Enzymatic Mechanism Controls Redox-mediated Protein-DNA Interactions at the Replication Origin of Kinetoplast DNA Minicircles*

Kinetoplast DNA (kDNA) is the mitochondrial DNA of trypanosomatids. Its major components are several thousand topologically interlocked DNA minicircles. Their replication origins are recognized by universal minicircle sequence-binding protein (UMSBP), a CCHC-type zinc finger protein, which has been implicated with minicircle replication initiation and kDNA segregation. Interactions of UMSBP with origin sequences in vitro have been found to be affected by the protein's redox state. Reduction of UMSBP activates its binding to the origin, whereas UMSBP oxidation impairs this activity. The role of redox in the regulation of UMSBP in vivo was studied here in synchronized cell cultures, monitoring both UMSBP origin binding activity and its redox state, throughout the trypanosomatid cell cycle. These studies indicated that UMSBP activity is regulated in vivo through the cell cycle dependent control of the protein's redox state. The hypothesis that UMSBP's redox state is controlled by an enzymatic mechanism, which mediates its direct reduction and oxidation, was challenged in a multienzyme reaction, reconstituted with pure enzymes of the trypanosomal major redox-regulating pathway. Coupling in vitro of this reaction with a UMSBP origin-binding reaction revealed the regulation of UMSBP activity through the opposing effects of tryparedoxin and tryparedoxin peroxidase. In the course of this reaction, tryparedoxin peroxidase directly oxidizes UMSBP, revealing a novel regulatory mechanism for the activation of an origin-binding protein, based on enzyme-mediated reversible modulation of the protein's redox state. This mode of regulation may represent a regulatory mechanism, functioning as an enzyme-mediated, redox-based biological switch.

gering of antioxidation responses has been reported for several cysteine-containing proteins. It has been shown that upon elevation of intracellular hydroperoxides level, the yeast transcription factor Yap1 receives a signal from the glutathione peroxidase-like enzyme Gpx3, which serves as sensor and transducer of hydroperoxide signal to Yap1 that consequently activates the expression of antioxidant genes. Both the sensor and the regulator are reduced by thioredoxin (24,25). Multimerization of Ask1 (apoptosis signal-regulated kinase-1), induced by hydrogen peroxide, and its reduction by thioredoxin regulates the H 2 O 2 -induced c-Jun NH 2 -terminal kinase activation and apoptosis (26). The Schizosaccharomyces pombe transcription factor Pap1 (27,28) was reported to regulate antioxidant gene transcription in response to a low concentration of H 2 O 2 , through the action of a Tpx1 peroxiredoxin, which functions as an upstream activator, transferring a redox signal to Pap1. It has been reported that activation of the function of Escherichia coli Hsp33, a redox-regulated molecular chaperone, requires the presence of reactive oxygen and hydroxyl radicals, which are sensed by the thiol-containing zinc center of the protein. Upon exposure to oxidative stress, the protein undergoes a conformational rearrangement and dimerizes to yield its functionally active structure in a process that is overall temperature-dependent (15,29,30).
In trypanosomatids, the major enzymatic pathway that regulates the intracellular redox state is based on tryparedoxin (TXN) and tryparedoxin peroxidase (TXNPx). In this pathway, glutathione is converted into an NЈ,N 8 -bis(glutathionyl) spermidine adduct, designated trypanothione (31,32). The oxidized form of trypanothione, which is reduced by trypanothione reductase (TR), reduces TXN. Reduced TXN reduces in turn TXNPx, a member of the peroxiredoxin family. TXN shares low homology with thioredoxins, displayed mainly in the active site, which contains the Cys-X-X-Cys motif (33). The known function of TXNPx is to detoxify hydrogen peroxide, defending the cell from oxidative stress (34,35).
We have previously reported that redox affects both the binding of UMSBP to the origin sequence and the protein oligomerization. Furthermore, we have shown that C. fasciculata tryparedoxin I and II (CfTXN I and II) activate UMSBP DNA binding activity in vitro and suggested that UMSBP activity may be regulated in the trypanosomatid cell through the cellular control of the protein's redox state (12). However, the mechanism that mediates the specific oxidation of UMSBP in such a regulatory model has remained unknown. In a recent report, Motyka et al. (36) have suggested that oligomerization of UMSBP and loss of kDNA in Trypanosoma brucei cells overexpressing TXNPx, have resulted from the depletion of reduced mitochondrial TXN, thus affecting the redox state of UMSBP indirectly, through the oxidation of TXN. Such a model would presume that inactivation of UMSBP in the cell results from its direct oxidation by hydroperoxides and the inhibition of its reactivation through the oxidation of TXN. Here we hypothesized that oxidation of TXN-reduced UMSBP is conducted through an enzymatic reaction in which UMSBP is utilized directly as a substrate for oxidation by TXNPx, providing both the efficiency and the specificity, required from a regulatory mechanism that controls the binding of UMSBP to the replica-tion origin. We found that the capacity of UMSBP to bind the origin sequence is controlled in vivo through a cell cycle-dependent regulation of the protein redox state. We further report here, for the first time, on the direct oxidation of UMSBP by TXNPx and the opposing regulatory effects of TXN and TXNPx on the interaction of UMSBP with the conserved origin sequence. Based on these observations, we suggest an enzymatic mechanism, which regulates the action of UMSBP at the kDNA replication origin.

Preparation of UMSBP
Recombinant UMSBP was prepared as we have previously described (12), except that elution from the phenyl-Sepahrose column was conducted with the following ammonium sulfate concentrations: 2 bed volumes of 1.0 M, followed by 6 bed volumes each of 0.8 and 0.7 M and 2 bed volumes each of 0.6, 0.5, 0.4, 0.3, 0.2, and 0 M ammonium sulfate concentrations.

Cell Synchronization
C. fasciculata cells were cultured in brain heart infusion (Difco) containing 10 g/ml hemin (Sigma), 100 g/ml ampicillin, and 100 units/ml streptomycin at 28°C with shaking (150 -200 rpm). Synchronization by hydroxyurea arrest was performed as previously described (37). Briefly, a late log cell culture (ϳ10 8 cells/ml) was diluted to 10 7 cells/ml in medium containing 0.2 mg/ml hydroxyurea (Amersham Biosciences), followed by 6 -10 h growth with shaking at 28°C. Synchronized cells were harvested by centrifugation (3,220 ϫ g, for 5 min, at 4°C), resuspended in fresh medium without hydroxyurea, and allowed to grow for an additional 6.5 h. A sample of cells, withdrawn prior to hydroxyurea release, serves as time point zero. Such samples were withdrawn at 30-min intervals, centrifuged, and washed twice with ice-cold PBS. Cells were centrifuged, frozen in liquid nitrogen, and stored at Ϫ80°C.

Preparation of Cell Lysates
Cleared cell lysates (Fraction I) were prepared by the gentle disruption of the C. fasciculata cell membrane, using a nonionic detergent in hypotonic solution, as described previously (38), except that 0.2% (w/v) Brij-58 (Sigma) was used (7) and dithiothreitol was omitted from the lysis buffer.

FACS Analysis
Samples, withdrawn at 30-min intervals (ϳ4 ϫ 10 6 cells) from hydroxyurea-synchronized C. fasciculata cultures, were resuspended in PBS, fixed in ice-cold methanol, and stored at Ϫ20°C. Before their analysis, the cells were centrifuged (3,220 ϫ g, for 10 min, at 4°C), resuspended in PBS, containing 50 g/ml RNase A, and incubated for 30 min on ice. In order to avoid cell aggregates, the samples were passed through 26-gauge syringe needle, followed by their filtration through a silk mesh. Propidium iodide was added to the sample (to final concentration of 25 g/ml) before its injection to the FACS analyzer. Analysis was carried out using a FACScan (BD Biosciences) FACS analyzer. The DNA content of 10,000 events was measured using an FL3 emission filter (FL3-H). Values of the x and y axis are in arbitrary units. Analysis of the results was performed using FCS express version 3 software (De-Novo Software).

Monitoring UMSBP Oxidation by Maleimide-Polyethylene Glycol (PEG) Assay
Lysates prepared from samples of synchronized cell cultures (500 g of protein) were treated with 10 mM N-ethylmaleimide (Sigma) for 1 h at 0°C. N-Ethylmaleimide was then diluted to 0.13 mM with 20 mM sodium phosphate buffer, pH 7.2, 5 mM EDTA solution and concentrated using a Microcon concentrator (Millipore). Dithiothreitol was added to a final concentration of 20 mM, and the lysates were incubated for 1 h at 0°C, followed by extensive dialysis against 20 mM sodium phosphate, pH 7.2, 5 mM EDTA solution. The lysates were concentrated using a Microcon concentrator, their protein content was determined by Bradford assay, and 50-g protein samples were reacted for 1 h at 0°C with a 0.3 mM final concentration of maleimide-polyethylene glycol (PEG; molecular mass of 2,385 Ϯ 500 Da; NEKTAR). The reaction products were analyzed by 5-15% Tris/glycine SDS-PAGE under reducing conditions. Gels were blotted onto nitrocellulose membranes, and protein bands were detected by ECL using anti-UMSBP antibodies.

Western Blot Analysis
Protein samples were analyzed by SDS-PAGE electrophoresis, as described above. Protein bands were transferred onto a Protran BA85 cellulose nitrate membrane (Schleicher & Schuell). Membranes were blocked by incubation for 30 min in 5% skim dry milk (Difco) in PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 , 1.7 mM KH 2 PO 4 , pH 7.4), containing 0.05% (v/v) Tween 20 (PBST) and probed for 90 min with a 1:3,000 dilution of mouse anti-UMSBP antibodies or 1:2000 rabbit anti-LimTXNPx antibody, as indicated. Membranes were probed for 40 min with a 1:20,000 dilution horseradish peroxidase-conjugated goat anti-mouse or goat anti-rabbit secondary antibodies, followed by ECL detection, as recommended by the manufacturer (Biological Industries).

Purification of Recombinant CfTXNPx
Pet15b plasmid (Novagene), expressing CfTXNPx was transformed into E. coli BL-21 (DE3). Cells were grown at 37°C with vigorous shaking in LB growth medium containing 100 g/ml ampicillin. Protein expression was induced by 1 mM isopropyl-␤-D-thiogalactopyranoside (at A 600 of 0.6). Affinity chromatography was conducted using nickel-nitrilotriacetate beads (Qiagen), as we have previously described (12), except that the elution buffer contained 350 mM imidazole, and no reducing agent was used throughout the entire procedure. Buffer exchange to 50 mM Tris-Cl, pH 7.5, was performed using ZEBA desalting spin columns (Pierce). When further purification was required, a sample of the affinity chromatography-eluted TXNPx (2 mg) was loaded on a 1-ml fast protein liquid chromatography HiTrap Q FF column (Amersham Biosciences), equilibrated with 50 mM Tris-Cl, pH 7.5. The column was washed with 2 bed volumes of the equilibration buffer and then consecutively with 1.5 bed volumes of equilibration buffer containing 100 mM and 200 mM KCl. TXNPx was eluted by a 10-bed volume 200 -700 mM KCl linear gradient in the same buffer. Chromatography was conducted in an AKTA purifier instrument (Amersham Biosciences), monitoring protein elution using a UV detector ( 214/220/280 nm ). The eluted TXNPx was then rechromatographed on a nickel affinity column, as described above.

CfTXNPx Assays
CfTXNPx Activity-CfTXNPx activity was assayed, by measuring the reduction of hydrogen peroxide or cumene hydroperoxide (Sigma), following the method of Tetaud and Fairlamb (40). Briefly, the reaction, performed in quartz cuvettes, contained 180 M NADPH, 1 mM EDTA, 50 mM Hepes-NaOH, pH 7.4, 130 M trypanothione, 17 nM TcTRH6, 20 M CfTXN2H6, 1-5 M CfTXNPx, and 70 M cumene hydroperoxide in 150 l at room temperature. The reaction was performed by steps as follows. First, trypanothione was added to allow its reduction, and then CfTXN2H6 was added, followed by CfTXNPx addition. The reaction was initiated by the addition of cumene hydroperoxide, and its reduction was monitored by spectrophotometer ( 340 nm ), recording the consumption of NADPH.
CfTXNPx Effect on UMSBP Binding-The effect of CfTXNPx on UMSBP binding to UMS was assayed by the incubation of 2.7 nM UMSBP in the presence of the indicated concentrations of CfTXNPx in an 8-l reaction mixture, containing 50 mM Tris-Cl, pH 7.5, 0.3 mg/ml BSA, and 70 M cumene hydroperoxide. The reactions were incubated at 30°C for 60 min, followed by the addition of 2 l of binding assay solution, containing 6.4 g of BSA, 10 mM MgCl 2 , and 12.5 fmol of 32 P-labeled UMS. The reactions were incubated at 30°C for 5 min, and their products were analyzed by EMSA and exposed to a Phos-phorImager, as described above.
TXNPx-UMSBP Interactions-To monitor TXNPx-UMSBP interactions as well as oxidation of UMSBP by TXNPx, 7.2 M UMSBP was incubated in a 10-l reaction mixture, containing 50 mM Tris-Cl, pH 7.5, 70 M cumene hydroperoxide, and 7.8 M CfTXNPx for 60 min at 30°C. The protein interaction was analyzed by 16.5% Tris-Tricine SDS-PAGE, under nonreducing conditions, blotted onto a nitrocellulose membrane, and subjected to Western analysis, using anti-UMSBP and anti-TXNPx antibodies. The ECL reaction was developed by LAS-3000 (Fuji) or by exposure to x-ray film and analysis by TINA software.
Oxidation of UMSBP by CfTXNPx (10 M) in this reaction was assayed by monitoring the complex of UMSBP with fluorescein-5-maleimide (Pierce), as follows. The reaction products were cooled on ice. Flourescein 5-maleimide was added to a final concentration of 1 mM, followed by incubation for 30 min on ice. The reaction products were analyzed by nonreducing 16.5% Tris-Tricine SDS-PAGE, as described above.

Coupling the TR-TXN-TXNPx-reconstituted Reaction to the UMSBP DNA-binding Reaction
CfTXN I was incubated at the indicated concentrations in a 5-l reaction mixture, containing 50 mM Tris-Cl, pH 7.5, 1 mg/ml BSA, 150 M NADPH, 100 ng of TcTRH6, 20 M trypanothione, and 7.2 nM UMSBP. Reaction mixtures were incubated for 30 min at 30°C, followed by the addition of CfTXNPx, at the indicated concentrations, in a 10-l assay mixture containing 70 M cumene hydroperoxide, 50 mM Tris-Cl, pH 7.5, and 1 mg/ml BSA. The reactions were further incubated for 60 min at 30°C. UMS (12.5 fmol) was added in a 5-l binding assay buffer, containing 5 g of BSA, 50 mM Tris-Cl, pH 7.5, and 8 mM MgCl 2 . The reaction mixtures were incubated for an additional 5 min at 30°C, and their products were analyzed by EMSA and quantified by phosphorimaging.

Immunofluorescence
C. fasciculata (5 ϫ 10 5 cells) were washed once with PBS, placed on poly-L-lysine-coated slides, and fixed with 4% paraformaldehyde for 5 min. This was followed by two 5-min incubations with 0.1 M glycine-PBS, a 5-min incubation with 0.1% Triton X-100 solution in PBS, a 5-min wash with PBS, and overnight fixation with 100% methanol at Ϫ20°C. The slides were washed three times for 5 min with PBS, followed by 30 min of blocking with 1% BSA in PBS. Rabbit anti-LimTXNPx antibody at a 1:300 dilution was added, and the slides were incubated for 90 min in a humidity chamber. This was followed by three 5-min washes with 0.05% Tween 20 in PBS and a 45-min incubation with goat anti-rabbit antibody conjugated to Alexa-594, at a 1:20,000 dilution. The slides were washed three times for 5 min with 0.05% Tween 20, stained for 1 min with 1 g/ml DAPI in PBS, washed 5 min with PBS, and dried for 1 h. Dabco antifade was added, and the slides were sealed. The cells were photographed using a 100-ms exposure time for the DAPI staining and 500 -1,000 ms for the Alexa fluorophore.

A Cell Cycle-dependent Control of the UMSBP Redox State
Regulates Its Binding to UMS-Previous studies have shown that the steady state levels of UMSBP mRNA and protein are apparently constant throughout the trypanosomatid cell cycle (41), suggesting that regulation of UMSBP activity in vivo is most likely of a post-translational nature. The observation that binding of UMSBP to an oligonucleotide representing the UMS, conserved at the minicircle replication origin, is redoxdependent (12) was in accord with such a mechanism. However, the role of redox in the regulation of UMSBP activity in vivo has not been studied. To address this question, UMSBP activity was correlated with the progress of the cell cycle, in hydroxyurea-synchronized cell cultures. UMSBP activity was measured in cell samples, withdrawn at time intervals from the synchronized culture (Fig. 1A), and the progress in the cell cycle was followed by monitoring the abundance of cells undergoing division, using phase microscopy (Fig. 1B), and by FACS analysis of propidium iodide-stained cells in these samples (Fig. 1C). These analyses revealed that when no further reduction of UMSBP is performed, its UMS-binding activity fluctuates in a cell cycle-dependent manner. Cycling of UMSBP activity was significantly suppressed when measured under reducing assay conditions (Fig. 1A) (41), indicating that it reflects the changes in the level of UMSBP reduction, during the various stages of the cell cycle. Hence, these results suggested that UMSBP activity is regulated in vivo through a cell cycle-mediated control of the protein redox state. As shown in Fig. 1A, UMSBP activity peaks during two distinct stages of the cell cycle: during S phase, NOVEMBER 14, 2008 • VOLUME 283 • NUMBER 46 when kDNA replication occurs, and during M-G 1 phase, when neither nuclear nor kinetoplast DNA replication takes place. A recent study has described the physiological significance of a postreplication peak of UMSBP activity, revealing an additional function of UMSBP during kDNA segregation and nuclear mitosis (11).

Redox Regulates UMSBP Binding at the Replication Origin
The observations described above imply that the redox state of UMSBP cycles throughout the cell cycle displaying maxima in the protein reduction level, which correlate with the observed peaks of its UMS-binding activity. To monitor directly the fluctuation in the UMSBP redox state throughout the trypanosomatid cell cycle, we have measured the relative abundance of oxidized thiol groups during the different stages of the cell cycle. For this purpose, samples of cells withdrawn from hydroxyurea-synchronized cultures were lysed and then treated with N-ethylmaleimide, which alkylates only those thiol groups in the protein that are reduced. The remaining thiol groups, which were oxidized and thereby nonreactive in the N-ethylmaleimide reaction, were then reduced by a high concentration of dithiothreitol. Following the removal of the reducing agent, the lysate was treated with a maleimide-polyethylene glycol (PEG) conjugate, which interacts only with the reduced thiol groups in the protein. The binding of each SH group in UMSBP by the maleimide-PEG conjugate increases its mass by ϳ5 kDa (42). This shift in the mass of the complexed protein is detectable by SDS-PAGE analysis, under denaturing and reducing conditions, followed by Western blot analysis, using anti-UMSBP antibodies (Fig. 2, A and B). Since the thiol groups, which were reduced and thereby were available for interaction with the maleimide-PEG conjugate, are those that were oxidized in the original lysate (and thereby were unavailable for the interaction with N-ethylmaleimide), then UMSBP molecules in the shifted protein band represent those UMSBP molecules that contained oxidized thiols at the specific stage in the cell cycle examined. UMSBP oxidation was correlated in these analyses with the progress of the cell cycle, as described above, by monitoring the changes in abundance of cells undergoing division, using phase microscopy (Fig. 2C), and by FACS analysis of propidium iodide-stained cell samples, withdrawn at time intervals from the synchronized culture (not shown).  2N and 4N, cells at G 1 or G 2 phase of the cell cycle, respectively. These analyses revealed that the abundance of oxidized/reduced UMSBP molecules fluctuates in a cell cycle-dependent manner. Furthermore, these analyses indicated that cycling of the UMSBP redox state during the trypanosomatid cell cycle correlates with the cycling of its UMS-binding activity (Fig. 1). Overall, the observations presented here suggest that UMSBP activity is regulated in vivo through a cell cycle-dependent control of the protein's redox state.
UMSBP-TXNPx Interactions Enhance UMSBP Oligomerization-UMSBP redox state fluctuates in vivo in a cell cycle-dependent manner, displaying relatively higher levels of reduction during S and M-G 1 phases and higher levels of oxidation during other stages of the cell cycle (Fig. 1), implying that both reduction and oxidation of UMSBP are tightly regulated in the trypanosomatid cell. We have previously shown that TXN acts upon an oxidized, inactive UMSBP substrate and activates its binding to UMS (12). To explore the possibility that the redox state of UMSBP may be controlled by enzyme-based machinery, we searched for a trypanosomal enzyme that would be capable of oxidizing UMSBP and consequently inhibit its binding to UMS. Several previous observations suggested that TXNPx, an enzyme that reduces hydroperoxides while oxidizing tryparedoxin in the trypanosomatid cell, is a likely candidate for playing a role in such a regulatory mechanism. First, TXNPx, a member of the 2-Cys peroxiredoxin group, is reduced by a conserved Cys-X-X-Cys motif in TXN (33). Sequence analysis (5) reveals that UMSBP contains five such Cys-X-X-Cys motifs, one for each of its five zinc fingers (Fig.  4A). A recent study, using UMSBP affinity chromatography of Leishmania major cell lysate, followed by mass spectrometry of the UMSBP-bound proteins, as well as yeast two-hybrid analysis, revealed the specific interactions between UMSBP and TXNPx. 4 Finally, although not interpreted as a direct effect on the redox state of UMSBP but rather as the result of the depletion of reduced TXN, it has been recently found that overexpression of TbTXNPx resulted in an increase in UMSBP oligomerization and loss of kDNA (36).
To examine the interaction between UMSBP and TXNPx under the oxidation reaction conditions, reaction products were analyzed by SDS-PAGE, under nonreducing conditions, and the resulting protein bands were detected by Western blot analysis, using antibodies raised against UMSBP and against LimTXNPx. As we have shown previously (12), in addition to UMSBP monomers, the anti-UMSBP antibodies detect UMSBP dimers and, in long exposures of the membrane, UMSBP trimers and traces of tetramers (Fig. 3, lane  a). Incubation of UMSBP with TXNPx under the TXNPx oxidation reaction conditions results in a significant overall increase (of ϳ1.5-fold) in the generation of UMSBP trimers and tetramers, with a concomitant decrease (of ϳ9%) in the relative abundance of UMSBP monomers and dimers (Fig. 3, lane b).
We have previously shown that oxidation of UMSBP monomers results in protein oligomerization (12). Hence, the observations presented here, demonstrating the effect of TXNPx on the oligomeric structure of UMSBP, imply an interaction between the two proteins, which may have affected the redox state of UMSBP. Following the TXNPx reaction, an additional high molecular mass protein band is identified by the anti-UMSBP antibodies, that constitutes 5% of the total UMSBP loaded on the gel (in comparison with near background levels of 0.2% measured in the absence of TXNPx) (Fig. 3, compare lanes  a and b). A high mass protein band is also identified by the anti-TXNPx antibodies, either in the absence (Fig. 3, lane f) or the presence of UMSBP (Fig. 3, lane e). The reason for the generation of the UMSBP-containing high mass protein band, as the result of the TXNPx reaction, is not clear. The possibility that this band may represent a UMSBP⅐TXNPx complex has yet to be studied.
TXNPx Inhibits the Binding of UMSBP to UMS DNA through Protein Oxidation-Considering the accumulating evidence for the interaction between UMSBP and TXNPx, the observed enhancement of UMSBP oligomerization by the TXNPx reaction, and the possible involvement of UMSBP SH groups in this interaction, the presumption was strong that UMSBP, which contains the required Cys-X-X-Cys motifs (Fig. 4A), could serve as a substrate for direct oxidation by TXNPx. We have tested this possibility using EMSA analysis, assaying for the effect of purified recombinant CfTXNPx on the binding of prereduced UMSBP to UMS. Indeed, as it is demonstrated in Fig. 4, action of CfTXNPx on UMSBP inhibited significantly its capacity to bind the UMS DNA (Fig. 4B, lanes a-g), indicating that UMSBP can serve as a substrate for the action of CfTXNPx. Quantitative analysis (Fig. 4C)   tions of TXNPx increase the inhibitory effect, with significantly (16-fold) lower slope (displaying approximately only an additional 22% inhibition of the DNA binding reaction, in the presence of 5 M TXNPx). The reason for this biphasic pattern of inhibition is yet unknown. One possible explanation may be that the first phase of extensive inhibition represents the interaction of TXNPX with more exposed and thereby more accessible SH groups in the protein. This may be followed in the second phase by a significant decrease in the concentration of the readily accessible SH groups and the availability of much less accessible reduced SH groups, located inside the folded protein structure, as substrates for oxidation by TXNPx.
Next we asked whether the effect of CfTXNPx on the DNA binding activity of UMSBP results from the oxidation of UMSBP by the enzyme. To address this question, we analyzed the complexes formed between maleimide-fluorescein conjugate (fluorescein 5-maleimide; Pierce), which reacts only with free SH groups in the protein, and UMSBP, which was either untreated or pretreated with CfTXNPx. Since UMSBP is a relatively small protein (13.7 kDa), which contains 15 cysteine residues, its complex with the maleimide conjugate is expected to yield a detectable change (of ϳ0.4 kDa for each bound maleimide conjugate molecule) in the protein mass. Thus, the interaction of UMSBP with the maleimide conjugate could differentiate between the reduced and oxidized forms of UMSBP. In the experiment described in Fig. 5, UMSBP was treated with CfTXNPx and then incubated with fluorescein 5-maleimide and subsequently submitted to SDS-PAGE analysis under non- FIGURE 4. CfTXNPx inhibits the binding of UMSBP to UMS. A, amino acids sequence of UMSBP, indicating the five Cys-X-X-Cys motifs (boldface type), within the five CCHC-type zinc finger domains (underlined). B, 2.7 nM UMSBP was incubated in a CfTXNPx reaction mixture, as described under "Experimental Procedures," with increasing concentration of CfTXNPx, as indicated below. Subsequently, the reaction products were assayed in a DNA binding reaction with a 32 P-labeled UMS ligand and subjected to EMSA, and the gel was exposed to phosphorimaging. Lanes a-g, 0, 0.15, 0.3, 0.6, 1.2, 2.4, and 5 M CfTXNPx. In lane h, no protein was added. C, phosphorimaging quantification of the EMSA data, presented in B. The percentage of complex formation measured is presented relative to the value measured in the absence of TXNPx, used as 100%. The CfTXNPx reaction, with prereduced UMSBP as a substrate, was conducted as described under "Experimental Procedures." Reaction products were incubated in the presence of flourescein 5-maleimide, and the protein-maleimide complexes formed were analyzed by SDS-PAGE under nonreducing conditions, transferred onto nitrocellulose membrane, and subjected to Western blot analysis, using anti-UMSBP antibodies, as described under "Experimental Procedures." Lane a, UMSBP prior to the interaction with CfTXNPx; lane b, products of the CfTXNPx-UMSBP reaction. R-UMSBP, reduced UMSBP; O-UMSBP, oxidized UMSBP. reducing conditions. Western analysis, using anti-UMSBP antibodies, indicates clearly that CfTXNPx-treated UMSBP acquired a higher electrophoretic mobility upon its interaction with the maleimide conjugate, compared with UMSBP that was not treated with CfTXNPx (Fig. 5, compare lanes a and b). Quantitative analysis of the data revealed the shift in mobility of 60% of the original UMSBP substrate used. Thus, the number of free SH groups available for the interaction with the maleimide conjugate decreased significantly as a result of the CfTXNPx reaction. In accord with this result is also an observation that treatment of UMSBP with CfTXNPx resulted in miscleavage by trypsin at several sites in the UMSBP sequence, as indicated by mass spectrometry analysis, 5 which could most probably be attributed to changes in the protein structure as a result of its oxidation by TXNPx.
Countereffects of TXN and TXNPx Actions on UMSBP-UMS Nucleoprotein Interactions-To examine the potential physiological significance of TXNPx effect on the binding of UMSBP to the origin sequence, we tested the opposing effects of TXN and TXNPx in turning "on" or "off" the binding of UMSBP to DNA. For this purpose, we have reconstituted the tryparedoxin redox cascade pathway of trypanosomatids, from the purified recombinant enzymes TR, TXN, and TXNPx, and coupled this reaction in vitro to the UMSBP-UMS binding reaction (Fig. 6). In this reaction, TR is being reduced, through the conversion of NADPH to NADP ϩ , and, in turn, reduces trypanothione, which subsequently reduces TXN. As shown in Fig. 6, when preoxidized UMSBP was added to the above reconstituted enzymatic system, which was depleted of TXN, relatively low levels of DNA binding activity, most probably the result of incomplete preoxidation of UMSBP, could be observed in the EMSA analysis (Fig.  6, A and B, lane a). The addition of CfTXN I significantly activated the binding of UMSBP to DNA (Fig. 6, A  and B, lanes b and c), as we have previously reported (12). The addition of CfTXNPx and cumene hydroperoxide to the reaction mixture, which contained CfTXN-activated UMSBP, turned off the TXN effect, resulting in the inhibition of the UMSBP-UMS interactions (Fig. 6, A  and B, lanes e and f). This inhibitory effect was not observed when CfTX-NPx was omitted from a reaction mixture, which contained all of the reaction components, including cumene hydroperoxide (Fig. 6, A  and B, lane d), indicating that the inhibitory effect on the UMSBP-DNA interaction has resulted from the action of the enzyme CfTXNPx. These observations demonstrated the capacity of TXN and TXNPx, through their opposing effects on UMSBP, to reciprocally control the interaction of the protein with the origin sequence in vitro. The fact that both up-and down-regulation of UMSBP activity is displayed in the same reconstituted reaction suggests a potential physiological role in vivo of this enzymatic pathway (Fig. 6C) in regulating the binding of UMSBP onto the minicircle replication origin.
Intramitochondrial Localization of C. fasciculata TXNPx Overlaps That of the kDNA Disk-Potential interaction between UMSBP and TXNPx would have been largely facilitated had the two proteins been localized in close proximity within the trypanosomatid cell. To address this question, we have localized TXNPx in C. fasciculata by immunofluorescence, using antibodies directed against LimTXNPx. These antibodies were found to cross-react with CfTXNPx, by both Western blot analysis and immunoprecipitation of CfTXNPx from C. fasciculata cell lysates (data not shown). Overlay images of DAPI-stained cells (Fig. 7B) with those of cells that were immunostained with anti-LimTXNPx antibodies (Fig. 7C) reveals that C. fasciculata mitochondrial TXNPx localization overlaps with the region occupied by the kDNA disc (Fig. 7D).  Previous studies have localized UMSBP to the kineto-flagellar zone near the edge of the kDNA disc facing the flagellum (41). The localization of both UMSBP and TXNPx, in close proximity to each other and to the kDNA network, is in accord with the notion that the two proteins may interact in vivo in the trypanosomatid cell.

DISCUSSION
Considering the essential role played by the zinc fingers of UMSBP in the protein interaction with DNA (12), the high sensitivity of zinc fingers to redox changes, and the reversible nature of the changes in the protein's redox state, we have previously examined the possibility that redox signaling may be involved in the regulation of UMSBP. Indeed, these studies have revealed that both UMSBP binding to the origin sequence and its oligomerization are affected in vitro by the protein's redox state (12). However, the question of whether UMSBP activity is regulated in vivo through the control of the protein redox state is being first challenged in the current study. Our observations reveal that UMSBP binding to the origin sequence cycles throughout the cell cycle, in correlation with the cycling of its redox state, suggesting that it is regulated through cell cycle-dependent control of the protein redox state (Figs. 1 and  2). These cell cycle-directed changes in the redox state of UMSBP, which are evident when assaying UMSBP activity under nonreducing conditions, explain our previous results that showed an approximately constant level of UMS binding activity throughout the entire cell cycle, when assayed under reducing assay conditions (Fig. 1) (41).
The correlation observed in the current study between the cycling of UMSBP redox state and its DNA binding activity, during the progress in the cell cycle, links the cell cycle-dependent regulation of UMSBP redox state to its functions during the different stages of the trypanosomatid cell cycle. In a recent study, using RNA interference analysis (11), we have found that the minicircle origin-binding protein, UMSBP, functions during the S phase of the cell cycle, playing a role in the initiation of minicircles replication. Unexpectedly, this study has also revealed an additional, postreplication function for this protein, during kDNA segregation and nuclear mitosis (11). These later observations suggested the possible involvement of UMSBP in the control of the trypanosomal cell cycle. The observations described in the current study imply a potential linkage between cell cycle control and the redox-mediated regulation of UMSBP. We hypothesized that UMSBP activity is regulated in the trypanosomatid cell through the action of enzymatic machinery, which directly changes the protein's redox state, thereby modulating the DNA binding properties of the protein. This hypothesis was examined in a multienzyme reaction, reconstituted with pure enzymes of the main enzymatic pathway controlling the redox state in trypanosomatids. Indeed, our results demonstrate that the redox state of UMSBP, and consequently its DNA binding activity, are modulated enzymatically. Although the action of TXN enhances the binding of UMSBP to DNA (12) (Fig. 6), that of TXNPx impairs UMSBP-UMS interactions (Figs. 4 and 6). The intramitochondrial co-localization of a TXNPx isoform in C. fasciculata with the kDNA disk, in close proximity to UMSBP (41), similarly to its reported localization in L. infantum and L. major (35,43), is in accord with a potential functional role for TXNPx in kDNA metabolism, through its effect on the recognition of the kDNA minicircle replication origin.
Reaction of TXNPx using UMSBP as a substrate was conducted here using equimolar concentrations of the two proteins or in excess of TXNPx. However, in studying the catalytic nature of the UMSBP oxidation reaction, considering this stoichiometry, one has to consider also the following factors. The assay used here is a functional assay, measuring the effect of the oxidation of UMSBP by TXNPx on the binding of UMSBP to DNA. As we have previously shown, UMSBP consists of five CCHC-type zinc finger domains, of which the four C-terminal ones are essential for its binding to the origin sequence and the stability of the nucleoprotein complex (12). 3 Thus, although it is still unknown how many of these 12 SH groups serve as substrates for oxidation by TXNPx, one has to consider the possible involvement of several cysteine residue in this reaction while calculating UMSBP-TXNPx stoichiometry. Another factor that has considerably affected the stoichiometry of the TXNPx-UMSBP reaction was the fact that the recombinant TXNPx preparation used in these experiments revealed, when assayed using a hydroperoxide substrate, about an order of magnitude lower activity than expected. This significant decrease in the enzyme activity has most probably resulted from an extensive aggregation observed of the purified protein.
Aggregation of peroxiredoxins has been discussed on several occasions (e.g. see Ref. 44 and 45). An alternative interpretation of these observations, that the inhibition of the binding of UMSBP onto the origin sequence results from the depletion of free UMSBP molecules, through the generation of complexes between UMSBP and TXNPx rather than by UMSBP oxidation, was also examined in this study. Indeed, a relatively small fraction, of 5% of the UMSBP in the TXNPx reaction, could be detected in a high mass protein band, containing both proteins and may represent TXNPx⅐UMSBP complexes (Fig. 3, lane b). This small fraction of UMSBP bound in the complex with TXNPx could not explain the extent of oxidation observed as a result of the TXNPx reaction (Fig. 5) and the significant FIGURE 7. A TXNPx isoform is localized to the region occupied by the kDNA disc in C. fasciculata. Localization of TXNPx in C. fasciculata was conducted by immunofluorescence, using antibodies raised against LimTXNPx, as described under "Experimental Procedures." A, phase image; B, DAPI staining; C, immunostaining with anti-LimTXNPx; D, overlay of B and C; K, kDNA; N, nucleus. Bar, 5 m. enhancement of UMSBP oligomerization by TXNPx (Fig. 3) but may rather represent the covalent intermediates in the oxidation of UMSBP by TXNPx. The possibility that oxidation of UMSBP by TXNPx occurs in vivo is supported by a previous report, demonstrating the oligomerization of UMSBP and loss of kDNA in procyclic T. brucei that overexpress the mitochondrial TbTXNPx (36).
Redox control of proteins activity has been described in various cellular processes, such as oxidative stress defense (18,46,47), cell cycle regulation (48), and DNA replication (12,49,50). Regulation of macrophage migration inhibitory factor, by direct interaction with the peroxiredoxin protein PAG, has recently been reported (51). Furthermore, involvement of peroxiredoxins in the sensing of hydroperoxide signal and its transmission to a regulator protein has been reported previously in the activation of the yeast transcription factors Yap1 (24,25) and Pap1 (27,28), both involved in the induction of antioxidant genes in response to oxidative stress. Similarly to the reported cases, our observations indicate that enzymatic machinery, based upon the opposing actions of TXN and TXNPx, is capable of regulating the activity of a regulatory protein. However, unlike in those cases, UMSBP, the regulated protein under study, was not reported to be involved in the cell defense against oxidative stress but rather to function in DNA replication and segregation (11).
The basic features of the enzymatic pathway that regulates redox in trypanosomatids are shared by similar pathways observed in most other living organisms studied (reviewed in Ref. 33). TXNPx is a member of the 2-Cys peroxiredoxin group, whose members, which are reduced by thioredoxins (or TXNs in trypanosomatids), are conserved throughout evolution. Hence, enzymatic mechanisms, which are capable of regulating the redox state of proteins, via the Cys-X-X-Cys motif, may be abundant in various other organisms. On the basis of the observations presented here, we suggest that the mechanism demonstrated in this study may represent a general mode of regulation of protein activities, through enzyme-mediated post-translational modification of their redox state. In such a model, the readily reversible oxidation-reduction of proteins, catalyzed by specific enzymes, would modify their structure and thus modulate their physiological function, similarly to the well documented regulation of proteins activity through their enzymemediated reversible phosphorylation.