Picomolar Concentrations of Free Zinc(II) Ions Regulate Receptor Protein-tyrosine Phosphatase β Activity*

Background: Protein-tyrosine phosphatases (PTPs) are critical regulators of phosphorylation signaling. Results: Zinc(II) ions are reversible inhibitors of receptor PTPβ with a remarkably low Ki of 21 ± 7 pm. Conclusion: The inhibition implicates cellular zinc(II) ions as physiological regulators of PTPs. Significance: The results suggest that zinc modulates signal transduction in endothelial cells affecting angiogenesis and wound healing. As key enzymes in the regulation of biological phosphorylations, protein-tyrosine phosphatases are central to the control of cellular signaling and metabolism. Zinc(II) ions are known to inhibit these enzymes, but the physiological significance of this inhibition has remained elusive. Employing metal buffering for strict metal control and performing a kinetic analysis, we now demonstrate that zinc(II) ions are reversible inhibitors of the cytoplasmic catalytic domain of the receptor protein-tyrosine phosphatase β (also known as vascular endothelial protein-tyrosine phosphatase). The Ki(Zn) value is 21 ± 7 pm, 6 orders of magnitude lower than zinc inhibition reported previously for this enzyme. It exceeds the affinity of the most potent synthetic small molecule inhibitors targeting these enzymes. Inhibition is in the range of cellular zinc(II) ion concentrations, suggesting that zinc regulates this enzyme, which is involved in vascular physiology and angiogenesis. Thus, for some enzymes that are not recognized as zinc metalloenzymes, zinc binding inhibits rather than activates as in classical zinc enzymes. Activation then requires removal of the inhibitory zinc.

As key enzymes in the regulation of biological phosphorylations, protein-tyrosine phosphatases are central to the control of cellular signaling and metabolism. Zinc(II) ions are known to inhibit these enzymes, but the physiological significance of this inhibition has remained elusive. Employing metal buffering for strict metal control and performing a kinetic analysis, we now demonstrate that zinc(II) ions are reversible inhibitors of the cytoplasmic catalytic domain of the receptor protein-tyrosine phosphatase ␤ (also known as vascular endothelial protein-tyrosine phosphatase). The K i(Zn) value is 21 ؎ 7 pM, 6 orders of magnitude lower than zinc inhibition reported previously for this enzyme. It exceeds the affinity of the most potent synthetic small molecule inhibitors targeting these enzymes. Inhibition is in the range of cellular zinc(II) ion concentrations, suggesting that zinc regulates this enzyme, which is involved in vascular physiology and angiogenesis. Thus, for some enzymes that are not recognized as zinc metalloenzymes, zinc binding inhibits rather than activates as in classical zinc enzymes. Activation then requires removal of the inhibitory zinc.
Phosphatases are major control enzymes in phosphorylation signaling, and they do not simply oppose the action of kinases (1). Protein-tyrosine phosphatases (PTPs) 2 are one subfamily of phosphatases. The human genome contains 107 PTPs that belong to families of soluble (nonreceptor-type) and membrane-resident (receptor-type) proteins (2). PTPs have pivotal functions in cellular processes and are extensively regulated. At the protein level, this regulation includes redox modulation, phosphorylation, sumoylation, dimerization, and proteolysis (3). With the exception of class IV PTPs, the catalytic group in their active sites is a cysteine with a remarkably low thiol pK a (about 4.4 -5.5 in PTP-1B). The cysteine forms a phosphoenzyme intermediate in the reaction with substrate. Redox modulation of PTP-1B involves oxidation of the catalytic cysteine to a sulfenic acid, which reacts with a backbone nitrogen of an adjacent serine to form a cyclic sulfenyl amide (4,5). Other PTPs are redox-modulated differently. They form a disulfide between the two "backdoor" cysteines or between the catalytic cysteine and one of the backdoor cysteines (6,7).
Vanadium compounds and zinc(II) ions are classical inhibitors of PTPs. Meta-and orthovanadate mimic the phosphate group in the transition states of the enzyme and therefore have been valuable in delineating the reaction mechanism (8). In contrast to vanadium, zinc is nutritionally essential with total cellular concentrations in the range of a few hundred micromolar. Because zinc(II) ions at 10 M concentrations completely inhibit PTP, it was suggested that they are a physiologically important inhibitor (9). However, cellular zinc(II) ions are not available at these concentrations. Free zinc(II) ions are orders of magnitudes lower than total cellular zinc concentrations. They are estimated to be in the picomolar range and to fluctuate in this range (10 -14). Because the observed zinc inhibition of PTPs is much higher than the available cellular zinc(II) ion concentrations, zinc inhibition was discussed as a consequence of toxic exposure to zinc or at elevated intracellular zinc(II) ion concentrations in oxidative neuronal death (15,16). More recent investigations, however, demonstrated that zinc binds significantly tighter: 200 nM (IC 50 ) for human T-cell PTP (17) versus 85% inhibition at 100 M (18); 17 nM (K i (app)) for human PTP-1B (19,20); and 93 nM for PTP1C (SHP-1) (19). Such tight inhibition suggests that zinc modulates PTP activity under physiological conditions in a variety of signaling pathways (21). For example, PTP-1B inhibition causes the IGF-1/insulin receptor to remain in a phosphorylated state. Thus, zinc inhibition of PTP-1B and the ensuing sustained insulin signaling have been suggested to account for the insulinomimetic effects of zinc(II) ions (22,23). Also, zinc inhibition of a MAPK tyrosine phosphatase in Erk signaling is involved in hippocampusdependent memory (24).
In this study, we investigated the effects of zinc on PTPs in more detail. Specifically, we optimized the assay conditions, employed a sensitive fluorometric assay, and performed a kinetic analysis of the inhibition. We demonstrate that zinc(II) ions reversibly inhibit the cytoplasmic catalytic domain of RPTP␤ (R3 subtype) with unparalleled potency. Zinc inhibition is 6 orders of magnitude tighter than reported previously for RPTP␤ (25). In the context of cellular zinc(II) ion concentrations and their fluctuations, the data suggest that cellular zinc inhibits RPTP␤ tonically and that the enzyme must be activated by zinc chelation.
PTP Assay-RPTP␤ was assayed fluorometrically for enzymatic activity at 23°C in a freshly prepared buffer containing final concentrations of 50 mM HEPES/Na ϩ , 100 M TCEP, and 1 mM NTA, pH 7.4. The enzyme was added to this buffer to yield final concentrations of 0.5-5 nM. The reaction was initiated by adding FDP, a fluorogenic phosphatase substrate (26), to a final concentration of 0.8 M. Assays were performed in triplicate in a total volume of 100 l in 96-well black optical bottom plates (p-clear plate black, reference number 655090, Greiner Bio-One Ltd., Stonehouse, UK). The increase in fluorescence due to the hydrolysis of FDP to fluorescein was measured at 485 Ϯ 20 nm excitation and 530 Ϯ 20 nm emission for up to 1 h, using a fluorescence plate reader (Synergy HT, BioTek, Winooski, VT). Initial velocities were determined from the linear portion of the progress curves. Kinetic parameters were obtained from nonlinear regression analyses of hyperbolic velocity versus [S] plots.
Zinc Inhibition of RPTP␤-RPTP␤ was assayed for enzymatic activity in the presence of free zinc(II) ions. Zinc was buffered in the pico-and nanomolar range (0.3 pM-1.7 M) using 1 mM NTA containing 0.1 M to 1 mM ZnSO 4 . Zinc was added to the buffer prior to enzyme and substrate. Free zinc(II) ion concentrations were calculated using the software programs Maxchelator (27) and MINEQL version 4.6 (Environmental Research Software, Hallowell, ME). PTP activity at each free zinc(II) ion concentration was compared with the maximum activity determined in the presence of 1 mM NTA. Data were plotted semilogarithmically and fitted to Hill's equation (20) using SigmaPlot version 12 (SYSTAT Software Inc.). Zinc inhibition at 0.8 M FDP was measured at three different enzyme concentrations (0.5, 2, and 5 nM). The data for the Dixon plots were obtained by single readings at each substrate and inhibitor concentration.
Determination of Protein and Metal Concentrations-Protein was quantitated spectrophotometrically (Synergy HT plate reader) using the Bio-Rad protein assay reagent (Bio-Rad) with bovine serum albumin as a standard, but using a modified Bradford assay with a 10-fold increased sensitivity (28). Inductively coupled plasma mass spectrometry (Perkin-Elmer Life Sciences, model Elan 6100 DRC plus) analysis was performed on all components of the assay to determine whether zinc contamination of reagents affects the free zinc(II) ion concentrations in the buffer system. Samples were diluted in a minimum volume of 1 ml of HNO 3 (1 or 5% v/v) in 15-ml polypropylene tubes. The final concentrations were the same as those used in the assay.

RESULTS
PTP Assays-A major issue when assaying PTPs, and a confounding factor in published data, is the control of the concentrations of metal ions and the redox state of the enzymes. The catalytic cysteine of PTPs is prone to oxidation and rapid inactivation. PTPs are prepared and supplied with rather high (mM) concentrations of the chelating agent EDTA and a reducing agent, typically DTT, which is also a relatively strong zincchelating agent (29). Removal of the agents results in rapid loss of activity. It is virtually impossible to dilute enzyme solutions in such a way that the agents do not interfere with assays of metal binding. Zinc is a ubiquitous contaminant of laboratory chemicals and biochemicals. Metal contamination in chemicals is usually given as ppm (parts per million). However, 1 ppb (parts per billion) corresponds to 15 nM zinc, a concentration that inhibits PTP-1B 50% and is clearly too high for investigating zinc inhibition of PTPs. Measuring zinc in the reagents used in the assay demonstrated that all the chemicals contained zinc at concentrations greater than 1 ppb. Notably, reagents, such as the substrate FDP, are usually not thought to be a source of zinc contamination and thought to be free of inhibitors (26). However, FDP has zinc concentrations far exceeding the concentrations that would cause 50% inhibition of PTP-1B (27.5 ppb at 0.42 M FDP). The RPTP␤ solution contains at least stoichiometric amounts of zinc. It was not feasible to remove zinc from all the solutions to control its concentrations at the sub-ppb level in the assay, although substantial efforts were made to avoid and control metal contamination. A solution to this problem is the use of metal buffering techniques, measurement of zinc concentrations in all solutions, and subsequently inclusion of measured zinc concentrations from the reagents in the calculations of free zinc(II) ion concentrations in the zinc-buffered solutions.
PTPs have pH optima at slightly acidic pH (30). Because protons compete with zinc at zinc-binding sites, enzymatic activity should not be measured at the pH optimum when investigating zinc inhibition. Also critical in the experiments is the choice of the reducing agent to protect PTP from oxidation and to maintain its enzymatic activity. Standard reducing agents, such as DTT, glutathione, or 2-mercaptoethanol, are thiols and are not suitable because they are also chelating or zinc-binding agents. Therefore, we used TCEP as a reducing agent with low zinc binding capacity (31). We chose HEPES as a pH buffer because it has a low complexing capacity for zinc at pH 7.4 (32). As a metal buffering system, we used NTA, which has a logKЈ of 8.3, at a sufficiently high zinc buffering capacity of 1 mM (33). NTA forms 1:1 and 2:1 complexes with zinc (34). Including the 2:1 complex in the calculations does not change the free zinc(II) ion concentrations significantly. Both TCEP and NTA are also a source of zinc with 16 and 5 ppb zinc at 244 and 76 nM, respectively. These concentrations were considered in the calculations of free zinc(II) ion concentrations.
PTP Inhibition-3,6-Fluorescein diphosphate was used as a PTP substrate because it allows assaying the enzyme at sub-nanomolar concentrations. Using this substrate and the optimized assay conditions, we found that zinc inhibits RPTP␤ with a K i(app) value of 98 Ϯ 9 pM (Fig. 1A). The inhibition is 5 orders of magnitude stronger than previously reported (50% inhibition between 1 and 25 M zinc) (25). The zinc inhibition is reversible as adding EDTA to the assay restores RPTP␤ activity (Fig. 1B).
The enzyme follows Michaelis-Menten kinetics with a K m value of 0.21 Ϯ 0.02 M for FDP and a V max value of 27.5 Ϯ 0.8 nM/min ( Fig. 2A). If substrate competes with zinc binding, the kinetically determined inhibition constant at a fixed substrate concentration (Fig. 1A) reflects an apparent K i only. Therefore, a kinetic analysis of zinc inhibition was performed for RPTP␤ by varying the substrate concentration below and above the K m value for FDP. Zinc indeed competes with substrate. Hence the zinc inhibition constant in the absence of substrate is even lower than the estimated value of 98 pM determined at about 4ϫ K m (FDP). Analysis of the data in the form of a Dixon plot (Fig. 2B) gives a K i value of 21 Ϯ 7 pM. The Dixon plot does not discriminate between purely competitive and mixed inhibition. When we analyzed the data with the SigmaPlot kinetics pro-gram, both competitive and mixed (intersecting, linear noncompetitive) inhibition passed the statistical tests. In both models, the slope effect, representative of the substrate competing with the inhibitor for the enzyme, is virtually the same (K i values of 21 and 28 pM). The small intercept effect reflecting inhibitor binding to the enzyme-substrate complex with a 14-fold weaker affinity needs confirmation with other substrates. In summary, using enzymatic activity has been the only feasible approach to investigating zinc inhibition of PTPs, but it has the inherent limitation that the substrate affects the measured zinc inhibition. Only a full kinetic analysis provides the correct inhibition constant for zinc(II) ions.

DISCUSSION
Zinc inhibits PTPs much more tightly than generally appreciated. Methodological issues have precluded observing the tight zinc inhibition of these important enzymes. To maintain activity, they are prepared and supplied with high concentrations of chelating agents, such as EDTA, reducing agents, such as DTT and/or glutathione, and pH buffers, such as Tris, all of which have metal-chelating capacity. These conditions are not  conducive for investigating metal binding. Critical for studies of metal inhibition is the use of buffers and a reducing agent with low zinc-complexing capacity and control of metal ion concentrations by metal buffering. When controlling these parameters, using a sensitive fluorometric assay, and performing a kinetic analysis of the inhibition, a remarkably strong zinc inhibition of RPTP␤ is revealed.
The strong zinc inhibition indicates physiological rather than pathophysiological or toxicological significance with regard to recent studies on cellular free zinc(II) ion concentrations and their regulation (10,13). The distinction between free and total zinc concentrations is not widely appreciated, and many investigations claim physiological significance at micromolar zinc(II) ion concentrations because total cellular zinc concentrations are a few hundred micromolar. However, this concentration of total cellular zinc is strongly buffered by cytosolic proteins binding zinc with high affinity, and thus, the free zinc(II) ion concentration and not the total zinc concentration is the important parameter to be considered for physiological significance. Investigators have employed different experimental conditions, zinc-chelating fluorophores, and methods, such as quantitation with a fluorescence microscope or a spectrofluorometer, and demonstrated that free zinc(II) ion concentrations in various cultured cells are in the range of tens to hundreds of picomolar (10 -12, 14). Critical for the physiological significance of zinc regulation of PTPs is the recent interest in a role of cellular zinc(II) ions in information transfer and regulation. Free zinc(II) ions fluctuate in the picomolar range of concentrations, and these fluctuations are influenced by various physiological stimuli (13,(35)(36)(37).
RPTP␤ is a receptor-type PTP, also known as also known as vascular endothelial PTP (VEPTP), that controls the phosphorylation state of the angiopoietin receptor Tie-2. Its expression is restricted to endothelial cells (38,39). Mice lacking the RPTP␤ gene die on embryonic day 10 because of effects on angiogenesis in the yolk sac and in the embryo, but not on vasculogenesis (39,40). It is well known that zinc is important for wound healing, but the molecular mechanism is not known (41). The results indicate a molecular basis for zinc in controlling growth arrest and maintenance of quiescent cells in angiogenesis, which occurs in the proliferative phase of wound healing. The only other instance where such a strong zinc inhibition has been observed is erythrocyte Ca 2ϩ -ATPase, which zinc inhibits with an IC 50 of 80 pM (42). Because the free zinc(II) ion concentration in erythrocytes is 24 pM and indeed in the range of this inhibition (43), the data also suggest physiological zinc inhibition of an enzyme that is not known to be a zinc enzyme.
Zinc inhibition of PTPs within the range of resting zinc(II) ion concentrations in the cell raises the possibility that zinc is a tonic inhibitor of some of these enzymes. Such inhibitory control has been postulated as one of the main functions of zinc in biology (44). A way of activating zinc-inhibited enzymes is the reaction with thionein, the apo-form of metallothionein (17). The redox state of the cell is one of the main regulators of zinc loading onto metallothionein. Given that cellular zinc(II) ions are not freely available, the induction, extensive regulation, and changing zinc load make metallothionein/thionein a candidate for controlling zinc activation and zinc inhibition of enzymes (45).
The structural basis of zinc inhibition of PTPs is not known. It could involve the sulfhydryl group of the catalytic cysteine (residue 1094 in RPTP␤) as a ligand for zinc, the general acid/ base aspartate (residue 1870), and/or an adjacent histidine (residue 1871). The latter two residues are on the conformationally flexible WPD loop (46).
Zinc ions are much better inhibitors than the classic vanadium complexes or any other drug. Experimental drugs bind with micromolar affinity, but exploitation of another substrate-binding site has resulted in nanomolar inhibition (47). Proper design of the ligand environment of zinc complexes or modulation of cellular zinc availability could provide additional avenues in the intensive ongoing efforts to develop therapeutic drugs for these important enzymes. The remarkable zinc inhibition also raises questions of how perturbed cellular zinc homeostasis affects major cellular signaling pathways.