Design, Synthesis, and Biological Evaluation of Tetrahydro-β-carboline Derivatives as Selective Sub-Nanomolar Gelatinase Inhibitors

Article abstract of DOI:10.1002/cmdc.201800237

Targeting matrix metalloproteinases (MMPs) is a pursued strategy for treating several pathological conditions, such as multiple sclerosis and cancer. Herein, a series of novel tetrahydro-β-carboline derivatives with outstanding inhibitory activity toward MMPs are present. In particular, compounds 9 f, 9 g, 9 h and 9 i show sub-nanomolar IC₅₀ values. Interestingly, compounds 9 g and 9 i also provide remarkable selectivity toward gelatinases; IC₅₀=0.15 nm for both toward MMP-2 and IC₅₀=0.63 and 0.58 nm, respectively, toward MMP-9. Molecular docking simulations, performed by employing quantum mechanics based partial charges, shed light on the rationale behind binding involving specific interactions with key residues of S1′ and S3′ domains. Taken together, these studies indicate that tetrahydro-β-carboline represents a promising scaffold for the design of novel inhibitors able to target MMPs and selectively bias gelatinases, over the desirable range of the pharmacokinetics spectrum.

Full text of DOI:10.1002/cmdc.201800237

DOI: 10.1002/cmdc.201800237
Full Papers
Design, Synthesis, and Biological Evaluation of Tetrahydro-
b-carboline Derivatives as Selective Sub-Nanomolar
Gelatinase Inhibitors
Giuseppe Felice Mangiatordi,[a, b] Tatiana Guzzo,[c] Eugenio Claudio Rossano,[c]
Daniela Trisciuzzi,[a] Domenico Alberga,[a] Giovanni Fasciglione,[d] Massimiliano Coletta,[d]
Alessandra Topai,*[c] and Orazio Nicolotti*[a]
Targeting matrix metalloproteinases (MMPs) is a pursued strat-
egy for treating several pathological conditions, such as multi-
ple sclerosis and cancer. Herein, a series of novel tetrahydro-b-
carboline derivatives with outstanding inhibitory activity
toward MMPs are present. In particular, compounds 9 f, 9g, 9h
and 9i show sub-nanomolar IC50 values. Interestingly, com-
pounds 9g and 9i also provide remarkable selectivity toward
gelatinases; IC50 =0.15 nm for both toward MMP-2 and IC50 =
0.63 and 0.58 nm, respectively, toward MMP-9. Molecular dock-
ing simulations, performed by employing quantum mechanics
based partial charges, shed light on the rationale behind bind-
ing involving specific interactions with key residues of S1’ and
S3’ domains. Taken together, these studies indicate that tetra-
hydro-b-carboline represents a promising scaffold for the
design of novel inhibitors able to target MMPs and selectively
bias gelatinases, over the desirable range of the pharmacoki-
netics spectrum.
Introduction
The matrix metalloproteinase (MMP) family includes over 25
zinc-/calcium-dependent endopeptidases, the primary role of
which consists of the degradation of the extracellular matrix
(ECM) and the release of growth factors, cytokines and chemo-
kines.[1–3] According to their in vivo substrate, they can be clas-
sified as membrane-type MMPs, matrilysins, gelatinases, colla-
genases and stromelysins.[4,5] These proteins are involved in
several physiological and pathological processes implicated in
inflammatory reactions. In particular, their over-activation is re-
sponsible for anomalous tissue degradation, leading to many
pathological conditions, such as osteoarthritis, multiple sclero-
sis, tumour growth and metastasis.[6–9] Such severe pathologies
prompted both academia and industry to search for new
matrix metalloproteinase inhibitors (MMPIs).[10] Since the late
1990s, some compounds entered clinical trials for the treat-
ment of both arthritis and cancer, but their development was
halted because of low bioavailability and the occurrence of rel-
evant side effects due to the lack of selectivity.[11,12] MMP family
members share common structural features, including a pro-
peptidic domain and a residue motif within the catalytic
domain, HExGHxxGxxH, with three histidine residues coordinat-
ing to the zinc ion. In addition to the catalytic pocket, MMPs
are provided with additional binding subsites, namely, S1, S2,
S3, S1’, S2’ and S3’, which cooperate for substrate recognition
and general functioning.[4,5] Because it is known that the S1’
subsite shows the highest variability of primary structure, and
thus, representing a convenient place for challenging MMP se-
lectivity,[13,14] huge efforts have been made in the design of
suitably shaped molecular moieties to fit the S1’ subsite, the
pocket of which can be shallow, intermediate and deep, and
thus, extremely varied among MMP evolutionary groups.[15] Ini-
tially, our over-riding target was that of biasing gelatinases A
(MMP-2) and B (MMP-9), the selective inhibition of which rep-
resents a highly efficient strategy for early stage cancer thera-
py.[16–20] Indeed, in patients with highly advanced tumours,
MMP-9 showed a host-protective activity; thus behaving as an
anti-target.[6,21] Gelatinases play a central role in the degrada-
tion of collagen types IV, V, VII and X and gelatin and are ex-
pressed in many cells and tissues, including skin, fibroblasts,
keratinocytes, chondrocytes, endothelial cells and mono-
cytes.[22] They are involved in almost all phases of tumour pro-
gression, from tumour development to metastasis. However,
our attention was slightly shifted to MMP-2, which is involved
in endometrial menstrual breakdown, regulation of vascularisa-
[a] Dr. G. F. Mangiatordi, D. Trisciuzzi, Dr. D. Alberga, Prof. Dr. O. Nicolotti
Dipartimento di Farmacia-Scienze del Farmaco, Universitꢀ di Bari “Aldo
Moro”, Via Orabona, 4, 70126, Bari (Italy)
E-mail: orazio.nicolotti@uniba.it
[b] Dr. G. F. Mangiatordi
Istituto Tumori IRCCS Giovanni Paolo II, Bari (Italy)
[c] Dr. T. Guzzo, E. C. Rossano, A. Topai
C4T S.r.l Colosseum Combinatorial Chemistry Centre for Technology, Via
della Ricerca Scientifica snc, Ed. PP2-Macroarea Scienze, 00133, Rome
(Italy)
E-mail: alessandra.topai@c4t.it
[d] Dr. G. Fasciglione, Prof. Dr. M. Coletta
Dipartimento di Scienze cliniche e Medicina Traslazionale, Universitꢀ di
Roma “Tor Vergata”, Via Montpellier, 1, 00133, Rome (Italy)
Supporting information and the ORCID identification number(s) for the
author(s) of this article can be found under:
https://doi.org/10.1002/cmdc.201800237.
ChemMedChem 2018, 13, 1 – 11
1
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
tion, inflammatory response[23] and significantly overexpressed
in the bone marrow of myeloma patients. In this respect, the
selective targeting of MMP-2 over that of closely related MMP-
9 has recently been suggested as a strategy for multiple mye-
loma treatment.[24] Bearing this in mind, our efforts were thus
addressed to the identification of a novel molecular scaffold in-
corporating zinc binding groups (ZBGs) and other structural
motifs retrieved from a retrospective analysis on MMPIs. The
affinity improvement provided through step-by-step structural
modifications allowed us to discover four derivatives, 9 f, 9g,
9h and 9i, with excellent activities in the sub-nanomolar
range; in particular, compounds 9g and 9i also showed a very
good selectivity toward both gelatinases MMP-2 and MMP-9.
Molecular docking simulations, which were performed by em-
ploying quantum mechanics (QM) based partial charges, al-
lowed the identification of key interactions responsible for the
high affinity shown by our derivatives toward their preferred
MMPs.
Table 1. (Continued)
Compd
R1
R2
IC50 [nm][a]
MMP-2
MMP-9
9b
11 c
5c
7Æ2.5
7Æ2
H
100ꢁ106
ND
62000Æ
2500Æ100
5000
5d
5e
1500000
50ꢁ106
14100000
ND
5 f
9 f
5g
9g
5h
9h
5i
600Æ100
0.15Æ0.1
2200Æ800
0.25Æ0.2
Results and Discussion
Molecular design
Based on existing knowledge, we conceived of novel MMPIs,
including three essential molecular determinants: 1) the hy-
droxamic or carboxylic acid as a ZBG to engage the zinc ion
within the catalytic domain; 2) the sulfonamide-based hook-
like fragments that tether the S1’ enzyme subsite to tune MMP
selectivity; and 3) the tetrahydro-1H-pyrido [3,4-b]indole, better
known as tetrahydro-b-carboline (Table 1), to anchor the S3’
enzyme subsite, which may offer additional subsite interac-
tions.[25]
10000Æ
15000Æ
2000
4000
0.15Æ0.1
0.63Æ0.2
4000Æ1000
0.25Æ0.2
2400Æ800
0.58Æ0.2
ND
28000Æ
1000
0.15Æ0.1
800Æ100
0.15Æ0.1
300000
Table 1. The inhibitory activities toward MMP-2 and MMP-9.
9i
5l
Compd
R1
R2
H
IC50 [nm][a]
MMP-2
prinomastat
<0.3
<0.5
MMP-9
0.5Æ0.2[b]
NA[b]
[a] Values are from at least six concentrations, each of which was deter-
mined in duplicate; prinomastat was used as a control compound.
[b] Values reported by Shalinsky et al.,[59] which are included for clarity.
ND: not determined; NA: not available.
11 a
5a
100ꢁ106
600Æ80
14Æ3
ND
7500Æ5000
15Æ2
9a
Hydroxamic acid has the advantage of chelating the zinc ion
in a bidentate fashion and of establishing hydrogen-bond rein-
forced ionic interactions with counterpart residues (i.e., gluta-
mates) acting as hydrogen-bond acceptors.[13] As expected, the
activity drops if hydroxamic acid is replaced by a carboxylic
acid as the ZBG (Table 1). Nevertheless, it is worth mentioning
that the sulfonamide-based hydroxamate moiety represents a
well-known structural determinant for targeting MMPs;[13,25,26]
the sulfonyl group is responsible for hydrogen-bonding inter-
9’a
>50000
>50000
11 b
5b
H
100ꢁ106
ND
24000Æ
680Æ200
1000
&
ChemMedChem 2018, 13, 1 – 11
www.chemmedchem.org
2
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
actions with the enzyme backbone. In this respect, intense re-
search over the last two decades has mainly aimed at properly
fitting the S1’ pocket to improve selectivity among different
MMPs.[14,27] To this end, a very good fitting was obtained by
probing the depth of the S1’ pocket with para-substituted
biaryl molecular fragments.[13,28–31] Based on this knowledge,
we investigated differently substituted P1’ moieties, including
three para-substituted phenoxyphenyl groups.
and MMP-9. Indeed, substitution at the R2 position with an
amide group leads to a loss of activity (e.g., 9a vs. 9’a). By fo-
cusing attention on different R1 substitutions (see Table 1), it
appears evident that the activity toward both gelatinases de-
pends on the length and hydrophobicity of the P1’ group. In
particular, the replacement of naphthyl (5c), biphenyl (5a) or
phenoxyphenyl (5 f) groups with 4-methylphenyl (5d) or 4-
tert-butylphenyl (5e) groups strongly reduces the inhibitory
potencies. More specifically, activities at the sub-nanomolar
IC50 level are reached toward both MMPs in the presence of a
phenoxyphenyl group at R2 with chlorine (9g), fluorine (9h),
methoxy (9 f) and cyano (9i) substituents in the para position.
This supports the idea that the hydrophobic moiety in position
P1’ should be equipped with some flexibility to efficiently fit
the S1’ pocket. Taken together, the obtained data indicate that
tetrahydro-b-carboline derivatives can bind the catalytic sites
of both MMP-2 and MMP-9 with high affinity; thus indicating
this novel scaffold as being very promising for the design of a
series of new gelatinase inhibitors. In particular, taking advant-
age of the already known ZBG (hydroxamic acid) and P1’ por-
tions (phenoxyphenyl derivatives), impressive IC50 values at the
sub-nanomolar level were reached.
Furthermore, our attention was engaged by the study of
subsite S3’, which proved to be crucial to enhance the affinity
toward several MMPs, including MMP-2 and MMP-9.[25] Notably,
these MMPs share an identical S3’ residue motif (DGLL), where-
as differences in sequence involving at least one residue can
be found in the S3’ subsites belonging to all other MMPs. In
this respect, we envisaged the tetrahydro-b-carboline ring as a
flat surface wide enough to fit the S3’ pocket through a specif-
ic interaction with a non-conserved leucine residue and capa-
ble of reducing the metabolic lability of the hydroxamic acid
group by steric hindrance.[32,33] On the other hand, tetrahydro-
b-carboline also emerged as a privileged substructure in me-
dicinal chemistry that was able to match several relevant phar-
maceutical targets, such as serotonin receptors,[34,35] an imida-
zoline receptor,[36] acetylcholinesterase,[37] mitogen-activated
protein kinase activated protein kinase 2,[38] and mitotic kinesin
spindle protein,[39] making it a promising candidate for cancer
therapy.[40] To the best of our knowledge, the potential of tetra-
hydro-b-carboline as a molecular scaffold to inhibit MMP-2 and
MMP-9 gelatinases was never assessed. The only exception is
represented by the work of Matter and Schwab, who reported
nanomolar inhibitory potencies against MMP-3 and MMP-8 for
some tetrahydro-b-carboline derivatives provided with a ZBG
(hydroxamic acid), but at a different position (i.e., position
3).[41] Starting from this evidence, we designed a panel of sulfo-
namide-based carboxylates and hydroxamates containing dif-
ferently substituted P1’ groups and tetrahydro-b-carboline as a
novel hydrophobic and bulky molecular scaffold to be accom-
modated into the S3’ subsites of MMP-2 and MMP-9.
To properly challenge the selectivity of our molecular series,
we tested the inhibitory potencies of the most active com-
pounds toward collagenases (MMP-1, MMP-8, MMP-13), stro-
melysin-1 (MMP-3), macrophage metalloelastase (MMP-12) and
membrane-type 1 MMP (MMP-14). Table 2 shows the obtained
IC50 values reported on the nanomolar scale.
Notably, among the top four active compounds, compounds
9g and 9i were the most selective, showing inhibitory poten-
cies for MMP-2 and MMP-9 at least one order of magnitude
higher, with respect to all considered MMPs. In other words,
steric hindrance by the para substituents of the phenoxyphen-
yl P1’ moiety seem to play a critical role in the detected selec-
tivity, as evident by looking at the different steric Verloop pa-
rameters B1,[43] which account for the minimal molecular
widths of para-O-CH3 (B1=1.35), para-F (B1=1.35), para-Cl
(B1=1.80) and para-CN (B1=1.60) that reflect the different af-
finities of the compounds (see Table 2). Conversely, no relevant
trend can be observed among Hansch hydrophobic[44] and
Hammett electronic[45] constants of the para substituents and
the detected activities toward gelatinases, which proves that
this feature is predominantly supported by the presence of
electron-withdrawing groups at the same position.[28] Consis-
tently, such a condition is fulfilled in all four of the most active
compounds; the Hammett s constants[45] corresponding to
para-O-CH3 (9 f), para-Cl (9g), para-F (9h) and para-CN (9i) are
0.27, 0.23, 0.06 and 0.66, respectively. Finally, it should be
noted that, for all compounds, a relevant selectivity is shown
mainly with respect to MMP-1 and MMP-3, the residues of the
S3’ motif (GGNL and GNVL, respectively) of which markedly
differ from those of MMP-2 and MMP-9 (DGLL).[4] This allows us
to speculate that a different affinity of the tetrahydro-b-carbo-
line scaffold toward the S3’ subsite might concur with the ob-
served selectivity.
Biological evaluation
Racemic mixtures of all synthesised compounds were tested to
evaluate the inhibitory effects on MMP-2 and MMP-9. In partic-
ular, a fluorogenic assay was performed by using human re-
combinant MMP-2 and MMP-9, purchased from Biomol Inter-
national (http://www.enzolifesciences.com/biomol/), and apply-
ing the protocol described by Knight et al.[42] (see the Experi-
mental Section for methodological details). All measured IC50
values are reported in Table 1. As expected, the replacement of
a ZBG by a hydrogen atom brings about a total loss of activity,
which is evident by comparing 11 a with 5a and 9a; 11 b with
5b and 9b; and 11 c with 5c. More specifically, the detected
inhibitory potencies strongly increase upon going from mono-
(carboxylic acid as the ZBG) to bidentate (hydroxamic acid as
the ZBG) coordination with zinc; this is confirmed in all cases
(e.g., 5a vs. 9a; 5b vs. 9b; 5 f vs. 9 f; 5g vs. 9g; 5h vs. 9h; 5i
vs. 9i). In addition, the obtained data confirm that the pres-
ence of a sulfonamide group is required for targeting MMP-2
ChemMedChem 2018, 13, 1 – 11
www.chemmedchem.org
3
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
Table 2. Selectivity profiles of 5a, 9a, 5b, 9b, 5 f, 9 f, 5g, 9g, 5h, 9h, 5i, 9i, and 5l.
Compd
IC50 [nm][a]
Collagenases
MMP-8
Stromelysin-1
MMP-3
Metalloelastase
MMP-12
Membrane-type 1
MMP-14
MMP-1
>50000
MMP-13
5a
9a
5b
9b
5 f
9 f
5g
9g
5h
9h
5i
9i
3000Æ700
4.7Æ0.6
160Æ50
11000Æ1
600
300Æ50
30000Æ2000
ND
3000Æ1000
>50000
20000Æ2000
>50000
350Æ100
>50000
>50000
>50000
340Æ100
>50000
>50000
>50000
7Æ1
3Æ0.4
2000Æ500
2Æ0.3
4Æ1.6
4500Æ1000
0.25Æ0.2
500Æ100
0.25Æ0.2
2000Æ200
70Æ30
3000000
ND
ND
0.25Æ0.2
1600Æ200
0.25Æ0.2
10000Æ1000
12Æ2
4000Æ1500
>50000
30Æ10
50Æ10
>50000
5Æ1
15000Æ5000
0.25Æ0.2
800Æ200
2Æ0.5
>50000
250Æ50
>50000
600Æ200
>50000
350Æ50
>50000
1.4Æ0.4
>50000
30Æ5
21000Æ2000
0.25Æ0.2
20000Æ1000
12.8Æ2.1
22000Æ2000
<0.5
>50000
0.25Æ0.2
1000Æ200
4.5Æ1.0
2000Æ300
ND
>50000
3.2Æ0.5
>50000
60Æ10
>50000
2000Æ200
110Æ30
2200Æ300
<0.5
5l
>50000
<0.5
prinomastat
83Æ8[b]
0.3Æ0.1[b]
3Æ1[b]
[a] Values are from at least six concentrations, each of which was determined in duplicate; prinomastat was used as a control compound. [b] Values report-
ed by Shalinsky et al.,[59] which are included for clarity. ND: not determined.
Molecular docking simulations
forced hydrogen bond with a glutamate residue (Glu404 in
MMP-2 and Glu402 in MMP-9). Notably, residues are numbered
according to UniProt[48] (codes P08253 and P14780 for human
MMP2 and MMP9, respectively). Furthermore, the sulfonamide
group establishes a hydrogen-bond interaction with the S3’
subsite and, in particular, with the backbone of Leu191
(Leu188) in MMP-2 (MMP-9), whereas the phenoxyphenyl
moiety (group P1’) is accommodated in the hydrophobic S1’
pocket in both MMPs. Notably, all of these interactions turned
out to be crucial for the binding of other sulfonamide hydroxa-
mates[13] with MMPs, and they are also detectable by looking
at the NMR spectroscopy (PDB ID: 1HOV)[46] and X-ray (PDB ID:
4H3X)[47] coordinates of the cognate ligands of the considered
MMP structures; hence again supporting the robustness of the
predicted binding mode. Important clues can be detected by
visually inspecting the binding conformation of the tetrahydro-
b-carboline ring engaging the S3’ subsite (see Figure 1). More
specifically, a hydrophobic interaction with Leu190 (Leu187) is
detected for all considered compounds in the catalytic site of
MMP-2 (MMP-9). Pairwise sequence alignment analysis[4] shows
that this residue belonging to the S3’ domain is not conserved
among other MMPs. The only exception is represented by
MMP-13 (i.e., SGLL), for which most of the tested compounds
do not show any relevant selectivity (see, for instance, com-
pounds 9 f and 9h). This highlights the structural relevance of
the difference with respect to MMP-1 and MMP-3, the S3’ do-
mains of which show, in the corresponding position, an aspara-
gine and a valine residue, respectively (see Table 2). Taken as a
whole, these results suggest that the tetrahydro-b-carboline
moiety could contribute to enhancing binding toward MMP-2
and MMP-9; thus indicating that this molecular scaffold is very
promising for the development of new potent and selective
gelatinase inhibitors. In particular, the observed selectivity
might benefit from the hydrophobic interaction with a Leu190
(Leu187) belonging to the S3’ subsite of MMP-2 (MMP-9).
To gain insight into the molecular interactions behind the high
affinity of our derivatives toward MMP-2 and MMP-9, QM po-
larised docking experiments were carried out, by selecting the
most active compounds, 9 f, 9g, 9h and 9i, with sub-nanomo-
lar activities as molecular probes. In particular, for each com-
pound both enantiomers were taken into account and docked
into the binding sites of MMP-2 (PDB ID: 1HOV)[46] and MMP-9
(PDB ID: 4H3X).[47] It is important to note that, for all consid-
ered ligands, the top-scored pose was that returned by the R-
enantiomer. In addition, it should be observed that docking
simulations showed that S-enantiomers failed, in most cases,
to properly engage the zinc ion, unlike R-enantiomers, the ZBG
of which was instead always well oriented. Based on docking
simulations, we could rationally assume that the interactions
taking place at the binding pocket would be stereoselective;
the R-enantiomers were preferred in terms of both scoring and
posing (Figure S12 of Supporting Information). Figure 1 shows
the top-scored posed of 9 f, 9g, 9h and 9i in both considered
MMPs. An almost equal posing is returned for both MMP-2
(Figure 1A, C, E and G) and MMP-9 (Figure 1B, D, F and H). Im-
portantly, the docking scores returned by 9 f (À8.522 and
À9.596 kcalmolÀ1 in 1HOV and 4H3X, respectively), 9g (À8.591
and À9.537 kcalmolÀ1 in 1HOV and 4H3X, respectively), 9h
(À8.897 and À9.148 kcalmolÀ1 in 1HOV and 4H3X, respectively)
and 9i (kcalmolÀ1 À8.624 and À9.538 kcalmolÀ1 in 1HOV and
4H3X, respectively) are very close to those given by the corre-
sponding cognate ligands (À8.546 and À10.317 kcalmolÀ1 in
1HOV and 4H3X, respectively); hence making us confident
about the robustness of the predicted binding modes. In addi-
tion to the bidentate coordination of the zinc ion, the hydroxa-
mic acid moiety engages a hydrogen-bond interaction with
the backbone of an alanine residue belonging to the S1 sub-
site (Ala192 in MMP-2 and Ala189 in MMP9) and a charge-rein-
&
ChemMedChem 2018, 13, 1 – 11
www.chemmedchem.org
4
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
the Lipinski rule of five without any exception (molecular
weight <500, logP<5, number of hydrogen-bond-donor
atoms ꢀ5, and number of hydrogen-bond-acceptor atoms
ꢀ10).[50,51] In addition, the percentage of oral absorption in
humans was also computed based on a quantitative linear re-
gression model available in QikProp. In particular, percentages
ranging from 70 (9i) to 87% (9g) were predicted; thus indicat-
ing high oral adsorption for all considered compounds. Finally,
the drug likeness of the candidate compounds was assessed
by comparing 24 representative molecular properties with
those of 95% of known drugs, as implemented in QikProp. It is
noteworthy that all predicted properties for 9 f, 9g and 9h fall
inside the range defined by known drugs, which supports the
indication that the most active tetrahydro-b-carboline deriva-
tives are characterised by promising drug-likeness properties.
Conclusions
A promising series of gelatinase (MMP-2 and MMP-9) inhibitors
containing a novel tetrahydro-b-carboline moiety has been
identified. Satisfactorily, racemic mixtures of compounds 9 f,
9g, 9h and 9i reach inhibitory potencies at the sub-nanomolar
IC50 level toward both targets, showing interesting selectivity
profiles with respect to other MMPs, as well as promising phar-
macokinetics, as indicated by in silico ADME testing. Sequence
alignment analysis and intensive docking experiments suggest-
ed that the proposed tetrahydro-b-carboline scaffold could
contribute to enhancing the selectively toward MMP-2 and
MMP-9 by taking advantage of hydrophobic interactions with
a leucine residue belonging to the S3’ subsite missing in other
MMPs. Taken together, these results indicate tetrahydro-b-car-
boline as a new molecular scaffold candidate to be further in-
vestigated for the design of potent and selective gelatinases
inhibitors with a fair ADME profile.
Figure 1. Top-scored poses of 9 f (A, B), 9g (C, D), 9h (E, F) and 9i (G, H) in
MMP-2 (A, C, E and G) and MMP-9 (B, D, F and H). Ligands and important
residues are shown in stick representation, whereas the protein is rendered
in surface representation. The detected hydrogen-bonding interactions are
depicted by a dotted line. For clarity, only polar hydrogen atoms are shown.
Residue numbering follows that of the generic MMP nomenclature.
Experimental Section
Chemistry
The investigated tetrahydro-b-carboline derivatives were synthes-
ised by using an original solid-phase approach to obtain either a 2-
chlorotrityl resin or a NovaSyn Tentagel resin, depending on
whether final cleavage should release a carboxylic (Scheme 1) or a
hydroxamic moiety (Scheme 2). In particular, compound 1 (amino
acid b-carboline) was synthesised from 2-(1H-indol-3-yl)ethan-1-
amine and 2-oxoacetic acid, according to procedures described in
the literature,[52] and used as racemic mixtures in the subsequent
acylation reaction with Fmoc-Cl to obtain a Fmoc-protected amino
acid (2). Attachment to the 2-chlorotrytil resin followed by removal
of the Fmoc group and subsequent coupling with appropriate sul-
fonyl chloride provided sulfonamide derivatives, which were subse-
quently finally cleaved from the resin to give compounds 5a–l in
good yields (Scheme 1). Notably, the degree of purity (>90% by
HPLC–MS) reached after cleavage from the resin was such that no
further purification was required. For the synthesis of compound 9
series, the fluoride loading method reported by Kaduk et al.[53] was
adopted (Scheme 2) to increase loading on the hydroxymethyl
resin and to limit side reactions on the particularly reactive b-trip-
tolinic nucleus. More specifically, the substrate was anchored on
Prediction of relevant ADME properties
It is acknowledged that the main reason behind the failure of
clinical trials involving MMPIs consists of the poor pharmacoki-
netics of the tested compounds, mainly due to high metabolic
instability and low oral bioavailability.[11,12,17] Based on this
knowledge, some absorption, distribution, metabolism and ex-
cretion (ADME) properties are nowadays considered to be cru-
cial for turning a potent MMPI into a promising drug candi-
date. In this respect, we predicted some relevant pharmacoki-
netics properties of the most active compounds belonging to
our molecular series (9 f, 9g, 9h and 9i) by using QikProp; a
computation tool available in the Schrçdinger suite.[49] The ob-
tained data are reported in Table S1 in the Supporting Informa-
tion, and indicate that all considered compounds accomplish
ChemMedChem 2018, 13, 1 – 11
www.chemmedchem.org
5
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
Scheme 1. Synthesis of compounds 5a–l: a) dioxane/H2O, Na2CO3, 9-fluorenylmethoxycarbonyl chloride (Fmoc-Cl), RT, 24 h; b) N,N-diisopropylethylamine
(DIEA), CH2Cl2, RT, 12 h; c) 1) 20% piperidine in DMF, 20 min, RT; 2) R-sulfonyl chloride/dimethylaminopyridine (DMAP), CH2Cl2, RT, 18 h; d) AcOH/2,2,2-trifluor-
oethanol (TFE)/CH2Cl2 (1:1:8), RT, 1 h.
Scheme 2. Synthesis of compounds 9a, 9b, 9’a and 9 f–i: a) CH2Cl2/DMF, diethylaminosulfur trifluoride (DAST), RT, 3 h; b) DMF, RT, 12 h; c) 1) 20% piperidine
in DMF, 20 min, RT; 2) R-sulfonyl chloride/DMAP, CH2Cl2, RT, 18 h; c’) 1) 20% piperidine in DMF, 20 min, RT; 2) R-acyl chloride/DMAP, CH2Cl2, RT, 18 h; d) NH2OH
(50 wt% in H2O), THF, RT, 5 h.
NovaSyn Tentagel resin and after Fmoc cleavage the amino group
reacted with several sulfonyl chlorides to give resin-bound sulfona-
mides 8a–i, or with [1,1’-biphenyl]-4-carbonyl chloride to obtain
resin-bound amide 8’a. Final cleavage from the resin
with (NH2OH) released compounds 9a–i and 9’a with
generally high purity (>90%) in significant yields (50–
60%). Compounds that did not show the target degree
of purity after cleavage were purified by semi-prepara-
tive reverse-phase (RP) HPLC. Finally, to synthesise com-
pounds 11 a–c, decarboxylation of the b-triptolinic scaf-
fold 1 was followed by coupling with the appropriate
Reagents and solvents were obtained from commercial suppliers
and used without further purification. Flash chromatography was
performed on E. Merck silica gel. RP purifications were performed
Scheme 3. Synthesis of compounds 11 a–c: a) HCl/dioxane, 24 h, b) R-sulfonyl chloride/
sulfonyl chloride (Scheme 3).
DMAP, CH2Cl2, RT, 18 h.
&
ChemMedChem 2018, 13, 1 – 11
www.chemmedchem.org
6
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
1
by using a Gilson HPLC-UV system (321/H2M pumps, UV/Vis 152,
Fraction collector 2020). 1H NMR spectra were recorded in the
specified deuterated solvents on a Bruker Avance 300 spectrome-
ter operating at 300 MHz. Chemical shifts are reported in ppm (d)
from the tetramethylsilane (TMS) resonance in the indicated sol-
vent (TMS: d=0.0 ppm). Compound purities and mass spectra
were determined by using an LC–MS platform (Gilson/ThermoFin-
nigan and Agilent 1260 Series) by means of the positive electro-
spray ionisation technique (+ES) with a mobile phase of acetoni-
trile/water containing 0.1% trifluoroacetic acid (TFA). Elementary
analyses were performed on a PerkinElmer 2400 Series II elemental
analyser and an AD-4 autobalance PerkinElmer microbalance. Sol-
vents were distilled for resin reactions. Unless otherwise stated, re-
actions were performed in 8 mL polypropylene cartridges with
70 mm polyethylene (PE) frits. Cartridges and stopcocks were ob-
tained from Applied Separations. All coupling reagents and resins
were purchased from Novabiochem and Aldrich. All compounds
were analysed by HPLC-UV-MS for purity assay and structure confir-
mation. With regard to analytical characterisation of the com-
pound, it should be noted that the presented molecules represent-
ed a sub-library of MMPIs synthesised by means of automated par-
allel methodology. As a result, NMR spectroscopy and HPLC–MS
analyses on a second instrument were performed for a representa-
tive test sample, by random selection of final compounds. More
afford 5a (29.3 mg, 70%). H NMR (300 MHz, CD3OD): d=2.58 (m,
2H), 3.91 (m, 1H), 4.17 (m, 1H), 5.55 (s, 1H), 6.91 (m, 1H), 7.04 (m,
1H), 7.25 (d, J=7.8 Hz, 1H), 7.31 (d, J=8.1 Hz, 1H), 7.40 (m, 3H),
7.52 (d, J=7.23 Hz, 2H), 7.61 (d, J=8.2 Hz, 2H), 7.9 ppm (d, J=
8.2 Hz, 2H); MS: m/z (%): 387.1 (100) [MÀCOOH]; elemental analy-
sis calcd (%) for C24H20N2O4S: C 66.65, H 4.66, N 6.48, O 14.80, S
7.41; found: C 66.65, H 4.63, N 6.51, O 14.87, S 7.45.
Compound 5b: MS: m/z (%): 429.2 (100) [M+H]+; elemental anal-
ysis calcd (%) for C22H24N2O5S: C 61.67, H 5.65, N 6.54, O 18.67, S
7.48; found: C 61.70, H 5.68, N 6.54, O 18.69, S 7.48.
Compound 5c: MS: m/z (%): 407.2 (100) [M+H]+; elemental analy-
sis calcd (%) for C22H18N2O4S: C 65.01, H 4.46, N 6.89, O 15.74, S
7.89; found: C 65.03, H 4.46, N 6.89, O 15.75, S 7.89.
Compound 5d: MS: m/z (%): 371.2 (100) [M+H]+; elemental anal-
ysis calcd (%) for C19H18N2O4S: C 61.61, H 4.90, N 7.56, O 17.28, S
8.65; found: C 61.63, H 4.90, N 7.56, O 17.29, S 8.66.
Compound 5e: MS: m/z (%): 413.2 (100) [M+H]+; elemental analy-
sis calcd (%) for C22H24N2O4S: C 64.06, H 5.86, N 6.79, O 15.51, S
7.77; found: C 64.09, H 5.86, N 6.80, O 15.52, S 7.77.
1
Compound 5 f: H NMR (300 MHz, CD3OD): d=2.59 (m, 2H), 3.71
(m, 1H,), 3.74 (m, 4H), 4.18 (m, 1H), 5.70 (s, 1H), 6.84 (m, 6H), 7.05
(m, 1H), 7.13 (m, 1H), 7.34 (d, J=8.2 Hz, 2H), 7.75 ppm (d, J=
8.1 Hz, 2H); MS: m/z (%): 479.2 (100) [M+H]+; elemental analysis
calcd (%) for C25H22N2O6S: C 62.75, H 4.63, N 5.85, O 20.06, S 6.70;
found: C 62.78, H 4.63, N 5.85, O 20.07, S 6.70.
1
specifically, H NMR spectra of 5a, 5 f, 9 f, 5i and 9i, and additional
HPLC–MS results for 9a, 5b, 9b, 5g and 9g, are available in the
Supporting Information.
Compound 5g: MS: m/z (%): 437.2 (100) [MÀCOOH]+; elemental
analysis calcd (%) for C24H19ClN2O5S: C 59.69, H 3.97, Cl 7.34, N 5.80,
O 16.56, S 6.64; found: C 59.71, H 3.97, Cl 7.34, N 5.80, O 16.57, S
6.64.
General procedures for the coupling of Fmoc-b-carboline (2)
to 2-chlorotrityl resin
Example: Synthesis of 3: A solution of 2 (65.8 mg; 0.15 mmol)
and DIEA (105 mL, 0.6 mmol) in dry CH2Cl2 (4 mL) was added to 2-
chlorotrityl resin (0.2 g; 0.24 mmol; loading 1.2 mmolgÀ1) pre-swol-
len in dry CH2Cl2. After shaking for 12 h at room temperature, the
resin was filtered and rinsed with CH2Cl2 (3ꢁ5 mL), DMF (3ꢁ5 mL),
CH2Cl2 (3ꢁ5 mL) and methyl tert-butyl ether (TBME; 3ꢁ5 mL) and
finally dried in vacuo. The efficiency of loading (loading yield:
53%) was determined on resin (3 mg) by Fmoc removal analysis
(treatment of the resin with 20% piperidine in DMF) and absorb-
ance measurement at l=301 nm (UV/Vis spectrophotometer; Shi-
madzu).
Compound 5h: MS: m/z (%): 467.2 (100) [M+H]+; elemental anal-
ysis calcd (%) for C24H19N2O5S: C 61.80, H 4.11, F 4.07, N 6.01, O
17.15, S 6.87; found: C 61.82, H 4.11, F 4.07, N 6.01, O 17.16, S 6.87.
Compound 5i: 1H NMR (300 MHz, [D6]DMSO): d=2.65 (m, 2H),
3.62 (m, 1H), 4.12 (m, 1H), 5.58 (s, 1H), 7.09 (m, 6H), 7.35 (m, 2H),
7.84 (m, 4H), 10.91 ppm (brs, 1H); MS: m/z (%); 428.2 (80)
[MÀCOOH]+; elemental analysis calcd (%) for C25H19N3O5S: C 63.42,
H 4.04, N 8.87, O 16.89, S 6.77; found: C 63.43, H 4.03, N 8.87, O
16.89, S 6.77.
Compound 5l: MS: m/z (%): 449.2 (100) [M+H]+; elemental analy-
sis calcd (%) for C24H20N2O5S: C 64.27, H 4.50, N 6.25, O 17.84, S
7.15; found: C 64.30, H 4.50, N 6.25, O 17.85, S 7.15.
General procedures for N-sulfonylation of polymer-bound b-
carboline 3 and 7
Example: Synthesis of 4a–l: After removal of Fmoc with 20% pi-
peridine in DMF for 20 min (ꢁ2), the resin (0.15 mmol) was washed
(2ꢁDMF, 2ꢁCH2Cl2, 2ꢁTBE) and dried. The resin was then allowed
to swell in THF with DMAP (2 equiv) for 1 h. A solution of bis-phe-
nylsulfonylchloride (0.45 mmol) in dry CH2Cl2 (2.5 mL) was added
and the cartridge was rocked for 18 h at room temperature. Then
the resin was filtered and washed with MeOH (ꢁ2), DMF (ꢁ2),
CH2Cl2 (ꢁ2) and TBME (ꢁ2).
General procedures for Fmoc-b-carboline fluoride coupling to
NovaSynTGOH
Example: Synthesis of 7: DAST (95 mL, 0.72 mmol) was added to a
solution of Fmoc-b-carboline 1 (263.1 mg; 0.6 mmol) in CH2Cl2/dry
DMF (4 mL; 7/1) and the reaction mixture was stirred for 3 h at
room temperature. The desired product was extracted with CH2Cl2/
H2O. The organic phases were combined, dried over Na2SO4 and
evaporated to dryness. Derivative 6 thus obtained was dissolved in
dry DMF (4 mL) and placed on NovaSyn TGOH (0.2 g,
0.3 mmolgÀ1), pre-swollen in dry CH2Cl2, with DMAP (0.7 mg;
6 mmol). The cartridge was gently rocked for 2 h at room tempera-
ture. After filtration, the resin was rinsed with DMF (3ꢁ5 mL),
CH2Cl2 (3ꢁ5 mL) and TBME (3ꢁ5 mL) and finally dried in vacuo.
The efficiency of loading (loading yield: 81%) was determined on
resin (5 mg) by Fmoc removal analysis (treatment of the resin with
General procedures for cleavage of carboxylate derivatives
5a–i
Example: Synthesis of 5a: After swelling functionalised resin 4,
the crude product (0.1 mmol) was cleaved by treatment with
AcOH/TFE/CH2Cl2 (2 mL; 1:1:8) for 1 h at room temperature. The
resin was filtered and the filtrate was evaporated under N2 to
ChemMedChem 2018, 13, 1 – 11
www.chemmedchem.org
7
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
20% piperidine in DMF) and absorbance measurement at l=
Procedure for the synthesis of 10
301 nm (UV/Vis spectrophotometer; Shimadzu).
b-Carboline 1 (1 g; 4,66 mmol) was dissolved in 4m HCl in dioxane
(20 mL). The mixture was heated at reflux and stirred for 24 h, until
complete decarboxylation occurred. Dioxane was removed under
reduced pressure and the mixture was diluted with ethyl acetate
and treated with a saturated aqueous solution of NaHCO3. The or-
ganic layer was washed with brine and dried over Na2SO4. After fil-
tration, the solvent was removed by evaporation to afford 10 as a
colourless oil, which was used as a crude product without charac-
terisation.
General procedure for the cleavage of hydroxamate deriva-
tives 9a–i
Example: Synthesis of 9a: Functionalised resin derivative 7h
(38 mmol) was swollen in THF (1.8 mL) and cleaved from the solid
support by treatment with 50% NH2OH in H2O (500 mL, 8 mmol).
After rocking for 2 h at room temperature, the resin was filtered
and the filtrate was evaporated to dryness. The crude product was
further purified by semi-preparative RPHPLC (C18, Inertisil.ODS-3,
300 mmꢁ25 mm, 3 mLminÀ1, detection at l=230 nm; eluent A:
water/TFA (99.92:0.08); eluent B: acetonitrile/TFA (99.92:0.08); gra-
dient B from 35 to 70% in 20 min, from 70 to 98% in 3 min, collec-
tion peak: 8 min) to afford 9a with high purity (>99%). MS: m/z
(%): 448.2 (100) [M+H]+; elemental analysis calcd (%) for
C24H21N3O4S: C 64.42, H 4.73, N 9.39, O 14.30, S 7.16; found: C
64.44, H 4.73, N 9.40, O 14.31, S 7.16.
Procedure for the synthesis of sulfonamides 11
Example: Synthesis of 11a: Crude 10 (50 mg, 0.32 mmol) was dis-
solved in dry CH2Cl2, DMAP (2 equiv) was added and the mixture
was stirred at room temperature for 15 min. A solution of bis-phe-
nylsulfonylchloride (0.45 mmol) in dry CH2Cl2 (2.5 mL) was added,
and the mixture was stirred at room temperature for 3 h. The
crude product was further purified by semi-preparative RP-HPLC
(C18, Inertisil.ODS-3, 300 mmꢁ25 mm, 3 mLminÀ1, detection at l=
230 nm; eluent A: water/TFA (99.92:0.08); eluent B acetonitrile/TFA
(99.92:0.08); gradient B from 25 to 70% in 20 min, from 70 to 98%
in 3 min, collection peak: 6 min) to afford 11 a with high purity (>
99%). MS: m/z (%): 389.2 (40) [M+H]+; elemental analysis calcd
(%) for C23H20N2O2S: C 71.11, H 5.19, N 7.21, O 8.24, S 8.25; found:
C 71.14, H 5.19, N 7.22, O 8.25, S 8.26.
Compound 9b: MS: m/z (%): 444.2 (100) [M+H]+; elemental anal-
ysis calcd (%) for C22H25N3O5S: C 59.58, H 5.68, N 9.47, O 18.04, S
7.23; found: C 59.60, H 5.68, N 9.47, O 18.05, S 7.23.
Compound 9 f: MS: m/z (%): 494.2 (100) [M+H]+; elemental analy-
sis calcd (%) for C25H23N3O6S: C 60.84, H 4.70, N 8.51, O 19.45, S
6.50; found: C 60.83, H 4.70, N 8.50, O 19.44, S 6.50.
Compound 11b: MS: m/z (%): 385.3 (100) [M+H]+; elemental
analysis calcd (%) for C21H24N2O3S: C 65.60, H 6.29, N 7.29, O 12.48,
S 8.34; found: C 65.61, H 6.29, N 7.29, O 12.49, S 8.35.
Compound 9g: MS: m/z (%): 498.2 (100) [M+H]+; elemental anal-
ysis calcd (%) for C24H20CllN3O5S: C 57.89, H 4.05, Cl 7.12, N 8.44, O
16.06, S 6.44; found: C 57.91, H 4.05, Cl 7.12, N 8.45, O 16.07, S
6.44.
Compound 11c: MS: m/z (%): 363.3 (100) [M+H]+; elemental anal-
ysis calcd (%) for C21H18N2O2S: C 69.59, H 5.01, N 7.73, O 8.83, S
8.85; found: C 69.60, H 5.00, N 7.73, O 8.84, S 8.86.
Compound 9h: MS: m/z (%): 482.2 (100) [M+H]+; elemental anal-
ysis calcd (%) for C24H20FN3O5S: C 59.87, H 4.19, F 3.95, N 8.73, O
16.61, S 6.66; found: C 59.88, H 4.19, F 3.95, N 8.73, O 16.62, S 6.66.
Compound 9i: 1H NMR (300 MHz, [D6]DMSO): d=2.62 (m, 2H),
3.71 (m, 1H,), 3.88 (m, 1H), 4.0 (m, 1H), 5.42 (s, 1H), 7.02 (m, 6H),
7.35 (m, 2H), 7.82 (m, 4H), 9.16 (s, 1H), 10.72 (s, 1H), 11.04 ppm (s,
1H); MS: m/z (%): 511.1 (40) [M+Na]+, 428.1 (80) [MÀCONHOH]+;
elemental analysis calcd (%) for C25H20N4O5S: C 61.47, H 4.13, N
11.47, O 16.38, S 6.56; found: C 61.49, H 4.13, N 11.48, O 16.39, S
6.56.
Procedure for the synthesis of 1
Synthesis of b-carboline 1 was accomplished according to a proce-
dure reported in the literature.[52] 1H NMR (300 MHz, [D6]DMSO):
d=2.7–2.6 (m, 2H), 3.3–3.2 (m, 2H), 4.8 (s, 1H), 6.95 (m, 1H), 7.04
(m, 1H), 7.38 (m, 1H), 7.46 (m, 1H), 9.03 (brs, 1H), 10.72 ppm (s,
1H); LC–MS: m/z (%): 217.1 (100) [M+H]+.
Procedure for the synthesis of prinomastat reference com-
pound
Procedure for the synthesis of 2
b-Carboline 1 (3 g; 14 mmol) and Na2CO3 (3 g; 28 mmol) were sus-
pended in a solution of H2O (180 mL) and dioxane (50 mL) at 08C
and a solution of Fmoc-Cl (4 g; 14 mmol) in dioxane (30 mL) was
added slowly. After stirring for 24 h at room temperature, dioxane
was removed by evaporation under reduced pressure. The reaction
mixture was diluted with diethyl ether and 2m HCl was added to
pH 1. The two phases were separated and the aqueous phase was
washed several times with diethyl ether. The organic phases were
combined, washed with brine and dried over Na2SO4. After filtra-
tion, the solvent was removed by evaporation to afford 2 as a col-
ourless oil. Subsequent treatment with pentane and filtration of
the white solid obtained gave the desired product as an amor-
phous solid (5.93 g, 70%). LC-UV purity: 98 (l=20), 96% (254 nm);
1H NMR (300 MHz, [D6]DMSO): d=2.78–3.0 (m, 2H), 3.3 (m, 1H),
4.31 (m, 4H), 5.60 (s, 1H,), 7.04 (m, 2H), 7.39 (m, 6H), 7.65 (m, 2H),
7.88 (m, 2H), 10.68 (s, 1H), 13.5 ppm (brs, 1H); LC–MS (ESI; H2O/
CH3CN,1% TFA): m/z (%): 439 (100) [M+H]+.
The synthesis of prinomastat was carried out as described in the
literature.[54] 1H NMR (300 MHz, [D6]DMSO): d=1.18 (s, 3H), 1.47 (s,
3H), 2.6 (d, J=13.72 Hz, 1H), 2.97 (dd, J=13.65, 12.45 Hz, 1H), 3.98
(m, 3H), 7.42 (d, J=6.21 Hz, 2H), 7.48 (d, J=8.58 Hz, 2H), 7.86 (d,
J=8.58 Hz, 2H), 8.74 (d, J=6.21 Hz, 2H), 10.67 ppm (brs, 1H); LC–
MS: m/z (%): 425.08 (100) [M+H]+.
Biological evaluation
All synthesised compounds were screened to evaluate the inhibito-
ry effects on gelatinases. A fluorigenic assay was performed by
using human recombinant MMP-2 and MMP-9 and applying the
protocol described by Knight et al.[42] From the resulting dose–re-
sponse curve, the IC50 values were calculated for each molecule.
The enzymatic assays for MMP1, -2, -3, -8, -9, -12 and -14 were
based on the protocol described by Knight et al.[42] Active catalytic
&
ChemMedChem 2018, 13, 1 – 11
www.chemmedchem.org
8
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
domains of MMP1, -3, -8, -9, -12 and -14 were purchased from
Biomol (Biomol International, Plymouth Meeting), whereas
proMMP2 (from R&D Systems, Inc.) was activated by incubation for
1 h at 378C with a solution of aminophenylmercuric acetate
(APMA; 1 mm final concentration). All enzymatic solutions were
then diluted in Tris-HCl assay buffer (Tris/HCl (50 mm), NaCl (0.1m),
CaCl2 (10 mm), 0.05% Brij 35, pH 7.5) to obtain final concentrations
of the enzymes used in the assay included between 0.5–3 nm. The
fluorogenic substrate was Mca-Pro-Leu-Gly-Leu-Dpa-Ala-Arg-NH2
(Biomol Research Laboratories, Inc.) for all tested MMPs. The final
substrate concentrations in the assay conditions were in the range
of 2–7 mm. Test compounds were dissolved in DMSO and diluted
with test buffer solution, so that no more than 1% DMSO was
present. Test compound and enzymes were added to each well of
a 96-well plate and allowed to equilibrate for 1 h at 378C prior to
the addition of substrate. Determinations of fluorescence was per-
formed by using a Varian Cary Eclipse spectrofluorimeter (lex =
328 nm, lem =393 nm) after the addition of fluorogenic substrate
(t0) and after 3 h incubation at 378C (t3h). The inhibitor potency
was evaluated from the amount of substrate cleavage calculated
from the change in fluorescence between the two measurements,
t0 and t3h, at a range of test compound concentrations (at least six
concentrations, each in duplicate). From the resulting dose–re-
sponse curve, the IC50 values were calculated from a three-parame-
ter logistic fit of the data (GraphPad Prism version 4.00 for Win-
dows, GraphPad Software, San Diego, CA, USA).
Conflict of interest
The authors declare no conflict of interest.
Keywords: metalloproteinases · gelatinases · b-carbolines ·
molecular docking · selectivity
[1] R. H. Holm, P. Kennepohl, E. I. Solomon, Chem. Rev. 1996, 96, 2239.
[2] J. D. Mott, Z. Werb, Curr. Opin. Cell Biol. 2004, 16, 558.
[3] H. Nagase, R. Visse, G. Murphy, Cardiovasc. Res. 2006, 69, 562.
[4] O. Nicolotti, T. F. Miscioscia, F. Leonetti, G. Muncipinto, A. Carotti, J.
Chem. Inf. Model. 2007, 47, 2439.
[5] I. Giangreco, O. Nicolotti, A. Carotti, F. De Carlo, G. Gargano, R. Bellotti,
BMC Bioinf. 2010, 11, 500.
[6] C. M. Overall, O. Kleifeld, Nat. Rev. Cancer 2006, 6, 227.
[7] J. E. Rundhaug, J. Cell. Mol. Med. 2005, 9, 267.
[8] E. I. Deryugina, J. P. Quigley, Cancer Metastasis Rev. 2006, 25, 9.
[9] J. Hu, P. E. Van den Steen, Q.-X. A. Sang, G. Opdenakker, Nat. Rev. Drug
Discovery 2007, 6, 480.
[10] C. M. Overall, O. Kleifeld, Br. J. Cancer 2006, 94, 941.
[11] L. M. Coussens, B. Fingleton, L. M. Matrisian, Science 2002, 295, 2387.
[12] M. Pavlaki, S. Zucker, Cancer Metastasis Rev. 2003, 22, 177.
[13] N. Adhikari, A. Mukherjee, A. Saha, T. Jha, Eur. J. Med. Chem. 2017, 129,
72.
[14] B. Fabre, A. Ramos, B. de Pascual-Teresa, J. Med. Chem. 2014, 57, 10205.
[15] S. P. Gupta, V. M. Patil, Exper. Suppl. 2012, 103, 35.
[16] E. Koivunen, W. Arap, H. Valtanen, A. Rainisalo, O. P. Medina, P. Heikkilꢃ,
C. Kantor, C. G. Gahmberg, T. Salo, Y. T. Konttinen, T. Sorsa, E. Ruoslahti,
R. Pasqualini, Nat. Biotechnol. 1999, 17, 768.
[17] C. M. Overall, C. Lꢄpez-Otꢅn, Nat. Rev. Cancer 2002, 2, 657.
[18] M. W. Roomi, J. C. Monterrey, T. Kalinovsky, M. Rath, A. Niedzwiecki,
Oncol. Rep. 2009, 21, 1323.
QM polarised ligand docking simulations
The 3D structures of MMP-2 (PDB ID: 1HOV)[46] and -9 (PDB ID:
4H3X)[47] were retrieved from the Protein Data Bank[55] and pre-
treated by using the protein preparation module available from
the Schrçdinger suite 2016-4,[56] which enables the addition of
missing hydrogen atoms, optimisation of missing residues, removal
of water molecules and determination of the optimal protonation
states for histidine residues. In the case of the 1HOV NMR-solved
structure, the first copy was selected because the protein structure
was essentially unchanged regardless of which NMR solution was
considered. For residues 12–163, the atomic root-mean-square de-
viations (RMSDs) from the mean coordinate positions were as
small as 1.02 and 1.62 ꢂ for the backbone and all heavy atoms, re-
spectively. The obtained protein structures were employed to per-
form highly accurate docking simulations by using the QM polar-
ised ligand docking protocol available from the Schrçdinger Suite
2016-4.[57,58] During the docking process, the receptor protein was
fixed, whereas full conformational flexibility was allowed for the li-
gands. In particular, a cubic grid centred on the centre of mass of
the cognate ligand with an automatic edge was used. The QM po-
larised ligand docking protocol consisted of different phases. In
the first step, ligands were docked with Glide by using a standard
precision (SP) protocol. Then, QM partial charges of the ligand
were computed in the field generated by the receptor at the BLYP/
3-21G level of theory. Finally, each ligand pose was re-docked to
account for the computed QM charges following the extra preci-
sion (XP) protocol. Such a protocol reproduced the binding confor-
mations of the co-crystallised ligands. In particular, RMSD values of
2.16 and 0.47 ꢂ were measured by comparing the experimental co-
ordinates with the top-scored docking poses of n-{4-[(1-hydroxy-
carbamoyl-2-methylpropyl)(2-morpholin-4-yl-ethyl)sulfamoyl]-4-
pentylbenzamide (PDB ID: 1HOV)[46] and N-2-(biphenyl-4-ylsulfon-
yl)-N-2-(isopropyloxy)acetohydroxamic acid (PDB ID: 4H3X),[47] re-
spectively.
[19] S. A. Amin, N. Adhikari, T. Jha, Pharmacol. Res. 2017, 122, 8.
[20] H. Li, Z. Qiu, F. Li, C. Wang, Oncol. Lett. 2017, 14, 5865.
[21] M. Egeblad, Z. Werb, Nat. Rev. Cancer 2002, 2, 161.
[22] N. I. Medeiros, R. C. G. Fares, E. P. Franco, G. R. Sousa, R. T. Mattos, A. T.
Chaves, M. D. C. P. Nunes, W. O. Dutra, R. Correa-Oliveira, M. O. C. Rocha,
J. A. S. Gomes, PLoS Neglected Trop. Dis. 2017, 11, e0005284.
[23] R. Reich, C. T. Supuran, E. Breuer in Annual Reports in Medicinal Chemis-
try, Vol. 49 (Ed.: M. C. Desai), Academic Press, San Diego, 2014, pp. 269–
284.
[24] G. Shay, M. Tauro, F. Loiodice, P. Tortorella, D. M. Sullivan, L. A. Hazle-
hurst, C. C. Lynch, Oncotarget 2017, 8, 41827.
[25] S. M. Yang, R. H. Scannevin, B. Wang, S. L. Burke, L. J. Wilson, P. Karnachi,
K. J. Rhodes, B. Lagu, W. V. Murray, Bioorg. Med. Chem. Lett. 2008, 18,
1135.
[26] P. Jain, C. Saravanan, S. K. Singh, Eur. J. Med. Chem. 2013, 60, 89.
[27] B. Fabre, K. Filipiak, N. Dꢅaz, J. M. Zapico, D. Suꢆrez, A. Ramos, B. de
Pascual-Teresa, ChemBioChem 2014, 15, 399.
[28] O. Nicolotti, M. Catto, I. Giangreco, M. Barletta, F. Leonetti, A. Stefanachi,
L. Pisani, S. Cellamare, P. Tortorella, F. Loiodice, A. Carotti, Eur. J. Med.
Chem. 2012, 58, 368.
[29] D. P. Becker, C. I. Villamil, T. E. Barta, L. J. Bedell, T. L. Boehm, G. A. De-
crescenzo, J. N. Freskos, D. P. Getman, S. Hockerman, R. Heintz, S. C.
Howard, M. H. Li, J. J. McDonald, C. P. Carron, C. L. Funckes-Shippy, P. P.
Mehta, G. E. Munie, C. A. Swearingen, J. Med. Chem. 2005, 48, 6713.
[30] D. Leung, G. Abbenante, D. P. Fairlie, J. Med. Chem. 2000, 43, 305.
[31] R. J. Cherney, R. Mo, D. T. Meyer, K. D. Hardman, R. Q. Liu, M. B. Coving-
ton, M. Qian, Z. R. Wasserman, D. D. Christ, J. M. Trzaskos, R. C. Newton,
C. P. Decicco, J. Med. Chem. 2004, 47, 2981.
[32] M. C. Noe, V. Natarajan, S. L. Snow, P. G. Mitchell, L. Lopresti-Morrow,
L. M. Reeves, S. A. Yocum, T. J. Carty, J. A. Barberia, F. J. Sweeney, J. L.
Liras, M. Vaughn, J. R. Hardink, J. M. Hawkins, C. Tokar, Bioorg. Med.
Chem. Lett. 2005, 15, 2808.
[33] L. A. Reiter, R. P. Robinson, K. F. McClure, C. S. Jones, M. R. Reese, P. G.
Mitchell, I. G. Otterness, M. L. Bliven, J. Liras, S. R. Cortina, K. M. Dona-
hue, J. D. Eskra, R. J. Griffiths, M. E. Lame, A. Lopez-Anaya, G. J. Martinel-
ChemMedChem 2018, 13, 1 – 11
www.chemmedchem.org
9
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
li, S. M. McGahee, S. A. Yocum, L. L. Lopresti-Morrow, L. M. Tobiassen,
M. L. Vaughn-Bowser, Bioorg. Med. Chem. Lett. 2004, 14, 3389.
[34] B. Grella, M. Teitler, C. Smith, K. Herrick-Davis, R. A. Glennon, Bioorg.
Med. Chem. Lett. 2003, 13, 4421.
[35] N. Khorana, C. Smith, K. Herrick-Davis, A. Purohit, M. Teitler, B. Grella, M.
Dukat, R. A. Glennon, J. Med. Chem. 2003, 46, 3930.
[48] The UniProt Consortium, Nucleic Acids Res. 2017, 45, D158.
[49] Schrçdinger Release 2016-4: QikProp, Schrçdinger, LLC, New York, NY,
2016.
[50] C. A. Lipinski, Drug Discovery Today Technol. 2004, 1, 337.
[51] C. A. Lipinski, F. Lombardo, B. W. Dominy, P. J. Feeney, Adv. Drug Delivery
Rev. 2001, 46, 3.
[36] R. A. Glennon, B. Grella, R. J. Tyacke, A. Lau, J. Westaway, A. L. Hudson,
Bioorg. Med. Chem. Lett. 2004, 14, 527.
[37] N. Khorana, K. Changwichit, K. Ingkaninan, M. Utsintong, Bioorg. Med.
Chem. Lett. 2012, 22, 2885.
[38] J. I. Trujillo, M. J. Meyers, D. R. Anderson, S. Hegde, M. W. Mahoney, W. F.
Vernier, I. P. Buchler, K. K. Wu, S. Yang, S. J. Hartmann, D. B. Reitz, Bioorg.
Med. Chem. Lett. 2007, 17, 4657.
[39] F. Liu, L.-Q. Yu, C. Jiang, L. Yang, W.-T. Wu, Q.-D. You, Bioorg. Med. Chem.
2010, 18, 4167.
[52] B. T. Ho, K. E. Walker, Org. Synth. 1971, 51, 136.
[53] C. Kaduk, H. Wenschuh, M. Beyermann, K. Forner, L. A. Carpino, M. Bien-
ert, Lett. Pept. Sci. 1996, 2, 285.
[54] S. E. Zook, R. Dagnino, M. Deason, S. Bender, M. Melnick, Eur. Pat. No.
EP0874830, 2003.
[55] H. M. Berman, J. Westbrook, Z. Feng, G. Gilliland, T. N. Bhat, H. Weissig,
I. N. Shindyalov, P. E. Bourne, Nucleic Acids Res. 2000, 28, 235.
[56] G. M. Sastry, M. Adzhigirey, T. Day, R. Annabhimoju, W. Sherman, J.
Comput.-Aided Mol. Des. 2013, 27, 221.
[40] H. Song, S. Zhou, R. Wang, S. Li, ChemMedChem 2013, 8, 1736.
[41] H. Matter, W. Schwab, J. Med. Chem. 1999, 42, 4506.
[42] C. G. Knight, F. Willenbrock, G. Murphy, FEBS Lett. 1992, 296, 263.
[43] “Development and application of new steric substituent parameters in
drug design”: A. Verloop, W. Hoogenstraaten, J. Tipker in Drug Design,
Vol. 7 (Ed.: C. J. Ariens), Academic Press, New York, 1976, pp. 165–207.
[44] C. Hansch, A. R. Steward, J. Iwasa, E. W. Deutsch, Mol. Pharmacol. 1965,
1, 205.
[57] Schrçdinger Release 2016-4, QSite, Schrçdinger, LLC, New York, NY,
2016.
[58] A. E. Cho, V. Guallar, B. J. Berne, R. Friesner, J. Comput. Chem. 2005, 26,
915.
[59] D. R. Shalinsky, J. Brekken, H. Zou, C. D. Mcdermott, P. Forsyth, D. Ed-
wards, S. Margosiak, S. Bender, G. Truitt, A. Wood, N. M. Varki, K. Appelt,
Ann. N. Y. Acad. Sci. 1999, 878, 236.
[45] L. P. Hammett, J. Am. Chem. Soc. 1937, 59, 96.
[46] Y. Feng, J. J. Likos, L. Zhu, H. Woodward, G. Munie, J. J. McDonald, A. M.
Stevens, C. P. Howard, G. A. De Crescenzo, D. Welsch, H. S. Shieh, W. C.
Stallings, Biochim. Biophys. Acta Proteins Proteomics 2002, 1598, 10.
[47] C. Antoni, L. Vera, L. Devel, M. P. Catalani, B. Czarny, E. Cassar-Lajeu-
nesse, E. Nuti, A. Rossello, V. Dive, E. A. Stura, J. Struct. Biol. 2013, 182,
246.
Manuscript received: April 11, 2018
Revised manuscript received: May 3, 2018
Accepted manuscript online: && &&, 0000
Version of record online: && &&, 0000
&
ChemMedChem 2018, 13, 1 – 11
www.chemmedchem.org
10
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

FULL PAPERS
G. F. Mangiatordi, T. Guzzo, E. C. Rossano,
D. Trisciuzzi, D. Alberga, G. Fasciglione,
M. Coletta, A. Topai,* O. Nicolotti*
&& – &&
Design, Synthesis, and Biological
Evaluation of Tetrahydro-b-carboline
Derivatives as Selective Sub-
Clear to dock! A new series of tetrahy-
dro-b-carboline derivatives with out-
standing inhibitory activity toward
matrix metalloproteinases (MMPs) is
presented. Four compounds show re-
markable selectivity and excellent activi-
ty in the sub-nanomolar range toward
both gelatinases MMP-2 and MMP-9.
Molecular docking simulations allow the
identification of the key interactions re-
sponsible for the detected high affinity
(see figure).
Nanomolar Gelatinase Inhibitors
ChemMedChem 2018, 13, 1 – 11
www.chemmedchem.org
11
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ