Introduction

Oxygen is required for life. However, high-energy forms of oxygen or reactive oxygen species are inevitably formed as electrons escape their intended molecular destinations, or upon exposure to radiation of various wavelengths. Oxidants damage all forms of biomolecules including DNA, proteins and lipids, and are involved in the etiology of various diseases such as cancer, cataracts, retinopathies, neurodegenerative diseases (Giasson et al., 2000) and perhaps aging per se (Taylor, 1999). Oxidative stress may also contribute to the phenotype of the pleiotropic recessive disorder ataxia-telangiectasia (A-T) (Rotman and Shiloh, 1997, 1999). A-T is characterized by cerebellar atrophy leading to neuromotor dysfunction, telangiectases (dilated blood vessels) in the eyes and/or facial skin, immunodeficiency, gonadal atrophy, genomic instability, profound predisposition to cancer, and acute radiosensitivity. Since many of these characteristics are observed upon normal aging, A-T may be regarded as a premature aging syndrome.

A-T is due to lack or inactivation of the ATM protein, a large protein kinase, which serves as a master controller of cellular responses to DNA double-stranded breaks (DSBs). Upon induction of these extremely toxic DNA lesions, a fraction of nuclear ATM adheres to the sites of the breaks, and ATM's kinase activity is enhanced, resulting in phosphorylation of many proteins, each of which in turn affects a specific signaling pathway (Andegeko et al., 2001; reviewed by Shiloh and Kastan, 2001). Thus, A-T cells are defective in a large array of cellular responses to DSBs.

Evidence that A-T cells may be under enhanced oxidative stress include: observations of reduced life span in culture (Shiloh and Becker, 1982), decreased survival when challenged with H2O2 (Applegate et al., 1994; Takao et al., 2000), increased NO-mediated damage to proteins in brains of ATM-deficient mice, higher levels of lipid peroxidation in ATM-deficient testes (Barlow et al., 1999), and elevated activity of heme oxygenase (HO) in ATM-deficient Purkinje cells (Barlow et al., 1999), cultured cells (Watters et al., 1999), and red blood cells from A-T patients (Rybczynska et al., 1996). Increased HO is a common response to oxidative stress. In addition, Kamsler et al. (2001) found reduced levels of thiol compounds, and a higher level of thioredoxin, another indication of oxidative stress (Tanaka et al., 2000). Some of these stress effects and the responses to the stress were blocked by antioxidants (Gatei et al., 2001; Takao et al., 2000). Shackelford et al. (2001) showed that the response of human cells to the oxidant t-butyl hydroperoxide is ATM-dependent. Accordingly, A-T fibroblasts were hypersensitive to this agent and exhibited defective activation of cell cycle checkpoints in response to this treatment.

Studies regarding cellular responses to DSBs utilize ionizing radiation or strand break-inducing chemicals. Typical of such radiomimetic compounds is the antibiotic dienediyne neocarzinostatin (NCS), an oxidant drug that still finds use in cancer therapy (Okada, 2000) and which results in activation of cell cycle checkpoints. Generally, the cell cycle is also regulated through a timed and coordinated degradation of cyclins and inhibitors of cyclin-dependent protein kinases via the ubiquitin pathway (for reviews see Hershko and Ciechanover, 1998; Hochstrasser, 1996; King et al., 1996).

Recently, we demonstrated that the ubiquitin pathway is also regulated by cellular redox status (Jahngen-Hodge et al., 1997; Obin et al., 1999; Shang et al., 1997b) and that it is the major cellular pathway for removal of oxidized proteins (Jahngen-Hodge et al., 1997; Shang et al., 1997a). Taken together with data which suggest elevated oxidant sensitivity in A-T cells, it would appear that sequelae due to the absence of a functional ATM protein might involve the ubiquitin pathway as well. Yet no reports characterize ubiquitination responses in A-T cells or the effects of NCS on the ubiquitination responses in any cells.

The ubiquitin proteolytic pathway is found in many if not all eukaryotic cells. Experiments have implicated ubiquitination or ubiquitin dependent protein degradation in virtually every facet of cellular regulation, including: cell cycle progression, differentiation, stress responses, DNA repair, signal transduction, gene regulation, control of proto-oncogenes, antigen presentation, and removal of damaged proteins (Haas and Siepmann, 1997a,b; Jahngen-Hodge et al., 1997; Johnston et al., 1998; Kornitzer and Ciechanover, 2000; Obin et al., 1999; Shang et al., 1997a; for reviews see Ciechanover, 1994; Ciechanover et al., 1985; Hochstrasser, 1996; Huang et al., 1995; Jentsch and Schlenker, 1995; Shang et al., 1997b; Shang and Taylor, 1995; Varshavsky, 1997). Most, but not all, of these functions are accomplished by degrading or preventing the degradation of relevant regulatory proteins.

In its simplest form, substrate recognition involves attachment of multiple molecules of ubiquitin to protein substrates. Generally, oxidative modifications render proteins better substrates for ubiquitination (Hershko et al., 1986; Huang et al., 1995; Shang and Taylor, 1995). In order to initiate this process, ubiquitin is ‘activated’ by the formation of a high-energy thiolester with ubiquitin activating enzyme, E1, the first enzyme in the ubiquitination pathway. The ubiquitin is then transferred to one of >11 E2s (also called ubiquitin conjugating enzymes), also via formation of a thiolester (Jentsch, 1992; Varshavsky, 1997). Subsequently, ubiquitin is transferred directly to substrates or is transferred to substrates after reaction with one of several E3s (also called ubiquitin ligases) (Scheffner et al., 1995). Multiple ubiquitins are usually attached to each other. Thus, ‘ubiquitin trees’ are found attached to substrates. Such ubiquitin conjugates generally attain high masses. These ubiquitin conjugates are recognized and degraded, by the 26S proteasome, at rates which are generally proportional to rates of ubiquitin-adduct formation (Haas, 1997; Haas and Siepmann, 1997a; Jabben et al., 1989; Shang and Taylor, 1995). The multiplicity of E2 and E3 enzymes allows for specificity within the system.

Each of the enzymes in the ubiquitin pathway has an active site cysteine. Thus it is not surprising that the activity of the enzymes involved in recognition of substrate proteins and their removal are inactivated when cellular redox capability is significantly diminished (Jahngen-Hodge et al., 1997; Obin et al., 1998).

Upon aging and/or exposure to high levels of oxidants there are increases in levels of oxidized proteins accompanied by decreased proteolytic activity (Cuervo and Dice, 2000; Jahngen et al., 1986; Stadtman, 1992; Taylor, 1999; Taylor and Davies, 1987). Milder physiological oxidation regimes (Shang et al., 1997b) also indicate elevations in levels of ubiquitin conjugates but show either unchanged or slightly increased activity of the ubiquitin conjugating machinery. In this study we compared four lymphoblastoid cell lines from healthy individuals and three A-T cell lines for indicators of oxidative stress associated with the ubiquitin system: levels of endogenous ubiquitin conjugates, ability to form ubiquitin conjugates de novo, thiol-ester forming ability of E1 and E2s, and levels of GSH and GSSG, with and without treatment with the NCS or the physiological oxidant H2O2. We show that NCS and H2O2 treatments have similar effects on protein ubiquitination, both resulting in transient increases in ubiquitin conjugates. However, A-T cells have a phenotype more characteristic of cells already aged and under oxidative burden. We also show that ubiquitination capabilities are among the earliest indicators of exposure to stress.

Results

Ubiquitin conjugates change upon treatment with NCS

The radiomimetic drug NCS has been used to distinguish A-T from normal cells but not with respect to effects of A-T mutations on function of the ubiquitin pathway. Thus, we examined the ubiquitination response of wt and A-T cell lines during NCS treatment. In lymphoblastoid cells from normal and A-T patients, levels of endogenous ubiquitin conjugates (which are usually observed at the higher mass regions of protein gels) increased 50–70% (based on average of densitometric evaluation of Western blots) within 0.5 h of NCS treatment and declined after 2 h, returning to basal levels by 4–8 h (Figure 1). The same was true with other normal and A-T lymphoblastoid cells and when the cells were treated with ionizing radiation (data not shown).

Figure 1
figure 1

A transient increase in endogenous ubiquitin conjugates upon exposure to NCS. Lymphoblastoid cells from a normal (L-40) and an A-T patient (A-T59) at a concentration of 0.5×106 cells/ml were treated with 80 ng/ml of NCS for 0–4 h. Endogenous ubiquitin conjugates were detected by Western blotting analysis of cellular extracts using an anti-ubiquitin antiserum. The membrane was reprobed using anti-actin antibody to monitor loads of the different lanes in the gel. (a) Wild type cells; (b) cells derived from an A-T patient. (c and d) Present densitometric analyses of the data in a and b, respectively. The data in c and d are the ratios relative to actin of the optical density of the area of the film at which ubiquitin conjugates are observed in the Western blot. The ratio was designated as 100 for the cells at time 0

Full size image

Increased levels of conjugates might be due to NCS-induced elevation of substrate levels and/or upregulated activities of the cellular ubiquitination machinery. In order to determine if ubiquitin conjugation capabilities per se were involved in the NCS-induced increase in ubiquitin conjugates, we determined the ability to form ubiquitin conjugates de novo using exogenous 125I-ubiquitin and endogenous conjugating enzymes and substrates. These assays revealed changes in de novo ubiquitin conjugates (Figure 2) similar to those noted for endogenous conjugates. By 4 h de novo formed ubiquitin conjugates increased by 75% (value normalized for protein load), and then gradually decreased, returning to basal levels by 8 h. A similar response was noticed in two wild type and two A-T cell lines. Taken together, these experiments indicate that NCS treatment results in enhanced ubiquitination of endogenous proteins as well as in more substrates for ubiquitination.

Figure 2
figure 2

A transient increase in de novo ubiquitin conjugate formation upon exposure of cells to NCS. Lymphoblastoid cells from a normal (L40) and an A-T patient (A-T59) were exposed to 500 ng/ml NCS for 0–8 h, lysates were prepared as described in Materials and methods. Ubiquitin conjugating activities were determined by the amount of ubiquitin conjugates formed with exogenous 125I-ubiquitin. Samples were resolved on SDS–PAGE and the gel exposed to film. (a) Wild type cells. (b) Cells from an A-T patient

Full size image

The extent and chronology of this response was reminiscent of responses in lens or retina cells which were under (or recovering from) mild oxidative stress (Jahngen-Hodge et al., 1997; Shang et al., 1997b). Thus, we sought to (1) further characterize and compare the ubiquitination response in A-T and wt cells, and (2) determine if changes in the ubiquitination response were associated with changes in redox status using a more frequently employed and physiologically relevant oxidant.

Endogenous and de novo 125I-ubiquitin conjugate formation in wt and A-T cells treated with H2O2

We used H2O2 in further experiments because (1) this would allow us to compare these results with prior information regarding the effects of H2O2 on ubiquitination capabilities in other types of cells (from both fast and slowly metabolizing tissues), (2) H2O2 is encountered in many physiological conditions, and (3) because there was little data which related NCS treatment to alterations in redox status (Adamo et al., 1999; Jahngen-Hodge et al., 1997; Obin et al., 1998; Ramanathan et al., 1999; Shang et al., 1997b). The experiments were done at three cell densities: 0.3×106, 0.6×106 and 1.0×106 cells/mL in order to assure that cell density alone did not cause changes in ubiquitin conjugates. At these densities SDS–PAGE indicated no differences between proteins in the soluble or insoluble fractions from 3 A-T cell lines (A-T59, A-T24, and L6) and four wt cell lines (L40, NL552, NL553, and 394RM) (data not shown). Thus, A-T and wt cells do not differ in levels of the most prominent proteins. Next, we monitored the ability of cells to form endogenous ubiquitin conjugates in response to a single bolus of 500 μM H2O2. Oxidation was indicated by an increase in protein carbonyls (data not shown). As with NCS treatment there was an ≅45% increase in endogenous ubiquitin conjugates by 0.5-4 h (Figure 3). Subsequently the levels of endogenous conjugates declined to pretreatment levels. Thus, in extent and chronology of the response, H2O2 induces changes that are similar to those induced upon NCS treatment. Comparison of the ubiquitination response in wt and A-T cells at 0 and 30 min indicated that both wt and A-T cells mount a response to the H2O2. However, higher levels of high mass conjugates in the A-T cells at t=0 and t=30 min in identically loaded gels suggested that A-T cells had constitutively higher levels of endogenous ubiquitin conjugates than wt cells (Figure 3c,d).

Figure 3
figure 3

A transient increase in endogenous ubiquitin conjugates upon treatment with H2O2. Lymphoblastoid cells from a normal (NL553) and an A-T patient (A-T59) were treated with a single bolus of 500 μM for 0–8 h. Cells were collected at time 0, 0.5, 2, 4 and 8 h. Endogenous ubiquitin conjugates were determined by Western blot analysis using an anti-ubiquitin antiserum. To normalize for any potential differences in protein loading, the blots were stripped and reprobed with an anti-actin antibody. (a) Normal cells; (b) A-T cells. (c) Comparison of endogenous ubiquitin conjugates between wild type cells and A-T cells at time 0 and 0.5 h after H2O2 exposure. (d) Densitometric analyses of data in c, in which ubiquitin conjugate levels were normalized to actin levels

Full size image

In order to explore origins of the differences between wt and A-T cells in endogenous conjugates in response to H2O2 treatment, we monitored via thiol ester assays of related enzymes, the de novo formation of ubiquitin conjugates in response to a single bolus of 500 μM H2O2 (Figure 4). These assays provide information about levels of substrates available, overall ubiquitin conjugating capabilities, as well as about activities of specific enzymes which are involved in the activation and/or conjugation of ubiquitin to substrates. E1 thiolesters, a measure of the ability to activate ubiquitin for conjugation, are observed in the 110 kDa range and E2 thiolesters are usually observed between 15–35 kDa (Hauser et al., 1998; Jentsch, 1992; Varshavsky, 1997) (Figure 4b).

Figure 4
figure 4

A transient increase in de novo formed ubiquitin conjugates and activity of ubiquitin-activating enyzme (E1) following H2O2 exposure. Lymphoblastoid cells from a normal (NL553) and an A-T patient (A-T24) were exposed to a single bolus 500 μM H2O2 for 0–8 h. Cells were lysed and de novo formed ubiquitin conjugates and E1 activity were determined by thiolester assays as indicated in the Materials and methods using 125I-ubiquitin. (a) Autoradiography showing changes in levels of de novo formed ubiquitin conjugates in wild type cells in response to H2O2 treatment. (b) Comparison between lymphoblastoid cells from a normal (L40) and an A-T (A-T24) patient of the levels of de novo formed ubiquitin conjugates and E1-ubiquitin thiol esters at 0 and 30 min after exposure to 500 μM H2O2. The asterisk indicates a polymer of ubiquitin. (c) Densitometric analysis of per cent of high mass ubiquitin conjugates shown in b. Bars indicate high mass ubiquitin conjugate levels (ub)n-P as per cent of t=0 for NL553 for simultaneous experiments after adjusting for the amount of protein loaded on the gel. The t=0 value is not 100% because all the assays were normalized to the assay that showed the lowest value at t=0. (d) Densitometric analysis of E1-ubiquitin thiol esters of b. Bars indicate the amount of E1-ubiquitin thiol esters as per cent of t=0 for NL553 for simultaneously performed experiments after adjusting for the amount of protein loaded on the gel. The t=0 value is not 100% because for the figure we summarized data from both wt cells and both A-T cells and then normalized to the assay that showed lowest activity at t=0. The results summarize two experiments, each of which was done in duplicate

Full size image

All cell types tested show a rapid 50–150% increase in de novo formed 125I-ubiquitin conjugates that reach maximal levels after 30 min of H2O2 treatment (Figure 4a). By 4 h of H2O2 treatment the conjugation capability is approximately equal to the level found in untreated cells, and after ≅8h H2O2 treatment ubiquitin conjugate formation is lower than the levels observed in the cells prior to treatment. By this time, all of the H2O2 (t1/2=5–15 min in most cells) has been destroyed.

Next the ability to form 125I-ubiquitin conjugates de novo was compared between wt and A-T cells at specific times after exposure to H2O2 (Figure 4b). As suggested in the assays of endogenous ubiquitin conjugates shown in Figure 3c, the ubiquitination capabilities in A-T cells before H2O2 treatment are 20% higher than in wt cells (Figure 4b, compare lanes 3 and 1 and Figure 4c). In addition, A-T and wt cells mount a ubiquitination response to H2O2 such that at 30 min de novo formed conjugate levels are 63% higher in wt cells and 40% higher in A-T cells (numbers based on two experiments, each done in duplicate; Figure 4b, compare lanes 4 and 3, and lanes 2 and 1). Similar results were obtained with A-T59 and NL553 cells (not shown). These data confirm that more substrates are available for conjugation upon stress. The enhanced capability to form ubiquitin conjugates in unstressed A-T cells, the higher level of ubiquitin conjugation capability in unstressed A-T cells, and attenuated ability of these cells to respond to stress are consistent with the A-T cells being under oxidative stress and with their having an ‘aged’ phenotype (Eisenhauer et al., 1988; Huang et al., 1995; Jahngen-Hodge et al., 1997; Obin et al., 1998; Shang et al., 1997a,b; Taylor et al., 1991).

Another possibility for the differential levels of ubiquitin conjugates observed in A-T vs wt cells could be different proteolytic capabilities between the cell types. As indicated in Figure 5, in the presence of endogenous and additional ATP, proteolytic activities in A-T and wt cells were indistinguishable. This makes it unlikely that the difference in ubiquitin conjugates observed between wt and A-T cells is due to differential proteolytic capabilities. The ATP-stimulated degradation in lysates from both cell types provides further evidence that ubiquitin-dependent processes are involved since a hallmark of the ubiquitin pathway is a requirement for ATP.

Figure 5
figure 5

Comparable proteolytic capability in wild type and A-T cell lines. Wild type (NL553 and L40) and A-T (AT59 and AT24) cells were collected at density of 8×105 cells/ml and lysed in 50 mM Tris-HCl buffer, pH 7.6. Proteolytic activity in the supernatants was determined using 125I-α-lactalbumin as a substrate as described in the Materials and methods section. The data are means±s.d. of two wild type cell lines or two A-T cell lines, each of which was determined in triplicate

Full size image

E1 activity

In order to identify components of the ubiquitination machinery which may be associated with the H2O2-induced change in ubiquitin conjugates, levels of E1-thiolesters of 125I-ubiquitin were examined. E1-thiolester formation capability was 30% higher in A-T than wt cells at t=0 and there was a corresponding increase in E1-thiolesters in the samples that showed elevated levels of ubiquitin conjugates (Figure 4b,d). Similar to changes in ubiquitin conjugates, the relative increase in ubiquitin thiolesters of E1 was lower in A-T than in wt cells (19 and 25%, between t=0 and 30 min, respectively).

E2 activities

Thiolesters of at least five E2s were noted. The similarity of the profile of E2 ubiquitin thiolesters indicates that both cell types use the same E2s (Figure 4b, compare lanes 1–4). In contrast to changes in E1 levels, E2 levels in the cells were not elevated upon H2O2 exposure. Together with the elevated E1 activity, these data imply that the rate-determining reaction in the elevated conjugation observed upon H2O2 stress is catalyzed by E1. In this respect, these cells are similar to lens (Shang et al., 1997b) and HeLa cells (Stephen et al., 1996). The data also suggest that E2 activity levels are sufficient to catalyze conjugation of ubiquitin to the additional substrates to which the ubiquitin is added upon oxidative stress.

Glutathione levels in untreated and in NCS-treated cells

We previously demonstrated that the GSSG/GSH ratio is related to the activity of enzymes which are involved in ubiquitin conjugation through direct effects on active site cysteines in ubiquitin conjugating enzymes (Jahngen-Hodge et al., 1997; Obin et al., 1998; Shang et al., 1997b; Shang and Taylor, 1995). To determine if the NCS or H2O2-induced changes in ubiquitin conjugates were associated with alterations in GSH, GSSG or GSSG/GSH ratios, we determined glutathione levels in the seven groups of cells at all three densities. The levels of GSH (range: 9.9–27.4 nmol GSH/mg protein) in these cells were comparable to GSH levels in brain of A-T and wt mice (10–30 nmoles GSH/mg protein) (Kamsler et al., 2001), human lymphocytes (22 nmoles GSH/mg protein) (Lenton et al., 2000), unstressed retina pigmented epithelial cells (40–50 nmol GSH/mg protein) (Jahngen-Hodge et al., 1997; Obin et al., 1998), retina tissue (8 nmol GSH/mg protein) (Jahngen-Hodge et al., 1997), and lens epithelial cells (25 nmol GSH/mg protein) (Shang and Taylor, 1995). They are somewhat higher than GSH levels in ODS rat liver (4.2 nmol/mg protein) and kidney (0.04 nmol/mg) (Smith et al., 1999), and Emory mouse lens (1 nmol/mg protein) (Taylor et al., 1995). However, some of these differences may be due to differences in techniques used by different groups to obtain various measures.

Repeated measures ANOVA indicated a density/group interaction (P=0.05). GSH and GSSG levels in the two types of cells were then compared at each density by using Student's t-test for independent samples. In addition, each cell type was compared across all three densities by using two separate repeated measures ANOVAs with Tukey's Honestly Significant Differences for pairwise comparisons of the density levels. GSH levels tended to be lower at the highest cell density as compared with the lowest cell density (P=0.09 for A-T cells) (Figure 6a). In contrast, GSSG levels tended to remain unchanged (∼0.6 nmol/mg protein) as cells achieved higher density (Figure 6b). As a result, the GSSG/GSH ratio increased as cells achieved higher density (P=0.009 for A-T cells and P=0.12 for wt cells). These data indicate that both cells tended to achieve a more oxidized status at higher density (Figure 6c), with the GSSG/GSH ratio being slightly higher at the high density for A-T cells. While the data suggest that A-T cells are under greater oxidative stress than wt cells, the low ratios of GSSG/GSH are within ranges normally found in unstressed cells and tissues. Thus, the data indicate that under these conditions neither the wt nor the A-T cells are under acute oxidative stress (Jahngen-Hodge et al., 1997; Shang and Taylor, 1995). This GSSG/GSH ratio is below the lowest ratios of GSSG/GSH (0.19 in retina tissue) (Jahngen-Hodge et al., 1997) which were shown to be associated with lower endogenous ubiquitin conjugate levels and attenuated de novo conjugate formation capabilities (Jahngen-Hodge et al., 1997; Obin et al., 1998; Shang et al., 1997b). But distinguishing ratios lower than this is difficult because of possible changes in GSSG during sample preparation, even under the stringent conditions used in this study. Thus, to the extent to which the increased level of conjugates has an oxidation-induced etiology, these results suggest that changes in ubiquitination provide another, more sensitive, measure of redox status than is obtained from measures of GSSG/GSH concentrations. In cells treated with 500 ng/ml NCS, the GSH and the GSSG values do not change significantly with time during the 8-h treatment period, and the GSSG/GSH ratio remains ≅0.04. Thus, they are also not under severe oxidative stress. The absence of significant effects of NCS on GSSG/GSH ratios is in keeping with very limited effects of NCS on protein carbonyl levels (data not shown).

Figure 6
figure 6

Levels of GSH, GSSG and the GSSG/GSH ratio in normal and A-T lymphoblastoid cells at different cell densities. GSH and GSSG levels were determined as indicated in the Materials and methods section

Full size image

Discussion

A well established difference between wt and A-T cells results from the inability of A-T cells to respond to NCS-induced double strand breaks (Shiloh and Kastan, 2001). However, there are also reports that NCS induces oxidative damage to membrane lipids (Schor et al., 1999) and there is a single report which indicates potential damage to proteins upon NCS treatment (Edo et al., 1991). The data presented here indicate that NCS treatment causes protein damage reflected in part as alterations in the ubiquitin proteolytic pathway. Accordingly, we hypothesized that NCS treatment affects the ubiquitin pathway in wt and A-T lymphoblastoid cells in a fashion similar to what had been previously described in H2O2 exposed eye tissues and cells (Jahngen-Hodge et al., 1997; Obin et al., 1998; Shang et al., 1997b). These oxidative effects of NCS are consistent with additional roles of NCS in altering progress through the cell cycle since (1) passage through the cell cycle is controlled, in part, by ubiquitin-dependent processes (King et al., 1996, 2) enzymes involved in catalyzing the ubiquitination of proteins are redox sensitive (Jahngen-Hodge et al., 1997; Obin et al., 1998), and (3) oxidation of proteins is related to their ability to enter the ubiquitin proteolytic pathway (Shang et al., 1997b; Shang and Taylor, 1995). The data also allow further characterization of the cellular A-T phenotype.

We demonstrate here that A-T and wt lymphoblastoid cells show transient increases in levels of endogenous ubiquitin conjugates and in ability to form conjugates de novo upon stress with NCS (Figures 1 and 2, respectively) and H2O2 (Figures 3 and 4). Prior work demonstrated that the increase in ubiquitin conjugates during recovery from stress was associated with elevated levels of oxidized proteins which become substrates for the ubiquitin pathway, and this appears to pertain here as well (Shang et al., 1997b, 2001). Thus, both stressors result in more protein substrates for ubiquitination in both types of cells. Furthermore, because recovery from both stressors is associated with enhanced activities of the cellular ubiquitination machinery in both types of cells, it is probable that NCS and H2O2 alter homeostasis or the ubiquitin pathway in comparable ways.

The enhanced ability to catalyze conjugation appears to involve upregulation of E1, since higher levels of E1 thiolesters are observed in assays which show elevated levels of de novo formed ubiquitin conjugates (Figure 4b). The up-regulation of E1 thiolester may also involve enhanced expression of ubiquitin upon stress (Muller-Taubenberger et al., 1988). This is in contrast with no obvious increases in E2 activity levels.

There are also important distinctions between wt and A-T cells with respect to the ubiquitination response to stress. Significantly, untreated A-T cells appear to produce higher levels of de novo formed ubiquitin conjugates than normal cells (Figure 4b, lanes 3 vs 1). In addition, more proteins are found in endogenous ubiquitin conjugates in non-stressed A-T cells (Figure 3). The observation that A-T cells ‘idle’ at high ubiquitination levels, and their greater ability to form conjugates (without stress), are consistent with their being under enhanced oxidative stress (Gatei et al., 2001; Shang et al., 1997b).

The data also allow corroboration of prior indications that A-T cells provide useful models for the study of normal aging. We previously demonstrated higher ‘idling’ levels but limited proteolytic responses to removal of serum in in vitro aged cells (Taylor et al., 1991). Attenuation of specific proteolytic activities was noted in old vs young in vitro aged lens cells (Eisenhauer et al., 1988). In lens tissue attenuation of ubiquitination pathway responses per se was also noted upon aging (Shang et al., 1997a). To the extent to which ubiquitination is involved in proteolytic processes, the combined observations that A-T cells ‘idle’ at higher ubiquitination levels (Figure 3c,d) but mount a proportionately more muted response to stress than is observed in wt lymphoblastoid cells, is consistent with their having an aging phenotype. Interestingly, in many aging syndromes accumulations of ubiquitinated proteins are observed (Mayer et al., 1991). These include inclusions in the substantia nigra (Masliah et al., 2000), an area of the brain which is thought to be affected in ATM. Conjugates may accumulate due to inhibition of the proteasome (Bence et al., 2001; Jahngen-Hodge et al., 1992), isopeptidases, or due to damage or inappropriate assembly of the conjugates.

Kamsler et al. (2001) measured GSH and antioxidant enzyme activities and found that cultured A-T cells and ATM-deficient tissues are under oxidative stress. They also observed in the cerebellum of Atm−/− mice a slight decrease in cysteine and GSH between 1 and 2 months of age (although Atm-deficient mice started with higher levels of GSH). Levels of GSH in these cells were comparable to those found in most other tissues (Adamo et al., 1999; Ramanathan et al., 1999). In this study, GSSG/GSH ratios were used because that ratio more accurately reflects cellular redox status (Figure 6) (Jahngen-Hodge et al., 1997; Obin et al., 1998; Shang and Taylor, 1995) and because that ratio is related to activities of the ubiquitination enzymes in lens and retina cells in tissues (Jahngen-Hodge et al., 1997; Obin et al., 1998; Shang et al., 1997b). We did not observe significant differences in this ratio in A-T vs wt lymphoblastoid cells. Ubiquitin conjugate levels and ubiquitin conjugation capabilities are enhanced upon oxidative stress without significant changes in GSSG/GSH ratios, indicating that levels of ubiquitin conjugates and assays which monitor ability to form ubiquitin conjugates de novo provide a sensitive measure of cellular stress. Moreover, they suggest that this measure may be an even more sensitive indicator of oxidative stress than the GSSG/GSH ratio.

It is of interest that lymphoblastoid cells seem to mount a more rapid ubiquitination response to H2O2 than was observed in lens epithelial cells and tissues (compare the present data with Shang et al. (1997b)). Lymphoblasts are a more metabolically active cell type than lens cells. However, the reason for the differential ability of lymphoblastoid cells to mount a ubiquitination response to oxidative stress remains a subject for future study.

In summary, human lymphoblasts join lens and retina tissues in demonstrating a potentiated ubiquitination response to mild oxidative stress, suggesting that this is a generalizable reaction to mild stress. The response involves two changes to homeostatic regulation of the pathway: evolution of more substrates and increased conjugation activity, including elevated ubiquitin activating enzyme activity. The data show that NCS and H2O2 share the ability to transiently enhance endogenous ubiquitin conjugates in both wt and A-T lymphoblastoid cells. This information may be relevant to other diseases with etiologies involving oxidation, aberrant protein metabolism and protein damage. This includes cataract, maculopathies, Alzheimer's, many premature aging syndromes, and perhaps aging itself.

Materials and methods

Reagents and cells

N-ethyl maleimide (NEM) was obtained from Aldrich Chemical Co. (Milwaukee, WI, USA). AEBSF was obtained from Calbiochem-Novabiochem Corp. (La Jolla, CA, USA). o-Phosphoric acid was from Fisher Scientific (Fair Lawn, NJ, USA). Perchloric acid was from JT Baker Inc. (Phillipsburg, NJ, USA). Bathophenanthroline-disulfonic acid (BPDS) was from Sigma Chemical Company (St. Louis, MO, USA). Neocarzinostatin was from Kayaku Co. (Tokyo, Japan). Anti-actin antibody was from Santa Cruz Biochemical (Santa Cruz, CA, USA). Anti-rabbit-HRP antibody was from Jackson ImmunoResearch (West Grove, PA, USA). Anti-DNP antibody was from Dako Corp. (Carpenteria, CA, USA). Super Signal was from Pierce (Rockford, IL, USA). 125I-Protein A was from NEN Life Science Products Inc. (Boston, MA, USA). SDS–PAGE reagents were from BIORAD (Hercules, CA, USA). All other chemicals were obtained from Sigma. All buffers and chemical reagents were of the highest purity available. Wild-type (L40, NL552, NL553, 394RM) and A-T lymphoblast cell lines (A-T59, A-T24, L6) were grown in RPMI 1640 medium supplemented with 10% fetal bovine serum.

Gel electrophoresis

In order to examine the cells for relative amounts of proteins and for levels of endogenous conjugates prior to stress, the cells were collected at three concentrations (0.3×106, 0.6×106 and 1.0×106 cells per ml). The cells were sonicated in death buffer (50 mM Tris-HCl pH 7.6, 1% NP-40, 0.1% SDS, 5 mM EDTA, 10 mM NEM, and 5 mM AEBSF), and separated into water-soluble and water-insoluble (2% SDS-soluble) fractions. The protein profiles of all the wt and A-T cells were indistinguishable. For most gels, 25 ug/lane of A-T and wild-type cell lysates were subjected to 12% SDS–PAGE and stained with Coomassie Blue.

All values derived from autoradiograms or Western analyses were normalized for protein load. This was done by dividing the densitometric value for images of specific bands on autoradiograms or Western analyses by the densitometric value of a specific, invariant band which was quantified by scanning of identically loaded and run Coomassie blue stained gels or by comparison to actin staining patterns when the Western blots were reprobed for actin. All experiments were done a minimum of two times and as many as 20 times.

Endogenous conjugates

Cells at a concentration of 0.6×106 cells per ml were treated with 80 ng/ml and 500 ng/ml NCS for 0, 15 and 30 min and 1, 2, 4 and 8 h or with a single bolus of 500 uM H2O2 for 0 and 30 min, and 2, 4 and 8 h. Then the cells were collected. Cells were sonicated in death buffer (50 mM Tris-HCl pH 7.6, 1% NP-40, 0.1% SDS, 5 mM EDTA, 10 mM NEM, and 5 mM AEBSF) and separated into soluble and insoluble (2% SDS-soluble) fractions, separated by 12% SDS–PAGE, transferred to nitrocellulose and treated with anti-ubiquitin antisera. Detection was done using either 125I-Protein A or anti-rabbit-HRP secondary antibody and Super Signal chemiluminescence (Shang et al., 1997a).

Activity of ubiquitin conjugating enzymes and ability to form ubiquitin conjugates de novo using exogenous ubiquitin

All known ubiquitin conjugating enzymes form thiolesters between the active site cysteine and the carbonyl terminus of ubiquitin. The activity of ubiquitin conjugating enzymes was determined by monitoring thiolester forming capability (Shang et al., 1997b). Saturating levels of exogenous 125I-ubiquitin are used in this assay and enzyme activity is monitored by autoradiography after samples are resolved on SDS–PAGE run in the absence of mercaptoethanol (BME). The ability to form ubiquitin conjugates using saturating levels of exogenous 125I-ubiquitin was determined by quantifying levels of high mass 125I-ub-containing species, usually on gels run in the presence of BME. Levels of ubiquitin conjugates can be monitored in the presence or absence of BME because ubiquitin is attached to substrates or to itself by isopeptide bonds which are stable in reducing agents.

Two A-T and two wild-type cell lines were treated with 500 ng/ml NCS or with 500 μM H2O2 for 0 and 30 min and 2, 4 and 8 h. Cells were lysed and thiolester assays were performed with 125I-ubiquitin. The samples were run on 12% SDS–PAGE.

Proteolytic activity assay

Wild type (NL553 and L40) and A-T (AT59 and AT24) cells were collected at a density of 8×105 cells/ml. The cell pellets were lysed in 50 mM Tris-HCl buffer (pH 7.6) containing 1 mM DTT. After centrifugation at 6000 g for 20 min, protein concentrations in the supernatants were determined and adjusted to 6 mg/ml. Proteolytic activity in the supernatants was determined using 125I-α-lactalbumin as a substrate. The degradation assays were performed in a buffer containing (final concentration) 50 mM Tris-HCl, pH 7.6, 5 mM MgCl2, 2 mM DTT. Each 25 μl assay contained 15 μl supernatant and 1 μg 125I- α-lactalbumin. To determine the effect of ATP on the proteolytic activity, 2 mM ATP, 10 mM creatine phosphate and 200 μg/ml creatine phosphokinase were added to the assay buffer. The reaction was carried out at 37°C for 2 h and then stopped by adding 200 μl cold BSA solution (1%, w/v) and 50 μl 100% (w/v) TCA. Degradation rates were expressed as the per cent 125I-α-lactalbumin which were digested to TCA soluble fragments.

Glutathione detection

Cells were collected at three concentrations (0.3×106, 0.6×106 and 1.0×106 cells per ml). Cells were collected into 10% phosphoric acid with 1 mM BPDS and γ-glutamyl glutamate as a standard (Smith et al., 1999). Samples were derivatized with fluorodinitro benzene (FDNB) and separated by HPLC using a Waters uBondapak NH2 column. Mobile phase A was 80% methanol and mobile phase B was 30 g potassium acetate, 320 ml methanol and enough glacial acetic acid to adjust the pH to 4.5. Values of glutathione (GSH), and oxidized glutathione (GSSG), were calculated based on the γ-glu-glu standard in each sample.

Cells at a concentration of 0.6×106 were treated with 80 ng/ml and 500 ng/ml neocarzinostatin for 0,15 and 30 min and 1, 2, 4 and 8 h. The cells were collected into 10% perchloric acid with 1mM BPDS and analysed for GSH and GSSG as indicated above.

Protein carbonyl

Two A-T and two wild-type cell lines were treated with 80 ng/ml and 500 ng NCS/ml for 0 and 30 min and 1, 2, 4 and 8 h or with 500 μM H2O2 for 0 and 30 min and 2, 4 and 8 h. The cells were lysed, derivatized with DNPH and levels of protein carbonyls were detected on dot blots with anti-DNP antibody from DAKO (Carpinteria, CA, USA) and visualized with 125I- protein A (Shang et al., 2001).