was almost identical to that of the methyl l-pyrrolecarboxylate used as starting material and showed no indication of deuteration. The recovered pyrrole was extensively deuterated and showed parent ions for all stages of deuteration up to pyrrole-d,. Accurate mass measurements on the main deuterated parent ions in the mass spectrum from the deuterated. 1,2'-dipyrrolyl ketone are summarized in Table 2. Reaction of the Pyrrole-d4 Grignard Reagent with I,a'- Dipyrrolyl Ketone Pyrrole-d4 Grignard reagent was allowed to react with 1,2'-dipyrrolyl ketone under conditions similar to those used in the preparation of 1,2'-dipyrrolyl ketone above. During the addition of the ether solution of 1,2'-dipyrrolyl ketone to the Grignard reagent the ether boiled and some solid was precipitated. However, on hydrolysis and work-up, the only product isolated was 1,2'-dipyrrolyl ketone which showed no deuteration in its mass spectrum. No ring interchange could be detected when 1,2'-dipyrrolyl ketone was refluxed in ether with pyrroled4 for several hours. Methyl I-Pyrrole-d4-carboxylate The ester was prepared by the method reported earlier (2) using pyrrole-d, Grignard reagent in the place of pyrrole Grignard reagent. The product consisted of a mixture of methyl 1-pyrrole-d3-carboxylate (30%) and methyl I -pyrrole-d4-carboxylate (70 %) as shown by mass spectrometry. The reasons for the formation of the large amount of the d3 ester have not yet been investigated. Reaction of Methyl I-Pyrrole-d4-carboxylate with the Pyrrole Grignard Reagent Pyrrole Grignard reagent was allowed to react with methyl 1-pyrrole-d4-carboxylate as for the preparation of 1,2'-dipyrrolyl ketone described above. A deuterated 1,2'-dipyrrolyl ketone was obtained from the reaction products, m.p. 60-61". The mass spectrum of this product is recorded in Table 1. 1,2'- Dipyrrolyl-1'-d Ketone The ketone was prepared by repeated deuterium exchange between 1,2'-dipyrrolyl ketone (500 mg) and mix-
LOADER AND ANDERSON: PYRROLE CHEMISTRY. XII 1069 tures of deuterium oxide (1 ml) and acetone (2 ml) (8). 2. C. E. LOADER and H. J. ANDERSON. Can. J. Chem. After four exchanges the product was about 75% 49, 45 (1971). deuterated as shown by the mass spectrum and was 3. G. P. BEAN. J. Heterocycl. Chem. 2, 473 (1965). suitable for the comparison spectrum used in this work. 4. P. S. SKEL~ and G. P. BEAN. J. Amer. Chem. Sot. 84. 4660 (1962). 5. E.' P. PAPADOPOULOS and H. S. HABIBY. J. Org. We are pleased to acknowledge the financial support Chem. 31, 327 (1966). E. P. PAPADOPOULOS. J. Org. of the National Research Council of Canada. We wish Chem. 31, 3060 (1966). to thank Dr. W. D. Jamieson of the Atlantic Regional 6. H. J. ANDERsoN and C. W. HUANG. Can. J. Chem. 459 Laboratory, N.R.C.C. for accurate mass measurements. 897 7. B. BAK. D. CHRISTENSEN. L. HANSEN. and J. RASTRUP- ANDERSEN. J- hem. Phys. 24, 720'(1956). 1. R. M. ACHESON. An introduction to the chemistry 8. E. FETZ and CH. TAMM. Helv. Chim. Acta, 49, 349 of heterocyclic compounds. Interscience, London, (1966). 1967. p. 71.