Exploring Oxidovanadium(IV) Complexes as YopH Inhibitors: Mechanism of Action and Modeling Studies.

YopH tyrosine phosphatase, a virulence factor produced by pathogenic species of Yersinia, is an attractive drug target. In this work, three oxidovanadium(IV) complexes were assayed against recombinant YopH and showed strong inhibition of the enzyme in the nanomolar range. Molecular modeling indicated that their binding is reinforced by H-bond, cation-π, and π-π interactions conferring specificity toward YopH. These complexes are thus interesting lead molecules for phosphatase inhibitor drug discovery.

Protein phosphorylation at histidine, serine, threonine, and tyrosine residues is controlled by two large families of enzymes: phosphatases and kinases. While the transfer of a γ-phosphate from ATP to target residues is operated by kinases, phosphatases act at the opposite way by restoring the unphosphorylated state of their substrates. This reversible phosphorylation can affect many different properties of target proteins, including interference in their activity, localization, interaction with other proteins, conformation, and function. Despite the fact that phosphorylation at tyrosine residues represents only about 0.01 to 0.05% of total protein phosphorylation,[1] it plays a critical role in controlling many physiological processes such as cell–cell communications, differentiation, cell cycle, cell growth, metabolism, and immune response. Due to the importance of tyrosine phosphorylation to cell homeostasis, any factor that could alter the physiological tyrosine dephosphorylation rate, either by excessive or decreased activity of protein tyrosine phosphatases (PTPs), could cause a variety of diseases such as cancer,[2] obesity, diabetes,[3] and neurodegenerative and autoimmune disorders.[4]

The metal oxoanion vanadate is reported as inhibitor of PTPs since 1980s. Although vanadate behaves as a phosphate mimetic, it displays higher inhibitory profile than the phosphate ion[5] and assumes a trigonal–bipyramidal structure within the active site of the enzyme.[6,7] It has been shown that vanadium compounds also present insulin-mimetic effect, prolonging this hormone action by increasing insulin receptor activation and downstream signaling. Bis-maltolato-oxidovanadium(IV) (BMOV), for example, is an organic oxidovanadium compound that decreases the level of plasma glucose in diabetic animals.[8] Bis(ethylmaltolato)oxidovanadium(IV) (BEOV), an analogue of BMOV, has been profiled for safety and efficacy by Akesis Pharmaceuticals, Inc. (La Jolla, CA), in human phase II clinical trial and presented a good tolerance, reduction of fast blood glucose comparing to placebo subjects, and an improvement in the response to an oral glucose load.[9] The insulin-like properties of vanadium compounds could partly be explained by its activity against PTPs. Among them, PTP1B aberrant activity has long been associated with diabetes, and its catalytic activity has been efficiently inhibited by organic vanadium complexes, including BMOV and BEOV.[10−12]

PTPs produced by pathogenic bacteria are also able to interfere with host signaling pathways. Yersinia outer protein H (YopH), for example, is a 51 kDa PTP that is secreted by the three pathogenic species of Yersinia genus: Yersinia pestis, Yersinia enterocolitica, and Yersinia pseudotuberculosis. These species cause a variety of pathologies from gastrointestinal syndromes to sepsis. YopH is the PTP that presents one of the highest catalytic efficiencies among all known PTPs (kcat/Km 514 mM–1·s–1).[13] YopH is injected into the host cell by a type III secretion apparatus and helps these bacteria to evade the host’s immune defense. YopH can dephosphorylate the focal adhesion proteins p130Cas and FAK of the host, leading to a subsequent disruption of the focal complex structures.[14] YopH protects the bacteria from phagocytic uptake by cultured macrophages and polymorphonuclear neutrophils (PMNs)[15] and interferes at macrophage response to infection by inactivation the PI3K/Akt pathway.[16] YopH can also impair the activation of T and B lymphocytes, thus allowing Yersinia to avoid the adaptive immune response.[17] Due to the importance of YopH for bacteria pathogenesis and based on the evidence that a catalytic mutation of YopH enzyme abolishes the virulence of Y. pseudotuberculosis in murine infection model,[18] this PTP is considered an attractive drug target.

In the last years, a variety of YopH inhibitors belonging to different chemical classes have been reported such as aurintricarboxylic acid,[19] salicylic acid derivatives,[20] peptides,[21] phosphonate derivatives of calixarenes,[22] oximes,[23] isothiazolidinones,[24] chalcones, and sulphonamides.[25] To further contribute to the field of YopH inhibition, the present work shows the potent inhibitory effect of three oxidovanadium(IV) complexes (compounds 1, 2,[26,27] and 3(28) in Table ).

Table 1

Inhibition and IC50 against YopH from Yersinia enterocolitica and Estimation of the Ligand Binding Affinity to YopH by the Chemscore Function of the GOLD Docking Program for the Selected Compounds and Complexesa

Data are expressed as means ± SD of three independent experiments.

α-dimensional value; higher the fitness, higher the affinity; n.d. ,not determined.

The inhibitory activity of these molecules was evaluated in vitro using a recombinant purified YopH and p-nitrophenylphosphate (pNPP) as substrate, as described in Supporting Information. Initial assays were performed using a single concentration (25 μM) of each molecule and showed more than 99% of inhibitory effect against YopH (Table ). Then, we further determined IC50 values, and the three oxidovanadium(IV) complexes showed extraordinary inhibition in the nanomolar range concentration. IC50 values for the ligands used to build the complexes (L1–L4) were also determined (Table ), and notably, they were at least 80-fold higher than those of their corresponding oxidovanadium(IV) complexes (1–3), demonstrating that the ligands per se do not exhibit a remarkable inhibitory activity against YopH, as could be expected.

To explore the specificity of oxidovanadium(IV) complexes for YopH, we measured their inhibitory activity toward several members of the PTP superfamily, including Mycobacterium tuberculosis PTPs (PtpA and PtpB) and human PTPs (PTP1B, LYP, and PTP–PEST). The results presented in Table demonstrate that the complexes are quite selective toward YopH. The most potent and selective inhibitor 1 showed an approximate 9-fold and 29-fold selectivity for YopH with respect to Mycobacterium tuberculosis PTPs (PtpA and PtpB, respectively). The sole difference between compounds 1 and 2 is an amino group substituted on the phenyl ring. Apparently, the presence of this group results in an increase in selectivity for YopH relative to PTP1B and LYP, whereas it decreases this selectivity in comparison to Mycobacterium tuberculosis PTPs.

Table 2

IC50 and Selectivity Index Values (SI) of Oxidovanadium(IV) Complexes 1, 2, and 3 Measured on Some Representative PTPsa

	1		2		3
PTP	IC50 (μM)	SI	IC50 (μM)	SI	IC50 (μM)	SI
YopH	0.06 ± 0.01	ref value	0.10 ± 0.01	ref value	0.29 ± 0.01	ref value
PtpA	0.56 ± 0.03	9.3	2.69 ± 0.27	26.9	9.12 ± 0.26	31.5
PtpB	1.73 ± 0.61	28.8	5.12 ± 0.78	51.2	52.55 ± 1.56	181.2
PTP1B	0.33 ± 0.05	5.5	0.27 ± 0.02	2.7	0.78 ± 0.02	2.7
LYP	5.42 ± 1.32	90.3	6.82 ± 1.93	68.2	28.30 ± 1.71	97.6
PTP–PEST	3.97 ± 0.23	66.2	4.28 ± 0.66	42.8	41.30 ± 3.19	142.4

The results are shown as the average of the individual mean ± SD for three independent experiments. The selective index (SI) is given by (IC50PTP/IC50YopH).

We also examined the potential reversibility of binding of the inhibitors by modification of our in vitro functional assay (see SI). After removal of unbound inhibitor, YopH activity was always restored since binding of the inhibitor was not irreversible (Figure S1). To further characterize the mechanism of YopH inhibition and evaluate the kinetic parameters, we tested three different concentrations of the vanadium complex and seven concentrations of pNPP (ranging from 0.2 to 12.8 mM). Kinetic analysis revealed that the three molecules are competitive inhibitors. The Lineweaver–Burk double reciprocal plots (Figure ) showed all lines converging at the y-axis (1/Vmax), whereas the slope (Kmapp/Vmax) and x-axis interception (1/Kmapp) vary according to the inhibitor concentration. The constant Vmax value and the increase in Kmapp values with increasing inhibitor concentration indicate a mutually exclusive binding mode, thus suggesting that these compete with pNPP for the active site of YopH. The Ki values obtained were 0.02 μM for complex 1 and 2, and 0.19 for complex 3.

Figure 1

Lineweaver–Burk double-reciprocal plots representing inhibitory profiles of oxidovanadium(IV) complexes 1, 2, and 3 against YopH. Kinetic experiments were conducted in the presence of increasing concentrations of inhibitors, and pNPP was used as the substrate in all experiments.

The possible binding mode of 1–3 and L1–L4 to YopH was investigated by molecular modeling. Small molecules were prepared as described in the Supporting Information and were then docked toward the catalytic active site of YopH phosphatase by using AutoDock, FRED, and GOLD docking programs.[29−33] The high resolution (1.42 Å) crystal structure of YopH in complex with a divanadate ion within the catalytic active site, which corresponds to PDB ID 3F9B, was selected as rigid receptor in this study[34,35]

All docking programs provided a highly similar binding mode for each compound, summarized in Figure for active inhibitors 1–3 and in Supporting Information Figure S2 for noncomplexed fragments L1–L4.

Figure 2

Predicted binding conformation of oxidovanadium complexes 1–3 to YopH. (A) Docking-based binding mode of 1, the most active inhibitor of the series; (B) docking-based binding mode of 2; (C) docking-based binding mode of 3. Ligands are shown as cyan sticks, vanadium as gray sphere. The crystallographic structure of YopH (PDB 3F9B) is shown as green cartoon and lines. Residues involved in binding to 1–3 are shown as orange sticks and are labeled. Ligand nonpolar H atoms were omitted. H-bond interactions are shown as black dashed lines.

All docking poses of 1 and 3 are very similar, with an RMSD lower than 1.5 Å, whereas two different poses were found for 2, one of which is highly similar to the binding mode predicted for 1 and was therefore accounted for in results analysis (for the other binding conformation of 2, see Supporting Information). Molecules 1–3 bind within the catalytic pocket of YopH in proximity to Cys403. Their binding to YopH is reinforced by a cation−π interaction with Arg409 and Lys447, whereas a H-bond interaction is established with the side chain of Gln446. It is worth mentioning that the interaction with Gln446 is also present in the crystallographic structure of YopH bound to a known inhibitor.[36] Compounds 1–3 further establish a π–π interaction with the side chain or Phe229 (T-shape geometry for 1 and 2, parallel sandwich-like geometry for 3), whereas the amino group of 1 is involved in a H-bond interaction with the side chain of Asp231. Notably, this interaction may account for the higher affinity of 1 compared to 2. One may note that Asp231 as well as other residues of the catalytic cleft are generally highly conserved among PTPs. However, as also recently reviewed by Böhmer et al., PTPs share different and peculiar folding.[37] Indeed, YopH shares a good structural similarity with PTP1B (RMSD = 3.5 Å by comparing PDB 3F9B and 2NT7), whereas a higher difference does exist with PtpA and PtpB (RMSD higher than 10 Å by comparing PDB 3F9B with 1U2P and 2OZ5, respectively). Moreover, PtpB is characterized by a peculiar folding and is therefore quite different from all other PTPs. These structural differences may account for the different inhibitory activity of 1–3 against PTPs tested in this work (Table ). Indeed, when compared to YopH inhibition, 1–3 are quite active on PTP1B, which is similar to YopH, but are less active against PtpA and PtpB, which have a different folding structure than YopH.

Molecular docking of noncomplexed fragments L1–L4 revealed that they interact within the same pocket of the catalytic site of YopH (for details see Supporting Information Figure S3). Accordingly, L1–L4 may compete for the same binding site on YopH surface, thus potentially explaining the lack of synergistic effects when tested in combination, as observed experimentally (Table ). Prediction of ligand binding affinity revealed that the Chemscore function (GOLD docking program) provided better agreement with experimental inhibitory data than other scoring functions, similar to what was outlined in recent computational studies on organometallic compounds (Table ).[38] Nevertheless, Chemscore slightly overestimated the predicted affinity of L2 and, particularly, of L3 compared to newly discovered and reference YopH inhibitors (see also Supporting Information).

In summary, the complexation with vanadium is thought to place functional group of 1–3 in correspondence of YopH pharmacophoric hotspots, thus resulting in a higher inhibitory efficacy than noncomplexed fragments L1–L4 alone. Besides complexation with vanadium, the H-bond interaction established by 1 with the side chain of Asp231 appears to be a key discriminant for potent inhibition of YopH phosphatase catalytic activity.

Despite highly chemically diverse inhibitors of YopH have been disclosed so far, only a limited number of them present inhibitory activity in the nanomolar range. Oxidovanadium(IV) complexes have demonstrated excellent inhibition against PTP1B[10−12] but have been poorly explored as YopH inhibitors.[39] Due to the fact that the organic bioactive ligands could be favorably modified, vanadium complexes of organic ligands offer an attractive prospect to develop new potent and selective inhibitors of YopH. In this work we presented three oxidovanadium(IV) complexes endowed with remarkable inhibition against YopH. These compounds have IC50 and Ki values in the nanomolar range, and the most potent is at least 5-fold more selective for YopH compared to other human PTPs.

To identify nanomolar inhibitors of YopH that experience also a certain degree of selectivity is in perspective with developing new drug candidates against Yersinia, and most notably, these molecules would be also useful as a tool in chemical biology investigations.