Iodocyclisation and rearrangement reactions of mono-protected allyl substituted guanidines

Article abstract of DOI:10.1016/j.tetlet.2010.10.050

The iodocyclisation of a range N-allyl and N-homoallylguanidines using I₂/K₂CO₃ has been found to lead to a series of novel heterocycles which undergo selective rearrangements on variation of the reaction conditions, and p

Full text of DOI:10.1016/j.tetlet.2010.10.050

Tetrahedron Letters 51 (2010) 6825–6829
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Iodocyclisation and rearrangement reactions of mono-protected allyl
substituted guanidines
Deiniol Davies a, Matthew D. Fletcher a, Herjan Franken a, Jackie Hollinshead b, Kristina Kähm a,
b
c
Patrick J. Murphy a, , Robert Nash , David M Potter
⇑
a School of Chemistry, Bangor University, Gwynedd, LL57 2UW, UK
b Phytoquest Limited, Plas Gogerddan, Aberystwyth, Ceredigion SY23 3EB, UK
c Menai Organics Ltd, MENTEC, Deiniol Road, Bangor, Gwynedd LL57 2UP, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
The iodocyclisation of a range N-allyl and N-homoallylguanidines using I2/K2CO3 has been found to lead
to a series of novel heterocycles which undergo selective rearrangements on variation of the reaction
conditions, and predictable protecting group migration in the presence of trifluoroacetic acid in
methanol.
Received 27 August 2010
Revised 29 September 2010
Accepted 8 October 2010
Available online 16 October 2010
Ó 2010 Elsevier Ltd. All rights reserved.
Keywords:
Iodocyclisation
Ring contraction
Rearrangement
Guanidine
We have previously reported1,2 on the iodocyclisation3 of bis-
Boc- and bis-Z-protected allylguanidines 1a and 1b resulting in
the formation of the corresponding five-membered heterocycles
2a and 2b in 85% and 79% yields, respectively, upon treatment with
I2 in acetonitrile (Scheme 1).
yield) in which the Boc group had migrated to one of the other two
guanidine nitrogens and a deprotected guanidine 10 (13% yield).
Interestingly the urea 7 was also isolated in a low 2% yield which
may have arisen via a sequence of Boc-transfer reactions and
hydrolysis of the C@NBoc moiety by water5 (Scheme 4).
We wanted to study the effect of the protecting groups on the
selectivity of the cyclisation, and to this end the three mono-
protected guanidines 5a–c were prepared by the reaction of mono-
protected 1H-pyrazole carboxamidines 3a,b with the corresponding
amine or amine salt 4 in acetonitrile4 (Scheme 2, Table 1).
Initial cyclisation of the mono-Boc-allylguanidine 5a led to for-
mation of product 6 in essentially quantitative yield as was evident
from the 1H NMR spectrum. Cyclisation occurred at the Boc-
substituted nitrogen of the guanidine as was apparent from HMBC
studies (Scheme 3).
Attempts at purifying the heterocycle by column chromatogra-
phy were problematic as several new products were formed. Previ-
ous work2 on the epoxidation of bis-protected guanidines had
demonstrated that Boc and Z-protecting groups were prone to
migration to hydroxy groups, and it was possible to mediate this
by stirring with moist silica gel. An attempt to mimic this with 6
was found to give a complex mixture of products, and after purifi-
cation, four compounds were isolated including 8 and 9 (3% and 4%
On further analysis it was apparent that on standing at À20 °C,
chloroform solutions of the crude product 6 underwent consider-
able decomposition/rearrangement, with NMR analysis indicating
the presence of at least four different compounds. It was thought
that trace moisture or acid might be mediating this and we thus
took a solution of crude 6 and treated it with a threefold excess
of trifluoroacetic acid (TFA) in dichloromethane for 16 h6 (Scheme
5). On NMR analysis it was apparent that the initial compound 6
had been converted into a single major compound which had a
1H NMR spectrum similar to that observed for structure 9 with
diagnostic signals at dY 1.56 (9H, s), 3.35 (1H, dd, J = 10.7, 1.8 Hz),
3.47 (1H, dd, J = 10.5, 3.4 Hz), 3.53 (1H, dd, J = 10.5, 7.0 Hz), 3.87
(1H, app. t, J = 10.5 Hz) and 4.39 (1H, m) ppm. Final determination
of the structure of 9 was obtained by HMBC. A strong correlation
between the ring CH2N signal and the Boc-carbonyl quaternary
carbon was observed. Pure compound 9 could be obtained as its
trifluoroacetate salt in 70% overall yield by recrystallisation from
dichloromethane/petroleum ether (5:1).
We further investigated this process by utilising a Z-protected
guanidine in the hope that this protecting group might be less
likely to undergo rearrangement. Thus cyclisation of mono-Z allyl-
guanidine 5b was attempted. Again the expected 5-exo-product 11
⇑
Corresponding author.
E-mail address: chs027@bangor.ac.uk (P.J. Murphy).
0040-4039/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2010.10.050

6826
D. Davies et al. / Tetrahedron Letters 51 (2010) 6825–6829
HN
HN
HN
P
I
I
(a)
N
N
P
HN
P
N
P
HN
P
N
P
2a: P = Boc
2b: P = Z
1a: P = Boc
1b: P = Z
Scheme 1. Reagents and conditions: (a) I2, CH3CN, K2CO3, 0 °C to rt, 16–24 h.
R
J = 6.5, 4.7, 5.7, 9.1 Hz), 5.35 (2H, d, J = 12.3 Hz) and 7.5 (5H, m).
However, as was observed with the Boc-protected analogue 6, at-
tempted purification led to decomposition and only two com-
pounds were obtained, the previously originally identified
product 10 in 22% yield and the rearranged compound 12 in 10%
yield. In order to ascertain if the Z-protecting group was migrating
in a similar manner to that observed in Boc-protected 5a, the reac-
tion was repeated and the resulting crude compound 11 treated
with TFA in methanol overnight. This resulted in the formation of
12, identical to the product isolated previously, as the major prod-
uct. Compound 12 gave 1H NMR signals at dH 3.32 (1H, br d,
J = 10.4 Hz), 3.45 (1H, m), 3.54 (1H, dd, J = 10.8, 4.1 Hz), 3.88 (1H,
dd, J = 10.8, 9.8 Hz), 4.47 (1H, m), 5.31 (1H, d, J = 11.7 Hz), 5.40
(1H, d, J = 11.7 Hz) and 7.38–7.45 (5H, m) ppm which correlate
very closely to 9, and as before, the position of the Z group was
confirmed by HMBC correlation studies. Pure compound 12 could
be obtained in 47% overall yield by crystallisation from dichloro-
methane/diethyl ether/petroleum ether (5:2:2) (Scheme 6).
We next investigated the cyclisation of the Z-dimethylallyl
guanidine 5c. In related work,2 it had been reported that the
bis-Z-dimethylallyl guanidine 13 underwent cyclisation to give
the six-membered product 14 in 64% yield and we wanted to see
if a similar reaction was possible with 5c (Scheme 7).
We thus treated 5c under the previously employed conditions,
and on work-up, obtained a crude product which appeared to be
composed of two compounds in a 63:37 ratio by 1H NMR. Further
analysis of the proton NMR spectrum indicated that the minor
compound had an ABX coupling pattern at dY 3.49 (1H, dd,
J = 9.5, 13.6 Hz), 3.60 (1H, dd, J = 4.9, 13.6 Hz) and 4.00 (1H, dd,
J = 4.9, 9.5 Hz) ppm whilst the major product had a similar pattern
at dY 3.43 (1H, dd, J = 9.2, 13.6 Hz), 3.55 (1H, dd, J = 4.6, 13.6 Hz)
and 3.97 (1H, dd, J = 4.6, 9.2 Hz) ppm. Analysis of the 13C NMR
spectrum indicated that the CH2 signals were at 46.4 and
5.8 ppm for the two species. The first of these was as expected
for a CH2N group and the structure was determined to be the
R
N
R
HN
R
H2N
N
4
H2N
N
P
5a-c
H2N
N
P
3a: P = Boc
3b: P = Z
Scheme 2. Method A: amine 4,
50 °C 6 h.
D
, 2.5 h. Method B: 4ÁHCl, NEt3, CH3CN, 16 h, rt then
Table 1
Product
P
R
Method
Yield (%)
Mp (°C)
5a
5b
5c
Boc
Z
Z
H
H
Me
A
A
B
80
61
74
75–78
119–121
81–83
HN
HN
I
(a)
HN
N
H2N
N
6
5a
Boc
Boc
Scheme 3. Reagents and conditions: (a) CH3CN, 4 equiv K2CO3, 4 equiv I2, À15 °C to
rt, 16 h.
was observed which was identified by 1H NMR signals at dY 3.28
(1H, dd, J = 6.5, 9.8 Hz), 3.38 (1H, dd, J = 4.7, 9.8 Hz), 3.70 (1H, dd,
J = 5.7, 10.7 Hz), 4.05 (1H, dd, J = 9.1, 10.5 Hz) 4.16 (1H, dddd,
Boc
HN
HN
N
I
(a)
I
I
Boc
HN
N
N
O
N
H
N
6
Boc
7
8
Boc
(2%)
(3%)
Boc
HN
N
I
I
HN
HN
N
H
N
H
10
9
(13%)
(4%)
Scheme 4. Reagents and conditions: (a) SiO2, CH2Cl2, stir, 5 d.

D. Davies et al. / Tetrahedron Letters 51 (2010) 6825–6829
6827
Boc
HN
HN
I
N
(b)
(a)
HN
I
N
H2N
N
HN
N
6
5a
H
Boc
.CF3CO2H
9
Boc
Scheme 5. Reagents and conditions: (a) CH3CN, 4 equiv K2CO3, 4 equiv I2, À15 °C to rt; (b) TFA, MeOH, 16 h, 70%.
HN
HN
I
(a)
(b)
HN
N
Z
H2N
N
Z
11
5b
(c)
Z
HN
N
+
I
I
HN
HN
N
N
H
H
12
10
Scheme 6. Reagents and conditions: (a) CH3CN, 4 equiv K2CO3, 4 equiv I2, À15 °C to rt, 16 h; (b) SiO2, CH2Cl2, stir, 5 d; (c) TFA, MeOH, 16 h, 47%.
to the presence of HI in the reaction medium. The results of a series
of experiments are shown in Scheme 8.
H
I
H
N
N
Interestingly, it was observed that in the reaction of 5c in the
absence of potassium carbonate the rearranged product 16 was
formed almost exclusively and only a trace amount of 15 was ob-
served. As the amount of base was increased a steady switch to the
formation of 15 was observed. This suggests that under acidic con-
ditions the cyclisation proceeds via the ZNH function, leading to
intermediate 18, possibly because of protonation of one of the
other nitrogens.
(a)
N
Z
N
Z
N
Z
NH
Z
13
14
Scheme 7. Reagents and conditions: (a) CH3CN, 4.5 equiv K2CO3, 4 equiv I2, À15 °C
to rt, 16 h, 64%.
In order to elucidate further information regarding this reaction
we reinvestigated the cyclisation of the bis-Z-dimethylallylguani-
dine 13 using varying amounts of potassium carbonate in the reac-
tion mixture. The results of a series of experiments are shown in
Scheme 9. It was found that in the absence of potassium carbonate
only the six-membered product 14 was observed in the crude reac-
tion mixture. However, when more than 4 equiv of base were uti-
lised a second product was observed, which was on isolation, and
was shown to be the five-membered product 19. It is possible that
this compound 19 arose through a base-mediated ring contraction
via the anion 20 and the aziridine 21. In order to test this proposal
we took a purified sample of 14 and treated it with 5 equiv of finely
powdered potassium carbonate in acetonitrile for 16 h at room
temperature and found that it rearranged to give the heterocycle
19 as the only product (Scheme 9).
The reason for the different conditions required to effect these
two ring contraction reactions might lie in the nature of the guani-
dine substituents. In compound 14 the delocalization of electron
density by the two Z-protecting groups reduces the nucleophilicity
of the remaining ring nitrogen, but leads to increased NH acidity
enabling the reaction to proceed under basic conditions. In the case
of compound 18 it is possible that the single Z-protecting group
does not have such a profound effect and it can be postulated that
sufficient nucleophilicity is present to enable a similar rearrange-
ment (Scheme 10).
six-membered product 15, however, the signal at 5.8 ppm suggests
that structure 16 containing a CH2I moiety had formed. As with
previous examples, both 15 and 16 decomposed when in contact
with silica gel and they were thus treated with methanolic TFA
to give a crude mixture containing two rearranged products.
NMR analysis of the crude mixture indicated that the signals for
the major product 15 remained unchanged whilst the minor
compound had a rearranged structure, guanidine 17, in which
the Z-group had migrated to the exocyclic nitrogen. This com-
pound displayed an ABX pattern at dY 3.39 (1H, dd, J = 7.6,
11.0 Hz), 3.56 (1H, dd, J = 4.4, 11.0 Hz) and 3.94 (1H, dd, J = 4.4,
7.6 Hz) ppm. Compound 15 could be separated from the mixtures
in 75–85% yield at either stage of this process by dissolution in
MeOH, followed by cooling (À15 °C), which resulted in its precipi-
tation. The precise structures of compound 15 and the rearranged
products 16 and 17 were confirmed by HMBC correlation studies
which demonstrated that 16 and 17 were two isomeric five-mem-
bered compounds differing only in the position of the Z-protecting
group (Scheme 8).
We wanted to investigate this process further and envisaged
that we might be able to influence the ratios of the two products
by varying the amount of potassium carbonate used in the reac-
tion, as it was postulated that the rearrangement might be due

6828
D. Davies et al. / Tetrahedron Letters 51 (2010) 6825–6829
H
H
I
H
I
N
N
N
(a)
+
HN
NH
Z
N
H
N
Z
N
Z
N
H
5c
18
15
?
Eq. K2CO3 Ratio 15:16
I
0.0
1.0
2.0
3.0
4.5
6.0
2:98
16:84
27:73
35:65
63:37
80:20
I
H
H
N
N
(b)
N
N
Z
HN
N
H
Z
17
16
Scheme 8. Reagents and conditions: (a) CH3CN, K2CO3, 4 equiv I2, À15 °C to rt, 16 h; (b) TFA, MeOH, 16 h, 49% (for 17, two steps).
I
H
H
I
H
N
N
N
(a)
+
N
Z
NH
Z
N
Z
N
Z
N
N
Z
13
19
Z
14
I-
K2CO3
I
Eq. K2CO3 Ratio 14:19
N
0.0
1.0
2.0
3.0
4.5
6.0
100:0
100:0
100:0
100:0
78:22
46:54
N
N
N
N
N
Z
21
20
Z
Z
Z
Scheme 9. Reagents and conditions: (a) CH3CN, K2CO3, 4 equiv I2, À15 °C to rt, 16 h.
I
H
I
H
N
H
N
I-
N
HN
N
Z
N
HN
N
HN
Z
18
22
16
Z
Scheme 10. Rearrangement of intermediate 18.

D. Davies et al. / Tetrahedron Letters 51 (2010) 6825–6829
6829
Yanada, K.; Takemoto, Y. Heterocycles 2005, 66, 101–106; (h) Arnold, M. A.;
Duron, S. G.; Gin, D. Y. J. Am. Chem. Soc. 2005, 127, 6924–6925.
In conclusion, we have demonstrated that the cyclisation of
mono-protected guanidines is a considerably more complex pro-
cess than for the bis-protected analogues and the products of these
reactions proved to be very sensitive to silica gel chromatography
and acid. However, the reactions of simple allyl-substituted guan-
idines do proceed in high yield and in a predictable manner. Both
the mono-Z-protected and the bis-Z-protected guanidines 5c and
13 undergo predictable ring contraction rearrangements which
can be controlled by carefully selecting the conditions and these
processes should be of synthetic interest in preparing highly
substituted cyclic guanidines.
4. Typical experimental procedure: N-Boc-1H-pyrazole-1-carboxamide (3a) (2.0 g,
9.51 mmol) was dissolved in excess allylamine (10 ml) and the mixture refluxed
for 2.5 h. The solvent was removed under reduced pressure and the residue
crystallised from CH2Cl2/petroleum ether to give 5a (1.51 g, 80%) as a white
solid. Mp 75–78 °C; dH (500 MHz, CDCl3) 1.46 (9H, s, 3 Â CH3), 3.83 (2H, d,
J = 4.9 Hz, CH2), 5.24 (1H, br d, J = 10.4 Hz, CH) 5.33 (1H, br d, J = 17.1 Hz, CH),
5.83 (1H, ddt, J = 17.1, 10.4, 4.9 Hz, CH), 6.34 (1H, br s, NH), 7.60 (2H, br s,
2 Â NH); dc (125 MHz, CDCl3) 28.4 (CH3), 43.9 (CH2), 78.1 (C), 117.3 (CH2), 133.5
(CH), 162.1 (C), 163.6 (C); HRMS (ES+) m/z: C9H18O2N3 ([M+H]+); requires
200.1394; found 200.1391.
5. We previously observed that attempts to protect the cyclic guanidine A using
excess NaH and (Boc)2O led to the formation of urea C as the main product in
30% yield. It is thought that this occurs via hydrolysis of intermediate B on
aqueous work-up or chromatography.7
Boc
Boc
H
N
Boc
N
N
N
N
(a)
+
NH2 -OAc
N
O
N
C
A
B
Boc
Boc
H
(a) NaH (excess), (Boc)2O (excess), THF, 0 ºC, 16 h.
Acknowledgements
6. Typical experimental procedure: N-Boc-allylguanidine 5a (0.30 g, 1.5 mmol) was
dissolved in acetonitrile (20 ml) and cooled (À15 °C) using an ice/salt bath.
K2CO3 (0.83 g, 6 mmol, 4 equiv) was added, followed by I2 (1.53 g, 6 mmol,
4 equiv). The mixture was stirred at rt for 16 h before being diluted with H2O
(50 ml). Na2S2O3 solution (saturated) was then added until the I2 colour had
dispersed and the mixture was extracted with EtOAc (3 Â 50 ml), dried (MgSO4)
and evaporated to dryness to give crude 6 (0.498 g). This product was dissolved
in MeOH (5 ml), TFA (0.38 ml, 4.9 mmol) was added and the mixture stirred for
16 h. On evaporation, a solid (0.65 g) was obtained which was dissolved in a
minimum volume of CH2Cl2 (ca. 5–10 ml), diluted with petroleum ether (1–
Thanks are given to the European Social Fund and Menai Organics
(D.H.D.) and the Leonardo Da Vinci programme (K.K. and H.F.) for
funding, and to the EPSRC Mass Spectrometry centre at Swansea
for invaluable assistance.
References and notes
1. Murphy, P. J.; Dennis, M.; Hall, L. H.; Thornhill, A. J.; Nash, R.; Winters, A. L.;
Hursthouse, M.; Light, M. E.; Horton, P. Tetrahedron Lett. 2003, 44, 3075–3080.
2. Murphy, P. J.; Albrecht, C.; Barnes, S.; Bockemeier, H.; Davies, D.; Dennis, M.;
Evans, D. M.; Fletcher, M. D.; Jones, I.; Leitmann, V.; Rowles, R.; Nash, R.;
Stephenson, R. A.; Horton, P. N.; Hursthouse, M. Tetrahedron Lett. 2008, 49, 185–
188.
3. For related iodocyclisations, see: (a) Bruni, E.; Cardillo, G.; Orena, M.; Sandri, S.;
Tomasini, C. Tetrahedron Lett. 1989, 30, 1679–1682; (b) Balko, T. W.; Brinkmeyer,
R. S.; Terando, N. H. Tetrahedron Lett. 1989, 30, 2045–2048; (c) Creeke, P. I.;
Mellor, J. M. Tetrahedron Lett. 1989, 30, 4435–4438; (d) Noguchi, M.; Okada, H.;
Watanabe, M.; Moriyama, H.; Nakamura, O.; Kakehi, A. Heterocycl. Commun.
1996, 2, 361–370; (e) Watanabe, M.; Okada, H.; Teshima, T.; Noguchi, M.;
Kakehi, A. Tetrahedron 1996, 52, 2827–2838; (f) Kitagawa, O.; Fujita, M.; Li, H.;
Taguchi, T. Tetrahedron Lett. 1997, 38, 615–618; (g) Yanada, R.; Kaieda, A.;
2 ml) and placed in
a freezer overnight to give a pale yellow solid. The
supernatant liquid was decanted and the solid washed with petroleum ether to
give 9ÁHCO2CF3 (0.34 g, 70%) as a pale yellow solid. Mp 90–92 °C; Rf: 0.29 (10%
MeOH/CHCl3); dH (500 MHz, CDCl3) 1.56 (9H, s, tBu), 3.35 (1H, dd, J = 10.7,
1.8 Hz, CH), 3.47 (1H, dd, J = 10.5, 3.4 Hz, CH), 3.53 (1H, dd, J = 10.5, 7.0 Hz, CH),
3.87 (1H, app. t, J = 10.5 Hz, CH), 4.39 (1H, m, CH), 7.75 (1H, br s, NH), 11.07 (1H,
br s, NH), 11.88 (1H, br s, NH); dC (125 MHz, CDCl3) 7.0 (CH2), 27.8 (3 Â CH3),
1
46.9 (CH22), 56.8 (CH), 87.4 (C), 116.5 (q, JC–F = 292 Hz, C), 149.9 (C), 156.9 (C),
163.2 (q, JC–C–F = 35 Hz, C); IR mmax 3306, 1752, 1679 cmÀ1; MS m/z: 326 (45%,
[M+H]+), 270 (100%); HRMS (ES+) m/z: C9H17N3IO2 [(M+H)+]; requires;
326.0360; found 326.0363.
7. Everall, E. MSc Thesis, Bangor University, 2008.