A concise building block approach to a diverse multi-arrayed library of the circumdatin family of natural products

Article abstract of DOI:10.1055/s-2003-40870

Using a polymer-supported phosphine-mediated intramolecular aza-Wittig reaction in a key step of a reaction sequence a diverse library of benzodiazepine-quinazolinone alkaloids (Circumdatins) has been prepared using a novel modified Eguchi protocol. The multi-arrayed library generation strategy commenced from readily accessible benzodiazepinedione derivatives (and anthranilic acids) and all library members were purified by reverse phase preparative HPLC.

Full text of DOI:10.1055/s-2003-40870

PAPER
1707
A Concise Building Block Approach to a Diverse Multi-Arrayed Library of the
Circumdatin Family of Natural Products
Concise Building
B
l
lock A
f
pproach
r
to
a
D
iv
e
e
rse
M
ulti-
d
A
rrayed Library Grieder, Andrew W. Thomas*
F. Hoffmann-La Roche AG, Pharmaceuticals Division, Discovery Chemistry, 4070 Basel, Switzerland
Fax +41(61)6888714; E-mail: andrew.thomas@roche.com
Received 28 April 2003
Dedicated to Dr. Leo Sternbach, the father of Valium, on the ocassion of his 95th birthday.
to the family of Circumdatins recently isolated. Our ob-
Abstract: Using a polymer-supported phosphine-mediated in-
tramolecular aza-Wittig reaction in a key step of a reaction se-
quence a diverse library of benzodiazepine-quinazolinone alkaloids
(Circumdatins) has been prepared using a novel modified Eguchi
jective was to perform the chemistry in solution employ-
ing solid-supported reagents in a key step of the reaction
sequence. Herein, we describe the realization of these
protocol. The multi-arrayed library generation strategy commenced goals in the development of a concise building block ap-
from readily accessible benzodiazepinedione derivatives (and an-
proach to a diverse natural product-like library of Circum-
thranilic acids) and all library members were purified by reverse
datins.
phase preparative HPLC.
Key words: aza-Wittig reaction, polymer supported triphenylphos-
phine, natural product libraries, privileged scaffolds
Circumdatins A–G (Figure 1) have recently been isolated
from a terrestrial isolate of the fungus Aspergillus ochra-
ceus and are suggested to be suitable chemotaxonomic
markers for this species.1–3 Related alkaloids include the
benzomalvins A–C, which have been isolated from a fun-
gus identified as Penicillium sp and showed potent inhib-
itory activity against substance P at the guineau pig, rat
and human neurokinin NK1 receptor;4 and the asperlicins
which are known as potent cholescystokinin antagonists.
A total synthesis of Benzolmalvin A and B was recently
reported by Eguchi5 which employs an intramolecular
aza-Wittig reaction in a key step and Snider6 has complet-
ed a total synthesis of asperlicin. A further example is
Sclerotigenin,7 the simplest member of the benzodiaz-
epine-quinazolinone family, which was isolated from the
sclerotia of Penicillium sclerotigenum and has shown
promising anti-insectan activity.
The benzodiazepine moiety has received much attention
in the synthetic community, in part due to its representa-
tion as a member of the family of ‘privileged scaffolds’.8
The unique characteristics that allow it to mimic endo-
Figure 1 Isolated members of the Circumdatins and related benzo-
diazepine-quinazolinone alkaloids
genous ligands towards a number of different receptor tar-
gets have promoted its selection as a key scaffold in lead
Our synthetic strategy relied on an efficient formation of
generation for many pharmaceutical programs through
the fused quinazolinone ring system using polymer-sup-
conventional solution phase chemistry and solid phase
ported phosphine as a key step. To establish this objective
chemistry.
Sclerotigenin7 (Scheme 1), the simplest member of the
We aimed to combine the attractive attributes and features
of the privileged benzodiazepine scaffold annelated to a
fused quinazolinone ring to prepare a diverse pure library
of natural product-like benzodiazepine alkaloids similar
benzodiazepine-quinazolinone family was prepared as a
model for later exploiting this strategic bond formation in
the construction of our library. Starting from anthranilic
acid, the o-azidoacid chloride9 was prepared and treated
with the known benzodiazepinedione 1 to afford the de-
sired product 2 in excellent yield. The key step, an in-
tramolecular aza-Wittig reaction, proved to be uneventful
and provided the desired product 3 in good yield (75%),
Synthesis 2003, No. 11, Print: 05 08 2003.
Art Id.1437-210X,E;2003,0,11,1707,1711,ftx,en;C02003SS.pdf.
© Georg Thieme Verlag Stuttgart · New York

1708
A. Grieder, A. W. Thomas
PAPER
using polymer-supported triphenylphosphine (2.5 equiva- azepinediones 6 (Figure 2) and acid chlorides 7 (Figure 3)
lents). Removal of the dimethoxybenzyl (DMB) protect- following well-established literature methods.
ing group under acidic conditions10 (TFA and TfOH in
In the first library prepared we chose to extend the reac-
CH2Cl2) at room temperature revealed the natural product
tion times of the two steps required to allow sufficient
Sclerotigenin7 (4) in 85% yield.
time to ensure all reactions had proceeded to completion.
Therefore, although the conditions established for Scle-
rotigenin were complete within 1–5 hours per step, we de-
cided to prolong the reaction times (up to 48 h) to account
for the different inherent reactivities of the building
blocks due to the electronic nature of the substituents.
Scheme 1 Total Synthesis of Sclerotigenin: Reagents and Conditi-
ons: a) o-Azidobenzoylchloride, Huenigs Base, CH2Cl2, r.t., 2 h,
88%; b) PS-PPh2 (2.5 equiv), 5 h, 60 ºC, CH2Cl2–toluene (1:1), 75%;
c) TFA (10 equiv), TfOH (0.1 equiv), CH2Cl2, r.t., 4 h, 85%
Given the success and efficiency of this approach we set
about to prepare a diverse natural product-like library of
Circumdatins using the strategy outlined in Scheme 2.
Figure 2 Diversity reagents for libraries 8a,b
Scheme 2 Retrosynthesis planning of the “Circumdatin“ library in-
corporating two anthranilic acids and one -amino acid
This retro-biosynthetic analysis revealed that only one
Figure 3 Diversity reagents for libraries 8a,b, 11, 12
main synthetic building block (substituted anthranilic ac-
ids 5) is required for preparation of the target library along
Following this route, two sub-libraries with 121 family
members were prepared in the 5–50 mg range (Scheme 3).
The diversity of reagents 6a–h and 6i–o were used to pre-
pare the sub-libraries 8a and 8b, respectively, by iterative
array preparation in combination with 7a–l.11 In general,
all combinations of 6 and 7 worked well. However, the
para-nitro derivative 7f performed poorly in all reactions
and was replaced with 7m in all subsequent reactions.
with -substituted amino acids for the preparation of 6.
This is consistent with a biosynthetic rational which, via
14C labelling experiments, has been postulated for the
preparation of family members in nature having chosen to
efficiently combine two units of a suitably substituted an-
thranilic acid with one unit of an -amino acid.
Thus far, only a few members of this natural product fam-
ily have been isolated but inevitably several new members
will be identified in the future, so that rapid analogue syn-
thetic access would be highly desirable. Within our corpo-
rate collection we had access to a wide variety of the
necessary acids 5 and these were used to prepare benzodi-
The yields of these reactions were usually in the range of
40–60% per step and it was decided to perform a further
purification step by preparative HPLC so that the overall
yield after purification was generally in the range of 30%.
The product integrity was established initially by analysis
Synthesis 2003, No. 11, 1707–1711 © Thieme Stuttgart · New York

PAPER
Concise Building Block Approach to a Diverse Multi-Arrayed Library
1709
tives (from anthranilic and -amino acids). All library
members were purified by reverse phase preparative
HPLC yielding a success rate of 74% (283 isolated pure
products from 384 individual reactions; 4 of 12 × 8 ar-
rays) over the two steps. We remain confident that the li-
brary members formed in these iteratively parallel
assembled arrays will produce new leads for several ongo-
ing research programs within our therapeutic areas and
further investigations will be reported in due course.
Scheme 3 Preparation of the two sub-libraries; iterative arrays of
6a–h with 7a–l; reactions performed in a Buechi syncore: a) 7a–l (1.0
equiv), Huenigs Base (1.5 equiv), dioxane, r.t., 24–36 h; b) PS-PPh2
(2.5 equiv), dioxane, r.t. to 60 ºC, 24–36 h, then preparative HPLC
All reagents and solvents were purchased from Fluka and used di-
rectly without further purification. All reactions were performed in
a Buechi syncore apparatus using PTFE reaction tubes and the reac-
tion solutions evaporated in situ. After dissolution in DMF–dioxane
reversed phase chromatography [Dynamax pumps and detector in
combination with a Gilson 215 liquid handler on a YMC-ODS-A
column (50 × 20 mm)] eluting with a gradient of CH3CN–H2O (20-
95%) followed by evaporation of the elution solvents yielded the
pure desired products (HPLC UV Detector at 230 nM). Mass spec-
tra were recorded on a Micromass Platform II mass spectrometer at
70 eV. All doublet and tripling coupling constants are 7.5 Hz.
of the MS spectra followed by 1H NMR analysis for ran-
dom library members. Purity was established by HPLC
and was generally in excess of 95% for all successfully
prepared compounds.
Following the same strategy, the libraries incorporating
tricyclic derivatives 9 and 10 (Figure 4), similar to the
Circumdatins D and E, were prepared and the library of
products 11,12 (162 members) was characterized after
preparative HPLC purification (Scheme 4).
6,7-Dihydro-6,8,13a-triaza-benzo[3,4]cyclohepta[1,2-b]naph-
thalene-5,13-dione (4, Sclerotigenin)
To a suspension of 1 (500 mg, 1.5 mmol) in CH2Cl2 (10 mL) was
added Huenigs base (297 mg, 2.2 mmol) followed by freshly pre-
pared o-azidobenzoylchloride (240 mg, 1.5 mmol) and the resulting
mixture was stirred at r.t. for 2 h. After standard aqueous work-up,
the solid residue (605 mg, 88%) was dissolved in CH2Cl2–toluene
(1:1, 10 mL), then polymer-supported diphenylphosphine (2.5
equiv, loading = 2.9 mmolg–1) was added and the resulting mixture
heated at 60 ºC for 5 h. After cooling to r.t., the mixture was filtered
and the resin washed with DMF (2 × 2 mL). The filtrate was evap-
orated and the solid residue (435 mg, 75%) was suspended in
CH2Cl2 (10 mL). Then TFA (0.1 mL) and TfOH (0.01 mL) were
added drop-wise and the reaction stirred vigorously for 4 h at r.t.
The mixture was then evaporated, H2O was added and the product
extracted with CH2Cl2. Purification on silica gel (CH2Cl2–EtOAc,
1:1) afforded the title compound (240 mg, 85%) as an off-white sol-
id.
1H NMR (300 MHz, CDCl3): = 8.30 (d, 1 H), 7.95 (d, 1 H), 7.80
(t, 1 H), 7.69 (d, 1 H), 7.65 (m, 1 H), 7.55 (m, 2 H), 7.05 (br s, 1 H),
4.32 (dd, 1 H), 4.22 (dd, 1 H).
Figure 4 Diversity of reagents for libraries 11, 12
MS (EI): m/z (%) = 277.1.
HPLC purity: 100%.
Representative Examples from Iterative Array: {[6a–h] + [7a–
l]}
[6a + 7b]
White solid; 22 mg; HPLC purity: 97%.
1H NMR (300 MHz, DMSO-d6): = 8.91 (d, 1 H), 8.30 (d, 1 H),
8.28 (d, 1 H), 8.12 (d, 1 H), 7.80 (d, 1 H), 7.71–7.56 (m, 5 H), 4.55
(d, 1 H), 4.21 (d, 1 H), 3.12 (s, 3 H).
Scheme 4 Preparation of the ‘Cirumdatin D/E’ libraries; iterative
arrays of 9a–h or 10a–h with 7a–g, i–m; reactions performed in a
Buechi syncore: a) 7a–l (1.0 equiv), Huenigs Base (1.5 equiv), diox-
ane, r.t., 24–36 h; b) PS-PPh2 (2.5 equiv), dioxane, r.t. to 60 ºC, 24–
36 h, then preparative HPLC
MS (EI): m/z (%) = 341.4.
9-Chloro-4-iodo-6-methyl-6,7-dihydro-6,8,13a-triaza-ben-
zo[3,4]cyclohepta[1,2-b]naphthalene-5,13-dione [6d + 7e]
White solid; 11 mg; HPLC purity: 100%.
1H NMR (300 MHz, DMSO-d6): = 8.15 (d, 1 H), 8.09 (d, 1 H),
In summary, we have exemplified an efficient multi-ar-
rayed library generation strategy to produce four sub-li-
braries of the Circumdatin family of natural products. The
8.05 (d, 1 H), 7.61 (d, 1 H), 7.58 (t, 1 H), 7.32 (t, 1 H), 4.55 (d, 1 H),
synthetic sequence is short, relies on a modified Eguchi 4.21 (d, 1 H), 3.08 (s, 3 H).
protocol using polymer-supported triphenylphosphine,
and used readily accessible benzodiazepinedione deriva-
MS (EI): m/z (%) = 451.7.
Synthesis 2003, No. 11, 1707–1711 © Thieme Stuttgart · New York

1710
A. Grieder, A. W. Thomas
PAPER
4-Bromo-9,11-dichloro-6-methyl-6,7-dihydro-6,8,13a-triaza-
benzo[3,4]cyclohepta[1,2-b]naphthalene-5,13-dione [6c + 7a]
White solid; 11 mg; HPLC purity: 100%.
1H NMR (300 MHz, DMSO-d6): = 8.20 (d, 1 H), 7.90 (t, 1 H),
7.85 (d, 1 H), 7.71 (d, 1 H), 7.65–7.51 (m, 4 H), 4.65 (d, 1 H), 3.62
(m,1 H), 3.48 (m, 1 H), 2.94 (m, 1 H), 2.12 (m, 2 H), 2.00 (m, 1 H).
1H NMR (300 MHz, DMSO-d6): = 8.22 (d, 1 H), 8.09 (d, 1 H),
7.89 (d, 1 H), 7.61 (d, 1 H), 7.52 (t, 1 H), 4.55 (d, 1 H), 4.20 (d, 1
H), 3.07 (s, 3 H).
MS (EI): m/z (%) = 317.4.
[9h + 7e]
White solid; 23 mg; HPLC purity: 96%.
MS (EI): m/z (%) = 439.1.
1H NMR (300 MHz, DMSO-d6): = 8.15 (d, 1 H), 8.05 (d, 1 H),
7.70–7.55 (m, 4 H), 4.72 (d, 1 H), 3.65 (m, 1 H), 3.42 (m, 1 H), 2.94
(m, 1 H), 2.20 (m, 2 H), 2.05 (m, 1 H).
4-Chloro-3-fluoro-6,9,11-trimethyl-6,7-dihydro-6,8,13a-triaza-
benzo[3,4]cyclohepta[1,2-b]naphthalene-5,13-dione [6h + 7j]
White solid; 9 mg; HPLC purity: 96%.
1H NMR (300 MHz, DMSO-d6): = 7.86 (d, 1 H), 7.68 (m, 2 H),
7.60 (d, 1 H), 4.68 (d, 1 H), 4.15 (d, 1 H), 3.08 (s, 3 H), 2.55 (s, 3
H), 2.42 (s, 3 H).
MS (EI): m/z (%) = 386.2.
[9h + 7a]
White solid; 32 mg; HPLC purity: 96%.
MS (EI): m/z (%) = 371.8.
1H NMR (300 MHz, DMSO-d6): = 8.22 (d, 1 H), 8.15 (d, 1 H),
7.85 (d, 1 H), 7.70–7.55 (m, 3 H), 4.66 (d, 1 H), 3.62 (m, 1 H), 3.43
(m, 1 H), 2.93 (m, 1 H), 2.15 (m, 2 H), 2.01 (m, 1 H).
Representative Examples from Iterative Array: {[6i–o] + [7a–
l]}
6-(2,4-Dimethoxybenzyl)-10,11-dimethoxy-6,7-dihydro-
6,8,13a-triaza-benzo[3,4]cyclohepta[1,2-b]naphthalene-5,13-di-
one [6i + 7d]
MS (EI): m/z (%) = 386.2.
[9d + 7l]
White solid; 26 mg; HPLC purity: 94%.
White solid; 47 mg; HPLC purity: 98%.
1H NMR (300 MHz, DMSO-d6): = 8.13 (m, 1 H), 7.85 (m, 1 H),
7.50–7.35 (m, 3 H), 4.60 (d, 1 H), 3.62 (m, 1 H), 3.42 (m, 1 H), 2.88
(m, 1 H), 2.11 (m, 2 H), 2.00 (m, 1 H).
1H NMR (300 MHz, DMSO-d6): = 7.82 (d, 1 H), 7.58 (m, 3 H),
7.49 (s, 1 H), 7.09 (d, 1 H), 7.00 (s, 1 H), 4.68 (q, 2 H), 4.39 (d, 1
H), 4.05 (d, 1 H), 3.93 (s, 3 H), 3.90 (s, 3 H), 3.70 (s, 3 H), 3.52 (s,
3 H).
MS (EI): m/z (%) = 367.4.
MS (EI): m/z (%) = 487.5.
[9g + 7k]
White solid; 17 mg; HPLC purity: 98%.
11-Bromo-3-chloro-6-(2,4-dimethoxybenzyl)-6,7-dihydro-
6,8,13a-triaza-benzo[3,4]cyclohepta[1,2-b]naphthalene-5,13-di-
one [6j + 7i]
1H NMR (300 MHz, DMSO-d6): = 7.71–7.60 (m, 3 H), 7.52–7.48
(m, 2 H), 4.80 (d, 1 H), 3.90 (s, 3 H), 3.63 (m, 1 H), 3.43 (m, 1 H),
2.89 (m, 1 H), 2.15 (m, 2 H), 2.00 (m, 1 H).
White solid; 29 mg; HPLC purity: 99%.
1H NMR (300 MHz, DMSO-d6): = 8.31 (d, 1 H), 8.00 (dd, 1 H),
7.83 (d, 1 H), 7.72 (m, 1 H), 7.67 (m, 1 H), 7.49 (d, 1 H), 7.09 (d, 1
H), 6.40 (d, 1 H), 6.35 (s, 1 H), 4.65 (q, 2 H), 4.50 (d, 1 H), 4.05 (d,
1 H), 3.73 (s, 3 H), 3.60 (s, 3 H).
MS (EI): m/z (%) = 399.8.
Representative Examples from Iterative Array: {[10a–h] + [7a–
g,i–m]}
[10c + 7b]
White solid; 15 mg; HPLC purity: 100%.
MS (EI): m/z (%) = 540.8.
1H NMR (300 MHz, DMSO-d6): = 9.00 (s, 1 H), 8.39 (s, 1 H),
8.29 (d, 1 H), 8.15 (d, 1 H), 7.70–7.60 (m, 3 H), 7.52–7.49 (m, 2
H), 5.18 (d, 1 H), 4.14 (q, 1 H), 3.97 (m, 1 H), 3.14 (m, 1 H), 2.80
(m, 1 H).
3-Chloro-6-(2,4-dimethoxybenzyl)-10,11-difluoro-6,7-dihydro-
6,8,13a-triaza-benzo[3,4]cyclohepta[1,2-b]naphthalene-5,13-di-
one [6j + 7l]
White solid; 32 mg; HPLC purity: 99%.
1H NMR (300 MHz, DMSO-d6): = 8.10 (m, 1 H), 7.80 (d, 1 H),
7.72 (m, 1 H), 7.62 (m, 2 H), 7.10 (d, 1 H), 6.40 (d, 1 H), 6.35 (s, 1
H), 4.65 (q, 2 H), 4.50 (d, 1 H), 4.05 (d, 1 H), 3.73 (s, 3 H), 3.62 (s,
3 H).
MS (EI): m/z (%) = 371.4.
[10e + 7k]
White solid; 14 mg; HPLC purity: 92%.
1H NMR (300 MHz, DMSO-d6): = 8.00 (d, 1 H), 7.90 (d, 1 H),
7.80 (d, 1 H), 7.70 (d, 1 H), 7.59 (d, 1 H), 7.55 (d, 1 H), 5.20 (d, 1
H), 4.13 (q, 1 H), 3.90 (s, 3 H), 3.89 (m, 1 H), 3.14 (m, 1 H), 2.82
(m, 1 H).
MS (EI): m/z (%) = 497.9.
4-Chloro-6-(2,4-dimethoxybenzyl)-10,11-difluoro-6,7-dihydro-
6,8,13a-triaza-benzo[3,4]cyclohepta[1,2-b]naphthalene-5,13-di-
one [6o + 7j]
MS (EI): m/z (%) = 401.3.
White solid; 18 mg; HPLC purity: 100%.
1H NMR (300 MHz, DMSO-d6): = 8.10 (t, 1 H), 7.70–7.58 (m, 4
H), 7.09 (d, 1 H), 6.40 (d, 1 H), 6.35 (dd, 1 H), 4.65 (q, 2 H), 4.50
(d, 1 H), 4.01 (d, 1 H), 3.70 (s, 3 H), 3.68 (s, 3 H).
[10f + 7d]
White solid; 34 mg; HPLC purity: 99%.
1H NMR (300 MHz, DMSO-d6): = 7.70 (d, 1 H), 7.60 (t, 1 H),
7.55–7.50 (m, 2 H), 5.09 (d, 1 H), 4.10 (q, 1 H), 3.95 (s, 3 H), 3.82
(s, 3 H), 3.81 (m, 1 H), 3.12 (m, 1 H), 2.79 (m, 1 H).
MS (EI): m/z (%) = 497.9.
Representative Examples from Iterative Array: {[9a–h] + [7a–
g,i–m]}
MS (EI): m/z (%) = 397.8.
[9a + 7c]
White solid; 32 mg; HPLC purity: 98%.
Synthesis 2003, No. 11, 1707–1711 © Thieme Stuttgart · New York

PAPER
Concise Building Block Approach to a Diverse Multi-Arrayed Library
1711
(5) (a) Sugimori, T.; Okawa, T.; Eguchi, S.; Yashima, E.;
Acknowledgments
Kakehi, A.; Okamoto, Y. Tetrahedron 1998, 54, 7997.
(b) Sugimori, T.; Okawa, T.; Eguchi, S.; Yashima, E.;
Okamoto, Y. Chem. Lett. 1997, 869.
We would like to thank Drs Geo Adam, Alex Alanine, Konrad Blei-
cher, Christian Hubschwerlen and Gerard Schmid for support of
this work.
(6) He, F.; Foxman, B. M.; Snider, B. B. J. Am. Chem. Soc.
1998, 120, 6417.
(7) (a) Joshi, B. K.; Gloer, J. B.; Wicklow, D. T.; Dowd, P. F. J.
Nat. Prod. 1999, 62, 650. (b) Total synthesis of
Circumdatin F and Sclerotigenin: Snider, B. B.; Busuyek,
M. N. Tetrahedron 2001, 57, 3301.
(8) Herpin, T. F.; Van Kirk, K. G.; Salvino, J. M.; Yu, S. T.;
Labaudiniere, R. F. J. Comb. Chem. 2000, 2, 513.
(9) Thomas, A. W. Bioorg. Med. Chem. Lett. 2002, 12, 1881.
(10) Masciadri, R.; Thomas, A. W.; Wichmann, J. PCT WO
200240487, 2002.
References
(1) Dai, J.-R.; Carte, B. K.; Sidebottom, P. J.; Yew, A. L. S.; Ng,
S.-W.; Huang, Y.; Butler, M. S. J. Nat. Prod. 2001, 64, 125.
(2) Raebaek, L.; Breinholt, J. J. Nat. Prod. 1999, 62, 904.
(3) Raebaek, L.; Breinholt, J.; Frisvald, J. C.; Christopherson, C.
J. Org. Chem. 1999, 64, 1689.
(4) Sun, H. H.; Barrow, C. J.; Sedlock, D. M.; Gillum, A. M.;
Cooper, R. J. Antibiotics 1994, 47, 515.
(11) For a more detailed discussion of workflow procedures and
automation technology utilized for synthetic arrays, see:
Nettekoven, M.; Thomas, A. W. Curr. Med. Chem. 2002, 9,
2179.
Synthesis 2003, No. 11, 1707–1711 © Thieme Stuttgart · New York