Combining a solution-phase derived library with in-situ cellular bioassay: Prompt screening of amide-forming minilibraries using MTT assay

Article abstract of DOI:10.1248/cpb.57.714

We constructed a minilibrary using a solution-phase synthesis through coupling of three core amino compounds (5′-amino-5′-deoxy uridine, 5′-amino-2′,5′-di-deoxy arabinosyl uridine, and butan-1-amine) with 30 carboxylic acids via amide bond formation. The

Full text of DOI:10.1248/cpb.57.714

714
Chem. Pharm. Bull. 57(7) 714—718 (2009)
Vol. 57, No. 7
Combining a Solution-Phase Derived Library with In-Situ Cellular
Bioassay: Prompt Screening of Amide-Forming Minilibraries Using MTT
Assay
Li-Wu CHIANG,a Kai PEI,a Shao-Wei CHEN,a Ho-Lien HUANG,a Kun-Ju LIN,b Tzu-Chen YEN,b and
,a,c
*
Chung-Shan YU
a Department of Biomedical Engineering and Environmental Sciences, National Tsing-Hua University; c Institute of
Nuclear Engineering and Science, National Tsing-Hua University; Hsinchu 30043, Taiwan: and b Department of Nuclear
Medicine, Chang-Gung Memorial Hospital; Taoyuan 33305, Taiwan.
Received February 22, 2009; accepted April 3, 2009; published online April 7, 2009
We constructed a minilibrary using a solution-phase synthesis through coupling of three core amino com-
pounds (5ꢀ-amino-5ꢀ-deoxy uridine, 5ꢀ-amino-2ꢀ,5ꢀ-di-deoxy arabinosyl uridine, and butan-1-amine) with 30 car-
boxylic acids via amide bond formation. The simplified structural core compound butan-1-amine was selectively
coupled with 9 carboxylic acids as control. 3-(4,5-Dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide
assay of the crude mixtures showed that analogues derived from fenbufen, butylfenbufen C15; ethacrynic acid,
butyl ethacrynic amide C18; and sphingosines, Sph-1, Sph-2 and U27 had an increased cytotoxicity against
MCF-7 cells as well as A549 cells. Structural elucidation with molecular docking suggested that cytotoxicity of
these compounds is mainly due to the inhibition of enzymes regulating cellular apoptosis.
Key words amide; assay; solution phase; in situ; library; molecular docking
There has been growing interest in a method combining centration could be determined since the cellular survival
a solution-phase derived library with an in-situ bioassay on ratio of most compounds acting as negative control is greater
microtiter plate.1) Construction of the library is initiated than 80%. The method is illustrated in Fig. 2.
through a core compound, either as a lead of natural product
As shown by crude assay results (Figs. 3a—c), the cellular
or a transition-state analog from mechanistic considerations, survival ratio for the majority of the products was over 60%,
followed by coupling with various carboxylic acids as build- suggesting the absence of cytotoxicity. On the other hand, the
ing blocks. The libraries were constructed on a microtiter coupling products U27, A15, C15, and C18 exhibited biolog-
plate or a set of centrifuge tubes. In each well or tube, the ical activity. These potential compounds were further pre-
products obtained were screened for their binding affinity for pared and purified on flash chromatography to be submitted
the enzyme of interest. The bioactivity derived from the mix- for a “clean” analysis. In contrast, we disregarded the evalua-
ture, in general, is consistent with that of the purified prod- tion of U15 due to the higher activity of its 2ꢀ-epimer A15
uct. Indeed, a number of potential substrates for numerous against both A549 and MCF7 cell lines.
enzymes including sulfotransferase,2) fucosidase,3) fucosyl-
Unexpectedly, a subtle cytotoxicity of U27 was evident
transferase,4) protease,5) and protease dimerization, have been (Fig. 3b and supporting information: S16, S17). As a control,
discovered by this approach.6) Since this method is mainly a referential core compound: butan-1-amine was coupled
focused on enzymatic assays, a further application in cellular with the corresponding acid moieties to provide C27 (Fig.
assays may be of importance. In this article, we sought to 3b). Since U27 shares structural similarities to sphingosine,
couple the amide-forming libraries with a cell-line based the sphingosine analogs Sph-1 and Sph-2 (carbon length of
assay. In addition, we elucidated the mechanisms related to 13 and 18, respectively), were prepared for comparison pur-
bioactivity of potential amides through molecular docking.
As part of our ongoing research, we are focusing on the
development of nucleoside analogs to be used as prodrugs to
target herpes simplex virus thymidine kinase gene.7—9) Thus,
5ꢀ-amino-5ꢀ-deoxy analogues of pyrimidine nucleosides set
as a core compound. To clarify whether the bioactivity is
mainly associated with the core 5ꢀ-amino nucleoside com-
pounds, a structurally simplified butan-1-amine was used as
control for comparison purposes. In addition to the commer-
cially available butan-1-amine, the preparation of these nu-
cleoside analogs has been reported.10) In this study, car-
boxylic acid moieties were structurally classified in four
types as follows: mono-aromatic rings (1—15), di-aromatic
rings (16—18), fused rings (19, 20), and aliphatic groups
containing heteroatoms such as phosphor and aza acids 21—
30 (Table 1).
The coupling product in each well was diluted before
transfering to the plate. Thus the appropriate working con-
Fig. 1. Libraries Constructed through a Coupling of Three Core Amines
with Carboxylic Acids
∗ To whom correspondence should be addressed. e-mail: csyu@mx.nthu.edu.tw
© 2009 Pharmaceutical Society of Japan

July 2009
715
Table 1. Carboxylic Acids as Building Blocks for Amide Bond Formation
D1
D7
D2
D8
D3
D9
D4
D5
D6
D10
D11
D12
D13
D14
D15
D16
D17
D18
D19
D25
D20
D26
D21
D27
D22
D28
D23
D29
D24
D30
role for the biological activity.16) Notably, the OH group at C-
1 forming a hydrogen bond with the guanidine group of argi-
nine-442 is crucial. Our docking results (Figs. 5a—c) sug-
gested a similar interaction formed among numerous amino
acid residues including hydrophilic and hydrophobic con-
tacts. Interestingly, the 4-OH group forms a hydrogen bond
to OH of tyrosine-576 closing to the tyrosine-553, which is
responsible for the hydrogen bond to the amide group of nat-
Fig. 2. Flowchart of the Dilution Process
poses (Fig. 4). While the bioactivity is mainly due to both the ural ceramide. Impressively, in the case of Sph-2, the 2-NH2
lipid part and the hydrophilic head, the hydrophilic moieties and 4-OH groups form hydrogen bonds to glutamine-467 and
of sphingosine are more effective in inducing cytotoxicity. tyrosine-553, respectively (Fig. 5c). Similarly, although the
The hydrophilic moieties of sphingosine are capable to in- 3ꢀ-OH group of U27 forms a hydrogen bond to the OH of
duce specific proapoptotic signals that may account for cyto- threonine-448, the cavity is still able to accommodate the
toxicity.11—14)
pyrimidine base ring (Fig. 5d). Our results suggest that the
In spite of the few data available on the membrane recep- START domain is the site potentially responsible for mediat-
tor and X-ray diffraction data, an attempt to clarify the prob- ing apoptosis through these lipid analogs.
able binding mechanisms through molecular docking was
During the crude assay of C-series compounds, an unex-
performed. The START, known as a domain of steroidogenic pected bioactivity of C18 against both A549 and MCF7 was
acute regulatory protein-related lipid transfer (StARD10), evident (Fig. 16 and 17 in S.I., Table 2). Acid moieties D18
has been reported to be overexpressed in breast cancer.15) In also known as fenbufen is a member of the non-steroid anti-
addition, the START domain of ceramide transport (CERT) inflammatory drugs (NSAIDs). As an inhibitor targeting cy-
protein is known for its transferring ability of natural D- clooxygenase, fenbufen is effective in rheumatoid arthritis
erythro ceramides, mainly used for transporting from ER and osteoarthritis.17) Although NSAIDs such as diclofenac
to Golgi apparatus.16) The two START domains have similar may inhibit tumor growth in vitro (IC50: 360 mM)18) and in
characteristics and consist of 210 and 250 amino acids, re- vivo,19) there was no significant improvement in cytotoxicity
spectively. The cavity of CERT START domain comprises a through modification of other NSAIDs, for instance fenopro-
line of hydrophobic and polar charged heads. According to fen.20) Therefore, fenbufen (D18) is an interesting probe for
the crystal structure of the complex formed by CERT domain studying the relationship between these two diseases through
of 2e3n with ceramides having acyl groups of various structural modification. Furthermore, C18 is a potential hit
lengths, the hydrogen bond formed by the hydrophilic head and may act as a starter for further screening of a new hit
of the substrate within the deep active site plays a critical compound or even a lead compound. The potential enzyme

716
Vol. 57, No. 7
Fig. 3. (a) Assay of Cytotoxicity against A549 Cell Line by the Libraries Prepared from Uridine Analogue 1 and Arabinosyl Uridine Analogue 2, (b)
Assay of Cytotoxicity against MCF-7 Cell Line by the Libraries Prepared from Uridine Analogue 1 and Arabinosyl Uridine Analogue 2 and (c) Assay of Cy-
totoxicity against MCF-7 Cell Lines by the Libraries Prepared from n-Butanamine (P: Cisplatin)
targeted by C18 has been suggested to be cyclooxygenase
(COX).21) COX-2 is an inducible cyclooxygenase isoform
that plays a major role in inflammation and has been found to
be hyperexpressed in several human tumors.22—24) COX-2
overexpression is involved in cancer growth and invasion.25)
Various NSAIDS have been studied as anticancer drugs.
Since the bioactive flurbiprofen mimics our butylfenbufen,
the 3-D crystal structure of the complex formed by
prostaglandin H2 synthase-1 from goat with flurbiprofen was
Fig. 4. Lipid-Containing Compounds Used in the Bioassay
chosen as model (Fig. 6a). When examining the inner struc-
ture formed by flurbiprofen within the active site, butyl fen-
bufen-rather than fenbufen-shared a similar environment
(Figs. 6b, c). This is likely to be ascribed to the hydrogen
bond formed between the oxo group and the terminal amino
group of arginine-120 and the van der Waals contact formed
between butyl group and Val 116 as well as Leu93.
Instead of a hydrogen bond formed between the oxo group
and arginine-120 as for butyl fenbufen, fenbufen formed a
hydrogen bond between the oxo group and tyrosine-355. The
slightly incline-to-the-left structure weakened the nonpolar
contact of the biphenyl ring with the deep pocket of active
site, thereby diminishing its activity. Nevertheless, in spite of
the success in the elucidation of the structure–activity rela-
tionship through molecular-docking, the apoptotic mecha-
nism in terms of the block of COX-2 might not fully account
for the subtle cytotoxicity of butylfenbufen. Accordingly, tar-
geting cyclooxygenase may be not the sole antiproliferative
mechanism of NSAIDs.22) We therefore examined the protein
data bank (PDB) to find the related report regarding the
probable human-derived enzyme responsible for the
proapoptotic activity. The screening results obtained by con-
fining to the substructure of diphenyl rings indicated that
proteins of the Bcl-x or Bcl-2 families overexpressed in
many cancers fit these criteria.26) However, a contradictory
docking result for butylfenbufen was obtained (Fig. 7a). In
Fig. 5. (a) Docking of SPh-1 to the Active Site Defined by the Ligand of
Ceramide, (b) Schematic Representation of the Electronic Clouds of Amino
Acid Residues Participating in the Binding to Sph-1, (c) Amino Acid
Residues Participating in the Binding to Sph-2 and (d) Amino Acid
Residues Participating in the Binding to U27
contrast to the report of the extension of the biphenyl ring of
4ꢀ-fluoro-1,1ꢀ-biphenyl-4-carboxylic acid (4-FC) into the
deep pocket of the active site (Fig. 7b), an inverted binding
of the butylfenbufen was evident. Furthermore, most of the
binding energy of this model is resulting from the coverage

July 2009
717
Table 2. IC50 of the Compounds Purified from Independent Synthesis
Samples
C18
U18
D18
C15
A15
D15
U27
Sph-1 Sph-2
A16
C16
X16
D16
U1 Cisplatin
Cell lines A549
100
87
ꢃ100 ꢃ100
ꢃ100 ꢃ100
18
8ꢄ5
ꢃ100 84ꢄ16 ꢃ100
ꢃ100 26ꢄ2 ca. 100
50
40
38
20
ꢃ100 ꢃ100 ꢃ100 ꢃ100 ꢃ100
ꢃ100 ꢃ100 ꢃ100 ꢃ100 ꢃ100 25ꢄ6
8ꢄ3
(mM)
MCF 7
Data are the mean of two or three experiments and are reported as the mean [standard error of the mean (S.E.M.)].
Fig. 8. (a) 2gss Docking with Ethacrynic Acid and (b) 2gss Docking with
Butyl Ethacrynic Amide C15
edema.17) Furthermore, D15 is an inhibitor of betaglutathion-
transferase (BGTT) whose upregulation has been associated
with drug resistance during chemotherapy of various can-
cers.27,28) Further analysis of the binding pattern between
C15 and BGTT was performed by using the DS program ac-
cording to 3D-data of 2gss from the PDB bank (Fig. 8a).29)
Compared to ethacrynic acid, the extra butyl group extends
into the deep pocket, an unidentified extra site freely avail-
able for extra stabilization (Fig. 8b). Furthermore, both
residues of tyrosine-107 and phenyl alanine-7 near the en-
trance of the pocket provided a primary stabilization for the
benzene ring of C15 through a pi–pi stackering, a mimic of a
sandwich complex. The terminal amido group in the gluta-
mine-51 residue may provide indirect polar contact with the
amido group of butyl ethacrynic amide. Interestingly, despite
the presence of the butyl group, there is a loose space sur-
rounded by the residues of glutamine-64, serine-65 (not
specified in Fig. 8b) and the terminal butyl group of butyl
ethacrynic amide. This may provide a base for future modifi-
cations of the butyl group in order to develop potential in-
hibitors.
Fig. 6. (a) The Amino Acid Residues of 1 CQE Participating in Direct
Binding to Flurbiprofen, (b) The Amino Acid Residues of 1 CQE Participat-
ing in Direct Binding to Butyl Fenbufen and (c) The Amino Acid Residues
of 1 CQE Participating in Direct Binding to Fenbufen
Interestingly, a recent study has shown that a butyl ester
derivative of ethacrynic acid may induce apoptosis in
leukemia cells.30) Our approach may thus provide a new av-
enue to explore new potential compounds based on the con-
cept of “old drug new use.”
In conclusion, we have described a simple method for
screening an in vitro solution-derived library using the MTT
assay at the cellular level. The probe screening of this mini-
library (comprising of 69 compounds derived from three core
Fig. 7. (a) Residues of Active Site of Bcl-x Responsible for the Binding to
Butylfenbufen and (b) Binding of 4-FC and TN1 to 1YSG (Bcl-x)
Biphenyl ring extending into the active site.
of the biphenyl group on the protein surface, a result that is amines) was helpful in the quest of novel hit compounds
not consistent with 4-FC’s. The proapoptotic activity of such as C18. It also shows enough sensitivity to detect bioac-
butylfenbufen is thus unlikely to occur via inhibition of Bcl- tive compounds based on a relative survival ratio of ꢃ60%
x proteins.
for controls. The hit compound C15 may be further modified
Acid moieties D15, known as ethacrynic acid, are classi- as a potential hit compound through a series of high through-
fied as a group of diuretics that can block Naꢁ–Kꢁ–2Clꢂ put screening (HTS). Investigation of the mechanism respon-
symporter in the thick ascending limb of the loop of Henle. sible for the inhibition of proliferation of cancer cells was
They are generally used in patients with acute pulmonary performed by molecular docking using the DS program. The

718
Vol. 57, No. 7
10) Winans K. A., Bertozzi C. R., Chem. Biol., 9, 113—129 (2002).
11) Teixeira-Clerc F., Julien B., Grenard P., Van Nhieu J. T., Deveaux V., Li
L. Y., Serriere-Lanneau V., Ledent C., Mallat A., Lotersztajn S., Nat.
Med., 12, 671—676 (2006).
cytotoxicity of U27, C18 and C15 discovered through the
combinatorial approach is likely to occur mainly via binding-
mediated inhibition of apoptosis-regulating enzymes.
12) Ogretmen B., Hannun Y. A., Nat. Rev. Cancer, 4, 604—616 (2004).
13) Coursol S., Fan L. M., Le Stunff H., Spiegel S., Gilroy S., Assmann S.
M., Nature (London), 423, 651—654 (2003).
14) Padrón J. M., Curr. Med. Chem., 13, 755—770 (2006).
15) Olayioye M. A., Vehring S., Müller P., Hermann A., Schiller J., Thiele
C., Lindeman G. J., Visvader J. E., Pomorski T., J. Biol. Chem., 280,
27436—27442 (2005).
16) Kudo N., Kumaga K., Tomishige N., Yamaji T., Wakatsuki S., Nishi-
jima M., Hanada K., Kato R., Proc. Natl. Acad. Sci. U.S.A., 105,
488—493 (2008).
17) Hardman J. G., Goodman G. A., Limbird L. E., “Goodman and
Gilman’s The Pharmacological Basis of Therapeutics,” 9th ed., Mc-
Graw-Hill Companies, Southern California, 1996.
18) Johnsen J. I., Lindskog M., Ponthan F., Pettersen I., Elfman L., Orrego
A., Sveinbjornsson B., Kogner P., Cancer Res., 64, 7210—7215
(2004).
19) Takada Y., Bhardwaj A., Potdar P., Aggarwal B. B., Oncogene, 23,
9247—9258 (2004).
20) Barbaric´ M., Kralj M., Marjanovic´ M., Husnjak I., Pavelic´ K., Fil-
ipovic-Grcic´ J., Zorc D., Zorc B., Eur. J. Med. Chem., 42, 20—29
(2007).
21) Thun M. J., Henley S. J., Patrono C., J. Natl. Cancer Inst., 94, 252—
266 (2002).
Experimental
The reagents used in the amide bond formation was core amine (1 mg,
4 mmol), carboxylic acid (1 eq), diisopropyl ethyl amine (DIEA) (1.2 eq) and
2-(1H-benzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate
(HBTU) (1.1 eq). Both the starting core amine and the carboxylic acids were
prepared as a stock solution of DMSO at a concentration of 0.1 mmol/200 ml
and 5 mmol/10 ml, respectively. HBTU and DIEA were dissolved in DMSO
as a concentration of 5 mmol/10 ml, respectively. Each of the acid portion
was firstly mixed with HBTU in a plastic tube for 30 s, followed by the addi-
tion of a mixture of core amine (10 ml) and DIEA (10 ml) in a total volume
of 40 ml. All vials were shaked for 1 min. A portion of the mixture (10 ml)
was transferred to a novel tube followed by addition of 990 ml of water. A
volume of 10 ml was pipetted out and added into the corresponding well of
microtiterplate planted with 100 ml of A549 or MCF7 cells in a concentra-
tion of 30000 cells/ml. After an incubation of 2 d, the supernatants were re-
moved through washing followed by MTT reagents and the absorbance at
580 nm was recorded according to the usual protocol conducted routinely in
the medicinal chemistry laboratory.
Acknowledgment We are grateful to the National Science Council of
Taiwan and CGMH_NTHU Joint Research for providing financial support
(NSC-95-2113-M-007-039 and CGTH96N2342E1).
22) Bock J. M., Menon S. G., Goswami P. C., Sinclair L. L., Bedford N. S.,
Domann F. E., Trask D. K., Mol. Carcinog., 46, 857—864 (2007).
23) Brown J. R., DuBois R. N., Clin. Cancer Res., 10, 4266s—4269s
(2004).
24) Lee D. W., Sung M. W., Park S. W., Seong W. J., Roh J. L., Park B.,
Heo D. S., Kim K. H., Anticancer Res., 22, 2089—2096 (2002).
25) Bottone F. G., Moon Y., Kim J. S., Alson-Mills B., Ishibashi M., Eling
T. E., Mol. Cancer Ther., 4, 693—703 (2005).
Supporting Information Available Details of the experimental proce-
dures including bioassay, characterization data and molecular docking for li-
1
brary members including H- and 13C-NMR and molecular docking by DS
program under ligandfit working model for library members.
References
1) Brik A., Wu C.-Y., Wong C.-H., Org. Biomol. Chem., 4, 1446—1457
(2006).
26) Oltersdorf T., Elmore S. W., Shoemaker A. R., Armstrong R. C.,
Augeri D. J., Belli B. A., Bruncko M., Deckwerth T. L., Dinges J.,
Hajduk P. J., Joseph M. K., Kitada S., Korsmeyer S. J., Kunzer A. R.,
Letai A., Li C., Mitten M. J., Nettesheim D. G., Ng S., Nimmer P. M.,
O’Connor J. M., Oleksijew A., Petros A. M., Reed J. C., Shen W.,
Tahir S. K., Thompson C. B., Tomaselli K. J., Wang B. L., Wendt M.
D., Zhang H. C., Fesik S. W., Rosenberg S. H., Nature (London), 435,
677—681 (2005).
27) Townsend D. M., Tew K. D., Oncogene, 22, 7369—7375 (2003).
28) Zhang K., Mack P., Wong K. P., Int. J. Oncol., 12, 871—882 (1998).
29) Oakley A. J., Rossjohn J., Lo Bello M., Caccuri A. M., Federici G.,
Parker M. W., Biochemistry, 36, 576—585 (1997).
2) Best M., Brik A., Chapman E., Lee L., Cheng W.-C., Wong C.-H.,
ChemBioChem, 5, 811—819 (2004).
3) Wu C.-Y., Chang, C.-F., Chen J. S. Y., Lee, S.-T., Wong C.-H., Lin, C.-
H., Angew. Chem.-Int. Edit., 42, 4661—4664 (2003).
4) Lee L. V., Mitchell M. L., Huang S. J., Fokin V. V., Sharpless K. B.,
Wong C. H., J. Am. Chem. Soc., 125, 9588—9589 (2003).
5) Brik A., Lin Y.-C., Elder J. Wong C.-H., Chem. Biol., 9, 891—896
(2002).
6) Lee S. G., Chmielewski J., Chem. Biol., 13, 421—426 (2006).
7) Yu C.-S., Chiang L.-W., Wu C.-H., Wang R.-T., Chen S.-W., Wang H.-
Y., Yeh C.-H., Nucl. Med. Biol., 33, 367—370 (2006).
8) Yu C.-S., Wu C.-H., Chiang L.-W., Pei K., Hsu Z.-K., Synthesis, 2006,
3835—3840 (2006).
30) Wang K., Li C., Song D., Zhao G., Zhao L., Jing Y., Cancer Res., 67,
7856—7864 (2007).
9) Yu C.-S., Wang R.-T., Chiang L.-W., Lee M.-S., Tetrahedron Lett., 48,
2979—2982 (2007).