Solution phase combinatorial chemistry. Discovery of novel polyazapyridinophanes with potent antibacterial activity by a solution phase simultaneous addition of functionalities approach

Article abstract of DOI:10.1021/ja964153r

Chemical modification of pre-formed asymmetric polyazaphane scaffolds by simultaneous addition of functionality (letters) in solution has been developed for the preparation of tertiary nitrogen-based combinatorial chemistry libraries. This approach has so

Full text of DOI:10.1021/ja964153r

3696
J. Am. Chem. Soc. 1997, 119, 3696-3708
Solution Phase Combinatorial Chemistry. Discovery of Novel
Polyazapyridinophanes with Potent Antibacterial Activity by a
Solution Phase Simultaneous Addition of Functionalities
Approach
Haoyun An, Lendell L. Cummins, Richard H. Griffey, Ramesh Bharadwaj,
Becky D. Haly, Allister S. Fraser, Laura Wilson-Lingardo, Lisa M. Risen,
Jacqueline R. Wyatt, and P. Dan Cook*
Contribution from the Department of Medicinal Chemistry, Isis Pharmaceuticals, Inc.,
2292 Faraday AVenue, Carlsbad, California 92008
ReceiVed December 2, 1996X
Abstract: Chemical modification of pre-formed asymmetric polyazaphane scaffolds by simultaneous addition of
functionality (letters) in solution has been developed for the preparation of tertiary nitrogen-based combinatorial
chemistry libraries. This approach has some significant advantages over the more commonly employed solid phase
bead splitting/reaction/mixing procedures for the preparation of libraries. Three novel, asymmetric polyazaphanes
32, 33, and 37 have been synthesized in high yields by an efficient cyclization of 2,6-bis(bromomethyl)pyridine (31)
with new orthogonally protected triamines 29, 30, and 35, respectively. Selective deprotection of 32, 33, and 37
provided mono-t-Boc-protected scaffolds 1-3 suitable for solution phase, simultaneous addition of functionalities.
Model studies of small libraries of scaffold 2 using CZE analyses indicated that simultaneous addition of 10 benzylic
bromide alkylating functionalities would result in libraries containing approximately equimolar amounts of all possible
compounds. Sixteen purified tertiary amine libraries 4-19 (total complexity of 1600 compounds) were generated
by this procedure from scaffold 2. A “fix-last” combinatorial method was devised to minimize chemical reactions.
Several first-round sublibraries of scaffold 2, containing a mixture of 100 compounds, exhibited potent antimicrobial
activities. Twenty single compounds 63-82 with uniform functionalities at the combinatorialized sites were
synthesized. Some of these pure compounds were more active, while others were less active, compared with the
parent mixtures 5 and 10.
Introduction
otides and carbopeptoids,6 oligoureas,7 oligonucleotides,8 and
oligosaccharides.9 More recent efforts are directed to chemical
Combinatorial chemistry can rapidly generate large numbers
of diverse compounds which, when combined with high-
throughput screening techniques, provide a very important
strategy for generating drug leads as well as for optimizing lead
compounds.1 Initial efforts have focused primarily on chemical
libraries of oligomeric materials such as peptides,2 peptoids,3
oligocarbamates,4 vinylogous sulfonyl peptides,5 carbonucle-
(3) (a) Simon, R. J.; Kania, R. S.; Zuckermann, R. N.; Huebner, V. D.;
Jewll, D. A.; Banville, S.; Ng, S.; Wang, L.; Rosenberg, S.; Marlowe, C.
K,; Spellmeyer, D. C.; Tan, R.; Frankel, A. D.; Santi, D V.; Cohen, F. E.;
Bartlett, P. A. Proc. Natl. Acad. Sci. U.S.A. 1992, 89, 9367. (b) Zuckermann,
R. N.; Martin, E. J.; Spellmeyer, D. C.; Stauber, G. B.; Shoemaker, K. R.;
Kerr, J. M.; Figliozzi, G. M.; Goff, D. A.; Siani, M. A.; Simon, R. J.;
Banville, S. C.; Brown, E. G.; Wang, L.; Richter, L. S.; Moos, W. H. J.
Med. Chem. 1994, 37, 2678.
(4) (a) Hsieh-Wilson, L. C.; Xiang, X-D; Schultz, P. G. Acc. Chem. Res.
1996, 29, 164. (b) Paikoff, S. J.; Wilson, T. E.; Cho, C. Y., Schulz, P. G.
Tetrahedron Lett. 1996, 37, 5653. (c) Cho, C. Y.; Moran, E. J.; Cherry, S.
R.; Stephans, J. C.; Fodor, S. P. A.; Adams, C. L.; Sundaram, A.; Jacobs,
J. W.; Schultz, P. G. Science 1993, 261, 1303.
(5) Gennari, C.; Nestler, H. P.; Sallom, B.; Still, W. C. Angew. Chem.,
Int. Ed. Engl. 1995, 34, 1763.
(6) (a) Burgess, K.; Linthicum, D. S.; Shin, H. Angew. Chem., Int. Ed.
Engl. 1995, 34, 907. (b) Kim, J.-M.; Bi. Y.; Paikoff, S. J.; Schultz, P. G.
Tetrahedron Lett. 1996, 37, 5305. (c) Kim, J.-M.; Wilson, T. C.; Norman,
T. C.; Schultz, P. G. Tetrahedron Lett. 1996, 37, 5309.
(7) Nicolaou, K. C.; Flo¨rke, H.; Egan, M. G.; Barth, T.; Estevez, V. A.
Tetrahedron Lett. 1995, 36, 1775.
(8) (a) Ecker, D. J.; Vickers, T. A.; Hanecak, R.; Driver, V.; Anderson,
K. Nucleic Acids Res. 1993, 21, 1853. (b) Wyatt, J. R.; Vickers, T. A.;
Roberson, J. L.; Buckheit, R. W., Jr.; Klimkait, T.; DeBaets, E.; Davis, P.
E.; Rayner, B.; Imbach, J. L.; Ecker, D. J. Proc. Natl. Acad. Sci. U.S.A.
1994, 91, 1356. (c) Hebert, N.; Davis, P. W.; DeBaets, E. L.; Acevedo, O.
L. Tetrahedron Lett. 1994, 35, 9509. (d) Hebert, N. In Molecular DiVersity
and Combinatorial Chemistry; Chaiken, I. M., Janda, K. D., Eds.; American
Chemical Society: Washington, DC, 1996; p 40. (e) Acevedo, O. L.;
Andrews, R. S. Tetrahedron Lett. 1996, 37, 3931. (f) Davis, P. W.; Vickers,
T. A.; Wilson-Lingardo, L.; Wyatt, J. R.; Guinosso, C. J.; Sanghvi, Y. S.;
DeBaets, E. A.; Acevedo, O. L.; Cook, P. D.; Ecker, D. J. J. Med. Chem.
1995, 38, 4363. (g) Fathi, R; Rudolph, M. J.; Gentles, R. G.; Patel, R.;
MacMillan, E. W.; Reitman, M. S.; Pelham, D.; Cook, A. F. J. Org. Chem.
1996, 61, 5600.
X Abstract published in AdVance ACS Abstracts, April 1, 1997.
(1) For recent reviews, see: (a) Still, W. C. Acc. Chem. Res. 1996, 29,
155. (b) Ohlmeyer, M. H. J.; Swanson, R. N.; Dillard, L. W.; Reader, J.
C.; Asouline, G.; Kobayashi, R.; Wigler, M.; Still, W. C. Proc. Natl. Acad.
Sci. U.S.A. 1993, 90, 10922. (c) Thompson, L. A.; Ellman, J. A. Chem.
ReV. 1996, 96 555. (d) Terrett, N. K.; Gardner, M.; Gordon, D. W.;
Kobylecki, R. J.; Steele, J. Tetrahedron 1995, 51, 8135. (e) Gordon, E.
M.; Barrett, R. W.; Dower, W. J.; Fodor, S. P. A.; Gallop, M. A. J. Med.
Chem. 1994, 37, 1385. (f) Gallop, M. A.; Barrett, R. W.; Dower, W. J.;
Fodor, S. P. A.; Gordon, E. M. J. Med. Chem. 1994, 37, 1233. (g) Holmes,
C. P.; Chinn, J. P.; Look, G. C.; Gordon, E. M.; Gallop, M. A. J. Org.
Chem. 1995, 60, 7328. (h) Lowe, G. Chem. Soc. ReV. 1995, 309. (i) For
recent highlights, see the following Special Issue: Acc. Chem. Res. 1996,
29, No. 3.
(2) (a) Houghten, R. A. Proc. Natl. Acad. Sci. U.S.A. 1985, 82, 5131.
(b) Houghten, R. A.; Pinilla, C.; Blondelle, S. E.; Appel, J. R.; Dooley, C.
T.; Cuervo, J. H. Nature 1991, 354, 84. (c) Blondelle, S. E.; Houghten, R.
A. Trends Biotechnol. 1996, 14, 60. (d) Lam, K. S.; Salmon, S. E.; Hersh,
E. M.; Hruby, V. J.; Kazmierski, W. M.; Knapp, R. J. Nature 1991, 354,
82. (e) Lam, K. S.; Hruby, V. J.; Lebl, M.; Knapp, R, J.; Kazmierski, W.
M.; Hersh, E. M.; Salmon, S. E. Bioorg. Med. Chem. Lett. 1993, 3, 419. (f)
Sepetov, N. F.; Krchnak, V.; Stankova, M.; Wade, S.; Lam, K. S.; Lebl,
M. Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 5426. (g) Cheng, Y.; Suenaga,
T.; Still, W. C. J. Am. Chem. Soc. 1996, 118, 1813. (h) Burger, M. T.;
Still, W. C. J. Org. Chem. 1995, 60, 7382. (i) Carrasco, M. R.; Still, W. C.
Chem. Biol. 1995, 2, 205.
S0002-7863(96)04153-4 CCC: $14.00 © 1997 American Chemical Society

Solution Phase Combinatorial Chemistry
J. Am. Chem. Soc., Vol. 119, No. 16, 1997 3697
libraries of small organic molecules10 generated by stepwise
additions (oligomerizations)11 or simultaneous addition of
reactants (multiple-component condensations, MCCs), utilizing
solid phase methods (solid phase organic synthesis, SPOS).12
Although solution phase combinatorial strategies for the prepa-
ration of chemical libraries remains a relatively unexplored
area,13 they may offer significant advantages over solid phase
synthesis.13ef,14
We have developed an approach to produce combinatorial
libraries by chemically modifying pre-formed asymmetric
scaffolds by simultaneous addition of functionalities (letters)
in solution. The advantages of a solution phase, simultaneous
addition of functionalities (SPSAF) approach over the more
often employed solid phase, bead splitting/reacting/mixing
procedures for the preparation of libraries are several-fold. (1)
Solution phase synthesis allows the opportunity for purification
and analysis of a combinatorial chemistry process at any reaction
steps, if needed. Thus, the high-yielding reactions usually
required on solid support are not absolute requirements for
reactions in solution. This should allow a broader scope of
combinatorial organic synthesis chemistry to be performed.
Searching an expanded scope of chemistry increases the
possibility of developing reaction conditions which will allow
simultaneous addition of functionality (letters).15 Thus, the usual
laborious bead-splitting procedures to prepare equimolar mix-
tures are avoided. (2) Solution phase syntheses will also allow
libraries to be prepared in greater quantities. Sufficient quanti-
ties of material that may be derived from solution synthesis will
allow testing the libraries in a broader range of assays for many
years. This is of particular importance in devising a serial
conversion of a library having a specific scaffold (footprint) to
one or more additional libraries with very different scaffolds
(footprints). (3) The ability to rapidly and continually change
the scaffold (footprint) is best accomplished by traditional
organic synthesis in solution rather than synthesis on solid
supports. As the scaffold has a direct bearing on the structure
space searched, the ability to readily change the scaffold is very
important. (4) A very significant synthesis advantage will be
realized by SPSAF for library synthesis compared to conven-
tional solid phase, bead splitting reaction/mixing procedures.
SPSAF for the preparation of chemical libraries may not be as
dependent on robotics as are solid support procedures. The
chemistry required to prepare libraries by a SPSAF approach
consists of the design and synthesis of appropriate scaffolds
and subsequent combinatorial reactions.
The selection of scaffolds to combinatorialize in this approach
presents a vast number of choices based on known pharma-
cophores (structure-based libraries) and unknown structures
(non-structure-based libraries). Important considerations for
selecting scaffolds include (1) ease of synthesis, (2) level of
conformational constraint, (3) minimization of the molecular
weight of repeating units as well as of the overall molecular
weight, (4) asymmetry of scaffolds, (5) number of positions to
be combinatorialized, (6) achirality or controlled chirality, (7)
selection of a combinatorial position that is orthogonally
protected, subsequently deprotected, and fixed last with func-
tionalities (fixed last concept) for an iterative screening process,
(8) versatility to readily change the size and shape of the scaffold
(footprint), and (9) options for chemical conversion of libraries
to libraries.16 We have identified a number of classes of
compounds which “fit” these scaffold selection considerations.
In particular, secondary nitrogens in a ring, which on substitution
will provide a combinatorialized tertiary nitrogen, seem exquis-
itely suited for scaffolds in combinatorial chemistry libraries;
alkylation and acylations of secondary amines presents a general,
simple, and high-yielding combinatorial chemistry.
Rebek and co-workers13a-d utilized tetracarboxylic acid
chloride substituted xanthene and cubane as scaffolds to prepare
amide libraries by simultaneously adding a mixture of amino
acids and employed a reverse iterative deconvolution procedure
to find active compounds in a relative short time frame. The
high symmetry of these scaffolds has considerable impact on
the nature of the resulting libraries. The number of distinct
compounds (complexity) in the libraries is greatly reduced. For
example, because of the two-fold symmetry of the xanthene
scaffold, four sites combinatorialized with 19 amino acids should
theoretically provide 65 341 distinct compounds assuming
similar chemical reactivity of each amino acid. Millions of
compounds (419) would be expected if the scaffold was
asymmetric. As the trend is to prepare less complex libraries,
this approach appears attractive. However, and surely more
important, is the fact that symmetry in the xanthene and, thus,
equivalency of the two positions results in distorted libraries.
Library compounds possessing any one of the 19 amino acids
two, three, or four times are obtained in a lower concentration
than the more heterogeneous compounds. McDevitt and
Lansbury13e employed a 1,3,5-benzenetricarbonyl trichloride
scaffold for the preparation of a library of glycotides. In this
example three glycosamino acids were added simultaneously
(9) (a) Rademann, J.; Schmidt, R. R. Tetrahedron Lett. 1996, 37, 3989.
(b) Danishefsky, S. J.; McClure, K. F.; Randolph, J. T.; Ruggeri, R. B.;
Science 1993, 260, 1307. (c) Yan, L.; Taylor, C. M.; Goodnow, R., Jr.;
Kahne, D. J. Am. Chem. Soc. 1994, 116, 6953.
(10) The terminology “analogous organic synthesis” has been used to
set small molecule diversity apart from peptide/oligosaccharide/nucleotide
diversity. Chen, C.; Randall, L. A. A.; Miller, R. B.; Jones, A. D.; Kurth,
M. J. J. Am. Chem. Soc. 1994, 116, 2661.
(11) For recent articles, see: (a) Hermkens, P. H. H.; Ottenheijm, H. C.
J.; Rees, D. Tetrahedron 1996, 52, 4527. (b) DeWitt, S. H.; Czarnik, A.
W. Acc. Chem. Res. 1996, 29, 114. (c) Ellman, J. A. Acc. Chem. Res. 1996,
29, 132. (d) Gordon, E. M.; Gallop, M. A.; Patel, D. V. Acc. Chem. Res.
1996, 29, 144. (e) Kobayashi, S.; Hachiya, I.; Suzuki, S.; Moriwaki, M.
Tetrahedron Lett. 1996, 37, 2809. (f) Bonnat, M.; Bradley, M.; J. D. Kilburn.
Tetrahedron Lett. 1996, 37, 5409. (g) Kobayashi, S.; Hachiya, I.; Yasuda,
M. Tetrahedron Lett. 1996, 37, 5569. (h) Marquais, S.; Arlt, M. Tetrahedron
Lett. 1996, 37, 5491. (i) Bolton, G. L. Tetrahedron Lett. 1996, 37, 3433.
(j) Brown, S. D.; Armstrong, R. W. J. Am. Chem. Soc. 1996, 118, 6331.
(k) O’Donnell, M. J.; Zhou, C.; Scott, W. L. J. Am. Chem. Soc. 1996, 118,
6070. (l) Hutchins, S. M.; Chapman, K. T. Tetrahedron Lett. 1995, 36,
2583.
(12) (a) Armstrong, R. W.; Combs, A. P.; Tempest, P. A.; Brown, S.
D.; Keating, T. A. Acc. Chem. Res. 1996, 29, 123 and referencs cited
therein. (b) Keating, T. A.; Armstrong, R. W. J. Am. Chem. Soc. 1996,
118, 2574. (c) Keating, T. A.; Armstrong, R. W. J. Am. Chem. Soc. 1995,
117, 7842. (d) Goebel, M.; Ugi, I. Tetrahedron Lett. 1995, 36, 6043.
(e) Wipf, P.; Cunningham, A. Tetrahedron Lett. 1995, 36, 7819. (f) Hamper,
B. C.; Dukesherer, D. R.; South, M. S. Tetrahedron Lett. 1996, 37,
3671.
(13) (a) Carell, T.; Wintner, E. A.; Bashir-Hashemi, A.; Rebek, J., Jr.
Angew. Chem., Int. Ed. Engl. 1994, 33, 2059. (b) Carell, T.; Wintner, A.;
Rebek, J., Jr. Angew. Chem., Int. Ed. Engl. 1994, 33, 2061. (c) Carell, T.;
Wintner, E. A.; Sutherland, A. J.; Rebek, J., Jr.; Dunayevskiy, Y. M.;
Vouros, P. Chem. Biol. 1995, 2, 171. (d) Shipps, G. W., Jr.; Spitz. U. P.;
Rebek, J., Jr. Bioorg. Med. Chem. 1996, 4, 655. (e) McDevitt, J. P.;
Lansbury, P. T., Jr. J. Am. Chem. Soc. 1996, 118, 3818. (f) Boger, D. L.;
Tarby, C. M.; Myers, P. L.; Caporale, L. H. J. Am. Chem. Soc. 1996, 118,
2109. (g) Cheng, S.; Comer, D. D.; Williams, J. P.; Myers, P. L.; Boger,
D. L. J. Am. Chem. Soc. 1996, 118, 2567. (h) Cheng, S.; Tarby, C. M.;
Comer, D. D.; Williams, J. P.; Caporale, L. H.; Myers, P. L.; Boger, D. L.
Bioorg. Med. Chem. 1996, 4, 727. (i) Pirrung, M. C.; Chen, J. J. Am. Chem.
Soc. 1995, 117, 1240. (j) Smith, P. W.; Lai, J. Y. Q.; Whittington, A. R.;
Cox, B.; Houston, J. G.; Stylli, C. H.; Banks, M. N.; Tiller, P. R. Bioorg.
Med. Chem. Lett. 1994, 4, 2821. (k) Nielsen, J.; Rusmussen, P. H.
Tetrahedron Lett. 1996, 37, 3351. (l) Hanessian, S.; Yang, R-Y. Tetrahedron
Lett. 1996, 37, 5835. (m) An, H.; Cook, P. D. Tetrahedron Lett. 1996, 37,
7233.
(14) (a) Curran, D. P. ChemTracts: Org. Chem. 1996, 9, 75. (b) Curran,
D. P.; Hoshino, M. J. Org. Chem. 1996, 61, 6480.
(15) Ostresh, J. M.; Winkle, J. H.; Hamashim, V. T.; Houghten, R. A.
Biopolymer 1994, 34, 1681.
(16) Ostresh, J. M.; Husar, G. M.; Blondelle, S. E.; Do¨rner, B.; Weber,
P. A.; Houghten, R. A. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 11138.

3698 J. Am. Chem. Soc., Vol. 119, No. 16, 1997
An et al.
Subsequently, 16 purified tertiary amine libraries 4-19 (total
complexity of 1600 compounds) (Figure 1) were generated by
this procedure from scaffold 2. Libraries were monitored by
TLC, purified by chromatographic techniques, and characterized
by 1H NMR and ESI MS spectral data. A “fix-last” combina-
torial method was devised to minimize chemical reactions and
to allow SAR studies to be performed on the mixtures. Several
first-round sublibraries of scaffold 2, each containing a mixture
of 100 compounds, exhibited potent antimicrobial activities.
Twenty single compounds from the active sublibraries were
synthesized; five of them exhibited potent antimicrobial activi-
ties.
Results and Discussion
We elected to examine conformationally constrained, asym-
metric polyazaphane scaffolds such as 1-3 for our initial
development work on a solution phase, simultaneous addition
of functionalities (SPSAF) approach. Polyazaphane scaffolds
1-3 are relatively small (average molecular weight 336) and
have three sites to combinatorialize. These types of scaffolds
are non-structure-based, that is, they are not modeled after a
known pharmacophore and thus libraries derived from them will
allow the search of novel structure space. Biological activities
discovered from non-structure-based scaffolds are more likely
to result from novel modes of actions rather than activity that
may be found from examining known pharmacophores.18 We
believe that, when considering library mixtures and deconvo-
lution processes, the minimization of “suboptimal binders” 19
(low-energy conformers) will facilitate discovery of active
compounds. Fewer suboptimal binders are expected from
examination of conformationally restricted molecules in a library
compared to a library of more flexible molecules in which a
composite activity may result from many suboptimal binders.20
Figure 1. Novel unsymmetric polyazaphane scaffolds 1-3 and libraries
4-19.
to provide, without purification, nine out of 10 expected
compounds. Again, symmetry of the scaffold reduced the
complexity (33 ) 27 to 10) and distorted the library.
In considering the chemical combinatorialization process, we
have chosen reactivity and reaction conditions to allow alkyla-
tions to go to completion in several hours. This should enhance
the likelihood of a near 1:1 ratio of electrophiles (functionalities,
letters) to nucleophiles (nucleophilic sites on the scaffolds)
which should minimize purification requirements and should
tend to minimize differential reactivity of nucleophiles on the
scaffold. A set of functionalities will require approximately
equal reactivity of each electrophile to provide near-equimolar
amounts of individual compounds in the pool. Nucleophilic
reaction at one site on the scaffold should not prevent near-
equal reactivity at other nucleophilic sites on the scaffold. This
conditions is also required in order to provide near-equimolar
amounts of individual compounds in the pool. We have
employed alkylations in the initial development work. Obvi-
ously, other electrophilic reactions, such as acylations, could
be utilized as well.
In this paper, we describe the design and synthesis in high
yields of three novel, asymmetric polyazaphanes17 32, 33, and
37 by an efficient cyclization of 2,6-bis(bromomethyl)pyridine
(31) with novel orthogonally protected triamines 29, 30, and
35, respectively. The key triamine intermediates 29, 30, and
35 were prepared from their corresponding mono-t-Boc-
protected triamines 26, 27, and 34, respectively. Selective
deprotection of 32, 33, and 37 by thiophenol provided mono-
t-Boc-protected scaffolds 1-3 suitable for solution phase,
simultaneous addition of functionalities. In model studies,
capillary zone electrophoresis (CZE) and electrospray ionization
mass spectrometry (ESI/MS) were utilized to determine product
formation in order to assure the relative chemical reactivities
of each electrophile with polyazaphane scaffold 2. These studies
indicated that simultaneous addition of 10 benzylic bromide
alkylating functionalities results in the libraries containing
approximately equimolar amounts of all possible compounds.
(18) For recent combinatorial libraries based on known pharmacophores,
see the following. Benzodiazepines: Bunin, B. A.; Ellman, J. A. J. Am.
Chem. Soc. 1992, 114, 10997. Bunin, B. A.; Plunkett, M. J.; Ellman, J. A.
Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 4708. Plunkett, M. J.; Ellman, J. A.
J. Am. Chem. Soc. 1995, 117, 3306. Hydantoins: DeWitt, S. H.; Kiely, J.
S.; Stankovic, C. J.; Schroeder, M. C.; Reynolds Cody, D. M.; Pavia, M.
R. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 6909. Dressman, B. A.; Spangle,
L. A.; Kaldor, S. W. Tetrahedron Lett. 1996, 37, 937. â-Lactams: Ruhland,
B.; Bhandari, A.; Gordon, E. M.; Gallop, M. A. J. Am. Chem. Soc. 1996,
118, 253. Piperazineidones: Gordon, D. W.; Steele, J. Bioorg. Med. Chem.
Lett. 1995, 5, 47. 1,4-Dihydropyridines: Gordeev, M. F.; Patel, D. V.;
Gordon, E. M. J. Org. Chem. 1996, 61, 924. Pyridines and pyrido-
[2,3]pyrimidines: Gordeev, M. F.; Patel, D. V.; Wu, J.; Gordon, E. M.
Tetrahedron Lett. 1996, 37, 4643. Quinolines: MacDonald, A. A.; DeWitt,
S. H.; Hogan, E. M.; Ramage, R. Tetrahedon Lett. 1996, 37, 4815. Buckman,
B. O.; Mohan, R. Tetrahedron Lett. 1996, 37, 4439. Smith, A. L.; Thomson,
C. G.; Leeson, P. D. Bioorg. Med. Chem. Lett. 1996, 6, 1483. Tetrahy-
droisoquinolines and tetrahydroimidazopyridines: Hutchins, S. M.; Chapman,
K. T. Tetrahedron Lett. 1996, 37, 4865. Pyrroles: Mjalli, A. M. M.; Sarshar,
S.; Baiga, T. J. Tetrahedron Lett. 1996, 37, 2943. Indoles: Hutchins, S.
M.; Chapman, K. T. Tetrahedron Lett. 1996, 37, 4869. Benzimidazoles:
Phillips, G. B.; Wei, G. P. Tetrahedron Lett. 1996, 37, 4887. Piperazines:
Dankwardt, S. M.; Newman, S. R.; Krstenansky, J. L. Tetrahedron Lett.
1995, 37, 4923. Kowalski, J.; Lipton, M. A. Tetrahedron Lett. 1996, 37,
5839. Phosphonates: Cao, X.; Mjalli, A. M. M. Tetrahedron Lett. 1996,
37, 6073. Zhang, C.; Mjalli, A. M. M. Tetrahedron Lett. 1996, 37, 5457.
Tropanes: Koh, J. S.; Ellman, J. A. J. Org. Chem. 1996, 61, 4494.
Tetrahydro-â-carbolines: Mayer, J. P.; Bankaitis-Davis, D.; Zhang, J.;
Beaton, G.; Bjergarde, K.; Andersen, C. M.; Goodman, B. A.; Herrera, C.
J. Tetrahedron Lett. 1996, 37, 5633. Yang, L.; Guo, L. Tetrahedron Lett.
1996, 37, 5041. Purines: Norman, T. C.; Gray, N. S.; Koh, J. T.; Schultz,
P. G. J. Am. Chem. Soc. 1996, 118, 7430. Dihydropyrimidinediones:
Kolodziej, S. A.; Hamper, B. C. Tetrahedron Lett. 1996, 37, 5277.
Rauwolfina alkaloids: Atuegbu, A.; Maclean, D.; Nguyen, C.; Gordon, E.
M.; Jacobs, J. W. Bioorg. Med. Chem. 1996, 4, 1097. Miconazole: Tortolani,
D. R.; Biller, S. A. Tetrahedron Lett. 1996, 37, 5687.
(17) Cyclophanes or phanes contain an aromatic nucleus (benzene or
arene) and an aliphatic bridge. Heterophanes contain heteroatoms in the
aromatic ring, while heteraphanes contain heteroatoms in the bridge. Our
compounds include one nitrogen atom in the aromatic ring (pyridine) and
three nitrogen atoms in the bridge. They are abbreviated as “polyazaphanes”.
For the definition of “phanes”, also see: Vo¨gtle F. Cyclophane Chem-
istrysSynthesis, Structures and Reactions; Vo¨gtle, F., Ed.; Jones, P. R.
Translated; John Wiley and Sons, Ltd.: Chichester, U.K., 1993; p 11.
(19) Freier, S. M.; Konings, D. A. M.; Wyatt, J. R.; Ecker, D. J. J. Med.
Chem. 1995, 38, 344.
(20) Geysen, H. M. Book of Abstracts, 209th ACS National Meeting,
Anaheim, CA, April 2, 1995; ORGN 25.

Solution Phase Combinatorial Chemistry
J. Am. Chem. Soc., Vol. 119, No. 16, 1997 3699
The oxyamine scaffold 3 is of interest because the two “benzyl”
secondary amines can be differentiated through their large
difference in basicity,21 and the N-O bond can be hydrolytically
cleaved to provide new sublibraries according to the libraries
from libraries concept.16 We were particularly interested in
avoiding the bead-splitting processes devised to provide equimo-
lar amounts of each compound in the library.22 Simultaneous
addition of functionalities would circumvent this approach and
would facilitate more rapid library production. Also important
in the selection of a scaffold is the concept of placing the
functionality that differentiates each pool last in the synthetic
scheme (fix last concept).23 This procedure greatly reduces the
required chemistry by diverging at the end of a scheme of steps
rather than diverging early in the scheme and repeating the same
synthetic sequence for each functionality. This is accomplished
in the present work by an orthogonal protection protocol
allowing the fix-last position to remain blocked during the
simultaneous addition of functionalities chemistry.
Scheme 1
The scaffolds were synthesized (Schemes 1 and 2) by a novel
cyclization of 2,6-bis(bromomethyl)pyridine (31) with fully
protected triamines 29, 30, and 35. This high-yielding, efficient
polyazaphane synthesis is very versatile in that the polyamine
synthon can readily be changed to provide a variety of secondary
amines and oxyamines in various chain lengths which will afford
macrocycles of various ring sizes. A great number of mono-
and polycyclic heterocycle platforms are available or could be
designed. The cyclization reaction provides orthogonal protec-
tion of the benzylamines. This will allow a sequential depro-
tection-reaction process to combinatorialize 1-3 in a single-
compound mode, or provide one position available for
combinatorialization with the others orthogonally protected for
sequential processing (fix first), or provide one of the three
positions protected for addition of functionalities simultaneously
to prepare mixtures with the other two available positions (fix
last). In the later case, the protected position would be liberated
(21) The Determination of Ionization Constants, A Laboratory Mannual,
3rd ed.; Albert, A., Serjeant, E. P., Eds.; Champman and Hall: New York,
1984; p 151.
(22) Furka, A.; Sebestyen, F.; Asgedom, M.; Dibo, G. Int. Pept. Protein
Res. 1991, 37, 487.
(23) For example, consider a polyazaphane scaffold with four sites to
combinatorialize and with one “fixed” or selectively protected site. In the
simplest situation (one step for deprotection), 48 reactions are required to
completely deconvolute (find the single active compound) for a fix-last
process. This consists of (1) a simultaneous addition reaction step using all
functional groups, (2) a deprotecting step, (3) a division of the material
into 10 pools (assuming 10 functionalities), and (4) reaction of each pool
with an individual functionality at the fixed position (the pools are all
screened). To do this for four combinatorial sites requires 48 reactions (4
× 12) assuming the scaffold is available as an orthogonally protected
species. If the same scaffold is deconvoluted by a fix-first process, 110
reactions are required. This requires (1) a deprotection step of the
orthogonally protected scaffold (note to start this approach requires the
synthesis of a fully orthogonally protected scaffold), (2) separation into 10
pools, (3) reaction of each pool with each functionality, and (4) repetition
of the entire process three times: thus, in total, (4 × 11) reactions (first
round) + (3 × 11) reactions (second round) + (2 × 11) reactions (third
round) + (1 × 11) reactions (fourth round) ) 110 reactions to completely
deconvolute by a fix-first process. The number of steps in the fix-first
process increases as the number of sites increases relative to the fix-last
process (e.g., for 10 functionalities and five sites: fix-first for first round
) 55 reactions Vs 12 for fix-last, and to completely deconvolute is 165
reactions Vs 60 reactions, and with six sites the first round is 66 Vs 12
reactions, and 231 Vs 72 reaction are needed to completely deconvolute).
Possibly more important is the likelihood that a high percentage of first-
round libraries will not be active in a particular assay, and thus subsequent
deconvolution is not required, a potential saving of many reaction steps. In
the case of fix-last, a completely orthogonally protected scaffold is not
initially required to obtain first-round libraries. In the case of a fix-first
process, considerable protection chemistry is required to start the process
since completely orthogonally protected scaffold is required and will be of
little value if activity is not found in the first round. Furthermore,
orthogonally protecting more than three sites can be synthetically chal-
lenging.
Scheme 2

3700 J. Am. Chem. Soc., Vol. 119, No. 16, 1997
An et al.
last and treated individually with the combinatorialization
functionalities (fix-last concept) for an iterative deconvolution
process. The amine/oxyamine scaffold 3 allows the use of the
concept of libraries from libraries16 in that hydrogenolytic
cleavage of the N-O bond of compounds in each sublibrary
would linearize polyazaphanes, yielding a series of new subli-
braries and greatly change the structure space (footprint)
searched.
Scheme 3
Various cyclization methods to prepare polyazaphanes have
been reported.24 Typically, N1,Nω-bis(amino) compounds pro-
tected with Cbz, COOEt, t-Boc, Tf, and Ts groups25 have been
employed in a cyclization step to prepare symmetric polyaza-
phane compounds. Most of these methods require strong basic
conditions at elevated temperature and suffer from low yields.
Deprotection of the resulting macrocycles are often low-yielding
reactions because of the requirements of strong acids or bases
or reducing conditions at elevated temperature to remove various
blocking groups. These conditions are not compatible with our
need to retain one nucleophilic site protected, such as t-Boc
group, for the fix-last concept and would cleave the N-O bond
in polyazaphane 3 as well as other oxyamine-containing
polyazaphanes. Therefore, it was necessary to find an efficient
cyclization method that also would be compatible with a suitable
combination of orthogonal protecting groups. We required a
selective deprotection scheme that would yield mono-protected
polyazaphane scaffolds for the SPSAF approach. We elected
to use the 2-nitrobenzenesulfonyl group26 for protection of the
primary amines of 26, 27, and 34. This process, which begins
with a t-Boc-protected secondary amine, provides fully protected
triamines 29, 30, and 35. Once cyclization has occurred,
removal of the sulfonamide protecting group by thiophenol is
readily accomplished without affecting the t-Boc group. The
t-Boc group is easily removed as required, and the secondary
amine can be fixed with a discriminating functionality at the
selected site (fix-last concept) for iterative screening.23
to provide two reactive secondary nitrogen groups for combi-
natorialization under mild conditions. The combination of
o-nitrosulfonate and t-Boc protecting groups on polyamines and
diamines provides a general synthetic approach to a variety of
polyazaphane scaffolds suitable for combinatorialization chem-
istry.
The 1:1 cyclizations of 2,6-bis(bromomethyl)pyridine (31)
with orthogonally protected key intermediates 29 and 30 were
carried out in DMF at room temperature for 24 h. Cesium
carbonate was used as a weak base with the Cs+ cation serving
as a cyclization template. The corresponding macrocyclic
compounds 32 and 33 were isolated as white foams in yields
of 80% and 73%, respectively. In comparison, the cyclizations
of polyamines protected by COOEt, Cbz, or t-Boc groups25
generally provide 20-50% yields. The 2-nitrobenzenesulfonyl
protecting groups in 32 and 33 were removed selectively by
thiophenol in the presence of K2CO3 to give the desired mono-
t-Boc-protected asymmetric polyazaphane scaffolds 1 and 2 in
94% and 72% yields, respectively. Notable features of this
cyclization and deprotection process were the remarkable
reaction efficiency, mild reaction conditions, and high yields.
This cyclization method and synthetic route should be useful
for the syntheses of other types of macrocyclic compounds for
combinatorials, molecular recognition, crown ether chemistry,
and other areas.
Schemes 1 and 2 illustrate the synthesis of fully protected,
asymmetric polyazaphane compounds 32, 33, and 37 and the
corresponding desired mono-protected polyazaphanes scaffolds
1-3. Spermidine (21), N-(2-aminoethyl)-1,3-propanediamine
(20), and N5-(tert-butoxycarbonyl)-1,7-diamino-1-oxa-5-aza-
heptane (34)27 were selected as initial polyamine synthons. They
are asymmetric and upon cyclization according to our procedure
provide novel, interesting 13-, 14-, and 15-membered polyaza-
phanes.17 The mono-t-Boc-protected spermidine 27 was pre-
pared from spermidine (21) through intermediates 23 and 25
by the reported procedures.28 Similar procedures were used to
prepare mono-t-Boc-protected triamine 26 from N-(2-amino-
ethyl)-1,3-propanediamine (20) through the corresponding in-
termediates 22 and 24. The mono-t-Boc-protected triamines
26 and 27 were reacted with 2-nitrobenzenesulfonyl chloride
(28) in dichloromethane using triethylamine as the base to afford
the corresponding triprotected triamines 29 and 30 in 67% and
91% yields, respectively. These key intermediates contain
orthogonal protecting groups that can be selectively removed
The same route was used for the synthesis of the polyaza-
phane scaffold 3 containing an N-O bond (Scheme 2). N5-
(tert-Butoxycarbonyl)-1,7-diamino-1-oxa-5-azaheptane (34)27
was treated with 28 under the same conditions as for the
preparation of 29 and 30. The desired triprotected triamine
product 35 was isolated in 33% yield, and the tetraprotected
triamine 36 also was isolated in 45% yield. The reaction
conditions have not been optimized for the synthesis of the
desired product 35. The cyclization of 35 with 31 under the
same conditions as described above gave the corresponding
triprotected polyazaphane compound 37 in 50% yield. The
lower yield compared to those of 32 and 33 can be attributed
to the different reactivities of two different nitrogenous nucleo-
philes CH2NH and ONH in the triamine 35. Fully protected
macrocycle 37 was selectively deprotected by thiophenol in the
presence of K2CO3, providing the desired polyazaphane scaffold
3 in 97% yield. Novel compounds 26, 29, 30, 32, 33, 35, 37,
1
and 1-3 were characterized by their H and 13C NMR data,
high-resolution (HR) MS spectroscopic techniques, and com-
1
(24) Bradshaw, J. S.; Krakowiak, K. E.; Izatt, R. M. Aza-Crown
Macrocycles; Taylor, E. C., Ed.; John Wiley and Sons, Inc.: New York,
1993.
(25) (a) Krakowiak, K. E.; Bradshaw, J. S. J. Heterocycl. Chem. 1995,
32, 1639. (b) Panetta, V.; Yaouanc, J. J.; Handel, H. Tetrahedron Lett. 1992,
33, 5505. (c) Pearson, D. P. J.; Leigh, S. J.; Sutherland, I. O. J. Chem.
Soc., Perkin Trans. 1 1979, 3113. (d) Hodgkinson, L. C.; Sutherland, I. O.
J. Chem. Soc., Perkin Trans. 1 1979, 1908.
(26) Fukuyama, T.; Jow, C.-K.; Cheung, M. Tetrahedron Lett. 1995, 36,
6373.
(27) Kung, P. P.; Fraser, A. S.; Guinosso, C. J.; Bharadwaj, R.; Cook,
P. D. Unpublished results.
bustion analyses. The structure of 36 was confirmed by H
NMR and ESI MS spectral data. The structure of 2 was further
confirmed by the preparation of its derivatives (Scheme 3).
Scaffold 2 was treated with trifluoroacetic acid (TFA) to give
the corresponding unprotected macrocyclic scaffold 38. The
reaction of 38 with benzyl bromide using K2CO3 as the base
gave the trisubstituted derivative 39. Compounds 38 and 39
were characterized by NMR and HRMS spectroscopic tech-
niques. The unprotected scaffold 38 can be used to generate
libraries directly, and a reverse iterative deconvolution process
similar to Rebek’s13a-c can be employed. However, the use of
(28) O’Sullivan, M. C.; Dalrymple, D. M. Tetrahedron Lett. 1995, 36,
3451.

Solution Phase Combinatorial Chemistry
J. Am. Chem. Soc., Vol. 119, No. 16, 1997 3701
Scheme 4
the corresponding mono-t-Boc-protected scaffold 2 as the
starting point allows the use of the “fix-last” strategy and a
subsequent iterative deconvolution.23
For the combinatorialization of scaffolds 1-3, by the
simultaneous addition of functionalities (electrophiles), an
approximately equal reactivity of each electrophile is required
to obtain undistorted libraries. We have selected meta-
substituted benzylic bromides as an initial class of electrophiles
to examine for the SPSAF approach. The differentiating
(diversity) groups are located in the meta position. These
electrophiles are expected to have approximately the same
reactivity because different substituents on the meta position
are expected to have a minimal contribution to the electrophi-
licity compared with substituents on the ortho and para
positions. The structural diversity demonstrated by meta-
substituted benzyl moieties would seem limited; however, these
type of electrophiles are attractive for initial model studies.
A set of 10 commercially available benzylic bromides R1-
Br, R2-Br ... R10-Br (see Figure 1 for structural details) were
selected as functionalities for simultaneous addition. The two
secondary, benzylic nitrogens on scaffold 2 are expected to have
approximately the same nucleophilicities. Therefore, the librar-
ies generated by simultaneous reaction of these electrophiles
with scaffold 2 are expected to contain all possible compounds
in approximately the same amounts. To confirm this hypothesis,
we performed model studies by synthesizing several small
libraries (Scheme 4) which could be carefully analyzed. An
equimolar amount of each electrophile R1-Br and R2-Br was
reacted simultaneously with scaffold 2 to give library 40
containing four compounds. The electrophile R1-Br was
chosen as the standard and combined separately with each of
the other electrophiles R3-Br, R4-Br, R5-Br, R6-Br, R7-
Br, R8-Br, R9-Br, and R10-Br to react with scaffold 2. In
this manner, nine different small libraries, 40-48, containing
four compounds each were obtained after chromatographic
purification.29 These pure libraries were confirmed by TLC,
1H NMR, and ESI MS spectral data30 (see the Experimental
Section). Libraries 40-48 were all analyzed by capillary zone
electrophoresis (CZE) techniques and detected by UV absorption
Figure 2. Representative capillary zone electophoresis profile for
libraries 45. CZE conditions: 20 mM ammonium formate buffer
containing 0.1% of acetic acid in methanol; detected at 214 nm. Area
percentages (peak number): 21.58% (1), 22.87% (2), 26.77% (3), and
28.78% (4) in the deviation of 7-15%. Height percentages (peak
number): 26.51% (1), 24.11% (2), 24.98% (3), and 24.38% (4).
at 214 nm. Figure 2 shows a representative CZE profile of
library 45. The four expected compounds were very well
separated with almost the same peak heights. The area
percentages of the four peaks are 21.58% (1), 22.87% (2),
26.77% (3), and 28.78% (4), respectively, with a deviation of
7-15%. This relative ratio is appropriate for our biological
screening purpose.31 Other small libraries gave similar CZE
profiles. These results indicated that R-bromo-m-xylene (R1-
Br) shows approximately the same reactivity as each of the other
electrophiles. Therefore, all electrophiles tested should have
approximately the same reactivity under our reaction conditions.
These data suggest that under these reaction conditions the meta
position of the benzyl group may be sufficiently isolated from
the benzylic reactive site such that a wide variety of more diverse
moieties could be placed in the meta position with the likelihood
of near-equal reactivities.
To substantiate the above conclusion, model libraries 49-
52 were prepared by the reaction of scaffold 2 with a set of
three, four, and five different electrophiles (Scheme 5) under
the same reaction, workup, and purification conditions. Librar-
1
ies 49-52 exhibited acceptable TLC, H NMR, and ESI MS
spectra (see the Experimental Section). Libraries 49 and 50
were analyzed with CZE, and the nine expected peaks were
well-separated. Figure 3 exhibits a CZE profile of library 49.
Area percentages of these peaks deviate 0.2-12% from the
expected values. CZE analysis of library 51 (supporting data)
(29) The column loaded with a crude library was eluted with a less polar
solvent to remove the excess benzylic bromides and then was eluted with
more polar solvents to completely remove the library. All fractions
containing library compounds were collected. The preparative thin layer
chromatography (PLC) loaded with crude library was developed by selected
solvents. The library band, which was generally wider than those of single
compounds, was collected to provide pure library without electrophiles and
other impurities.
(31) Our current requirement for the libraries is that the molar differences
of theoretical amounts and the detected amounts of all compounds in one
library are within 20%. The active compounds within this deviation would
not be missed using our antibacterial assays.
1
(30) The H NMR spectra of libraries show the correct ratio of certain
protons. The ESI mass spectra show the peaks ranging from the smallest
molecular weight to the largest molecular weight in libraries.

3702 J. Am. Chem. Soc., Vol. 119, No. 16, 1997
An et al.
Scheme 6
crude library was purified by flash chromatography on a silica
gel column to afford pure library 4 in 94% yield.32 Combina-
torialization of the two different nucleophilic sites on scaffold
2 with 10 functionalities resulted in library 4 containing 100
Figure 3. Capillary zone electrophoresis profile of library 49. CZE
conditions: 30 mM ammonium formate buffer containing 0.3% acetic
acid in methanol; detected at 214 nm. Area percentages (peak
number): 11.08% (1), 11.33% (2), 10.59% (3), 12.19% (4), 11.84%
(5), 10.96% (6), 10.94% (7), 11.33% (8), and 9.74% (9) in the deviation
of 0.2-12%.
1
compounds (102 ) 100). Library 4 was verified by H NMR
and ESI MS spectral data.30 Deprotection of library 4 with TFA
gave the corresponding library 5 in 93% yield after chromato-
graphic purification. The secondary amine in intermediate
library 5 was reacted sequentially with 10 different function-
alities Rx-Br (x ) 1-10) under similar conditions as described
above for the preparation of library 4. This afforded 10 different
sublibraries 6-15 (Scheme 6 and Figure 1) in 60-98% yields
after preparative thin layer chromatographic (PLC) purifica-
tion.29
Scheme 5
With the key intermediate library 5 containing one reactive
site in hand, initial structure activity relationship (SAR) studies
were performed on the mixture by introducing several different
functionalities at the fixed site (Scheme 7). Adding new
functionalities in this manner would “positionally bias” the
sublibraries in that the new groups are not combinatorialized at
the other positions. The reaction of 5 with methyl R-bromoace-
tate, bromoacetonitrile, and R-bromoacetamide under similar
reaction and purification conditions afforded the corresponding
libraries 16, 17, and 18, respectively, in 95-98% yields. The
commercially available 2-chloro-N-methoxy-N-methylacetamide
was reacted with 5 at 60-70 °C for 6 h and then at room
temperature for 10 h, giving library 19 in 93% yield. The PLC
conditions for the purification of libraries 18 and 19 differed
from the others because of the more polar functionalities. All
libraries described above were prepared in large quantities
(200-3000 mg), monitored by TLC, purified by chromato-
containing 16 compounds exhibits 12 well-resolved peaks and
one broad, later migrating peak representing incomplete separa-
tion of four compounds. The CZE analysis of library 52
containing 25 compounds indicated that 20 peaks out of 25 were
detected. Five of the 20 peaks represent the incomplete
separations of two compounds each (Supporting Information).
The above results further indicated that these electrophiles have
similar reactivities and that the reactivities of these electrophiles
exhibit little influence on the final libraries. Therefore, we were
confident that high-quality libraries containing approximately
equimolar amounts of all possible compounds would result by
simultaneous addition, in solution, of selected functionalities
under our reaction conditions.
1
graphic techniques, and confirmed by their H NMR and ESI
MS spectral data. The availability of key intermediate libraries
as well as of final sublibraries for broader screening and further
reactions is a major advantage of SPSAF process Vs the usual
bead-splitting, solid support synthesis.
A mixture containing equimolar amounts of 10 benzylic
bromides Rx-Br (x ) 1-10, total 2.4 equiv) was added to a
stirred mixture of scaffold 2 in acetonitrile using K2CO3 as the
base (Scheme 6). After an overnight reaction and workup, the
(32) The theoretical yield was calculated according to the average
molecular weight. The yield was obtained by comparing the amount of
isolated library with the theoretical yield.

Solution Phase Combinatorial Chemistry
J. Am. Chem. Soc., Vol. 119, No. 16, 1997 3703
Figure 4. Electrospray ionization mass spectrum of library 7. Molecular weights range from 518 to 676; 49 distinct peaks33 corresponding to 100
different compounds.
Scheme 7
ESI taken of the purified libraries. Figure 4 shows the
representative electrospray ionization mass spectrum of library
7. The (M + H)+ peak range 519-677 corresponds to the
molecular weight range 518-676 for 100 compounds in library
7. There were no obvious impurities beyond this range. Forty-
nine theoretical molecular weights33 out of 100 different
compounds were all represented by 49 distinct peaks. The
simulated electrospray mass spectrum (Figure 5) of library 7
was generated by individually calculating the formula, mass,
and isotope distribution for each library member. The masses
were tabulated as a seven-point fit to a Gaussian line shape with
a 0.15 Da width, and overlapping masses were summed prior
to storage in a flat file. A graphical representation of the mass
table was inspected visually, and the relative abundance of each
functionality was adjusted iteratively to fit the observed mass
spectrum. As shown in Figure 5, the mass distribution seen in
the computer-simulated spectrum with functionality weights
ranging from 0.88 to 1.05 is similar to the observed data in
Figure 4.
Initial screening results (Table 1) indicated that sublibraries
5, 9, and 10, containing 100 compounds each, exhibited
antimicrobial activities against E. coli imp- and S. pyogenes in
minimum inhibitory concentration (MIC) in the low micromolar
range. Sublibraries 17 and 19 also show activities. Other
sublibraries did not show activity at the concentration of 100
µM. The potent activities of sublibraries 5 and 10 prompted
us to make 20 single compounds of these libraries. Scaffold 2
The alkylation reactions, performed simultaneously under the
described conditions, completely consume the scaffold. There-
fore, the possible impurities that may result from these alkyla-
tion reactions may be (1) scaffold having an unreacted second-
ary amine (incomplete alkylations that would provide mono-
alkylation instead of dialkylation), (2) scaffold having a
quaternary nitrogen (overalkylation to yield three or four benzyl
substitutions), (3) scaffold with an unreacted secondary amine
and a quaternary amine, and (4) benzylic bromides or decom-
posed benzylic bromides. The initial aqueous extraction would
remove the ionic quarternary species and chromatography would
remove benzyl species. Mono-, tri-, and tetra-alkylated species
were not present in the initial ESI taken before extraction
procedures. Obviously, these species were not observed in the
(33) Library 7 has the following 49 theoretically different molecular
weights: 518, 532, 536, 543, 544, 546, 550, 552, 554, 557, 558, 561, 562,
563, 566, 568, 569, 570, 576, 577, 578, 581, 586, 588, 589, 590, 594, 597,
600, 601, 602, 604, 608, 610, 611, 612, 614, 620, 621, 622, 623, 631, 634,
641, 644, 654, 655, 664, 676.

3704 J. Am. Chem. Soc., Vol. 119, No. 16, 1997
An et al.
Figure 5. Computer-simulated mass spectrum of library 7.
Table 1. Activities of Libraries in Growth Inhibition Assays
(MIC, µM)a
Scheme 8
Library
Complexity
S. pyogenes
E. coli imp-
4
5
6
7
8
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
1
>100
5-10
>100
>100
>100
5-10
10-20
>100
>100
>100
>100
>100
>100
10-20
>100
50-100
>100
>100
>100
>100
10-20
>100
>100
>100
5-10
<2.5
>100
>100
>100
>100
>100
>100
10-20
>100
50-100
>100
>100
>100
9
10
11
12
13
14
15
16
17
18
19
2
38
39
1
1
a The MIC (minimum inhibitory concentration) value is given as a
range of library concentration (total concentration of compounds in
the libraries). After 24 h, the complete growth was observed at lower
bound of the given MIC and no growth was observed at the upper
bound.
was reacted with each of the 10 benzylic bromides to give
compounds 53-62 (Scheme 8). Deprotection of 53-62 with
TFA afforded products 63-72. Each of the compounds 63-
72 was reacted with cinnamyl bromide to afford 10 final
compounds 73-82. Compounds 63-72 and 73-82 belong to
sublibraries 5 and 10, respectively. Each of these compounds
has the same functionalities in the two combinatorialized
positions. Compounds 63-82 were screened against imp- and
S. pyogenes, and the MIC results were listed in Table 2. Six
active compounds 69, 71-73, 77, and 81 were further tested
against S. aureus, E. coli, K. pneumoniae, and E. faecalis.
Compounds 69, 71, and 72 showed low MIC values in most of
the assays, while compounds 72, 76, and 81 showed high
selectivity among these assays. Scaffold 2 and compound 38
(Table 1) did not show activity at 100 µM which suggests that
some benzylic groups on polyazaphane 38 are essential for
activity. Tribenzyl-subsituted compound 39 and a series of other
compounds in Table 2 did not show activities. These results
suggest that, even though the diversity of substituted benzylic

Solution Phase Combinatorial Chemistry
J. Am. Chem. Soc., Vol. 119, No. 16, 1997 3705
Table 2. Activities of Compounds 63-82 in Growth Inhibition
Assays (MIC, µM)a
the presence of library or relevant antibiotic or antifungal controls.34
Growth was monitored as a function of time by measuring absorbance
at 595 nm. The S. pyogenes strain was ATCC #14289 and was grown
in 1x Todd-Hewitt broth. The E. coli imp- strain was a kind gift of
Spencer Bensen and was grown in 1/2x LB.35 For tier two screening,
S. pyogenes (ATCC#49399), S. aureus (ATCC# 25923), E. coli
(ATCC# 11775), K. pneumoniae (ATCC# 13883), and E. faecalis
(ATCC# 29212) strains were grown in the standard media. A computer
program has been developed to generate simulated mass spectra of
combinatorial libraries from input molecular formulas and relative
abundances for the scaffold and the functionalities. The nominal
molecular mass and isotope distributions (N + 1 and N + 2 for 13C,
15N, and 2H; all for Cl, Br, and S) are calculated for each library member
and converted to a seven-point fit to a Gaussian line shape with a line
width specified as input. The program has been written and compiled
in C using a Silicon Graphics Crimson computer.
N1,N6-Bis(trifluoroacetyl)-1,6-diamino-3-azahexane (22). Ethyl
trifluoroacetate (16.7 mL, 19.89 g, 0.14 mol, 3.5 equiv) was added to
a solution of N-(2-aminoethyl)-1,3-propanediamine (20) (4.70 g, 40
mmol) in 25 mL of CH3CN followed by H2O (0.72 g, 40 mmol, 1
equiv). The resulting reaction mixture was refluxed overnight, and
the solvent was evaporated under reduced pressure. The residue was
dried under high vacuum to give product 22 as a white solid: mp 131-
133 °C, yield 16.8 g (99.4%); silica gel TLC Rf 0.45 (100% EtOH);
1H NMR (CD3OD + D2O) δ 1.88-2.05 (m, 2 H), 3.08 (t, 2 H, J )
7.6 Hz), 3.23 (t, 2 H, J ) 6.0 Hz), 3.40 (t, 2 H, J ) 6.8 Hz), 3.63 (t,
2 H, J ) 6.0 Hz) (product thus obtained was carried out to the next
step without further purification).
N3-(tert-Butoxycarbonyl)-N1,N6-bis(trifluoroacetyl)-1,6-diamino-
3-azahexane (24). Di-tert-butyl dicarbonate (13.2 g, 60 mmol, 1.5
equiv) was added to a cooled solution of 22 (16.8 g, 40 mmol) and
Et3N (10.9 g, 108 mmol, 2.7 equiv) in 140 mL of anhydrous THF.
The resulting reaction mixture was stirred at room temperature
overnight. Saturated NH4Cl solution (270 mL) was added, and the
resulting mixture was extracted with CHCl3. The CHCl3 extract was
washed with brine and dried (Na2SO4). The solvent was evaporated,
and the residue was purified by flash chromatography on a silica gel
column (17 × 7 cm). Elution with 100% CH2Cl2 and then 20:1 CH2-
Cl2-MeOH afforded product 24 as a white solid: mp 113-114 °C,
yield 16.4 g (100%); silica gel TLC Rf 0.37 (50:1 CH2Cl2-MeOH);
1H NMR (CDCl3) δ 1.49 (s, 9 H), 1.65-1.88 (m, 2 H), 3.20-3.37 (m,
4 H), 3.40-3.55 (m, 4 H), 7.00 (bs, 1 H), 8.10 (bs, 1 H).
compd E. coli
no.
S.
S.
E.
K.
E.
imp- pyogenes aureus
coli pneumonie faecalis
63 >20
64 >20
65 >20
66 >20
10-20
>20
>20
>20
>50
>50
>50
>50
67
10-20
5-10
68 >20
>20
2.5-5
>20
69
2.5-5
3-6
3-6
6-12
3-6
70 >20
71
72
73
74
75
2.5-5 <2.5
2.5-5 <2.5
2.5-5 >10
3-6
3-6
>50
3-6
3-6
12-25
3-6
6-12
1.6-3
3-6
5-10
5-10
5-10 >20
76 >20
77
>20
2.5-5 >50
>20
>20
2.5-5
3-6
3-6 >50
78 >20
79 >20
80
81
10-20 >20
2.5-5 5-10
>20
6-12 12-25 >50
25-50
82 >20
a See footnote a to Table 1.
functionalities appears limited, significant discriminating di-
versity in our antimicrobial assays is provided by substitution
of functional groups in the meta position of the benzyls. The
iterative deconvolution of the active sublibraries is in progress.
Further research planned in this area includes the examination
of the effect of scaffold ring size on biological activity. Thus,
scaffolds 1 and 3, which are 13- and 15-membered polyaza-
phanes (scaffold 2 is 14-membered), will be combinatorialized
with the same functionality set as described. This should have
a significant effect on the size and volume of structural space
searched, but interacting functionalities will be the same as in
libraries constructed from scaffold 2.
In conclusion, a highly efficient, versatile synthetic route has
been developed for the preparation of polyazaphane scaffolds
suitable for solution phase combinatorialization procedures.
Three novel asymmetric polyazaphanes scaffolds 1-3 were
synthesized using this process. A novel solution phase,
simultaneous addition of functionalities (SPSAF) process was
developed to generate large quantities of libraries containing
approximately equimolar amounts of all possible compounds.
Sixteen high-quality libraries were prepared from scaffold 2.
N3-(tert-Butoxycarbonyl)-1,6-diamino-3-azahexane (26). Com-
pound 24 (16.4 g, 40 mmol) was treated with a mixture of MeOH-
30% NH4OH (270 mL, 1:2) and refluxed for 20 h. The solvent was
evaporated under reduced pressure, and the residue was purified by
flash chromatography on a silica gel column (25 × 7 cm). Gradient
elution with 50:1, 20:1, and then 5:1 MeOH-30% NH4OH gave product
26 as a colorless oil: yield 6.0 g (69%); silica gel TLC Rf 0.47 (10:1
MeOH-30% NH4OH); 1H NMR (CDCl3) δ 1.45 (s, 9 H), 1.54 (bs, 4
H, ex D2O), 1.56-1.72 (m, 2 H), 2.68 (t, 2 H, J ) 6.6 Hz), 2.81 (t, 2
H, J ) 6.4 Hz), 3.18-3.35 (m, 4 H); 13C NMR (CDCl3) δ 28.22, 31.75,
38.96, 40.43, 44.71, 49.81, 79.32, 155.73; HRMS (FAB) m/z 218.186
(M + 1)+ (C10H24N3O2 requires 218.187). Anal. Calcd for
C10H23N3O2: C, 55.27; H, 10.67; N, 19.33. Found: C, 55.39; H, 10.31;
N, 18.76.
N3-(tert-Butoxycarbonyl)-N1,N6-bis(2-nitrobenzenesulfonyl)-1,6-
diamino-3-azahexane (29). A solution of 2-nitrobenzenesulfonyl
chloride (28) (15.2 g, 68.2 mmol, 2.33 equiv) in 90 mL of CH2Cl2 was
added dropwise to a stirred solution of 26 (6.35 g, 29.2 mmol) and 24
mL of Et3N in 90 mL of CH2Cl2 at 0 °C. The resulting reaction mixture
was allowed to warm to room temperature and further stirred for 1 h.
The reaction mixture was diluted with CHCl3 and washed with H2O
and then brine. The organic phase was dried (Na2SO4), and the solvent
was evaporated under the reduced pressure. The residue was purified
by flash chromatography on a silica gel column (20 × 5 cm). Elution
with 2:1 and then 1:1 hexanes-EtOAc afforded product 29 as a pale
yellow viscous oil: yield 11.5 g (67%); silica gel TLC Rf 0.52 (1:2
1
The purity of the libraries was confirmed by TLC, CZE, H
NMR, and ESI/MS spectroscopic techniques. A “fix-last”
concept was developed to allow initial SAR studies to be
performed on the mixtures. Several of the sublibraries as
mixtures of 100 compounds exhibited potent antimicrobial
activities. Five single compounds showed promising activities
against several assays, and three compounds showed high
selectivities in some assays.
Experimental Section
General Methods. Capillary zone electrophoresis (CZE) analyses
were performed using a 37 cm fused silica column (50 µm i.d.) and an
applied voltage of 30 kV at 25 °C. Columns were flushed with 0.2 M
NaOH solution for 1.5 min, then buffered for 2.0 min prior to sample
injection. Samples were pressure injected for 3-8 s at 0.5 psi.
Compounds 23, 25, and 27 were prepared according to the literature
procedures.28 N5-(tert-Butoxycarbonyl)-1,7-diamino-1-oxa-5-azahep-
tane (34) was synthesized according to our procedure.27 Starting
materials and anhydrous DMF, acetonitrile, and THF were purchased
from Aldrich and used directly. The bacterial and yeast anti-growth
assays were performed in 96-well plate format in 150 µL volume in
(34) National Committee for Clinical Laboratory Standards. Methods for
Dilution Antimicrobial Susceptibility Tests for Bacterial that Grow Aerobi-
cally, 1993; MCCLS Document M7-A3, Vol. 13, Villanova, PA.
(35) Sampson, B. A.; Misra, R.; Benson, S. A. Genetics 1989, 122, 491.

3706 J. Am. Chem. Soc., Vol. 119, No. 16, 1997
An et al.
hexanes-EtOAc); 1H NMR (CDCl3) δ 1.40 (s, 9 H), 1.56-1.75 (m, 2
H), 2.98-3.10 (m, 2 H), 3.14-3.36 (m, 6 H), 5.60 (bs, 1 H, ex D2O),
residue was purified by flash chromatography on a silica gel column
(20 × 3 cm). Elution with 100% MeOH, 100:1 and then 50:1 MeOH-
30% NH4OH, afforded product 1 as a colorless oil: yield 1.81 g (94%);
silica gel TLC Rf 0.42 (50:1 MeOH-30% NH4OH); 1H NMR (CDCl3)
δ 1.35 (s, 9 H), 1.48-1.65 (m, 2 H), 2.19 (bs, 2 H, ex in D2O), 2.62
(t, 2 H, J ) 6.4 Hz), 2.78 (t, 2 H, J ) 6.2 Hz), 3.00 (t, 2 H, J ) 6.2
Hz), 3.28 (t, 2 H, J ) 6.4 Hz), 3.90 (s, 4 H), 7.04 (t, 2 H, J ) 7.0 Hz),
7.56 (t, 1 H, J ) 7.0 Hz); 13C NMR (CDCl3) δ 28.43, 29.89, 46.09,
46.39, 48.31, 48.60, 54.25, 54.67, 79.25, 120.91, 121.31, 137.14, 156.04,
159.95, 160.11; HRMS (FAB) m/z 321.230 (M + H)+ (C17H29N4O2
requires 321.229). Anal. Calcd for C17H28N4O2: C, 63.72; H, 8.80;
N, 17.48. Found: C, 63.24; H, 8.46; N, 17.13.
7-(tert-Butoxycarbonyl)-3,7,12,18-tetraazabicyclo[12.3.1]octadeca-
1(18),14,16-triene (2). The mono-protected polyazaphane scaffold 2
was synthesized as above for 1 from 1.44 g (2.0 mmol) of 33, 500 µL
(0.53 g, 4.8 mmol) of thiophenol, and 2.21 g (16.0 mmol) of anhydrous
K2CO3 in 30 mL of anhydrous DMF. The crude product was purified
by flash chromatography to afford product 2 as a colorless oil: yield
0.50 g (72%); silica gel TLC Rf 0.44 (50:1 MeOH-30% NH4OH); 1H
NMR (CDCl3) δ 1.37 (s, 9 H), 1.42-1.58 (m, 4 H), 1.65-1.80 (m, 2
H), 2.22 (bs, 2 H, ex D2O), 2.53 (t, 2 H, J ) 5.2 Hz), 2.62 (t, 2 H, J
) 6.2 Hz), 2.95-3.09 (m, 2 H), 3.21 (t, 2 H, J ) 6.2 Hz), 3.85 (s, 4
H), 6.99 (t, 2 H, J ) 7.4 Hz), 7.50 (t, 1 H, J ) 7.4 Hz); 13C NMR
(CDCl3) δ 24.90, 26.22, 28.45, 29.41, 45.00, 46.16, 46.50, 47.35, 54.10,
54.46, 79.03, 120.81, 121.06, 136.56, 155.57, 159.42; HRMS (FAB)
m/z 349.261 (M + H)+ (C19H33N4O2 requires 349.260). Anal. Calcd
for C19H32N4O2: C, 65.49; H, 9.26; N, 16.07. Found: C, 65.61; H,
9.45; N, 15.77.
N5-(tert-Butoxycarbonyl)-N1,N7-bis(2-nitrobenzenesulfonyl)-1,7-
diamino-1-oxa-5-azaheptane (35) and N5-(tert-Butoxycarbonyl)-
N1,N1,N7-tris(2-nitrobenzenesulfonyl)-1,7-diamino-1-oxa-5-azahep-
tane (36). Compound 35 was prepared as above for 29 from 10.64 g
(47.0 mmol) of 28, 4.67 g (20.0 mmol) of 34,27 and 16 mL of Et3N in
140 mL of CH2Cl2. Flash chromatographic purification afforded
product 35 as a white foam: yield 3.95 g (33%); silica gel TLC Rf
0.62 (1:2 hexanes-EtOAc); 1H NMR (CDCl3) δ 1.40 (s, 9 H),
1.72-1.90 (m, 2 H), 3.18-3.41 (m, 6 H), 3.98-4.11 (m, 2 H),
7.68-7.90 (m, 6 H), 8.10-8.25 (m, 2 H); HRMS (FAB) m/z 736.035
(M + Cs)+ (C22H29N5O11S2Cs requires 736.035). Anal. Calcd for
C22H29N5O11S2: C, 43.78; H, 4.84; N, 11.60. Found: C, 43.69; H,
4.82; N, 11.34.
6.20 (bs, 1 H, ex D2O), 7.66-7.88 (m, 6 H), 8.02-8.13 (m, 2 H); 13
C
NMR (CDCl3) δ 23.25, 40.68, 42.43, 44.18, 45.17, 46.89, 80.64, 125.32,
130.07, 132.94, 133.13, 133.99, 147.85, 156.07; HRMS (FAB) m/z
720.039 (M + Cs)+ (C22H29N5O10S2Cs requires 720.041). Anal. Calcd
for C22H29N5O10S2: C, 44.97; H, 4.97; N, 11.92. Found: C, 44.78; H,
5.03; N, 11.58.
N4-(tert-Butoxycarbonyl)-N1,N8-bis(2-nitrobenzenesulfonyl)sper-
midine (30). Compound 30 was prepared as above for 29 from 5.32
g (24.0 mmol) of 28, 2.45 g (10.0 mmol) of N4-(tert-butoxycarbonyl)-
spermidine (27)28 and 8 mL of Et3N in 60 mL of CH2Cl2. The resulting
reaction mixture was allowed to warm to room temperature and further
stirred for 1 h. Flash chromatographic purification of the crude product
gave 30 as a pale yellow viscous oil: yield 5.62 g (91%); silica gel
TLC Rf 0.52 (1:2 hexanes-EtOAc); 1H NMR (CDCl3) δ 1.36 (s, 9 H),
1.40-1.52 (m, 4 H), 1.58-1.74 (m, 2 H), 2.98-3.25 (m, 8 H), 5.45
(bs, 1 H, ex D2O), 6.31 (bs, 1 H, ex D2O), 7.65-7.85 (m, 6 H),
8.00-8.11 (m, 2 H); 13C NMR (CDCl3) δ 25.37, 26.87, 28.35,
40.81, 43.45, 46.39, 79.92, 125.18, 125.33, 130.78, 130.94, 132.77,
132.92, 133.51, 133.79, 148.02, 156.00; HRMS (FAB) m/z 748.075
(M + Cs)+ (C24H33N5O10S2Cs requires 748.072). Anal. Calcd for
C24H33N5O10S2: C, 46.82; H, 5.40; N, 11.37. Found: C, 46.66; H,
5.40; N, 11.00.
6-(tert-Butoxycarbonyl)-3,10-bis(2-nitrobenzenesulfonyl)-3,6,10,-
16-tetraazabicyclo[10.3.1]hexadeca-1(16),12,14-triene (32). A mix-
ture of 2.6-bis(bromomethyl)pyridine (31) (4.93 g, 18.6 mmol),
anhydrous Cs2CO3 (24.0 g, 73.6 mmol, 4 equiv), and 29 (10.98 g, 18.6
mmol, 1 equiv) in 500 mL of anhydrous DMF was stirred at room
temperature for 24 h. The solvent was evaporated under vacuum, and
the residue was dissolved in a mixture of H2O and CHCl3. The layers
were separated, and the aqueous phase was extracted with CHCl3. The
combined organic phase was washed with brine, dried (Na2SO4), and
concentrated. The residue was purified by flash chromatography on a
silica gel column (18 × 5 cm). Elution with 1:1, 1:2, and then 1:4
hexanes-EtOAc gave product 32 as a white foam: yield 10.3 g (80%);
silica gel TLC Rf 0.50 (1:4 hexanes-EtOAc); 1H NMR (CDCl3) δ 1.40
(s, 9 H), 1.67-1.85 (m, 2 H), 2.60-2.80 (m, 2 H), 3.04-3.15 (m, 2
H), 3.25-3.46 (m, 4 H), 4.55 (s, 4 H), 7.48-7.80 (m, 9 H), 7.96-
8.10 (m, 2 H); 13C NMR (CDCl3) δ 26.98, 28.34, 45.36, 46.40, 47.07,
49.26, 55.24, 55.65, 79.83, 123.97, 124.25, 124.41, 130.67, 130.92,
131.91, 133.98, 138.64, 148.13, 148.36, 155.06, 155.67; HRMS (FAB)
m/z 823.086 (M + Cs)+ (C29H34N6O10S2Cs requires 823.083). Anal.
Calcd for C29H34N6O10S2: C, 50.42; H, 4.96; N, 12.16. Found: C,
49.95; H, 4.96; N, 12.07.
The tetra-protected triamine 36 also was isolated as a white foam:
1
yield 7.13 g (45%); silica gel TLC Rf 0.49 (1:2 hexanes-EtOAc); H
NMR (CDCl3) δ 1.42 (s, 9 H), 1.75-1.92 (m, 2 H), 3.10-3.42 (m, 6
H), 4.10-4.22 (m, 2 H), 7.68-7.90 (m, 9 H), 8.05-8.25 (m, 3 H); MS
(ESI) m/z 787 (M - 1)+.
7-(tert-Butoxycarbonyl)-3,12-bis(2-nitrobenzenesulfonyl)-3,7,12,-
18-tetraazabicyclo[12.3.1]octadeca-1(18),14,16-triene (33). Macro-
cyclic compound 33 was synthesized as above for 32 from 1.33 g (5.0
mmol) of 31, 6.52 g (29 mmol) of anhydrous Cs2CO3, and 3.23 g (5.0
mmol) of 30 in 160 mL of anhydrous DMF. Purification of the crude
product by flash chromatography gave product 33 as a white foam:
8-(tert-Butoxycarbonyl)-3,11-bis(2-nitrobenzenesulfonyl)-4-oxa-
3,8,11,17-tetraazabicyclo[11.3.1]heptadeca-1(17),13,15-triene (37).
Compound 37 was synthesized as above for 32 from 1.66 g (6.2
mmol) of 31, 8.20 g (25.0 mmol) of anhydrous Cs2CO3, and 3.80 g
(6.2 mmol) of 35 in 200 mL of anhydrous DMF. Flash chromato-
graphic purification afforded product 37 as a white foam: yield 2.20
g (50%): silica gel TLC Rf 0.55 (1:4 hexanes-EtOAc); 1H NMR
(CDCl3) δ 1.20-1.30 (m, 2 H), 1.38 (s, 9 H), 2.60-3.00 (m, 2 H),
3.16-3.40 (m, 2 H), 3.78-3.95 (m, 2 H), 4.30 (s, 2 H), 4.55 (s, 2 H),
7.50-7.78 (m, 9 H), 7.85-8.10 (m, 2 H); HRMS (FAB) m/z 839.078
(M + Cs)+ (C29H34N6O11S2Cs requires 839.078). Anal. Calcd for
C29H34N6O11S2: C, 49.28; H, 4.84; N, 11.89. Found: C, 49.27; H,
4.59; N, 11.68.
1
yield 2.63 g (73%); silica gel TLC Rf 0.48 (1:4 hexanes-EtOAc); H
NMR (CDCl3) δ 1.15-1.30 (m, 4 H), 1.36 (s, 9 H), 1.58-1.74 (m, 2
H), 2.98-3.10 (m, 4 H), 3.16-3.40 (m, 4 H), 4.48 (s, 2 H), 4.62 (s, 2
H), 7.43 (d, 2 H, J ) 7.6 Hz), 7.61-7.78 (m, 7 H), 7.98-8.10 (m, 2
H); 13C NMR (CDCl3) δ 26.14, 26.68, 28.37, 46.72, 47.00, 48.67, 48.84,
53.69, 54.42, 79,60, 122.58, 122.85, 124.27, 130.77, 131.83, 133.77,
138.06, 148.24, 155.67, 156.00, 156.12; HRMS (FAB) m/z 851.118
(M + Cs)+ (C31H38N6O10S2Cs requires 851.115). Anal. Calcd for
C31H38N6O10S2: C, 51.80; H, 5.33; N, 11.69. Found: C, 51.81; H,
5.48; N, 11.50.
8-(tert-Butoxycarbonyl)-4-oxa-3,8,11,17-tetraazabicyclo[11.3.1]-
heptadeca-1(17),13,15-triene (3). Compound 3 was synthesized as
above for 1 from 495 mg (0.70 mmol) of 37, 158 µL (169 mg, 1.53
mmol) of thiophenol, and 0.80 g (5.7 mmol) of anhydrous in 10 mL of
anhydrous DMF. Flash chromatographic purification afforded product
3 as a colorless oil: yield 230 mg (97%); silica gel TLC Rf 0.50 (100%
6-(tert-Butoxycarbonyl)-3,6,10,16-tetraazabicyclo[10.3.1]hexadeca-
1(16),12,14-triene (1). Thiophenol (1.5 mL, 1.60 g, 14.5 mmol, 2.46
equiv) was added to a stirred mixture of fully protected macrocyclic
compound 32 (4.15 g, 6.0 mmol) and anhydrous K2CO3 (6.63 g, 48
mmol, 8 equiv) in 80 mL of anhydrous DMF. The resulting blue
mixture was stirred at room temperature for 2 h, and the color of the
reaction mixture gradually changed to yellow. The yellow reaction
mixture thus obtained was concentrated under vacuum, and the residue
was dissolved in H2O. The solution was adjusted to pH 13-14 with
NaOH solution and was extracted with CHCl3. The combined organic
phase was washed with brine, dried (Na2SO4), and concentrated. The
1
MeOH); H NMR (CDCl3) δ 1.34 (s, 9 H), 1.45-1.65 (m, 2 H), 2.65
(t, 2 H, J ) 6.8 Hz), 2.97 (t, 2 H, J ) 7.2 Hz), 3.14 (t, 2 H, J ) 6.8
Hz), 3.58 (t, 2 H, J ) 6.0 Hz), 3.92 (s, 2 H), 4.05 (s, 2 H), 7.09 (t, 2
H, J ) 7.2 Hz), 7.56 (t, 1 H, J ) 7.2 Hz); 13C NMR (CDCl3) δ 27.78,
28.43, 43.97, 45.88, 47.32, 54.39, 57.11, 70.57, 79.27, 121.72, 121.87,
137.01, 155.72, 157.88, 159.25; HRMS (FAB) m/z 337.224 (M + H)+
(C17H29N4O3 requires 337.224).

Solution Phase Combinatorial Chemistry
J. Am. Chem. Soc., Vol. 119, No. 16, 1997 3707
3,7,12,18-Tetraazabicyclo[12,3,1]octadeca-1(18),14,16-triene (38).
Trifluoroacetic acid (TFA) (8 mL) was added to a stirred solution of 2
(349 mg, 1.0 mmol) in 3 mL of CHCl3 at 0 °C. The resulting reaction
mixture was stirred at room temperature for 4 h. The solvent and excess
TFA were evaporated under vacuum. The residue was dissolved in
H2O and adjusted to pH 14 by addition of NaOH solution. The mixture
was extracted with CHCl3. The combined organic phase was washed
with brine, dried (Na2SO4), and concentrated. The residue was purified
by flash chromatography on a silica gel column (11 × 2 cm). Elution
with 100% MeOH and then 10:1 MeOH-30% NH4OH afforded
product 38 as a colorless oil: yield 180 mg (73%); silica gel TLC Rf
0.41 (5:1 MeOH-30% NH4OH); 1H NMR (CDCl3) δ 1.24-1.38 (m, 2
H), 1.39-1.58 (m, 4 H), 2.10 (bs, 3 H, ex D2O), 2.30-2.57 (m, 8 H),
3.64 (s, 2 H), 3.71 (s, 2 H), 6.83 (d, 2 H, J ) 7.6 Hz), 7.35 (t, 1 H, J
) 7.6 Hz); 13C NMR (CDCl3) δ 26.76, 27.21, 29.34, 48.16, 48.63,
48.84, 53.99, 55.20, 120.47, 136.32, 159.04, 159.75; HRMS (FAB)
m/z 249.207 (M + 1)+ (C14H25N4 requires 249.207). Anal. Calcd for
C14H24N4: C, 67.70; H, 9.74; N, 22.55. Found: C, 67.02; H, 9.67; N,
22.09.
3,7,12-Tribenzyl-3,7,12,18-tetraazabicyclo[12,3,1]octadeca-1(18),-
14,16-triene (39). A mixture of 38 (75 mg, 0.30 mmol), anhydrous
K2CO3 (0.50 g, 3.6 mmol), and benzyl bromide (120 µL, 169 mg, 0.99
mmol, 3.3 equiv) in anhydrous CH3CN (5 mL) was stirred at room
temperature overnight. The solvent was evaporated, and the residue
was dissolved in H2O-CHCl3. The organic phase was separated off,
and the aqueous phase was extracted with CHCl3. The combined
organic phase was washed with brine, dried (Na2SO4), and concentrated.
The residue was purified by preparative thin layer chromatography
(PLC) on a silica gel plate by using 1:3 hexanes-EtOAc as the
developing agent. Product 39 was obtained as a pale yellow oil: yield
60 mg (39%); silica gel TLC Rf 0.52 (1:4 hexanes-EtOAc); 1H NMR
(CDCl3) δ 1.18-1.38 (m, 4 H), 1.45-1.62 (m, 2 H), 2.09-2.28 (m, 4
H), 2.47-2.60 (m, 4 H), 3.38 (s, 2 H), 3.62 (s, 2 H), 3.67 (s, 2 H),
3.70 (s, 2 H), 3.75 (s, 2 H), 7.12-7.48 (m, 17 H), 7.60 (t, 1 H, J ) 7.6
Hz); HRMS (FAB) m/z 519.347 (M + 1)+ (C35H43N4 requires 519.348).
General Procedure for the Preparation of Libraries 40-52 and
4. A solution containing equimolar amounts of selected benzylic
bromides (total 2.4 equiv) in anhydrous CH3CN was added to a stirred
mixture of scaffold 2 (1.0 equiv) and anhydrous K2CO3 (15 equiv) in
CH3CN. The resulting reaction mixture was stirred at room temperature
overnight. The solvent was evaporated and the residue was dissolved
in H2O and CHCl3. The organic phase was separated off, and the
aqueous phase was extracted with CHCl3. The combined organic phase
was washed with brine, dried (Na2SO4), and concentrated. The residue
was purified by flash chromatography on a silica gel column or by
preparative thin layer chromatography (PLC) on a silica gel plate to
afford desired libraries.
Elution with 100% MeOH and then 50:1 and 20:1 MeOH-30% NH4-
OH afforded library 5 as a pale yellow oil: yield 2.62 g (93%); silica
gel TLC Rf 0.32-0.53 (20:1 MeOH-30% NH4OH); 1H NMR (CDCl3)
δ 1.30-1.75 (m, 6 H), 2.28-2.72 (m, 8.3 H), 3.35-3.82 (m, 8 H),
3.92 (s, 0.3 H), 6.30-6.70 (m, 0.3 H), 6.90-8.50 (m, 11.1 H); MS
(ESI) m/z 429-587 (M + 1)+.
General Procedure for the Preparation of Libraries 6-19. A
selected benzylic bromide or functionality (1.3 equiv) was added to a
stirred mixture of library 5 (1.0 equiv) and anhydrous K2CO3 (15 equiv)
in anhydrous CH3CN. The resulting reaction mixture was stirred at
room temperature overnight. The solvent was evaporated, and the
residue was dissolved in H2O and CHCl3. The organic phase was
separated off, and the aqueous phase was extracted with CHCl3. The
combined organic phase was washed with brine, dried (Na2SO4), and
concentrated. The residue was purified by preparative thin layer
chromatography (PLC) on a silica gel plate to afford the desired
libraries.
Library 12: pale yellow oil; yield 106 mg (85%); 1H NMR (CDCl3)
δ 1.15-1.40 (m, 4 H), 1.42-1.65 (m, 2 H), 2.05-2.29 (m, 4 H), 2.36
(s, 0.3 H), 2.46-2.63 (m, 4 H), 3.34 (s, 2 H), 3.58-3.86 (m, 8 H),
3.93 (s, 0.3 H), 6.50-8.45 (m, 15.3 H); MS (ESI) m/z 553-712 (M +
1)+.
Library 19. Library 19 was prepared according to the general
procedure by stirring the reaction mixture at 60-70 °C for 6 h and
then at room temperature for 10 h: colorless oil; yield 112 mg (93%);
1H NMR (CDCl3) δ 1.15-1.42 (m, 4 H), 1.43-1.62 (m, 2 H), 2.28-
2.65 (m, 8.3 H), 3.11 (s, 3 H), 3.26 (s, 2 H), 3.55-3.84 (m, 11 H),
3.91 (s, 0.3 H), 6.35-8.42 (m, 11.3 H); MS (ESI) m/z 530-688 (M +
1)+.
General Procedure for the Preparation of Compounds 53-62.
A mixture of benzylic bromide (2.4 equiv), scaffold 2 (1.0 equiv), and
anhydrous K2CO3 (15 equiv) in anhydrous CH3CN was stirred at room
temperature overnight. The reaction was worked up by the same
procedure as for the preparation of libraries 40-52. Flash chromato-
graphic purification on silica gel column afforded the desired products.
Compound 54: colorless oil; yield 750 mg (97%); silica gel TLC
Rf 0.41 (3:2 hexanes-EtOAc); 1H NMR (CDCl3) δ 1.33-1.55 (m, 13
H), 1.57-1.77 (m, 2 H), 2.50-2.68 (m, 4 H), 2.75-2.98 (m, 2 H),
3.00-3.14 (m, 2 H), 3.61 (s, 2 H), 3.69 (s, 2 H), 3.74 (s, 2 H), 3.76 (s,
2 H), 7.03 (d, 1 H, J ) 7.4 Hz), 7.12 (d, 1 H, J ) 7.4 Hz), 7.19-7.56
(m, 11 H); 13C NMR (CDCl3) δ 23.62, 26.27, 28.54, 46.38, 47.22,
51.94, 59.84, 60.44, 60.71, 78.77, 122.23, 122.54, 126.97, 128.23,
129.05, 136.03, 139.66, 139.83, 155.70, 159.12; MS (FAB) m/z 529
(M + H)+; HRMS (FAB) m/z 661.250 (M + Cs)+ (C33H44N4O2Cs
requires 661.251).
Compound 55: colorless oil; yield 1.60 g (94%); silica gel TLC Rf
1
0.69 (3:2 hexanes-EtOAc); H NMR (CDCl3) δ 1.30-1.55 (m, 13
Library 40: colorless oil; yield 115 mg (71%); silica gel TLC Rf
H), 1.57-1.76 (m, 2 H), 2.49-2.67 (m, 4 H), 2.72-2.96 (m, 2 H),
3.07 (t, 2 H, J ) 6.8 Hz), 3.59 (s, 2 H), 3.68 (s, 2 H), 3.70 (s, 2 H),
3.74 (s, 2 H), 6.86-7.32 (m, 10 H), 7.51 (t, 1 H, J ) 7.6 Hz); 13C
NMR (CDCl3) δ 23.69, 26.25, 28.51, 46.34, 47.21, 51.98, 59.24, 59.89,
60.12, 60.51, 78.80, 113.60, 113.96, 115.44, 115.86, 122.18, 122.48,
124.30, 129.50, 129.66, 136.12, 142.75, 142.99, 155.63, 158.93, 159.14,
160.63, 165.51; MS (FAB) m/z 697 (M + Cs)+; HRMS (FAB) m/z
565.335 (M + H)+ (C33H43N4O2F2 requires 565.335).
1
0.47 (3:2 hexanes-EtOAc); H NMR (CDCl3) δ 1.38-1.52 (m, 13
H), 1.60-1.80 (m, 2 H), 2.33 (s, 3 H), 2.51-2.68 (m, 4 H), 2.77-
3.00 (m, 2 H), 3.02-3.15 (m, 2 H), 3.63 (s, 2 H), 3.68-3.81 (m, 6 H),
6.85-7.60 (m, 12 H); MS (ESI) m/z 529, 543, 557 (M + 1)+.
Library 41: colorless oil; yield 117 mg (70%); silica gel TLC Rf
1
0.48 (3:2 hexanes-EtOAc); H NMR (CDCl3) δ 1.35-1.52 (m, 13
H), 1.59-1.78 (m, 2 H), 2.31 (s, 1.5 H), 2.38 (s, 1.5 H), 2.51-2.68 (m,
4 H), 2.74-2.98 (m, 2 H), 3.00-3.14 (m, 2 H), 3.61 (s, 2 H), 3.65-
3.70 (m, 6 H), 6.85-7.33 (m, 10 H), 7.52 (t, 1 H, J ) 7.6 Hz); MS
(ESI) m/z 557, 561, 565 (M + 1)+.
General Procedure for the Preparation of Compounds 63-72.
Compounds 63-72 were prepared from the corresponding compounds
53-62 by the same procedure as for the preparation of library 5. Flash
chromatographic purification provided the desired products.
Library 4: pale yellow oil; yield 3.40 g (94%); silica gel TLC Rf
0.23-0.78 (1:2 hexanes-EtOAc); 1H NMR (CDCl3) δ 1.30-1.55 (m,
13 H), 1.57-1.80 (m, 2 H), 2.35 (s, 0.15), 2.36 (s, 0.15 H), 2.50-2.70
(m, 4 H), 2.78-2.98 (m, 2 H), 3.00-3.18 (m, 2 H), 3.57-3.88 (m, 8
H), 3.92 (s, 0.3 H), 6.30-6.70 (m, 0.2 H), 6.90-8.50 (m, 11.16 H);
MS (ESI) m/z 529-687 (M + 1)+.
Preparation of Library 5. Trifluoroacetic acid (TFA) (40 mL) was
added to a stirred solution of library 4 (3.37 g, 5.6 mmol) in 8 mL of
CHCl3 at 0 °C. The resulting reaction mixture was stirred at room
temperature for 4 h. The solvent and excess TFA were evaporated,
and the residue was dissolved in 400 mL of CHCl3. The CHCl3 solution
was washed with K2CO3 solution (200 mL × 2) and brine. After drying
(Na2SO4) the organic phase was concentrated and the residue was
purified by flash chromatography on a silica gel column (12 × 3 cm).
Compound 63: pale yellow oil; yield 570 mg (96%); silica gel TLC
1
Rf 0.28 (20:1 MeOH-30% NH4OH); H NMR (CDCl3) δ 1.37-1.51
(m, 2 H), 1.52-1.75 (m, 4 H), 2.32-2.45 (m, 8 H), 2.52-2.68 (m, 6
H), 3.61 (s, 2 H), 3.64 (s, 2 H), 3.71 (s, 2 H), 3.73 (s, 2 H), 7.03-7.31
(m, 10 H), 7.54 (t, 1 H, J ) 7.6 Hz); 13C NMR (CDCl3) δ 21.52, 23.00,
25.89, 26.77, 47.54, 47.84, 51.45, 52.96, 59.40, 59.63, 60.13, 60.41,
122.57, 126.19, 127.71, 128.14, 129.80, 136.11, 137.75, 139.67, 139.83,
158.67, 159.37; MS (FAB) m/z 589 (M + Cs)+; HRMS (FAB) m/z
457.332 (M + H)+ (C30H41N4 requires 457.333). Anal. Calcd for
C30H40N4: C, 78.90; H, 8.82; N, 12.27. Found: C, 78.68; H, 8.98; N,
12.06.
Compound 64: pale yellow oil; yield 564 mg (93%); silica gel TLC
1
Rf 0.31 (20:1 MeOH-30% NH4OH); H NMR (CDCl3) δ 1.35-1.49

3708 J. Am. Chem. Soc., Vol. 119, No. 16, 1997
An et al.
(m, 2 H), 1.50-1.75 (m, 4 H), 2.36 (t, 2 H, J ) 5.6 Hz), 2.50-2.70
(m, 6 H), 3.59 (s, 2 H), 3.67 (s, 2 H), 3.69 (s, 2 H), 3.74 (s, 2 H),
7.00-7.15 (m, 2 H), 7.20-7.57 (m, 11 H); 13C NMR (CDCl3) δ 23.05,
25.85, 26.66, 47.47, 47.72, 51.39, 52.88, 59.44, 59.67, 60.11, 60.38,
122.59, 126.98, 128.27, 129.07, 136.17, 139.72, 139.92, 158.65, 159.36;
MS (FAB) m/z 561 (M + Cs)+; HRMS (FAB) m/z 429.303 (M + H)+
(C28H37N4 requires 429.301). Anal. Calcd for C28H36N4: C, 78.46;
H, 8.45; N, 13.07. Found: C, 78.20; H, 8.21; N, 12.82.
General Procedure for the Preparation of Compounds 73-82.
A mixture of cinnamyl bromide (1.1 equiv), anhydrous K2CO3 (15
equiv), and the corresponding compounds 63-72 in anhydrous CH3-
CN was stirred at room temperature for 2-5 h. The same work-up
procedure as for the preparation of libraries 6-19 was used and was
followed by chromatographic purification on a silica gel column to
provide the desired product.
Compound 79: colorless oil; yield 362 mg (30%); silica gel TLC
Rf 0.55 (5:1 EtOAc-MeOH); H NMR (CDCl3) δ 1.25-1.49 (m, 4
1
H), 1.50-1.69 (m, 2 H), 2.22 (t, 2 H, J ) 5.8 Hz), 2.34 (t, 2 H, J )
6.3 Hz), 2.50-2.70 (m, 4 H), 3.09 (d, 2 H, J ) 6.3 Hz), 3.66 (s, 2 H),
3.68 (s, 2 H), 3.71 (s, 2 H), 3.73 (s, 2 H), 6.09-6.25 (m, 1 H), 6.46 (d,
1 H, J ) 15.8 Hz), 7.12-7.68 (m, 16 H); 13C NMR (CDCl3) δ 24.40,
24.72, 51.72, 52.37, 52.63, 53.94, 57.14, 59.69, 60.09, 60.31, 60.43,
122.50, 126.32, 127.14, 128.38, 128.59, 129.06, 129.55, 131.80, 134.24,
136.25, 137.40, 142.35, 158.87, 158.99; MS (FAB) m/z 745 (M + Cs)+;
HRMS (FAB) m/z 613.284 (M + H)+ (C37H43N4Cl2 requires 613.286).
Anal. Calcd for C37H42N4Cl2: C, 72.42; H, 6.90; N, 9.13. Found: C,
72.42; H, 6.68; N, 8.99.
Acknowledgment. The authors thank Professor Michael E.
Jung and Dr. Thomas W. Bruice for their stimulating scientific
discussions. High-resolution (FAB) and electrospray ionization
mass spectrometry were provided by The Scripps Research
Institute Mass Spectrometry.
Compound 74: colorless oil; yield 124 mg (41%); silica gel TLC
1
Rf 0.45 (5:1 EtOAc-MeOH); H NMR (CDCl3) δ 1.24-1.48 (m, 4
H), 1.50-1.67 (m, 2 H), 2.21 (t, 2 H, J ) 5.8 Hz), 2.33 (t, 2 H, J )
6.3 Hz), 2.53-2.68 (m, 4 H), 3.10 (d, 2 H, J ) 6.3 Hz), 3.68 (s, 4 H),
3.76 (s, 2 H), 3.79 (s, 2 H), 6.09-6.24 (m, 1 H), 6.46 (d, 1 H, J )
16.0 Hz), 7.15-7.61 (m, 18 H); 13C NMR (CDCl3) δ 24.54, 24.76,
24.96, 51.86, 52.42, 52.58, 54.13, 57.06, 60.29, 60.86, 122.48, 126.32,
126.97, 127.23, 128.29, 128.58, 129.11, 131.66, 136.10, 137.47, 140.02,
159.09, 159.35; MS(FAB) m/z 545 (M + H)+; HRMS (FAB) m/z
677.264 (M + Cs)+ (C37H44N4Cs requires 677.262). Anal. Calcd for
C37H44N4: C, 81.57; H, 8.14; N, 10.28. Found: C, 81.55; H, 8.09; N,
10.13.
Supporting Information Available: Characterization of
libraries 42-52, 6-11, and 13-18 and compounds 53, 56-
62, 65-73, 75-78, 80-82, capillary zone electrophoresis
profiles of libraries 51 and 52, and NMR spectra of compounds
22, 24, 3, and 39 (21 pages). See any current masthead page
for ordering and Internet access instructions.
JA964153R