A concise route to triazolobenzodiazepine derivatives via a one-pot alkyne-azide cycloaddition reaction.

Article abstract of DOI:10.1016/S0960-894X(02)00262-7

A new and efficient one-pot synthesis of [1,2,3]triazolo[1,5-a][1,4] benzodiazepin-6(4H)-ones is described starting from readily available anthranilic acids. A small array of the title compounds were assembled via a four-step sequence involving diazotisat

Full text of DOI:10.1016/S0960-894X(02)00262-7

Bioorganic & Medicinal Chemistry Letters 12 (2002) 1881–1884
A Concise Route to Triazolobenzodiazepine Derivatives Via a
One-Pot Alkyne-Azide Cycloaddition Reaction
Andrew W. Thomas*
F. Hoffmann-La Roche AG, Pharmaceuticals Division, Discovery Chemistry, Basel, CH 4070, Switzerland
Received 10 November 2001; accepted 10 January 2002
Abstract—A new and efficient one-pot synthesis of [1,2,3]triazolo[1,5-a][1,4] benzodiazepin-6(4H)-ones is described starting from
readily available anthranilic acids. A small array of the title compounds were assembled via a four-step sequence involving diazo-
tisation, azide addition followed by amide bond formation employing polymer supported carbodiimide and subsequent 1,3-dipolar
cycloaddition reaction. # 2002 Elsevier Science Ltd. All rights reserved.
Benzodiazepines have made a tremendous impact on the
quality of human life since their introduction, over 40
years ago, for the treatment of central nervous system
(CNS) disorders.1 During this time elegant and practical
synthesis of 1,4-benzodiazepines have been developed
that allow a diverse array of heterocyclic moieties to be
annelated to the benzodiazepine framework.2 Flumaze-
nil 1 and Estazolam 2, for example, possess an imida-
zole and 1,2,4-triazole moiety, respectively, and have
found both clinical and commercial success. In addition,
G-7453 3, which has an annulated tetrazole sub-unit,
has recently been reported to have potential utility as a
fibrinogen antagonist (Fig. 1).
annulated 1,2,3-triazoles.3 A route to [1,2,3]triazolo[1,5-
a][1,4] benzodiazepin-6(4H)-ones 4, from isatoic anhy-
drides 5, has recently been reported.4 However this
method was found unsuitable for parallel synthesis, due
in part, to the need to carefully extract the intermediate
aniline products 6 at alkaline pH. To complement
this route it was chosen to optimise for the direct
formation of derivative 4 proceeding to the desired targets
through an intramolecular azide-alkyne 1,3-dipolar
cycloaddition reaction. Additionally, due to the recent
popularisation of parallel methodologies in synthetic
chemistry it was aimed to design an operationally
simplistic protocol amenable to using semi-automated
synthesis as well as benefit from the advantages of
polymer supported reagents.5 Herein, we report the
realisation of these approaches in the development of
a short and reliable synthesis of a small array of tri-
cyclic triazolo-1,4-benzodiazepine derivatives. Two
methods explored are outlined in more detail below for
comparison.
Figure 1. Selected examples of benzodiazepines.
Chemistry
During recent lead evaluation studies within the CNS
disease area an efficient synthesis of [1,2,3]triazolo-
[1,5-a][1,4] benzodiazepin-6(4H)-ones 4, amenable to a
parallel chemistry approach was required.
Isatoic anhydride route (Scheme 1)
In our hands, the treatment of isatoic anhydrides 5 to
afford the anilines 6 was best achieved by running the
reaction in the presence of catalytic quantities of
DMAP according to literature procedures.6 Following
this method, the ortho-aminobenzamides 6 were rou-
tinely obtained in yields in excess of 85%, along with
traces of the urea-acid 8. In a parallel experiment it was
The intramolecular azide-alkyne 1,3-dipolar cycloaddi-
tion reaction is a well established method to prepare
*Fax: +61-688-6459; e-mail: andrew.thomas@roche.com
0960-894X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(02)00262-7

1882
A. W. Thomas / Bioorg. Med. Chem. Lett. 12 (2002) 1881–1884
Polymer-supported carbodiimides (Method A: Scheme
3). Surprisingly, employing polymer supported carbo-
diimides 14 led to a superior reaction.8 Indeed the
desired products 13 (R=Me) could be obtained in high
yield and purity by simply filtering the reaction mixture
followed by evaporation without the need for chroma-
tographic purification. This is an example where the use
of a polymer supported version of a reagent offers a dis-
tinct advantage in both the reaction and in the purification
step. In the event, it was found that the use of 2.0 equiv
of polymer supported carbodiimide along with 1.1 equiv
of freshly distilled N-methyl propargylamine gave the
best results. When the reaction was conducted in di-
chloromethane, the reaction proceeded instantaneously,
at room temperature, to form the 1,3-dipolar cycload-
dition product 15, which could be isolated in pure form.
The subsequent amide bond formation reaction
required reaction times from 30 min to 3 h in order to
proceed to completion.9
Scheme 1. Route to 4 via isatoic anhydride: Reagents and conditions:
(a) PS-DMAP, 0.1 equiv, MeNHCH2CꢁCH, 2.0 equiv, toluene,
reflux, 3 h, SiO2 chromatography, 60%; (b) NaNO2, HCl, NaN3, tol-
uene, reflux, 2 h, SiO2 chromatograhphy, 50%.
found that treatment of isatoic anhydride 5 with 2 equiv
of N-methyl propargylamine and 0.2 equiv of polymer
supported DMAP afforded the desired product 6 in
60% yield after chromatography. However substantial
quantities of the undesired acid 8 and two additional
unidentified impurities were detected by HPLC. Sub-
sequent diazotisation of 6 and in situ azide formation by
addition of sodium azide, followed by heating the result-
ing mixture under reflux in toluene, afforded the tricyclic
product 4 in 50% yield, after chromatography, as repor-
ted previously.4 However, this route was not considered
satisfactory for parallel synthetic exploitation.
Product isolation required removal of the reaction
solution by filtration and evaporation to afford the
desired compounds 13 (R=Me) in a pure form. How-
ever, when propargylamine was employed a different
reaction course took place via 16 and 17. When the
reaction was conducted in the presence of 2.0 equiv of
polymer supported carbodiimide along with 1.1 equiv of
freshly distilled propargylamine none of the desired
product was detected by HPLC. In contrast, it was
found that only a small amount of the starting ortho-
azido acid was recovered. This suggested that the poly-
mer supported carbodiimide was acting as a sequestering
agent for the acid (in the form of 16). However, when
using excess of the amine (3.5 equiv was optimal) the
amide bond formation reaction proceeded smoothly in
dichloromethane or toluene at 40 ꢀC to afford 17, which
can be isolated, in quantitative yield, if desired. The
subsequent dipolar cycloaddition step [17!13 (R=H)]
was sluggish when conducted in dichloromethane
requiring reaction times up to 36 h to proceed to com-
pletion. In contrast, when carrying out the reaction, in
the same pot, in toluene at 70 ꢀC the reaction proceeded
to completion within 3 h to afford 13 (R=H). The usual
filtration of the reaction mixture, after cooling, followed
by evaporation of the filtrate allowed the desired com-
pounds to be isolated in near quantitative yields.
Anthranilic acid route
Solution-phase carbodiimides (Scheme 2). Alternative
retrosynthetic analysis revealed the intermediate ortho-
azidobenzoic acid 9. This can be prepared by diazotisa-
tion and an azide displacement reaction of anthranilic
acids 10 which proceeded very efficiently following lit-
erature precedent.7 Purification of the products was
effected by simply acidifying the reaction mixture
resulting in precipitation of the ortho-azido benzoic acid
products 10, which were obtained after filtration and
drying, in greater than 95% purity and in near quanti-
tative yields, in each case, after removal of the aqueous
solution by filtration. Initial attempts to activate the
ortho-azido acids 9, towards amide bond formation
using conventional solution phase methods (DCCI,
EDCI, HOBt, etc.) met with little success and in no case
was the desired adduct isolated in pure form. In most
instances the activated ester 11 was the main product,
isolated in up to 60% yield along with traces of the
dipolar cycloaddition product 12, even after using
excess amine and elevated temperatures in attempts to
force the reaction to completion.
Scheme 2. Problematic solution-phase route.
Scheme 3. One-pot supported carbodiimide route.

A. W. Thomas / Bioorg. Med. Chem. Lett. 12 (2002) 1881–1884
Table 1. Representative products from 66 examples prepared
1883
afforded the reaction products 13 (R=Me, H) in high
yield after 15–30 min at room temperature via adducts 7
or 17, respectively. However, in this case the product
purity was somewhat lower (<80%) therefore the route
outlined in Scheme 3 was used to prepare a small array
of products starting from the appropriate anthranilic
acids (Table 1).
Entry
1
Structure
Yield
MS
94%a 75%b
200.16
2
3
4
5
6
7
8
96%
96%
97%
369.34
277.14
428.01
214.43
232.42
283.11
228.64
Conclusion
In conclusion, a conceptually simple, short and efficient
route to the triclyclic [1,2,3]triazolo[1,5-a][1,4] benzo-
diazepin-6(4H)-ones 13, has been developed employing
the use of polymer supported carbodiimide in a key
step. The approach has been shown to be suitable for
parallel chemical assembly and all products were iso-
lated in high yield and purity without the involvement
of chromatographic methods. Application of this meth-
odology to lead generation within CNS projects will be
reported elsewhere in due course.
99%a
60%b
Acknowledgements
Special thanks are extended to Drs Raffaello Masciadri
and Jurgen Wichmann for encouragement and helpful
discusions.
99%
96%
99%
References and Notes
1. Sternbach, L. H. In The Benzodiazepine Story 2nd ed.;
Hoffmann-La Roche, F., Ed.; Roche Scientific Services: Basel,
1983; p 5.
2. Fryer, R. I. In Comprehensive Heterocyclic Chemistry;
Katritzky, A. R., Rees, C. W., Eds. Pergamon: Oxford, 1984;
Vol. 3, p 539.
3. Padwa, A. In: 1,3-Dipolar Cycloaddition Chemistry; Padwa,
A., Ed.; Wiley-Interscience; New York, 1984; Vol. 2, Chapter
2.
4. Broggini, G.; Molteni, G.; Terraneo, A.; Zecchi, G. Tetra-
hedron 1999, 55, 14803.
aReaction yield following method A.
bReaction yield following method B.
5. Ley, S. V.; Baxendale, I. R.; Bream, R. N.; Jackson, P. S.;
Leach, A. G.; Longbottom, D. A.; Nesi, M.; Scott, J. S.;
Storer, R. I.; Taylor, S. J. J. Chem. Soc., Perkin Trans. 1 2000,
3815.
6. Venuti, M. C. Synthesis 1982, 266.
7. Rao, K. A. N.; Venkataraman, P. R. J. Ind. Chem. Soc.
1938, 15, 194.
Acid chlorides (Method B: Scheme 4). To complement
this route it was further shown that starting from the
ortho-azido acids 9 it was possible to prepare the desired
adducts 13, via 18 (Scheme 3). The required acid chlo-
ride 18 was formed by treatment of 10 with excess thio-
nyl chloride in dichloromethane (in the presence of
catalytic dimethylformamide) at ambient temperature
for 1 h. Evaporation of the reaction mixture followed by
further addition of dichloromethane, 2 equiv of the
amine (propargylamine of N-methyl propargylamine)
and two equivalents of the polymer supported base 19
8. All supported reagents used in this study are commercially
available from Argonaut. PS-DMAP (P/N 800288, 1.48 mmol
g
À1, PS-DCCI (P/N 800371 1.35 mmol gÀ1) and PS-DIEA (P/
N 800279 3.83 mmol gÀ1). The polymer supported EDCI
available from Fluka did not effect the desired tansformation.
9. Representative procedure for the synthesis of [1,2,3]tri-
azolo[1,5-a][1,4] benzodiazepin-6(4H)-ones. Method A: Reac-
tions performed in parallel using the Aronaut Quest 210: To a
mixture of the anthranilic acid 10 (1 mmol) was added an
aqueous solution HCl (50%, 4 mL) followed by a solution of
sodium nitrite (1.05 mmol) in water (1 mL). After 30 min, a
solution of sodium acetate (11.1 mmol) in water (2 mL) con-
taining sodium azide (1.1 mmol) was added dropwise and after
a further 15 min, the reaction mixture was acidified with HCl
(concd 3 drops) whereupon a precipitate formed. The reaction
solution was filtered off and the product was dried by purging
Scheme 4. One-pot acid chloride route.

1884
A. W. Thomas / Bioorg. Med. Chem. Lett. 12 (2002) 1881–1884
with an Argon flow for 15 min. The solid 9 (1 mmol) was then
dissolved in dichloromethane (4 mL) and 14 was added (2
mmol). After resin swelling, N-methyl propargylamine (1.1
mmol) was added and the reaction agitated for 90 min. Reac-
tion completion was judged by TLC and HPLC. Filtration of
the reaction mixture followed by evaporation afforded 5-
methyl[1,2,3]triazolo[1,5-a][1,4]benzodiazepine-6-(4H)-one 13
(R=Me) (0.99 mmol, 99%) as a white solid, mp 165–167 ꢀC;
1H NMR (CDCl3) d 8.1–8.0 (2H, m), 7.7 (1H, s, ¼NCH¼C),
7.7–7.6 (3H, m), 4.3 (2H, s, CH2) and 3.3 (3H, s, NCH3); ES–
MS 214.43 (M+). Propargylamine (3.5 mmol) can be used in
place of N-methyl propargylamine and toluene (5 mL) was
used as solvent (instead of DCM) and the reaction mixture
heated at 70 ꢀC for 3 h. Method B: Reactions performed in
parallel using the Radleys Carousel: To a mixture of the azido
acid 9 (1 mmol) was added thionyl chloride (5 mmol), DMF (1
drop) and the resulting mixture stirred at rt for 1 h. The mix-
ture was then evaporated by attachment of vaccuo to the
Argon manifold. To the resulting oil was added dichloro-
methane (5 mL) followed by propargylamine (2 mmol) and 19
(3 mmol). After 30 min, the reaction mixture was filtered and
evaporated to afford [1,2,3]triazolo[1,5-a][1,4]benzodiazepine-
6-(4H)-one 13 (R=H) as a white solid (0.75 mmol, 75%), mp
125–127 ꢀC; 1H NMR (CDCl3) d 8.2–8.0 (2H, m), 7.8–7.6 (2H,
m), 7.7 (1H, s,=NCH=C), 4.4 (2H, d, CH2); ES–MS 200.16
(M+).