Sulfonamido carboranes as highly selective inhibitors of cancer-specific carbonic anhydrase IX

Article abstract of DOI:10.1016/j.ejmech.2020.112460

Carbonic anhydrase IX (CA IX) is a transmembrane enzyme overexpressed in hypoxic tumors, where it plays an important role in tumor progression. Specific CA IX inhibitors potentially could serve as anti-cancer drugs. We designed a series of sulfonamide inhibitors containing carborane clusters based on prior structural knowledge of carborane binding into the enzyme active site. Two types of carborane clusters, 12-vertex dicarba-closo-dodecaborane and 11-vertex 7,8-dicarba-nido-undecaborate (dicarbollide), were connected to a sulfonamide moiety via aliphatic linkers of varying lengths (1–4 carbon atoms; n = 1–4). In vitro testing of CA inhibitory potencies revealed that the optimal linker length for selective inhibition of CA IX was n = 3. A 1-sulfamidopropyl-1,2-dicarba-closo-dodecaborane (3) emerged as the strongest CA IX inhibitor from this series, with a K_i value of 0.5 nM and roughly 1230-fold selectivity towards CA IX over CA II. X-ray studies of 3 yielded structural insights into their binding modes within the CA IX active site. Compound 3 exhibited moderate cytotoxicity against cancer cell lines and primary cell lines in 2D cultures. Cytotoxicity towards multicellular spheroids was also observed. Moreover, 3 significantly lowered the amount of CA IX on the cell surface both in 2D cultures and spheroids and facilitated penetration of doxorubicin. Although 3 had only a moderate effect on tumor size in mice, we observed favorable ADME properties and pharmacokinetics in mice, and preferential presence in brain over serum.

Full text of DOI:10.1016/j.ejmech.2020.112460

Journal Pre-proof
Sulfonamido carboranes as highly selective inhibitors of cancer-specific carbonic
anhydrase IX
Jana Štěpánková, Michael Kugler, Josef Holub, Václav Šícha, Viswanath Das, Jan
Nekvinda, Suzan El Anwar, Miroslav Havránek, Klára Pospíšilová, Milan Fábry,
Vlastimil Král, Martina Medvedíková, Stanislava Matějková, Barbora Lišková,
Soňa Gurská, Petr Džubák, Jiří Brynda, Marián Hajdúch, Bohumír Grüner, Pavlína
Řezáčová
PII:
S0223-5234(20)30431-1
DOI:
https://doi.org/10.1016/j.ejmech.2020.112460
EJMECH 112460
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 1 April 2020
Revised Date: 8 May 2020
Accepted Date: 11 May 2020
Please cite this article as: J. Štěpánková, M. Kugler, J. Holub, Vá. Šícha, V. Das, J. Nekvinda, S. El
Anwar, M. Havránek, Klá. Pospíšilová, M. Fábry, V. Král, M. Medvedíková, S. Matějková, B. Lišková,
Soň. Gurská, P. Džubák, Jiří. Brynda, Mariá. Hajdúch, Bohumí. Grüner, Pavlí. Řezáčová, Sulfonamido
carboranes as highly selective inhibitors of cancer-specific carbonic anhydrase IX, European Journal of
Medicinal Chemistry (2020), doi: https://doi.org/10.1016/j.ejmech.2020.112460.
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.
© 2020 Published by Elsevier Masson SAS.

Ki = 0.5 nM

Sulfonamido carboranes as highly selective inhibitors of cancer-specific carbonic
anhydrase IX
Jana Štěpánková1,2,¥, Michael Kugler3,4¥, Josef Holub5¥, Václav Šícha5, Viswanath Das1,2, Jan
Nekvinda5,6, Suzan El Anwar5, Miroslav Havránek7, Klára Pospíšilová4, Milan Fábry3, Vlastimil
Král3, Martina Medvedíková1, Stanislava Matějková4, Barbora Lišková1, Soňa Gurská1, Petr
Džubák1,2, Jiří Brynda3*, Marián Hajdúch1,2,*, Bohumír Grüner5*, Pavlína Řezáčová4*
Affiliations
1
Institute of Molecular and Translational Medicine, Olomouc, Hněvotínská 1333/5, 77900
Olomouc, Czech Republic
2 Cancer Research Czech Republic, Hněvotínská 5, 77900 Olomouc, Czech Republic
3 Institute of Molecular Genetics of the Czech Academy of Sciences, Flemingovo nam. 2, 16610
Prague, Czech Republic
4
Institute of Organic Chemistry and Biochemistry of the Czech Academy of Sciences,
Flemingovo nam. 2, 16610 Prague, Czech Republic
5
Institute of Inorganic Chemistry of the Czech Academy of Sciences, 250 68 Řež, Czech
Republic
6
Department of Organic Chemistry, Faculty of Natural Science, Charles University, Hlavova
2030, 12800 Prague 2, Czech Republic.
7 Apigenex, s.r.o. Poděbradská 59/ 186, 180 66 Prague 9, Czech Republic.
Corresponding Authors (*)
Dr. P. Řezáčová, E-mail: pavlina.rezacova@uochb.cas.cz, Phone: +420 220 183 144
Dr. B. Grüner, E-mail: gruner@iic.cas.cz, Phone: +420 311 236 942
Dr. M. Hajdúch, E-mail: marian.hajduch@upol.cz, Phone: +420 585 632 082
Dr. J. Brynda, E-mail: brynda@img.cas.cz, Phone: +420 220 183 210
(¥) These authors contributed equally.
Declaration of interest
The authors declare the following competing financial interest(s): B. G., J. B., V. S., J. H., P. D.,
M. H., and P. R. are inventors of a United States Patent, Pat. No. 9,290,529 B2, issued on Mar.
22, 2016, that covers the title compounds.

Abstract
Carbonic anhydrase IX (CA IX) is a transmembrane enzyme overexpressed in hypoxic
tumors, where it plays an important role in tumor progression. Specific CA IX inhibitors
potentially could serve as anti-cancer drugs. We designed a series of sulfonamide inhibitors
containing carborane clusters based on prior structural knowledge of carborane binding into the
enzyme active site. Two types of carborane clusters, 12-vertex dicarba-closo-dodecaborane and
11-vertex 7,8-dicarba-nido-undecaborate (dicarbollide), were connected to a sulfonamide moiety
via aliphatic linkers of varying lengths (1 to 4 carbon atoms; n = 1-4). In vitro testing of CA
inhibitory potencies revealed that the optimal linker length for selective inhibition of CA IX was
n=3. A 1-sulfamidopropyl-1,2-dicarba-closo-dodecaborane (3) emerged as the strongest CA IX
inhibitor from this series, with a Ki value of 0.5 nM and roughly 1,230-fold selectivity towards
CA IX over CA II. X-ray studies of 3 yielded structural insights into their binding modes within
the CA IX active site. Compound 3 exhibited moderate cytotoxicity against cancer cell lines and
primary cell lines in 2D cultures. Cytotoxicity towards multicellular spheroids was also
observed. Moreover, 3 significantly lowered the amount of CA IX on the cell surface both in 2D
cultures and spheroids and facilitated penetration of doxorubicin. Although 3 had only a
moderate effect on tumor size in mice, we observed favorable ADME properties and
pharmacokinetics in mice, and preferential presence in brain over serum.
Keywords
anti-tumor agents; carbonic anhydrase IX; carboranes; dicarbollide; enzyme inhibitors; drug
penetration; multicellular spheroids

Introduction
Hypoxia is a major contributor to tumor development and progression [1]. Growing tumors have
an insufficient supply of oxygen, which leads to changes in cell metabolism such that glycolysis
becomes the preferred pathway to metabolize glucose. This phenomenon, known as the Warburg
effect [2], results in acidification of the extracellular environment, which promotes tumor cell
invasiveness and increases resistance to chemotherapy and radiotherapy [3], [4]. Hypoxia
stimulates overexpression of many proteins critical for cancer cell survival and metastases.
Carbonic anhydrase IX (CA IX) is one of the key proteins controlled by hypoxia-induced
transcription factor (HIF). This membrane-bound enzyme catalyzes reversible hydration of
carbon dioxide to bicarbonate in the extracellular space and plays an important role in tumor
progression [5-7]. While physiological expression of CA IX is limited to the gastrointestinal tract
[8], overexpression of CA IX has been reported in a broad spectrum of malignancies. CA IX
overexpression indicates poor prognosis in lung, breast, cervical, brain, renal and neck cancers
[9-14]. CA IX thus represents a clinically relevant biomarker and a target for development of
anticancer drugs.
CA IX belongs to large group of human carbonic anhydrases, are ubiquitous zinc
metalloenzymes with different subcellular localizations and tissue expression profiles. In
addition to CA IX, several other isoforms are also recognized as therapeutic targets for the
treatment of various diseases [15]. Development of inhibitors that selectively target CA IX has
been a subject of intensive investigation [16], [17] [18]. Sulfonamides have shown potential as
anticancer agents in preclinical experiments and clinical trials [19], [20]. A current challenge in
the development of therapeutic agents targeting CA IX is designing compounds that specifically
inhibit this isoenzyme and not the 11 other catalytically active human CA isoforms [21].
The isomeric dicarba-closo-dodecaboranes (with broadly used trivial name carboranes)
are a newly emerging three-dimensional pharmacophores, providing hydrophobic interactions
between biologically active molecules coupled to the boron cluster and their receptors [22].
Carboranes not only act as a space-filling moiety but also can increase a compound’s interaction
energy, in vivo stability and bioavailability [23, 24]. The use of carboranes as components of
pharmacologically relevant molecules has rapidly been increasing [22-25, 26-28]. Previously, we
introduced icosahedral dicarba-closo-dodecaborane, 11-vertex dicarba-nido-undecaborate
(dicarbollide) and cobalt bis(dicarbollide) scaffolds into structures of sulfamide-type compounds

[29-31]. These cluster compounds exhibited inhibitory activity against CA IX in the micromolar
to subnanomolar concentration range and showed promising selectivity [29-31]. Structural
information about the active-site binding mode of these carborane inhibitors determined by X-
ray diffraction aided further inhibitor design, leading to the compounds presented in this work.
Experiments with an organic series of inhibitors revealed that a sulfonamide head group
can bind to the zinc atom present in the active center of the enzyme with higher interaction
energy than a sulfamide [32]. Therefore, when considering further improvements in the
structures of active carborane species, we replaced the sulfamide group [29] with a sulfonamide.
Here, we present the synthetic procedure based on insertion of a newly prepared family of
organic alkyne-1-sulfonamides into the opencage decaborane derivative, 6,9-(Me2S)2-arachno-
B10H12. This yielded a series of 1,2-dicarba-closo-dodecaboranes substituted at the cage with
alkylsulfonamide functions and enabled us to finely tune the length of the alkyl linker connecting
both moieties. Also a series of 7,8-dicarba-nido-undecaborates was prepared in parallel by
degradation of the cage of the respective 1,2-dicarba-closo-dodecaboranedodecaborane
derivatives.
The aim of this work was to evaluate closo- and nido- dicarbaboranescarboranes
substituted at the cage with alkylsulfonamide function as specific inhibitors of CA IX in vitro, in
cellulo and in vivo. Our results show that the presence of alkylsulfonamide group led to
significant improvements in in vitro inhibitory activity compared to a series of similar sulfamide
compounds [30]. We also confirmed that the new compounds act as potent and specific
inhibitors of CA IX in vivo. The selected lead compound of formula 1-H2NSO2C3H6-1,2-
C2B10H11 (3) exhibited a tumor-specific growth inhibitory effect in both two-dimensional (2D)
and multicellular spheroid (MCS) cultures of CA IX-positive cell lines. Compound 3also showed
favorable pharmacologic properties, as demonstrated using in vitro absorption, distribution,
metabolism and excretion (ADME) assays and in vivo pharmacokinetics methods. In mouse
models, this compound displayed anti-cancer effect in both syngenic breast (4T1-12B) and
human xenografted HT-29 colorectal tumors.

Results
Compound design and synthesis
General pathways to substituted 1,2-dicarba-closo-dodecaboranes usually involve
incorporation of substituted alkynes into ten-vertex open-cage precursors [typically bis(ligand)
derivatives of nido-decaborane(14), i.e. 6,9-(CH3CN)2- arachno-B10H12 or 6,9-(Me2S)2- arachno-
B10H12 or lithiation of the CH groups in the 1,2-dicarba-closo-dodecaboranecage followed by
reaction with numerous reagents [22] [25]. To our knowledge, no 1,2-dicarba-closo-
dodecaborane derivatives with sulfonamide groups had been prepared to date. The still-limited
availability of alkynes bearing terminal sulfonamide groups for insertion reactions, as well as the
absence of reaction pathways for direct reactions, may account for the absence of such
derivatives. The presence of polar sulfonamide groups in the starting reagents could in principle
lead to degradations of the cage, producing open-cage products or interfering with the terminal
H2N- group using lithiated carborane. Surprisingly, insertion of alkyne-1-sulfonamides of the
general formulation H2NSO2(CH2)nC≡C (n=1-4) into the open-cage, 10-vertex 6,9-(Me2S)2-
arachno-B10H12 derivative proceeded smoothly. The reactions, carried out in toluene at high
temperature, resulted in high yield conversion, and the respective 1-alkylsulfonamide-1,2dicarba-
closo-dodecaboranesdicarbaboranes of general formula 1-H2NSO2(CH2)n-1,2-C2B10H11 (1-4)
were isolated in good yields after simple work-up. This consisted of extraction of crude products
into diethyl ether, degradation of minor open-cage products with methanol under slightly acidic
conditions, and final purification by chromatography (see Scheme 1; for details, see
Experimental or Supporting Information). A key factor in the successful synthesis of these
carboranes is the availability of alkyne-1-sulfonamides. Synthesis of these compounds from
suitable building blocks, such as sodium prop-2-yne-1-sulfonate and iodo-alk-1-yne, was enabled
by recent progress in chemistry of acetylenes. These precursors were converted in several
reaction steps to acetylenes with terminal sulfonamide groups (for full experimental details, see
Supporting Information).
[Scheme 1, color print]

Scheme 1: Overall reaction scheme leading to neutral 1-H2NSO2(CH2)n-1,2-dicarba-closo-
dodecaboranes and 7-H2NSO2(CH2)n-7,8-dicarba-nido-undecaborate ions.
O
NH2
S
O
(H C)
n
2
1
SMe2
Me2S
9
4
6
2
6
2
5
9
8
7
HCC-(CH2)n-S(O)2NH2
Tol. reflux
3
4
10
5
11
10
3
1
7
8
12
[arachno-6,9-(Me2S)2-B10H12
1, n= 1
2, n= 2
3, n= 3
4, n= 4
-
O
H2N
H2
C
H
10
9
S
7
n
11
8
O
6
5
5-, n= 1
4
2
3
1
6-, n= 2
7-, n= 3
8-, n= 4
The chemical stability of the alkyl sulfonamide group in the 1,2-dicarba-closo-
dodecaboranedicarbaborane derivatives resembles that of organic sulfonamide compounds. The
terminal group withstands attacks by strongly acidic or basic reagents with no noticeable
decomposition. This feature was employed to produce a series of substituted 11-vertex 7,8-
dicarba-nido-undecaborate ions of the general formulation [7-H2NSO2(CH2)n-7,8-C2B9H11]- (n =
1-4, 5-8-) by degradation of the cage with excess of potassium hydroxide in methanol. The
degradation proceeds far more easily than in the case of the parent 1,2-dicarba-closo-
dodecaborane cage and can be accomplished even in the presence of water in methanol (typically
used for quenching of these reactions) or with aqueous alkaline solutions. For example,
degradation of dicarbaborane 4 to the respective potassium salt of 5- could be performed with 6%
o
aqueous potassium hydroxide at 60 C. Experimentally observed rates of degradation increased
with the length of the linker. The products were isolated by dilution of the reaction mixtures with
water, evaporation of methanol, washing of the resulting aqueous solution with diethyl ether, and
extraction of the products into ethyl acetate. Alternatively, the 1,2-dicarba-closo-dodecaborane
cage in these species could be also degraded by treatment with triethylamine in methanol or by
potassium or cesium fluoride in methanol, following the procedures described in the literature

[33]. Notably, no derivatives of 7,8-dicarba-nido-7,8-undecaborate ion were observed (by LC-
MS) in mouse serum after administration of substituted 1,2-dicarba-closo-
dodecaboranesdodecaborane (2 and 3) for pharmacokinetics studies. Substitution with terminal
alkylsulfonamide groups appreciably increased aqueous solubility compared to the parent series
of of 7,8-dicarba-nido-undecaborate(1-) ions. For biological assays, compounds were used in the
form of potassium salts from synthesis or converted to sodium salts by metathesis.
Inhibition of CA activity
Candidate inhibitors were tested in vitro using a stopped-flow carbon dioxide hydration
assay. To assess selectivity, we compared inhibition of two CA isoforms, the cancer-associated
CA IX and the widespread CA II (Table 1, Figures 1, S1 and S2).
Table 1. In vitro inhibition of selected carbonic anhydrase isoenzymes.
Compound
Ki (CA II)
[nM]
Ki (CA IX)
[nM]
Selectivity
Index[a]
0.44
1
2
494.4 ± 113.3
1136.0 ± 140.7
0.92
1229.3
238.4
9.6
21.0 ± 2.3
622.0 ± 177.4
870.1 ± 91.04
63007 ± 28409
60.69 ± 22.36
1546 ± 385.9
333 ± 36.59
22.8 ± 3.4
0.506 ± 0.11
3.65 ± 0.64
6545 ± 1625
45.16 ± 5.73
1.178 ± 0.12
1.605 ± 0.38
3
4
5-
6-
7-
8-
1.34
1312.4
207.5
[a] Selectivity index is the ratio between Ki (CA II) and Ki (CA IX)
Inhibitory constants (Ki values) against CA IX ranged from micromolar to low nanomolar.
The selectivity for CA IX over CA II also differed substantially among compounds. Most
compounds were selective toward CA IX, with the exception of 1 and 2. Compounds with long
alkyl linkers connecting the sulfonamide moiety to the cluster showed high inhibitory activity
against CA IX (with Ki values in the low nanomolar range). The optimal linker length for CA IX

affinity was n= 3; compounds 3 and 7- inhibited CA IX with subnanomolar or nanomolar Ki
values and selectivity indices over 1,000. Compounds with n= 4 linker (4 and 8-) retained
nanomolar inhibitory potency toward CA IX but lost selectivity. In contrast, the optimal length
for CA II affinity was n=2, although none of the compounds in our series was a low-nanomolar
inhibitor of CA II.
[Figure 1, grayscale print]
Figure 1. Inhibition of CA II (A) and CA IX (B) by rising concentration of compounds 3
and 7-. Plots show the dose dependent decrease of CA catalytic activity of given. vi is initial
reaction rate (reaction rate is determined as 1st derivative of the absorbance with respect to time),
each dot represents one measurement (some points are overlaid), black line shows fitted curves
by the nonlinear least-squares method using Williams-Morrison equation. Raw data of
absorbance change in the stopped-flow carbon dioxide hydration assay are shown in Figure S3.

Binding pose and interactions within the CA IX active site
The two compounds with the highest affinity and selectivity toward CA IX, 3 and 7-,
were selected for structural studies to explore their interactions with the CA IX active site.
Compounds were co-crystallized with a CA IX mimic - CA II containing seven amino acid
substitutions in the active site (A65S, N67Q, E69T, I91L, F130V, K169E, and L203A). This
variant is often used in structural studies because it retains the good crystallization properties of
CA II while the active site resembles that of CA IX [34]. The Ki values for inhibition of CA IX
mimic by 3 and 7- were 2.949 ± 0.48 nM and 17.25 ± 2.54 nM, respectively. These values are
comparable to Ki values determined for inhibition of CA IX and differ substantially from Ki
values for inhibition of CA II, indicating that CA IX mimic is indeed applicable to explore
interactions of 3 and 7- with the CA IX active site. Two crystal structures determined at atomic
resolutions of 1.1-1.2 Å confirmed specific binding of 3 and 7- in the CA IX active site (Figure
2A, B, Figure S4).
The sulfonamide moiety is deeply buried in the active site, where it makes polar
interactions with the zinc ion and residues located at the bottom of the active site cavity (His94,
His96, and His119). This binding mode is similar to interactions with the enzyme active site
reported for organic sulfonamide-containing inhibitors [35-37]. In addition to interactions with
catalytic residues, 3 made 19 interatomic interactions with H64, Q92, V121, L198, T199-200 and
W209. Two additional contacts were observed for 7-, for a total of 21 interatomic interactions
with Q92, V121, V131, V135, L198 and T199-200. The hydrophobic carborane clusters of 3 and
7- interact with a hydrophobic pocket formed by L91, V121, V131, V135, L141 and L198.
The linker moiety of 7- adopts an extended conformation, with second carbon in the
aliphatic chain is in transposition to N atom of sulfonamide group. The second and third carbon
atoms in the aliphatic chain of 3 adopt a gauche conformation, and the second carbon atom in the
chain is in trans to the O2 atom of the sulfonamide group (Figure 2C). We speculate that the
extended conformation of 7- is caused by a polar interaction between the positive hydrogens of
the amino groups of Q67, Q92 and negative charge of 7,8-dicarba-nido-undecaborate(1-) cage
[38, 39].
[Figure 2, color print]

Figure 2. Structure of 3 (A) and 7- (B) bound to the CA IX active site. (A), (B) Compounds
are depicted in stick representation, and the protein is shown in green cartoon representation with
interacting residues highlighted as sticks and labelled. The Zn2+ ion is represented by a gray
sphere. Dashed lines indicate polar interactions. (C) Overlay of the binding modes of 3 and 7- in
the CA IX active site, which is represented by its solvent-accessible surface colored by
electrostatic potential (red for negative, blue for positive).
Cytotoxicity of 3 in tissue cultures and spheroids
We assessed the toxicity of 3 in the form of a sodium salt on a panel of 11 cancer cell lines and
two non-malignant fibroblast cell lines using a standard MTT assay under normoxic cell culture
conditions. Compound 3 showed moderate cytotoxicity across the cell line panel, with no
selectivity for malignant over non-malignant MRC-5 or BJ fibroblasts, possibly due to off-target
effects (Table 2).
Table 2. Cytotoxicity of 3 in a panel of 13 cell lines. IC50 (µM) values are presented as mean ±
SD of four independent experiments.

CCRF-
CEM
MDCK-
CAIX
MRC-5
BJ
CEM-DNR K562 K562-TAX A549 HCT116 HT-29 HeLa 4T1-luc
MDCK-NEO
89±11 44±32
32±7
58±4
50±4
60±5
72±5 45±17 57±6 56±4 77±6
82±18
92±12
As spheroids of cancer cells mimic many characteristics of solid tumors [40], we next
determined the effect of 3 on spheroids of HCT116 and HT-29 cells over a period of 3-7 days. In
parallel, we determined the cytotoxicity of 3 after 7 days of treatment under normoxic conditions
in 2D cultures of HCT116 and HT-29 cells by standard MTT assay. We observed a dose-
dependent decrease in the size of MCSs of both the cell lines; the fold change in IC50 values was
more pronounced in MCSs of HT-29 cells (Figure 3). Similar to the 3D culture effect, longer
treatment with 3 resulted in greater cytotoxicity in 2D cultures.
[Figure 3, color print]
Figure 3. Effect of 3 on spheroid size. (A) IC50 values of 3 in 2D and spheroid cultures of
HCT116 and HT-29 cells following 3-7 days of treatment under normoxic conditions. Data are

the mean ± SEM, n = 3-4 independent experiments. (B) Images showing the dose-dependent
effect of 3 in spheroids. Scale bar: 500 µm.

Inhibition of CA IX expression in hypoxic 2D cultures and spheroids of HT-29
Previously, we found that some carborane sulfamide compounds affect both the amount and
localization of intracellular CA IX. Specifically, CA IX was detected in high amounts in the
cytoplasm and nuclei following treatment of HT-29 cells with 1-sulfamidoethyl derivative of
cobalt bis(1,2-dicarbollide) ion [31].
To determine if 3 inhibits CA IX expression, 2D cultures of HT-29 cells were first grown
under 1% O2 for 4-5 days and then treated with different concentrations of 3 for 12 h in hypoxic
conditions. We observed dose-dependent inhibition of transmembrane expression of CA IX by 3
(Figure 4A).
MTT assay indicated that the cytotoxic effect of 3 was more pronounced under when cells were
treated for 72 h of treatment in hypoxia (Figure 4B). Although there was some cytotoxicity at
100 µM after 12 h, the decrease presence of cells in Fig. 4A after treatment with 50-100 µM 3
could be due to cell lifting from slides as CA IX downregulation has been reported to affect cell
attachment [41]. Nevertheless, nuclear shrinkage presumably indicates the induction of cell
death.
Similar to our findings in 2D cultures, treatment of HT-29 spheroids for 3 days with an IC50
concentration of 3 inhibited the transmembrane expression of CA IX (Figure 4C). Furthermore,
Western blot analysis of protein lysates from drug-treated spheroids indicated concentration-
dependent inhibition of CA IX by 3 (Figure 4D).
Redistribution of CA IX may affect the composition of the extracellular matrix via
modulation of matrix metalloproteinases activities [42] resulting in aberrant penetration of
cytotoxic drugs into tissues [43].
[Figure 4, color print]

Figure 4. Treatment with 3 decreases CA IX expression in HT-29 cells. (A) Images showing
CA IX expression in 2D cultures of HT-29 cells treated with 3 (100 µM) for 12 h under hypoxic
(1% O2) conditions. Scale bar: 10 µm. (B) MTT cell proliferation assay after treatment of 2D
cultures of HT-29 cells with the indicated concentrations of 3 for 12-72 h in hypoxia. Data are
mean ± SEM of 3 independent experiments, *** p < 0.001, ** p < 0.01, * p < 0.05 comparing
drug-treated to untreated (0 µM), one-way ANOVA with Dunnett's multiple comparison test; ɸɸ
p < 0.01 comparing 12 h to 72 h, unpaired t-test with Welch's correction. (C) Ten-day-old
spheroids of HT-29 cells treated with 3 (42 µM) under standard cell culture conditions for 72 h
show a decrease in CA IX expression. Scale bar: 50 µm. (D) Western blot showing a
concentration-dependent decrease in the expression of CA IX in HT-29 spheroids following 72 h
of treatment under standard conditions. Data are mean ± SEM of 3 independent experiments, ***
p < 0.001, ** p < 0.01 comparing treated to untreated, one-way ANOVA with Dunnett's multiple
comparisons test.

Compound 3 facilitates doxorubicin accumulation in spheroids
Our recent study showed that ionic sulfamide derivatives of cobalt bis(1,2-dicarbollide) causing
redistribution of CA IX also facilitated penetration of anti-cancer drugs such as doxorubicin
(DOX) into spheroids [31]. Therefore, we next evaluated the penetration and accumulation of
fluorescent DOX in spheroids following co-treatment with DOX and 3 for 140 min. Our live-
spheroid imaging data revealed that also neutral compound 3 significantly increases the time-
dependent penetration of DOX in co-treated spheroids compared with spheroids treated with
DOX only (Figure 5A, B). Analysis of spheroids at a depth of 50 µm z-plane height showed
increased accumulation of DOX in the interior of spheroids co-treated with DOX and 3 (Figure
5C). Tumor pH has been suggested to affect the intake of weakly basic anti-cancer drugs such as
DOX [44], [40]. Immunostaining of drug-treated spheroids showed a marked decrease in CA IX
level in the interior of spheroids at a depth of around 50 µm z-plane height (Figure 5D),
suggesting that 3-mediated inhibition of CA IX and potential alterations in pH facilitated DOX
penetration into spheroids. The accumulation of DOX could also result due to cell death by 3,
causing increased permeabilization of DOX into dying cells. Additionally, the effect of 3 on CA
IX may alter the composition of the extracellular matrix via modulation of matrix
metalloproteinases activities [45], facilitating DOX penetration. These potential mechanisms
need to be investigated in future studies.
[Figure 5, color print]

Figure 5. Concomitant treatment with 3 increases DOX accumulation in spheroids. (A)
Images of HT-29 spheroids at a depth of 50 µm z-plane height showing the accumulation of
DOX at the indicated time points in the absence and presence of 3. Scale bar: 50 µm, Objective:
20×. (B) Graph showing the rate of DOX accumulation in spheroids treated with DOX only
(blue) and DOX in the presence of 3 (red). Data are mean ± SEM of at least 4 spheroids per
treatment from 3 independent experiments, *** p < 0.001, ** p < 0.01 comparing DOX + 3 to
DOX only, two-way ANOVA with Sidak's multiple comparisons test. (C) Line-profile plot from
images from (A) showing an increased accumulation of DOX after 140 min of treatment in the
interior of spheroids treated with 3 compared with DOX-only-treated spheroids. (D)
Downregulation of CA IX expression in 3-treated spheroids at a depth of 50 µm z-plane height.
A 2.5D plot of spheroids shows a decrease in the CA IX expression in the interior of 3-treated
spheroids compared with controls. Scale bar: 20 µm, Objective: 60×.

Pharmacological properties of compound 3
ADME parameters obtained from in vitro and in vivo assays indicated that the 3 is stable in
plasma and binds more than 90% of plasma proteins (Table 3). The intrinsic clearance of 3 was
classified as medium (in microsomal stability assay), and the compound was metabolized by
microsomes with a half-time of 110 min. The transcellular permeation of 3 was classified as
medium compared with other commonly used drugs in a parallel artificial membrane
permeability assay (PAMPA) used as an in vitro model of passive diffusion. Furthermore, the
medium permeability of 3 in Caco-2 cells indicates relatively good human intestinal
permeability, suggesting that 3 would not be subject to drug efflux. Additionally, 3 showed a
positive threshold for the central nervous system (CNS) and did not permeate MDCK-MDR1
cells, indicating a low potential for penetration through the blood-brain barrier.
Table 3. Pharmacological properties of 3 determined by in vitro and in vivo ADME assays
Metabolism
Permeability
in vitro
in vitro
in vivo
MDR1-MDCK
PAMPA
Plasma protein
binding (%
bound)
Microsomal
stability
(clearance)
Caco-2
Plasma stability
(category)
Category
%recovery
CNS (-ive/+ive)
Category
Moderate
CNS +ive
Papp (x10-6):
10.3
Papp (x10-6): 10
Effluxratio: 0.2
Medium
log Pe:
-5.8
Clearance
category: Medium
Stable
93.3%
110.0
Effluxratio: 7.9
Active efflux: Yes Active efflux: No
% recovery: 38.8 % recovery: 82.4
Pharmacokinetics
The concentration-time profiles of 3 following intraperitoneal administration to NMRI mice at a
maximum tolerated dose (MTD) of 125 mg/kg and ½ MTD of 62.5 mg/kg were determined over
36 h in plasma and brain. Besides initial symptoms of piloerection, there were no visible signs of
drug toxicity on animal behavior, appearance, or body weight. Interestingly, we observed
preferential penetration of 3 into the brain, with a peak 1 h after administration and a half-life of

2 h. While the maximum concentration (Cmax) of boron in serum was 75.1 μM, the Cmax in brain
was 186 μM 1 h after administration of the MTD of 3 (Figure S5A, Table S2).
Brain penetration was not observed for previously studied family of metallacarborane
sulfamide compounds (Figure S5B, C). Thus, we set out to decipher whether the nature of the
cluster is crucial for penetration through the blood-brain barrier by assessing the
pharmacokinetics of three other compounds containing closo- or nido-dicarbaborane clusters (2,
6- and 7-). Compound 2 exhibited preferential brain penetration, to an even higher extent than 3.
The Cmax of 2 in brain was 695 μM at 1 h, while Cmax in serum was 462 μM (Figure S5D). The
half-life of 2 in brain was 12 h.
On the other hand, concentrations of 6- and 7- in the brain were
low or negligible, and these compounds were found mainly in serum (Figure S5E, F, Table S2).
For 6-, the Cmax in serum was 2,997 μM at 0.5 h, while Cmax in brain was only 4.5 μM.
Compound 7- reached a maximal serum concentration of 2,707 μM at 0.5 h, while Cmax in brain
was 36 μM. These findings suggest that the presence of a dicarba-closo-dodecaborane cluster is
essential for active transport into brain.
Antitumor activity of 3 in mouse models of mammary and colorectal tumors
We explored the in vivo anti-tumor activity of the sodium salt of 3 in two mouse models: (1)
BALB/c mice bearing syngeneic mammary tumors of 4T1-12B cells and (2) SCID mice
xenotransplanted with HT-29 cells (Figure 6). The effect of 3 on tumor size was evident in the
first 10 days and after 29-30 days in BALB/c mice. In SCID mice, the effect of 3 on tumor size
persisted throughout 24 days.
[Figure 6, grayscale print]

Figure 6. In vivo antitumor activity of 3. Graphs showing the effect of 3 on tumor size in
female (A) BALB/c mice orthotopically transplanted with 4T1-12B cells for 33 days and (B)
SCID mice bearing subcutaneously transplanted HT-29 cell tumors for 24 days. Mice were
administered 3 twice per day at a dose of 62.5 mg/kg. Dosing schedules are indicated by the
black solid lines on the x-axes. Data in (A) and (B) are mean ± SEM of typically 20 tumors
transplanted to 10 animals per group, *** p < 0.001, ** p < 0.01, * p < 0.05 comparing vehicle-
treated to drug-treated tumors, Student’s t-test, paired.
Conclusions
We developed highly potent inhibitors of human carbonic anhydrase by using a structure-assisted
design strategy and applying our previous knowledge of sulfamide compounds containing
icosahedral dicarbaborane, 11-vertex dicarba-nido-undecaborate(1-) [30] and cobalt
bis(dicarbollide)(1-) scaffolds [31]. By replacing the head group with alkylsulfonamide and
varying the linker between the head group and carborane clusters, we fine-tuned inhibitor
selectivity toward the cancer-specific CA IX isoform. These efforts led to an inhibitor with a
subnanomolar inhibition constant and high selectivity (Ki value of 0.5 nM and selectivity index
of 1,230 for CA IX over CA II). Atomic-resolution X-ray crystallographic structures uncovered
interactions in active site of CA IX that are crucial for this high affinity and selectivity.
Of the various compounds generated in this series, we identified 3 as a promising
candidate for further evaluation of activity in tissue cultures and spheroids. This compound
exhibited moderate cytotoxicity against cancer cell lines and primary cell lines in 2D cultures, as

well as MCSs. Moreover, 3 significantly lowered the amount of CA IX on the cell surface both
in 2D cultures and MCSs, facilitating the penetration of doxorubicin.
Compound 3 showed favorable ADME properties and pharmacokinetics in mice, CNS-
positive properties in MDCK-MDR1 permeability assay indicating a greater potential for uptake
across the blood-brain barrier and presence in the brain over serum. Although just moderate
effect on tumor size in two mouse models (syngeneic and xenotransplanted tumors) was
observed, we believe that compound 3 has a potential for future synergistic combinations to
combat human cancers.
In conclusion, 1-sulfonamidopropyl-1,2-dicarbadodecaboranes are highly selective
inhibitors of cancer-specific carbonic anhydrase IX in vitro as well as in vivo and show thus the
potential for further development of antitumor agents.
EXPERIMENTAL
Chemical syntheses. Nido-decaborane(14) (14) was purchased from Katchem Ltd., Czech
Republic. The arachno-6,9-(Me2S)2-arachno-B10H12 was prepared by reaction of decaborane in
dimethyl sulfide, as described in the literature [46]. Toluene was dried with sodium metal and
distilled prior to use. Acetonitrile was dried using 4 Å molecular sieves (Fluka). Other chemicals
were purchased from Aldrich. Solvents were obtained from Aldrich, Lachema a.s. and Penta
Ltd., Czech Republic, and used without further purification. Analytical TLC was carried out on
Silufol® (silica gel on aluminum foil, starch as binder, Kavalier, Czech Republic). Unless
otherwise specified, column chromatography was performed on high-purity silica gel (Merck
Grade, Type 7754, 70-230 mesh, 60 Å).
All reactions were performed with standard vacuum or inert-atmosphere techniques. Some
operations, including flash chromatography and crystallization, were carried out in air. Melting
points were determined in sealed capillaries on a BŰCHI Melting Point B-545 apparatus and are
uncorrected.
1
11
Instrumental techniques. H and B NMR spectroscopy was performed on a Varian Mercury
400Plus Instrument. The spectra of all compounds were measured immediately after dissolution.
NMR chemical shifts are given in ppm to high-frequency (low field) relative to F3B OEt2 as the

1
external reference. Residual solvent H resonances were used as internal secondary standards.
1
11
Coupling constants J(11B–1H) were measured by resolution-enhanced B spectra with a digital
resolution of 2 Hz and are given in Hz.
11
11
NMR data are presented in the following format: B NMR: B chemical shifts δ(11B)
(ppm), multiplicity, J(11B–1H) coupling constants in Hz. Peak assignment is based on {11B-11B}
COSY NMR spectroscopy and comparison with the spectrum of the parent ion 1 (for assignment
1
of the unsubstituted ligand). H NMR: chemical shifts δ(1H) are given in ppm, coupling
constants J(H,H) in Hz. δ (11B{11B}) data are also presented, assignment is based on selectively
decoupled δ (1H)-{11B selective}NMR experiments.
Mass spectrometry. MS measurements were performed on a Thermo-Finnigan LCQ-Fleet Ion
Trap instrument using Atmospheric Pressure Chemical Ionization of Electrospray Ionization
(ESI) techniques. Negative ions were detected. Samples dissolved in acetonitrile (concentrations
approx. 100 ng.ml–1) were introduced to the ion source by infusion (5 μL.min–1). Molecular ions
[M]- were detected for all univalent anions and [M-H]- for neutral compounds as the base peaks
in the spectra. Full agreement between the experimental and calculated isotopic distribution
patterns was observed for all compounds. HRMS were measured using an Agilent 6230 LC-
TOFMS spectrometer with ESI ionization. The isotopic distribution in the boron plot of all peaks
was in perfect agreement with the calculated spectral pattern. The data are presented for the most
abundant mass in the boron distribution plot (100%) and for the peak corresponding to the m/z
value.
Elemental analyses. Elemental analyses were performed on a Thermo Scientific FlashSmart
Organic Elemental Analyzer using a V2O5 catalyst weighted with the sample for combustion of
the samples in oxygen. All compounds were dried for 12 h in vaccum at 80 oC before elemental
analyses.
General procedure used for the synthesis of 1-(sulfonamido)alkyl-1,2-dicarba-closo
-
dodecaboranes (1-4). To a mixture of the corresponding (C3-C7) alkyne-1-sulfonamide (A-D)
(36 mmol) and arachno-6,9-(Me2S)2-arachno-B10H12 (9.8 g, 40.0 mmol), toluene (50 mL) was
added with a syringe. The slurry was heated under stirring and refluxed for 24 h. After cooling to

room temperature, solvent was removed under reduced pressure, and products were extracted
with diethyl ether (3x 40 mL). The organic extracts were separated by filtration or decantation,
and the combined fractions were evaporated under reduced pressure. The crude products were
treated overnight with MeOH (50 mL) acidified with few drops of HCl (3 M) under stirring. The
solvent was then evaporated to dryness. The methyl ester of boric acid formed from degradable
open-cage impurities was removed by evaporation in this step. Pure products were isolated by
liquid chromatography on a silica gel column (25 x 3.5 cm I.D.) using diethyl ether as a solvent.
Fractions containing the product (according to NMR) were combined, evaporated under reduced
pressure, and dried in a vacuum.
o
11
1-H2NSO2CH2-closo-1,2-C2B10H11; (1) white solid, yield: 5.71 g (67%); m.p. 143-145 C, B
NMR (128 MHz, CD3CN, 25 °C, BF3.Et2O): = -2.86 d (1B, J = 153, B(12)H), -4.67 d (1B,
δ
J = 153, B(9)H), -9.71 d (2B, J = 153, B(8,10)H), -11.00 d (2B, J = 104, B(7,11)H), -11.75 d
(2B, J = 107, B(3,6)H), -13.04 d (2B, J = 162, B(4,5)H); 1H {11B} NMR (400 MHz, CD3CN, 25
°C, TMS):
B(3,6)H), 2.26 s (2H, B(7,11)H), 2.22 s (1H, B(12)H), 2.14 s (1H, B(9)H), 2.12 s (4H,
B(4,5,8,10)H); 13C NMR (100 MHz, CD3CN, 25 °C, TMS):
= 68.24 s (1C, (1)carborane),
δ= 5.69 br. s (2H, NH2), 4.46 s (1H, C(2)Hcarborane), 3.99 s (2H, CH2S), 2.51 s (2H,
δ
61.80 d (1C, J = 194 ,C(2)Hcarborane), 59.23 t (1C, J = 142, CH2S); MS (HR ESI-), m/z 236.1750
(100%) [M-H]-, 238.1677 (38%), calcd. 238.1681 for C3H14O2N1B10S (38%); Analysis: Found C
15.53, H 6.35, N 5.78 Calcd. for B10C3H15O2NS: C 15.18; H 6.37, N 5.90.
1-H2NSO2(CH2)2closo-1,2- C2B10H11; (2) white solid, yield: 5.18 g (57%) m.p. 153-159 oC;
11B NMR (128 MHz, CD3CN, 25 °C, BF3.Et2O):
δ = -3.03 d (1B, J = 146, B(12)H), -6.00 d (1B,
J = 143, B(9)H), -9.69 d (2B, J = 150, B(8,10)H), -11.90 d (4B, J = 150, B(3,6,7,11)H), -13.04 d
1
(2B, J = 156, B(4,5)H); H {11B} NMR (400 MHz, CD3CN, 25 °C, TMS):
δ= 5.37 br. s (2H,
NH2), 4.28 s (1H, C(2)Hcarborane), 3.21 2d (2H, J = 12.8, CH2S), 2.71 2d (2H, J = 12.4, CH2), 2.33
s (2H, B(3,6)H), 2.19 s (1H, B(12)H), 2.15 s (2H, B(7,11)H), 2.09 s (4H, B(4,5,8,10)H), 2.05 s
(1H, B(9)H); 13C NMR (100 MHz, CD3CN, 25 °C, TMS):
δ = 74.07 s (1C, C(1)carborane), 63.56 d
(1C, J = 197, C(2)Hcarb), 53.96 t (1C, J = 138, CH2S), 32.44 t (1C, J = 133, C(1)CH2); MS
(HR ESI-), m/z 250.1907 (100%) [M-H]-, 252.1834 (30%), calcd. 252.1838 (30% for

C4H16O2N1B10S; Analysis: Found C 19.34, H 6.73, N 5.56; Calcd. for B10C3H15O2NS: C 19.11,
H 6.82, N 5.57.
1-H2NSO2C3H6-closo-1,2-C2B10H11; (3). Compound was prepared starting from pent-4-yne-1-
o
sulfonamide, white solid, yield: 5.2 g (54%), m. p. 130-140 C; 11B NMR (128 MHz, CD3CN,
25 °C, BF3.Et2O):
δ
= -3.10 d (1B, J = 150, B(12)H), -6.24 d (1B, J = 146, B(9)H), -9.90 d (2B,
1
J = 153, B(8,10)H), -11.92 d (4B, J = 168, B(3,6,7,11)H), -13.20 d (2B, J = 165, B(4,5)H); H
{11B} NMR (400 MHz, CD3CN, 25 °C, TMS):
δ= 5.30 br. s (2H, NH2), 4.21 s (1H, C(2)Hcarb),
2.98 t (2H, J = 8.0, CH2S), 2.57 s (1H, B(12)H), 2.39 t (2H, J = 8.8, CH2), 2.32 s (2H, B(3,6)H),
13
2.18 s (2H, B(7,11)H), 2.07 s (4H, 4,5,8,10)H), 2.03 s (1H, B(9)H), 1.92 m (2H, C(1)CH2); C
NMR (100 MHz, CD3CN, 25 °C, TMS):
δ = 76.26 s (1C, C(1)carb), 63.33 d (1C, J = 195,
C(2)Hcarb), 54.00 t (1C, J = 137, CH2S), 36.13 t (1C, J= 133, CH2), 24.60 t (1C, J = 132,
C(1)CH2); HR MS: m/z 264.2063 (100%) [M-H]-, 266.1990 (25%), calcd. 266.1994 (25% for
C5H18O2N1B10S); Analysis: Found C 22.86, H 6.95, N 5.68 Calcd. for B10C3H15O2NS: C 22.63,
H 7.22, N 5.28.
1-H2NSO2C4H8-closo-1,2-C2B10H11 (4). Compound was prepared at a 10-fold lower scale
starting from hex-5-yne-1-sulfonamide (0.55 g, 3.41 mmol) and 6,9-(Me2S)2-B10H12 (1.0 g, 4.08
mmol) in toluene (25 mL). The same general procedure was used for product isolation; white
solid, yield: 0.60 g (63%), m.p. 115-118 oC, 11B NMR (128 MHz, CD3CN, 25 °C, BF3.Et2O):
δ
= -3.22 d (1B, J = 156, B(12)H), -6.38 d (1B, J = 146, B(9)H), -9.95 d (2B, J = 156, B(8,10)H), -
1
11.78 d (4B, J = 178, B(3,6,7,11)H), -13.23 d (2B, J = 162, B(4,5)H); H {11B} NMR (400
MHz, CD3CN, 25 °C, TMS): δ= 5.22 br. s (2H, NH2), 4.16 s (1H, C(2)Hcarborane), 3.00 m (2H,
CH2S), 2.29 m (2H, J = 8.2, CH2), 2.17 s (1H, B(12)H), 2.16 s (4H, 3,6,7,11)H), 2.07 s (4H,
13
B(4,5,8.10)H), 2.02 s (1H, B(9)H), 1.70 m (2H, J = 8.0 CH2), 1.59 t (2H, J = 8.0, CH2); C
NMR (100 MHz, CD3CN, 25 °C, TMS):
δ = 68.15 s (1C, C(1)carb), 59.69 d (1C, J = 195,
C(2)Hcarb), 59.50 t (1C, J = 137, CH2S), 42.28 t (1C, J = 137, CH2), 33.21 t (1C, J = 133, CH2),
28.67 t (1C, J = 132, C(1)CH2); MS (ESI-), m/z 278.2222 (100%), 281.2151 (35 %), calcd.
281.2224 for C6H20O2N1B10S (38%); Analysis: Found C 26.08, H 7.24, N 5.28 Calcd. for
B10C3H15O2NS: C 25.79, H 7.58, N 5.01.

General procedure for the synthesis of potassium salts of 7-(sulfonamido)alkyl--dicarba-
nido-7,8-undecaborates (5-7). To the respective 1-(sulfonamido)alkyl-1,2-dicarba-closo-
dodecaborane (10 mmol), MeOH (50 mL) was added, followed by addition of solid KOH (0.40
g, 100 mmol) in several portions under vigorous stirring. The methanol solution was stirred and
heated at 60 oC for 6 h under reflux. Then, water (50 mL) was added and MeOH was evaporated
under reduced pressure. The resulting aqueous solution was diluted with water (60 mL) and
extracted with diethyl ether (2 x 20 mL) and ethyl acetate (4 x 25 mL). The ethyl acetate extracts
were combined, water (10 mL) was added, and the solvents were evaporated under reduced
pressure. Final products were purified by liquid chromatography on a silica gel column (25 x 2.5
cm I.D.) using a CH2Cl2-CH3CN solvent mixture (3:1 to 2:1 b.v.) for elution. Fractions
containing the product (according to NMR) were combined, evaporated under reduced pressure,
and dried in a vacuum for 5 h at 45 oC.
o
[7-H2NSO2CH2-nido-7,8-C2B9H11]K (5-). white solid, yield: 2.28 g (86%), m. p. 120 C
(decomp.); 11B (128 MHz, CD3CN, 25 °C, BF3.Et2O):
δ = -10.85 d (2B, J = 134, B(9,11)H), -
14.01 d (1B, J = 174, B(4)H), -15.37 d (1B, J = 143, B(6)H), -19.22 d (2B, J = 119, B(2,5)H), -
21.29 d (1B, J = 204, B(3)H), -33.00 d (1B, J = 125, B(10)H), -37.44 d (1B, J = 140 , B(1)H);
1H {11B} NMR (400 MHz, CD3CN, 25 °C, TMS):
δ= 3.76 br. s (2H, NH2), 3.32 d (1H,
J = 14.8, CH2S), 3.05 d (1H, J = 14.8, CH2S), 2.09 s (1H, B(9)H), 2.08 s (2H, CH2S), 2.06 s (1H,
C(8)Hcarborane), 1.95 s (1H, B(11)H), 1.64 s (1H, B(4)H), 1.28 s (1H, B(6)H), 1.22 s (1H, B(3)H),
1.13 s (1H, B(5)H), 1.01 s (1H, B(2)H), 0.44 s (2H, B(1)H), -0.01 s (1H, B(10)H), -2.69 s (1H,
13
µ-H); C NMR (100 MHz, CD3CN, 25 °C, TMS):
δ = 65.05 t (1C, J = 139, CH2S), 52.92 s
(1C, C(7)carb), 45.63 d (1C, J = 149, C(8)Hcarb); HRMS (ESI-), m/z [M]-: 226.1736 (100%),
228.1664 (32%), calcd. 228.1667 for C3H15O2N1B9S (50%); Analysis for Et3NH.5: Found C
32.56, H 9.24, N 8.76, Calcd. for B10C3H15O2NS: C 32.88, H 9.51, N 8.52.
[7-H2NSO2-(CH2)2- nido-7,8-C2B9H11]K (6-). white solid, yield: 2.24 g (80%), m. p. 229-232 oC
11
(decomp.); B (128 MHz, CD3CN, 25 °C, BF3.Et2O):
δ = -11.33 d (2B, J = 134, B(9,11)H), -
14.23 d (1B, J = 159, B(4)H), -16.61 d (1B, J = 134, B(6)H), -19.32 d (2B, J = 140, B(2,5)H), -
22.00 d (1B, J = 150, B(3)H), -33.51 d (1B, J = 128, B(10)H), -37.39 d (1B, J = 140 , B(1)H); 1H
{11B} NMR (400 MHz, CD3CN, 25 °C, TMS):
δ= 5.16 s (2H, NH2), 3.13 m (2H, CH2) and 1.89

m (2H, CH2C), 1.81 s (1H, B(9)H), 1.71 br. s (1H, C(8)Hcarborane), 1.71 s (1H, B(11)H), 1.63 s
(2H, CH2), 1.28 s (1H, B(5)H), 1.15 s (2H, B(3,6)H), 0.98 s (1H, B(2)H), 0.79 s (1H, B(4)H),
13
0.39 s (1H, B(1)H), -0.06 s (1H, B(10)H), -2.82 s (1H, µ-H); CNMR (100 MHz, CD3CN, 25
°C, TMS):
δ = 57.76 s (1C, C(7)carborane), 56.29 t (1C, J = 137, CH2S), 46.28 d (1C, J = 108,
C(8)Hcarborane), 33.74 t (1C, J = 132, CH2C); HRMS (ESI-), m/z [M]-: 241.1855 (100%),
242.1819 (47%), calcd. 242.1823 for C4H17O2N1B9S (47%); Analysis for Et3NH.6: Found C
34.66, H 9.42, N 8.48 Calcd. for B10C3H15O2NS: C 35.04, H 9.70, N 8.17.
[7-H2NSO2-(CH2)3-nido-7,8-C2B9H11]K (7-). white solid, yield: 2.71 g (92%), m. p. 312-314
decomp.; 11B (128 MHz, CD3CN, 25 °C, BF3.Et2O):
δ = -11.42 d (2B, J = 134, B(9,11)H), -14.32
d (1B, J = 156, B(4)H), -17.25 d (1B, J = 156, B(6)H), -18.86 d (2B, J = 186, B(2,5)H), -22.22 d
1
(1B, J = 146, B(3)H), -33.63 d (1B, J = 125, B(10)H), -37.53 d (1B, J = 137 , B(1)H); H {11B}
NMR (400 MHz, CD3CN, 25 °C, TMS): δ= 5.23 s (2H, NH2), 2.94 m (2H, CH2S), 2.30 s (1H,
C(8)Hcarborane), 1.85 (2H, B(9,11)H), 1.81 m (2H, J = 0.02, CH2), 1.67 m (2H, J = 0.02, CH2),
1.53 s (1H, B(4)H), 1.12 s (1H, B(3)H), 1.11 s (1H, B(6)H), 1.00 s (2H, B(2,5)H), 0.37 s (2H,
B(1)H), -0.08 s (1H, B(10)H), -2.82 s (1H, µ-H); 13C NMR (100 MHz, CD3CN, 25 °C, TMS):
δ
= 60.14 s (1C, C(7)carb), 55.58 t (1C, J = 136, CH2S), 47.48 d (1C, J = 152, C(8)Hcarborane), 38.39 t
(1C, J = 127, CH2), 26.30 t (1C, J = 128, C(7)CH2); HRMS (ESI-), m/z [M]-: 256.1979 (50%),
calcd. 256.1980 (for C5H19O2N1B9S); Analysis for Et3NH.7: Found C 36.62, H 9.52, N 8.18
Calcd. for B10C3H15O2NS: C 37.03, H 9.89, N 7.85.
[7-H2NSO2-(CH2)4-nido-7,8-C2B9H11]Et3NH (8-). MeOH (15 mL) was added to 1-
(sulfonamido)butyl-1,2-dicarba-closo-dodecaborane (4, 1.0 mmol) followed with an excess of
o
Et3N (0.5 mL, 70 mmol). The methanol solution was stirred and heated at 60 C for 10 h under
reflux. After cooling, volatiles were removed in vacuo, and the product was extracted with a
ether-ethylacetate mixture (1:1 b.v., 3 x 15 mL). Water (5 mL) was added, and the organic
solvents were removed under vacuum pressure. The product was decanted, washed with water,
dried, and chromatographed on a silica gel column (25 x 1.5 cm I.D.) using a CH2Cl2-CH3CN
solvent mixture (3:1 to 2:1 b.v.) for elution. The final product was a white solid, yield: 0.325 mg
o
11
(88%), m. p. 118-123 C; B NMR (128 MHz, CD3CN, 25 °C, BF3.Et2O):
δ = -11.4 d (2B,
J = 131, B(9,11)H), -14.23 d (1B, J = 153, B(3)H), -17.77 d (1B, J = 116, B(5)H), -18.48 d (2B,

J = 107, B(2,6)H), -22.38 d (1B, J = 137, B(4)H); -33.73 d (1B, J = 119, B(10)H); -37.58 d (1B,
J = 134, B(1)H); 1H {11B} NMR (400 MHz, CD3CN, 25 °C, TMS):
= 5.38 br. s (2H, NH2), 3.25
δ
s (1H, C(7)Hcarborane), 3.09 q (6H, CH2 Et3NH(+)), 3.00 m (2H, CH2S), 1.68 m (2H, CH2), 1.64 t
(2H, CH2), 1.46 s (2H, CH2), 1.85 s (2H, B(9,11)H), 1.48 s (1H, (3)H), 1.21 t (9H, CH3 Et3NH(+)
), 1.15 s (1H, (5)H), 1.02 s (2H, (2,6)H), 1.75 s (1H, (4)H), -0.09 s (1H, (10)H), -0.34 s (1H,
(1)H), -2.81 s (1H, µH); 13C NMR (100 MHz, CD3CN, 25 °C, TMS):
δ = 60.80 s (1C, C(7)carb),
46.20 d (1C, C(8)Hcarb), 54.76 t (1C, CH2S), 39.55 t (1C, CH2), 29.75 t (1C, CH2), 23.98 t (1C,
C(7)CH2), 47.68 t (3C, CH3 Et3NH(+)), 9.23 t (3C, CH2 Et3NH(+)); HRMS (ESI-) m/z [M]-:
268.2210 (100%) 270.2136 (52%), calcd. for C6H20O2N1B10S: 270.2131 (52%); Analysis for
Et3NH.8: Found C 38.49, H 9.74, N 7.92 Calcd. for B10C3H15O2NS: C 38.87, H 10.06, N 7.55.
For biological activity tests, the triethylammonium cation present in this salt was replaced with
sodium by methathesis.
Protein cloning, expression, and purification. Recombinant CA II and CA IX mimic (CA II
containing amino acid substitutuions A65S, N67Q, E69T, I91L, F130V, K169E, and L203A)
were prepared by heterologous expression in E. coli and purified as previously described [34].
The extracellular part of CA IX comprising the PG and CA domains (residues 38-391) and
including the amino acid substitution C174S was expressed in HEK 293 cells and purified as
previously described [47].
CA inhibition assay. The stopped-flow instrument (Applied Photophysics) has been used for
measuring the CA-catalysed CO2 hydration activity in the presence of inhibitors [48]. The assay
buffer consists of 0.2 mM phenol red (pH indicator used in absorbance maximum of 557 nm), 20
mM HEPES-Na (pH 7.5) and 20 mM Na2SO4. Concentration of CA II and CA IX in the enzyme
assay was 2.5 nM and 0.5 nM, respectively. To stabilize CA IX, 0.0025 % dodecyl-β-D-
maltopyranoside (DDM, Anatrace) was included in the reaction mixture. The substrate (CO2)
concentration in the reaction was 8.5 mM. Rates of the CA-catalysed CO2 hydration reaction
were followed for a period of 30 s at 25 °C. Four traces of the initial 5−10% of the reaction have
been used for determining the initial velocity for each inhibitor. The uncatalysed rates were
determined in the same manner and subtracted from the total observed rates. Stock solutions of
inhibitors (100 mM) were prepared in dimethyl sulfoxide (DMSO) and dilutions up to 100 nM

were made thereafter in DMSO. Ki' values were obtained from dose-response curves recorded
for at least six different concentrations of the test compound by the nonlinear least-squares
method using an EXCEL spreadsheet fitting the Williams-Morrison equation [49][49]. Ki values
were then derived using the Cheng-Prusoff equation [50]. KM values used in Cheng-Prusoff
equation were 9.3 mM for CA II and 7.5 mM for CA IX [51, 52].
Protein crystallization and X-ray data collection. CA IX mimic was used for X-ray studies.
Complexes with 3 and 7 were prepared by addition of a 1.1-fold molar excess of inhibitor
(dissolved in pure DMSO) to a 25 mg/mL protein solution in 50 mM Tris, pH 7.8. The best
crystals were prepared by the vapor-diffusion hanging drop method at 18 °C using aprecipitation
solution containing 1.6 M sodium citrate, 50 mM Tris-HCl, pH 7.8. Drops containing 2 μL
complex solution were mixed with 1 μL precipitant solution and equilibrated over a reservoir
containing 1 mL precipitant solution. The final DMSO concentration in the drop did not exceed
5% (v/v). Crystals with dimensions of 0.5 mm × 0.3 mm × 0.2 mm typically grew within 1-4
weeks. Before data collection, the crystals were soaked for 5−10 s in reservoir solution
supplemented with 20% (v/v) sucrose and stored in liquid N2. Diffraction data at 100 K were
collected on BL14.1 operated by the Helmholtz-Zentrum Berlin (HZB) at the BESSY II electron
storage ring (Berlin-Adlershof, Germany) [53]. Diffraction data were processed using the XDS
suite of programs [54]. Crystal parameters and data collection statistics are summarized in
Supplementary Table S1.
Structure determination, refinement, and analyses. Crystal structures of the CA IX mimic in
complexes with 3 and 7- were determined by the difference Fourier technique. Coordinates from
PDB entry 5OGP were used as a model. Atomic coordinates of inhibitor molecules were
generated by quantum mechanical (QM) optimizations in the Turbomole package [55] with the
density functional theory (DFT) method using the B-LYP functional and the SVP basis set,
augmented with empirical dispersion correction [56]. The geometric library for the inhibitors
was generated using the Libcheck program, part of the CCP4 package [57]. Coot [58] was used
for inhibitor fitting, model rebuilding, and the addition of water molecules. Refinement was
carried out with Refmac5 [59], with 5% of reflections reserved for cross-validation [60].