An efficient and general synthesis of pyrroles

Article abstract of DOI:10.1016/j.tetlet.2011.02.085

We have discovered that a wide range of 3-alkynyl-hydroxyalkanamine derivatives undergo 5-endo-dig cyclisations when exposed to silver nitrate supported on silica gel. Subsequent in situ dehydration of the resulting and sometimes isolable hydroxy-dihydrop

Full text of DOI:10.1016/j.tetlet.2011.02.085

Tetrahedron Letters 52 (2011) 2320–2323
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
An efficient and general synthesis of pyrroles
Christopher M. Sharland, Jirada Singkhonrat, Muhammad NajeebUllah, Simon J. Hayes,
⇑
David W. Knight , Damian G. Dunford
School of Chemistry, Cardiff University, Main College, Park Place, Cardiff CF10 3AT, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
We have discovered that a wide range of 3-alkynyl-hydroxyalkanamine derivatives undergo 5-endo-dig
cyclisations when exposed to silver nitrate supported on silica gel. Subsequent in situ dehydration of the
resulting and sometimes isolable hydroxy-dihydropyrroles leads to pyrroles in essentially quantitative
yields using this recoverable and reusable heterogeneous catalyst.
Received 9 December 2010
Revised 26 January 2011
Accepted 18 February 2011
Available online 4 March 2011
Ó 2011 Elsevier Ltd. All rights reserved.
Keywords:
Pyrroles
Synthesis
5-endo-Dig
Silver(I)
Heterogeneous catalyst
Our discovery that 3-alkyne-1,2-diols 1 could be very efficiently
cyclised and dehydrated to give outstanding yields of furans 2
(Scheme 1) when exposed to 10% w/w silver nitrate adsorbed on
silica gel led us to speculate that other nucleophiles could be sim-
ilarly induced to undergo such intramolecular additions to suitably
positioned alkyne groups.1 Herein, we report that a number of
modified amino groups, in particular sulfonamides, are especially
suitable nucleophiles in this respect and thus we have been able
to extend this idea to a new and fairly general synthesis of pyrroles.
Pyrroles have occupied a central position in modern chemistry
in general since its inception over one hundred years ago, initially
by reason of their occurrence in the key molecules of life,2 chloro-
phyll and haemoglobin and, more recently, because of their role as
components of pharmaceuticals (currently, one of the best selling
drugs, Lipitor, has a pyrrole at its core) and of many important
materials, such as organic semi-conductors.3 Despite this length
of time and the development of many ‘classical’ and more specia-
lised ways of synthesising pyrroles, there remain many deficiencies
in these and more modern methods of synthesis.4 It seems remark-
able how often, even now, that a report of an elegant application of
a particular pyrrole is preceded by a rather inefficient substrate
synthesis, which often delivers an initial product having unwanted
functionality, the presence of which is essential to the success of
the pyrrole synthesis rather than being desired in the target. This
alone provided sufficient motivation to attempt to develop an effi-
cient and hopefully widely applicable pyrrole synthesis, based on
the chemistry shown in Scheme 1. In addition, if successful, such
methodology could turn out to satisfy many of the requirements
of an environmentally friendly procedure.
In general, Baldwin-favoured 5-endo-dig cyclisation reactions
have proven useful synthetic tactics in a number of areas.5 The
most notable feature of the furan synthesis shown in Scheme 1 is
the fact that the cyclisations are triggered by a recoverable and
reusable heterogeneous catalyst. We naturally chose to attempt
to employ similar catalysts for the projected pyrrole syntheses, ini-
tially using the anti-propargylic glycinates 4, which we had a con-
siderable supply of from our previous work on their synthesis
using the Kazmaier method of condensing tin(II) enolates of glyci-
nates with aldehydes and ketones, in the present cases, the corre-
sponding alkynyl derivatives 3 (Scheme 2).6
R2
OH
R2
Initial attempts using various palladium-based catalysts were
partly successful in producing pyrroles, but in variable yields as
R3
Ag(I)
H2O
+
R1
>95%
R1
R3
O
OH
1
2
OH
R2
CO2Me
O
ref. 6
R1
R1
R2
Scheme 1.
TsHN
3
4
⇑
Corresponding author. Tel.: +44 2920 874210; fax: +44 2920 874030.
E-mail address: knightdw@cardiff.ac.uk (D.W. Knight).
Scheme 2.
0040-4039/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2011.02.085

C. M. Sharland et al. / Tetrahedron Letters 52 (2011) 2320–2323
2321
rather impure products; similar observations have been reported
by a Dutch group.7 Mercury(II) salts were very successful in pro-
viding b-mercuriopyrroles and thence metal-free pyrroles, an
unsurprising result in view of a past use of this methodology in a
synthesis of the natural product (À)-Preussin.8 Of course, the use
of stoichiometric mercury salts is unacceptable these days. How-
ever, more recent publications show that mercury(II) triflate is
emerging as an extremely viable and versatile catalyst for trigger-
ing cyclisations of this and many other types; a recent develop-
ment of a supported version is especially notable.9 However, we
then wondered why we had tried any alternatives to the silver(I)
method, as this had performed so well in the furan synthesis
(Scheme 1). In the event, we were not disappointed: upon expo-
sure to 10 mol % of 10% w/w silver nitrate on silica gel in dichloro-
methane at ambient temperature overnight in the dark (!),
essentially quantitative yields of the corresponding pyrroles 5
were obtained after sequential filtration through a plug of silica
gel and evaporation (Table 1).
the alternative of adding protection and deprotection steps for,
however efficient and simple these might be, there would still be
an inevitable requirement for additional quantities of reagents
and solvents. Although hardly relevant here, on a larger scale, it
would conceivably be possible to recover the extra equivalent of
alkyne before or after the cyclisation. Oddly, a few attempts to
deprotonate the keto-sulfonamides 7 by prior treatment with
one equivalent of base (NaH or BuLi), followed by the addition of
one equivalent of the acetylide 6 were not successful. It is unclear
why this was; this aspect evidently requires further investigation.
The amino-ketones 7 were obtained using Jones or PCC oxidations
of the corresponding hydroxytosylamides, which were in turn
readily obtained by amino-hydroxylations of the related alkenes
using chloramine-T.10
To our delight, cyclisations of this series of substrates 8 proved
in general to be equally as efficient and well-behaved: the results
are collected in Table 2. Both initial precursors (entries a and b),
ultimately derived from stilbene, provided excellent yields of the
expected pyrroles 9a,b, with the minor complication that the distal
O-silyl group was partly removed. However, as the two subsequent
examples reveal (entries c and d), such silyl groups, when present
as derivatives of propargylic alcohol, are not so sensitive and are
thus fully retained. The exact reason(s) for this are unclear at
present.
Happily, products 5 were all remarkably clean: given a chemi-
cally pure precursor 4, we were unable to detect the generation
of any impurities during the cyclisation. Evidently, both alkyl and
aryl substituents are acceptable in the eventual 5-position (entries
a and b), as is a more sensitive propenyl substituent (entry c);
hence, this method should be applicable to the synthesis of
a-
vinylpyrroles in general. Despite being much more crowded, the
two ketone-derived precursors (entries d and e) both underwent
equally smooth conversions into the corresponding trisubstituted
pyrroles 5d,e.
Cyclisations onto terminal alkynes are also perfectly viable (en-
tries e–g), leading to equally excellent yields of the corresponding
2,3-disubstituted pyrroles 9e–g. This is in marked contrast to the
related furan syntheses (Scheme 1)1 wherein such cyclisations
were entirely unsuccessful in examples having terminal alkynes.
A possible explanation is the comparatively lower pKa value of a
tosylamide group, which would allow for easier proton loss and
hence relatively more facile cyclisations. The last two examples
(entries h and i) were derived from commercial 2-amino-1-phe-
nyl-1-propanol and emphasise a possible drawback to the method,
that of precursor synthesis: not all of the necessary amino alcohols
or ketones will be especially easy to synthesise. On the positive
side, it was readily possible to recover the catalyst and to reuse
it, by simple filtration. Taking account of the obvious small losses
involved in such recovery, three subsequent runs showed no dim-
inution in catalytic activity.
Hence for these examples at least, this would seem to be an
exceptionally efficient and simple pyrrole synthesis. In order to ex-
tend the methodology and to check that the ester group was not
essential to the progress of the cyclisation, we prepared further
examples of suitable alkynyl tosylamides 8 by the addition of
either lithio- or magnesio-acetylides 6 to the protected amino-
aldehydes or ketones 7 (Scheme 3). These generally proceeded in
good to excellent yields; the stereochemical aspects of these and
related additions, along with the influence of precursor stereo-
chemistries on the actual cyclisations are discussed below. We
chose not to replace the labile sulfonamide N–H proton by a pro-
tecting group, but rather to use two equivalents of the alkyne as
this led to the desired products 8 in a single step. Further, it is argu-
able that this is a more atom-efficient tactic when compared with
One simple exception to any difficulties with precursor synthe-
sis can be used to prepare pseudo-symmetrical pyrroles 12 and
13.11 Thus, direct addition of acetylides, using over three equiva-
lents, to N-tosyl amino esters 10 gives excellent yields (90–95%)
of the expected bis-alkyne adducts 11 (Scheme 4).
Table 1
Initial cyclisations of propargylic glycinates 4
R2
CO2Me
OH
R2
CO2Me
R1
10% w/w AgNO3-SiO2
CH2Cl2, 20 °C, 16 h
Table 2
A generalised synthesis of N-tosylpyrroles 9
TsHN
R1
N
R2
R3
Ts
OH
R2
R3
R1
10% w/w AgNO3-SiO2
CH2Cl2, 20 °C, 16 h
4
5
R1
R1
R2
Yield (%)
TsHN
Entry
N
Ts
a
b
c
d
e
Bu
Ph
H
H
H
97
95
91
94
96
9
8
R1
Entry
R2
R3
Yield (%)
Me(:CH2)C
Ph
Ph
Me
a
b
c
d
e
f
g
h
i
Bu
RO(CH2)4
TBSOCH2
Bu
H
H
H
Bu
Ph
Ph
Ph
Ph
Ph
Me
TBSOCH2
Ph
Me
Me
Me
Me
100
96
93
98
100
95
95
97
97
iPr
a
Me
TBSOCH2
Ph
Me
Ph
OH
R2
R3
R2
R1
R1
Li(MgBr)
(2.2 equiv.)
O
+
R3
Ph
Ph
TsHN
TsHN
a
R = TBS or H: Pyrrole 9b was isolated as a ca. 1:1 mixture of the O-silyl deriv-
6
7
8
ative and the free alcohol; 96% refers to the combined yield after separation. TBS = t-
butyldimethylsilyl.
Scheme 3.

2322
C. M. Sharland et al. / Tetrahedron Letters 52 (2011) 2320–2323
R2
R2
Table 3
CO2Me
R1
OH
Alternative N-protecting groups
3.2 R2
Li(MgBr)
R2
R3
TsHN
OH
R1
R2
10% w/w AgNO3-SiO2
CH2Cl2, 20 °C, 16 h
R1
TsHN
10a) R1 = Me; R2 = Bu;
10b) R1 = H; R2 = Ph;
10c) R1 = H; R2 = Bu.
R3
R1
RHN
N
11
R
20
19
Scheme 4.
Entry
R1
R2
R3
R
Yield (%)
a
b
c
d
e
f
Bu
Bu
Bu
Bu
Bu
Ph
Ph
Ph
H
Ph
H
Me
Me
Et
Me
Et
4-Nsa
CO2Me
CO2Me
Boc
Boc
Boc
96
98
98
90
96
98
Upon exposure to the usual cyclisation conditions (Tables 1 and
2), these are then converted into the corresponding pyrroles 12 and
13 in essentially quantitative yields (Fig. 1).
Unsurprisingly, not every example tested was successful. In the
case of silylated alkynes 14 (Scheme 5), cyclisation appeared to oc-
cur only after desilylation, to give the 2,3-disubstituted pyrroles
15; no trace of the 5-silyl derivatives 16 was ever observed.
It may be that this is due to excessive steric hindrance, as the
similarly sterically impeded t-butyl acetylene 17 also failed to un-
dergo cyclisation when exposed to the silver reagent (Scheme 6).
Thus far, these are the only limitations which have come to light.
In contrast to most other N-tosyl derivatives, wherein cleavage
of this group can be quite challenging, in the case of pyrroles such
removals are essentially equivalent to an ester hydrolysis by rea-
son of the typical pKa value of this heteroaromatic system being
around 17.5. Hence, it is quite simple to convert all of the foregoing
products into the corresponding ‘free’ pyrroles.12 Despite this, we
have investigated the use of alternative N-protecting groups, in or-
der to extend the methodology to other systems, to give greater
precursor flexibility and also perhaps to define more atom efficient
groups, in terms of the by-products generated upon their removal.
Representative examples (Table 3) were readily prepared: the 4-
nosyl (Ns), methoxycarbonyl and Boc precursors 19a,b and d from
commercially available 2-amino-1-phenyl-1-propanol following
N-derivatisation, oxidation to the ketone and acetylide addition,
and the remainder 19c,e and f by the same three steps, but starting
from 2-amino-1-butanol. Despite some misgivings about the sta-
bility of these more acid-sensitive functions during prolonged
H
Et
a
4-NS = 4-Nitrophenylsulfonyl.
exposure to silica gel and silver nitrate, these all proved perfectly
stable and all delivered essentially quantitative yields of the ex-
pected pyrroles 20a–f.
It is not immediately obvious how these cyclisations work.
While it seems most likely that a 5-endo-dig cyclisation is initially
induced by silver to give a hydroxy-dihydropyrrole 21 (Scheme 7),
which then undergoes a presumably very facile dehydration, it is
conceivable that these steps could be reversed and that dehydra-
tion is the first step and is followed by cyclisation of the resulting
enyne or perhaps some other rearranged species such as an allene.
In an attempt to delineate this aspect, we took samples of the
cyclisation mixture leading to the N-nosyl derivative 20a and ana-
lysed these by proton NMR after a simple filtration. In this way, we
were able to confirm that the pathway shown in Scheme 7 is most
likely the sole route leading to the pyrroles 20; the relevant NMR
data are summarized in Figure 2. As much as 35–40% of the inter-
mediate 21 was observed in some instances.13 No doubt, the hy-
droxy group adjacent to the alkyne group assists complexation
with the silver(I) atoms, as suggested by the Pale group.14
In subsequent studies, it proved possible to isolate the surpris-
ingly stable hydroxy-dihydropyrroles (e.g., 21);13 this aspect will
be reported separately.
Bu
R
HO
OH
Ph
Me
Ph
Me
Ph
Me
-H2O
Bu
Bu
Me
R
N
N
Ts
Ts
NsHN
Bu
Bu
N
N
Ns
Ns
12 (97%)
13a) R = Ph (98%);
13b) R = Bu (98%).
19a
21
20a
Figure 1.
Scheme 7.
δ
H 5.0 (t)
R1
R2
R1
R2
OH
R1
δ
H 1.3 (d)
Pr
OH
Ph
Me
δH 3.8 (q)
HO
Me3Si
H
Ph
and not
R2
Me3Si
N
N
H
TsHN
Pr
H 2.4 and 2.6
Ts
Ts
NsHN
H
δ
H 2.2 (t)
Me
N
Ns
14
15
16
δ
H 4.55 (br. d) δH 3.55 (br. pent)
δ
δH 1.5 (d)
(both dt)
21
19a
Scheme 5.
δH 6.1 (~s)
H
Ph
Ph
OH
Ph
Pr
Me
N
Me
Ns
Me
N
TsHN
δH 1.55 (~s)
Ts
δ
H 2.7
(dist. t)
20a
17
18
Figure 2. Summary of the 1H NMR data for precursor 19a, hydroxy-dihydropyrrole
21 and pyrrole 20a (400 MHz; CDCl3).
Scheme 6.

C. M. Sharland et al. / Tetrahedron Letters 52 (2011) 2320–2323
2323
3. Lipitor: Oldfield, E. Acc. Chem. Res. 2010, 43, 1216–1226; Lindsley, C. W. ACS
Chem. Neurosci. 2010, 1, 407–408; Polypyrroles: Huang, J. X. Pure Appl. Chem.
2006, 78, 367–376; Huang, J. H.; Kaner, R. B. Chem. Commun. 2006, 367–376.
4. Bergman, J.; Janosik, T. In Comprehensive Heterocyclic Chemistry III(1995–2007);
Katritzky, A. R., Ramsden, C. A., Scriven, E. F. V., Taylor, R. J. K., Eds.; Elsevier: St.
Louis and Oxford, 2008. vol. 3.03, pp 269–351; Sundberg, R. J. In Comprehensive
Heterocyclic Chemistry II(1984–1996); Katritzky, A. R., Rees, C. W., Scriven, E. F.
V., Eds.; Elsevier: St. Louis and Oxford, 1996. Vol. 2.03, pp 119–206 and
references therein.; For an extensive listing of very recent methods of pyrrole
synthesis, see: Merkul, E.; Boersch, C.; Frank, W.; Müller, T. J. J. Org. Lett. 2009,
Although not pursued during the present studies, it is evident
from the excellent yields of pyrroles obtained that the two possible
diastereoisomers of the precursors 4, 8 and 19 undergo smooth
cyclisations, but probably at slightly different rates. This aspect
has been discussed previously in the context of our related furan
synthesis.1,15
Overall, this present pyrrole synthesis is arguably competitive
with any other currently available by reason that it proceeds in
essentially quantitative yields at ambient temperature, the only
by-product is water and both the solvent and catalyst can easily
be recycled and reused. It, therefore, satisfies many of the criteria
of a ‘green’ procedure. Of course, this statement does not take into
account all of the approach work required to prepare the necessary
precursors, much of which suffers from all of the usual drawbacks
and waste associated with many synthetic steps in current use. Of
all the alternative catalyst systems recently reported, probably
those based on the use of gold(III) are the most comparable.16
However, these are generally homogeneous species not easily
recovered and reused, in direct contrast to the present heteroge-
neous system. The present method also has clear potential for
use in various types of flow systems; this aspect will be reported
separately.
11, 2269–2272; For
a review of multi-component pyrrole syntheses, see:
Estevez, V.; Villacampa, M.; Carlos, J. Chem. Soc. Rev. 2010, 39, 4402–4421.
5. Knight, D. W. Prog. Heterocycl. Chem. 2002, 14, 19–51.
6. Gridley, J. J.; Coogan, M. P.; Knight, D. W.; Malik, K. M. A.; Sharland, C. M.;
Singkhonrat, J.; Williams, S. Chem. Commun. 2003, 2550–2551; Grandel, R.;
Kazmaier, U. Eur. J. Org. Chem. 1998, 1833–1840.
7. Van Esseveldt, B. C. J.; Van Delft, F. L.; Smits, J. M. M.; De Gelder, R.;
Schoemaker, H. E.; Rutjes, F. P. J. T. Adv. Synth. Catal. 2004, 346, 823–834; Van
Esseveldt, B. C. J.; Van Delft, F. L.; Smits, J. M. M.; De Gelder, R.; Rutjes, F. P. J. T.
Synlett 2003, 2354–2358; Esseveldt, B. C. J. V.; Vervoort, P. W. H.; Delft, F. L. V.;
Rutjes, F. P. G. T. J. Org. Chem. 2005, 70, 1791–1795.
8. Overhand, M.; Hecht, S. M. J. Org. Chem. 1994, 59, 4721–4722; For an Au(III)-
catalysed version, see: Gouault, N.; Le Roch, M.; Cornée, C.; David, M.; Uriac, P.
J. Org. Chem. 2009, 74, 5614–5617.
9. Yamamoto, H.; Sasaki, I.; Hirai, Y.; Namba, K.; Imagawa, H.; Nishizawa, M.
Angew. Chem., Int. Ed. 2009, 48, 1244–1246; Nishizawa, M.; Imagawa, H.;
Yamamoto, H. Org. Biomol. Chem. 2010, 8, 511–521. and references therein.
10. Herranz, E.; Sharpless, K. B. Org. Synth. Coll. Vol. 1990, 7, 375–381.
11. El-Taeb, G. M. M.; Evans, A. B.; Knight, D. W.; Jones, S. Tetrahedron Lett. 2001,
42, 5945–5948.
12. Haskins, C. M.; Knight, D. W. Tetrahedron Lett. 2004, 45, 599–601. and
references therein..
Acknowledgements
13. We have previously observed such intermediates in copper(I)-catalysed
cyclisations leading to pyrroles, see: Knight, D. W.; Sharland, C. M. Synlett
2004, 119–121; and Sharland, C. M. unpublished results, and from
iodocyclisations, see: Knight, D. W.; Rost, H. C.; Sharland, C. M.; Singkhonrat,
J. Tetrahedron Lett. 2007, 48, 7906–7910.
We are grateful to the EPSRC, to GSK Ltd. and the Government
of Thailand for financial support.
14. For an excellent summary of such silver-catalysed processes, see: Weibel, J.-
M.; Blanc, A.; Pale, P. Chem. Rev. 2008, 108, 3149–3173.
References and notes
15. Dunford, D.; Guyader, M.; Jones, S.; Knight, D. W.; Hursthouse, M. B.; Coles, S. J.
Tetrahedron Lett. 2008, 49, 2240–2242.
16. For an extensive overview of the use of ‘coinage’ metals in heterocyclic
synthesis, see: Patil, N. T.; Yamamoto, Y. Chem. Rev. 2008, 108, 3395–3442.
1. Hayes, S. J.; Knight, D. W.; Menzies, M. D.; O’Halloran, M.; Tan, W.-F.
Tetrahedron Lett. 2007, 48, 7709–7712.
2. Battersby, A. R. Proc. Royal. Soc. London., Series B-BioSci. 1985, 225,
1–26.