Synthesis of tylocrebrine and related phenanthroindolizidines by VOF 3-mediated oxidative aryl-alkene coupling

Article abstract of DOI:10.1021/ol1023954

A highly convergent strategy to prepare phenanthroindolizidines is reported involving three consecutive C-C coupling reactions. This sequence features a novel VOF₃-mediated aryl-alkene coupling in the final step, which enables regioselective preparation of C5-substituted phenanthroindolizidines for the first time. This strategy has been applied to the synthesis of eight natural and unnatural members in this class to investigate the scope of this chemistry and to explore structure-activity relationships.

Full text of DOI:10.1021/ol1023954

ORGANIC
LETTERS
2011
Vol. 13, No. 2
196-199
Synthesis of Tylocrebrine and Related
Phenanthroindolizidines by
VOF3-Mediated Oxidative Aryl-Alkene
Coupling
Micah J. Niphakis†,‡ and Gunda I. Georg*,‡
Department of Medicinal Chemistry, UniVersity of Kansas, 1251 Wescoe Hall DriVe,
Lawrence, Kansas 66045, United States, and Department of Medicinal Chemistry and
The Institute for Therapeutics DiscoVery & DeVelopment, UniVersity of Minnesota, 717
Delaware Street SE, Minneapolis, Minnesota 55414, United States
georg@umn.edu
Received October 18, 2010
ABSTRACT
A highly convergent strategy to prepare phenanthroindolizidines is reported involving three consecutive C-C coupling reactions. This sequence
features a novel VOF3-mediated aryl-alkene coupling in the final step, which enables regioselective preparation of C5-substituted
phenanthroindolizidines for the first time. This strategy has been applied to the synthesis of eight natural and unnatural members in this class
to investigate the scope of this chemistry and to explore structure-activity relationships.
The phenanthroindolizidine and phenanthroquinolizidine
alkaloids have received increasing attention over the past
decade for their newfound potent anticancer1 and anti-
inflammatory activity.2 Salient members of this class
tylophorine,3 antofine,4 and boehmeriasin A,5 for example
inhibit tumor growth with IC50 values in the low nanomolar
to picomolar range, on par with currently used anticancer
drugs. These medicinal properties have inspired numerous
total syntheses of the alkaloids and their unnatural analogs.
Yet despite a plethora of synthetic strategies,1,6 several
members of this class cannot be prepared efficiently.
The prevailing strategy to construct the phenanthrene
system in these natural products whether used early or later
† Department of Medicinal Chemistry, University of Kansas.
‡ Department of Medicinal Chemistry and The Institute for Therapeutics
Discovery & Development, University of Minnesota.
(1) Chemler, S. R. Curr. Bioact. Compd. 2009, 5, 2–19, and references
therein.
(3) (a) Ganguly, T.; Khar, A. Phytomedicine 2002, 9, 288–295. (b) Paull,
K. D.; Hamel, E.; Malspeis, L. In Cancer Chemotherapeutic Agents; Foye,
W. O., Ed.; American Chemical Society: Washington, DC, 1995; pp 9-45.
(c) Gao, W.; Lam, W.; Zhong, S.; Kaczmarek, C.; Baker, D. C.; Cheng,
Y. C. Cancer Res. 2004, 64, 678–688.
(2) (a) Gopalakrishnan, C.; Shankaranarayan, D.; Kameswarn, L.;
Natarajan, S. J. Med. Res. 1979, 69 (3), 513–520. (b) Gopalakrishnan, C.;
Shankaranarayanan, D.; Nazimudeen, S. K.; Kameswaran, L. Indian J. Med.
Res. 1980, 71, 940–948. (c) You, X.; Pan, M.; Gao, W.; Shiah, H. S.; Tao,
J.; Zhang, D.; Koumpouras, F.; Wang, S.; Zhao, H.; Madri, J. A.; Baker,
D.; Cheng, Y. C.; Yin, Z. Arthritis Rheum. 2006, 54 (3), 877–886. (d) Yang,
C. W.; Chen, W. L.; Wu, P. L.; Tseng, H. Y.; Lee, S. J. Mol. Pharmacol.
2006, 69 (3), 749–758.
(4) (a) Lee, S. K.; Nam, K. A.; Heo, Y. H. Planta Med. 2003, 69 (1),
21–25. (b) Stærk, D.; Lykkeberg, A. K.; Christensen, J.; Budnik, B. A.;
Abe, F.; Jaroszewski, J. W. J. Nat. Prod. 2002, 65, 1299–1302. (c) Fu, Y.;
Lee, S. K.; Min, H. Y.; Lee, T.; Lee, J.; Cheng, M.; Kim, S. Bioorg. Med.
Chem. Lett. 2007, 17, 97–100.
(5) (a) Luo, Y.; Liu, Y.; Luo, D.; Gao, X.; Li, B.; Zhang, G. Planta
Med. 2003, 69, 842–845. (b) Yan, J.; Luo, D.; Luo, Y.; Gao, X.; Zhang, G.
Int. J. Gynecol. Cancer 2006, 16, 165–170.
10.1021/ol1023954  2011 American Chemical Society
Published on Web 12/14/2010

in the reaction sequence is through dehydrogenative biaryl
coupling (Scheme 1). Although this approach is flexible with
poses regioselective control by exploiting the problematic
C5-substituent (R3) to block coupling at C5, thereby,
preventing regioisomer formation. The requisite precursor
2, in turn, could be derived from a Negishi reaction of
biphenylzinc reagent 3 and a triflate derived from indolizidine
4. Having recently developed a concise route to six-
membered cyclic enaminones involving the cyclization of
amino acid derived ynones,7 we thought these versatile
synthons would be apt precursors to enantiomerically pure
phenanthropiperidines. The third C-C bond could be realized
through a Suzuki-Miyaura biaryl coupling reaction from
iodobromobenzene (5) and boronic acid (6) precursors.
First looking to the indolizidine fragment, we began our
synthesis with commercially available Boc-D-ꢀ-homoproline
(Scheme 3). Weinreb amide R-7 could be formed directly
Scheme 1. Typical Strategy for Synthesis of Phenanthrene Moiety
respect to the aryl substituents being installed, it poses a
noteworthy limitation on the accessibility of particular
phenanthrene substitution patterns. Specifically, these reac-
tions cannot accommodate sterically disfavored substituents,
such as the obtrusive C5-methoxy group seen in tylocrebrine
(1a), with any practical degree of selectivity.
Scheme 3. Synthesis of Weinreb Amide
To avoid the problem of regioselectivity, we designed a
new approach for the phenanthrene construction that involved
three sequential C-C bond forming steps (Scheme 2). The
Scheme 2. Phenanthroindolizidine Retrosynthetic Analysis
using 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDCI).
In order to facilitate library development, a scalable approach
was developed for the synthesis of the racemic Weinreb
amide (()-7. Boc-pyrrolidinone was reduced to hemiaminal
9 and then subjected to HWE conditions. The resultant ring-
opened ester spontaneously cyclized to provide the racemic
ꢀ-proline derivative 10. Amide (()-7 was then obtained from
ester 10 using conditions reported by Williams et al.8
Although this sequence is more lengthy than that of amide
R-7, the payoff is the ease and economy of scale-up.
Constructing the indolizidine fragment was our next goal.
We have conducted extensive studies into the formation of
cyclic enaminones from ꢀ-amino ynones.7 Grignard addition
of ethynylmagnesium bromide to Weinreb amide R-7 provided
the desired ynone 11 in excellent yield (Scheme 4).
Indolizidine formation could be achieved using our
one-pot deprotection/cyclization strategy. Either 4 N HCl
in dioxane or NaI in HCO2H can be used for Boc-
deprotection.7b Although the yields of the latter conditions
(i.e., NaI/HCO2H) were generally higher, workup following
the cyclization was easier using the HCl method. It should
be noted that the use of HCl promotes retro-Michael/
Michael-type racemization of the ꢀ-stereocenter and, there-
fore, was not used for constructing indolizidine R-4. With
key step is the final ring closure, involving an aryl-alkene
dehydrogenative coupling. Unlike the standard biaryl-
coupling strategy to phenanthrenes, this disconnection im-
(6) For selected recent examples, see: (a) Yang, X.; Shi, Q.; Bastow,
K. F.; Lee, K. H. Org. Lett. 2010, 12, 1416–1419. (b) Wang, K. L.; Su, B.;
Wang, Z. W.; Wu, M.; Li, Z.; Hu, Y. N.; Fan, Z. J.; Mi, N.; Wang, Q. M.
J. Agric. Food. Chem. 2010, 58, 2703–2709. (c) Stoye, A.; Opatz, T. Org.
Lett. 2010, 12, 2140–2141. (d) Dumoulin, D.; Lebrun, S.; Couture, A.;
Deniau, E.; Grandclaudon, P. Eur. J. Org. Chem. 2010, 1943–1950. (e)
Yamashita, S.; Kurono, N.; Senboku, H.; Tokuda, M.; Orito, K. Eur. J.
Org. Chem. 2009, 1173–1180. (f) Wang, K.; Su, B.; Wang, Z.; Wu, M.;
Li, Z.; Hu, Y.; Fan, Z.; Mi, N.; Wang, Q. J. Agric. Food Chem. 2009, 58,
2703–2709. (g) Li, S. T.; Han, L.; Sun, L.; Zheng, D.; Liu, J.; Fu, Y. B.;
Huang, X. S.; Wang, Z. Y. Molecules 2009, 14, 5042–5053. (h) Cui, M. B.;
Wang, Q. M. Eur. J. Org. Chem. 2009, 5445–5451.
(7) (a) Turunen, B. J.; Georg, G. I. J. Am. Chem. Soc. 2006, 128, 8702–
8703. (b) Niphakis, M. J.; Turunen, B. J.; Georg, G. I. J. Org. Chem. 2010,
75, 6793–6805.
(8) Williams, J. M.; Jobson, R. B.; Yasuda, N.; Marchesini, G.; Dolling,
U. H.; Grabowski, E. J. J. Tetrahedron Lett. 1995, 36, 5461–5464.
Org. Lett., Vol. 13, No. 2, 2011
197

biarylzinc coupling partner. Only two biarenes, 13a and 13b,
were appended to the enantiopure indolizidine core as a proof
of concept and to complete the synthesis of (R)-tylocrebrine
for the first time. We were pleased to find that these cross-
couplings proceeded in under an hour at room temperature
providing near quantitative yields of indolizidines 2a-h
without significant variation (Table 2).
Scheme 4. Synthesis of Indolizidine Fragment
Table 2. Coupling of Indolizidine and Biaryl Fragments
the indolizidine core assembled, triflate 12 was prepared
using a 1,4-reduction and enolate trapping with Comins’
reagent.
We then turned to the synthesis of the biaryl fragment via
Suzuki-Miyaura cross-coupling. Fu¨rstner and Kennedy have
reported a straightforward route to biarylbromides that was
used for the synthesis of several phenanthropiperidine
alkaloids.9 This method was well-suited for our purposes
(Table 1). Although biarenes with ortho substituents had not
entry
R1
R2
R3
R4
R5
yield (%)a
1
2
3
4
5
6
7
8
OMe
OMe
OMe
OMe
OMe
H
OMe
OMe
OMe
OMe
OMe
OMe
H
OMe
H
OMe
H
H
H
OMe
OMe
H
OMe
H
OMe
OMe
H
H
OMe
H
95 (2a)
98 (2b)
97 (2c)
95 (2d)
98 (2e)
98 (2f)
95 (2g)
92 (2h)
H
OMe
OMe
OMe
H
H
OMe
H
H
Table 1. Synthesis of Biaryl Fragment
OMe
a Isolated yield.
Having all the desired carbons in place, we explored the
oxidative aryl-alkene coupling. To our knowledge there has
only been only one report of such a reaction.11 Concerned
by this lack of precedent, we undertook a systematic
investigation using reagents that have found use in the related
biaryl coupling reaction.12 Using indolizidine 2a as a model
system, we identified VOF3, which gave the desired product
in 70% yield by NMR, as a suitable reagent for this
transformation.
With this initial success, we applied this reaction to our
small collection of biarylindolizidines 2a-h to prepare the
corresponding phenanthroindolizidines 1a-1h (Table 3).
Notably, we did not observe the formation of regioisomers
in substrates lacking the C5-methoxy group. However, there
were noticeable differences between the reactivity of sub-
strates 2a and 2c, where R3 ) OMe, compared with the rest,
and as such, two variations of the oxidative coupling method
were employed. Substrates 2a and 2c required warming the
reaction mixture from -78 to -10 °C and addition of up to
4.5 equiv of VOF3 to drive the reaction to completion
(Method A). For all other substrates, where the C5-methoxy
group was absent, the reactions were kept at -78 °C and
only 2.2-2.5 equiv of VOF3 were required for complete
entry
R1
R2
R3
R4
R5
yield (%)a
1
2
3
4
5
6
7
8
OMe
OMe
OMe
OMe
OMe
H
OMe
OMe
OMe
OMe
OMe
OMe
H
OMe
H
OMe
H
H
H
OMe
OMe
H
OMe
H
OMe
OMe
H
H
OMe
H
72 (13a)
67 (13b)
77 (13c)
68 (13d)
70 (13e)
70 (13f)
75 (13g)
40b (13h)
H
OMe
OMe
OMe
H
H
OMe
H
H
OMe
a Isolated yield. b Isolated along with 4′,5′-dimethoxy-1,1′:2′,1′′-terphenyl
as an inseparable mixture.
been previously synthesized by this method, the reported
conditions worked well with these hindered boronic acids
(entries 1 and 3, Table 1). Coupling phenyl boronic acid to
1-bromo-2-iodo-4,5-dimethoxybenzene (entry 8), however,
was problematic due to the formation of an inseparable
impurity.10 Nevertheless, the mixture could still be used in
subsequent reactions without detrimental effects.
We next employed a Negishi cross-coupling reaction
between either triflate (()-12 or (R)-12 and the appropriate
(10) The impurity is presumably 4′,5′-dimethoxy-1,1′:2′,1′′-terphenyl result-
ing from a bis-phenylation of 1-bromo-2-iodo-4,5-dimethoxybenzene.
(11) Lapouyade, R.; Villeneuve, P.; Nourmamode, A.; Morand, J. P.
J. Chem. Soc., Chem. Commun. 1987, 776–778.
(9) Fu¨rstner, A.; Kennedy, J. W. J. Chem.sEur. J. 2006, 12, 7398–
7410.
(12) See Supporting Information.
198
Org. Lett., Vol. 13, No. 2, 2011

Table 3. Aryl-Alkene Biaryl Couplings
Table 4. Antiproliferative Activity of Phenanthroindolizidinesa
IC50 (nM)
product
paclitaxel
(R)-tylocrebrine (1a)
(()-tylocrebrine (1a)
(R)-tylophorine (1b)
(()-tylophorine (1b)
(()-1c
(()-antofine (1d)
(()-1e
(()-1f
(()-1g
(()-1h
COLO-205
MCF-7
NCI/ADR-RES
5.3
1.3
12
17
10
9.5
2.8
270
8.5
48
2.3
15
39
32
42
22
21
530
5.8
25
>6400
26
46
160
50
16
23
1300
200
180
180
29
21
a COLO-205 ) human colorectal adenocarcinoma; MCF-7 ) human
breast carcinoma; NCI-ADR-RES
adenocarcinoma.
) drug-resistant human ovarian
racemate shows greater activity in the NCI/ADR-RES cell
line than pure (R)-tylophorine. On the other hand, (R)-
tylocrebrine exhibits greater activity than its racemate, which
suggests that the (R)-enantiomer is more active. With the
exception of analog 1f, the removal or relocation of each of
the methoxy groups has surprisingly little effect on the
alkaloids’ activity. The NCI/ADR-RES cell line, however,
appears to be more sensitive to these structural modifications,
particularly in analogs 1e, 1f, 1g, and 1h. The retention of
activity across this series is an indication that the phenan-
throindolizidines are well suited for further lead development
in order to improve their potential for use in the clinic.
In summary, we have developed a highly convergent route
to phenanthroindolizidine alkaloids by aid of a novel aryl-
alkene oxidative coupling reaction. Most notably, this method
gives access to C5-substituted analogues, such as tylocrebrine
(1a), with complete regioselectivity for the first time. The
modularity of this approach allows for late stage diversifica-
tion and has enabled the synthesis and biological evaluation
of eight alkaloids with varying phenanthrene substitution
patterns.
a Isolated yield. b Method A: VOF3 (2.5 equiv), CH2Cl2, EtOAc, TFA,
TFAA, -78 °C; then VOF3 (1.5-2.0 equiv), -10 °C. Method B: VOF3
(2.5 equiv), CH2Cl2, EtOAc, TFA, TFAA, -78 °C.
conversion (Method B). We suggest that these observations
can be explained by the rates of intramolecular versus
intermolecular coupling. The C5-substituent evidently slows
the progression of this reaction resulting in the starting
material, product or both being shunted into competing
intermolecular coupling pathways. The end result is non-
productive consumption of VOF3, hence the requirement for
excess amounts of this reagent to prepare tylocrebrine (1a)
and its unnatural analogue 1c. In spite of the consequential
loss in yield, all of the desired products could be obtained
in sufficient quantities to assess their antiproliferative proper-
ties.
As seen in Table 4, these alkaloids are very potent
antiproliferative agents in COLO-205, MCF-7, and drug-
resistant NCI/ADR-RES cell lines. Most of these analogs
have comparable activity with paclitaxel for COLO-205 and
MCF-7 cells and retain their antiproliferative effects in the
paclitaxel-resistant NCI/ADR-RES cell line. It has been
previously shown that (S)-tylophorine is more active than
its R-enantiomer.13 Although this is not clear from the present
data in the COLO-205 and MCF-7 cells, tylophorine’s
Acknowledgment. The authors are grateful to Defang Tian
(University of Minnesota) for carrying out the antiprolifera-
tive assays and to Matthew W. Leighty (University of
Minnesota) for useful discussions about aryl-alkene coupling
conditions. This work was supported by the University of
Minnesota through the Vince and McKnight Endowed
Chairs. M.J.N. gratefully acknowledges the ACS-Division
of Medicinal Chemistry for a predoctoral fellowship.
Supporting Information Available: Experimental pro-
cedures and full spectroscopic data for all new compounds.
This material is available free of charge via the Internet at
http://pubs.acs.org.
(13) Gao, W.; Chen, A. P. C.; Leung, C. H.; Gullen, E. A.; Fu¨rstner,
A.; Shi, Q.; Wei, L.; Lee, K. H.; Cheng, Y. C. Bioorg. Med. Chem. Lett.
2008, 18, 704–709.
OL1023954
Org. Lett., Vol. 13, No. 2, 2011
199