Rational design and structure-activity relationship studies of quercetin-amino acid hybrids targeting the anti-apoptotic protein Bcl-xL

Article abstract of DOI:10.1039/c7ob02045g

Anti-apoptotic proteins, like the Bcl-2 family proteins, present an important therapeutic cancer drug target. Their activity is orchestrated through neutralization upon interaction of pro-apoptotic protein counterparts that leads to immortality of cancer

Full text of DOI:10.1039/c7ob02045g

View Article Online
View Journal
Organic &
Biomolecular
Chemistry
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: T. F. Kellici, M.
Chatziathanasiadou, M. Lee, N. Sayyad, E. Geromichalou, E. I. Vrettos, A. D. Tsiailanis, S. Chi, G.
Geromichalos, T. Mavromoustakos and A. G. Tzakos, Org. Biomol. Chem., 2017, DOI:
10.1039/C7OB02045G.
This is an Accepted Manuscript, which has been through the
Volume 14 Number
1 7 January 2016 Pages 1–372
Royal Society of Chemistry peer review process and has been
accepted for publication.
Organic &
Biomolecular
Chemistry
www.rsc.org/obc
Accepted Manuscripts are published online shortly after
acceptance, before technical editing, formatting and proof reading.
Using this free service, authors can make their results available
to the community, in citable form, before we publish the edited
article. We will replace this Accepted Manuscript with the edited
and formatted Advance Article as soon as it is available.
You can find more information about Accepted Manuscripts in the
author guidelines.
Please note that technical editing may introduce minor changes
to the text and/or graphics, which may alter content. The journal’s
standard Terms & Conditions and the ethical guidelines, outlined
in our author and reviewer resource centre, still apply. In no
event shall the Royal Society of Chemistry be held responsible
for any errors or omissions in this Accepted Manuscript or any
consequences arising from the use of any information it contains.
ISSN 1477-0520
COMMUNICATION
Takeharu Haino et al.
Solvent-induced emission of organogels based on tris(phenylisoxazolyl)
benzene
rsc.li/obc

Page 1 of 44
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C7OB02045G
Rational design and structure-activity relationship studies of quercetin-
amino acid hybrids targeting the anti-apoptotic protein Bcl-xL
Tahsin F. Kellicia,b, Maria V. Chatziathanasiadoua, Min-Sung Leec, Nisar Sayyada,
Geromichalou EGd, Eirinaios I. Vrettosa, Antonis D. Tsiailanisa, Seung-Wook Chic, George D
Geromichalose,f Thomas Mavromoustakosb, Andreas G. Tzakosa,*
aDepartment of Chemistry, University of Ioannina, Ioannina 45110, Greece
bDepartment of Chemistry, National and Kapodistrian University of Athens,
Panepistimiopolis Zografou 15771, Greece
cDisease Target Structure Research Center, KRIBB, Daejeon 34141, Republic of Korea
dLaboratory of Pharmacology, Medical School National and Kapodistrian University of
Athens, Athens 11527, Greece
eLaboratory of Inorganic Chemistry, Faculty of Chemistry, Aristotle University of
Thessaloniki, GR-54124 Thessaloniki, Greece
fCell Culture, Molecular Modelling and Drug Design Lab., Symeonidion Research Center,
Theagenion Cancer Hospital, Thessaloniki, Greece
Corresponding Authors
* Tel.: +3026510 08387, E-mail: agtzakos@gmail.com (A. Tzakos)

Organic & Biomolecular Chemistry
Page 2 of 44
View Article Online
DOI: 10.1039/C7OB02045G
Abstract
Anti-apoptotic proteins, like the Bcl-2 family proteins, present an important therapeutic
cancer drug target. Their activity is orchestrated through neutralization upon interaction of
pro-apoptotic protein counterparts that leads to immortality of cancer cells. Therefore,
generating compounds targeting these proteins is of immense therapeutic importance. Herein,
Induced Fit Docking (IFD) and Molecular Dynamics (MD) simulations were performed to
rationally design quercetin analogues that bind in the BH3 site of the Bcl-xL protein. IFD
calculations determined their binding cavity while Molecular Mechanics Poisson Boltzmann
Surface Area (MM-PBSA) and Molecular Mechanics Generalised Born Surface Area (MM-
GBSA) calculations provided an insight into the binding enthalpies of the analogues. The
quercetin analogues were synthesized and their binding to Bcl-xL was verified with
fluorescence spectroscopy. The binding affinity and the thermodynamic parameters between
Bcl-xL and quercetin-glutamic acid were estimated through Isothermal Titration Calorimetry.
1
2D H−15N HSQC NMR chemical shift perturbation mapping was used to map the binding
site of the quercetin analogues in the Bcl-xL that overlapped with the predicted poses
generated by both IFD and MD calculations. Furthermore, evaluation of the four conjugates
against the prostate DU-145 and PC-3 cancer cell lines, revealed quercetin-glutamic acid and
quercetin-alanine as the most potent conjugates bearing the higher cytostatic activity. This
pinpoints that the chemical space of natural products can be tailored to exploit new hits for
difficult tractable targets such as protein-protein interactions.
Keywords: Bcl-xL; Quercetin analogues; Induced fit docking; Molecular Dynamics; Protein-
protein interactions; fluorescence spectroscopy; 2D HSQC NMR

Page 3 of 44
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C7OB02045G
Introduction
Programmed cell death (apoptosis) plays a pivotal role in the normal growth, homeostasis
of tissues and removal of infected or dangerous cells. The alteration of apoptosis results in the
occurrence of different diseases such as inflammation, tumor development, tumor cell
resistance to chemotherapy, neoplasia, neurodegeneration or autoimmune disorders. As a
result, the molecular mechanisms responsible for the apoptotic cell death are highly
conserved 1-4
.
It is now well-established that the B-cell lymphoma-2 family proteins directly regulate all
5
major types of cell death, including necrosis, autophagy and apoptosis . This family is
subdivided into the anti-apoptotic proteins (i.e. Bcl-2, Bcl-xL, Bcl-W, Mcl-1 and Α1) and the
pro-apoptotic proteins (i.e. Bax, Bak, Bad, Bim, PUMA, Noxa)1,6. Bcl-xL binds the
proapoptotic proteins Bax and Bak in the third B-cell homology domain (BH3), antagonizing
cell apoptosis 7-10. The X-ray structure of Bcl-xL with the BH3 domain of Bad and Bim
provided a basis for the rationalization of the binding mode of these pro-apoptotic proteins to
the hydrophobic groove of Bcl-xL 11, 12. Bcl-xL is comprised of eight α-helices connected by
loops of different lengths 4. The homology domains ΒΗ1, ΒΗ2 and ΒΗ3 form a hydrophobic
cavity in the surface of the protein. This cleft is the binding site of the BH3 domains of the
proapoptotic proteins and the main target for small molecules acting as inhibitors of these
protein-protein interactions (PPI)13. The BH3 inhibitor WEHI-539 induces a significant
conformational change around the p2 pocket to accommodate their benzothiazole moiety. It
was suggested that WEHI-539’s selectivity for Bcl-xL may be due to an array of hydrogen
bonds taking place in this enlarged p2 pocket between the benzothiazole hydrazone group
and backbone NH and carbonyls. The hydrophobic compound ABT-737 binds in two out of
four key hydrophobic pockets on the surface of their pro-survival protein targets (p2 and p4)
inducing an important remodeling of the p2 pocket14.
The interface implicated in protein-protein interactions is flat and extensive and thus
triggers a special challenge in developing small molecules targeting this biological space15, 16

Organic & Biomolecular Chemistry
Page 4 of 44
View Article Online
DOI: 10.1039/C7OB02045G
17. One of the major breakthroughs in this direction was made by the pharmaceutical
company Abbott, which disclosed in 2005 ABT-737, a small molecule that binds to Bcl-xL,
Bcl-2 and Bcl-W with subnanomolar affinity16. This compound has a large molecular weight
(813 Da) and presents low oral bioavailability16. Recently, the same company disclosed the
compounds Navitoclax (ABT-263) [discontinued in phase II clinical trials] and Venetoclax
(ABT-199). Finally, on April 2016 Vevetoclax was approved by FDA as a second line
treatment of chronic lymphocytic leukemia (CLL). However, this was the only synthetic BH3
mimetic that managed to pass the three phases of clinical trials, whereas many other like
Oblimersen, Obatoclax and Navitoclax were rejected18. Therefore, the quest to identify novel
compounds targeting these therapeutic targets is of importance and the relevant efforts are in
the forefront.
Natural products have served as a rich and profitable source for drugs19. Our group recently
reported that the plant flavonoid quercetin directly binds to the BCL-2 family proteins and
triggers their proapoptotic activity20. Through this study we unveiled the interaction profile
developed by quercetin and the target protein. In our effort to enhance the solubility of the
parent compound, quercetin, as also to construct analogues able to develop favorable
interactions with the BH3-binding site of Bcl-xL we rationally designed several quercetin
hybrids (Figure 1) using Induced Fit Docking (IFD) and Molecular Dynamics (MD)
simulations. Our inspiration was based on the fact that cancer cells adopt an increased
demand for nutrients and amino acids and selective amino acid transporters have been
identified to be upregulated in cancer cells in response to this increased demand 21. Peptide
transporters have been identified to be overexpressed in several tumors and have been
characterized as promising targets for tumor-specific drug delivery22. Specifically, the human
peptide transporter, hPEPT1, has been involved in the transportation of amino acids into the
intestine. hPEPT1 has been also implicated to recognize amino acids conjugated to a drug
and has thus been exploited to transport drug-amino acid conjugates23, 24. Therefore, for drugs
with poor permeability, their absorption can be enhanced by attaching to them amino acids
and thus recognized and transported by hPEPT1. Along these lines, several different prodrugs
based on drug-amino acid conjugates have been reported targeting such receptors allowing
enhanced cell permeability profile 25-29. Given the facts that: (i) quercetin is practically
insoluble in aqueous solutions, (ii) we have already charted the interaction architecture of
quercetin with BCl-2 family proteins that can be utilized for further rational design of new
analogues, (iii) its conjugation with amino acids has amplified the bioavailability of the
30
parent compound
and (iv) amino acid transporters are up-regulated in cancer cells, we

Page 5 of 44
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C7OB02045G
rationally designed four quercetin-amino acid conjugates. These conjugates are achieved at 3´
or 4´ positions of quercetin and involve the different degree lipophilic amino acids Ala, Phe,
Glu and Leu. In addition, these conjugates are designed to allow increase of multiple
interactions contacts in terms of quercetin and therefore increase the binding affinity.
Molecular Mechanics Poisson Boltzmann Surface Area (MM-PBSA) and Molecular
Mechanics Generalised Born Surface Area (MM-GBSA) calculations provided an insight into
the binding enthalpies of the analogues. The compounds were then synthesized and their
binding to Bcl-xL was evaluated with fluorescence spectroscopy and the thermodynamics of
the interaction between Bcl-xL and Que-Glu were determined with isothermal titration
calorimetry. To experimentally map the Bcl-xL binding surface implicated in the interaction
of the different quercetin analogues we conducted chemical shift perturbation mapping
1
through 2D H-15N HSQC NMR experiments of the 15N-labeled Bcl-xL. In order to evaluate
both the cytostatic and cytotoxic effects of the quercetin conjugates as also the implication on
their activity of peptide transporters, we conducted in vitro studies in the prostate cancer cell
line PC-3, that has been established as a cell culture model for hPEPT122, and DU-145 as a
negative control.

Organic & Biomolecular Chemistry
Page 6 of 44
View Article Online
DOI: 10.1039/C7OB02045G
Figure 1: The newly synthesized quercetin analogues as Bcl-xL inhibitors.
Materials and Methods
All chemicals were of the best grade commercially available (Acros Organics, Sigma
Aldrich, Alfa-Aesar, Fluka and Merck) and were used without further purification. All
solvents were purified according to standard procedures; dry solvents were obtained
according to methods previously described in the literature and stored over molecular sieves.
HPLC grade acetonitrile and water, purchased from Fischer Scientific, were used for the RP-
HPLC. Analytical thin-layer chromatography was performed on glass plates coated with
silica gel containing calcium sulfate binder. Column chromatography was performed using
1
SiO2 and Dichloromethane-Methanol mixtures were used for elution. The H NMR and 13C
NMR spectra were recorded at 400 MHz in DMSO-d6 and 298 K. The samples were titrated

Page 7 of 44
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C7OB02045G
with 1 M solution of picric acid in DMSO-d6 to make the peaks derived from the phenol
groups sharper (the peak broadening of these hydroxyl protons is due to rapid intermolecular
exchange which is minimized upon addition of picric acid as our group has recently
1
shown31). Chemical shifts for H NMR spectra are reported as δ units of parts per million
(ppm) and relative to the signal of DMSO (δ 2,50 multiplet). Multiplicities are given as
follows: s (singlet); d (doublet); dd (doublet of doublets); t (triplet); bs (broad singlet); m
(multiplet). Coupling constants are reported as J values in Hz (Supplementary Information 1,
2-1 – 2.8). For the MS analysis, each quercetin derivative was diluted in ACN:H2O (50:50
v/v) and injected directly to the EVOQ™ Elite ER LC-TQ system (Bruker, Germany). LC-
MS grade acetonitrile and water were obtained from Fischer. The analyses were performed
using an ESI probe in negative ionization with a spray voltage (-) set to 4000 V. The cone
and heated probe temperature were set to 300°C and 25°C, respectively. A nebulizer gas flow
equal to 20 units and a cone gas flow of 25 units were applied for the solvent evaporation.
MS spectra were collected and analyzed through the MSWS 8.2 software, provided by
Bruker.
Preparation of the quercetin - amino acid analogues
The quercetin conjugates were synthesized based on the previous publication by Kim et al.
30
with few modifications to successfully enhance the overall yield. Briefly, the addition of
DIPEA (2eq) was conducted every day for 3-4 days instead of a single addition and as a
result the yield of this step was increased by 10-15% for each conjugate. Reaction was
monitored by thin layer chromatography, which indicated that Quercetin as a starting material
was completely consumed at the final day. Moreover, the final compounds were obtained via
precipitation with hexane instead of recrystallization with acetone/CH2Cl2 which resulted in
the same yield as reported but in minimum time and effort.
Specifically, amino acids were converted to their tert-butyl esters through reaction with
perchloric acid (HClO4) and tert-butyl acetate as a solvent in airtight closed vessel. These
tert-butyl esters of amino acids were transformed into their active urethane with the usage of
Bis(4-nitrophenyl)carbonate (BNPC) and N,N-Diisopropylethylamine (DIPEA) in anhydrous
THF, followed by alcoholysis of quercetin. Final deprotection of the tert-butyl protecting
group was achieved by non-aqueous hydrolysis (TFA/CH2Cl2), resulting to the expected
quercetin-amino acid conjugate and characterization was conducted with NMR, mass and IR
spectroscopic techniques (Supplementary Fig. S1-Fig. S13)

Organic & Biomolecular Chemistry
Page 8 of 44
View Article Online
DOI: 10.1039/C7OB02045G
Rapid intramolecular transesterification causes the formation of two inseparable
1
regioisomers in a 3:1 ratio32 (supplementary information) as seen in Figure 1. H NMR
spectra confirmed these results and showed that the major isomer was the 3'-O-substituted
quercetin (a, c) and the minor was 4'-O-substituted quercetin (b, d).
Figure 1: General synthetic scheme for Quercetin-amino acids. i) BNPC, DIPEA, THF, rt, 12h;
ii) Quercetin, rt, 12h; iii) TFA/CH2Cl2 (1/10, v/v), 0 0C to rt, 4h.
Synthesis of 3'-O-CO-(Amino acid-OtBu)-Quercetin (a) and 4'-O-CO-(Amino acid-
OtBu)-Quercetin (b).
DIPEA (1.1 mmol) was added to the solution of OtBu-amino acid (0.550 mmol) and Bis(4-
nitrophenyl)carbonate (0.550 mmol) in anhydrous THF (20 ml) at room temperature under
nitrogen atmosphere. After 12 hours, Quercetin (0.550 mmol) was added and the reaction
mixture was stirred at room temperature for another 12 hours. DIPEA (1.1 mmol) was being
added for 3-4 days until complete consumption of quercetin was verified by TLC. The
reaction’s solvent was evaporated on rotary evaporator and the crude compound was purified
by column chromatography (dichloromethane/methanol, 8:2) to afford inseparable mixture of
the regioisomers (a and b) as yellow solid.
NMR and MS characterization of 3'-O-CO-(Ala-OtBu)-Quercetin (1a) and 4'-O-CO-
(Ala-OtBu)-Quercetin (1b) (76%).
1H-NMR characterization of 1a (500 MHz, DMSO-d6, 25οC): δ = 12.47 (s, 1 H, 5-OH), 10.84
(s, 1 H, 7-OH), 10.40 (s, 1 H, 4’-OH), 9.58 (s,1 H, 3-OH), 8.13 (d, J = 7.3 Hz, 1 H, 8’-NH),
7.94 (dd, J = 8.7 Hz, 2.2 Hz, 1 H, 6’-H), 7.88 (d, J = 2.2 Hz, 1 H, 2’-H), 7.09 (d, J = 8.6 Hz, 1
H, 5’-H), 6.48 (d, J = 2.0 Hz, 1 H, 8-H), 6.23 (d, J = 2.0 Hz, 1 H, 6-H), 4.05- 3.99 (m, 1 H,
9΄-H), 1.46 (s, 9 H, 12΄-H), 1.35 (d, J = 7.2, 3 H, 13’-H) ppm; 13C-NMR (500 MHz, DMSO-

Page 9 of 44
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C7OB02045G
d6, 25οC): δ = 177.04 (C4), 172.64 (C10’), 164.87 (C7), 161.51 (C5), 157.04 (C9), 154.54
(C7’), 152.5 (C4’), 146.64 (C2), 126.62 (C6΄), 123.79 (C2΄), 122.50 (C1’), 117.33 (C5΄),
103.86 (C10), 98.91 (C6), 94.22 (C8), 81.34(C11’), 50.97 (C9’), 28.40 (C12’), 17.74 (C13’)
ppm; 1H-NMR characterization of 1b (500 MHz, DMSO-d6, 25οC): δ = 12.41 (s, 1 H, 5-OH),
10.89 (s, 1 H, 7-OH), 10.00 (s, 1 H, 3’-OH), 9.69 (s, 1 H, 3-OH), 8.12 (d, J = 7.3 Hz, 1 H, 8’-
NH), 7.81 (d, J = 2.0 Hz, 1 H, 2’-H), 7.61 (dd, J = 8.6, 2.0 Hz, 1 H, 6’-H), 7.18 (d, J = 8.6Hz,
1 H, 5’-H), 6.45 (d, J = 2.0 Hz, 1 H, 8-H), 6.24 (d, J = 2.0 Hz, 1 H, 6-H), 4.05- 3.99 (m, 1 H,
9΄-H), 1.46 (s, 9 H, 12΄-H), 1.35 (d, J = 7.2 Hz, 3 H, 13’-H) ppm; 13C-NMR (500 MHz,
DMSO-d6, 25οC): δ = 177.02 (C4), 172.64 (C10’), 164.86 (C7), 161.65 (C5), 157.00 (C9),
154.50 (C7’), 150.31 (C3’), 146.32 (C2), 140.77 (C4’), 137.75 (C3), 129.4 (C1’), 124.20
(C5΄), 119.24 (C6’), 116.55 (C2΄), 103.98 (C10), 98.97 (C6), 94.07 (C8), 81.34 (C11’), 50.97
(C9’), 28.40 (C12’),17.74 (C13’) ppm; Mass characterization of 1a and 1b regioisomers:
Mass: MS (ESI) m/z: [M-H]- for C23H23NO10: calcd: 473.13, found: 472.1.
NMR and MS characterization of 3'-O-CO-(Leu-OtBu)-Quercetin (2a) and 4'-O-CO-
(Leu-OtBu)-Quercetin (2b) (75%):
1H-NMR characterization 2a (500 MHz, DMSO-d6, 25οC): δ = 12.47 (s, 1 H, 5-OH), 10.83
(s, 1H, 7-OH), 10.40 (s, 1 H, 4’-OH), 9.57 (s, 1 H, 3-OH), 8.09 (d, J = 7.8 Hz, 1 H, 8’-NH),
7.94 (dd, J = 8.6 Hz, 2.2 Hz, 1 H, 6’-H), 7.88 (d, J = 2.2 Hz, 1 H, 2’-H), 7.08 (d, J = 8.6Hz, 1
H, 5’-H), 6.48 (d, J = 2.0 Hz, 1 H, 8-H), 6.23 (d, J = 2.0 Hz, 1 H, 6-H), 4.01- 3.96 (m, 1 H,
9΄-H), 1.82-1.74 (m, 1 H, 14’-H), 1.67-1.60 (m, 1 H, 13’a-H), 1.56-1.52 (m, 1 H, 13΄b-H),
1.46 (s, 9 H, 12’H), 0.93 (d, J = 6.5, 6 H, 15’H) ppm; 13C-NMR (500 MHz, DMSO-d6, 25οC):
δ = 176.7 (C4), 172.5 (C10’), 164.75 (C7), 161.62 (C5), 156.95 (C9), 154.80 (C7’), 152.63
(C4’), 146.58 (C2), 139.90 (C3’), 132.45 (C3), 126.63 (C6’), 123.63 (C2΄), 122.62 (C1’),
117.26 (C5΄), 103.86 (C10), 99.09 (C6), 94.21 (C8), 81.49 (C11’), 53.89 (C9’), 40.47 (C13’),
28.50 (C12’), 24.96 (C14’), 22.09 (C15’) ppm.
1H-NMR characterization 2b (500 MHz, DMSO-d6, 25οC): δ = 12.41 (s, 1 H, 5-OH), 10.89
(s, 1 H, 7-OH), 10.00 (s, 1 H, 3’-OH), 9.68 (s, 1 H, 3-OH), 8.10 (d, J = 7.8 Hz, 1 H, 8’-NH),
7.81 (d, J = 2.2 Hz, 1 H, 2’-H), 7.61 (dd, J = 8.6Hz, 2.2 Hz, 1 H, 6’-H), 7.18 (d, J = 8.6Hz, 1
H, 5’-H), 6.45 (d, J = 2.0 Hz, 1 H, 8-H), 6.24 (d, J = 2.0 Hz, 1 H, 6-H), 4.01- 3.96 (m, 1 H,
9΄-H), 1.82-1.74 (m, 1 H, 14’-H), 1.67-1.60 (m, 1 H, 13’a-H), 1.56-1.52 (m, 1 H, 13΄b-H),
1.46 (s, 9 H, 12’H), 0.93 (d, J = 6.5, 6 H, 15’H) ppm; 13C-NMR (500 MHz, DMSO-d6, 25οC):
δ = 172.50 (C10’), 164.93 (C7), 161.73 (C5), 154.80 (C7’), 157.09 (C9), 150.31 (C3’),

Organic & Biomolecular Chemistry
Page 10 of 44
View Article Online
DOI: 10.1039/C7OB02045G
146.20 (C2), 141.00 (C4’), 122.62 (C1’), 119.34 (C6΄), 117.26 (C5΄), 116.56 (C2’), 103.90
(C10), 99.19 (C6), 94.06 (C8), 81.49 (C11’), 53.89 (C9’), 40.47 (C13’), 28.50 (C12’), 24.96
(C14’), 22.09 (C15’) ppm. Mass characterization of 2a and 2b regioisomers:
Mass: MS (ESI) m/z: [M-H]- for C26H29NO10: calcd: 515.18; found: 514.2.
NMR and MS characterization of 3'-O-CO-(Glu-OtBu)-Quercetin (3a) and 4'-O-CO-
(Glu-OtBu)-Quercetin (3b) (71%):
1H-NMR characterization 3a (500 MHz, DMSO-d6, 25οC): δ = 12.47 (s, 1 H, 5-OH), 10.84
(s, 1 H, 7-OH), 10.36 (s, 1 H, 4’-OH), 9.55 (s, 1 H, 3-OH), 8.11 (d, J = 7.5 Hz, 1 H, 8’-NH),
7.95 (dd, J = 8.6 Hz, 2.2 Hz, 1 H, 6’-H), 7.8 (d, J = 2.2 Hz, 1 H, 2’-H), 7.09 (d, J = 8.6Hz, 1
H, 5’-H), 6.48 (d, J = 2.0 Hz, 1 H, 8-H), 6.23 (d, J = 2.0 Hz, 1 H, 6-H), 4.04- 3.99 (m, 1 H,
9΄-H), 2.00 (m, 2 H, 14’-H), 1.87 (m, 2 H, 13’-H), 1.47 (s, 9 H, 17΄-H), 1.45 (s, 9 H, 12’-H)
ppm; 13C-NMR (500 MHz, DMSO-d6, 25οC): δ = 176.81 (C4), 172.31 (C15’), 172.31 (C10’),
164.80 (C7), 161.60 (C5), 157.00 (C9), 154.75 (C7’), 152.44 (C4’), 146.59 (C2), 139.59
(C3’), 126.72 (C6΄), 123.73 (C2΄), 122.63 (C1’), 117.30 (C5΄), 103.92 (C10), 98.96 (C6),
94.25 (C8), 81.7 (C11’), 80.72 (C16’), 54.62 (C9’), 28.48 (C17’), 28.48 (C12’), 27.00 (C14’),
27.00 (C13’) ppm.
1H-NMR characterization 3b (500 MHz, DMSO-d6, 25οC): δ = 12.41 (s, 1 H, 5-OH), 10.88
(s, 1 H, 7-OH), 9.97 (s, 1 H, 3’-OH), 9.68 (s, 1 H, 3-OH), 8.11 (d, J = 7.5 Hz, 1 H, 8’-NH),
7.81 (d, J = 2.2 Hz, 1 H, 2’-H), 7.62 (dd, J = 8.6, 2.2 Hz, 1 H, 6’-H), 7.18 (d, J = 8.6Hz, 1 H,
5’-H), 6.45 (d, J = 2.0 Hz, 1 H, 8-H), 6.24 (d, J = 2.0 Hz, 1 H, 6-H), 4.04- 3.99 (m, 1 H, 9΄-
H), 2.00 (m, 2 H, 14’-H), 1.87 (m, 2 H, 13’-H), 1.47 (s, 9 H, 17΄-H), 1.45 (s, 9 H, 12’-H)
ppm; 13C-NMR (500 MHz, DMSO-d6, 25οC): δ = 177.03 (C4), 172.31 (C10’), 172.30 (C15’),
164.80 (C7), 161.60 (C5), 157.00 (C9), 154.75 (C7’), 150.21 (C3’), 146.36 (C2), 140.86
(C4’), 129.55 (C1΄), 124.15 (C5΄), 119.33 (C6’), 116.61 (C2΄), 103.92 (C10), 98.96 (C6),
94.14 (C8), 81.70 (C11’), 80.72 (C16’), 54.62 (C9’), 28.48 (C17’), 28.48 (C12’), 27.05
(C14’), 27.05 (C13’) ppm. Mass characterization of 3a and 3b regioisomers: Mass: MS
(ESI) m/z: [M-H]- for C29H33NO12: calcd: 587.20; found: 586.3.
NMR and MS characterization 4'-O-CO-(Phe-OtBu)-Quercetin (4a) and 4'-O-CO-
(Phe-OtBu)-Quercetin (4b) (67%):

Page 11 of 44
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C7OB02045G
1H-NMR characterization 4a (500 MHz, DMSO-d6, 25οC): δ = 12.47 (s, 1 H, 5-OH), 10.84
(s, 1H, 7-OH), 10.39 (s, 1 H, 4’-OH), 9.57 (s, 1 H, 3-OH), 8.20 (d, J = 7.7 Hz, 1 H, 8’-NH),
7.93 (dd, J = 8.7 Hz, 2.2 Hz, 1 H, 6’-H), 7.82 (d, J = 2.2 Hz, 1 H, 2’-H), 7.36-7.34 (m, 4 H,
15’ and 16’-H), 7.30-7.26 (m, 1 H, 17’-H), 7.08 (d, J = 8.7Hz, 1 H, 5’-H), 6.48 (d, J = 2.0 Hz,
1 H, 8-H), 6.23 (d, J = 2.0 Hz, 1 H, 6-H), 4.22- 4.17 (m, 1 H, 9΄-H), 3.08-2.98 (m, 2 H, 13’-
H), 1.39 (s, 9 H, 12’-H) ppm; 13C-NMR (500 MHz, DMSO-d6, 25οC): δ = 176.81 (C4),
171.42 (C10’), 164.76 (C7), 161.63 (C5), 156.98 (C9), 154.59 (C7’), 152.45 (C4’), 146.56
(C2), 139.54 (C3’), 138.2 (C14’), 136.89 (C3), 130.03 (C15’), 129.05 (C16’), 127.32 (C17’),
126.69 (C6’), 123.69 (C2΄), 122.45 (C1’), 117.25 (C5΄), 103.91 (C10), 98.98 (C6), 94.25
(C8), 81.66 (C11’), 57.08 (C9’), 37.56 (C13’), 28.43 (C12’) ppm.
1H-NMR characterization 4b (500 MHz, DMSO-d6, 25οC): δ = 12.41 (s, 1 H, 5-OH), 10.88
(s, 1 H, 7-OH), 9.99 (s, 1 H, 3’-OH), 9.68 (s, 1 H, 3-OH), 8.21 (d, J = 7.8 Hz, 1 H, 8’-NH),
7.80 (d, J = 2.1 Hz, 1 H, 2’-H), 7.59 (dd, J = 8.6 Hz, 2.1 Hz, 1 H, 6’-H), 7.36-7.34 (m, 4 H,
15’ and 16’-H), 7.30-7.26 (m, 1 H, 17’-H), 7.10 (d, J = 8.6Hz, 1 H, 5’-H), 6.44 (d, J = 2.0 Hz,
1 H, 8-H), 6.24 (d, J = 2.0 Hz, 1 H, 6-H), 4.22- 4.17 (m, 1 H, 9΄-H), 3.08-2.98 (m, 2 H, 13’-
H), 1.39 (s, 9 H, 12’-H) ppm; 13C-NMR (500 MHz, DMSO-d6, 25οC): δ = 177.07 (C4),
171.46 (C10’), 164.95 (C7), 161.74 (C5), 157.04 (C9), 154.59 (C7’), 150.27 (C3’), 146.27
(C2), 140.87 (C4’), 138.20 (C14’), 130.03 (C15’), 129.05 (C16’), 127.32 (C17’), 124.05
(C5’), 122.45 (C1’), 119.32 (C6΄), 116.59 (C2΄), 104.02 (C10), 98.98 (C6), 94.14 (C8),
81.66 (C11’), 57.08 (C9’), 37.56 (C13’), 28.43 (C12’) ppm. Mass characterization of 4a and
4b regioisomers:
Mass: MS (ESI) m/z: [M-H]- for C29H27NO10: calcd: 549.16; found: 548.2.
Synthesis of 3'-O-CO-Amino acid-Quercetin (c) and 4'-O-CO-Amino acid-Quercetin
(d):
To a solution of a and b (1mmol) in anhydrous dichloromethane (10 ml), TFA (1 ml) was
added at 0oC and the reaction was stirred at room temperature for 4 hours. After the
completion of the reaction, anhydrous hexane (5 ml) was added to the reaction mixture and
the yellow precipitate was filtered, washed twice with anhydrous hexane (5 ml X 2) and dried
well on high vacuum to afford inseparable mixture of the regioisomers (c and d) as yellow
solid.

Organic & Biomolecular Chemistry
Page 12 of 44
View Article Online
DOI: 10.1039/C7OB02045G
NMR and MS characterization of 3'-O-CO-Ala-Quercetin (1c) and 4'-O-CO-Ala-
Quercetin (1d) (94%).
1H-NMR characterization 1c (500 MHz, DMSO-d6, 25οC): δ = 12.66 (bs, 1 H, 11’-OH),
12.47 (s, 1 H, 5OH), 10.85 (s, 1 H, 7–OH), 10.39 (s, 1 H, 4’-OH), 9.56 (s, 1 H, 3-OH), 8.10
(d, J = 7.4 Hz, 1 H, 8’-NH), 7.93 (dd, J = 8.6 Hz, 2.2 Hz, 1 H, 6’-H), 7.90 (d, J = 2.2 Hz, 1 H,
2’-H), 7.09 (d, J = 8.6Hz, 1 H, 5’-H), 6.50 (d, J = 2.0 Hz, 1 H, 8-H), 6.23 (d, J = 2.0 Hz, 1 H,
6-H), 4.12- 4.09 (m, 1 H, 9΄-H), 1.39- 1.37 (m, 3 H, 11΄-H) ppm; 13C-NMR (500 MHz,
DMSO-d6, 25οC): δ = 176.5 (C4), 174.7 (C10’), 164.83 (C7), 161.6 (C5), 157.04 (C9),
154.46 (C7’), 152.63 (C4’), 146.57 (C2), 139.4 (C3’), 126.45 (C6΄), 123.73 (C2΄), 122.77
(C3), 122.66 (C1΄), 117.19 (C5΄), 103.95 (C10), 103.91 (C8), 99.05 (C6), 50.33 (C9’), 18.13
(C11’) ppm.
1H-NMR characterization 1d (500 MHz, DMSO-d6, 25οC): δ = 12.66 (bs, 1 H, 11’-OH),
12.41 (s, 1 H, 5OH), 10.87 (s, 1 H, 7–OH), 9.99 (s, 1H, 3’-OH), 9.68 (s, 1H, 3-OH), 8.09 (d, J
= 7.4 Hz, 1 H, 8’-NH), 7.81 (d, J= 2.1 Hz, 1 H, 2’-H), 7.61 (dd, J = 8.6 Hz, 2.2 Hz, 1 H, 6’-
H), 7.19 (d, J = 8.6Hz, 1 H, 5’-H), 6.45 (d, J = 2.0 Hz, 1 H, 8-H), 6.24 (d, J = 2.0 Hz, 1 H, 6-
H), 4.12- 4.06 (m, 1 H, 9΄-H), 1.39- 1.37 (m, 3 H, 11΄-H) ppm; 13C-NMR (500 MHz,
DMSO-d6, 25οC): δ = 176.5 (C4), 174.7 (C10’), 164.83 (C7), 161.7 (C5), 156.92 (C9),
154.46 (C7’), 150.22 (C3’), 146.39 (C2), 140.97 (C4’), 129.29 (C1΄), 124.10 (C5΄), 122.77
(C3), 119.23 (C6΄), 116.54 (C2΄), 104.06 (C10), 99.12 (C6), 94.02 (C8), 50.1 (C9’), 18.13
(C11’) ppm. Mass characterization of 1c and 1d regioisomers:
Mass: MS (ESI) m/z: [M+H]+ for C19H15NO10: calcd: 417.07; found: 417.9, 438.0 [M+Na]+.
NMR and MS characterization of 3'-O-CO-Leu-Quercetin (2c) and 4'-O-CO-Leu-
Quercetin (2d) (90%):
1H-NMR characterization 2c (500 MHz, DMSO-d6, 25οC): δ = 12.67 (bs, 1 H, 10’-OH),
12.47 (s, 1 H, 5OH), 10.83 (s, 1H, 7-OH), 10.39 (s, 1 H, 4’-OH), 9.57 (s, 1 H, 3-OH), 8.07 (d,
J = 8.1 Hz, 1 H, 8’-NH), 7.94 (dd, J = 8.6 Hz, 2.0 Hz, 1 H, 6’-H), 7.89 (d, J = 2.0 Hz, 1 H, 2’-
H), 7.08 (d, J = 8.6Hz, 1 H, 5’-H), 6.49 (d, J = 1.8 Hz, 1 H, 8-H), 6.23 (d, J = 1.8 Hz, 1 H, 6-
H), 4.08- 4.03 (m, 1 H, 9΄-H), 1.82-1.76 (m, 1 H, 12’-H), 1.69-1.62 (m, 1 H, 11’a-H), 1.60-
1.54 (m, 1 H, 11΄b-H), 0.94 (d, J= 6.5Hz, 6 H, 13’-H) ppm; 13C-NMR (500 MHz, DMSO-d6,
25οC): δ = 176.76 (C4), 174.77 (C10’), 164.82 (C7), 161.58 (C5), 157.95 (C9), 152.55 (C4’),

Page 13 of 44
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C7OB02045G
146.69 (C2), 139.7 (C3’), 136.77 (C3), 126.54 (C6’), 123.63 (C2΄), 117.22 (C5΄), 103.91
(C10), 99.09 (C6), 94.27 (C8), 53.08 (C9’), 40.57 (C11’), 24.95 (C12’), 22.00 (C13’) ppm.
1H-NMR characterization 2d (500 MHz, DMSO-d6, 25οC): δ = 12.67 (bs, 1 H, 10’-OH),
12.41 (s, 1 H, 5OH), 10.89 (s, 1 H, 7-OH), 9.96 (s, 1 H, 3’-OH), 9.68 (s, 1 H, 3-OH), 8.07 (d,
J = 8.1 Hz, 1 H, 8’-NH), 7.80 (d, J = 2.0 Hz, 1 H, 2’-H), 7.62 (dd, J = 8.6Hz, 2.0 Hz, 1 H, 6’-
H), 7.18 (d, J = 8.6 Hz, 1 H, 5’-H), 6.45 (d, J = 1.8 Hz, 1 H, 8-H), 6.24 (d, J = 1.8 Hz, 1 H, 6-
H), 4.08- 4.03 (m, 1 H, 9΄-H), 1.82-1.76 (m, 1 H, 12’-H), 1.69-1.62 (m, 1 H, 11’a-H), 1.60-
1.54 (m, 1 H, 11΄b-H), 0.94 (d, J = 6.5, 6 H, 13’H) ppm; 13C-NMR (500 MHz, DMSO-d6,
25οC): δ = 177.01 (C4), 174.77 (C10’), 165.09 (C7), 161.64 (C5), 157.15 (C9), 150.20 (C3’),
146.40 (C2), 141.00 (C4’), 137.76 (C3), 124.08 (C5’), 119.35 (C6΄), 116.51 (C2΄), 103.95
(C10), 99.24 (C6), 94.02 (C8), 53.08 (C9’), 40.57 (C11’), 24.95 (C12’), 22.00 (C13’) ppm.
Mass characterization of 2c and 2d regioisomers: Mass: MS (ESI) m/z: [M+H]+ for
C22H21NO10:
calcd:
459.12;
found:
459.4.
NMR and MS characterization 3'-O-CO-Glu-Quercetin (3c) and 4'-O-CO-Glu-
Quercetin (3d) (86%):
1H-NMR characterization 3c (500 MHz, DMSO-d6, 25οC): δ = 12.85 (bs), 12.47 (s, 1 H, 5-
OH), 10.84 (s, 1 H, 7-OH), 10.36 (s, 1 H, 4’-OH), 9.55 (s, 1 H, 3-OH), 8.14 (d, J = 8.2 Hz, 1
H, 8’-NH), 7.92 (dd, J = 8.6 Hz, 2.2 Hz, 1 H, 6’-H), 7.79 (d, J = 2.2 Hz, 1 H, 2’-H), 7.06 (d, J
= 8.6Hz, 1 H, 5’-H), 6.48 (d, J = 2.0 Hz, 1 H, 8-H), 6.23 (d, J = 2.0 Hz, 1 H, 6-H), 4.27 - 4.22
(m, 1 H, 9΄H), 3.16 - 3.11 (m, 2 H, 12’-H), 3.01 - 3.29 (m, 2 H, 11’-H) ppm; 13C-NMR (500
MHz, DMSO-d6, 25οC): δ = 177.01 (C4), 173.77 (C13’), 173.44 (C10’), 164.81 (C7), 161.67
(C5), 156.88 (C9), 154.99 (C7’), 152.26 (C4’), 146.70 (C2), 139.42 (C3’), 126.7 (C6΄),
123.83 (C1΄), 123.73 (C2’), 117.31 (C5΄), 103.99 (C10), 98.98 (C6), 94.25 (C8), 56.46 (C9’),
37.36 (C11’), 37.36 (C12’) ppm.
1H-NMR characterization 3d (500 MHz, DMSO-d6, 25οC): δ = 12.85 (bs), 12.41 (s, 1 H, 5-
OH), 10.88 (s, 1 H, 7-OH), 9.97 (s, 1 H, 3’-OH), 9.68 (s, 1 H, 3-OH), 8.15 (d, J = 8.2 Hz, 1
H, 8’-NH), 7.78 (dd, J = 2.2 Hz, 1 H, 2’-H), 7.58 (d, J = 8.6, 2.2 Hz, 1 H, 6’-H), 7.07 (d, J =
8.4Hz, 1 H, 5’H), 6.44 (d, J = 2.0 Hz, 1 H, 8-H), 6.23 (d, J = 2.0 Hz, 1 H, 6-H), 4.27 - 4.22
(m, 1 H, 9΄-H), 3.16 - 3.11 (m, 2 H, 12’-H), 3.01 - 3.29 (m, 2 H, 11’-H) ppm; 13C-NMR (500
MHz, DMSO-d6, 25οC): δ = 177.01 (C4), 173.77 (C13’), 173.44 (C10’), 165.07 (C7), 161.67
(C5), 157.23 (C9), 154.99 (C7’), 150.28 (C3’), 146.50 (C2), 140.89 (C4’), 129.47 (C1΄),

Organic & Biomolecular Chemistry
Page 14 of 44
View Article Online
DOI: 10.1039/C7OB02045G
124.01 (C5΄), 119.30 (C6’), 116.54 (C2΄), 104.03 (C10), 98.98 (C6), 94.25 (C8), 56.46 (C9’),
37.36 (C11’), 37.36 (C12’) ppm. Mass characterization of 3c and 3d regioisomers:
Mass: MS (ESI) m/z: [M-H]- for C21H17NO12: calcd: 475.08; found: 474.0.
NMR and MS characterization of 3'-O-CO-Phe-Quercetin (4c) and 4'-O-CO-Phe-
Quercetin (4d) (90%):
1H-NMR characterization 4c (500 MHz, DMSO-d6, 25οC): δ = 12.88 (bs, 1H, 11’-OH), 12.47
(s, 1 H, 5OH), 10.85 (s, 1H, 7-OH), 10.37 (s, 1 H, 4’-OH), 9.55 (s, 1 H, 3-OH), 8.14 (d, J = 8
Hz, 1 H, 8’-NH), 7.92 (dd, J = 8.7 Hz, 2.2 Hz, 1 H, 6’-H), 7.79 (d, J = 2.2 Hz, 1 H, 2’-H),
7.36-7.35 (m, 4 H, 14’ and 15’-H), 7.31-7.26 (m, 1 H, 16’-H), 7.06 (d, J = 8.6Hz, 1 H, 5’-H),
6.48 (d, J = 2.0 Hz, 1 H, 8-H), 6.23 (d, J = 2.0 Hz, 1 H, 6-H), 4.27- 4.22 (m, 1 H, 9΄-H), 3.14
(dd, J = 14.0 Hz, 4.4Hz, 1 H, 12a’-H), 2.99 (dd, J = 9.6Hz, 4.2Hz, 1 H, 12b’-H) ppm; 13C-
NMR (500 MHz, DMSO-d6, 25οC): δ = 174.04 (C10’), 165.03 (C7), 161.77 (C5), 157.07
(C9), 152.45 (C4’), 146.74 (C2), 139.57 (C3’), 138.64 (C13’), 130.02 (C14’), 129.01 (C15’),
127.3 (C16’), 126.70 (C6’), 123.83 (C2΄), 122.63 (C1’), 117.40 (C5΄), 104.00 (C10), 99.10
(C6), 94.34 (C8), 56.54 (C9’), 37.39 (C12’) ppm.
1H-NMR characterization 4d (500 MHz, DMSO-d6, 25οC): δ = 12.88 (bs, 1 H, 11’-OH),
12.40 (s, 1 H, 5OH), 10.89 (s, 1 H, 7-OH), 9.97 (s, 1 H, 3’-OH), 9.67 (s, 1 H, 3-OH), 8.15 (d,
J = 7.8 Hz, 1 H, 8’-NH), 7.78 (d, J = 2. Hz, 1 H, 2’-H), 7.58 (dd, J = 8.6 Hz, 2.1 Hz, 1 H, 6’-
H), 7.36-7.35 (m, 4 H, 14’ and 15’-H), 7.32-7.26 (m, 1 H, 16’-H), 7.07 (d, J = 8.6Hz, 1 H, 5’-
H), 6.44 (d, J = 2.2 Hz, 1 H, 8-H), 6.23 (d, J = 2.2 Hz, 1 H, 6-H), 4.27- 4.22 (m, 1 H, 9΄-H),
3.14 (dd, J = 14.0 Hz, 4.4Hz, 1 H, 12a’-H), 2.99 (dd, J = 9.6Hz, 4.2Hz, 1 H, 12b’-H) ppm;
13C-NMR (500 MHz, DMSO-d6, 25οC): δ = 174.04 (C10’), 165.09 (C7), 161.50 (C5), 157.26
(C9), 150.15 (C3’), 146.74 (C2), 140.84 (C4’), 138.64 (C13’), 130.02 (C14’), 129.64 (C1’),
129.01 (C15’), 127.30 (C16’), 124.16 (C5’), 119.50 (C6΄), 116.66 (C2΄), 104.00 (C10), 99.10
(C6), 94.23 (C8), 56.54 (C9’), 37.39 (C12’) ppm. Mass characterization of 4c and 4d
regioisomers:
Mass: MS (ESI) m/z: [M+H]+ for C25H19NO10: calcd, 493.10; found: 516.0 [M+Na]+.
Structures, molar ratios and 2D NMR spectra of the four quercetin – amino acid analogues
are given in the Supplementary Information (Section 2, Supplementary Figures S2-S5).

Page 15 of 44
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C7OB02045G
Fluorescence spectroscopy studies
Recombinant, human protein Bcl-xL (with Δ197-233 deletions in the C-terminus and Δ45-
84 internal loop between helix R1 and helix R2) was, firstly, expressed as a 6xHis-tagged
protein in E. coli strain BL21 and then, purified by Ni2+-affinity chromatography by Äkta
Prime FPLC (GE Healthcare). Fluorescence spectroscopy studies were performed and
analyzed as described in previous work20. Fluorescence spectra were recorded on a Perkin-
Elmer LS-55 spectrofluorometer at 25 °C. Gradual concentrations of the quercetin analogues
were titrated in Bcl-xL solution (2 uM) and diluted in sodium phosphate buffer (pH= 6.7). A
1.0 cm quartz cuvette was used for measurements. Bcl-xL fluorescence was collected after
excitation at 285 nm. Both emission and excitation slits were set at 10 nm. The maximum
emission wavelength was occurred approximately at 350 nm.
Quercetin analogues absorb at the maximum excitation and emission wavelength of Bcl-
xL, due to the phenol groups that they contain. This absorption can cause quenching of the
Bcl-xL fluorescence and lead to false measurements. Thus, each measurement was corrected
according to the following equation 33:
∗
ꢈ
ꢐ
ꢎ ꢉ
ꢑ ꢒ
ꢄ ꢀꢂꢅꢆ10ꢇ ꢉ
ꢓ
ꢊꢋꢌꢍ
ꢊꢋꢏ
ꢀ
ꢁꢂꢃ
Where Fobs is the maximum measured fluorescence of Bcl-xL and Fcor the corrected value
of the maximum fluorescence of Bcl-xL; ελexc or ελem are the molar extinction factors of
Quercetin analogues at the excitation and emission wavelengths of Bcl-xL; L0 is the total
concentration of the bound and unbound Quercetin analogue; and l is the path length in the
measuring cell. Data analysis achieved by hyperbola curve fitting, using Origin 8.0,
according to the following equation 33, 34
:
ꢈ
ꢐ
ꢀꢔ ꢕ ꢀꢖ ꢗꢔ
ꢀꢔ ꢕ ꢀ ꢄ
ꢘꢙ ꢚ ꢗꢔ
Where, F0 and F are the fluorescence emission intensities in the absence and presence of
the Quercetin analogue, respectively; F∞ is a theoretically occurred value corresponding to
the Bcl-xL maximum fluorescence when it is saturated; and Kd represents the binding affinity
between Bcl-xL and the Quercetin analogue. Each measurement was repeated at least three
times.
Isothermal Titration Calorimetry
The thermodynamic profile of the interaction between the most soluble quercetin-
aminoacid conjugate, Que-Glu, and Bcl-xL was explored through isothermal titration

Organic & Biomolecular Chemistry
Page 16 of 44
View Article Online
DOI: 10.1039/C7OB02045G
calorimetry. Syringe and sample cell were loaded with Que-Glu (1 mM) and Bcl-xL (0.05
mM), respectively. Recombinant human protein Bcl-xL was expressed and purified, as it is
described above. Both molecules were diluted in sodium phosphate buffer containing 4%
DMSO at 298 K. Each titration injected 2 μL of Que-Glu solution in the sample cell and
lasted for 0.8 sec, except from the initial injection that was 0.4 μL. Overall, 17 injections
were performed and the initial injection was excluded from the data analysis. A 60-sec
spacing was set between the titrations. The reference power was set to 6 μcal/sec and a
stirring speed equal to 1000 rpm was used for the syringe. Finally, a 3 sec filter period was
applied. In order to correct the heat dilution effect, one control experiment was conducted
where que-glu was titrated into the buffer and subtracted from the initial experiment. A
Microcal ITC200 was used for the aforementioned measurements and the data were analyzed
through the Origin 7.0 software provided by Microcal. The Gibbs free energy change (ΔG)
was calculated using the equation ΔG=ΔH-TΔS.
2D NMR Spectroscopy studies
Truncated Bcl-xL was expressed and purified for NMR experiments, as we previously
1
reported 20. The 2D H−15N HSQC spectra of Bcl-xL were obtained without and with the
addition of 1:1 ratio of the quercetin analogues.
Molecular Modeling
The computational workflow for the design of the new quercetin inhibitors is described in
scheme 1.

Page 17 of 44
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C7OB02045G
Scheme 1: Workflow for the rational design and synthesis of the quercetin analogues.
Induced Fit Docking (IFD) Simulations: The crystallographic structures of the proteins
shown in Table 1 were used during the IFD simulations of quercetin. The crystal structures of
the protein were prepared using the Protein Preparation Wizard35, missing loops and side
chains were prepared using Prime, ionized at a pH 7.4 using PROPKA36, and minimized
using the OPLS3 force field 37. The structures of quercetin and its analogues were designed
using Maestro 10.2, ionized at the target pH 7.4 using the Ligprep module 38, and their
conformational analysis was performed using ConfGen39. These conformations were then
submitted as starting geometries for the Induced Fit Docking calculations 40-42. At the initial
stage of docking side chains of residues that are within 5 Å of the ligand were trimmed. Three
residues that are within 5 Å of the ligand and have the highest B-factors (above 40) were
refined using Prime 4.0 40. The ligands were redocked using the extra precision mode.
Table 1: Crystallographic structures of Bcl-xL used in the IFD simulations of quercetin.
PDB Id’s
2YXJ43
3ZLR44
Year
2007
2013
Resolution (Å)
2.20
Co-crystallized inhibitor
ΑΒΤ-737
2.03
WEHI-539

Organic & Biomolecular Chemistry
Page 18 of 44
View Article Online
DOI: 10.1039/C7OB02045G
3SPF45
3QKD46
4EHR47
4C5248
4BPK49
3ΙΝQ50
2012
2011
2012
2014
2013
2009
1.70
2.02
2.09
2.05
1.76
2.00
BM-501
Quinazoline sulfonamide
Pyrimidine phenylsulfonamide
Benzoylurea compound (39b)
Alpha beta Puma BH3 peptide 5
W1191542
Molecular Dynamics Simulations. The poses with best Induced Fit score for the eight
analogues docked in the crystal structure 3ZLRwere selected as initial structures for the MD
simulations. The MD simulations were carried out with the GPU version of the PMEMD
51, 52
53,
module
from the AMBER 14 simulation package
54. The geometry of every
compound was optimized with the HF/6-31G* basis set (Gaussian 09) 55. The general
AMBER force field (GAFF) was used to obtain force field parameters for the compounds
with RESP charges 56, 57. The system was next solvated in a truncated octahedron with 10,170
TIP3P water molecules. The minimum distance between the edge of the periodic box and
each atom of the system was set at 14 Å. The minimization of the system was carried out in
four stages using restraints of 500, 10 and 2 kcal mol-1 Å-2 while in the final stage no
restraints were applied. During these stages, the system was minimized for 15000 steps, using
a nonbonded cutoff of 20 Å. The solvated complex was then heated under constant volume
for 100 ps from 0 to 310 K using a Langevin thermostat 58. Next, the system was equilibrated
under constant pressure in four stages with gradually released restraints of 100, 50, 10 and 2
kcal mol-1 Å-2 for 400 ps for another 100 ps. The bonds involving hydrogen atoms were
constrained at their equilibrium distance using SHAKE 59. The collision frequency (γ) was set
at 2 ps-1. The simulation of each system was run for 100 ns. The analysis of the trajectories
was done with the ptraj module of AMBER 60.
MM-PBSA. The molecular mechanics Poisson–Boltzmann Surface Area (MM-PBSA)
method was used for the calculation of the binding energy between the eight analogues of
quercetin and Bcl-xL 61-64
.
10000 MD snapshots from the MD run of the complex were considered for the calculation
of the enthalpy contribution (ΔH). Topology files for the complex (stripped of water), the
ligands, and Bcl-xL were created using the ante-mmpbsa utility. For each snapshot, a free
energy for the complex, the quercetin analogues, and Bcl-xL is calculated.
The enthalpy term is calculated as:
ΔH = ΔEgas + ΔGsolv, (1)

Page 19 of 44
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C7OB02045G
where ΔEgas is defined as the molecular mechanical (MM) free energy change upon
complexation of Bcl-xL with the quercetin analogues in the gas phase, while ΔGsolv is the
solvation free energy. ΔEgas is further divided into Coulomb interactions (ΔEele) and van der
Waals interaction terms (ΔEvdW):
ΔEgas = ΔEele + ΔEvdW (2)
The solvation term of eq. 1 is defined as a sum of polar (ΔEPB) and nonpolar (ΔEcavity
)
contributions:
ΔGsolv = ΔGPB + ΔGcavity (3)
Cell Culture
Two well established human prostate cell lines, DU-145 (human prostate cancer
hypotriploid cell line) and PC-3 (human prostate cancer cell line) were treated with the four
compounds (Que-Ala, Que-Leu, Que-Glu and Que-Phe) at concentrations ranging from 1 to
100 μM to test their cytostatic and cytotoxic activity. The cell lines were obtained from the
American Type Culture Collection (ATCC) and were grown in different culture medium
according to the instructions. Cells were routinely grown as monolayer cell cultures in T-75
flasks (Costar) in an atmosphere containing 5% CO2 in air, and 100% relative humidity at 37
°C and subcultured twice a week. The culture medium used were Dulbecco's modified
Eagle's medium, DMEM (Gibco, Glasgow, UK) (for DU-145 cell line) and RPMI 1640
medium (for PC-3 cell line) (Roswell Park Memorial Institute medium) (Gibco, Glasgow,
UK), both supplemented with 10% fetal bovine serum (Gibco, Glasgow, UK), 2 mM
glutamine (Sigma), 100 μg/mL streptomycin and 100 IU/mL penicillin. Adherent cells at a
logarithmic growth phase were detached by addition of 2–3 mL of a 0.05% trypsin (Gibco) –
0.02% EDTA mixture and incubation for 2–5 min at 37 ˚C. The loss of membrane integrity,
as a morphological characteristic for cell death, was assayed by Trypan Blue exclusion. All
chemicals and solvents used were of high purity and purchased from Sigma or Merck.
MTT assay
The MTT ((3-(4, 5-imethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide) assay is a well-
established and standard method for evaluating the cytostatic and cytotoxic activity of drugs
and chemicals. Briefly, the cells were plated in 96-well flat-bottom microplates (Costar-
Corning, Cambridge) (100 μL containing 5,000 cells/well for PC-3 and 8,000 cells/well for

Organic & Biomolecular Chemistry
Page 20 of 44
View Article Online
DOI: 10.1039/C7OB02045G
the rest cell lines) and maintained for 72 h at 37 °C in a 5% CO2 incubator and grown as
monolayers. After 24 h, cells were treated with 1–100 μmol/l of the compounds for 48 h. The
viability of cultured cells was estimated using MTT (Sigma, St Louis, Missouri, USA)
metabolic assay as described previously. MTT was dissolved in PBS in a concentration of 5
mg/mL, filter sterilized, and stored at 4°C. MTT (50 μL of stock solution) was added to each
culture and incubated for 3 h at 37 °C to allow metabolization. Formazan crystals were
solubilized by DMSO (100 μL) Absorbance of the converted dye was measured at a
wavelength of 540 nm on an ELISA reader (Versamax, Orleans, USA)65.
The mean concentrations of each drug that generated 50% or total (100%) growth
inhibition (GI50 and TGI, respectively) as well as the drug concentrations that produced
cytotoxicity against 50% of the cultured cells [(half maximal cytotoxic concentration (IC50)]
were calculated using the linear regression method. GI50 and TGI concentrations calculated
after 48ꢀh exposure. Results are mean of three independent experiments. GI50 was calculated
as the drug concentration that reduced the number of cells to 50% of the number of cells
before each compound addition. TGI is the drug concentration that achieved total growth
inhibition of the cells. Using seven absorbance measurements [time 24 h (Ct24), control
growth 72 h (Ct72), and test growth in the presence of drug at five concentration levels
(Tt72x)], the percentage of growth was calculated at each level of the drug concentrations.
The percentage growth inhibition was calculated according to National Cancer Institute
(NCI) as: [(Tt72x) - (Ct24)/(Ct72) - (Ct24)] × 100 for concentrations for which Tt72x>Ct24
and [(Tt72x) - (Ct24)/Ct24] × 100 for concentrations for which Tt72x< Ct24; GI50 was
calculated from [(Tt72x)-(Ct24)/(Ct72) - (Ct24)] × 100 = 50,TGI from [(Tt72x) -
(Ct24)/(Ct72) - (Ct24)] × 100 = 0, and IC50 from [(Tt72x) - (Ct24)/Ct24] × 100 = 50. All the
experiments were carried out in triplicate.
Results and Discussion
Computational Approaches
IFD simulations and rational design: To rationally design the quercetin-amino acid
analogues able to bind to the BH3 binding site of Bcl-xL we used for the IFD process the
different available crystal structures shown in Table 1. Since the preferred binding site of
quercetin for Bcl-xL has already been described by Tzakos et al.20 this information was used
to validate the docking poses of quercetin in the eight different crystal structures. Although
all eight crystal structures predicted the binding site for quercetin20 only the structures with

Page 21 of 44
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C7OB02045G
PDB ID 3ZLR 44 and 2YXJ 43 provided the binding poses with the highest GlideScore and the
1
best correlation with our previously published 2D H-15N HSQC NMR data20. As can be
observed in Figure 2Α quercetin is inserted in the hydrophobic pocket P2 with its B ring
directed towards the surface of the protein. Although quercetin develops multiple interactions
with Bcl-xL it cannot interact with polar residues such as Tyr195, Arg139, Tyr101 (Figure
2C) that are located in the neighborhood of ring B and play an important role in the binding
of ABT737 and WEHI-539 43, 44. Addition of a complementary negatively charged group to
the ring B of quercetin and specifically at positions 3´ or 4´ could develop favorable
interactions with the neighborhood positively charged residues of Bcl-xL such as Arg139 and
Asn136 (Figure 2B). To achieve this we envisaged conjugating different amino acids at the 3´
or 4´ positions of quercetin via their α-amino group through carbamate bonds (Figure 2D). To
evaluate this, substitutions at both positions were investigated computationally. Quercetin
was conjugated with 4 representative residues: alanine (the simplest residue after glycine),
leucine (hydrophobic), phenylalanine (aromatic) and glutamic acid (negatively charged). All
eight generated compounds have a negatively charged carboxyl group that can opperate as a
hydrogen bond acceptor for the positively charged residues of the surface of Bcl-xL.

Organic & Biomolecular Chemistry
Page 22 of 44
View Article Online
DOI: 10.1039/C7OB02045G
Figure 2: (A) Binding pose of quercetin44. (B) Rational design of quercetin analogues. (C)
The average distance from the 4´ and 3´ positions and the polar residues Asn136, Tyr195 and
Tyr101. (D) Quercetin substituted at the 4´ position with the residue leucine (top) and at the
3´ position by glutamic acid (bottom). The exposed carboxyl group is situated at a distance of
5 Å from the positions 3´ and 4´, in order to develop hydrogen bonds with Asn136. This
suggests the linking of the amino acids via carbamate bonds. The second carboxyl group in
the case of quercetin –glutamic acid conjugate is located in the appropriate position to
develop interactions with the polar residues Tyr195 and Tyr101.
Figure 3 illustrates the binding poses of the quercetin analogues substituted at the 3´
position (defined through the text as Que-amino acid3). The carboxyl group of Que-Ala3
develops multiple hydrogen bonds with Arg139 and Asn136. The hydroxyl groups at
positions 3 and 5 develop hydrogen bonds with Ser106 and Leu108. Glu98 develops a
hydrogen bond with the hydroxyl group at position 7 of quercetin segment (Figure 3A).
Multiple hydrophobic interactions are also developed with quercetin segment and Bcl-xL
residues Ala142, Phe146, Ala149, Phe97, Trp137 and Tyr101 (Figure S14A). The two
carboxyl groups of the Que-Glu3 analogue develop hydrogen bonds with residues Arg139,
Asn136, Tyr101 and Tyr195. Hydrogen bonds are also developed between the hydroxyl
group at position 5 of the compound and Leu108 and Ser106 (Figure 3B and Figure S14B).
Identical hydrogen bonds are also developed by Que-Phe3. Furthermore, the development of
π-π stacking interactions of quercetin segment with Phe105 and Tyr195 (Figure 3C and
S14C) were observed. The analogue Que-Leu3 develops multiple hydrogen bonds with
residues Asn136, Arg139, Asp133, Ala142 and Ser106 (Figure 3D and S14D). The binding
poses of these compounds in the crystal structure 2YXJ illustrate very similar interactions
(Figure S16 and S17). The Que-Ala3 analogue develops multiple hydrogen bonds with
Ala142, Leu130, Arg139 and Asn136 (Figure S16A and S17A). The rest of the analogues
develop the same described interactions and additionally a hydrogen bond with Tyr101
(Figure S16 and S17).

Page 23 of 44
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C7OB02045G
Figure 3: Binding poses of the quercetin analogues: A) Que-Ala3; B) Que-Glu3; C) Que-
Phe3 and D) Que-Leu3 in the crystal structure of Bcl-xL with PDB ID 3ZLR.
The binding poses of the quercetin analogues substituted at 4´ (defined through the text as
Que-amino acid4) are illustrated in Figure 4. The critical residues that develop interactions
remain mostly the same. Que-Ala4 develops hydrogen bonds with Ser106, Leu108, Arg139,
Asn136, Tyr101 and Glu98 (Figure 4A and S15A). The residues Ser106, Ser145, Asn136,
Arg139, Tyr195, Gly138 develop hydrogen bonds with Que-Glu4. Furthermore, a π-π
stacking interaction is developed with Phe105 (Figure 4B and S15B). Similar interactions are
developed between the other two analogues (Que-Phe4 and Que-Leu4) and Bcl-xL (Figure
4C, D and S15C, D). Similar binding poses are predicted by the IFD calculations of the four
compounds in the 2YXJ crystal structure (Figures S18 and S19).
To better evaluate the binding poses of the newly designed compounds they were
superimposed with the crystal structures of the established Bcl-xL inhibitors ABT-737 and
WEHI-539 (in Figure 5 is shown the Que-Glu3 analogue). As it can be observed in Figure 5A
both ABT-737 and the newly synthetic conjugate bind in the same manner to the Bcl-xL
protein. The rings A, B and C of the Que-Glu3 analogue bind to the hydrophobic core P2 as
does the 2-(4-chlorophenyl)phenyl]methyl moiety of ABT-737 (Figure 5A,B) and the 1,3-
benzothiazol-2-ylhydrazinylidene moiety of WEHI-539 (Figure 5 C,D). The carboxyl part of

Organic & Biomolecular Chemistry
Page 24 of 44
View Article Online
DOI: 10.1039/C7OB02045G
Que-Glu3 analogue overlaps with the sulfonyl moiety of ABT-737 (red circle in Figure 5B)
and the carboxyl moiety of WEHI-539 (red circle in Figure 5D). These superimpositions
suggest that the designed quercetin analogues possess the appropriate hydrophobic moiety to
interact with the P2 core and the negatively charged moieties to interact with the positively
charged residues located in the protein surface.
Figure 4: Binding poses of the quercetin analogues: A) Que-Ala4; B) Que-Glu4; C) Que-
Phe4 and D) Que-Leu4 in the crystal structure of Bcl-xL with PDB ID 3ZLR.

Page 25 of 44
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C7OB02045G
Figure 5: A, B) Superimposition of the Que-Glu3 analogue (carbon atoms are in blue color)
to the ABT-737 structure (carbon atoms are in green color). C, D) Superimposition of the
Que-Glu3 analogue to the WEHI-539. In panels A and C the surface of the protein is showed
using an electrostatic potential map (red negative, blue positive, white neutral).
Molecular Dynamics Simulations: To assess the stability of the quercetin-analogues – Bcl-
xL complexes MD simulations of 100 ns were run for all analogues using the coordinates of
Bcl-xL from the PDB ID 3ZLR 44. The root mean squared deviations (RMSDs) of the protein
and ligands are shown in Figure 6. As previously reported by Guo et al. 66 the RMSDs of the
complexes reach a stable equilibrium after 20 ns in most cases. The average RMSD for the
simulations goes from 5 Å (in the case of Que-Glu4) to 12 Å for Que-Leu3. These numbers
67
are larger compared to other simulations reported by Novak et al.
probably due to the
shorter simulation time in the latter case. The lower RMSDs of the protein in its bound states
manifest a ligand stabilization effect on the flexible region of the active site 66. As a result the
analogues Que-Glu3, Que-Glu4 and Que-Phe4 that display the lowest RMSDs display the
largest ligand stabilization effects.
The root mean squared fluctuations (RMSF) of the Cα atoms of all eight simulation
systems over the MD runs are shown in Figure 6. The more flexible regions are highlighted
by higher RMSF values. The connecting loop between α1 and α2 helices (residues 29-42) is
the most flexible region as previously reported by Guo et al.66. The loop residues fluctuate at
an average RMSF of around 5Å. The disordered nature of this segment can explain these high

Organic & Biomolecular Chemistry
Page 26 of 44
View Article Online
DOI: 10.1039/C7OB02045G
fluctuations. The residues around the binding pocket (95-150) show an average RMSF of 1-2
Å and do not display the significant deviations of an average RMSF of 4 Å reported by Guo
66
et al. signifying the further stabilization of the newly synthesized analogs to the highly
flexible BH3 binding site.
Hydrogen bond analyses of all trajectories are shown in Table 2. The analogue Que-Ala3
develops multiple H-bond interactions with the residues Arg103, Asn136 and Arg139. These
interactions are very stable over time. Two further interactions are developed with Ala104
and Ser106. Que-Ala4 does not develop interactions with the residues Arg103, Ala104 and
Asn136 but present a very stable H-bond with Ser106 (94% of the simulation time).
Furthermore, interactions are developed with residues Thr109, Arg132 and Ser145. The Que-
Glu3 analogue develops multiple interactions with the residues Arg100, Tyr101, Asn136, and
Arg139. A stable H-bond (75.54% of the simulation time) is developed with Ser106. The
Que-Glu4 analogue develops multiple H-bonds with Tyr101, Arg139 and Ser145. Stable H-
bonds are also developed with Glu98, Ser106, Thr109, Arg132 and Asn136. The Que-Phe3
and Que-Phe4 analogues develop few H-bonds mainly with Ser106, Asn136 and Arg139. The
same interactions are developed for the Que-Leu analogues. From these analyses it became
evident that the hydrogen bonds developed among the quercetin analogues and the positive
patch (localized in the neighborhood of ring B of quercetin and consisted of the polar
residues Asn136 and Arg139), are stable (> 50% of the simulation time) for all analogues.
This evidence aligns with our initial hypothesis that incorporating a negative charge close to
ring B of quercetin could lead to analogues developing favorable interactions with the
positively charged patch localized in the neighborhood of the Bcl-xL protein surface.

Page 27 of 44
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C7OB02045G
Figure 6: RMSD of the backbone of the protein Bcl-xL (in blue) and the ligands (in red): A)
Que-Ala3; B) Que-Ala4; C) Que-Glu3; D) Que-Glu4; E) Que-Phe3; F) Que-Phe4; G) Que-
Leu3 and H) Que-Leu4.

Organic & Biomolecular Chemistry
Page 28 of 44
View Article Online
DOI: 10.1039/C7OB02045G
Figure 7: RMSF of the protein’s residues in the Cα atoms for the analogues A) Que-Ala3; B)
Que-Ala4; C) Que-Glu3; D) Que-Glu4; E) Que-Phe3; F) Que-Phe4; G) Que-Leu3 and H)
Que-Leu4.

Page 29 of 44
Organic & Biomolecular Chemistry
View Article Online
DOI: 10.1039/C7OB02045G
Table 2: Occurrence of hydrogen bonds between the quercetin analogues and the Bcl-xL
protein.
Occurrence a
Residues
in H-
Que-Ala3
Que-Ala4
Que-Glu3
Que-Glu4
Que-Phe3
Que-Phe4
Que-Leu3
Que-Leu4
bonds
Glu98
27.00
Arg100
67.89/59.07
/57.40/34.94
66.39/24.46
/19.63
Tyr101
34.38/10.50
Arg102
Arg103
24.12
44.50/17.37
/15.87/14.88
b
Ala104
Ser106
Thr109
Arg132
Asn136
Arg139
12.75
10.37
94.00
75.54
82.50
13.25
88.00
27.00
77.00
41.43
22.88
48.88
15.13
42.88/38.37
49.25/41.50
/14.88/12.00
50.29/44.76
57.40/52.5
11.25
32.88/25.00
40.38/33.75
/28.12/26.37
34.75/18.87
36.50/31.00
/26.12/16.50
25.25/23.25
44.25/39.88
/32.88/27.88
20.87/17.25
25.46/23.96
46.09/45.09
/43.76/41.26
70.63/69.00
/46.13/38.00
89.88
68.88/62.50
/52.88/37.87
65.50/13.00
/43.59/38.27
Ser145
aOccurrence is defined as the percentage of simulation time that a specific interaction
b
exists; interactions occurring less than 10% of the simulation are not shown. Multiple
hydrogen bonds are formed between different atoms of the inhibitor and the specific residue.
MM-PBSA and MM-GBSA calculations: To assess the theoretical binding affinities of
all complexes MM-PBSA and MM-GBSA calculations were run on all eight simulation
trajectories to measure the binding enthalpies for all eight compounds. The average values for
all analogues are shown in Tables 3 and 4. The MM-PBSA calculations pinpoint that the
Que-Glu3 analogue displays the highest binding enthalpy (-47.16 kcal/mol). The change from
Que-Glu3 to Que-Glu4 causes a higher enthalpy with a difference of 16 kcal/mol (ΔHQue-Glu4
= -31.05 kcal/mol). The most favored enthalpies after the Que-Glu3 analogue are displayed
by the Que-Ala3 analogue (-36.65 kcal/mol) and the Que-Phe4 analogue (-36.48 kcal/mol).