International Journal of
Molecular Sciences Article
Seeking for Non-Zinc-Binding MMP-2 Inhibitors: Synthesis, Biological Evaluation and Molecular Modelling Studies Alessandra Ammazzalorso 1,† , Barbara De Filippis 1,† , Cristina Campestre 1 , Antonio Laghezza 2 , Alessandro Marrone 1 , Rosa Amoroso 1 , Paolo Tortorella 2 and Mariangela Agamennone 1, * 1
2
* †
Dipartimento di Farmacia, Università “G. d’Annunzio” Chieti, Via dei Vestini 31, 66100 Chieti, Italy; [email protected] (A.A.); [email protected] (B.D.F.); [email protected] (C.C.); [email protected] (A.M.); [email protected] (R.A.) Dipartimento di Farmacia-Scienze del Farmaco, Università “A. Moro” Bari, Via Orabona 4, 70125 Bari, Italy; [email protected] (A.L.); [email protected] (P.T.) Correspondence: [email protected]; Tel.: +39-0871-355-4585 These authors contributed equally to the work.
Academic Editors: Claudiu T. Supuran and Koji Sode Received: 29 July 2016; Accepted: 14 October 2016; Published: 22 October 2016
Abstract: Matrix metalloproteinases (MMPs) are an important family of zinc-containing enzymes with a central role in many physiological and pathological processes. Although several MMP inhibitors have been synthesized over the years, none reached the market because of off-target effects, due to the presence of a zinc binding group in the inhibitor structure. To overcome this problem non-zinc-binding inhibitors (NZIs) have been recently designed. In a previous article, a virtual screening campaign identified some hydroxynaphtyridine and hydroxyquinoline as MMP-2 non-zinc-binding inhibitors. In the present work, simplified analogues of previously-identified hits have been synthesized and tested in enzyme inhibition assays. Docking and molecular dynamics studies were carried out to rationalize the activity data. Keywords: MMPs; inhibitors; molecular dynamics; docking; quinoline; isoquinoline; ureas; non-zinc-binding group
1. Introduction Matrix metalloproteinases are calcium and zinc-containing proteases (24 in humans) involved in the hydrolysis of the extracellular matrix (ECM) components and, therefore, responsible for tissue remodeling in physiological conditions [1]. Unbalanced matrix metalloproteinase (MMP) activity has been observed in many pathologies, such as cancer, neurodegeneration, cardiovascular diseases and others [2–11]; for this reason, they represent a well-studied therapeutic target. More recently, it has been demonstrated that MMP involvement in several diseases can be due to the hydrolysis of substrates other than ECM components, in particular chemokines involved in inflammation and immune response [1,12]. Ever since the first ligands had been disclosed almost 30 years ago, several MMP inhibitors (MMPIs) have been discovered, but none reached the market because of relevant side effects emerging in long-term treatment (musculoskeletal syndrome (MSS)) [13]. Traditional MMPIs are constituted by a functional group binding the zinc ion (zinc binding group (ZBG)) and chemical functions mimicking the substrate and interacting with the MMP subsites, in particular with the hydrophobic S1’ site. The ZBG is represented in most of the cases by the hydroxamate function, that binds very efficiently in the MMP active site, but can possibly bind other divalent cations and cause an unselective inhibition of other metallo-enzymes.
Int. J. Mol. Sci. 2016, 17, 1768; doi:10.3390/ijms17101768
www.mdpi.com/journal/ijms
Int. J. Mol. Sci. 2016, 17, 1768
2 of 19
Another important issue of MMP inhibition is represented by the unselective inhibition of the studied MMPIs. In fact, clinically-tested compounds (e.g., batimastat and marimastat) were unselective Int. J. Mol. Sci. 2016, 17, 1768 2 of 19 MMP inhibitors not discriminating among different isoforms; while, during the years, the diverse role played byAnother each MMP has been better understood [14]. In particular, anticancer therapy, it has important issue of MMP inhibition is represented byinthe unselective inhibition of been the that studied clinically-tested compounds (e.g., batimastat and marimastat) weresome defined someMMPIs. MMPs In canfact, be considered therapeutic targets (MMP-1, MMP-2, MMP-7), while MMPasinhibitors not (MMP-8) discriminating othersunselective are classified anti-targets [15]. among different isoforms; while, during the years, the diverse role played by each MMP has been betterapproaches understood [6]. In particular, in anticancer These issues prompted the search for alternative to the MMP inhibition [16], such therapy, it has been defined that some MMPs can be considered therapeutic targets (MMP-1, as the substitution of the ZBG with less potent, more selective zinc binders [17–22], the use of MMP-2, MMP-7), while some others are classified as anti-targets (MMP-8) [7]. MMP-directed antibody [23] and the design of non-zinc binding inhibitors (NZIs) [24]. These ligands, These issues prompted the search for alternative approaches to the MMP inhibition [8], such as commonly reported as third-generation MMPIs, have been developed for MMP-13 [25], MMP-8 [26] the substitution of the ZBG with less potent, more selective zinc binders [9,18–22], the use of and MMP-12 [27] and can reach a better selectivity by binding missinginhibitors the ZBG. Recently, we attempted MMP-directed antibody [23] and the design of non-zinc (NZIs) [24]. These ligands, to identify MMP-2 NZIs characterized by enhanced interactions with the S1’ pocket [28]. Our interest commonly reported as third-generation MMPIs, have been developed for MMP-13 [25], MMP-8 [26] toward MMP-2 selective inhibition is due to the demonstrated role of MMP-2 in tumors. MMP-2, and MMP-12 [27] and can reach a better selectivity by missing the ZBG. Recently, we attempted to identifygelatinase MMP-2 NZIs characterized by enhanced interactions with the to S1’tissue pockethomeostasis. [28]. Our interest also called A, is constitutively expressed and contributes Even if toward MMP-2MMP-2 selectivehas inhibition is dueoverexpressed to the demonstrated role of tumors MMP-2 playing in tumors. poorly inducible, been found in several a MMP-2, role in both also calledand gelatinase is constitutively tumorigenesis cancerA, progression [29]. expressed and contributes to tissue homeostasis. Even if poorly inducible, MMP-2 has been found overexpressed in several tumors playing a role in both MMP-8 and MMP-13 are classified as collagenases because they mainly degrade collagen; MMP-13 tumorigenesis and cancer progression [29]. is able to degrade the components of the basal membrane and is involved in tumor dissemination and MMP-8 and MMP-13 are classified as collagenases because they mainly degrade collagen; progression. the contrary, has a confirmed protective role in tumor progression, and its MMP-13 On is able to degradeMMP-8 the components of the basal membrane and is involved in tumor inhibition should be avoided [29]. The final aim of our study is, therefore, to identify selective dissemination and progression. On the contrary, MMP-8 has a confirmed protective role in tumorNZIs blocking MMP-2 and and its MMP-13 activity, sparing theThe antitarget MMP-8. progression, inhibition shouldwhile be avoided [29]. final aim of our study is, therefore, to In the previous work,blocking we moved from analysis of the superposition betweenMMP-8. MMP-2 and identify selective NZIs MMP-2 andthe MMP-13 activity, while sparing the antitarget the previous we moved from the analysis of the superposition between MMP-2inhibitors and MMP-8 inIncomplex with work, a non-zinc-binding ligand that highlighted the possibility to design MMP-8 in complex with a non-zinc-binding ligand that highlighted the possibility to design tailored to (potential) additional interactions in the MMP-2 S1’ site. The following virtual screening inhibitors tailored (potential) additional interactions in the MMP-2 site. The following virtual campaign afforded onetohit, the hydroxynaphtyridine derivative Hit 1, S1’ binding MMP-2 at a micromolar screening campaign afforded one hit, the hydroxynaphtyridine derivative Hit 1, binding MMP-2 at concentration (IC50 = 5 µM (MMP-2), 75 µM (MMP-8) and 36 µM (MMP-13)). An analogue testing a micromolar concentration (IC50 = 5 µM (MMP-2), 75 µM (MMP-8) and 36 µM (MMP-13)). An allowed identifying also hydroxyquinoline compounds (i.e., Hit 2, IC50 = 15 µM (MMP-2), 105 µM analogue testing allowed identifying also hydroxyquinoline compounds (i.e., Hit 2, IC50 = 15 µM (MMP-8) and 5105 µM (MMP-13)), almost active thanequally the prior hitthan (Figure 1). hit (Figure 1). (MMP-2), µM (MMP-8) and 5 µMequally (MMP-13)), almost active the prior
Figure 1. Chemical structures of Hits 1 and 2.
Figure 1. Chemical structures of Hits 1 and 2.
Moving from these data, in the present work we investigated the structural determinants essential from for the inhibition non-zinc-binding ligands. Activity determinants data and docking Moving these data, in of theMMP-2 presentby work we investigated the structural essential studies of hydroxynaphtyridine/hydroxyquinoline derivatives suggested the role of relevant for the inhibition of MMP-2 by non-zinc-binding ligands. Activity data and docking studies of structural features: hydroxynaphtyridine/hydroxyquinoline derivatives suggested the role of relevant structural features:
• • •
•
the furyl/aryl group that provides the π–π stacking with His201 (MMP-2 numbering);
the furyl/aryl that provides the π–π stacking with His201 (MMP-2 numbering); • the CHgroup 2 spacer between the furyl/aryl group and the urea function; the CH between furyl/aryl andinthe • 2 spacer the urea functionthe that provides group H-bonds theurea S1’ function; site with NH of Thr227 and CO of Leu218. the urea function that provides H-bonds in the S1’ site with NH of Thr227 and CO of Leu218. Starting from these evidences, we described the synthesis and the biological evaluation of new
Starting from these evidences, we described thestructural synthesis and the biological evaluation of new MMP inhibitors, based on the combination of three portions in which the lead compound MMPcould inhibitors, based on theincombination of three structural portions in which the leadofcompound be ideally divided, order to verify their importance for the selective inhibition MMP couldisoforms. be ideally divided, in order to verify importance for the selective inhibition In all compounds, the furyl group their was replaced by a more common benzyl function. of In aMMP isoforms. In allofcompounds, was replaced by a unaltered, more common benzyl function. first series compounds, the we furyl kept group the benzyl-ureidic scaffold by modifying the hydroxyquinoline portion with heterocycles (quinoline, isoquinoline, indole, indazole), aromatic In a first series of compounds, we kept the benzyl-ureidic scaffold unaltered, by modifying the rings (naphthalene, benzene, pyridine), alkyl chains isoquinoline, or saturated heterocycles (piperidine, hydroxyquinoline portion with heterocycles (quinoline, indole, indazole), aromatic piperazine, pyrrolidine) (Figure 2; Table 1). rings (naphthalene, benzene, pyridine), alkyl chains or saturated heterocycles (piperidine, piperazine, pyrrolidine) (Figure 2; Table 1).
Int. J. Mol. Sci. 2016, 17, 1768
3 of 19
Int. J. Mol. Sci. 2016, 17, 1768
3 of 19
Int. J. Mol. Sci. 2016,Int. 17,J.1768 Mol. Sci. 2016, 17, 1768 Int. J. J. Mol. Mol. Sci. Sci. 2016, 2016, 17, 17, 1768 Int. J.J.J. Mol. Int. Mol. Sci. Sci. 2016, 2016, 17, 17, 1768 1768 Int. 1768 Int. Mol. Sci. 2016, 17, 1768 Int. J.J. Mol. Mol. Sci. Sci. 2016, 2016, 17, 17, 1768 Int. 1768 Int. Int. J. J. Mol. Mol. Sci. Sci. 2016, 2016, 17, 17, 1768 1768 Int. J. Mol. Sci. 2016, 17, 1768
3 of 19 3 of of 19 19 of 19 19 333 of
3 of 19 333 of of 19 19 of 19 33 of of 19 19 3 of 19
4 of 19
4 of 19
Figure 2. Structural modifications performed on Hit 2.
Figure 2. Structural modifications performed on Hit 2.
Figure 2. Structural modifications performed on Hitperformed 2. Figure 2. Structural modifications on Hit 2. Figure 2. 2. Structural Structural modifications performed on Hit Hit 2. 2.performed 2. Structural modifications Figure 2.derivatives Structural modifications performed on on Hit Hit 2. 2. Figure modifications performed on TableFigure 1. Urea 1a–α.on Figure 2. 2. Structural Structural modifications performed on Hit Hit 2. 2.performed Figure modifications performed Figure 2. Structural modifications on Hit 2. Figure 2. Structural modifications performed on Hit 2. Table 1. Urea derivatives 1a–α. Structural modifications performed on Hit 2. TableFigure 1. Urea2.derivatives 1a–α.derivatives Table Urea 1a–α. O 1. Table 1. 1. Urea Urea derivatives derivatives 1a–α. Table Urea Table 1a–α. Table 1. 1. Urea derivatives derivatives 1a–α. 1a–α. Table 1. 1. Urea Urea derivatives derivatives 1a–α. Table 1a–α. Table Table 1. 1.RUrea Urea derivatives derivatives 1a–α. 1a–α. O N1. Urea derivatives N Table 1a–α. O O
ID ID
1a
R R RN
ID ID1a ID ID 1a 1a 1a 1a 1b
1b
1b 1b 1b 1b 1c
1c
1c 1c 1c 1d
1d
1d 1d 1d 1d 1e
1e
1e 1e 1e 1e 1f
1f
1f 1f 1f 1g
1g
1g 1g 1g 1g 1h
1h
1h 1h 1h 1h 1i
1i
1i 1i 1i 1j
R R R
N N N N N
N N N
N N N N N
N
N N N N N
N
ID ID ID1n ID IDID ID IDID ID 1n 1a 1n 1a 1n 1a1n 1a 1a 1a 1o 1o 1b 1o 1b 1b1o 1b 1b1o 1b 1p 1p 1c 1p 1c 1c 1c1p 1c 1c 1q 1q 1d 1q 1d 1d1q 1q 1d 1d N 1d 1r
1r N 1e 1r N 1e 1r 1r N 1e N N 1e N1e 1e 1s N N 1s N N N 1f 1s 1f 1s N1f 1f 1f 1f 1t N N N N N
N N N N
N N N N N
N N H
1l N
N H
N
1t 1g 1t 1g 1g 1t N 1t 1g 1g 1g 1u N 1u 1h 1u 1h 1h1u 1u 1h 1h N 1h 1v N 1v 1i 1v 1i 1i 1v 1i 1i 1i 1w 1w 1j 1w N 1j 1w 1j1w N 1j N N 1j H N N H H 1j 1xHH H N H
ID R R R R R R R RN R N N RN N 1n N N N N N N
1o
1o 1o 1o 1o 1o 1o
1p N
1p 1p 1p 1p 1p 1p
1q
1q 1q 1q 1q 1q 1q
N
N
N N N NN N N N NN N N N N
N N NN N NN N N N N
N N N N N N
N H
HO N N N N N HON HO N HO HO HO1s
N H
R R R R R R
R R R R R R R
N N N N N N N
N N N N N N N
1s 1s 1s 1s 1s 1s3 OCH
HO HO HO HO HO HO HO
N OCH N N 3 N N 1t 3 N OCH 1t 1t OCH 3 OCH 31t N OCH 33 OCH OCH 3 3 OCH 331t OCH OCH 1t 1t3CO3333333 OCH OCH H
OCH3 OCH OCH3333 OCH 33 OCH3OCH OCH OCH3333OCH3 OCH OCH OCH33OCH3
N
3
N N OCH3 N H3CON 1u N 1u 1u H33CO CO N H 1u 3 OCH 3 H CO H 3 H333CO CO OCH331u OCH OCH333331u OCH 1u OCH N N N N N N 1v N
1v 1v 1v 1v 1v 1v
1w
1w 1w 1w 1w 1w 1w
1x
1x 1x 1x 1x 1x 1x
OCH3 H3CO H H3333CO CO H CO H H333CO CO OCH OCH OCH33333 OCH OCH H3CO OCH333 OCH3
N
N
1y 1y N
1m 1z N N O
N N N N N H N H H H H N H H
1r 1r 1r 1r 1r 1r
1r N
H
1l
ID ID ID ID ID 1n 1n 1n 1n 1n 1n
N N
1x 1k 1x N 1k 1k1x 1x N 1k N N 1k H N N H H H 1k H H N 1y N
O O O O O N N N O N N H N H H H H N H H ID
R
N
N H N N N NH N N H H H H N 1j H 1k N 1k 1k N H 1k N Int. J. Mol. Sci. 2016, 17, 1768 N N Int. J. Mol. Sci. 2016, 17, 1768 N H H H H H 1k
1m
N
N N N N N
1j 1j 1j 1j 1k
1l
O H HR O O R N N R R R NH N NH N N N N N H H ID H H H H H H
N
1z
H
O N N
N H
N N
Int. J. Mol. Sci. 2016, 17, 1768
4 of 19
Figure 2. Structural modifications performed on Hit 2.
Table 1. Cont.
Table 1. Urea derivatives 1a–α. O
Mol. Sci. 2016, 17, 1768 Int. J. Mol.Int. Sci.J. 2016, 17, 1768 Int. J. Mol. Sci. 2016, 17, 1768
1l
N
Int. J. J. Mol. Mol. Sci. Sci. 2016, 2016, 17, 17, 1768 1768 Int.
1m 1m
N
1l ID
1l
N 1m 1a 1lN N N N 1l HNN NN H H 1l N
H H N N N N N NH
1b1m 1m 1m
N
ID R 1y
H
NN 1z 1m 1z1n N N 1y N N 1y N 1y 1z H
4 of 19
4 of 19
4 of 19
R
of 19 19 44 of
N
1z
N
O 1o 1z N N N 1z H N1z
O
NH H H
1α
N
1y N IDN
N
4 of 19
R
N H N
1y RR
N H
N H
Int. J. Mol. Sci. 2016, 17, 1768 ID
N H
N N N
N
O
N
O
N
N
N
H
N
O O 1c N1p N Hstarting N A second group of derivatives was obtained from the isoquinolinyl derivative N from N the A second group of derivatives was obtained starting isoquinolinyl derivative 1h, by 1h, byderivative 1h, by N N A derivatives second group of derivatives was obtained starting from the isoquinolinyl A second group of was obtained from the isoquinolinyl derivative H starting H N replacing the ureidic linker with an amide or sulfonamide group (2a–e; Figure 3; Table 2). Finally, 1h, by N Figure 3; Table 2). Finally, replacing the ureidic linker with an amide orlinker sulfonamide group N (2a–e; Nwith an amide or sulfonamide group (2a–e; Figure 3; Table 2). Finally, replacing the ureidic replacing the ureidic linker with an amide or sulfonamide group (2a–e; Figure 3; Table 1d 1q A second group of derivatives was obtained starting from the isoquinolinyl derivative 1h,the by2). novel ureidic derivatives were designed by adding para substituents on the ring, benzyl ring, aimFinally, novel ureidic derivatives were ureidic designed by adding paradesigned substituents on the benzyl with thewith aim novel derivatives were by adding para substituents on the benzyl ring, with the aim replacing the ureidic linker with an amide or sulfonamide group (2a–e; Figure 3; Table 2). Finally, novel ureidic derivatives were designed by adding para substituents on the benzyl ring, with A second group of derivatives was obtained starting from the isoquinolinyl derivative 1h, bythe aim to explore of thefunctional functional group, potentially able the to interact with residues at the top to explore the effect ofeffect the group, able togroup, interact with residues at interact the top Athe second group ofpolar derivatives obtained starting from isoquinolinyl derivative 1h, by topolar explore the effect of was thepotentially polar functional potentially able to with residues at the top novel ureidic derivatives were designed by adding para substituents on the benzyl ring, with the aim replacing the ureidic linker with an amide or sulfonamide group (2a–e; Figure 3; Table 2). Finally, of site thereplacing S1’ site (5 and Figure 3;3). Table 3). group, to S1’ explore effect of7;the polar functional potentially able(2a–e; to interact residues at the top of the (5the and 7; Figure 3; Table the linker with an or sulfonamide group Figure 3;with Table 2). Finally, ofureidic the S1’ 7; amide Figuregroup, 3). N3; Table to explore the effect ofsite the (5 polar potentially able to interact with residues at the top 1r 1eandfunctional
novel ureidic derivatives were designed by adding adding para para substituents substituents on on the the benzyl benzyl ring, ring, with with the the aim aim novel ureidic derivatives were designed of the S1’ site (5S1’ and 3; Table 3). 3).by of the site7; (5 Figure and 7; Figure 3; Table
O O to interact O group, potentially able O to explore explore the the effect effect of of the the polar polar functional functional with residues at the top to group, potentially O able to interact with residuesOat the top HN O N HN N ofLinker the S1’ S1’Linker site (5 (5 and and 7; 7; Figure FigureHN 3; Table Table 3). NN N HN of the site 3; 3). O HN N Linker 1f
n
n
Linker
N
n
N N
2a-e
Linker Linker
2a-e
N amide N amide
N
sulfonamide
N
1g
2a-e
2a-e
inverse amide amide inverse amide inverse sulfonamide 2a-eamide sulfonamide 2a-e
N
N
n n
HN
amide inverse amide 1h sulfonamide
N O H O
N
N
NN N H 1h H
HN HN
1h
HN
1s
H
H
n
H
H
HN
N
N H
N
OH O
HN HN
N
1t
X
N N H H 5,7
OCH35,7
1h N OCH3 N
1h
N N
N
HO H
N
NN H
5,7
amide amide sulfonamide amide sulfonamide
N 1h 1h
1u
X
X NX N H H
5,7 sulfonamide 5,7
H3CO
amide amide
OCH3
H
N H
H X
X
5,7 amide sulfonamide
amide
amide 3. From isoquinolinyl urea 1h1h to novel derivatives 2a–e and 5, 7. inverse amide isoquinolinyl Figure 3.Figure From urea 1h to novel derivatives 2a–e and2a–e 5, 7.and sulfonamide inverse amide Figure 3. FromFigure isoquinolinyl urea to novel derivatives 5, 7. sulfonamide 3. From isoquinolinyl 1h to novel derivatives sulfonamide Figure 3. From isoquinolinyl urea 1h to novelurea derivatives 2a–e and 5, 7.2a–e and 5, 7. sulfonamide
N 1i 1v To provide some rational explanation activity data, structure-based computational To provide some rational explanation of rational the of activity data, structure-based To provide some explanation ofthe theto activity data, Figure 3.rational From some isoquinolinyl urea 1h novel derivatives derivatives 2a–e and andcomputational 5, 7. computational To provide explanation of thestructure-based activity data, structure-based computational Figure 3. From isoquinolinyl urea 1h to novel 2a–e 5, 7. approaches were applied to the study of the binding process. approaches were applied to the study of the binding process. approaches applied to the study ofexplanation the binding process. To were provide some rational ofstudy the of activity data,process. structure-based computational approaches were applied to the the binding
To provide some rational explanation of 1w the process. activity data, data, structure-based structure-based computational computational 1j explanation provide some rational of the activity approaches To were applied toTable the study of binding N the 2. Isoquinoline amide and sulfonamide derivatives 2a–2e. 2. Isoquinoline amide and sulfonamide derivatives Table 2.Table Isoquinoline sulfonamide derivatives 2a–2e. 2a–2e. approaches were applied to the theamide study of the the binding process. Hand approaches were applied to study of binding process. Table 2. Isoquinoline amide and sulfonamide derivatives 2a–2e.
Table 2. Isoquinoline amide and sulfonamide derivatives 2a–2e. Table 2. 2.1k Isoquinoline amide amideLinker and1x sulfonamide derivatives derivatives 2a–2e. 2a–2e. Table Isoquinoline and sulfonamide N n Linker
Linker H
Linker Linker
N
N
Linker
n
n
N
n
n
Cpd LinkernN n Linker n1 NHCO Cpd Cpd Linker n Cpd N 2a N Linker n 2b NHCO NHCO 2a 12 2a NHCO 1 2a NHCO 1 Cpd Linker n Cpd Linker n Cpd Linker 2c NHCO CONH 2b 21n 2b NHCO 2 2b NHCO 2 2a NHCO 1 2d CONH 2a NHCO 1 2a NHCO 2c 121 2c CONH CONH 1 2c CONH 1 2b NHCO 2 2b NHCO 2 2e NHSO2 2b NHCO 2d CONH 22 2d2c CONH 2 CONH 1 2c CONH 1 2d CONH 2 2c CONH 1 CONH 2 urea 5 and 7. 2e - isoquinolinyl 2e2d NHSO 2NHSO - 2 2e Table 3. Amide and sulfonamide derivatives of 2d CONH 2 NHSO 2 2d CONH 2 2e NHSO2 -
2e O 2e
NHSO22 NHSO
--
Amide and sulfonamide derivatives of isoquinolinyl urea Table 3.Table Amide3.and sulfonamide derivatives of isoquinolinyl urea 5 and 7. 5 and 7. Table 3. Amide and sulfonamide derivatives of isoquinolinyl urea 5 and 7. HN N
TableTable 3. Amide andand sulfonamide isoquinolinyl 3. Amide sulfonamideHderivatives derivatives ofofisoquinolinyl ureaurea 5 and57.and 7. Table 3. Amide and derivatives of X isoquinolinyl urea 5 and 7. O O sulfonamide HN N
NN HN H
N
Cpd 5 7
N N
N H
O N O H
HN NN X Cpd N HN H 5 HH
7
Cpd Cpd Cpd 5Cpd 7 555 7 77
N
X SO2 CO
O
HN N X N X H X2 H SO X N N HCO H
X Cpd X 2 SOX X 5SO SO222 CO SO 7CO CO CO
N H
X
X SO2 CO
Int. J. Mol. Sci. 2016, 17, 1768
5 of 19
Int. J. Mol. Sci. 2016, 17, 1768
5 of 19
2. Results
2. Int.Results J. Mol. Sci. 2016, 17, 1768 5 of 19 hits The present work moves from a previous study that identified MMP-2 non-zinc-binding Int. J. Mol. Sci. 2016, 17, 1768 of 19 (NZIs) [28]. already stated, this astudy, structural modifications were non-zinc-binding applied to these5 hits NZIs to TheAs present work movesinfrom previous study that identified MMP-2 2. Results (NZIs) [28]. As already stated, in this study, structural modifications were applied to these NZIs probe2.the structural features determining the binding at and the selective inhibition of MMP-2,toone of Results The work from atherapy. previous study that identified MMP-2 hits was probe thepresent structural features determining the binding and thefinal selective of MMP-2, one the most relevant MMPs inmoves anticancer In this at line, the goalinhibition of non-zinc-binding our research project The present work moves from a previous study that identified MMP-2 non-zinc-binding hits (NZIs) [28]. As already stated, in this study, structural modifications were applied to these NZIs totarget of the most relevant MMPs in anticancer therapy. In this line, the final goal of our research project the identification of NZIs particularly suitable for anticancer therapy, being selective toward (NZIs) [28]. As already stated, in this study, structural modifications were applied to these NZIs to probe the structural features determining the binding at and the selective inhibition of MMP-2, one was the identification of NZIs particularly suitable for anticancer therapy, being selective toward MMPs (such as MMP-2 and MMP-13) while sparing anti-target MMPs (such as MMP-8).
probe the structural the binding at and the inhibition of MMP-2, one of the MMPs most relevant MMPs indetermining anticancer therapy. In this line, theselective final goal of ourasresearch project target (such asfeatures MMP-2 and MMP-13) while sparing anti-target MMPs (such MMP-8). of the most relevant MMPs in anticancer therapy. Infor thisanticancer line, the final goal being of ourselective researchtoward project was the identification of NZIs particularly suitable therapy, 2.1. Chemistry wasChemistry the identification of NZIsand particularly for anticancer therapy, being as selective toward 2.1. target MMPs (such as MMP-2 MMP-13) suitable while sparing anti-target MMPs (such MMP-8). targetderivatives MMPs (such1a–α as MMP-2 MMP-13) while by sparing anti-target (suchto as the MMP-8). Urea (Tableand 1) were prepared adding benzyl MMPs isocyanate corresponding Urea derivatives 1a–α (Table 1) were prepared by adding benzyl isocyanate to the 2.1. Chemistry in methanol, or chloroform, or dichloromethane, stirring the reaction at room or amino derivatives corresponding amino derivatives in methanol, or chloroform, or dichloromethane, stirring the 2.1. Chemistry refluxing temperature for 1–24 [30,31].1)The synthesis is depicted in Scheme 1.isocyanate Urea derivatives 1a–αh (Table were by synthesis adding benzyl reaction at room or refluxing temperature for 1–24prepared h [30,31]. The is depicted in Schemeto1. the Urea derivatives 1a–α (Table 1) were prepared by adding benzyl isocyanate to the corresponding amino derivatives in methanol, or chloroform, or dichloromethane, stirring O corresponding amino derivatives in methanol, or chloroform, or dichloromethane, stirring reaction at room or refluxing temperature for 1–24 h [30,31]. The synthesis is depicted in Scheme 1. the R or (b)h [30,31]. The synthesis is depicted reaction at room or refluxing temperature for(a) 1–24 in Scheme 1. NCO N N H O H O R NCO N N R (a) or (b) H H NCO 1a-α N N H H Scheme 1. Reagents and conditions: benzyl isocyanate (1 equnivalent (eq)), amine, or aniline (1 eq), Scheme 1. Reagents and conditions: benzyl isocyanate (1 equnivalent (eq)), amine, or aniline (1 eq), 1a-α MeOH (Method a), or CHCl3, or dichloromethane (DCM) (Method b), room temperature (r.t.) or MeOH (Method a), or CHCl3 , or dichloromethane (DCM) (Method1a-α b), room temperature (r.t.) or refluxing, h, 41%–99%. Scheme 1.1–24 Reagents and conditions: benzyl isocyanate (1 equnivalent (eq)), amine, or aniline (1 eq), refluxing, 1–24 h, 41%–99%. Scheme(Method 1. Reagents andCHCl conditions: benzyl isocyanate (1 equnivalent (eq)), amine, or aniline(r.t.) (1 eq), 3, or dichloromethane (DCM) (Method b), room temperature or MeOH a), or The synthesis of amides (Table 2) was carried out b), byroom adding to a solution or dichloromethane (DCM) (Method temperature (r.t.) or of MeOH (Method or CHCl3, 2a,b refluxing, 1–24 h, a), 41%–99%. 5-aminoisoquinoline and triethylamine (TEA), a solution of out suitable in refluxing, 1–24of h, 41%–99%. The synthesis amides 2a,b (Table 2) was carried by phenylalkyl-chloride adding to a solution (a) or (b)
of
The synthesis of amides 2a,b (Table 2) wasatcarried out hby adding2) to solution2c,d of dichloromethane atand room temperature, while stirring r.t. for 2–24 (Scheme [32].a Amides 5-aminoisoquinoline triethylamine (TEA), a solution of suitable phenylalkyl-chloride in The synthesis of amides 2a,b (Table 2) was carried out by adding to a solution of 5-aminoisoquinoline and triethylamine (TEA), a solution of suitable phenylalkyl-chloride in were obtained by coupling 5-carboxyisoquinoline with benzylamine (for 2c) or 2-phenylethylamine dichloromethane at room temperature, while stirring at r.t. for 2–24 h (Scheme 2) [32]. Amides 2c,d 5-aminoisoquinoline andoftemperature, triethylamine (TEA), a (HOBt), solution of suitable phenylalkyl-chloride in dichloromethane at room while stirring at r.t. for 2–24 h (Scheme 2) [32]. Amides 2c,d (for 2d), in the presence 1-hydroxybenzotriazole N,N-dicyclohexylcarbodiimide (DCC) were were obtained by coupling 5-carboxyisoquinoline with benzylamine (for 2c) or 2-phenylethylamine dichloromethane at room temperature, while stirring at r.t. for 2–24 h (Scheme 2) [32]. Amides 2c,d obtained by coupling 5-carboxyisoquinoline with benzylamine (for 2c) or 2-phenylethylamine and N-methylmorpholine (NMM), in DMF at 0 °C (Scheme 3). (for 2d), in theinpresence of 1-hydroxybenzotriazole (HOBt), N,N-dicyclohexylcarbodiimide (DCC) and were obtained coupling with benzylamine (for 2c) or 2-phenylethylamine (for 2d), the by presence of 5-carboxyisoquinoline 1-hydroxybenzotriazole (HOBt), N,N-dicyclohexylcarbodiimide (DCC) ◦ C (Scheme 3). N-methylmorpholine (NMM), in DMF at 0 (for 2d), in the presence of 1-hydroxybenzotriazole (HOBt), N,N-dicyclohexylcarbodiimide (DCC) and N-methylmorpholine (NMM), in DMF at 0 °C (Scheme 3). O and N-methylmorpholine (NMM), in DMF at 0 °C (Scheme 3). NH2 NH
N
NH2 NH2
N
NH NH
O O
1-2
1-2 1-2
N N 2a-b N N Scheme 2. Reagents and conditions: 5-aminoisoquinoline (1 eq), phenylalkyl-chloride (1 eq), TEA 2a-b (1.5 eq), dry DCM, N2, r.t., 2–24 h, 55%–63%; n = 1, 2a; n = 2, 2b. 2a-b Scheme 2. Reagents and conditions: 5-aminoisoquinoline (1 eq), phenylalkyl-chloride (1 eq), TEA Scheme 2. Reagents and conditions: 5-aminoisoquinoline (1 eq), phenylalkyl-chloride (1 eq), TEA Scheme 2. Reagents 5-aminoisoquinoline , r.t., conditions: 2–24 h, 55%–63%; n = 1, 2a; n = 2, (1 2b.eq), phenylalkyl-chloride (1 eq), TEA (1.5 eq), dry DCM, N2and (1.5 eq), dry DCM, N2 , N r.t., 2–24 h, 55%–63%; n = 1, 2a; n = 2, 2b. H 2, r.t., 2–24 h, 55%–63%; n = 1, 2a; n = 2, 2b.O (1.5 eq), dry DCM, N COOH n H O N H n COOH N N O N COOH n
N N Reagents
N 2c-d N Scheme 3. and conditions: 5-carboxyisoquinoline (1 eq), amine (1 eq), 2c-d 1-hydroxybenzotriazole (HOBt) (1 eq), N,N-dicyclohexylcarbodiimide (DCC) (1 eq), 2c-d N-methylmorpholine (NMM) DMF, 0 °C,5-carboxyisoquinoline r.t., 24 h, 58%–70%. Scheme 3. Reagents and(1 eq), conditions: (1 eq), amine (1 eq), Scheme 3. Reagents (HOBt) and conditions: (1 eq), (DCC) amine (1 (1 eq), eq), 1-hydroxybenzotriazole (1 eq), 5-carboxyisoquinoline N,N-dicyclohexylcarbodiimide Scheme 3. Reagents and conditions: 5-carboxyisoquinoline (1 eq), amine (1 eq), 1-hydroxybenzotriazole 1-hydroxybenzotriazole (HOBt) (1 eq), N,N-dicyclohexylcarbodiimide (DCC) (1 eq), N-methylmorpholine (NMM) (1 eq), DMF, 0 °C, r.t., 24 h, 58%–70%.
(HOBt) (1 eq), N,N-dicyclohexylcarbodiimide (DCC) (1 eq), N-methylmorpholine (NMM) (1 eq), DMF, N-methylmorpholine (NMM) (1 eq), DMF, 0 °C, r.t., 24 h, 58%–70%. 0 ◦ C, r.t., 24 h, 58%–70%.
Int. J. Mol. Sci. 2016, 17, 1768
6 of 19
Int. J. Mol. Sci. 2016, 17, 1768
6 of 19
The sulfonamide 2e was easily obtained by the reaction of 5-aminoisoquinoline
Int. J. Mol. Sci. 2016, 17, 1768 6 of 19 with The sulfonamide 2e was easily obtained by the reaction of 5-aminoisoquinoline with phenylsulfonyl chloride in pyridine at room temperature according to the literature phenylsulfonyl (Scheme 4) chloride [33,34]. in pyridine at room temperature according to the literature (Scheme 4) [33,34].
The sulfonamide 2e was easily obtained by the reaction of 5-aminoisoquinoline with phenylsulfonyl chloride in pyridine at room temperature according to the literature O O (Scheme 4) [33,34]. S NH2
NH
NH2
N
NH
N
2e
O N N
S
O
Scheme 4. Reagents and conditions: 5-aminoisoquinoline (1 eq), phenylsulfonyl chloride (1.5 eq),
Scheme 4. Reagents and conditions: 5-aminoisoquinoline (1 eq), phenylsulfonyl chloride (1.5 eq), 2e pyridine, 0 °C, 3 h, 61%. pyridine, 0 ◦ C, 3 h, 61%. Scheme 4. Reagents and conditions: 5-aminoisoquinoline (1 eq), phenylsulfonyl chloride (1.5 eq),
The synthetic pyridine, 0 °C, 3 h,route 61%. to derivatives 5 and 7 (Table 3) is illustrated in Scheme 5. The 4-aminobenzylamine wastomonoprotected tert-butyloxycarbonyl in THF,inat Scheme room The synthetic route derivatives 5by and 7 (Table 3) is (Boc) illustrated 5. temperature; the intermediate 3 was then reacted with phenylsulfonyl chloride, in pyridine at °C.room The synthetic route to monoprotected derivatives 5 and (Table 3) is illustrated(Boc) in Scheme 5. 0at The The 4-aminobenzylamine was by7 tert-butyloxycarbonyl in THF, The resulting sulfonamide acidicphenylsulfonyl conditions, followed by treatment with 4-aminobenzylamine was 4monoprotected byin tert-butyloxycarbonyl (Boc) in a THF, at room temperature; the intermediate 3was wasdeprotected then reacted with chloride, in pyridine at 0 ◦ C. triphosgene and 5-aminoisoquinoline, in the presence of TEA. In this step, the intermediate temperature; the intermediate 3 was then reacted phenylsulfonyl pyridine at 0 °C. with The resulting sulfonamide 4 was deprotected in with acidic conditions, chloride, followedinby a treatment isocyanate, formed by treatment with triphosgene, was conditions, not isolated,followed but it was reacted in situ to The resulting sulfonamide 4 was deprotected in acidic by a treatment with triphosgene and 5-aminoisoquinoline, in the presence of TEA. In this step, the intermediate isocyanate, obtain the desired urea 5. triphosgene and 5-aminoisoquinoline, in the presence of TEA. In this step, the intermediate formed byThe treatment with triphosgene, was not isolated, but it was intermediate reacted in situ towith obtain the desired amide derivative 7 was obtained by starting the same isocyanate, formed by treatment with triphosgene, was notfrom isolated, but it was3,reacted in situ to urea 5. synthetic obtain theprocedure. desired urea 5.
The The amide derivative fromintermediate intermediate 3, with amide derivative7 7was wasobtained obtained by by starting starting from 3, with the the samesame H N synthetic 2 procedure. synthetic procedure. NH2 H2N
Boc
Boc
a
N H
Boc
O
a
Boc b NH2
N H
Boc
NH2
e
3 NH 2
e
3
N H
Boc b
N H
O N H O 4N
S
S
O
HN
c,d
O O
N
c,d
H
HN
N H N H
O N H O 5
N
4
N H
S
S
O
O
5
O O
N H
N H
N H 6
HN
c,d
O O
c,d
N
N H
HN
O
N H
N H
N H 7
N
O
N H
Scheme 5. Reagents and conditions: (a) 4-aminobenzylamine (1 eq), di-tert-butyl-dicarbonate (Boc2O) 6 (0.8 eq), THF, r.t., 2 h, 66%; (b) phenylsulfonyl chloride (1.27 eq), pyridine, 0 °C, r.t., 24 h, 86%; (c) 4 N HCl, dioxane, r.t., 4 h; (d) TEA (1 eq), triphosgene (0.5 eq), toluene, reflux, h; Scheme 5. Reagents and conditions: (a) (i) 4-aminobenzylamine (1 eq), di-tert-butyl-dicarbonate (Boc42O) Scheme 5. Reagents and conditions: (a)04-aminobenzylamine (1 eq), di-tert-butyl-dicarbonate (Boc2 O) (ii) 5-aminoisoquinoline (1 eq), DCM, °C, r.t., 18 h, 45%–48%; (e) benzoyl chloride (1.2 eq), pyridine, (0.8 eq), THF, r.t., 2 h, 66%; (b) phenylsulfonyl chloride (1.2 eq), pyridine, 0 °C, r.t., 24 h, 86%; (0.8 eq), THF, r.t.,h,2 72%. h, 66%; (b) phenylsulfonyl chloride (1.2 eq), pyridine, 0 ◦ C, r.t., 24 h, 86%; (c) 4 N HCl, 0(c)°C, 4 r.t., N 24 HCl, dioxane, r.t., 4 h; (d) (i) TEA (1 eq), triphosgene (0.5 eq), toluene, reflux, 4 h;
dioxane, r.t., 4 h; (d) (i) TEA (1 eq), triphosgene (0.5 eq), toluene, reflux, 4 h; (ii) 5-aminoisoquinoline (ii) 5-aminoisoquinoline (1 eq), DCM, 0 °C, r.t., 18 h, 45%–48%; (e) benzoyl chloride (1.2 eq), pyridine, ◦ C, r.t., 18 h, 45%–48%; (e) benzoyl chloride (1.2 eq), pyridine, 0 ◦ C, r.t., 24 h, 72%. 2.2. Biochemical (1 eq), 0 DCM, °C, r.t., 024 Assays h, 72%.
The synthesized compounds were submitted to MMP inhibition assays. A dose/response
2.2. Biochemical Assays was performed at 100 µM by registering the residual enzyme activity on 2.2. Biochemical Assays preliminary screening
MMP-2, MMP-8 and MMP-13 (Figure 4). Compounds showing percentageassays. of inhibition higher than The synthesized compounds were submitted to to MMP ainhibition A A dose/response The synthesized compounds were submitted MMP inhibition assays. dose/response 50% were submitted to IC 50 calculation. Activity data are reported in Table 4. preliminary screening was performed at 100 µM by registering the residual enzyme activity on on preliminary screening was performed at 100 µM by registering the residual enzyme activity MMP-2, MMP-8 and MMP-13 (Figure 4). Compounds showing a percentage of inhibition higher than MMP-2, MMP-8 and MMP-13 (Figure 4). Compounds showing a percentage of inhibition higher than 50% were submitted to IC50 calculation. Activity data are reported in Table 4. 50% were submitted to IC50 calculation. Activity data are reported in Table 4.
Int. J. Mol. Sci. 2016, 17, 1768
7 of 19
Int. J. Mol. Sci. 2016, 17, 1768
7 of 19
Figure 4. Residual enzyme activity on MMP-2, MMP-8 and MMP-13 of Compounds 1a–α, 2a–e, 5 and
Figure 4. Residual enzyme activity on MMP-2, MMP-8 and MMP-13 of Compounds 1a–α, 2a–e, 5 and 7 at 100 µM. 7 at 100 µM. Table 4. Activity data (IC50, µM) on MMP-2, MMP-8 and MMP-13.
Table 4. Activity data (IC50 , µM) on MMP-2, MMP-8 and MMP-13. ID MMP-2 MMP-8 MMP-13 1c 55 ± 1 >100 58 ± 1 MMP-13 ID MMP-2 MMP-8 1g 74 ± 3 >100 73 ± 11 1c 55 ± 1 >100 58 ± 1 1h 15 ± 4 >100 14 ± 4 1g 74 ± 3 >100 73 ± 11 1l 86 ± 2 >100 98 ± 2 1h 15 ± 4 >100 14 ± 4 5 54 ± 7 88 ± 5 55 ± 7 98 ± 2 1l 86 ± 2 >100 7 7.4±± 70.8 48.588 ± 1.6 6.6 ± 0.4 55 ± 7 5 54 ±5 7 7.4 ± 0.8 48.5 ± 1.6 6.6 ± 0.4 The 1h resulted the most active compound among the first two series of compounds (1a–α, 2a–e), while tight analogues, such as 1i, were inactive. The active compounds showed an interesting The 1h resulted the most active compound among the first two series of compounds (1a–α, selectivity with respect to MMP-8. 2a–e), while tight analogues, such as 1i, inactive. The active compounds an interesting The inhibition assays allowed to were evidencing in detail the role played by showed specific structural selectivity with respect to MMP-8. features characterizing the set of NZIs: (i) the hydroxyl group on the quinoline ring (Hit 2) does not lead inhibition to detectable improvement MMP inhibition; urea →byamide simplification or The assays allowed toofevidencing in detail(ii) theeither role played specific structural features decreasing the the set bicycle hindrance to inactive or less active compounds; presence characterizing of NZIs: (i) theled hydroxyl group on the quinoline ring (iii) (Hitthe 2) does not of lead to nitrogen on the bicyclic ring is a requirement for activity, while only definite positions in the detectable improvement of MMP inhibition; (ii) either urea → amide simplification or decreasing quinoline moiety led to active ligands. the bicycle hindrance led to inactive or less active compounds; (iii) the presence of nitrogen on the In the attempt to improve inhibition activity, we extended the molecular structure of these bicyclic ring is a requirement for activity, while only definite positions in the quinoline moiety led to inhibitors, to tentatively approach the residues on the top of the S1’ site, which are usually involved activeinligands. binding with H-bond acceptor groups (generally the sulfonamide oxygen atoms) [35]. To this aim, In the attempt to improve inhibitionwere activity, we extended the group molecular structure of these p-phenylsulfonylamide or p-phenylamide introduced on the benzyl to yield Compounds inhibitors, tentatively(Table approach thecompounds residues on the top ofactive the S1’ site, which are usually involved 5 and 7,torespectively 4). These proved to be toward all of the assayed MMPs. in binding withto H-bond acceptor 7groups (generally the sulfonamide oxygen toward atoms) MMP-2 [35]. To and this aim, Compared 1h, Compound presented an improved inhibition potency MMP-13, although with a reduced selectivity MMP-8. p-phenylsulfonylamide or p-phenylamide weretoward introduced on the benzyl group to yield Compounds 5 In order to(Table elucidate the inhibition assay outcomes the structure of assayed MMP-NZI and 7, respectively 4). These compounds proved to bebased activeontoward all of the MMPs. complexes, computational studies were carried out by using either docking or Compared to 1h, Compound 7 presented an improved inhibition potency toward MMP-2molecular and MMP-13, dynamics procedures.
although with a reduced selectivity toward MMP-8. In to elucidate the inhibition assay outcomes based on the structure of MMP-NZI complexes, 2.3.order Docking computational studies were carried out by using either docking or molecular dynamics procedures. Docking calculations of the synthesized ligands in the MMP-2, MMP-8 and MMP-13 catalytic site were 2.3. Docking performed using Glide. Receptor structures were selected through a cross-docking study (Supplementary Materials), and docking calculations were conducted at both standard precision (SP) Docking calculations the synthesized ligands in themolecule MMP-2,onMMP-8 and extra precision (XP)of settings, with and without a water the zincand ion. MMP-13 Above all, catalytic the
site were performed using Glide. Receptor structures were selected through a cross-docking study (Supplementary Materials), and docking calculations were conducted at both standard precision (SP) and extra precision (XP) settings, with and without a water molecule on the zinc ion. Above all, the
Int. J. Mol. Sci. 2016, 17, 1768
8 of 19
binding mode of the synthesized ligands is similar to the one described for the Hits 1 and 2. As an J. Mol. Sci. 2016, pose 17, 1768of 1h in the MMP-2 binding site showed that the benzyl ring 8 of 19 example,Int.the docked stacks on the His201 imidazole ring; the urea NHs form H-bonds with Leu218 CO; while the carbonyl oxygen binding mode of the synthesized ligands is similar to the one described for the Hits 1 and 2. As an Int. J. Mol. 2016, 17, 1768pose 8 ofthe 19 is connected toSci. the Thr227 NH.ofThe ring is only partially example, the docked 1h inisoquinoline the MMP-2 binding sitemainly showedsolvent-exposed that the benzyl ringand stacks on buried byHis201 hydrophobic residues in the lower part of the S1’ site (Leu218, Phe232) (Figure 5A). A imidazole ring; the urea NHs form H-bonds with Leu218 CO; while the carbonyl oxygen issimilar binding mode of the synthesized ligands is similar to the one described for the Hits 1 and 2. As an connected to the was Thr227 NH. in Theallisoquinoline ringdocking is mainlycomplexes, solvent-exposed and only partially pattern of interactions found MMP-2-NZI including both active and example, the docked pose of 1h in the MMP-2 binding site showed that the benzyl ring stacks on the buried by hydrophobic residues in the lower part of the S1’ site (Leu218, Phe232) (Figure 5A). inactiveHis201 ligands. The binding of indazole derivatives (1l and 1m) is inverted in all receptors with the imidazole ring; the urea NHs form H-bonds with Leu218 CO; while the carbonyl oxygen is A similar pattern of interactions was found in all MMP-2-NZI docking complexes, including both heterocycle NH H-bonded to Pro221 (Figure 5B). connected to the Thr227 NH. TheCO isoquinoline ring is mainly solvent-exposed and only partially active and inactive ligands. The binding of indazole derivatives (1l and 1m) is inverted in all receptors buried by hydrophobic residues in the of lower part of the S1’chain site (Leu218, In the MMP-8 active site, the presence the Arg222 hindersPhe232) the S1’ (Figure site and5A). limits the with the heterocycle NH H-bonded to Pro221 CO (Figureside 5B). A similar pattern ofisoquinoline interactions was found in alla MMP-2-NZI docking complexes, including both 5C). positioning of the bulky ring, giving reason for the observed selectivity (Figure In the MMP-8 active site, the presence of the Arg222 side chain hinders the S1’ site and limits the active and inactive ligands. The binding of indazole derivatives (1l and 1m) is inverted in all receptors of the isoquinoline ring, giving a reason for the observed selectivity (Figure 5C). On positioning the contrary, thebulky larger specificity site of MMP-13 allows easily accommodating the studied with the heterocycle NH H-bonded to Pro221 CO (Figure 5B). On the contrary, the larger specificity site of MMP-13 allows easily accommodating the studied ligands in aIn binding mode similar to that of well-known NZIs [25]: the benzyl ring forms a π–π the MMP-8 active site, the presence of the Arg222 side chain hinders the S1’ site and limits the ligands in a binding mode similar to that of well-known NZIs [25]: the benzyl ring forms a π–π stackingpositioning with His201; the urea NHs are engaged in H-bonding with Thr224 CO; while the isoquinoline of the bulky isoquinoline ring, giving a reason for the observed selectivity (Figure 5C). stacking with His201; the urea NHs are engaged in H-bonding with Thr224 CO; while the N forms isoquinoline a H-bond with thethe Met232 NH (Figure 5D). On the contrary, larger specificity site of MMP-13 allows easily accommodating the studied N forms a H-bond with the Met232 NH (Figure 5D). ligands in a binding mode similar to that of well-known NZIs [25]: the benzyl ring forms a π–π stacking with His201; the urea NHs are engaged in H-bonding with Thr224 CO; while the isoquinoline N forms a H-bond with the Met232 NH (Figure 5D).
Figure 5. Docking poses of: (A) 1h in the MMP-2; (B) 1m in the MMP-13; (C) 1h in the MMP-8; and
Figure 5.(D) Docking poses of: (A) 1h in the MMP-2; (B) 1m in the MMP-13; (C) 1h in the MMP-8; and 1h in the MMP-13. MMPs are represented as grey (MMP-2), pale cyan (MMP-8) and pale green (D) 1h in(MMP-13) the MMP-13. MMPs areion represented as greymost (MMP-2), cyan and pale green cartoons. The zinc is a purple sphere; relevantpale residues are(MMP-8) represented as lines Figure 5. Docking poses of: (A) 1h in the sphere; MMP-2; (B) 1m in the MMP-13; (C)are 1h in the MMP-8;as and (MMP-13) cartoons. zinc ion is aare purple most relevant residues represented lines and and ligands asThe sticks. H-bonds depicted as yellow dashed lines. (D) 1h in the MMP-13. MMPs are represented as grey (MMP-2), pale cyan (MMP-8) and pale green ligands as sticks. H-bonds are depicted as yellow dashed lines.
(MMP-13) cartoons. The zinc ion is a purple sphere; relevant residues are represented Additional interactions were evidenced by themost docking of larger ligands (5 and 7):asinlines MMP-2, ligands as sticks. H-bonds depicted yellowNH, dashed lines.in MMP-13, they occupy the S1’ site theseand compounds interact with are Leu164 andasAla165 while Additional interactions wereofevidenced by(Figure the docking of larger ligands (5 and 7): in MMP-2, with a banana-shape, typical MMP-13 NZIs 6). Additional interactions were evidenced by the NH, docking of larger ligands (5they and 7): in MMP-2, these compounds interact with Leu164 and Ala165 while in MMP-13, occupy the S1’ site these compounds interact with Leu164 and Ala165 NH, while in MMP-13, they occupy the S1’ site with a banana-shape, typical of MMP-13 NZIs (Figure 6). with a banana-shape, typical of MMP-13 NZIs (Figure 6).
Figure 6. Docking poses of: (A) ligand 5 in the MMP-2; (B) and ligand 7 in the MMP-13. MMPs are represented as grey (MMP-2), and pale green (MMP-13) cartoons. The zinc ion is a purple sphere; most relevant residues are represented as lines and ligands as sticks. H-bonds are depicted as yellow Figure 6.lined. Docking poses of: (A) ligand 5 in the MMP-2; (B) and ligand 7 in the MMP-13. MMPs are dashed Figure 6. Docking poses of: (A) ligand 5 in the MMP-2; (B) and ligand 7 in the MMP-13. MMPs are represented as grey (MMP-2), and pale green (MMP-13) cartoons. The zinc ion is a purple sphere; represented as grey (MMP-2), and pale green The zinc ion is aaspurple most relevant residues are represented as lines(MMP-13) and ligands cartoons. as sticks. H-bonds are depicted yellow sphere; most relevant are represented as lines and ligands as sticks. H-bonds are depicted as yellow dashed residues lined.
dashed lined.
Int. J. Mol. Sci. 2016, 17, 1768
9 of 19
Int. J. Mol. Sci. 2016, 17, 1768
9 of 19
a better comprehension thedescribed described SAR, of the theof studied To haveToa have better comprehension of of the SAR,the thesurface surface of S1’ thesite S1’ofsite the studied MMPs has been represented, showing the different shapes of the three hydrophobic pockets MMPs has been represented, showing the different shapes of the three hydrophobic pockets (Figure 7). (Figure 7).
Int. J. Mol. Sci. 2016, 17, 1768
9 of 19
To have a better comprehension of the described SAR, the surface of the S1’ site of the studied MMPs has been represented, showing the different shapes of the three hydrophobic pockets (Figure 7).
Figure 7. The molecular surface of the S1’ site of, (A) MMP-2 (pale yellow), (B) MMP-8 (pale cyan)
Figure 7. The molecular surface of the S1’ site of, (A) MMP-2 (pale yellow), (B) MMP-8 (pale cyan) and and (C) MMP-13 (pale green) is depicted. MMP-2 has a tunnel-like S1’ site, while the MMP-8 and (C) MMP-13 (pale green) is depicted. MMP-2 has a tunnel-like S1’ site, while the MMP-8 and MMP-13 MMP-13 S1’ loop moves, opening an accessory pocket called the S1’* site. However, the MMP-13 site S1’ loopismoves, opening an accessory called the S1’* site. However, the MMP-13 site is much much larger, the MMP-8 site beingpocket hindered by the Arg222 side chain. larger, the MMP-8 site being hindered by the Arg222 side chain. From a quantitative point of view, no correlation between the experimental activity and scoring Figure molecular of calculations. the S1’ site of,Moreover, (A) MMP-2the (pale yellow),potency (B) MMP-8 (pale cyan) values can 7. beThe provided by surface docking different observed for close From a quantitative of view, no correlation the activity and (C) MMP-13 point (pale green) is depicted. MMP-2 has between a tunnel-like S1’experimental site, while the MMP-8 andand scoring analogues, i.e., the 1h and 1i ligands, cannot be easily explained based on their similar binding mode. MMP-13 S1’ loopby moves, opening an accessory pocket called thethe S1’*different site. However, the MMP-13 site values can be provided docking calculations. Moreover, observed for close In the latter example, it is not clear the role played by the nitrogen atom inpotency the inhibition process, is much larger, the MMP-8 site being hindered by the Arg222 side chain. analogues, i.e., itthe 1h and 1iappreciably ligands, cannot easily explained onmodels their similar because seems to not interactbe with the enzyme in thebased docking of eitherbinding 1h or 1i mode. boundexample, complexes. Thus, molecular dynamics (MD)by simulations wereatom carried out toinhibition include theprocess, In the latter it is not clear the role played the nitrogen in the From a quantitative point of view, no correlation between the experimental activity and scoring effects of induced and the interaction with explicit water molecules into the binding model. becausevalues it seems toprovided not fit appreciably with the enzyme in thepotency docking models either 1h can be by dockinginteract calculations. Moreover, the different observed for of close In particular, we focused on MMP-2, as it is an important therapeutic target, and just a few NZIs analogues, i.e., the 1hThus, and 1i ligands, cannot be easily explained based on their similar binding mode. or 1i bound complexes. molecular dynamics (MD) simulations were carried out to include have been identified so far [28,36,37]. The binding of the above-mentioned 1h and 1i ligands was In the latter example, it is the not clear the role with playedexplicit by the nitrogen atom in the into inhibition process, model. the effects of induced fit and interaction water molecules the binding studied, being representative of the case of highly similar ligands characterized by highly different because it seems to not appreciably interact with the enzyme in the docking models of either 1h or 1i In particular, weactivities. focused on MMP-2, as it is an important therapeutic target, and just a few NZIs inhibition bound complexes. Thus, molecular dynamics (MD) simulations were carried out to include the have been identified so far [28,36,37]. The binding of the above-mentioned 1h and 1i ligands was effects of induced fit and the interaction with explicit water molecules into the binding model. 2.4. Molecular Dynamics of the case of highly similar ligands characterized by highly different studied,In being representative particular, we focused on MMP-2, as it is an important therapeutic target, and just a few NZIs The complexes of 1h and were subjected to molecular dynamics simulations as inhibition activities. have beendocking identified so far [28,36,37]. The1ibinding of the above-mentioned 1h and 1i ligands was reported in the Materials and Methods section. In spite of the high similarity in the ligand structures studied, being representative of the case of highly similar ligands characterized by highly different and starting poses, the 1h and 1i bound complexes required about 100 and 10 ns, respectively, to gain 2.4. Molecular Dynamics inhibition activities. trajectory stabilization (Figure S1). On the one hand, the starting (obtained by docking) pose of 1i was The2.4. docking complexes of 1hthus, and induced 1i were asubjected to molecular simulations as 1i reported essentially maintained softer adaptation of the dynamics MMP-2 conformation. The Molecular Dynamics and, in the Materials and observed Methodsinsection. In spite of the(last high90similarity in the ligand structures binding pose the stabilized trajectory ns) is characterized by essentially the and samestarting The docking complexes of 1h and 1i were subjected to molecular dynamics simulations as interactions with MMP-2 detected in the corresponding docking pose (Figure 8B), although the poses, the 1h and 1i bound complexes required about 100 and 10 ns, respectively, to gain trajectory reported in the Materials and Methods section. In spite of the high similarity in the ligand structures presence of explicit waters allowed improving the description of the 1i binding pose domain. stabilization (Figure S1). On the one hand, the starting (obtained by docking) of 1i was essentially and starting poses, the 1h and 1i bound complexes required about 100 and 10 ns, respectively, to gain trajectory (Figure S1). Onadaptation the one hand,of thethe starting (obtained by docking)The pose1i of binding 1i was maintained and,stabilization thus, induced a softer MMP-2 conformation. pose essentially maintainedtrajectory and, thus, (last induced a softer adaptation of the conformation. 1i observed in the stabilized 90 ns) is characterized by MMP-2 essentially the sameThe interactions binding pose observed in the stabilized trajectory (last 90 ns) is characterized by essentially the same with MMP-2 detected in the corresponding docking pose (Figure 8B), although the presence of explicit interactions with MMP-2 detected in the corresponding docking pose (Figure 8B), although the waters allowed improving the description of the 1i binding domain. presence of explicit waters allowed improving the description of the 1i binding domain.
Figure 8. Binding poses of 1h (A) and 1i (B) obtained from the last 90 ns of the MD trajectory. MMP-2 is represented as a grey cartoon and ligands as sticks. H-bonds are depicted as yellow dashed lines.
Figure 8. Binding poses of 1h (A) and 1i (B) obtained from the last 90 ns of the MD trajectory.
Figure 8. Binding poses of 1h andcartoon 1i (B) obtained from the last 90 ns of MD trajectory. MMP-2 is represented as (A) a grey and ligands as sticks. H-bonds arethe depicted as yellow MMP-2 is represented as a grey cartoon and ligands as sticks. H-bonds are depicted as yellow dashed lines. dashed lines.
Int. J. Mol. Sci. 2016, 17, 1768
10 of 19
On the other hand, we detected a marked rearrangement of the 1h conformation consisting of the trans → cis conversion of the amide bond between the ureidic C=O and the 5-aminoisoquinoline (Figure 8A). This significant conformational change is responsible for the longer simulation time required for the MMP-2 adaptation to the new ligand pose (Figure S1). To corroborate the occurrence of amide bond rearrangement in the binding of 1h at MMP-2, density functional theory (DFT) calculations were performed to investigate the thermodynamics and kinetics of this isomerization process. Calculations indicated that, as expected, the trans is more stable than the cis conformation by 0.84 kcal/mol (0.38 kcal/mol in the gas phase), corresponding to a trans/cis ratio of 96:4. Thus, although the trans conformation is predominant, the amount of cis conformer is not negligible at equilibrium. Therefore, we estimated a very low kinetic barrier of 7.51 kcal/mol (7.50 kcal/mol in the gas phase), indicating that cis/trans equilibrium is rapidly gained. Above all, calculations evidenced that the formation of a MMP-2:1h complex with the ligand in the cis conformation cannot be excluded or that the trans → cis interconversion may feature the binding pose of these ligands that would be represented by time-averaged structures having the cis and trans conformations as limiting terms. MD calculations provided for additional information on the bound structure of MMP-2 and on the occurrence of trans → cis amide rearrangement potentially affecting the binding of these inhibitors. Thus, the stable trajectories of 1h and 1i complexes were processed to extract representative bound conformations of MMP-2 by means of a clustering analysis tool. The target conformations obtained from clustering are shaped by the binding of ligands in either trans (1i) or cis (1h) conformation and by the interaction with explicit water molecules. The two plus two protein conformations gained from the clustering of the two bound complexes show that MMP-2 maintains a global similarity in the shape of the ligand binding pocket, warranting the π–π stacking of the benzyl ring at His201 and the same number of hydrogen bond contacts of the ureidic moiety for both 1h and 1i ligands, with the main rearrangement involving the Arg233 side chain that, in the two complexes with 1i, bends toward the ligand to provide a cation–π interaction. Another difference evidenced in the representative structures of bound complexes was found in the number of hydrogen bond contacts being 138 and 129 for the MMP-2:1h complex structures and 135 and 124 for the MMP-2:1i complex, highlighting that also non-local effects may contribute to differentiating the affinity of these ligands for MMP-2. A newly-directed docking study was then performed by applying two changes to the standard protocol based on the MD results: (i) MMP-2 receptor structures obtained through the clustering of both 1h and 1i MD trajectories were assembled to have a multi-conformational model of the target (ensemble docking), hence encoding a higher adaptation to the synthesized compounds, either active or inactive; (ii) ligand conformations affected by the trans → cis rearrangement of the ureidic moiety were assigned with no energy penalty, i.e., increasing the probability of harvesting docking poses in the cis conformation. Ensemble docking calculations were thus carried out by using the four-conformational model of MMP-2 and by probing all studied ligands in each receptor conformation to finally assign one pose per ligand with the optimal scoring and interaction mapping. The binding mode of 1h and 1i ligands resulting from the ensemble docking calculations are similar to the ones obtained in the first docking campaign; no trans conformation was in fact detected among the best poses of ensemble docking. The ensemble docking procedure was not able to provide evidence for the binding of the ligand in the cis conformation. On the other hand, the use of ensemble docking allowed enhancing the correlation between the activity and the score and, more specifically, to gain a higher enrichment, as indicated by the ROC curve reporting the ability of our model to provide a higher score for more active compounds (AUC = 0.91) (Figure 9).
Int. J. Mol. Sci. 2016, 17, 1768
11 of 19
Int. J. Mol. Sci. 2016, 17, 1768
11 of 19
MMP-2 ensemble docking 1
0.8
0.6
Se 0.4
0.2
0 0
0.2
0.4
1-Sp
0.6
0.8
1
Figure 9. ROC plots obtained reportingthe theranking ranking of after ensemble docking. Figure 9. ROC plots obtained reporting ofactive activecompounds compounds after ensemble docking.
3. Discussion
3. Discussion
The objective of the present study was the identification of new NZIs selective toward MMP-2
The objectivewhich of theare present study was the identification of new NZIs selectivewas toward and MMP-13, involved in cancer development. The approach we followed aimedMMP-2 to and MMP-13, areissues involved in cancer development. approach aimed to overcome which two main connected to MMP inhibition: The (i) the presencewe of afollowed chemicalwas function overcome twothe main issues connected to MMP inhibition: (i) the presence of a chemical binding binding catalytic zinc ion, which could be responsible, in several cases, forfunction some side effects emerging in clinical trials, as already discussed above; (ii) the need for isoform selective the catalytic zinc ion, which could be responsible, in several cases, for some side effects emerging in ligands, each MMP plays aabove; different possibly pathological conditions, as underlined clinical trials,because as already discussed (ii)role theinneed for isoform selective ligands, because each MMPrecently plays a[6]. different role in possibly pathological conditions, as underlined recently [6]. While MMP-13 NZIs have been widely investigated [25], there are only a few examples of While MMP-13 NZIs have been widely investigated [25], there are only a few examples of MMP-2 non-zinc-binding ligands [28,36,37]; therefore, our focus was to depict the binding process of MMP-2 non-zinc-binding ligands [28,36,37]; therefore, our focus was to depict the binding process newly-designed inhibitors, gaining more insight into the essential structural features for MMP-2 of newly-designed inhibitors, gaining more insight into the essential structural features for MMP-2 non-zinc binding inhibition, in order to drive the design of the selective and most potent ligands. non-zinc binding inhibition, in order drive the design of were the selective andthrough most potent ligands. The structural determinants ofto isoform-selective NZIs investigated the synthesis The structural determinants of isoform-selective NZIs were investigated through the synthesis and the biological evaluation of 34 newly-designed ligands tested against MMP-2, MMP-8 and and the biological evaluation of 34 newly-designed ligands tested against MMP-2, and MMP-13. MMP-13. In the previous study [28], a virtual screening campaign identified newMMP-8 inhibitors able to the MMP-2 not containing the ZBG. Docking studies suggested the role of the aryl the In theblock previous studyactivity [28], a virtual screening campaign identified new inhibitors able to block group facingnot His201, such as the the necessary CH2 spacer between this aromatic and thegroup urea. In MMP-2 activity containing ZBG. Docking studies suggested the rolegroup of the aryl facing this work, by simplifying the Hit 2 structure, we confirmed previous observations about the role of His201, such as the necessary CH2 spacer between this aromatic group and the urea. In this work, the urea function, which is that simplification did not enhance the inhibition activity. Moreover, we by simplifying the Hit 2 structure, we confirmed previous observations about the role of the urea verified that the hydroxyl group of Hit 2 is not necessary for inhibition, but the bicyclic hindrance function, which is that simplification did not enhance the inhibition activity. Moreover, we verified must be maintained, and the nitrogen atom on a good definite position of the quinoline moiety is that the hydroxyl group of Hit 2 is not necessary for inhibition, but the bicyclic hindrance must be required to ensure inhibition. maintained, and the inhibitors, nitrogen atom on ap-phenylsulfonylamide good definite positionorofp-phenylamide the quinolinesubstitution moiety is required More sized including of the to ensure inhibition. benzyl ring, i.e., Compounds 5 and 7, respectively, were also synthesized to promote interactions More sized inhibitors, or p-phenylamide substitution with residues on the top ofincluding the S1’ site,p-phenylsulfonylamide as this interaction is frequently observed in traditional MMPIs.of the Results that 5our objective has been were possibly because compounds were with benzyl ring, demonstrated i.e., Compounds and 7, respectively, alsoreached synthesized tothese promote interactions both active toward all of the assayed MMPs, and more specifically, Compound 7 resulted in being a residues on the top of the S1’ site, as this interaction is frequently observed in traditional MMPIs. more potent inhibitor of MMP-2 and MMP-13 compared to 1h, although with a reduced selectivity Results demonstrated that our objective has been possibly reached because these compounds were both MMP-8. This result can be considered as a confirmation of the binding mode of these activetoward toward all of the assayed MMPs, and more specifically, Compound 7 resulted in being a more compounds that has been suggested by docking calculations. potent inhibitor of MMP-2 and MMP-13 compared to 1h, although with a reduced selectivity toward MMP-8. This result can be considered as a confirmation of the binding mode of these compounds that has been suggested by docking calculations.
Int. J. Mol. Sci. 2016, 17, 1768
12 of 19
Computational studies were then carried out by the use of either docking or molecular dynamics procedures to elucidate the inhibition assay outcomes and shed some light on the structure-activity relationships affecting the set of investigated NZIs. Docking calculations provided a sketch of the binding pose at MMP-2, MMP-8 and MMP-13 for the synthesized ligands. All compounds, either active or inactive, shared a similar binding pose, showing the stacking of the benzyl ring on the His201 imidazole ring, the same donating and accepting H-bond contacts of urea with S1’ site residues and the solvent exposure of the bicyclic ring (see Docking subsection in the Results section). However, no explanation of the activity data was provided by the docking results that assigned a similar score both to active and inactive ligands. These results evidenced that if, on the one hand, our docking approach provided for an insight into the target-ligand binding geometries, on the other hand, it may lead to an approximate estimation of the target-ligand affinities, probably because some significant aspects of the binding process are neglected or less accurately included in the model. Indeed, although reputed important determinants of the binding process, either receptor flexibility and adaptability to the ligands (induced fit effects) or explicit water molecule interactions are usually omitted in docking calculations. Molecular dynamics (MD) simulations were then carried out to include these latter ones in the binding model of 1h and 1i at MMP-2. These calculations aimed mostly at providing an explanation for the drastic response on the MMP-2 inhibition activity of 1h and 1i caused by changing only the position of the nitrogen in the quinoline ring. From the stabilized trajectory of the MMP-2:1i complex, we detected essentially the same target-ligand interactions obtained from docking, although the presence of explicit waters may potentially influence the accommodation of the ligand. An unexpected trans → cis rearrangement of the ligand conformation, affecting the amide bond connecting the ureidic moiety to the isoquinoline ring, resulted from the MD simulation of the MMP-2:1h complex. Density functional theory (DFT) calculations indicated that although the trans asset is more stable than the cis one, the trans/cis ratio can be estimated in 96:4 and, thus, determines a non-negligible amount of cis conformer at equilibrium. These calculations evidenced that the investigated NZIs could be subjected to a trans → cis interconversion in the binding of MMP-2 or, at least, that this hypothesis cannot be ruled out. This cis, less stable, conformation, therefore, can be stabilized by H-bonding interaction with water molecules, not observed for the 1i ligand. This result suggests the role of water molecules in determining the activity of quinoline/isoquinoline derivatives depending on the position of the nitrogen atom. Further interesting data emerging from MD are the number of intramolecular H-bonds of the complex: complexes with ligand 1h have a higher number of H-bonds accounting for a stabilization of the whole system. A similar result has been observed in a previous study that applied docking and molecular dynamics to an analogue case-study on MMP-2 [38]. Furthermore, in the present study, we can confirm that the MMP-2 inhibition process cannot be studied by applying simple docking calculations. The stable trajectories of 1h and 1i complexes were clustered to extract representative bound conformations of MMP-2, and a newly-directed, ensemble docking study was then performed. The ensemble docking procedure was not able to provide evidence for the binding of ligand in cis conformation, possibly for two reasons: (i) the trans → cis interconversion may be an intrinsic feature of a ligand bound at MMP-2, so that a single representative pose would be an average of a cis/trans bundle of structures; docking poses are instead calculated as minimum structures, thus making it rather unlikely to obtain a binding pose intermediate between cis and trans; (ii) explicit water molecules, which are automatically removed by the standard receptor preparation tool (see the Materials and Methods), might be essential to prompt the ligand into a cis or trans conformation, and their exclusion from the binding model may turn out to be a too marked approximation. On the other hand, the ensemble docking allowed enhancing the correlation between activity and score, gaining a higher enrichment in the ROC curve and enabling the application of this model to study the binding of new derivatives.
Int. J. Mol. Sci. 2016, 17, 1768
13 of 19
4. Materials and Methods 4.1. Chemistry 4.1.1. General Melting points were determined with a Büchi Melting Point B-450 (Büchi Labortechnik, Flawil, Switzerland). NMR spectra were recorded on a Varian Mercury 300 spectrometer (Varian, Palo Alto, CA, USA) with 1 H at 300.060 MHz and 13 C at 75.475 MHz. Proton chemical shifts were referenced to the tetramethylsilane (TMS) internal standard. Chemical shifts are reported in parts per million (ppm, δ units). Coupling constants are reported in Hertz (Hz). Splitting patterns are designed as: s, singlet; d, doublet; t, triplet; q, quartet; ql, quartet like; dd, double doublet; m, multiplet; b, broad. All commercial chemicals and solvents are reagent grade and were used without further purification unless otherwise specified. Reactions were monitored by thin-layer chromatography on silica gel plates (60F-254, E. Merck, Merck Group, Darmstadt, Germany) and visualized with UV light, iodine vapors or alkaline KMnO4 aqueous solution. The following solvents and reagents have been abbreviated: triethylamine (TEA), tetrahydrofuran (THF), dimethyl sulfoxide (DMSO), ethyl acetate (EtOAc), dichloromethane (DCM), dimethyl formamide (DMF), methanol (MeOH), ethanol (EtOH), chloroform (CHCl3 ). All reactions were carried out with the use of standard techniques. 4.1.2. General Procedure for the Synthesis of Ureas 1a–α Benzyl isocyanate (1 eq) was added to a solution of the suitable amine or aniline (1 eq) in 10 mL of MeOH (Method a), or CHCl3 , or DCM (Method b), and the mixture was allowed to react at r.t. or refluxing temperature from 1–24 h. The solvent was removed under reduced pressure, and crude products were purified by crystallization or column chromatography. 4.1.3. General Procedure for the Synthesis of Amides 2a,b To a solution of 5-aminoisoquinoline (1 eq) and TEA (1.05 eq) in DCM (3.0 mL) at room temperature was added a solution of phenylacetyl chloride (Compound 2a, 1.05 eq) or 3-phenylpropanoyl chloride (Compound 2b, 1.05 eq), respectively. The mixture was allowed to react at r.t. for 2–24 h. After this time, the solvent was removed under reduced pressure; the crude material was treated with saturated NaHCO3 (20 mL) and extracted with DCM (3 × 20 mL). The organic phase was washed with brine (2 × 60 mL), dried on sodium sulfate and evaporated under reduced pressure. Crude products were purified by crystallization. 4.1.4. General Procedure for the Synthesis of Amides 2c,d 1-Hydroxybenzotriazole (HOBt) (1 eq) and N,N-dicyclohexylcarbodiimide (DCC) (1 eq) were added to a solution of 5-carboxyisoquinoline (1 eq) in DMF (4 mL) at 0 ◦ C. After 10 min, N-methylmorpholine (1 eq) and the proper amine (1 eq) (benzylamine or phenylethylamine) were added, and the reaction was allowed to reach room temperature. After stirring overnight, the solvent was removed under reduced pressure; the crude material was treated with saturated NaHCO3 (20 mL) and extracted with DCM (3 × 20 mL). The organic phase was washed with brine (2 × 60 mL), dried on sodium sulfate and evaporated under reduced pressure. Crude products were purified by column chromatography. 4.1.5. General Procedure for the Synthesis of Carbamates 4 and 6 To a solution of 3 (1.2 g, 5.4 mmol) in pyridine (8 mL) was carefully added phenylsulfonyl chloride or benzoyl chloride (6.5 eq) at 0 ◦ C, and the temperature was allowed to reach room temperature. After 24 h, the mixture was neutralized by 2 N HCl at 0 ◦ C, then extracted with DCM (3 × 20 mL). The organic phase was dried on sodium sulfate and the solvent evaporated under reduced pressure. Crude materials were purified by column chromatography.
Int. J. Mol. Sci. 2016, 17, 1768
14 of 19
4.1.6. General Procedure for the Synthesis of Ureas 5 and 7 A solution of the suitable carbamate (4 or 6, 1 eq) in dioxane was treated with HCl 4 N (5 mL) and stirred at room temperature for 4 h. After removal of volatiles, the deprotected hydrochloric salt was added to a solution of triphosgene (0.5 eq) in toluene (10 mL), in the presence of TEA (1 eq), and the mixture was refluxed for 4 h. After evaporation of toluene under reduced pressure, the solid obtained was dissolved in DCM (10 mL) and treated with the 5-aminoisoquinoline (1 eq) at 0 ◦ C. The reaction was stirred at room temperature overnight, then concentrated to dryness, diluted with EtOAc (15 mL) and washed with HCl 2N (3 × 15 mL) and brine (3 × 15 mL). The organic solution was dried over sodium sulfate and evaporated under reduced pressure. Crude materials were purified by crystallization. 4.2. Enzyme Inhibition Assay The catalytic domains of MMP-2, MMP-8 and MMP-13 were purchased from Enzo Life Sciences (Lörrach, Germany). The assays were performed in triplicate in 96-well white microtiter plates (Corning, NBS; Corning, NY, USA). For assay measurements, inhibitor stock solutions (DMSO, 10 mM) were diluted to six different concentrations (1 nM–100 µM) in fluorometric assay buffer (50 mM Tris·HCl pH 7.5, 200 mM NaCl, 1 mM CaCl2 , 1 µM ZnCl2 , 0.05% Brij-35 and 1% DMSO). Enzyme and inhibitor solutions were incubated in the assay buffer for 15 min at room temperature before the addition of the fluorogenic substrate solution (OmniMMP® = Mca-Pro-Leu-Gly-Leu-Dpa-Ala-Arg-NH2 , Enzo Life Sciences, 2.5 µM final concentration or OmniMMP® RED = TQ3-GABA-Pro-Cha-Abu-Smc-His-Ala-Dab(6’-TAMRA)-Ala-Lys-NH2 , (TAMRA: tetramethylrhodamine) Enzo Life Sciences, 1 µM final concentration). After further incubation for 2–4 h at 37 ◦ C, fluorescence was measured (λex = 340 nm, λem = 405 nm or λex = 545 nm, λem = 572 nm) using a Perkin-Elmer Victor V3 plate reader (Waltham, MA, USA). Control wells lacked inhibitor. The MMP inhibition activity was expressed in relative fluorescence units (RFU). Percent inhibition was calculated from control reactions without inhibitor; IC50 values were determined using GraphPad Prism version 5.0f and are expressed as the mean ± SEM of at least three independent measurements in triplicate. 4.3. Computational Methods 4.3.1. Ligand Preparation The structures of synthesized ligands were manually built exploiting the Build facility in Maestro version 10.2 [39]. The protonation state at pH 7.4 and possible tautomers were generated using the Ligand preparation routine in Maestroand applying the following parameters: force field OPLS2005, generation of the protonation state using Epik version 3.2. The obtained structures were minimized to a derivative convergence of 0.01 kJ·Å−1 ·mol−1 , using the Polak–Ribiere conjugate gradient (PRCG) minimization algorithm, the OPLS2005 force field and the eneralized Born/Surface Area (GB/SA) water solvation model implemented in MacroModel version 10.8 [39]. The conformational search was carried out on all minimized ligand structures, applying the mixed-torsional/low-mode sampling, the OPLS2005 force field, the water solvation model GB/SA and the automatic setup protocol, in order to obtain the global minimum geometry of each molecule, which was then submitted to docking calculations. 4.3.2. Docking 3D structures of MMP-2, MMP-8 and MMP-13 were retrieved from the Protein Data Bank repository [40]. Only complexes with resolution ≤2.5 Å were selected (Supplementary Material, Table S1). Each MMP structure was submitted to the Protein Preparation wizard in Maestro [39] that allows eliminating ligands and water molecules, fix bond orders, add hydrogen atoms, compute the residue protonation state, optimize the H-bond network and relax the structure with a constrained
Int. J. Mol. Sci. 2016, 17, 1768
15 of 19
minimization. All structures were superimposed to have the same frame of reference. The Glide grid generation protocol (Glide version 3.9 [39]) was applied to each structure using the same coordinates for the grid box positioning (x = 67.1815; y = 25.4097; z = 27.9457). The so-prepared structures were submitted to a preliminary cross-docking study (cross-docking details and results are reported in the Supplementary Material, Tables S2–S4) that allowed the selection of the best performing 3D structures, used for the following docking calculations. PDB IDs of selected 3D structures are: 3AYU and 1QIB for MMP-2, 1KBC and 3DPF for MMP-8; and 2OZR for MMP-13. The global minimum conformations of the synthesized ligands were used as the input geometries for docking calculations in the previously-mentioned MMP catalytic sites. The docking calculations were carried out using the Virtual Screening Workflow. Both Glide SP (standard precision) and XP (extra precision) methods were used applying the default parameters [39]. Ten poses were saved for each ligand. The same protocol (protein preparation, grid generation and docking) was applied to the MMP-2 structures obtained from MD trajectory clustering. A total of four MMP-2 structures (corresponding to the two clusters for each complex; see below) were used for the final docking runs. 4.3.3. Molecular Dynamics The docking structures of either or MMP2:1h or 1i complexes used as starting models in subsequent molecular dynamics (MD) simulations with the Gromacs package [41]. Each complex was placed in a cubic box whose dimensions prevent self-interaction and then solvated with up to 9893 water molecules, at the typical density of water (298 K, 1.0 atm) and by employing the 4-site transferable intermolecular potential function (TIP4) water model [42,43]. The molecular systems were simulated in the Amber99 force field [44]. Sodium and chloride ions were added as counterions, to ensure electrical neutrality. Simulations were initially performed by the following scheme: (i) (ii) (iii)
Local energy minimization, heating up to 300 K in six MD runs (100 ps) with protein temperature set at 0, 100, 150, 200, 250 and 300 K; Production runs of 100 ns at 300 K in an isothermal–isobaric (NPT) ensemble, by using the velocity rescaling scheme (temperature) and the isotropic Berendsen coupling scheme (pressure) [45]; The bound proteins were simulated in the Amber force field [46,47]; the linear constrain solver (LINCS) algorithm was adopted to constrain all bond lengths [48], and the long range electrostatics were computed by the particle mesh Ewald method [49].
Such a procedure provided 90 ns of stable trajectory only for the MMP-2:1i complex, whereas longer equilibration times were required by MMP-2:1h simulation, thus corresponding to a different scheme: (i) (ii) (iii)
Local energy minimization, heating up to 300 K in six runs with protein temperature set at 0 (100 ps), 100 (200 ps), 150 (200 ps), 200 (200 ps), 250 (200 ps) and 300 K (200 ps); Production runs of 200 ns at 300 K in an NPT ensemble, by using the velocity rescaling scheme (temperature) and the isotropic Berendsen coupling scheme (pressure); Same setting of MMP-2:1i simulation.
Bonded and Van der Waals parameters for the description of the Zn2+ ions’ first coordination sphere were retrieved from [50], whereas atomic charges were calculated at the density functional theory (DFT) level of theory by using Gaussian 09 [51] (See below). The Amber parameters for the description of ligand 1i structures were mostly implemented in the Gromacs parameter file, and only atomic charges were calculated quantum mechanically with Gaussian 09 (See below). Trajectory analyses were carried out by using the Gromacs utilities, VMD [52] and Maestro [39] graphical interfaces. Ensembles of 900 snapshots per trajectory (one for 100 ps) were extracted, analyzed and clusterized by using the g_cluster utility. Clustering was employed to sample representative configurations of the bound proteins to be used in subsequent docking calculations. Either MMP-2:1h
Int. J. Mol. Sci. 2016, 17, 1768
16 of 19
or MMP-2:1i trajectory ensembles were clusterized by using the method proposed by Daura et al. [53] and a cut-off of 1.5 Å by the comparison of the binding pose and the protein residues or water molecules within 3.0 Å of the ligand. The middle structures for the clusters covering at least 80% of the whole ensemble were eventually selected as representative. 4.3.4. DFT Calculations The local minimum geometries of free 1h and 1i in the gas phase were calculated at the B3LYP/6-31G* [54–56] level of theory by using Jaguar version 8.8 [39]. Atomic charges were then obtained by the fitting of the electrostatic potential by the restrained electrostatic potential (RESP) method [57] and subsequently employed in the MD simulation of the 1h and 1i bound complexes. The same computational scheme was also used for the calculation of the charges of the atoms within the first coordination sphere of the Zn2+ ions. The investigation of the kinetics and thermodynamics of the cis/trans conversion of the 1h ligand (vide infra) was performed by using the Gaussian 09 programs package [51]. The geometry of either the cis or trans conformation of 1h was optimized in the gas phase at the B3LYP/6-31+G* [54–56] level of theory. A relaxed scan of the torsion around the ureidic-aminoisoquinoline amide bond was performed at the same level of theory to obtain a guess structure of the transition state for cis/trans conversion. The transition state structure was then obtained by optimizing the geometry with the maximum energy in the scan as a first-order saddle point by using the Berny algorithm [58,59] implemented in Gaussian 09. Vibrational frequency analyses were then performed at the same level of theory to confirm the correct nature of the optimized stationary points and to calculate the zero-point energy and thermal corrections (under the hypothesis of an ideal gas behavior) for free energy estimations. The integral equation formalism within the polarizable continuum model (IEFPCM) [60]) was used to correct the gas phase free energies of trans-1h, cis-1h and the corresponding trans-to-cis transition state structure (ts-1h) by the effect of water solvation. 5. Conclusions In conclusion, with respect to previous work [28], in the present study, we obtained: (i) the synthesis of 34 new compounds with simplified structures; (ii) a better understanding of the SAR of these quinoline/isoquinoline derivatives, in particular concerning the role of the urea and quinoline ring; (iii) the validation of the hypothesized binding mode, suggesting the missing interaction with the zinc ion; (iv) a computational model based on the combination of docking and molecular dynamics procedures able to enhance the correlation between the activity and docking score in the computer-assisted design of new MMP-2 NZIs. Above all, the combination of experimental and computational outcomes provided by the present investigation shed some lights on the structure-activity and structure-selectivity affecting the investigated compounds and paves the way to future efforts in the development of newly-designed NZIs. Supplementary Materials: Supplementary materials can be found at www.mdpi.com/1422-0067/17/10/1768/s1. Acknowledgments: This study was supported by grants from the Ministry of Education, University and Research of Italy (University “Gabriele d’Annunzio”, FAR funds to Alessandra Ammazzalorso, Barbara De Filippis, Cristina Campestre, Rosa Amoroso, Alessandro Marrone and Mariangela Agamennone). Author Contributions: Alessandra Ammazzalorso, Barbara De Filippis, Cristina Campestre and Rosa Amoroso performed the chemical synthesis and contributed reagents/materials. Antonio Laghezza and Paolo Tortorella performed the biological assays. Alessandro Marrone and Mariangela Agamennone performed the computational studies. Mariangela Agamennone and Paolo Tortorella conceived of and designed the experiments. All authors contributed to writing the paper. Conflicts of Interest: The authors declare no conflict of interest.
Int. J. Mol. Sci. 2016, 17, 1768
17 of 19
References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.
13. 14. 15. 16. 17. 18. 19. 20.
21.
22.
23.
Apte, S.S.; Parks, W.C. Metalloproteinases: A parade of functions in matrix biology and an outlook for the future. Matrix Bio. 2015, 44–46, 1–6. [CrossRef] [PubMed] Burrage, P.S.; Mix, K.S.; Brinckerhoff, C.E. Matrix metalloproteinases: Role in arthritis. Front Biosci. 2006, 11, 529–543. [CrossRef] [PubMed] Lohmander, L.S.; Hoerrner, L.A.; Lark, M.W. Metalloproteinases, tissue inhibitor, and proteoglycan fragments in knee synovial fluid in human osteoarthritis. Arthritis Rheum. 1993, 36, 181–189. [CrossRef] [PubMed] Brauer, P.R. MMPs: Role in cardiovascular development and disease. Front Biosci. 2006, 11, 447–478. [CrossRef] [PubMed] Hu, J.; Van den Steen, P.E.; Sang, Q.A.; Opdenakker, G. Matrix metalloproteinase inhibitors as therapy for inflammatory and vascular diseases. Nat. Rev. Drug Discov. 2007, 6, 480–498. [CrossRef] [PubMed] Yong, V.W. Metalloproteinases: Mediators of pathology and regeneration in the CNS. Nat. Rev. Neurosci. 2005, 6, 931–944. [CrossRef] [PubMed] Peress, N.; Perillo, E.; Zucker, S. Localization of tissue inhibitor of matrix metalloproteinases in Alzheimer's disease and normal brain. J. Neuropathol. Exp. Neurol. 1995, 54, 16–22. [CrossRef] [PubMed] Chandler, S.; Coates, R.; Gearing, A.; Lury, J.; Wells, G.; Bone, E. Matrix metalloproteinases degrade myelin basic protein. Neurosci. Lett. 1995, 201, 223–226. [CrossRef] Demers, M.; Couillard, J.; Bélanger, S.; St-Pierre, Y. New roles for matrix metalloproteinases in metastasis. Crit. Rev. Immunol. 2005, 25, 493–523. [CrossRef] [PubMed] Coussens, L.M.; Fingleton, B.; Matrisian, L.M. Matrix metalloproteinase inhibitors and cancer: Trials and tribulations. Science 2002, 295, 2387–2392. [CrossRef] [PubMed] Chang, C.; Werb, Z. The many faces of metalloproteases: Cell growth, invasion, angiogenesis and metastasis. Trends Cell. Biol. 2001, 11, S37–S43. [CrossRef] Cox, J.H.; Overall, C.M. Cytokine substrates: MMP regulation of inflammatory signaling molecules. In The Cancer Degradome; Edwards, D., Hoyer-Hansen, G., Blasi, F., Sloane, B.F., Eds.; Springer: New York, NY, USA, 2008; pp. 519–539. Vandenbroucke, R.E.; Libert, C. Is there new hope for therapeutic matrix metalloproteinase inhibition? Nat. Rev. Drug Discov. 2014, 13, 904–927. [CrossRef] [PubMed] Dufour, A.; Overall, C.M. Missing the target: Matrix metalloproteinase antitargets in inflammation and cancer. Trends Pharmacol. Sci. 2013, 34, 233–242. [CrossRef] [PubMed] Overall, C. M.; Kleifeld, O. Tumour microenvironment-opinion: Validating matrix metalloproteinases as drug targets and anti-targets for cancer therapy. Nat. Rev. Cancer 2006, 6, 227–239. [CrossRef] [PubMed] Fields, G.B. New strategies for targeting matrix metalloproteinases. Matrix Biology 2015, 44–46, 239–246. [CrossRef] [PubMed] Campestre, C.; Agamennone, M.; Tauro, M.; Tortorella, P. Phosphonate emerging zinc binding group in matrix metalloproteinase inhibitors. Curr. Drug Targets 2015, 16, 1634–1644. [CrossRef] [PubMed] Agrawal, A.; Romero-Perez, D.; Jacobsen, J.A.; Villarreal, F.J.; Cohen, S.M. Zinc-binding groups modulate selective inhibition of MMPs. ChemMedChem 2008, 3, 812–820. [CrossRef] [PubMed] Tanakit, A.; Rouffet, M.; Martin, D.P.; Cohen, S.M. Investigating chelating sulfonamides and their use in metalloproteinase inhibitors. Dalton Trans. 2012, 41, 6507–6515. [CrossRef] [PubMed] Campestre, C.; Tortorella, P.; Agamennone, M.; Preziuso, S.; Biasone, A.; Nuti, E.; Rossello, A.; Gallina, C. Peptidyl 3-substituted 1-hydroxyureas as isosteric analogues of succinylhydroxamate MMP inhibitors. Eur. J. Med. Chem. 2008, 43, 1008–1014. [CrossRef] [PubMed] Rubino, M.T.; Agamennone, M.; Campestre, C.; Fracchiolla, G.; Laghezza, A.; Loiodice, F.; Nuti, E.; Rossello, A.; Tortorella, P. Synthesis, SAR, and biological evaluation of α-sulfonylphosphonic acids as selective matrix metalloproteinase inhibitors. ChemMedChem 2009, 4, 352–362. [CrossRef] [PubMed] Tauro, M.; Laghezza, A.; Loiodice, F.; Agamennone, M.; Campestre, C.; Tortorella, P. Arylamino methylene bisphosphonate derivatives as bone seeking matrix metalloproteinase inhibitors. Bioorg. Med. Chem. 2013, 21, 6456–6465. [CrossRef] [PubMed] Devy, L.; Dransfield, D.T. New strategies for the next generation of matrix-metalloproteinase inhibitors: Selectively targeting membrane-anchored MMPs with therapeutic antibodies. Biochem. Res. Int. 2011, 2011, 191670. [CrossRef] [PubMed]
Int. J. Mol. Sci. 2016, 17, 1768
24.
25. 26.
27.
28. 29.
30. 31. 32. 33.
34.
35. 36.
37.
38. 39. 40. 41. 42.
43.
18 of 19
Jacobsen, J.A.; Major Jourden, J.L.; Miller, M.T.; Cohen, M.T. To bind zinc or not to bind zinc: An examination of innovative approaches to improved metalloproteinase inhibition. Biochim. Biophys. Acta 2010, 1803, 72–94. [CrossRef] [PubMed] Di Pizio, A.; Agamennone, M.; Tortorella, P. Non-zinc-binding inhibitors of MMP-13: GRID-based approaches to rationalize the binding process. Curr. Top. Med. Chem. 2016, 16, 449–459. [CrossRef] [PubMed] Pochetti, G.; Montanari, R.; Gege, C.; Chevrier, C.; Taveras, A.G.; Mazza, F. Extra binding region induced by non-zinc chelating inhibitors into the S1 subsite of matrix metalloproteinase 8 (MMP-8). J. Med. Chem. 2009, 52, 1040–1049. [CrossRef] [PubMed] Dublanchet, A.C.; Ducrot, P.; Andrianjara, C.; O’Gara, M.; Morales, R.; Compère, D.; Denis, A.; Blais, S.; Cluzeau, P.; Courté, K.; et al. Structure-based design and synthesis of novel non-zinc chelating MMP-12 inhibitors. Bioorg. Med. Chem. Lett. 2005, 15, 3787–3790. [CrossRef] [PubMed] Di Pizio, A.; Laghezza, A.; Tortorella, P.; Agamennone, M. Probing the S10 Site for the Identification of Non-Zinc-Binding MMP-2 Inhibitor. ChemMedChem 2013, 8, 1475–1482. [CrossRef] [PubMed] Sbardella, D.; Fasciglione, G.F.; Gioia, M.; Ciaccio, C.; Tundo, G.R.; Marini, S.; Coletta, M. Human matrix metalloproteinases: An ubiquitarian class of enzymes involved in several pathological processes. Mol. Asp. Med. 2012, 33, 119–208. [CrossRef] [PubMed] Blanc, M.; Cussac, M.; Boucherle, A.; Leclerc, G. Synthesis and immunomodulating activity of 1-amino-2-thiohydantoin derivatives. Eur. J. Med. Chem. 1992, 27, 839–843. [CrossRef] Albrecht, M.; Witt, K.; Frohlich, R.; Kataeva, O. Inter- and intramolecular hydrogen bonding in amide- and urea-substituted 8-hydroxyquinoline derivatives. Tetrahedron 2002, 58, 561–567. [CrossRef] Zaitsev, V.G.; Shabashov, D.; Daugulis, O. Highly regioselective arylation of sp3 C–H bonds catalyzed by palladium acetate. J. Am. Chem. Soc. 2005, 127, 13154–13155. [CrossRef] [PubMed] Dea-Ayuela, A.M.; Castillo, E.; Gonzalez-Alvarez, M.; Vega, C.; Rolón, M.; Bolás-Fernández, F.; Borrás, J.; González-Rosende, M.E. In vivo and in vitro anti-leishmanial activities of 4-nitro-N-pyrimidinand N-pyrazin-2-ylbenzenesulfonamides, and N2-(4-nitrophenyl)-N1-propylglycinamide. Bioorg. Med. Chem. 2009, 17, 7449–7456. [CrossRef] [PubMed] Acero-Alarcón, A.; Armero-Alarte, T.; Jordá-Gregori, J.M.; Rojas-Argudo, C.; Zaballos-García, E.; Server-Carrio, J.; Ahjyaje, F.Z.; Sepúlveda-Arques, J. Unusual ring closure reaction of amides with pyrimidines: Novel stereoselective synthesis of hexahydroimidazo[1,2-c]pyrimidines. Synthesis 1999, 12, 2124–2130. [CrossRef] Jain, P.; Saravanan, C.; Singh, S.K. Sulphonamides: Deserving class as MMP inhibitors? Eur. J. Med. Chem. 2013, 60, 89–100. [CrossRef] [PubMed] Nanjan, P.; Nambiar, J.; Nair, B.G.; Banerji, A. Synthesis and discovery of (I-3,II-3)-biacacetin as a novel non-zinc binding inhibitor of MMP-2 and MMP-9. Bioorg. Med. Chem. 2015, 23, 3781–3787. [CrossRef] [PubMed] Song, J.; Peng, P.; Chang, J.; Liu, M.M.; Yu, J.M.; Zhou, L.; Sun, X. Selective non-zinc binding MMP-2 inhibitors: Novel benzamide ilomastat analogs with anti-tumor metastasis. Bioorg. Med. Chem. 2016, 26, 2174–2178. [CrossRef] [PubMed] Di Pizio, A.; Agamennone, M.; Aschi, M. An integrated computational approach to rationalize the activity of non-zinc-binding MMP-2 inhibitors. PLoS ONE 2012, 7, e47774. [CrossRef] [PubMed] Schrödinger LLC. Maestro Version 10.2, LigPrep Version 3.4, Epik, Version 3.2, MacroModel, Version 10.8, Jaguar, Version 8.8, Glide Version 6.7; Schrödinger LLC: New York, NY, USA, 2015. Berman, H.M.; Westbrook, J.; Feng, Z.; Gilliland, G.; Bhat, T.N.; Weissig, H.; Shindyalov, I.N.; Bourne, P.E. The Protein Data Bank. Nucleic Acids Res. 2000, 28, 235–242. [CrossRef] [PubMed] Van Der Spoel, D.; Lindahl, E.; Hess, B.; Groenhof, G.; Mark, A.E.; Berendsen, H.J. GROMACS: Fast, flexible, and free. J. Comput. Chem. 2005, 26, 1701–1718. [CrossRef] [PubMed] Berendsen, H.J.C.; Postma, J.P.M.; van Gunsteren, W.F.; Hermans, J. Interaction models for water in relation to protein hydration. In Intermolecular Forces; Pullmann, B., Ed.; Springer Netherlands: Dordrecht, The Netherlands, 1981; pp. 331–342. Jorgensen, W.L.; Chandrasekhar, J.; Madura, J.D.; Impey, R.W.; Klein, M.L. Comparison of simple potential functions for simulating liquid water. J. Chem. Phys. 1983, 79, 926–935. [CrossRef]
Int. J. Mol. Sci. 2016, 17, 1768
44.
45. 46.
47.
48. 49. 50. 51. 52. 53. 54. 55. 56.
57.
58. 59. 60.
19 of 19
Wang, J.; Cieplak, P.; Kollman, P.A. How well does a restrained electrostatic potential (RESP) model perform in calculating conformational energies of organic and biological molecules? J. Comp. Chem. 2000, 21, 1049–1074. [CrossRef] Berendsen, H.J.C.; Postma, J.P.M.; van Gunsteren, W.F.; di Nola, A.; Haak, J.R. Molecular dynamics with coupling to an external bath. J. Chem. Phys. 1984, 81, 3684–3690. [CrossRef] Jorgensen, W.L.; Maxwell, D.S.; Tirado-Rives, J. Development and testing of the OPLS all-atom force field on conformational energetics and properties of organic liquids. J. Am. Chem. Soc. 1996, 118, 11225–11236. [CrossRef] Kaminski, G.A.; Friesner, R.A.; Tirado-Rives, J.; Jorgensen, W.J. Evaluation and reparametrization of the OPLS-AA force field for proteins via comparison with accurate quantum chemical calculations on peptides. J. Phys. Chem. B 2001, 105, 6474–6487. [CrossRef] Hess, B.; Bekker, H.; Berendsen, H.J.C.; Frajie, J.C.E.M. LINCS: A linear constraint solver for molecular simulations. J. Comp. Chem. 1997, 18, 1463–1472. [CrossRef] Darden, T.A.; York, D.M.; Pedersen, L.G. Particle mesh ewald: An N log(N) method for ewald sums in large systems. J. Chem. Phys. 1993, 98, 10089–10092. [CrossRef] Lin, F.; Wang, R. Systematic derivation of AMBER force field parameters applicable to zinc-containing systems. J. Chem. Theory Comput. 2010, 6, 1852–1870. [CrossRef] [PubMed] Gaussian Inc. Gaussian 09, Revision D.01; Gaussian Inc.: Wallingford, CT, USA, 2013. Humphrey, W.; Dalke, A.; Schulten, K. VMD: Visual Molecular Dynamics. J. Mol. Graph. 1996, 14, 33–38. [CrossRef] Daura, X.; Gademann, K.; Jaun, B.; Seebach, D.; van Gunsteren, W.F.; Mark, A.E. Peptide folding: When simulation meets experiment. Angew. Chem. Int. Ed. 1999, 38, 236–240. [CrossRef] Becke, A.D. Density-functional thermochemistry. III. The role of exact exchange. J. Chem. Phys. 1993, 98, 5648–5452. [CrossRef] Ditchfield, R.; Hehre, W.J.; Pople, J.A. Self-consistent molecular orbital methods. IX. extended Gaussian-type basis for molecular-orbital studies of organic molecules. J. Chem. Phys. 1971, 54, 724–728. [CrossRef] Francl, M.M.; Pietro, W.J.; Hehre, W.J.; Binkley, J.S.; DeFrees, D.J.; Pople, J.A.; Gordon, M.S. Self-consistent molecular orbital methods. XXIII. A polarization-type basis set for 2nd-row elements. J. Chem. Phys. 1982, 77, 3654–3665. [CrossRef] Breneman, C.M.; Wiberg, K.B. Determining atom-centered monopoles from molecular electrostatic potentials. The need for high sampling density in formamide conformational analysis. J. Comput. Chem. 1990, 11, 361–373. [CrossRef] Li, X.; Frisch, M.J. Energy-represented DIIS within a hybrid geometry optimization method. J. Comput. Chem. 2006, 2, 835–839. Peng, C.; Ayala, P.Y.; Schlegel, H.B.; Frisch, M.J. Using redundant internal coordinates to optimize equilibrium geometries and transition states. J. Comput. Chem. 1996, 17, 49–56. [CrossRef] Tomasi, J.; Mennucci, B.; Cammi, R. Quantum mechanical continuum solvation models. Chem. Rev. 2005, 105, 2999–3093. [CrossRef] [PubMed] © 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).
Molecular Sciences Article
Seeking for Non-Zinc-Binding MMP-2 Inhibitors: Synthesis, Biological Evaluation and Molecular Modelling Studies Alessandra Ammazzalorso 1,† , Barbara De Filippis 1,† , Cristina Campestre 1 , Antonio Laghezza 2 , Alessandro Marrone 1 , Rosa Amoroso 1 , Paolo Tortorella 2 and Mariangela Agamennone 1, * 1
2
* †
Dipartimento di Farmacia, Università “G. d’Annunzio” Chieti, Via dei Vestini 31, 66100 Chieti, Italy; [email protected] (A.A.); [email protected] (B.D.F.); [email protected] (C.C.); [email protected] (A.M.); [email protected] (R.A.) Dipartimento di Farmacia-Scienze del Farmaco, Università “A. Moro” Bari, Via Orabona 4, 70125 Bari, Italy; [email protected] (A.L.); [email protected] (P.T.) Correspondence: [email protected]; Tel.: +39-0871-355-4585 These authors contributed equally to the work.
Academic Editors: Claudiu T. Supuran and Koji Sode Received: 29 July 2016; Accepted: 14 October 2016; Published: 22 October 2016
Abstract: Matrix metalloproteinases (MMPs) are an important family of zinc-containing enzymes with a central role in many physiological and pathological processes. Although several MMP inhibitors have been synthesized over the years, none reached the market because of off-target effects, due to the presence of a zinc binding group in the inhibitor structure. To overcome this problem non-zinc-binding inhibitors (NZIs) have been recently designed. In a previous article, a virtual screening campaign identified some hydroxynaphtyridine and hydroxyquinoline as MMP-2 non-zinc-binding inhibitors. In the present work, simplified analogues of previously-identified hits have been synthesized and tested in enzyme inhibition assays. Docking and molecular dynamics studies were carried out to rationalize the activity data. Keywords: MMPs; inhibitors; molecular dynamics; docking; quinoline; isoquinoline; ureas; non-zinc-binding group
1. Introduction Matrix metalloproteinases are calcium and zinc-containing proteases (24 in humans) involved in the hydrolysis of the extracellular matrix (ECM) components and, therefore, responsible for tissue remodeling in physiological conditions [1]. Unbalanced matrix metalloproteinase (MMP) activity has been observed in many pathologies, such as cancer, neurodegeneration, cardiovascular diseases and others [2–11]; for this reason, they represent a well-studied therapeutic target. More recently, it has been demonstrated that MMP involvement in several diseases can be due to the hydrolysis of substrates other than ECM components, in particular chemokines involved in inflammation and immune response [1,12]. Ever since the first ligands had been disclosed almost 30 years ago, several MMP inhibitors (MMPIs) have been discovered, but none reached the market because of relevant side effects emerging in long-term treatment (musculoskeletal syndrome (MSS)) [13]. Traditional MMPIs are constituted by a functional group binding the zinc ion (zinc binding group (ZBG)) and chemical functions mimicking the substrate and interacting with the MMP subsites, in particular with the hydrophobic S1’ site. The ZBG is represented in most of the cases by the hydroxamate function, that binds very efficiently in the MMP active site, but can possibly bind other divalent cations and cause an unselective inhibition of other metallo-enzymes.
Int. J. Mol. Sci. 2016, 17, 1768; doi:10.3390/ijms17101768
www.mdpi.com/journal/ijms
Int. J. Mol. Sci. 2016, 17, 1768
2 of 19
Another important issue of MMP inhibition is represented by the unselective inhibition of the studied MMPIs. In fact, clinically-tested compounds (e.g., batimastat and marimastat) were unselective Int. J. Mol. Sci. 2016, 17, 1768 2 of 19 MMP inhibitors not discriminating among different isoforms; while, during the years, the diverse role played byAnother each MMP has been better understood [14]. In particular, anticancer therapy, it has important issue of MMP inhibition is represented byinthe unselective inhibition of been the that studied clinically-tested compounds (e.g., batimastat and marimastat) weresome defined someMMPIs. MMPs In canfact, be considered therapeutic targets (MMP-1, MMP-2, MMP-7), while MMPasinhibitors not (MMP-8) discriminating othersunselective are classified anti-targets [15]. among different isoforms; while, during the years, the diverse role played by each MMP has been betterapproaches understood [6]. In particular, in anticancer These issues prompted the search for alternative to the MMP inhibition [16], such therapy, it has been defined that some MMPs can be considered therapeutic targets (MMP-1, as the substitution of the ZBG with less potent, more selective zinc binders [17–22], the use of MMP-2, MMP-7), while some others are classified as anti-targets (MMP-8) [7]. MMP-directed antibody [23] and the design of non-zinc binding inhibitors (NZIs) [24]. These ligands, These issues prompted the search for alternative approaches to the MMP inhibition [8], such as commonly reported as third-generation MMPIs, have been developed for MMP-13 [25], MMP-8 [26] the substitution of the ZBG with less potent, more selective zinc binders [9,18–22], the use of and MMP-12 [27] and can reach a better selectivity by binding missinginhibitors the ZBG. Recently, we attempted MMP-directed antibody [23] and the design of non-zinc (NZIs) [24]. These ligands, to identify MMP-2 NZIs characterized by enhanced interactions with the S1’ pocket [28]. Our interest commonly reported as third-generation MMPIs, have been developed for MMP-13 [25], MMP-8 [26] toward MMP-2 selective inhibition is due to the demonstrated role of MMP-2 in tumors. MMP-2, and MMP-12 [27] and can reach a better selectivity by missing the ZBG. Recently, we attempted to identifygelatinase MMP-2 NZIs characterized by enhanced interactions with the to S1’tissue pockethomeostasis. [28]. Our interest also called A, is constitutively expressed and contributes Even if toward MMP-2MMP-2 selectivehas inhibition is dueoverexpressed to the demonstrated role of tumors MMP-2 playing in tumors. poorly inducible, been found in several a MMP-2, role in both also calledand gelatinase is constitutively tumorigenesis cancerA, progression [29]. expressed and contributes to tissue homeostasis. Even if poorly inducible, MMP-2 has been found overexpressed in several tumors playing a role in both MMP-8 and MMP-13 are classified as collagenases because they mainly degrade collagen; MMP-13 tumorigenesis and cancer progression [29]. is able to degrade the components of the basal membrane and is involved in tumor dissemination and MMP-8 and MMP-13 are classified as collagenases because they mainly degrade collagen; progression. the contrary, has a confirmed protective role in tumor progression, and its MMP-13 On is able to degradeMMP-8 the components of the basal membrane and is involved in tumor inhibition should be avoided [29]. The final aim of our study is, therefore, to identify selective dissemination and progression. On the contrary, MMP-8 has a confirmed protective role in tumorNZIs blocking MMP-2 and and its MMP-13 activity, sparing theThe antitarget MMP-8. progression, inhibition shouldwhile be avoided [29]. final aim of our study is, therefore, to In the previous work,blocking we moved from analysis of the superposition betweenMMP-8. MMP-2 and identify selective NZIs MMP-2 andthe MMP-13 activity, while sparing the antitarget the previous we moved from the analysis of the superposition between MMP-2inhibitors and MMP-8 inIncomplex with work, a non-zinc-binding ligand that highlighted the possibility to design MMP-8 in complex with a non-zinc-binding ligand that highlighted the possibility to design tailored to (potential) additional interactions in the MMP-2 S1’ site. The following virtual screening inhibitors tailored (potential) additional interactions in the MMP-2 site. The following virtual campaign afforded onetohit, the hydroxynaphtyridine derivative Hit 1, S1’ binding MMP-2 at a micromolar screening campaign afforded one hit, the hydroxynaphtyridine derivative Hit 1, binding MMP-2 at concentration (IC50 = 5 µM (MMP-2), 75 µM (MMP-8) and 36 µM (MMP-13)). An analogue testing a micromolar concentration (IC50 = 5 µM (MMP-2), 75 µM (MMP-8) and 36 µM (MMP-13)). An allowed identifying also hydroxyquinoline compounds (i.e., Hit 2, IC50 = 15 µM (MMP-2), 105 µM analogue testing allowed identifying also hydroxyquinoline compounds (i.e., Hit 2, IC50 = 15 µM (MMP-8) and 5105 µM (MMP-13)), almost active thanequally the prior hitthan (Figure 1). hit (Figure 1). (MMP-2), µM (MMP-8) and 5 µMequally (MMP-13)), almost active the prior
Figure 1. Chemical structures of Hits 1 and 2.
Figure 1. Chemical structures of Hits 1 and 2.
Moving from these data, in the present work we investigated the structural determinants essential from for the inhibition non-zinc-binding ligands. Activity determinants data and docking Moving these data, in of theMMP-2 presentby work we investigated the structural essential studies of hydroxynaphtyridine/hydroxyquinoline derivatives suggested the role of relevant for the inhibition of MMP-2 by non-zinc-binding ligands. Activity data and docking studies of structural features: hydroxynaphtyridine/hydroxyquinoline derivatives suggested the role of relevant structural features:
• • •
•
the furyl/aryl group that provides the π–π stacking with His201 (MMP-2 numbering);
the furyl/aryl that provides the π–π stacking with His201 (MMP-2 numbering); • the CHgroup 2 spacer between the furyl/aryl group and the urea function; the CH between furyl/aryl andinthe • 2 spacer the urea functionthe that provides group H-bonds theurea S1’ function; site with NH of Thr227 and CO of Leu218. the urea function that provides H-bonds in the S1’ site with NH of Thr227 and CO of Leu218. Starting from these evidences, we described the synthesis and the biological evaluation of new
Starting from these evidences, we described thestructural synthesis and the biological evaluation of new MMP inhibitors, based on the combination of three portions in which the lead compound MMPcould inhibitors, based on theincombination of three structural portions in which the leadofcompound be ideally divided, order to verify their importance for the selective inhibition MMP couldisoforms. be ideally divided, in order to verify importance for the selective inhibition In all compounds, the furyl group their was replaced by a more common benzyl function. of In aMMP isoforms. In allofcompounds, was replaced by a unaltered, more common benzyl function. first series compounds, the we furyl kept group the benzyl-ureidic scaffold by modifying the hydroxyquinoline portion with heterocycles (quinoline, isoquinoline, indole, indazole), aromatic In a first series of compounds, we kept the benzyl-ureidic scaffold unaltered, by modifying the rings (naphthalene, benzene, pyridine), alkyl chains isoquinoline, or saturated heterocycles (piperidine, hydroxyquinoline portion with heterocycles (quinoline, indole, indazole), aromatic piperazine, pyrrolidine) (Figure 2; Table 1). rings (naphthalene, benzene, pyridine), alkyl chains or saturated heterocycles (piperidine, piperazine, pyrrolidine) (Figure 2; Table 1).
Int. J. Mol. Sci. 2016, 17, 1768
3 of 19
Int. J. Mol. Sci. 2016, 17, 1768
3 of 19
Int. J. Mol. Sci. 2016,Int. 17,J.1768 Mol. Sci. 2016, 17, 1768 Int. J. J. Mol. Mol. Sci. Sci. 2016, 2016, 17, 17, 1768 Int. J.J.J. Mol. Int. Mol. Sci. Sci. 2016, 2016, 17, 17, 1768 1768 Int. 1768 Int. Mol. Sci. 2016, 17, 1768 Int. J.J. Mol. Mol. Sci. Sci. 2016, 2016, 17, 17, 1768 Int. 1768 Int. Int. J. J. Mol. Mol. Sci. Sci. 2016, 2016, 17, 17, 1768 1768 Int. J. Mol. Sci. 2016, 17, 1768
3 of 19 3 of of 19 19 of 19 19 333 of
3 of 19 333 of of 19 19 of 19 33 of of 19 19 3 of 19
4 of 19
4 of 19
Figure 2. Structural modifications performed on Hit 2.
Figure 2. Structural modifications performed on Hit 2.
Figure 2. Structural modifications performed on Hitperformed 2. Figure 2. Structural modifications on Hit 2. Figure 2. 2. Structural Structural modifications performed on Hit Hit 2. 2.performed 2. Structural modifications Figure 2.derivatives Structural modifications performed on on Hit Hit 2. 2. Figure modifications performed on TableFigure 1. Urea 1a–α.on Figure 2. 2. Structural Structural modifications performed on Hit Hit 2. 2.performed Figure modifications performed Figure 2. Structural modifications on Hit 2. Figure 2. Structural modifications performed on Hit 2. Table 1. Urea derivatives 1a–α. Structural modifications performed on Hit 2. TableFigure 1. Urea2.derivatives 1a–α.derivatives Table Urea 1a–α. O 1. Table 1. 1. Urea Urea derivatives derivatives 1a–α. Table Urea Table 1a–α. Table 1. 1. Urea derivatives derivatives 1a–α. 1a–α. Table 1. 1. Urea Urea derivatives derivatives 1a–α. Table 1a–α. Table Table 1. 1.RUrea Urea derivatives derivatives 1a–α. 1a–α. O N1. Urea derivatives N Table 1a–α. O O
ID ID
1a
R R RN
ID ID1a ID ID 1a 1a 1a 1a 1b
1b
1b 1b 1b 1b 1c
1c
1c 1c 1c 1d
1d
1d 1d 1d 1d 1e
1e
1e 1e 1e 1e 1f
1f
1f 1f 1f 1g
1g
1g 1g 1g 1g 1h
1h
1h 1h 1h 1h 1i
1i
1i 1i 1i 1j
R R R
N N N N N
N N N
N N N N N
N
N N N N N
N
ID ID ID1n ID IDID ID IDID ID 1n 1a 1n 1a 1n 1a1n 1a 1a 1a 1o 1o 1b 1o 1b 1b1o 1b 1b1o 1b 1p 1p 1c 1p 1c 1c 1c1p 1c 1c 1q 1q 1d 1q 1d 1d1q 1q 1d 1d N 1d 1r
1r N 1e 1r N 1e 1r 1r N 1e N N 1e N1e 1e 1s N N 1s N N N 1f 1s 1f 1s N1f 1f 1f 1f 1t N N N N N
N N N N
N N N N N
N N H
1l N
N H
N
1t 1g 1t 1g 1g 1t N 1t 1g 1g 1g 1u N 1u 1h 1u 1h 1h1u 1u 1h 1h N 1h 1v N 1v 1i 1v 1i 1i 1v 1i 1i 1i 1w 1w 1j 1w N 1j 1w 1j1w N 1j N N 1j H N N H H 1j 1xHH H N H
ID R R R R R R R RN R N N RN N 1n N N N N N N
1o
1o 1o 1o 1o 1o 1o
1p N
1p 1p 1p 1p 1p 1p
1q
1q 1q 1q 1q 1q 1q
N
N
N N N NN N N N NN N N N N
N N NN N NN N N N N
N N N N N N
N H
HO N N N N N HON HO N HO HO HO1s
N H
R R R R R R
R R R R R R R
N N N N N N N
N N N N N N N
1s 1s 1s 1s 1s 1s3 OCH
HO HO HO HO HO HO HO
N OCH N N 3 N N 1t 3 N OCH 1t 1t OCH 3 OCH 31t N OCH 33 OCH OCH 3 3 OCH 331t OCH OCH 1t 1t3CO3333333 OCH OCH H
OCH3 OCH OCH3333 OCH 33 OCH3OCH OCH OCH3333OCH3 OCH OCH OCH33OCH3
N
3
N N OCH3 N H3CON 1u N 1u 1u H33CO CO N H 1u 3 OCH 3 H CO H 3 H333CO CO OCH331u OCH OCH333331u OCH 1u OCH N N N N N N 1v N
1v 1v 1v 1v 1v 1v
1w
1w 1w 1w 1w 1w 1w
1x
1x 1x 1x 1x 1x 1x
OCH3 H3CO H H3333CO CO H CO H H333CO CO OCH OCH OCH33333 OCH OCH H3CO OCH333 OCH3
N
N
1y 1y N
1m 1z N N O
N N N N N H N H H H H N H H
1r 1r 1r 1r 1r 1r
1r N
H
1l
ID ID ID ID ID 1n 1n 1n 1n 1n 1n
N N
1x 1k 1x N 1k 1k1x 1x N 1k N N 1k H N N H H H 1k H H N 1y N
O O O O O N N N O N N H N H H H H N H H ID
R
N
N H N N N NH N N H H H H N 1j H 1k N 1k 1k N H 1k N Int. J. Mol. Sci. 2016, 17, 1768 N N Int. J. Mol. Sci. 2016, 17, 1768 N H H H H H 1k
1m
N
N N N N N
1j 1j 1j 1j 1k
1l
O H HR O O R N N R R R NH N NH N N N N N H H ID H H H H H H
N
1z
H
O N N
N H
N N
Int. J. Mol. Sci. 2016, 17, 1768
4 of 19
Figure 2. Structural modifications performed on Hit 2.
Table 1. Cont.
Table 1. Urea derivatives 1a–α. O
Mol. Sci. 2016, 17, 1768 Int. J. Mol.Int. Sci.J. 2016, 17, 1768 Int. J. Mol. Sci. 2016, 17, 1768
1l
N
Int. J. J. Mol. Mol. Sci. Sci. 2016, 2016, 17, 17, 1768 1768 Int.
1m 1m
N
1l ID
1l
N 1m 1a 1lN N N N 1l HNN NN H H 1l N
H H N N N N N NH
1b1m 1m 1m
N
ID R 1y
H
NN 1z 1m 1z1n N N 1y N N 1y N 1y 1z H
4 of 19
4 of 19
4 of 19
R
of 19 19 44 of
N
1z
N
O 1o 1z N N N 1z H N1z
O
NH H H
1α
N
1y N IDN
N
4 of 19
R
N H N
1y RR
N H
N H
Int. J. Mol. Sci. 2016, 17, 1768 ID
N H
N N N
N
O
N
O
N
N
N
H
N
O O 1c N1p N Hstarting N A second group of derivatives was obtained from the isoquinolinyl derivative N from N the A second group of derivatives was obtained starting isoquinolinyl derivative 1h, by 1h, byderivative 1h, by N N A derivatives second group of derivatives was obtained starting from the isoquinolinyl A second group of was obtained from the isoquinolinyl derivative H starting H N replacing the ureidic linker with an amide or sulfonamide group (2a–e; Figure 3; Table 2). Finally, 1h, by N Figure 3; Table 2). Finally, replacing the ureidic linker with an amide orlinker sulfonamide group N (2a–e; Nwith an amide or sulfonamide group (2a–e; Figure 3; Table 2). Finally, replacing the ureidic replacing the ureidic linker with an amide or sulfonamide group (2a–e; Figure 3; Table 1d 1q A second group of derivatives was obtained starting from the isoquinolinyl derivative 1h,the by2). novel ureidic derivatives were designed by adding para substituents on the ring, benzyl ring, aimFinally, novel ureidic derivatives were ureidic designed by adding paradesigned substituents on the benzyl with thewith aim novel derivatives were by adding para substituents on the benzyl ring, with the aim replacing the ureidic linker with an amide or sulfonamide group (2a–e; Figure 3; Table 2). Finally, novel ureidic derivatives were designed by adding para substituents on the benzyl ring, with A second group of derivatives was obtained starting from the isoquinolinyl derivative 1h, bythe aim to explore of thefunctional functional group, potentially able the to interact with residues at the top to explore the effect ofeffect the group, able togroup, interact with residues at interact the top Athe second group ofpolar derivatives obtained starting from isoquinolinyl derivative 1h, by topolar explore the effect of was thepotentially polar functional potentially able to with residues at the top novel ureidic derivatives were designed by adding para substituents on the benzyl ring, with the aim replacing the ureidic linker with an amide or sulfonamide group (2a–e; Figure 3; Table 2). Finally, of site thereplacing S1’ site (5 and Figure 3;3). Table 3). group, to S1’ explore effect of7;the polar functional potentially able(2a–e; to interact residues at the top of the (5the and 7; Figure 3; Table the linker with an or sulfonamide group Figure 3;with Table 2). Finally, ofureidic the S1’ 7; amide Figuregroup, 3). N3; Table to explore the effect ofsite the (5 polar potentially able to interact with residues at the top 1r 1eandfunctional
novel ureidic derivatives were designed by adding adding para para substituents substituents on on the the benzyl benzyl ring, ring, with with the the aim aim novel ureidic derivatives were designed of the S1’ site (5S1’ and 3; Table 3). 3).by of the site7; (5 Figure and 7; Figure 3; Table
O O to interact O group, potentially able O to explore explore the the effect effect of of the the polar polar functional functional with residues at the top to group, potentially O able to interact with residuesOat the top HN O N HN N ofLinker the S1’ S1’Linker site (5 (5 and and 7; 7; Figure FigureHN 3; Table Table 3). NN N HN of the site 3; 3). O HN N Linker 1f
n
n
Linker
N
n
N N
2a-e
Linker Linker
2a-e
N amide N amide
N
sulfonamide
N
1g
2a-e
2a-e
inverse amide amide inverse amide inverse sulfonamide 2a-eamide sulfonamide 2a-e
N
N
n n
HN
amide inverse amide 1h sulfonamide
N O H O
N
N
NN N H 1h H
HN HN
1h
HN
1s
H
H
n
H
H
HN
N
N H
N
OH O
HN HN
N
1t
X
N N H H 5,7
OCH35,7
1h N OCH3 N
1h
N N
N
HO H
N
NN H
5,7
amide amide sulfonamide amide sulfonamide
N 1h 1h
1u
X
X NX N H H
5,7 sulfonamide 5,7
H3CO
amide amide
OCH3
H
N H
H X
X
5,7 amide sulfonamide
amide
amide 3. From isoquinolinyl urea 1h1h to novel derivatives 2a–e and 5, 7. inverse amide isoquinolinyl Figure 3.Figure From urea 1h to novel derivatives 2a–e and2a–e 5, 7.and sulfonamide inverse amide Figure 3. FromFigure isoquinolinyl urea to novel derivatives 5, 7. sulfonamide 3. From isoquinolinyl 1h to novel derivatives sulfonamide Figure 3. From isoquinolinyl urea 1h to novelurea derivatives 2a–e and 5, 7.2a–e and 5, 7. sulfonamide
N 1i 1v To provide some rational explanation activity data, structure-based computational To provide some rational explanation of rational the of activity data, structure-based To provide some explanation ofthe theto activity data, Figure 3.rational From some isoquinolinyl urea 1h novel derivatives derivatives 2a–e and andcomputational 5, 7. computational To provide explanation of thestructure-based activity data, structure-based computational Figure 3. From isoquinolinyl urea 1h to novel 2a–e 5, 7. approaches were applied to the study of the binding process. approaches were applied to the study of the binding process. approaches applied to the study ofexplanation the binding process. To were provide some rational ofstudy the of activity data,process. structure-based computational approaches were applied to the the binding
To provide some rational explanation of 1w the process. activity data, data, structure-based structure-based computational computational 1j explanation provide some rational of the activity approaches To were applied toTable the study of binding N the 2. Isoquinoline amide and sulfonamide derivatives 2a–2e. 2. Isoquinoline amide and sulfonamide derivatives Table 2.Table Isoquinoline sulfonamide derivatives 2a–2e. 2a–2e. approaches were applied to the theamide study of the the binding process. Hand approaches were applied to study of binding process. Table 2. Isoquinoline amide and sulfonamide derivatives 2a–2e.
Table 2. Isoquinoline amide and sulfonamide derivatives 2a–2e. Table 2. 2.1k Isoquinoline amide amideLinker and1x sulfonamide derivatives derivatives 2a–2e. 2a–2e. Table Isoquinoline and sulfonamide N n Linker
Linker H
Linker Linker
N
N
Linker
n
n
N
n
n
Cpd LinkernN n Linker n1 NHCO Cpd Cpd Linker n Cpd N 2a N Linker n 2b NHCO NHCO 2a 12 2a NHCO 1 2a NHCO 1 Cpd Linker n Cpd Linker n Cpd Linker 2c NHCO CONH 2b 21n 2b NHCO 2 2b NHCO 2 2a NHCO 1 2d CONH 2a NHCO 1 2a NHCO 2c 121 2c CONH CONH 1 2c CONH 1 2b NHCO 2 2b NHCO 2 2e NHSO2 2b NHCO 2d CONH 22 2d2c CONH 2 CONH 1 2c CONH 1 2d CONH 2 2c CONH 1 CONH 2 urea 5 and 7. 2e - isoquinolinyl 2e2d NHSO 2NHSO - 2 2e Table 3. Amide and sulfonamide derivatives of 2d CONH 2 NHSO 2 2d CONH 2 2e NHSO2 -
2e O 2e
NHSO22 NHSO
--
Amide and sulfonamide derivatives of isoquinolinyl urea Table 3.Table Amide3.and sulfonamide derivatives of isoquinolinyl urea 5 and 7. 5 and 7. Table 3. Amide and sulfonamide derivatives of isoquinolinyl urea 5 and 7. HN N
TableTable 3. Amide andand sulfonamide isoquinolinyl 3. Amide sulfonamideHderivatives derivatives ofofisoquinolinyl ureaurea 5 and57.and 7. Table 3. Amide and derivatives of X isoquinolinyl urea 5 and 7. O O sulfonamide HN N
NN HN H
N
Cpd 5 7
N N
N H
O N O H
HN NN X Cpd N HN H 5 HH
7
Cpd Cpd Cpd 5Cpd 7 555 7 77
N
X SO2 CO
O
HN N X N X H X2 H SO X N N HCO H
X Cpd X 2 SOX X 5SO SO222 CO SO 7CO CO CO
N H
X
X SO2 CO
Int. J. Mol. Sci. 2016, 17, 1768
5 of 19
Int. J. Mol. Sci. 2016, 17, 1768
5 of 19
2. Results
2. Int.Results J. Mol. Sci. 2016, 17, 1768 5 of 19 hits The present work moves from a previous study that identified MMP-2 non-zinc-binding Int. J. Mol. Sci. 2016, 17, 1768 of 19 (NZIs) [28]. already stated, this astudy, structural modifications were non-zinc-binding applied to these5 hits NZIs to TheAs present work movesinfrom previous study that identified MMP-2 2. Results (NZIs) [28]. As already stated, in this study, structural modifications were applied to these NZIs probe2.the structural features determining the binding at and the selective inhibition of MMP-2,toone of Results The work from atherapy. previous study that identified MMP-2 hits was probe thepresent structural features determining the binding and thefinal selective of MMP-2, one the most relevant MMPs inmoves anticancer In this at line, the goalinhibition of non-zinc-binding our research project The present work moves from a previous study that identified MMP-2 non-zinc-binding hits (NZIs) [28]. As already stated, in this study, structural modifications were applied to these NZIs totarget of the most relevant MMPs in anticancer therapy. In this line, the final goal of our research project the identification of NZIs particularly suitable for anticancer therapy, being selective toward (NZIs) [28]. As already stated, in this study, structural modifications were applied to these NZIs to probe the structural features determining the binding at and the selective inhibition of MMP-2, one was the identification of NZIs particularly suitable for anticancer therapy, being selective toward MMPs (such as MMP-2 and MMP-13) while sparing anti-target MMPs (such as MMP-8).
probe the structural the binding at and the inhibition of MMP-2, one of the MMPs most relevant MMPs indetermining anticancer therapy. In this line, theselective final goal of ourasresearch project target (such asfeatures MMP-2 and MMP-13) while sparing anti-target MMPs (such MMP-8). of the most relevant MMPs in anticancer therapy. Infor thisanticancer line, the final goal being of ourselective researchtoward project was the identification of NZIs particularly suitable therapy, 2.1. Chemistry wasChemistry the identification of NZIsand particularly for anticancer therapy, being as selective toward 2.1. target MMPs (such as MMP-2 MMP-13) suitable while sparing anti-target MMPs (such MMP-8). targetderivatives MMPs (such1a–α as MMP-2 MMP-13) while by sparing anti-target (suchto as the MMP-8). Urea (Tableand 1) were prepared adding benzyl MMPs isocyanate corresponding Urea derivatives 1a–α (Table 1) were prepared by adding benzyl isocyanate to the 2.1. Chemistry in methanol, or chloroform, or dichloromethane, stirring the reaction at room or amino derivatives corresponding amino derivatives in methanol, or chloroform, or dichloromethane, stirring the 2.1. Chemistry refluxing temperature for 1–24 [30,31].1)The synthesis is depicted in Scheme 1.isocyanate Urea derivatives 1a–αh (Table were by synthesis adding benzyl reaction at room or refluxing temperature for 1–24prepared h [30,31]. The is depicted in Schemeto1. the Urea derivatives 1a–α (Table 1) were prepared by adding benzyl isocyanate to the corresponding amino derivatives in methanol, or chloroform, or dichloromethane, stirring O corresponding amino derivatives in methanol, or chloroform, or dichloromethane, stirring reaction at room or refluxing temperature for 1–24 h [30,31]. The synthesis is depicted in Scheme 1. the R or (b)h [30,31]. The synthesis is depicted reaction at room or refluxing temperature for(a) 1–24 in Scheme 1. NCO N N H O H O R NCO N N R (a) or (b) H H NCO 1a-α N N H H Scheme 1. Reagents and conditions: benzyl isocyanate (1 equnivalent (eq)), amine, or aniline (1 eq), Scheme 1. Reagents and conditions: benzyl isocyanate (1 equnivalent (eq)), amine, or aniline (1 eq), 1a-α MeOH (Method a), or CHCl3, or dichloromethane (DCM) (Method b), room temperature (r.t.) or MeOH (Method a), or CHCl3 , or dichloromethane (DCM) (Method1a-α b), room temperature (r.t.) or refluxing, h, 41%–99%. Scheme 1.1–24 Reagents and conditions: benzyl isocyanate (1 equnivalent (eq)), amine, or aniline (1 eq), refluxing, 1–24 h, 41%–99%. Scheme(Method 1. Reagents andCHCl conditions: benzyl isocyanate (1 equnivalent (eq)), amine, or aniline(r.t.) (1 eq), 3, or dichloromethane (DCM) (Method b), room temperature or MeOH a), or The synthesis of amides (Table 2) was carried out b), byroom adding to a solution or dichloromethane (DCM) (Method temperature (r.t.) or of MeOH (Method or CHCl3, 2a,b refluxing, 1–24 h, a), 41%–99%. 5-aminoisoquinoline and triethylamine (TEA), a solution of out suitable in refluxing, 1–24of h, 41%–99%. The synthesis amides 2a,b (Table 2) was carried by phenylalkyl-chloride adding to a solution (a) or (b)
of
The synthesis of amides 2a,b (Table 2) wasatcarried out hby adding2) to solution2c,d of dichloromethane atand room temperature, while stirring r.t. for 2–24 (Scheme [32].a Amides 5-aminoisoquinoline triethylamine (TEA), a solution of suitable phenylalkyl-chloride in The synthesis of amides 2a,b (Table 2) was carried out by adding to a solution of 5-aminoisoquinoline and triethylamine (TEA), a solution of suitable phenylalkyl-chloride in were obtained by coupling 5-carboxyisoquinoline with benzylamine (for 2c) or 2-phenylethylamine dichloromethane at room temperature, while stirring at r.t. for 2–24 h (Scheme 2) [32]. Amides 2c,d 5-aminoisoquinoline andoftemperature, triethylamine (TEA), a (HOBt), solution of suitable phenylalkyl-chloride in dichloromethane at room while stirring at r.t. for 2–24 h (Scheme 2) [32]. Amides 2c,d (for 2d), in the presence 1-hydroxybenzotriazole N,N-dicyclohexylcarbodiimide (DCC) were were obtained by coupling 5-carboxyisoquinoline with benzylamine (for 2c) or 2-phenylethylamine dichloromethane at room temperature, while stirring at r.t. for 2–24 h (Scheme 2) [32]. Amides 2c,d obtained by coupling 5-carboxyisoquinoline with benzylamine (for 2c) or 2-phenylethylamine and N-methylmorpholine (NMM), in DMF at 0 °C (Scheme 3). (for 2d), in theinpresence of 1-hydroxybenzotriazole (HOBt), N,N-dicyclohexylcarbodiimide (DCC) and were obtained coupling with benzylamine (for 2c) or 2-phenylethylamine (for 2d), the by presence of 5-carboxyisoquinoline 1-hydroxybenzotriazole (HOBt), N,N-dicyclohexylcarbodiimide (DCC) ◦ C (Scheme 3). N-methylmorpholine (NMM), in DMF at 0 (for 2d), in the presence of 1-hydroxybenzotriazole (HOBt), N,N-dicyclohexylcarbodiimide (DCC) and N-methylmorpholine (NMM), in DMF at 0 °C (Scheme 3). O and N-methylmorpholine (NMM), in DMF at 0 °C (Scheme 3). NH2 NH
N
NH2 NH2
N
NH NH
O O
1-2
1-2 1-2
N N 2a-b N N Scheme 2. Reagents and conditions: 5-aminoisoquinoline (1 eq), phenylalkyl-chloride (1 eq), TEA 2a-b (1.5 eq), dry DCM, N2, r.t., 2–24 h, 55%–63%; n = 1, 2a; n = 2, 2b. 2a-b Scheme 2. Reagents and conditions: 5-aminoisoquinoline (1 eq), phenylalkyl-chloride (1 eq), TEA Scheme 2. Reagents and conditions: 5-aminoisoquinoline (1 eq), phenylalkyl-chloride (1 eq), TEA Scheme 2. Reagents 5-aminoisoquinoline , r.t., conditions: 2–24 h, 55%–63%; n = 1, 2a; n = 2, (1 2b.eq), phenylalkyl-chloride (1 eq), TEA (1.5 eq), dry DCM, N2and (1.5 eq), dry DCM, N2 , N r.t., 2–24 h, 55%–63%; n = 1, 2a; n = 2, 2b. H 2, r.t., 2–24 h, 55%–63%; n = 1, 2a; n = 2, 2b.O (1.5 eq), dry DCM, N COOH n H O N H n COOH N N O N COOH n
N N Reagents
N 2c-d N Scheme 3. and conditions: 5-carboxyisoquinoline (1 eq), amine (1 eq), 2c-d 1-hydroxybenzotriazole (HOBt) (1 eq), N,N-dicyclohexylcarbodiimide (DCC) (1 eq), 2c-d N-methylmorpholine (NMM) DMF, 0 °C,5-carboxyisoquinoline r.t., 24 h, 58%–70%. Scheme 3. Reagents and(1 eq), conditions: (1 eq), amine (1 eq), Scheme 3. Reagents (HOBt) and conditions: (1 eq), (DCC) amine (1 (1 eq), eq), 1-hydroxybenzotriazole (1 eq), 5-carboxyisoquinoline N,N-dicyclohexylcarbodiimide Scheme 3. Reagents and conditions: 5-carboxyisoquinoline (1 eq), amine (1 eq), 1-hydroxybenzotriazole 1-hydroxybenzotriazole (HOBt) (1 eq), N,N-dicyclohexylcarbodiimide (DCC) (1 eq), N-methylmorpholine (NMM) (1 eq), DMF, 0 °C, r.t., 24 h, 58%–70%.
(HOBt) (1 eq), N,N-dicyclohexylcarbodiimide (DCC) (1 eq), N-methylmorpholine (NMM) (1 eq), DMF, N-methylmorpholine (NMM) (1 eq), DMF, 0 °C, r.t., 24 h, 58%–70%. 0 ◦ C, r.t., 24 h, 58%–70%.
Int. J. Mol. Sci. 2016, 17, 1768
6 of 19
Int. J. Mol. Sci. 2016, 17, 1768
6 of 19
The sulfonamide 2e was easily obtained by the reaction of 5-aminoisoquinoline
Int. J. Mol. Sci. 2016, 17, 1768 6 of 19 with The sulfonamide 2e was easily obtained by the reaction of 5-aminoisoquinoline with phenylsulfonyl chloride in pyridine at room temperature according to the literature phenylsulfonyl (Scheme 4) chloride [33,34]. in pyridine at room temperature according to the literature (Scheme 4) [33,34].
The sulfonamide 2e was easily obtained by the reaction of 5-aminoisoquinoline with phenylsulfonyl chloride in pyridine at room temperature according to the literature O O (Scheme 4) [33,34]. S NH2
NH
NH2
N
NH
N
2e
O N N
S
O
Scheme 4. Reagents and conditions: 5-aminoisoquinoline (1 eq), phenylsulfonyl chloride (1.5 eq),
Scheme 4. Reagents and conditions: 5-aminoisoquinoline (1 eq), phenylsulfonyl chloride (1.5 eq), 2e pyridine, 0 °C, 3 h, 61%. pyridine, 0 ◦ C, 3 h, 61%. Scheme 4. Reagents and conditions: 5-aminoisoquinoline (1 eq), phenylsulfonyl chloride (1.5 eq),
The synthetic pyridine, 0 °C, 3 h,route 61%. to derivatives 5 and 7 (Table 3) is illustrated in Scheme 5. The 4-aminobenzylamine wastomonoprotected tert-butyloxycarbonyl in THF,inat Scheme room The synthetic route derivatives 5by and 7 (Table 3) is (Boc) illustrated 5. temperature; the intermediate 3 was then reacted with phenylsulfonyl chloride, in pyridine at °C.room The synthetic route to monoprotected derivatives 5 and (Table 3) is illustrated(Boc) in Scheme 5. 0at The The 4-aminobenzylamine was by7 tert-butyloxycarbonyl in THF, The resulting sulfonamide acidicphenylsulfonyl conditions, followed by treatment with 4-aminobenzylamine was 4monoprotected byin tert-butyloxycarbonyl (Boc) in a THF, at room temperature; the intermediate 3was wasdeprotected then reacted with chloride, in pyridine at 0 ◦ C. triphosgene and 5-aminoisoquinoline, in the presence of TEA. In this step, the intermediate temperature; the intermediate 3 was then reacted phenylsulfonyl pyridine at 0 °C. with The resulting sulfonamide 4 was deprotected in with acidic conditions, chloride, followedinby a treatment isocyanate, formed by treatment with triphosgene, was conditions, not isolated,followed but it was reacted in situ to The resulting sulfonamide 4 was deprotected in acidic by a treatment with triphosgene and 5-aminoisoquinoline, in the presence of TEA. In this step, the intermediate isocyanate, obtain the desired urea 5. triphosgene and 5-aminoisoquinoline, in the presence of TEA. In this step, the intermediate formed byThe treatment with triphosgene, was not isolated, but it was intermediate reacted in situ towith obtain the desired amide derivative 7 was obtained by starting the same isocyanate, formed by treatment with triphosgene, was notfrom isolated, but it was3,reacted in situ to urea 5. synthetic obtain theprocedure. desired urea 5.
The The amide derivative fromintermediate intermediate 3, with amide derivative7 7was wasobtained obtained by by starting starting from 3, with the the samesame H N synthetic 2 procedure. synthetic procedure. NH2 H2N
Boc
Boc
a
N H
Boc
O
a
Boc b NH2
N H
Boc
NH2
e
3 NH 2
e
3
N H
Boc b
N H
O N H O 4N
S
S
O
HN
c,d
O O
N
c,d
H
HN
N H N H
O N H O 5
N
4
N H
S
S
O
O
5
O O
N H
N H
N H 6
HN
c,d
O O
c,d
N
N H
HN
O
N H
N H
N H 7
N
O
N H
Scheme 5. Reagents and conditions: (a) 4-aminobenzylamine (1 eq), di-tert-butyl-dicarbonate (Boc2O) 6 (0.8 eq), THF, r.t., 2 h, 66%; (b) phenylsulfonyl chloride (1.27 eq), pyridine, 0 °C, r.t., 24 h, 86%; (c) 4 N HCl, dioxane, r.t., 4 h; (d) TEA (1 eq), triphosgene (0.5 eq), toluene, reflux, h; Scheme 5. Reagents and conditions: (a) (i) 4-aminobenzylamine (1 eq), di-tert-butyl-dicarbonate (Boc42O) Scheme 5. Reagents and conditions: (a)04-aminobenzylamine (1 eq), di-tert-butyl-dicarbonate (Boc2 O) (ii) 5-aminoisoquinoline (1 eq), DCM, °C, r.t., 18 h, 45%–48%; (e) benzoyl chloride (1.2 eq), pyridine, (0.8 eq), THF, r.t., 2 h, 66%; (b) phenylsulfonyl chloride (1.2 eq), pyridine, 0 °C, r.t., 24 h, 86%; (0.8 eq), THF, r.t.,h,2 72%. h, 66%; (b) phenylsulfonyl chloride (1.2 eq), pyridine, 0 ◦ C, r.t., 24 h, 86%; (c) 4 N HCl, 0(c)°C, 4 r.t., N 24 HCl, dioxane, r.t., 4 h; (d) (i) TEA (1 eq), triphosgene (0.5 eq), toluene, reflux, 4 h;
dioxane, r.t., 4 h; (d) (i) TEA (1 eq), triphosgene (0.5 eq), toluene, reflux, 4 h; (ii) 5-aminoisoquinoline (ii) 5-aminoisoquinoline (1 eq), DCM, 0 °C, r.t., 18 h, 45%–48%; (e) benzoyl chloride (1.2 eq), pyridine, ◦ C, r.t., 18 h, 45%–48%; (e) benzoyl chloride (1.2 eq), pyridine, 0 ◦ C, r.t., 24 h, 72%. 2.2. Biochemical (1 eq), 0 DCM, °C, r.t., 024 Assays h, 72%.
The synthesized compounds were submitted to MMP inhibition assays. A dose/response
2.2. Biochemical Assays was performed at 100 µM by registering the residual enzyme activity on 2.2. Biochemical Assays preliminary screening
MMP-2, MMP-8 and MMP-13 (Figure 4). Compounds showing percentageassays. of inhibition higher than The synthesized compounds were submitted to to MMP ainhibition A A dose/response The synthesized compounds were submitted MMP inhibition assays. dose/response 50% were submitted to IC 50 calculation. Activity data are reported in Table 4. preliminary screening was performed at 100 µM by registering the residual enzyme activity on on preliminary screening was performed at 100 µM by registering the residual enzyme activity MMP-2, MMP-8 and MMP-13 (Figure 4). Compounds showing a percentage of inhibition higher than MMP-2, MMP-8 and MMP-13 (Figure 4). Compounds showing a percentage of inhibition higher than 50% were submitted to IC50 calculation. Activity data are reported in Table 4. 50% were submitted to IC50 calculation. Activity data are reported in Table 4.
Int. J. Mol. Sci. 2016, 17, 1768
7 of 19
Int. J. Mol. Sci. 2016, 17, 1768
7 of 19
Figure 4. Residual enzyme activity on MMP-2, MMP-8 and MMP-13 of Compounds 1a–α, 2a–e, 5 and
Figure 4. Residual enzyme activity on MMP-2, MMP-8 and MMP-13 of Compounds 1a–α, 2a–e, 5 and 7 at 100 µM. 7 at 100 µM. Table 4. Activity data (IC50, µM) on MMP-2, MMP-8 and MMP-13.
Table 4. Activity data (IC50 , µM) on MMP-2, MMP-8 and MMP-13. ID MMP-2 MMP-8 MMP-13 1c 55 ± 1 >100 58 ± 1 MMP-13 ID MMP-2 MMP-8 1g 74 ± 3 >100 73 ± 11 1c 55 ± 1 >100 58 ± 1 1h 15 ± 4 >100 14 ± 4 1g 74 ± 3 >100 73 ± 11 1l 86 ± 2 >100 98 ± 2 1h 15 ± 4 >100 14 ± 4 5 54 ± 7 88 ± 5 55 ± 7 98 ± 2 1l 86 ± 2 >100 7 7.4±± 70.8 48.588 ± 1.6 6.6 ± 0.4 55 ± 7 5 54 ±5 7 7.4 ± 0.8 48.5 ± 1.6 6.6 ± 0.4 The 1h resulted the most active compound among the first two series of compounds (1a–α, 2a–e), while tight analogues, such as 1i, were inactive. The active compounds showed an interesting The 1h resulted the most active compound among the first two series of compounds (1a–α, selectivity with respect to MMP-8. 2a–e), while tight analogues, such as 1i, inactive. The active compounds an interesting The inhibition assays allowed to were evidencing in detail the role played by showed specific structural selectivity with respect to MMP-8. features characterizing the set of NZIs: (i) the hydroxyl group on the quinoline ring (Hit 2) does not lead inhibition to detectable improvement MMP inhibition; urea →byamide simplification or The assays allowed toofevidencing in detail(ii) theeither role played specific structural features decreasing the the set bicycle hindrance to inactive or less active compounds; presence characterizing of NZIs: (i) theled hydroxyl group on the quinoline ring (iii) (Hitthe 2) does not of lead to nitrogen on the bicyclic ring is a requirement for activity, while only definite positions in the detectable improvement of MMP inhibition; (ii) either urea → amide simplification or decreasing quinoline moiety led to active ligands. the bicycle hindrance led to inactive or less active compounds; (iii) the presence of nitrogen on the In the attempt to improve inhibition activity, we extended the molecular structure of these bicyclic ring is a requirement for activity, while only definite positions in the quinoline moiety led to inhibitors, to tentatively approach the residues on the top of the S1’ site, which are usually involved activeinligands. binding with H-bond acceptor groups (generally the sulfonamide oxygen atoms) [35]. To this aim, In the attempt to improve inhibitionwere activity, we extended the group molecular structure of these p-phenylsulfonylamide or p-phenylamide introduced on the benzyl to yield Compounds inhibitors, tentatively(Table approach thecompounds residues on the top ofactive the S1’ site, which are usually involved 5 and 7,torespectively 4). These proved to be toward all of the assayed MMPs. in binding withto H-bond acceptor 7groups (generally the sulfonamide oxygen toward atoms) MMP-2 [35]. To and this aim, Compared 1h, Compound presented an improved inhibition potency MMP-13, although with a reduced selectivity MMP-8. p-phenylsulfonylamide or p-phenylamide weretoward introduced on the benzyl group to yield Compounds 5 In order to(Table elucidate the inhibition assay outcomes the structure of assayed MMP-NZI and 7, respectively 4). These compounds proved to bebased activeontoward all of the MMPs. complexes, computational studies were carried out by using either docking or Compared to 1h, Compound 7 presented an improved inhibition potency toward MMP-2molecular and MMP-13, dynamics procedures.
although with a reduced selectivity toward MMP-8. In to elucidate the inhibition assay outcomes based on the structure of MMP-NZI complexes, 2.3.order Docking computational studies were carried out by using either docking or molecular dynamics procedures. Docking calculations of the synthesized ligands in the MMP-2, MMP-8 and MMP-13 catalytic site were 2.3. Docking performed using Glide. Receptor structures were selected through a cross-docking study (Supplementary Materials), and docking calculations were conducted at both standard precision (SP) Docking calculations the synthesized ligands in themolecule MMP-2,onMMP-8 and extra precision (XP)of settings, with and without a water the zincand ion. MMP-13 Above all, catalytic the
site were performed using Glide. Receptor structures were selected through a cross-docking study (Supplementary Materials), and docking calculations were conducted at both standard precision (SP) and extra precision (XP) settings, with and without a water molecule on the zinc ion. Above all, the
Int. J. Mol. Sci. 2016, 17, 1768
8 of 19
binding mode of the synthesized ligands is similar to the one described for the Hits 1 and 2. As an J. Mol. Sci. 2016, pose 17, 1768of 1h in the MMP-2 binding site showed that the benzyl ring 8 of 19 example,Int.the docked stacks on the His201 imidazole ring; the urea NHs form H-bonds with Leu218 CO; while the carbonyl oxygen binding mode of the synthesized ligands is similar to the one described for the Hits 1 and 2. As an Int. J. Mol. 2016, 17, 1768pose 8 ofthe 19 is connected toSci. the Thr227 NH.ofThe ring is only partially example, the docked 1h inisoquinoline the MMP-2 binding sitemainly showedsolvent-exposed that the benzyl ringand stacks on buried byHis201 hydrophobic residues in the lower part of the S1’ site (Leu218, Phe232) (Figure 5A). A imidazole ring; the urea NHs form H-bonds with Leu218 CO; while the carbonyl oxygen issimilar binding mode of the synthesized ligands is similar to the one described for the Hits 1 and 2. As an connected to the was Thr227 NH. in Theallisoquinoline ringdocking is mainlycomplexes, solvent-exposed and only partially pattern of interactions found MMP-2-NZI including both active and example, the docked pose of 1h in the MMP-2 binding site showed that the benzyl ring stacks on the buried by hydrophobic residues in the lower part of the S1’ site (Leu218, Phe232) (Figure 5A). inactiveHis201 ligands. The binding of indazole derivatives (1l and 1m) is inverted in all receptors with the imidazole ring; the urea NHs form H-bonds with Leu218 CO; while the carbonyl oxygen is A similar pattern of interactions was found in all MMP-2-NZI docking complexes, including both heterocycle NH H-bonded to Pro221 (Figure 5B). connected to the Thr227 NH. TheCO isoquinoline ring is mainly solvent-exposed and only partially active and inactive ligands. The binding of indazole derivatives (1l and 1m) is inverted in all receptors buried by hydrophobic residues in the of lower part of the S1’chain site (Leu218, In the MMP-8 active site, the presence the Arg222 hindersPhe232) the S1’ (Figure site and5A). limits the with the heterocycle NH H-bonded to Pro221 CO (Figureside 5B). A similar pattern ofisoquinoline interactions was found in alla MMP-2-NZI docking complexes, including both 5C). positioning of the bulky ring, giving reason for the observed selectivity (Figure In the MMP-8 active site, the presence of the Arg222 side chain hinders the S1’ site and limits the active and inactive ligands. The binding of indazole derivatives (1l and 1m) is inverted in all receptors of the isoquinoline ring, giving a reason for the observed selectivity (Figure 5C). On positioning the contrary, thebulky larger specificity site of MMP-13 allows easily accommodating the studied with the heterocycle NH H-bonded to Pro221 CO (Figure 5B). On the contrary, the larger specificity site of MMP-13 allows easily accommodating the studied ligands in aIn binding mode similar to that of well-known NZIs [25]: the benzyl ring forms a π–π the MMP-8 active site, the presence of the Arg222 side chain hinders the S1’ site and limits the ligands in a binding mode similar to that of well-known NZIs [25]: the benzyl ring forms a π–π stackingpositioning with His201; the urea NHs are engaged in H-bonding with Thr224 CO; while the isoquinoline of the bulky isoquinoline ring, giving a reason for the observed selectivity (Figure 5C). stacking with His201; the urea NHs are engaged in H-bonding with Thr224 CO; while the N forms isoquinoline a H-bond with thethe Met232 NH (Figure 5D). On the contrary, larger specificity site of MMP-13 allows easily accommodating the studied N forms a H-bond with the Met232 NH (Figure 5D). ligands in a binding mode similar to that of well-known NZIs [25]: the benzyl ring forms a π–π stacking with His201; the urea NHs are engaged in H-bonding with Thr224 CO; while the isoquinoline N forms a H-bond with the Met232 NH (Figure 5D).
Figure 5. Docking poses of: (A) 1h in the MMP-2; (B) 1m in the MMP-13; (C) 1h in the MMP-8; and
Figure 5.(D) Docking poses of: (A) 1h in the MMP-2; (B) 1m in the MMP-13; (C) 1h in the MMP-8; and 1h in the MMP-13. MMPs are represented as grey (MMP-2), pale cyan (MMP-8) and pale green (D) 1h in(MMP-13) the MMP-13. MMPs areion represented as greymost (MMP-2), cyan and pale green cartoons. The zinc is a purple sphere; relevantpale residues are(MMP-8) represented as lines Figure 5. Docking poses of: (A) 1h in the sphere; MMP-2; (B) 1m in the MMP-13; (C)are 1h in the MMP-8;as and (MMP-13) cartoons. zinc ion is aare purple most relevant residues represented lines and and ligands asThe sticks. H-bonds depicted as yellow dashed lines. (D) 1h in the MMP-13. MMPs are represented as grey (MMP-2), pale cyan (MMP-8) and pale green ligands as sticks. H-bonds are depicted as yellow dashed lines.
(MMP-13) cartoons. The zinc ion is a purple sphere; relevant residues are represented Additional interactions were evidenced by themost docking of larger ligands (5 and 7):asinlines MMP-2, ligands as sticks. H-bonds depicted yellowNH, dashed lines.in MMP-13, they occupy the S1’ site theseand compounds interact with are Leu164 andasAla165 while Additional interactions wereofevidenced by(Figure the docking of larger ligands (5 and 7): in MMP-2, with a banana-shape, typical MMP-13 NZIs 6). Additional interactions were evidenced by the NH, docking of larger ligands (5they and 7): in MMP-2, these compounds interact with Leu164 and Ala165 while in MMP-13, occupy the S1’ site these compounds interact with Leu164 and Ala165 NH, while in MMP-13, they occupy the S1’ site with a banana-shape, typical of MMP-13 NZIs (Figure 6). with a banana-shape, typical of MMP-13 NZIs (Figure 6).
Figure 6. Docking poses of: (A) ligand 5 in the MMP-2; (B) and ligand 7 in the MMP-13. MMPs are represented as grey (MMP-2), and pale green (MMP-13) cartoons. The zinc ion is a purple sphere; most relevant residues are represented as lines and ligands as sticks. H-bonds are depicted as yellow Figure 6.lined. Docking poses of: (A) ligand 5 in the MMP-2; (B) and ligand 7 in the MMP-13. MMPs are dashed Figure 6. Docking poses of: (A) ligand 5 in the MMP-2; (B) and ligand 7 in the MMP-13. MMPs are represented as grey (MMP-2), and pale green (MMP-13) cartoons. The zinc ion is a purple sphere; represented as grey (MMP-2), and pale green The zinc ion is aaspurple most relevant residues are represented as lines(MMP-13) and ligands cartoons. as sticks. H-bonds are depicted yellow sphere; most relevant are represented as lines and ligands as sticks. H-bonds are depicted as yellow dashed residues lined.
dashed lined.
Int. J. Mol. Sci. 2016, 17, 1768
9 of 19
Int. J. Mol. Sci. 2016, 17, 1768
9 of 19
a better comprehension thedescribed described SAR, of the theof studied To haveToa have better comprehension of of the SAR,the thesurface surface of S1’ thesite S1’ofsite the studied MMPs has been represented, showing the different shapes of the three hydrophobic pockets MMPs has been represented, showing the different shapes of the three hydrophobic pockets (Figure 7). (Figure 7).
Int. J. Mol. Sci. 2016, 17, 1768
9 of 19
To have a better comprehension of the described SAR, the surface of the S1’ site of the studied MMPs has been represented, showing the different shapes of the three hydrophobic pockets (Figure 7).
Figure 7. The molecular surface of the S1’ site of, (A) MMP-2 (pale yellow), (B) MMP-8 (pale cyan)
Figure 7. The molecular surface of the S1’ site of, (A) MMP-2 (pale yellow), (B) MMP-8 (pale cyan) and and (C) MMP-13 (pale green) is depicted. MMP-2 has a tunnel-like S1’ site, while the MMP-8 and (C) MMP-13 (pale green) is depicted. MMP-2 has a tunnel-like S1’ site, while the MMP-8 and MMP-13 MMP-13 S1’ loop moves, opening an accessory pocket called the S1’* site. However, the MMP-13 site S1’ loopismoves, opening an accessory called the S1’* site. However, the MMP-13 site is much much larger, the MMP-8 site beingpocket hindered by the Arg222 side chain. larger, the MMP-8 site being hindered by the Arg222 side chain. From a quantitative point of view, no correlation between the experimental activity and scoring Figure molecular of calculations. the S1’ site of,Moreover, (A) MMP-2the (pale yellow),potency (B) MMP-8 (pale cyan) values can 7. beThe provided by surface docking different observed for close From a quantitative of view, no correlation the activity and (C) MMP-13 point (pale green) is depicted. MMP-2 has between a tunnel-like S1’experimental site, while the MMP-8 andand scoring analogues, i.e., the 1h and 1i ligands, cannot be easily explained based on their similar binding mode. MMP-13 S1’ loopby moves, opening an accessory pocket called thethe S1’*different site. However, the MMP-13 site values can be provided docking calculations. Moreover, observed for close In the latter example, it is not clear the role played by the nitrogen atom inpotency the inhibition process, is much larger, the MMP-8 site being hindered by the Arg222 side chain. analogues, i.e., itthe 1h and 1iappreciably ligands, cannot easily explained onmodels their similar because seems to not interactbe with the enzyme in thebased docking of eitherbinding 1h or 1i mode. boundexample, complexes. Thus, molecular dynamics (MD)by simulations wereatom carried out toinhibition include theprocess, In the latter it is not clear the role played the nitrogen in the From a quantitative point of view, no correlation between the experimental activity and scoring effects of induced and the interaction with explicit water molecules into the binding model. becausevalues it seems toprovided not fit appreciably with the enzyme in thepotency docking models either 1h can be by dockinginteract calculations. Moreover, the different observed for of close In particular, we focused on MMP-2, as it is an important therapeutic target, and just a few NZIs analogues, i.e., the 1hThus, and 1i ligands, cannot be easily explained based on their similar binding mode. or 1i bound complexes. molecular dynamics (MD) simulations were carried out to include have been identified so far [28,36,37]. The binding of the above-mentioned 1h and 1i ligands was In the latter example, it is the not clear the role with playedexplicit by the nitrogen atom in the into inhibition process, model. the effects of induced fit and interaction water molecules the binding studied, being representative of the case of highly similar ligands characterized by highly different because it seems to not appreciably interact with the enzyme in the docking models of either 1h or 1i In particular, weactivities. focused on MMP-2, as it is an important therapeutic target, and just a few NZIs inhibition bound complexes. Thus, molecular dynamics (MD) simulations were carried out to include the have been identified so far [28,36,37]. The binding of the above-mentioned 1h and 1i ligands was effects of induced fit and the interaction with explicit water molecules into the binding model. 2.4. Molecular Dynamics of the case of highly similar ligands characterized by highly different studied,In being representative particular, we focused on MMP-2, as it is an important therapeutic target, and just a few NZIs The complexes of 1h and were subjected to molecular dynamics simulations as inhibition activities. have beendocking identified so far [28,36,37]. The1ibinding of the above-mentioned 1h and 1i ligands was reported in the Materials and Methods section. In spite of the high similarity in the ligand structures studied, being representative of the case of highly similar ligands characterized by highly different and starting poses, the 1h and 1i bound complexes required about 100 and 10 ns, respectively, to gain 2.4. Molecular Dynamics inhibition activities. trajectory stabilization (Figure S1). On the one hand, the starting (obtained by docking) pose of 1i was The2.4. docking complexes of 1hthus, and induced 1i were asubjected to molecular simulations as 1i reported essentially maintained softer adaptation of the dynamics MMP-2 conformation. The Molecular Dynamics and, in the Materials and observed Methodsinsection. In spite of the(last high90similarity in the ligand structures binding pose the stabilized trajectory ns) is characterized by essentially the and samestarting The docking complexes of 1h and 1i were subjected to molecular dynamics simulations as interactions with MMP-2 detected in the corresponding docking pose (Figure 8B), although the poses, the 1h and 1i bound complexes required about 100 and 10 ns, respectively, to gain trajectory reported in the Materials and Methods section. In spite of the high similarity in the ligand structures presence of explicit waters allowed improving the description of the 1i binding pose domain. stabilization (Figure S1). On the one hand, the starting (obtained by docking) of 1i was essentially and starting poses, the 1h and 1i bound complexes required about 100 and 10 ns, respectively, to gain trajectory (Figure S1). Onadaptation the one hand,of thethe starting (obtained by docking)The pose1i of binding 1i was maintained and,stabilization thus, induced a softer MMP-2 conformation. pose essentially maintainedtrajectory and, thus, (last induced a softer adaptation of the conformation. 1i observed in the stabilized 90 ns) is characterized by MMP-2 essentially the sameThe interactions binding pose observed in the stabilized trajectory (last 90 ns) is characterized by essentially the same with MMP-2 detected in the corresponding docking pose (Figure 8B), although the presence of explicit interactions with MMP-2 detected in the corresponding docking pose (Figure 8B), although the waters allowed improving the description of the 1i binding domain. presence of explicit waters allowed improving the description of the 1i binding domain.
Figure 8. Binding poses of 1h (A) and 1i (B) obtained from the last 90 ns of the MD trajectory. MMP-2 is represented as a grey cartoon and ligands as sticks. H-bonds are depicted as yellow dashed lines.
Figure 8. Binding poses of 1h (A) and 1i (B) obtained from the last 90 ns of the MD trajectory.
Figure 8. Binding poses of 1h andcartoon 1i (B) obtained from the last 90 ns of MD trajectory. MMP-2 is represented as (A) a grey and ligands as sticks. H-bonds arethe depicted as yellow MMP-2 is represented as a grey cartoon and ligands as sticks. H-bonds are depicted as yellow dashed lines. dashed lines.
Int. J. Mol. Sci. 2016, 17, 1768
10 of 19
On the other hand, we detected a marked rearrangement of the 1h conformation consisting of the trans → cis conversion of the amide bond between the ureidic C=O and the 5-aminoisoquinoline (Figure 8A). This significant conformational change is responsible for the longer simulation time required for the MMP-2 adaptation to the new ligand pose (Figure S1). To corroborate the occurrence of amide bond rearrangement in the binding of 1h at MMP-2, density functional theory (DFT) calculations were performed to investigate the thermodynamics and kinetics of this isomerization process. Calculations indicated that, as expected, the trans is more stable than the cis conformation by 0.84 kcal/mol (0.38 kcal/mol in the gas phase), corresponding to a trans/cis ratio of 96:4. Thus, although the trans conformation is predominant, the amount of cis conformer is not negligible at equilibrium. Therefore, we estimated a very low kinetic barrier of 7.51 kcal/mol (7.50 kcal/mol in the gas phase), indicating that cis/trans equilibrium is rapidly gained. Above all, calculations evidenced that the formation of a MMP-2:1h complex with the ligand in the cis conformation cannot be excluded or that the trans → cis interconversion may feature the binding pose of these ligands that would be represented by time-averaged structures having the cis and trans conformations as limiting terms. MD calculations provided for additional information on the bound structure of MMP-2 and on the occurrence of trans → cis amide rearrangement potentially affecting the binding of these inhibitors. Thus, the stable trajectories of 1h and 1i complexes were processed to extract representative bound conformations of MMP-2 by means of a clustering analysis tool. The target conformations obtained from clustering are shaped by the binding of ligands in either trans (1i) or cis (1h) conformation and by the interaction with explicit water molecules. The two plus two protein conformations gained from the clustering of the two bound complexes show that MMP-2 maintains a global similarity in the shape of the ligand binding pocket, warranting the π–π stacking of the benzyl ring at His201 and the same number of hydrogen bond contacts of the ureidic moiety for both 1h and 1i ligands, with the main rearrangement involving the Arg233 side chain that, in the two complexes with 1i, bends toward the ligand to provide a cation–π interaction. Another difference evidenced in the representative structures of bound complexes was found in the number of hydrogen bond contacts being 138 and 129 for the MMP-2:1h complex structures and 135 and 124 for the MMP-2:1i complex, highlighting that also non-local effects may contribute to differentiating the affinity of these ligands for MMP-2. A newly-directed docking study was then performed by applying two changes to the standard protocol based on the MD results: (i) MMP-2 receptor structures obtained through the clustering of both 1h and 1i MD trajectories were assembled to have a multi-conformational model of the target (ensemble docking), hence encoding a higher adaptation to the synthesized compounds, either active or inactive; (ii) ligand conformations affected by the trans → cis rearrangement of the ureidic moiety were assigned with no energy penalty, i.e., increasing the probability of harvesting docking poses in the cis conformation. Ensemble docking calculations were thus carried out by using the four-conformational model of MMP-2 and by probing all studied ligands in each receptor conformation to finally assign one pose per ligand with the optimal scoring and interaction mapping. The binding mode of 1h and 1i ligands resulting from the ensemble docking calculations are similar to the ones obtained in the first docking campaign; no trans conformation was in fact detected among the best poses of ensemble docking. The ensemble docking procedure was not able to provide evidence for the binding of the ligand in the cis conformation. On the other hand, the use of ensemble docking allowed enhancing the correlation between the activity and the score and, more specifically, to gain a higher enrichment, as indicated by the ROC curve reporting the ability of our model to provide a higher score for more active compounds (AUC = 0.91) (Figure 9).
Int. J. Mol. Sci. 2016, 17, 1768
11 of 19
Int. J. Mol. Sci. 2016, 17, 1768
11 of 19
MMP-2 ensemble docking 1
0.8
0.6
Se 0.4
0.2
0 0
0.2
0.4
1-Sp
0.6
0.8
1
Figure 9. ROC plots obtained reportingthe theranking ranking of after ensemble docking. Figure 9. ROC plots obtained reporting ofactive activecompounds compounds after ensemble docking.
3. Discussion
3. Discussion
The objective of the present study was the identification of new NZIs selective toward MMP-2
The objectivewhich of theare present study was the identification of new NZIs selectivewas toward and MMP-13, involved in cancer development. The approach we followed aimedMMP-2 to and MMP-13, areissues involved in cancer development. approach aimed to overcome which two main connected to MMP inhibition: The (i) the presencewe of afollowed chemicalwas function overcome twothe main issues connected to MMP inhibition: (i) the presence of a chemical binding binding catalytic zinc ion, which could be responsible, in several cases, forfunction some side effects emerging in clinical trials, as already discussed above; (ii) the need for isoform selective the catalytic zinc ion, which could be responsible, in several cases, for some side effects emerging in ligands, each MMP plays aabove; different possibly pathological conditions, as underlined clinical trials,because as already discussed (ii)role theinneed for isoform selective ligands, because each MMPrecently plays a[6]. different role in possibly pathological conditions, as underlined recently [6]. While MMP-13 NZIs have been widely investigated [25], there are only a few examples of While MMP-13 NZIs have been widely investigated [25], there are only a few examples of MMP-2 non-zinc-binding ligands [28,36,37]; therefore, our focus was to depict the binding process of MMP-2 non-zinc-binding ligands [28,36,37]; therefore, our focus was to depict the binding process newly-designed inhibitors, gaining more insight into the essential structural features for MMP-2 of newly-designed inhibitors, gaining more insight into the essential structural features for MMP-2 non-zinc binding inhibition, in order to drive the design of the selective and most potent ligands. non-zinc binding inhibition, in order drive the design of were the selective andthrough most potent ligands. The structural determinants ofto isoform-selective NZIs investigated the synthesis The structural determinants of isoform-selective NZIs were investigated through the synthesis and the biological evaluation of 34 newly-designed ligands tested against MMP-2, MMP-8 and and the biological evaluation of 34 newly-designed ligands tested against MMP-2, and MMP-13. MMP-13. In the previous study [28], a virtual screening campaign identified newMMP-8 inhibitors able to the MMP-2 not containing the ZBG. Docking studies suggested the role of the aryl the In theblock previous studyactivity [28], a virtual screening campaign identified new inhibitors able to block group facingnot His201, such as the the necessary CH2 spacer between this aromatic and thegroup urea. In MMP-2 activity containing ZBG. Docking studies suggested the rolegroup of the aryl facing this work, by simplifying the Hit 2 structure, we confirmed previous observations about the role of His201, such as the necessary CH2 spacer between this aromatic group and the urea. In this work, the urea function, which is that simplification did not enhance the inhibition activity. Moreover, we by simplifying the Hit 2 structure, we confirmed previous observations about the role of the urea verified that the hydroxyl group of Hit 2 is not necessary for inhibition, but the bicyclic hindrance function, which is that simplification did not enhance the inhibition activity. Moreover, we verified must be maintained, and the nitrogen atom on a good definite position of the quinoline moiety is that the hydroxyl group of Hit 2 is not necessary for inhibition, but the bicyclic hindrance must be required to ensure inhibition. maintained, and the inhibitors, nitrogen atom on ap-phenylsulfonylamide good definite positionorofp-phenylamide the quinolinesubstitution moiety is required More sized including of the to ensure inhibition. benzyl ring, i.e., Compounds 5 and 7, respectively, were also synthesized to promote interactions More sized inhibitors, or p-phenylamide substitution with residues on the top ofincluding the S1’ site,p-phenylsulfonylamide as this interaction is frequently observed in traditional MMPIs.of the Results that 5our objective has been were possibly because compounds were with benzyl ring, demonstrated i.e., Compounds and 7, respectively, alsoreached synthesized tothese promote interactions both active toward all of the assayed MMPs, and more specifically, Compound 7 resulted in being a residues on the top of the S1’ site, as this interaction is frequently observed in traditional MMPIs. more potent inhibitor of MMP-2 and MMP-13 compared to 1h, although with a reduced selectivity Results demonstrated that our objective has been possibly reached because these compounds were both MMP-8. This result can be considered as a confirmation of the binding mode of these activetoward toward all of the assayed MMPs, and more specifically, Compound 7 resulted in being a more compounds that has been suggested by docking calculations. potent inhibitor of MMP-2 and MMP-13 compared to 1h, although with a reduced selectivity toward MMP-8. This result can be considered as a confirmation of the binding mode of these compounds that has been suggested by docking calculations.
Int. J. Mol. Sci. 2016, 17, 1768
12 of 19
Computational studies were then carried out by the use of either docking or molecular dynamics procedures to elucidate the inhibition assay outcomes and shed some light on the structure-activity relationships affecting the set of investigated NZIs. Docking calculations provided a sketch of the binding pose at MMP-2, MMP-8 and MMP-13 for the synthesized ligands. All compounds, either active or inactive, shared a similar binding pose, showing the stacking of the benzyl ring on the His201 imidazole ring, the same donating and accepting H-bond contacts of urea with S1’ site residues and the solvent exposure of the bicyclic ring (see Docking subsection in the Results section). However, no explanation of the activity data was provided by the docking results that assigned a similar score both to active and inactive ligands. These results evidenced that if, on the one hand, our docking approach provided for an insight into the target-ligand binding geometries, on the other hand, it may lead to an approximate estimation of the target-ligand affinities, probably because some significant aspects of the binding process are neglected or less accurately included in the model. Indeed, although reputed important determinants of the binding process, either receptor flexibility and adaptability to the ligands (induced fit effects) or explicit water molecule interactions are usually omitted in docking calculations. Molecular dynamics (MD) simulations were then carried out to include these latter ones in the binding model of 1h and 1i at MMP-2. These calculations aimed mostly at providing an explanation for the drastic response on the MMP-2 inhibition activity of 1h and 1i caused by changing only the position of the nitrogen in the quinoline ring. From the stabilized trajectory of the MMP-2:1i complex, we detected essentially the same target-ligand interactions obtained from docking, although the presence of explicit waters may potentially influence the accommodation of the ligand. An unexpected trans → cis rearrangement of the ligand conformation, affecting the amide bond connecting the ureidic moiety to the isoquinoline ring, resulted from the MD simulation of the MMP-2:1h complex. Density functional theory (DFT) calculations indicated that although the trans asset is more stable than the cis one, the trans/cis ratio can be estimated in 96:4 and, thus, determines a non-negligible amount of cis conformer at equilibrium. These calculations evidenced that the investigated NZIs could be subjected to a trans → cis interconversion in the binding of MMP-2 or, at least, that this hypothesis cannot be ruled out. This cis, less stable, conformation, therefore, can be stabilized by H-bonding interaction with water molecules, not observed for the 1i ligand. This result suggests the role of water molecules in determining the activity of quinoline/isoquinoline derivatives depending on the position of the nitrogen atom. Further interesting data emerging from MD are the number of intramolecular H-bonds of the complex: complexes with ligand 1h have a higher number of H-bonds accounting for a stabilization of the whole system. A similar result has been observed in a previous study that applied docking and molecular dynamics to an analogue case-study on MMP-2 [38]. Furthermore, in the present study, we can confirm that the MMP-2 inhibition process cannot be studied by applying simple docking calculations. The stable trajectories of 1h and 1i complexes were clustered to extract representative bound conformations of MMP-2, and a newly-directed, ensemble docking study was then performed. The ensemble docking procedure was not able to provide evidence for the binding of ligand in cis conformation, possibly for two reasons: (i) the trans → cis interconversion may be an intrinsic feature of a ligand bound at MMP-2, so that a single representative pose would be an average of a cis/trans bundle of structures; docking poses are instead calculated as minimum structures, thus making it rather unlikely to obtain a binding pose intermediate between cis and trans; (ii) explicit water molecules, which are automatically removed by the standard receptor preparation tool (see the Materials and Methods), might be essential to prompt the ligand into a cis or trans conformation, and their exclusion from the binding model may turn out to be a too marked approximation. On the other hand, the ensemble docking allowed enhancing the correlation between activity and score, gaining a higher enrichment in the ROC curve and enabling the application of this model to study the binding of new derivatives.
Int. J. Mol. Sci. 2016, 17, 1768
13 of 19
4. Materials and Methods 4.1. Chemistry 4.1.1. General Melting points were determined with a Büchi Melting Point B-450 (Büchi Labortechnik, Flawil, Switzerland). NMR spectra were recorded on a Varian Mercury 300 spectrometer (Varian, Palo Alto, CA, USA) with 1 H at 300.060 MHz and 13 C at 75.475 MHz. Proton chemical shifts were referenced to the tetramethylsilane (TMS) internal standard. Chemical shifts are reported in parts per million (ppm, δ units). Coupling constants are reported in Hertz (Hz). Splitting patterns are designed as: s, singlet; d, doublet; t, triplet; q, quartet; ql, quartet like; dd, double doublet; m, multiplet; b, broad. All commercial chemicals and solvents are reagent grade and were used without further purification unless otherwise specified. Reactions were monitored by thin-layer chromatography on silica gel plates (60F-254, E. Merck, Merck Group, Darmstadt, Germany) and visualized with UV light, iodine vapors or alkaline KMnO4 aqueous solution. The following solvents and reagents have been abbreviated: triethylamine (TEA), tetrahydrofuran (THF), dimethyl sulfoxide (DMSO), ethyl acetate (EtOAc), dichloromethane (DCM), dimethyl formamide (DMF), methanol (MeOH), ethanol (EtOH), chloroform (CHCl3 ). All reactions were carried out with the use of standard techniques. 4.1.2. General Procedure for the Synthesis of Ureas 1a–α Benzyl isocyanate (1 eq) was added to a solution of the suitable amine or aniline (1 eq) in 10 mL of MeOH (Method a), or CHCl3 , or DCM (Method b), and the mixture was allowed to react at r.t. or refluxing temperature from 1–24 h. The solvent was removed under reduced pressure, and crude products were purified by crystallization or column chromatography. 4.1.3. General Procedure for the Synthesis of Amides 2a,b To a solution of 5-aminoisoquinoline (1 eq) and TEA (1.05 eq) in DCM (3.0 mL) at room temperature was added a solution of phenylacetyl chloride (Compound 2a, 1.05 eq) or 3-phenylpropanoyl chloride (Compound 2b, 1.05 eq), respectively. The mixture was allowed to react at r.t. for 2–24 h. After this time, the solvent was removed under reduced pressure; the crude material was treated with saturated NaHCO3 (20 mL) and extracted with DCM (3 × 20 mL). The organic phase was washed with brine (2 × 60 mL), dried on sodium sulfate and evaporated under reduced pressure. Crude products were purified by crystallization. 4.1.4. General Procedure for the Synthesis of Amides 2c,d 1-Hydroxybenzotriazole (HOBt) (1 eq) and N,N-dicyclohexylcarbodiimide (DCC) (1 eq) were added to a solution of 5-carboxyisoquinoline (1 eq) in DMF (4 mL) at 0 ◦ C. After 10 min, N-methylmorpholine (1 eq) and the proper amine (1 eq) (benzylamine or phenylethylamine) were added, and the reaction was allowed to reach room temperature. After stirring overnight, the solvent was removed under reduced pressure; the crude material was treated with saturated NaHCO3 (20 mL) and extracted with DCM (3 × 20 mL). The organic phase was washed with brine (2 × 60 mL), dried on sodium sulfate and evaporated under reduced pressure. Crude products were purified by column chromatography. 4.1.5. General Procedure for the Synthesis of Carbamates 4 and 6 To a solution of 3 (1.2 g, 5.4 mmol) in pyridine (8 mL) was carefully added phenylsulfonyl chloride or benzoyl chloride (6.5 eq) at 0 ◦ C, and the temperature was allowed to reach room temperature. After 24 h, the mixture was neutralized by 2 N HCl at 0 ◦ C, then extracted with DCM (3 × 20 mL). The organic phase was dried on sodium sulfate and the solvent evaporated under reduced pressure. Crude materials were purified by column chromatography.
Int. J. Mol. Sci. 2016, 17, 1768
14 of 19
4.1.6. General Procedure for the Synthesis of Ureas 5 and 7 A solution of the suitable carbamate (4 or 6, 1 eq) in dioxane was treated with HCl 4 N (5 mL) and stirred at room temperature for 4 h. After removal of volatiles, the deprotected hydrochloric salt was added to a solution of triphosgene (0.5 eq) in toluene (10 mL), in the presence of TEA (1 eq), and the mixture was refluxed for 4 h. After evaporation of toluene under reduced pressure, the solid obtained was dissolved in DCM (10 mL) and treated with the 5-aminoisoquinoline (1 eq) at 0 ◦ C. The reaction was stirred at room temperature overnight, then concentrated to dryness, diluted with EtOAc (15 mL) and washed with HCl 2N (3 × 15 mL) and brine (3 × 15 mL). The organic solution was dried over sodium sulfate and evaporated under reduced pressure. Crude materials were purified by crystallization. 4.2. Enzyme Inhibition Assay The catalytic domains of MMP-2, MMP-8 and MMP-13 were purchased from Enzo Life Sciences (Lörrach, Germany). The assays were performed in triplicate in 96-well white microtiter plates (Corning, NBS; Corning, NY, USA). For assay measurements, inhibitor stock solutions (DMSO, 10 mM) were diluted to six different concentrations (1 nM–100 µM) in fluorometric assay buffer (50 mM Tris·HCl pH 7.5, 200 mM NaCl, 1 mM CaCl2 , 1 µM ZnCl2 , 0.05% Brij-35 and 1% DMSO). Enzyme and inhibitor solutions were incubated in the assay buffer for 15 min at room temperature before the addition of the fluorogenic substrate solution (OmniMMP® = Mca-Pro-Leu-Gly-Leu-Dpa-Ala-Arg-NH2 , Enzo Life Sciences, 2.5 µM final concentration or OmniMMP® RED = TQ3-GABA-Pro-Cha-Abu-Smc-His-Ala-Dab(6’-TAMRA)-Ala-Lys-NH2 , (TAMRA: tetramethylrhodamine) Enzo Life Sciences, 1 µM final concentration). After further incubation for 2–4 h at 37 ◦ C, fluorescence was measured (λex = 340 nm, λem = 405 nm or λex = 545 nm, λem = 572 nm) using a Perkin-Elmer Victor V3 plate reader (Waltham, MA, USA). Control wells lacked inhibitor. The MMP inhibition activity was expressed in relative fluorescence units (RFU). Percent inhibition was calculated from control reactions without inhibitor; IC50 values were determined using GraphPad Prism version 5.0f and are expressed as the mean ± SEM of at least three independent measurements in triplicate. 4.3. Computational Methods 4.3.1. Ligand Preparation The structures of synthesized ligands were manually built exploiting the Build facility in Maestro version 10.2 [39]. The protonation state at pH 7.4 and possible tautomers were generated using the Ligand preparation routine in Maestroand applying the following parameters: force field OPLS2005, generation of the protonation state using Epik version 3.2. The obtained structures were minimized to a derivative convergence of 0.01 kJ·Å−1 ·mol−1 , using the Polak–Ribiere conjugate gradient (PRCG) minimization algorithm, the OPLS2005 force field and the eneralized Born/Surface Area (GB/SA) water solvation model implemented in MacroModel version 10.8 [39]. The conformational search was carried out on all minimized ligand structures, applying the mixed-torsional/low-mode sampling, the OPLS2005 force field, the water solvation model GB/SA and the automatic setup protocol, in order to obtain the global minimum geometry of each molecule, which was then submitted to docking calculations. 4.3.2. Docking 3D structures of MMP-2, MMP-8 and MMP-13 were retrieved from the Protein Data Bank repository [40]. Only complexes with resolution ≤2.5 Å were selected (Supplementary Material, Table S1). Each MMP structure was submitted to the Protein Preparation wizard in Maestro [39] that allows eliminating ligands and water molecules, fix bond orders, add hydrogen atoms, compute the residue protonation state, optimize the H-bond network and relax the structure with a constrained
Int. J. Mol. Sci. 2016, 17, 1768
15 of 19
minimization. All structures were superimposed to have the same frame of reference. The Glide grid generation protocol (Glide version 3.9 [39]) was applied to each structure using the same coordinates for the grid box positioning (x = 67.1815; y = 25.4097; z = 27.9457). The so-prepared structures were submitted to a preliminary cross-docking study (cross-docking details and results are reported in the Supplementary Material, Tables S2–S4) that allowed the selection of the best performing 3D structures, used for the following docking calculations. PDB IDs of selected 3D structures are: 3AYU and 1QIB for MMP-2, 1KBC and 3DPF for MMP-8; and 2OZR for MMP-13. The global minimum conformations of the synthesized ligands were used as the input geometries for docking calculations in the previously-mentioned MMP catalytic sites. The docking calculations were carried out using the Virtual Screening Workflow. Both Glide SP (standard precision) and XP (extra precision) methods were used applying the default parameters [39]. Ten poses were saved for each ligand. The same protocol (protein preparation, grid generation and docking) was applied to the MMP-2 structures obtained from MD trajectory clustering. A total of four MMP-2 structures (corresponding to the two clusters for each complex; see below) were used for the final docking runs. 4.3.3. Molecular Dynamics The docking structures of either or MMP2:1h or 1i complexes used as starting models in subsequent molecular dynamics (MD) simulations with the Gromacs package [41]. Each complex was placed in a cubic box whose dimensions prevent self-interaction and then solvated with up to 9893 water molecules, at the typical density of water (298 K, 1.0 atm) and by employing the 4-site transferable intermolecular potential function (TIP4) water model [42,43]. The molecular systems were simulated in the Amber99 force field [44]. Sodium and chloride ions were added as counterions, to ensure electrical neutrality. Simulations were initially performed by the following scheme: (i) (ii) (iii)
Local energy minimization, heating up to 300 K in six MD runs (100 ps) with protein temperature set at 0, 100, 150, 200, 250 and 300 K; Production runs of 100 ns at 300 K in an isothermal–isobaric (NPT) ensemble, by using the velocity rescaling scheme (temperature) and the isotropic Berendsen coupling scheme (pressure) [45]; The bound proteins were simulated in the Amber force field [46,47]; the linear constrain solver (LINCS) algorithm was adopted to constrain all bond lengths [48], and the long range electrostatics were computed by the particle mesh Ewald method [49].
Such a procedure provided 90 ns of stable trajectory only for the MMP-2:1i complex, whereas longer equilibration times were required by MMP-2:1h simulation, thus corresponding to a different scheme: (i) (ii) (iii)
Local energy minimization, heating up to 300 K in six runs with protein temperature set at 0 (100 ps), 100 (200 ps), 150 (200 ps), 200 (200 ps), 250 (200 ps) and 300 K (200 ps); Production runs of 200 ns at 300 K in an NPT ensemble, by using the velocity rescaling scheme (temperature) and the isotropic Berendsen coupling scheme (pressure); Same setting of MMP-2:1i simulation.
Bonded and Van der Waals parameters for the description of the Zn2+ ions’ first coordination sphere were retrieved from [50], whereas atomic charges were calculated at the density functional theory (DFT) level of theory by using Gaussian 09 [51] (See below). The Amber parameters for the description of ligand 1i structures were mostly implemented in the Gromacs parameter file, and only atomic charges were calculated quantum mechanically with Gaussian 09 (See below). Trajectory analyses were carried out by using the Gromacs utilities, VMD [52] and Maestro [39] graphical interfaces. Ensembles of 900 snapshots per trajectory (one for 100 ps) were extracted, analyzed and clusterized by using the g_cluster utility. Clustering was employed to sample representative configurations of the bound proteins to be used in subsequent docking calculations. Either MMP-2:1h
Int. J. Mol. Sci. 2016, 17, 1768
16 of 19
or MMP-2:1i trajectory ensembles were clusterized by using the method proposed by Daura et al. [53] and a cut-off of 1.5 Å by the comparison of the binding pose and the protein residues or water molecules within 3.0 Å of the ligand. The middle structures for the clusters covering at least 80% of the whole ensemble were eventually selected as representative. 4.3.4. DFT Calculations The local minimum geometries of free 1h and 1i in the gas phase were calculated at the B3LYP/6-31G* [54–56] level of theory by using Jaguar version 8.8 [39]. Atomic charges were then obtained by the fitting of the electrostatic potential by the restrained electrostatic potential (RESP) method [57] and subsequently employed in the MD simulation of the 1h and 1i bound complexes. The same computational scheme was also used for the calculation of the charges of the atoms within the first coordination sphere of the Zn2+ ions. The investigation of the kinetics and thermodynamics of the cis/trans conversion of the 1h ligand (vide infra) was performed by using the Gaussian 09 programs package [51]. The geometry of either the cis or trans conformation of 1h was optimized in the gas phase at the B3LYP/6-31+G* [54–56] level of theory. A relaxed scan of the torsion around the ureidic-aminoisoquinoline amide bond was performed at the same level of theory to obtain a guess structure of the transition state for cis/trans conversion. The transition state structure was then obtained by optimizing the geometry with the maximum energy in the scan as a first-order saddle point by using the Berny algorithm [58,59] implemented in Gaussian 09. Vibrational frequency analyses were then performed at the same level of theory to confirm the correct nature of the optimized stationary points and to calculate the zero-point energy and thermal corrections (under the hypothesis of an ideal gas behavior) for free energy estimations. The integral equation formalism within the polarizable continuum model (IEFPCM) [60]) was used to correct the gas phase free energies of trans-1h, cis-1h and the corresponding trans-to-cis transition state structure (ts-1h) by the effect of water solvation. 5. Conclusions In conclusion, with respect to previous work [28], in the present study, we obtained: (i) the synthesis of 34 new compounds with simplified structures; (ii) a better understanding of the SAR of these quinoline/isoquinoline derivatives, in particular concerning the role of the urea and quinoline ring; (iii) the validation of the hypothesized binding mode, suggesting the missing interaction with the zinc ion; (iv) a computational model based on the combination of docking and molecular dynamics procedures able to enhance the correlation between the activity and docking score in the computer-assisted design of new MMP-2 NZIs. Above all, the combination of experimental and computational outcomes provided by the present investigation shed some lights on the structure-activity and structure-selectivity affecting the investigated compounds and paves the way to future efforts in the development of newly-designed NZIs. Supplementary Materials: Supplementary materials can be found at www.mdpi.com/1422-0067/17/10/1768/s1. Acknowledgments: This study was supported by grants from the Ministry of Education, University and Research of Italy (University “Gabriele d’Annunzio”, FAR funds to Alessandra Ammazzalorso, Barbara De Filippis, Cristina Campestre, Rosa Amoroso, Alessandro Marrone and Mariangela Agamennone). Author Contributions: Alessandra Ammazzalorso, Barbara De Filippis, Cristina Campestre and Rosa Amoroso performed the chemical synthesis and contributed reagents/materials. Antonio Laghezza and Paolo Tortorella performed the biological assays. Alessandro Marrone and Mariangela Agamennone performed the computational studies. Mariangela Agamennone and Paolo Tortorella conceived of and designed the experiments. All authors contributed to writing the paper. Conflicts of Interest: The authors declare no conflict of interest.
Int. J. Mol. Sci. 2016, 17, 1768
17 of 19
References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.
13. 14. 15. 16. 17. 18. 19. 20.
21.
22.
23.
Apte, S.S.; Parks, W.C. Metalloproteinases: A parade of functions in matrix biology and an outlook for the future. Matrix Bio. 2015, 44–46, 1–6. [CrossRef] [PubMed] Burrage, P.S.; Mix, K.S.; Brinckerhoff, C.E. Matrix metalloproteinases: Role in arthritis. Front Biosci. 2006, 11, 529–543. [CrossRef] [PubMed] Lohmander, L.S.; Hoerrner, L.A.; Lark, M.W. Metalloproteinases, tissue inhibitor, and proteoglycan fragments in knee synovial fluid in human osteoarthritis. Arthritis Rheum. 1993, 36, 181–189. [CrossRef] [PubMed] Brauer, P.R. MMPs: Role in cardiovascular development and disease. Front Biosci. 2006, 11, 447–478. [CrossRef] [PubMed] Hu, J.; Van den Steen, P.E.; Sang, Q.A.; Opdenakker, G. Matrix metalloproteinase inhibitors as therapy for inflammatory and vascular diseases. Nat. Rev. Drug Discov. 2007, 6, 480–498. [CrossRef] [PubMed] Yong, V.W. Metalloproteinases: Mediators of pathology and regeneration in the CNS. Nat. Rev. Neurosci. 2005, 6, 931–944. [CrossRef] [PubMed] Peress, N.; Perillo, E.; Zucker, S. Localization of tissue inhibitor of matrix metalloproteinases in Alzheimer's disease and normal brain. J. Neuropathol. Exp. Neurol. 1995, 54, 16–22. [CrossRef] [PubMed] Chandler, S.; Coates, R.; Gearing, A.; Lury, J.; Wells, G.; Bone, E. Matrix metalloproteinases degrade myelin basic protein. Neurosci. Lett. 1995, 201, 223–226. [CrossRef] Demers, M.; Couillard, J.; Bélanger, S.; St-Pierre, Y. New roles for matrix metalloproteinases in metastasis. Crit. Rev. Immunol. 2005, 25, 493–523. [CrossRef] [PubMed] Coussens, L.M.; Fingleton, B.; Matrisian, L.M. Matrix metalloproteinase inhibitors and cancer: Trials and tribulations. Science 2002, 295, 2387–2392. [CrossRef] [PubMed] Chang, C.; Werb, Z. The many faces of metalloproteases: Cell growth, invasion, angiogenesis and metastasis. Trends Cell. Biol. 2001, 11, S37–S43. [CrossRef] Cox, J.H.; Overall, C.M. Cytokine substrates: MMP regulation of inflammatory signaling molecules. In The Cancer Degradome; Edwards, D., Hoyer-Hansen, G., Blasi, F., Sloane, B.F., Eds.; Springer: New York, NY, USA, 2008; pp. 519–539. Vandenbroucke, R.E.; Libert, C. Is there new hope for therapeutic matrix metalloproteinase inhibition? Nat. Rev. Drug Discov. 2014, 13, 904–927. [CrossRef] [PubMed] Dufour, A.; Overall, C.M. Missing the target: Matrix metalloproteinase antitargets in inflammation and cancer. Trends Pharmacol. Sci. 2013, 34, 233–242. [CrossRef] [PubMed] Overall, C. M.; Kleifeld, O. Tumour microenvironment-opinion: Validating matrix metalloproteinases as drug targets and anti-targets for cancer therapy. Nat. Rev. Cancer 2006, 6, 227–239. [CrossRef] [PubMed] Fields, G.B. New strategies for targeting matrix metalloproteinases. Matrix Biology 2015, 44–46, 239–246. [CrossRef] [PubMed] Campestre, C.; Agamennone, M.; Tauro, M.; Tortorella, P. Phosphonate emerging zinc binding group in matrix metalloproteinase inhibitors. Curr. Drug Targets 2015, 16, 1634–1644. [CrossRef] [PubMed] Agrawal, A.; Romero-Perez, D.; Jacobsen, J.A.; Villarreal, F.J.; Cohen, S.M. Zinc-binding groups modulate selective inhibition of MMPs. ChemMedChem 2008, 3, 812–820. [CrossRef] [PubMed] Tanakit, A.; Rouffet, M.; Martin, D.P.; Cohen, S.M. Investigating chelating sulfonamides and their use in metalloproteinase inhibitors. Dalton Trans. 2012, 41, 6507–6515. [CrossRef] [PubMed] Campestre, C.; Tortorella, P.; Agamennone, M.; Preziuso, S.; Biasone, A.; Nuti, E.; Rossello, A.; Gallina, C. Peptidyl 3-substituted 1-hydroxyureas as isosteric analogues of succinylhydroxamate MMP inhibitors. Eur. J. Med. Chem. 2008, 43, 1008–1014. [CrossRef] [PubMed] Rubino, M.T.; Agamennone, M.; Campestre, C.; Fracchiolla, G.; Laghezza, A.; Loiodice, F.; Nuti, E.; Rossello, A.; Tortorella, P. Synthesis, SAR, and biological evaluation of α-sulfonylphosphonic acids as selective matrix metalloproteinase inhibitors. ChemMedChem 2009, 4, 352–362. [CrossRef] [PubMed] Tauro, M.; Laghezza, A.; Loiodice, F.; Agamennone, M.; Campestre, C.; Tortorella, P. Arylamino methylene bisphosphonate derivatives as bone seeking matrix metalloproteinase inhibitors. Bioorg. Med. Chem. 2013, 21, 6456–6465. [CrossRef] [PubMed] Devy, L.; Dransfield, D.T. New strategies for the next generation of matrix-metalloproteinase inhibitors: Selectively targeting membrane-anchored MMPs with therapeutic antibodies. Biochem. Res. Int. 2011, 2011, 191670. [CrossRef] [PubMed]
Int. J. Mol. Sci. 2016, 17, 1768
24.
25. 26.
27.
28. 29.
30. 31. 32. 33.
34.
35. 36.
37.
38. 39. 40. 41. 42.
43.
18 of 19
Jacobsen, J.A.; Major Jourden, J.L.; Miller, M.T.; Cohen, M.T. To bind zinc or not to bind zinc: An examination of innovative approaches to improved metalloproteinase inhibition. Biochim. Biophys. Acta 2010, 1803, 72–94. [CrossRef] [PubMed] Di Pizio, A.; Agamennone, M.; Tortorella, P. Non-zinc-binding inhibitors of MMP-13: GRID-based approaches to rationalize the binding process. Curr. Top. Med. Chem. 2016, 16, 449–459. [CrossRef] [PubMed] Pochetti, G.; Montanari, R.; Gege, C.; Chevrier, C.; Taveras, A.G.; Mazza, F. Extra binding region induced by non-zinc chelating inhibitors into the S1 subsite of matrix metalloproteinase 8 (MMP-8). J. Med. Chem. 2009, 52, 1040–1049. [CrossRef] [PubMed] Dublanchet, A.C.; Ducrot, P.; Andrianjara, C.; O’Gara, M.; Morales, R.; Compère, D.; Denis, A.; Blais, S.; Cluzeau, P.; Courté, K.; et al. Structure-based design and synthesis of novel non-zinc chelating MMP-12 inhibitors. Bioorg. Med. Chem. Lett. 2005, 15, 3787–3790. [CrossRef] [PubMed] Di Pizio, A.; Laghezza, A.; Tortorella, P.; Agamennone, M. Probing the S10 Site for the Identification of Non-Zinc-Binding MMP-2 Inhibitor. ChemMedChem 2013, 8, 1475–1482. [CrossRef] [PubMed] Sbardella, D.; Fasciglione, G.F.; Gioia, M.; Ciaccio, C.; Tundo, G.R.; Marini, S.; Coletta, M. Human matrix metalloproteinases: An ubiquitarian class of enzymes involved in several pathological processes. Mol. Asp. Med. 2012, 33, 119–208. [CrossRef] [PubMed] Blanc, M.; Cussac, M.; Boucherle, A.; Leclerc, G. Synthesis and immunomodulating activity of 1-amino-2-thiohydantoin derivatives. Eur. J. Med. Chem. 1992, 27, 839–843. [CrossRef] Albrecht, M.; Witt, K.; Frohlich, R.; Kataeva, O. Inter- and intramolecular hydrogen bonding in amide- and urea-substituted 8-hydroxyquinoline derivatives. Tetrahedron 2002, 58, 561–567. [CrossRef] Zaitsev, V.G.; Shabashov, D.; Daugulis, O. Highly regioselective arylation of sp3 C–H bonds catalyzed by palladium acetate. J. Am. Chem. Soc. 2005, 127, 13154–13155. [CrossRef] [PubMed] Dea-Ayuela, A.M.; Castillo, E.; Gonzalez-Alvarez, M.; Vega, C.; Rolón, M.; Bolás-Fernández, F.; Borrás, J.; González-Rosende, M.E. In vivo and in vitro anti-leishmanial activities of 4-nitro-N-pyrimidinand N-pyrazin-2-ylbenzenesulfonamides, and N2-(4-nitrophenyl)-N1-propylglycinamide. Bioorg. Med. Chem. 2009, 17, 7449–7456. [CrossRef] [PubMed] Acero-Alarcón, A.; Armero-Alarte, T.; Jordá-Gregori, J.M.; Rojas-Argudo, C.; Zaballos-García, E.; Server-Carrio, J.; Ahjyaje, F.Z.; Sepúlveda-Arques, J. Unusual ring closure reaction of amides with pyrimidines: Novel stereoselective synthesis of hexahydroimidazo[1,2-c]pyrimidines. Synthesis 1999, 12, 2124–2130. [CrossRef] Jain, P.; Saravanan, C.; Singh, S.K. Sulphonamides: Deserving class as MMP inhibitors? Eur. J. Med. Chem. 2013, 60, 89–100. [CrossRef] [PubMed] Nanjan, P.; Nambiar, J.; Nair, B.G.; Banerji, A. Synthesis and discovery of (I-3,II-3)-biacacetin as a novel non-zinc binding inhibitor of MMP-2 and MMP-9. Bioorg. Med. Chem. 2015, 23, 3781–3787. [CrossRef] [PubMed] Song, J.; Peng, P.; Chang, J.; Liu, M.M.; Yu, J.M.; Zhou, L.; Sun, X. Selective non-zinc binding MMP-2 inhibitors: Novel benzamide ilomastat analogs with anti-tumor metastasis. Bioorg. Med. Chem. 2016, 26, 2174–2178. [CrossRef] [PubMed] Di Pizio, A.; Agamennone, M.; Aschi, M. An integrated computational approach to rationalize the activity of non-zinc-binding MMP-2 inhibitors. PLoS ONE 2012, 7, e47774. [CrossRef] [PubMed] Schrödinger LLC. Maestro Version 10.2, LigPrep Version 3.4, Epik, Version 3.2, MacroModel, Version 10.8, Jaguar, Version 8.8, Glide Version 6.7; Schrödinger LLC: New York, NY, USA, 2015. Berman, H.M.; Westbrook, J.; Feng, Z.; Gilliland, G.; Bhat, T.N.; Weissig, H.; Shindyalov, I.N.; Bourne, P.E. The Protein Data Bank. Nucleic Acids Res. 2000, 28, 235–242. [CrossRef] [PubMed] Van Der Spoel, D.; Lindahl, E.; Hess, B.; Groenhof, G.; Mark, A.E.; Berendsen, H.J. GROMACS: Fast, flexible, and free. J. Comput. Chem. 2005, 26, 1701–1718. [CrossRef] [PubMed] Berendsen, H.J.C.; Postma, J.P.M.; van Gunsteren, W.F.; Hermans, J. Interaction models for water in relation to protein hydration. In Intermolecular Forces; Pullmann, B., Ed.; Springer Netherlands: Dordrecht, The Netherlands, 1981; pp. 331–342. Jorgensen, W.L.; Chandrasekhar, J.; Madura, J.D.; Impey, R.W.; Klein, M.L. Comparison of simple potential functions for simulating liquid water. J. Chem. Phys. 1983, 79, 926–935. [CrossRef]
Int. J. Mol. Sci. 2016, 17, 1768
44.
45. 46.
47.
48. 49. 50. 51. 52. 53. 54. 55. 56.
57.
58. 59. 60.
19 of 19
Wang, J.; Cieplak, P.; Kollman, P.A. How well does a restrained electrostatic potential (RESP) model perform in calculating conformational energies of organic and biological molecules? J. Comp. Chem. 2000, 21, 1049–1074. [CrossRef] Berendsen, H.J.C.; Postma, J.P.M.; van Gunsteren, W.F.; di Nola, A.; Haak, J.R. Molecular dynamics with coupling to an external bath. J. Chem. Phys. 1984, 81, 3684–3690. [CrossRef] Jorgensen, W.L.; Maxwell, D.S.; Tirado-Rives, J. Development and testing of the OPLS all-atom force field on conformational energetics and properties of organic liquids. J. Am. Chem. Soc. 1996, 118, 11225–11236. [CrossRef] Kaminski, G.A.; Friesner, R.A.; Tirado-Rives, J.; Jorgensen, W.J. Evaluation and reparametrization of the OPLS-AA force field for proteins via comparison with accurate quantum chemical calculations on peptides. J. Phys. Chem. B 2001, 105, 6474–6487. [CrossRef] Hess, B.; Bekker, H.; Berendsen, H.J.C.; Frajie, J.C.E.M. LINCS: A linear constraint solver for molecular simulations. J. Comp. Chem. 1997, 18, 1463–1472. [CrossRef] Darden, T.A.; York, D.M.; Pedersen, L.G. Particle mesh ewald: An N log(N) method for ewald sums in large systems. J. Chem. Phys. 1993, 98, 10089–10092. [CrossRef] Lin, F.; Wang, R. Systematic derivation of AMBER force field parameters applicable to zinc-containing systems. J. Chem. Theory Comput. 2010, 6, 1852–1870. [CrossRef] [PubMed] Gaussian Inc. Gaussian 09, Revision D.01; Gaussian Inc.: Wallingford, CT, USA, 2013. Humphrey, W.; Dalke, A.; Schulten, K. VMD: Visual Molecular Dynamics. J. Mol. Graph. 1996, 14, 33–38. [CrossRef] Daura, X.; Gademann, K.; Jaun, B.; Seebach, D.; van Gunsteren, W.F.; Mark, A.E. Peptide folding: When simulation meets experiment. Angew. Chem. Int. Ed. 1999, 38, 236–240. [CrossRef] Becke, A.D. Density-functional thermochemistry. III. The role of exact exchange. J. Chem. Phys. 1993, 98, 5648–5452. [CrossRef] Ditchfield, R.; Hehre, W.J.; Pople, J.A. Self-consistent molecular orbital methods. IX. extended Gaussian-type basis for molecular-orbital studies of organic molecules. J. Chem. Phys. 1971, 54, 724–728. [CrossRef] Francl, M.M.; Pietro, W.J.; Hehre, W.J.; Binkley, J.S.; DeFrees, D.J.; Pople, J.A.; Gordon, M.S. Self-consistent molecular orbital methods. XXIII. A polarization-type basis set for 2nd-row elements. J. Chem. Phys. 1982, 77, 3654–3665. [CrossRef] Breneman, C.M.; Wiberg, K.B. Determining atom-centered monopoles from molecular electrostatic potentials. The need for high sampling density in formamide conformational analysis. J. Comput. Chem. 1990, 11, 361–373. [CrossRef] Li, X.; Frisch, M.J. Energy-represented DIIS within a hybrid geometry optimization method. J. Comput. Chem. 2006, 2, 835–839. Peng, C.; Ayala, P.Y.; Schlegel, H.B.; Frisch, M.J. Using redundant internal coordinates to optimize equilibrium geometries and transition states. J. Comput. Chem. 1996, 17, 49–56. [CrossRef] Tomasi, J.; Mennucci, B.; Cammi, R. Quantum mechanical continuum solvation models. Chem. Rev. 2005, 105, 2999–3093. [CrossRef] [PubMed] © 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).
