Thioredoxin is required for S-nitrosation of procaspase-3 and the inhibition of apoptosis in Jurkat cells

Abstract

S-nitrosation is a posttranslational, oxidative addition of NO to cysteine residues of proteins that has been proposed as a cGMP-independent signaling pathway [Hess DT, Matsumoto A, Kim SO, Marshall HE, Stamler JS (2005) Nat Rev Mol Cell Biol 6:150–166]. A paradox of S-nitrosation is that only a small set of reactive cysteines are modified in vivo despite the promiscuous reactivity NO exhibits with thiols, precluding the reaction of free NO as the primary mechanism of S-nitrosation. Here we show that a specific transnitrosation reaction between procaspase-3 and thioredoxin-1 (Trx) occurs in cultured human T cells and prevents apoptosis. Trx participation in catalyzing transnitrosation reactions in cells may be general because this protein has numerous protein–protein interactions and plays a key role in cellular redox homeostasis [Powis G, Montfort WR (2001) Annu Rev Pharmacol Toxicol 41:261–295], nitrosothiol content in cells [Haendeler J, Hoffmann J, Tischler V, Berk BC, Zeiher AM, Dimmeler S (2002) Nat Cell Biol 4:743–749], and antiapoptotic signaling.

The function of NO as a signaling molecule in neurotransmission and smooth muscle relaxation via the second messenger cGMP has been firmly established (1). This occurs through selective activation of the soluble isoform of guanylate cyclase in target cells. However, increasing attention is being focused on cGMP-independent NO signaling pathways involving S-nitrosation, a posttranslational, oxidative addition of NO to cysteine residues of proteins (2). S-nitrosation converts the thiol of cysteine residues of proteins to a nitrosothiol, and this simple modification alters protein function both in vitro and in vivo (3). The outcome is especially noteworthy when the modified cysteine directly participates in the catalytic chemistry of an enzyme, as is the case with the caspase family of cysteine proteases (4–6).

A paradox of S-nitrosation is that only a small set of reactive cysteines are modified in vivo despite the promiscuous reactivity NO exhibits with thiols (7). This precludes the reaction of free NO as the primary mechanism of S-nitrosation in vivo. Although S-nitrosation has been compared with phosphorylation (8), several key questions remain to be answered to justify this comparison. In phosphorylation, families of kinases and phosphatases tightly regulate protein modification with spatial and temporal specificity. Additionally, ATP serves as the common phosphoryl donor. Given the indiscriminate reactivity of NO in solution, the requisite specificity depends on proteins that can capture NO to form a common nitrosothiol donor and proteins that can transfer and remove NO. Such proteins have not yet been described, and, until they are found to function in cellular settings, our understanding of cGMP-independent NO signaling will be missing key molecular details.

The studies reported here provide the first of the missing in vivo details with a focus on thioredoxin (Trx) and caspase-3. There are several reports regarding NO inhibition of apoptosis, and one mechanism to account for this effect on apoptosis is S-nitrosation of the caspase proteases (9–13). To support this, S-nitrosation of the catalytic cysteine of procaspase-3 and procaspase-9 has been reported to occur in human cell lines (4, 14), and treatment of T cells with a caspase-3/7 selective transnitrosating agent led to inhibition of caspase-3 activity and apoptosis (15). The potential involvement of Trx in this signaling cascade also has literature support. S-nitrosated Trx (Trx-SNO) was shown to be responsible for an increase in the overall protein SNO content of endothelial cells and for the redox and antiapoptotic functions of Trx (16). Trx-SNO also exhibits cardioprotective properties after ischemia–reperfusion events, which induce apoptosis (17). Herein evidence is presented that (pro)caspase-3 and Trx-SNO interact and that this interaction is critical to caspase-3 activity and apoptosis in a human T cell lymphoma line (Jurkat). During the course of the protein–protein interaction, the catalytic cysteine of procaspase-3 undergoes a transnitrosation reaction with a surface-accessible cysteine of Trx-SNO. This is the first demonstration of a specific transnitrosation reaction between endogenous proteins in cells and is consistent with the type of specificity expected in a signaling pathway (8). Although control over apoptosis is significant, Trx is ubiquitous and known to interact with many proteins, suggesting a general role for Trx in cGMP-independent signaling.

Results and Discussion

Although Trx-SNO has been detected and previously shown to be antiapoptotic in human endothelial cells (16), there is no precedent for the presence of Trx-SNO in T cells. Cultured human T cells have been shown to express two isoforms of NO synthase, inducible NO synthase and endothelial NO synthase (18), and addition of a NO synthase inhibitor induces apoptosis in primary T cell lymphoma cells (19). This supports the presence of a NO-dependent antiapoptotic mechanism in T cells. After treatment with FasL, etoposide, and IFN-γ, Trx-SNO was detected, but not in untreated cells (Fig. 1A). Procaspase-3–SNO was found in unstimulated Jurkat cells, as previously reported (4), in addition to cells treated with IFN-γ and etoposide (Fig. 1B). Treatment with the IFN-γ and etoposide led to a marked decrease in procaspase-3–SNO, relative to total procaspase-3. Processed caspase-3–SNO was also detected, but the Western blot band intensity was too low to allow for reliable quantification (data not shown). To probe molecular details of the Trx-dependent nitrosothiol transfer to procaspase-3, a strategy was devised to reduce the high level of endogenous Trx without inducing apoptosis (20, 21). Three V5 epitope-tagged constructs (Trx-WT, Trx-C73S, and Trx-E70A/K72A) were stably transfected into Jurkat cells [supporting information (SI) Fig. 5]. The Trx-C73S mutation should act as a dominant negative that binds to procaspase-3 but is unable to participate in S-nitrosation chemistry (6). Nearby polar residues, E70 and K72, were expected to be involved in the Trx/procaspase-3 interaction; thus, the double alanine mutant should maintain S-nitrosation with a loss of contact with procaspase-3. Jurkat cells transfected with the dominant negative and loss-of-contact Trx mutants were as viable as cells transfected with Trx-WT (trypan blue exclusion; data not shown). Importantly, these Trx proteins retained efficient insulin reduction activity in vitro (Table 1).

Fig. 1.

Trx transnitrosates procaspase-3 in Jurkat cells. (A) Trx is S-nitrosated in stimulated Jurkat cells. The Bradford assay was used to ensure equal protein loading. Trx_red (C32/C35) and Trx_ox are visible. Lane 1, no treatment; lane 2, FasL; lane 3, etoposide; lane 4, IFN-γ. (B) Procaspase-3 is S-nitrosated under basal conditions; the method was the same as in A but using anti-caspase-3. Lane 1, no treatment; lane 2, etoposide; lane 3, IFN-γ. (C) Trx is S-nitrosated on C73 in Jurkat cells. Cells were stably transfected with V5-tagged Trx before immunoprecipitation using anti-V5 and the biotin switch. Lane 1, V5-Trx-WT; lane 2, V5-Trx-E70A/K72A (loss of contact); lane 3, V5-Trx-C73S (dominant negative); lanes 4–6, same order as lanes 1–3. + and − indicate etoposide (etop.) treatment. (D) Trx is involved in the formation of procaspase-3–SNO. Jurkat cells were stably transfected with V5-Trx mutants before knockdown of endogenous Trx using shRNA (3′ UTR). Densitometric analysis was performed on blots probed with HRP-conjugated neutravidin (NA-HRP) and anti-caspase-3 after an anti-caspase-3 immunoprecipitation followed by the biotin switch method. Black bars, nontransfected Jurkat cells; blue bars, cells transfected with shRNA; gray bars, cells transfected with V5-Trx-WT and shRNA; red bars, V5-Trx-C73S and shRNA; purple bars, V5-Trx-E70A/K72A and shRNA. + and − indicate etoposide (etop.) treatment. Asterisks indicate statistical significance between samples that were untreated (∗) and treated with etoposide (∗∗). P value was calculated by ANOVA. (E) Same as in D, but lysate was tested for AcDEVD-ase activity. Values were normalized to a sample that did not receive etoposide. The color scheme identical to that of D. Asterisks indicate significance between Jurkat cells that were not transfected with V5-Trx (∗, Student's t test) and cells that received a V5-Trx construct in addition to the shRNA plasmid (∗∗, ANOVA).

Table 1.

Characterization of Trx mutants

Trx	Activity	% SNO	Location	kCasp inact	KD Casp3
WT	1.00 ± 0.06	88 ± 11	C73	77.1 ± 4.5	2.5 ± 0.5
D64A	0.74 ± 0.07	89 ± 7	ND	5.3 ± 0.8	>100
C69S	1.08 ± 0.07	96 ± 12	C73	11.5 ± 1.5	ND
E70A	0.86 ± 0.10	100 ± 8	C73	6.5 ± 0.4	60 ± 10
E70A/K72A	0.78 ± 0.04	88 ± 14	ND	4.6 ± 0.3	>100
C73D	0.49 ± 0.05	30 ± 23	NA	NA	30 ± 10*
C73S	1.02 ± 0.06	5 ± 6	NA	NA	2.4 ± 0.9

Human Trx and mutants were characterized by activity [insulin reduction assay, 0.5 mM DTT (pH 7.5), 25°C, normalized to WT], percentage of nitrosothiol formed after reaction with S-nitrosoglutathione [10 equivalents, 20 min, 37°C, reduced light, normalized to total protein, as measured by the 2,2′-azino-bis(3-ethylbenzthiazoline-6-sulphonic acid)/HgCl₂ and BCA assays], location of nitrosothiol (biotin switch method, trypsin/MALDI), rate of caspase-3-C163 transnitrosation [fluorimetry, 75 μM AcDEVD-amc (pH 7.5), 25°C, reduced light, M⁻¹·s⁻¹], and affinity for caspase-3 (ITC, μM). The ITC detection limit is estimated at 100 μM. Error is SD; NA, not applicable; ND, not determined.

*Fit only to a two-site binding model (others were fit to a one-site model).

Immunoprecipitation followed by the biotin switch method allowed for the relative amount of Trx-SNO and procaspase-3 to be determined for each of the V5 transfectants. Immunoprecipitation followed by the biotin switch method using the anti-V5 antibody indicated that Trx residue C73 is S-nitrosated in etoposide-stimulated Jurkat cells (Fig. 1C). These cells were then transiently transfected with a plasmid encoding a short hairpin RNA (shRNA) that targeted the 3′ UTR of the Trx mRNA, allowing the specific knockdown of endogenous Trx. Background Trx was reduced by ≈65% (SI Fig. 5), similar to a previous report (21). The V5-Trx transfectants were then assayed for relative procaspase-3 processing. Apoptosis was induced 36 h after knockdown, followed by an anti-caspase-3 Western blot, which revealed that Trx-WT and Trx-C73S transfected Jurkat cells both processed ≈40% of procaspase-3 to caspase-3 whereas Trx-E70A/K72A transfected cells converted 25–30% of procaspase-3 (data not shown). The extent of S-nitrosation of procaspase-3 was determined in similarly transfected cells that were also stimulated to undergo apoptosis (Fig. 1D). We observed the largest amount of procaspase-3–SNO in nontransfected Jurkat cells. The smallest quantity was found in the cells transfected with the dominant negative (V5-Trx-C73S) construct, and levels were intermediate in the loss-of-contact (V5-Trx-E70A/K72A) mutant.

Similarly treated cells were assayed for caspase-3 activity (AcDEVD-ase) after etoposide treatment. The activity correlated negatively with the relative amount of procaspase-3–SNO (Fig. 1E and SI Fig. 6). Nontransfected cells exhibited the lowest AcDEVD-ase activity. Addition of the shRNA plasmid to cells resulted in a modest increase in AcDEVD-ase activity, which is characteristic of the antiapoptotic actions of Trx (22). Knockdown of endogenous Trx in cells expressing the dominant negative mutant yielded the highest activity whereas the loss-of-contact mutation was again intermediate. The lower AcDEVD-ase activity in cells expressing Trx-E70A/K72A, relative to Trx-C73S, likely originates from less extensive zymogen processing, as mentioned above. Relative to Trx-WT, the E70A/K72A mutation displayed more rapid AcDEVD-amc turnover, indicative of a weaker affinity for procaspase-3. Importantly, cells transfected with Trx-WT and the knockdown plasmid had little effect on AcDEVD-ase activity. Flow cytometry was used to quantify the extent of apoptosis in cells not transfected with the shRNA plasmid. In accord with AcDEVD-ase activity, Trx-WT transfected Jurkat cells treated with etoposide displayed a marked resistance to apoptosis (29%) compared with cells expressing Trx-C73S and Trx-E70A/K72A (43% and 57%, respectively).

In vitro experiments were carried out by using recombinant proteins to place quantitative parameters on the activity of the transfected Trx mutants. Additional mutants of Trx were also constructed (SI Table 2) to further explore the protein–protein interaction and chemistry between Trx and procaspase-3. During bacterial overexpression, procaspase-3 undergoes autocatalytic processing to caspase-3 (6). The use of procaspase-3 for in vitro characterization was not possible because the addition of inhibitors to prevent autocatalytic processing was not practical. Caspase-3 was used for these experiments, which was fortuitous because it permitted the use of an established activity assay to quantify rates of nitrosothiol transfer (6). Trx mutants were constructed in which the mutations were in close proximity to the cysteine involved in nitrosothiol transfer with caspase-3, Trx-C73 (Table 1). The mutants were characterized by an activity assay (insulin reduction), extent and location of nitrosothiol formation (S-nitrosoglutathione), rate of caspase-3 inactivation, and affinity for caspase-3 [isothermal calorimetry (ITC)]. Four mutants exhibited a modest decrease in reductive capabilities (D64A, E70A, E70A/K72A, and C73D) whereas the others were essentially the same as WT. Upon reaction with S-nitrosoglutathione, all Trx mutants formed 1 mole equivalent of nitrosothiol on C73 except the C73 mutants, in which case nitrosothiols were not formed. Employing a previously described fluorimetry assay (6), Trx mutants that formed nitrosothiols were assayed for caspase-3 transnitrosation. Comparison to Trx-WT showed that all mutants tested exhibited a decrease in the rate of caspase deactivation with Trx-C69S being the most efficient Trx mutant that serves as a caspase-3 nitrosating agent.

A specific, catalyzed transnitrosation reaction between Trx and (pro)caspase-3 requires the formation of a protein–protein complex. To establish a biologically relevant association, the Trx/(pro)caspase-3 interaction was first assessed by coimmunoprecipitation. Procaspase-3 coimmunoprecipitated from both etoposide and vehicle-treated Jurkat cells when using an anti-Trx antibody (SI Fig. 7). As demonstrated in a silver-stained duplicate gel, many proteins coimmunoprecipitate with Trx, consistent with reports showing many Trx–protein interactions (22). The presence of procaspase-3 was verified by mass spectrometry. The reverse coimmunoprecipitation experiment was also performed (SI Fig. 7). Caspase-3 was not purified by coimmunoprecipitation within the limit of detection, suggesting that procaspase-3 may have a tighter affinity to Trx.

The association was further characterized by using fluorescence polarization and ITC with recombinant Trx and caspase-3. The Oregon green fluorophore was installed on the N terminus of Trx for fluorescence polarization measurements (SI Fig. 8). Binding to caspase-3 was detected with Trx-WT but not Trx-E70A or Trx-E70A/K72A. ITC measurements confirmed these observations; the K_D for caspase-3 with the single Trx-E70A mutant sustained an ≈30-fold loss in affinity whereas the association was not measurable with Trx-E70A/K72A (Table 1). This established that the double Trx mutant indeed served as a loss-of-contact mutation (Fig. 2). The location of H-bonding moieties was certainly critical, as a nonconservative surface mutation, Trx-C73D, led to a weaker interaction whereas the conservative Trx-C73S mutation had an affinity equal to WT (Table 1 and SI Fig. 9), supporting the use of Trx-C73S as a dominant negative (Fig. 4). Attempts to quantify the association between Trx and procaspase-3 failed because of autocatalytic processing at the high concentration required for ITC. Further confirmation of the loss-of-contact mutation was obtained by using recombinant His-tagged Trx protein to affinity-purify endogenous caspase-3 from apoptotic Jurkat cell lysate (Fig. 3A). The p17 subunit of caspase-3 was detected by anti-caspase-3 Western blot when using His-Trx-WT as bait. Conversely, if His-Trx-E70A/K72A was used as bait or if no His-Trx was immobilized to Ni-NTA, caspase-3 was not detected, indicating that residues E70 and K72 of Trx contribute to the affinity for caspase-3 in the cellular milieu. An identical experiment was attempted using His-Trx-C69S and His-Trx-C73S as bait. Caspase-3 was detected in both cases (data not shown), but a Coomassie-stained duplicate gel indicated the Trx-C69S construct contained as little as 25% of the Trx input after completion of the experiment (Fig. 3B) whereas the Trx-WT (Fig. 3B) and Trx-E70A/K72A (data not shown) samples contained ≈100% of the Trx input. Trx-C69S possibly underwent proteolysis into fragments not resolvable by 10–20% PAGE (<4 kDa) during the course of the experiment by virtue of the location of the mutation. Residue C69 is buried in a hydrophobic pocket of Trx (23–26). Mutation to serine, a much more polar residue, disrupted the hydrophobic packing, as confirmed by determining the ΔG_fold°. Trx-WT and Trx-C73S had an equal ΔG_fold° (CD) whereas the ΔG_fold° for Trx-C69S was destabilized by ≈25%, as measured by CD and tryptophan fluorescence (Fig. 4). This property of the Trx-C69S mutant may render it unsuitable for cellular studies (16). Interestingly, a recent structure clearly shows a nitrosothiol on Trx-C69; however, this Trx-SNO could only be stably formed at pH 9.0. The greatly enhanced reactivity of thiols at this pH requires some caution in the extension of these results to physiological function.

Fig. 2.

Model for nitrosothiol transfer between Trx and caspase-3. After apoptotic stimulation, endogenous WT Trx binds to and transfers ⁺NO (blue and red spheres) to caspase-3, which leads to inhibition of apoptosis (healthy cell is shown in green). Cells transfected with a dominant negative (D.N.) or loss-of-contact (L.C.) Trx mutant do not inhibit apoptosis (apoptotic cell is shown in blue). Residues of Trx involved in caspase binding and transfer are shown in stick. PyMol and Protein Data Bank entries 1ERT and 1NME were used to construct this figure.

Fig. 4.

Urea denaturation of Trx proteins. Trx-C69S (open circles) has a weaker fold relative to WT (filled squares) and Trx-C73S (filled circles). Differences in the baseline CD signals, visible at lower urea concentration, were due to variance in the protein concentration (confirmed by the BCA method, quantitative amino acid analysis, and dry weight). Measurement of ΔG_fold° is independent of protein concentration. Intrinsic tryptophan fluorescence gave similar results (data not shown).

Fig. 3.

Trx and (pro)caspase-3 interact. (A) Anti-caspase-3 Western blot from Ni-NTA magnetocapture assay. WT His-Trx purifies caspase-3 from apoptotic Jurkat lysate. Lane 1, nontreated; lane 2, TNF-α; lane 3, staurosporine (His-Trx-E70A/K72A does not purify caspase-3); lanes 4–6, same as WT; lane 7, no His-Trx loaded (n.l.). (B) Coomassie-stained gel from an experiment identical to that in A. Lanes 1–3, same treatment as in A; lanes 4–6, same treatment as in A but using His-Trx-C69S as bait. (C) Sequence of His-Trx. Caspase-3 cut site is highlighted. (D) Coomassie-stained gel of caspase-3 proteolysis assay. Leftmost lane, molecular weight marker; lane 1, Trx-WT; lane 2, Trx-C69S; lane 3, Trx-D64A.

The observation that Trx-C69S was proteolytically processed led us to pursue the possibility that the Trx/caspase-3 interaction could be due to Trx binding to the active site of caspase-3 with the potential for being a proteolytic substrate. Human Trx contains a caspase-3 recognition motif, DxxD (27) (Fig. 3C). All Trx proteins were screened in an in vitro caspase-3 proteolysis assay, and all were processed by recombinant caspase-3 at a single site to the same low efficiency as Trx-WT except for Trx-C69S and Trx-D64A (Fig. 3D). Trx-D64A was not a caspase-3 substrate, and the C69S mutant was processed ≈5-fold faster than WT. Mass spectrometry and N-terminal sequencing were used to verify the site of proteolysis, ₆₁DCQD₆₄/(V) (data not shown). A plausible interpretation of this result is that lower stability of the Trx-C69S mutant has increased the accessibility of the proteolytic site. It is important to note that Trx-C62 is completely buried in several published Trx structures (23–26), including a structure were a nitrosothiol was found to be located on Trx-C62 (28). In this structure, Trx-C32/C35 were found in an intramolecular disulfide and Trx-C73 was found in an intermolecular disulfide, suggesting that nitrosothiols are likely formed on C32/35 and C73 under kinetic control.

The above-mentioned proteolysis of Trx by caspase-3 complicates the study of the Trx/caspase-3 interaction because there is potentially more than one interaction mode: both Trx and (pro)caspase-3 have peptide-binding clefts. The contribution of Trx binding to the caspase-3 active site was expected to be minimal because Trx was a poor substrate for caspase-3 in vitro (Fig. 3D) and no previous data in the literature suggested that Trx is degraded during apoptosis. Trx also has a less restrictive binding site, which allows a myriad of other proteins to be substrates for oxidoreductase (22) and, possibly, transnitrosation chemistry (16). Additional ITC measurements were performed to elucidate the relative contribution of the two possible interaction modes. First, the active site of caspase-3 was fully alkylated with the irreversible chloromethylketone inhibitor AcDEVD-cmk. The alkylated protein had an affinity for Trx identical to that of unmodified caspase-3 (SI Fig. 9). Also, Trx still bound a synthetic peptide comprising residues 159–176 of caspase-3 with nominal loss of affinity, whereas Trx-E70A/K72A bound this peptide much more weakly, near the detection limit (SI Fig. 9).

The results reported here strongly suggest that Trx residues E70 and K72 are required for procaspase-3 affinity and that C73 is the residue involved in nitrosothiol formation and transfer to C163 of procaspase-3 (Fig. 4). Unfortunately, attempts to map the modified cysteine of Trx in Jurkat cells by a method other than mutagenesis (i.e., mass spectrometry) were unsuccessful. Because of an incomplete knockdown of endogenous Trx, the data presented here represent the minimum contribution of Trx residues C73, E70, and K72 in the association and transnitrosation reaction with procaspase-3. It is likely that their contribution is greater because, after the knockdown, the transfected cells express an ≈1:1 ratio of endogenous Trx to V5-Trx mutant (anti-Trx Western blot; data not shown).

In summary, we have characterized the interaction and transnitrosation reaction between two oncogenic proteins, Trx and procaspase-3. In cultured human T cells procaspase-3 is found to be extensively S-nitrosated on the catalytic cysteine (4). Evidence is presented that shows that this species is formed by transnitrosation with Trx-C73-SNO. Loss of this residue or mutation of nearby residues in Trx that confer affinity for procaspase-3 results in a decrease in the relative amount of procaspase-3–SNO and a concurrent increase in AcDEVD-ase activity and apoptosis upon etoposide stimulation (Fig. 4). Future work is necessary to fully uncover the role of Trx in other transnitrosation reactions of biological significance. Trx is involved in the nitrosothiol content of endothelial cells (16) and HeLa cells (29), and thus we anticipate that Trx and related proteins such as protein disulfide isomerase (30–32) will mediate specific nitrosothiol relay in many cell types for the purposes of signal transduction.

Materials and Methods

Additional Details.

More details can be found in SI Methods.

Protein Purification and Characterization.

Recombinant human Trx, Trx-C73S, and caspase-3 were expressed and purified as previously described (6). Recombinant expression of procaspase-3 leads to self-processing (6) in Escherichia coli, which makes purification of the zymogen protein difficult without the addition of tightly bound inhibitors or resorting to mutagenesis (33). Procaspase-3 was purified in the same manner as caspase-3, but we included a high concentration of malonate (30 mM), which has been described to inhibit autocatalytic processing (34, 35). A 3-h expression led to an ≈50/50 mixture of p32/p29:p17/p12, which can be partially separated by S75 gel filtration (data not shown; purity 80–85%). Site-directed mutants of Trx were introduced by using the QuikChange method (Stratagene, La Jolla, CA). Primers used are given in SI Table 2. The insulin reduction Trx activity assay (36, 37), 2,2′-azino-bis(3-ethylbenzthiazoline-6-sulphonic acid) (ABTS) nitrosothiol quantification assay (38), biotin switch method (39), trpysin/MALDI nitrosothiol peptide mapping, and rate of caspase-3 deactivation were performed as described (6).

Cell Culture and Treatments.

Jurkat cells were maintained under reduced light conditions in a 5% CO₂, water-saturated atmosphere at 37°C in RPMI medium 1640 with glutamine (Gibco, Carlsbad, CA) supplemented with 10% FBS (Invitrogen, Carlsbad, CA), 100 units/ml penicillin, and 100 μg/ml streptomycin. Apoptosis was initiated by harvesting the cells via centrifugation and resuspending in media containing reduced serum (1% FBS) and the following proapoptotic agents for 10 h: 1 μg/ml etoposide (Sigma, St. Louis, MO), 250 units/ml recombinant IFN-γ (Sigma), 1 μM staurosporine (Cayman Chemical, Ann Arbor, MI), or 1 ng/ml recombinant human FasL_134–281 (FasL; Sigma). Cells were then harvested by centrifugation and washed twice (PBS) before further experimentation.

Coimmunoprecipitations.

Jurkat cells were lysed in RIPA/HEN buffer (100 mM Hepes/150 mM NaCl/1 mM EDTA/1 mM NaF/100 μM neocuproine/5% glycerol/1% Triton X-100/1% sodium deoxycholate, pH 7.4) containing Complete (Roche, Indianapolis, IN) protease inhibitor mixture with rocking at 4°C. The insoluble fraction was removed by centrifugation (16,000 × g for 15 min at 4°C). The cellular supernatants in a volume of 500 μl were then precleared by addition of 30 μl of protein G-Sepharose 4B (equilibrated in RIPA/HEN; Invitrogen) for 30 min at 4°C with rocking. The beads were collected by centrifugation, and the supernatant was transferred to a clean tube containing either 40 μl of the rabbit E87 monoclonal anti-caspase-3 antibody (Abcam, Cambridge, MA) or 25 μl of goat polyclonal anti-Trx antibody (R & D Systems, Minneapolis, MN) and rocked for 90 min at 4°C, at which point 40 μl of protein G-Sepharose was added (equilibrated in RIPA/HEN) and rocked for an additional 90 min. The immunoprecipitate was centrifuged (400 × g for 2 min), and the supernatant was decanted. The beads were rapidly washed three times with 1 ml of RIPA/HEN before elution of the immunocomplex from the beads using SDS/PAGE loading buffer. The eluate was separated on duplicate SDS/PAGE gels with the first gel silver-stained (SilverQuest; Invitrogen) and the second transferred to nitrocellulose. The blot was blocked for 15 h (3% BSA in PBS/Tween) and subsequently probed for the presence of the partner protein by using either the goat polyclonal anti-Trx antibody or the rabbit polyclonal anti-caspase-3 antibody (Cayman Chemical). Note that Trx did not coimmunoprecipitate when using the Cayman Chemical antibody.

AcDEVD-ase Activity.

Under reduced light conditions, Jurkat cells were lysed as described above. Total soluble protein was quantified by the Bradford assay and normalized to 1.0–2.0 mg/ml by dilution into RIPA/HEN and the solution supplemented with AcDEVD-amc (75 μM) (27). AcDEVD-ase (caspase-3/7) activity was measured on a Molecular Devices Spectra Max Gemini XS (λ_ex = 380 nm, λ_em = 460 nm) using 96-well black fluorimetry plates (Corning, Corning, NY) in triplicate.

Immunoprecipitation for Detection of S-Nitrosation.

Under reduced light, Jurkat cells were lysed and precleared as above. The beads were collected by centrifugation, and the supernatant was transferred to a clean tube containing 25 μl of goat polyclonal anti-Trx (R & D Systems), 6 μl of mouse monoclonal anti-V5 (Invitrogen), or 30 μl of rabbit polyclonal anti-caspase-3 (Cayman Chemical). The antibody was allowed to bind for 90 min at 4°C with rocking, after which the protein G beads were again added (40 μl, equilibrated in RIPA/HEN) for an additional 90 min binding at the same temperature (reduced light). The immunoprecipitate was harvested by centrifugation (400 × g for 2 min), and the supernatant was decanted. The beads were washed three times with 1 ml of RIPA/HEN before elution of the immunocomplex from the beads using the biotin switch blocking buffer (contains 2.5% wt/wt SDS and 30 mM NEM). The elution proceeded under reduced light at room temperature for 15 min with frequent vortexing. After removal of the beads by centrifugation, the supernatant was subjected to the biotin switch method, starting with the blocking step (39, 40).

Transfections.

Low-passage Jurkat cells were harvested by centrifugation and washed with PBS. Cells (5 × 10⁶) were transfected with 10 μl of DMRIE-c Reagent (Invitrogen) containing purified plasmid (2 μg) as recommended by the manufacturer. Selection with 2–5 μg/ml blasticidin (Invivogen) was initiated 48 h later. Stable expression of the V5 epitope-tagged Trx was assayed by Western blot (anti-V5-HRP; Invitrogen) after 2 weeks of selection. Knockdown of endogenous Trx was then performed on the cells stably expressing V5-tagged Trx and nontransfected cells. Approximately 1 × 10⁶ cells were pelleted and washed with PBS. Transient transfection with 1.5 μg of shRNA harboring vector was performed by using the Nucleofector Kit V (Amaxa, Gaithersburg, MD) and electroporation setting S-18.

Caspase-3 Proteolysis Assay.

All of the recombinant Trx proteins described in this work were tested for the ability to be proteolytically digested by recombinant caspase-3. Trx (25 μg) and caspase-3 (1 μg, 1:50 molar ratio) were allowed to react for 3 h at 37°C (PBS/5 mM DTT, pH 7.4) before quenching with loading buffer, separation by SDS/PAGE, and staining by Coomassie. The cut site was verified to be C-terminal to D64 of Trx, which lies in a DxxD caspase-3 recognition motif (27) and was verified by MALDI mass spectrometry and N-terminal sequencing (data not shown). A subsequent experiment was set up as above, but aliquots were removed from the reaction at 30-min intervals, and rate of cleavage was analyzed by densitometry.

CD.

Trx-WT, Trx-C69S, and Trx-C73 (40–50 μg/ml) were equilibrated in CD buffer (25 mM KH₂PO₄/75 mM NaCl, pH 7.0) for 15 h with varying urea concentrations (0–10 M). CD measurements were acquired at 25°C on an Aviv 62DS spectropolarimeter with a Peltier temperature-controlled sample holder using a 1-cm cuvette. The CD signal of each sample was monitored (222 nm) with respect to increasing urea concentration. The final urea concentration of each sample was verified by refractive index. The data were fit to a two-state folding model (Kaleidagraph), and the ΔG of folding in water (ΔG_fold°) was calculated.

Ni-NTA Magnetocapture Assay.

His-tagged Trx (WT, C69S, and E70A/K72A) were used as bait to purify endogenous caspase-3 from apoptotic Jurkat cells using Ni-NTA magnetic agarose beads (Qiagen, Valencia, CA). A typical experiment was as follows: Jurkat cells were serum-deprived (1% FBS) and treated with 50 ng/ml TNF-α (Calbiochem, San Diego, CA), 1 μM staurosporine (Cayman Chemical), or buffer for 15 h. The cells were harvested by centrifugation (400 × g for 2 min at 4°C) and subjected to two freeze–thaw cycles in N₂(l). The lysed cells were further centrifuged at 16,000 × g for 15 min at 4°C to remove the insoluble fraction. The supernatant was precleared using Ni-NTA Superflow (50 μl; Qiagen) that was equilibrated in binding buffer (25 mM NaH₂PO₄/250 mM NaCl/10 mM imidazole/5% vol/vol glycerol/0.05% wt/wt CHAPS, pH 8.0) for 30 min at 4°C with rocking. The beads were collected by centrifugation (1,500 × g for 2 min at 4°C), and the cleared supernatant was transferred to a clean tube. Recombinant His-Trx-WT, His-Trx-C69S, and His-Trx-E70A/K72A (20 μg) were added to separate samples with binding buffer added to serve as a control. The His-Trx proteins were allowed to bind for 1 h at 4°C with rocking before addition of magnetic Ni-NTA beads (30 μl equilibrated in binding buffer for 1 h at 4°C with rocking). Washes (three times, 1 ml) were performed as quickly as possible, to minimize loss of caspase-3, using the Qiagen magnetic chamber for microcentrifuge tubes. The proteins were eluted by using binding buffer supplemented with imidazole to a final concentration of 0.5 M, separated by SDS/PAGE, transferred to nitrocellulose, and blocked for 15 h (3% BSA and PBS/Tween). The blots were probed by using a rabbit polyclonal anti-caspase-3 antibody (1:1,000; Cayman Chemical) and the HRP-conjugated goat anti-rabbit antibody (Pierce, Rockford, IL).

Flow Cytometry.

Flow cytometry was performed on etoposide-treated (2.5 μg/ml for 10 h) Jurkat cells stably expressing the V5-tagged Trx constructs: WT, C73S, and E70A/K72A. Apoptosis and necrosis were assayed by using 50,000 intact cells labeled with propidium iodide and Alexa Fluor 488-conjugated annexin V, as suggested by the manufacturer's instructions (Vybrant Kit #2; Invitrogen). Annexin V binds specifically to phosphatidylserine and is downstream of caspase activation in Jurkat cells (41). Results were quantified with a FACSCalibur flow cytometer (BD Biosciences, San Jose, CA) using CellQuest software. It is important to note that we were unable to quantify apoptosis by using the Vybrant assay after transfection of the cells with the Trx knockdown plasmid because this plasmid contains a GFP:zeocin resistance fusion protein. GFP has nearly identical fluorescence properties as Alexa Fluor 488. In these cases, AcDEVD-ase activity was used to estimate the extent of apoptosis.

Abbreviations

Trx

thioredoxin

Trx-SNO

S-nitrosated Trx

shRNA

short hairpin RNA

ITC

isothermal calorimetry.