Relative role of halogen bonds and hydrophobic interactions in inhibition of human protein kinase CK2α by tetrabromobenzotriazole and some C(5)-substituted analogues

Article abstract of DOI:10.1021/jp102848y

To examine the relative role of halogen bonding and hydrophobic interactions in the inhibition of human CK2α by 4,5,6,7- tetrabromobenzotriazole (TBBt), we have synthesized a series of 5-substituted benzotriazoles (Bt) and the corresponding 5-substituted 4,6,7- tribromobenzotriazoles (Br₃Bt) and examined their inhibition of human CK2α relative to that of TBBt. The various C(5) substituents differ in size (H and CH₃), electronegativity (NH₂ and NO ₂), and hydrophobicity (COOH and Cl). Some substituents were halogen bond donors (Cl, Br), while others were fluorine bond donors (F and CF ₃). Most of the 5-substituted analogues of Br₃Bt (with the exception of COOH and NH₂) exhibited inhibitory activity comparable to that of TBBt, whereas the 5-substituted analogues of the parent Bt were only weakly active (Br, Cl, NO₂, CF₃) or inactive. The observed effect of the volume of a ligand molecule pointed to its predominant role in inhibitory activity, indicating that presumed halogen bonding, identified in crystal structures and by molecular modeling, is dominated by hydrophobic interactions. Extended QSAR analysis additionally pointed to the monoanion and a preference for the N(1)-H protomer of the neutral ligand as parameters crucial for prediction of inhibitory activity. This suggests that the monoanions of TBBt and its congeners are the active forms that efficiently bind to CK2α, and the binding affinity is coupled with protomeric equilibrium of the neutral ligand.

Full text of DOI:10.1021/jp102848y

J. Phys. Chem. B 2010, 114, 10601–10611
10601
Relative Role of Halogen Bonds and Hydrophobic Interactions in Inhibition of Human
Protein Kinase CK2r by Tetrabromobenzotriazole and Some C(5)-Substituted Analogues
Romualda Wa˛sik,† Maja Łebska,† Krzysztof Felczak,†,§ Jarosław Poznan´ski,† and
David Shugar*,†,‡
Institute of Biochemistry and Biophysics, Polish Academy of Sciences, Pawin´skiego 5a, 02-106 Warszawa,
Poland, and DiVision of Biophysics, Institute of Experimental Physics, UniVersity of Warsaw,
˙
93 Zwirki i Wigury St., 02-089 Warszawa, Poland
ReceiVed: March 30, 2010; ReVised Manuscript ReceiVed: June 28, 2010
To examine the relative role of halogen bonding and hydrophobic interactions in the inhibition of human
CK2R by 4,5,6,7-tetrabromobenzotriazole (TBBt), we have synthesized a series of 5-substituted benzotriazoles
(Bt) and the corresponding 5-substituted 4,6,7-tribromobenzotriazoles (Br3Bt) and examined their inhibition
of human CK2R relative to that of TBBt. The various C(5) substituents differ in size (H and CH3),
electronegativity (NH2 and NO2), and hydrophobicity (COOH and Cl). Some substituents were halogen bond
donors (Cl, Br), while others were fluorine bond donors (F and CF3). Most of the 5-substituted analogues of
Br3Bt (with the exception of COOH and NH2) exhibited inhibitory activity comparable to that of TBBt,
whereas the 5-substituted analogues of the parent Bt were only weakly active (Br, Cl, NO2, CF3) or inactive.
The observed effect of the volume of a ligand molecule pointed to its predominant role in inhibitory activity,
indicating that presumed halogen bonding, identified in crystal structures and by molecular modeling, is
dominated by hydrophobic interactions. Extended QSAR analysis additionally pointed to the monoanion and
a preference for the N(1)-H protomer of the neutral ligand as parameters crucial for prediction of inhibitory
activity. This suggests that the monoanions of TBBt and its congeners are the active forms that efficiently
bind to CK2R, and the binding affinity is coupled with protomeric equilibrium of the neutral ligand.
Introduction
off-targets for TBBz and also showed that TBBt is more se-
lective toward CK2R.5
Protein kinase CK2 (previously known as casein kinase 2,
albeit casein is not an in vivo substrate), the most pleiotropic
of known protein kinases, is a Ser/Thr kinase, also known to
phosphorylate Tyr residues. It is constitutively active, but
attempts to clarify its mode of regulation remain ongoing.1 It
plays a key role as an antiapoptopic factor and is, consequently,
a highly promising drug target, particularly since its level is
markedly elevated in all human cancers and experimental
tumors.1,2 This underlines the importance of development of
permeable low molecular weight selective inhibitors of this
enzyme, as well as its intrinsically catalytically active subunits
CK2R and CK2R′.2
The first reported promising low molecular weight inhibitor,
4,5,6,7-tetrabromotriazole, now known as TBBt (or TBB), and
early on shown to be cell-permeable,3 was shortly thereafter
followed by the finding that the corresponding 4,5,6,7-tetra-
bromo-1H-benzimidazole (TBBz) is a comparably good inhibi-
tor of CK2.4 Although many more potent inhibitors of CK2 have
since been reported,5 TBBt and TBBz continue to be widely
applied.
A characteristic feature of the crystal structures of complexes
of CK2R (and one other kinase; see below) with TBBt and some
of its analogues substituted on the triazole ring is the presence
of what are now widely known as halogen bonds, that is,
noncovalent electrostatic interactions between a classical hy-
drogen bond acceptor (electronegative O, N, or S) and a
polarizable, partially electropositive halogen atom (Cl, Br, I).
Statistical analysis of crystal structures of many proteins
complexed with halogenated ligands7 demonstrated that such
halogen bonds exhibit a geometry similar to that of hydrogen
bonds,includingadirectionalpreference,butwithadonor-acceptor
distance significantly shorter than the sum of their van der Waals
radii. Fluorine, because of its high electronegativity, is a very
poor halogen bond acceptor but is a good acceptor from standard
hydrogen bond donors.8
However, despite the clear evidence of short halogen bonding
contacts in protein structures, their energetic contribution is still
under active debate.9 Thus, analysis of the effect of a halogen
bond on the isomerization of a four-stranded DNA junction led
to estimation of the halogen bond contribution as 2-5 kcal/
mol stronger than a hydrogen bond in an analogous environ-
ment.10 However, it should be noted that the energy contribution
was calculated on the assumption that the observed conforma-
tional switch is not accompanied by a change in DNA hydration
entropy.
Screening of TBBt against a small panel of over 30 different
kinases demonstrated that it is reasonably selective for CK2.6
Subsequent screening of a number of analogues of TBBt and
TBBz versus a larger panel of 80 kinases revealed three major
From a design perspective, it would appear that one may
substitute a halogen bond for a hydrogen bond in the active
site of an enzyme to increase both affinity and/or specificity. A
recent example of such an approach, involving the enzyme
ketosteroid isomerase, was based on replacement of the tyrosine
* Corresponding author: Phone: +48 22 592 3511. Fax: +48 22 592
2190. E-mail: shugar@ibb.waw.pl.
† Polish Academy of Sciences.
‡ University of Warsaw.
§ Present address: Center for Drug Design, University of Minnesota,
Minneapolis, Minnesota 55455.
10.1021/jp102848y  2010 American Chemical Society
Published on Web 07/23/2010

10602 J. Phys. Chem. B, Vol. 114, No. 32, 2010
Wa˛sik et al.
SCHEME 1
hydrogen bond donor in the oxyanion hole of the enzyme by
semisynthetic enzymes containg para-halogenated phenylalanine
derivatives in place of the tyrosine hydrogen bond donor. All
of the halogenated enzymes exhibited activity comparable to
that of the phenylalanine mutant, leading to the conclusion that
a halogen bond could not, in this system, functionally replace
a hydrogen bond in its active site.9
Furthermore, analysis of structures of T4 lysozyme binding
various halogenated benzene derivatives demonstrated that (with
the exception of an iodine substituent) pentafluorobenzene and
its bromo and chloro congeners display heterogeneity of binding
sites, strongly suggesting predominance of van der Waals
interactions rather than halogen bonding per se. Moreover,
comparison of association constants led to a rough estimate of
halogen bond strength as 0.5-0.7 kcal/mol, hence much less
than that for a canonical hydrogen bond.11
SCHEME 2
The most important factor enabling appropriate estimation
of halogen bond strength is the decomposition of the observed
change in free energy to the accompanying changes in entropy
and enthalpy. As demonstrated for halogenated anesthetic
compounds either by van’t Hoff analysis of the temperature
dependence of binding halothane to BSA12 or by calorimetric
studies on binding of halothane and isoflurane to the four-helix
bundle motif of Ferritin,13 both entropic and enthalpic effects
contribute comparably to binding. Similarly, binding of chlo-
rohydroxyphenyl acetal by CprK protein, monitored by isother-
mal titration calorimetry, was reported to be largely entropically
driven.14 A systematic study of binding of rationally designed
halogenated peptides by a small modular domain of a mam-
malian neuronal protein demonstrated that, for all compounds
studied (with the exception of fluorinated peptides), contributions
of enthalpic and entropic (e.g., hydrophobic) effects were
virtually equal.15
Results and Discussion
Synthetic Procedures. The starting compound, 4-Br-1,2-
phenylenediamine (1b, see Scheme 1) was obtained in good
yield, according to a previously published procedure,21 by
reduction of 4-Br-2-nitroaniline22 with SnCl2. The monosubsti-
tuted 5-F- (2a), 5-Br- (2b), 5-CF3- (2c), and 5-NO2- (2d)
benzotriazoles (Bt) were prepared from the appropriate o-
phenylenediamines by nitrous-acid-promoted cyclization23 (see
Scheme 1), while 5-NH2-benzotriazole (2i) was obtained in 60%
yield by HI reduction24 of the parent 5-nitro analogue (2d) or
with the use of SnCl2 in HCl (see Scheme 2).
To our knowledge, there are no reports on direct bromination
of monosubstituted benzotriazoles, with the exception of 5-Cl-
Bt (2f) to 5-Cl-Br3Bt (3f).23 The 5-F- and 5-CF3-, 5-CH3-, 4,6,7-
tribromobenzotriazoles (Br3Bt) (3a, 3c, 3e) were prepared from
the 5-monosubstituted benzotriazoles by direct bromination in
concentrated nitric acid.23 However, direct bromination of
5-COOH- and 5-NO2- benzotriazoles (2g, 2d) led to a complex
mixture of dibrominated products and a low yield of the required
tribromo analogues.
To circumvent difficult separations and low yields of the
desired tribromo analogues, we have developed new straight-
forward procedures leading to only one product. Oxidation of
5-CH3-Br3Bt (3e) with the use of KMnO4 led to 5-COOH-Br3Bt
(3g) in 72% yield and high purity (see Scheme 3). Then 5-NO2-
Br3Bt (3d) was obtained in 70% yield by nitration of Br3Bt (3h),
prepared by decarboxylation in quinoline of the parent 5-COOH-
Br3Bt in 68% yield.
In attempts to synthesize 5-NH2-Br3Bt (3i), reduction of
5-NO2-Br3Bt (3d) by SnCl2 led to partial debromination, with
dibromo isomers as main products. This drew our attention to
a recent report on the use of HI for selective reduction of
aromatic nitro compounds to amines and claimed to proceed
with excellent chemoselectivity, without affecting other func-
tional groups, including halides.24 In our hands, its application
to reduction of 5-NO2-Br3Bt (3d) led to a mixture of monobromo
and bromo-iodo analogues of 5-NH2-Bt (2i), as determined by
high-resolution mass spectroscopy (see Scheme 3). We have,
however, successfully applied this procedure to preparation of
5-NH2-Bt (2i) from its 5-nitro congener (2d). The foregoing
difficulties also prompted us to attempt direct reduction of
5-NO2-Br3Bt (3d) with use of Fe powder in acetic acid25 and
resulted in the desired 5-NH2 congener (3i, see Scheme 3) in
74% yield.
With a view to better understanding the selectivity, affinity,
and mode of binding of TBBt to CK2, we have synthesized a
series of analogues in which one of the Br atoms is replaced by
other substituents and examined the activities of these ligands
as inhibitors relative to that of TBBt. In some respects, this
approach is similar to that adopted by Kraut et al.9 (see above).
The bromine atom selected for replacement is that at C(5), which
was found to be involved in halogen bonding in the crystal
structure of TBBt bound to maize CK2R [pdb1j91]16 and human
pCDK2 [pdb1P5E],17 as well as in other TBBt analogues bound
to maize CK2R [pdb2oxd, pdb2oxx, pdb2oxy]18 [pdb1zoe,
pdb1zog].19 It should be noted that the TBBt molecule in the
two complexes differs by its prototropic state. In the complex
with CK2R, the ligand molecule is asymmetric with C4a-C7a,
C7a-N1, N1-N2, N2-N3, and N3-C4a bond lengths of 1.38,
1.40, 1.24, 1.42, and 1.38 Å, respectively, suggesting that, in
contrast to the monoanionic form highly predominant in aqueous
medium, the TBBt molecule is in the neutral N(1)-H/N(3)-H
form (see Scheme 4). In the complex with pCDK2, TBBt is
symmetric, and C4a-C7a, C7a-N1, N1-N2, N2-N3, and
N3-C4a bond lengths are 1.45, 1.38, 1.43, 1.43, and 1.38 Å,
respectively. This indicates that N1-N2, N2-N3 (1.43 Å), and
C4a-C7a (1.45 Å) are single chemical bonds, indicative of
TBBt in the neutral N(2)-H form. Note that, for the symmetrical
anionic form, the expected length of the nitrogen-nitrogen bond
is 1.33 Å, which is an average value of single (1.43 Å) and
double (1.24 Å) chemical bonds. On the other hand, QM/MM
calculations suggested that TBBt analogues bind efficiently to
CK2R in the monoanionic form,20 and thus both hydrophobic
forces and electrostatic interactions together drive inhibition of
CK2R activity.

5-Substituted TBBt Inhibitors of CK2R
J. Phys. Chem. B, Vol. 114, No. 32, 2010 10603
SCHEME 3
TABLE 1: Inhibition (IC50, and Standard Error in
Parentheses, in µM) of Human CK2r by 5-Substituted
Benzotriazoles and 5-Substituted-4,6,7-Tribromobenzotriazoles
SCHEME 4: Prototropy of the Neutral Forms of
5-Substituted Benzotriazoles (X ) H) and Their
4,6,7-Tribromo Analogues (X ) Br)
Br3Bt analogues are largely dissociated, whereas C(5)-substituted
Bt analogues are generally neutral. The fraction of the anionic
form in neutral solution was estimated assuming that the spectra
recorded at pH 7 are a superposition of those recorded for neutral
(pH 2) and anionic (pH 12) forms. This approach was successful
for most of the compounds, and the estimated populations of
neutral forms at physiological pH are collected in Table 2. Ab
initio calculations further demonstrated the predominance of
these two forms, while the third possible, the symmetrical
N(2)-H, is barely detectable, if at all.
The above method failed for 5-NH2-Bt (2i), indicating that
this compound experiences an additional ionic equilibriumsmost
likely a cationic form at low pH. The spectra of TBBt, 5-Cl-
Br3Bt, and 5-CH3-Br3Bt (3f, 3e) in acidic medium display a
long-wavelength tail arising from light scattering of aggregated
particles. This indicates that these three compounds, which are
monoanions at pH 7, are poorly soluble in their neutral forms.
In parallel, standard UV-monitored titrations versus pH were
performed to determine the pKa values for dissociation of the
triazole ring proton (Table 3).
* Roughly estimated since titration was performed for ligand
concentrations in the range of 0-100:M.
Inhibitory Activity against Human CK2r. The inhibitory
activities of the various analogues against CK2R are presented
in Table 1. The 5-substituted 4,6,7-tribromobenzotriazoles with
a hydrophobic group in the 5-position (H, Me, Cl, Br, NO2)
exhibit similar inhibitory activities, with IC50 values in the low
micromolar range, but with the 5-Br significantly more active.
For the analogues with hydrophilic groups in the 5-position,
COOH (dissociated in neutral aqueous medium) and NH2
activity is substantially decreased. With a 5-fluorine substituent,
which may play a role in hydrogen bonding, the appropriately
substituted analogue has activity similar to the parent Br3Bt.
All of the experimental data leading to the determination of
IC50 are presented in Supporting Information Figure 1.
Physicochemical Properties of TBBt Derivatives. Unlike
the neutral form of TBBt which, because of symmetry, may
exist as two prototropic tautomers, the 5-substituted analogues
may exist as three prototropic forms (see Scheme 4). Moreover,
the TBBt analogues in aqueous solution at pH close to neutral
may display a protonation equilibrium. Comparison of the
UV-vis spectra recorded in aqueous solution at three pH values
(2, 7, 12) demonstrated that, at neutral pH, C(5)-substituted
The prototropic equilibrium of the neutral forms, monitored with
the aid of 13C NMR spectroscopy in anhydrous DMSO, demon-
strated existence of a relatively slow exchange between at least
two prototropic forms, resulting in significant broadening of 13C
resonance lines observed for most of the compounds (Figure 1).

10604 J. Phys. Chem. B, Vol. 114, No. 32, 2010
Wa˛sik et al.
TABLE 2: Structural Parameters of the Benzotriazole Derivativesa
dipole [D]
rmsd #X
#H ASA ∆ASA ASAap ∆ASAap
∆Gsolv
∆G1
∆G2
∆G3
∆Gdiss
ligand
H(N1) H(N2) H(N3)
[Å2] bonds bonds [Å2] [Å2]
[Å2]
[Å2]
[kcal mol-1
]
[kcal mol-1
]
[kcal mol-1
]
[kcal mol-1
]
[kcal mol-1
]
Bt
1.2
1.3
1.3
0.9
1.5
1.5
1.7
1.5
0.8
1.3
1.3
1.2
0.9
1.6
1.9
0
0
0
0
0.5
0
0
0.3
0
1.7
1.5
1.8
1.8
1.9
1.2
1.5
2.3
2.6
4.6 226
3.4 249
3.8 240
2.5 252
3.5 259
3.4 279
3.1 277
3.8 282
3.8 282
3.4 362
3.6 371
3.7 366
3.0 372
3.1 379
3.3 378
4.2 381
3.5 391
3.3 389
193
205
199
209
223
227
246
243
232
310
311
317
325
310
312
330
322
329
131
101
111
157
165
94
101
188
97
286
264
272
297
302
246
254
313
249
102
79
87
118
131
74
89
152
79
237
213
230
251
239
198
216
250
210
5.6
6.8
6.0
5.8
5.8
7.9
16.1
5.8
6.1
3.7
4.5
4.8
4.1
4.5
7.0
16.5
4.4
4.9
0.7
2.5
2.0
1.2
2.4
6.3
15.9
2.5
3.4
2.0
1.1
3.5
1.3
3.5
6.7
15.7
3.4
1.1
5.7
6.9
3.4
6.2
3.2
3.2
11.9
3.2
2.9
5.3
6.3
4.0
5.8
4.3
4.9
11.1
4.4
6.3
-3.3
-5.7
-2.7
-1.0
-1.0
-6.2
-63.3
-1.0
-3.8
4.3
3.2
3.9
5.9
5.8
1.7
-55.4
5.3
0.0
2.8
0.4
0.5
0.3
0.0
0.0
0.0
2.2
0.6
2.7
0.6
1.2
0.0
0.0
3.0
0.0
1.0
5.1
5.9
5.4
3.6
5.6
7.0
6.7
5.4
6.1
3.4
4.8
3.6
5.3
4.1
4.4
5.1
5.5
5.2
0.0
0.0
0.0
0.0
0.0
2.4
0.0
0.3
0.0
0.0
0.0
0.0
0.0
0.4
0.7
0.0
0.0
0.0
59.5
62.5
57.8
59.9
56.7
54.3
53.1*
57.4
58.4
51.4
53.9
51.1
52.8
49.7
45.8
48.3*
49.7
50.4
5-NH2-Bt
5-F-Bt
5-CH3-Bt
5-Cl-Bt
5-NO2-Bt
5-COOH-Bt
5-Br-Bt
5-CF3-Bt
Br3Bt
5-NH2-Br3Bt
5-F-Br3Bt
5-CH3-Br3Bt
5-Cl-Br3Bt
5-NO2-Br3Bt
5-COOH-Br3Bt 1.6
TBBt
5-CF3-Br3Bt
1.6
1.3
2.7
a The rmsd is a measure of the average displacement of a ligand in the structure of the complex; #X bonds, # Hbonds are the average
number of halogen and hydrogen bonds, estimated from the calculated structures of the complexes; ASA is the average solvent-accessible area,
∆ASA its change upon complex formation; and ASAap and ∆ASAap are the values calculated only for the apolar atoms (e.g., excluding N, O,
HN, and HO). All of these parameters were estimated as average values calculated for the 20 lowest energy structures of the complexes. The
population of the major protomeric forms of the neutral ligands, P, and their dipole moments, D, were estimated with the aid of
quantum-mechanical calculations, using the PCM model of aqueous solvent; solvation free energy, ∆Gsolv was calculated ab initio for the
lowest energy protomeric form of the neutral ligand; ∆G1, ∆G2, and ∆G3 are the QM-derived relative free energies of the three possible
protomeric forms, and ∆Gdiss is the estimated free energy of dissociation of the triazole proton (cf. eq 3 for details). *Values calculated for the
nondissociated carboxyl group.
TABLE 3: Parameters Employed for Prediction of Inhibitory Activities of the Various Ligandsa
IC50 [µM ]
ligand
pKa
Xma
log P
log S Vmol [Å3] #F experimental R (Vmol) QSAR R(log D#) R(log S)
Bt
8.56 (0.04) 0.03
2i 9.23 (0.04) 0.01
2a 7.96 (0.05) 0.10
2e 8.79 (0.05) 0.02
2f 7.62 (0.08) 0.19
2d 6.31 (0.07) 0.83
1.21 (0.17) -1.38
0.52 (0.40) -1.52
1.49 (0.35) -1.44
1.62 (0.21) -1.53
1.93 (0.37) -1.96
1.20 (0.33) -1.66
103
115
106
120
119
125
128
125
129
168
179
171
185
182
189
192
188
192
0
0
1
0
0
0
0
0
3
0
0
1
0
0
0
0
0
3
nonactive
nonactive
180 (30)
440 (140)
72 (10)
54 (8)
190 (80)
51 (7)
109 (21)
6.6 (0.7)
73 (21)
3.51 (0.17)
2.49 (0.43)
1.20 (0.09)
2.26 (0.18)
nonactive
0.61 (0.09)
2.28 (0.21)
228
113
191
84
90
62
55
62
52
4.9
2.6
4.3
1.9
2.2
1.5
1.2
1.5
1.3
170
120
170
110
75
68
68
48
110
456
2.8 × 105
178
476
66
42
175
38
220
176
200
173
88
141
220
66
5-NH2-Bt
5-F-Bt
5-CH3-Bt
5-Cl-Bt
5-NO2-Bt
5-COOH-Bt
5-Br-Bt
5-CF3-Bt
Br3Bt
5-NH2-Br3Bt
5-F-Br3Bt
5-CH3-Br3Bt
5-Cl-Br3Bt
5-NO2-Br3Bt
2g 8.08 (0.06) 0.92* 0.40 (1.15) -1.38
2b 7.55 (0.07) 0.22
2c 7.23 (0.05) 0.37
3h 5.38 (0.04) 0.98
3i 5.95 (0.05) 0.92
3a 4.77 (0.04) 0.99
3e 6.20 (0.30) 0.86
3f 5.20 (0.30) 0.98
3d 4.49 (0.07) 1.00
2.03 (0.37) -2.14
2.09 (0.23) -2.08
3.39 (0.31) -3.95
2.71 (0.23) -3.78
3.55 (0.36) -4.06
3.78 (0.35) -4.33
4.09 (0.31) -4.58
3.28 (0.29) -4.05
171
72
3.7
3.7 4.2
2.5 72
3.5 3.3
2.3 4.3
1.5 1.0
1.6 2.6
1.5 1815
1.0 0.9
2.3 1.9
4.9
3.1
2.0
1.4
3.2
4.2
1.0
1.8
5-COOH-Br3Bt 3g 5.19 (0.05) 0.02* 2.32 (0.96) -3.87
TBBt
5-CF3-Br3Bt
5.10 (0.30) 0.99
3c 4.58 (0.04) 1.00
4.24 (0.32) -4.80
4.22 (0.46) -4.42
a Experimental pKa; Xma, population of monoanionic form at neutral pH; log P and log S are consensus values of calculated partition and
solubility coefficients; Vmol, calculated molecular volume; IC50, experimental values and those estimated from the four best single-parameter
nonlinear regression models, and, finally, by QSAR. *Corrected for ionization of carboxyl group.
Preliminary QM calculations pointed to the N(1)-H and
N(3)-H neutral forms as the main interexchanging states (cf.
∆G1, ∆G2, and ∆G3 in Table 2). The calculated free energy of
dissociation of the triazole proton (∆Gdiss in Table 2) was found
to be dependent on the nature of the 5-substituent, being lowest
for the neutral carboxyl and highest for the amino derivatives.
This is in line with the resonance patterns observed in the 13C
NMR spectra, where 5-NH2-Br3Bt (3i) exhibits two separate
sets of slightly broadened resonance lines, corresponding to the
N(1)-H and N(3)-H forms, while 5-COOH-Br3Bt (3g) displays
a set of seven narrow resonance lines. For the other compounds,
with C(5) substituents displaying intermediate electronic proper-
ties, the patterns of the resonance lines fall in the medium
exchange/coalescence regime, leading to extremely broad lines
with a width exceeding 1 kHz. Addition of a small amount of
water (∼1% v/v) led to significant broadening of the resonance
lines, pointing to an increase in proton transfer rates via the
more efficient water-mediated one. An analogous effect has been
reported for benzotriazole derivatives symmetrically substituted
on the benzene ring.26 The appropriate structure-related data are
summarized in Tables 2 and 3.
Molecular Modeling of Inhibitor Binding to Human CK2r.
Two Modes of Binding TBBt by CK2r. For each of the Bt and
Br3Bt analogues, a set of 10 lowest energy structures, selected
from 150 models of their complexes with human CK2R

5-Substituted TBBt Inhibitors of CK2R
J. Phys. Chem. B, Vol. 114, No. 32, 2010 10605
Figure 2. Lowest energy structure of the modeled complex of human
CK2R with 5-NO2-Br3Bt. The protein backbone is represented as a
ribbon model, the benzotriazole derivative by thick lines, and its volume
by a transparent yellow surface. The side chains of the residues found
interacting with the ligand in most of the low-energy structures are
identified; their apolar bonds are in black, while those with oxygen
and nitrogen are in red and blue, respectively. Backbone carbonyl
groups of V45, G46, and S51, as well as side chain oxygen atoms of
S51, E114, N118, and D175, make relatively highly populated halogen
bond-like contacts. The 5-NO2 substituent (denoted by an arrow) is
located in the peripheral region of the enzyme binding pocket and is
therefore partially accessible to solvent.
Figure 1. 13C spectra of 5-substituted derivatives of (A) benzotriazole
and (B) 4,6,7-tribromobenzotriazole in anhydrous DMSO, and (C) after
addition of 1% (v/v) water to the DMSO solutions of B.
complexed with either maize CK2R [pdb1j91] or human pCDK2
[pdb1p5e] (see Supporting Information Figure 2).
Common Pattern of Interactions. For each TBBt analogue,
the close contacts with CK2R residues found in at least 7 of
the 10 lowest energy structures of the complexes were analyzed.
This pointed to a common pattern of intermolecular interactions
stabilizing location of Bt and Br3Bt congeners in the CK2R
binding pocket. Thus the side chain carboxyl of Asp175
recognizes the N(2)-N(1)-H triazole ring configuration, and
this interaction was found for all of the ligands. Additionally,
the Ser51 hydroxyl is hydrogen bonded to the triazole N(2).
Halogen atoms of the Bt benzene ring make numerous contacts
with aliphatic side chains of the residues Val45, Val53, Ile66,
Phe113, Met163, and Ile174. This stabilizes the complex by
protecting the hydrophobic halogen atoms from solvent acces-
sibility. Additionally, some intermolecular contacts between
ligand bromine atoms and CK2R backbone (Val45, Gly46,
Ser51) and side chain (Ser51, Asn118, Asp175) oxygen atoms
are observed. These may be identified as halogen bond interac-
tions that stabilize the complex (see Figure 2). The pattern of
interactions involving Val45, Ser51, Glu114, and Asp175 was
obtained from SA simulations, were subjected to further
analysis. Initially, location of a potential inhibitor with respect
to the CK2R binding site was tested in terms of the rmsd values
estimated for the heavy atoms of the heterocyclic ring. All
compounds were found flexibly located in the protein pocket,
which readily adapts to small inhibitor movement. The rmsd
of their average locations varied from 1 Å (5-CF3-Bt (2c),
5-CH3-Bt (2e), 5-CH3-Br3Bt (3e)) up to 2 Å (5-NO2-Br3Bt (3d),
5-COOH-Bt (2g), 5-NH2-Br3Bt (3i)). Comparison of the en-
sembles of lowest energy structures obtained for the various
ligands points to a weak, but noticeable, ligand-specific prefer-
ence in its location inside the CK2R pocket. Both observations
clearly indicate that the CK2R binding site is flexible enough
to bind a broad class of heteroaromatic compounds. In particular,
for TBBt, two alternative binding modes were found by the
SA procedure (see Experimental Section), both closely corre-
sponding to those found in the crystal structures of TBBt

10606 J. Phys. Chem. B, Vol. 114, No. 32, 2010
Wa˛sik et al.
found to be common to all of the complexes modeled with the
tribromobenzotriazole ligands.
Substituents on the Benzene Ring Modify Stability of the
Complex. The flexible location of the Bt analogues in the CK2R
pocket permits a C(5) substituent to be either partially solvent-
exposed or fully protected from solvent accessibility, so that
both polar and apolar C(5)-substituted derivatives could be
efficiently bound by the enzyme.
Both theoretical (QM calculations) and experimental (13C
NMR and UV titration) data demonstrate that a C(5) substituent
on the benzene ring modulates the distribution of electron
density on the triazole ring, hence affecting both the relative
protomer populations and the pKa, which may result in modula-
tion of the crucial interaction with the Asp175 side chain; for
example, 5-NO2-Br3Bt (3d) activity is close to that of the parent
TBBt or 5-Cl-Br3Bt (3f), whereas 5-CH3-Br3Bt (3e) and Br3Bt
(3h) are somewhat less active.
Structure-Activity Relationships. Correlation of Inhibi-
tor ActiWity with Its Structure. For each TBBt congener, its
molar volume (Vmol), surface area (Smol), as well as the solvent-
accessible area (ASA) and its apolar part (ASAap) were estimated
(see Experimental Section). The populations of the three neutral
prototropic forms obtained by the ab initio method, the disorder
of a ligand location in the CK2R pocket estimated from
molecular modeling simulations, and the experimentally derived
population of neutral and charged forms at pH 7 were also
considered as potential factors influencing ligand activity.
Analysis was restricted to active ligands which exhibited IC50
values below 100 µM. The best single-parameter correlation
was obtained for ln(IC50) versus Vmol, which, according to LCW
theory,27 is an exact descriptor of hydrophobic interactions (see
Table 3 and Figure 3A). This strongly points to hydrophobic
interactions as the dominant factor influencing inhibitory
activity.
Correlation of Inhibitor ActiWity with Structure of the
Complex. The 10 lowest energy structures of the complexes
were then subjected to more detailed analysis. For each complex,
the average number of identified halogen, hydrogen, and fluorine
bonds was estimated. Reduction of the solvent-accessible area
and its apolar part, upon complex formation, was calculated
for each structure of the complex and then averaged. The mean
movement of a ligand in the complex (rmsd) was also estimated.
Following the concept of atomic solvation parameterization
(ASP),28 intermolecular hydrophobic interactions may also be
evaluated in terms of changes in the solvent-accessible surface
upon complex formation. Reduction of the apolar part of the
solvent-accessible surface area of a ligand upon complex
formation is also significantly correlated with its inhibitory
activity (not shown).
In general, there is no correlation between inhibitor volume,
or its activity, with its flexibility in the bound state, estimated
as the rmsd describing the average displacement within the
complex. This again points to the large flexibity of the CK2R
pocket, enabling it to accommodate a wide class of compounds
of various sizes. The effect of electron density on a C(5)
substituent in direct electrostatic interactions requires further
analysis.
QSAR Analysis. All physical and structural parameters for
the compounds collected in Tables 2 and 3 were subjected to
predictions of inhibitory activity, and in accord with the QSAR
approach,29 we assumed that ln(IC50) could be expressed as a
linear combination of the analyzed parameters. Following the
method of principal component analysis (PCA),30,31 we stepwise
iteratively reduced the number of parameters required for
Figure 3. Experimental IC50 values as a function of (A) Vmol, molecular
volumes of the ligands; (B) log[Pcons], calculated coefficients of ligand
partition between polar and apolar phases; (C) log[S], calculated
solubility coefficients. The high correlation coefficients of 0.94 (0.99
for TBBt, 3a, 3f, 3h: dashed line), 0.89, and 0.96, respectively, indicate
that the mechanism of inhibition by benzotriazole derivatives substituted
on the benzene ring is largely driven by hydrophobic interactions.
According to the presented analysis, the IC50 for the hitherto unknown
TIBt (o) was estimated as 0.12 ÷ 0.5 µM. All models overestimate
activity of 2e and 3i.
prediction of activity. The final, optimal model underlines the
role of the molecular volume of the ligand, the number of
fluorine atoms attached, and the fraction of the neutral form
estimated at pH 7. The average number of halogen bonds was
found to be virtually nonsignificant at the 10% significance level.
The model points to predominance of the volume effect of the
ligand, which, located deeply within the enzyme binding site,
is thus protected from solvent accessibility. Polar substituents
(COOH, NH2) attached to the benzene ring destabilize the

5-Substituted TBBt Inhibitors of CK2R
J. Phys. Chem. B, Vol. 114, No. 32, 2010 10607
Effects of Fluorine and Halogen Bonding. According to the
final two-parameter QSAR model, a fluorine atom attached to
the benzene ring of a Bt derivative increased the IC50 by a factor
∼1.4 (see Table 4). This is largely due to the unfavorable change
in free energy upon transfer of a fluorine atom from the polar
solvent to the apolar environment of the protein pocket.
Consequently, derivatives with a fluorine atom on the benzene
ring are weaker inhibitors than expected according to their
molecular volume, Vmol. Less clear is the role of halogen bonds.
In contrast to literature data,33 a halogen bond (defined as the
close contact between a halogen and O or N electron-rich atoms)
was not found to be significant in stabilization of the ligand in
the binding pocket. This conclusion must be judged bearing in
mind the predominance of hydrophobic interactions in stabiliza-
tion of the protein-ligand complexes, which are directly
correlated with the number of halogen atoms in the ligand.
Analysis of the parameters of the models that include the effect
of halogen bonding indicates that the contribution of hydro-
phobic interactions to the final free energy of complex formation
(∼-5 kcal/mol) is an order of magnitude larger than the average
energy estimated for a putative halogen bond (∼0.2 kcal/mol).
This may be readily verified independently by examination of
ligand aqueous solubility.
Figure 4. Optimal model that can be used for prediction of inhibitory
activities of benzotriazole and its analogues substituted on the benzene
ring. IC50 is the experimentally measured value, log P is the predicted
polarity; pKa values were determined experimentally, and XN(1)-H is
the population of the N(1)-H protomer, estimated by QM calculations.
This model is the only one which properly predicts the activities of 2e
and 3i and points to Bt, 2i, and 3g as inactive compounds (cf. Table 3
for the details).
Solubilities in neutral aqueous medium of benzotriazole and
some mono-, di-, tri-, and tetrabromobenzotriazoles are shown
in Figure 5. Hydrophobic effects are responsible for the observed
decrease of solubility with increasing molecular volume of the
compounds.
complexes, shifting the equilibrium toward the dissociated, free
state of the ligand. Contact of a fluorine atom on the benzene
ring with a proximal protein carbonyl/carboxyl oxygen increases
IC50 1.5- to 5-fold, due to unfavorable electrostatic interactions
between negatively charged centers. The final results of regres-
sion analysis (Figure 4) are summarized in Table 4.
Bearing in mind that QSAR analysis pointed to hydrophobic
interactions involving the monoanionic form of a ligand as the
dominant driving factor in its inhibitory activity, we tested the
combinations of descriptors related to the hydrophobic properties
of the ligand. The best results were obtained for the log P and
log S coefficients, which estimate partition of the molecule in
an octanol-water two-phase system and its solubility in water,
respectively (Figure 3B,C). Since the QSAR method pointed
to the populations of the monoanion and the QM-derived
populations of the neutral N(1)-H protomer as important for
prediction of ligand inhibitory activity, we introduce, following
the method for correction of the partition coefficient of ionizable
compounds,32 the corrections of log P as follows:
Since each water molecule exposes an oxygen atom capable
of participating in a halogen bond, it appears that the eventual
halogen bonding effect is masked by existence of dominant
hydrophobic solvation. Hence, the observed halogen bond
interaction may orientate the ligand molecule in the apolar
protein cavity, thus slightly decreasing the free energy of
binding, but the ligand-binding process is generally driven by
hydrophobic interactions that are represented as the molecular
volume, or log D, as shown in Figure 3A,C. This is also strongly
supported by the reported 20-fold decrease of inhibitory activity
of TBBt against a CK2RV66A/I174A double mutant.2
We are presently examining possible extension of the
structure-activity relationship described above to predict inhibi-
tory properties of other structurally related analogues of TBBt,
for example, those in which all four bromine atoms are replaced
by other substituents, with preliminary promising results. For
example, tetramethylbenzotriazole (TMBt) is predicted to exhibit
very low activity (consensus log P ) 2.04 ( 0.57; pKa ) 8.85;26
IC50,calcd ) 140 µM), in agreement with early experimental
results.4 Tetrachlorobenzotriazole (TCBt), with consensus log
P ) 3.71 ( 0.27; pKa ) 4.57;26 IC50,calcd ) 2.6 µM, is predicted
as more than 2-fold less active than TBBt (IC50,calcd ) 1.2 µM),
again in agreement with experimental results.4 The hitherto
unknown tetraiodobenzotriazole (TIBt; consensus log P ) 4.77
( 1.12; pKa < 4.5; IC50,calcd ) 0.6 µM) was estimated to be at
least two times more active than TBBt. This is in accord with
analysis of a series of 4,5,6,7-tetrahalogeno-1H-isoindole-1,3-
(2H)-dione inhibitors,34 where replacement of the four bromines
on the benzene ring by iodines decreased IC50 by a factor of
3-10, depending on the nature of the substituent on N(2). It is
also consistent with the report5 that tetraiodobenzimidazole is
up to 10-fold more active than tetrabromobenzimidazole. It
should be noted that, whereas the hydrophobicity of the mono-
anionic forms of the ligands is the dominant factor describing
their inhibitory activity, the propensity for protonation of the
N(1) nitrogen is also significant.
The final model, which accurately predicts IC50 values of all
the active compounds, with an R2 value of 0.98, is presented in
Figure 4. This model, based on the log D# descriptor appears
to be the best because it not only predicts the inhibitory activity
of all TBBt derivatives displaying IC50 < 500 µM but also points
to Bt, 5-COOH-Br3Bt (3g), and 5NH2-Bt (2a) as inactive
compounds.
In order to judge the statistical significance of all the pa-
rameters in the proposed model, an additional procedure was
carried out to test independently the significance of log P,
log(Xma), and log(XN(1)-H) parameters used in eq 1. The p
statistics values of 9 × 10-10, 1.5 × 10-5, and 5.9 × 10-6
obtained for log P, log(Xma), and log(XN(1)-H), respectively,
clearly indicate that, despite the ligand polarity being the
dominant factor, both protomeric and dissociation equilibria
almost equally drive its binding to CK2R.

10608 J. Phys. Chem. B, Vol. 114, No. 32, 2010
Wa˛sik et al.
TABLE 4: Parameters of the Four Best Models Describing the Inhibitory Activities of a Series of Benzotriazoles Substituted on
the Benzene Ring, According to the Equation ln(IC50) ) ∑wi ·xi, Where the Corresponding Structural Parameters are Listed in
Table 3a
weight, wi
model
value (SE)
intercept
log P/log D
Vmol
#F
R2 (df)
F (R)
85.1 (3.3 × 10-6
rank
5
R(log P)
R(log D#)
R(log S)
7.2 (0.6)
12.7
-1.68 (0.18)
9.2
0.89 (11)
)
t-stat
P
value (SE)
t-stat
1.7 × 10-7
5.2 (0.1)
45.0
3.3 × 10-6
-1.35 (0.05)
27.0
0.98 (13)
0.95 (11)
0.94 (11)
0.96 (11)
730 (8.3 × 10-13
)
1
2
4
3
P
1.2 × 10-15
7.6 (0.4)
20.2
8.3 × 10-13
1.59 (0.11)
14.3
value (SE)
t-stat
P
value (SE)
t-stat
P
value (SE)
t-stat
P
206 (1.8 × 10-8
154 (2.1 × 10-7
112 (4.3 × 10-7
)
)
)
4.9 × 10-10
11.5 (0.8)
15.2
1.8 × 10-8
R (Vmol
)
-0.059 (0.005)
12.4
3.1 × 10-8
11.3 (0.6)
17.6
2.1 × 10-7
-0.059 (0.004)
14.7
QSAR
0.26 (0.11)
2.3
2.8 × 10-8
1.3 × 10-7
4.7 × 10-2
a For each model, the value of R2, the number of degrees of freedom (df), Snedoker statistics (F), and the corresponding significance level
(R) are presented. The models were ranked according to their significance level, R.
(1a, 1d), benzotriazole, 5-CH3-, 5-COOH-, and 5-Cl-benzot-
riazoles (2e, 2g, 2f) were products of Aldrich or Lancaster
(USA), and 4,5,6,7-tetrabromobenzotriazole (TBBt) was syn-
thesized as previously reported.23
Synthesis. 5-Fluoro-1H-benzotriazole (2a). To a cooled (0
°C) and stirred solution of 1a (126 mg, 1 mmol) in 2 mL of
AcOH and 0.8 mL of H2O was added 114 mg (1.65 mmol) of
NaNO2 dissolved in 0.7 mL of H2O. Stirring was continued at
room temperature for 20-30 min; 0.5 mL of HCl was then
added, and the resulting solution was poured into a mixture of
EtOAc and brine (1:1 v/v, 50 mL). Organic layers were
separated, dried over Na2SO4, filtered, and excess was solvent
removed in vacuo. The residue was dissolved in EtOH/H2O,
treated with activated charcoal, filtered, and evaporated to give
116 mg (85% yield) of 2a; mp 148-149 °C, lit. 144-145.8
°C,36 Rf (A) 0.63, (B) 0.58. UV λmax (ε): (pH 2) 258.5 nm (5780),
Figure 5. Correlation of experimentally measured solubilities (cx) of
273.5 (5595); (pH 7) 260 nm (5960), 274 nm (6130); (pH 12)
275 nm (9690). 1H NMR δ: 8.03 (dd, 1H, J ) 4.7 Hz, J ) 8.9
Hz), 7.73 (dd, 1H, J ) 2.3 Hz, J ) 8.8 Hz), 7.35 (td, 1H, J )
2.3 Hz, J ) 9.2 Hz). 13C NMR δ: 160.31 (J ) 241.9 Hz), 137.61
(J ) 13.3 Hz), 137.36, 117.63 (J ) 10.7 Hz), 114.63 (J ) 27.5
Hz), 99.04 (J ) 26.7 Hz). MS for C6H5FN3 [M + H]+: found
137.9920, calcd 138.0468.
benzotriazole and its mono-, di-, tri-, and tetrabromo analogues with
their molecular volumes (Vmol). Bearing in mind the correlation between
IC50 and Vmol (0.92, cf. Table 4 and Figure 3A), the high correlation
between Vmol and cx (0.998) supports the mechanism of inhibition via
hydrophobic interactions.
Experimental Section
5-Bromo-1H-benzotriazole (2b). To a cooled (0 °C) and
stirred solution of the dihydrochloride of 1b (260 mg, 1 mmol)
in 6.5 mL of H2O was added 0.26 mL of concentrated HCl to
a solution of NaNO2 (114 mg, 1.65 mmol) in 0.7 mL of H2O.
Stirring was continued at room temperature for 30 min. The
solid product was filtered and washed with H2O. Recrystalli-
zation from EtOH/H2O gave 163 mg (83% yield) of 2b; mp
152.2-153.9 °C, lit. 150 °C,37 Rf (A) 0.61, (B) 0.6. UV λmax
(ε): (pH 2) 265 nm (5372), 282.5 nm (5525); (pH 7) 282.5 nm
(6190); (pH 12) 282 nm (6967). 1H NMR δ: 8.17 (dd, 1H, J )
0.6 Hz, J ) 1.7 Hz), 7.89 (dd, 1H, J ) 0.6 Hz, J ) 8.7 Hz),
7.5 (dd, 1H, J ) 1.7 Hz, J ) 8.7 Hz). 13C NMR δ: 137.86,
133.38, 131.7, 128.48, 117.46, 116.95. MS for C6H4Br1N3 [M
+ H]+: found 199.9303, calcd 199.9646.
Chemical Methods. Melting points (uncorrected) were
determined in open capillary tubes, using a Bu¨chi apparatus
B504. UV absorption spectra were recorded on a Cary 300
instrument. All compounds were checked by thin-layer chro-
matography (TLC) on 0.2 mm Merck silica gel 60 F254 plates.
Preparative separations were carried out by column chroma-
tography, using Merck silica gel (230-400 mesh) or on
preparative glass plates (2 mm, Merck silica gel 60 F254), using
the following solvent systems: (A) CHCl3/MeOH (9:1), (B)
EtOAc/n-heptan/AcOH (2:2:0.2). Mass spectrometry was per-
formed with a Micromass ESI Q-TOF spectrometer. 1H and 13
C
NMR spectra were recorded in DMSO-d6 on a Varian 500 or
Varian 400 (5-bromobenzotriazole) instrument. 13C spectra of
5-substituted 4,6,7-tribromobenzotriazoles and 4,6,7-tribro-
mobenzotriazole were recorded in aqueous medium.26 Spectra
were analyzed with the aid of the MestRe-C (version 2.3a)
program (Cobas, Cruces Mestre-C 2.3a). Chemical shifts are
in parts per million relative to tetramethylsilane (Me4Si, δ )
0). Starting materials: 5-F- and 5-NO2-1,2-phenylenediamines
5-Trifluoromethyl-1H-benzotriazole (2c). To a cooled (0 °C)
and stirred solution of 1c (176 mg, 1 mmol) in 2 mL of AcOH
and 0.8 mL of H2O was added 114 mg (1.65 mmol) of NaNO2
dissolved in 0.7 mL of H2O. Stirring was continued at room
temperature for 20-30 min; 0.5 mL of HCl was then added,

5-Substituted TBBt Inhibitors of CK2R
J. Phys. Chem. B, Vol. 114, No. 32, 2010 10609
and the resulting solution was poured into a mixture of EtOAc
and brine (1:1 v/v, 50 mL). Organic layers were separated, dried
over Na2SO4, filtered, and excess solvent was removed in vacuo.
The residue was dissolved in EtOH/H2O, treated with activated
charcoal, filtered, and evaporated to give 182 mg (97% yield)
of 2c; mp 125.2-127.1 °C, lit. 132 °C.38 UV λmax (ε): (pH 2)
259.5 nm (5001); (pH 7) 274.5 nm (5236); (pH 12) 276 nm
(7226). 1H NMR δ: 8.4 (s, 1H,), 8.09 (d, 1H, J ) 8.5 Hz), 7.72
(d, 1H, J ) 8.5 Hz). 13C NMR δ: 140.74, 139.65, 125.87 (q, J
) 32.2 Hz), 125.14 (q, J ) 272.5 Hz), 122.54 (q, J ) 3.9 Hz),
116.2, 115.4 (q, J ) 3.9 Hz). MS for C7H5F3N3 [M + H]+:
found 188.0430, calcd 188.0436.
5-Nitro-1H-benzotriazole (2d). To a mixture of 1d (153 mg,
1 mmol) in 2 mL of AcOH and 0.8 mL of water at 0 °C was
added a solution of NaNO2 (114 mg, 1.65 mmol) in 0.7 mL of
water. The solid was collected by filtration and washed with
H2O to afford 149 mg (91% yield) of pure light tan 2d; mp
212.4-213.7 °C, lit. 215-216 °C,39 Rf (B) 0.48. UV λmax (ε):
(pH 2) 290 nm (7527); (pH 7) 307.5 nm (6905); (pH 12) 313
nm (7471). 1H NMR δ: 8.93 (d, 1H, J ) 2.0 Hz), 8.28 (dd, 1H,
J ) 2.0 Hz, J ) 9.0 Hz), 8.10 (d, 1H, J ) 9.0 Hz). 13C NMR
δ: 144.60, 140.55, 138.69, 121.01, 114.25, 114.12. MS for
C6H5N4O2 [M + H]+: found 165.0361, calcd 165.0413.
5-Amino-1H-benzotriazole (2i). (a) To a solution of 2d (164
mg, 1 mmol) in 1 mL of concentrated HCl was added dropwise
a solution of SnCl2 (569 mg, 3 mmol) in 2 mL of concentrated
HCl. The resulting mixture was stirred for 1 h at room
temperature. The pale blue product was filtered off, washed with
EtOAc, and dried to afford 162 mg (95% yield) of the
hydrochloride 2i. (b) Reduction with HI was as follows: A
suspension of 2d (164 mg, 1 mmol) in 3 mL of aqueous HI
(57%, unstabilized) was heated at 90 °C for 3 h. After cooling
to room temperature, it was diluted with EtOAc (50 mL) and
washed with saturated Na2S2O3 until all formed iodine was
removed, as indicated by disappearance of the dark purple color.
The reaction mixture was brought to neutrality with saturated
NaHCO3, and the organic layer was dried over Na2SO4 and
evaporated. The residue was dissolved in n-hexane/EtOH (2:1)
and concentreated HCl added. The resulting precipitate of the
hydrochloride of 2i was filtered off and washed with n-hexane
to give 102 mg (60% yield).
silica gel column chromatography (0-15% methanol in chlo-
roform) gave 85 mg (20% yield) of pure 3c: mp >300 °C (dec.),
Rf (B) 0.65. UV λmax (ε): (pH 2) 279.5 nm (2630); (pH 7) 290.5
nm (2580), 303 nm (2530); (pH 12) 290.5 nm (2640), 302.5
nm (2590). 13C NMR δ: 144.74, 143.52, 123.14 (q, J ) 275.9
Hz), 118.14 (q, J ) 28.8 Hz), 115.186, 112.03, 110.21 (q, J )
2.44 Hz). MS for C7H2Br3F3N3 [M + H]+: found 423.7738,
calcd 423.7730.
5-Nitro-4,6,7-tribromo-1H-benzotriazole (3d). 4,6,7-tribro-
mobenzotriazole (3h) (356 mg, 1 mmol) was dissolved in
concentrated H2SO4 (1.75 mL) and stirred magnetically at room
temperature. NaNO3 (850 mg, 10 mmol) was added stepwise
over a period of 10 min. The resulting suspension was heated
with stirring at 70 °C until the reaction was complete (∼4 h);
the cooled mixture was poured into a mixture of ice and water
and the resulting precipitate filtered off, washed with H2O, and
dried. Recrystallization from AcOH and H2O gave 277 mg (69%
yield) pure 3d: mp 242.5-245.5 °C, Rf (B) 0.52. UV λmax (ε):
(pH 2) 277.5 nm (8770); (pH 7) 289.5 nm (9640); (pH 12) 289
nm (9920). 13C NMR δ: 149.23, 139.96, 139.80, 113.55, 112.78,
102.72. MS for C6H2Br3N4O2 [M + H]+: found 400.7722, calcd
400.7707.
5-Methyl-4,6,7-tribromo-1H-benzotriazole (3e). This com-
pound was prepared in 76% yield (282 mg) starting from 2e
(133 mg, 1 mmol) as described for 3a, but with more prolonged
stirring (24 h): mp 238.2-241.1 °C, Rf (A) 0.68, (B) 0.77. UV
λ
max (ε): (pH 7) 292.5 nm (10460); (pH 12) 292.5 nm (10980);
1
(MeOH) 281 nm (10120), 294.5 nm (8900). H NMR δ: 2.64
(s, 3H). 13C NMR δ: 139.51, 138.6, 136.53, 111.34, 125.31,
108.25, 25.32.MS for C7H5Br3N3 [M + H]+: found 369.7946,
calcd 369.8013.
5-Chloro-4,6,7-tribromo-1H-benzotriazole (3f). This com-
pound was prepared in 70% yield from 2f as described for 3e,
but with more prolonged stirring (41 h). Repeated recrystalli-
zation from MeOH gave 272 mg of pure 3f: mp 250.2-252.7
°C, lit. 249-251 °C,23 Rf (A) 0.39, (B) 0.73. UV λmax (ε): (pH
7) 291.5 nm (10440), 299.5 nm (10670); (pH 12) 291 nm
(10610), 299.5 nm (10850) (MeOH) 286 nm (8920), 300 nm
(8420), 310.5 nm (6040). 13C NMR δ: 139.30, 132.40, 123.37,
112.15, 108.90. MS for C6H2Br3N3 [M + H]+: found 389.7514,
calcd 389.7467.
5-Carboxy-4,6,7-tribromo-1H-benzotriazole (3g). A suspen-
sion of 3e (370 mg, 1 mmol) in 3.32 mL of 0.3 M KOH was
stirred and heated to ∼90 °C. KMnO4 was added in three
portions (207, 146, and 67 mg) over 3-4 h. Stirring and heating
was continued for another 2-3 h at 100 °C, and the resulting
precipitate of MnO2 was filtered off with the aid of Celite Hyflo
and washed with hot water. The clear colorless solution was
acidified with concentrated HCl, and the resulting precipitate
was collected by filtration and washed with H2O to afford 260
mg (72%) pure 3g: mp 297.4-300.5 °C, Rf (B) 0.14. UV λmax
(ε): (pH 2) 283.5 nm (10480); (pH 7) 296 nm (14890); (pH 12)
296 nm (13480). 13C NMR δ: 168.17, 140.03, 139.56, 138.30,
119.30, 111.78, 105.38. MS for C7H3Br3N3O2 [M + H]+: found
399.7423, calcd 399.7755.
4,6,7-Tribromo-1H-benzotriazole (3h). A suspension of 3g
(400 mg, 1 mmol) in 0.666 mL (5.6 mmol) of quinoline was
heated until the temperature reached 210-220 °C and kept at
this temperature for 10-15 min, cooled, and 30 mL of hexane
was added. The resulting precipitate was collected and washed
several times with hexane to remove residual quinoline. Re-
crystallization from EtOH/H2O gave 242 mg (68%) of pure 3h:
mp 239.2-241.9 °C, Rf (A) 0.55, (B) 0.75. UV λmax (ε): (pH 2)
279 nm (11000); (pH 7) 289 nm (12730); (pH 12) 289 nm
Rf (A): 0.22. UV λmax (ε): (pH 2) 258.5 nm (3550), 274.5 nm
(2970); (pH 7) 302 nm (3910); (pH 12) 276.5 nm (3270), 303.5
1
nm (3400). H NMR δ: 10.0 (br s, NH2), 8.02 (dd, 1H, J )
0.5, J ) 8.8 Hz), 7.89 (s, 1H), 7.39 (dd, 1H, J ) 1.9, J ) 8.8).
13C NMR δ: 138.06, 131.04, 120.69, 116.81, 107.71 MS for
C6H7N4 [M + H]+: found 135.0768, calcd 135.0671.
5-Fluoro-4,6,7-tribromo-1H-benzotriazole (3a). To a solution
of 2a (137 mg, 1 mmol) in 1.6 mL of 70% HNO3 was added
0.221 mL (4.3 mmol) of Br2, and the solution was heated at
130 °C with stirring for 15 h. The resulting crystalline product
was collected on a filter and washed with HNO3 and H2O.
Recrystallization from EtOH give 82 mg (22% yield) of 3a:
mp 239.3-242.5 °C, Rf (A) 0.41, (B) 0.73. UV λmax (ε): (pH 2)
280.5 nm (12840); (pH 7) 291 nm (13585); (pH 12) 292 nm
(13570). 13C NMR δ: 154.22 (J ) 243.62 Hz), 138.16, 137.77,
112.6 (J ) 27.6 Hz), 112.28, 93.53 (J ) 28.9 Hz). MS for
C6H2Br3N3F [M + H]+: found 373.7695, calcd 373.7762.
5-Trifluoromethyl-4,6,7-tribromo-1H-benzotriazole (3c). To
a solution of 2c (187 mg, 1 mmol) in 7 mL of 70% HNO3 was
added 0.410 mL (8 mmol) of Br2, and the solution was heated
under reflux with stirring for 64 h. The solution was cooled to
room temperature, and the resulting crystalline product collected
on a filter and washed with HNO3 and H2O. Purification by

10610 J. Phys. Chem. B, Vol. 114, No. 32, 2010
Wa˛sik et al.
(12760). 1H NMR δ: 8.06 (s, 1H). 13C NMR δ: 140.61, 139.61,
132.72, 122.89, 110.09, 108.84. MS for C6H2Br3N3 [M + H]+:
found 355.8037, calcd 355.7857.
solute-solvent interaction was estimated using the polarizable
continuum model (PCM) approach,43 including the electrostatic
partition of solute interaction with the apparent surface charge
distribution (Gelec), dispersion (Gdisp) and repulsion (Grep)
contributions to the solvation free energy,44 and the Claverie-
Pierotti cavitation energy45 (Ecav, cf. ref 46 for the most recent
review). The zero-point energies (Ezp) and their thermal cor-
rections for translational (Gtrans), rotational (Grot), and vibrational
(Gvib), calculated using ideal gas rigid rotor and harmonic normal
mode approximations, were used to convert the internal energies
to the Gibbs free energies at 298.15 K.
In principle, changes of free energy were analyzed for three
different processes: (a) the protomeric equilibria, ∆G1, ∆G2,
∆G3, respectively, for the forms presented in Scheme 4, (b)
dissociation of the triazole N-H proton (∆Gdiss), and (c)
solvation (∆Gsolv) according to the following formulas:
5-Amino-4,6,7-tribromo-1H-benzotriazole (3i). A mixture of
3d (401 mg, 1 mmol) and Fe powder (298 mg, 5.34 mmol) in
AcOH (20 mL) was stirred for 6 h at room temperature, then
for 18 h at 110 °C, cooled to room temperature, and 279 mg (5
mmol) of Fe powder was added, and stirring was continued for
18 h at 110 °C. The reaction mixture was cooled, diluted with
40 mL of EtOAc, filtered, and the resulting solid was washed
with EtOAc. The filtrate and washings were combined and
washed with H2O. The organic layer was separated, dried over
Na2SO4, filtered, and volatiles were removed in vacuo. Purifica-
tion of the crude product by silica gel column chromatography
(0-15% MeOH/CHCl3) gave 274 mg (74%) of 3i: mp (dec.)
249.6-251.9 °C, Rf (B) 0.62. UV λmax (ε): (pH 2) 290 nm
(5860), 320 nm (6440); (pH 7) 294 nm (7430), 320 (6350);
(pH 12) 294 nm (8190), 320 nm (6620). 13C NMR δ: 142.93,
137.04, 136.87, 113.56, 111.65, 84.61. MS for C6H4Br3N4 [M
+ H]+: found 370.8145, calcd 370.7966.
Enzyme and Activity Measurements. The cDNA of CK2R
was a kind gift of Prof. Lorenzo A. Pinna (Padua University).
Protein was expressed in Escherichia coli and then purified as
described elsewhere.35 Protein kinase activity was determined
in a medium (50 µL total volume) containing CK2R (0,1 µg),
casein (75 µg), γ-[32P]ATP (10 µM icpm/pmol), Tris buffer pH
7.5 (50 mM), MgCl2 (10 mM), and appropriate concentrations
of inhibitor in 1 µL of DMSO. After 20 min incubation at 30
°C, the reaction was terminated by spotting 40 µL of the reaction
mixture onto a 3MM filter strip and washing five times for 10
min with 5% TCA. Incorporation of 32P was measured in an
LKB Flexi-Vial scintillation counter (activities are expressed
against controls with DMSO).
Gsolv ) Eelec + Grep + Gdisp + Ecav + Ezp + Gtrans + Grot + Gvib
∆Gi ) Gsolv(state i) - min (Gsolv(state j))
j)1,2,3
∆Gdiss ) Gsolv(monoanion) - Gsolv(neutral) + Ghydr(H+)
∆Gsolv ) Gsolv - G0
(3)
where G0 is the QM-derived Gibbs free energy in vacuo. It
should be noted that both carboxy derivatives carry the COO-
group, hence ∆Gdiss was calculated as the difference in free
energy of the doubly ionized molecule and its monoanionic
form. The proton hydration free energy, Ghydr(H+), was taken
as -262.3 kcal/mol.47
The molecular surface, its contribution from polar and apolar
atoms, their changes upon binding to CK2R, and the volume
for each compound were calculated according to the algorithm
of tessellation implemented in the GEPOL package.48
Simulated Annealing Procedure. The 141 residue fragment
(Tyr39-Ala179) of the CK2R subunit of the human CK2
holoenzyme [pdb1JWH],49 comprising all of the amino acid
residues proximal to the postulated binding site for TBBt and
its derivatives, was used as template of the protein target.
Structural calculations were performed with the aid of the
simulated annealing (SA) protocol implemented in X-PLOR.50
The original parallhdg force-field parametrization was extended
for the various derivatives, based on ab initio calculated
geometries and ESP charges.
Coordinates for Bt derivatives in the complex were also
adopted from the structure of TBBt bound to maize CK2R
[pdb1j91],16 and the initial structures were relaxed upon 1000
steps of energy minimization, with all protein backbone atoms
kept fixed, while other protein atoms were allowed to move.
This was followed by the SA procedure, comprising 6 ps
evolution at 3000 K, 3 ps cooling down to 300 K, and final
energy minimization. To preserve the general protein fold, all
backbone residues located outside a 6 Å sphere around the TBBt
binding site were fixed in high-temperature cycles and released
upon final minimization. Finally, for each derivative, 150
random SA calculations were performed, and the 10 lowest
energy structures were subjected to further analysis.
CK2R activity was measured for the pure enzyme and in the
presence of 0.5, 1, 2, 5, 10, 20, 30, 40, and 50 µM ligand
solution. The additional ligand concentrations were tested for
less active compounds (75 and 100 µM for 5COOH-Br3Bt (3g),
5CF3-Bt (2c), 5F-Bt (2a), and 5NH2-Br3Bt (3i)), and the most
active TBBt was also tested in 0.2, 0.4, 0.6, and 0.8 µM
concentration. The dose-response data were analyzed according
to the three-parameter Hill-Slope logistic model in the form
A0
A(x) )
(2)
z
x
1 +
(
)
IC50
where the activity of the enzyme, A0, and ligand IC50 were
optimized, and the slope coefficient, z, was constrained to 1,
reproducing the expected form of a 1:1 complex. Independent
series of experiments were analyzed together, assuming for each
an individual enzyme activity, A0, and the same value of the
IC50 parameter.
Molecular Modeling of Inhibitor Binding. Ab Initio Calcula-
tions. The initial coordinates of Bt and TBBt were adopted from
the crystal structure of TBBt complexed with maize CK2R,16
and their C(5) derivatives were built with the aid of homemade
software. This was followed by ab initio analysis with the aid
of the GAMESS 6.0 program.40 The DFT calculations were
performed with the B3LYP functional,41 using the 6-31G(d,p)
basis set,42 previously found applicable in the analysis of
symmetrical Bt systems.26 Geometry preoptimizations were done
in vacuo, and ESP charges were calculated to mimic the
electrostatic properties of the molecules. Correction for the
Structural analysis was performed with the aid of MolMol.51
The affinity of an inhibitor for CK2R was judged in terms of
preferred complex organization, defined as the rmsd of a
derivative, calculated when the protein backbone atoms were
superimposed. Additionally, according the concept of atomic
solvation parameters (ASP),52 the average solvent-accessible area
and its reduction upon binding to enzyme were estimated for

5-Substituted TBBt Inhibitors of CK2R
J. Phys. Chem. B, Vol. 114, No. 32, 2010 10611
(23) Wiley, R. H.; Hussung, K. F. J. Am. Chem. Soc. 1957, 79, 4395.
(24) Kumar, J. S. D.; Ho, M. M.; Toyokuni, T. Tetrahedron Lett. 2001,
42, 5601.
(25) Zou, R.; Drach, J. C.; Townsend, L. B. J. Med. Chem. 1997, 40,
811.
(26) Poznan´ski, J.; Najda, A.; Bretner, M.; Shugar, D. J. Phys. Chem.
A 2007, 111, 6501.
(27) Lum, K.; Chandler, D.; Weeks, J. D. J. Phys. Chem. B 1999, 103,
4570.
(28) Juffer, A. H.; Eisenhaber, F.; Hubbard, S. J.; Walther, D.; Argos,
P. Protein Sci. 1995, 4, 2499.
each ligand as a measure of hydration-related free energy
partition in the binding process.
CompoundLipophilicity.Sincetheexperimentalwater-octanol
partition coefficients, log P, were accessible only for two
compounds, Bt and 5-NO2-Bt (2d),53 they were estimated as
the consensus values of the results obtained from the various
methods with the aid of the AlogPS 2.1 program54 from the
Virtual Computational Chemistry Laboratory server.55 In paral-
lel, the theoretical coefficients describing compound solubility,
log S, were also obtained.
(29) Patani, G. A.; LaVoie, E. J. Chem. ReV. 1996, 96, 3147.
(30) Pearson, K. Philos. Magn. 2 1901, 6, 559.
(31) Hotelling, H. J. Educ. Psychol. 1933, 24, 417–498.
(32) Scherrer, R. A.; Howard, S. M. J. Med. Chem. 1977, 20, 53.
(33) Voth, A. R.; Ho, P. S. Curr. Top. Med. Chem. 2007, 7, 1336.
(34) Golub, A. G.; Yakovenko, O. Y.; Prykhodko, A. O.; Lukashov,
S. S.; Bdzhola, V. G.; Yarmoluk, S. M. Biochim. Biophys. Acta 2008, 1784,
143.
Supporting Information Available: Additional figures. This
material is available free of charge via the Internet at http://
pubs.acs.org.
(35) Sarno, S.; Vaglio, P.; Meggio, F.; Issinger, O. G.; Pinna, L. A.
J. Biol. Chem. 1996, 271, 10595.
References and Notes
(1) Unger, G. M.; Davis, A. T.; Slaton, J. W.; Ahmed, K. Curr. Cancer
Drug Targets 2004, 4, 77.
(2) Zien, P.; Duncan, J. S.; Skierski, J.; Bretner, M.; Litchfield, D. W.;
Shugar, D. Biochim. Biophys. Acta 2005, 1754, 271.
(3) Szyszka, R.; Boguszewska, A.; Shugar, D.; Grankowski, N. Acta
Biochim. Pol. 1996, 43, 389.
(4) Zien´, P.; Bretner, M.; Zastapiło, K.; Szyszka, R.; Shugar, D.
Biochem. Biophys. Res. Commun. 2003, 306, 129.
(5) Pagano, M. A.; Bain, J.; Kazimierczuk, Z.; Sarno, S.; Ruzzene,
M.; Di Maira, G.; Elliott, M.; Orzeszko, A.; Cozza, G.; Meggio, F.; Pinna,
L. A. Biochem. J. 2008, 415, 353.
(6) Sarno, S.; Reddy, H.; Meggio, F.; Ruzzene, M.; Davies, S. P.;
Donella-Deana, A.; Shugar, D.; Pinna, L. A. FEBS Lett. 2001, 496, 44.
(7) Auffinger, P.; Hays, F. A.; Westhof, E.; Ho, P. S. Proc. Natl. Acad.
Sci. U.S.A. 2004, 101, 16789.
(8) Politzer, P.; Murray, J. S.; Concha, M. C. J. Mol. Model. 2007, 13,
643.
(9) Kraut, D. A.; Churchill, M. J.; Dawson, P. E.; Herschlag, D. ACS
Chem. Biol. 2009, 4, 269.
(10) Voth, A. R.; Hays, F. A.; Ho, P. S. Proc. Natl. Acad. Sci. U.S.A.
2007, 104, 6188.
(11) Liu, L.; Baase, W. A.; Matthews, B. W. J. Mol. Biol. 2009, 385,
595.
(12) Eckenhoff, R. G.; Johansson, J. S. Pharmacol. ReV. 1997, 49, 343.
(13) Liu, R.; Loll, P. J.; Eckenhoff, R. G. FASEB J. 2005, 19, 567.
(14) Pop, S. M.; Gupta, N.; Raza, A. S.; Ragsdale, S. W. J. Biol. Chem.
2006, 281, 26382.
(15) Memic, A.; Spaller, M. R. ChemBioChem 2008, 9, 2793.
(16) Battistutta, R.; De Moliner, E.; Sarno, S.; Zanotti, G.; Pinna, L. A.
Protein Sci. 2001, 10, 2200.
(17) De Moliner, E.; Brown, N. R.; Johnson, L. N. Eur. J. Biochem.
2003, 270, 1.
(18) Battistutta, R.; Mazzorana, M.; Cendron, L.; Bortolato, A.; Sarno,
S.; Kazimierczuk, Z.; Zanotti, G.; Moro, S.; Pinna, L. A. ChemBioChem
2007, 8, 1804.
(19) Battistutta, R.; Mazzorana, M.; Sarno, S.; Kazimierczuk, Z.; Zanotti,
G.; Pinna, L. A. Chem. Biol. 2005, 12, 1211.
(20) Retegan, M.; Milet, A.; Jamet, H. J. Chem. Inf. Model. 2009, 49,
963.
(21) Rangarajan, M.; Kim, J. S.; Sim, S.-P.; Liu, A.; Liu, L. F.; La Voie,
E. J. Med. Chem. 2000, 8, 2591.
(22) Gershon, H.; Clarke, D. D.; Gershon, M. Monatsh. Chem. 1994,
125, 723.
(36) Burton, D. E.; Lambie, A. J.; Lane, D. W. J.; Newbold, G. T.
J. Chem. Soc (C) 1968, 1268.
(37) Phillips, M. A. J. Chem. Soc. 1931, 1143.
(38) McHugh, C. J.; Tackley, D. R.; Graham, D. Heterocycles 2002,
57, 1461.
(39) Miller, N. L.; Wagner, E. C. J. Am. Chem. Soc. 1954, 76, 1847.
(40) Schmidt, M. W.; Baldridge, K. K.; Boatz, J. A.; Elbert, S. T.;
Gordon, M. S.; Jensen, J. J.; Koseki, S.; Matsunaga, N.; Nguyen, K. A.;
Su, S.; Windus, T. L.; Dupuis, M.; Montgomery, J. A. J. Comput. Chem.
1993, 14, 1347.
(41) Hertwig, R. H.; Koch, W. Chem. Phys. Lett. 1997, 268, 345.
(42) Francl, M. M.; Pietro, W. J.; Hehre, W. J.; Binkley, J. S.; Gordon,
M. S.; DeFrees, D. J.; Pople, J. A. J. Chem. Phys. 1982, 77, 3654.
(43) Cammi, R.; Tomasi, J. J. Comput. Chem. 1995, 16, 1449.
(44) Floris, F.; Tomasi, J. J. Comput. Chem. 1989, 10, 616.
(45) Pierotti, R. A. Chem. ReV. 1976, 76, 717.
(46) Tomasi, J.; Mennucci, B.; Cammi, R. Chem. ReV. 2005, 105, 2999.
(47) Tawa, G. J.; Topol, I. A.; Burt, S. K.; Caldwell, R. A.; Rashin,
A. A. J. Chem. Phys. 1998, 109, 4852.
(48) Pascual-Ahuir, J. L.; Silla, E.; Tomasi, J.; Bonaccorsi, R. J. Comput.
Chem. 1987, 8, 778.
(49) Niefind, K. B.; Guerra, B.; Ermakowa, I.; Issinger, O. G. EMBO
J. 2001, 20, 5320.
(50) (a) Nilges, M.; Clore, G. M.; Gronenborn, A. M. FEBS Lett. 1988,
239, 129. (b) Nilges, M.; Kuszewski, J. ; Brunger, A. T. In Computational
Aspects of the Study of Biological Macromolecules by NMR; Hoch, J. H.,
Ed.; Plenum Press: New York, 1991.
(51) Koradi, R.; Billeter, M.; Wu¨thrich, K. J. Mol. Graphics 1996, 14,
51.
(52) (a) Juffer, A. H.; Eisenhaber, F.; Hubbard, S. J.; Walther, D.; Argos,
P. Protein Sci. 1995, 4, 2499. (b) Wang, J. M.; Wang, W.; Huo, S. H.;
Lee, M.; Kollman, P. A. J. Phys. Chem. B 2001, 105, 5055.
(53) Hansch, C.; Leo, A.; Hoekman, D. H. Exploring QSAR: Hydro-
phobic, Electronic, and Steric Constants; American Chemical Society:
Washington, DC, 1995; pp 219-304.
(54) Tetko, I. V.; Tanchuk, V. Y. J. Chem. Inf. Comput. Sci. 2002, 42,
1136.
(55) Tetko, I. V.; Gasteiger, J.; Todeschini, R.; Mauri, A.; Livingstone,
D.; Ertl, P.; Palyulin, V. A.; Radchenko, E. V.; Zefirov, N. S.; Makarenko,
A. S.; Tanchuk, V. Y.; Prokopenko, V. V. J. Comput.-Aided Mol. Des.
2005, 19, 453.
JP102848Y