Synthesis and serum protein binding of novel ring-substituted harmine derivatives

Article abstract of DOI:10.1039/c5ra06426k

A series of new derivatives of the natural β-carboline alkaloid harmine, introducing hydrophobic substituents into positions 7 and 9 were synthesized as potential anticancer agents. Their binding affinities for human serum albumin (HSA) and α₁-acid glycoprotein (AAG) were investigated by affinity chromatography combined with fluorescence, circular dichroism (CD) and UV absorption spectroscopy. The weak binding of harmine to both proteins (K_a ～ 3 × 10⁴ M^-1) was highly increased by aromatic substitutions (K_a ～ 10⁵-10⁶ M^-1). Derivatives having a substituted benzyl group in the N⁹-position of the β-carboline nucleus showed about tenfold and hundredfold affinity enhancement for HSA and AAG, respectively. Such a strong plasma protein interaction would be of pharmacokinetic relevance for these potential drug candidates. Induced CD spectra indicated the variant selective, dimeric binding of the 7-pyridylethoxy derivative to AAG. Absorbance and fluorescence spectra refer to the binding preference of the neutral form of the studied β-carbolines for both proteins.

Full text of DOI:10.1039/c5ra06426k

RSC Advances
View Article Online
View Journal | View Issue
PAPER
Synthesis and serum protein binding of novel ring-
substituted harmine derivatives†
Cite this: RSC Adv., 2015, 5, 53809
a
a
Celesztina Domonkos, Ferenc Zsila, Ilona Fitos,a Julia Visy,a Rudolf Kassai,b
*
´
Balazs Balintb and Andras Kotschyb
´
´
´
A series of new derivatives of the natural b-carboline alkaloid harmine, introducing hydrophobic
substituents into positions 7 and 9 were synthesized as potential anticancer agents. Their binding
affinities for human serum albumin (HSA) and a1-acid glycoprotein (AAG) were investigated by affinity
chromatography combined with fluorescence, circular dichroism (CD) and UV absorption spectroscopy.
The weak binding of harmine to both proteins (Ka ꢀ 3 ꢁ 104
M
ꢂ1) was highly increased by aromatic
M
ꢂ1). Derivatives having a substituted benzyl group in the N9-position of the
substitutions (Ka ꢀ 105–106
b-carboline nucleus showed about tenfold and hundredfold affinity enhancement for HSA and AAG,
respectively. Such a strong plasma protein interaction would be of pharmacokinetic relevance for these
potential drug candidates. Induced CD spectra indicated the variant selective, dimeric binding of the
7-pyridylethoxy derivative to AAG. Absorbance and fluorescence spectra refer to the binding preference
of the neutral form of the studied b-carbolines for both proteins.
Received 10th April 2015
Accepted 9th June 2015
DOI: 10.1039/c5ra06426k
www.rsc.org/advances
skeleton at positions 1, 2, 6, 7, and 9. Structure–activity rela-
tionship studies indicated that upon incorporation of appro-
Introduction
priate alkyl and aralkyl groups in position 9, the antitumor
effect of harmine against Lewis lung cancer and sarcoma 180
was dramatically enhanced. It was also proved8,9 that the
neurotoxic side effect of these compounds can be decreased by
replacing methoxy group at position 7. We performed the
synthesis of a series of 7- and 9-substituted harmine derivatives
(Scheme 1) in order to study their interaction with human
serum proteins.
Binding of drugs10 and natural products11,12 to blood plasma
proteins inuences their pharmacokinetic and pharmacody-
namic action. Among human serum proteins, the most abun-
dant albumin component (HSA) is of the highest
importance,13,14 but no data are available on the harmine–HSA
interaction. HSA binding of the related b-carboline compound
harmane (R7: H) was found to be weak15 (Ka ꢀ 2.4 ꢁ 104 Mꢂ1),
while the binding of norharmane, lacking 1-methyl substitu-
tion, is much more pronounced (Ka ꢀ 1.7 ꢁ 105 Mꢂ1). Besides
HSA, a1-acid glycoprotein (AAG) also plays a decisive role in the
binding and transportation of a broad array of basic and neutral
Harmine, possessing a tricyclic pyrido[3,4-b]indole ring struc-
ture, is a representative member of the naturally occurring
b-carboline alkaloids. It was originally isolated from the seed of
Peganum harmala and widely utilized for hundreds of years in
traditional Chinese medicine to treat alimentary tract cancer
and malaria.1 This herbal substance is of great interest due to its
diverse biochemical activities including DNA intercalation,1,2
inhibition of topoisomerase1 and cyclin-dependent kinases.1,3
Several studies demonstrated that it inhibits selectively mono-
amine oxidases,4 which play a key role in psychiatric and
neurological disorders5 (depression and Parkinson's diseases).
Harmine was also identied as a potential inhibitor of the
kinase Dyrk1A that is implicated in the pathogenesis of Down
syndrome,6 a common hereditary disorder.
In the past, numerous studies investigated the effect of
harmine on the central nervous system such as its interaction
with benzodiazepine, 5-HT2A, 5HT2C and imidazoline receptors.
Recently, however, harmine and its ring-substituted derivatives
attracted attention as potential cancer drugs. Ishida et al. eval-
uated7 the change in the antitumor activity of this alkaloid on
the introduction of various substituents onto the b-carboline
aResearch Group of Chemical Biology, Institute of Organic Chemistry, Research Centre
for Natural Sciences, Hungarian Academy of Sciences, POB 289, H-1519, Budapest,
Hungary. E-mail: zsila.ferenc@ttk.mta.hu
b
´
Servier Research Institute of Medicinal Chemistry, Zahony utca 7, H-1031, Budapest,
Scheme 1 General chemical structure of harmine (R7: Me and R9: H)
and its derivatives showing the equilibrium between the protonated
and nonprotonated forms (the pKa of harmine is 7.45). The R7 and R9
notations represent the frequently substituted positions.
Hungary
† Electronic supplementary information (ESI) available: Predicted pKa values of
harmine and its synthetic derivatives. See DOI: 10.1039/c5ra06426k
This journal is © The Royal Society of Chemistry 2015
RSC Adv., 2015, 5, 53809–53818 | 53809

View Article Online
RSC Advances
Paper
drug molecules.16,17 AAG association of b-carbolines has not Hz, 1H), 7.79 (d, J ¼ 5.3 Hz, 1H), 6.98 (d, J ¼ 2.2 Hz, 1H), 6.81
been evaluated though these molecules exist as a mixture of (dd, J ¼ 8.5, 2.2 Hz, 1H), 4.27 (sp, J ¼ 6.0 Hz, 1H), 2.71 (s, 3H),
neutral and cationic forms at physiological pH. Thus, in this 1.33 (d, J ¼ 6.0 Hz, 6H); 13C NMR (125 MHz, DMSO-d6) d 158.6,
work we studied the binding of harmine and its derivatives to 142.4, 141.7, 138.2, 135.0, 127.7, 123.1, 115.2, 112.4, 110.7, 97.1,
HSA and AAG. Since plasma AAG is a mixture of two main 70.0, 23.3, 20.8; HRMS calculated for C15H16N2O: 240.1263;
genetic variants with different drug binding abilities,18–21 the found 241.1329 (M + H).
separated F1/S and A genetic variants were tested as well. The
[2]: 1-methyl-7-(1-phenylethoxy)-9H-pyrido[3,4-b]indole. Using
applied experimental methodologies include affinity chroma- 244 mg rac-1-phenylethanol and 3 equivalent cyanomethylene-
tography, as well as uorescence, circular dichroism, and UV tributylphosphorane instead of PPh3 and diethyl azodicarbox-
absorption spectroscopic techniques.
ylate 91 mg 2 (60%) was obtained. 50 mg of the racemate was
separated via chiral chromatography using Chiralpak AD
column and ethanol/heptane (5 : 95) eluent. 2a refers to the
earlier eluting enantiomer.
Experimental details
General methods
2a. 24 mg; retention time: 76 min; 1H NMR (500 MHz, DMSO-
d6) d 11.30 (s, 1H), 8.11 (d, J ¼ 5.3 Hz, 1H), 7.99 (d, J ¼ 8.5 Hz,
1H), 7.75 (d, J ¼ 5.3 Hz, 1H), 7.46 (d, J ¼ 8.3 Hz, 2H), 7.36 (t, J ¼
7.7 Hz, 2H), 7.25 (t, J ¼ 7.3 Hz, 1H), 6.89 (d, J ¼ 2.1 Hz, 1H), 6.87
(dd, J ¼ 8.5, 2.1 Hz, 1H), 5.60 (q, J ¼ 6.4 Hz, 1H), 2.66 (s, 3H),
1.61 (d, J ¼ 6.4 Hz, 3H); 13C NMR (125 MHz, DMSO-d6) d 158.6,
143.6, 142.0, 141.7, 138.2, 134.9, 129.1, 127.9, 127.5, 126.1,
123.0, 115.3, 112.4, 110.8, 97.5, 75.7, 25.0, 20.7.
1H, 13C and 19F NMR spectra were recorded on a Bruker Avance
III 500 instrument. Chemical shis are given in parts per
million (ppm) and spectra are obtained as DMSO-d6 or CDCl3
solutions, using chloroform (7.26 ppm) or DMSO-d6 (2.50 ppm)
as the reference standard. The following abbreviations are used
to denote signal multiplicities: s ¼ singlet, d ¼ doublet, t ¼
triplet, m ¼ multiplet, and br ¼ broadened. All coupling
constants (J) are given in hertz (Hz). Analytical HPLC was per-
formed on Agilent 1200 Series utilizing Waters Acquity UPLC
BEH C18, 2.1 ꢁ 30 mm, 1.7 mm column (solvent A: 100 : 2 : 0.1
H2O/ACN/MeSO3H, solvent B: 100 : 2 : 0.1 ACN/H2O/MeSO3H,
gradient: 0 min 0% B, 6 min 100%, 7 min 100%, 7.5 min 0%,
9 min 0%, ow: 0.8 mL minꢂ1) and UV detection at 210 nm.
High resolution mass spectrometric identication of
compounds was performed using SHIMADZU LCMS-IT-TOF
ESI, WATERS SUNFIRE C18, 2.1 ꢁ 50 mm, 2.5 mm column
(solvent A: 100 : 3 H2O/2-PrOH + 0.05% HCOOH, solvent B:
95 : 5 : 3 H2O/ACN/2-PrOH + 0.05% HCOOH, gradient: 0 min
0% B, 9 min 100%, 11 min 100%, 11.1 min 0%, 15 min 0%, ow:
0.7 mL minꢂ1). Flash chromatography was performed on Tele-
dyne ISCO CombiFlashRF instrument, using RediSept RF pre-
packed silica gel columns and UV detection at 254 or 210 nm.
Reagents and solvents were used as obtained from commercial
suppliers without further purication. Yields refer to puried
products and are not optimized. All tested compounds were
2b. 23 mg; retention time: 112 min; 1H NMR (500 MHz,
DMSO-d6) d 11.30 (s, 1H), 8.11 (d, J ¼ 5.3 Hz, 1H), 7.99 (d, J ¼ 8.5
Hz, 1H), 7.75 (d, J ¼ 5.3 Hz, 1H), 7.46 (d, J ¼ 8.3 Hz, 2H), 7.36 (t, J
¼ 7.7 Hz, 2H), 7.25 (t, J ¼ 7.3 Hz, 1H), 6.89 (d, J ¼ 2.1 Hz, 1H),
6.87 (dd, J ¼ 8.5, 2.1 Hz, 1H), 5.60 (q, J ¼ 6.4 Hz, 1H), 2.66 (s,
3H), 1.61 (d, J ¼ 6.4 Hz, 3H); 13C NMR (125 MHz, DMSO-d6) d
158.6, 143.6, 142.0, 141.7, 138.2, 134.9, 129.1, 127.9, 127.5,
126.1, 123.0, 115.3, 112.4, 110.8, 97.5, 75.7, 25.0, 20.7.
[3]: 1-methyl-7-(2-pyridylmethoxy)-9H-pyrido[3,4-b]indole. Using
1
218 mg 2-pyridinemethanol 120 mg 3 (92%) was obtained. H
NMR (500 MHz, DMSO-d6) d 11.40 (s, 1H), 8.60 (d, J ¼ 4.6 Hz,
1H), 8.13 (d, J ¼ 5.3 Hz, 1H), 8.06 (d, J ¼ 8.7 Hz, 1H), 7.83 (t, J ¼
7.6 Hz, 1H), 7.79 (d, J ¼ 5.3 Hz, 1H), 7.55 (d, J ¼ 7.9 Hz, 1H), 7.35
(t, J ¼ 6.2 Hz, 1H), 7.06 (d, J ¼ 1.6 Hz, 1H), 6.94 (dd, J ¼ 8.7, 1.6
Hz, 1H), 5.30 (s, 2H), 2.69 (s, 3H); 13C NMR (125 MHz, DMSO-d6)
d 159.3, 157.2, 149.6, 142.2, 141.8, 138.2, 137.5, 135, 127.6, 123.5,
123.2, 122.1, 115.6, 112.5, 110.1, 96.3, 71.0, 20.7; HRMS calcu-
lated for C18H15N3O: 289.1215; found 290.1287 (M + H).
1
>95% pure as assessed by LCMS and H NMR.
[4]: 1-methyl-7-[2-(2-pyridyl)ethoxy]-9H-pyrido[3,4-b]indole. Using
244 mg 2-(2-hydroxyethyl)pyridine 121 mg 4 (79%) was obtained.
1H NMR (500 MHz, DMSO-d6) d 11.36 (s, 1H), 8.52 (d, J ¼ 4.7 Hz,
Synthesis of compounds
General procedure A [1–4]. 99 mg harmol (0.5 mmol) and 1H), 8.22 (m, 1H), 8.12 (d, J ¼ 5.3 Hz, 1H), 8.01 (d, J ¼ 8.6 Hz, 1H),
2 mmol of the appropriate alcohol were dissolved in 2 mL dry 7.78 (d, J ¼ 5.3 Hz, 1H), 7.74 (td, J ¼ 7.6, 1.7 Hz, 1H), 7.40 (d, J ¼
tetrahydrofuran, then 0.50 g triphenylphosphine on polymer 7.6 Hz, 1H), 7.25 (dd, J ¼ 7.3, 5.3 Hz, 1H), 7.01 (d, J ¼ 2.1 Hz, 1H),
(3.0 mmol gꢂ1, 1.5 mmol) was added to the solution. Aer 6.80 (dd, J ¼ 8.6, 2.1 Hz, 1H), 4.47 (t, J ¼ 6.5 Hz, 2H), 3.25 (t, J ¼ 6.5
10 minutes 653 mg diethyl azodicarboxylate (1.5 mmol, 653 mL, Hz, 2H), 2.70 (s, 3H). 13C NMR (125 MHz, DMSO-d6) d 159.6,
ꢀ40% in toluene) was added. The mixture was stirred at rt under 149.6, 142.3, 141.7, 138.2, 137.0, 135.0, 127.6, 124.1, 123.1, 122.2,
nitrogen until no further conversion was observed. The mixture 115.4, 112.4, 109.8, 95.8, 67.4, 37.6, 20.8; HRMS calculated for
was ltered, the polymer was washed with tetrahydrofuran, and C19H17N3O: 303.1372; found 304.1437 (M + H).
the combined organic phases were evaporated under reduced
General procedure B [5–8]. 106 mg harmine (0.5 mmol) was
pressure. The residue was puried via ash chromatography dissolved in 2 mL dry DMF at 0 C. 13.2 mg NaH (0.55 mmol)
ꢃ
using hydrophilic interaction liquid chromatography.22
was added to the solution in small portions and the mixture was
[1]: 7-isopropoxy-1-methyl-9H-pyrido[3,4-b]indole. Using 120 stirred at 0 ꢃC until it became clear. Then 0.55 mmol of the
mg 2-propanol 44 mg 1 (37%) was obtained. 1H NMR (500 MHz, appropriate halide was added to the solution and the mixture
DMSO-d6) d 11.35 (s, 1H), 8.13 (d, J ¼ 5.3 Hz, 1H), 8.03 (d, J ¼ 8.5 was stirred further at same temperature until no further
53810 | RSC Adv., 2015, 5, 53809–53818
This journal is © The Royal Society of Chemistry 2015

View Article Online
Paper
RSC Advances
conversion was observed. The mixture was poured onto water, 2H), 6.84 (dd, J ¼ 8.7, 2.1 Hz, 1H), 6.65 (q, J ¼ 7.0 Hz, 1H), 6.4
extracted with ethyl acetate, and the combined organic layers (br, 1H), 3.62 (s, 3H), 2.91 (br, 3H), 2.01 (d, J ¼ 7.0 Hz, 3H); 13
C
were dried over Na2SO4, ltered and concentrated under NMR (125 MHz, DMSO-d6) d 160.0, 146.5, 141.2, 140.9, 138.7,
reduced pressure and the residue was puried via preparative 135.9, 129.2, 128.3, 127.8, 126.0, 123.1, 116.1, 112.9, 109.2, 96.9,
reversed phase chromatography using
NH4HCO3 solution and acetonitrile as eluents.
5
mM aqueous 55.6, 53.6, 24.7, 18.1;
[9]: 7-methoxy-1-methyl-9-[2-[4-(triuoromethyl)phenyl]ethyl]pyrido-
[5]: 9-benzyl-7-methoxy-1-methyl-pyrido[3,4-b]indole. Using 94 [3,4-b]indole. 106 mg harmine (0.50 mmol), 85 mg 40-tri-
mg benzyl bromide 110 mg 5 (73%) was obtained; 1H NMR (500 uoromethylphenyl acetylene (0.50 mmol) and 3 mg (0.05
MHz, DMSO-d6) d 8.18 (d, J ¼ 5.1 Hz, 1H), 8.15 (d, J ¼ 8.6 Hz, mmol) KOH were dissolved in 1 mL dry dimethyl sulfoxide. The
ꢃ
1H), 7.93 (d, J ¼ 5.1 Hz, 1H), 7.28 (t, J ¼ 7.4 Hz, 2H), 7.22 (t, J ¼ solution was stirred at 130 C until no further conversion was
7.4 Hz, 1H), 7.19 (d, J ¼ 2.1 Hz, 1H), 6.92 (d, J ¼ 7.4 Hz, 2H), 6.90 observed. Aer the completion of the reaction the product was
(dd, J ¼ 8.6, 2.1 Hz, 1H), 5.89 (s, 2H), 3.82 (s, 3H), 2.73 (s, 3H); puried via preparative reversed phase chromatography using 5
13C NMR (125 MHz, DMSO-d6) d 161.2, 143.8, 141.3, 139.5, mM aqueous NH4HCO3 solution and acetonitrile as eluents.
138.6, 135.4, 129.3, 129.0, 127.6, 125.8, 123.0, 114.7, 112.8, The obtained intermediate (117 mg) was dissolved in ethanol
109.9, 94.2, 56.1, 47.7, 23.2. HRMS calculated for C20H18N2O: and 10 w/w% Pd/C was added. The reaction was run until
302.1419; found 303.1487 (M + H).
completion in a pressure-boiler under 5 bar H2 at rt. The
[6]: 7-methoxy-9-[(3-methoxyphenyl)methyl]-1-methyl-pyrido- mixture was ltered and the ltrate evaporated to give 77 mg 9
[3,4-b]indole. Using 111 mg 3-methoxybenzyl bromide 107 mg 6 (40% for 2 steps) in sufficient purity.1H NMR (500 MHz, DMSO-
(64%) was obtained; 1H NMR (500 MHz, DMSO-d6) d 8.18 (d, J ¼ d6) d 8.19 (d, J ¼ 5.3 Hz, 1H), 8.09 (d, J ¼ 8.5 Hz, 1H), 7.93 (d, J ¼
5.2 Hz, 1H), 8.15 (d, J ¼ 8.6 Hz, 1H), 7.93 (d, J ¼ 5.2 Hz, 1H), 7.18 5.3 Hz, 1H), 7.59 (d, J ¼ 8.0 Hz, 2H), 7.43 (d, J ¼ 8.0 Hz, 2H), 7.04
(d, J ¼ 2.1 Hz, 1H), 7.18 (m, 1H), 6.90 (dd, J ¼ 8.6, 2.1 Hz, 1H), (d, J ¼ 2.1 Hz, 1H), 6.84 (dd, J ¼ 8.2, 2.1 Hz, 1H), 4.84 (t, J ¼ 7.4
6.80 (m, 1H), 6.51 (m, 1H), 6.41 (m, 1H), 5.85 (s, 2H), 3.83 (s, Hz, 2H), 3.84 (s, 3H), 3.17 (t, J ¼ 7.4 Hz, 2H), 2.97 (s, 3H); 13C
3H), 3.65 (s, 3H), 2.74 (s, 3H); 13C NMR (125 MHz, DMSO-d6) d NMR (125 MHz, DMSO-d6) d 161.1, 143.8, 143.3, 140.9, 137.8,
161.2, 160.0, 143.8, 141.3, 141.2, 138.6, 135.4, 130.5, 129.0, 134.9, 130.4, 129.3, 127.7, 125.6, 122.9, 114.5, 112.9, 110.1, 94.1,
123.0, 117.8, 114.6, 112.8, 112.4, 112.0, 109.9, 94.2, 56.1, 55.4, 55.9, 45.6, 36.3, 23.3; HRMS calculated for C22H19F3N2O:
47.6, 23.2; HRMS calculated for C21H20N2O2: 332.1525; found 384.1449; found 385.1520 (M + H).
333.1583 (M + H).
[7]: 7-methoxy-1-methyl-9-[[4-(triuoromethyl)phenyl]methyl]- Preparation of ligand and protein sample solutions
pyrido[3,4-b]indole. Using 131 mg 4-(triuoromethyl)benzyl
The 2 mM stock solutions of b-carbolines were prepared freshly
before each measurement in dioxane or in the EtOH/water
mixture. The volume of dioxane or EtOH added into sample
bromide 153 mg 7 (83%) was obtained; 1H NMR (500 MHz,
DMSO-d6) d 8.20 (d, J ¼ 5.2 Hz, 1H), 8.17 (d, J ¼ 8.6 Hz, 1H), 7.95
(d, J ¼ 5.2 Hz, 1H), 7.68 (d, J ¼ 8.1 Hz, 2H), 7.21 (d, J ¼ 2.1 Hz,
solutions never exceeded 5% (v/v) and caused negligible effects
1H), 7.16 (d, J ¼ 8.1 Hz, 2H), 6.92 (dd, J ¼ 8.6, 2.1 Hz, 1H), 6.01
on the CD or uorescence spectra. HSA (Sigma, 97%, essentially
(s, 2H), 3.82 (s, 3H), 2.70 (s, 3H); 13C NMR (125 MHz, DMSO-d6)
fatty acid-free) and AAG (Sigma) samples were dissolved in pH
d 161.3, 144.5, 143.7, 141.2, 138.8, 135.2, 129.2, 128.2, 126.7,
7.4 Ringer buffer solution (8.1 mM Na2HPO4$12H2O, 1.5 mM
126.3, 124.7, 123.1, 114.7, 112.9, 110.1, 94.1, 56.1, 47.4, 23.1;
KH2PO4, 137 mM NaCl, 2.7 mM KCl, 0.8 mM CaCl2, 1.1 mM
HRMS calculated for C21H17F3N2O: 370.1293; found 371.1362
MgCl2). Genetic variants of AAG were separated following the
(M + H).
´
chromatographic method of Herve et al. as described
[8]: 7-methoxy-1-methyl-9-[1-[4-(triuoromethyl)phenyl]ethyl]pyrido-
[3,4-b]indole. Using 139 mg rac-1-(1-bromoethyl)-4-(triuoro-
methyl)benzene 180 mg 8 was obtained. HRMS calculated for
previously.18
HSA binding test by affinity chromatography
C
22H19F3N2O: 384.1449; found 385.1513 (M + H). The racemate
Chromatographic experiments on HSA–Sepharose gel were
performed as described previously, using oxazepam acetate,
lorazepam acetate enantiomers and diazepam as reference
compounds.23 Elution volume of harmine and its derivatives
were detected at 280 nm. Control experiments were performed
on a gel containing no HSA. Compounds 2 and 7–9 showed
some non-specic adsorption.
was separated via chiral chromatography using Chiralcel OK
column and EtOH/heptane (30 : 70) eluent. 8a was isolated as
the rst eluting enantiomer.
8a. 85 mg (44%); retention time: 32 min; 1H NMR (500 MHz,
DMSO-d6) d 8.26 (d, J ¼ 5.2 Hz, 1H), 8.13 (d, J ¼ 8.7 Hz, 1H), 7.95
(d, J ¼ 5.2 Hz, 1H), 7.74 (d, J ¼ 8.2 Hz, 2H), 7.52 (d, J ¼ 8.2 Hz,
2H), 6.84 (dd, J ¼ 8.7, 2.1 Hz, 1H), 6.65 (q, J ¼ 7.0 Hz, 1H), 6.4
(br, 1H), 3.62 (s, 3H), 2.91 (br, 3H), 2.01 (d, J ¼ 7.0 Hz, 3H); 13
C
Fluorescence spectroscopic measurements
NMR (125 MHz, DMSO-d6) d 160.0, 146.5, 141.2, 140.9, 138.7,
135.9, 129.2, 128.3, 127.8, 126.0, 123.1, 116.1, 112.9, 109.2, 96.9,
55.6, 53.6, 24.7, 18.1;
Fluorescence measurements were carried out in a JASCO FP
8300 spectrouorimeter at 23 ꢄ 1 ꢃC, using a quartz cuvette with
1 cm optical path length; both excitation and emission band-
8b. 87 mg (45%); retention time: 42 min; 1H NMR (500 MHz, widths were set at 5 nm. The b-carbolines were excited at
DMSO-d6) d 8.26 (d, J ¼ 5.2 Hz, 1H), 8.13 (d, J ¼ 8.7 Hz, 1H), 7.95 320 nm. Intensities were corrected for the inner lter effect
(d, J ¼ 5.2 Hz, 1H), 7.74 (d, J ¼ 8.2 Hz, 2H), 7.52 (d, J ¼ 8.2 Hz, according to the absorbance of the added alkaloids at both the
This journal is © The Royal Society of Chemistry 2015
RSC Adv., 2015, 5, 53809–53818 | 53811

View Article Online
RSC Advances
Paper
excitation and the emission wavelengths. The association and distance-dependent dielectric functions were used in the
binding constants at 1 : 1 stoichiometry were calculated from calculation of the van der Waals and the electrostatic terms,
the uorescence emission increase of the protein bound alka- respectively. Docking simulations were performed using the
loids using the following equation (non-linear regression anal- Lamarckian genetic algorithm and the Solis & Wets local search
ysis with Microcal Origin ver. 8.6):
method. Each docking experiment was derived from 100 different
runs which were set to terminate aer a maximum of 2 500 000
energy evaluations. The population size was set to 150. During the
ꢀ
ꢁ
ꢀ
ꢁ
ꢀ
fp
fl
F ¼ ½Pꢅ ꢂ ½Lꢅ ꢂ ½Kꢅ
ꢁ
fpl
þ ½Lꢅ ꢂ ½Pꢅ ꢂ ½Kꢅ
ꢁ
þ ½Pꢅ
t
t
D
t
t
D
t
2
2
˚
search, a translational step of 0.2 A, and quaternions and torsion
ꢁ
ꢀ
ꢁ
þ ½Lꢅ þ ½Kꢅ
ꢁ
þ fp þ fl ꢂ fpl
t
D
steps of 5 were applied. The outputs of docking calculations were
rendered with PyMOL (The PyMOL Molecular Graphics System,
DeLano Scientic LLC, Palo Alto, CA, USA. http://www.pymol.org).
2
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ꢀ
ꢁ
2
½Pꢅ þ ½Lꢅ þ ½Kꢅ
t
t
D
ꢁ
ꢂ ½Pꢅ ꢁ ½Lꢅ
t
t
4
pKa prediction
where [P]t,[L]t are the total concentration of the protein and
ligand; fp, fl, fpl are the specic uorescence of the protein,
ligand and complex.24
Calculator Plugin of the soware MarvinSketch version 6.3.0
(2014, ChemAxon, http://www.chemaxon.com) was used to
estimate the pKa values of harmine and its synthetic derivatives.
Circular dichroism and UV absorption spectroscopy
measurements
Results and discussion
CD and UV absorption spectra were recorded on a JASCO J-715
spectropolarimeter at 25 ꢄ 0.2 ꢃC. Temperature control was
provided by a Peltier thermostat equipped with magnetic stirring.
Rectangular quartz cells of a 1 cm optical path length (Hellma,
USA) were used. Each spectrum represents the average of three
scans obtained by collecting data at a scan speed of 100 nm
minꢂ1. Absorption spectra were obtained by conversion of the
high voltage (HT) values of the photomultiplier tube of the CD
equipment into absorbance units. CD and absorption curves of
ligand–protein mixtures were corrected by subtracting the spectra
of ligand-free protein solutions. JASCO CD spectropolarimeters
record CD data as ellipticity (Q) in units of millidegrees (mdeg).
Details of the estimation of the association constant (Ka) and the
number of binding sites (n) using CD spectroscopic data have
been described elsewhere.25 Non-linear regression analysis of the
induced CD values measured at increasing [ligand]/[protein]
molar ratios was performed by Microcal Origin 8.6 Pro (Origin-
Lab Corporation, Northhampton, MA).
HSA binding studies by affinity chromatography
HSA interaction of harmine and its nine synthetic analogues were
evaluated from their retentions on a HSA–Sepharose column,
where the elution volume is characteristic of the overall binding
P
affinity ( nKa). Results showed that harmine binds weakly to
HSA (Ka ꢀ 2.5 ꢁ 104 Mꢂ1) and the methoxy-isopropyloxy group
replacement (compound 1) caused only a slight affinity increase
(Table 1). A similar result (Ka ꢀ 2.4 ꢁ 104 Mꢂ1) was found for the
HSA binding of harmane15 which lacks a methoxy substituent of
harmine. For harmine–bovine serum albumin interaction also
weak binding was reported (Ka ꢀ 2.0 ꢁ 104 Mꢂ1).27 Introducing
aromatic substituents in position 7, however, resulted in consid-
erably enhanced albumin binding affinities. Compound 2 showed
about ten times stronger association compared to harmine,
without signicant enantioselectivity. In case of compounds 3
and 4 having a pyridine ring in their R7 substituent, the Ka values
are only about four times higher than that of harmine. Aromatic
substitution on the pyrrole nitrogen provoked even stronger
enhancement, association constant values of compounds 5–9 are
about 4.0 ꢁ 105 Mꢂ1. The enantiomers of compound 8 showed no
stereospecic binding, either. The enhanced HSA affinity of the
more hydrophobic harmine derivatives may be benecial since
the overall solubility of the drugs in plasma will increase.
Furthermore, tight albumin association may improve the chem-
ical as well as the metabolic stability of the alkaloids28 and can
facilitate their uptake in growing tumor tissues where albumin
molecules are increasingly taken up by cancer cells.29 However,
the efficacy of drugs can be compromised by extensive HSA
binding since only the unbound fraction exhibits pharmacologic
effects. The highest HSA affinity constants obtained in this study
predict no serious limitation for the pharmacological activity of
the harmine derivates that is also impacted by other factors such
as tissue binding, stability, and clearance.30,31
Molecular docking calculations
All calculations were performed using DockingServer.26 PM6
semi-empirical method (MOPAC2009) was used for energy
minimization and partial charges calculation to the neutral form
of compound 4. X-ray structure of the A variant of AAG (PDB code
3APU) was selected. All water molecules were removed from the
protein coordinates prior to docking calculations. Hydrogen
atoms were added to the PDB structure using AutoDockTools.
The total charge of AAG and partial charges of the atoms were
calculated by the Mozyme function of MOPAC2009 soware. The
calculated partial charges were applied for further calculations.
Sequential ligand docking was carried out, initial position and
orientation of compound 4 were set randomly. The rst molecule
was docked using a pre-evaluated interaction grid based on
interactions with atoms in the protein alone. The second ligand
was then docked employing an interaction grid including inter-
Protein binding studies by uorescence spectroscopy
actions with the protein and bound congurations of the rst
˚
ligand. Affinity (grid) maps of 25 ꢁ 25 ꢁ 25 A grid points were Harmine and its derivatives, like other b-carbolines, are highly
generated using the Autogrid program. AutoDock parameter set- uorescent compounds. Fluorescence properties of the cationic
53812 | RSC Adv., 2015, 5, 53809–53818
This journal is © The Royal Society of Chemistry 2015

View Article Online
Paper
RSC Advances
and neutral forms are different, and this phenomenon can be experimentally determined pKa of harmine is 7.45.32 The use of
utilized in protein binding studies, as shown previously for the this value in the Henderson–Hasselbalch equation at pH 7.4
HSA association of harmane and norharmane.15,28 The cationic gives rise to 53 and 47 per cent for the neutral and cationic form
species uoresces strongly with a peak maximum centred of harmine, respectively. Since the protonation constants of the
ꢀ430 nm while the blue shied emission of the neutral mole- synthetic derivatives are unknown, calculated pKa values are
cules is much weaker centered around 350 nm. The shown in the ESI.† As expected, the substitutions made in the 7
Table 1 HSA and AAG association constants of harmine and its synthetic derivatives estimated by affinity chromatography (ac), fluorescence (fl),
and circular dichroism (cd) spectroscopy methods. n.d.: not determined
Substituent
Ka (Mꢂ1
)
Compound
Harmine
R7
R9
H
HSA (ac)
AAG ()
Me
2.5 ꢁ 104
3.3 ꢁ 104
1
H
3.5 ꢁ 104
n.d.
2aa
H
2.2 ꢁ 105
1.3 ꢁ 105
2ba
H
2.7 ꢁ 105
1.3 ꢁ 105
3.3 ꢁ 105
3
H
H
9.0 ꢁ 104, 8.4 ꢁ 104 ()
4.1 ꢁ 105 (F1/S) (cd),
3.4 ꢁ 105 (A) (cd)
4
5
9.0 ꢁ 104, 9.4 ꢁ 104 ()
4.0 ꢁ 105, 7.2 ꢁ 105 ()
Me
Me
1.6 ꢁ 106
8.0 ꢁ 105
6
4.3 ꢁ 105, 4.0 ꢁ 105 ()
7
Me
Me
Me
4.0 ꢁ 105, 3.1 ꢁ 105 ()
3.9 ꢁ 105, 2.1 ꢁ 105 ()
3.5 ꢁ 105
4.1 ꢁ 106 (F1/S), 3.9 ꢁ 105 (A)
8aa
8ba
1.2 ꢁ 106
7.0 ꢁ 105
9
Me
3.2 ꢁ 105, 1.1 ꢁ 105 ()
2.8 ꢁ 106
a
The assignment of the chirality is arbitrary and it is only used to denote the stereochemical relationship of compounds 2a–2b and 8a–8b.
This journal is © The Royal Society of Chemistry 2015
RSC Adv., 2015, 5, 53809–53818 | 53813

View Article Online
RSC Advances
Paper
and 9 positions do not alter signicantly the basicity of the
pyridine nitrogen suggesting very similar neutral–cationic
equilibrium for the derivatives. The uorescence approach was
employed to investigate the AAG binding properties of harmine
and its derivatives. Fig. 1 shows the change of the uorescence
emission of harmine upon addition of AAG. The ligand solution
at pH 7.4 displays a single, unstructured band at 421.5 nm,
assigned to the emission of the cationic form due to the well-
studied excited state pKa shi of b-carbolines.33 Increasing
concentration of AAG induces intensity decrease with simulta-
neous development of a new band with a maximum around 349
nm, indicating the bound neutral form of harmine. An iso-
emission point appears around 379 nm suggesting the co-
existence of the bound and free species at equilibrium. Based
on the emission data at 349 nm, a value of Ka ꢀ 3.3 ꢁ 104 Mꢂ1
could be calculated, which is close to that measured for HSA
binding of harmine (Table 1).
Fig. 2 Fluorescence emission spectral changes of compound 7 as a
function of HSA and AAG concentration (2 mL of 2 mM ligand solution
was titrated with aliquots of 80 mM HSA and AAG dissolved in the ligand
stock solution).
Similar experiments, comparing the binding of compound 7
to AAG and HSA can be seen in Fig. 2. Analysis of the low
intensity emission data of the HSA bound neutral species at 366
nm yielded a Ka value of 3.1 ꢁ 105 Mꢂ1 for HSA, which is in good
agreement with the affinity constant obtained on the HSA
column. The uorescence changes measured with AAG,
however, prove about eight times stronger AAG binding (Ka ꢀ
2.4 ꢁ 106 Mꢂ1). It is to be noted that the satisfactory concor-
dance between the HSA affinity constants of additional harmine
derivatives estimated by affinity chromatography method as
well as from uorescence emission of the protein bound neutral
species (Table 1) refers to the decisive contribution of the
neutral form in the binding process.
Due to genetic polymorphism, AAG can be isolated as a
70 : 30 mixture of two genetic variants, called F1/S and A, having
different drug binding properties. The binding of compounds 7
and 4 to the separated F1/S and A variants was examined, too. In
case of 7 we observed (Fig. 3) that addition of the F1/S variant
form to the ligand solution gave rise to more pronounced uo-
rescence changes than the same amount of the A variant,
corresponding to ten times higher affinity on the F1/S variant
compared to the A (Table 1). By monitoring the uorescence
spectral changes of 4 upon addition of the genetic variants, the
descending emission of the peak at 422 nm could be observed,
however, the short-wavelength band corresponding to the neutral
form evolved only in presence of the F1/S variant (Fig. 3). There-
fore, in case of the A form the uorescence method was not
adequate to calculate the association constant, but it demonstrates
different affinity of this ligand for the genetic variants of AAG.
AAG affinity data derived for harmine and its derivatives are
summarized in Table 1. Similarly to HSA, the AAG binding of
harmine is also weak. Substitution in position 9 with aralkyl
groups, besides enhancing their cytotoxic activity in general, also
increased the interaction with plasma proteins. The replacement
of the methoxy group by alkoxy and (het)aryl alkoxy substituents
in position 7 has a similar effect. Compared to harmine, this
modication resulted in a four to ten times higher AAG affinity
(Table 1). It is to be noted, that the HSA interaction of
compounds possessing a pyridine moiety was weaker. The
conguration of the stereocenter of 2 or 8 had negligible effect
on protein binding. The AAG affinity of compounds 5–9 bearing
a benzyl-type substituent on the indole nitrogen (R9) was high
(Ka ꢀ 106 Mꢂ1). The substitution of the phenyl ring had only
minor inuence on the binding. The highest affinities were
measured for the 4-triuoromethyl substituted analogues, while
the 3-methoxy group reduced the binding.
Circular dichroism and UV absorption spectroscopic
investigation on the binding of derivatives 4 and 7 to AAG
HSA binding of the optically inactive b-carboline derivatives
studied in this work induced no or poor quality CD bands in the
respective light absorption region of the ligands (data not
shown). Therefore, the induced CD spectroscopic approach15,34,35
could not be employed for estimation of the HSA association
constants. In contrast to this, entrapment of some alkaloids
within the central drug binding pocket of AAG induced salient
CD activity. Addition of the 7-pyridylethoxy derivative of harmine
Fig. 1 Emission spectra of harmine measured in AAG solution at
increasing protein/ligand molar ratios (2 mL of harmine solution was
titrated with 200 mM AAG dissolved in the ligand solution). Fluores-
cence intensities plotted against the AAG concentration of the sample
solution. The solid line is the result of non-linear fitting analysis.
53814 | RSC Adv., 2015, 5, 53809–53818
This journal is © The Royal Society of Chemistry 2015

View Article Online
Paper
RSC Advances
Fig. 3 Comparison of the emission spectral changes of compounds 7 and 4 in the presence of the F1/S and a genetic variant of AAG (2 mL of
ligand solutions were titrated with 80 mM AAG variants dissolved in the ligand stock solution).
(4) to the A genetic variant of AAG induced multiple CD signals A variant of AAG shows the relative dominance of the neutral
between 240 and 400 nm (Fig. 4). Two weaker, opposite signals species but binding in the charged form also occurs (Fig. 5).
can be seen at 303 and 360 nm which are allied to the broad Binding preference for the unprotonated ligand is more
absorption band displayed above 280 nm. In the short- pronounced in the presence of the F1/S variant, where the UV
wavelength region a more intense, negative–positive CD band prole of the bound ligand is very similar to that observed in
pair was resolved with a zero cross-over point around 253 nm. alkaline solution (Fig. 5). The weaker hypochromism and the
Magnitudes of these induced CD peaks increased proportionally less intense induced CD bands (data not shown) suggest that
with the increasing concentration of the ligand but neither their the F1/S protein also binds two alkaloid molecules but with a
shape nor their position showed variations during the titration. larger intermolecular distance between them.
Comparison of the UV spectra of the free and AAG bound form of
Taking into consideration these data, the very low uores-
4 recorded at the same concentration indicates large intensity cence emission of the neutral species of compound 4 obtained
reduction of the absorption values (green and black curves in with the A variant (Fig. 3) might in part be attributed to its
Fig. 5). This phenomenon, called hypochromism, suggests the mixed, neutral–cationic binding and also to a dimerization
p–p stacking of two carboline rings at the binding site of AAG. caused self-quenching mechanism.37
Due to the asymmetry of the protein environment, the stacked,
Interestingly, the AAG binding behaviour of 4 is highly
planar aromatic moieties cannot be arranged in a completely reminiscent to that of acridine orange, the association of which
parallel fashion but their long axis are rotated relative to each to the A variant also occurs in a dimeric form inducing a
other, forming a helical array. This situation results in a chiral characteristic biphasic CD pattern.19 Similarly to compound 4,
intermolecular exciton coupling between the p–p* transition F1/S binding of the acridine orange results in much weaker
moments of the proximal carboline chromophores generating exciton signals. Comparison of the crystal structures of the
two intense, opposite CD bands associated to the respective UV genetic variants revealed that the central drug binding pocket is
absorption peak (Fig. 4). As it follows from the exciton chirality narrower in the A form.38 Instead of the three sub-chambers
rule,36 the negative–positive order of these signals predicts the present in F1/S, the A variant possesses only two binding
M-helicity of the stacked dimer accommodated within the central lobes. It follows that due to steric restrictions, co-binding of two
cavity of AAG. The less intense ellipticity bands of 4 measured molecules of compound 4 can occur only in a tightly stacked,
above 290 nm may also be of excitonic origin. The dimeric dimeric conguration while the larger cle of the F1/S variant
binding mode of 4 inferred from the spectroscopic results is allows greater room and separation of the carboline moieties.
supported by the analysis of the CD titration data. The best t on The pyridine side chain seems to be important for the dimeric
the experimental values was obtained using 2 : 1 ligand : protein binding since the harmine–AAG interaction did not produce the
binding stoichiometry. The estimated affinity constants refer to above spectroscopic features (hypochromism, exciton CD
tight interaction of 4 with both genetic variants of AAG (Table 1). coupling).
It is to be noted that apart from the hypochromism, the
As the uorescence measurements showed, compound 7
overall UV absorption prole of compound 4 measured with the bearing a triuorophenyl substituent attached to the indole
This journal is © The Royal Society of Chemistry 2015
RSC Adv., 2015, 5, 53809–53818 | 53815

View Article Online
RSC Advances
Paper
provoked limited changes only suggesting that its cationic–
neutral discrimination ability is much weaker than that of the
F1/S form. Due to the more uniform, tighter association of
compound 7 to the F1/S variant, higher intensity ellipticity
signals were induced than that measured with the A form
(Fig. 6). Distinctly from 4, the bulky triuoromethyl-phenyl
moiety sterically restricts the dimeric binding of 7. Similarly
to various AAG ligands having a rigid, planar ring system, the
observed CD activity of 7 may be induced through non-
degenerate exciton coupling with high-energy p–p* transi-
tions of adjacent aromatic side chains of the binding
pocket.39,40
Sequential molecular docking of compound 4 to the A variant
of AAG
Computational docking calculation was also performed to
obtain insight into the molecular details of the AAG associa-
tion of compound 4. Taking into account the CD and UV
spectroscopic results, two molecules of compound 4 were
docked sequentially into the binding room of AAG. The most
frequent and best energy result shows a pair of partially
p-stacked ligand molecules in the binding crevice stabilized
by four intermolecular H-bonds (Fig. 7). The interplanar
˚
distance between the b-carboline skeletons is about 4 A and
their long axes close a counterclockwise, le-handed angle.
Fig. 4 Difference CD and UV absorption spectra of compound 4
added to 25 mM buffer solution of the a genetic variant of AAG.
Fig. 5 Comparison of the UV absorption spectra of compound 4
measured in protein-free solutions at different pH values and in the
presence of the A and F1/S genetic variant of AAG (Ringer buffer).
nitrogen also exhibits selective AAG binding in favour of the
F1/S variant (Fig. 3). This behaviour is reected in the distinct
CD and absorption spectroscopic changes obtained with the
genetic variants, too (Fig. 6). The F1/S form binds preferen-
tially the neutral species of the molecule which is attested by
the great similarity between the UV curves recorded in 0.1 M
NaOH and in pH 7.4 buffer solution of the protein. In relation
Fig. 6 Difference CD and UV absorption spectra of compound 7
measured with the F1/S and A genetic variant of AAG. UV curves of the
ligand recorded in protein-free solutions at different pH values are also
to the absorption spectrum measured in buffer, the A variant shown.
53816 | RSC Adv., 2015, 5, 53809–53818
This journal is © The Royal Society of Chemistry 2015

View Article Online
Paper
RSC Advances
Acknowledgements
¨
Financial support of the ‘Lendulet’ Program of the Hungarian
Academy of Sciences (LP2013-55/2013) is acknowledged.
References
1 R. Cao, W. Peng, Z. Wang and A. Xu, Curr. Med. Chem., 2007,
14, 479–500.
2 R. Cao, W. Peng, H. Chen, Y. Ma, X. Liu, X. Hou, H. Guan and
A. Xu, Biochem. Biophys. Res. Commun., 2005, 338, 1557–1563.
3 J. Bain, L. Plater, M. Elliott, N. Shpiro, C. J. Hastie,
H. McLauchlan, I. Klevernic, J. S. Arthur, D. R. Alessi and
P. Cohen, Biochem. J., 2007, 408, 297–315.
4 A. Miralles, S. Esteban, A. Sastre-Coll, D. Moranta,
V. J. Asensio and J. A. Garcia-Sevilla, Eur. J. Pharmacol.,
2005, 518, 234–242.
5 D. Frost, B. Meechoovet, T. Wang, S. Gately, M. Giorgetti,
I. Shcherbakova and T. Dunckley, PLoS One, 2011, 6, e19264.
6 F. K. Wiseman, K. A. Alford, V. L. Tybulewicz and
E. M. Fisher, Hum. Mol. Genet., 2009, 18, R75–R83.
7 J. Ishida, H. K. Wang, K. F. Bastow, C. Q. Hu and K. H. Lee,
Bioorg. Med. Chem. Lett., 1999, 9, 3319–3324.
Fig. 7 Two compound 4 molecules docked sequentially into the
binding cavity of the A variant of AAG. Dotted lines indicate intermo-
lecular H-bonds.
8 R. Cao, Q. Chen, X. Hou, H. Chen, H. Guan, Y. Ma, W. Peng
and A. Xu, Bioorg. Med. Chem., 2004, 12, 4613–4623.
9 R. Cao, W. Fan, L. Guo, Q. Ma, G. Zhang, J. Li, X. Chen, Z. Ren
and L. Qiu, Eur. J. Med. Chem., 2013, 60, 135–143.
10 R. Calvo, J. C. Lukas, M. Rodriguez, N. Leal and E. Suarez,
Curr. Pharm. Des., 2006, 12, 977–987.
11 A. Das and G. S. Kumar, RSC Adv., 2014, 4, 33082–33090.
12 M. Hossain, A. Y. Khan and G. S. Kumar, J. Chem.
Thermodyn., 2012, 47, 90–99.
This binding geometry is favourable for dipole–dipole
coupling between the p–p* transitions of the b-carboline
chromophores, which is consistent with the induced CD
exciton couplet and the UV hypochromism of the AAG bound
species (Fig. 4). Considering the exciton chirality rule,36 the
le-handed (M-helical) orientation of the ligand molecules
predicts
a longer-wavelength negative and a shorter-
wavelength positive CD band pair which is in full accor-
dance with the spectroscopic data.
13 F. Zsila, Mol. Pharm., 2013, 10, 1668–1682.
14 A. Varshney, P. Sen, E. Ahmad, M. Rehan, N. Subbarao and
R. H. Khan, Chirality, 2010, 22, 77–87.
15 C. Domonkos, I. Fitos, J. Visy and F. Zsila, Mol. Pharm., 2013,
10, 4706–4716.
Conclusions
In the present study, the synthesis of a series of novel harmine 16 Z. H. Israili and P. G. Dayton, Drug Metab. Rev., 2001, 33,
derivatives bearing an arylated alkoxy substituent in position 161–235.
7, or a modied benzyl moiety in position 9 was described. 17 M. Otagiri, Drug Metab. Pharmacokinet., 2005, 20, 309–323.
´
Plasma protein binding of these b-carboline analogues was 18 F. Herve, G. Caron, J. C. Duche, P. Gaillard, N. Abd Rahman,
found to be substituent-dependent. Incorporation of the
aromatic pharmacophore group into the positions 7 or 9 of the
carboline skeleton brought about considerably enhanced
A. Tsantili-Kakoulidou, P. A. Carrupt, P. d'Athis,
J. P. Tillement and B. Testa, Mol. Pharmacol., 1998, 54,
129–138.
´
´
affinity to both HSA and AAG. These results imply that the 19 I. Fitos, J. Visy, F. Zsila, Z. Bikadi, G. Mady and M. Simonyi,
pharmacologically active, free serum levels of these potential Biochem. Pharmacol., 2004, 67, 679–688.
therapeutic agents are supposed to be very low that should be 20 K. Nishi, M. Ueno, Y. Murakami, N. Fukunaga, T. Akuta,
taken into consideration during further drug development
phases. In contrast to the well documented cationic drug
D. Kadowaki, H. Watanabe, A. Suenaga, T. Maruyama and
M. Otagiri, J. Pharm. Sci., 2009, 98, 4316–4326.
´
binding preference of AAG, the spectroscopic results proved 21 I. Fitos, J. Visy, M. Simonyi, G. Mady and F. Zsila, Biochim.
that the AAG association of b-carbolines possessing basic Biophys. Acta, 2010, 1800, 367–372.
pyridine nitrogen takes place mainly in non-protonated form. 22 P. Jandera, Anal. Chim. Acta, 2011, 692, 1–25.
Moreover, the 7-pyridylethoxy derivative of harmine exhibited 23 I. Fitos, Z. Tegyey, M. Simonyi, I. Sjoholm, T. Larsson and
p-stacked, dimeric binding to the A but not to the F1/S variant
C. Lagercrantz, Biochem. Pharmacol., 1986, 35, 263–269.
of AAG that is related to the distinct architecture of the 24 D. A. Breustedt, D. L. Schonfeld and A. Skerra, Biochim.
internal cavity of their b-barrels.
Biophys. Acta, 2006, 1764, 161–173.
This journal is © The Royal Society of Chemistry 2015
RSC Adv., 2015, 5, 53809–53818 | 53817

View Article Online
RSC Advances
Paper
´
25 F. Zsila, Z. Bikadi and M. Simonyi, Biochem. Pharmacol., 33 A. Mallick, P. Das and N. Chattopadhyay, J. Photochem.
2003, 65, 447–456.
Photobiol., C, 2010, 11, 62–72.
26 Z. Bikadi and E. Hazai, J. Cheminf., 2009, 1, 15.
27 A. P. Shohreh Nasi and G. B. Sadeghi, J. Lumin., 2012, 132,
2361–2366.
28 H. Greige-Gerges, Y. Diab, J. Farah, J. Magdalou, C. Haddad
and N. Ouaini, Biopharm. Drug Dispos., 2008, 29, 83–89.
29 A. M. Merlot, S. Sahni, D. J. Lane, A. M. Fordham, N. Pantarat,
34 D. Tedesco and C. Bertucci, J. Pharm. Biomed. Anal., 2015,
DOI: 10.1016/j.jpba.2015.02.024.
35 C. Domonkos, I. Fitos, J. Visy and F. Zsila, Phys. Chem. Chem.
Phys., 2014, 16, 22632–22642.
36 N. Berova, L. Di Bari and G. Pescitelli, Chem. Soc. Rev., 2007,
36, 914–931.
D. E. Hibbs, V. Richardson, M. R. Doddareddy, J. A. Ong, 37 J. P. Knemeyer and N. Marme, Recent Pat. DNA Gene
M. L. Huang, D. R. Richardson and D. S. Kalinowski,
OncoTargets Ther., 2015, 6, 10374–10398.
30 A. Sparreboom, K. Nooter, W. J. Loos and J. Verweij, Neth. J.
Med., 2001, 59, 196–207.
31 J. Heuberger, S. Schmidt and H. Derendorf, J. Pharm. Sci.,
2013, 102, 3458–3467.
32 M. Balon, J. Hidalgo, P. Guardado, M. A. Munoz and
C. Carmona, J. Chem. Soc., Perkin Trans. 2, 1993, 99–104.
Sequences, 2007, 1, 145–147.
38 K. Nishi, T. Ono, T. Nakamura, N. Fukunaga, M. Izumi,
H. Watanabe, A. Suenaga, T. Maruyama, Y. Yamagata,
S. Curry and M. Otagiri, J. Biol. Chem., 2011, 286, 14427–
14434.
39 F. Zsila and Y. Iwao, Biochim. Biophys. Acta, 2007, 1770, 797–
809.
40 F. Zsila, J. Mol. Recognit., 2011, 24, 995–1006.
53818 | RSC Adv., 2015, 5, 53809–53818
This journal is © The Royal Society of Chemistry 2015