Restoration of Peroxisomal Catalase Import in a Model of Human Cellular Aging

Authors

  • Jay I. Koepke,

    1. Department of Pharmacology, Wayne State University School of Medicine, 540 East Canfield Avenue, Detroit, Michigan 48201, USA
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Kerry-Ann Nakrieko,

    1. Department of Anatomy and Cell Biology, The University of Western Ontario, London, Ontario, Canada N6A 5C1
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Christopher S. Wood,

    1. Department of Pharmacology, Wayne State University School of Medicine, 540 East Canfield Avenue, Detroit, Michigan 48201, USA
    Search for more papers by this author
  • Krissy K. Boucher,

    1. Department of Anatomy and Cell Biology, The University of Western Ontario, London, Ontario, Canada N6A 5C1
    Search for more papers by this author
  • Laura J. Terlecky,

    1. Department of Pharmacology, Wayne State University School of Medicine, 540 East Canfield Avenue, Detroit, Michigan 48201, USA
    Search for more papers by this author
  • Paul A. Walton,

    1. Department of Anatomy and Cell Biology, The University of Western Ontario, London, Ontario, Canada N6A 5C1
    Search for more papers by this author
  • Stanley R. Terlecky

    Corresponding author
    1. Department of Pharmacology, Wayne State University School of Medicine, 540 East Canfield Avenue, Detroit, Michigan 48201, USA
    Search for more papers by this author

Stanley R. Terlecky, srterlecky@med.wayne.edu

Abstract

Peroxisomes play an important role in human cellular metabolism by housing enzymes involved in a number of essential biochemical pathways. Many of these enzymes are oxidases that transfer hydrogen atoms to molecular oxygen forming hydrogen peroxide. The organelle also contains catalase, which readily decomposes the hydrogen peroxide, a potentially damaging oxidant. Previous work has demonstrated that aging compromises peroxisomal protein import with catalase being particularly affected. The resultant imbalance in the relative ratio of oxidases to catalase was seen as a potential contributor to cellular oxidative stress and aging. Here we report that altering the peroxisomal targeting signal of catalase to the more effective serine-lysine-leucine (SKL) sequence results in a catalase molecule that more strongly interacts with its receptor and is more efficiently imported in both in vitro and in vivo assays. Furthermore, catalase-SKL monomers expressed in cells interact with endogenous catalase subunits resulting in altered trafficking of the latter molecules. A dramatic reduction in cellular hydrogen peroxide levels accompanies this increased peroxisomal import of catalase. Finally, we show that catalase-SKL stably expressed in cells by retroviral-mediated transduction repolarizes mitochondria and reduces the number of senescent cells in a population. These results demonstrate the utility of a catalase-SKL therapy for the restoration of a normal oxidative state in aging cells.

Human peroxisomes generate and metabolize reactive oxygen species (ROS) produced by a number of organelle-specific metabolic reactions. These include the β-oxidation of various fatty acids, prostaglandins and leukotrienes, the metabolism of specific amino acids and polyamines and the elimination of certain xenobiotics and glyoxylate (1). The primary by-product of peroxisomal oxidations is hydrogen peroxide, a metabolite readily decomposed by the tetrameric, heme-containing enzyme, catalase (2). This balance in the production and degradation of hydrogen peroxide limits accumulation of these potentially harmful reactants and mitigates the downstream effects on cellular constituents. In contrast, perturbations in peroxisomal catalase levels, as observed in certain pathological situations (3–6) or known to accompany aging (7–11), result in oxidative damage to the cell. What role such oxidative stress plays in the initiation or progression of disease or in the process of aging is only beginning to be explored.

Peroxisomes import enzymes into their lumens by employing a cellular machinery whose components have been largely defined (12,13) and whose mechanisms have begun to emerge (12,14). Most peroxisomal enzymes, including all known human oxidases that produce hydrogen peroxide, contain a carboxy-terminal, type 1 peroxisomal targeting signal (PTS1) related to the consensus, serine-lysine-leucine (SKL) (15). These PTS1 sequences are recognized by tetratricopeptide repeat domains found in the cycling import receptor, Pex5p (16,17). Amino acids just upstream of the carboxy-terminus may also contribute to this interaction through their effects on accessibility/non-accessibility of the PTS1 (see http://mendel.imp.ac.at/mendeljsp/sat/pts1/PTS1predictor.jsp, an in silico program designed to evaluate PTS1 binding strength based on these and other parameters). The generally accepted view is that for a given enzyme, the stronger the interaction with Pex5p, the greater the capacity for peroxisomal import.

In contrast to the peroxisomal oxidases, the antioxidant enzyme catalase contains a PTS1 distantly related to the consensus, specifically lysine-alanine-asparagine-leucine (KANL) (18). In this report, we carefully measured the import of catalase, comparing its efficiency to that of an SKL-containing model enzyme. We examined its import capacity in young and old human cells and determined the effects of switching its PTS1 to the canonical SKL sequence. We provide evidence that in cells expressing both SKL-containing and endogenous KANL-containing catalase monomers, mixed oligomers form with the potential for piggybacked import of the less efficiently targeted subunits. We also show that in aged cells, more efficiently targeting catalase to peroxisomes dramatically diminishes cellular hydrogen peroxide levels. Concomitantly, these cells re-establish a membrane potential across their mitochondrial inner membranes and reduce expression of a senescence marker.

Results

Catalase is a weak substrate for import into peroxisomes

Our previous results documented that the PTS1 of catalase, KANL, only poorly targeted reporter proteins to the peroxisome (7). In contrast, we observed far more robust import of proteins containing the more canonical SKL targeting signal. In order to more directly measure peroxisomal import of catalase, an in vitro assay was modified to accommodate this critically important antioxidant enzyme. This ELISA-based assay, (described in 7,19,20), employs biotinylated import substrates and semi-permeabilized human tissue culture cells. Import in this system has been shown to be time and temperature dependent and to be stimulated by ATP and cytosol (19).

Employing the assay here, we confirmed that catalase is imported far less efficiently than the SKL-containing model enzyme, luciferase (Figure 1). Note that these experiments were performed with human A431 cells, an immortalized non-senescent cell line.

Figure 1.

Figure 1.

Quantitative analysis of peroxisomal protein import. Biotinylated luciferase (-SKL) and catalase (-KANL) were incubated with semi-intact human (A431) cells at 37°C for the times indicated in an in vitro import reaction as described in the Materials and Methods . The actual amount of reporter protein imported was determined by comparing the obtained absorbance units (at 490 nm minus the time zero values) to a standard curve of known (reporter) protein amounts. Values presented are means and ranges of duplicate samples. Note that both catalase and luciferase are enzymes with a molecular mass of 60 kilodaltons. Results are also plotted as a percentage of the maximum imported, which is arbitrarily set at 100 to permit comparison.

Our previous studies (7) have demonstrated that peroxisomal protein import is reduced in human cells of advancing replicative age and that catalase is especially affected. To examine the latter point more thoroughly, we compared the import of catalase in young/early passage cells versus old/late passage cells. Figure 2 demonstrates that the already weak import of catalase is further compromised in aging cells.

Figure 2.

Figure 2.

Catalase import in early and late passage cells. Catalase import into peroxisomes was measured at 37°C for the times indicated, using equal numbers of semi-intact early (E) or late (L) passage (IMR90) human diploid fibroblasts. Values presented are absorbance units at 490 nm (×103) and represent the means ± 1 SD of triplicate samples with time zero values subtracted. Results are also plotted as a percentage of the maximum imported, which is arbitrarily set at 100 to permit comparison.

Replacement of catalase’s endogenous targeting signal with SKL results in increased recognition and import efficiency

In an attempt to generate an efficiently imported catalase derivative, we replaced the naturally occurring KANL PTS1 with the canonical SKL sequence. This resulted in a catalase molecule that was better recognized by the PTS1-import receptor, Pex5p, in a real-time binding assay employing surface plasmon resonance (Figure 3). These results confirm and extend prior observations made using a ligand-blot/overlay assay approach (7); specifically that catalase engineered with an SKL PTS1 is better recognized by Pex5p.

Figure 3.

Figure 3.

Binding of Pex5p to catalase molecules. Surface plasmon resonance was used to examine binding of Pex5p to catalase with an endogenous PTS1 (KANL) as well as with an altered PTS1 (SKL). Note that the results presented are ‘Response Units’ and have been corrected for non-specific binding to a control surface. One thousand response units of His-tagged catalase (KANL or SKL) were coupled to a CM5 sensor chip surface by amine coupling according to Biacore’s protocols and 10 μM of GST-tagged human Pex5p was injected over the chip surface at 10 μL/min for 1 min. A Biacore biosensor 3000 was used in all experiments.

In order to determine whether the enhanced recognition of catalase-SKL by Pex5p translated to increased import competence, these molecules were examined using the in vitro import assay described above. We found that catalase-SKL was imported into peroxisomes of human A431 cells with a dramatically increased efficiency (Figure 4). Similar results were obtained with late passage Hs68 fibroblasts (Figure 5); specifically, catalase targeted using the SKL PTS1 sequence was imported more than three times as efficiently (in 30 min) as the endogenous, KANL-targeted molecule.

Figure 4.

Figure 4.

Peroxisomal import of catalase derivatives. Biotinylated catalase (-KANL) and catalase (-SKL) were incubated with semi-intact human (A431) cells at 37°C for the times indicated and import measured. The actual amount of reporter protein imported was determined by comparing the obtained absorbance values (at 490 nm minus the time zero values) to a standard curve of known (reporter) protein amounts. Results presented are a composite of three experiments; maximal import is arbitrarily set at 100 to permit comparison.

Figure 5.

Figure 5.

Peroxisomal import of catalase derivatives in late passage cells. Biotinylated catalase (KANL) and catalase (SKL) were incubated with semi-intact late passage human (Hs68) diploid fibroblasts for 30 min at 37°C. Values presented are absorbance units at 490 nm (×103) and represent the means and one ± 1 SD of triplicate samples with time zero values subtracted. Results are also plotted as a percentage of the maximum imported, which is arbitrarily set at 100 to permit comparison.

To confirm that catalase-SKL is actually trafficking properly to peroxisomes, we performed microinjection experiments depicted in Figure 6. Plasmids encoding an epitope-tagged catalase-SKL construct were nuclear microinjected into late passage cells and the distribution of the resultant molecule examined immunologically 48 h later. In panel A, antibodies to the epitope tag identify the localization of catalase-SKL. In panel B, peroxisomes are identified by staining for the membrane marker protein, PMP70. In panel C, a confocal overlay of the peroxisomes (red) and catalase-SKL distribution (green) reveals the extent of overlap (yellow/orange). Catalase-SKL is largely peroxisomal in these late passage cells, with only modest amounts seen in the cytosol. As a control, catalase bearing the KANL PTS1 was expressed in these late passage cells and examined for its distribution (Figure 6, panels D–F). As expected, this version of catalase is predominantly cytosolic in the late passage/aged fibroblasts. Therefore, modifying catalase with an SKL PTS1 enhances its engagement with the peroxisomal protein import apparatus and results in its increased uptake into the organelle, even in aged cells.

Figure 6.

Figure 6.

Localization of catalase-SKL and PMP70 in late passage cells. Late passage human (Hs27) diploid fibroblasts were nuclear microinjected with plasmids encoding Xpress™ epitope-tagged catalase-SKL (A, B, C), or catalase-KANL (D, E, F). Following incubation for 48 h, cells were fixed, stained with rabbit anti-Xpress™ antibodies followed by fluorescein isothiocyanate-conjugated goat anti-rabbit antibodies and mouse anti-PMP70 antibodies followed by rhodamine-conjugated goat anti-mouse antibodies and imaged. The cellular distribution of catalase-SKL appears in panel A and catalase-KANL in panel D; PMP70 appears in panels B and E; and an overlay of the appropriate confocal images appears in C and F. Scale bar, 10 μm.

Expression of catalase-SKL results in the formation of mixed oligomers and the co-import of endogenous catalase

As catalase is a tetrameric holoenzyme, we sought to determine whether the synthesis and import of catalase-SKL could facilitate co-import of the weakly targeted endogenous catalase. The putative first step in this co-import process would be the formation of mixed oligomers of the two molecules. To test for the presence of such hetero-oligomers, CHO cells were transfected with a (His)6-tagged catalase-SKL expression vector and incubated for 48 h. We took advantage of the (His)6 tag to isolate the expressed catalase-SKL on nickel (His)6-affinity columns, and assayed for the association of endogenous catalase, which would have bound to the column by virtue of its association with catalase-SKL. Results indicate that while the higher molecular weight catalase-SKL cannot be observed in untransfected cells (Figure 7, first lane), catalase-SKL accounts for an appreciable percentage of the total catalase in transfected cells (Figure 7, second lane). When similar lysates were applied to the nickel (His)6-affinity columns, washed and eluted, no catalase can be seen to have bound from the untransfected cell lysate (Figure 7, third lane), indicating that endogenous catalase does not bind to the nickel (His)6-affinity column. However, catalase-SKL from the lysate of the transfected cells does bind to the nickel (His)6-affinity columns (Figure 7, fourth lane) and endogenous catalase can be observed to have also bound under these conditions. Of the catalase molecules bound to the nickel (His)6-affinity columns, endogenous catalase accounted for a small, but clearly discernible portion of the total. It is worth noting that the stoichiometry of the association between the two forms of catalase would depend upon their relative expression levels during the time of synthesis and assembly. As catalase-SKL is being expressed using the cytomegalovirus promoter, its expression and thus its levels in the newly synthesized and assembled oligomers would be expected to be higher than that of endogenous catalase. This may explain the relative ratios of catalase subunits seen in Figure 7. However, this result does demonstrate the formation of mixed oligomers of endogenous catalase and catalase-SKL, a necessary first step in the co-import of endogenous catalase by catalase-SKL.

Figure 7.

Figure 7.

Association of catalase-SKL with endogenous catalase in cells. Lysates (first and second lanes) obtained from CHO cells transfected (where indicated) with (His)6-catalase-SKL and eluates from the (His)6-affinity column (third and fourth lanes) were separated by SDS–PAGE and immunoblotted with anti-catalase antibodies. Note the presence of both (His)6-catalase-SKL (upper band) and endogenous catalase (lower band) in the nickel-resin pull down after transfection (fourth lane).

In order to determine whether or not the newly synthesized catalase-SKL has an effect on the subcellular distribution of endogenous catalase, we performed immunofluorescence analysis to localize the catalase molecules in microinjected late passage human cells. Both catalase-SKL and endogenous catalase were found contained within punctate peroxisomes (Figure S1, panels A–C). This in marked contrast to catalase’s localization in uninjected late passage cells where we previously reported that the molecule was only poorly compartmentalized and appeared largely cytosolic [(7); see also Figure 6 for an example of catalase-KANL trafficking in aged cells]. Thus, it appears possible that the presence of at least one SKL-tagged catalase monomer in the tetrameric complex may be sufficient to facilitate efficient import of the holoenzyme into the peroxisomal matrix. Note that such ‘targeting in trans’ or ‘piggybacking’ has been described as an extant mechanism for both the PTS1 (21) and PTS2 (22) mediated import pathways.

Catalase-SKL expression reduces cellular hydrogen peroxide levels

Work from many laboratories has demonstrated that aging cells accumulate hydrogen peroxide (7,23,24). As this increase is coincident with a decrease in peroxisomal catalase (7), we sought to determine whether or not restoration of catalase to the peroxisome, by virtue of the engineered SKL targeting signal, resulted in a decrease in cellular hydrogen peroxide levels. In order to unambiguously identify the microinjected cells in the live-cell hydrogen peroxide assay, we co-injected plasmids expressing catalase-SKL or catalase-KANL with DsRed-SKL at equal concentrations. At 48 h after microinjection, cells expressing catalase-SKL demonstrated a clear reduction in the amount of intracellular hydrogen peroxide, as measured by 2′,7′-dichlorofluorescein staining (Figure 8, panels A and B). DsRed-SKL in these cells was primarily localized in punctate peroxisomes. Expression of catalase-KANL and DsRed-SKL after 48 h of expression yielded different results (Figure 8, panels E and F), with DsRed-SKL having a peroxisomal and cytosolic localization, indicative of inefficient import and little demonstrable reduction in the cellular hydrogen peroxide levels. After 72 hours, catalase-SKL expression continued to reduce cellular hydrogen peroxide concentrations (Figure 8, panels C and D) and even the expression of catalase-KANL resulted in a decrease in levels of the ROS (Figure 8, panels G and H). However, significant amounts of (cytosolic) hydrogen peroxide were still detectable in the catalase-KANL expressing cells (Figure 8, panel H). Cytosolic hydrogen peroxide was not observed in the catalase-SKL expressing cells (Figure 8, panel D); 2′,7′-dichlorofluorescein staining in these cells was primarily limited to DsRed-SKL containing peroxisomes. As a control, we also showed that expression of DsRed-SKL alone had no effect on cellular hydrogen peroxide levels (Figure 8, panels I and J), with levels of 2′,7′-dichlorofluorescein staining equivalent to that of uninjected cells (Figure 8, panel B or D).

Figure 8.

Figure 8.

Effect of catalase-SKL expression on hydrogen peroxide levels in late passage cells. Late passage human (Hs27) diploid fibroblasts were nuclear microinjected with plasmids encoding DsRed-SKL and catalase-SKL (panels A and B, C and D) or DsRed-SKL and catalase-KANL (panels E and F, G and H) and incubated for 48 h (panels A, B, E, F), or 72 h (C, D, G, H) at 37°C. Control experiments, which expressed DsRed-SKL alone (panels I and J), were incubated for 72 h. Left column panels identify living microinjected cells expressing the DsRed-SKL. Right column panels indicate the relative levels of hydrogen peroxide in cells as revealed by 2′,7′-dichlorofluorescein staining, imaged under identical conditions. Panel K reflects quantitation (mean fluorescence intensities per unit area using National Institutes of Health’s Image J (v1.31) public domain shareware) of the cellular hydrogen peroxide levels seen in the variously treated late passage cells. Cells expressing catalase-SKL contain less hydrogen peroxide than cells expressing catalase-KANL, both at 48 and 72 h. Scale bar, 10 μm.

Retrovirus-mediated expression of catalase-SKL delays the appearance of a senescence marker and reverses mitochondrial depolarization

To extend our analysis of catalase’s effects on aging cells, late passage (Hs27) human diploid fibroblasts were transduced with retroviruses engineered to induce expression of catalase-KANL, catalase-SKL, or a null (vector) control. The epitope-tagged catalase molecules were expressed to similar extents as determined by Western blotting (Figure S2A). To confirm their characteristic trafficking behavior, we examined their localization by indirect immunofluorescence microscopy. Consistent with results shown in Figure 6, epitope-tagged catalase-SKL localized to peroxisomes whereas catalase-KANL was largely cytosolic (S2B).

Peroxisomal catalase-SKL efficiently metabolized hydrogen peroxide and related ROS as evidenced by the dramatic reduction in 2′,7′-dichlorofluorescein fluorescence (Figure S3). Catalase-KANL showed no similar effects; indeed, hydrogen peroxide levels appeared similar to those seen in vector control-treated cells.

We next asked whether or not the presence of catalase-SKL would reduce or delay expression of the aging marker, senescence-associated β-galactosidase (Table 1). Cells were cultured to various population-doubling levels, including those falling into early, middle and late passage categories. Early and middle passage cells expressed only little senescence-associated β-galactosidase, whereas some 75% of late passage cells were positive for the marker activity. Importantly, retrovirally introduced catalase-SKL reduced this staining by over 29% (from 75 per 100, to 53 per 100); in contrast, similarly expressed catalase-KANL had only a modest 8% (75 per 100, to 69 per 100) effect.

Table 1.  Senescent cells present in various age groups
Cell passage numberSA-β-galactosidase-a positive cellsb
  • a

    Senescence-associated β-galactosidase, a histochemcial biomarker for aging (36).

  • b

    Cells used were human (Hs27) diploid fibroblasts cultured to the passage numbers indicated. Where noted, catalase-KANL or catalase-SKL was introduced by expression of appropriately engineered recombinant retroviruses as described in the Materials and Methods . Values presented are the mean number of senescence-associated β-galactosidase-positive cells (±1 SD) per hundred examined. At least three separate populations were analyzed per condition.

22 (Early)0
31 (Middle)2
75 (Late)75 ± 3
75 (Late) + catalase-KANL69 ± 4
75 (Late) + catalase-SKL53 ± 2

With oxidative balance restored and appearance of an age-related marker reduced, we considered that catalase-SKL-expressing late passage cells might similarly delay or reverse other aging phenomena. We turned to an examination of mitochondria. Mitochondria are critically important cellular organelles whose proper functioning requires maintenance of a potential gradient across its inner membrane. Mitochondrial aging is associated with a loss of this polarization. To determine whether or not catalase-SKL and the resultant redox balance created had any effect on mitochondrial polarization/depolarization, we employed the potential sensor dye, 5, 5′,6, 6′-tetrochloro-1, 1′,3, 3′-tetraethyl-benzimidazolycarbocyanine iodide, known as JC-1. Our results (Figure 9) show that early passage cells display the largely red/orange staining pattern associated with fully polarized mitochondria. In contrast, mitochondria of late passage cells are largely depolarized and stain fluorescent green. Importantly, late passage cells expressing catalase-SKL reveal a restored mitochondrial inner membrane potential as the organelle once again stains fluorescent red. Thus, a renewed coupling of peroxisomal pro-oxidants and antioxidants re-establishes an oxidative equilibrium in cells, reduces the number of senescent cells present and restores mitochondrial integrity. The implications of these findings are discussed below.

Figure 9.

Figure 9.

Restoration of mitochondrial inner membrane potential by catalase-SKL expression in late passage human cells. Hs27 cells were treated with the mitochondrial potential sensor 5,5′,6,6′,-tetrachloro-1,1′,3,3′-tetraethyl-benzimidazolyl-carbocyanine iodide (JC-1) as per manufacturer’s instructions (Molecular Probes) and the formation of J-aggregates, indicative of depolarization, examined. Live-cell images were collected with a laser-scanning confocal microscope using fluorescein isothiocyanate or rhodamine filter sets and were illuminated and collected under identical conditions. Images represent mitochondrial inner membrane potentials of early passage cells (A), late passage cells (B) and late passage cells expressing catalase-SKL (C). Scale bar, 10 μm.

Discussion

Peroxisomes carry out a variety of metabolic reactions that produce hydrogen peroxide and other ROS as by-products. Under most circumstances, the organelle maintains oxidative equilibrium through its ability to rapidly metabolize these compounds in situ. In the case of hydrogen peroxide, degradation is the responsibility of catalase, a peroxisomal enzyme whose structure and mechanism of action are well understood (25,26).

Peroxisomal hypocatalasemia, secondary to alterations in the enzyme’s synthesis or turnover rates, is associated with increased stress in cells and with premature appearance of age-related pathologies in human patients (4,5). Many of the hallmarks of the late-passage cell, including an accumulation of hydrogen peroxide, oxidative damage to DNA and proteins and a reduced rate of mitosis can be observed in hypocatalasemic cells (6). We have previously demonstrated that the restoration of catalase-SKL, using a cell-penetrating peptide-based methodology decreases the levels of hydrogen peroxide in these cells (6). Peroxisomal catalase insufficiency is also manifest in aging cells where the enzyme appears to be only poorly imported into the organelle (7). Based on these and other observations in a variety of organisms, the suggestion has been made that peroxisomal catalase and life span may be linked (27). Our goal in the present study was to carefully characterize catalase trafficking in human cells and attempt to restore its levels in aging peroxisomes.

Consistent with our previous observations, catalase was imported only poorly when compared with the SKL-containing luciferase molecule in an in vitro import assay (Figure 1). That catalase is an especially poor substrate in aged cells (Figure 2) may also seem predictable based on recent work (7). However, our present results are, to our knowledge, the first time catalase has directly been shown to be poorly imported in aging cells. Previous work that led to the same conclusion was based largely on mislocalization of the endogenous enzyme in late passage cells or the age-dependent altered trafficking of fluorescent reporter molecules (7).

Using the web-based PTS1 prediction analysis software developed by Mandel and colleagues (28), the PTS1 of catalase scores a 1.5, whereas the SKL sequence of luciferase scores a 5.2. Interestingly, our use of luciferase may only underestimate the differences in import capacities between peroxisomal oxidases and catalase. The oxidase enzymes (which include three acyl-CoA oxidases, d-amino acid oxidase, d-aspartate oxidase, α-hydroxyacid oxidase, pipecolic oxidase and polyamine oxidase; see 1 for more information) have an average PTS1 predictor score of 10.2 – nearly seven times that of catalase. Replacing catalase’s PTS1 sequence (KANL) with SKL increases its PTS1 predictor score to 11.3. Surface plasmon resonance measurements of the binding between Pex5p and the catalase derivatives confirm that such a change increases the molecule’s recognition by Pex5p (Figure 3). Consistent with the increased interaction with Pex5p, replacement of the PTS1 with an SKL sequence results in more efficient import of the catalase derivative (Figure 4). This is true in early passage cells and in late passage cells (Figure 5). We verified correct trafficking of the catalase-SKL molecule to peroxisomes by immunofluorescence analysis of aged cells microinjected with the corresponding plasmid (Figure 6). This peroxisomal import did not occur when catalase-KANL was expressed in late passage cells. Thus, there remains little doubt that catalase-SKL is a vigorous substrate for peroxisomal protein import – even in aging cells.

Catalase, enzymatically active as a tetramer, appears to assemble in the cytosol and be imported without loss of quaternary structure. Of course, this is not an issue from the standpoint of the peroxisomal protein import machinery, which can accommodate oligomeric structures (29). What is created by these properties is the possibility that in cells expressing the SKL as well as the naturally occurring KANL catalase derivative, hetero-oligomers may form. We tested for these, and confirmed their existence in catalase-SKL transfected cells (Figure 7). Importantly, the presence of catalase-SKL had a ‘clearing’ effect on the remainder of cellular catalase – largely removing mislocalized enzyme from the cytosolic compartment and having it amass in peroxisomes (Figure S1). Catalase-KANL expression in these cells had no such effect (data not shown). It should be noted that an alternative explanation for peroxisomal localization of endogenous catalase is that the newly restored oxidative state allows import of the poorly targeted catalase molecule; this cannot be ruled out at this point. Therefore, perhaps the most appropriate explanation based on the information currently available is that both the formation of mixed oligomers and the decrease in cellular hydrogen peroxide levels contribute to the enhanced import of catalase.

Cells expressing catalase-SKL possess a largely peroxisomal pool of the antioxidant enzyme and are well equipped to process the hydrogen peroxide produced by the organelle. Thus we observed a dramatic reduction in cellular hydrogen peroxide levels in late passage cells (Figure 8). Localization within the peroxisome was an important factor, as this reduction was less effective when catalase-KANL was expressed in these cells. Targeting catalase to the source of hydrogen peroxide production increases the effective concentrations of both the enzyme and its substrate by a factor of approximately 100. In addition, keeping catalase in a relatively high concentration of hydrogen peroxide favors the enzyme’s catalatic reaction, that being the disproportionation of two molecules of hydrogen peroxide to water. Catalase is also capable of peroxidatic reactions, which use one molecule of hydrogen peroxide to reduce cellular alcohols, acids and aldehydes. Thus, low levels of catalase and hydrogen peroxide in the cytosol may actually contribute to cellular damage. Furthermore, transient low levels of hydrogen peroxide are now recognized as important mediators of signal transduction (30). Elevated levels of cytosolic catalase might be expected to compromise this redox signaling. However, it should be noted that high levels of cytosolic hydrogen peroxide have also been shown to inhibit mitogenic signals in late passage cells (31). Because the localization of antioxidants is important to their function, restoration of catalase to the peroxisome by virtue of the SKL targeting signal allows the cell to eliminate (peroxisomal) hydrogen peroxide while not disturbing cellular redox signaling cascades. Thus, ours is a strategy based on targeted antioxidant prophylaxis.

To extend our analysis of catalase-SKL’s effects, we examined late passage/aged cells stably expressing the enzyme after infection with specifically constructed/assembled recombinant retroviruses. Not unexpectedly, the molecule successfully trafficked to peroxisomes and quenched a significant amount of hydrogen peroxide and related ROS (Figures S2 and S3). As importantly, it also reduced the number of senescent cells present in the culture (Table 1). What effects the continuous presence of catalase-SKL has on cells when introduced in early or middle passage remains to be determined; perhaps their replicative life span will be extended.

Continuing with this notion, any attempts to extend a cell’s life span will have to include a strategy to maintain the functional integrity of mitochondria. These organelles are critical to the cell by virtue of their ability to synthesize ATP and participate in various metabolic activities. Indeed, many assays for cell viability/compound toxicity rely on mitochondrial indices. Like the peroxisome, mitochondria produce ROS and are affected by them. Therefore, the hydrogen peroxide and related ROS-laden environment of late passage/aged cells would be expected to compromise their activity. Perhaps this explains why mitochondria of such cells have largely lost their ability to maintain a membrane potential across their inner membranes [(32) and Figure 9]. Our observation that expression of catalase-SKL in such cells repolarizes mitochondria (Figure 9) is also consistent with this view.

In conclusion, this work emphasizes the cytoprotective effects of catalase-SKL. Restoring an oxidative balance in peroxisomes extends beyond the confines of the organelle. Total cellular hydrogen peroxide and related ROS are reduced, mitochondrial integrity is restored and the appearance of a senescent phenotype is at least delayed. Our efforts going forward will focus on more thoroughly examining the protective effects of catalase-SKL, both in a cellular and organismal context.

Materials and Methods

Reagents

Firefly (Photinus pyralis) luciferase and human erythrocyte catalase were purchased from Sigma. Three recombinant hexahistidine (His)6-tagged human catalase proteins with different carboxy-terminal PTS1 sequences were created as described in (7), expressed in bacteria and purified with nickel-nitriloacetic acid agarose as per manufacturer’s protocols (Qiagen). These molecules were 95% pure by Coomassie blue staining. Glutathione S-transferase (GST)-tagged human Pex5p was expressed and purified as described (33). Rabbit polyclonal antibodies directed against human catalase were obtained from Calbiochem or custom ordered from Sigma-Genosys; rabbit polyclonal antibodies directed against PMP70 and glyceraldeyde-3-phosphate dehydrogenase (GADPH) were purchased from Affinity Bioreagents; and mouse monoclonal antibodies to the Xpress™ epitope were procured from Invitrogen. Fluorescently conjugated secondary antibodies were purchased from Jackson ImmunoResearch Laboratories Inc. or KPL. 2′,7′-dichlorofluorescin diacetate was purchased from Acros Organics (Fisher Scientific) and used to detect cellular hydrogen peroxide as described (7,34,35). Other reagents were obtained from standard sources.

Cell culture

Hs27 and Hs68 human diploid fibroblasts, Chinese hamster ovary (CHO) cells and human epithelial carcinoma (A431) cells were obtained from American Type Culture Collection. Early passage IMR90 human diploid fibroblasts were obtained from the National Institutes of Aging, Aging Cell Repository/Coriell Institute for Medical Research. Cells were cultured, maintained, and passaged as previously detailed (7). Senescence-associated β-galactosidase staining (36) was used to confirm the late passage/aged phenotype.

In vitro import assays

Peroxisomal protein import was examined in semi-intact cells using an enzyme-linked immunosorbent assay (ELISA)-based in vitro assay previously described (7,19,20). A related ELISA system was developed here to measure catalase import. For these studies, both human erythrocyte catalase as well as the three recombinant catalase derivatives described above were employed. Biotinylation of each (2 mg) protein was achieved using 6-((6-(biotinoyl) amino)hexanoyl)amino)-hexanoic acid, succinimidyl ester (biotin-XX SE) in a reaction for 45 min at room temperature. Unconjugated biotin-XX SE was removed from the biotinylated species by gel filtration using a G25 spin-desalt column pre-equilibrated with PBS/0.2% BSA. These proteins were used at the same specific activity. Semi-intact cells were incubated with a given biotinylated catalase derivative (or biotinylated luciferase) (typically at 10 μg/mL) at 37oC in a 40-μL reaction volume. The amount of substrate used is designed to not be limiting for the transport step. The import reaction buffer contained 40-mM HEPES-sodium hydroxide (pH 7.4), 85 mM sucrose, 2 mM magnesium acetate, 100 mM potassium acetate, 100 μM zinc chloride, 1 mM adenosine 5′-triphosphate, 5 mM creatine phosphate and 0.4 IU creatine phosphokinase. Import was determined at specific time points by masking accessible biotin sites on non-imported protein with avidin (Calbiochem), quenching excess avidin with biocytin (Calbiochem), and solubilizing the cells with detergent. The reactions were plated on microtiter plates coated with anticatalase polyclonal antibodies (diluted 1:1000). The extent of unmasked biotin, the hallmark of import, was then quantitated with horseradish peroxidase-labeled streptavidin in a microplate absorbance reader set at 490 nm and corrected for background at 650 nm. The maximal absorbance signal obtained in a given experiment depended on such factors as the concentration of ligand added and the number of cells used (typically 1.4–2.7 × 105 per reaction). DNA content was measured to determine relative cell numbers. The procedure used may be found in (37).

Surface plasmon resonance bimolecular interaction measurements

A Biacore 3000 (Biacore AB) was used to evaluate the binding of Pex5p to catalase with its own PTS1 (catalase-KANL) as well as with an altered PTS1 (catalase-SKL). The purified recombinant catalase proteins were immobilized on individual flow cells of a Biacore AB CM5 chip by amine coupling according to the manufacturer’s protocols. The sensor surface was first activated with 0.2 M N-ethyl-N′-(diethylaminopropyl) carbodiimide and 0.05 M N-hydroxy-succinimide. The proteins were injected in acetate buffer (10 mM, pH 4.0) (Biacore AB), using the surface preparation wizard to bind 1000 response units of each. The chip was then blocked with 1 M ethanolamine hydrochloride pH 8.5 (Biacore AB). A 10 μM solution of the GST-Pex5p was injected over the chip surface at 10 μL/min for 1 min in HEPES buffer (10 mM, pH 7.4, containing 0.15 M NaCl and 0.005% (v/v) P20 surfactant). Sensorgrams from a control surface were subtracted from sensorgrams obtained with immobilized catalase to yield true net binding responses.

Preparation of plasmids

The pJK22 and pJK23 mammalian expression vectors, expressing catalase-SKL and catalase-KANL, respectively, were created by ligation of a human catalase PCR fragment into pcDNA3.1/HIS A (Invitrogen). To generate the fragment, human catalase was PCR-amplified from a full-length cDNA clone (Invitrogen). The forward primer, 5′-CGGggtaccTATGGCTGACAGCCGGGATCCCGCC-3′ complemented the 5′-sequence of human catalase cDNA along with a KpnI restriction site (lower-case letters) for cloning purposes. The reverse primer, 5′-GGGCGCgcggccgcTCACAGTTTCGATTTCTCCCTTGCCGCCAAGTG-3′ comple-mented the 3′-end of human catalase with nucleotide changes that code for a carboxy-terminus of SKL as opposed to the naturally occurring ANL. The reverse primer, 5′-GGGCGCgcggccgcTCACAGATTTGCCTTCTCC-3′ complemented the 3′ end of human catalase. In both reverse primers, a NotI restriction site (lower case) was also incorporated downstream of the stop codon. The ligation product was transformed into a DH5α bacterial strain. Transformants were selected and the recovered plasmid was confirmed to be correct by restriction analysis and DNA sequencing (Applied Genomics Technology Center, Wayne State University). The pDsRed2-SKL mammalian expression vector was created as described previously (7).

Transfection

Exponentially growing CHO cells at 80% confluency were transfected with the pJK22 mammalian expression vector using a MBS mammalian transfection kit as described by the manufacturer (Stratagene). Stable transformants were selected by incubating cells with the antibiotic G418 (400 μg/mL) for 2 weeks.

Nickel-resin pulldown

One confluent 100 mm plate of transfected cells (and one of non-transfected cells) was harvested in 500 μL of lysis buffer (50 mM NaH2PO4, 300 mM NaCl, 20 mM imidazole, pH 8.0). Cell lysates were prepared by sonication of the cell suspension and clarified by removal of cellular debris by centrifugation at 10 000× gfor 10 min. An aliquot of each was saved for Western blotting and the remaining sample incubated with 10 μL of nickel-nitriloacetic acid agarose for 1 h at 4°C. The beads were washed twice with lysis buffer followed by elution with 250 mM imidazole. The eluate was analyzed by Western blotting. Enhanced chemiluminescence was used to visualize immunodecorated (catalase) bands. Various exposures were examined to assure blots were not overexposed.

Nuclear microinjection/expression of tagged constructs/imaging

Nuclear microinjection of late passage/old Hs27 cells with plasmids (25 μg/mL) encoding DsRed-SKL or Xpress™-epitope tagged/(His)6-tagged catalase-SKL was carried out as outlined (7). Immunostaining for the Xpress™ epitope of the (His)6-tagged catalase-SKL construct was carried out with anti-Xpress™ antibodies at a 1:200 dilution. Cells were imaged using a Zeiss Axiovert fluorescence microscope, equipped with a CCD camera or with a Zeiss LSM 420 fluorescence confocal microscope. As a result of the small size of the peroxisomes, small numbers of cells were photographed at high magnification. These images were always representative of a large number of microinjected cells.

Recombinant retrovirus engineering and transduction

Xpress™ epitope-tagged catalase-SKL (pJK22) and catalase-KANL (pJK23) were subcloned using HindIII and NotI restriction sites, into the retroviral vector pLNCX2 (Clonetech), creating pKN10 and pKN11, respectively. Recombinant retroviruses encoding Xpress™-tagged catalase-KANL, catalase-SKL, or a vector control, were generated in packaging cells transfected with Superfect (Qiagen) using standard procedures (38). Retroviruses were microfilter purified and titered by dilution assay in early passage human (Hs27) diploid fibroblasts. Viral infections were performed by incubating late passage Hs27 cells in Dulbecco’s modified Eagle’s medium (Invitrogen) containing 10% (v/v) fetal bovine serum (Invitrogen) and 10 mg/mL polybrene (Sigma-Aldrich) and retrovirus for 3 h, followed by culturing in normal growth medium for 48 h. Post-infection, transgenic pKN10-, pKN11-, and vector control-Hs27 cells were selected in normal growth medium treated with the antibiotic G418 (500 μg/mL) for 21 days. Cells were then passaged three times prior to use.

Acknowledgments

The authors acknowledge the Center for Molecular and Cellular Toxicology with Human Applications in Michigan (funded by NIEHS grant P30 ES06639). These studies were supported by grants from the National Institutes of Health (DK56299) to S. R. T. and the Canadian Institutes of Health Research to P. A. W. S. R. T. and P. A. W. are co-founders of EXT Life Sciences Inc.

Ancillary