1240 J. Org. Chem.

1997, 62, 1240-1256

Solid-Phase Synthesis of 1,4-Benzodiazepine-2,5-diones. Library

Preparation and Demonstration of Synthesis Generality

Constantine G. Boojamra,† Kristina M. Burow, Lorin A. Thompson, and Jonathan A. Ellman*

Department of Chemistry, University of California, Berkeley, California 94720

Received December 9, 1996X

A general and expedient method has been developed for the solid-phase synthesis of 1,4-
benzodiazepine-2,5-diones 1 from three commercially available components; anthranilic acids,
R-amino esters, and alkylating agents. Reaction conditions have been developed to prepare either
racemic compounds for lead identification efforts or optically pure compounds for lead optimization
efforts. The incorporation of diverse functionality into the benzodiazepine products has also been
demonstrated. On the basis of the scope and generality of the synthesis sequence, a library of
1,4-benzodiazepine-2,5-diones has been prepared from 11 alkylating agents, 12 anthranilic acids,
and 19 R-amino esters (nine sets of enantiomeric pairs and glycine methyl ester hydrochloride).
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

The library was prepared in a spatially separate format using a microtiter-based apparatus that
is inexpensive and straightforward to construct from ordinary items found in an organic or bioorganic
Downloaded via UNIV DE GUADALAJARA on November 11, 2023 at 22:54:27 (UTC).

laboratory. The high quality of the 1,4-benzodiazepine-2,5-dione library has been demonstrated
by evaluating representative compounds by HPLC analysis and 1H NMR.

Introduction

The preparation and evaluation of small-molecule

libraries has become an important part of the drug
discovery process.1-5 In these efforts, particular empha-
sis has been placed upon the preparation and evaluation
of libraries based upon priviledged templates. The
display of different functionality upon these templates
Figure 1. Benzodiazepine templates for library synthesis.
has previously provided a number of potent and specific
drugs or candidates toward different therapeutic targets.6 antibiotics12-19 as well as to the benzodiazepine receptor
The 1,4-benzodiazepines are an important class of antagonist, flumazenil.20 Structures related to flumaze-
priviledged templates, and numerous derivatives have nil have experimental applications as ethanol-intoxica-
been identified that have selective activities against a tion antagonists.21-24 The 1,4-benzodiazepine-2,5-diones
diverse array of biological targets.7 A subset of the 1,4- have also shown promise as herbicides.25 In addition, the
benzodiazepines, the 1,4-benzodiazepine-2,5-diones 1 1,4-benzodiazepine-2,5-dione core appears in a number
(Figure 1), is the focus of this work. Derivatives have of natural products.26-29
shown promise as antithrombotics.8-11 They serve as
precursors to the anthramycin family of antitumor (9) Blackburn, B. K.; Lee, A.; Baier, M Atropisomeric GPIIbIIIa
antagonists. Presented at the 209th National Meeting of the American
Chemical Society, Anaheim, CA; American Chemical Society: Wash-
† Present Address: Microcide Pharmaceuticals Inc., 850 Maude ington, DC, 1995; ORGN 280.
Avenue, Mountain View, CA 94043. (10) McDowell, R. S.; Blackburn, B. K.; Gadek, T. R.; McGee, L. R.;
X Abstract published in Advance ACS Abstracts, February 1, 1997. Rawson, T.; Reynolds, M.; Robarge, K. D.; Somers, T. C.; Thorsett, E.
(1) Thompson, L. A.; Ellman, J. A. Chem. Rev. 1996, 96, 555-600. D.; Tischler, M.; Webb, R. R.; Venuti, M. C. J. Am. Chem. Soc. 1994,
(2) Choong, I. C., Ellman, J. A. Annu. Rep. Med. Chem. 1996, 31, 116, 5077-5083.
309-318. (11) McDowell, R. S.; Gadek, T. R.; Barker, P. L.; Burdick, D. J.;
(3) Terrett, N. K.; Gardner, M.; Gordon, D. W.; Kobylecki, R. J.; Chan, K. S.; Quan, C. L.; Skelton, N.; Strumble, M.; Thorsett, E. D.;
Steele, J. Tetrahedron 1995, 51, 8135-8173. Tischler, M.; Tom, J. Y. K.; Webb, T. R.; Burnier, J. J. Am. Chem. Soc.
(4) Gordon, E. M.; Barrett, R. W.; Dower, W. J.; Fodor, S. P. A.; 1994, 116, 5069-5076.
Gallop, M. A. J. Med. Chem. 1994, 37, 1385-1401. (12) Ishikura, M.; Mori, M.; Terashima, M.; Ban, Y. J. Chem. Soc.,
(5) Früchtel, J. S.; Jung, G. Angew. Chem., Int. Ed. Engl. 1996, 35, Chem. Commun. 1982, 741-742.
17-42. (13) Mori, M.; Uozumi, Y.; Kimura, M.; Ban, Y. Tetrahedron 1986,
(6) Ariens, E. J.; Beld, A. J.; Rodrigues de Miranda, J. F.; Simonis, 42, 3793-3806.
A. M. In The Receptors: A Comprehensive Treaty; O’Brien, R. D., Ed.; (14) Reed, J. N.; Snieckus, V. Tetrahedron Lett. 1984, 25, 5505.
Plenum: New York, 1979; Vol. 1; pp 33-41. (15) Kaneko, T.; Wong, H.; Doyle, T. W. Tetrahedron Lett. 1983, 24,
(7) Selected 1,4-benzodiazepine therapeutic agents or promising 5165-5168.
candidates: (a) Anxiolytics, hypnotics, and antiepileptics: Sternbach, (16) Stevens, R. V.; Cory, R. M.; Rossen, S. J. Chem. Soc., Chem.
L. H. J. Med. Chem. 1979, 22, 1-7. (b) Cholecystokinin antagonists: Commun. 1975, 742.
Evans, B. E. et al. J. Med. Chem. 1988, 31, 2235-2246. (c) Opioid (17) Confalone, P. N.; Huie, E. M.; Ko, S. S.; Cole, G. M. J. Org.
receptor antagonist: Romer, D. et al. Nature 1982, 298, 759-760. (d) Chem. 1988, 53, 482-487.
HIV Tat antagonist: Hsu, M.-C. et al. Science 1991, 254, 1799-1802. (18) Peña, M. R.; Stille, J. K. Tetrahedron Lett. 1987, 28, 6573-
(e) HIV reverse transcriptase inhibitor: Pauwels, R. et al. Nature 1990, 6576.
343, 470-474. (f) Platelet aggregating factor antagonist: Miyazawa, (19) Peña, M. R.; Stille, J. K. J. Am. Chem. Soc. 1989, 111, 5417-
S. et al. Chem. Pharm. Bull. 1992, 40(2), 521-523 (thienodiazepine 5424.
derivative). (g) Antipsychotics: Liégeois, J.-F. F.; Bruhwyler, J.; Damas, (20) Hunkeler, W. Chimia 1993, 47, 141-147.
J.; Nguyen, T. P.; Chleide, E. M. G.; Mercier, M. G. A.; Rogister, F. A.; (21) Suzdak, P. D.; Glowa, J. R.; Crawley, J. N.; Schwartz, R. D.;
Delarge, J. E. J. Med. Chem. 1993, 36, 2107-2114. (h) Oxytocin Skolnick, P.; Paul, S. M. Science 1986, 234, 1243-1247.
antagonist: MDDR MDL Drug Data Report data base. MDL Informa- (22) Polc, P.; Richards, J. G. Br. J. Pharmacol. 1985, 86, 465P.
tion Systems, Inc., 14600 Catalina St., San Leandro, CA. (23) Malminiemi, O.; Korpi, E. R. Eur. J. Pharmacol. 1989, 169, 53-
(8) Webb, R. R.; Barker, P. R.; Baier, M.; Reynolds, M. E.; Robarge, 60.
K. D.; Blackburn, B. K.; Tichler, M. H.; Weese, K. J. Tetrahedron Lett. (24) Bonetti, E. P.; Burkard, W. P.; Gabl, M.; Mohler, H.; Richards,
1994, 35, 2113-2116. J. G. Br. J. Pharmacol. 1985, 86, 463P.

S0022-3263(96)02284-0 CCC: $14.00 © 1997 American Chemical Society

Synthesis of 1,4-Benzodiazepine-2,5-diones J. Org. Chem., Vol. 62, No. 5, 1997 1241

We30-32 and others33 have previously reported the (Figure 1) is derived from glycine (R3 ) H). We did not
preparation of libraries of 1,4-benzodiazepines that be- consider palladium-catalyzed carbonylative insertion due
long to the 1,4-benzodiazepin-2-one class 2 (Figure 1). to the need for high temperatures and high pressures of
Recently, we have also reported a high-yielding and CO and poor yields.12,13,49-52 Finally, we chose not to
general method for the solid-phase synthesis of 1,4- employ the aza-Wittig cyclization53 because this approach
benzodiazepine-2,5-diones 1.34 The strategy comple- would require the use of 2-azidobenzoic acids as one of
ments our synthesis of structures 2 because it allows the components, few of which are commercially available.
much greater diversity to be introduced at the R1 site. Upon selecting lactamization as the cyclization method,
Two alternative solid-phase approaches to 1,4-benzodi- we then designed a sequence that relied on the incorpo-
azepine-2,5-diones have now also been reported by other ration of three different components, anthranilic acids
researchers.35,36 Here, we describe further exploration (R1), alkylating agents (R2), and R-amino esters (R3). Over
of the generality of our chemistry and application of the 30 anthranilic acids, 50 R-amino esters with the ap-
chemistry to the preparation of a 2508-member library propriate side-chain protection, and 100 alkylating agents
of benzodiazepines 1. are commercially available.54 To expedite library genera-
tion, we required that incorporation of the anthranilic
Results and Discussion acid derivatives be accomplished without prior protection
We sought to achieve three goals in designing an of either the aniline or carboxylic acid functionality.
approach to prepare libraries of 1,4-benzodiazepine- The optimized synthesis approach is shown in Scheme
2,5-diones: (1) Several different building block sets 1 and begins with the attachment of the R-amino ester
should be incorporated to provide rapid access to a large to the solid support, followed by acylation with an
number of diverse compounds. (2) The building blocks anthranilic acid, base-catalyzed lactamization, alkylation,
used in the synthesis of the library should be readily and cleavage. A critical feature of the strategy is the
accessible and ideally commercially available to facilitate attachment to the support as an N-alkyl substituent 5,
rapid library synthesis. (3) The chemistry should be which provides a tertiary amide lactamization precursor
compatible with the display of as much functionality as 6 (Scheme 1). Literature pertaining to the cyclizations
possible including reactive functionality that is commonly of small peptides indicates that the presence of backbone
found in drugs. amide N-alkyl groups55-60 enhances the equilibrium cis/
In order to acheive these goals, literature regarding trans ratio about these tertiary amides and, in a small
the classical synthesis of 1,4-benzodiazepine-2,5-diones peptide cyclization precursor, greatly enhances the rate
indicated that closure of the seven-membered ring would of formation of cyclic peptide over dimer or oligomer.60-63
be the cornerstone of any approach. Ring closure via 1,4-Benzodiazepine-2,5-dione Synthesis Sequence.
lactamization was the most often used method and According to this strategy, (chloromethyl)polystyrene
clearly provided the highest yields and the most (Merrifield resin) is derivatized by alkylation with the
generality.37-46 We rejected alkylative cyclizations,10,47,48 sodium salt of 4-hydroxy-2,6-dimethoxybenzaldehyde (3)
since such closures are only precedented in the case to provide resin-bound aldehyde 4 (Scheme 1).64,65 An
where the amino acid portion of the benzodiazepine 1 R-amino ester is then loaded onto the support by reduc-
tive amination employing NaBH(OAc)3 in 1% acetic acid
(25) Karp, G. M. J. Org. Chem. 1995, 60, 5814-5819.
(26) White, J. D.; Haefliger, W. E.; Dimsdale, M. J. Tetrahedron
1970, 26, 233-242. (47) Yamamoto, H.; Inaba, S.; Nakao, M.; Maruyama, I. Chem.
(27) Smith, H.; Wegfahrt, P.; Rapoport, H. J. Am. Chem. Soc. 1968, Pharm. Bull. 1969, 17, 400-403.
90, 1668-1669. (48) Lee, C.-M. J. Heterocycl. Chem. 1964, 1, 235-238.
(28) Richter, H.; Winter, K.; Kousy, S. E.; Luckner, M. Pharmazie (49) Ishikura, M.; Mori, M.; Ikeda, T.; Terashima, M.; Ban, Y. J.
1974, 29, 506-510. Org. Chem. 1982, 47, 2456-2461.
(29) Martin, P. K.; Rapoport, H.; Smith, H. W.; Wong, J. L. J. Org. (50) Mori, M.; Ishikura, M.; Ikeda, T.; Ban, Y. Heterocycles 1981,
Chem. 1969, 34, 1359-1363. 16, 1491-1494.
(30) Bunin, B. A.; Plunkett, M. J.; Ellman, J. A. Methods Enzymol. (51) Mori, M.; Purvaneckas, G. E.; Ishikura, M.; Ban, Y. Chem.
1996, 267, 448-465. Pharm. Bull. 1984, 32, 3840-3847.
(31) Bunin, B. A.; Plunkett, M. J.; Ellman, J. A. Proc. Natl. Acad. (52) Mori, M.; Kimura, M.; Uozumi, Y.; Ban, Y. Tetrahedron Lett.
Sci. U.S.A. 1994, 91, 4708-4712. 1985, 26, 5947-5950.
(32) Bunin, B. A.; Ellman, J. A. J. Am. Chem. Soc. 1992, 114, (53) Eguchi, S.; Yamashita, K.; Matsushita, Y. Synlett 1992, 295-
10997-10998. 296.
(33) DeWitt, S. H.; Kiely, J. S.; Stankovic, C. J.; Schroeder, M. C.; (54) Structure searches were performed using the Available Chemi-
Cody, D. M. R.; Pavia, M. R. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, cals Directory (ACD), MDL Information Systems, Inc., 14600 Catalina
6909-6913. St., San Leandro, CA 94577.
(34) Boojamra, C. G.; Burow, K. M.; Ellman, J. A. J. Org. Chem. (55) Cavelier-Frontin, F.; Pèpe, G.; Verducci, J.; Siri, D.; Jacquier,
1995, 60, 5742-5743. R. J. Am. Chem. Soc. 1992, 114, 8885-8890.
(35) Goff, D. A.; Zuckermann, R. N. J. Org. Chem. 1995, 60, 5744- (56) Dale, J.; Titlestad, K. J. Chem. Soc., Chem. Commun. 1969,
5745. 656-659.
(36) Moroder, L.; Lutz, J.; Grams, F.; Rudolph-Böhner, S.; Ösapay, (57) Pastuszak, J.; Gardner, J. H.; Singh, J.; Rich, D. H. J. Org.
G.; Goodman, M.; Kolbeck, W. Biopolymers 1996, 38, 295-300. Chem. 1982, 47, 2982-2987.
(37) Cho, N. S.; Song, K. Y.; Párkányi, C. J. Heterocycl. Chem. 1989, (58) Steiner, J. R.; Barnes, C. L. Int. J. Pept. Protein Res. 1988, 31,
26, 1807-1810. 212-219.
(38) Hoffmann, E.; Jagnicinski, B. J. Heterocycl. Chem. 1966, 3, (59) Titlestad, K. Acta Chem. Scand., B 1977, 31, 641-661.
348-351. (60) Jacquier, R.; Lazaro, R.; Raniriseheno, H.; Viallefont, P. Int. J.
(39) Carabateas, P. M.; Harris, L. S. J. Med. Chem. 1966, 9, 6-10. Pept. Protein Res. 1987, 30, 22-32.
(40) Carabateas, P. M. U.S. Patent 3 415 814, 1968. (61) Cavelier-Frontin, F.; Achmad, S.; Verducci, J.; Jacquier, R.;
(41) Carabateas, P. M. U.S. Patent 3 384 635, 1968. Pèpe, G. THEOCHEM 1993, 286, 125-130.
(42) Uskokovic, M.; Iacobelli, J.; Wenner, W. J. Org. Chem. 1962, (62) Stewart, W. E.; T. H. Siddall, I. Chem. Rev. 1970, 70, 517-
27, 3606-3608. 551.
(43) Uskokovic, M. R., Wenner, W. U.S. Patent 3 261 828, 1966. (63) Vitoux, B.; Aubry, A.; Cung, M. T.; Marraud, M. Int. J. Pept.
(44) Akssira, M.; Boumzebra, M.; Kasmi, H.; Dahdouh, A.; Roumes- Protein Res. 1986, 27, 617-632.
tant, M. L.; Viallefont, P. Tetrahedron 1994, 50, 9051-9060. (64) Landi, J. J.; Ramig, K. Synth. Commun. 1991, 21, 167-171.
(45) Akssira, M.; Kasmi, H.; Dahdouh, A.; Boumzebra, M. Tetrahe- (65) In initial studies, a monomethoxy analog of the dimethoxy
dron Lett. 1992, 33, 1887-1888. linker described herein was initially employed. However, at the end
(46) Akssira, M.; Boumzebra, M.; Kasmi, H.; Dahdouh, A.; Roumes- of the synthesis treatment with TFA resulted in release of the linker
tant, M. L.; Viallefont, P. Synth. Commun. 1993, 23, 2265-2272. from the support still covalently bound to the benzodiazepine.
1242 J. Org. Chem., Vol. 62, No. 5, 1997 Boojamra et al.

Scheme 1

in DMF. In order to minimize racemization, it is es- in DMF/THF (1:1) for 30 h followed by addition of an
sential in this step that preequilibration of the resin- appropriate alkylating agent provides a fully derivatized,
bound aldehyde and the R-amino ester be minimized support-bound benzodiazepine 8. Complete cyclization
(vide infra) before addition of the reducing agent to the and alkylation (>95%) are observed according to this
reaction mixture. We commonly employ two different reaction sequence as determined after cleavage of ben-
protocols for racemization free reductive amination. The zodiazepines 8 from support. Cleavage is accomplished
NaBH(OAc)3 can be premixed with resin-bound aldehyde in good yields and high purity by final treatment with
4 in 1% acetic acid in DMF followed by addition of an TFA/Me2S/H2O (90:5:5).
R-amino ester. Alternatively, the R-amino ester and According to the above-described synthesis sequence,
reductant can be premixed in 1% acetic acid in DMF the benzodiazepine products 1 are obtained without any
followed by addition of resin-bound aldehyde 4. loss of optical purity at the stereocenter introduced by
Acylation of the resulting secondary amine 5 with the R-amino ester. Racemization in the reductive ami-
commercially available unprotected anthranilic acids nation, anthranilic acid acylation, and cyclization steps
then provides the support-bound tertiary amide 6. Op- could readily be detected by chiral HPLC analysis. Less
timization of this acylation step required considerable than 1% racemization was observed for either benzodi-
experimentation. For example, even the highly activated azepine 1b or benzodiazepine 1c (Figure 2).
azabenzotriazole-based reagents recently developed by It was also necessary to show that no racemization
Carpino, such as HATU,66 gave poor reaction conversion. occurs in the alkylation step. Unfortunately, since enan-
Carbodiimides were the only coupling agents found to tiomers of alkylated benzodiazepines separate poorly on
efficiently effect this transformation. Furthermore, good Pirkle chiral HPLC columns,68 it was necessary instead
yields of acylated material were obtained only when the to hydrolyze the benzodiazepine and analyze the chiral
carbodiimides were employed in conjunction with the integrity of the liberated amino acid. A sample of
hydrochloride salt of a tertiary amine. EDC (1-ethyl-3- benzodiazepine 1o (Figure 2) was heated at reflux in 6
[3-(dimethylamino)propyl] carbodiimide‚hydrochloride) N aqueous hydrochloric acid in order to liberate leucine.
proved to be the most efficient reagent since it contains The methyl ester was formed using thionyl chloride in
an internal tertiary amine hydrochloride salt. To ensure methanol, and separate aliquots of the amino ester were
that complete acylation occurs, the resin is subjected then acylated with (+)-MTPA chloride and with (-)-
twice to this coupling procedure. MTPA chloride. Gas chromatographic analysis indicated
Solution studies indicated that base-catalyzed lactam- a 97/3 ratio of amino acid enantiomers. The 3% minor
ization would be the most general way of producing the enantiomer most likely arose during the benzodiazepine
support-bound cyclic product 8 (R2 ) H). Ideally, cy- hydrolysis step69 and not in the course of benzodiazepine
clization would be accomplished under conditions suf- synthesis.
ficiently basic to provide the anilide anion 7 for subse- Chirality is derived from the commercially available
quent alkylation to introduce the R2 components of R-amino ester starting materials. In some cases, the
compounds 8 in the same reaction step. The lithium salt D-enantiomer of the starting material is much more
of acetanilide proved to be an optimal base for effecting expensive than the naturally occurring enantiomer or
these transformations. Because acetanilide has a pKa of may not be available from commercial sources. For lead
21.5 in DMSO,67 deprotonation and alkylation of amides, identification efforts it would be important to access both
esters, or carbamates does not occur since these func- enantiomers of the 1,4-benzodiazepine-2,5-dione products
tionalities are all considerably more basic, with pKa’s in order to maximize the diverse display of functionality.
greater than 24 in DMSO. Treatment of 6 with this base
(68) Pirkle, W. H.; Tsipouras, A. J. Chromatogr. 1984, 291, 291-
298.
(66) Carpino, L. A.; Elfaham, A.; Albericio, F. Tetrahedron Lett. (69) Benoiton, N. L. Quantitation and Sequence Dependence of
1994, 35, 2279-2282. Racemization in Peptide Synthesis. In The Peptides; Gross, E.; Meien-
(67) Bordwell, F. G. Acc. Chem. Res. 1988, 21, 456-463. hofer, J., Eds.; Academic: New York, 1983; Vol. 5, pp 221-222.
Synthesis of 1,4-Benzodiazepine-2,5-diones J. Org. Chem., Vol. 62, No. 5, 1997 1243

Figure 2. Benzodiazepine products. Yields are of purified material and are based upon the amino ester loading levels of the
resin as determined according to the procedure described in the Experimental Section: (*) products formed under racemizing
conditions described in text from enantiopure starting materials; (†) racemic products from commercially available racemic material.

We therefore chose to develop conditions whereby racemic Products and Yields. We have synthesized a diverse
benzodiazepine products could be obtained under mild array of 1,4-benzodiazepine-2,5-diones to demonstrate the
conditions from the enantiomerically pure R-amino ester versatility of the described synthesis sequence and to
starting materials. Complete racemization could be determine which derivatives should be incorporated into
accomplished, presumably by imine tautomerization, by a library. We have been able to incorporate an array of
preequilibrating the R-amino ester‚HCl salt, 0.3 equiv of sterically and electronically diverse anthranilic acids,
i-Pr2EtN, and the resin-bound aldehyde 4 for 6 h before both unfunctionalized and functionalized R-amino esters,
addition of the NaBH(OAc)3. Chiral HPLC analysis of and a variety of alkylating agents.
benzodiazepine product 1c prepared using these initial Simple amino acids work well in this protocol, as well
reductive amination conditions confirmed that complete as those with functionalized side chains: Benzodiaz-
racemization had in fact occurred. epines 1f, 1i, 1v, and 1n are derived from glutamic acid,
1244 J. Org. Chem., Vol. 62, No. 5, 1997 Boojamra et al.

thienylalanine, tyrosine, and lysine, respectively. Com- epine 1d was cleaved from support, an aliquot of resin
pound 1g is derived from glutamine. Glutamine methyl was subjected to cross-coupling with p-methoxyben-
ester, when incorporated without side-chain protection, zeneboronic acid employing Pd(PPh3)4 as the catalyst in
provided a succinimide side product resulting from side- THF at reflux to furnish 1u after cleavage from support.
chain cyclization with the methyl ester. When the Benzodiazepine 1v is an example of coupling reaction
dimethoxybenzhydrol group70 was used as a side-chain with an alkylborane derived from B-hexyl-9-BBN (pre-
protecting group, glutamine provided products in com- pared in situ by hydroboration) with a benzodiazepine
parable yield to other amino acids. derived from 4-bromoanthranilic acid. We have also
Attempts to incorporate serine met with failure, as the explored Suzuki reactions employing conditions reported
large majority of material isolated after attempting to by Johnson in the synthesis of prostaglandin analogs75
synthesize a serine-containing benzodiazepine was de- because the conditions reportedly allow reactions to be
rived from elimination of the β-hydroxyl group.71 The performed at significantly lower temperatures. Product
incorporation of valine proceeds with low conversion due 1w was synthesized in 78% overall yield employing a
to incomplete acylation in the anthranilic acid acylation DMF/THF/H2O solvent system and bis(diphenylphosphi-
step as determined by recovery of significant amounts no)ferrocenepalladium(II) chloride [PdCl2(dppf)] as cata-
of unacylated valine methyl ester after cleavage from lyst. Heating the reaction mixture to 50 °C was required
support. in order to ensure reaction completion.
We have been able to incorporate anthranilic acids Library Synthesis. The library was prepared pre-
containing electron-rich (e.g., 1m) and electron-poor dominantly to demonstrate that our synthetic method
functionality (e.g. 1i), a host of halogenated structures proceeds with high fidelity, even when executed in a
(e.g., 1c,d,n,j), as well as heterocyclic analogs (1p,t). parallel format. To this end, rigorous characterization
Unfortunately, attempts to alkylate 1t led to product that of representative members of the library was required,
was only 60% alkylated, presumably because the pKa of and thus, a spatially separate parallel synthesis was
the anilide proton of the thienodiazepine 1t is slightly employed. The development of detailed structure-activ-
higher than that of the other benzodiazepines synthe- ity relationships upon biological evaluation of the spa-
sized. tially separate library should also be straightforward.
According to our synthesis strategy, the anthranilic The key to producing a high-quality library in a
acids are employed without aniline protection. For the parallel format requires the identification of the ap-
compounds shown in Figure 2, no aniline acylation was propriate solid support on which to synthesize the library,
observed as determined by 1H NMR or TLC analysis of and to a similar extent, the apparatus used to hold and
the unpurified benzodiazepine products from support. filter the support so that the synthesis can be performed
The presence of electron-donating functionality on the efficiently. There are a number of paradigms reported
anthranilic acid para with respect to the aniline does in the literature for the spatially separate manipulation
result in significant acylation of the unprotected aniline. of solid supports in the context of parallel synthesis.76-78
In the case of 5-iodoanthranilic acid, the desired product Our most important selection criteria was that we employ
was contaminated with approximately 15-20% of higher apparatus configured in a standard 96-well microtiter
order oligomers, and 5-acetamidoanthranilic acid pro- format. A great deal of instrumentation has been
vided only oligomeric products. developed for the manipulation of materials in a micro-
In order to further increase the diversity of the ben- titer format, e.g., multichannel pipet devices, filtration
zodiazepine products, we have demonstrated the func- units, and microtiter-based concentrators.
tionalization of aryl bromide-containing benzodiazepines Library Synthesis Employing Chiron Mimotopes
with the Suzuki coupling reaction.72 This reaction is an Pin Apparatus. We intially employed the Chiron
ideal carbon-carbon bond-forming method for the syn- Mimotopes pin apparatus, originally developed by Geysen
thesis of compounds by combinatorial methods due to its for peptide epitope mapping.79-81 In this apparatus, 96
compatibility with a wide range of functionality and high polyethylene pins are placed into a supporting block so
reaction yields. In addition, there are many commer- that each pin fits into a separate well of a 96-well
cially available arylboronic acids, and alkylboranes can microtiter plate, thus allowing use of microtiter-based
readily be accessed through in situ hydroboration meth- instrumentation. In addition, we had previously pre-
ods. There are now also a number of reports in the pared a number of 1,4-benzodiazepin-2-one libraries30,31
literature of solid-phase applications of these reac- and a β-turn mimetic library82 employing this apparatus.
tions.2,73,74 Unfortunately, we encountered two significant problems
Benzodiazepines synthesized from 4- and 5-bromoan- with the Mimitopes apparatus in the synthesis of librar-
thranilic acids have been successfully subjected to Suzuki ies of 1,4-benzodiazepine-2,5-diones. First, the pins were
cross-coupling reaction conditions. Before benzodiaz-
(75) Johnson, C. R.; Braun, M. P. J. Am. Chem. Soc. 1993, 115,
11014-11015.
(70) König, W.; Geiger, R. Chem. Ber. 1970, 103, 2041-2051. (76) Dewitt, S. H.; Kiely, J. S.; Stankovic, C. J.; Schroeder, M. C.;
(71) In preliminary studies, good yields of 1,4-benzodazepine-2,5- Cody, D. M. R.; Pavia, M. R. Proc. Natl. Acad. Sci. U.S.A. 1993, 90,
diones incorporating serine were obtained by performing the cyclization 6909-6913.
with pyridinium hydrochloride in DMF at 90 °C for 48 h. However, (77) Geysen, H. M.; Meloen, R. H.; Barteling, S. J. Proc. Natl. Acad.
the trifluoroacetic acid-mediated cleavage step resulted in significant Sci. U.S.A. 1984, 81, 3998-4002.
trifluoroacetylation of the primary hydroxyl. We have not previously (78) Meyers, H. V.; Dilley, G. J.; Durgin, T. L.; Powers, T. S.;
observed trifluoroacetylation of secondary alcohols or phenols during Winssinger, N. A.; Zhu, H.; Pavia, M. R. Mol. Diversity 1995, 1, 13-
the cleavage step. Benzodiazepines incorporating phenylalanine and 20.
glycine were also prepared in good yield using these modified cycliza- (79) Valerio, R. M.; Bray, A. M.; Maeji, N. J. Int. J. Pept. Protein
tion conditions. Res. 1994, 44, 158-165.
(72) Miyaura, N.; Suzuki, A. Chem. Rev. 1995, 95, 2457-2481. (80) Geysen, H. M.; Rodda, S. J.; Mason, T. J.; Tribbick, G.; Schoofs,
(73) Backes, B. J.; Ellman, J. A. J. Am. Chem. Soc. 1994, 116, P. G. J. Immun. Methods 1987, 102, 259.
11171-11172. (81) Maeji, N. J.; Bray, A. M.; Valerio, R. M.; Wang, W. Peptide Res.
(74) Frenette, R.; Friesen, R. W. Tetrahedron Lett. 1994, 35, 9177- 1995, 8, 33-38.
9180. (82) Virgilio, A. A.; Schurer, S.; Ellman, J. A. Tetrahedron, in press.
Synthesis of 1,4-Benzodiazepine-2,5-diones J. Org. Chem., Vol. 62, No. 5, 1997 1245

acids, 11 alkylating agents, and 10 R-amino esters (19 if

both enantiomers are counted). These are shown in
Figure 4.
The synthesis of the library was then carried out in a
split-split format. Equal portions of resin were first
derivatized with the 10 R-amino esters. The R-amino
esters derived from alanine, naphthylalanine, and thi-
enylalanine were loaded under the nonracemizing condi-
tions, since they were available in racemic form. Glycine
was also loaded under these conditions. The remaining
R-amino esters were loaded employing the racemizing
conditions. In every case, infrared spectroscopy indicated
complete disappearance of the aldehyde, which was
Figure 3. Multitube parallel synthesis apparatus. replaced by a strong ester carbonyl stretch. The flasks
were weighed, and equal portions of the product resin
only available derivatized with amine functional groups. mass was removed from each flask and placed into each
This required that we reengineer our linker for attach- of 12 12-mL filter cartridges fitted with 70-µm hydro-
ment through amine functionality. Of more serious phobic polyethylene frits (available from Applied Separa-
concern, we observed variable yields in 1,4-benzodiaz- tions). At the end of this splitting operation, 120 filter
epine-2,5-dione synthesis on pins, indicating that sig- cartridges had been charged with resin.
nificant chemistry reoptimization would be necessary. We After the EDC-mediated acylations were performed
therefore decided to explore alternative library synthesis and the resins rinsed clean, the beads were suspended
apparatus that would allow us to use commercially as an isopycnic slurry83 in 3:2 dichloroethane/DMF.
available resin. Under isopycnic conditions, the solvated beads and the
Library Synthesis on Beads. The Diversomer ap- solvent have the same density, and therefore, the beads
paratus of DeWitt and co-workers allows the use of do not settle after agitation and remain evenly distrib-
polystyrene beads;33 however, it is not compatible with uted for resin transfer. All of the resins in all 120 filter
a microtiter format. We therefore decided to design a cartridges were taken up in the same volume of this
high-throughput parallel synthesis apparatus that could mixture, and aliquots of equal volume were removed from
be made quickly with relatively little expense and that the cartridges and placed in the appropriately addressed
would be compatible with a microtiter format; it is shown tubes in the multitube reaction apparatus. In this way,
in schematic form in Figure 3. It consists of three parts, approximately 15 mg dry weight of resin was transfered
a 96-hole (8 × 12) bracket, a series of 96 7-mm (outer to each tube. Given that the mixture was isopycnic and
diameter) glass tubes, and hydrophobic polyethylene frits, equal volumes were added to all tubes, we expected the
which have been sealed into one end of each of the tubes number of moles of benzodiazepine precursor delivered
by heating. Reagents are delivered into the tubes by to each tube to be the same. A small portion of resin
submerging them in a 96 2-mL well microtiter block. A was saved from each of the 120 filter cartridges in case
detailed account of the construction and use of this a particular benzodiazepine needed to be remade.
apparatus is documented in the Experimental Section. Two columns in every multitube plate were left empty,
For the sake of simplicity, this will be referred to as the providing an apparatus that was effectively 80 (10 × 8)
multitube apparatus. tubes per plate. Thus, after cleavage into a 96-well
microtiter plate, two columns would be left blank for
Several features of this apparatus make it attractive
biological controls during assays. Sixteen multitube
for use in the context of parallel synthesis. Construction
plates, plus one that was only half full, were required.
is easy. It is made from ordinary items found in an
These plates were transferred to small polypropylene bins
organic or bioorganic chemistry lab. In several days, 17
in a nitrogen bag (former pipet rack covers), and a 0.25
blocks (∼1500 discrete reaction tubes) can be constructed
M solution of lithium acetanilide in 1:1 THF/DMF was
for a total cost of under $500. Part or all of the apparatus
added to induce cyclization. After 48 h, each plate was
may be reused, depending on the conditions to which it
transferred to another bin that had been charged with a
is subjected. Commercially available Beckman 2 mL
DMF solution of 0.4 M alkylating agent. After 6 h, the
deep-well microtiter plates serve as reaction vessels.
contents of the tubes were rinsed repeatedly and the
Finally, the apparatus seems to be amenable to many
product benzodiazepines were cleaved into microtiter
conditions that would be encountered over the course of wells.
organic synthesis, e.g., acidic and basic conditions, as well
Library Evaluation. The integrity of the library was
as compatibility with a range of protic and aprotic
evaluated in two ways. First, exact yields were obtained
solvents. Elevated temperatures may be a problem since
by HPLC analysis for multiple benzodiazepines. For
the polyethylene frits begin to deform around 60 °C.
these benzodiazepines the relative extinction coefficient
Synthesis of a Library of 1,4-Benzodiazepine-2,5- of the benzodiazepine to the internal standard nitrotolu-
diones. Each amino acid and each anthranilic acid that ene was used to determine the quantity of material.
were incorporated into our library were rigorously dem- Second, 1H NMR spectra were obtained from compounds
onstrated to provide 1,4-benzodiazepine-2,5-diones in in 36 randomly chosen wells.
good to high yield (see, e.g., Figure 2). Every alkylating The first set of benzodiazepines evaluated were from
agent was similarly evaluated, with the exception of the wells corresponding to structures whose UV absorbance
2-(bromomethyl)naphthalene, which we considered equiva- relative to the internal standard p-nitrotoluene had been
lent in reactivity to the other benzylic halides. Many
components were successfully tested in more than one (83) Zuckermann, R. N.; Kerr, J. M.; Siani, M. A.; Banville, S. C.
combination. Ultimately, we selected 12 anthranilic Int. J. Pept. Protein. Res. 1992, 40, 497-506.
1246 J. Org. Chem., Vol. 62, No. 5, 1997 Boojamra et al.

Table 1. Yields of Selected Benzodiazepinesa

entry R1 R2 R3 yield (µmol)
1 7-Cl benzyl (CH2)2CO2H 4.8
2 9-pyridyl c-(C3H5)CH2 CH2C6H5 3.6
3 8-Me piperonyl 2-thienylmethyl 3.3
4 7-Cl piperonyl CH2C6H5 2.4
5 9-Br, 7-Me methyl CH2CH(CH3)2 2.8
6 6-F NH2COCH2 CH2C6H5 1.8
7 9-Br, 7-Me H CH2CH(CH3)2 not soluble
8 8-Me 4-phenylBn CH2C6H4OH 2.4b
9 8-O2N c-(C3H5)CH2 2-thienylmethyl 2.0
10 7-Cl ethyl (CH2)2CONH2 2.4
11 7-Cl NH2COCH2 CH2C6H4OH 2.3
12 ethyl CH2C6H5 2.4c
13 6-F 3,5-diMeBn 2-thienylmethyl 1.1
14 7-Cl methyl CH2-2-(C10H7) 2.6
15 8-O2N ethyl CH2CH(CH3)2 2.4
16 8-OMe 2-MeOBn H 4.0
17 8-OMe benzyl CH3 9.7
a UV absorbances of the listed derivatives were determined

relative to the internal standard (p-nitrotoluene) were predeter-

mined (see Table 8). b This yield is based on the corresponding R1
) 8-hexyl derivative, entry 13, Table 8 (Experimental Section).
c The benzodiazepine in entry 12 coeluted with p-nitrotoluene.

Thus, 9-fluoreneone was used as an internal standard.

Table 2. Evaluation of Benzodiazepines

UV ratio yield
entry R1 R2 R3 of Bz/Stda (µmol)
1 7-Br c-(C3H5)CH2 H 3.59 2.7
2 7-Br c-(C3H5)CH2 (CH2)4NH2 4.12 3.1
3 7-Br c-(C3H5)CH2 (CH2)2CO2H 3.35 2.5
4 7-Br c-(C3H5)CH2 CH2C6H4OH 2.87 2.2
5 7-Br c-(C3H5)CH2 CH2CH(CH3)2 3.03 2.3
6 7-Br c-(C3H5)CH2 CH2-(2-C10H7) 3.65 2.7
7 7-Br c-(C3H5)CH2 CH2C6H5 3.47 2.6
8 7-Br c-(C3H5)CH2 2-thienylmethyl 5.31 4.0
9 7-Br c-(C3H5)CH2 (CH2)2CONH2 1.13 0.8
10 7-Br c-(C3H5)CH2 CH3 12.5 9.4
a UV ratio of benzodiazepine to the standard 4-nitrotoluene was

based upon the relative chromophore of a related benzodiazepine

prepared from 5-bromoanthranilic acid (entry 14, Table 8).

tions are most likely due to variations in the equimolar

distribution techniques for partitioning resin as described
previously rather than synthesis fidelity. Excluding
alanine, the average quantity produced is 2.7 µmol with
a standard deviation (σn) of 0.9 µmol.
Figure 4. Components used in the synthesis of the 1,4- We then evaluated a series of benzodiazepines from a
benzodiazepine-2,5-dione library. randomly chosen plate. All benzodiazepines in this plate
determined for purified derivatives prepared on a large had been alkylated with (bromomethyl)cyclopropane. In
scale (see the Experimental Section). Yields were cal- order to directly compare benzodiazepines incorporating
culated by delivery of a stock solution of the nitrotoluene each of the different amino acids, we examined all the
internal standard to the contents of a particular well, benzodiazepines in this plate substituted at position 7
followed by HPLC analysis. Results are presented in with a bromide (see Table 2). The chromophore of these
Table 1. Entry 7 corresponds to a well that contained benzodiazepine-2,5-diones are dominated by the anthra-
insoluble product; the yield for this benzodiazepine is nilic acid portion of the molecules; we therefore assumed
omitted. However, an alkylated version of this benzodi- that the extinction coefficient of each derivative would
azepine was obtained in 2.8 µmol yield (Table 1, entry be approximately equal. The quantity of each benzodi-
5). azepine produced, with the exception of the benzodiaz-
Relatively small quantities of the benzodiazepines with epine derived from glutamine (Table 2, entry 9) and
a 6-fluoro substituent (Table 1, entries 6 and 13) were alanine (Table 2, entry 10), fall within a factor of 2. The
produced. This is not unexpected given that the yield of average quantity produced is 2.5 µmol with a standard
benzodiazepines made on a large scale with this anthra- deviation of 0.8 µmol.
nilic acid were also low. The benzodiazepine derived from For each of the compounds evaluated (with the excep-
alanine, Table 1, entry 17, was recovered in almost four tion of the glutamine-derived compounds), HPLC with
times the average quantity. Other alanine-containing detection at 252 nm indicated that the desired product
benzodiazepines appeared in equally high yield relative was in all cases the major UV active peak and in most
to the nonalanine benzodiazepines. The benzodiazepine cases (∼80%) the only observed UV active peak aside
represented by entry 1 (Table 1) was also produced in from the internal standard. The dimethoxybenzhydrol
greater than expected quantities. The observed varia- protecting group used to protect glutamine during ben-
Synthesis of 1,4-Benzodiazepine-2,5-diones J. Org. Chem., Vol. 62, No. 5, 1997 1247

zodiazepine synthesis to some extent complicated the Novabiochem, or Advanced Chemtech. The anthranilic acids
evaluation of these benzodiazepines by HPLC. were purchased from Aldrich, Lancaster, TCI America, or
1H NMR spectra of compounds from 36 randomly Maybridge, through their American distributor, Ryan Scien-
tific. The synthesis of two anthranilic acids that are not
chosen wells were obtained in order to evaluate the purity
commercially available is described directly below this para-
of the components of the library and to characterize them graph. Anhydrous N,N-dimethylformamide (DMF) and an-
as well.84 On the basis of the 1H NMR spectra, the correct hydrous 1-methyl-2-pyrrolidinone (NMP) were purchased from
benzodiazepine could be unequivocally assigned for 35 Aldrich. Tetrahydrofuran was distilled under N2 from sodium/
of the 36 wells. For 31 of the 35 wells the desired benzophenone immediately prior to use unless sieve-dried THF
benzodiazepine product was not contaminated with any was used (indicated when used). It was not necessary to use
related material. For two of the wells the desired distilled solvents for rinsing of resin. Flash column chroma-
benzodiazepine was contaminated with ∼25-35% of a tography was carried out using Merck 60 230-400-mesh silica
gel. Thin layer chromatography (TLC) analyses were per-
related compound, and for the final two wells the desired formed with Uniplate 250 µm silica gel plates from Analtech,
benzodiazepine was contaminated with an equal amount Newark, DE (catalog no. 21521). 1H NMR spectra were
of a related impurity. obtained with a UCB Bruker AM-400 or AM-500 FT spec-
While in almost every case the only product related to trometer. Proton-decoupled 13C spectra were obtained at 101
our benzodiazepine synthesis was the desired compound, or 126 MHz with the same instruments with a line broadening
we found varying amounts of the plasticizer dioctylphtha- of 1.5 Hz. Chemical shifts are reported in ppm. Coupling
late in some wells and varying amounts of another constants are reported in Hz. Unless otherwise noted, spectra
were obtained in CDCl3 with residual CHCl3 as an internal
unrelated aliphatic impurity in some wells. Efforts to standard at 7.25 ppm; spectra obtained in DMSO-d6 were
isolate and characterize this second impurity by extrac- referenced to the residual DMSO-d5, 2.49 ppm; spectra ob-
tive and chromatographic methods were unsuccessful. tained in acetonitrile-d3 were referenced to residual acetoni-
trile-d2 at 1.94 ppm. NMR samples from the library were
Conclusion obtained by dissolving the entire contents of the well in 0.75
mL of acetonitrile-d3 or methanol-d4. Elemental analyses were
A general and expedient method for the solid-phase performed by M-H-W Labs, Phoenix, AZ, or by the University
synthesis of 1,4-benzodiazepine-2,5-diones has been de- of California, Department of Chemistry Microanalytical Labo-
veloped. Three commerically available components, an- ratory. Chloromethylpolystyrene (1% cross-linked) resin was
thranilic acids, R-amino esters, and alkylating agents, are obtained from Novabiochem, catalog no. 01-64-0007, 200-400
employed to introduce functionality. Reaction conditions mesh size. (Aminomethyl)polystyrene was obtained from
Novabiochem or Bachem. A filtration cannula (Pharmacia,
were developed to prepare either racemic compounds for
Uppsala, Sweden) is useful for filtration of resins in round-
lead identification efforts or optically pure compounds for bottom flasks.
lead optimization efforts. The incorporation of diverse 4-Methoxyanthranilic Acid. This compound was pre-
functionality into the benzodiazepine products was dem- pared according to a modified literature procedure from
onstrated, including the incorporation of amines, amides, 4-methyl-3-nitroanisole (available from Aldrich) through the
carboxylic acids, phenols, ethers, thiophenes, pyridines, intermediate 4-methoxy-2-nitrobenzoic acid.85 In 150 mL of
halogens, sulfones, and nitro groups. Limitations to the HOAc was suspended 9.74 g of 4-methoxy-2-nitrobenzoic acid
(49.4 mmol) in a round-bottom flask. To this was added 75
chemistry were also identified.
mL of EtOAc in order to help dissolve the starting material.
On the basis of the scope and generality of the The solution was subsequently degassed by purging the
synthesis sequence, a library of 1,4-benzodiazepine-2,5- contents of the flask with N2 for 30 min. A catalytic amount
diones was prepared from 11 alkylating agents, 12 of 10% Pd/C was added, and with vigorous stirring, the flask
anthranilic acids, and 19 R-amino esters (nine sets of was fitted with an H2 balloon and purged rapidly with this
enantiomeric pairs and glycine methyl ester hydrochlo- gas. The contents of the flask were stirred at rt for 12 h under
ride) to give 2508 total compounds, or 1320 spatially 1 atm of H2. The product and starting material had identical
separate compounds. In order to prepare the library, a Rf values when TLC plates were eluted with 1:1 hexanes/
EtOAc. Completion was determined by the drastically differ-
microtiter-based apparatus was developed that is inex- ent appearance of the two compounds when TLC plates were
pensive and straightforward to prepare from ordinary held under mixed short-wave UV light. The aniline was
items found in an organic or bioorganic laboratory. The extremely fluorescent, glowing blue on the TLC plate, while
high quality of the benzodiazepine-2,5-dione library the starting material was not. The heterogeneous mixture was
prepared with this apparatus was demonstrated by filtered through a plug of Celite, and the clear fluorescent
evaluating representative compounds by HPLC analysis liquid was concentrated in vacuo. As the volume reduced, 5.85
and 1H NMR. g (35.0 mmol, 71%) of off-white needles precipitated, mp 165-
167 °C (color change/decomposition begins, lit.18 mp 166 °C).
The reported solid-phase synthesis sequence, demon- A second crop of product could be isolated by further reducing
stration of scope and generality of the chemistry, and the volume of the supernatant. These combined crops of
methods for library synthesis should significantly expe- product were used directly in the benzodiazepine synthesis
dite lead identification or optimization efforts based upon described below.
the 1,4-benzodiazepine-2,5-dione structure. In addition, 4-Bromoanthranilic Acid. This compound was prepared
the described multitube apparatus should have broad from 4-amino-2-nitrobenzoic acid (available from Research
applicability to parallel synthesis efforts towards a range Plus) through the intermediate 4-bromo-2-nitrobenzoic acid.
To a solution of 4-amino-2-nitrobenzoic acid (1.00 g, 5.49 mmol)
of different applications.
in 48% aqueous HBr, which was cooled in an ice bath to 0 °C,
was added 16.4 mL of H2O. A solution of NaNO2 (0.38 g, 5.5
Experimental Section mmol) in 13.7 mL of H2O was prepared. The solution of
General Methods. Unless otherwise noted, materials were NaNO2 was added dropwise to the solution of 4-amino-2-
obtained from commercial suppliers and used without further nitrobenzoic acid (still at 0 °C). The reaction mixture was
purification. The R-amino esters were purchased from Bachem, initially cloudy but after 5 min turned clear orange. In a
separate 500-mL, round-bottom flask was dissolved CuBr (1.04
g, 7.25 mmol) in 18 mL of 48% aqueous HBr. The flask was
(84) Glutamine-containing benzodiazepines were not evaluated since
they were contaminated with the dimethoxybenzhyrol cleavage side
product. (85) Ullman, F.; Dootson, P. Ber. Dtsch. Chem. Ges. 1918, 51, 9-24.
1248 J. Org. Chem., Vol. 62, No. 5, 1997 Boojamra et al.

fitted with a magnetic stir bar and was cooled to 0 °C. The mmol) was dissolved, also with gentle stirring. Alternatively,
diazonium salt of 4-amino-2-nitrobenzoic acid was transferred the reductant and the R-amino ester may be premixed in the
to an addition funnel and was added dropwise with vigorous same volume of 1% HOAc in DMF and the same mass of dry
stirring to the solution of CuBr, generating a deep purple resin 4. An aliquot of this resin was removed after 20 min
solution. Caution must be used when working up diazonium- and rinsed with DMF (7 × 20 mL), CH2Cl2 (7 × 20 mL), and
forming reactions. HNO2 is not stable in aqueous solution and finally, CH3OH (3 × 10 mL). The aldehyde stretch was no
may decompose, producing NO2. NO2 is an insidious poison longer observed. This procedure was repeated at 60 min to
and exposure to it may be deadly. Appropriate precautions ensure that there were no additional changes in the IR
should be taken. The contents of the reaction flask were spectrum. The bulk of the resin was then rinsed with CH3-
concentrated in vacuo (not quite to dryness) and slurried in OH (1 × 10 mL), DMF (7 × 20 mL), CH2Cl2 (7 × 20 mL), and
200 mL of EtOAc and 1 N aqueous HCl. The phases were finally, CH3OH (3 × 10 mL). The resin was dried in vacuo to
separated, and the aqueous layer was extracted again with a constant weight IR (KBr) 1736 cm-1 (s).
EtOAc (2 × 100 mL). The combined organic layers were dried Racemizing Procedure for the Loading of r-Amino
and concentrated in vacuo to a green solid (the color presum- Esters 5. A general procedure is described for loading of any
ably due to residual copper salts). The green solid was R-amino ester hydrochloride to support-bound aldehyde 4
dissolved in 50 mL of EtOAc and decolorized on a Florisil plug when racemic product is desired from scalemic R-amino ester.
(eluting with excess EtOAc). Concentration in vacuo yielded In a three-neck, 250-mL, round-bottom flask was placed a
4-bromo-2-nitrobenzoic acid as a yellow solid (0.80 g, 60% small magnetic stir bar and 1% HOAc in DMF (50 mL). In
yield). This diazotization was scaled up, and the product this solvent mixture was suspended 6.00 g of aldehyde-
bromide was carried on without further purification to gener- derivatized resin 4. To this was added the R-amino ester
ate 4-bromoanthranilic acid according to literature procedure.86 hydrochloride of choice (22.2 mmol) and distilled i-Pr2EtN
In a large, round-bottom flask charged with 130 mL of (0.859 g, 6.66 mmol). The system was stirred gently for 6 h,
concentrated NH4OH was dissolved 4-bromo-2-nitrobenzoic after which time NaBH(OAc)3 (4.70 g, 22.2 mmol) was added.
acid (6.52 g, 26.5 mmol). In a separate conical flask was It may be necessary to pulverize the chunks that form in this
dissolved 65 g (160 mmol) of ammonium iron(II) sulfate in 130 reagent before addition. The suspension was stirred gently
mL of H2O. This light green solution was added to the solution for 36 h, at which point CH3OH was added to the resin to
of 4-bromo-2-nitrobenzoic acid with magnetic stirring. On quench the excess reductant and dissolve the borate salts that
contact, the light green solution became dark green and then were present. The solution was then removed from the resin
rust colored. The mixture was heated at reflux for 2 min and via filtration cannula and the resin rinsed with DMF (7 × 50
cooled. The rust-colored suspension was filtered though a pad mL), CH2Cl2 (7 × 50 mL), and finally, CH3OH (3 × 20 mL).
of Celite, which was then rinsed with water and aqueous NH4- The resin was dried to a constant weight in vacuo, and IR
OH. The dark brown filtrate was acidified with concentrated (KBr) indicated that the aldehyde carbonyl stretch had been
aqueous HCl, and the resulting pink turbid suspension was replaced by an ester carboxyl stretch (∼1740 cm-1).
extracted with EtOAc (3 × 200 mL). The combined organic Determination of Resin Substitution Level. The load-
layers were washed with H2O, dried, and concentrated in ing of resin-bound R-amino esters is quantitated by acetylation
vacuo. The resulting residue was decolorized on a silica plug, of the free amine of 5 followed by cleavage with subsequent
eluting with 1% HOAc in 2:1 hexanes/EtOAc. Concentration determination of the mass balance of the silica gel purified
in vacuo of the eluant provided 3.3 g (62% yield) of a slightly product. A typical experiment follows: A 50-mL flame-dried
brown solid, mp 222-223 °C (lit.86 mp 222 °C). Anal. Calcd round-bottom flask was charged with a magnetic stir bar and
for C7H6NO2Br: C, 38.92; H, 2.80; N, 6.48. Found: C, 39.17; 20 mL of a mixture of pyr/Ac2O (2:1) and a catalytic amount
H, 2.80; N, 6.59. of 4-(dimethylamino)pyridine (DMAP). To this flask, with
Aldehyde-Derived Polystyrene 4. To a flame-dried, 2-L, gentle stirring, was added ∼0.500 g of R-amino ester resin.
three-neck, round-bottom flask fitted with a mechanical stirrer The flask was fitted with a rubber septum, equipped with an
and an Ar inlet were added 15.3 g (84.2 mmol) of 4-hydroxy- Ar inlet, and gently heated to no more than 50 °C. After 12
2,6-dimethoxybenzaldehyde87 and DMF (800 mL). The solu- h, the solution was removed from the resin by filtration
tion was degassed by bubbling Ar through it for 30 min. Under cannula and the resin was washed with DMF (7 × 20 mL),
a gentle positive Ar flow, 1.92 g (79.9 mmol) of 95% NaH was CH2Cl2 (7 × 20 mL), and finally, CH3OH (3 × 10 mL), which
added slowly in ∼0.5 g aliquots. The addition of NaH was desolvates the resin. The resin was then dried in vacuo to a
complete in 15 min. An off-white/yellow solid slowly precipi- constant weight, and a known mass was presolvated in 10 mL
tated as the solution turned dark red. As H2 evolved, the of CH2Cl2 and stirred with a magnetic stir bar for 5 min. A
solution was continually purged with a stream of Ar. After filtration cannula was then used to remove the CH2Cl2, and
30 min, H2 evolution ceased, and Merrifield resin (chloro- care was taken not to accidentally remove any resin from the
methylpolystyrene, 1% divinylbenzene, 31.7 g, loading level flask as the cannula was removed. The resin was then treated
0.76 mmol g-1, 24.1 mmol) added, and the Ar purge continued with 20 mL of a solution of TFA/Me2S/H2O (90:5:5) and stirred
for 15 min. The contents were stirred mechanically for 36 h for a minimum of 12 h. After 12 h, the contents of the flask
while being heated at 50 °C. The DMF suspension was diluted were filtered into a round-bottom flask, and the resin and filter
with CH3OH (200 mL) and filtered with a filtration cannula. paper were repeatedly washed with CH2Cl2 and then CH3OH.
The resin was then rinsed with DMF/CH3OH 1:1 (2 × 500 mL), The filtrate was concentrated in vacuo, and the product
DMF (5 × 500 mL), CH2Cl2 (7 × 500 mL), and CH3OH (4 × N-acetyl R-amino ester was purified by flash column chroma-
400 mL). The salmon-colored resin was dried in vacuo to a tography and the mass balance obtained. The above acetyla-
constant weight of 33.4 g. A strong aldehyde carbonyl stretch tion procedure was performed on phenylalanine methyl ester
was observed in the IR (KBr) spectrum of this resin: 1690 resin (0.500 g), and 0.420 g of the acetylated resin was cleaved
cm-1. as described above. Silica gel purification eluting with 2:1-
Racemization-Free Loading of an r-Amino Ester To 1:1 hexanes/EtOAc yielded 34.5 mg (0.156 mmol) of clear oil
Obtain Solid Support 5. This procedure may be used for (that later solidified). The loading is therefore determined to
racemization-free loading of an R-amino ester hydrochloride be 0.156 mmol/0.420 g resin or 0.37 mmol g-1 resin. An
to support-bound aldehyde 4. A 50-mL round-bottom flask was identical experiment performed on leucine-derived resin from
charged with 1% HOAc in DMF (20 mL) and a magnetic stir the same lot of 4 showed the resin to be substituted at 0.36
bar. To the flask was added 0.500 g of aldehyde-derived resin mmol g-1 resin. Such loadings are typical when starting with
4 and NaBH(OAc)3 (0.636 g, 3.00 mmol), generating a white, commercially available Merrifield resin that is functionalized
turbid suspension. It may be necessary to pulverize the at a loading of ∼0.7 mmol g-1. Such a loading experiment was
chunks in commercially available NaBH(OAc)3 before addition. performed every time a new lot of aldehyde-derived resin was
The suspension was stirred gently, and an R-amino ester (3.00 synthesized, and all yields are based on loading levels obtained
in this manner.
(86) Leggate, P.; Dunn, G. E. Can. J. Chem. 1965, 43, 1158-1174. General Procedure for Synthesis of 1,4-Benzodiaz-
(87) Landi, J. J.; Ramig, K. Synth. Commun. 1991, 21, 167-171. epine-2,5-diones 1. A typical procedure for the generation
Synthesis of 1,4-Benzodiazepine-2,5-diones J. Org. Chem., Vol. 62, No. 5, 1997 1249

of 1,4-benzodiazepine-2,5-diones follows. In an oven-dried, 50- ride, and acetic acid. The scalemic R-amino ester was loaded
mL, round-bottom flask with a dry magnetic stir bar was according to the nonracemizing conditions. Cleavage of 0.170
placed 0.500 g (0.185 mmol) of R-amino ester resin 5. The flask g (0.0595 mmol) of this resin by stirring in 20 mL of 90:5:5
was fitted with a rubber septum, and 5 mL of N-methylpyr- TFA/Me2S/H2O for 36 h yielded a crude yellow residue after
rolidinone (NMP) was added. Once the resin was solvated (3 gravity filtration, rinsing of the resin and filter paper with
min), 1-ethyl-3-[3-(dimethylamino)propyl]carbodiimide‚HCl CH2Cl2 (4 × 15 mL) and CH3OH (2 × 10 mL), and concentra-
(EDC, 0.425 g, 2.22 mmol) was added. It may be necessary to tion. Silica gel chromatography eluting with 3:1 hexanes/
pulverize the chunks that form in this reagent before addition. EtOAc provided 15.9 mg (0.053 mmol, 89% yield) of an
After the solution became saturated in EDC (most of the EDC amorphous glass. IR (KBr): 3425, 3232, 3067, 2932, 1685,
will dissolve over the course of 5 min), the anthranilic acid of 1652, 1479, 1432, 1353, 1221, 1128, 828, 700 cm-1. 1H NMR
choice was slowly added (1.85 mmol). It is imperative that the (400 MHz, DMSO-d6): δ 2.85 (dd, 1, J ) 9.2, 14.3), 3.11 (dd,
EDC and resin be allowed to mix thoroughly before addition 1, J ) 5.1, 14.3), 3.92-3.98 (m, 1), 7.11 (d, 1, J ) 8.9), 7.18 (t,
of anthranilic acid and that the anthranilic acid be added 1, J ) 7.2), 7.24 (t, 2, J ) 7.2), 7.30 (d, 2, J ) 7.4), 7.57 (dd, 1,
slowly. This minimizes side reactions that can occur with the J ) 2.3, 8.7), 7.61 (d, 1, J ) 2.5), 8.65 (d, 1, J ) 6.1), 10.51 (s,
unprotected aniline functionality. The reaction mixture was 1). 13C NMR (101 MHz, CDCl3, one drop CD3OD): δ 34.0, 53.7,
stirred gently for 8-12 h. For some anthranilic acids (4-Cl, 122.5, 126.7, 127.0, 128.6, 129.1, 130.5, 133.0, 134.5, 136.1,
4-NO2), a precipitate formed as the reaction proceeded. The 167.7, 171.1. HRMS (FAB+) m/e: 301.0741 (MH+ C16H14N2O2-
precipitate was so fine that it was easily removed during the Cl requires 301.0744). Anal. Calcd for C16H13O2N2Cl: C,
rinses with the filtration cannula. The resin was rinsed with 63.91; H, 4.36; N, 9.32. Found: C, 63.71; H, 4.45; N, 9.15.
DMF (7 × 20 mL), CH2Cl2 (7 × 20 mL), and finally CH3OH (3 Benzodiazepine 1c. The support-bound benzodiazepine
× 10 mL). The acylation and rinsing were repeated (2 h is was prepared according to the above procedure from 4-chlo-
adequate) to ensure reaction completion, and the resin was roanthranilic acid, (S)-leucine methyl ester hydrochloride, and
dried to a constant weight in vacuo. Nitro-containing anthra- acetic acid. The scalemic R-amino ester was loaded according
nilic acids may require a third subjection. Cyclization and to the nonracemizing conditions. Cleavage of 0.350 g (0.123
alkylation were then accomplished in one pot: The acylated mmol) of this resin by stirring in 20 mL of 90:5:5 TFA/Me2S/
resin 6 was placed in a 50-mL, round-bottom flask fitted with H2O for 36 h yielded a crude yellow residue after gravity
a small, magnetic stir bar and purged gently with Ar for 5 filtration, rinsing of the resin and filter paper with CH2Cl2 (4
min. In a separate flame-dried, 50-mL, round-bottom flask × 15 mL) and CH3OH (2 × 10 mL), and concentration. Silica
was placed a magnetic stir bar and acetanilide (4.44 mmol), gel chromatography eluting with 2:1 hexanes/EtOAc provided
followed by 7.5 mL of THF. This flask was purged with Ar 29 mg (0.109 mmol, 89% yield) of a white powder, mp 244-
for 5 min. The flask was then cooled in a dry ice/acetone bath 248 °C (decomposition begins). IR (KBr): 3427, 3164, 2967,
to -78 °C. A fraction of the compound may precipitate on 1664, 1604, 1473, 1427, 1216 (w), 1098 (w), 828 cm-1. 1H NMR
cooling (this does not affect the reaction). A hexanes solution (400 MHz, DMSO-d6): δ 0.76 (d, 3, J ) 6.5), 0.85 (d, 3, J )
of n-BuLi (1.48 mL, 2.5 M, 3.7 mmol) was added dropwise over 6.5), 1.54 (t, 2, J ) 7.0), 1.65-1.70 (m, 1), 3.62-3.67 (m, 1),
10 min with rapid stirring. The suspension became slightly 7.13 (s, 1), 7.26 (m, 1), 7.74 (d, 1, 8.5), 8.49 (d, 1, J ) 5.6),
yellow and sometimes turned into a gelatin over the course of 10.45 (s, 1). 13C NMR (101 MHz, DMSO-d6): δ 21.5, 22.8, 23.8,
30 min. DMF (7.5 mL) was then added to homogenize the 36.1, 50.1, 120.2, 123.8, 125.0, 132.3, 136.3, 138.1, 166.8, 171.5.
solution. After all solids dissolved, the mixture was stirred Anal. Calcd for C13H15O2N2Cl: C, 58.55; H, 5.67; N, 10.50.
at -78 °C for another 15 min, warmed to rt, and then Found: C, 58.67; H, 5.81; N, 10.46.
transferred via metal cannula into the flask containing the Benzodiazepine 1d. The benzodiazepine was prepared
resin (which was under Ar at rt). The suspension was stirred according to the above procedure from 5-bromoanthranilic acid,
gently at rt for 30 h under an Ar atmosphere. Alkylating agent (S)-leucine methyl ester hydrochloride, and EtI. The scalemic
(7.40 mmol) was added via syringe, and stirring was continued R-amino ester was loaded according to the racemizing condi-
until the suspension no longer turned pH paper dark green tions. Cleavage of 0.500 g (0.170 mmol) of this resin by stirring
or blue. This typically took 3-6 h. The alkylating solution in 20 mL of 90:5:5 TFA/Me2S/H2O for 36 h yielded a crude
was removed via filtration cannula and the resin was rinsed yellow residue after gravity filtration, rinsing of the resin and
with DMF (7 × 20 mL), CH2Cl2 (7 × 20 mL), and finally CH3- filter paper with CH2Cl2 (4 × 15 mL) and CH3OH (2 × 10 mL),
OH (3 × 10 mL) and dried in vacuo to a constant weight. The and concentration. Silica gel chromatography, eluting with
benzodiazepines were then cleaved from support. A known 4:1-3:1 hexanes/EtOAc yielded, 41.0 mg (0.121 mmol) of an
mass of resin was swelled with CH2Cl2. The CH2Cl2 was amorphous white solid (71% yield). IR (KBr): 3441 (b), 2967,
removed with a filtration cannula, taking care not to lose any 1670, 1598 (w), 1440, 1396 (w), 1256 (w), 1124 (w) cm-1. 1H
resin on the cannula as it was removed from the flask. The NMR (400 MHz, DMSO-d6): δ 0.73 (d, 3, J ) 6.3), 0.82 (d, 3,
benzodiazepines made by this method are cleaved and char- J ) 6.3), 0.99 (t, 3, J ) 6.6), 1.57-1.74 (m, 3), 3.59-3.71 (m,
acterized below. 2), 4.03-4.10 (m, 1), 7.44 (d, 1, J ) 9.2), 7.74-7.80 (m, 2), 8.67
Benzodiazepine 1a. The support-bound benzodiazepine (d, 1, J ) 5.8). 13C NMR (101 MHz, DMSO-d6): δ 13.1, 21.7,
was prepared according to the above procedure from anthra- 22.8, 23.8, 36.3, 42.2, 50.1, 117.7, 125.0, 131.6, 131.7, 134.8,
nilic acid and (R)-alanine methyl ester hydrochloride, and 138.6, 166.2, 169.3. Anal. Calcd for C15H19O2N2Br: C, 53.11;
acetic acid was substituted for an alkylating agent. The H, 5.65; N, 8.26. Found: C, 52.89; H, 5.61; N, 8.12.
scalemic R-amino ester was loaded according to the nonrace- Benzodiazepine 1e. The support-bound benzodiazepine
mizing conditions. Cleavage of 0.314 g (0.107 mmol) of this was prepared according to the above procedure from 4-meth-
resin in 20 mL of 90:5:5 TFA/Me2S/H2O for 36 h yielded a crude oxyanthranilic acid (vide supra), glycine methyl ester hydro-
yellow residue after gravity filtration, rinsing of the resin and chloride, and 2-methoxybenzyl chloride. The R-amino ester
filter paper with CH2Cl2 (4 × 15 mL) and CH3OH (2 × 10 mL), was loaded according to the nonracemizing conditions. Cleav-
and concentration. Silica gel chromatography eluting with 2:1 age of 0.392 g (0.128 mmol) of this resin by stirring in 20 mL
EtOAc/hexanes provided 15.7 mg (0.083 mmol, 77% yield) of of 90:5:5 TFA/Me2S/H2O for 36 h yielded a crude yellow residue
a fine, white, amorphous powder. IR (KBr): 3272, 3173, 3060, after gravity filtration, rinsing of the resin and filter paper
2919, 1705, 1675, 1478, 1408, 755 cm-1. 1H NMR (400 MHz, with CH2Cl2 (4 × 15 mL) and CH3OH (2 × 10 mL), and
DMSO-d6): δ 1.21 (d, 3, J ) 6.8), 3.76-3.82 (m, 1), 7.08 (d, 1, concentration. Silica gel chromatography eluting with 5:1
8.0), 7.20 (t, 1, J ) 7.1), 7.47-7.51 (m, 1), 7.71 (dd, 1, J ) 1.5, EtOAc/hexanes followed by neat ethyl acetate after product
7.9), 8.39 (d, 1, J ) 5.1), 10.3 (s, 1). 13C NMR (101 MHz, began to elute provided 16.9 mg (0.052 mmol, 40% yield) of a
DMSO-d6): δ 13.8, 47.3, 120.9, 123.8, 126.2, 130.4, 132.2, white fluffy powder, mp 99-100 °C. IR (KBr): 3435, 2935,
136.7, 167.7, 172.2. HRMS (FAB+) m/e: 191.0822 (MH+ 1642, 1608, 1660, 1248, 1118, 1096 cm-1. 1H NMR (400 MHz,
C10H11N2O2 requires 191.0821). DMSO-d6): δ 3.52 (b s, 1), 3.67 (s, 3), 3.73 (s, 3), 3.80 (b s, 1),
Benzodiazepine 1b. The support-bound benzodiazepine 4.85 (d, 1, J ) 15.6), 5.20 (d, 1, J ) 16.6), 6.77-6.83 (m, 2),
was prepared according to the above procedure from 5-chlo- 6.87-6.89 (m, 2), 7.02 (d, 1, J ) 7.4), 7.16 (d, 1, J ) 7.6), 7.56
roanthranilic acid, (S)-phenylalanine methyl ester hydrochlo- (d, 1, J ) 8.5), 8.56 (t, 1, J ) 5.1). 13C NMR (101 MHz, DMSO-
1250 J. Org. Chem., Vol. 62, No. 5, 1997 Boojamra et al.

d6): δ 45.0, 45.7,55.1, 55.5, 107.6, 110.6, 111.4, 119.9, 121.7, residue after gravity filtration, rinsing of the resin and filter
124.3, 128.1, 128.4, 131.4, 141.3, 156.7, 161.5, 167.7, 169.1. paper with CH2Cl2 (4 × 15 mL) and CH3OH (2 × 10 mL), and
Anal. Calcd for C18H18O4N2: C, 66.25; H, 5.56; N, 8.58. concentration. Silica gel chromatography eluting with 3:1
Found: C, 66.20; H, 5.61; N, 8.32. hexanes/EtOAc provided 43.6 mg (0.121 mmol, 81% yield) of
Benzodiazepine 1f. The support-bound benzodiazepine an amorphous yellow powder. A small aliquot of this product
was prepared according to the above procedure from 5-chlo- was taken up in a drop of ethyl acetate and precipitated by
roanthranilic acid, H-(S)-Glu(t-Bu)-OMe‚HCl, and benzyl bro- addition of hexanes. The supernatant was decanted from the
mide. The scalemic R-amino ester was loaded according to the product yielding a fine yellow crystals, mp 173 °C (decomposi-
nonracemizing conditions. Cleavage of 0.528 g (0.169 mmol) tion begins). IR (KBr): 3472, 3260, 3096, 2934, 1669, 1527,
of this resin by stirring in 20 mL of 90:5:5 TFA/Me2S/H2O for 1442, 1347, 1232 (w), 1155 (w), 697 cm-1. 1H NMR (400 MHz,
36 h yielded a crude yellow residue after gravity filtration, DMSO-d6): δ 0.08-0.15 (m, 2), 0.31-0.35 (m, 2), 0.80-0.88
rinsing of the resin and filter paper with CH2Cl2 (4 × 15 mL) (m, 1), 3.19 (dd, 1, J ) 8.1, 15), 3.29-3.34 (m, 1), 3.73 (dd, 1,
and CH3OH (2 × 10 mL), and concentration. Silica gel J ) 6.5, 14.5), 3.97 (dd, 1, J ) 6.2, 14.2), 4.06 (dd, 1, J ) 7.5,
chromatography eluting with 60:40:1 hexanes/EtOAc/HOAc 14.6), 6.87-6.90 (m, 1), 6.92-6.94 (m, 1), 7.28 (dd, 1, J ) 1.2,
provided 32.5 mg (0.087 mmol, 52% yield) of an amorphous 5.1), 7.90 (d, 1, J ) 8.6), 8.10 (dd, 1, J ) 2.1, 8.6), 8.33 (d, 1,
white powder. IR (KBr): 3432, 3272, 2919, 2849, 1670, 1441, J ) 2.0), 9.08 (d, 1, J ) 6.4). 13C NMR (101 MHz, DMSO-d6):
1207, 730 cm-1. 1H NMR (400 MHz, DMSO-d6): δ 1.85-1.92 δ 3.1, 3.9, 9.8, 28.1, 50.9, 53.7, 118.9, 120.1, 124.6, 126.7, 131.1,
(m, 1), 2.00-2.07 (m, 1), 2.28-2.34 (m, 2), 3.85-3.90 (m, 1), 135.2, 139.3, 140.4, 149.5, 166.1, 169.3. HRMS (FAB+) m/e:
4.93 (d, 1, J ) 16.0), 5.33 (d, 1, J ) 16.0), 7.06 (d, 2, J ) 7.1), 372.1027 (MH+ C18H18N3O4S requires 372.1018).
7.18-7.25 (m, 3), 7.49 (d, 1, J ) 8.4), 7.56-7.59 (m, 2), 8.80 Benzodiazepine 1j. The support-bound benzodiazepine
(d, 1, J ) 5.8). 13C NMR (101 MHz, DMSO-d6): δ 23.5, 29.8, was prepared according to the above procedure from 6-fluo-
49.5, 51.0, 124.7, 126.7, 127.1, 128.5, 128.9, 129.7, 131.4, 131.8, roanthranilic acid, (S)-phenylalanine methyl ester hydrochlo-
137.0, 138.0, 166.3, 170.0, 174.0. Anal. Calcd for C19H17O4N2- ride, and iodoacetamide. The scalemic R-amino ester was
Cl: C, 61.22; H, 4.60; N, 7.52. Found: C, 61.22; H, 4.76; N, loaded according to the nonracemizing conditions. Cleavage
7.35. of 0.340 g (0.120 mmol) of this resin by stirring in 20 mL of
Benzodiazepine 1g. The support-bound benzodiazepine 90:5:5 TFA/Me2S/H2O for 36 h yielded a crude yellow residue
was prepared according to the above procedure from 5-chlo- after gravity filtration, rinsing of the resin and filter paper
roanthranilic acid, H-(S)-Gln(Dod)-OMe‚HCl,70 and ethyl io- with CH2Cl2 (4 × 15 mL) and CH3OH (2 × 10 mL), and
dide. The scalemic R-amino ester was loaded according to the concentration. Silica gel chromatography, eluting with 3:100
nonracemizing conditions. Cleavage of 0.464 g (0.142 mmol) i-PrOH/EtOAc, yielded 25.0 mg (0.073 mmol) of an amorphous
of this resin by stirring in 20 mL of 90:5:5 TFA/Me2S/H2O for yellow solid (62% yield). IR (KBr): 3276, 3157, 1678, 1618,
36 h yielded a crude yellow residue after gravity filtration, 1473, 1394, 1242, 1032, 1006, 755, 703, 624 cm-1. 1H NMR
rinsing of the resin and filter paper with CH2Cl2 (4 × 15 mL) (400 MHz, DMSO-d6): δ 2.91 (dd, 1, J ) 9.4, 14.3), 3.07 (dd,
and CH3OH (2 × 10 mL), and concentration. Silica gel 1, J ) 4.9, 14.3), 4.02-4.07 (m, 1), 4.24 (d, 1, J ) 16.7), 4.43
chromatography eluting with 2:1 hexanes/EtOAc provided 31.0 (d, 1, J ) 16.8), 7.11-7.18 (m, 4), 7.24 (t, 2, J ) 7.1), 7.29 (d,
mg (0.100 mmol, 70% yield) of an amber powder. It was 2, J ) 7.0), 7.52-7.57 (m, 2), 8.82 (d, 1, J ) 7.3). 13C NMR
necessary to weight this powder out immediately; it was (101 MHz, DMSO-d6): δ 33.3, 51.0, 53.4, 113.1 (d, JFC ) 21.4),
extraordinarily hygroscopic and visibly liquefied with atmo- 117.8 (d, JFC ) 15.3), 118.1, 126.3, 128.1, 129.2, 132.3 (d, JFC
spheric water within minutes after isolation. IR (KBr): 3427, ) 10.4), 137.7, 141.6, 159.5 (d, JFC ) 258), 163.3, 169.1, 170.1.
3208, 2933, 1669, 1602, 1444, 1207, 1133, 1028, 826, 801, 721 HRMS (FAB+) m/e: 342.1252 (MH+ C18H17N3O3F requires
cm-1. 1H NMR (400 MHz, DMSO-d6): δ 1.00 (t, 3, J ) 7.0), 342.1254).
1.79-1.84 (m, 1), 1.93-1.97 (m, 1), 2.13 (t, 2, J ) 7.2), 3.64- Benzodiazepine 1k. The support-bound benzodiazepine
3.72 (m, 2), 4.06-4.12 (m, 1), 6.72 (b s, 1), 7.25 (b s, 1), 7.52 was prepared according to the above procedure from 5-fluo-
(d, 1, J ) 9.3), 7.63-7.66 (m, 2), 8.72 (d, 1, J ) 5.8). 13C NMR roanthranilic acid, (R)-alanine methyl ester hydrochloride, and
(101 MHz, DMSO-d6): δ 13.1, 23.6, 30.6, 42.1, 51.4, 124.8, 3,5-dimethylbenzyl bromide. The scalemic R-amino ester was
128.7, 129.7, 131.5, 131.9, 138.0, 166.2, 169.3, 173.6. HRMS loaded according to the nonracemizing conditions. Cleavage
(FAB+) m/e: 310.0961 (MH+ C14H17N3O3Cl requires 310.0958). of 0.433 g (0.143 mmol) of this resin by stirring in 20 mL of
Benzodiazepine 1h. The support-bound benzodiazepine 90:5:5 TFA/Me2S/H2O for 36 h yielded a crude yellow residue
was prepared from 4-nitroanthranilic acid, (S)-leucine methyl after gravity filtration, rinsing of the resin and filter paper
ester hydrochloride, and EtI according to the above procedure. with CH2Cl2 (4 × 15 mL) and CH3OH (2 × 10 mL), and
The scalemic R-amino ester was loaded according to the concentration. Silica gel chromatography eluting with 2:1
racemizing conditions. This benzodiazepine required a third hexanes/EtOAc provided 29.1 mg (0.892 mmol, 62% yield) of
subjection in the EDC-mediated acylation step. Cleavage of an amorphous off-white powder. IR (KBr): 3434, 2925, 1665,
0.450 g (0.153 mmol) of this resin by stirring in 20 mL of 90: 1492, 1451, 1197 cm-1. 1H NMR (400 MHz, DMSO-d6): δ 1.26
5:5 TFA/Me2S/H2O for 36 h yielded a crude yellow residue after (d, 3, J ) 6.7), 2.14 (s, 6), 3.93-3.98 (m, 1), 4.80 (d, 1, J )
gravity filtration, rinsing of the resin and filter paper with 16.0), 5.30 (d, 1, J ) 16.0), 6.65 (s, 2), 6.78 (s, 1), 7.35-7.41
CH2Cl2 (4 × 15 mL) and CH3OH (2 × 10 mL), and concentra- (m, 2), 7.52 (dd, 1, J ) 4.6, 8.9), 8.78 (d, 1, J ) 5.7). 13C NMR
tion. Silica gel chromatography, eluting with 3.5:1 hexanes/ (101 MHz, DMSO-d6): δ 14.0, 20.8, 47.4, 49.3, 115.2 (d, JFC )
EtOAc, yielded 43.0 mg (0.141 mmol, 92% yield) of a white 23.8), 119.0 (d, JFC ) 23.2), 124.4, 125.1 (d, JFC ) 7.6), 128.4,
solid: mp 179.5-180.5 °C. IR (KBr): 3427 (b), 3204, 3098, 132.0 (d, JFC ) 7.6), 135.7, 137.0, 137.4, 158.9 (d, JFC ) 243),
2960, 2868, 1703, 1670, 1539, 1440, 1354, 1249, 1223, 1131, 166.3, 170.8. HRMS (FAB+) m/e: 327.1502 (MH+ C19H20N2O2F
802, 742 cm-1. 1H NMR (400 MHz, DMSO-d6): δ 0.72 (d, 3, J requires 327.1509).
) 6.1), 0.83 (d, 3, J ) 6.1), 1.03 (t, 3, J ) 7.0), 1.57-1.65 (m, Benzodiazepine 1l. The support-bound benzodiazepine
3), 3.65-3.68 (m, 1), 3.75-3.82 (m, 1), 4.15-4.20 (m, 1), 7.93 was prepared according to the above procedure from 5-chlo-
(d, 1, J ) 8.5), 8.11 (d, 1, J ) 8.5), 8.22 (s, 1), 8.85 (d, 1, J ) roanthranilic acid, (S)-phenylalanine methyl ester hydrochlo-
5.9). 13C NMR (101 MHz, DMSO-d6): δ 13.0, 21.5, 22.7, 23.7, ride, and allyl bromide. The scalemic R-amino ester was
36.1, 42.1, 49.9, 117.9, 119.8, 131.1, 135.0, 140.0, 149.4, 165.1, loaded according to the nonracemizing conditions. Cleavage
169.3. Anal. Calcd for C15H19O4N3: C, 59.00; H, 6.27; N, of 0.319 g (0.111 mmol) of this resin by stirring in 20 mL of
13.76. Found: C, 58.97; H, 6.48; N, 13.69. 90:5:5 TFA/Me2S/H2O for 36 h yielded a crude yellow residue
Benzodiazepine 1i. The support-bound benzodiazepine after gravity filtration, rinsing of the resin and filter paper
was prepared according to the above procedure from 4-ni- with CH2Cl2 (4 × 15 mL) and CH3OH (2 × 10 mL), and
troanthranilic acid, 3-(2-thienyl)-D,L-alanine methyl ester concentration. Silica gel chromatography, eluting with 4:1
hydrochloride, and (bromomethyl)cyclopropane. Racemic R-ami- hexanes/EtOAc, yielded 33.6 mg (0.098 mmol) of an amorphous
no ester was loaded according to the nonracemizing conditions. white solid (89% yield). IR (KBr): 3447 (b), 3072, 3033, 2940,
Cleavage of 0.390 g (0.150 mmol) of this resin by stirring in 1670, 1479, 1440, 1354 cm-1. 1H NMR (400 MHz, DMSO-d6):
20 mL of 90:5:5 TFA/Me2S/H2O for 36 h yielded a crude yellow δ 2.91 (dd, 1, J ) 8.9. 14.2), 3.12 (dd, 1, J ) 5.45, 14.2), 4.01-
Synthesis of 1,4-Benzodiazepine-2,5-diones J. Org. Chem., Vol. 62, No. 5, 1997 1251

4.06 (m, 1), 4.40 (dd, 1, J ) 5.4, 16.5), 4.59 (dd, 1, J ) 4.7, stirring in 20 mL of 90:5:5 TFA/Me2S/H2O for 36 h yielded a
16.5), 5.01-5.09 (m, 2), 5.68-5.77 (m, 1), 7.14-7.18 (m, 1), crude yellow residue after gravity filtration, rinsing of the resin
7.23 (t, 2, J ) 7.1), 7.29 (d, 2, J ) 7.1), 7.47 ( d, 1, J ) 8.8), and filter paper with CH2Cl2 (4 × 15 mL) and CH3OH (2 × 10
7.57 (d, 1, J ) 2.6), 7.62 (dd, 1, J ) 2.6, 8.8), 8.86 (d, 1, J ) mL), and concentration. Silica gel chromatography eluting
6.4). 13C NMR (101 MHz, DMSO-d6): δ 33.6, 49.1, 53.6, 116.7, with 2:1 hexanes/EtOAc provided 28.8 mg (0.090 mmol, 69%
124.5, 126.3, 128.1, 128.7, 129.3, 129.6, 131.1, 131.8, 133.0, yield) of an amorphous white powder. IR (KBr): 3432, 3072,
137.6, 138.3, 166.3, 169.4. Anal. Calcd for C19H17O2N2Cl: C, 2919, 1664, 1559, 1426, 1401, 1226, 1097, 786, 752, 700 cm-1.
66.96; H, 5.03; N, 8.22. Found: C, 66.83; H, 5.20; N, 8.18. 1
H NMR (400 MHz, DMSO-d6): δ 0.00-0.02 (m, 1), 0.15-0.18
Benzodiazepine 1m. The support-bound benzodiazepine (m, 1), 0.28-0.31 (m, 2), 0.92-0.96 (m, 1), 2.94 (dd, 1, J ) 9.0,
was prepared according to the above procedure from 4-meth- 14.0), 3.16 (dd, 1, J ) 4.9, 14.0), 3.98-4.05 (m, 3), 7.16-7.18
oxyanthranilic acid (vide supra), (S)-leucine methyl ester (m, 1), 7.23 (t, 2, J ) 7.1), 7.30-7.36 (m, 3), 8.07 (d, 1, J )
hydrochloride, and (bromomethyl)cyclopropane. The scalemic 6.7), 8.63 (d, 1, J ) 2.9), 8.85 (d, 1, J ) 6.1). 13C NMR (101
R-amino ester was loaded according to the racemizing condi- MHz, DMSO-d6): δ 2.9, 3.6, 10.1, 33.6, 48.4, 53.7, 120.9, 124.3,
tions. Cleavage of 0.428 g (0.145 mmol) of this resin by stirring 126.3, 128.1, 129.4, 137.6, 139.5, 150.5, 151.2, 166.3, 169.7.
in 20 mL of 90:5:5 TFA/Me2S/H2O for 36 h yielded a crude Anal. Calcd for C19H19O2N3: C, 71.10; H, 5.96; N, 13.07.
yellow residue after gravity filtration, rinsing of the resin and Found: C, 71.10; H, 6.18; N, 12.92.
filter paper with CH2Cl2 (4 × 15 mL) and CH3OH (2 × 10 mL),
Benzodiazepine 1q. The support-bound benzodiazepine
and concentration. Silica gel chromatography, eluting with
was prepared according to the above procedure from 4-methyl-
4:1-2:1 hexanes/EtOAc, yielded 36.0 mg (0.114 mmol) of a
anthranilic acid, 3-(2-thienyl)-D,L-alanine methyl ester hydro-
white solid (79% yield), mp 133-135 °C. IR (KBr): 3434 (b),
3184, 3079, 2954, 1664, 1611, 1453, 1381, 1262, 1223, 1124, chloride, and piperonyl chloride. The racemic R-amino ester
1032, 841, 795 cm-1. 1H NMR (400 MHz, DMSO-d6): δ 0.05- was loaded according to the nonracemizing conditions. Cleav-
0.10 (m, 2), 0.27-0.31 (m, 2), 0.72 (d, 3, J ) 6.3), 0.81 (d, 4, J age of 0.332 g (0.085 mmol) of this resin by stirring in 20 mL
) 6.3), 1.57-1.68 (m, 3), 33.57-3.67 (m, 2), 3.83 (s, 3), 3.99 of 90:5:5 TFA/Me2S/H2O for 36 h yielded a crude yellow residue
(dd, 1, J ) 7.4, 14.5), 6.91 (dd, 1, J ) 2.1, 8.7), 7.01 (d, 1, J ) after gravity filtration, rinsing of the resin and filter paper
2.1), 7.62 (d, 1, J ) 8.6), 8.41 (d, 1, J ) 5.8). 13C NMR (101 with CH2Cl2 (4 × 15 mL) and CH3OH (2 × 10 mL), and
MHz, DMSO-d6): δ 13.0, 13.7, 19.9, 31.7, 32.7, 33.8, 46.4, 60.2, concentration. Silica gel chromatography eluting with 3:1-
60.8, 65.5, 118.1, 121.9, 132.8, 140.9, 151.2, 171.8, 177.4, 179.9. 2:1 hexanes/EtOAc provided 23.5 mg (0.056 mmol, 66% yield)
HRMS (FAB+) m/e: 317.1858 (MH+ C18H25N2O3 requires of an off-white amorphous powder. IR (KBr): 3448, 3283,
317.1865). 2927, 1665, 1649, 1612, 1500, 1438, 1248, 1041, 696 cm-1. 1H
Benzodiazepine 1n. The support-bound benzodiazepine NMR (400 MHz, DMSO-d6): δ 2.31 (s, 3), 3.17 (dd, 1, J ) 8.2,
was prepared according to the above procedure from 5-chlo- 15.0), 3.33-3.37 (m, 1), 3.90-3.95 (m, 1), 4.83 (d, 1, J ) 15.5),
roanthranilic acid, H-Lys(Boc)-OMe‚HCl, and allyl bromide. 5.25 (d, 1, J ) 15.5), 5.94 (s, 2), 6.56 (b d, 1, J ) 8.0), 6.60 (d,
The scalemic R-amino ester was loaded according to the 1, J ) 1.4), 6.74 (d, 1, J ) 7.9), 6.89 (dd, 1, J ) 3.5, 5.0), 6.93
racemizing conditions. Cleavage of 0.405 g (0.137 mmol) of (b s, 1), 7.09 (b d, 1, J ) 7.9), 7.29 (dd, 1, J ) 1.2, 5.1), 7.32 (b
this resin by stirring in 20 mL 90:5:5 of TFA/Me2S/H2O for 36 s, 1), 7.46 (d, 1, J ) 7.98), 8.74, (d, 1, J ) 6.2). 13C NMR (101
h yielded a crude yellow residue after gravity filtration, rinsing MHz, DMSO-d6): δ 20.9, 28.2, 49.2, 54.0, 100.9, 107.3, 108.1,
of the resin and filter paper with CH2Cl2 (4 × 15 mL) and CH3- 120.3, 123.0, 124.6, 126.6, 126.7, 127.1, 129.3, 131.0, 137.7,
OH (2 × 10 mL), and concentration. Silica gel chromatography 138.9, 139.6, 142.3, 146.2, 147.3, 167.4, 169.7. HRMS (FAB+)
eluting with 90:10:1 CH2Cl2/CH3OH/NH4OH provided 28 mg m/e: 420.1136 (M+ C23H20N2O4S requires 420.1144).
(0.087 mmol, 63% yield) of a viscous glass. IR (thin film): Benzodiazepine 1r. The support-bound benzodiazepine
3177, 3079, 2940, 1670, 1598, 1572, 1486, 1446, 1367, 1210, was prepared according to the above procedure from 3-methyl-
736 cm-1. 1H NMR (400 MHz, CD3OD): δ 1.37-1.40 (m, 1), anthranilic acid, (S)-leucine methyl ester hydrochloride, and
1.40-1.48 (m, 1), 1.48-1.52 (m, 2), 1.72-1.79 (m, 1), 1.90- 4-bromobenzyl bromide. The scalemic R-amino ester was
1.97 (m, 1), 2.66 (t, 2, J ) 7.0), 3.80 (t, 1, J ) 6.5), 4.46 (dd, 1, loaded according to the nonracemizing conditions. Cleavage
J ) 5.5, 16.3), 4.60 (dd, 1, J ) 5.0, 16.3), 5.11-5.15 (m, 2). of 0.355 g (0.112 mmol) of this resin by stirring in 20 mL of
5.78-5.84 (m, 1), 7.47 (d, 1, J ) 8.8), 7.58 (dd, 1, J ) 2.6, 8.8), 90:5:5 TFA/Me2S/H2O for 36 h yielded a crude yellow residue
7.73 (d, 1, J ) 2.5). 13C NMR (101 MHz, DMSO-d6): δ 24.0, after gravity filtration, rinsing of the resin and filter paper
28.9, 31.9, 41.5, 51.5, 53.8, 117.7, 125.7, 130.3, 132.1, 132.5, with CH2Cl2 (4 × 15 mL) and CH3OH (2 × 10 mL), and
133.6, 134.0, 140.1, 169.4, 171.6. HRMS (FAB+) m/e: 322.1319 concentration. Silica gel chromatography eluting with 4.5:1
(MH+ C16H21N3O2Cl requires 322.1322). hexanes/EtOAc provided 34.5 mg (0.083 mmol, 74% yield) of
Benzodiazepine 1o. The support-bound benzodiazepine an amorphous white powder. IR (KBr): 2966, 1654, 1637,
was prepared according to the above procedure from 4-chlo- 1560, 1542, 1508, 1090, 832, 803 cm-1. 1H NMR (400 MHz,
roanthranilic acid, (S)-leucine methyl ester hydrochloride, and DMSO-d6): δ 0.68 (d, 3, J ) 6.4), 0.78 (d, 3, J ) 6.4), 1.54 (t,
EtI. The scalemic R-amino ester was loaded according to the 2, 6.9), 1.59-1.66 (m, 1), 2.40 (s, 3), 3.56 (q, 1, J ) 6.5), 4.22
racemizing conditions. Cleavage of 0.440 g (0.150 mmol) of (d, 1, J ) 14.7), 5.31 (d, 1, J ) 14.6), 6.94 (d, 2, J ) 8.2), 7.31
this resin by stirring in 20 mL of 90:5:5 TFA/Me2S/H2O for 36
(t, 1, J ) 7.5), 7.33 (d, 2, J ) 8.2), 7.43 (d, 1, J ) 7.9), 7.49 (d,
h yielded a crude yellow residue after gravity filtration, rinsing
1, J ) 7.6), 8.47 (d, 1, J ) 6.0). Anal. Calcd for C21H23N2O2-
of the resin and filter paper with CH2Cl2 (4 × 15 mL) and CH3-
Br: C, 60.73; H, 5.58; N, 6.74. Found: C, 60.90; H, 5.69; N,
OH (2 × 10 mL), and concentration. Silica gel chromatogra-
6.65.
phy, eluting with 3:1 hexanes/EtOAc, yielded 33.0 mg (0.112
mmol) of an amorphous off-white powder (75% yield). IR Benzodiazepine 1s. The support-bound benzodiazepine
(KBr): 3441 (b), 2967, 1689, 1657, 1598, 1440 cm-1. 1H NMR was prepared according to the above procedure from 3-bromo-
(400 MHz, DMSO-d6): δ 0.73 (d, 3, J ) 6.3), 0.82 (d, 3, J ) 5-methylanthranilic acid, leucine methyl ester hydrochloride,
6.3), 0.98 (t, 3, J ) 7.0), 1.45-1.65 (m, 3), 3.61 (m, 1), 3.70 (m, and iodomethane. The scalemic R-amino ester was loaded
1), 4.13 (m, 1), 7.39 (d, 1, J ) 8.3), 7.58 (s, 1), 7.68 (d, 1, 8.3), according to the nonracemizing conditions. Cleavage of 0.248
8.62 (d, 1, 5.8). 13C NMR (101 MHz, DMSO-d6): δ 13.0, 21.6, g (0.073 mmol) of this resin by stirring in 20 mL of 90:5:5 TFA/
22.8, 23.8, 36.3, 42.0, 50.1, 122.5, 125.7, 128.8, 131.1, 136.5, Me2S/H2O for 36 h yielded a crude yellow residue after gravity
140.4, 166.8, 169.4. HRMS (FAB+) m/e: 295.1212 (MH+ filtration, rinsing of the resin and filter paper with CH2Cl2 (4
C15H20N2O2Cl requires 295.1213). × 15 mL) and CH3OH (2 × 10 mL), and concentration. Silica
Pyridodiazepine 1p. The support-bound benzodiazepine gel chromatography eluting with 3:1-2:1 hexanes/EtOAc
was prepared according to the above procedure from 2-ami- provided 21.5 mg (0.064 mmol, 88% yield) of an amorphous
nopyridine-3-carboxylic acid, (S)-phenylalanine methyl ester white powder. Two compounds eluted separately from each
hydrochloride, and (bromomethyl)cyclopropane. The scalemic other. On standing, the second to elute converted into the first.
R-amino ester was loaded according to the nonracemizing The two compounds are atropisomers, which results from
conditions. Cleavage of 0.398 g (0.130 mmol) of this resin by substitution at the 3-position. Equilibration occurs to provide
1252 J. Org. Chem., Vol. 62, No. 5, 1997 Boojamra et al.

the most stable atropisomer. Gilman88,89 and Blackburn9 have dried, round-bottom flask was dissolved 9-BBN dimer (0.506
previously detailed atropisomerism in benzodiazepine deriva- g, 2.02 mmol) in 7 mL of THF. A dry stir bar was introduced
tives. IR (KBr): 3271, 2955, 1667, 1447, 1108, 779 cm-1. 1H into the flask, and the contents were cooled with stirring to 0
NMR (500 MHz, DMSO-d6): δ 0.73 (d, 3, J ) 5.5), 0.79 (d, 3, °C. Over the course of 7 min was added 1-hexene (0.65 mL,
J ) 5.5), 1.59 (m, 3), 2.35 (s, 3), 3.16 (s, 3), 3.61 (m, 1), 7.50 (s, 5.2 mmol) via syringe. The reaction mixture was warmed to
1), 7.77 (s, 1), 8.65 (d, 1, J ) 5.5). 13C NMR (125 MHz, DMSO- rt and stirred for 6 h. This solution was then transferred via
d6): δ 20.3, 22.2, 23.0, 24.4, 36.3, 37.6, 50.7, 119.1, 129.3, 132.9, cannula into a 50-mL, round-bottom flask containing the resin
137.4, 137.6, 139.1, 167.2, 171.2. HRMS (FAB+) m/e: 339.0705 derivatized with the benzodiazepine (0.70 g, 0.24 mmol) that
(MH+ C15H20N2O2Br requires 339.0708). was presolvated in 7 mL of THF and 5 mL of 2 N aqueous
Thienodiazepine 1t. The support-bound benzodiazepine K2CO3. The flask was fitted with a reflux condenser and a
was prepared according to the above procedure from 3-amino- needle passed through the top of the condenser into the
4-(4-chlorobenzenesulfonyl)thiophene-2-carboxylic acid, D,L- solution. The solution was then degassed by bubbling Ar
homophenylalanine methyl ester hydrochloride, and HOAc. through it for 20 min. A catalytic amount of Pd(PPh3)4 was
The racemic R-amino ester was loaded according to the then added and purging continued for 10 min. The reaction
nonracemizing conditions. Cleavage of 0.323 g (0.101 mmol) mixture was stirred with a magnetic stirrer for 20 h while
of this resin by stirring in 20 mL of 90:5:5 TFA/Me2S/H2O for being heated at reflux. The solution was removed from the
36 h yielded a crude yellow residue after gravity filtration, resin by filtration cannula, and the resin was rinsed with DMF
rinsing of the resin and filter paper with CH2Cl2 (4 × 15 mL) (7 x 20 mL), CH2Cl2 (7 × 20 mL), and finally CH3OH (3 × 10
and CH3OH (2 × 10 mL), and concentration. Silica gel mL). The resin was dried in vacuo to a constant weight, and
chromatography eluting with 2:1 hexanes/EtOAc provided 42.3 0.418 g (0.137 mmol) was cleaved by stirring in 20 mL of 90:
mg (0.092 mmol, 90% yield) of an amorphous white powder. 5:5 TFA/Me2S/H2O for 36 h. Gravity filtration and rinsing of
IR (KBr): 2900, 1715, 1654, 1561, 1502, 1355, 1220, 1155, the resin and filter paper with excess CH2Cl2 and CH3OH
1091, 750, 667, 615 cm-1. 1H NMR (400 MHz, DMSO-d6): δ yielded a crude yellow solid upon evaporation in vacuo. Silica
1.76-1.85 (m, 1), 1.96-2.05 (m, 1), 2.51-2.64 (m, 2), 3.63- gel chromatography 2:1 hexanes/EtOAc yielded 55.2 mg (0.106
3.69 (m, 1), 7.10 (d, 2, J ) 7.0), 7.15 (d, 1, J ) 7.1), 7.23 (t, 2, mmol, 77% yield) of white solid, mp 176-178 °C. IR (KBr):
J ) 7.1), 7.73 (d, 2, J ) 8.6), 7.99 (d, 2, J ) 8.6), 8.68 (d, 1, J 3430 (b), 3100, 3026, 2927, 1657, 1612, 1517, 1443, 1385, 1224,
) 4.7), 8.86 (s, 1), 9.43 (s, 1). 13C NMR (101 MHz, DMSO-d6) 1174, 1130, 1008, 834, 757, 698 cm-1. 1H NMR (400 MHz,
δ 30.0, 31.2, 52.2, 125.9, 126.2, 128.2, 128.3, 129.1, 130.0, 130.5, DMSO-d6): δ 0.79 (t, 3, J ) 6.5), 1.17-1.22 (m, 6), 1.50 (t, 2,
133.4, 139.1, 139.4, 139.5, 141.1, 162.3, 168.8. Anal. Calcd J ) 6.5), 2.56 (t, 2, J ) 7.5), 2.83 (dd, 1, J ) 8.6, 14.2), 3.07
for C21H17N2O4S2Cl: C, 54.72; H, 3.72; N, 6.08. Found: C, (dd, 1, J ) 5.5, 14.3), 3.89 (q, 1, J ) 6.0), 5.00 (d, 1, J ) 15.9),
54.88; H, 3.90; N, 5.87. 5.39 (d, 1, J ) 15.9), 6.62 (d, 2, J ) 8.4), 7.06 (d, 1, J ) 8.0),
Benzodiazepine 1u. This is a Suzuki cross-coupling 7.10 (d, 2, J ) 8.4), 7.15 (d, 2, J ) 8.1), 7.30-7.34 (m, 2), 7.42
product. In a 100-mL round-bottom flask was suspended 0.500 (t, 2, J ) 7.6), 7.46 (d, 1, J ) 7.9), 7.51 (d, 2, J ) 8.2), 7.59 (d,
g (0.180 mmol) of the resin from which 1d was cleaved in 30 2, J ) 7.2), 8.66 (d, 1, J ) 6.2), 9.2 (s, 1). 13C NMR (101 MHz,
mL of THF. After the resin had completely solvated (5 min), DMSO-d6): δ 13.8, 21.9, 28.1, 30.2, 31.0, 32.9, 34.8, 49.1, 54.2,
2 N aqueous K2CO3 (5 mL) was added to the flask generating 114.9, 122.1, 125.7, 126.4, 126.5, 127.3, 127.4, 127.5, 127.8,
a biphasic liquid. With magnetic stirring, 4-methoxyben- 128.8, 129.4, 130.3, 136.6, 138.8, 139.0, 139.5, 146.9, 155.8,
zeneboronic acid (0.281 g, 1.85 mmol) was added, and the flask 167.5, 170.2. Anal. Calcd for C35H36O3N2: C, 78.92; H, 6.81;
was fitted with a reflux condenser. A long needle was dropped N, 5.26. Found: C, 78.74; H, 6.84; N, 5.07.
through the reflux condenser, and the system was purged with Benzodiazepine 1w. This is a Suzuki cross-coupling
Ar. After 30 min, Pd(PPh3)4 (0.100 g, 0.086 mmol) was added, product. The support-bound benzodiazepine, prepared accord-
and the suspension was purged with Ar for another 15 min. ing to the above procedure from 5-bromoanthranilic acid,
The purge needle was removed, and the system was heated leucine methyl ester hydrochloride (nonracemizing reductive
at reflux for 18 h under 1 atm of Ar. The yellow/orange amination conditions), and (bromomethyl)cyclopropane, was
solution was removed from the resin by filtration cannula, and suspended (0.500 g, 0.180 mmol) in 4 mL of THF, 4 mL of
the resin was rinsed with DMF (7 × 20 mL), CH2Cl2 (7 × 20 DMF, and 1 mL of H2O. After the resin had completely
mL), and finally CH3OH (3 × 10 mL). The resin was dried in solvated (5 min), CsCO3 (0.61 g, 1.87 mmol) was added to the
vacuo to a constant weight, and 0.320 g (0.106 mmol) was flask. With magnetic stirring, 4-methoxybenzeneboronic acid
cleaved by stirring in 20 mL of 90:5:5 TFA/Me2S/H2O for 36 (0.577 g, 3.80 mmol) was added, followed by addition of
h. Gravity filtration and rinsing of the resin and filter paper [PdCl2(dppf)]75 (0.100 g, 0.122 mmol). The solution was heated
with excess CH2Cl2 and CH3OH yielded a crude yellow solid gently to 55 °C for 12 h under 1 atm of N2. The black solution
upon evaporation in vacuo. Silica gel chromatography 2:1 was removed from the resin by filtration cannula and was
hexanes/EtOAc yielded 24.0 mg (0.066 mmol, 62% yield) of rinsed with a solution of saturated KCN in DMSO (in order
white solid, mp 202-204 °C (phase change and subsequent to remove the Pd) followed by DMF (7 × 20 mL), CH2Cl2 (7 ×
decomposition). IR (KBr): 3454 (b), 3177, 3046, 2960, 1690, 20 mL), and finally CH3OH (3 × 10 mL). Caution! Extreme
1664, 1611, 1493, 1453, 1256, 1183, 1117, 1032, 828 cm-1. 1H care should be used during rinsing with KCN in DMSO. This
NMR (400 MHz, DMSO-d6): δ 0.74 (d, 3, J ) 6.1), 0.82 (d, 3, mixture is toxic and may be deadly on skin contact. Appropri-
J ) 6.0), 1.03 (t, 3, J ) 6.8), 1.60-1.69 (m, 3), 3.62-3.68 (m, ate precautions should be taken. Rinses should be disposed of
1), 3.71-3.79 (m, 1), 3.80 (s, 3), 4.07-4.14 (m, 1), 7.04 (d, 2, J accordingly. The resin was dried in vacuo to a constant
) 8.1), 7.52 (d, 1, J ) 8.3), 7.67 (d, 2, J ) 8.1), 7.84-7.87 (m, weight, and 0.450 g (0.144 mmol) was cleaved by stirring in
2), 8.59 (d, 1, J ) 5.8). 13C NMR (101 MHz, DMSO-d6): δ 13.1, 20 mL of 90:5:5 TFA/Me2S/H2O for 36 h yielded a crude yellow
21.7, 22.7, 23.8, 36.4, 42.1, 50.2, 55.1, 114.5, 123.1, 126.4, 127.7, residue after gravity filtration, rinsing of the resin and filter
129.6, 130.2, 130.6, 136.7, 137.8, 159.2, 167.5, 169.3. Anal. paper with CH2Cl2 (4 × 15 mL) and CH3OH (2 × 10 mL), and
Calcd for C22H26O3N2: C, 72.11; H, 7.15; N, 7.64. Found: C, concentration. Silica gel chromatography eluting with 2:1
71.99; H, 7.00; N, 7.43. hexanes/EtOAc provided 44.0 mg (0.112 mmol, 78% yield) of
Benzodiazepine 1v. The support-bound benzodiazepine a white powder, mp 85-86 °C. IR (KBr): 2953, 1654, 1555,
was prepared according to the above procedure from 4-bro- 1486, 1444, 1364, 1246, 1180, 1025, 870, 828, 668, 553 cm-1.
1H NMR (400 MHz, DMSO-d ): δ 0.10 (m, 2), 0.31 (m, 2), 0.74
moanthranilic acid, H-(S)-Tyr(t-Bu)-OMe‚HCl, and 4-phenyl- 6
benzyl bromide. The scalemic R-amino ester was loaded (d, 3, J ) 6.3), 0.82 (m, 4), 1.59-1.67 (m, 3), 3.61-3.80 (m, 2),
according to the nonracemizing conditions. A Suzuki reaction 3.80 (s, 3), 4.04 (dd, 1, J ) 14.4, 7.6), 7.03, (d, 2, J ) 8.8), 7.57
was then performed with B-hexyl-9-BBN. In a 10-mL, flame- (d, 1, J ) 8.5), 7.68 (d, 2, J ) 8.8), 7.86 (dd, 1, J ) 8.5, 2.4),
7.88 (d, 1, J ) 2.25), 8.61 (d, 1, J ) 6.1). 13C NMR (101 MHz,
DMSO-d6): δ 3.5, 4.3, 10.4, 22.2, 23.2, 24.3, 36.9, 50.6, 51.1,
(88) Gilman, N. W.; Rosen, P.; Early, J. V.; Cook, C.; Todaro, L. J.
J. Am. Chem. Soc. 1990, 112, 3969-3978. 55.7, 115.0, 124.2, 126.9, 128.2, 129.9, 130.9, 131.1, 137.2,
(89) Gilman, N. W.; Rosen, P.; Earley, J. V.; Cook, C. M.; Blount, J. 138.6, 159.7, 168.1, 170.3. Anal. Calcd for C24H28N2O3: C,
F.; Todaro, L. J. J. Org. Chem. 1993, 58, 3285-3298. 73.44; H, 7.19; N, 7.14. Found: C, 73.19; H, 6.96; N, 7.14.
Synthesis of 1,4-Benzodiazepine-2,5-diones J. Org. Chem., Vol. 62, No. 5, 1997 1253

Racemization Study. Benzodiazepine 1c was prepared Table 3. Amounts Used for the Loading r-Amino Esters
according to the racemization-free procedure described previ- under Racemization Conditions
ously. HPLC analysis68 (chiral column, D-phenylglycine de- total
rivatized silica, 5% i-PrOH in isooctane eluant, 4 mL/min flow amino FW concn vol mmol mass i-Pr2EtN reductant
rate, UV detection at 262 nm) of benzodiazepine 1c, prepared label ester (g/mol) (M) (mL) ester (g) (mL) (g)
under racemization-free conditions, indicated a single enan- L Leu 181.7 0.15 50 7.50 1.36 0.393 1.59
tiomer, tR ) 45.4 min. A fully racemized sample of 1c was F Phe 215.7 0.15 50 7.50 1.62 0.393 1.59
also prepared by allowing preequilibration in the reductive Y Tyr 287.8 0.15 50 7.50 2.16 0.393 1.59
amination step for 6 h in the presence of i-Pr2EtN. Analysis E Glu 253.8 0.15 50 7.50 1.90 0.393 1.59
of the racemized sample under the same HPLC conditions Q Gln 422.9 0.15 50 7.50 3.17 0.393 1.59
indicated two well-separated peaks of approximate equal area, K Lys 296.8 0.15 50 7.50 2.23 0.393 1.59
tR ) 41.8, 44.8 min, confirming that complete racemization
had occurred. holes. Before the tubes cooled, a thermally resistant cloth was
An additional HPLC study was performed with 1b, synthe- used to wipe away the excess polyethylene from around the
sized from the more racemization-prone phenylalanine.90 Ben- outside of the tube. After these tubes cooled, they were
zodiazepine 1b was prepared according to the racemization- introduced manually into the bracket creating a reactor of 12
free procedure described previously. A racemized sample of × 8 (96) spatially separate reaction tubes per bracket, leaving
1b was also prepared. This racemized sample served as an most of the tube below the bracket. Only ∼1/4 in. was left
HPLC standard. HPLC analysis5 (chiral column, D-phenyl- above the bracket. It was useful to let the brackets sit in a
glycine derivatized silica, 4% i-PrOH in hexanes eluant, 9.99 desiccator that had been charged with 50 mL of THF for 30
mL/min flow rate, UV detection at 252 nm) of benzodiazepine min prior to inserting the filter tubes. This helped to swell
1b, prepared under racemization-free conditions indicated a the plastic, allowing for easy insertion. On standing for 2 h
major enantiomer, tR ) 29.5 min and a trace of a second (< on the bench top, the plastic shrinks and seals around the
1%), tR ) 35.7 min. Analysis of the racemized sample under tubes. This apparatus is shown in schematic form in Figure
the same conditions indicated two well-separated peaks, tR ) 3.
30.5 min and tR ) 35.8 min. Library Synthesis. Coupling of Linker to Support. In
Benzodiazepine 1o was also prepared employing racemiza- an oven-dried, two-neck, 2000-mL round-bottom flask, fitted
tion-free conditions. To verify that during and after alkylation with an oven-dried mechanical stirring arm, was dissolved
significant racemization was not able to occur, we heated 18.56 g (182.2 g mol-1, 102 mmol) of 4-hydroxy-2,6-dimethoxy-
benzodiazepine 1o in 6 N aqueous HCl at reflux for 12 h in benzaldehyde87 in 400 mL of dry DMF. The solution was
order to liberate the amino acid portion of this benzodiazepine. purged with Ar for 20 min. To this solution, with cooling to 0
The liquid was evaporated in vacuo to a yellow residue that °C, under constant positive Ar pressure, was added 2.45 g
was subsequently dissolved in 10 mL of CH3OH. To this (24.00 g mol-1, 97 mmol) of 95% NaH in four aliquots. The
mixture was added SOCl2 (28 mL, 0.40 mmol). The mixture dark orange solution evolved H2 (g) rapidly and turned a turbid
was heated at reflux for 3 h, and the volatile contents of the blood red color and then finally became cloudy orange in
flask were removed in vacuo. The yellow residue was then appearance. After the evolution of H2 ceased, the suspension
dissolved in 10 mL of CH2Cl2 and partitioned in equal volumes was purged continuously with Ar under vigorous mechanical
into two round-bottom flasks. One portion was treated with stirring and was simultaneously allowed to warm to rt. After
5 drops of (R)-R-methoxy-R-(trifluoromethyl)phenylacetic acid 20 min, 50.0 g (34.5 mmol of chloride) of Merrifield resin (200-
chloride, (R)-MTPA chloride, and 5 drops of Et3N, and the 400 mesh chloromethylpolystyrene, 1% divinylbenzene cross-
other portion was treated with 5 drops of (S)-MTPA chloride link, 0.69 mmol Cl-/g resin) was added with constant gentle
and 5 drops of Et3N. After the reactions were allowed to mechanical stirring under gentle positive pressure of Ar,
proceed for 12 h, the reaction solutions were each washed with taking care not to let the Ar flow blow the resin out of the
1 N aqueous HCl (3 × 2 mL), 5% aqueous K2CO3 (2 × 2 mL), flask as it was being added. The mixture was then heated to
saturated aqueous NaHCO3 (2 mL), and brine (2 mL). The 100 °C for 30 min and cooled to 50 °C where it was allowed to
organic layers were dried over MgSO4, and the crude residues stir under an atmosphere of Ar for 36 h. The resin was rinsed
were purified on a silica plug, eluting with 1:1 hexanes/EtOAc. with 2:1 DMF/MeOH (3 × 500 mL), DMF (7 × 375 mL), CH2-
The diastereomeric ratio of the products formed was evaluated Cl2 (7 × 400 mL), and finally MeOH (4 × 300 mL). After being
by capillary GC analysis (HP-1, 150-200 °C, 1 deg/min ramp, dried in vacuo, the resin had assumed the tan color of the
20 psi). The sample prepared from the (R)-MTPA chloride linker aldehyde. A strong aldehyde carbonyl stretch was
indicated a major diastereomer (tR ) 26.0 min, 97%) of (R)- observed in the IR (KBr) spectrum of this resin: 1690 cm-1.
MTPA-(S)-Leu-OMe, and a minor diastereomer (tR ) 27.0 min, Reductive Aminations, Racemizing Procedure. In
3%). The sample prepared from (S)-MTPA chloride indicated each of six tared 100-mL round-bottom flasks was placed 3.5
a minor diastereomer of (R)-MTPA-(S)-Leu-OMe (tR ) 26.0 g of aldehyde resin. To this was added 50 mL of 1% HOAc in
min, 3%) and a major diastereomer (tR ) 27.1 min, 97%). This DMF. After the resin had solvated (5 min), one R-amino ester
experiment demonstrates that racemization occurred to a (amounts are found in Table 3) was added to each flask, which
maximum extent of 3% in this leucine-derived benzodiazepine. was then labeled with the code for that particular R-amino
It is most likely that the racemization occurs in the benzodi- ester. To each flask was then added a catalytic amount of
azepine hydrolysis step69 and not during the benzodiazepine i-Pr2EtN (0.291 g, 2.25 mmol). The mixtures were secured on
synthesis. a shaker table and allowed to agitate for 9 h. The reductant,
Construction of Multitube Apparatus. The apparatus NaBH(OAc)3 (1.59 g, 7.50 mmol), was then added. After 6 h,
consists of three parts: a bracket, a tube, and a filter frit. The an IR spectrum of the resin containing H-Leu-OMe‚HCl
bracket, which secured the individual reactor tubes in place, indicated that the reaction had proceeded to completion. The
was made by sawing off the top ∼1/6 in. of 96 1-mL well resins were washed with 1:1 DMF/MeOH (1 × 30 mL), DMF
microtiter plates. The individual reaction tubes were made (7 × 30 mL), CH2Cl2 (7 × 30 mL), and MeOH (4 × 20 mL). A
by cutting 1500 identical 2 in. pieces of 7 mm (outer diameter) Kim Wipe was inserted into the mouth of each flask and the
glass tubing using a diamond-bladed glass saw with a 2-in. resins dried in a vacuum desiccator for 18 h. Infrared spectra
stop to ensure uniformity. The filter frits were introduced onto (KBr) were obtained, and all resins showed a strong ester
each of these by heating one end in a Bunsen burner until carbonyl absorbance at ∼1740 cm-1.
yellow-hot and melting it directly into a 70 µm hydrophobic Reductive Aminations, Nonracemizing Loading of
polyethylene filter disk (purchased from Applied Separations, Racemic r-Amino Esters and Glycine. Into four tared 100-
catalog no. 2427), taking the time to ensure that the entire mL round-bottom flasks was aliquoted 3.5 g of the aldehyde
bottom of the tube was sealed, with no obvious large pores or resin. To each of them was added 50 mL of DMF and 0.5 mL
of HOAc. Upon addition of the acetic acid, the resin turned
(90) Kovács, J. In The Peptides; Gross, E., Meienhofer, J., Eds.; from bright orange to yellow. NaBH(OAc)3 (1.59 g, 7.50 mmol)
Academic: New York, 1980; Vol. 2, pp 485-539. was added and dissolved, followed immediately by the R-amino
1254 J. Org. Chem., Vol. 62, No. 5, 1997 Boojamra et al.

Table 4. Amounts Used for the Loading r-Amino Esters top and placed upright in a box on a shaker table. After 12 h
under Racemization-Free Conditions of agitation, the filter cartridges were rinsed, 12 at a time,
amino FW concn total vol mass reductant using a Visiprep filter cartridge filtration apparatus obtained
label ester (g/mol) (M) (mL) mmol (g) (g) from Supelco (catalog no. 5-7030). The rinses consisted of
soaking the resin in each cartridge with DMF (5 mL) as many
G Gly 125.6 0.15 50 7.50 0.94 1.59 times as was needed until the supernatant that was removed
A Ala 139.6 0.15 50 7.50 1.05 1.59 was colorless. Each resin sample was then soaked for 5 min
U 2-Nal 245.3 0.15 50 7.50 1.84 1.59 in CH2Cl2 (5 × 5 mL). The acylation and rinsing was repeated
X 2-Thi 221.7 0.15 50 7.50 1.66 1.59
in an identical manner, with the exception that an additional
Table 5. Distribution of r-Amino Ester-Derivatized rinse with MeOH (3 × 5 mL) was performed at the end. All
Resin by Mass 120 filter cartridges were sealed at the top with a 14/20 rubber
septum and dried in a vacuum desiccator overnight.
amino FW recovered 0.9/12 of resin Resin Distribution, Cyclization, and Alkylation. The
label ester (g/mol) mass (g) to each cartridge (g) cartridges were removed from the vacuum desiccator, and a
L Leu 181.7 3.55 0.266 mixture of 3:2 1,2-dichloroethane/DMF was found to be isopyc-
F Phe 215.7 3.79 0.284 nic83 with respect to the resins in a randomly chosen cartridge.
Y Tyr 287.8 3.95 0.296 Since each cartridge had a structurally unique resin-bound
E Glu 253.8 3.92 0.294 benzodiazepine precursor, there was concern that 3:2 dichlo-
Q Gln 422.9 4.21 0.316 roethane/DMF might not be isopycnic with respect to the
K Lys 296.8 3.88 0.291 contents of all the cartridges. This was found not to be the
G Gly 125.6 3.59 0.268 case. In particular, for each cartridge after dilution, no visible
A Ala 139.6 3.64 0.273 settling of the resin was observed over 10 s. We felt that this
U 2-Nal 245.3 3.93 0.295 was sufficient to ensure that removal of equal volume aliquots
X 2-Thi 221.7 3.84 0.288 from 120 equal total volume slurries would allow for equimolar
distribution.
Table 6. Amounts Used for the Anthranilic Acid
Acylation Step The contents of every cartridge were suspended in excess
3:2 dichloroethane/DMF and agitated on a vortex mixer until
anthranilic completely solvated. The liquid was removed by suction, the
acid FW concn vol mass EDC
label (abbr Ant) (g/mol) (M) (mL) mmol (mg) (mg, M) cartridge bottom was stoppered with an 8 mm white rubber
septum, and 2 mL of 3:2 dichloroethane/DMF was delivered
A anthranilic 137.1 0.35 3.0 1.05 144 247, 0.43 by a Pipetman. Aliquots of 240 µL of the contents of each
acid (Ant)
B 4-Cl Ant 171.6 0.35 3.0 1.05 180 247, 0.43
cartridge were transfered to 10 separate reaction tubes. The
C 5-Cl Ant 171.6 0.35 3.0 1.05 180 247, 0.43 volume was delivered as 2 × 120 µL portions to help average
D 6-F Ant 155.1 0.35 3.0 1.05 162 247, 0.43 out any variation. Aliquoting was performed with a Pipetman,
E 3-Me Ant 151.7 0.35 3.0 1.05 159 247, 0.43 fitted with tips whose ends had been cropped to provide a
F 4-Br Ant 216.0 0.35 3.0 1.05 227 247, 0.43 wider bore to accommodate free passage of polymer beads.
G 5-Br Ant 216.0 0.35 3.0 1.05 227 247, 0.43 Enough extra resin in each cartridge remained to remake at
H 4-MeO Ant 167.2 0.35 3.0 1.05 175 247, 0.43 least one benzodiazepine if necessary.
I 4-nitro Ant 182.1 0.35 3.0 1.05 191 247, 0.43 A total of 16.5 multitube plates were filled, and each of these
J 4-Me Ant 151.7 0.35 3.0 1.05 159 247, 0.43 plates was blotted on a terry cloth towel, which helped to draw
K 3-Br, 5-Me Ant 230.1 0.35 3.0 1.05 242 247, 0.43
L 3-pyridine Ant 138.1 0.35 3.0 1.05 145 247, 0.43
solvent out of the tubes. Several polypropylene covers from
pipet tip racks (referred to as polypropylene bins) were then
used as small solvent reservoirs and filled with DMF. Each
esters (amounts are shown in Table 4). The flasks were multitube plate was submerged in DMF 1 in. high (2 × 1 min).
secured on the shaker table. As the reduction proceeded, the The filter bottoms were blotted on terry cloth towels in between
color of the resin became even lighter. After 3 h, the resins rinses. Several large polypropylene tubs were then filled 1
were given the same rinse sequence described in the preceding in. high with THF, and the tubes were submerged and soaked
experimental procedure and were dried in vacuo for 12 h, and (3 × 30 min), allowing the tubes to drain in between soakings.
IR (KBr) spectra were obtained. All resins showed a strong The multitube plates were then submerged quickly in THF
ester carbonyl stretch at ∼1740 cm-1. (3×) and blotted on terry cloth towels in between rinses. The
Acylations were performed in 120 12-mL filter cartridges plates were then soaked in 4-Å sieve-dried THF for 36 h in an
with 70 µm hydrophobic polyethylene frits, purchased from N2 glove bag over 4-Å sieves.
Applied Separations (catalog no. 2427). In order to deliver an Cyclizations were performed using lithiated acetanilide in
equimolar amount of resin-bound R-amino ester into each of a THF/DMF solution: In a 2-L, oven-dried, round-bottom flask,
these reaction vessels, the total mass of each of the R-amino fitted with a large magnetic stir bar, was placed 500 mL of
ester-derivatized resins was obtained based on the tare of the 4-Å sieve-dried THF. To this was added 55.0 g (407 mmol,
round-bottom flasks. This total mass was then divided by 12 135.2 g mol-1) of acetanilide. The acetanilide had been dried
(corresponding to acylation with 12 anthranilic acids). In order in vacuo overnight. This solution was cooled to -78 °C, and
to have a small amount of resin left over, this quantity was the flask was purged with dry N2. With rapid stirring,
then multiplied by 0.9. These masses are shown in the last n-butyllithium (2.5 M in hexanes, 150 mL, 375 mmol) was
column of Table 5. The resin corresponding to each R-amino added dropwise by syringe over the course of 1.5 h. As the
ester was then distributed into 12 filter cartridges according addition proceeded, a white solid precipitated from the reaction
to the mass in the right most column of Table 5. The total mixture. After n-BuLi addition was complete, the contents of
number of filter cartridges was 120; each was labeled in the flask were warmed to rt as solids continued to precipitate.
duplicate according to the R-amino ester resin contained within Once at rt, 1 L of dry DMF was added over 30 min. All solids
and the anthranilic acid to be acylated. dissolved. The flask containing the 0.25 M cyclization cocktail
Acylation of Resin-Bound r-Amino Esters. The bottom was moved into an N2 glovebag.
of each of the 120 12-mL filter cartridges was sealed with an Into each of 17 of the 40 small polypropylene bins was
8 mm rubber septum and to each of them was added 1-methyl- poured between 75 and 100 mL of the cyclization cocktail. Each
2-pyrrolidinone (NMP, 3 mL) and the coupling agent EDC (see of the multitube plates was removed from the 4-Å sieve/THF
Table 6 for amounts). Some EDC remained undissolved until bin, blotted on a terry cloth towel, which had dried over night
the anthranilic acid was added. Before addition of the an- in a N2 glovebag, and transferred into the small polypropylene
thranilic acid, the mixture of EDC, NMP, and resin was bins containing the 75-100 mL of cyclization cocktail. A
agitated on a vortex mixer for 10 s. The anthranilic acid (Table multichannel pipet device was then used to remove cyclization
6) was then added and immediately vortexed to homogeneity. cocktail from the bottom of the polypropylene reaction bins
The cartridges were sealed with a 14/20 rubber septum at the and drop it down from the tops of the reactor tubes in order
Synthesis of 1,4-Benzodiazepine-2,5-diones J. Org. Chem., Vol. 62, No. 5, 1997 1255

Table 7. Amounts Used for the Alkylation Stepa Table 8. Relative Chromophores of Benzodiazepines to
Nitrotoluene Standarda
alkylating FW d concn mass, vol
label agent (g/mol) (g/mL) (M) mol (g, mL) entry R1 R2 R3 factor
A none (AcOH) 60.05 1.049 0.40 0.030 1.80, 1.72 1 7-Cl ethyl (CH2)2CONH2 0.37
B methyl iodide 141.94 2.280 0.40 0.030 4.26, 1.86 2 7-Cl benzyl (CH2)2CO2H 0.38
C ethyl iodide 155.97 1.513 0.40 0.030 4.68, 3.09 3 9-pyridyl c-(C3H5)CH2 CH2C6H5 0.55
D 4-Phbenzyl bromide 247.13 0.40 0.030 7.41 g 4 8-MeO 2-MeOBn H 0.39
E piperonyl chloride 170.60 0.40 0.030 5.18 g
5 8-NO2 ethyl CH2CH(CH3)2 0.24
F 3,5-dimethylBnBr 199.10 0.40 0.030 5.97 g
6 7-Cl H CH2C6H5 0.35
G o-MeOBnCl 156.61 0.40 0.030 4.70 g
H (bromomethyl)- 135.00 1.410 0.40 0.030 4.05, 2.87
7 8-Me piperonyl 2-thienylmethyl 0.30
cyclopropane 8 7-Cl piperonyl CH2C6H5 0.25
I 2-(bromomethyl)- 221.10 0.40 0.030 6.63 g 9 8-O2N c-(C3H5)CH2 2-thienylmethyl 0.26
naphthalene 10 9-Br, 7-Me methyl CH2CH(CH3)2 0.40
J benzyl bromide 171.04 1.438 0.40 0.030 5.13, 3.57 11 6-F NH2COCH2 CH2C6H5 0.53
K iodoacetamide 187.96 0.40 0.030 5.64 g 12 9-Br, 7-Me H CH2CH(CH3)2 0.37
a Amounts are calculated on the basis of a volume of stock 13 8-hexyl 4-phenylBn CH2C6H4OH 0.20
14 7-Br ethyl CH2CH(CH3)2 0.25
solution of 75 mL. 15 7-Cl NH2COCH2 CH2C6H4OH 0.28
16 - ethyl CH2C6H5 12.4b
to ensure that the resin was properly bathed in the cyclization 17 8-MeO benzyl CH3 0.42
cocktail. These cyclization conditions were maintained for 40 18 6-F 3,5-diMeBn 2-thienylmethyl 0.46
h, over which time the volume of the cyclization cocktail had 19 7-Cl methyl CH2-2-(C10H7) 0.24
reduced and some white solid had precipitated. a Correction factors by which HPLC ratios of benzodiazepine
Solutions of the alkylating agents (0.4 M, 75 mL) were to internal standard (p-nitrotoluene) would need to be multiplied
prepared in the remaining dry polypropylene bins employing to obtain the actual ratio of benzodiazepine to standard. UV
the amounts shown in Table 7. With the exception of 3,4- absorbance detection was monitored at 252 nm. A sample calcula-
(methylenedioxy)benzyl (piperonyl) chloride and 2-methoxy- tion appears in the Experimental Section. b The benzodiazepine
benzyl chloride, the liquid alkylating agents were eluted from in entry 16 coeluted with p-nitrotoluene. This factor was mea-
a plug of basic alumina prior to use. The solid alkylating sured relative to 9-fluoreneone.
agents were used without any further purification or prepara-
tion, except 3,5-dimethylbenzyl bromide. This solid was device and reconcentrated. When brought to dryness, the
dissolved in a minimal volume of hexanes and eluted from a bottom of each well had a small orange or yellow pellet. The
plug of basic alumina prior to use. The 3,4-(methylenedioxy)- wells with glutamine-containing benzodiazepines had red
benzyl chloride (piperonyl chloride) and 2-methoxybenzyl pellets at the bottom, the color presumably derived from the
chloride were obtained from Trans World Chemicals as 50%- dimethoxybenzhydrol protecting group.
by-weight solutions in CH2Cl2, stabilized with CaCO3. These Determination of Yields by 1H NMR Combined with
solutions were passed through a 0.2-µm nylon syringe tip filter HPLC with UV Detection. The correction factors found in
and concentrated in vacuo without heating prior to use. Table 8 were all calculated according to the following protocol.
The resin blocks were removed from the cyclization cocktail, p-Nitrotoluene (∼2 mg) and the selected benzodiazepine (∼4
blotted on a terry cloth towel, and transferred to the bins mg) were dissolved in DMSO-d6. The exact ratio of the
containing the alkylating agents. A multichannel pipet device benzodiazepine (Bz) to the internal standard (Std) was deter-
was used to bathe each of the wells in the solution. The mined by 1H NMR integration. A portion of the mixture was
multitube plates were configured such that only one alkylating then injected onto an analytical RP-HPLC column (ramp 40-
agent was used for each of eight multitube plates. For the 100% MeOH in H2O over 40 min, UV detection at 252 nm) to
remaining eight plates only two alkylating agents were used. obtain the ratio of UV responses of the benzodiazepine relative
Alkylation reactions were therefore straightforward to perform to the internal standard. With this information the correction
since they were carried out in bins rather than requiring that factor (C) could be calculated:
the stock solutions be aliquoted into individual wells of
microtiter plates. Alkylations were allowed to continue for C ) (Std/Bz)HPLC × (Bz/Std)NMR
6-8 h. The multitube plates were blotted on towels until all
the alkylation cocktails had drained and transferred to bins In order to determine the exact quantity of a benzodiazepine
containing DMF to soak. in the library, the following protocol was used. The well from
The multitube plates were then repeatedly submerged in the library corresponding to the desired benzodiazepine was
at least 1 in. of DMF and allowed to soak for at least 5 min. diluted with 1.00 mL of a solution of 1:2:0.5 MeOH/CH3CN/
This was repeated five times with blotting in between each 1,2-dichloroethane. This solvent mixture was used instead of
rinse. The plates were then rinsed in the same manner with DMSO to allow for easy reconcentration. The contents of the
THF and then allowed to soak overnight in THF-filled polypro- well were repeatedly drawn up into a pipet to ensure that
pylene bins in an N2 glovebag. After 12 h, they were blotted everything had dissolved. A Pipetman was used to transfer
on terry cloth towels and given three quick rinses by submerg- 10 µL from the well to a small vial. A solution of the internal
ing them in THF, 1 in. deep. Although probably not necessary, standard p-nitrotoluene (0.1 mg mL-1) was added so as to
they were soaked overnight again in THF. After 12 h, they deliver 0.03 µmol (41.1 µL) of the internal standard. This
were all soaked in CH2Cl2 (7 × 20 min) to remove the THF. sample was mixed thoroughly after addition of five drops of
Cleavages. Into every well (except in the first and last DMF. Injection onto RP-HPLC, with ramp and UV detection
columns, which were left blank for controls during biological as above, provides the ratio of benzodiazepine to standard. The
assay) of 17, 96 2-mL well microtiter plates was added ∼1 mL ratio of benzodiazepine to standard by UV absorbance and the
of 90:5:5 TFA/(CH3)2S/H2O by glass pipet. The tubes of the correction factor (C) are then used to calculate the quantity of
multitube plates were slowly submerged into the cleavage the benzodiazepine: (Bz/Std)HPLC × C (correction factor) × 0.03
cocktail, taking care not to accidentally force TFA up over the µmol (quantity of Std).
edges of each well as the tubes filled. Cleavages were allowed Full characterization has previously been provided for most
to proceed for 36-40 h. The plates were removed slowly, of the benzodiazepines evaluated above. Complete character-
allowing liquid to drain into the wells, and immediately ization for an additional five benzodiazepines that had been
submerged into a second plate, charged with a rinse cocktail synthesized on large scale during the development of this
of 1:1 dichloroethane/CH3CN. The plate with the TFA mixture synthetic methodology and that were used in library yield
were concentrated in a Jouan microtiter plate concentrator. calculation follows.
After all the TFA was removed, the rinse cocktails from the Benzodiazepine (Table 8, Entry 16). IR (KBr): 3272,
appropriate wells were combined using a multichannel pipet 2943, 1654, 1602, 1460, 1400, 1240, 1129, 761, 703 cm-1. 1H
1256 J. Org. Chem., Vol. 62, No. 5, 1997 Boojamra et al.

NMR (400 MHz, DMSO-d6): δ 0.99 (t, 3, J ) 7.0), 2.90 (dd, 1, Benzodiazepine (Table 8, Entry 18). IR (thin film):
J ) 8.9, 14.2), 3.11 (dd, 1, J ) 5.4), 14.1), 3.68-3.74 (m, 1), 2916, 1672, 1611, 1468, 1390, 1233 cm-1. 1H NMR (400 MHz,
3.79-3.85 (m, 1), 4.11-4.16 (m, 1), 7.14-7.16 (m, 1), 7.22 (dd, DMSO-d6): δ 2.14 (s, 6) 3.20 (dd, 1, J ) 8.0, 14.9), 3.33-3.39
2, J ) 7.1, 7.5), 7.27-7.31 (m, 3), 7.47 (d, 1, J ) 7.9), 7.56- (m, 1), 4.08-4.13 (m, 1), 4.89 (d, 1,J ) 16.0), 5.27 (d, 1, J )
7.60 (m, 2), 8.68 (d, 1, J ) 6.2). 13C NMR (101 MHz, DMSO- 16.0), 6.69 (b s, 2), 6.80 (b s, 1), 6.90-6.92 (m, 1), 6.94-6.95
d6): δ 13.0, 33.6, 42.1, 53.9, 122.7, 125.6, 126.3, 128.1, 129.3 (m, 1), 7.11 (t, 1, J ) 9.0), 7.28-7.31 (m, 2), 7.46-7.51 (m, 1),
(b), 129.9, 132.2, 137.9, 139.1, 167.5, 169.3. Anal. Calcd for 9.07 (d, 1, J ) 7.1). 13C NMR (126 MHz, DMSO-d6): δ 21.1,
C18H18N2O2: C, 73.45; H, 6.16; N, 9.52. Found: C, 73.40; H, 28.4, 49.8, 54.2, 113.8 (d, JFC ) 21.5), 118.9, 119.1 (d, JFC )
5.98; N, 9.39. 15.0), 124.7, 125.0, 127.1, 128.8, 132.7 (d, JFC ) 10.0), 137.1,
Benzodiazepine (Table 8, Entry 15). IR (KBr): 2965, 137.8, 139.8, 141.0 (b), 159.9 (d, JFC ) 252.9), 163.7, 170.3.
1677, 1519, 1440, 1368, 1210, 1137, 841 cm-1. 1H NMR (400 HRMS (FAB+) m/e: 409.1377 (MH+ C23H22FN2O2S requires
MHz, DMSO-d6): δ 2.78 (dd, 1, J ) 9.3, 14.3), 2.98 (dd, 1, J ) 409.1386).
5.0, 14.4), 3.87-3.90 (m, 1), 4.22 (d, 1, J ) 16.8), 4.41 (d, 1, J HPLC Conditions. The same conditions were used both
) 16.8), 6.61 (d, 2, J ) 8.4), 7.06 (d, 2, J ) 8.4), 7.17 (b s, 1), when evaluating yields and when determining the correction
7.35 (d, 1, J ) 8.8), 7.56 (d, 1, J ) 2.6), 7.59 (b s, 1), 7.62 (dd, factors (C). The samples were prepared as described im-
1, J ) 2.6, 8.8), 8.72 (d, 1, J ) 6.3), 9.22 (s, 1). 13C NMR (101 mediately above. Samples were injected at 40% MeOH in H2O
MHz, DMSO-d6): δ 32.7, 50.8, 53.7, 114.9, 124.5, 127.5, 128.7, on an analytical C-18 reversed-phase HPLC column, ramp 40-
129.5, 130.2, 130.7, 131.8, 139.2, 155.9, 166.2, 169.3, 169.9. 100% MeOH in H2O over 40 min, UV detection at 252 nm.
HRMS (FAB+) m/e: 374.0910 (MH+ C18H17N3O4Cl requires However, an isocratic solvent system of 70% MeOH in H2O
374.0908). was used in order to minimize slight chromophore variation
compared with entry 13, Tables 3 and 4. The correction factors
Benzodiazepine (Table 8, Entry 8). IR (KBr): 3256,
of these benzodiazepines were used interchangeably in yield
2883, 1669 (s), 1560, 1488, 1438, 1351, 1244, 1190, 1101, 1042,
calculations since they only differ in the length of the C-8 alkyl
932, 812 cm-1. 1H NMR (400 MHz, DMSO-d6): δ 2.94 (dd, 1,
group.
J ) 8.9, 14.0), 3.15 (dd, 1, J ) 5.2, 14.0), 4.04-4.09 (m, 1),
4.81, (d, 1, J ) 15.6), 5.27 (d, 1, J ) 15.5), 5.94 (b s, 2), 6.54 (d, Acknowledgment. The NIH (GM50353) and Bur-
1, J ) 7.7), 6.59 (s, 1), 6.75, (d, 1, J ) 7.8), 7.17 (t, 1, J ) 7.1), roughs Wellcome Fund provided support for this re-
7.23 (t, 2, J ) 7.1), 7.31 (d, 2, J ) 7.1), 7.51-7.53 (m, 2), 7.58 search. J.A.E. is an Alfred E. Sloan Foundation Fellow
(d, 1, J, 8.5), 8.90 (d, 1, J ) 6.3). 13C NMR (101 MHz, DMSO-
and Eli Lilly Grantee. K.M.B. also gratefully acknowl-
d6): δ 33.6, 49.1, 53.6, 101.0, 107.3, 108.2, 120.4, 124.9, 126.4,
128.1, 128.8, 129.4, 129.8, 130.7, 131.5, 131.8, 137.7, 137.9, edges the Jean Dreyfus Boissevain Undergraduate
146.3, 147.4, 166.2, 169.9. HRMS (FAB+) m/e: 434.1038 (M+ Scholarship for Excellence in Chemistry, sponsored by
C24H19N2O4Cl requires 434.1033). the Camille and Henry Dreyfus Foundation. Reaction
Benzodiazepine (Table 8, Entry 19). IR (KBr): 3060, database searches were performed with ISIS (from
2931, 1664, 1432, 1123, 826 cm-1. 1H NMR (400 MHz, DMSO- MDL, Inc., 14600 Catalina St., San Leandro, CA 94577).
d6): δ 3.05 (dd, 1, J ) 8.74, 14.2), 3.28-3.31 (m, 4), 4.05-4.10 Supporting Information Available: 1H NMR spectra for
(m, 1), 7.44-7.46 (m, 4), 7.53-7.56 (m, 1), 7.63 (dd, 1, J ) 3.3, library members (3 pages). This material is contained in
8.8), 7.76-7.83 (m, 4), 8.88 (d, 1, J ) 6.3). 13C NMR (126 MHz, libraries on microfiche, immediately follows this article in the
DMSO-d6): δ 33.9, 35.0, 53.6, 124.2, 125.5, 126.0, 127.3, 127.4, microfilm version of the journal, and can be ordered from the
127.6, 127.7, 127.8, 128.7, 129.3, 130.2, 131.7, 131.8, 132.8, ACS; see any current masthead page for ordering information.
135.4, 139.6, 166.3, 170.1. HRMS (FAB+) m/e: 365.1049
(MH+ C21H18N2O2Cl requires 365.1057). JO9622845