Novel phosphorus radical-based routes to horsfiline

Article abstract of DOI:10.1021/ol051095i

(Chemical Equation Presented) Radicals derived from the phosphorus reagents, ethylpiperidine hypophosphite (EPHP) and diethylphosphine oxide (DEPO), are used in two related approaches to synthesis of the alkaloid horsfiline (1). In particular, DEPO proves

Full text of DOI:10.1021/ol051095i

ORGANIC
LETTERS
2005
Vol. 7, No. 15
3287-3289
Novel Phosphorus Radical-Based
Routes to Horsfiline
John A. Murphy,* Re´gis Tripoli, Tanweer A. Khan, and Umesh W. Mali
WestCHEM, Department of Pure and Applied Chemistry, UniVersity of Strathclyde,
295 Cathedral Street, Glasgow G1 1XL, United Kingdom
john.murphy@strath.ac.uk
Received May 11, 2005
ABSTRACT
Radicals derived from the phosphorus reagents, ethylpiperidine hypophosphite (EPHP) and diethylphosphine oxide (DEPO), are used in two
related approaches to synthesis of the alkaloid horsfiline (1). In particular, DEPO proves beneficial for effecting cyclizations at 80
difficult or impossible with Bu3SnH.
°C that are
Horsfiline 1 is a member of the spiropyrrolidinyloxindole
family of alkaloids and was first isolated1 by Bodo et al.
from the roots of Horsfieldia superba, a small tree native to
Malaysia. The spiropyrrolidinyloxindole molecules, and
particularly horsfiline, have been popular targets for synthesis
with quite a variety of approaches being recorded.2
line, suggesting either different mechanisms of action or an
insensitivity to stereochemical configuration at this center.
Compound 3 bears only the three basic rings of the
spiropyrrolidinoindolones that are present in horsfiline (1)
and coerulescine (2). Accordingly, new synthetic approaches
to spiropyrrolidinoindolones are welcome.
These and related molecules are of interest in relation to
cancer treatment. The findings3 of Danishefsky et al. are
particularly interesting in this regard; they tested compounds
3 and 4 (both diastereomers) (Figure 1) and reported
significant activity against human breast cancer cells.
In defining our new approach to horsfiline, we were
attracted by the possibility of using a 1,5-hydrogen atom
translocation (1,5-HAT) as the key step,4 starting from 5 (X
) I or Br) (Scheme 1).
Beckwith and Storey had already studied5 the preparation
of oxindoles, including 7 [R, R′ ) (CH2)n, n ) 2-5] by
(2) (a) Jones, K.; Wilkinson, J. J. Chem. Soc., Chem. Commun. 1992,
1767. (b) Pellegrini, C.; Stra¨ssler, C.; Weber, M.; Borschberg, H.-J.
Tetrahedron: Asymmetry 1994, 5, 1979. (c) Bascop, S.-I.; Sapi, J.; Laronze,
J.-Y.; Levy, J. Heterocycles 1994, 38, 725. (d) Palmisano, G.; Annunziata,
R.; Papeo, G.; Sisti, M. Tetrahedron: Asymmetry 1996, 7, 1. (e) Laksh-
maiah, G.; Kawabata, T.; Shang, M.; Fuji, K. J. Org. Chem. 1999, 64, 1699.
(f) Fischer, C.; Meyers, C.; Carreira, E. M. HelV. Chim. Acta 2000, 83,
1175. (g) Kumar, U. K. S.; Ila, H.; Junjappa, H. Org. Lett. 2001, 3, 4193.
(h) Lizos, D.; Tripoli, R.; Murphy, J. A. Chem. Commun. 2001, 2732. (i)
Selvakumar, N.; Azhagan, A. M.; Srinivas, D.; Krishna, G. G. Tetrahedron
Lett. 2002, 43, 9175. (j) Lizos, D. E.; Murphy, J. A. Org. Biomol. Chem.
2003, 1, 117.
Figure 1. Natural and synthetic spiropyrrolidinoxindoles.
(3) Edmondson, S.; Danishefsky, S. J.; Sepp-Lorenzino, L.; Rosen, N.
J. Am. Chem. Soc. 1999, 121, 2147.
(4) Khan, T. A.; Tripoli, R.; Crawford, J. J.; Martin, C. G.; Murphy, J.
A. Org. Lett. 2003, 5, 2971.
(5) (a) Beckwith, A. L. J.; Storey, J. M. D. J. Chem. Soc., Chem. Com-
mun. 1995, 977. (b) For recent work on the mechanism of the final rearo-
matization step, see: Beckwith, A. L. J.; Bowry, V. W.; Bowman, W. R.;
Mann, E.; Parr, J.; Storey, J. M. D. Angew. Chem., Int. Ed. 2004, 43, 95.
The two diastereoisomers 4, differing in the configuration
at the spiro-center, are equally active against the quoted cell
(1) Jossang, A.; Jossang, P.; Hadi, A.; Sevenet, T.; Bodo, B. J. Org.
Chem. 1991, 56, 6527.
10.1021/ol051095i CCC: $30.25
© 2005 American Chemical Society
Published on Web 06/30/2005

Scheme 1. Radical Translocation Approach to Horsfiline
Scheme 2. Route to Horsfiline Skeleton Using EPHP
hydrogen atom abstraction, using tributyltin hydride. At 80
°C in benzene and using AIBN as initiator, significant
amounts of uncyclized, reduced compounds 8 (X ) H) had
been formed. However, reaction at 160 °C with t-Bu2O2 as
initiator was much more successful. The formation of the
desired cyclized product was, therefore, favored by the
increase of temperature.
Recently, as part of the general drive to develop alterna-
tives to toxic organotin reagents, we have explored the
chemistry of modern phosphorus-based radical precursors.
The chosen reagents, ethylpiperidine hypophosphite (EPHP)6
and diethylphosphine oxide (DEPO),4 provide easy isolation
of products as the phosphorus byproducts can be washed
into water.7 The phosphorus-based reagents appear to have
relatively strong P-H bonds, facilitating cyclizations where
Bu3SnH might, in contrast, encourage premature reduction
of radicals.6,7 They are thus poorer chain-transfer reagents,
but low cost and easy reaction workup make the reagents
attractive.
with 2-iodo-4-methoxybenzenamine 11,2h,10 and the amide
12a was then protected to give the desired precursor 12b (R
) SEM) for the cyclization.2a The best conditions for the
radical cyclization were found to use EPHP and AIBN in
refluxing toluene. However, this gave the desired cyclized
product 13 in a disappointing yield (33%) along with re-
duced compound 14a as a dominant side product (51%)
(Scheme 2).
When the cyclization experiment was carried out with
substrate 12b using EPHP-d and AIBN in refluxing benzene,
a much improved yield (60%) of cyclized product 13, along
with 36% reduced product 14b, was isolated. This compound
showed three aromatic protons in the 1H NMR spectrum and
Our initial cyclizations in this study featured ethylpiperi-
dine hypophosphite (EPHP) (Scheme 2).
Pyrrolidine ester 10a was prepared from the pyrrolidine
9a8 in 80% overall yield. This pyrrolidine was then coupled9a
2
a single deuterium resonance at 7.16 ppm in the H NMR
spectrum.
The completion of the synthesis consisted in removal of
the protecting groups from 13 and N-methylation in situ using
formaldehyde and NaCNBH3 in CH2Cl2 to afford horsfiline
(1) (47% in first step; 89% yield over the last two steps)
(Scheme 3) with spectroscopic data identical to an authentic
sample.2f
The results obtained from EPHP-d in this case were
welcome since the cyclization proceeded in much higher
yield and also because no HAT was identified from the SEM
protecting group or from undesired positions on the pyrro-
lidine to the aryl radical. However, the position of the
(6) (a) Barton, D. H. R.; Jang, D. O.; Jaszberenyi, J. Cs. Tetrahedron
Lett. 1992, 33, 5709. (b) Calderon, J. M. B.; Chicharro, G. J.; Fiandhorn,
R. J.; Huss, S.; Ward, R. A. EP2284 970506, 1997. (c) McCague, R.;
Pritchard, R. G.; Stoodley, R. J.; Williamson, D. S. Chem. Commun. 1998,
2691. (d) Graham, S. R.; Murphy, J. A.; Coates, D. Tetrahedron Lett. 1999,
40, 2415. (e) Graham, S. R.; Murphy, J. A.; Kennedy, A. R. J. Chem. Soc.,
Perkin Trans. 1 1999, 3071. (f) Kita, Y.; Nambu, H.; Ramesh, N. G.;
Anilkumar, G.; Matsugi, M. Org. Lett. 2001, 3, 1157. (g) Roy, S. C.; Guin,
C.; Kumar, K. K.; Maiti, G. Tetrahedron 2002, 58, 2435. (h) Jang, D. O.;
Cho, D. H. Tetrahedron Lett. 2002, 43, 5921. (i) Thomson, D. W.;
Commeureuc, A. G. J.; Berlin, S.; Murphy, J. A. Synth. Commun. 2003,
33, 3631. (j) Caddick, S.; Hamza, D.; Judd, D. B.; Reich, M. T.; Wadman,
S. N.; Wilden, J. D. Tetrahedron Lett. 2004, 45, 2363. (k) Cho, D. Y.;
Jang, D. O. Synlett 2005, 59. For a complete list of references, see ref 4.
(7) For other phosphorus reagents under development, see: (a)Jessop
C. M.; Parsons, A. F.; Routledge, A.; Irvine, D. Tetrahedron Lett. 2004,
45, 5095. (b) Jessop C. M.; Parsons, A. F.; Routledge, A.; Irvine, D.
Tetrahedron Lett. 2003, 44, 479.
(9) (a) Cossy, J.; Cases, M.; Pardo, D. G. Tetrahedron Lett. 1998, 39,
2331. (b) A. G. Wright and J. A. Murphy, unpublished results.
(10) Kondo, Y.; Kojima, S.; Sakamoto, T. J. Org. Chem. 1997, 62, 6507.
(8) Collins, C. J.; Lanz, M.; Singaram, B. Tetrahedron Lett. 1999, 40,
3673.
3288
Org. Lett., Vol. 7, No. 15, 2005

deuterium label in the reduced byproduct 14b was surprising.
The outcome contrasted with results reported by Beckwith
et al. in a cyclization reaction on a simpler substrate [6, R,
R′ ) (CH2)2] using tributyltin deuteride.5 In that case, the
entire isotope found in the reduced product 8 resided in the
side chain at C-1 in the cyclopropane ring [8, R, R′ ) (CH2)2,
X ) D] (Scheme 1). That result indicated that the radical
translocation step was much faster than the deuterium atom
transfer from Bu3SnD to the aryl radical.5 In turn, that meant
that the later cyclization step was the less efficient of the
two key steps since it suffered competition by radical
reduction.
Scheme 4. Concise Route to Horsfiline Using DEPO
However, our observation of no deuterium incorporation
at the aliphatic carbon atom in 14 and the high yield of 13
implies that, under our conditions and with our substrate,
the radical generated after 1,5-hydrogen atom abstraction in
a reaction of 12b does not abstract deuterium from EPHP-d
but cyclizes with total efficiency (Scheme 2).
4).14 [This is also in complete contrast to the findings of
Beckwith and Storey5 for a related bromide 8 (R ) R′ )
Me, X ) Br), which under Bu3SnH conditions at 80 °C
underwent 98% reduction and only 2% cyclization to
indolone.]
As mentioned, 19 is an advanced precursor in the Fuji
synthesis2e of horsfiline (1). We thus completed a formal
synthesis of the horsfiline (1) using the newly developed
phosphorus radical precursor, diethylphosphine oxide (DEPO).
In particular, this work shows the advantages of DEPO in
synthesis, compared to either tributyltin hydride or EPHP.
This surprising observation persuaded us to investigate an
alternative and more direct route to the indolones, avoiding
HAT and featuring, instead, direct radical generation in the
alkyl side chain, followed by cyclization. The most direct
route to a formal total synthesis of horsfiline would be by
synthesis of 19, an advanced precursor in the Fuji synthesis2e
of horsfiline. Accordingly, we prepared precursor 17 for
horsfiline synthesis, as illustrated in Scheme 4.
Scheme 3. Completion of EPHP Route to Horsfiline
Acknowledgment. We thank WestCHEM, EPSRC, Uni-
versities UK and the University of Strathclyde for funding,
and the EPSRC National Mass Spectrometry Service for high
resolution mass measurements.
Supporting Information Available: Experimental meth-
ods and spectroscopic data for selected compounds are
provided. This material is available free of charge via the
Internet at http://pubs.acs.org.
Ester 10b was prepared easily from pyrrolidine 9b11
(Scheme 2), and the phenylseleno group was introduced to
give selenide 15. This was directly reacted with p-anisidine
in the presence of trimethylaluminum9 to give amide 16. The
amide nitrogen of 16 was protected to produce the desired
precursor 17 in 85% yield (Scheme 4). This was then treated
with EPHP and AIBN in refluxing toluene, but only reduced
product 18 was obtained.12
We have recently established that DEPO4 has all of the
advantages of EPHP, but that it is significantly slower to
quench carbon radicals.13 Hence, it was tried in this reaction.
When precursor 17 was treated with DEPO and AIBN in
refluxing dry benzene, the substrate underwent smooth
radical cyclization and the only product obtained was the
desired cyclized product 19 in 85% isolated yield (Scheme
OL051095I
(11) (a) Cravotto, G.; Giovenzana, G. B.; Pilati, T.; Sisti, M.; Palmisano,
G. J. Org. Chem. 2001, 66, 8447. (b) Joucla, M.; Mortier, J. J. Chem. Soc.,
Chem. Commun. 1985, 1566.
(12) This may result from generation of PhSeH from attack by nucleo-
philes on small amounts of PhSeP(O)H(O)- formed from radical reaction
in situ. Protonation of the leaving PhSe- would generate PhSeH. The
excellent radical quenching ability of PhSeH has been established. See:
(a) Newcomb, M.; Manek, M. B.; Glenn, A. G. J. Am. Chem. Soc. 1991,
113, 949. (b) Crich, D.; Yao, Q. W. J. Org. Chem. 1995, 60, 84.
(13) Although the kinetics of hydrogen atom abstraction for DEPO and
EPHP appear not to have been studied quantitatively, kinetic data for
hydrogen abstraction from other H-P bonds (e.g., dialkyl phosphites) show
them to be relatively poor hydrogen donors. See: Chatgilialoglu, C.;
Timokhin, V. I.; Ballestri, M. J. Org. Chem. 1998, 63, 1327.
(14) A blank experiment was conducted in which DEPO was not added.
This led solely to recovery of starting material 17; no cyclization was
observed.
Org. Lett., Vol. 7, No. 15, 2005
3289