Solid-phase synthesis and screening of N-acylated polyamine (NAPA) combinatorial libraries for protein binding

Article abstract of DOI:10.1016/j.bmcl.2010.09.054

Inhibitors for protein-protein interactions are challenging to design, in part due to the unique and complex architectures of each protein's interaction domain. Most approaches to develop inhibitors for these interactions rely on rational design, which re

Full text of DOI:10.1016/j.bmcl.2010.09.054

Bioorganic & Medicinal Chemistry Letters 20 (2010) 6500–6503
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Solid-phase synthesis and screening of N-acylated polyamine (NAPA)
combinatorial libraries for protein binding
Jaclyn A. Iera a, Lisa M. Miller Jenkins b, Hiroshi Kajiyama a, Jeffrey B. Kopp a, Daniel H. Appella a,
⇑
a Laboratory of Bioorganic Chemistry and Kidney Disease Section, NIDDK, NIH, DHHS, Bethesda, MD 20892, United States
b Laboratory of Cell Biology, NCI, NIH, DHHS, Bethesda, MD 20892, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 21 July 2010
Revised 8 September 2010
Accepted 10 September 2010
Available online 17 September 2010
Inhibitors for protein–protein interactions are challenging to design, in part due to the unique and com-
plex architectures of each protein’s interaction domain. Most approaches to develop inhibitors for these
interactions rely on rational design, which requires prior structural knowledge of the target and its
ligands. In the absence of structural information, a combinatorial approach may be the best alternative
to finding inhibitors of a protein–protein interaction. Current chemical libraries, however, consist mostly
of molecules designed to inhibit enzymes. In this manuscript, we report the synthesis and screening of a
library based on an N-acylated polyamine (NAPA) scaffold that we designed to have specific molecular
features necessary to inhibit protein–protein interactions. Screens of the library identified a member with
favorable binding properties to the HIV viral protein R (Vpr), a regulatory protein from HIV, that is
involved in numerous interactions with other proteins critical for viral replication.
Keywords:
Protein–protein interactions
Combinatorial library
High-throughput screening
Vpr
Published by Elsevier Ltd.
The development of synthetic molecules to inhibit protein–pro-
tein interactions has made great progress in the past few years, but
to date there are only inhibitors for a handful of targets in this
area.1 Most approaches to identify inhibitors of these interactions
rely on rational design based upon a three-dimensional structure
of the target complex.2,3 In cases where a structure of the complex
is available, there is a range of peptide mimetic scaffolds that may
be used to begin targeting the interface of a protein–protein inter-
action.4–6 The best inhibitors of such interactions tend to be signif-
icantly larger molecules than the average small-molecule drugs;
this difference likely reflects the increased binding area involved
in most protein–protein interactions as compared with the area
of enzyme active sites where small molecules bind.7 In many cases,
however, there is no convenient three-dimensional structure of the
complex to use as a baseline for inhibitor design. Therefore, ran-
dom screening of a chemical library may be the best approach to
find inhibitors of a structurally uncharacterized protein–protein
interaction. While there are numerous types of chemical libraries
available for screening, most consist of classes of molecules de-
signed to bind to the active sites of enzymes.8 Since the molecular
architecture of protein–protein interactions is different from en-
zyme active sites, new libraries must be developed that more clo-
sely represent the types of molecular architectures associated with
inhibitors of protein–protein interactions.
Recently, we introduced a scaffold we have termed N-acylated
polyamines (NAPAs, Fig. 1) as a new type of molecular architecture
to develop inhibitors for protein–protein interactions.9 As illus-
trated in Figure 1, our general strategy is to project sidechains from
the tertiary amide (R2, R4, R6) and the carbon backbone (R1, R3, R5)
to achieve a densely functionalized array of sidechains. In our ori-
ginal report, we optimized the sidechains projecting from the
NAPA scaffold to inhibit the protein–protein interaction between
HDM and p53, and inhibitors with IC50 values around 2 lM were
identified.9 We report in this manuscript the methods to make a
combinatorial library of these molecules and the associated proce-
dures to screen for hits that bind to biotin-labeled proteins. These
methods successfully identified a lead NAPA that binds to viral
protein R (Vpr), a regulatory protein important in the life cycle of
HIV-1, that is known to engage in multiple interactions with other
proteins.10 The procedures we report can be extended to make
larger NAPA libraries for screening.
Although we previously developed a solid-phase synthesis of
NAPAs on Rink-amide resin,9 the original synthetic route had to
be adapted to allow for efficient synthesis of a library in 96-well
R4
R1
O
O
CH3
R3
H
N
N
H2N
N
N
R5
O
R6
O
R2
O
NO2
⇑
Corresponding author. Tel.: +1 3014511052; fax: +1 3014804977.
E-mail address: appellad@niddk.nih.gov (D.H. Appella).
Figure 1. General NAPA structure.
0960-894X/$ - see front matter Published by Elsevier Ltd.
doi:10.1016/j.bmcl.2010.09.054

J. A. Iera et al. / Bioorg. Med. Chem. Lett. 20 (2010) 6500–6503
6501
microtiter plates in a parallel format such that on-resin screening
of the library could be performed. Therefore, the NAPA library
was synthesized on Tentagel-NH2 macrobeads, which swell in both
water and organic solvents and are resistant to acidic cleavage. Fol-
lowing synthesis of the NAPA on the resin, deprotection of the side-
chains under acidic conditions resulted in a Tentagel-bound NAPA
that was ready for screening under aqueous conditions. The larger
resin size of the macrobeads further facilitated direct visualization
of library hits by a fluorescent microscope as well as allowing
physical manipulation of small numbers of beads.
oligomer lengths; for example, in the current trimeric library, the
number of molecules in the library equals the number of different
amino acids used to make the library raised to the third power
times the number of different acylating groups also raised to the
third power. Monomers used to build the library were selected
to maximize diversity, and included a selection of hydrophilic,
hydrophobic, charged, and aromatic sidechains.
We began with smaller test libraries to ensure the feasibility of
the synthesis and screening. In developing these libraries, we iden-
tified two specific problems: (1) sterically hindered acyl chlorides
(such as isobutyryl chloride) were not very reactive with the sec-
ondary amines in the acylation steps and, (2) while histidine could
be incorporated into the library, it resulted in false positives in the
screens for protein binding. Therefore, the main library was con-
structed from the four amino acids and two acid chlorides listed
in Table 1, resulting in a library of 512 [(4)3 Â (2)3] different NAPA
molecules. To identify different NAPAs within the library, we used
a one-letter amino acid code for the amino acid-derived sidechains
(K, Y, J, or Ff) in addition to a numbered code for the acid chlorides
(2 or 3). For example, a NAPA consisting of three tyrosine residues
and three hydrocinnamoyl residues would be referred to as
Y2Y2Y2.
Following our previously described procedures,9 NAPAs were
synthesized using a series of reductive alkylations, acylations,
and Fmoc deprotections on Tentagel resin (Scheme 1). In contrast
to the previous procedure, which used b-alanine as the first resi-
due, a b-homoalanine was used to reduce over-alkylation for the
first reductive amination in the NAPA synthesis (Supplementary
data). Aldehydes used for reductive amination were made via lith-
ium aluminum hydride reduction of the Weinreb amides of Fmoc-
amino acids.11 For the reductive amination step, imine formation
between a primary amine on the resin and the Fmoc-amino alde-
hyde was promoted by excess trimethylorthoformate (TMOF),12
and the intermediate imine was immediately reduced with a solu-
tion of NaBH(OAc)3 in acetic acid.13 Success of the reductive ami-
To screen for protein binding after completion of the library
synthesis, we modified a procedure originally reported by Kodadek
and coworkers15 that relies on the use of a streptavidin-coated
quantum dot (Qdot 605) to signal the presence of a NAPA that
interacts with a biotinylated protein target ( Fig. 2). The use of
quantum dots was necessary since TentaGel resin exhibits auto-
fluorescence over a broad range of wavelengths that makes visual-
ization of standard fluorescent dyes difficult. Fluorescence
emission of quantum dots occurs at wavelengths that do not over-
lap with TentaGel’s autofluorescence.16 The first step of the screen-
ing protocol involved manually transferring a few beads from each
well of the 96-well microtiter plates that were used for synthesis to
a 384-well microtiter filter plate that would be used for screening.
After transfer, the NAPA-containing Tentagel resin was swelled in
water, then washed with an appropriate buffer, and then a biotin-
ylated protein target was added. After thorough washing to re-
move any unbound protein, a solution of streptavidin-coated
quantum dots (Qdot 605) was added, followed by additional
washes. The plate was then examined using a fluorescent micro-
nation was determined by
a positive chloranil test, which
confirms the presence of a secondary amine on the Tentagel
resin.14
The resulting secondary amine was then acylated with the de-
sired acid chloride in N-methylpyrrolidone (NMP) and an excess
of diisopropylethylamine (DIEA). Completion of this reaction was
monitored by a negative result for the same chloranil amine test.
The final amine was capped with 4-nitrophenylacetic acid, which
aided in HPLC visualization when the hits from the screens of the
library were resynthesized. All these steps were performed in par-
allel using a 96-well microtiter plate.
We designed a NAPA library that was trimeric in amino acid
composition, as we felt that this represented a reasonable size
for inhibitors of protein–protein interactions. Within this type of
NAPA library, there are a total of six positions that may be random-
ized: three positions for amino acid incorporation and an addi-
tional three positions for acylation. One advantage of the NAPA
library is the rapid diversity that may be achieved at short
1) 20% Piperidine/DMF
O
O
CH3
NHFmoc
Fmoc- -homoalanine
NH2
, TMOF, CH2Cl2,10 min
2)
R1
N
H
HATU, HOBt, DIEA, 1 h
Tentagel-NH2 Resin
H
3) NaBH(OAc)3, HOAc, CH2Cl2, 20 min
CH3
O
O
CH3
O
R2
Cl
1) 20% Piperidine/DMF
NHFmoc
NHFmoc
N
H
N
N
H
N
2) 4-nitrophenylacetic acid
HATU, HOBt, DIEA, 1.5 h
H
DIEA, NMP, 2 x 1.5 h
R1
R4
R1
R2
O
O
O
CH3
R3
3
H
N
N
95% TFA/1%TES/CH2Cl2
N
H
N
N
R1
R5
O
R2
O
R6
O
NO2
R4
R1
O
O
CH3
R3
H
N
N
N
H
N
N
R5
O
R2
O
R6
O
NO2
DEPROTECTED
Scheme 1.

6502
J. A. Iera et al. / Bioorg. Med. Chem. Lett. 20 (2010) 6500–6503
Table 1
Initially, screens were performed using 175 nM protein and
incubating it with the NAPA beads for 2 h at 4 °C. Surprisingly, a
large number of the wells had very bright red beads while all the
control wells were green. The brightest wells from this initial
screen, 80 in all, were re-plated into another 384-well plate and
re-screened with 75 nM protein for 30 min at room temperature.
This time there were about 10-fold fewer hits and the controls
were again green. The two hits with the highest fluorescence inten-
sity were chosen, K2K2Y2 and K2K3Y2, and resynthesized. The
resynthesis followed similar procedures, except that Rink-amide
resin was used as the solid support and the acid chlorides were
coupled to the secondary amine in dichloromethane instead of
NMP (acylation proceeds more quickly in dichloromethane, but it
is not used for this step in the library synthesis due to chemical
incompatibility with the filter plate). K2K2Y2 was synthesized
cleanly with one major HPLC peak, whereas K2K3Y2 produced a
large amount of a by-product, which, by mass, consisted of the de-
sired molecule plus an additional isovaleryl group. Therefore,
K2K2Y2 was selected for additional studies (Fig. 3).
Amino aldehydes and acid chlorides used to make the NAPA
library
Aldehyde or acid chloride
Label
Tyrosine aldehyde
Lysine aldehyde
Norvaline aldehyde
4-Fluorophenylalanine aldehyde
Hydrocinnamoyl chloride
Isovaleryl chloride
Y
K
J
Ff
2
3
scope equipped with a triple bandpass filter to determine if protein
binding to members of the NAPA library had occurred. In this sce-
nario, beads that appeared bright red were deemed to have a NAPA
bound to the protein target while the beads that appeared green
were considered to have a NAPA that was not able to bind the pro-
tein target. Therefore, the NAPAs on the bright red beads were
deemed preliminary ‘hits’ and resynthesized for validation and fur-
ther binding studies.
A protein derived from HIV (viral protein R, or Vpr) was selected
for screening against this NAPA library. Vpr is a 96 amino acid pro-
tein (14 kDa) consisting of three
To verify that K2K2Y2 did indeed bind to Vpr, we performed
experiments using isothermal titration calorimetry (ITC). Using a
a helices connected by two
Vpr concentration of 1000
observed as the NAPA was titrated into the protein solution. From
these data, the binding of K2K2Y2 was found to have a Kd of 25 M,
lM, a pronounced binding curve was
loops.17 There are approximately 200 copies of Vpr per virion,
and it has a variety of functions, including the regulation of the
transcription of the HIV-1 long terminal repeat, nuclear transloca-
tion of the HIV-1 preintegration complex, and induction of cell cy-
cle arrest.10 Investigations into the roles of Vpr in the replication
cycle of HIV are ongoing, and molecules that selectively bind to this
protein could be useful in such studies.18 Currently, there are only
a few small molecules that inhibit Vpr activity in cell-based as-
says.19,20 For this study, synthetic Vpr containing a single biotin
group at the N terminus was purchased from Dr. Ulrich Schubert
and Dr. Peter Heinklein (University of Erlangen-Nürnberg,
Germany).
l
indicating modest binding of the NAPA to Vpr ( Fig. 3). Similar
screens were conducted with two other biotinylated proteins,
MDMX21 and FKBP52.22 For both proteins, hits from the library
were identified (K3K2K2 for MDMX, and Y3K2J3 for FKBP52) and
resynthesized, but unfortunately neither NAPA showed any activ-
ity in binding assays to these proteins.
In conclusion, we have developed the procedures to make and
screen a NAPA library for new protein-binding ligands. The scaffold
can be used to generate very diverse libraries with short oligomer
lengths. From our initial studies on the NAPA library, we were able
Figure 2. (A) Schematic for screening of NAPA library. Red-colored beads derive from retention of the streptavidin-labeled quantum dot and is therefore indicative of binding
of the NAPA to the protein target. (B) Green bead color under a fluorescent microscope indicating no retention of streptavidin-labeled Qdot 605 and therefore no NAPA
binding to the protein target. (C) Red bead color observed under a fluorescent microscope indicating a NAPA-protein-binding interaction.

J. A. Iera et al. / Bioorg. Med. Chem. Lett. 20 (2010) 6500–6503
6503
A
B
NH2
Ph
O
O
O
CH3
H
N
N
H2N
N
N
O
O
NO2
Ph
Ph
OH
NH2
K2K2Y2, MW=1039
Figure 3. (A) Chemical structure of K2K2Y2. (B) Isothermal titration data (ITC) of K2K2Y2 binding to Vpr.
to identify a lead candidate, K2K2Y2, that binds to the protein Vpr.
Due to the variety of complex roles this protein plays in the HIV life
cycle, there are a large number of protein–protein interactions
involving Vpr and, at the present time, there is no readily available
in vitro assay to test whether K2K2Y2 directly inhibits a protein–
protein interaction involving Vpr. Confirmatory tests of the biolog-
ical activity will be the subjects of future collaborative research.
While we were not successful at finding NAPA binders for other
protein targets, we believe this could be an indication that our li-
brary lacks promiscuity in its interactions with proteins. The man-
ual aspects for synthesis and screening of the previous NAPA
library preclude expanding this technique to larger libraries at this
time. In future studies, we plan to automate the synthesis and
screening process so that larger NAPA libraries may be examined
across a range of protein targets.
References and notes
1. Wilson, A. J. Chem. Soc. Rev. 2009, 38, 3289.
2. Cummings, C. G.; Hamilton, A. D. Curr. Opin. Chem. Biol. 2010, 14, 341.
3. Bautista, A. D.; Appelbaum, J. S.; Craig, C. J.; Michel, J.; Schepartz, A. J. Am. Chem.
Soc. 2010, 132, 2904.
4. Robinson, J. A. ChemBioChem 2009, 10, 971.
5. Horne, W. S.; Johnson, L. M.; Ketas, T. J.; Klasse, P. J.; Lu, M.; Moore, J. P.;
Gellman, S. H. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 14751.
6. Patgiri, A.; Jochim, A. L.; Arora, P. S. Acc. Chem. Res. 2008, 41, 1289.
7. Wells, J. A.; McClendon, C. L. Nature 2007, 450, 1001.
8. Chuprina, A.; Lukin, O.; Demoiseaux, R.; Buzko, A.; Shivanyuk, A. J. Chem. Inf.
Model. 2010, 50, 470.
9. Hayashi, R.; Wang, D.; Hara, T.; Iera, J. A.; Durell, S. R.; Appella, D. H. Bioorg.
Med. Chem. 2009, 17, 7884.
10. Kino, T.; Pavlakis, G. N. DNA Cell Biol. 2004, 23, 193.
11. Wen, J. J.; Crews, C. M. Tetrahedron: Asymmetry 1998, 9, 1855.
12. Look, G. C.; Murphy, M. M.; Campbell, D. A.; Gallop, M. A. Tetrahedron Lett.
1995, 36, 2937.
13. Abdel-Magid, A. F.; Carson, K. G.; Harris, B. D.; Maryanoff, C. A.; Shah, R. D. J.
Org. Chem. 1996, 61, 3849.
14. Vojkovsky, T. Pept. Res. 1995, 8, 236.
Acknowledgements
15. Kodadek, T.; Bachhawat-Sikder, K. Mol. BioSyst. 2006, 2, 25.
16. Garske, A. L.; Denu, J. M. Biochemistry 2006, 45, 94.
17. Henklein, P.; Bruns, K.; Sherman, M. P.; Tessmer, U.; Licha, K.; Kopp, J.; de
Noronha, C. M.; Greene, W. C.; Wray, V.; Schubert, U. J. Biol. Chem. 2000, 275,
32016.
18. Morellet, N.; Roques, B. P.; Bouaziz, S. Curr. HIV Res. 2009, 7, 184.
19. Richter, S. N.; Frasson, I.; Palu, G. Curr. Med. Chem. 2009, 16, 267.
20. Shimura, M.; Zhou, Y.; Asada, Y.; Yoshikawa, T.; Hatake, K.; Takaku, F.; Ishizaka,
Y. Biochem. Biophys. Res. Commun. 1999, 261, 308.
The authors thank Dr. Ulrich Schubert for the supply of biotin-
ylated Vpr. This research was supported by the Intramural
Research Programs (IRPs) of NIDDK and NCI, and the Intramural
AIDS Targeted Antiviral Program (IATAP) at NIH.
Supplementary data
21. Pazgier, M.; Liu, M.; Zou, G.; Yuan, W.; Li, C.; Li, J.; Monbo, J.; Zella, D.; Tarasov,
S. G.; Lu, W. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 4665.
22. Davies, T. H.; Sanchez, E. R. Int. J. Biochem. Cell Biol. 2005, 37, 42.
Supplementary data (synthetic procedures and data for all as-
says) associated with this article can be found, in the online ver-
sion, at doi:10.1016/j.bmcl.2010.09.054.