Identification of Peroxiredoxin 1 as a novel interaction partner for the lifespan regulator protein p66Shc

Introduction

Aging is usually accompanied by diseases like cardiovascular dysfunctions or cancer [1], indicating that they are caused by related molecular processes [2-4]. A mechanism recognized to contribute to many aging processes and accompanying diseases is the accumulation of damage caused by reactive oxygen species (ROS) [5,6], which are mainly formed by the mitochondrial respiratory chain. Aging tissues exhibit elevated levels of oxidized lipids and proteins [5] and show increased expression of stress response genes [7]. Consistently, lifespan extension through Sirtuin activation or caloric restriction appears to be due to an increase of stress resistance and the alleviation of diseases [6,8,9].

Under normal conditions, overall cellular ROS levels are kept low by antioxidants, such as glutathione (GSH). Besides their destructive role, however, ROS also act as signaling mediators [10,11]. They can, e.g., initiate apoptosis by inducing permeability transition of mitochondrial membranes [12], resulting in cytochrome c (Cyt c) release and apoptosome activation [13]. ROS can represent oxidative stress upstream of this step, but cellular ROS levels can also be actively increased after pro-apoptotic stimuli [6]. ROS can thus establish a positive feedback loop, as described for the oxidant-induced activation of p53 that leads to ROS formation [6,14,15]. Consistently, antioxidant treatment protects cells from apoptosis not only if initiated by oxidants, but also if initiated by a wide variety of stressors and stimuli [16].

p66Shc appears to be a central player in stress-induced apoptosis and ROS amplification, in diseases caused by dysfunction of this system, and in life-span regulation [17,18]. Knocking out the shcA (Src homologous and collagen A) gene coding for p66Shc results in decreased ROS levels and a 30 % extended lifespan of rodents [19]. p66Shc has been implicated in promoting cancer cell growth and in aging-associated arterial dysfunctions [2,20], and has thus been suggested as a novel drug target [21]. p66Shc is the largest polypeptide encoded by the shcA locus, besides p46Shc and p52Shc. All isoforms share a phosphotyrosine-binding (PTB), a collagen homology (CH), and a Src homology 2 (SH2) domain; p52Shc shares an additional Cytochrome c binding domain (CB) with p66Shc, which carries a unique CH2 domain at its N-terminus. The mainly cytosolic isoforms p46Shc and p52Shc act as adaptor proteins regulating, e.g., elements of the Ras signaling pathway [18]. In contrast, p66Shc is located in the cytosol and the mitochondrial intermembrane space (IMS) [22-24]. Upon induction by stress factors, expression of the p66Shc protein increases [25], existing p66Shc is stabilized [26], and cytosolic p66Shc gets phosphorylated by protein kinase Cβ (PKCβ, followed by an interaction with the prolyl isomerase Pin1 and translocation to the IMS [24]. In this study, p66Shc has been suggested to directly form H₂O₂ [22], which is assumed to activate the permeability transition pore (PTP), ultimately causing rupture of mitochondria. Consistently, overexpression of p66Shc increases ROS level [27] and p66Shc-/- fibroblasts are resistant to apoptosis induced by several stressors [19,28,29]. Thus, p66Shc acts as sensor for ROS, but also for other stress factors, and on the other hand increases ROS formation leading to apoptosis initiation, presumably under overwhelming stress when repair systems cannot cope with the ROS anymore. However, the N-terminal p66Shc domain responsible for apoptosis induction and H₂O₂ formation could be shown to be a better apoptosis inducer in an oxidized, tetrameric form, but to produce more ROS in its reduced, dimeric form [30], indicating that the molecular details of p66Shc-mediated apoptosis are still not fully understood.

We showed previously that p66Shc can interact with thioredoxins (Trx), which reduce and thereby inactivate the N-terminal domains, CH2 and CB (p66CH2CB), that carry the apoptosis inducing and H₂O₂-forming activity of p66Shc [30]. Another prominent family of proteins contributing to redox metabolism and signaling are peroxiredoxins (Prx) [31]. Prx degrade H₂O₂ using a conserved cysteine residue. Mammals have six isoforms, Prx1-6, divided into classes depending on whether they exhibit just the catalytic (peroxidasic) cysteine or also an additional, resolving cysteine, and whether both cysteines are on the same polypeptide [32]. Prx are weak scavengers, but due to their abundance and broad substrate specificity they effectively degrade peroxides at low concentrations. Increased H₂O₂ levels, in contrast, inactivate Prx through overoxidation of the peroxidasic cysteine, making them ideal sensors for H₂O₂-mediated signaling [31]. Many Prx further exist in different oligomeric forms, a dimeric form with significant peroxidase activity, which was proposed to be stabilized by oxidation at the catalytic cysteine, and an oligomeric form with low peroxidase activity that appears to act as chaperone. Furthermore, Prx can interact with other proteins apparently for regulating their activity [31], such as the inhibitory interaction of Prx1 with the apoptosis signal-regulating kinase 1 (Ask1) [33].

Here, we describe the identification of Prx1 as a novel interaction partner for the unique N-terminal domain of p66Shc, and the biochemical characterization of this complex. Prx1 was identified in affinity experiments as dominant interaction partner. Complex formation leads to disassembly of the decameric form of Prx1, which is known to increase its peroxidase activity. The interaction leads to reduction of the p66CH2CB tetramer, which reduces its ability to induce mitochondrial rupture. These results indicate that p66CH2CB and Prx1 form a stress-sensing complex that keeps p66Shc inactive as long as stress levels remain moderate.

Results

Discussion

p66Shc is a key player in stress sensing and the stress-induced mitochondrial apoptosis pathway. It acts as both, stress sensor and inducer of ROS formation, in a complex regulatory circuit [18,38]. Despite a wealth of in vivo observations, several molecular mechanisms connecting p66Shc and other players of this network remain to be fully understood, such as the exact mechanism of p66Shc-dependent ROS generation and the mechanism of ROS-dependent permeability transition. Such an understanding would also support efforts to exploit p66Shc as a therapeutic target. We reasoned that identifying novel interaction partners of the p66Shc-specific N-terminal domain could reveal insights into the regulation of its deadly activity. We could indeed show here that p66CH2CB interacts with the peroxiredoxin family of antioxidant enzymes that also contribute to signal transduction, e.g., of the p38 protein kinase [31,39]. In particular, our results show that p66Shc can interact with Prx1, and possibly with Prx3, in a physiological environment and in vitro. We observed a dominant enrichment of Prx1 in affinity experiments, and a much weaker signal from Prx3 as well as the characterized p66Shc interaction partner Hsp70 [28], highlighting the value of our mass spectrometry approach which allowed the identification of an unexpected and dominant but so far overlooked interaction partner.

Prx can act as H₂O₂ scavengers, but many studies have shown that they are also major contributors to a complex ROS signaling system [31,40], most prominently in apoptosis initiation [41-43]. A Prx3 knock-down shows increased sensitivity to apoptosis inducing signals [44], and Prx1 is involved in signaling by the p38 kinase [39] and acts as tumor suppressor [45,46]. Prx1 is mainly found in the cytosol, although a recent determination of the mitochondrial proteome also identified Prx1 in this organelle [47]. The fact that Prx1 was more prominent in affinity experiment eluates than Prx3 even when mitochondrial lysates were used, which should contain more Prx3, indicates a pronounced preference of p66Shc for the Prx1 isoform. In fact, Prx3 is mainly located in the mitochondrial matrix, whereas the mitochondrial IMS, where p66Shc appears to induce apoptosis [22], is generally assumed to be strongly connected to the cytosol. Thus, Prx1 appears to be the major isoform encountered by p66Shc and seems to be bound preferentially, even in presence of other peroxiredoxins, indicating a pronounced level of specificity for this interaction. The fact that even a cysteine-lacking mutant of p66CH2CB could bind to Prx1 and disassembles its decamer form also reinforces the notion that the interaction is rather specific and does not resemble the redox interactions often encountered with thioredoxins, which appear to keep surface cysteines of cellular proteins reduced with little specificity. However, the observation that Prx1 reduces and thereby inactivates p66CH2CB indicates that Prx1 might functionally act like a Trx in this system. Protein/protein interactions have been described for Prx and are assumed to contribute to their function, in addition to H₂O₂ scavenging [31]. A regulatory role for an interaction with Prx1 has been described before for Ask1 [33], for example. In this system, the interaction rather than a redox reaction with Ask1 appears to be relevant for the apoptosis inhibiting effect of complex formation. It remains to be shown for the p66Shc/Prx1 complex whether purely forming the interaction, or the reciprocal reduction/oxidation, or both processes are relevant for its physiological function. At present we hypothize that both elements contribute to making this complex of two intrinsic ROS sensors [30,31] a highly regulated ROS sensing system (Figure 4): Complex formation keeps both players close in space, able to exchange information. The Prx, an efficient H₂O₂ scavenger at low concentrations [31], will protect cells from moderate oxidative stress, from external sources or formed by p66Shc. It will thereby keep p66Shc reduced and unable to induce apoptosis. The peroxidase activity of Prx1 will be inactivated through overoxidation, however, if excessive oxidative stress is encountered. This szenario would favor Prx1 decamerization, p66Shc release and oxidation as well as release of additional ROS by p66Shc now lacking its ROS-scavenging partner. This change will finally lead to the ultimate cellular stress response to overwhelming stress, apoptosis induction.

Prx show a versatile quaternary structure, apparently corresponding to different activities and or activation states [32,48]. Two interaction interfaces of a Prx monomer contribute to oligomerization of Prx of the A-type, which comprises mammalian Prx1 and Prx3 [32]. The first one is oriented parallel to the central β-sheet and mediates formation of a Prx1 dimer as its stable building block. A second one is oriented perpendicular to the central β-sheet and enables higher oligomerization of dimers. A "fully folded" conformation of the peroxidasic cysteine loop is observed in the reduced Prx1 form and stabilizes the perpendicular interface, promoting oligomerization. The "locally unfolded" conformation of the oxidized form destabilizes this interface and thus the decamer. The disassembling effect of p66CH2CB on Prx1 even in absence of the p66Shc Cys, and thus thiol redox reactions, suggest this interface as likely binding site for p66Shc with a direct communication contact to the Prx active site (Figure 4). Our results with added Trx indicate that an additional site on Prx can mediate further interactions, enabling formation of various complex forms, which can be reduced by Trx either through reduction or competition. The Prx1 binding site, however, does not serve as Trx binding site. The exact architecture of the p66Shc complexes formed with Prx1, however, will require future structural studies.

Another protein contributing to p66Shc regulation is Pin1 [24], a phosphoSer/Thr-Pro recognizing protein that can catalyze cis-trans isomerization within this motif. Pin1 has previously been implicated in aging processes and cellular stress responses [49], and the interaction with p66Shc is likely to be a mechanism contributing to these effects. Pin1 can bind to proteins and influence their stability, for example to p53 and cyclin D1 [49-52]. Pin1 has been reported to bind to phosphorylated p66Shc prior to transport into the IMS [24]. Our results indicate that Pin1 might act in the p66Shc system as a cytosolic inactivator, or stabilizer of a redox inactive p66Shc form (Figure 4). The exact nature of the p66CH2CB-dependent H₂O₂-metabolizing redox activity that can be silenced by Pin1 remains to be revealed. It might help to understand the discrepancy between p66Shc's apoptosis induction activity and its previously reported redox activity, H₂O₂ formation [22,30]. However, Pin1 was reported to be inactivated by oxidative stress [49,53], which would be consistent with an p66Shc-inhibiting function. Excessive cellular stress leading to Pin1 inactivation would promote the cytotoxic effect of p66Shc by allowing IMS localization and ROS formation.

Our results revealed novel regulatory effects and interactions within a p66Shc-centered signaling network. They should stimulate future studies on characterizing the coordination of these mechanisms in vivo and on the molecular architecture of the complexes formed by p66Shc. Such studies should lead to a further refined understanding of this complex signaling system, which should eventually enable development of drugs exploiting these mechanisms and interactions.

Materials and Methods

Cloning and protein purification. Cloning of full length p66Shc and of residues 1-150 (p66CH2CB) were described previously [30]. The gene for Prx1 was PCR-amplified from mouse colon cDNA and cloned into pET151/D-Topo (Invitrogen), resulting in a construct with an N-terminal 6x His-tag. Site-directed mutagenesis of p66CH2CB and Prx1 was done using the QuickChange protocol (Stratagene). Expression of WT proteins and variants was done as described for p66CH2CB [30]. Shortly, proteins were expressed in E. coli BL21(DE3)Rosetta2 cells cultured at 37 °C until the OD₆₀₀ reached 0.6. Protein expression was induced by using 0.5 mM isopropyl β-D-thiogalactosid and cell culturing continued for 18 h at 20 °C. Harvested cells were disrupted using a French Press, and cell debris was removed by centrifugation (45 min at 18000 rpm, HFA22.50 rotor). Affinity chromatography was done with Talon resin (Clontech) (washing buffer 1: 20 mM Tris pH 7.8 + 500 mM NaCl; washing buffer 2: buffer A (20 mM Tris pH 7.8 + 150 mM NaCl) + 20 mM imidazole; elution buffer: buffer A + 100 mM imidazole). The proteins and their different oligomeric states were then further purified by size exclusion chromatography using a superose 12 column (GE Healthcare) in buffer A.

Non-reducing SDS-PAGE, blue-native (BN)-PAGE and 2D-PAGE. SDS-PAGE was performed according to Laemmli et al.[54] with a T=15 % separating gel. The samples were incubated with loading buffer devoid of reducing agents for 5 min at 95 °C. BN-PAGE according to Schaegger et al.[55] was performed using a separating gel consisting of a T=10 % and a T=14 % layer overlaid with a 5 % stacking gel. The samples were supplied with 10 % (v/v) Glycerol instead of sample buffer. In order to run a second, denaturing dimension of the natively separated complexes, the respective gel bands were cut out of the BN-gel and equilibrated with 1 % (w/v) SDS and 1 % (v/v) β-mercaptoethanol first for 20 min at room temperature and second for 10 minutes at 50 °C. Subsequently, the gel pieces were washed extensively with distilled water and finally embedded into a SDS-PAGE stacking gel.

Mitochondria swelling assay and fluorescence-based ROS assay . Rat liver mitochondria were prepared freshly before use [30]. Swelling and rupture of mitochondria pretreated with 7 μM CaCl₂ was monitored photometrically as a decrease of the OD₆₂₀ as described previously [22]. The ROS-generating activity of p66CH2CB was measured using H₂DDFDA (Invitrogen). For details see [22]. Changes in fluorescence (λ(ex)=498 nm, λ(em)=525 nm) of 10 μM H₂DDFDA in 100 μl buffer A were recorded using a Perkin Elmer LS50B spectrofluorimeter.

Pull-down experiment and MudPIT analysis. 40 μl of Talon resin were saturated by incubation with 400 μg of His-tagged protein in the presence of 10 mM imidazole for 30 min at room temperature. Mitochondrial lysate (3 mg total mitochondrial protein) was added and incubated in the presence of 10 mM imidazole for 1 h at room temperature. The supernatant was removed and the affinity material first washed twice with washing buffer 1, and then twice with washing buffer 2. Finally, the His-tagged protein was eluted by addition of 30 μl elution buffer. Trypsin (Promega) was added to the eluate in a ratio of 1:100 and the proteins digested at 37 °C over night. Digestion was stopped with 0.25 % TFA and the peptide mixture analysed by MudPIT [34].