Two comparisons of the performance of positional scanning and deletion synthesis for the identification of active constituents in mixture combinatorial libraries

Article abstract of DOI:10.1021/jo9916481

Two libraries of 120 compounds each were prepared as individual compounds and as full mixtures. The corresponding scanning and deletion synthesis deconvolution libraries were prepared and tested (L-1210, IC₅₀) alongside the individual compounds and mixture libraries. This testing, where the properties of each compound in the mixtures were known, was used to compare the performance of scanning and deletion deconvolution libraries. Each has its own intrinsic strengths, with the former being capable of identifying multiple hits at the expense of accurately identifying the most potent library member, while the latter typically is more sensitive to identifying the most potent hit but at the expense of differentiating weaker activities. The protocols complement one another and together more thoroughly identify potent library members.

Full text of DOI:10.1021/jo9916481

J . Org. Chem. 2000, 65, 1467-1474
1467
Tw o Com p a r ison s of th e P er for m a n ce of P osition a l Sca n n in g a n d
Deletion Syn th esis for th e Id en tifica tion of Active Con stitu en ts in
Mixtu r e Com bin a tor ia l Libr a r ies
Dale L. Boger,* J ae Kyoo Lee, J oel Goldberg, and Qing J in
Department of Chemistry and The Skaggs Institute for Chemical Biology, The Scripps Research Institute,
10550 North Torrey Pines Road, La J olla, California 92037
Received October 21, 1999
Two libraries of 120 compounds each were prepared as individual compounds and as full mixtures.
The corresponding scanning and deletion synthesis deconvolution libraries were prepared and tested
(L-1210, IC50) alongside the individual compounds and mixture libraries. This testing, where the
properties of each compound in the mixtures were known, was used to compare the performance
of scanning and deletion deconvolution libraries. Each has its own intrinsic strengths, with the
former being capable of identifying multiple hits at the expense of accurately identifying the most
potent library member, while the latter typically is more sensitive to identifying the most potent
hit but at the expense of differentiating weaker activities. The protocols complement one another
and together more thoroughly identify potent library members.
In tr od u ction
Integrins are cell surface receptors that recognize
extracellular matrix adhesive proteins such as fibrinogen,
fibronectin, vitronectin, and VCAM-1 (vascular cell adhe-
sion molecule-1).1-3 These important biological targets
are membrane-bound, heterodimeric glycoproteins, each
consisting of an R subunit (approximately 1100 residues)
and a â subunit (approximately 800 residues). The
relative affinity and specificity for ligand binding are
determined by the unique combinations of the different
R and â subunits.3 Of the members of this family of
receptors, RIIbâ3 (GPIIb-IIIa), Rvâ3, Rvâ5, R5â1, and R4â1
F igu r e 1.
continue to be extensively studied. Some of the disease
states that have a strong â integrin component in their
availability and platelet aggregation inhibitory activity
were enhanced by introduction of the sulfonamide func-
tionality found in 2.10 Similarly, a benzodiazepine scaffold
incorporating a piperidine, 3 (SB 214857),11 exhibited
etiologies are thrombosis (integrin RIIbâ3), unstable an-
gina (integrin RIIbâ3), osteoporosis (Rvâ5), and tumor
metastasis (Rvâ3 and Rvâ5). These integrins bind the
sequence Arg-Gly-Asp (RGD) as a common recognition
motif within their putative ligands.4-8
R
IIbâ3 selectivity, while that incorporating a properly
spaced benzimidazole, 4 (SB 223245), exhibited Rvâ3
Because of the crucial role that platelet RIIbâ3 plays in
thrombosis, antagonists which compete for binding to
selectivity (Figure 1).12
The vitronectin receptor (Rvâ3) is involved in many cell
adhesion processes. Cheresh and co-workers have shown
that in vivo inhibition of binding of these integrins to
their native ligands interferes with angiogenesis and
induces tumor regression.13 In addition to its relevance
to angiogenesis, Rvâ3 has also been shown to play a role
in mediating adhesion of osteoclasts to the bone matrix
and in the migration of vascular smooth muscle cells.
Therefore, antagonists of Rvâ3 are envisioned as potential
R
IIbâ3 are useful for the treatment of artial thrombotic
conditions including unstable angina, myocardial infarc-
tion, and stroke. Nonpeptide RGD-based antagonists
have been developed employing heterocyclic scaffolds or
constrained linkers as templates for attachment of the
crucial Arg and Asp side chain surrogates. Examples
include the isoindolone-based agent 1,9 whose oral bio-
(1) Ruoslahti, E. Annu. Rev. Cell Dev. Biol. 1996, 12, 697.
(2) Cox, D. Drug News Perspective 1995, 8, 197.
(3) Cox, D.; Aoki, T. Seki, J .; Motoyama, Y.; Yoshida, K. Med. Res.
Rev. 1994, 14, 195.
(10) Egbertson, M. S.; Hartman, G. D.; Gould, R. J .; Bednar, B.;
Bednar, R. A.; Cook, J . J .; Gaul, S. L.; Holahan, M. A.; Libby, L. A.;
Lynch, J . J ., J r.; Lynch, R. J .; Sitko, G. R.; Stranieri, M. T.; Vassallo,
L. M. Bioorg. Med. Chem. Lett. 1996, 6, 2519.
(4) Ruoslahti, E.; Pierschbacher, M. D. Science 1987, 238, 491.
(5) Springer, T. A. Nature 1990, 356, 425.
(6) D’Souza, S. E.; Ginsberg, M. H.; Plow, E. F. Trends Biochem.
Sci. 1991, 16, 246.
(7) Tomasini, B. R.; Mosher, D. F. Prog. Hemostasis Thromb. 1990,
10, 269.
(8) Ruoslahti, E.; Pierschbacher, M. Cell 1986, 44, 517.
(9) Egbertson, M. S.; Naylor, A. M.; Hartman, G. D.; Cook, J . J .;
Gould, R. J .; Holahan, M. A.; Lynch. J . J ., J r.; Lynch, R. J .; Stranieri,
M. T.; Vassallo, L. M. Bioorg. Med. Chem. Lett. 1994, 4, 1835.
(11) Samanen, J . M.; Ali, F. E.; Barton, L. S. et al. J . Med. Chem.
1996, 39, 4867.
(12) Keenan, R. M.; Miller, W. H.; Kwon, C.; Ali, F. E.; Callahan, J .
F.; Calvo, R. R.; Hwang, S.-M.; Kopple, K. D.; Peishoff, C. E.; Samanen,
J . M.; Wong, A. S.; Yuan, C.-K.; Huffman, W. F. J . Med. Chem. 1997,
40, 2289.
(13) Brooks, P. C.; Montgomery, A. M. P.; Rosenfeld, M.; Reisfeld,
R. A.; Hu, T.; Klier, G.; Cheresh, D. A. Cell 1994, 79, 1157.
10.1021/jo9916481 CCC: $19.00 © 2000 American Chemical Society
Published on Web 01/28/2000

1468 J . Org. Chem., Vol. 65, No. 5, 2000
Boger et al.
F igu r e 3.
F igu r e 2.
multiple assays, can be used to identify new leads.
Notably, the simple protocol of mixture synthesis typi-
cally cannot be conducted on the solid phase. Unlike
solid-phase synthesis where the polymer-bound sub-
strates must be the stoichiometry-limiting partner, either
the substrates or the reacting attachment groups may
be limiting in solution-phase chemistry. This dictates the
use of mix and split synthesis for the solid-phase to
accommodate differential reaction rates, whereas the
simpler procedure of mixture synthesis with limiting
reagent stoichiometry may be used in solution to ensure
all library members are generated. This only requires the
ability to remove unreacted starting substrates. Although
not possible with solid-phase synthesis, this was ac-
complished by aqueous acid/base extractions in each of
the steps, which also served to remove reactants, re-
agents, and reagent byproducts, providing clean products.
When appropriately functionalized, 7 contains a rigid
bicyclic core which enables it to function as a Arg-Gly-
Asp (RGD) mimic. Its symmetrical structure contains
three positions which can be functionalized with a variety
of nucleophiles and acylating agents, enabling the syn-
thesis of libraries with three points of diversity (Figure
3).
therapeutic agents for the treatment of numerous dis-
eases including diabetic retinopathy, cancer, osteoporosis,
and restenosis. Recently a number of high-affinity ligands
for Rvâ3 possessing structures different from classical
peptide frameworks (e.g., 5 and 6, Figure 2) have been
disclosed.12,14,15
In conjuction with an effort to evaluate a new class of
RGD mimetics, we elected to use the opportunity to
compare the performance of positional scanning and
deletion synthesis to identify active constituents in
mixture combinatorial libraries.16 We prepared two small
libraries of 120 compounds each as individual compounds
along with the two full mixture libraries and the ac-
companying scanning17 and deletion synthesis18 decon-
volution libraries. In this manner, the activity of each
library member could be established and compared to the
results derived from the scanning and deletion synthesis
libraries. These libraries will be subjected to multiple
screens including those designed to identify RGD an-
tagonists, and herein we report the first of these results
obtained in a cytotoxic assay. The results from the
comparisons serve to highlight the complementary nature
of the two deconvolution protocols.
Removal of the protecting group on the nitrogen of the
starting template allows for the first functionalization
by N-acylation and provides monoamides 8, which can
be purified by simple acid/base liquid-liquid extraction,
removing unreacted starting material, reagents, and
reaction byproducts. For the second functionalization of
the diester, hydrolysis affords the corresponding diacid,
which is activated as the cyclic anhydride. Suitable
amines (R2NH2) can be added to open the anhydride to
provide the diamide with release of a third functional-
ization site (-CO2H). The released carboxylic acid func-
tionality may be used for the purification of the expected
products, allowing removal of starting material, reagents,
and reaction byproducts by simple liquid-liquid extrac-
tion. Functionalization of the released acid, utilizing a
protocol similar to that of the second functionalization,
allows for additional diversity to be introduced onto the
rigid template. In each step of the sequence the reactants,
unreacted starting materials, and reagents and their
byproducts can be removed by simple extractions, provid-
ing the intermediates and the final compounds in high
purities (Scheme 1).
The template synthesis required N-Boc protection of
propargylamine and subsequent alkylation, effected by
treatment with NaH (1.1 equiv, DMF, 25 °C, 30 min)
followed by propargyl bromide (1.2 equiv, 0-25 °C, 5 h),
to generate 14 (85%, two steps), Scheme 2. Treatment of
N-Boc-4-aza-1,6-heptadiyne (13) and dimethyl acetylene-
dicarboxylate (4 equiv) with (PPh3)3RhCl (0.02 equiv,
EtOH, reflux, 18 h) followed by deprotection of the Boc
group afforded the desired isoindoline template 7 in 26%
overall yield.22
Syn th esis
Adopting a technically nondemanding multistep, solu-
tion-phase strategy for the preparation of chemical
libraries which relies on the removal of excess reactants
and reagents by liquid-liquid or liquid-solid extrac-
tions,19-21 two small libraries based on the template 7
were prepared. The approach highlights the ease with
which solution-phase mixture synthesis coupled with
scanning and deletion synthesis deconvolution, which is
conducted upfront for depository libraries subjected to
(14) Duggan, M. E.; Fisher, J . E.; Gentile, M. A.; Hartman, G. D.;
Hoffman, W. F.; Huff, J . R.; Ihle, N. C.; Krause, A. E.; Leu, T. C.; Nagy,
R. M.; Perkins, J . J .; Rodan, G. A.; Rodan, S. B.; Wesolowski, G.;
Whitman, D. B. Abstracts of papers, 211th ACS National Meeting, New
Orleans, LA, March 1996; American Chemical Society: Washington,
DC, 1996; MEDI 234.
(15) Corbett, J . W.; Graciani, N. R.; Mousa, S. A.; DeGrado, W. F.
Bioorg. Med. Chem. Lett. 1997, 7, 1371.
(16) Houghten, R. A.; Pinilla, C.; Appel, J . R.; Blondelle, S. E.;
Dooley, C. T.; Eichler, J .; Nefzi, A.; Ostresh, J . M. J . Med. Chem. 1999,
42, 3743.
(17) Dooley, C. T.; Houghten, R. A. Life Sci. 1993, 52, 1509. Pinilla,
C.; Appel, J . R.; Blanc, P.; Houghten, R. A. Biotechniques 1992, 13,
901.
(18) Boger, D. L.; Chai, W.; J in, Q. J . Am. Chem. Soc. 1998, 120,
7220.
(19) Boger, D. L.; Tarby, C. M.; Myers, P. L.; Caporale, L. H. J . Am.
Chem. Soc. 1996, 118, 2109.
(20) Cheng, S.; Comer, D. D.; Williams, J . P.; Myers, P. L.; Boger,
D. L. J . Am. Chem. Soc. 1996, 118, 2567.
(21) For recent applications: Boger, D. L.; Chai, W.; Ozer, R. S.;
Andersson, C.-M. Bioorg. Med. Chem. Lett. 1997, 7, 463. Boger, D. L.;
Ozer, R. S.; Andersson, C.-M. Bioorg. Med. Chem. Lett. 1997, 7, 1903.
Boger, D. L.; Chai, W. Tetrahedron 1998, 54, 3955. Boger, D. L.;
Ducray, P.; Chai, W.; J iang, W.; Goldberg, J . Bioorg. Med. Chem. Lett.
1998, 8, 2339. Boger, D. L.; Goldberg, J .; J iang, W.; Chai, W.; Ducray,
P.; Lee, J . K.; Ozer, R. S.; Andersson, C.-M. Bioorg. Med. Chem. 1998,
6, 1347. Boger, D. L.; Goldberg, J .; Andersson, C.-M. J . Org. Chem.
1999, 64, 2422. Boger, D. L.; J iang, W.; Goldberg, J . J . Org. Chem.
1999, 64, 7094.
(22) Grigg, R.; Scott, R.; Stevenson, P. J . Chem. Soc., Perkin. Trans
1 1988, 1357.

Positional Scanning and Deletion Synthesis Libraries
J . Org. Chem., Vol. 65, No. 5, 2000 1469
Sch em e 1
Sch em e 2
The approach outlined in Scheme 1 was shown to
dependably deliver pure individual compounds in large
quantities (25-60 mg), and two small libraries of 120
individual members were assembled. Each library mem-
ber contains a free carboxylic acid that mimics the
aspartic acid (D) in the RGD tripeptide and an amine at
the opposite end of the structure to mimic the arginine
(R), Figure 4. The varying length, rigidity, and structural
properties introduced by the template and the linkages
as well as variations in the basicity of the amino group
provided a rich array of potential RGD mimetics. Simul-
taneous deprotection of the amine and carboxylic acid
protecting groups provided 120 individual compounds
(library I), 6 scanning sublibraries for R1, 20 scanning
sublibraries for R2, 6 deletion sublibraries for R1, 20
deletion sublibraries for R2, and 1 full mixture of 120
compounds. Library II was prepared by the same method
but with the additional coupling of methylamine to the
remaining carboxylic acid of library I, affording another
set of 120 individual compounds (library II), 6 scanning
sublibraries for R1, 20 scanning sublibraries for R2, 6
deletion sublibraries for R1, 20 deletion sublibraries for
R2, and 1 full mixture of 120 compounds.
F igu r e 4.
1). The triamides 11 were prepared with sequential
addition of MeNH2‚HCl (2 equiv), EDCI (2.2 equiv, 2-2.2
equiv of i-Pr2NEt, DMF, 25 °C, 4 h, 25-95%) to 9,
affording another set of 120 individual compounds 11
denoted as library II (30-95%, Table 2). The lower
yielding reactions were compromised by the water solu-
bility of the product. The resulting diamides 9 and the
triamides 11 were treated with 3.9 M HCl in EtOAc to
deprotect simultaneously the amine and carboxylic acid
protecting groups with the formation of a white precipi-
tate, which was collected by filtration or evaporation of
solvent. Irrespective of the reaction efficiency or product
recovery and without deliberate reaction optimization,
the purities of the resulting individual compounds were
uniformly satisfactory (g80-95%) and the identities of
the products were confirmed by matrix characterization
To mimic the arginine, six different amino acids (A1-
A6) (R1CO2H) were attached to 7 with the water-soluble
coupling reagent EDCI (1-1.1 equiv, DMF, 15 h). Suc-
cessive washing of the crude products diluted in EtOAc
with aqueous acid (10% aqueous HCl) and saturated
aqueous NaHCO3 served to remove unreacted amine,
R1CO2H, and EDCI and its byproducts and provided the
1
(HRMS, H NMR, and IR).
For the formation of the triamides 11, we compared
the in situ formation without isolation of the diamide 9
with the stepwise process. Both afforded satisfactory
results, and the in situ process was more convenient.
Following the preparation of the individual compounds
(Tables 1 and 2), the full mixture libraries, 6 positional
scanning sublibraries of 9 and 11 which contain only A1,
A2, A3, A4, A5, or A6 (sca n A) and an equimolar mixture
of the 20 amines R2NH2 (Table 3), and 20 positional
scanning sublibraries, each of which contains one B
subunit from B1-B20 (sca n B) and an equimolar mix-
ture of the 6 acids R1CO2H (A1-A6) were produced
(Table 3). Deletion sublibraries (d eleteA) were also
1
pure momoamides 8 (>95% pure by H NMR). Each of
the six monoamides was converted to its dicarboxylic acid
by treatment with LiOH in THF/MeOH/H2O (3/1/1), and
then partitioned into 20 portions. Each portion was
treated with 1 of 20 different amines (B1-B20) (1.0
equiv) and PyBROP (1.05 equiv, 2-2.2 equiv of i-Pr2NEt,
DMF, 25 °C, 16 h, 25-100%) to afford the 120 individual
diamides 9 (library I), which were purified by sequential
10% aqueous HCl and saturated aqueous NaCl extrac-
tions from EtOAc to remove unreacted amine, unreacted
starting material, and PyBROP and its byproducts (Table

1470 J . Org. Chem., Vol. 65, No. 5, 2000
Ta ble 1. Yield s (%) a n d Am ou n ts (m g) for Com p ou n d s 8 a n d Dia m id es 9 (Libr a r y I)
Boger et al.
8 (A1) (88%)
8 (A2) (92%)
8 (A3) (87%)
8 (A4) (94%)
8 (A5) (99%)
8 (A6) (92%)
B1
B2
B3
B4
B5
B6
B7
B8
28 mg, 79%
19 mg, 53%
24 mg, 52%
29 mg, 57%
34 mg, 72%
31 mg, 60%
43 mg, 88%
46 mg, 92%
46 mg, 90%
35 mg, 67%
19 mg, 36%
33 mg, 59%
52 mg, 88%
27 mg, 50%
47 mg, 82%
53 mg, 90%
51 mg, 76%
31 mg, 56%
30 mg, 52%
13 mg, 22%
33 mg, 87%
30 mg, 81%
46 mg, 97%
38 mg, 72%
44 mg, 95%
22 mg, 40%
44 mg, 86%
40 mg, 79%
53 mg, 99%
41 mg, 75%
19 mg, 34%
39 mg, 68%
60 mg, 99%
24 mg, 47%
39 mg, 67%
48 mg, 79%
40 mg, 59%
30 mg, 53%
35 mg, 59%
33 mg, 53%
28 mg, 71%
29 mg, 73%
37 mg, 83%
32 mg, 60%
35 mg, 70%
43 mg, 78%
43 mg, 85%
46 mg, 88%
52 mg, 95%
38 mg, 68%
26 mg, 47%
42 mg, 71%
42 mg, 68%
37 mg, 65%
31 mg, 52%
49 mg, 79%
55 mg, 79%
33 mg, 58%
35 mg, 58%
25 mg, 39%
33 mg, 88%
35 mg, 90%
37 mg, 77%
31 mg, 57%
44 mg, 89%
44 mg, 81%
36 mg, 71%
32 mg, 62%
47 mg, 87%
46 mg, 83%
24 mg, 43%
43 mg, 74%
45 mg, 72%
34 mg, 61%
35 mg, 58%
45 mg, 72%
68 mg, 97%
36 mg, 62%,
37 mg, 61%
20 mg, 32%
28 mg, 71%
36 mg, 89%
46 mg, 92%
33 mg, 61%
43 mg, 84%
24 mg, 43%
34 mg, 65%
37 mg, 68%
54 mg, 97%
47 mg, 82%
35 mg, 62%
49 mg, 81%
58 mg, 91%
39 mg, 70%
31 mg, 51%
56 mg, 89%
58 mg, 82%
42 mg, 71%
43 mg, 70%
18 mg, 28%
36 mg, 89%
35 mg, 84%
40 mg, 79%
37 mg, 66%
41 mg, 79%
42 mg, 72%
47 mg, 88%
45 mg, 82%
54 mg, 95%
49 mg, 83%
35 mg, 59%
51 mg, 83%
57 mg, 87%
46 mg, 78%
46 mg, 73%
58 mg, 90%
61 mg, 84%
49 mg, 82%
49 mg, 78%
24 mg, 36%
B9
B10
B11
B12
B13
B14
B15
B16
B17
B18
B19
B20
Ta ble 2. Yield s (%) a n d Am ou n ts (m g) for Tr ia m id es 11 (Libr a r y II)
A1
A2
A3
A4
A5
A6
B1
B2
B3
B4
B5
B6
B7
B8
19 mg, 53%
26 mg, 71%
27 mg, 48%
35 mg, 69%
13 mg, 27%
32 mg, 61%
25 mg, 51%
32 mg, 64%
50 mg, 97%
34 mg, 64%
27 mg, 49%
45 mg, 78%
48 mg, 78%
40 mg, 72%
18 mg, 31%
47 mg, 77%
51 mg, 75%
25 mg, 43%
37 mg, 50%
14 mg, 22%
17 mg, 46%
36 mg, 96%
33 mg, 70%
34 mg, 65%
27 mg, 56%
34 mg, 64%
28 mg, 57%
35 mg, 69%
46 mg, 86%
37 mg, 68%
33 mg, 59%
39 mg, 67%
52 mg, 83%
30 mg, 53%
25 mg, 58%
48 mg, 77%
39 mg, 56%
34 mg, 57%
37 mg, 49%
14 mg, 21%
26 mg, 70%
27 mg, 68%
24 mg, 49%
36 mg, 67%
15 mg, 30%
34 mg, 61%
40 mg, 78%
38 mg, 72%
35 mg, 64%
40 mg, 71%
36 mg, 63%
46 mg, 77%
47 mg, 73%
31 mg, 53%
24 mg, 39%
51 mg, 81%
59 mg, 83%
34 mg, 57%
39 mg, 50%
25 mg, 38%
18 mg, 49%
17 mg, 43%
31 mg, 64%
34 mg, 64%
19 mg, 38%
31 mg, 56%
42 mg, 82%
44 mg, 84%,
46 mg, 84%
39 mg, 70%
32 mg, 56%
48 mg, 78%
46 mg, 72%
33 mg, 56%
20 mg, 33%,
51 mg, 80%
59 mg, 83%
33 mg, 56%
35 mg, 45%
19 mg, 29%
33 mg, 86%
31 mg, 78%
28 mg, 56%
36 mg, 65%
20 mg, 39%
40 mg, 71%
35 mg, 66%
38 mg, 71%
51 mg, 91%
39 mg, 68%
40 mg, 68%
54 mg, 86%
56 mg, 86%
37 mg, 63%
25 mg, 46%
52 mg, 81%
61 mg, 85%
37 mg, 60%
42 mg, 56%
17 mg, 26%
37 mg, 92%
40 mg, 96%
22 mg, 44%
42 mg, 74%
13 mg, 25%
36 mg, 63%
39 mg, 73%
46 mg, 83%
33 mg, 58%
44 mg, 74%
41 mg, 69%
51 mg, 80%
55 mg, 83%
25 mg, 42%
27 mg, 42%
60 mg, 91%
52 mg, 70%
47 mg, 75%
49 mg, 61%
29 mg, 43%
B9
B10
B11
B12
B13
B14
B15
B16
B17
B18
B19
B20
Ta ble 3. Sca n n in g Su blibr a r ies of 9 a n d 11, Syn th esis Yield sa (%) a n d Am ou n ts (m g)
sca n A1b
sca n A2
sca n A3
sca n A4
sca n A5
sca n A6
Am ix
library I
library II
32 mg, 60%
30 mg, 55%
44 mg, 80%
34 mg, 61%
39 mg, 67%
41 mg, 72%
29 mg, 56%
34 mg, 60%
37 mg, 64%
33 mg, 56%
48 mg, 83%
21 mg, 35%
35 mg, 63%
38 mg, 66%
sca n B1c
sca n B2
sca n B3
sca n B4
sca n B5
sca n B6
sca n B7
library I
library II
16 mg, 80%
15 mg, 72%
19 mg, 91%
17 mg, 79%
18 mg, 89%
12 mg, 58%
20 mg, 72%
41 mg, 73%
23 mg, 87%
9 mg, 32%
23 mg, 80%
20 mg, 76%
23 mg, 85%
23 mg, 85%
sca n B8
sca n B9
sca n B10
sca n B11
sca n B12
sca n B13
sca n B14
library I
library II
23 mg, 86%
23 mg, 84%
25 mg, 91%
23 mg, 81%
23 mg, 80%
18 mg, 63%
19 mg, 63%
19 mg, 62%
23 mg, 74%
23 mg, 73%
24 mg, 79%
27 mg, 82%
18 mg, 61%
19 mg, 65%
sca n B15
sca n B16
sca n B17
sca n B18
sca n B19
sca n B20
library I
library II
21 mg, 69%
14 mg, 46%
24 mg, 77%
23 mg, 72%
30 mg, 84%
26 mg, 73%
20 mg, 67%
18 mg, 58%
22 mg, 71%
18 mg, 56%
15 mg, 46%
13 mg, 38%
a
b
Calculated on the basis of the average molecular weight for the mixture (reaction scale, 0.1 mmol). Mixture synthesis but including
only Xn for sca n Xn ; e.g. sca n A1 includes only A1 and B1-B20 (reaction scale, 0.05 mmol). c Mixture synthesis but including only Xn
for sca n Xn ; e.g., sca n B1 includes only B1 and A1-A6 (reaction scale, 0.05 mmol).
prepared using the same protocol by the reaction of an
equimolar mixture of 5 of 6 monoamides 8 and an
equimolar mixture of the 20 amines R2NH2 (B1-B20)
producing 6 sublibraries of 100 compounds, each lacking
only one of the A subunits. The 20 B deletion sublibraries
(d eleteB) were prepared by the reaction of an equimolar
mixture of the six monoamides 8 and 19 of the 20 amines
R2NH2, producing 20 sublibraries of 114 compounds, each
lacking only one of the B subunits (Table 4).
Biologica l Assa y: Cytotoxic Activity. The 240
individual compounds 9 and 11 constituting the protected
precursors to libraries I and II as well as the correspond-
ing mixture libraries and the scanning and deletion
synthesis deconvolution libraries were assayed for in

Positional Scanning and Deletion Synthesis Libraries
J . Org. Chem., Vol. 65, No. 5, 2000 1471
Ta ble 4. Deletion Su blibr a r ies of Libr a r ies I a n d II, Syn th esis Yield sa (%) a n d Am ou n ts (m g)
d eleteA1b
d eleteA2
d eleteA3
d eleteA4
d eleteA5
d elete A6
Am ix
library I
library II
20 mg, 70%
19 mg, 66%
19 mg, 67%
21 mg, 75%
18 mg, 65%
15 mg, 68%
17 mg, 60%
19 mg, 65%
20 mg, 70%
16 mg, 60%
22 mg, 78%
21 mg, 75%
35 mg, 63%
38 mg, 66%
d eleteB1
d eleteB2
d eleteB3
d eleteB4
d eleteB5
d eleteB6
d eleteB7
library I
library II
22 mg, 77%
21 mg, 74%
17 mg, 60%
26 mg, 93%
22 mg, 78%
18 mg, 65%
21 mg, 75%
18 mg, 63%
26 mg, 90%
20 mg, 70%
24 mg, 86%
19 mg, 67%
22 mg, 79%
23 mg, 80%
d eleteB8
d eleteB9
d eleteB10
d eleteB11
d eleteB12
d eleteB13
d eleteB14
library I
library II
22 mg, 78%
17 mg, 60%
20 mg, 70%
23 mg, 79%
24 mg, 86%
23 mg, 73%
18 mg, 65%
23 mg, 73%
20 mg, 72%
18 mg, 64%
24 mg, 86%
21 mg, 73%
23 mg, 83%
19 mg, 66%
d eleteB15
d eleteB16
d eleteB17
d eleteB18
d eleteB19
d eleteB20
library I
library II
21 mg, 76%
11 mg, 39%
19 mg, 69%
23 mg, 80%
18 mg, 65%
19 mg, 68%
15 mg, 55%
17 mg, 60%
17 mg, 63%
18 mg, 13%
14 mg, 49%
15 mg, 54%
a
b
Calculated on the basis of the average molecular weight for the mixture (reaction scale, 0.1 mmol). Mixture synthesis but deleting
only Xn for d eleteXn ; e.g., d eleteA1 includes A2-A6 and B1-B20 (reaction scale, 0.06 mmol).
Ta ble 5. Cytotoxic Activity (L-1210 IC50) of Mixtu r e, Sca n n in g, a n d Deletion Decon volu tion Su blibr a r ies for 9
a
b
L-1210 (mouse leukemia) cytotoxic activity. Structure 9, sca n A1 mixture includes A1 and B1-B20. c Structure 9, d eleteA1 mixture
includes A2-A6 and B1-B20.
vitro cytotoxic activity in a L-1210 assay,18,23 Tables 5-8.
Active compounds may be recognized from a gain in
activity within the scanning deconvolution sublibraries
or a loss in activity in the deletion synthesis deconvolu-
tion sublibraries relative to the full mixtures. Several
elements of the comparisons are worth noting. First, the
IC50 values established for the mixtures corresponded
beautifully to the average values calculated from the
components of the mixtures where this was possible.
Thus, in this particular assay, the components of the
mixtures are not acting independently. Rather, and
perhaps because of their structural similarity and likeli-
hood of acting at a common target with near equivalent
potencies, they act cooperatively. With library I, scanning
deconvolution revealed the second and fifth most potent
agents, which were essentially indistinguishable from the
third and fourth most potent agents. It did not, however,
adequately detect the most potent agent 9, A4B7. None-
theless, potent leads relative to the remainder of the
library were detected. In the case of library I, deletion
synthesis deconvolution uniquely identified A4B7 as the
most potent agent and further identified A4B13, the
fourth most potent agent. Thus, deletion synthesis de-
convolution proved more sensitive to identifying the most
potent library member. More significantly, the two
protocols together revealed the identity of four of the top
five most potent members of the library, and both
protocols identified B7 and B13 as key residues. This
level of lead identification from the mixture libraries
would have required the synthesis of three compounds
based on the scanning deconvolution results (A5B7,
A513, and A5B17, two of which would have been active)
(23) The assay was run 2-3 times each in duplicate, and the results
are reported as an average IC50, standard deviation (10% with each
assay. Boger, D. L.; Yasuda, M.; Mitscher, L. A.; Drake, S. D.; Kitos,
P. A.; Thompson, S. C. J . Med. Chem. 1987, 30, 1918.

1472 J . Org. Chem., Vol. 65, No. 5, 2000
Boger et al.
Ta ble 6. Cytotoxic Activity (IC50, µM) for In d ivid u a l
the scanning deconvolution results from library I would
have been required before A4B7 would have been identi-
fied. The screening of 9-12 in other assays is in progress,
and the results of the comparative performance of the
scanning and deletion synthesis deconvolution will be
forthcoming.
Com p ou n d s 9
Exp er im en ta l Section
N-(ter t-Bu tyloxy)ca r bon yl Dip r op a r gyla m in e (14). A
solution of N-(tert-butyloxy)carbonyl propargylamine (13; 33.36
g, 215 mmol) in 50 mL of DMF was treated portionwise (4×)
with 60% NaH (10.4 g) at 0 °C. After being stirred for 30 min
at 25 °C, 39 mL of an 80% solution of propargyl bromide in
toluene was added. The reaction mixture was stirred for an
additional 5 h at 25 °C, and then quenched with the addition
of ice-water. The mixture was extracted with Et2O (3 × 200
mL), and the combined extracts were washed with saturated
aqueous NaCl, dried (Na2SO4), and concentrated in vacuo. The
resulting residue was fractionally distilled under high vacuum
(57-60 °C/0.05 Torr or 76 °C/0.1 Torr) to give 34.5 g of 14 as
a colorless liquid: 1H NMR (CDCl3, 500 MHz) δ 4.16 (s, 4H),
2.21 (t, J ) 2.0 Hz, 2H), 1.47 (s, 9H); IR (film) νmax 2976, 2962,
1725 cm-1, FABHRMS (NBA-NaI) m/z 216.1006 (M + Na+,
C
11H15NO2 requires 216.1000).
Dim eth yl Isoin d olin e-5,6-d ica r boxyla te (7). A solution
of 14 (10.4 g, 53.9 mmol) and dimethyl acetylenedicarboxylate
(30.7 g, 216 mmol) in 110 mL of absolute EtOH was degassed
by bubbling N2 through the solution for 10 min. To this was
added 1.0 g (0.02 equiv) of Wilkinson’s catalyst [(Ph3P)3RhCl]
at 25 °C. After being warmed at reflux for 18 h, the reaction
mixture was cooled to 25 °C and then concentrated in vacuo.
The resulting brown residue was diluted in 200 mL of Et2O,
and the precipitate was removed by filtration over Celite. The
filtrate was concentrated and the crude product purified by
column chromatography (SiO2, 20% EtOAc/hexane) to give 4.60
g (26%) of the dimethyl N-(tert-butyloxycarbonyl) isoindoline-
1
5,6-dicarboxylate as a white solid: mp 121-123 °C; H NMR
(CDCl3, 400 MHz) δ 7.62 and 7.56 (two s, 2H), 4.70 and 4.67
(two s, 4H), 3.88 (s, 6H), 1.50 (s, 9H); IR (film) νmax 2976, 2962,
1731, 1698 cm-1; FABHRMS (NBA-NaI) m/z 358.1277 (M +
Na+, C17H21NO6 requires 358.1267). A solution of the N-Boc-
isoindoline (4.6 g, 13.7 mmol) in 50 mL of CH2Cl2 was treated
with 20 mL of 3.9 M HCl solution in EtOAc at 25 °C, resulting
in a pale gray precipitate. After being stirred for 5 h at 25 °C,
the resulting precipitate was collected by filtration, rinsed with
CH2Cl2, and then dried under vacuum to give 3.5 g (95%) of
the HCl salt of 7 as a gray powder: 1H NMR (CDCl3, 400 MHz)
δ 7.71 (s, 2H), 4.61 (s, 4H), 3.79 (s, 6H); IR (film) νmax 3310,
2985, 2967, 1698 cm-1; FABHRMS (NBA-NaI) m/z 236.0923
(M + Na+, C12H13NO4 requires 236.0923).
and two compounds from the deletion synthesis decon-
volution (A4B7 and A4B13, both of which would have
been active).
Gen er a l P r oced u r e for t h e F ir st F u n ct ion a liza t ion ,
P r ep a r a tion of 8 (A1). A solution of 7 (0.81 g, 3.0 mmol),
R1CO2H [A1, 4-((tert-butyloxycarbonyl)amino)butanoic acid
(0.60 g, 3 mmol)], and i-Pr2NEt (1.23 mL, 2 equiv) in anhydrous
DMF (30 mL) was treated with EDCI (0.58 g, 1 equiv). After
being stirred for 16 h at 25 °C, the reaction mixture was poured
into a separatory funnel containing 50 mL of 10% aqueous
HCl. Extraction with EtOAc (2 × 50 mL) followed by washing
of the combined organic extracts with 10% aqueous HCl,
saturated aqueous NaHCO3 (2 × 50 mL), and saturated
aqueous NaCl, drying over Na2SO4, and evaporation provided
1.10 g (87%) of the monoamide 8 (A1)24 as a foamy solid.
Gen er a l P r oced u r e for P r ep a r a tion of th e Dia m id es
9 a n d 10 (Libr a r y I) (A1B1). A solution of 8 (A1; 1.10 g, 2.6
mmol) in 16 mL of 25% MeOH-THF was treated with a
solution of LiOH‚H2O (0.43 g, 4 equiv) in 4 mL of H2O at 25
°C. After being stirred for 2-3 h at 25 °C, the reaction mixture
was diluted with H2O (20 mL) and poured into a separatory
funnel. The mixture was washed with EtOAc to remove the
unreacted starting material, and then the aqueous phase was
With library II, both scanning and deletion synthesis
deconvolution would have identified the most potent
library member 11, A5B17. In this instance, scanning
deconvolution uniquely identified this agent as the most
potent, while the deletion deconvolution identification
was not unique. The deconvolution of the leads from the
library II scanning (one compound, A5B17) and deletion
synthesis (three compounds, A5B6, A5B14, and the same
A5B17) libraries would have required the individual
synthesis of only three compounds.
Although a more complete assessment of the more
weakly active leads could be derived from an examination
of the weaker scanning or more active deletion synthesis
libraries, they would not have improved the performance
at identifying the most active compounds in the mixtures.
More importantly, the same level of confidence in the
coverage provided by either deconvolution protocol alone
was not observed and would require a more extensive
deconvolution to ensure adequate identification. For
example, the preparation of 9-12 compounds based on
(24) Characterization data are provided in the Supporting Informa-
tion.

Positional Scanning and Deletion Synthesis Libraries
J . Org. Chem., Vol. 65, No. 5, 2000 1473
Ta ble 7. Cytotoxic Activity of Mixtu r e, Sca n n in g, a n d Deletion Decon volu tion Su blibr a r ies for 11
acidified with 10% aqueous HCl. Extraction into 70 mL of
CHCl3/i-PrOH (1/1) followed by washing of the organic phase
with saturated aqueous NaCl, drying over Na2SO4, and
evaporation provided 0.92 g (90%) of the diacid as a pale yellow
powder: 1H NMR (CD3OD, 400 MHz) δ 7.71 (s, 1H), 7.70 (s,
1H), 4.94 (s, 2H), 4.79 (s, 2H), 3.12 (d, J ) 6.8 Hz, 2H), 2.47 (t,
J ) 7.2 Hz, 2H), 1.75 (quintet, J ) 7.0 Hz, 2H), 1.41 (s, 9H);
IR (film) νmax 3336, 2952, 2867, 1724, 1646 cm-1; FABHRMS
(NBA-CsI) m/z 553.0929 (M + Cs+, C19H24N2O7 requires
553.0951). A solution of the diacid (39 mg, 0.1 mmol) and i-Pr2-
NEt (65 µL, 4 equiv) in anhydrous DMF (2 mL) was treated
with PyBROP (50 mg, 1.1 equiv) at 25 °C. After being stirred
for 1 h at 25 °C, Gly-Ot-Bu (B1; 19 mg, 1.1 equiv) was added,
and the mixture was stirred for an additional 15 h at 25 °C.
The reaction mixture was poured into a separatory funnel
containing 20 mL of 10% aqueous HCl. Extraction into EtOAc
(2 × 20 mL) followed by washing of the combined organic
portions with 10% aqueous HCl (4 × 20 mL) and saturated
aqueous NaCl, drying over Na2SO4, and evaporation provided
46 mg (92%) of 9 (A1B1) as a foamy solid: 1H NMR (CD3OD,
500 MHz) δ 7.91 (d, J ) 9.2 Hz, 1H), 7.50 (d, J ) 9.2 Hz, 1H),
4.93 (d, J ) 4.3 Hz, 2H), 4.79 (d, J ) 4.8 Hz, 2H), 3.97 (s, 2H),
3.12 (t, J ) 6.8 Hz, 2H), 2.45 (t, J ) 7.2 Hz, 2H), 1.82 (quintet,
J ) 7.0 Hz, 2H), 1.48 (s, 9H), 1.40 (s, 9H); IR (film) νmax 3325,
2978, 2933, 1711, 1654 cm-1; FABHRMS (NBA-CsI) m/z
638.1499 (M + Cs+, C25H39N3O8 requires 638.1478). A solution
of 9 (A1B1; 46 mg, 0.92 mmol) in 0.5 mL of EtOAc was treated
with 1.5 mL of a 3.9 M HCl solution in EtOAc. After being
stirred for 16 h at 25 °C, 2 mL of Et2O was added to the
resulting white suspension to precipitate the product. The
mixture was allowed to stand for 30 min before the solvent
was removed by decantation. Evaporation of the residual
solvent under a N2 stream followed by drying under vacuum
provided 1024 (A1B1; 32 mg, 100%) as a fine powder.
added i-Pr2NEt (65 µL, 4 equiv), methylamine hydrochloride
(13 mg, 2 equiv), and EDCI (38 mg, 2 equiv) at 25 °C. The
reaction mixture was stirred for an additional 5 h and then
poured into a separatory funnel containing 10 mL of 10%
aqueous HCl. Extraction into EtOAc (2 × 20 mL) followed by
washing of the combined organic portions with 10% aqueous
HCl (4 × 20 mL), saturated aqueous NaHCO3 (2 × 20 mL),
and saturated aqueous NaCl, drying over Na2SO4, and evapo-
ration provided 45 mg (86%) of 11 (A1B1) as a foam: 1H NMR
(CD3OD, 500 MHz) δ 7.87-7.49 (m, 2H), 4.95 (d, J ) 4.3 Hz,
2H), 4.79 (d, J ) 4.8 Hz, 2H), 3.97 (s, 2H), 3.14 (t, J ) 6.8 Hz,
2H), 2.86 (s, 3H), 2.49 (t, J ) 7.2 Hz, 2H), 1.84 (quintet, J )
7.0 Hz, 2H), 1.48 (s, 9H), 1.40 (s, 9H); IR (film) νmax 3325, 2978,
2933, 1711, 1654 cm-1; FABHRMS (NBA-CsI) m/z 651.1445
(M + Cs+, C26H42N4O7 requires 638.1468). A solution of 11
(A1B1; 40 mg, 0.08 mmol) in 0.3 mL of EtOAc was treated
with 1 mL of a 3.9 M HCl solution in EtOAc at 25 °C, resulting
in a white precipitate. After being stirred for 16 h at 25 °C,
the solvent and excess HCl were removed by evaporation under
a N2 stream to give 1224 (A1B1; 28 mg, 100%) as a powder.
Gen er a l P r oced u r e for th e A-Sca n (Libr a r y I). A
solution of the diacid derived from 8 (A1; 39 mg, 0.1 mmol) in
DMF was treated with PyBROP (57 mg, 1.1 equiv) and i-Pr2-
NEt (65 µL, 4 equiv) at 25 °C. After being stirred for 30 min,
the reaction mixture was treated with 1 mL of a 0.1 M solution
of the 20 amines B1-B20 in DMF. After being stirred for 16
h at 25 °C, the reaction mixture was poured into a separatory
funnel containing 20 mL of 10% aqueous HCl. Extraction into
EtOAc (2 × 20 mL) followed by washing of the combined
organic portions with 10% aqueous HCl (4 × 20 mL) and
saturated aqueous NaCl, drying over Na2SO4, and evaporation
provided 43 mg of 9, sca n A1 mixture as a foamy solid. A
solution of this bisprotected sca n A1 mixture (43 mg, 0.62
mmol, average MW 700) in 0.5 mL of EtOAc was treated with
1.5 mL of a 3.9 M HCl solution in EtOAc. After being stirred
for 16 h at 25 °C, 2 mL of Et2O was added into the resulting
white suspension to precipitate the product. The mixture was
allowed to stand for 30 min before the solvent was removed
by decantation. Evaporation of the residual solvent under a
N2 stream followed by drying under vacuum provided the 10,
sca n A1 library (32 mg, 100%) as a fine powder.
Gen er a l P r oced u r e for P r ep a r a tion of th e Tr ia m id es
11 a n d 12 (Libr a r y II) (A1B1). A solution of the diacid
prepared from 8 (A1; 39 mg, 0.1 mmol), and i-Pr2NEt (65 µL,
4 equiv) in anhydrous DMF (2 mL) was treated with PyBROP
(50 mg, 1.1 equiv). After being stirred for 1 h at 25 °C, Gly-
Ot-Bu (B1; 19 mg, 1.1 equiv) was added, and the mixture was
stirred for an additional 15 h at 25 °C. To this mixture were

1474 J . Org. Chem., Vol. 65, No. 5, 2000
Boger et al.
Ta ble 8. Cytotoxic Activity (IC50, µM) for In d ivid u a l
sion to precipitate the product. The mixture was allowed to
stand for 30 min followed by the removal of the solvent by
decantation. Evaporation of the residual solvent under a N2
stream followed by drying under vacuum provided the desired
product (16 mg) as a fine powder.
Com p ou n d s 11
Gen er a l P r oced u r e for th e Deletion Decon volu tion
Libr a r y, DeleteA1 (Libr a r y I). A stock solution (0.05 M) of
B1-B20 was prepared in DMF, and a 0.1 M solution in DMF
of each diacid derived from monoamide 8 was prepared. A 1
dram vial was charged with 0.1 mL of each of the five diacid-
stock solutions (A2-A6), i-Pr2NEt (30 µL, 4 equiv), and
PyBROP (28 mg, 1.1 equiv). After 30 min, it was treated with
1 mL (0.05 M) of the B1-B20 mixture solution. After being
stirred for 16 h at 25 °C, the reaction mixture was poured into
a separatory funnel containing 20 mL of 10% aqueous HCl.
Extraction into EtOAc (2 × 20 mL) followed by washing of the
combined organic portions with 10% aqueous HCl (4 × 20 mL)
and saturated aqueous NaCl, drying over Na2SO4, and evapo-
ration provided 25 mg (70%, average MW 716) of the bispro-
tected d eleteA1 mixture of library I as a foamy solid. This
mixture was treated with 1.5 mL of a 3.9 M HCl solution in
EtOAc. After being stirred for 16 h at 25 °C, 2 mL of Et2O
was added into the resulting white suspension to precipitate
the product. The mixture was allowed to stand for 30 min
before the solvent was removed by decantation. Evaporation
of the residual solvent by a N2 stream followed by drying under
vacuum provided the desired product (20 mg) as a fine powder.
Gen er a l P r oced u r e for th e Deletion Decon volu tion
Libr a r y, DeleteB1 (Libr a r y I). A stock solution (0.05 M) of
each amine from B1 to B20 was prepared by dilution in DMF,
and a 0.1 M solution in DMF of each diacid derived from
monoamide 8 was also prepared. A 1 dram vial was charged
with 80 µL of the 0.1 M solution of each of the six diacids A1-
A6, i-Pr2NEt (30 µL, 4 equiv), and PyBROP (28 mg, 1.1 equiv).
After 30 min, 50 µL (2.5 µmol) of each of the 19 amines B2-
B20 was added to the mixture. After being stirred for 16 h at
25 °C, the reaction mixture was poured into a separatory
funnel containing 20 mL of 10% aqueous HCl. Extraction into
EtOAc (2 × 20 mL) followed by washing of the combined
organic portions with 10% aqueous HCl (4 × 20 mL) and
saturated aqueous NaCl, drying over Na2SO4, and evaporation
provided 25 mg (70%, average MW 716) of the 9, d eleteA1
library I as a foam. This product was treated with 1.5 mL of
3.9 M HCl solution in EtOAc. After being stirred for 16 h at
25 °C, 2 mL of Et2O was added into the resulting white
suspension to precipitate the product. The mixture was allowed
to stand for 30 min before the solvent was removed by
decantation. Evaporation of the residual solvent by a N2
stream followed by drying under vacuum provided 10, deleteA1
(20 mg) as a fine powder.
Gen er a l P r oced u r e for th e B-Sca n (Libr a r y I). A
solution which contains 0.01 mmol of each of the six diacids
(total 0.06 mmol) derived from 8 (A1-A6) in DMF (1.2 mL)
was treated with i-Pr2NEt (32 µL, 4 equiv) and PyBROP (30
mg, 1.1 equiv) at 25 °C. After 30 min, it was treated with amine
B1 (10 mg, 1.1 equiv) and then stirred for 16 h at 25 °C. The
reaction mixture was poured into a separatory funnel contain-
ing 20 mL of 10% aqueous HCl. Extraction into EtOAc (2 ×
20 mL) followed by washing of the combined organic portions
with 10% aqueous HCl (4 × 20 mL) and saturated aqueous
NaCl, drying over Na2SO4, and evaporation provided 22 mg
of the 9, sca n B1 mixture of library I as a foamy solid. A
solution of this bisprotected sca n B1 mixture (22 mg, average
MW 560) in 0.5 mL of EtOAc was treated with 1 mL of a 3.9
M HCl solution in EtOAc. After being stirred for 16 h at 25
°C, 2 mL of Et2O was added into the resulting white suspen-
Ack n ow led gm en t. We gratefully acknowledge the
financial support of the National Institutes of Health
(Grant CA 78045).
Su p p or tin g In for m a tion Ava ila ble: 1H NMR spectra of
7 and characterization data and 1H NMR spectra for 8 (A1-
A6) and 9-12 [A1B1, A2B2, A3B3, A4B4, A5B5, A6B6,
A1B7, A2B8, A3B9, A4B10, A5B11, A6B12, A1B13, A2B14,
A3B15, A4B16, A5B17 (9, A3B17), A6B18 (9, A2B18), A1B19
(9, A5B19), A2B20]. This material is available free of charge
via the Internet at http://pubs.acs.org.
J O9916481