Scaffold hopping via a transannular rearrangement-encompassing cascade

Article abstract of DOI:10.1021/ol4000444

A novel reaction cascade for converting benzodiazepinediones into oxazoloquinolinones using carboxylic acid anhydrides in the presence of base has been developed using flow methods. Products are obtained in yields up to 98%.

Full text of DOI:10.1021/ol4000444

ORGANIC
LETTERS
2013
Vol. 15, No. 5
1052–1055
Scaffold Hopping via a Transannular
RearrangementꢀEncompassing Cascade
Johannes L. Vrijdag, An M. Van den Bogaert, and Wim M. De Borggraeve*
Department of Chemistry, KU Leuven, Celestijnenlaan 200F, 3001 Heverlee, Belgium
Wim.DeBorggraeve@chem.kuleuven.be
Received January 7, 2013
ABSTRACT
A novel reaction cascade for converting benzodiazepinediones into oxazoloquinolinones using carboxylic acid anhydrides in the presence of
base has been developed using flow methods. Products are obtained in yields up to 98%.
In the course of previous work in our group dealing with
the construction of (hetero) benzodiazepinediones (BDPs),1
two different compounds were unexpectedly isolated when
treating 3 with acetic anhydride (Scheme 1). Based on the
work of Pigeon and co-workers,2 we were expecting the
in situ generated open-chain precursor 3 to undergo ring
closure in the presence of acetic anhydride and K2CO3,
yielding its corresponding BDP. Imide 4a and compound 5a
were obtained instead from the reaction. Furthermore,
imide 4a, when brought back under the same reaction
conditions, gave rise to compound 5a, indicating its inter-
mediacy in a yet unreported cascade reaction. Not only the
supposed cascade reaction but also the compound obtained
(5a) caught our attention, as this scaffold is brought into
relation with numerous biological activities, including
Gly/NMDA receptor antagonism,3 modulating GABA
receptor inverse agonism,4 and inhibition of multidrug
resistance-associated protein 1 (MRP1).5
The literature hinted toward a plausible mechanism
for the observed cascade reaction:2,6ꢀ10 first, we assume
that BDP 6, our original compound of interest, is indeed
being formed during the reaction (Scheme 2). Subsequent
acylation of 6 yields the isolated imide 4. Comparable
cyclic imides (activated lactams) are known to undergo
a transannularrearrangement6ꢀ9 reactionunderbasiccon-
ditions, giving rise to acylated R-amino ketones 9.
Finally, compounds 9are cyclodehydrated in a Robinsonꢀ
Gabriel type reaction10 by the excess of carboxylic acid
anhydride present, yielding oxazoloquinolinones (OQOs) 5.
As this cascade reaction is most likely taking place via
a BDP intermediate 6, and also the synthesis of these
classical ‘privileged scaffolds’ has already been investi-
gated extensively,11 we decided to use BDPs as starting
compounds in our further optimization of the Transannular
Rearrangement of an Activated Lactam Encompassing
Cascade (TRALEC) reaction for the synthesis of OQOs.
(1) Van den Bogaert, A. M.; Nelissen, J.; Ovaere, M.; Van Meervelt,
L.; Compernolle, F.; De Borggraeve, W. M. Eur. J. Org. Chem. 2010,
2010, 5397.
(7) Farran, D.; Parrot, I.; Toupet, L.; Martinez, J.; Dewynter, G.
Org. Biomol. Chem. 2008, 6, 3989.
(8) Coursindel, T.; Restouin, A.; Dewynter, G.; Martinez, J.; Collette,
(2) Pigeon, P.; Othman, M.; Netchitailo, P.; Decroix, B. Tetrahedron
1998, 54, 1497.
(3) Calabri, F. R.; Colotta, V.; Catarzi, D.; Varano, F.; Lenzi, O.;
Filaccioni, G.; Costagli, C.; Galli, A. Eur. J. Med. Chem. 2005, 40, 897.
(4) Albaugh, P. U.S. Patent 5 182 290, 1993.
(5) Bonjouklian, R. W.O. Patent 0 127 116 A2, 2000.
(6) Farran, D.; Parrot, I.; Martinez, J.; Dewynter, G. Angew. Chem.,
Int. Ed. 2007, 46, 7488.
Y.; Parrot, I. Bioorg. Chem. 2010, 38, 210.
(9) Farran, D.; Archirel, P.; Toupet, L.; Martinez, J.; Dewynter, G.
Eur. J. Org. Chem. 2011, 2011, 2043.
(10) Robinson, R. J. Chem. Soc. 1909, 95, 2167.
(11) For recent comprehensive works, see: (a) Alvarez-Boilla, J.;
Vaguero, J. J.; Barluenga, J. Modern Heterocyclic Chemistry; Wiley-
VCH: 2011. (b) Katrizky, A. R.; Ramsden, C. A.; Scriven, E. F. V.; Taylor,
R. J. K. Comprehensive Heterocyclic Chemistry; Elsevier: 2008.
r
10.1021/ol4000444
Published on Web 02/11/2013
2013 American Chemical Society

Scheme 1. Conversion of 2-Aminobenzoic Derivative 1
Scheme 2. Mechanism Proposed for the Formation of 5
The N1-functionalized BDP starting materials were
conveniently prepared from isatoic anhydride using a
two-step procedure (Scheme 3). First, a [4 þ 3] cyclocon-
densation reaction of isatoic anhydride 11 with glycine was
conducted in DMSO, largely following the procedure as
described by Clark and co-workers.12 In the second step
unsubstituted BDP 12 was selectively alkylated on the
N1-position, using a procedure based on the reaction
protocol used by Cheng and co-workers.13 DMSO was
used as solvent in this series of alkylation reactions, as
to our knowledge it is the only common solvent in which
BDP 12 dissolves sufficiently. Apart from the reaction in
which benzyl bromide was used as an alkylating reagent,
N1-functionalized BDPs 6 were generally obtained in fair
yields. Indeed, benzyl bromide is known to undergo side
reactions with DMSO.14
Scheme 3. Preparation of BDPs 6
Optimization of the TRALEC reaction which converts
BDPs 6 to OQOs 5 (Scheme 4) started with the previously
applied reaction protocol, originally intended to convert
an open-chain precursor 3 to its corresponding BDP. BDP
6 and 3 equiv of K2CO3 were added to acetic anhydride
(solvent), which, after being stirred for 30 min at room
temperature, was refluxed for 24h. Unfortunately, wewere
facing two considerable problems while exploring afore-
mentioned reaction conditions. First, acceptable conver-
sions (>40%) were never achieved, and second, persistent
side reactions were observed giving rise to unidentified side
products of which the share increased linearly with reac-
tion time. We were to address both of these problems in a
systematic way.
To reduce the share of side reactions, first the reaction
mixture itself was modified. As we believed that the large
excess of acetic anhydride was mainly responsible for these
side reactions, a logical first adjustment was to reduce the
amount of acetic anhydride in the reaction. Although
stoichiometrically only 2 equiv of acetic anhydride are
required in the reaction cascade (Scheme 2), we decided
still to use 10 equiv toensure proper conversion of the BDP
starting material. o-Xylene was chosenas a suitable solvent
in which the reaction can be carried out, as it is known
as both a high boiling and a relatively inert solvent. Also,
to avoid excessive gas evolution, the K2CO3 base was
(12) Clark, R. L.; Carter, K. C.; Mullen, A. B.; Coxon, G. D.; Owusu-
Dapaah, G.; McFarlane, E.; Thi, M. D. D.; Grant, M. H.; Tettey,
J. N. A.; Mackay, S. P. Bioorg. Med. Chem. Lett. 2007, 17, 624.
(13) Cheng, M.-F.; Yu, H.-M.; Ko, B.-W.; Chang, Y.; Chen, M.-Y.;
Ho, T.-I.; Tsai, Y.-M.; Fang, J.-M. Org. Biomol. Chem. 2006, 4, 510.
(14) Park, K. H.; So, M. S.; Kim, Y. W. Bull. Korean Chem. Soc.
2005, 26, 1491.
Org. Lett., Vol. 15, No. 5, 2013
1053

Table 1. Synthesis of OQOs from BDPs 6 and Carboxylic Acid Anhydrides 13a
OQO
R1
C5H11
R2
reactor conditions
yield (%)b
14a
14b
14c
14d
14e
14f
14g
14h
5a
Me
nPr
iPr
Ph
300 °C, 200 μL/minc
300 °C, 200 μL/minc
300 °C, 200 μL/minc
300 °C, 200 μL/minc
300 °C, 100 μL/mind
300 °C, 100 μL/mind
300 °C, 100 μL/mind
220 °C, 400 μL/mine
220 °C, 400 μL/mine
220 °C, 400 μL/mine
220 °C, 400 μL/mine
220 °C, 400 μL/mine
94
98
94
69
83
87
86
69
69
79
79
64
C5H11
C5H11
C5H11
PhCH2CH2
PhCH2CH2
PhCH2CH2
PhCH2CH2
Bn
Me
nPr
iPr
Ph
Me
nPr
iPr
Ph
14i
14j
Bn
Bn
14k
Bn
a Reagents and conditions: BDP 6 (60 mg), anhydride 13 (10 equiv), DBU (5 equiv), o-xylene (5 mL), gently heated until complete dissolution of
reagents. The residence time in the flow reactor was 2 min. b Isolated yield after workup and silica gel chromatography. c Reactor 1: length (2 m), diameter
(0.51 mm). d Reactor 2: length (1 m), diameter (0.51 mm). e Reactor 3: length (1 m), diameter (1.02 mm).
substituted for potassium acetate, as this base is generated
in situ from acetic anhydride during previous reactions.
Implementation of this modified reaction protocol showed
that indeed the amount of side reactions was strikingly
reduced. However, similar conversions were observed as
before.
poor conversions observed in previous experiments indi-
cate that the RDS involves a considerably high activation
energy, thus making efficient heating of the reaction again
required to counteract side reactions (see Supporting
Information). The development of a microwave-assisted
reaction protocol was not undertaken because of specia-
lized experimental equipment demands for upscaling,
although several authors show this approach is actually
feasible.15ꢀ17
Scheme 4. Exploration of the TRALEC Reaction on BDPs 6
As in conventional batch approaches the reaction tem-
perature is usually limited by the boiling point of the
solvent, the focus of our research shifted toward more
efficient heating techniques to obtain higher conversions.
In our reaction cascade, the key transannular rearrange-
ment was also found to be the rate-determining step
(RDS). Indeed, after reaction, imides 4 were always found
in the reaction mixture while compounds 9 were not
(Scheme 2), indicating that both acylation of BDP 6 and
cyclodehydration of compound 9 to OQO 5 proceed faster
than the transannular rearrangement reaction. Also, the
Figure 1. Schematic overview of experimental flow setup.
On the other hand it was continuous flow chemistry in
particular that caught our attention. Flow technology not
only allow us to exceed the boiling point of the reaction
(15) Razzaq, T.; Kappe, C. O. Chem.;Asian J. 2010, 5, 1274.
€
(16) Ohrngren, P.; Fardost, A.; Russo, F.; Schanche, J.-S.; Fagrell,
M.; Larhed, M. Org. Process Res. Dev. 2012, 16, 1053.
(17) Sadler, S.; Moeller, A. R.; Jones, G. B. Expert Opin. Drug
Discov. 2012, 7, 1107.
1054
Org. Lett., Vol. 15, No. 5, 2013

solvent but also allows short retention times, avoiding
prolonged heating of the reaction mixture.
as it introduces a conveniently adjustable point of diversity
on the C2-position of the OQO skeleton.
For this purpose, a flow reactor was built in-house from
commercially available stainless steel HPLC tubing. The
reactor coil is heated by means of an HPLC column oven.
A schematic overview of the experimental setup is depicted
in Figure 1. Continuous flow chemistry, however, requires
homogeneous reaction conditions, which made it crucial
to identify an organic base that dissolves readily in the
reaction mixture before injection. After a screening of
different bases under batch conditions (see Supporting
Information), 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU)
was found to be the suitable base, as it both catalyzes the
TRALEC reaction and is soluble in the reaction mixture.
Also, the concentration of reactants had to be lowered to
ensure homogeneous conditions upon injection into the
flow reactor. As OQOs 14a, 14e, and 5a were isolated in
yields varying from 69% to 94%, we were encouraged to
also give alternative carboxylic acid anhydrides a try in the
TRALEC reaction. Indeed, the utilization of (iso)butyric
anhydride and benzoic anhydride analogously gave
rise to the corresponding OQOs in yields varying from
64% to 98% (Table 1). These interesting results show
that the TRALEC reaction might be implemented for
the preparation of more complex OQO based scaffolds,
In conclusion, an efficient reaction protocol has been
developed using a flow method for converting BDPs 6 to
pharmaceutically interesting OQOs in the presence of a
carboxylic acid anhydride 13 and base. Using this proto-
col, OQOs have been isolated in yields up to 98%. To
extend the scope of the TRALEC reaction even more, in
future work we are planning its implementation on other
diazepine systems as well as alternative lactam systems,
such as diketopiperazines.
Acknowledgment. This work was financially supported
by KU Leuven Project OT/11/047. A.M.V.D.B. thanks
FWO-Vlaanderen for a scholarship received. We are grate-
ful to K. Duerinckx (KU Leuven) for the assistance with
NMR measurements and to Ir. B. Demarsin (KU Leuven)
for HRMS measurements.
Supporting Information Available. Experimental pro-
cedures, characterization data, and copies of spectra for
all new compounds. This material is available free of
charge via the Internet at http://pubs.acs.org.
The authors declare no competing financial interest.
Org. Lett., Vol. 15, No. 5, 2013
1055