Small molecule macroarray construction via Ugi four-component reactions

Article abstract of DOI:10.1021/ol051684o

(Chemical Equation Presented) We report the construction of small molecule macroarrays via Ugi four-component reactions on planar cellulose supports. Array synthesis was enabled by the development of a high efficiency photocleavable linker system and the

Full text of DOI:10.1021/ol051684o

ORGANIC
LETTERS
2005
Vol. 7, No. 20
4455-4458
Small Molecule Macroarray Construction
via Ugi Four-Component Reactions
Qi Lin, Jennifer C. O’Neill, and Helen E. Blackwell*
Department of Chemistry, UniVersity of WisconsinsMadison, 1101 UniVersity AVenue,
Madison, Wisconsin 53706-1322
blackwell@chem.wisc.edu
Received July 18, 2005
ABSTRACT
We report the construction of small molecule macroarrays via Ugi four-component reactions on planar cellulose supports. Array synthesis
was enabled by the development of a high efficiency photocleavable linker system and the strategic use of both water- and microwave-
assisted organic reactions.
Research in genomics and proteomics is delivering an
increasing number of new biological targets.1 The use of
small molecule probes to study and further annotate these
targets has attracted considerable interest.2 Access to such
probes is contingent on the development of robust methods
for their synthesis and biological evaluation. Combinatorial
synthesis on polymeric solid supports remains one of the
most common routes for the construction of small molecule
libraries.3 Standard polystyrene resins, however, are often
expensive and difficult to manipulate, and reactions per-
formed on these supports can exhibit reduced rates. Further,
these hydrophobic supports are frequently incompatible with
biological assays performed directly on immobilized com-
pounds. Recently, we developed a library synthesis platform
using planar cellulose supports that addresses these chal-
lenges.4 This synthetic approach is based on microwave-
assisted SPOT-synthesis5 and permits the rapid construction
of small molecule “macroarrays”. These lower density arrays
are complementary to small molecule microarrays,6 yet differ
in two ways: (1) compound synthesis is performed directly
on the planar support, and (2) sufficient compound is
generated per spot for characterization and biological evalu-
ation post-cleavage. Here, we report the use of multiple-
component reactions for the efficient construction of small
molecule macroarrays. Further, we demonstrate that the
pretreatment of the support with small volumes of water in
these reactions,7 as opposed to the use of microwave
irradiation, dramatically increases product conversion and
purity.
We were interested in evaluating multiple-component
reactions (MCRs) for macroarray construction because these
reactions can introduce structural diversity with high ef-
ficiency, joining three or more reactants in a single step.
(5) Frank, R. Tetrahedron 1992, 48, 9217-9232.
(6) For a review, see: Uttamchandani, M.; Walsh, D. P.; Yao, S. Q.;
Chang, Y.-T. Curr. Opin. Chem. Biol. 2005, 9, 4-13.
(7) (a) For a recent report of water-promoted reactions, see: Narayan,
S.; Muldoon, J.; Finn, M. G.; Fokin, V. V.; Kolb, H. C.; Sharpless, K. B.
Angew. Chem., Int. Ed. 2005, 44, 3275-3279. (b) For reviews, see:
Breslow, R. Acc. Chem. Res. 1991, 24, 159-164. (c) Lubineau, A.; Auge,
J. Top. Curr. Chem. 1999, 206, 1-39.
(1) (a) Kramer, R.; Cohen, D. Nat. ReV. Drug DiscoVery 2004, 3, 965-
972. (b) Lindsay, M. A. Nat. ReV. Drug DiscoVery 2003, 2, 831-838.
(2) (a) Kaiser, J. Science 2004, 304, 1728. (b) Schreiber, S. L. Chem.
Eng. News 2003, 81, 51-61.
(3) Yu, Z.; Bradley, M. Curr. Opin. Chem. Biol. 2002, 6, 347-352.
(4) Bowman, M. D.; Jeske, R. C.; Blackwell, H. E. Org. Lett. 2004, 6,
2019-2022.
10.1021/ol051684o CCC: $30.25
© 2005 American Chemical Society
Published on Web 09/03/2005

MCRs are suited uniquely for combinatorial library synthesis,
as evidenced by considerable recent work in this area.8 The
Ugi 4CR represents one of the best studied MCRs and has
been performed on traditional solid supports and in numerous
combinatorial applications.9 In the Ugi 4CR, an amine
component reacts with an acid, an aldehyde or ketone, and
an isocyanide to produce an R-acylamino amide. While
commonly high yielding, this reaction can suffer from
exceptionally long reaction times, especially when performed
on solid-phase bound substrates (i.e., on the order of days).10
To accelerate the reaction rate, Hoel and Nielsen recently
applied microwave (MW) irradiation during Ugi 4CRs on
solid-phase bound substrates and observed dramatically
shortened reaction times (e5 min).11 MW irradiation con-
tinues to emerge as a powerful method to accelerate a broad
range of solid-phase reactions.12 Therefore, on the basis of
these past reports, we chose to examine MW-assisted reaction
conditions for Ugi 4CRs performed in a macroarray format.
linker 4 to planar cellulose, we derivatized the cellulose
(Whatman 1Chr filter paper, 1) with a flexible diamine spacer
unit (2) to enhance its reactivity (Scheme 1).4,14 The spacer
was attached via tosylation of cellulose membrane 1 followed
by submersion of the membrane in neat diamine 2 and
subjection to MW heating (300 W, 15 min) in a commercial
MW reactor.15 This MW-assisted displacement reaction
generated support 3 with reproducible amine spacer loadings
of ca. 4 µmol/cm2 (as determined by UV Fmoc quantita-
tion).16
Next, photolabile linker 4 was attached to amine spacer
support 3 via a standard carbodiimide-mediated coupling
reaction (Scheme 1). We applied linker 4 to support 3 in an
arrayed “spot” format (0.3 cm2/spot), as opposed to submers-
ing the entire membrane in a solution of 4, to minimize the
use of this expensive reagent. Immediately after linker
application, the membrane was placed between two conduc-
tive, polymer composite plates (Weflon)17 and subjected to
MW irradiation for 8 min at 80 °C.18 We have found that
this Weflon sandwiching procedure is a convenient method
to heat planar polymeric membranes and can enhance
conversion in numerous MW-assisted reactions. Using this
method, linker loadings of ca. 450 nmol/cm2 were generated
(as determined by UV Fmoc quantitation).16,19 Following
linker loading, the membrane was acetylated (to “cap” any
unreacted spacer amines) and subjected to Fmoc deprotection
to generate 5.
We selected the photolabile linker 4-{4-[1-(Fmoc-amino)-
ethyl]-2-methoxy-5-nitrophenoxy}butanoic acid (4, Scheme
1) for macroarray construction, as this linker has been well
Scheme 1. Support Modification and Linker Installationa
We chose to immobilize the amine building block on
support 5 for the Ugi 4CR and found N-Fmoc-4-nitro-L-
phenylalanine (Fmoc-Nph (6), Scheme 2) to be an excellent
Scheme 2. Loading of Fmoc-Nph onto Support 5
a Reagents and conditions: TsCl ) tosyl chloride; MW )
microwave irradiation; DIC ) N,N′-diisopropylcarbodiimide; HOSu
) hydroxy succinimide; DIPEA ) diisopropylethylamine; NMP
) 1-methyl-2-pyrrolidinone; Ac2O ) acetic anhydride; DMF )
dimethylformamide.
amine substrate, as its strong absorbance at 280 nm allowed
for straightforward UV quantification. Fmoc-Nph (6) was
coupled efficiently to the linker spots of support 5 using a
MW-assisted protocol similar to that used for linker attach-
studied and has proven to be effective in prior SPOT-
syntheses.13 In addition, photolabile linkers provide advan-
tages for postsynthetic macroarray screening, as library
members remain spatially addressed after solvent-free cleav-
age with UV light. For example, this “dry-state” cleavage
permits bacteriological agar overlay assays to be performed
directly on the array. Prior to the attachment of photolabile
(14) See Supporting Information for full experimental details of MW-
assisted reactions, array construction, and compound characterization.
(15) A multimodal MW reactor was used throughout this work.
(16) Carpino, L. A.; Han, G. Y. J. Org. Chem. 1972, 37, 3404-3409.
(17) Plate size: 20 × 20 × 0.5 cm. Weflon is a composite of Teflon
and carbon black particles and commercially available from Milestone, Inc.
(18) This protocol was developed using temperature-controlled MW
heating methods by sandwiching a fiber optic temperature probe between
two Weflon plates (one with a groove that fit the probe), ramping to a
desired temperature using MW irradiation, and then holding at that
temperature for a specified amount of time. The average wattages obtained
during the ramp and hold times were then used in SPOT reactions to
reproduce these heating curves.
(8) For a review, see: Ugi, I. Pure Appl. Chem. 2001, 73, 187-191.
(9) Armstrong, R. W.; Combs, A. P.; Tempest, P. A.; Brown, S. D.;
Keating, T. A. Acc. Chem. Res. 1996, 29, 123-131.
(10) (a) Tempest, P. A.; Brown, S. D.; Armstrong, R. W. Angew. Chem.,
Int. Ed. Engl. 1996, 35, 640-642. (b) Short, K. M.; Ching, B. W.; Mjalli,
A. M. M. Tetrahedron 1997, 53, 6653-6679.
(11) Hoel, A. M. L.; Nielsen, J. Tetrahedron Lett. 1999, 40, 3941-3944.
(12) Blackwell, H. E. Org. Biomol. Chem. 2003, 1, 1251-1255.
(13) See: Scharn, D.; Germeroth, L.; Schneider-Mergener, J.; Wenschuh,
H. J. Org. Chem. 2001, 66, 507-513 and references therein.
(19) Control reactions performed at 80 °C in an oven for 10 min gave
ca. 50% lower loadings.
4456
Org. Lett., Vol. 7, No. 20, 2005

ment (10 min at 100 °C, on Weflon). This extended MW-
assisted coupling reaction gave consistent loadings of 6 at
ca. 400 nmol/cm2.16,18,19 Subsequent acetylation of unreacted
amine linker and Fmoc deprotection of Nph yielded amine
support 7.
Table 1. Selected Reaction Conditions for Ugi 4CR on Nph
Support 7 to Generate Compound 8
Prior to macroarray construction, tests were performed to
examine the efficiency of H-Nph-NH2 photocleavage from
planar support 7. We examined the photocleavage of
N-acetyl-Nph-NH2, as this derivative exhibited enhanced
stability to cleavage conditions relative to that of H-Nph-
NH2. Various solvents, times, and light sources were
evaluated; optimal cleavage conditions were achieved when
“punched-out” spots (isolated with a desktop hole punch)
were irradiated at 366 nm in methanol for 16 h using an
upturned hand-held UV lamp.20 Using this solution-phase
method, cleavage efficiencies of 85-93% were achieved for
support 7 (ca. 100 nmol/spot, as determined by UV).14
Performing cleavage in the absence of methanol, that is, “dry
state” cleavage, provided a reduced yet acceptable amount
of material (ca. 55%). Both cleavage efficiencies are notable,
as photocleavage from polymeric supports is commonly low
yielding.21 For example, we observed only 18% cleavage of
N-acetyl-Nph-NH2 from traditional polystyrene (PS) resin
(100-200 mesh) with the identical photolinker using our
solution-phase cleavage method.22 We speculate that the
higher photocleavage efficiency for planar supports could
be due, in part, to the reduced light scattering caused by
planar supports relative to PS beads. This high cleavage
efficiency further underscores the value of planar supports
for solid-phase synthesis.4,5,13
We selected cyclohexane carboxaldehyde, propionic acid,
and cyclohexylisocyanide as components to test the Ugi 4CR
on amine support 7, as these three reactants have been shown
previously to give high conversion to Ugi product 8 on solid
support (Table 1).11 We initially attempted the synthesis of
model compound 8 by individually spotting the three
components neat onto amine support 7 and allowing the
reaction to proceed for 10 min at room temperature. These
reaction conditions gave poor conversion and low purity
(13%; Table 1, entry 1). We found that product purity
quadrupled upon premixing the aldehyde and carboxylic acid
components in DMF prior to spotting (entry 2).23 Longer
reaction times or multiple applications of reagents failed to
improve conversion to 8, however (data not shown). Surpris-
ingly, subjecting our model Ugi 4CR to MW irradiation also
failed to increase product conversion (entry 3).17,18 This result
differed from those of Hoel and Nielsen11 and prompted the
examination of alternate Ugi 4CR conditions for macroarray
construction.
temp
(°C)
pre-spot
H2O
mol ratiob
R1:R2:R3
purity
(%)c
entry
solventa
1
2
3
4
5
6
7
8
25
25
MW: 80f
25
MW: 80f
25
25
4g
d,e
-
-
-
-
-
+
+
+
1:1:2
1:1:2
1:1:2
1:1:2
1:1:2
1:1:2
1:6:3
1:6:3
13
51
47
91
37
72
95
93
DMF
DMF
H2O
H2O
d
d
d
a Used to dissolve or suspend building blocks prior to spotting (1 M
aldehyde; 4 M acid; 4 M isocyanide). Aldehyde and acid premixed. Reaction
times were 10 min unless noted. b R1 ) aldehyde; R2 ) carboxylic acid;
R3 ) isocyanide. c Based on integration of LC spectra with UV detection
(280 nm); error ) (2%. d No solvent used. e No premixing. f MW condi-
tions: 8 min at 80 °C on Weflon. g Reaction mixture spotted at 4 °C and
moved to 25 °C for 15 min.
safety and environmental benefits. Reagent heterogeneity
appears to play a prominent, yet poorly understood, role in
rate acceleration.7a Recently, Pirrung and Das Sarma reported
that water can appreciably accelerate the rates of isocyanide-
based MCRs in solution.24 We reasoned that these conditions
could be applied to heterogeneous, solid-phase Ugi 4CRs
on support 7, and we found that the use of water as a solvent
gave dramatically improved product purity (91%; Table 1,
entry 4). Applying MW irradiation during this reaction
caused a substantial decrease in conversion to 8 (entry 5);
this correlates with data that show MCR rates are slower in
water at elevated temperatures.24 The poor solubility of the
building blocks in water, however, hindered the reproducible
delivery of reagents from aqueous suspensions during SPOT-
syntheses. Further adjustment of reagent ratios and conditions
revealed that simply prespotting support 7 with water (3 µL)
prior to spotting the neat building blocks in a 1:6:3 ratio
gave the highest product purity (95%, entry 7). Finally, to
minimize the evaporation of reagents during macroarray
construction, we found that spotting water and reagents at 4
°C (in a cold room) and then moving the array to room
temperature for 15 min gave comparable product purity
(entry 8). Ongoing studies in our laboratory are focused on
fully characterizing the advantageous role of water in these
solid-phase reactions.25
The use of water as a solvent in organic reactions has
received considerable attention over the past 25 years,7 due
to impressive reaction rate enhancements and its obvious
Using these aqueous Ugi 4CR conditions, we systemati-
cally examined library building blocks to determine the most
reactive set for macroarray synthesis. As others have
(20) C. Entela Mineralight Lamp; Model UVGL-58.
(21) James, I. W. Tetrahedron 1999, 55, 4855-4946.
(22) Our cleavage method is similar to photocleavage protocols used
for PS beads. See ref 22.
(23) The isocyanide was not premixed with the other components, to
avoid formation of the Passerini 3CR product, and was applied to the support
after the aldehyde/carboxylic acid mixture. We attribute the improved
product purity to better mixing and the order of addition.
(24) Pirrung, M. C.; Sarma, K. D. J. Am. Chem. Soc. 2004, 126, 444-
445.
(25) These water-accelerated Ugi 4CR conditions were directly translat-
able to PS resin, generating compound 8 in excellent purity (ca. 95%). See
Supporting Information for details.
Org. Lett., Vol. 7, No. 20, 2005
4457

reported, we found that the reaction was most sensitive to
the nature of the aldehyde component;10a aliphatic-type
aldehydes gave the purest products, while aromatic aldehydes
routinely gave reduced purities. The three sets of building
blocks chosen for the construction of a 96-compound
macroarray are shown in Figure 1; the amine building block
was maintained as Nph.
Table 2. Purity Data for Selected Ugi 4CR Library Membersa
purity
(%)b
purity
(%)b
entry R1 R2 R3
entry R1 R2 R3
8/9a
9b
9c
9d
9e
9f
9g
9h
9i
9j
9k
9l
9m
9n
A
A
A
A
A
A
B
B
B
B
B
B
C
C
a
b
c
d
e
f
a
b
c
d
e
f
1
3
2
1
3
4
4
1
4
2
3
1
2
4
93
97
96
97
90
96
96
88
97
95
90
93
93
87
9o
9p
9q
9r
9s
9t
9u
9v
9w
9x
9y
9z
9a′
9b′
C
C
C
C
C
D
D
D
D
D
D
D
D
D
c
d
e
f
2
3
4
2
3
1
4
1
3
2
3
4
2
3
98
89
97
93
93
92
97
98
92
90
92
93
89
96
f
a
a
b
c
d
d
e
f
a
b
Figure 1. The three sets of building blocks used for the construc-
tion of a 96-member Ugi 4CR library.
f
a Reaction conditions: see Table 1, entry 8. b Based on integration of
LC spectra with UV detection at 280 nm; error ) (2%.
The final macroarray was constructed on a 15 × 18 cm
piece of support 7, with 0.3 cm2 spots spaced 1.2 × 1.8 cm
apart in a grid. Multichannel pipets were used to expedite
the manual delivery of reagents onto the planar support, and
the total setup and synthesis time was less than 30 min. A
subset of the library (30%) was cleaved and analyzed by
LC-MS, and all of these compounds showed purities greater
than 85%, with 37% of the samples tested having purities
greater than 95% (Table 2). This array format provided ca.
50 µg of compound per spot at ca. 90% cleavage efficiency.
A representative LC trace used for analysis is shown in
Figure 2. For products where the diastereomeric products
may play a role in the reaction. Future efforts are directed
at improving these diastereoselectivities, generating Ugi
products with increased structural diversity, and producing
larger macroarrays.
In conclusion, we have shown that Ugi 4CRs are compat-
ible with small molecule macroarray construction on planar
cellulose supports. We have developed a robust photolabile
linker/support system that gives exceptionally high cleavage
efficiencies relative to traditional polymeric supports. Further,
we have shown that water also can accelerate heterogeneous
solid-phase MCR rates in analogy to those of solution-phase
MCRs. This approach permitted the rapid production of a
96-member small molecule macroarray in excellent purity.
The incorporation of MCRs further expands the versatility
of the SPOT-synthesis technique and should find use in
numerous small molecule synthesis and screening efforts.
Acknowledgment. We thank the Donors of the ACS
Petroleum Research Fund, the NSF (CHE-0449959), and the
University of WisconsinsMadison for financial support of
this work. We thank Matthew D. Bowman for helpful
discussions and technical assistance.
Figure 2. Representative LC trace of macroarray member 9a
cleaved from a single spot. UV detection at 280 nm.
Supporting Information Available: Full experimental
details for macroarray construction. This material is available
free of charge via the Internet at http://pubs.acs.org.
could be resolved by LC, their ratios ranged from 3:1 to
6:1. These diastereoselectivities could be inverted when the
Ugi 4CR was performed on PS support with an alternate
linker (Rink amide),14 suggesting that the linker environment
OL051684O
4458
Org. Lett., Vol. 7, No. 20, 2005