NMR Structure of Carcinoscorpius rotundicauda Thioredoxin-related Protein 16 and Its Role in Regulating Transcription Factor NF-κB Activity*

Background: Cr-TRP16 is a thioredoxin like protein that regulates NF-κB activity. Results: NMR structure of Cr-TRP16, along with its regulation of NF-κB activity, is reported. Conclusion: Our findings elucidate molecular mechanism by which NF-κB activation is regulated through Cr-TRP16. Significance: This study helps further in understanding the interactions of Trxs and Trx-like proteins with other metabolic pathways and will lead to promising therapeutic development. Thioredoxins (Trxs), which play a key role in maintaining a redox environment in the cell, are found in almost all organisms. Trxs act as potential reducing agents of disulfide bonds and contain two vicinal cysteines in a CXXC motif at the active site. Trx is also known to activate the DNA binding activity of NF-κB, an important transcription factor. Previously, Trx-related protein 16 from Carcinoscorpius rotundicauda (Cr-TRP16), a 16-kDa Trx-like protein that contains a WCPPC motif, was reported. Here we present the NMR structure of the reduced form of Cr-TRP16, along with its regulation of NF-κB activity. Unlike other 16-kDa Trx-like proteins, Cr-TRP16 contains an additional Cys residue (Cys-15, at the N terminus), through which it forms a homodimer. Moreover, we have explored the molecular basis of Cr-TRP16-mediated activation of NF-κB and showed that Cr-TRP16 exists as a dimer under physiological conditions, and only the dimeric form binds to NF-κB and enhances its DNA binding activity by directly reducing the cysteines in the DNA-binding motif of NF-κB. The C15S mutant of Cr-TRP16 was unable to dimerize and hence does not bind to NF-κB. Based on our finding and combined with the literature, we propose a model of how Cr-TRP16 is likely to bind to NF-κB. These findings elucidate the molecular mechanism by which NF-κB activation is regulated through Cr-TRP16.

Thioredoxins (Trxs), which play a key role in maintaining a redox environment in the cell, are found in almost all organisms. Trxs act as potential reducing agents of disulfide bonds and contain two vicinal cysteines in a CXXC motif at the active site. Trx is also known to activate the DNA binding activity of NF-B, an important transcription factor. Previously, Trx-related protein 16 from Carcinoscorpius rotundicauda (Cr-TRP16), a 16-kDa Trx-like protein that contains a WCPPC motif, was reported. Here we present the NMR structure of the reduced form of Cr-TRP16, along with its regulation of NF-B activity. Unlike other 16-kDa Trx-like proteins, Cr-TRP16 contains an additional Cys residue (Cys-15, at the N terminus), through which it forms a homodimer. Moreover, we have explored the molecular basis of Cr-TRP16-mediated activation of NF-B and showed that Cr-TRP16 exists as a dimer under physiological conditions, and only the dimeric form binds to NF-B and enhances its DNA binding activity by directly reducing the cysteines in the DNAbinding motif of NF-B. The C15S mutant of Cr-TRP16 was unable to dimerize and hence does not bind to NF-B. Based on our finding and combined with the literature, we propose a model of how Cr-TRP16 is likely to bind to NF-B. These findings elucidate the molecular mechanism by which NF-B activation is regulated through Cr-TRP16.
Reactive oxygen species are natural by-products of several oxidative metabolic pathways and are known to mediate intra-cellular signaling pathways (1). However, excessive production of reactive oxygen species is detrimental to the cell, because of increased oxidative stress (2). Thioredoxin (Trx) 4 plays a key role in maintaining an intracellular redox state, which is essential for protecting cells from oxidative damage (3). Trx is a 12-kDa ubiquitous and evolutionarily conserved protein in almost all organisms and functions as a disulfide reductase (4,5). A conserved Cys-Xaa-Xaa-Cys sequence at the active site of Trx is involved in either reducing disulfide bonds or oxidizing sulfhydryls, both intra-and extracellularly (6). The two cysteines are maintained in an active reduced form by Trx reductase, a selenocysteine-containing protein that uses the reducing power of NADPH (6). The disulfide reductase activity and reduction potential of the cysteines depend largely on the identity of the two XX residues (7).
Trx proteins also function as a redox sensor and transducer that impart information on the cellular redox status to proteins that do not possess their own redox-sensitive residues. For instance, the reduced form of Trx1, a mammalian cytosolic isoform, binds to and thereby inhibits the activity of apoptosis signal-regulating kinase 1 (ASK1), an upstream activator of the JNK and p38 MAPK signaling pathways (8). Their alterations have been implicated in cardiovascular diseases (9), diabetes (10), hepatic and renal diseases (11), Alzheimer disease (12), Parkinson disease (13), and rheumatoid arthritis (14). Some human cancers show greatly increased Trx expression (15,16), indicating a potential role for Trx in tumorigenesis. Its interactions with different proteins of the metabolic and signaling pathways make Trx an attractive target for therapeutic interventions.
There are several Trx-like proteins, which possess the thioredoxin fold and function and exist in different sizes . Recently, a 16-kDa protein that contains a WCPPC motif has been isolated and characterized from a species of horseshoe crab, Carcinoscorpius rotundicauda, and named Trx-related protein 16 (Cr-TRP16) (17). Notably, the 16-kDa Trx-like pro-teins are functionally similar to the 12-kDa Trxs, and each monomer contains two Cys residues at the active site. Unlike other 16-kDa Trx like proteins, Cr-TRP16 contains an additional Cys residue (Cys-15, at the N terminus), but it lacks the extra C-terminal Cys residue, which is found in mammalian Trxs (18).
A subtle balance between oxidants and antioxidants is crucial for homeostasis. There is emerging evidence that the highly conserved intracellular redox state regulates the mechanism of signal transduction and gene expression (19). NF-B is one of the redox-regulated proteins. It is a member of the Rel family of transcription factors and exists in the cytoplasm in complex with its inhibitor protein, IB (20). A wide variety of stimuli, including TNF-␣, phorbol ester, bacterial lipopolysaccharide, and virus infection, can activate NF-B (21). NF-B activation involves site-specific phosphorylation of IB-␣, which results in the dissociation of the complex, unmasking of the NF-B nuclear localization signal, and nuclear entry of NF-B to bind to its cognate DNA (22). Thus, for the stimuli that potently and rapidly modulate the nuclear activity of NF-B, the IB-␣ may represent a critical activation target (20). Several studies have suggested that Trx is a specifically potent antioxidant for NF-B activation (23). Similar to the human Trx1, Cr-TRP16 also up-regulates the TNF␣-induced NF-B activation.
Cr-TRP16 exists as a dimer in the oxidized state. However, the underlying mechanism of dimerization and the regulation of NF-B by Cr-TRP16 are so far unknown. For example, what is the physiological role of the Trx-dimer? Is Cys-15 involved in Cr-TRP16 dimer formation? How is NF-B regulated under oxidized and reduced cellular environments? It is possible that the oxidation-induced dimerization facilitates sensing cellular oxidative stress. Dimerization might remove the Trx from the redox cycle, which is catalyzed by Trx reductase because the dimer is not a suitable substrate for Trx reductase. In addition, dimerization of secreted Trx in the relatively oxidizing extracellular environment could be a way of limiting the growthstimulating effects of Trx (24). Further mutational and structural studies will define the function of the additional N-terminal Cys-15 residue in Cr-TRP16. To understand the NF-B regulation by Cr-TRP16 and to address the above questions, we had undertaken the structural and functional characterization of Cr-TRP16, the first Trx fold structure from an ancient arthropod. Here we report the NMR structure of the reduced form of wild type Cr-TRP16. Moreover, the role of the Cys-15 residue in dimerization and NF-B regulation were studied by site-directed mutagenesis, biophysical, and immunoprecipitation techniques.

EXPERIMENTAL PROCEDURES
Cloning-The full-length Cr-TRP16 gene (encoding residues Met-1-Arg-143) was PCR-amplified using a forward (CAG-CATATGATGGAATTTATCCAAGGAAT) and a reverse (CCTCTCGAGTCTTGCCCAGTT) primer with an NdeI restriction site at the 5Ј end and an XhoI site at the 3Ј end (underlined), respectively, and inserted into the pET-22b(ϩ) vector (Invitrogen), which was previously linearized by the corresponding restriction enzymes. For mammalian expression, the Cr-TRP16 gene was PCR-amplified with the following primers: forward, GGAGGATCCATGGAATTTATCCAA-GGA, and reverse, ATTCTCGAGTCATTTGTCGTCATCG-TCCTTATAGTCTCTTGCCCAGTTCTGGAA, which contained BamHI and XhoI restriction sites (underlined), respectively. The reverse primer introduced a FLAG tag at the C terminus of the Cr-TRP16 protein. The insert was ligated into the previously linearized expression vector pcDNA3.1/V5-His/ lacZ (Invitrogen).
Protein Expression and Purification-The pET22b:Cr-TRP16 construct was transformed into the bacterial strain Escherichia coli BL21(DE3) for expression. Optimal expression of the Cr-TRP16 protein was achieved by induction of 1-liter culture with 0.3 mM isopropyl ␤-D-1-thiogalactopyranoside at 20°C. The cells were disrupted by a French press, and the supernatant was collected after centrifuging at 10,000 ϫ g for 1 h at 4°C. His-tagged Cr-TRP16 proteins were purified in two steps using nickel-nitrilotriacetic acid (Qiagen) affinity chromatography followed by a Superdex 75 gel filtration column on Äkta Express (GE Healthcare). The buffer was exchanged to a solution containing 10 mM Tris (pH 7.0), 100 mM NaCl, 5 mM dithiothreitol, 5% glycerol, 1 mM EDTA, and finally the protein was concentrated to 18 mg/ml. For NMR study, 15 N-labeled and 13 C, 15 N-double labeled samples were prepared using M9 minimal medium with 15 NH 4 Cl and [ 13 C]glucose as a sole source of nitrogen and carbon, respectively. Weakly aligned 15 N-labeled sample was prepared by the addition of 8 mg ml Ϫ1 of filamentous phage Pf1 (from ASLA Biotech) for residue dipolar coupling measurement.
NMR Experiments and Structure Determination-All NMR experiments were carried out at 25°C on a Bruker Avance 800 MHz spectrometer equipped with a TXI cryogenic probe. 1 H, 13 C, and 15 N resonance assignments were performed by measuring the three-dimensional HNCACB, three-dimensional CBCA(CO)NH (25), and three-dimensional CCH-TOCSY (26) spectra. Interproton distance restraints for structural calculation were obtained from three-dimensional 13 C-edited NOESY-HSQC, three-dimensional 15 N-edited NOESY-HSQC, and two-dimensional NOESY spectra using a 100-ms mixing time. 1 D NH residue dipolar couplings were measured using the IPAP method (27). The residue dipolar coupling values were obtained by subtracting the reference value in isotropic solution. Two-and three-dimensional NMR spectra were processed using the NMRPipe program (28), and data analysis was performed with the help of the Sparky program (29).
The structure was calculated using the Xplor-NIH 2.24 software package (30). A total of 1731 NOE distance restraints, 76 hydrogen bonds, and 124 dihedral angle restraints were predicted by the TALOS program (31). Residue dipolar couplings were used in the final cooling stage. 100 structures were calculated using 30,000 steps of simulated annealing, and a final ensemble of 20 lowest energy structures was selected for figure preparation.
Site-directed Mutagenesis-On the NMR structure of Cr-TRP16, Cys-15 is located on the surface and has been hypothesized to be involved in dimerization. This residue was mutated to serine in both the pET22b and pcDNA3.1/ V5-His/lacZ constructs by the use of the inverse PCR-based mutagenesis protocol (32). The C15S mutant protein was expressed following the steps that were described for the wild type protein.
Analytical Ultracentrifugation (AUC)-The dimeric state of WT and C15S mutant Cr-TRP16 proteins purified by affinity and gel filtration followed by reverse phase HPLC (data not shown) was investigated by monitoring their sedimentation properties in sedimentation velocity experiments. 500 l of samples at A 280 nm of 0.8 in 10 mM Tris-HCl (pH 7.0), 100 mM NaCl, and 5% glycerol were used. The experiments were carried out in the absence of DTT in the buffer. Sedimentation velocity profiles were collected by monitoring the absorbance at 280 nm. The samples were centrifuged at 40,000 rpm at 20°C in a Beckman Optima XL-I centrifuge fitted with a six-hole AN-60 rotor and double-sector aluminum centerpieces and equipped with absorbance optics. The scans were analyzed using the Sedfit program (33).
Western Blotting-The human cervical epithelial carcinoma (HeLa) cell line was obtained from American Type Culture Collection (Manassas, VA). It has to be noted that no horseshoe crab cell line is available, and HeLa are the most commonly used cells for thioredoxin related (including Cr-TRP16) ex vivo studies (17,23,34,35). The cells were cultured in RPMI 1640 medium containing 1ϫ antibiotic-antimycotic solution with 10% FBS. HeLa cells transfected with the Cr-TRP16 or C15S Cr-TRP16 plasmids were lysed in lysis buffer (20 mM Tris (pH 7.4), 250 mM NaCl, 2 mM EDTA (pH 8.0), 0.1% Triton X-100, 0.01 mg/ml aprotinin, 0.005 mg/ml leupeptin, 1 mM PMSF, and 4 mM NaVO 4 ). The lysates were then spun on a microcentrifuge at 14,000 rpm for 10 min to remove insoluble material and resolved on 10% SDS gel. Nuclear protein extracts were prepared using a nuclear extraction kit (Active Motif, Carlsbad, CA) according to the manufacturer's instructions. Protein concentration was determined with the Bradford reagent, and 30 g of nuclear or cytoplasmic extract was taken for immunoblotting. After electrophoresis, the proteins were electrotransferred to a nitrocellulose membrane, blocked with 5% nonfat milk, and probed overnight with the antibodies of interest at 4°C. The blot was washed, exposed to HRP-conjugated secondary antibodies for 1 h, and finally examined by ECL (Amersham Biosciences).
NF-B DNA Binding Assay-To determine NF-B activation, we performed DNA binding assay using the TransAM NF-B kit (Active Motif) according to the manufacturer's instructions and as previously described (36). Briefly, HeLa cells were transfected with the Cr-TRP16 or C15S Cr-TRP16 plasmids for 30 h and then stimulated with TNF-␣ (1 nM) for 6 h. Nuclear protein extracts were prepared using a nuclear extraction kit (Active Motif) according to the manufacturer's instructions. Protein concentration was determined with the Bradford reagent, and 20 g of nuclear extract was added to the plates precoated with an NF-B consensus oligonucleotide sequence and incubated at room temperature for 1 h with continuous shaking at 100 rpm. The plates were washed and incubated with a p65 primary antibody for 1 h at room temperature. After the final incubation with a secondary antibody, the wells were washed with washing buffer; color was developed by the addition of substrate solution and read after 10 min at 450 nm against a blank, which was read at a reference wavelength of 655 nm.
NF-B-dependent Luciferase Reporter Assay-NF-B dependent reporter gene expression assay was performed as described previously (37). Briefly, ϳ10 5 /well HeLa cells were plated in 96-well plates in RPMI buffer containing 10% FBS. The NF-B-responsive elements, linked to a luciferase reporter gene plasmid, were transfected with the wild type Cr-TRP16 or C15S Cr-TRP16 (mutant) plasmids. Transfection was performed using Lipofectamine (Invitrogen) according to the manufacturer's protocols. After 30 h of transfection, HeLa cells were treated with TNF-␣ (1 nM) for 6 h and then washed and lysed in luciferase lysis buffer (Promega). Luciferase activity was measured by Tecan (Durham, NC) plate reader, which is equipped with a function to measure luminescence by using a luciferase assay kit (Promega) and was normalized to ␤-galactosidase activity. All of the luciferase experiments were performed in triplicate and repeated three or more times.
DNA Binding Assay of NF-B/p65 C38S Mutant-To further explore the mechanism of interaction between Cr-TRP16 and NF-B, we undertook an ex vivo study to demonstrate that Cr-TRP16 exists as a dimer inside cell and directly reduces the Cys-38 of NF-B p65 subunit. We used expression plasmids that encode p65 wild type and C38S mutated p65. The advantage of using transfected cells overexpressing p65 is that there is no TNF-␣ stimulation needed to observe the direct effect of Cr-TRP16 on NF-B DNA binding activity. Briefly, HeLa cells were transfected with the plasmids pcDNA3.1, pcDNA-p65, and pcDNA-p65C38S (a gift from Dr. Gilmore, Boston University), along with Cr-TRP16 or C15S Cr-TRP16 for 36 h.
Nuclear protein extracts were prepared using a nuclear extraction kit (Active Motif) according to the manufacturer's instructions. Protein concentration was determined with the Bradford reagent, and 20 g of nuclear extract was added to the plates that were precoated with NF-B consensus sequence oligonucleotide and incubated at room temperature for 1 h with continuous shaking at 100 rpm. The plates were washed and incubated with p65 primary antibody for 1 h at room temperature. After the final incubation with secondary antibody, the wells were washed with washing buffer, and color was developed by the addition of substrate solution, which was read at 450 nM against the blank, reference wavelength of 655 nM after 10 -15 min. Western blot analysis was performed to determine the expression levels of p65 and p65C38S. The experiments were repeated three times with almost identical outcome.
Sequence and Structural Homology-Proteins with the Trx fold and active site sequence are classified under the Trx superfamily (38). Comparison of Trxs and Trx like protein sequences ( Fig. 2A) show that the Trx core domain is well conserved with the central five ␤-strands, surrounded by four ␣-helices (39). Cr-TRP16 has an overall sequence identity of 18% with the 12-kDa Trxs from mammalian, insect, and bacterial species. Most invariant residues are clustered around the active site. The 12-kDa Trxs are characterized by a conserved WCGPC motif in the active site, whereas Cr-TRP16 has a WCPPC motif.
So far, no Trx structure is reported from the subphylum (Celicerata) to which horseshoe crab belongs (noninsect species of arthropods). A search for topologically similar domains in the Protein Data Bank using the DALI program (40) reveals that the structural features of Cr-TRP16 resemble a typical Trx (41). The closest match is observed between the Trx-related protein tryparedoxin II (Protein Data Bank code 1FG4), a 16-kDa Trx-like protein from Crithidia fasciculate with 26% sequence identity, and an RMSD of 2.9 Å for 126 C␣ atoms. The structure-based sequence alignment reveals that most of the structurally invariant residues are located at the N-terminal loop region between ␤2 and ␤3 (Fig. 2).  Dimerization of Cr-TRP16-Human 12-kDa Trx1 forms a homodimer through an intermolecular disulfide bond via Cys-73 (18). Interestingly, Cr-TRP16 can also form a stable dimer. Despite low sequence identity (18%) between Cr-TRP16 and hTrx1, Cr-TRP16 could also enhance TNF␣-induced NF-B transcriptional activation by increasing the NF-B DNA binding activity. In addition to the two conserved Cys residues (Cys-38 and Cys-41) at the active site, Cr-TRP16 has an additional Cys residue at position 15, which is involved in the formation of an intermolecular disulfide bond. The previous mass spectrometry report (17) has shown that only the two conserved Cys residues at the active site are redox active, similar to the human Trx1. To further explore the role of Cys-15 in dimer formation, we have performed two-dimensional HSQC experiments for oxidized and reduced wild type Cr-TRP16 and analytical ultracentrifuge experiments on oxidized and reduced wild type and C15S mutant Cr-TRP16. 1 H- 15 N-HSQC NMR Spectroscopy-To characterize the mode of dimerization and to identify the critical residues involved in dimer formation, the 1 H-15 N-HSQC NMR spectra of reduced and oxidized wild type Cr-TRP16 were compared. Sequential assignments were completed by using the standard triple-resonance methods. The assigned spectra of reduced and the overlay 1 H-15 N-HSQC NMR spectra of reduced and oxidized Cr-TRP16 are shown in Fig. 3, respectively. Two-dimensional NMR experiments show that the majority of the amides are unaffected by dimerization. Fig. 3B shows that a total of 26 amide signal chemical shifts have changed. Nearly all of these residues are located in the surface-exposed region around the N terminus Cys-15 and the second ␤-sheet (␤1␤2). This suggests that this part of the molecule is involved in dimerization.
Analytical Ultracentrifugation-The C15S (C15S) mutant was generated to elucidate the role of the additional Cys at position 15 (Cys-15) in dimer formation and its physiological significance in the NF-B function. Dynamic light scattering measurements on reduced and oxidized C15S Cr-TRP16 show an apparent molecular mass corresponding to only a monomer. Nonreducing and SDS-PAGE for C15S Cr-TRP16 show a single FIGURE 2. Comparison of Cr-TRP16 with Trypaerdoxin. A, structure-based sequence comparison between Cr-TRP16 and other 16-and 12-kDa thioredoxins. Alignment was done by Clustal W. Identical and conserved residues are highlighted in red and outlined in blue, respectively. This figure was created by using the program ESPript (60). B, comparison of Cr-TRP16 (red) and Trypaerdoxin II (green) (Protein Data Bank code 1FG4) from C. fasciculate. The alignment was performed using the program COOT (61). The RMSD between is 2.9 Å for 126 C␣ atoms.
band at an apparent molecular mass of 17 kDa (Fig. 4C), whereas nonreducing PAGE for the wild type protein shows an apparent dimer band (at 34 kDa). In the AUC analysis of oxidized wild type Cr-TRP16 protein (A 280 nm ϭ 0.8), there is only one peak corresponding to an apparent molecular mass of a dimeric Cr-TRP16 (34 kDa; Fig. 4A, panel i), whereas the reduced Cr-TRP16 protein shows a peak at 17 kDa, corresponding to monomeric Cr-TRP16 (Fig. 4A, panel ii). On the other hand, we observed only one peak with an apparent molecular mass of 17 kDa for both the oxidized (Fig. 4B, panel i) and reduced (Fig. 4B, panel ii) C15S mutant of Cr-TRP16. These results are in agreement with our hypothesis that Cys-15 is responsible for dimerization.

Cr-TRP16 Increases TNF-␣-induced Nuclear
Translocation of p65 and p50-Biophysical and mutational studies have confirmed that Cr-TRP16 exists as a dimer, and Cys-15 is responsible for its dimerization. However, it is unclear whether Cr-TRP16 exists as a dimer under physiological conditions. To address this, we expressed the wild type and C15S mutant Cr-TRP16 proteins and stimulated the cells with TNF-␣, one of the most potent inducers of NF-B activation (43). Western blot analysis with anti-p50 antibody revealed minimal change in the NF-B p50 protein, suggesting that the overexpression of Cr-TRP16 does not change the expression of p50 (Fig. 5A). Similarly it does not show any significant effect on the TNF-␣ induced IB␣ degradation (Fig. 5C). Subsequently, we investi- gated the effect of Cr-TRP16 overexpression on nuclear translocation of NF-B. Western blot analysis with nuclear extract using anti-p50, anti-p65, and anti-phosphorylated p65 antibodies revealed that the overexpression of wild type Cr-TRP16 induced the nuclear translocation of the p50-p65 heterodimer and increased the level of phosphorylated p65, whereas no appreciable increase was observed in C15S Cr-TRP16-transfected cells (Fig. 5, B and C).
Cr-TRP16 Augments TNF␣-induced NF-B DNA Binding Activity-Next we investigated the effect of wild type and C15S mutant Cr-TRP16 on NF-B DNA binding activity. Nuclear extracts were prepared and tested for NF-B activity using an ELISA-based DNA binding assay. Fig. 6A demonstrates that wild type Cr-TRP16 increases TNF-␣ induced NF-B DNA binding activity, whereas there is no significant change observed in the C15S Cr-TRP16-transfected cells. These results clearly indicate that overexpression of Cr-TRP16 significantly enhances TNF-␣ induced NF-B activation in HeLa cells.

Cr-TRP16 Augments TNF-␣-induced NF-B-dependent
Reporter Gene Expression-Although we observe that Cr-TRP16 increases NF-B activation in the NF-B DNA binding assay, DNA binding alone does not always correlate with NF-B-dependent gene transcription, indicating there could be some additional regulatory steps (44). To determine the effects of Cr-TRP16 on TNF-␣-induced NF-B-dependent reporter gene expression, we transiently transfected HeLa cells with Cr-TRP16 or C15S Cr-TRP16 along with NF-B to regulate a luciferase reporter construct and co-transfected with a ␤-galactosidase vector. A 2-fold increase in luciferase activity was observed after stimulation with TNF-␣ alone (Fig. 6B), whereas the wild type Cr-TRP16-transfected HeLa cells show a 6-fold increase in TNF-␣-mediated NF-B-dependent luciferase expression. Interestingly, no significant effect was observed in the cells transfected with C15S Cr-TRP16. These results show that Cr-TRP16 can potentiate TNF-␣-inducible NF-B-dependent reporter gene expression. Effect of Cr-TRP16 on DNA Binding of NF-B/p65 C38S Mutant-The mammalian NF-B family includes five related proteins (p50, p52, p65, c-Rel, and RelB) that form various combinations of homo-and heterodimers to control the activity of numerous genes. Both p50 NF-B and p65 NF-B have redoxsensitive Cys residues that must be in the reduced state for the proteins to participate in transcriptional regulation. The binding of NF-B to DNA requires that the Cys-62 of the NF-B p50 subunit or Cys-38 of the NF-B p65 subunit be reduced (43). Previous studies showed that the C38S mutant of p65 NF-B shows severalfold higher DNA binding activity (43,(45)(46)(47).
Our biophysical studies (AUC, nonreducing SDS, and two-dimensional NMR experiments), mutation, and TNF-␣ induced NF-B DNA binding activity confirm that only wild type Cr-TRP16 forms a dimer under an oxidized condition and not the C15S mutant. Furthermore our experiment on the effect of Cr-TRP16 on DNA binding of the NF-B-p65 C38S mutant shows that Cr-TRP16 directly reduces Cys-38 of the p65 NF-B subunit and exists as dimer inside cells. As shown in Fig. 7A, in the presence of Cr-TRP16 there is a 3-fold increase in the DNA binding activity of wild type p65. This experiment indicates that the enhancement of the DNA binding activity of wild type p65 is only due to reduction of Cys-38 of p65 by Cr-TRP16 (thereby increasing its DNA binding activity). However, C15S Cr-TRP16 could not enhance the wild type p65 activity. Overall, our results clearly demonstrate that the Cys-15 residue is crucial for the dimerization of Cr-TRP16, and only the dimeric form plays a pivotal role under physiological conditions in the regulation of NF-B activation.

DISCUSSION
Sequence alignment reveals that the catalytic WCPPC motif is evolutionarily conserved and indicates the potential  Luciferase activity was measured using luciferase assay kit (Promega). Relative light units were measured by a Tecan (Durham, NC) plate reader with the function of measuring luminescence, and the data were normalized to ␤-galactosidase activity. All of the luciferase experiments were done in triplicate and repeated three or more times. FIGURE 7. Cr-TRP16 enhances NF-B DNA binding activity by reducing cysteine residues on p65. A, effect of wild type or C15S Cr-TRP16 on NF-B activation. HeLa cells were transfected with 1 g of pcDNA3.1 or pcDNA-p65 or pcDNA-p65C38S expression plasmid and co-transfected with indicated amounts of wild type and mutant Cr-TRP16 for 36 h, and then nuclear protein extracts were prepared using a nuclear extraction kit (Active Motif) according to the manufacturer's instructions. Nuclear extracts from HeLa cells were assayed for NF-B p65 activation using the TransAM NF-B p65 kit. B, HeLa cells were transfected with 1 g of pcDNA3.1 or pcDNA-p65 or pcDNA-p65C38S expression plasmid. The nuclear extracts were subjected to Western blot analysis. The level of poly(ADP-ribose) polymerase served as loading control.
importance of this motif even in 16-kDa Trx-like proteins. Although the active site of the 16-kDa Trx-like proteins in mammals has undergone marked changes, several other regions remain well conserved, when compared with the 12-kDa Trxs. This observation suggests that the members of the 16-kDa Trx-like protein family have evolutionarily diverged from the 12-kDa Trxs at an early stage, from Caenorhabditis elegans to human (17). Several multi-domain Trx-like proteins have evolved, retaining the Trx fold (49). Although the Trx superfamily has received much attention, only a few structural and functional studies have been undertaken to analyze the multi-domain Trx proteins. Cr-TRP16 has a fold similar to that of tryparedoxin, a 16-kDa Trx-like protein from the parasitic protozoa C. fasciculate (Trypanosoma) (50). In particular, the active site motif, CPPC, is homologous to that of the corresponding motif in the protozoan Trx-like protein (51). However, functionally the Cr-TRP16 is quite different from the tryparedoxin. In contrast, despite notable differences in the amino acid sequence of the catalytic site, Cr-TRP16 appears to be functionally similar to human Trx (52) and regulates NF-B activity (17).
Although most of the 16-kDa Trx-like proteins do not have an extra cysteine, Cr-TRP16 possesses an additional Cys residue at the N terminus. Therefore, it is important to identify which Cys residue is redox-sensitive in Cr-TRP16. Our previous studies have shown that only Cys-38 and Cys-41 are redoxsensitive (17). However, Cys-15 could not react with iodoacetamide and remains inactive toward redox reactions. This prompted us to investigate the functional significance of this Cys and hypothesize that Cr-TRP16 could form a homodimer and regulate NF-B, like hTrx1. Mutation, AUC, nonreducing SDS, and two-dimensional NMR experiments all confirm that only wild type Cr-TRP16 can form a dimer under an oxidized condition and not the C15S mutant. However, it was unclear whether the dimer or monomer is the active form in vivo.
To clarify the physiological role of the N-terminal Cys-15 residue of Cr-TRP16, we overexpressed the wild type and C15S mutant proteins in HeLa cells, stimulated the cells with TNF-␣, and measured NF-B activation. We found that wild type Cr-TRP16 could increase the TNF-␣ induced NF-B DNA binding activity but not the C15S mutant-transfected cells (Fig. 6). Moreover, we observed a similar pattern of results in TNF-␣induced NF-B-dependent luciferase expression in the cells that were transfected with wild type and C15S mutant (Fig. 7).
The Cys-38 residue in the p65 subunits of NF-B has been identified to be crucial for DNA binding (47). We investigated the molecular mechanism by which Cr-TRP16 directly reduces Cys-38 of the NF-B p65 subunit and exists as dimer inside the cell. We found that in the presence of wild type Cr-TRP16, there is a 3-fold increase in the DNA binding activity of wild type p65 (Fig. 7A). However, C15S could not enhance the p65 activity. Thus, our results suggest that the dimeric form of Cr-TRP16 must modify this cysteine residue of p65, leading to NF-B DNA binding activity.
Even though the C15S mutant contains intact redox-active Cys-38 and Cys-41 residues, it could not regulate the NF-B activity, suggesting that Cr-TRP16 exists as a dimer under the physiological condition and that Cys-15 is the crucial residue for the dimer formation. Similar to human Trx1, Cr-TRP16 also up-regulates TNF␣-induced NF-B activation. Interestingly, although Cr-TRP16 lacks the C-terminal extra Cys residue, it forms a dimer with a molecular mass of 32 kDa via N-terminal Cys-15. In addition, human Trx1 is a protein highly homologous to Cr-TRP16 and shares 54% sequence similarity. Notably, it has been reported that the NF-B-modulating activity is rather conserved in human TXN-6 (a thioredoxin like protein-6) because it could cross-enhance the horseshoe crab NF-B DNA binding activity as well (34), suggesting the functional conservation among the 16-kDa Cr-TRP16 and 24-kDa human Trxs. Therefore, it is reasonable to postulate that similar to the human Trx1 dimer (23,53), dimeric Cr-TRP16 reduces Cys-38 in the human p65 N-B DNA-binding motif to enhance its DNA binding activity. These results are consistent with our biophysical studies and the present evidence that the dimer of Cr-TRP16 is the naturally occurring form.
Studies by Matthews et al. (23) and Meyer et al. (54) show that the DNA binding activity of NF-B follows redox regulation through its modulatory Cys residues. Furthermore, the NF-B subunits, p50 and p65, contain well conserved cysteine residues in their DNA-binding loops (55,56). Activation of the DNA binding properties of human NF-B requires that the Cys-62 of the p50 subunit and Cys-38 of p65 subunit be reduced. In the p50 homodimer-DNA complex structure (55,56), the p65 homodimer binds to DNA in a channel, wherein one end is closed by the C-terminal dimerization domain, and Based on the previously published human NF-B p50 homodimer-DNA complex and human Trx-NF-B p50 peptide complex structures and our structural, biophysical, and biochemical results, we propose that Cr-TRP16 forms a homodimer by making a disulfide link through Cys-15 and is likely to occupy the location where DNA binds to NF-B. The active site motif WCPPC of Cr-TRP16 (through Cys-38 and Cys-41) is likely to interact with the redox active Cys-62 of NF-B dimer. For the ease of residue numbering, a p50 homodimer is assumed for NF-B, and the C-terminal cysteines of NF-B that promote dimerization of NF-B are not shown for simplicity. the front is closed by two loops that contain Cys-38, one from each subunit. If a disulfide is formed between the two Cys-38 residues in uncomplexed NF-B, then DNA cannot bind to the p65 homodimer. Similarly, if a disulfide bridge is formed in the p65 homodimer-DNA complex, then DNA cannot be released (52). This implies that the activation of NF-B DNA binding activity begins with IB degradation, reduction of NF-B by Trx followed by translocation of NF-B to the nucleus (57,58) for effective DNA binding and transactivation (58). Based on this argument and our structure-function analyses, we propose a model on how Cr-TRP16 could bind to the p65 homodimer or p50/p65 heterodimer in vivo (Fig. 8).
Several diseases, including cancer, autoimmune diseases, and infectious diseases, are mediated through the effects of Trx and Trx reductase and imbalance in NF-B activation. The structure-based functional studies clearly demonstrated that 1) Cr-TRP16 regulates NF-B similar to the 12 kDa human Trx, 2) Cys-15 is the key residue for dimerization, and 3) only the dimeric Cr-TRP16 regulates NF-B activity. This study helps in further understanding of the interactions of Trxs and Trx-like proteins with other metabolic pathways and their physiological relevance, which will lead to promising therapeutic development to modulate the actions of Trx and NF-B.