On the viability of 5-endo-dig cyclisations of O-propargylic hydroxylamine derivatives, leading to 2,5-dihydroisoxazoles (3-isoxazolines)

Article abstract of DOI:10.1016/j.tetlet.2006.11.114

O-Propargylic hydroxylamines undergo smooth 5-endo-dig cyclisations upon exposure to 3 equiv of molecular iodine to give respectable yields of the corresponding 4-iodo-2,5-dihydroisoxazoles, which should find a number of applications as intermediates for syntheses amongst this useful class of heterocycles.

Full text of DOI:10.1016/j.tetlet.2006.11.114

Tetrahedron Letters 48 (2007) 647–650
On the viability of 5-endo-dig cyclisations of
O-propargylic hydroxylamine derivatives, leading to
2,5-dihydroisoxazoles (3-isoxazolines)
Oliver F. Foot, David W. Knight,* Ai Cheng Lilian Low and YingFa Li
Centre for Heterocyclic Synthesis, School of Chemistry, Cardiff University, Main Building, Park Place, Cardiff CF10 3AT, UK
Received 14 September 2006; revised 3 November 2006; accepted 17 November 2006
Available online 8 December 2006
Abstract—O-Propargylic hydroxylamines undergo smooth 5-endo-dig cyclisations upon exposure to 3 equiv of molecular iodine
to give respectable yields of the corresponding 4-iodo-2,5-dihydroisoxazoles, which should find a number of applications as inter-
mediates for syntheses amongst this useful class of heterocycles.
Ó 2006 Published by Elsevier Ltd.
Despite being a favoured process according to Baldwin’s
rules,1 the 5-endo-dig cyclisation mode has only enjoyed
the prominence afforded to related annulation reactions
relatively recently. Some time ago, encouraged by a scat-
tering of isolated examples in the older Russian litera-
ture,2 we discovered that such cyclisations could be
used to a great effect in the synthesis of 3-iodofurans 2
from 3-alkyne-1,2-diols 1 by treatment with 3 equiv of
iodine in the presence of anhydrous potassium carbon-
ate (Scheme 1).3 Subsequently, we have shown that sim-
ilar cyclisations can be highly effective in the synthesis of
b-halopyrrole derivatives.4 Furthermore, many other
halide-induced cyclisations, both 5-endo-dig and other
modes, in which nucleophilic centres are induced to
bond to acetylenic carbons, have been used to elaborate
an impressive range of heterocycles5 and even
carbocycles.6
tem but also leaves the latter endowed with a halide
atom, which is usually ideally set up for the incorpora-
tion of an additional substituent by using one of the
plethora of palladium-catalysed coupling reactions
now available. The presence of such halide atoms can
also give the option of selective carbonylation reactions
leading to ester formation, along with the potential for
the generation of useful metallated intermediates by
halogen–metal exchange and homologation by radical
formation. It occurred to us that such cyclisation
methodology might be extended to include examples
of alkynyl hydroxylamine derivatives 3 (Scheme 2).
R3
N
NHR3
I2 ?
O
O
R2
R1
R1
I
R2
R3
3
4
I
OH
R2
R1
R2
R1
3.I2
Scheme 2.
R3
O
K2CO3, CH2Cl2
OH
If successful, this would lead to halo-isoxazoline deriva-
tives 4 and, subsequently perhaps, various other oxida-
tion states of these useful heterocycles, along with
open-chain 1,3-amino alcohol derivatives, following
cleavage of the relatively weak N–O bond.7 In addition,
this approach could provide a useful complementarity to
the ubiquitous [1,3]-dipolar cycloaddition chemistry,
which is most often deployed to obtain this type of
heterocycle.7,8 Initially, we wondered whether such
1
2
Scheme 1.
Of course, such chemistry offers a double advantage as it
is effective both in forming a particular heterocyclic sys-
*
Corresponding author. E-mail: knightdw@cardiff.ac.uk
0040-4039/$ - see front matter Ó 2006 Published by Elsevier Ltd.
doi:10.1016/j.tetlet.2006.11.114

648
O. F. Foot et al. / Tetrahedron Letters 48 (2007) 647–650
cyclisations would be viable because of the weakness of
the N–O bond. However, we were reassured that this
might well not be a problem by the recent reports of
the successful 5-exo cyclisation of various unsaturated
hydroxylamine derivatives.9 Our isolation of iodo-dihy-
dropyrroles 54 also indicated both that such 5-endo-dig
cyclisations were viable when a sulphonamide acted as
the nucleophile and that dehydration resulting in aro-
matisation (as shown in Scheme 1) was not necessary
to drive the cyclisation to completion. Herein, we report
on a successful model study, which does indeed demon-
strate the viability of the proposed cyclisation shown in
Scheme 2.
these modified reagents are generally not commercially
available. This encouraged us to develop the first chro-
matography-free Mitsunobu work-up method which
employs the more commonplace reagents, DEAD or
DIAD, and triphenylphosphine.12 When using this
new procedure, yields of at least 90% of O-alkylated
hydroxyphthalimides 8 were routinely secured. While
removal of the phthalimide protecting group could be
effected reasonably efficiently using the classical
Ing–Manske procedure (hydrazine),13 a much milder
method14 consisted of exposure to aqueous methylamine
in ether at ambient temperature (1–4 h) followed simply
by filtration through celite, to remove highly insoluble
N,N0-dimethylphthalic diamide, and evaporation. The
resulting O-alkyl hydroxylamines 9 were then directly
N-tosylated to give the required sulfonamides 10 in
reasonable yields.15
I
CO2Me
N
R
Ts
We were delighted to find that these precursors 10
responded well to a typical set of conditions for carrying
out 5-endo iodocyclisations, amongst other types of ring
generation, and gave the desired iodo-oxazolines 11 in
excellent isolated yields (Scheme 4).16 An exception
was the l-alkyne derivative 10e, which failed to give
any of the 3-unsubstituted derivative [11e; R2 = H],
but instead an unidentified mixture of products.
5
The required precursors 3 were synthesised using the
four step sequence shown in Scheme 3 which, after some
optimisation, turned out to be highly efficient. Firstly,
an aldehyde, R1CHO 6 was condensed with either an
alkynyl lithium or an alkynyl Grignard reagent to give
excellent yields of propargylic alcohols 7. These were
then subjected to a Mitsunobu reaction with N-hydroxy
phthalimide using a standard set of reagents, DEAD or
DIAD and Ph3P.10 Initially, this was a very poor step
when carried out in solvents such as ethyl acetate or
dichloromethane or mixed systems, delivering yields
which were usually much less than 50%. It was only
when we employed anhydrous tetrahydrofuran as the
solvent that viable yields, usually around 85–90%, were
obtained.10 However, to secure these, the usual relatively
demanding level of column chromatographic separation
was required. This necessity is such a drawback to an
otherwise extremely useful reaction that many deriva-
tives of the reagents or sophisticated solvent systems
have been developed in efforts to alleviate this and ren-
der the method ‘chromatography-free’.11 Unfortunately,
Ts
N
(a) 89% [20 oC, 0.5 h]
NHTs
3.I2, 3.K2CO3
MeCN
O
(b) 95% [0 oC, 1 h]
O
R2
(c) 95% [20 oC, 0.7 h]
(d) 91% [20 oC, 0.5 h]
(e) 0% [20oC, 0.5 h]
R1
R1
I
R2
10
11
Scheme 4.
In general, the cyclisations were relatively rapid and very
clean. This was particularly fortunate, as products 11
are sensitive to silica gel chromatography over a
prolonged period, undergoing variable amounts of elim-
ination of toluenesulfinic acid to give the corresponding
isoxazoles, along with other unidentified decomposition
products. Rapid filtration through silica using ethyl
acetate–petrol mixtures served to secure pure products,
which were also unstable when kept in solution but
which could be stored below 0 °C for prolonged periods.
O
N
OH
O
R2
O
O
i
R1
R1
R1
In order to exemplify further these successful cyclisa-
tions, we chose to incorporate more reactive functional
groups into the isoxazoline side chains, which could
potentially compete with the 5-endo-dig pathway and
which would also offer possibilities for subsequent trans-
formations of the initial products. In all cases, the
approach to these precursors was the same as that out-
lined in Scheme 3. For example, by starting with 3-meth-
ylbut-3-en-l-yne, hydroxylamine derivative 12 was
prepared in a good overall yield. Subsequent exposure
to the usual iodocyclisation conditions for 2.5 h led to
the isolation of the expected product 13 in 64% yield
(Scheme 5). There was no evidence which suggested
the formation of addition products involving the alkene
group.
R2
R2
6
7
8
(a) R1 = iPr, R2 = Bu;
(b) R1 = tBu, R2 = Ph;
(c) R1 = R2 = Ph;
ii
NH2
NHTs
(d) R1 = Bn, R2 = Bu;
(e) R1 = Me, R2 = H.
O
O
iii
R1
R1
R2
R2
10
9
Scheme 3. Reagents and conditions: (i) Ph3P (1.2 equiv), THF, 0 °C,
add DIAD (1.1 equiv), 0.25 h, add alcohol 7 (1 equiv), 20 min, add N-
hydroxy phthalimide (1.1 equiv), 16 h, 20 °C, >90% [Ref. 12]; (ii) aq
MeNH2, Et2O, 20 °C, 3 h; (iii) TsCl (1.1 equiv), py, DMAP (cat.),
CH2Cl2, 0 °C, 16 h, 60–70%.

O. F. Foot et al. / Tetrahedron Letters 48 (2007) 647–650
649
Ts
N
N
NHTs
O
O
O
Bu
Bu
3.I2, 3.K2CO3
O
Ts
N
MeCN, 20 oC, 2.5 h
64%
i)
18
I
O
Bu
13
12
NHTs
ii)
I
Ts
N
N
11a
Bu
OTBS
O
O
3.I2, 3.K2CO3
( )n
OTBS
MeCN, 20 oC, 2.5 h
a) 67%; b) 60%
( )n
I
OH
14
15
a) n = 1; b) n = 2
19
Ts
NHTs
Scheme 6. Reagents and conditions: (i) (E)-1-hexenylboronic acid,
10 mol % (Ph3P)4Pd, 2 equiv Na2CO3, aq EtOH, reflux, 16 h, 81%; (ii)
2 equiv but-1-yn-3-ol, 5 mol % Pd(PPh3)2Cl2, 5 mol % CuI, DMF/
Et3N, 60 °C, 16 h, 37%.
TBSO
O
N
TBSO
O
3.I2, 3.K2CO3
Bu
MeCN, 20 oC, 2.5 h
75%
Bu
I
16
17
perhaps ominously, examples of such couplings do not
feature in a recent and extensive treatise on the subject
of such couplings involving heteroaromatics.18
Scheme 5.
Starting with the O-TBS derivative of 3-butyn-1-ol, the
required precursor [14a, n = 1] was also obtained in a
good yield and subsequently underwent smooth cyclisa-
tion to provide the hydroxyethyl isoxazoline 15a in 67%
isolated yield, after a reaction period of 2.5 h at ambient
temperature. No products arising from competing 5-
endo cyclisation by the oxygen atom were observed;
the slightly lower yield was due to some loss of the silyl
group to give the corresponding alcohol (10–15% iso-
lated). In the same way, homologous hydroxylamine
14b, derived from 4-pentyn-l-ol, was converted into
isoxazoline 15b in a similar yield and again with some
loss of the silyl group. In this case, we did not observe
the formation of any products from either 5-exo or 6-
endo cyclisations of the oxygen onto the alkyne function.
Finally, precursor 16 was obtained from the correspond-
ing silyl ether of 3-hydroxypropanal and found to also
undergo smooth cyclisation to give iodo-isoxazoline 17
without competition from either of these two possible
alternative modes involving oxygen as the nucleophile.
Hence, the prospects for the incorporation of at least
some functionalised side chains look viable.
Finally, we were able to demonstrate that selenocyclisa-
tion is also possible in this area. Exposure of hydroxyl-
amine derivative 11a to phenylselanyl chloride in the
presence of potassium carbonate, initially at low temper-
ature, led to an excellent yield of selanyl derivative 20
(Scheme 7). The claimed yield of 95% is based solely
1
on the weight of the product and H NMR analysis
which suggested >96% purity. This was because all at-
tempts to fully purify this product, especially using silica
gel chromatography, resulted in rapid elimination to
give selanyl-isoxazole 21, which was isolated in a pure
state, typically in overall yields of 70–75%.
Ts
N
NHTs
N
O
O
O
i)
ii)
Bu
Bu
SePh
20
SePh
21
Bu
11a
Scheme 7. Reagents and conditions: (i) 1.05 equiv PhSeCl, 1.1 equiv
K2CO3, CH2Cl2, ꢀ78 °C to +20 °C, 2.5 h, ca. 95%; (ii) SiO2, EtOAc–
petrol, ꢁ75%.
The 4-iodo-isoxazolines reported in this letter, to the
best of our knowledge, appear to be a novel class of
compounds which, by reason of the presence of an
iodine atom bonded to an sp2-carbon should have some
potential in palladium-catalysed coupling chemistry. A
very brief survey in the present study showed that the
derivatives reported herein are very prone to elimination
to give the corresponding isoxazoles. For example, an
attempted Suzuki coupling reaction between iodide
11a and (E)-hexenylboronic acid resulted in the isolation
of a good yield of vinyl isoxazole 18 (Scheme 6). Simi-
larly, a Sonogashira coupling using the same iodo-isox-
azoline gave, in albeit lower yield, alkynyl isoxazole 19.
It is not clear if the elimination of p-toluenesulfinic acid
to give the corresponding 4-iodo-isoxazoles precedes the
coupling step. A few examples of both types of coupling
reactions using iodo-isoxazoles are known,17 although,
Acknowledgements
We are very grateful to the EPSRC Mass Spectrometry
Centre, University College Swansea for the provision of
high resolution mass spectrometric data and to the
EPSRC for financial support.
References and notes
1. Baldwin, J. E. J. Chem. Soc., Chem. Commun. 1976, 734–
736, and 739–741.
2. Fabrycy, A.; Kubala, J. Zh. Obshch. Khim. 1961, 31, 476–
479; Fabrycy, A.; Wichert, Z. Zh. Obshch. Khim. 1979, 49,
2499–2504.

650
O. F. Foot et al. / Tetrahedron Letters 48 (2007) 647–650
3. Bew, S. P.; Knight, D. W. J. Chem. Soc., Chem. Commun.
2333; Bates, R. W.; Sa-Ei, K. Org. Lett. 2002, 4, 4225–
4227; Janza, B.; Studer, A. Synthesis 2002, 2117–2123;
Waldo, J. P.; Larock, R. C. Org. Lett. 2005, 7, 5203–5205.
10. For reviews, see: Mitsunobu, O. Synthesis 1981, 1–28;
Hughes, D. L. Org. Prep. Proced. Int. 1996, 28, 127–164;
Hughes, D. L. Org. React. 1992, 42, 335–656.
11. These methods have been highlighted in a recent review:
Dembinski, R. Eur. J. Org. Chem. 2004, 2763–2772.
12. Proctor, A. J.; Beautement, K.; Clough, J. M.; Knight, D.
W.; Li, Y.-F. Tetrahedron Lett. 2006, 47, 5151–5154.
13. Ing, H. R.; Manske, R. H. F. J. Chem. Soc. 1926, 2348–
2351.
1996, 1007–1008; For alternative approaches to the
precursors 1, see: El-Taeb, G. M. M.; Evans, A. B.; Jones,
S.; Knight, D. W. Tetrahedron Lett. 2000, 41, 5945–5948.
4. Knight, D. W.; Redfern, A. L.; Gilmore, J. J. Chem. Soc.,
Perkin Trans. 1 2001, 2874–2883; Knight, D. W.; Redfern,
A. L.; Gilmore, J. J. Chem. Soc., Perkin Trans. 1 2002,
622–628.
5. Rao, M. S.; Esho, N.; Sergeant, C.; Dembinski, R. J. Org.
Chem. 2003, 68, 6788–6790; Sniady, A.; Wheeler, K. A.;
Dembinski, R. Org. Lett. 2005, 7, 1769–1772; Yue, D.;
Yao, T.; Larock, R. C. J. Org. Chem. 2005, 70, 10292–
10296; Flynn, B. L.; Verdier-Pinard, P.; Hamel, E. Org.
Lett. 2001, 3, 651–654; van Esseveldt, B. C. J.; Vervoort,
P. W. H.; van Delft, F. L.; Rutjes, F. P. J. T. J. Org. Chem.
2005, 70, 1791–1795; Ramanathan, B.; Keith, A. J.;
Armstrong, D.; Odom, A. L. Org. Lett. 2004, 6, 2957–
2960; Knight, D. W.; Sharland, C. M. Synlett 2003, 2258–
2260; Knight, D. W.; Sharland, C. M. Synlett 2004, 119–
121; Yue, D.; Larock, R. C. Org. Lett. 2004, 6, 1037–1040;
Amjad, M.; Knight, D. W. Tetrahedron Lett. 2006, 47,
2825–2828; Sromek, A. W.; Rubina, M.; Gevorgyan, V. J.
Am. Chem. Soc. 2005, 127, 10500–10501, For a compre-
hensive listing, see Yao et al. cited in Ref. 6.
14. Wolfe, S.; Hasan, S. K. Can. J. Chem. 1970, 48, 3572–
3576.
15. The main loss of material under the conditions shown in
Scheme 3 appeared to be due to the formation of bis-tosyl
derivatives. Subsequent experiments, performed after
completion of the present project, have indicated that a
superior method is to mix the reactants at ꢀ78 °C in
dichloromethane and then allow the reaction mixture to
slowly warm to ambient temperature during 3 h—Proctor,
A. J., unpublished observations.
16. Full analytical and spectroscopic data consistent with the
proposed structures have been obtained for all compounds
reported herein.
´
6. Barluenga, J.; Gonzalez, J. M.; Campos, P. J.; Asensio, G.
Angew. Chem., Int. Ed. Engl. 1988, 27, 1546; Barluenga, J.;
Trincado, M.; Marco-Arias, M.; Ballesteros, A.; Rubio,
E.; Gonzalez, J. M. Chem. Commun. 2005, 2008–2010;
Yao, T.; Campo, M. A.; Larock, R. C. J. Org. Chem.
2005, 70, 3511–3517.
17. For some Stille and Suzuki couplings of iodo-isoxazoles,
see: Labadie, S. S. Synth. Commun. 1994, 24, 709–719; For
various Heck and Sonogashira couplings of similar
substrates, see: Yamanaka, H.; Shiraiwa, M.; Yamamoto,
E.; Sakamoto, T. Chem. Pharm. Bull. 1981, 29, 3543–3547;
For palladium-catalysed couplings with phenylsulfonyl-
acetonitrile, see: Sakamoto, T.; Kondo, Y.; Suginome, T.;
Ohba, S.; Yamanaka, H. Synthesis 1992, 552–554; For
condensations of metallated derivatives with cyclic
ketones, see: Antequera, T.; Ramos, V.; Vincente, T.
Heterocycles 1986, 24, 3203–3211; Polo, C.; Ramos, V.;
Torroba, T.; Rodriguez, M. L.; Bossio, R.; Marcaccini, S.;
Pepino, R. Heterocycles 1991, 32, 1757–1764.
7. For recent applications, see: Fuller, A. A.; Chen, B.;
Minter, A. R.; Mapp, A. K. J. Am. Chem. Soc. 2005, 127,
5376–5383; Buhrlage, S. J.; Brennan, B. B.; Minter, A. R.;
Mapp, A. K. J. Am. Chem. Soc. 2005, 127, 12456–12457,
and references therein.
8. For recent applications of nitrile oxide chemistry, see:
Lohse-Fraefel, N.; Carreira, E. M. Org. Lett. 2005, 7,
2011–2014, and references therein.
9. Tiecco, M.; Testaferri, L.; Tingoli, M.; Bagnoli, L. J.
Chem. Soc., Chem. Commun. 1995, 235–236; Aschwanden,
P.; Frantz, D. E.; Carreira, E. M. Org. Lett. 2000, 2, 2331–
18. Li, J. J.; Gribble, G. W. Palladium in Heterocyclic
Chemistry. Tetrahedron Organic Chemistry Series;
Pergamon Press: Oxford, 2000; Vol. 20.