Sulfonamide Inhibitors of Human Carbonic Anhydrases Designed through a Three-Tails Approach: Improving Ligand/Isoform Matching and Selectivity of Action

Article abstract of DOI:10.1021/acs.jmedchem.0c00733

The "tail approach"has become a milestone in human carbonic anhydrase inhibitor (hCAI) design for various therapeutics, including antiglaucoma agents. Besides the classical hydrophobic/hydrophilic division of hCAs active site, several subpockets have been

Full text of DOI:10.1021/acs.jmedchem.0c00733

Subscriber access provided by the University of Exeter
Article
Sulfonamide Inhibitors of Human Carbonic Anhydrases
Designed through a Three-Tails Approach: Improving
Ligand/Isoform Matching and Selectivity of Action
Alessandro Bonardi, Alessio Nocentini, Silvia Bua, Jacob Combs, Carrie L. Lomelino, Jacob Andring, Laura
Lucarini, Silvia Sgambellone, Emanuela Masini, Robert McKenna, Paola Gratteri, and Claudiu T. Supuran
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.0c00733 • Publication Date (Web): 10 Jun 2020
Downloaded from pubs.acs.org on June 13, 2020
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the
course of their duties.

Page 1 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Sulfonamide Inhibitors of Human Carbonic Anhydrases Designed through a Three-Tails
Approach: Improving Ligand/Isoform Matching and Selectivity of Action
Alessandro Bonardi,a,b Alessio Nocentini,a,b* Silvia Bua, Jacob Combs, Carrie Lomelino, Jacob
a
c
c
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
c
d
d
d
c
Andring, Laura Lucarini, Silvia Sgambellone, Emanuela Masini, Robert McKenna, Paola
a,b
a,*
Gratteri, Claudiu T. Supuran
a
Department NEUROFARBA – Pharmaceutical and nutraceutical section, University of Firenze, via Ugo Schiff 6, 50019 Sesto Fiorentino (Florence),
Italy.
b
Department NEUROFARBA – Pharmaceutical and nutraceutical section; Laboratory of Molecular Modeling Cheminformatics & QSAR, University
of Firenze, via Ugo Schiff 6, 50019 Sesto Fiorentino (Florence), Italy.
c
Department of Biochemistry and Molecular Biology, College of Medicine, University of Florida, Box 100245, Gainesville, FL 32610, USA.
d
Department NEUROFARBA – Pharmaceutical and nutraceutical section, University of Firenze, viale Gaetano Pieraccini 6, 50139 Firenze (Florence),
Italy.
Abstract
The “tail-approach” has become a milestone in human carbonic anhydrase inhibitor (hCAI) design
for various therapeutics, including antiglaucoma agents. Beside the classical hydrophobic/hydrophilic
division of hCAs active site, several subpockets have been importantly identified at the medium/outer
active sites rim which could be targeted to increase CAI isoform-selectivity. This postulate is explored
here by three-tailed benzenesulfonamide CAIs (TTI) to fully exploit the such amino acid differences
among hCAs. In this proof-of-concept study, an extensive structure-activity-relationship (SAR) study
was carried out with thirty-two such benzenesulfonamides differing for tails combination, that were
assayed for hCAs I, II, IV, and XII inhibition. A structural study was undertaken by X-ray
crystallography and in silico tools to assess the ligand/target interaction mode. The most active and
selective inhibitors against isoforms implicated in glaucoma were assessed in a rabbit model of the
disease achieving an intraocular pressure lowering action comparable to the clinically used
dorzolamide.
Keywords: metalloenzyme, drug-design, tail-approach, selectivity, inhibition, X-ray
crystallography, glaucoma, rabbit model.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
INTRODUCTION
Carbonic anhydrases (CAs, EC 4.2.1.1) are among the most efficient catalysts, speeding up the simple
yet physiologically essential reaction in all kingdoms: the reversible hydration of carbon dioxide to
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
bicarbonate and protons. Among the eight genetically unrelated CA families α, β, γ, δ, η, ζ, θ and
ι,2-9 α-CAs are uniquely present in higher vertebrates.2,10 In particular, humans express fifteen α-CA
isoforms (hCAs) which differ in catalytic activity, subcellular/tissue localization, and physiological
11
role. Therefore, hCAs are involved in multiple physiological processes and their levels of activities
are linked to many human disorders such as glaucoma, retinal/cerebral edema, retinitis pigmentosa,
other retinopathies, stroke, epilepsy, sterility, osteoporosis, altitude sickness, cariogenesis,
neurodegeneration, obesity, and cancer.12-14 As a result, almost all catalytically active hCAs have
generated great interest for the design of inhibitors (CAIs) or activators (CAAs) with biomedical
1
5
applications. Whereas initially CAIs were used as diuretics, antiglaucoma agents, antiepileptics, and
2
for the management of altitude sickness, a new generation of CAIs are being developed for the
treatment of cancers, obesity, inflammation, neuropathic pain, infections, and neurodegenerative
disorders.16-21 CAAs are also of interest in the field of cognition, aging, and neurodegeneration.22
Nevertheless, the use as antiglaucoma agents is still the main therapeutic application of CAIs. In
fixed-drugs combinations (mainly with prostaglandin analogues and beta-blockers) CAIs continue to
be marketed worldwide and widely used.23 Acetazolamide (AAZ), methazolamide (MTZ), and
dichlorophenamide (DCP) are first-generation CAIs used as systemic drugs for the management of
this disease (Figure 1). Dorzolamide (DRZ) and brinzolamide (BRZ) represent second-generation
inhibitors used topically, as eye drops, with less side effects compared to the first-generation drugs.24
However, none of these drugs possess a selective inhibition profiles against the hCA isoforms mainly
implicated in the disease, that are hCA II (main isoform), IV, and XII. Considering that the current
therapies are overall often inadequate given that multiple classes of medications have to be
2
5
coadministered to control IOP efficiently, it might be of crucial importance to optimize the single
ACS Paragon Plus Environment

Page 3 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
CAI agents, by increasing their efficacy (against the target CAs), and decreasing adverse events
(improving their selectivity of action).
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
H NO S
2
2
SO NH
2
2
H
S
S
N
SO NH
2 2
N
SO NH2
2
N N
Cl
N N
O
O
Cl
AAZ
MTZ
DCP
HN
HN
O
SO NH
SO NH
2 2
2
2
N
S
S
S
S
O O
O O
DZA
BRZ
Figure 1. Clinically used antiglaucoma CAIs.
The 12 catalytically active hCAs (isoforms VIII, X and XI are catalytically inactive) are characterized
by a Zn(II) ion, which is tetrahedrally coordinated by three histidine residues, and a solvent molecule
that are situated at the base of a 13Å deep conical cavity portioned into a hydrophobic and hydrophilic
side.11,15,26 As the hCAs catalytic domains are structurally homologus and conserved in amino acid
sequence identity, it is rather challenging to achieve targeted inhibition of a specific hCA isozyme
over others. Despite this, many new approaches have been developed for this purpose, especially over
the last two decades.15
So far, four unique CA inhibition mechanisms have been validated by both kinetic and structural
assessment:15,27 1) zinc-binding, which consists in the direct coordination of the catalytical Zn(II) ion
with a tetrahedral or trigonal bipyramidal coordination geometry (sulfonamides, sulfamides,
sulfonates, anions, mono-dithiocarbamates, xanthates, thioxanthates, carboxylates, hydroxamates,
benzoxaboroles, selenols); 2) anchorage to the zinc-bound water molecule/hydroxide ion (phenols,
thiophenols, polyphenols, carboxylates, polyamines, 2-thioxocoumarins, sulfocoumarins); 3)
occlusion of the active site entrance (coumarins and bioisosters); and 4) binding out of the active site
(a unique carboxylic acid derivative exhibited this inhibition mode to date).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Undoubtedly, the zinc binders, such as sulfonamides and their bioisosters sulfamates and sulfamides
in a prominent position, are among the most effective and investigated derivatives in the field of CA
inhibition as well as in the related clinical context.11,15
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
In fact, most efforts have been made on this class of CAIs to achieve isozyme selectivity of action, in
2
8
order to lower the side effects consequent to promiscuous inhibition. As simple as effective, the so
called “tail approach” made its appearance in the field of CA inhibition in 1999 and led to the
development of a large number of studies and compounds that expanded the database of CA isoform
selective inhibitors by appending a wide spectrum of chemical functionalities, named tails, to the
main zinc-binding scaffold.29-35 The original aim was to increase the water solubility29 and
3
2
subsequently membrane (im)permeability of aromatic sulfonamide derivatives. Afterwards, the
design was shifted toward the modulation of the interactions between the ligand and the middle and
outer rim of the hCAs active sites, which contain the most variable polypeptide regions among the
various isoforms, in order to increase isoform specificity. Simple tailed CAIs are composed of the
following elements: (i) a zinc-binding function, (ii) a main scaffold that can include a linker, and (iii)
the tail (Figure 2A).
3
6
An extension of this approach was proposed in 2015 by Tanpure et al., with the simultaneous
inclusion of two tails of diverse nature onto aromatic sulfonamide scaffolds, at a nitrogen atom
branching point, allowing distinct binding to the hydrophobic and hydrophilic sections of the hCAs
active site (Figure 2B). However, a limited number of compounds was reported (three) and an in vitro
assay was performed solely on hCA II, which makes this pioneering study rather unfulfilled. More
recently, Fares et al. used a similar approach proposing a diverse type of dual tails to
benzenesulfonamide CAIs.37
The detailed knowledge of the active site composition and architecture of hCAs (mostly available by
X-ray crystallographic studies, except for CAs VA and VB) derived from many previous studies38-40
led to the conclusion that the simple hydrophobic/hydrophilic division of the isoforms binding pocket
may no longer be sufficient. In fact, some CA isozymes do not exhibit such a precise distinction as
ACS Paragon Plus Environment

Page 5 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
13
originally noticed in hCA I, II and IX, and a bulk of accessory subpockets exist which differentiate
the various CA isoforms. Here, the inclusion of a third tail is proposed as an approach to improve the
matching and fitting of the target-ligand interaction within the different hCAs active sites (Figure
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
C).
As a first proof-of-concept of this improved approach, a diverse array of tail combinations were
investigated with the aim of identifying suitable isoform-imprints. Described here is the screening of
hCA isozymes I, II, IV, and XII with thirty-two benzenesulfonamide derivatives incorporating three
tails. In the context of the antiglaucoma CAI application, hCA I is the main off-target isoform as it is
2
widespread in red blood cells and many other tissues. A comprehensive structural study was also
undertaken by X-ray crystallography with hCA II and in silico with isozymes hCA I, IV and XII, to
assess the ligand-/target interaction modes. A selection of the three-tailed inhibitors most active
against hCAs implicated in glaucoma were assessed in vivo in a rabbit model of the diseases and
compared to classical clinically used CAIs.
B
C
A
Hydrophilic half
of active site
Hydrophilic half
of active site
Hydrophilic half
of active site
TAIL
TAIL
middle rim
middle rim
middle rim
TAIL
TAIL
outer rim
TAIL
outer rim
outer rim
Glu106
Glu106
Glu106
TAIL
O
O
O
O
LD
LD
LD
O
O
O
O
O
SCAFF
Hydrophobic half
of active site
SCAFF
Hydrophobic half
of active site
SCAFF
Hydrophobic half
of active site
H
O
H
O
H
O
ZBG
ZBG
ZBG
H
N
inner rim
H
inner rim
H
inner rim
N
N
Thr199
Thr199
Thr1
99
Zn2+
Zn2+
Zn2+
His119
His96
His119
His119
His96
His94
His94
His94
His96
Figure 2. Schematic representation of the (A) “tail”, (B) “two-tails” and (C) “three-tails” approach
for the design of zinc-binding CAIs.
RESULT AND DISCUSSION
Drug-design and chemistry
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Currently, the tail approach has been a focus of CAIs research area with most design studies adopting
the p-substituted benzenesulfonamide scaffold as a main foothold to include a variety of chemical
1
5
frameworks. In fact, avoiding heteroaromatic sulfonamide scaffolds markedly eases the synthetic
3
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
procedures, moving the focus on the inclusion of pendants on the inhibitor structure. Likewise, to
converge efforts and attention on studying the three-tailing effects on CA inhibition, a p-substituted
benzenesulfonamide was here adopted as a CAI scaffold.
It should be stressed that it is not possible to easily include three chemically diverse tails on a single
branching atom (e.g. a nitrogen atom, as proposed by Tanpure et al. in the two-tails approach),36
unless obtaining an ammonium salt or a chiral center. As a result, among several identified
alternatives to branch a spacer attached to the main scaffold into three tails, the general structure TTI
(Figure 3) was selected to combine easy and versatile chemistry with the possibility to extend it to
many diverse chemical groups, which is relevant for producing a range of tail combinations. As a
result, TTI was designed in the following manner: (i) a benzenesulfonamide scaffold (blue) which
assures the interaction with the zinc ion and the bottom of the active site; (ii) an ethylenic spacer (red)
which has the function to allow sufficient space between the main scaffold and the tails; (iii) a first
ramification point (N atom, in black) from which the first tail T (green) branches off; (iv) an amide-
1
based spacer (red); (v) a second intersection point (N atom, in black) by which T and T (green)
2
3
branch off. Having the benzenesulfonamide bound to the Zn(II) at the bottom of the active site, the
linkers in red (Figure 3) were chosen in such a way as to explore a vast chemical space at the medium
and outer rims of the binding clefts.
H NO S
2
2
T2
O
N
n
N
T3
T1
TTI
Figure 3. General structure of the designed three-tailed inhibitors (TTIs).
The synthetic strategies adopted to yield the TTI derivatives are reported in Schemes 1-3. T was
1
introduced on the 4-(2-aminoethyl)benzenesulfonamide by reductive amination with the proper
ACS Paragon Plus Environment

Page 7 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
aromatic aldehyde and sodium borohydride in MeOH or, alternatively, by nucleophilic substitution
with the appropriate halides in anhydrous DMF and in presence of TEA to furnish secondary amines
1-5 and 6 and 7, respectively. The latter were reacted with chloroacetyl chloride or chloropropionyl
chloride in acetone and in presence of K CO to provide amides 8-17. T and T were finally included
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
3
2
3
through a nucleophilic substitution with commercially available or synthesized secondary amines in
anhydrous ACN and TEA as a base to produce TTIs 18-33. The nitrile derivatives 33-37 were further
converted to the corresponding amines 40-44 through a Ni/Raney catalyzed hydrogenation or
hydrolyzed in NaOH(aq) into the corresponding carboxylic acids 45-49 (Scheme 2). Additionally, the
markedly hydrophobic oleylamide derivative 50 was yielded by coupling the carboxylic acid 46 with
oleylamine in presence of EDC and DMAP in anhydrous DMF.
All derivatives were purified by silica gel chromatography eluting with MeOH/DCM gradients and
1
13
fully characterized by H-NMR, C-NMR, and HRMS (Supporting Information).
1: R =C H
6 5
1
H2NO2S
H NO S
2: R =4-NO -C H
1 2 6 4
2
2
a+b or c
3: R =4-F-C H
1 6 4
4
5
: R =2-Naph
: R =Fu
1
NH2
NH
R1
1
6: R =CH CN
7
1
2
1-7
1 2 6 5
: R =CH C H
8: n=1, R =C H
1
6
5
5
d
9
1
1
1 6
: n=2, R =C H
0: n=1, R =4-NO -C H
6 4
1
2
1: n=1, R =4-F-C H
6 4
1
H2NO2S
H2NO2S
O
N
R2
N
n
O
12: n=1, R1=2-Naph
e
13: n=1, R =Fu
Cl
n
1
R3
N
14: n=2, R1=Fu
1
1
1
5: n=1, R =CH CN
1 2
R1
R1
6: n=1, R =CH C H
2 6 5
18-39
8-17
1
7: n=2, R =CH C H
2 6 5
1
1
8: n=1, R =C H , R =CH CH , R =CH CH
29: n=2, R =C H , R =(CH ) CH , R =(CH ) CH
1 6 5 2 2 7 3 3 2 7 3
1
6
5
2
2
3
3
2
3
19: n=1, R =C H , R =CH CH , R =CH C H
30: n=2, R =CH C H , R =(CH ) CH , R =(CH ) CH
1
6
5
2
2
3
3
2
6
5
1 2 6 5 2 2 4 3 3 2 4 3
2
2
2
0: n=1, R =C H , R =CH C H , R =CH C H
31: n=2, R =Fu, R =(CH ) CH , R =(CH ) CH
1 2 2 4 3 3 2 4 3
1
6
5
2
2
6
5
3
2
6
5
1: n=1, R =C H , R =(CH ) CH , R =(CH ) CH
32: n=1, R =2-Naph, R =(CH ) CH , R =(CH ) CH
1
6
5
2
2 4
3
3
2 4
3
3
3
1 2 2 4 3 3 2 4 3
2: n=1, R =C H , R =(CH ) CH , R =(CH ) CH
33: n=1, R =CH CN, R =(CH ) CH , R =(CH ) CH
1 2 2 2 4 3 3 2 4 3
1
6
5
2
2 5
3
3
2 5
23: n=1, R =C H , R =(CH ) CH , R =(CH ) CH
34: n=1, R =CH C H , R =(CH ) C H , R =(CH ) CN
1
6
5
2
2 7
3
3
2 7
1 2 6 5 2 2 2 6 5 3 2 2
2
4: n=2, R =C H , R =CH CH , R =CH CH
35: n=1, R =Fu, R =(CH ) C H , R =(CH ) CN
1 2 2 2 6 5 3 2 2
1
6
5
2
2
3
3
2
3
2
5: n=2, R =C H , R =CH CH , R =CH C H
36: n=1, R =4-F-C H , R =(CH ) C H , R =(CH ) CN
1
6
5
2
2
3
3
2
6
5
1 6 4 2 2 2 6 5 3 2 2
26: n=2, R =C H , R =CH C H , R =CH C H
37: n=1, R =2-Naph, R =(CH ) C H , R =(CH ) CN
1 2 2 2 6 5 3 2 2
1
6
5
2
2
6
5
3
2
6
5
2
7: n=2, R =C H , R =(CH ) CH , R =(CH ) CH
38: n=1, R =4-NO -C H , R =(CH ) C H , R =(CH ) CN
39: n=1, R =CH CN, R =(CH ) C H , R =(CH ) CN
1 2 2 2 2 6 5 3 2 2
1
6
5
2
2 4
3
3
2 4
3
3
1 2 6 4 2 2 2 6 5 3 2 2
28: n=2, R =C H , R =(CH ) CH , R =(CH ) CH
1
6
5
2
2 5
3
3
2 5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Scheme 1. Reagents and conditions: a) R CHO, anhydrous MeOH, reflux, 4h; b) NaBH , anhydrous
1
4
MeOH, reflux, 0.5-2h; c) R CH X, TEA, anhydrous DMF; d) ClCO(CH ) Cl, K CO , acetone, r.t,
1
2
2 n
2
3
1h; e) R R NH, TEA, anhydrous ACN, reflux, 4-24h.
2
3
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
R1
40: R1=CH2C6H5, R2=(CH2)2C6H5, R3=(CH2)3NH2
41: R1=Fu, R2=(CH2)2C6H5, R3=(CH2)3NH2
N
R3
a
N
42: R1=4-F-C6H4, R2=(CH2)2C6H5, R3=(CH2)3NH2
O
R2
H2NO2S
43: R =2-Naph, R =(CH ) C H , R =(CH ) NH
1
2
2 2
6
5
3
2 3
2
44: R1=(CH2)2NH2, R2=(CH2)5CH3, R3=(CH2)5CH3
40-44
R1
R1: (CH2)2NH2 and/or (CH2)3NH2
N
R3
N
O
R2
H2NO2S
R1
33-37
R1: CH2CN and/or (CH2)2CN
4
4
4
4
5: R1=CH2C6H5, R2=(CH2)2C6H5, R3=(CH2)2COOH
6: R1=Fu, R2=(CH2)2C6H5, R3=(CH2)2COOH
7: R1=4-F-C6H4, R2=(CH2)2C6H5, R3=(CH2)2COOH
8: R1=2-Naph, R2=(CH2)2C6H5, R3=(CH2)2COOH
N
R3
b
N
O
R2
H2NO2S
49: R1=CH2COOH, R2=(CH2)5CH3, R3=(CH2)5CH3
4
5-49
R1: CH2COOH and/or (CH2)2COOH
Scheme 2. Reagents and conditions: a) H , Ni/Raney, NaOH, EtOH, r.t., o.n.; b) NaOH, EtOH, reflux,
2
o.n.
O
O
N
N
O
N
N
a
O
H2NO2S
NH
O
H2NO2S
46
COOH
50
Scheme 3. Reagents and conditions: a) oleylamine, EDC٠HCl, DMAP, anhydrous DMF, r.t, o.n.
Carbonic Anhydrase Inhibition
In this first screening, mono-tailed (1-7) and three-tailed (18-50) compounds were analyzed by a
41
stopped-flow kinetic assay with hCA isoforms I, II, IV, and XII. HCAs II, IV, and XII are involved
in glaucoma with the last isoform being reported to be upregulated in the eyes of glaucoma patients.
Thus, all of them are involved in this disease, both in the elevation of intraocular pressure (IOP) and
4
2
the decrease of blood flow and oxygen supply within the hypoxic neovascular retinic tissues. HCA
IV was reported to be involved in stroke, glaucoma, retinitis pigmentosa, astrocytomas, and
1
2
gliomas. HCA XII is also validated as an anticancer target, (being overexpressed on the membrane
ACS Paragon Plus Environment

Page 9 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
of hypoxic tumor cells)17 and recently, overexpression of this isoform has also been linked to
19
inflammation. HCA I is a main off-target isoform for the therapeutic application of CAIs in ocular
diseases, as this isoform is widespread in red blood cells and many other tissues.2
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 1. Inhibition data of human CA isoforms CA I, II, IV, and XII with sulfonamides 1-7, 18-50
reported here and the standard sulfonamide inhibitor acetazolamide (AAZ) by a stopped-flow CO2
hydrase assay.41
a
K
I
(nM)
Cmpd n
R
1
R
-
2
R
-
3
CA I CA II CA IV CA XII
1
2
3
4
5
6
7
C
6
H
5
95.3
98.4 2854.4
65.4
77.4
60.1
78.6
55.4
-
- 4-NO
-C
H
-
-
-
-
224.3 120.9 1685.3
112.8 78.5 1196.7
458.1 87.1 6248.1
2
6
4
-
-
-
-
-
4-F-C H
6
4
-
-
-
-
-
-
-
-
2-Naph
Fu
68.4
62.8 1584.5
CH
CH
2
CN
105.3 153.7 5547.2 113.2
278.4 89.1 3587.4 104.3
2
6 5
C H
1
8
9
0
1
1
1
1
1
1
2
2
2
2
2
2
2
2
1
1
1
1
1
1
C
C
C
C
C
C
C
C
C
C
C
C
6
6
6
6
6
6
6
6
6
6
6
6
H
H
H
H
H
H
H
H
H
H
H
H
5
CH
CH
2
2
CH
CH
3
3
CH
2
CH
3
786.6
8.3
4147.5
43.9
1
2
5
5
5
5
5
5
5
5
5
5
5
CH
CH
2
C
6
6
H
5
5
4210.4 391.6 >10000 82.6
865.9 412.3 >10000 98.8
506.1 124.5 >10000 69.4
CH
(CH
(CH
(CH
2
C
6
H
5
2
2
2
2
C
H
21
2
)
4
)
5
)
7
CH
CH
CH
3
3
3
(CH
(CH
(CH
)
4
)
5
)
7
CH
CH
CH
3
3
3
2
2
2
3
2
2
878.7
946.7 843.8 >10000 99.4
184.7 8.9 3928.8 61.1
237 >10000 92.8
24
CH
CH
2
2
CH
CH
3
3
2 3
CH CH
2
2
5
6
CH
CH
2
C
6
6
H
5
5
544.3 79.6 >10000 90.4
692.3 559.2 4640.8 302.5
563.6 522.6 3244.8 100.3
CH
(CH
(CH
(CH
(CH
(CH
(CH
(CH
2
C
6
H
5
2
2
2
2
2
2
2
2
C
H
27
2
)
4
)
5
)
7
)
5
)
5
)
5
)
5
CH
CH
CH
CH
CH
CH
CH
3
3
3
3
3
3
3
(CH
(CH
(CH
(CH
(CH
(CH
(CH
)
4
)
5
)
7
)
5
)
5
)
5
)
5
CH
CH
CH
CH
CH
CH
CH
3
3
3
3
3
3
3
2
2
8
9
2
2
2
2
2
2
308.2 578.4 3455.4
209.3 778.8 >10000
518.4 780.8 3413.2
220.1 60.4 3153.7
77.8
280
62.5
9.7
30
CH
Fu
2-Naph
CH CN
CH
Fu
4-F-C
2
C
6
H
5
3
3
3
3
1
2
3
4
541.4 4562.9 >10000 61.7
395.9 52.5 3478.3 8.6
777.3 368.5 >10000 75.5
2
2
C
6
H
5
4
(CH
(CH
(CH
(CH
(CH
(CH
(CH
(CH
2
)
2
C
C
C
C
C
C
C
C
6
6
6
6
6
6
6
6
H
5
5
5
5
5
5
5
5
(CH
(CH
(CH
(CH
(CH
(CH
2
)
2
2
2
2
2
2
CN
CN
CN
CN
CN
CN
35
2
2
2
2
2
2
2
)
2
)
2
)
2
)
2
)
2
)
2
)
2
H
H
H
H
H
H
H
2
2
2
2
2
)
)
)
)
)
300.8 73.2
457.4
8.7
9.8
3
3
6
7
6
H
676.4
685
133 4133.8
247.5 3812.9
2-Naph
-C
64.9
89.5
8.9
38 1 4-NO
2
6
H
4
407.5 264.2 2421.5
61.6 0.7 726.6
242.4 367.3 2149.2
246.7 57 374.1
3
4
4
9
0
1
1
1
1
CH
CH
2
CN
2
C
6
H
5
(CH
(CH
2
)
3
NH
NH
2
2
83.7
42.7
Fu
2
)
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
4
4
2
3
1
1
1
1
1
1
1
1
1
-
4-F-C
2-Naph
(CH NH
CH
Fu
4-F-C
6
H
4
(CH
(CH
2
2
)
2
2
C
C
6
6
H
H
5
5
(CH
(CH
2
2
)
3
3
NH
NH
2
2
451.4 30.4
506.7 5.6
365.3
819.2
0.6
10.5
32.5
29.7
7.1
)
)
44
2
)
2
2
(CH
2
)
5
CH
3
(CH
2
)
5
CH
3
435.8 2924.8 913.9
4
4
4
4
4
5
5
6
7
8
9
0
2
C
6
H
5
4
(CH
(CH
(CH
(CH
2
)
2
)
2
)
2
)
2
C
C
C
C
6
6
6
6
H
H
H
H
5
5
5
5
(CH
2
2
2
2
)
2
COOH
COOH
COOH
COOH
203.5
79.5
95.8
197
72
2330.5
335.5
419.3
680.6
45.8
2
2
2
(CH
(CH
(CH
)
)
)
2
2
2
2.4
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
H
23.5
72.5
8.8
6
2-Naph
CH
6.8
2
COOH (CH
2
)
5
CH
3
(CH
(CH CONHoleyl 737.9
250
2
)
5
CH
3
285.5 585.7
9.9
Fu
-
(CH
2
)
2
C
6
H
5
2
)
2
132 1807.1
12 74
5.5
AAZ
-
-
5.7
a. Mean from three different assays, by a stopped-flow technique (errors were in the range of ±5−10% of the reported values). Fu =
furyl; Naph = naphthyl
Generally, the inhibition data reported in Table 1 highlighted that mono-tailed compounds 1-7 were
medium to high nanomolar inhibitors of hCA I (K = 68.4 - 458.1 nM), II (K = 62.8 - 153.7 nM) and
I
I
XII (K = 55.4-113.2 nM), and weak inhibitors of hCA IV with inhibition constant (K ) values in the
I
I
low micromolar range (1.1 - 6.2 μM).
In detail, compounds 1 (R = C H ) and 5 (R = Fu) inhibited the off-target hCA I in the medium
1
6
5
1
nanomolar range (K = 95.3 and 68.4 nM, respectively), while compounds 2, 4 and 7 acted as weaker
I
inhibitors (K = 224.3 - 458.1 nM). In fact, the introduction of bulky substituents (2 and 4, K s of
I
I
224.3 and 458.1 nM) or the elongation of the chain (7, K of 278.4 nM) in R decreased the action
I
1
against hCA I when compared to compound 1.
The aryl-tailed compounds 1-6 acted as medium nanomolar inhibitors (K = 62.8 - 120.9 nM) against
I
hCA II, with compound 5 (R = Fu) being the single-tail isoform inhibitor. Compound 7 (R = CH CN)
1
1
2
reported instead the worst inhibition of action against hCA II (K = 153.7 nM).
I
HCA IV was the least inhibited by compounds 1-7. In this context, derivatives 2 (K = 1.6 μM), 3
I
(
K = 1.1 μM) and 5 (K = 1.5 μM) resulted to be significantly better inhibitors than the bulkier
I I
derivative 4 (R = 2-Naph, K value of 6.2 μM).
1
I
HCA XII was inhibited almost similarly by the single-tail compounds 1-7. Nonetheless, again
derivative 5 (R = Fu) stood out as the best inhibitor (K = 55.4 nM), whereas the cyanoalkyl- and
1
I
phenethyl-tailed compounds 6 and 7 exhibit K s above 100 nM.
I
ACS Paragon Plus Environment

Page 11 of 68
Journal of Medicinal Chemistry
1
2
3
Data in Table 1 showed that the development of 1-7 upon inclusion of two other tails to synthesize
compounds 18-50 significantly impacted the inhibition profiles against the panel of CA isoforms. In
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
fact, TTIs showed lightly decreased or markedly improved inhibition of hCA XII (K s = 0.6 - 302.5
I
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
nM). HCA IV remained the less inhibited isozyme, though inhibition improvement of one or two
orders of magnitude were testified for some compounds (K s = 45.8 - >10000 nM). On the whole, no
I
significant improvement of hCA I inhibiton was detected with TTIs (K s = 79.5 - 4210.4 nM). HCA
I
II showed the inhibition profiles most affected, both positively and negatively, upon inclusion of
additional tails on the scaffold of 1-7 (K s = 0.7 - 4562.9 nM).
I
To better discuss TTIs structure-activity relationship (SAR) from Table 1, compounds and related
data were distinguished in five subsets: (i) 18-29 (with R = C H ), (ii) 30-33, 44, 49 (with R = R =
1
6
5
2
3
(
CH ) CH ), (iii) 34-39 (R = (CH ) C H and R = (CH ) CN), (iv) 40-43 (R = (CH ) C H and R
2 5 3 2 2 2 6 5 3 2 2 2 2 2 6 5 3
=
(CH ) NH ), and (v) 45-48 (R = (CH ) C H and R = (CH ) COOH).
2 3 2 2 2 2 6 5 3 2 2
(i) In the first subset, compounds 18 and 20-29 were high nanomolar inhibitors of the ubiquitous off-
target hCA I with K values between 184.7 and 946.7 nM, while derivative 24 (R = R = CH CH )
I
2
3
2
3
showed the best inhibitory profile (K = 184.7 nM). Instead, compound 19 (R = CH CH and R =
I
2
2
3
3
CH C H ) resulted in the worst hCA I inhibitor among all synthesized compounds (K = 4210.4 nM).
2
6
5
I
The glaucoma-implicated isoform hCA II was inhibited in the nanomolar range (K = 8.3 - 843.8 nM)
I
and, in particular, the introduction of R = R = CH CH for compounds 18 (n = 1) and 24 (n = 2)
2
3
2
3
and R = CH CH and R = CH C H for derivative 25 (n = 2) increased the inhibition profile against
2
2
3
3
2
6
5
this isoform (K = 8.3, 8.9 and 79.6 nM, respectively). Thus, derivative 18 is the most hCA II selective
I
compound (CA I/ CA II = 94).
Only compounds 18, 24, and 26-28 inhibited hCA IV with K values in the range of 3.2 - 4.6 μM,
I
while the other compounds of this series showed no activity below 10 μM.
All derivatives potently inhibited the other glaucoma-associated isoform, hCA XII, with K values
I
below 100 nM, except for compounds 26 and 29 that were also the worst inhibitors among all the
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
synthesized compounds against this isoform (K = 280.0 and 302.5 nM). Compound 18 showed the
I
best inhibitory profile of this series (K = 43.9 nM).
I
The importance of the linker length (n =1,2) is pointed out from the activity analysis of this first
subset. In fact, the elongation of the chain between R and R /R increased the activity against hCA
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
I, II and IV which possess the smallest binding cavities, as a longer linker (n=2) can shift the tails
R /R towards the rim of the active site, removing the ligand-target steric encumbrance. On the other
2
3
hand, the larger active sites of hCA XII are able to host bulky substituents and the introduction of the
linker n=2, which drives the tails R /R out from the active site, may decrease the activity by
2
3
weakening the ligand-target interactions.
(ii) Comparing the second subset (30-33, 44, 49 with R = R = (CH ) CH ) compounds) with the first
2
3
2 5
3
subset R /R -analogues 22 and 28, it was highlighted that the introduction of Fu and CH CN in R
1
2
3
2
increased the activity against the off-target hCA I and hCA II, such as observed in compounds 31
hCA I K = 220.1 nM; hCA II K = 60.4 nM) and 33 (hCA I K = 395.9 nM; hCA II K = 52.5 nM).
(
I
I
I
I
On the other hand, for R = CH C H (30) and 2-Naph (32), the activity on hCA II strongly decreased
1
2
6
5
for both substituents (K = 780.8 nM and 4.5 μM, respectively), while a weak increase in inhibition
I
was observed for compound 30 (K = 518.4 nM) and a decrement for 32 (K = 541.4 nM) against
I
I
hCA I.
HCA IV was weakly inhibited by 30-32 with K values in the micromolar range of 3.1 - 3.4 μM.
I
Furthermore, the tail R = CH CN reduction of compounds 33 into amine 44 decreased the activity
1
2
on hCA II of 55 times (K = 2.9 μM) and increased the activity on hCA IV of 3 times (K = 913.9
I
I
nM). Instead, the swap of 33 nitrile into carboxylic acid 49 worsened the activity against hCA II by
1 times (K = 585.7 nM), but increased the inhibition profile against hCA IV by 76 times (K = 45.8
1
I
I
nM), obtaining the most potent and selective compounds against this isozyme (CA I/ CA IV = 6.2).
In the case of hCA XII, all compounds showed a good activity against the target and, in particular,
compounds 31 (K = 9.7 nM), 33 (K = 8.6 nM) and 49 (K = 9.9 nM) inhibited this isoform with K
I
I
I
I
ACS Paragon Plus Environment

Page 13 of 68
Journal of Medicinal Chemistry
1
2
3
in the low nanomolar range while 30, 32, and 44 acted as medium nanomolar inhibitors (K = 32.5 -
I
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
6
2.5 nM).
Generally, for this subset it was observed that the concomitant presence of R = R = (CH ) CH with
2
3
2 5
3
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a 2-Naph in R (32) worsened the activity by 19 times against hCA II (K = 4.5 μM) and increased
1
I
the activity by 1.5 times against hCA XII (K = 61.7 nM) with respect to the analog 22 (R = C H ),
I
1
6
5
improving the CA II/ CA XII selectivity from 2.5 to 74 times. Of note, the presence of a potentially
charged moiety in R such as (CH ) NH (44) or better CH COOH (49) increased the activity against
1
2 2
2
2
hCA IV, which possesses a wider hydrophilic half in the active site with respect to the other hCAs
with many acidic/basic residues at the middle rim of the cavity.
(iii) The third subset (34-39) is characterized by the introduction of a hydrophobic tail R =
2
(CH ) C H , a polar one R = (CH ) CN and a variable pendant R . Only compound 39 (R =
2 2 6 5 3 2 2 1 1
(CH CN) was a medium nanomolar inhibitor (K = 61.6 nM), which resulted to be the most potent
2
I
agent against the off-target hCA I, whereas 34-38 acted in the high nanomolar range (K = 300.8 -
I
7
77.3 nM).
The glaucoma-associated hCA II was potently inhibited by derivative 39 with a K in the
I
subnanomolar range (0.7 nM), resulting the most potent and third selective inhibitor against this
isozyme (CA I/ CA II = 88.0), while 35 (R = Fu) acted in the medium nanomolar range with K =
1
I
7
3.2 nM and derivatives 34, 36-38 showed K values between 133.0 - 368.5 nM.
I
The best inhibitors against hCA IV within this subset were 35 and 39 with K in the high nanomolar
I
range (457.4 and 726.6 nM, respectively) whereas 36-38 were low micromolar inhibitors with KI
values between 2.4-4.1 μM and derivative 34 (R = CH C H ) acted with K > 10 μM.
1
2
6
5
I
The target hCA XII was strongly inhibited by all compounds of the subset with compounds 35, 36
and 39 acting in a low nanomolar range (K = 8.7, 9.8 and 8.9 nM, respectively), while 34, 37 and 38
I
were medium nanomolar inhibitors (K = 75.5, 64.9 and 89.5 nM, respectively). In this case,
I
derivative 36 resulted in the third most selective inhibitor against hCA XII (CA I/ CA XII = 69.8).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The comparison of compounds 37 and 39 from subset iii with the second subset analogs 32 and 33
R = R = (CH ) CH ) pointed out that the substitution of R and R with the tails (CH ) C H and
(
2
3
2 5
3
2
3
2 2
6
5
(CH ) CN, respectively, generally increased the activity against hCA II and IV, with the opposite
2
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
effect against hCA I and no significant effect against hCA XII.
iv) The fourth series (40-43) was obtained by reducing R = (CH ) CN to obtain primary amine tails
(
3
2 2
in the aforesaid derivatives 34-37, introducing a potentially positively charged pendant. This
structural modification led to a general increment of the activity against hCA I, II, IV, and XII,
suggesting that a strong polar interaction is favorable for the binding and might take place in all five
active sites.
In detail, the four compounds resulted to be high nanomolar inhibitors of hCA I with K in the 242.4-
I
5
06.7 nM range. Moreover it is observed that 40 (R = CH C H ) and 41 (R = Fu) inhibited this
1 2 6 5 1
isoform with a two-fold potency (K = 242.4 and 246.7 nM, respectively) with respect to 42 (R = 4-
I
1
F-C H ) and 43 (R = 2-Naph), which showed a K of 451.4 and 506.7 nM, respectively.
6
5
1
I
Derivatives 40-43 were good inhibitors of the glaucoma-associated hCA II with K s in the high
I
nanomolar for 40 (K = 367.3 nM), medium nanomolar for 41 and 42 (K = 57.0 and 30.4 nM,
I
I
respectively) and low nanomolar range for 43 (K = 5.6 nM), which was the second most selective
I
obtained inhibitor against this isoform (CA I/ CA II = 90.5).
Interestingly, it was observed that the introduction of a positively charged tail increased the activity
against hCA IV at least 4 times for 40 (K = 2.1 μM), 1.2 times for 41 (K = 374.1 nM), 11 times for
I
I
42 (K = 365.3 nM), and 4.5 times for 43 (K = 819.2 nM) with respect to their analogs of the third
I
I
subset (34-37).
The glaucoma-related hCA XII was strongly inhibited by 42 with a subnanomolar K of 0.6 nM that
I
makes it the most potent and selective compound against this isoform (selectivity ratio CA I/ CA XII
=
752.3), whereas 40 (K = 83.7 nM), 41 (K = 42.7 nM) and 43 (K = 10.5 nM) acted with a K in the
I
I
I
I
medium nanomolar range.
ACS Paragon Plus Environment

Page 15 of 68
Journal of Medicinal Chemistry
1
2
3
(v) The fifth subset (45-48) obtained by the introduction of a potentially negatively charged tail in R3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
showed a general increment of the inhibition activity against hCA I, II, IV, and XII compared to their
analogs 34-37.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
In detail, compounds 46 (K = 79.5 nM) and 47 (K = 95.8 nM) acted as medium nanomolar inhibitors
I
I
against the off-target CA I, whereas the introduction of a more encumbering R (CH C H and 2-
1
2
6
5
Naph), such as in 45 and 48, lightly decreased the activity to the high nanomolar range (K = 203.5
I
and 197.0 nM, respectively).
The target hCA II was inhibited in the low nanomolar range by compound 46 (K = 2.4 nM), the
I
second most potent inhibitor against this isozyme, and in the medium nanomolar range by 45 (K =
I
72.0 nM), 47 (K = 23.5 nM) and 48 (K = 72.5 nM).
I
I
The inhibition profile against hCA IV was in the high nanomolar range for derivatives 46-47 (K =
I
3
35.5, 419.3 and 680.6 nM, respectively) and decreased for compound 45 with a K value of 2.3 μM.
I
Moreover, derivatives 46-48 were low nanomolar inhibitors of hCA XII (K = 7.1, 8.8 and 6.8 nM,
I
respectively), whereas 48 and 46 resulted to be the second and third most potent inhibitors of this
glaucoma-associated isoform, while compound 45 acted with a K of 29.7 nM.
I
Comparing the fourth (40-43) and the fifth subset (45-48), it was detected that the presence of R =
3
(CH ) COOH in place of amine tails shifted the activity against hCA I.
2
2
Finally, the loss of the hydrophilic tail R in 50 decreased the activity against hCA I (K = 737.9 nM),
3
I
II (K = 132.0 nM) and IV (K = 1.8 μM) without effects against hCA XII (K = 5.5 nM), obtaining
I
I
I
the second most potent and selective compound against this isoform (CA I/ CA XII = 134.2).
As pointed out by data in Table 1, single-tail inhibitors 1-7 showed rather flat inhibition profiles
against all tested hCAs and no marked isoform selectivity was detected. In contrast, the selectivity of
action is often enhanced with TTIs 18-50 (selectivity index, SI in Table S1, Supporting Information).
For instance, starting from compound 1 (R = C H ; SI CA I/ CA II = 1.0; CA I/ CA XII = 1.5; CA
1
6
5
II/ CA XII = 1.5; CA IV/ CA XII = 43.6), the introduction of various lipophilic pendants in R and
2
R (as in 18-29) decreased the activity against all isoforms, except for derivatives 18 and 24 where
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
the inhibition profile against hCA II and XII was increased. Interestingly, CA I/ CA II selectivity of
8-27 was improved up to a SI of 94.8 for compound 18. Compound 28 and 29 (CA I/ CA II = 0.5
1
and 0.3, respectively) were instead the most selective hCA I inhibitors of this subset.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Derivatives 18-28 also exhibited improved selectivity for hCA XII over hCA I (SI 2.3 - 51.0), whereas
2
9 showed a greater action against hCA I (SI I/XII = 0.7). Moreover, derivatives 19-23 and 26-29
showed an increased selectivity for hCA XII over CA II (SI 1.8 - 9.5), in contrast with 24 and 25
more active against hCA II (SI 0.14 and 0.9). Within this subset, improved selectivity profiles for
hCA IV over hCA I and II were not detected. CA IV/ XII selectivity increased up to 64.3 - >144.1
for the subset 18-25.
In comparison to the single-tail derivative 2, compound 38 (R = 4-NO -C H , R = (CH ) C H , R
3
1
2
6
4
2
2 2
6
5
=
(CH ) CN) showed an increased selectivity for hCA XII over hCA I, II and IV (SI I/XII 4.6, II/XII
2 2
3
.0, IV/XII 27.1), while I/II selectivity showed a decrease (SI 1.5).
While derivative 3 (R = 4-F-C H ) showed SIs equal to I/II 1.4, I/ IV 0.1, I/ XII 1.9, II/XII 1.3, IV/
1
6
4
XII 19.9, the addition in R of (CH ) C H and (CH ) CN (36), (CH ) NH (42) and (CH ) COOH
2
2 2
6
5
2 2
2 3
2
2 2
(
47) in R led to remarkable results in terms of selectivity of action. In detail, selectivity was increased
3
for hCA II over hCA I and for hCA XII over hCA I, II and IV for compounds 36 (I/ II = 5.1, I/XII =
69.8, II/XII = 13.6, IV/XII = 421.8), 47 (I/ II = 4.1, I/ XII = 10.9, II/XII = 2.7, IV/XII = 47.6) and
even more in derivative 42 (I/ II = 14.9, I/XII = 753.3, II/XII = 50.7, IV/XII = 608.8). Notably, the
introduction of an amine moiety in R significantly shifted the selectivity toward hCA XII making
3
compound 42 752.3 times more active against the glaucoma-associated isoform hCA XII than the
off-target hCA I. 42 also showed the best CA IV/ CA XII selectivity index with a ratio of 608.8. The
nature of R can also be assumed to be responsible for a >1 SI for hCA IV over I.
3
Variable outcomes in terms of selectivity of action were observed appending R and R tails on the
2
3
2
-Naph single-tail 4 (SIs CA I/ CA II = 5.3, CA I/ CA XII = 5.8, CA II/ CA XII = 1.1, CA IV/ CA
XII = 79.5) yielding 32 (R = R (CH ) CH , R ), 37 (R = (CH ) C H , R = (CH ) CN), 43 (R =
2
3
2 5
3
3
2
2 2
6
5
3
2 2
2
(
CH ) C H , R = (CH ) NH ) and 48 (R = (CH ) C H , R = (CH ) COOH). In fact, I/II selectivity
2 2 6 5 3 2 3 2 2 2 2 6 5 3 2 2
ACS Paragon Plus Environment

Page 17 of 68
Journal of Medicinal Chemistry
1
2
3
decreased for derivative 37 (I/II SI 2.8) and 48 (I/II SI 2.7) up to the inversion displayed by 32 (SI
.1). In contrast it strongly increased with amine 43 (I/II SI 90.5). I/XII selectivity was improved for
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
all these derivatives in the order 32 (I/XII SI = 8.8), 37 (I/XII SI = 10.6), 48 (I/XII SI = 29.0) 43 (I/XII
SI = 48.3). The lipophilic TTI 32 showed great selectivity for hCA XII over hCA II and hCA IV (SI
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
=
74.0 and > 162.1, respectively). Likewise acted carboxylic acid 48 (SI II/ XII = 10.7, IV/ XII =
00.1). A low II/XII SI increase was observed for nitrile 37 (SI 3.8), while an inversion was detected
for amine 43 (SI II/XII = 0.5). The latter also showed an improved I/IV SI value (0.6) with respect to
(0.07).
1
4
The TTI development of derivative 5 (SIs I/II = 1.1, I/IV = 0.04, I/XII = 1.2, II/XII = 1.1, IV/XII =
28.6), to give 31 (R = R (CH ) CH , R ), 35 (R = (CH ) C H , R = (CH ) CN), 41 (R =
2
3
2 5
3
3
2
2 2
6
5
3
2 2
2
(
CH ) C H , R = (CH ) NH ), 46 (R = (CH ) C H , R = (CH ) COOH) and 50 (R = (CH ) C H ,
2 2 6 5 3 2 3 2 2 2 2 6 5 3 2 2 2 2 2 6 5
R = (CH ) CONHoleyl), overall increased the selectivity for hCA II over hCA I (3.6, 4.1, 4.3, 33.1
3
2 2
and 5.6, respectively). CA I/IV SIs were overall improved (0.2-0.7) with respect to 5 (except 31) but
not reversed. Interestingly, the reduction of nitrile 35 into amine 41 did not lead to variations in the
I/IV selectivity, while the hydrolysis to carboxylic acid 46 decreased it by 3 times. I/IV SI increased
instead twice upon formation of amide 50.
The selective index for hCA XII over hCA I increased for all five derivatives, greatly with 31 (SI
2
2.7), nitrile 35 (SI 34.6) and amide 50 (SI 134.1), less with amine 41 (SI 5.8) and carboxylic acid 46
(I/XII = 11.2). These compounds also showed selectivity for hCA XII over hCA II with SI of 6.2 (31),
8.4 (35), 1.3 (41), 24.0 (24), except for the carboxylic acid 46 (SI 0.3).
Except for derivative 41 (SI 8.8), selectivity for hCA XII over IV ratio was enhanced for compound
3
1 (SI 325.1), 35 (SI 52.6), 46 (SI 47.3) and 50 (SI 328.6) respect to the lead 5.
The functionalization of 6 (I/ II = 0.7, I/IV = 0.02, II/IV = 0.03, I/XII = 0.9, II/XII = 1.4, IV/XII =
9.0) with R and R produced derivatives 33 (R = R (CH ) CH ) and 39 (R = (CH ) C H , R =
4
2
3
2
3
2 5
3
2
2 2
6
5
3
(
CH ) CN), that acted 7.5 and 88.0 times more efficiently against hCA II over hCA I. Moreover,
2 2
compound 33 showed an increment of SI I/XII (46.0), II/XII (6.1) and IV/XII (404.5). Instead,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
derivative 39 showed a drastically improved action against hCA II over hCA XII (CA II/ CA XII =
.1), and improved SIs I/XII (6.9) and IV/XII (81.6). The reduction and hydrolysis of the nitrile of
derivative 33 to give amine 44 and carboxylic acid 49 led to a selectivity against hCA I over hCA II
CA I/ CA II = 0.2 and 0.5, respectively). Interestingly, 49 was the first-in-class selective hCA IV
0
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
inhibitor over CA I (SI 6.2) and hCA II (SI 12.8) and also showed the lowest IV/XII SI (4.6). Finally,
amine and carboxylic acid 44 and 49 showed increased II/XII SI (90.0 and 59.2, respectively).
The R /R development of compound 7 (R = CH C H , I/II = 3.1, I/XII = 2.7, II/XII = 0.9, IV/ XII
2
3
1
2
6
5
=
34.4) to give 30 (R = R (CH ) CH , R ), 34 (R = (CH ) C H , R = (CH ) CN), 40 (R =
2 3 2 5 3 3 2 2 2 6 5 3 2 2 2
(
CH ) C H , R = (CH ) NH ) and 45 (R = (CH ) C H , R = (CH ) COOH) decreased I/II
2 2 6 5 3 2 3 2 2 2 2 6 5 3 2 2
selectivity up to a total inversion with derivatives 30 (SI 0.7) and 40 (SI 0.7). On the contrary, an
improvement was detected in the selectivity against hCA XII over hCA I (SI 2.9 - 10.3), hCA II (SI
2
.4 - 12.5) and hCA IV (SI 54.6 - >137.5), except for compound 45 that showed a worsening in the
IV/XII selectivity (SI 25.9) when compared to the lead 7.
X-ray Crystallography
Co-crystallization of hCA II with selected three-tailed inhibitors resulted in solved structures with
resolutions between 1.35-1.62 Å (Figures 4-6, Table 2). For all the inhibitors studied the
benzenesulfonamide was orientated with the zinc binding group displacing the active site zinc-bound
water (ZBW) and forming a hydrogen bond between the amide backbone of Thr199 and oxygen of
sulfonamide (2.8-3.0Å). Therefore, with the benzenesulfonamide binding in identical fashion, any
differences in observed binding affinity most likely result from differences in the tail regions.
ACS Paragon Plus Environment

Page 19 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4. X-ray crystallography: surface representation of hCA II with inhibitors 34, 41, 42, 46, and
4
8 bound within the active site (PDBs 6WQ4, 6WQ5, 6WQ7, 6WQ8 and 6WQ9, respectively).
Figure 5. Electron density of A) 34, B) 41, C) 42, D) 46, and E) 48 in hCA II active site with a sigma
of 1.0.
Compound 34 showed a well observed omit map electron density, indicating good binding with a
high binding occupancy (PDB 6WQ4, Figure 5B). The T phenethyl was accommodated in the
1
lipophilic pocket lined by Val135, Leu198, Pro202, and Leu204, whereas the phenethyl in T lied
2
above Phe131, forming contacts with the α-helix portion constituted by residues 130-136 (Figure 6B).
A water bridged H-bond took place between the ligand amide carbonyl group and Gln92 side chain
NH . The hydrophilic CN tail extended into bulk solvent.
2
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Compound 41 exhibited a weaker observed omit map electron density around the phenethyl and
aminopropyl tails (PDB 6WQ5, Figure 5C). While the furyl ring took the place occupied by the T1
benzene ring of compound 34, the phenethyl tail in T formed again interactions with Phe131 and
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
residues nearby (Figure 6C). The amino group in T protonated at physiological pH was exposed to
3
bulk solvent.
Compound 42 showed a weak observed omit map electron density at the end of the aminopropyl tail
(PDB 6WQ7, Figure 5D). The switch from a furyl (41) to a 4-F-benzyl (42) in T markedly shifted
1
the tails of the ligand within the CA II active site, probably because the pocket hosting the furyl ring
cannot accommodate additional steric hindrance. This did not occur with the T phenethyl group of
1
34 as the presence of an additional carbon unit allowed a torsion preserving a 41-like binding mode.
The 4-F-benzyl formed hydrophobic contacts with Val135, Leu198, and Phe131, and an edge-to-face
π-π stacking with the latter residue benzene ring (Figure 6D). The T phenethyl of 42 lodged over the
2
lipophilic portion composed by Trp5, Phe20, Pro201, and Pro202, while the protonated amino group
in T was again exposed to bulk solvent.
3
Compound 46 showed a strong observed omit map electron density, which is indicative of a high
binding occupancy (PDB 6WQ8, Figure 5E). The T and T tails of the ligand adopted analog
1
2
positions within the active site to those of compound 41 (Figure 6E). The carboxylic tail was oriented
-
toward the hydrophilic region within the active site, where the COOH, presumably as COO , is
involved in a water-mediated H-bond network with Asn62, Asn67, Glu69, and Gln92.
Compound 48 had a weaker observed omit map electron density near the carboxylic acid tail (PDB
6
WQ9, Figure 5F). As it occurred with compound 42, the additional steric hindrance in T moved the
1
naphthyl ring away from the pocket occupied by the furyl core of 41 (Figure 6F). Nonetheless, the
-
intense H-bond network between the COO moiety in T and Asn62, Asn67, Glu69, and Gln92
3
prevented the T /T branching N atom to move toward Trp5. This produced a switch between the
2
3
positioning of T and T tails for 48 with respect to 42. The naphthyl portion in T accommodated
1
2
1
ACS Paragon Plus Environment

Page 21 of 68
Journal of Medicinal Chemistry
1
2
3
above the lipophilic pocket lined by Leu198, Pro201, Pro202, and Leu204, whereas the phenethyl in
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
T2 interacted with Phe131 and other α-helix composing residues by Van der Waals contacts.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 6. X-ray crystallography: active site view of hCA II in adduct with A) no inhibitor (PDB
3
KKX), B) 34 (PDB 6WQ4), C) 41(PDB 6WQ5), D) 42(PDB 6WQ7), E) 46(PDB 6WQ8), and F)
8 (PDB 6WQ9). H-bonds and π-π stackings are represented as black and red dashed lines,
4
respectively. Water molecules involved in water-bridged H-bonds are shown as red spheres. Amino
acids are labelled with one letter symbols: D, Asp; E, Glu; F, Phe; H, His; I, Ile; L, Leu; N, Asn; P,
Pro; Q, Gln; T, Thr; V, Val; W, Trp.
It can be noted that the binding mode exhibited by compound 46 was the most efficient for promoting
hCA II inhibition because of a 10-fold better K (2.4 nM) than the second-best derivative among those
I
co-crystallized (42, K of 30.4 nM). Considering the similar interactions observed for tails T and T
2
I
1
with respect to compounds 34 and 41, this enhanced efficacy might be consequent of the extended
water-mediated H-bond network the carboxyethyl pendant formed with the hydrophilic portion of the
binding cleft. Interestingly, the binding mode exhibited by 42, though most departed from those of
the other co-crystallized ligands, produced the second-best inhibition of hCA II. Swapping the furyl
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ring of 46 with the naphthyl of 48 significantly lowered the efficiency of the binding mode, as the bi-
cycle cannot be accommodated in the Leu198, Pro201, Pro202, and Leu204 pocket and was partially
exposed to bulk solvent. The exposure of the markedly less hydrophilic cyanoethyl tails of 34 to bulk
solvent is the presumable reason of the drop of CA II inhibition exhibited by the ligand.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 2. X-ray crystallography data collection and refinement statistics of inhibitors bound hCA II
crystal structures
Inhibitor
34
41
42
46
48
PDB
6WQ4
6WQ5
6WQ7
P21
6WQ8
6WQ9
Space group
Cell dimensions:
a, b, c, β (Å, deg)
Resolution (Å)
Highest Resolution Shell (Å)
Total reflections
I/σ(I)
42.4, 41.5,
42.3, 41.4,
42.1, 41.3
42.4, 41.3,
42.3, 41.3,
72.3, 104.3 72.3, 104.4 72.1, 104.3 72.4, 104.4 72.3, 104.4
29.19 - 1.35 25.34 - 1.30 25.28 - 1.30 28.75 - 1.41 21.13 - 1.30
(1.40 - 1.35) (1.35 - 1.30) (1.35 - 1.30) (1.46 - 1.41) (1.35 - 1.30)
9536
8627
8885
14181
12.5 (2.4)
3.3 (3.1)
8927
16.3 (2.7)
3.1 (2.2)
14.5 (1.6)
3.1 (2.2)
15.5 (1.7)
3.1 (1.9)
20.6 (2.5)
3.2 (2.3)
Redundancy
Completeness (%)
98.0 (82.5) 95.8 (68.1) 97.4 (81.5) 99.4 (97.8) 94.0 (66.6)
4.10 (25.6) 4.33 (49.0) 3.87 (40.7) 5.44 (39.3) 3.29 (35.7)
Rsyma
Rcrysb
Rfreec
Rpimd
15.4 (22.7) 16.0 (29.3) 16.0 (26.9) 14.6 (20.5) 14.9 (22.8)
17.3 (25.2) 18.1 (31.7) 18.1 (30.5)
17. (22.4)
17.3 (28.5)
2.67 (19.2) 2.82 (37.1) 2.53 (34.6) 3.52 (26.0) 2.12 (26.5)
#
of atoms: protein, ligand,
2049, 52,
209
2075, 44,
239
2076, 46,
239
2073, 87,
235
2080, 56,
248
water
Protein residues
Ramachandran stats (%):
favored, allowed
257
257
257
258
257
9
6.1, 3.9
96.9, 3.1
96.9, 3.1
96.1, 3.9
96.5, 3.5
2
Avg. B-factors (Å ): main-,
13.9, 14.7
16.7, 21.9
15.4, 16.5
25.7, 24.0
16.4, 17.4
27.5, 24.5
15.1, 16.5
29.3, 24.1
14.9, 16.4
31.5, 25.2
side-chain, inhibitor, solvent
Rmsd for bond lengths,
0
.008, 1.05 0.008, 1.04 0.008, 1.04 0.009, 1.09 0.008, 1.07
angles (Å, deg)
a
b
c
Rsym = (Σ|I – ⟨I⟩|/Σ ⟨I⟩) × 100. Rcryst = (Σ|F – F |/Σ|F |) × 100. Rf is calculated in the same way as Rcryst except it is for
o
c
o
ree
d
e
data omitted from refinement (5% of reflections for all data sets). Rpim = [(Σ√1/N – 1)Σ|I – ⟨I⟩|/Σ ⟨I⟩] × 100. Values in
parentheses correspond to those of the highest resolution shell.
In Silico Study
The crystallographic screening was complemented with docking calculations to also study hCA
43
44
isoforms not included in the crystallographic study; hCA I (PDB 2NMX), hCA IV (PDB 1ZNC),
ACS Paragon Plus Environment

Page 23 of 68
Journal of Medicinal Chemistry
1
2
3
45
and hCA XII (PDB 1JD0). The in silico study was performed on the single-tail derivatives 1-7 and,
among TTIs, the most potent compounds against each isoform and cocrystallized ligands assembling
a subset of 7 derivatives (34, 39, 41, 42, 46, 48 and 49) and predicting their binding to hCAs I, IV
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
4
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
and XII as well as CA II (PDB 5LJT) when missing (Figures S1-S6, Supporting Information). The
binding orientations resulting from docking were refined with a MM-GBSA method simulating a
water media (VSGB method) for improving the comparison with the crystallographic outcomes. The
efficiency of the adopted protocol with three-tail compounds was validated by application to the
crystallographic target/inhibitor adducts described above. Despite the absence of water molecules,
crystallographic/simulated ligand RMSDs were computed below 1.0 Å, with a main deviation at the
level of aliphatic tails (e.g. the carboxylate pendant in compound 46, Figure S1, Supporting
Information).
Predictably, derivatives 1-7 showed interactions within the hCA I, II and XII active sites limited to a
portion of the hydrophobic half of the cavity (Figure S2, Supporting Information). As a result, this
produces inhibition profiles devoid of selectivity and thus promiscuous. The absence of a
hydrophobic half in the active site of hCA IV led the tails of 1-7 towards alternative pockets according
to the nature of the pendants, and on the whole, reduces the inhibition efficacy up to a micromolar
range.
Panels A and B of Figure 7 depict the predicted binding mode of 39 to hCA I and II, respectively, as
the most active inhibitor against these two isoforms. HCA I shows a narrower active site than hCA II
because of specific amino acid mutations such as Thr/His200, Asn/His67, Leu/Tyr204 and, mostly,
Ile/Phe91 (Figure S3, Supporting Information). As a main result of the latter mutation, T and T are
2
3
shifted towards the lipophilic pocket lined by Trp5, Val62 (solely present in hCA I), His64, and
Pro201, where the cyanoethyl moiety receives a H-bond by Trp5 NH. The cyanoethyl in T engages
1
interactions with the hydrophilic half of the binding cavity among which a H-bond with Asn69 forms.
As for hCA II, the tail of the ligand occupies on the whole a region nearer to the hydrophobic half of
the active site. In fact, the phenethyl in T , as observed in crystallography with similar ligands, lies
2
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
above Phe131 interacting with residues 13-135 of the α-helix. The position of the two cyanoethyl
portions is almost inverted when compared to hCA I: the moiety in T receives H-bond by His64 NH,
1
whereas the nitrile group in T3 is in H-bond distance with Asn67. As compound 39 uniquely
possesses, among the selected derivatives (Figure S4, Supporting Information), two aliphatic,
partially polar but non-protic tails (cyanoethyl), it can be supposed a favorable complementarity with
the narrow and rather lipophilic active sites of hCA I and II, which drives the most potent action here
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
reported against the two ubiquitous isoforms (K s of 61.6 and 0.7 nM respectively).
I
In fact, the greater steric hindrance produced by another phenethyl in T (compound 34, Figure 6B
1
and S4A, Supporting Information) lowered the inhibition potency by 10 and 500 times against hCA
I and II, respectively. Solely the presence of a carboxyethyl tail in T of compound 46 (but not 48,
3
presumably because of the unwieldy naphthyl ring in T ) leads the inhibitor action against isoform I
1
(
K of 79.5 nM) and II (K of 2.4 nM) to the level of compound 39, likely because of the interactions
I
I
of the carboxylate with the hydrophilic portion of the binding cavity (Figures 6E-F and S4B,
Supporting Information).
The active site of hCA IV is the most particular among those of hCAs as largely losing the
hydrophibic/hydrophilic division common to most other catalytically active isoforms. In fact, α-helix
130-135 is absent and replaced by an extended loop which protrudes to bulk solvent. At the same
time, the hydrophobic half of the binding cavity is replaced by a region rich in polar amino acids such
as Lys91, Glu123, Thr202, Asp204, Lys206, and Glu138 (Figure 7C and S5, Supporting
Information). As a result, this isoform is the less inhibited by TTIs, with K s above 100 nM, except
I
for derivative 49, that solely possesses a carboxylate function in T . As shown in Figure 7C, the latter
1
forms a salt bridge with Lys206, and this conformation also leads the protonated T /T N branching
2
3
atom in salt bridge with Asp204. Other ligands, such as 41, 42 and 48, were also predicted to form
salt bridges within the hCA IV active site (Figure S5, Supporting Information), but involving
carboxylate or amine moieties in T . As a result, the ligands adopt conformation which do not allow
3
the formation of two salt bridges with the polar pocket of the active site, as observed for 49. As the
ACS Paragon Plus Environment

Page 25 of 68
Journal of Medicinal Chemistry
1
2
3
latter shows a K of 45.8 nM despite two hexyl groups protruding to bulk solvent (Figure 7C), it can
I
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
be supposed that their replacement with less lipophilic groups might even increase the inhibition
efficiency of this membrane associated CA.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Isoform hCA XII maintains an overall hydro/lipophilic partition in its wide active site (Phe/Ala131
with respect to hCA II), but specific mutations with respect to CA II, that are Asn/Lys67, Ile/Thr91,
Gly/Ser132, Val/Ser135, and Leu/Asn204 significantly enhance the hydrophilicity of the binding
cavity (Figure 7D and S6, Supporting Information). It should be noted that compound 42 shows the
unique subnanomolar K value against a tumor associated CA (K of 0.6 nM against CA XII). The
I
I
peculiar active site architecture of hCA XII indeed drives a favorable disposition of the three tails of
the ligands: the T 4-F-phenyl accommodates in the pocket lined by Trp5, His64, Asn62, and Lys67;
1
the T phenethyl lies over the most lipophilic cleft of the binding pocket, made by Val121, Thr91 and
2
Ala131; the propylamine pendant in T is involved in a bifurcated H-bonds system with the side chain
3
Asn136 and Ser132 backbone CO (Figure 7D). In contrast, compound 41, having a furyl ring in place
of the 4-F-phenyl of 42, exhibits a very different binding orientation in the hCA XII active site (Figure
S6, Supporting Information), as it occurred with CA II as well (Figure 6 and 7).
Figure 7. In silico predicted binding conformations for the adducts A) 39/hCA I, B) 39/hCA II, C)
4
9/hCA IV, and D) 42/hCA XII. H-bond and salt bridge interactions are depicted as black and red
dashed lines, respectively.
Intraocular Pressure Lowering Activity
For a first pharmacological application of the proposed approach, we selected the inhibitors showing
the best concomitant action against hCA II, IV and XII (39, 46 and 47) for evaluating their intraocular
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
pressure (IOP) lowering activity in a rabbit model of glaucoma (Figure 8). The compounds showed a
sufficient water solubility to be formulated as 1% eye drops and DRZ hydrochloride 1% was used as
reference compound and hydroxypropylcellulose 0.05% as vehicle in the experimental setting. The
compounds were formulated and administered as 1% eye drops to rabbits with high IOP, induced by
the injection of 0.05 mL of hypertonic saline solution (5% in distilled water) into the vitreous of both
eyes. As depicted in Figure 4, at 30 min post-instillation only compounds 39 and 47 decreased the
IOP by 1.0 and 1.3 mmHg, respectively such as DRZ (-1.0 mmHg), while 46 was inactive. At 60 min
after administration all compounds triggered the maximum IOP reduction, where 39 and 46 showed
a maximal IOP lowering activity of 3.0 and 3.3 mmHg, respectively. Instead, compound 47 resulted
the most effective, decreasing the IOP of 4.8 mmHg in a comparable manner of DRZ (-5.4 mmHg).
After 120 min, a decrement of the effect was observed for all compounds with 39 and 46 that
decreased IOP by 0.3 and 1.0 mmHg, while the standard DRZ showed to be less effective than 47 (-
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
.2 mmHg) with an IOP reduction of 2.8 mmHg. Uniquely, compounds 46 (-1.2 mmHg) and 47 (-2.4
mmHg) protracted their action at 240 min post-instillation, whereas compound 39 was inactive. In
particular, 47 showed a similar profile to the standard DRZ (-3.0 mmHg).
Figure 8. Drop of intraocular pressure (ΔIOP, mmHg) versus time (min) in hypertonic saline induced
ocular hypertension in rabbits, treated with 50 μL of 1% solution of compounds 39, 46 and 47, and
DRZ as the standard. Hydroxypropylcellulose at 0.05% was used as vehicle. Data are analyzed with
ACS Paragon Plus Environment

Page 27 of 68
Journal of Medicinal Chemistry
1
2
3
4
2
-way ANOVA followed by Bonferroni multiple comparison test. * p < 0.05 47 vs vehicle at 60’; **
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
p < 0.01 DRZ vs vehicle at 60’.
CONCLUSIONS
The tail approach was proposed already in 1999 and progressively developed with a variety of
chemical scaffolds up to the first report of dual-tail design in 2015 to target both the hydrophobic and
hydrophilic halves of hCAs’ active sites. Such an undoubtedly favorable approach in the field of
hCAs is still the main strategy used for obtaining CAIs and led us to propose its further development
by the incorporation of three tails onto a benzenesulfonamide CA inhibitory scaffold. In fact, we
deem the simple hydrophobic/hydrophilic division of hCAs binding pocket not totally sufficient
anymore, because of many accessory pockets existing in each hCA isoform. This proof-of-concept
study reported here was carried out by the design and synthesis of thirty-two benzenesulfonamide
derivatives of the TTI type (Figure 3) screened against a first set of hCAs, that are I, II, IV, and XII,
and comparing the results with the corresponding single-tail derivatives 1-7 (Table 1).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Our results showed that the development of 1-7 upon inclusion of two other tails to give compounds
1
8-50 significantly impacted the inhibition profiles in terms of potency and selectivity of action. On
the whole, it should be noted that the inclusion of three lipophilic tails in the TTI structure, such in
compounds 18-32, did not produce noteworthy outcomes in terms of potency and selectivity against
the tested hCAs, with very flat SAR within the subset, except for a few exceptions. In contrast,
increasing the polarity of at least one tail (starting from compound 33) resulted in a great variability
of potencies and selectivities according to the type of tails included in T , T and T .
1
2
3
The structural study made by X-ray crystallography with hCA II and in silico tools with the other
isozymes pointed out the limited and almost superimposable interactions that the tail of 1-7 can
establish within the CAs active site. In contrast, we demonstrated that the TTI derivatives show a
greater occupancy of the binding cavities with a great variability among isoforms, that contribute to
the development of improved selectivity of action.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The structural study together with the SAR formed a profile of the combinations of tail features which
can be favorable for the interaction with the distinct hCA isoforms. As an outcome of this preliminary
investigation, we can infer that inhibition of hCA I and II, possessing narrow and markedly lipophilic
active sites, can be promoted by inclusion of a lipophilic tail and two small half polarity pendants in
the TTI structure (e.g. compound 39) or alternatively two not too bulky tails and a polar one (e.g.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
compound 46). The marked polarity of hCA IV active site makes a significantly polar tail in T1
(nearby the main CA inhibitory scaffold) necessary to attain a low nanomolar inhibition (e.g.
compound 49). HCA XII and its wide hydrophobic binding pocket better accommodate almost all
ligands with respect to hCA II, and thus most tail combinations produce efficient inhibition of the
isozyme. The combination of lipophilic and polar tails coexisting with a medium polarity one even
led to a subnanomolar hCA XII inhibitor (compound 42).
For a first pharmacological application of the proposed approach, three TTIs were selected because
of their potent and concomitant inhibition of hCA II, IV and XII (CAs implicated in glaucoma) and
were assessed in vivo in a rabbit model of the disease. Compound 47 showed the capability of
lowering IOP as efficiently as the clinically used DRZ up to 120 min post administration.
The outcomes of this proof-of-concept study represent a firm starting point for optimizing the general
TTI design as well as to produce a wider set of tail combinations to even improve the ligand/isoforms
matching in search of new CAI candidates in the treatment of spreading diseases such as glaucoma,
tumors, neuropathic pain and inflammation. It should be stressed that an analogous approach might
be extended to other multi-isoform metalloenzymes to improve the outcomes in terms of selectivity
of action.
EXPERIMENTAL SECTION
Chemistry. Anhydrous solvents and all reagents were purchased from Sigma-Aldrich, Fluorochem
and TCI. All reactions involving air- or moisture-sensitive compounds were performed under a
nitrogen atmosphere using dried glassware and syringes techniques to transfer solutions. Nuclear
ACS Paragon Plus Environment

Page 29 of 68
Journal of Medicinal Chemistry
1
2
3
1
13
magnetic resonance ( H-NMR, C-NMR) spectra were recorded using a Bruker Advance III 400
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
MHz spectrometer in DMSO-d . Chemical shifts are reported in parts per million (ppm) and the
6
coupling constants (J) are expressed in Hertz (Hz). Splitting patterns are designated as follows: s,
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
singlet; d, doublet; t, triplet; q, quadruplet; m, multiplet; bs, broad singlet; dd, double of doublets. The
assignment of exchangeable protons (OH and NH) was confirmed by the addition of D O. Two
2
tautomeric forms of the amide bond were detected for compounds 8-50 which partially double the
1
13
signals in the H- and C-NMR spectra. Analytical thin-layer chromatography (TLC) was carried out
on Sigma Aldrich silica gel F-254 plates. Flash chromatography purifications were performed on
Sigma Aldrich Silica gel 60 (230-400 mesh ASTM) as the stationary phase and ethyl acetate/n-hexane
or MeOH/DCM were used as eluents. Melting points (mp) were measured in open capillary tubes
with a Gallenkamp MPD350.BM3.5 apparatus and are uncorrected.
Compounds 1-7, 18-50 were ≥95% pure. The purity of the final compounds was determined by HPLC
analysis performed by using an Agilent 1200 Series equipped by autosampler, binary pump system
and diode array detector (DAD). The column used was a Luna PFP 30 mm length, 2 mm internal
–
1
diameter and 3 μm particle size (Phenomenex, Bologna, Italy) at constant flow of 0.25 mL min ,
employing a binary mobile phase elution gradient. The eluents used were 10 mM formic acid and 5
mM ammonium formate in mQ water solution (solvent A) and 10 mM formic acid and 5 mM
ammonium formate in methanol (solvent B) according to the elution gradient as follows: initial at
9
0% solvent A, which was then decreased to 10% in 8 min, kept for 3 min, returned to initial
conditions in 0.1 min and maintained for 3 min for reconditioning, to a total run time of 14 min. The
-1
stock solution of each analyte was prepared in methanol at 1.0 mg mL and stored at 4 °C. The sample
solution of the analyte was freshly prepared by diluting its stock solution up to a concentration of 10
-1
μg mL in mixture of mQ water:methanol 50:50 (v/v) and 5 μL was injected into the HPLC system.
The solvents used in HPLC measures were methanol (Chromasolv grade), purchased from Sigma-
Aldrich (Milan - Italy), and mQ water 18 MΩ cm, obtained from Millipore's Simplicity system
(Milan-Italy).
ACS Paragon Plus Environment