3-Hydroxypyrroles and 1H-pyrrol-3(2H)-ones. Part 14. Pyrolysis of oxazolidinylmethylene derivatives of Meldrum's acid - Synthesis of N-alkenyl-3-hydroxypyrroles and related reactions

Article abstract of DOI:10.1039/b109788a

Flash vacuum pyrolysis (FVP) of the title compounds 14 and 17 at 600-625°C (0.005 Torr) gives the N-alkenyl-pyrrolones 22 and 23 respectively. The mechanism is shown to involve hydrogen transfer and cyclisation of the methyleneketene intermediate (e.g. 25) to a fused pyrrolone (e.g. 28). This species fragments to create an azomethine ylide which provides the alkenyl substituent by a further hydrogen transfer. Similar reactions are shown by the thiazolidine derivatives 20 and 21, though in the latter case the initial bicycles 35 and 37 can be observed by NMR spectroscopy. When the normal sites for hydrogen transfer are blocked by substituents (as in compound 16) an alternative hydrogen transfer-cycloaddition sequence leads to the fused pyridin-4-one 43.

Full text of DOI:10.1039/b109788a

View Online / Journal Homepage / Table of Contents for this issue
3-Hydroxypyrroles and 1H-pyrrol-3(2H)-ones. Part 14.1,2
Pyrolysis of oxazolidinylmethylene derivatives of Meldrum’s acid –
synthesis of N-alkenyl-3-hydroxypyrroles and related reactions
1
Abd El-Aal M. Gaber,a,b Gordon A. Huntera and Hamish McNab*a
a Department of Chemistry, The University of Edinburgh, West Mains Road, Edinburgh,
UK EH9 3JJ
b Chemistry Department, Faculty of Science, Assiut University, Assiut 71516, Egypt
Received (in Cambridge, UK) 25th October 2001, Accepted 3rd December 2001
First published as an Advance Article on the web 28th January 2002
Flash vacuum pyrolysis (FVP) of the title compounds 14 and 17 at 600–625 ЊC (0.005 Torr) gives the N-alkenyl-
pyrrolones 22 and 23 respectively. The mechanism is shown to involve hydrogen transfer and cyclisation of the
methyleneketene intermediate (e.g. 25) to a fused pyrrolone (e.g. 28). This species fragments to create an azomethine
ylide which provides the alkenyl substituent by a further hydrogen transfer. Similar reactions are shown by the
thiazolidine derivatives 20 and 21, though in the latter case the initial bicycles 35 and 37 can be observed by NMR
spectroscopy. When the normal sites for hydrogen transfer are blocked by substituents (as in compound 16) an
alternative hydrogen transfer–cycloaddition sequence leads to the fused pyridin-4-one 43.
be N-alkenyl-3-hydroxypyrroles rather than the anticipated
Introduction
bicycles.2 Studies of the mechanism of this transformation and
The ‘Meldrum’s acid route’ to 3-hydroxypyrroles and 1H-
pyrrol-3(2H )-ones 1 (Scheme 1) has proved to be a flexible
the syntheses and pyrolyses of some related precursors are also
reported.
Results and discussion
The oxazolidines 5–8 and the related tetrahydro-1,3-oxazine 9
were prepared by standard methods (see Experimental section).
Similarly, the hexahydropyrimidine 10 and the thiazolidines 11
and 12 were synthesised so that the effect of the second hetero-
Scheme 1
atom on the thermal cyclisation could be assessed. The sub-
strates 5–7 and 9–12 were reacted with methoxymethylene
Meldrum’s acid 13 in acetonitrile solution at room temperature
to give the products 14–16 and 18–21 respectively, often in
>80% yield. 2,2-Dimethyloxazolidine 8 was hydrolytically
unstable under these conditions, but could be prepared and
reacted with 13 in situ to give a good yield of the Meldrum’s
derivative 17, and a similar method was used to make the
deuterium-labelled compound 17D. The products 14–21 were
characterised by their spectra. In some cases, the NMR spectra
showed that two rotamers were present in solution, as found
previously for N,N-disubstituted aminomethylene Meldrum’s
acid derivatives in which the two N-substituents have rather
similar steric requirements.3 In other cases broad peaks were
observed due to restricted rotation about the C–N bond.
Molecular ions were observed in the electron impact mass
spectra of each of the derivatives 14–21 (though some were of
low intensity) and peaks due to some or all of the normal
Meldrum’s acid breakdown pattern (sequential loss of acetone,
carbon dioxide and carbon monoxide) were detected in most
cases.
method for the preparation of 1-substituted or 1,2-disubsti-
tuted derivatives of this highly sensitive ring system.3,4 Using
flash vacuum pyrolysis (FVP) conditions, the products are
collected at low temperature in the absence of air and so the
high reactivity of the system to electrophilic reagents1 and to
oxidising agents5 can be controlled. N-Unsubstituted pyrrol-
ones cannot be made directly by this gas-phase route;6 neverthe-
less, these are attractive synthetic targets, because of their
relationship to the 3-alkoxypyrrole core of the prodigiosin
series of antibiotics.7,8 Although the pyrolysis of some simple
N-protected Meldrum’s acid precursors proved unsuccessful,9
we anticipated that oxazolidinyl derivatives 2 could provide the
bicycles 3 as primary pyrolysis products which could release the
N-unsubstituted pyrrolone 4 functionality upon hydrolysis
(Scheme 2). As we report here, pyrolyses of these substrates
proceeded cleanly, but unexpectedly, and the products proved to
Pyrolysis of the oxazolidine derivative 14 under FVP condi-
tions at 600–625 ЊC proceeded relatively cleanly to give an
unstable involatile oil which condensed at the mouth of the
trap. The 1H NMR spectrum of the crude pyrolysate (deuterio-
chloroform) clearly showed that a pyrrolone had been formed
(from the characteristic10 widely spaced doublets at δH 7.94 and
5.15) but a two-proton singlet at δH 3.82 was indicative of the
product being a 2-unsubstituted pyrrolone. This assignment
Scheme 2
548
J. Chem. Soc., Perkin Trans. 1, 2002, 548–554
This journal is © The Royal Society of Chemistry 2002
DOI: 10.1039/b109788a

View Online
Scheme 3
first mechanism (route a) involves hydrogen abstraction by the
methyleneketene 25 from the methyl group followed by con-
certed loss of CH2O to generate the azomethine ylide 26 with
the alkenyl substituent already in place. The product 23 is
formed by electrocyclisation of 26. Alternatively, pyrrolone
formation by route b could take place by the standard hydro-
gen transfer mechanism3 to give an alternative azomethine
ylide 27 and hence the expected bicycle 28, which can give the
final product 23 by subsequent loss of CH2O and consoli-
dation of the azomethine ylide 29. There is precedent for
the final step of this sequence since loss of formaldehyde
from oxazolidines under FVP conditions is a known route to
azomethine ylides.12
The two mechanisms (Scheme 3) can be distinguished by
identifying the source of the hydrogen atom in the 4-position of
the final product (specified in the Scheme 3); in route a, it comes
from the methyl substituent of the oxazolidine, whereas in route
b it is transferred from the 4-position of the oxazolidine ring.
The deuteriated substrate 17D was therefore synthesised and
the labelled 23 which was obtained from the pyrolysis revealed
no deuterium incorporation at the 4-position of the ring by 1H
NMR spectroscopy. In addition the signal due to the proton(s)
at the 2-position integrated to ca. 1H. Hence the concerted
mechanism (route a) can be discounted and instead the product
is apparently formed via collapse of the bicycle 28 (route b). In
the experiment using the labelled substrate, the ratio of minor
product 24 had substantially increased (deuteriated 23 : 24 = 1.3
: 1.0) which suggests that both products may be formed from
the same intermediate and that the final hydrogen shift to gen-
erate 23 from 29 may be rate determining. In this interpretation,
kH/kD = 2.1, which is in line with other values of kinetic isotope
effects under FVP conditions.13 The route to 24 could therefore
involve collapse of the azomethine ylide 29 to the aziridine 30
followed by ring opening (Scheme 4). Thermal partition
between azomethine ylides, aziridines and enamines has been
reported.12
was confirmed by recording the spectrum in [2H6]DMSO, which
promoted tautomerisation to the enol (hydroxypyrrole) form
and showed three signals for the pyrrole ring protons. An
unexpected one proton alkene signal was present in spectra
recorded in both solvents. The mass spectrum showed a parent
ion at m/z 163 (C10H13NO) consistent with the loss of CH2O
during the pyrolysis over and above the expected Meldrum’s
acid breakdown. This information is consistent with the prod-
uct being the N-alkenylpyrrolone (hydroxypyrrole) 22, obtained
in ca. 70% yield. N-Alkenylpyrroles are generally obtained by
reaction of ketoximes with acetylenes under basic conditions,11
but the present method provides the first route to N-alkenyl-
pyrrolones.
Pyrolysis of the dimethyloxazolidine 17 also provided the
corresponding alkenylpyrrolone 23 (ca. 50%) (see Experimental
section) but in this case a second product was identified as the
2-methylene isomer 24 (ca. 20%). The ratio of the two products
(2.7 : 1) was essentially unchanged over the pyrolysis temper-
1
ature range 500–750 ЊC. The H NMR spectrum of the minor
product showed typical pyrrolone signals but instead of the
3
usual characteristic doublets, a triplet (δH 7.59; J 3.6 Hz) and
double doublet (δH 5.33; 3J 3.6 and 4J 1.6 Hz) was obtained as
well as a one-proton broad signal (NH); the extra couplings are
explained by the presence of the NH.
Assuming that the pyrolyses of the Meldrum’s acid deriv-
atives give methyleneketene intermediates4 there are two reason-
able mechanisms for the formation of the alkenylpyrroles,
exemplified in Scheme 3 for the case of the formation of 23. The
No tangible products were obtained from the pyrolysis of the
phenyl-substituted compound 15. It is possible that a 2-
substituted N-alkenylpyrrolone was formed, but that it was too
J. Chem. Soc., Perkin Trans. 1, 2002, 548–554
549

View Online
Scheme 4
readily oxidised5 to allow its isolation under our usual
conditions.
In view of these results, the six-membered ring tetrahydro-
oxazinyl derivative 18 was pyrolysed. In this case, elimination of
formaldehyde cannot occur from the initial cyclisation product
to give an azomethine ylide and indeed the initial cyclisation
product 31 was thermally stable and could be identified from
the characteristic pyrrolone and hydroxypyrrole peaks in the
complex 1H NMR spectrum of the crude pyrolysate. Best
results were obtained when the pyrolysate was condensed on a
‘cold-finger’ which was washed with [2H6]acetone and trans-
ferred directly to an NMR tube in the absence of air. Such
2-substituted pyrrolones are readily oxidised and both the
pyrrolone 31 and its oxidation product 32 were subsequently
characterised by mass spectrometry (see Experimental section).
In contrast, pyrolysis of the tetrahydropyrimidine 19 was
unsuccessful.
Pyrolysis of the thiazolidine 20 gave, as the major product,
the same alkenylpyrrolone 22 which was obtained from the
corresponding oxazolidine 14. Clearly extrusion of thioform-
aldehyde from the bicyclic intermediate corresponding to 28
can take place by a similar mechanism. However, this process
is less facile than in the oxazolidine series, since a second
(unidentified) N-substituted pyrrolone (ca. half the amount of
the major product) was also formed in competition. In addi-
tion, bicyclic products could be characterised from the spectra
of the products formed by pyrolysis at 600 ЊC of the Meldrum’s
acid derivative 21 of the parent thiazolidine. Use of the ‘cold-
finger’ trap was essential and the unstable products were identi-
fied by their NMR spectra in [2H6]acetone at Ϫ20 ЊC. Three
products 35, 37 and 40 were formed in 55 : 18 : 27 ratio (Scheme
5). The major bicycle was identified as the pyrrolo[2,1-b]-
thiazole 35 by the characteristic pyrrolone resonances (δH 7.97
Scheme 5
favourable than hydrogen transfer from the bridgehead, as
found in the formation of 24 from 30. The regioselectivity
observed in the formation of 35 and 37 may be due to addi-
tional stabilisation of the dipolar intermediate by interaction
with the adjacent sulfur atom (as in 34A) but preferential cleav-
age of 37 to 38 cannot be ruled out at this stage. The relative
stability of 35 and 37 over the oxazolidine-derived products
reported in this paper, may be due to the longer C–S bonds
which impose less strain on the bicycle. It is of interest that
none of the hydroxy tautomers of 35 and 37 were observed in
[2H6]acetone solution, which may also be due to the increase in
strain of the more planar bicycle.
Next, the substrate 16 was designed to discover whether the
hydrogen transfer mechanism of Scheme 3 route a, could take
place when no hydrogen atoms α- to the nitrogen atom are
present. FVP of 16 at 625 ЊC, gave a single product (Mϩ 221
Da) in 50% yield, which was isomeric with the methyleneketene
intermediate 41 (i.e. no loss of CH2O had taken place). The 1H
NMR spectrum showed the presence of an enaminone unit
(δH 6.98 and 4.99) whose vicinal coupling constant (3J 7.1 Hz) is
indicative of a six-membered ring (rather than a five-membered
ring) framework. The 13C NMR (DEPT) spectrum showed the
presence of three quaternaries, three methine, five methylene
and the two methyl groups, which suggests that hydrogen atom
transfer to the methyleneketene from the cyclohexane ring has
taken place to generate the enone unit and that subsequent
3
and 5.25; J 3.6 Hz) the singlet due to the 7a-proton (δH 4.71)
and the four signals due to the non-equivalent protons at the
2- and 3-positions (δH 2.7–4.1). The minor bicycle showed simi-
3
lar alkene signals (δH 8.10 and 5.35, J 3.6) but the pattern of
aliphatics was different, consistent with the isomeric pyrrolo-
[1,2-c]thiazole structure 37. The third product showed only two
signals in the 1H NMR spectrum and three in the 13C NMR and
was identified as pyridin-4-one 40 by comparison with literature
data. This compound was the sole product at 750 ЊC.
The likely mechanism for the formation of these products is
also shown in Scheme 5. Hydrogen transfer from the methyl-
eneketene 33 can take place from either the 2- or the 4-positions
of the thiazolidine to give the azomethine ylides 34 and 36
respectively which then cyclise to the isomeric products 35 and
37. Loss of CH2S from either of these provides another dipolar
structure, 38, which can collapse to 40 via the aziridine 39. This
mode of breakdown from the aziridine is apparently more
550
J. Chem. Soc., Perkin Trans. 1, 2002, 548–554

View Online
cyclisation at the site of hydrogen atom removal has created the
third methine group. The remainder of the carbon skeleton of
the methyleneketene is unchanged. The most likely structure
consistent with these features is the tricycle 43, which could be
formed by an initial retro-ene-type fragmentation to give the
imidoylketene intermediate 42 followed by intramolecular
cycloaddition with the co-formed enol ether moiety (Scheme 6).
Fig. 2 NOE data for the hydroxypyrrole form of 22.
about the N-cyclohexenyl bond, since both the methine and the
adjacent methylene signals of the cyclohexenyl substituent are
enhanced by irradiation of the hydroxypyrrole 2- and 5-proton
resonances.
Both the alkenylpyrrolones 22 and 23 could be regio-
selectively O-alkylated directly from the pyrolysis in moderate
yield under conditions we have previously developed,15 to pro-
vide the methoxypyrroles 46 and 47. The pyrrolones are there-
fore stable enough to survive basic conditions (NaH) in a
dipolar aprotic solvent (dimethylimidazolidinone). In contrast
to the corresponding pyrrolones, the O-alkyl compounds are
stable oils which can be purified and handled easily at room
temperature. A number of preliminary attempts were made to
hydrolyse the enamine unit of the methoxypyrroles 46 or 47
under dilute acid conditions to provide the parent 3-methoxy-
pyrrole but no products could be isolated.
Scheme 6
The regioselectivity of hydrogen transfer from the cyclohexane
ring rather than the methyl groups (despite a 6 : 4 statistical
ratio in the opposite direction) would then be explained by the
reactivity of the enol ether 42 compared with the isomeric
alkene 44 in the inverse electron demand cycloaddition with the
electron deficient imidoylketene component.
If this mechanism is correct, the stereochemistry at the fusion
of the two six-membered rings should be cis, owing to the
suprafacial nature of the cycloaddition. This was confirmed in
two ways. First the single proton at the ring junction (identified
by proton–carbon correlation) shows no large couplings and is
therefore unlikely to be in an axial position as required for a
trans fusion. Second, NOE experiments (Fig. 1) confirm that the
Conclusions
This work has provided the first synthetic route to N-alkenyl-
pyrrol-3(2H )-ones (1-alkenyl-3-hydroxypyrroles), by employ-
ing the standard Meldrum’s acid route to the pyrrolone and the
known thermal breakdown of oxazolidine rings to give azo-
methine ylide intermediates. After formation of the methyl-
eneketene by Meldrum’s acid fragmentation, the overall process
requires a hydrogen transfer (to generate an azomethine ylide),
electrocyclisation (to a fused pyrrolone), fragmentation of the
oxazolidine ring (to create a second azomethine ylide) and
hydrogen transfer to give the final product. Corresponding
thiazolidines may be more stable to fragmentation. The
N-alkenylpyrrol-3(2H )-ones were shown to behave in a broadly
similar way as other pyrrol-3(2H )-ones with respect to tauto-
merism and O-alkylation reactions. When pyrrolone formation
is blocked, an alternative hydrogen transfer to the methyl-
eneketene by a retro-ene reaction can take place, followed by
an intramolecular reverse electron demand cycloaddition to
provide the fused tricyclic pyridin-4-one 43.
Fig. 1 Selected NOE data for 43.
fused oxazolidine ring is on the same side of the molecule as the
hydrogen atom at the bridgehead.
Finally the chemistry of the new 1-alkenylpyrrolones 22 and
23 was briefly explored. As previously mentioned, tautomerisa-
tion to the hydroxypyrrole form is promoted by polar solvents
such as DMSO, but this is relatively disfavoured by comparison
with simple model compounds. Thus, whereas the N-phenyl-
pyrrolone 45 exists to the extent of 95% in the enol form in
DMSO, 22 and 23 remain partially in their keto tautomers
(20% and 37% respectively).
A series of NOE experiments on the 3-hydroxypyrrole
tautomer of 22 (Fig. 2) has shown that the chemical shifts of the
three pyrrole ring protons are in the same order and the coup-
ling constants are of similar magnitudes as previously observed
for simple 3-hydroxypyrroles [22: δH 6.69 (5-position) > 6.43
(2-position) > 5.65 (4-position); 3J4,5 = 2.9; 4J2,5 = 2.5; 4J2,4 = 1.9
Hz].14 The NOE results also prove that there is facile rotation
Experimental
1H And 13C NMR spectra were recorded at 250 (or 200) and 63
(or 50) MHz respectively for solutions in [2H]chloroform unless
otherwise stated. Coupling constants are quoted in Hz. 13C
NMR signals refer to CH resonances unless otherwise stated; in
most cases assignments were confirmed by appropriate DEPT
J. Chem. Soc., Perkin Trans. 1, 2002, 548–554
551

View Online
3
2
experiments. Mass spectra were obtained under electron impact
conditions.
3.87 (1H, d of d, J 5.8 and J 9.2), 2.19–1.16 (10H, m), 1.48
(3H, s) and 0.77 (3H, s); δC 165.41 (quat), 159.90 (quat), 149.41,
136.15 (quat), 128.77, 128.21, 127.42, 102.81 (quat), 100.31
(quat), 87.16 (quat), 71.57, 63.99, 37.48, 33.22, 27.37, 24.30,
23.92, and 23.04 (2C’s overlapping; full assignments not con-
firmed by DEPT); m/z 313 (M Ϫ C3H6O, 23%), 215 (100), 143
(48), 104 (66), and 91 (13), (Mϩ not detected).
Oxazolidines and related compounds
The general method of Hancock and Cope was used,16 whereby
the appropriate ethanolamine or diamine (1 equivalent) was
heated with cyclohexanone (1.1 equivalent) in toluene solution
in the presence of a small amount of toluene-p-sulfonic acid
using a Dean and Stark trap. The following products were
prepared.
2,2-Dimethyl-5-(4,4-dimethyl-2,2-pentamethyleneoxazolidin-
3-ylmethylene)-1,3-dioxane-4,6-dione 16. (82%) mp 185–187 ЊC
(from ethanol) (Found: C, 63.4; H, 8.0; N, 4.45. C17H25NO5
requires C, 63.2; H, 7.75; N, 4.35%); δH (all signals broad) 8.00
(1H, s), 3.66 (2H, s), 1.59–1.35 [22H, m including 1.58 (s), and
1.51 (s)]; δC (signals broadened) 163.58 (two quat), 150.18,
102.30 (quat), 102.13 (quat), 84.39 (quat), 77.72 (CH2), 65.50
(quat), 36.14 (CH2), 26.50 (CH3), 23.91 (CH2) and 22.97 (CH2)
(some signals overlapping); m/z 323 (Mϩ, 25%), 266 (27), 265
(74), 193 (29), 178 (58), 169 (60), 166 (28), 154 (55) and 150 (88).
1-Oxa-4-azaspiro[4.5]decane
(2,2-pentamethyleneoxazol-
idine) 5. (82%), bp 93–95 ЊC (15 Torr) [lit.,17 89–90 ЊC (16 Torr)].
3-Phenyl-1-oxa-4-azaspiro[4.5]decane (2,2-pentamethylene-4-
phenyloxazolidine) 6. (78%), bp 140–142 ЊC (0.2 Torr) [charac-
terised as its Meldrum’s acid derivative 15 (see below)].
3,3-Dimethyl-1-oxa-4-azaspiro[4.5]decane (4,4-dimethyl-2,2-
pentamethyleneoxazolidine) 7. (62%), bp 106–109 ЊC (34 Torr)
[lit.,16 95–97.5 ЊC (20 Torr)].
2,2-Dimethyl-5-(2,2-pentamethylenetetrahydrooxazinyl-
methylene)-1,3-dioxane-4,6-dione 18. (85%) mp 155–156 ЊC
(from ethanol) (Found: C, 61.7; H, 7.5; N, 4.3. C16H23NO5
requires C, 62.15; H, 7.45; N, 4.55%); δH 8.34 (1H, s), 3.8–4.0
(4H, m), 2.28 (2H, m), 1.94 (2H, m) and 1.7–1.5 (14H, m);
δC 156.04, 103.24 (quat), 92.34 (quat), 84.17 (quat), 58.79
(CH2), 47.65 (CH2), 34.76 (2CH2), 27.12 (2CH3), 25.24 (CH2),
25.17 (CH2) and 22.64 (2CH2) (2 quat signals expected at ca. δC
160 are broadened by exchange); m/z 309 (Mϩ, 7%), 251 (66),
153 (41), 137 (44), 127 (100), 109 (70) and 81 (84).
1-Oxa-5-azaspiro[5.5]undecane
(2,2-pentamethylenetetra-
hydro-1,3-oxazine) 9. (80%) mp 44–45 ЊC (lit.,18 44–45 ЊC);
δH 3.75 (2H, m), 2.87 (2H, m) and 1.3–1.8 (12H, m); δC 83.96
(quat), 60.21 (CH2), 38.55 (CH2), 34.54 (2CH2), 27.64 (CH2),
26.34 (CH2) and 22.30 (2CH2).
1-Methyl-1,5-diazaspiro[5.5]undecane (1-methyl-2,2-penta-
methylenehexahydropyrimidine) 10. (85%) bp 50–60 ЊC (0.9
Torr) [lit.,19 112 ЊC (70 Torr)].
2,2-Dimethyl-5-(3-methyl-2,2-pentamethylenehexahydro-
pyrimidin-1-ylmethylene)-1,3-dioxane-4,6-dione
19.
(82%
crude), decomposed on attempted crystallisation and was not
fully characterised δH 8.03 (1H, s), 3.50 (2H, t), 2.64 (2H, t),
2.34 (3H, s), 2.24 (3H, m), 1.76 (6H, m), 1.65 (3H, m) and 1.60
(6H, s); δC 165.63 (quat), 164.55 (quat), 159.74, 116.93 (quat),
104.64 (quat), 84.29 (quat), 49.90 (CH2), 49.47 (CH2), 42.28
(2CH2), 36.45 (CH3), 29.53 (CH2), 27.34 (2CH2), 27.13 (2CH3)
and 25.27 (CH2); m/z 322 (Mϩ, 2%), 264 (3), 221 (5), 168 (49),
139 (26) and 125 (100).
2,2-Pentamethylene-1,3-thiazolidine 11. Made by the general
method of Schmolka and Spoerri20 bp 106–107 ЊC (11 Torr)
[lit.,21 125–127 ЊC (20 Torr)], δH 3.18 (2H, t), 2.78 (2H, t) and
1.3–1.8 (10H, m); δC 81.75 (quat), 49.93 (CH2), 39.93 (2CH2),
34.96 (CH2), 25.00 (CH2) and 24.33 (2CH2).
1,3-Thiazolidine 12. Compound 12 was made in 75% yield by
the standard route22 bp 60–65 ЊC (10 Torr) [lit.,22 60–62 ЊC
(10 Torr)]; δH 4.01 (2H, s), 3.03 (2H, t) and 2.71 (2H, t);
δC 55.73, 53.12 and 34.30 (all CH2).
2,2-Dimethyl-5-(2,2-pentamethylenethiazolidin-3-ylmethyl-
ene)-1,3-dioxane-4,6-dione 20. (85%) mp 134–136 ЊC (from
ethanol) (Found: C, 57.05; H, 6.9; N, 4.25. C15H21NO4Sؒ0.25
H2O requires C, 57.05; H, 6.8; N, 4.45%); δH 8.25 (1H, s), 4.13
(2H, t, 3J 6.4), 2.98 (2H, br d, 3J 6.4), 2.2–1.4 (16H, m);
δC 165.64 (quat), 160.10 (quat), 152.54, 102.83 (quat), 85.39
(quat), 82.21 (quat), 57.58 (CH2), 39.11 (2CH2), 27.58 (CH2),
26.45 (2CH3), 24.88 (2CH2) and 24.22 (CH2); m/z 253 (Mϩ, 2%),
181 (15), 168 (100), 152 (21), 138 (65), 125 (26) and 110 (46).
Meldrum’s acid derivatives
A solution of the relevant oxazolidine or related substrate
(22 mmol) in acetonitrile (5 cm3) was added to a stirred solution
of 2,2-dimethyl-5-(methoxymethylene)-1,3-dioxane-4,6-dione
13 (3.72 g, 20 mmol) in acetonitrile (25 cm3) and the reaction
was allowed to stir at room temperature for 3 h. The product
was isolated as a precipitate or by removal of the solvent under
reduced pressure. The following derivatives were prepared by
this method.
2,2-Dimethyl-5-(thiazolidin-3-ylmethylene)-1,3-dioxane-4,6-
dione 21. (80%) mp 114–116 ЊC (from ethanol) (Found: C, 49.4;
H, 5.25; N, 5.7. C10H13NO4S requires C, 49.4; H, 5.35; N,
5.75%); δH (2 rotamers present in 62 : 38 ratio) 8.21 (minor, 1H,
2,2-Dimethyl-5-(2,2-pentamethyleneoxazolidin-3-ylmethyl-
ene)-1,3-dioxane-4,6-dione 14. (87%) mp 147–149 ЊC (from
ethanol) (Found: C, 61.3; H, 7.4; N, 4.7. C15H21NO5 requires C,
61.0; H, 7.1; N, 4.75%); δH 8.05 (1H, s), 4.00 (2H, br d, 3J 5.6),
4
s), 8.17 (major, 1H, t, J 0.8), 4.71 (minor, 2H, s), 4.58 (major,
2H, d, 4J 0.8), 4.01 (major, 2H, t, 3J 6.4), 3.89 (minor, 2H, t, 3J
3
3
6.5), 3.10 (minor, 2H, t, J 6.5), 3.02 (major, 2H, t, J 6.4) and
1.63 (both rotamers, 6H, s); δC 165.02 (both rotamers, quat),
160.33 (major, quat), 160.20 (minor, quat), 155.63 (major),
155.56 (minor), 102.96 (major, quat), 102.88 (minor, quat),
86.19 (major, quat), 85.83 (minor, quat), 59.20 (major, CH2),
58.70 (minor, CH2), 54.35 (minor, CH2), 53.51 (major, CH2),
30.63 (minor, CH2), 26.36 (major, CH2) and 26.38 (both rotam-
ers, CH3); m/z 243 (Mϩ, 2%), 185 (100), 157 (37), 129 (23), 113
(28), 87 (25) and 53 (47).
3
3.94 (2H, br d, J 5.6), 1.77–1.55 (10H, m) and 1.63 (6H, s);
δC 165.60 (quat), 160.13 (quat), 150.87, 102.81 (quat), 99.05
(quat), 85.07 (quat), 63.38 (CH2), 50.51 (CH2), 35.17 (2CH2),
26.47 (2CH3), 24.14 (CH2) and 22.89 (2CH2); m/z 295 (Mϩ,
15%), 238 (28), 237 (89), 194 (30), 150 (40), 141 (36), 140 (21),
139 (100), 122 (50), 96 (25) and 95 (27).
2,2-Dimethyl-5-(2,2-pentamethylene-4-phenyloxazolidin-3-
ylmethylene)-1,3-dioxane-4,6-dione 15. (27%) mp 141–143 ЊC
(from cyclohexane) (Found: C, 68.2; H, 6.8; N, 3.8. C21H25NO5
requires, C, 67.9; H, 6.75; N, 3.8%); δH 8.08 (1H, d, 4J 1.4), 7.31–
7.10 (5H, m), 5.93 (1H, m), 4.48 (1H, d of d, 3J 7.4 and 2J 9.2),
2,2-Dimethyl-5-(2,2-dimethyloxazolidinylmethylene)-1,3-
dioxane-4,6-dione 17. The acetone-derived oxazolidines showed
variable amounts of hydrolysis product using the standard
552
J. Chem. Soc., Perkin Trans. 1, 2002, 548–554

View Online
conditions and so an alternative ‘one-pot’ method was devised
as follows. Ethanolamine (2-aminoethanol) (2.44 g, 40 mmol)
was dissolved in acetone (50 cm3) and a catalytic amount of
toluene-p-sulfonic acid was added. The solution was then
heated at reflux for 3 h. After the solution had cooled a solution
of freshly prepared 2,2-dimethyl-5-(methoxymethylene)-1,3-
dioxane-4,6-dione 13 (7.44 g, 40 mmol) in acetone (25 cm3) was
added. The reaction mixture was then stirred overnight at room
temperature. The product was obtained by removal of solvent
under reduced pressure to give 2,2-dimethyl-5-(2,2-dimethyl-
oxazolidinylmethylene)-1,3-dioxane-4,6-dione 17 (85%) mp 135–
137 ЊC (from ethanol) (Found: C, 56.3; H, 6.7; N, 5.6.
C12H17NO5 requires C, 56.5; H, 6.7; N, 5.5%); δH (all signals
broad), 8.93 (1H, s), 4.11–3.98 (4H, m), 1.69 (6H, s) and 1.56
(6H, s); δC 165.52 (quat), 160.10 (quat), 151.03, 102.95 (quat),
97.96 (quat), 85.40 (quat), 63.63 (CH2), 50.34 (CH2), 26.54 and
26.31 (both 2CH3); m/z 255 (Mϩ, 10%), 198 (17), 197 (26), 139
(100), 138 (46), 96 (23), 95 (45), 68 (23) and 67 (89).
formation of the fused pyrrolone 31 δH ([2H6]acetone)
(pyrrolone tautomer) 8.29 (1H, d, 3J 3.4) and 5.00 (1H, d,
3J 3.4); (hydroxypyrrole tautomer) 6.44 (1H, d, 3J 3.0) and 5.65
3
(1H, d, J 3.0). The keto and enol tautomers were present in
47 : 53 ratio under these conditions. The assignment was con-
firmed by mass spectrometry (31, Found: Mϩ, 207.1256.
C12H17NO2 requires M 207.1259) which also showed the pres-
ence of the oxidation product 32 (Found: Mϩ, 223.1206.
C12H17NO3 requires M 223.1208). This oxidation behaviour is
characteristic of 2-substituted pyrrolones.5
FVP of the 5-(2,2-pentamethylenethiazolidinyl)- derivative 20.
(200 mg, T f 600 ЊC, T i 185 ЊC, P 10Ϫ2 Torr, t 15 min) gave
1-(cyclohex-1-enyl)-1H-pyrrol-3(2H)-one 22 as above, (ca.
30%) δH 7.97 (1H, d, 3J 3.4), 5.21 (1H, d, 3J 3.4), 5.04 (1H, br t,
3J 1.2), 3.88 (2H, s) and 2.30–1.5 (8H, m); δC 198.44 (quat),
157.87, 134.59 (quat), 105.65, 100.91, 55.23 (CH2), 24.50 (CH2),
23.71 (CH2), 22.07 (CH2) and 21.84 (CH2); a second unidenti-
fied component (ca. 20%) showed characteristic pyrrolone
Under similar conditions, 2,2-dimethyl-5-(2,2-[2H6]dimethyl-
oxazolidinylmethylene)-1,3-dioxane-4,6-dione, 17D m/z 261
(Mϩ, 2%), 203 (10), 139 (100), 111 (11), 95 (35), 68 (16) and 67
(64) was prepared in 80% yield.
3
3
signals at δH 7.91 (1H, d, J 3.9) and 5.18 (1H, d, J 3.9); δC
163.47 and 101.07.
FVP of the 5-(thiazolidinyl) derivative 21. (100 mg, T f 600 ЊC,
T i 165 ЊC, P 10Ϫ2 Torr, t 15 min) gave no significant products
except acetone. However, when a ‘cold-finger’ trap was used as
described for compound 18 above, three products 35, 37 and 40
formed in 55 : 18 : 27 ratio could be identified from the NMR
spectra of an aliquot of the mixture. The major product was
2,3-dihydropyrrolo[2,1-b]thiazol-7(7aH )-one 35 (55% of the
Flash vacuum pyrolysis experiments
Conditions for the pyrolyses were established in small scale
experiments in which the product(s) were dissolved in a deuteri-
ated solvent and analysed immediately by H NMR spectro-
scopy. On a larger scale the crude products could not be further
purified because of decomposition. The precursor, pyrolysis
conditions [quantity of precursor, furnace temperature (T f),
inlet temperature (T i), pressure range (P) and pyrolysis time (t)]
and, where appropriate, approximate yields are given.
1
3
mixture) δH ([2H6]acetone) 7.97 (1H, d, J 3.6), 5.25 (1H, d,
3J 3.6), 4.71 (1H, s), 4.03 (1H, dd, 2J 12.1 and 3J 5.6), 3.29 (1H,
td, 2J and 3J 11.8, 3J 5.3), 3.05 (1H, dd, 2J 10.5 and 3J 5.4) and
2
3
3
2.82 (1H, td, J and J 11.0, J 5.7); δC ([2H6]acetone) 202.73
(quat), 170.29, 105.68, 66.70, 53.14 (CH2) and 32.44 (CH2). The
minor bicycle was 1,3-dihydropyrrolo[1,2-c]thiazol-7(7aH )-one
FVP of the 5-(2,2-pentamethyleneoxazolidinyl)- derivative 14.
(158 mg, T f 600 ЊC, T i 140 ЊC, P 5 × 10Ϫ3 Torr, t 2 h) gave
1-(cyclohex-1-enyl)-1H-pyrrol-3(2H)-one, 22 (ca. 70%) decom-
poses on attempted distillation [140 ЊC (0.2 Torr)] (Found: Mϩ
163.0997. C10H13NO requires M 163.0997); δH 7.94 (1H, d,
3J 3.4), 5.15 (1H, d, 3J 3.4), 5.01 (1H, br t, 3J 1.2), 3.82 (2H, s)
and 2.20–1.45 (8H, m); δC 198.16 (quat), 158.20, 134.49 (quat),
105.55, 100.49, 55.07 (CH2), 24.27 (CH2), 23.54 (CH2), 21.88
(CH2) and 21.65 (CH2); m/z 163 (Mϩ, 100%), 162 (27) and 135
(23).
3
37 (18% of the mixture) δH ([2H6]acetone) 8.10 (1H, d, J 3.6),
3
2
5.35 (1H, d, J 3.6), 4.70 (1H, d, J 10.5), 4.37 (1H, d,2J 10.5),
3.97 (1H, overlapping m), 3.38 (1H, overlapping m) and 2.97
(1H, overlapping m); δC ([2H6]acetone) 204.19 (quat), 171.54,
109.20, 69.37, 54.52 (CH2) and 31.63 (CH2). The presence of
the pyrrolothiazoles was confirmed by mass spectrometry
(Found: Mϩ, 141.0249. C6H7NOS requires M 141.0248). The
third product was 1H-pyridin-4-one 40 (27% of the mixture) δH
3
3
([2H6]acetone) 7.87 (1H, d, J 7.4) and 6.37 (1H, d, J 7.4);
δC ([2H6]acetone) (CH’s only) 140.36 and 117.70 (Found: Mϩ
95.0370. C5H5NO requires M 95.0371). FVP at 750 ЊC gave
pyridin-4-one 40 δH ([2H6]acetone) 7.83 (1H, d, 3J 7.4) and 6.39
(1H, d, 3J 7.4) as the only product.
FVP of the 5-(2,2-dimethyloxazolidinyl)- derivative 17. (90
mg, T f 625 ЊC, T i 140 ЊC, P 1 × 10Ϫ3 Torr, t 30 min) gave
1-(1-methylethenyl)-1H-pyrrol-3(2H)-one 23 (ca. 50%),
decomposes on attempted distillation [100 ЊC (0.2 Torr)]
(Found: Mϩ, 123.0683. C7H9NO requires M 123.0684); δH 9.45
(1H, d, 3J 3.6), 5.22 (1H, d, 3J 3.6), 4.11 (2H, s), 3.84 (2H, s) and
1.96 (3H, s); δC 198.50 (quat), 158.62, 139.24 (quat), 102.71,
91.84 (CH2), 55.17 (CH2) and 18.30 (CH3); m/z 123 (Mϩ, 100%),
108 (37), 95 (11), 94 (15), 83 (56), 80 (14) and 67 (11); a second
product formed during this pyrolysis, (ca. 20%), which decom-
posed on attempted recrystallisation. A small sample of this
compound was adventitiously obtained in reasonable purity,
allowing it to be identified as 2-(dimethylmethylene)-1H-pyrrol-
3(2H)-one 24 (Found: Mϩ, 123.0690. C7H9NO requires M,
123.0684); δH 7.59 (1H, t, 3J 3.6), 7.25 (1H, br s), 5.33 (1H, d of
d, 3J 3.6 and 4J 1.6), 2.35 (3H, s) and 1.97 (3H, s); m/z 123 (Mϩ,
100%), 108 (54), 80 (20), 53 (14) and 41 (20).
FVP of the 5-(4,4-dimethyl-2,2-pentamethyleneoxazolidinyl-)
compound 16. (148 mg, T f 625 ЊC, T i 150 ЊC, P 10Ϫ3 Torr, t 1 h),
gave 3,3-dimethyl-2,3,8,9,10,11-hexahydro-7aH-oxazolo[2,3-j]-
quinolin-7-one 43 (50%) bp 165–167 ЊC (0.2 Torr) (Found: Mϩ,
221.1416. C13H19NO2 requires M 221.1414); δH 6.98 (1H, d,
3J 7.1), 4.99 (1H, br d, 3J 7.1), 3.95 (1H, d, 2J 9.0), 3.87 (1H, br
2
d, J 9.0), 2.63 (1H, m), 2.43 (1H, m), 1.70 (1H, m), 1.88–1.28
(6H, m), 1.38 (3H, s) and 1.35 (3H, s); δC 193.29 (quat), 142.66,
97.77, 95.79 (quat), 76.71 (CH2), 59.91 (quat), 50.07, 29.06
(CH3), 26.51 (CH2), 24.68 (CH3), 22.56 (CH2) and 21.28 (2 CH2
signals superimposed); m/z 221 (Mϩ, 28%), 206 (43), 192 (39),
179 (100), 150 (14), 124 (22) and 55 (26).
FVP of the 5-(2,2-pentamethylene-4-phenyloxazolidinyl-) deriv-
ative 15 and the hexahydropyrimidinyl compound 19 failed
repeatedly to give any identifiable products.
FVP of the 5-(2,2-pentamethylenetetrahydrooxazinyl) deriv-
ative 18. (200 mg, T f 600 ЊC, T i 185 ЊC, P 10Ϫ2 Torr, t 15 min)
with standard work-up gave a mixture whose spectra showed
only decomposition products. However, when the pyrolysate
was trapped on a ‘cold-finger’ at Ϫ80 ЊC and an aliquot was
washed directly into an NMR tube under a nitrogen atmos-
phere, the 1H NMR spectrum at Ϫ20 ЊC showed evidence of the
1-Cyclohex-1-enyl-3-methoxypyrrole 46
The pyrrolone precursor 14 (0.5 g, 1.7 mmol) was pyrolysed
under FVP conditions as described above. The entire pyrolysate
J. Chem. Soc., Perkin Trans. 1, 2002, 548–554
553

View Online
was dissolved in dimethylimidazolidinone (10 cm3), sodium
hydride (50%, 0.24 g, ca. 3 × excess) was added and the mixture
was stirred for 15 min under an atmosphere of nitrogen. Methyl
toluene-p-sulfonate (0.32 g, 1.7 mmol) was added and the stir-
ring was continued at room temperature for 2 h. Standard
work-up14 gave a dark oil which was distilled (Kugelrohr) to
give 1-cyclohex-1-enyl-3-methoxypyrrole 46 (0.090 g, 30% over-
all for the two steps), bp 138–140 ЊC (0.1 Torr) (Found: Mϩ,
177.1148. C11H15NO requires M 177.1154); δH 6.67 (1H, dd,
References
1 Part 13. P. A. Derbyshire, G. A. Hunter, H. McNab and L. C.
Monahan, J. Chem. Soc., Perkin Trans. 1, 1993, 2017.
2 Preliminary communication, G. A. Hunter and H. McNab, J. Chem.
Soc., Chem. Commun., 1993, 794.
3 H. McNab and L. C. Monahan, J. Chem. Soc., Perkin Trans. 1, 1988,
863.
4 For a review, see A. M. Gaber and H. McNab, Synthesis, 2001, 2059.
5 H. McNab, L. C. Monahan and J. C. Walton, J. Chem. Soc., Perkin
Trans. 2, 1988, 759.
6 H. J. Gordon, J. C. Martin and H. McNab, J. Chem. Soc., Perkin
Trans. 1, 1984, 2129.
7 For example, A. J. Blake, G. A. Hunter and H. McNab, J. Chem.
Soc., Chem. Commun., 1990, 734.
8 R. D’Alessio, A. Bargiotti, O. Carlini, F. Colotta, M. Ferrari,
P. Gnocchi, A. Isetta, N. Mongelli, P. Motta, A. Rossi, M. Rossi,
M. Tibolla and E. Vanotti, J. Med. Chem., 2000, 43, 2557 and
references therein.
4
4
3J 3.1 and J 2.5), 6.47 (1H, dd, J 2.5 and 2.0), 5.90 (1H, dd,
3J 3.1 and 4J 2.0), 5.63 (1H, m), 3.72 (3H, s), 2.44–2.37 (2H, m),
2.21–2.13 (2H, m) and 1.83–1.61 (4H, m); δC 149.40 (quat),
135.92 (quat), 115.50, 110.78, 99.41, 98.04, 57.53 (CH3), 26.41
(CH2), 24.00 (CH2), 22.50 (CH2) and 21.95 (CH2); m/z 177 (Mϩ,
100%), 162 (58), 134 (16) and 81 (20).
1-(1-Methylethenyl)-3-methoxypyrrole 47
9 H. McNab and K. Withell, ARKIVOC, 2000, 1, 806.
10 H. McNab and L. C. Monahan, J. Chem. Soc., Perkin Trans. 2, 1988,
1459.
11 B. A. Trofimov, in Pyrroles Part 2, ed. R. A. Jones, Wiley, New York,
1992, p. 131.
12 (a) M. Joucla and J. Mortier, Tetrahedron Lett., 1987, 28, 2973;
(b) M. Joucla, J. Mortier and R. Bureau, Tetrahedron Lett., 1987, 28,
2975; (c) R. Bureau and M. Joucla, Tetrahedron Lett., 1990, 31,
6017; (d ) R. Bureau, J. Mortier and M. Joucla, Tetrahedron, 1992,
48, 8947; (e) R. Bureau, J. Mortier and M. Joucla, Bull. Soc. Chim.
Fr., 1993, 130, 584.
13 C. L. Hickson, E. M. Keith, J. C. Martin, H. McNab, L. C.
Monahan and M. D. Walkinshaw, J. Chem. Soc., Perkin Trans. 1,
1986, 1465 and references therein.
The pyrrolone precursor 17 (0.37 g, 1.5 mmol) was pyrolysed
under FVP conditions as described above. The entire pyrolysate
was dissolved in dimethylimidazolidinone (8 cm3) containing a
suspension of sodium hydride (50%, 0.22 g, ca. 3 × excess) and
the mixture was stirred for 15 min under an atmosphere of
nitrogen. Methyl toluene-p-sulfonate (0.19 g, 1 mmol) was
added and the stirring was continued at room temperature for
1 h. Standard work-up14 gave a brown oil which was distilled
(Kugelrohr) to give 1-(methylethenyl)-3-methoxypyrrole 47
(0.066 g, 32% overall for the two steps), bp 142–145 ЊC (20 Torr)
(Found: Mϩ, 137.0843. C8H11NO requires M 137.0841); δH 6.74
14 H. McNab and L. C. Monahan, J. Chem. Soc., Perkin Trans. 2, 1991,
1999.
15 G. A. Hunter, H. McNab, L. C. Monahan and A. J. Blake, J. Chem.
Soc., Perkin Trans. 1, 1991, 3245.
3
4
4
(1H, dd, J 3.1 and J 2.5), 6.53 (1H, dd, J 2.5 and 2.0), 5.94
(1H, dd, 3J 3.1 and 4J 2.0), 4.81 (1H, br s), 4.46 (1H, br s), 3.74
(3H, s) and 2.17 (3H, d, 4J 1.2); δC 150.01 (quat), 140.44 (quat),
116.27, 99.42 (2CH), 95.88 (CH2), 57.52 (CH3) and 19.64
(CH3); m/z 137 (Mϩ, 100%), 122 (70), 82 (75) and 41 (62).
16 E. M. Hancock and A. C. Cope, J. Am. Chem. Soc., 1944, 66,
1738.
17 A. C. Cope and E. M. Hancock, J. Am. Chem. Soc., 1942, 64,
1503.
18 E. M. Hancock, E. M. Hardy, D. Heyl, M. E. Wright and A. C.
Cope, J. Am. Chem. Soc., 1944, 66, 1747.
Acknowledgements
19 G. Parrinello and R. Mülhaupt, J. Org. Chem., 1990, 55, 1772.
20 I. R. Schmolka and P. E. Spoerri, J. Am. Chem. Soc., 1957, 79, 4716.
21 E. D. Bergmann and A. Kaluszyner, Recl. Trav. Chim. Pays-Bas,
1959, 78, 289.
We are grateful to the University of Edinburgh for the award of
the Colin and Ethel Gordon Scholarship (to G. A. H.), to the
Royal Society for a Developing World Fellowship (to A. M. G.)
and to Lonza Ltd. for a generous gift of Meldrum’s acid.
22 G. W. Stacy and P. L. Strong, J. Heterocycl. Chem., 1968, 5, 101.
554
J. Chem. Soc., Perkin Trans. 1, 2002, 548–554