Redox regulation of cardiac protein kinase C

Abstract

The two extremes of redox stress imposed on cardiac tissue under ischemia and reperfusion change the redox potential of the cells and affect numerous redox-sensitive molecules, including the ones involved in intracellular communication. Protein kinase C (PKC), a key signalling kinase, is one of those subject to redox control. Activation of PKC by oxidation represents a new paradigm of the alternate signalling principle. Reactive oxygen species act directly on PKC, releasing chelated Zn²⁺ ions from the zinc finger of the regulatory domain. Zn²⁺ release from PKC by oxidative stress has been shown at the level of isolated protein fragments, PKC immune complexes and single cells. Zn²⁺ movements have been further characterized in cryosections prepared from adult rat hearts subjected to in vivo stress by global ischemia followed by reperfusion. The morphology of labile zinc in cardiac tissue and zinc release following PKC stimulation with lipid activator are described. The studies lead to an unexpected and intriguing result, suggesting that in addition to serving a structural function, Zn²⁺ ions are likely to play a dynamic regulatory role in PKC. The cysteine-rich domains of the serine/threonine kinases are identified as redox sensors. Thus, being an integrated composite of redox signalling systems, free Zn²⁺ reflects the protein redox status and serves as a valid biomarker of stressed tissue and its capacity to respond to stimuli.

The area of redox signalling research is exponentially growing due to the vital role reactive oxygen species (ROS) play in both physiology and in the development of pathological conditions. ROS are produced as a part of physiological signalling by hormones (1,2), lymphokines (3,4) and growth factors (5,6), and thus, are considered by many researchers to be second messengers. The generation of ROS by ligand-activated membrane receptors can be mediated by triggering NADPH oxidase (7–9), the mitochondrial respiratory chain (10,11) or hemoproteins that sense oxygen tension (12). In addition, ROS can also be produced by stressing physical and chemical sensors, such as shear flow (13–15) or endoplasmic reticulum overload (16,17). However, under conditions of metabolic or exogenous stress (eg, mitochondrial dysfunction, inflammation, the presence of toxins and irradiation), ROS formation may be excessive and lead to triggering of the signalling cascade, leading to cell death (18–20). In general, oxidative stress may be perceived as an imbalance between ROS production and the state of glutathione redox buffer and antioxidant defence. In the present report, experimental data are presented that validate another component that plays a decisive role in the equilibrium, ie, the redox status of individual redox-sensitive molecules. The past several years have been marked by reporting the effects of oxidative stress on different proteins, including members of classic signalling networks. The list includes but is not limited to ion channels, serine/threonine kinases, receptor and Src family tyrosine kinases, phosphatases and transcription factors (21). However, despite all the accumulated descriptive knowledge, the molecular mechanism of the possible direct action of ROS on the signalling proteins has not been uncovered. The present report suggests a possible mechanism of redox sensing by zinc finger-containing proteins by using the example of serine/threonine kinases. Furthermore, the report discusses how oxidative stress affects zinc fluctuations in myocardial tissue.

REDOX ACTIVATION OF PROTEIN KINASE C REQUIRES COFACTORS

Protein kinase C (PKC) (22,23), a key signalling kinase, is an enzyme that is subject to redox control. The ability of PKC to regulate many cardiovascular functions is supported by the fact that many cardiovasotropic growth factors (eg, angiotensin, endothelin, and vascular endothelial growth and permeability factor) target PKC (24–28). The physiological importance of PKC can be surmised by the existence of its multiple isoforms, which are usually arranged into groups according to their structure and cofactor requirements. Conventional PKCs (alpha [α], beta_1/2 and gamma) are Ca²⁺-dependent and activated by diacylglycerol (DAG); novel PKCs (delta [δ], epsilon [ε], eta and theta) are Ca²⁺-independent but activated by DAG; atypical PKCs (zeta [ζ] and iota/lamda) are Ca²⁺- and DAG-independent (25,29–31). It has been shown that restricted physiological functions are limited to specific isoforms. For example, cardiac PKC-δ is activated by ischemia and PKC-ε is involved in ischemic preconditioning linking PKC function to redox control (26,27,30,31). Experimental evidence suggests that, independent of the classic pathway, PKC is also controlled by a redox mechanism, ie, oxidation converts the protein to a catalytically competent form (32–34), whereas reduction reverses this process (35). In addition, it has been found that redox activation of the kinase requires cofactors (ie, retinol or its metabolites). Bound retinol causes no intrinsic PKC activation on its own, nor does it affect the PKC function triggered by growth factor or phorbol ester. However, redox activation is significantly more effective in the presence of retinol (35). This link between retinol effect and redox regulation is not surprising because retinoids have long been known to possess both pro-oxidant and antioxidant properties, with the former dominating in PKC catalysis.

IMAGING OF INTRACELLULAR FREE Zn²⁺ TRIGGERED BY PKC ACTIVATORS

Redox chemistry involves electron exchange between molecules with different oxidation (redox) potential. One manner in which such preferential oxidation may occur is by the facilitation of electron transfer by retinol. Because the high-affinity retinoid binding site has been mapped to cysteine-rich regions in the regulatory domains of the serine/threonine kinases Raf and the isoforms of PKC (35), one could assume that these 50 amino acid-long highly homologous stretches containing six conserved cysteine and two conserved histidine residues tetrahedrally coordinated by two Zn²⁺ ions into a composite zinc finger (36,37) would be targets of redox control (Figure 1A). In this scenario, oxidation of even one of the cysteines would compromise the chelation of Zn²⁺ ions that could be readily captured by Zn²⁺-sensitive probes. The assays based on Zn²⁺ release would allow for the deduction of the direct action of ROS on PKC molecules. The search for specific and sensitive probes led to the use of the membrane-permeable Zn²⁺-sensitive probe N-(6-methoxy)-8-quinolyl-toluene sulfonamide (TSQ). TSQ was first used by Frederickson et al (38) for assessment of zinc levels in frozen brain slices. This highly specific probe is sensitive to nanomolar doses of Zn²⁺.

Figure 1).

Imaging of intracellular Zn²⁺ release. A Hypothetical model of Zn²⁺ release from the zinc finger of protein kinase C (PKC). The oxidation of even one of the three cysteines will compromise chelation of tetrahedrically coordinated Zn²⁺, which can then be readily captured by a corresponding selective Zn²⁺ probe. Retinol serves as a bridge to transfer electrons (e⁻) to sensitive cysteines. B Zn²⁺ release in neonatal rat cardiac myocytes visualized with N-(6-methoxy)-8-quinolyl-toluene sulphonamide (TSQ): (1) TSQ-loaded cardiomyocytes; (2) same field after 2 min treatment with 0.1 μM phorbol 12-myristate 13-acetate (PMA). The bar graph shows the mean ± SD for fluorescence intensity acquired from three sections. The increase in TSQ fluorescence was 30% to 50% above the basal value and was reproducible in every experiment, triggered either by oxidation or by phorbol ester application. C Real-time confocal imaging of Zn²⁺ release in insect cells overexpressing PKC-alpha. The images (adapted from reference 39) were acquired serially at 35 s intervals. Every second image of PKC-alpha-expressing cells is presented from a1 to g2. PMA (0.1 μM) was added before acquisition of the c1 image. The TSQ fluorescence increase was eight- to 10-fold higher than that in vector-infected cells. GSH Glutathione; TPEN N,N,N’,N’-tetrakis(2-pyridylmethyl)ethylenediamine

Figure 1B shows previously unpublished data on Zn²⁺ release in neonatal cardiomyocytes treated with phorbol 12-myristate 13-acetate (PMA) and with H₂O₂. Substantial amounts of Zn²⁺ could be detected within minutes of treatment, as visualized by confocal microscopy with TSQ as an indicator (Figure 1B). The selective Zn²⁺-chelating agent N,N,N′,N′-tetrakis(2-pyridylmethyl)ethylenediamine quenched the TSQ fluorescence, indicating Zn²⁺ specificity. The increase in TSQ fluorescence was 30% to 50% above the basal value and was reproducible in every experiment, triggered either by oxidation or by phorbol ester application. The latter indicated that the source of free Zn²⁺ may have been PKC. To trace Zn²⁺ release to PKC, similar experiments were performed in Trichoplusia ni insect cells infected with baculovirus constructs that harboured either the full-length human PKC-α or PKC-ζ genes, or an empty vector (39). PMA produced a strong increase in fluorescence in the cells that overexpressed PKC-α but not in the cells that expressed PKC-ζ or the empty vector (Figure 1C). In the PKC-α-producing cells, the TSQ fluorescence increase was eight-to 10-fold. Both the PKC-α and the PKC-ζ isoforms responded to H₂O₂. Thus, overexpression of PKC resulted in increased Zn²⁺ release triggered by both stimuli. These results suggest that zinc plays a dynamic role in cell function, and that PKC or PKC-triggered processes can be a source of free intracellular Zn²⁺.

Zn²⁺ RELEASE FROM THE PURIFIED PROTEIN FRAGMENTS — ZINC FINGERS

To strengthen the finding that PKC was the principal source of liberated Zn²⁺, and to trace Zn²⁺ release to the zinc finger domain, an in vitro evaluation of glutathione S-transferase fusion proteins, containing human PKC-δ C1A, PKC-δ C1B, PKC-ζ C1 or cRaf C1 peptides, was performed using a spectrofluorimetric quantitative assay with TSQ as the probe (39). PKC-α and PKC-δ cysteine-rich domains released Zn²⁺ after PMA treatment, whereas PKC-ε and cRaf cysteine-rich domains did not. However, all proteins dissociated Zn²⁺ after H₂O₂ treatment. Detailed titrations of the effect are presented for the PKC-δ C1B peptide (Figure 2A). Both 1,3-diolein (Sigma, USA) and PMA triggered the release of stoichiometric amounts of Zn²⁺, with each mole of lipid generating one equivalent mole of free Zn²⁺. Saturation was reached when approximately one-half of the available Zn²⁺ was released. The thiol-reactive compound p-hydroxy-mercuri-phenylsulphonate (PMPS) (40) was used to determine the total Zn²⁺ content in the glutathione S-transferase fusion protein. Oxidation by high-dose H₂O₂ also led to Zn²⁺ release from the protein that, as with PMA treatment, reached saturation within a few minutes, but released almost twice as much Zn²⁺. Using the cRaf C1 domain as a control protein not known to bind PMA (41), no Zn²⁺ release with PMA treatment was observed, whereas oxidation was effective (39). The results achieved with TSQ fluorimetry were confirmed by a commonly used colorimetric assay with 4-(2-pyridylazo)-resorcinol. To clarify whether PMA and oxidation with H₂O₂ targeted the same or different Zn²⁺ coordination centres, these two agents were tested alone and in combination. The results confirmed that each agent displaced one-half of the protein-bound Zn²⁺, whereas combined action caused total Zn²⁺ release. In both instances, these experiments suggest that the inactive kinase binds Zn²⁺ with an affinity exceeding that of the probes TSQ and 4-(2-pyridylazo)-resorcinol. On oxidation or PMA binding to the cysteine-rich domain, the chelation status of Zn²⁺ is changed, allowing Zn²⁺ to partition to the probes. The concept that zinc movements are important for PKC to gain enzymatic capacity has been validated by examining the Zn²⁺ contents in resting versus active PKC, with the expectation that the former should contain more Zn²⁺. Zinc content correlated inversely with PKC activity (39). These results are in complete agreement with those of Knapp and Klann (34), who had previously underscored the importance of Zn²⁺ release during PKC activation by the redox mechanism.

Figure 2).

Zn²⁺ release from the protein kinase C (PKC) zinc finger in vitro. A Dose dependences of Zn²⁺ release from the PKC-delta C1B peptide. Both 1,3-diolein and phorbol 12-myristate 13-acetate (PMA) triggered the release of stoichiometric amounts of Zn²⁺, with each mole of lipid generating one equivalent of free Zn²⁺. B Additive effect of PMA and H₂O₂ on Zn²⁺ release as measured by the 4-(2-pyridylazo)-resorcinol assay. PMPS p-Hydroxy-mercuri-phenylsulphonate. Adapted from reference 39

LABILE ZINC IN RAT HEART TISSUE SECTIONS

To show that the phenomenon of Zn²⁺ release takes place in vivo and to show the effect of stress on Zn²⁺ release, I have developed a new approach to investigate functional zinc in Langendorff-perfused hearts. Adult rat hearts subjected to 15 min global ischemia followed by 20 min reperfusion and control hearts subjected to Langendorff retrograde perfusion were quickly frozen using a standard protocol to preserve tissue elements (42). Briefly, the whole hearts, soaked in OCT (10.4% polyvinyl alcohol and 4.26% polyethelene glycol), were transferred to isopentane precooled in liquid nitrogen, and stored after freezing on dry ice. Cryosections were obtained and placed onto a dish with a glass bottom for confocal measurements. The thawed sections were immediately loaded with either TSQ or ZinPyr-1 (ZP-1) (NeuroBioTex, USA), a fluorescent probe synthesized on the basis of fluorescein by Burdette et al (43), and the experiments were further conducted as previously described for use with isolated cells (39). Control experiments revealed that OCT and isopentane did not alter subsequent loading of the tissues with fluorescent probes, nor did these procedures affect tissue response to stimuli.

Figure 3 summarizes the major findings from the experiments described above. Nomarski images in Figure 3A show a regular pattern of rod-shaped cells with a sarcomeric ultra-structure characteristic of cardiomyocytes revealed on the level of single cells. TSQ fluorescence originated from sarcomeric units, cell periphery and intercalated disks, where it was most prominent. Both sarcomeric staining and decoration of intercalated disks were confirmed by use of ZP-1 staining. In tissue sections obtained from hearts subjected to ischemia/reperfusion (I/R), sarcomeres were also distinguishable; in addition, areas with irregular morphology were observed, most likely the result of damaged cells. Because the quinolyl derivative TSQ proved to be the most sensitive probe for free Zn²⁺, it was used to compare stressed with control myocardial tissues. In I/R tissues, TSQ-decorated zones of cell contacts were found, as well as patchy areas containing granulated vesicle-like structures, reminiscent of the vesicle-associated zinc accumulation seen in neurons (44). Overall, fluorescence intensity was significantly lower in oxidatively stressed hearts than in control hearts (Figure 3B). PMA treatment for 3 min significantly increased TSQ fluorescence of tissue sections obtained from control hearts (90% increase) but not from hearts after I/R (14% increase). Intracellular Zn²⁺, however, could still be released by treating the latter tissue further with the thiol-active agent PMPS (data not shown). Fluorescent signals were quenched by the Zn²⁺ chelator N,N,N′,N′-tetrakis(2-pyridylmethyl)ethylenediamine. Quantitative data are summarized in Figure 3C. Each value represents the mean ± SD obtained from three sections. The results suggest that heart tissue becomes zinc-depleted after I/R. More important, the capacity to liberate labile zinc is significantly diminished. Altogether, the data show technical feasibility of the new approach to investigate functional zinc and its movements in these in vivo models, including human myocardial tissue.

Figure 3).

Functional Zn²⁺ in rat myocardial tissue sections. A Morphology of free Zn²⁺ in Langendorff-perfused adult rat hearts. Nomarski images (left) show cell morphology characteristic of cardiomyocytes. The upper right panel shows N-(6-methoxy)-8-quinolyl-toluene sulphonamide (TSQ) fluorescence and the lower right panel shows ZinPyr-1 (NeuroBioTex, USA) fluorescence from corresponding sections. B Cryosections obtained from control hearts (top) and hearts after 15 min global ischemia followed by 20 min reperfusion (bottom) were treated with 100 nM phorbol 12-myristate 13-acetate (PMA) while on the stage of the microscope. Fluorescence intensity was recorded online at 630× original magnification. C Quantitative analysis of TSQ fluorescence intensity in tissue sections. The bar graph shows the mean ± SD for fluorescence intensity acquired from three sections. A decrease in fluorescence intensity after N,N,N′,N′-tetrakis(2-pyridylmethyl)ethylenediamine (TPEN) treatment proved the specificity of TSQ as a reporter for zinc. D Oxidation of cysteines in protein kinase C (PKC) isoforms isolated from adult rat hearts after ischemia/reperfusion (I/R). Lane 1 shows stabilization, lanes 2 and 3 show ischemia, and lanes 4 and 5 show reperfusion. Although cell death of myocytes usually occurs after I/R, cysteines of both PKC isoforms (PKC-delta [PKC-δ] and PKC-epsilon [PKC-ε ]) were oxidized under ischemia alone. PMPS p-Hydroxy-mercuri-phenylsulphonate

CONCLUSIONS

The two extremes of redox stress imposed on cardiac tissue under I/R change the redox potential of the cells and affect numerous redox-sensitive molecules, including the ones involved in intracellular communication. As it has been shown, the key mechanism of the alternate signalling principle may link to zinc movements within the protein molecule. My colleagues and I were among the first to establish that ROS directly act on PKC, releasing chelated Zn²⁺ from the zinc finger of the regulatory domain (39). Our studies suggested the unexpected and intriguing result that, in addition to serving a structural function, Zn²⁺ likely plays a dynamic regulatory role. Also, it was confirmed that the activated form of native PKC contained significantly less Zn²⁺ than the resting form. The evidence obtained from these studies clearly defines the cysteine-rich domain as a redox sensor and a reversible redox switch. We have also shown that lipid activators cause, in principle, the same effect (eg, the release of Zn²⁺), thus stressing the importance of Zn²⁺ in PKC function (39). Figure 4 shows a model that underscores the convergence of the two signalling pathways.

Figure 4).

Hypothetical model of convergence of two signalling pathways of protein kinase C activation. Due to oxidation of cysteines, or by affecting the chain of hydrogen bonds (in the case of phorbol 12-myristate 13-acetate), the two different stimuli cause principally the same effect, ie, the release of Zn²⁺ ions (38). Loss of even one of these stimuli would destabilize the C3H1 zinc finger and allow precisely the kind of conformational changes that have been postulated by others for the initiation of protein kinase C catalytic activity. These changes involve removal of the regulatory N-terminus from the catalytic domain, exposure of hydrophobic surfaces that facilitate translocation to membranes and accessibility to substrate and cofactors (47,48). Cat Catalytic site; DAG Diacylglycerol

The redox regulation of signalling kinases appears to involve conformational changes in the regulatory zinc finger. The hypothesis is that far from cementing rigid structures, zinc is actually the linchpin that orchestrates dynamic changes in response to specific signals, allowing kinase activity to be turned on or off. The function of zinc finger structures as redox-regulated, reversible hinges has been described for the bacterial chaperone Hsp33 (45). In an oxidizing microenvironment, thiols are converted to disulphide, Zn²⁺ becomes uncoupled and unfolding of the protein leads to enzyme activation (45,46). This reversible activation could be an adaptive mechanism in response to stress, whereas irreversible overoxidation could trigger protein degradation and, consequently, apoptosis.

Detailed knowledge of the mechanisms that control PKC function under the conditions of oxidative stress and the potential targets of redox control will help to direct future therapeutic modalities for I/R and related diseases. In the equilibrium of oxidative stress, the redox status of individual molecules (eg, cysteine-rich domain status and zinc content) may become the most valid biomarker of stressed tissue and its capacity to respond to stimuli. Possibly the most effective protection from ischemic damage, preconditioning sets up the redox switch on the regulatory molecules. I believe that better knowledge of the operation of this switch will help to find clues to the mechanism of such protection.