Myotrophin/V-1, a Protein Up-regulated in the Failing Human Heart and in Postnatal Cerebellum, Converts NFκB p50-p65 Heterodimers to p50-p50 and p65-p65 Homodimers*

Myotrophin/V-1 is a cytosolic protein found at elevated levels in failing human hearts and in postnatal cerebellum. We have previously shown that it disrupts nuclear factor of κB (NFκB)-DNA complexes in vitro. In this study, we demonstrated that in HeLa cells native myotrophin/V-1 is predominantly present in the cytoplasm and translocates to the nucleus during sustained NFκB activation. Three-dimensional alignment studies indicate that myotrophin/V-1 resembles a truncated IκBα without the signal response domain (SRD) and PEST domains. Co-immunoprecipitation studies reveal that myotrophin/V-1 interacts with NFκB proteinsin vitro; however, it remains physically associated only with p65 and c-Rel proteins in vivo during NFκB activation. In vitro studies indicate that myotrophin/V-1 can promote the formation of p50-p50 homodimers from monomeric p50 proteins and can convert the preformed p50-p65 heterodimers into p50-p50 and p65-p65 homodimers. Furthermore, adenovirus-mediated overexpression of myotrophin/V-1 resulted in elevated levels of both p50-p50 and p65-p65 homodimers exceeding the levels of p50-p65 heterodimers compared with Adβgal-infected cells, where the levels of p50-p65 heterodimers exceeded the levels of p50-p50 and p65-p65 homodimers. Thus, overexpression of myotrophin/V-1 during NFκB activation resulted in a qualitative shift by quantitatively reducing the level of transactivating heterodimers while elevating the levels of repressive p50-p50 homodimers. Correspondingly, overexpression of myotrophin/V-1 resulted in significantly reduced κB-luciferase reporter activity. Because myotrophin/V-1 is found at elevated levels during NFκB activation in postnatal cerebellum and in failing human hearts, this study cumulatively suggests that myotrophin/V-1 is a regulatory protein for modulating the levels of activated NFκB dimers during this period.

Myotrophin/V-1 (Myo/V1) 1 protein was initially characterized in the mammalian heart, where it was called myotrophin (1), and in the rat cerebellum, where it was called V-1 (2). It was later found to be ubiquitously expressed in all mammalian tissues (3,4). Myo/V1 is a 12-kDa ankyrin repeat-containing intracellular protein that has been found at elevated levels in failing human hearts (5) as well as in the hearts of spontaneously hypertensive rats (6). Although Myo/V1 was originally described as a trophic protein (myotrophin) exhibiting growth properties exogenously on rat neonatal myocytes (1), other studies showed that this protein was only present in intracellular space (2,(7)(8)(9) and its trophic growth properties on neonatal myocytes were not confirmed (10). Moreover, this protein was originally identified and isolated only from an intracellular location (1), and a transcriptional regulatory function has been proposed (3,9).
Since its discovery, investigators have proposed various functions for Myo/V1 protein (1)(2)(3). In the postnatal rat cerebellum, the cellular level of soluble Myo/V1 was found to be transiently up-regulated immediately after birth and later declined displaying a unique pattern of expression among 120 soluble proteins, implicating its role during postnatal cerebellum development (2). Because of its aberrant expression in genetically defective cerebellar granular cells, this protein was proposed to play a role in granular cell differentiation process (Unigene Mm.4123) (7). To date, however, the molecular function of Myo/V1 protein is still lacking. We recently reported that Myo/V1 protein exhibits significant homology to IB␣ protein and that Myo/V1 can disrupt the NFB-DNA complexes in vitro (3). Utilizing our recombinant Myo/V1 protein, the NMR structure of Myo/V1 was determined (11,12), and the ankyrin repeats of Myo/V1 exhibited structural features similar to those of IB␣ (13,14) at the three-dimensional level.
In response to a variety of pathophysiological and developmental signals, the NFB/Rel family of transcription factors are activated and form different types of hetero-and homodimers among themselves to regulate the expression of target genes containing B-specific binding sites (15,16). Among the activated NFB dimers, the p50-p65 heterodimers are known to be involved in enhancing the transcription of target genes and the p50-p50 homodimers in transcriptional repression (17)(18)(19)(20)(21)(22). However, the p65-p65 homodimers are known for both transcriptional activation and repressive activity against target genes (23)(24)(25)(26)(27)(28)(29)(30)(31). NFB activation is regulated at multiple levels. The dynamic shuttling of the inactive NFB dimers between the cytoplasm and nucleus by IB proteins (32)(33)(34)(35) and its termination by phosphorylation and proteasomal degradation (36,37), direct phosphorylation (38), acetylation of NFB factors (39), and dynamic reorganization of NFB subunits among the activated NFB dimers (40 -42) have all been identified as key regulatory steps in NFB-mediated transcription process. B DNA binding sites with varied affinities to different NFB dimers (43) have been discovered in the promoters of several eukaryotic genes (16,20,22,44,45), and the balance between activated NFB homo-and heterodimers ultimately determines the nature and level of gene expression within the cell (18,22). However, thus far the underlying molecular mechanism for the generation and dynamic reorganization of NFB dimers during chronic activation is unknown. Here, for the first time, we show that Myo/V1 acts as a "zipper chaperone" protein to generate NFB homodimers from monomeric p50 proteins and with its "unzipping" function converts the transcriptionally active p50-p65 heterodimers to transcriptionally repressive homodimers in HeLa cells, thus attenuating NFB-mediated transcription.

Recombinant Expression Plasmids and Adenoviruses-Recombinant
Myo/V1 protein was expressed in Escherichia coli in two different forms (ϳ12-kDa full-length and histidine-tagged ϳ14-kDa fusion protein) using pET expression vectors as described before (3). The following mammalian expression plasmids were constructed by recombinant DNA methods. For pcDNA-AM1.1-Myo/V1, because of the poor translation initiation Kozak site in the Myo/V1 mRNA, we engineered a heterologous highly efficient Kozak site into 5Ј-untranslated region of Myo/V1 so that Myo/V1 is expressed at high levels in mammalian cells. In vitro transcription and translation with pcDNA3-AM1.1-Myo/V1 template DNA confirmed the synthesis of 12-kDa Myo/V1 protein at higher levels than the native Myo/V1 mRNA (data not shown). For pB-tk-luc, the parent chloramphenicol acetyltransferase reporter plasmid containing a minimal thymidine kinase promoter and two B enhancer sites was obtained and replaced with the coding region of luciferase enzyme. pRSV-RelA vector expressing p65 was obtained through the AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, National Institutes of Health. The expression plasmid is from Dr. Gary Nabel and Dr. Neil Perkins. The recombinant adenoviruses expressing Myo/V1 and ␤-galactosidase were constructed as follows. For AdMyo/ V1/Ad␤gal, the respective expression plasmids pcDNA3-AM1.1-Myo/V1 and pcDNA3-␤gal were incorporated into Ad5 adenovirus through allelic recombination. Recombinant adenoviruses were propagated, purified, and titered as previously reported (46).
Cell Biology Techniques-HeLa cells (ATCC-CCL2) were maintained in minimal essential medium. Cell fixation and indirect immunofluorescence studies were performed as previously described (47). Cytoplasmic and nuclear extracts were prepared using NE-PER nuclear and cytoplasmic extraction reagents (Pierce). Plasmid DNA transient transfection experiments were performed using FuGENE6 reagent (Roche). Luciferase assays were conducted with reagents from Promega Inc. The protein concentration was determined by BCA method using Pierce reagents.
HeLa cells were treated for 2 h with TNF (50 ng/ml) to induce NFB. For superinduction of NFB, cells were further treated with cycloheximide (10 g/ml). After 2 h of treatment, cells were harvested and subcellular fractionation for cytoplasmic and nuclear extracts were carried out as previously described (48). Briefly, HeLa cells were harvested at 150 ϫ g and Dounce-homogenized in a hypotonic lysis buffer (buffer A; Ref. 49), and nuclei were collected at 4300 ϫ g. The nuclei were extracted with buffer C (49) and were used for GSA and Western blot analysis. After the removal of the nuclei, the supernatant was further centrifuged at 20,000 ϫ g for isolating mitochondrial fractions. The remaining cytosolic supernatant was further concentrated by acetone precipitation. Identical amounts of cytosolic (40 g), nuclear extract (30 g), and mitochondrial proteins (35 g) were fractionated on a 10% Tris-Tricine SDS-PAGE and processed for Myo/V1 ECL immunoblotting.
Quantification of Individual NFB Dimers in HeLa Cell Nuclear Extracts-We developed a more accurate method to quantify NFB dimers in mammalian cell nuclear extracts. We chose three B sites from native genes, which have been previously well characterized to exhibit high affinity to individual NFB dimers. The rationale behind this approach is with varied amounts of activated NFB dimers in the mammalian cell, using a single B oligo, which exhibits high affinity to one dimer, and not to others and using it to quantify the various NFB dimers will yield erroneous results. Moreover, the ideal B oligo, which exhibits equal affinity toward all NFB dimers with varied mobilities on GSA to distinguish them, does not exist. Therefore, the conventional B-Ig/HIV oligonucleotide (5Ј-AGTTGAGGGGACTTTCCCAGGC-3Ј from Santa Cruz Biotechnology, Inc.) was used to quantify only p50-p65 heterodimers because it exhibited high affinity only toward p50-p65 heterodimers. Because this B site exhibited lower affinities toward p50-p50 and p65-p65 homodimers (5-and 15-fold, respectively) (53), we chose two other B oligonucleotides B#SeqB (44) and B#u-iNOS (45), which were previously shown to exhibit high affinity and exclusive specificity toward p50-p50 and p65-p65 homodimers (44,45). The B#SeqB (5Ј-GTAGGGGGCCTCCCCGGCTCGAGATCCTATG-3Ј) and B#u-iNOS (5Ј-GTACCGGAAATTCCGGGCTCGAGATCCTATG-3Ј) oligonucleotides were custom synthesized (Synthetic Genetics, CA) and used for quantifying respective NFB homodimers. To determine the relative levels of various NFB dimers, nuclear extracts (ϳ20 g) from AdMyo/V1-and Ad␤gal-infected cells were incubated with radiolabeled individual B oligonucleotides (B-Ig, #SeqB; #u-iNOS; 25,000 cpm with similar specific activities), and the resulting DNA-protein complexes were fractionated on the same 4% PAGE. Radiographic images of the gels were captured using a STORM 860 imager (Molecular Dynamics, CA) and quantified using ImageQuant 4.2 software. NFB antibody supershifts were done as described previously; however, p65 antibody from Geneka Biotechnology (Montreal, Canada) was used in these experiments. Additionally, heterologous chase experiments were done to further confirm the nature of the NFB dimers.

Myo/V1 Is Localized in the Cytoplasm and Nucleus-Indirect
immunofluorescence studies were conducted to locate the native Myo/V1 protein in HeLa cells. Under basal conditions (Fig.  1A), Myo/V1 was predominantly observed in a wide area of the cytoplasm surrounding the nucleus. Further analysis with confocal microscopy revealed that native Myo/V1 was also present in the nucleus to a lesser extent (spotted green fluorescence in Fig. 1C). Upon treatment with TNF for 1 h, Myo/V1 was found to cluster around the perinuclear region in the cytoplasm. Additionally, Myo/V1 was found to increase slightly within the nucleus (green FITC masking the blue DAPI nuclear staining, resulting in pale blue nucleus) suggesting migration of Myo/V1 to the nucleus (Fig. 1B). Because TNF is known to reorganize the cytoskeleton (54 -56), the present observation of increased perinuclear clustering and nuclear migration of Myo/V1 in TNF-treated HeLa cells (Fig. 1B) suggests that Myo/V1 might be associated with the cytoskeleton and its associated organelles and might participate in the signal transduction process from these locations.
Myo/V1 Translocates to Nucleus during Sustained Induction of NFB-Cellular partitioning studies were performed to determine the location of Myo/V1 under the NFB inducing conditions. Cells were treated with TNF or TNF plus CHX for 2 h. Cytoplasmic, mitochondrial, and nuclear extracts were prepared and immunoblotted for Myo/V1 protein (Fig. 2). Under basal conditions the majority of Myo/V1 was present in the cytoplasmic fraction (lane 5 in Fig. 2B) and only a small quantity in the nuclear fraction (lane 3 in Fig. 2B). After TNF stimulation for 2 h, Myo/V1 levels in the cytoplasmic fraction did not change significantly (data not shown). However, with TNF plus CHX stimulation, the levels of Myo/V1 reduced significantly in the cytoplasmic fraction compared with control (lanes 5 and 6 in Fig. 2B). Additionally, a simultaneous increase in the levels of Myo/V1 was observed in the nucleus (lane 4 in Fig. 2B), indicating that the translocated Myo/V1 migrated from cytoplasm. We did not observe any Myo/V1 in mitochondrial fractions (lanes 1 and 2 in Fig. 2B). These data suggest that, under acute stress conditions, a major relocalization of Myo/V1 first occurs within the cytoplasm (Fig. 1B), following which Myo/V1 is translocated to the nucleus.
Myo/V1 Physically Interacts with NFB Proteins-To confirm that Myo/V1 physically interacts with NFB proteins, co-immunoprecipitation studies were conducted both in vitro and in vivo. Fig. 3A shows that, in vitro, p50, p65, and c-Rel proteins strongly interacted with Myo/V1 protein (lanes 2-4). To further confirm these physical interactions in vivo, HeLa cells were treated with diluent (lanes 8 -10 in Fig. 3B), TNF plus CHX (lanes 5-7 in Fig. 3B), or phorbol ester (lanes 11-13 in Fig. 3C) for 2 h or infected with AdMyo/V1 for 12 h (lanes 14 and 15 in Fig. 3C), cellular extracts were prepared, and coimmunoprecipitation experiments were conducted. In control diluent-treated cells, Myo/V1 did not associate with any of the NFB proteins (lanes 8 -10 in Fig. 3B). However, in TNF plus CHX-or AdMyo/V1-treated cells, Myo/V1 predominantly asso- Myo/V1 Promotes the Formation of p50-p50 Homodimers from Monomeric p50 Proteins in Vitro-Because our earlier studies indicated that Myo/V1 disrupts and induces the formation of new NFB-DNA complexes (3) and can physically interact with NFB proteins in vitro and in vivo (Fig. 3), we designed a detailed study to identify the role of Myo/V1 in altering the specific NFB dimers in vitro. For this purpose, we initially studied the effect of Myo/V1 on monomeric NFB proteins. NFB proteins, p50, and p65 have low affinity to form homodimers among themselves, and a majority of the proteins remain monomeric without binding to B DNA. However, when incubated together, they exhibit high affinity to its heterologous partner and they form heterodimers very easily. Purified recombinant p50 protein and truncated p65⌬ protein (from Dr. Gaurishankar Ghosh) was used for these studies. First, monomeric p50 proteins were mixed at very low concentrations (2.5 ng in 50 l; 2 nM) with Myo/V1 (0 -100 ng) and the formation of p50-p50 homodimers were studied using gel-shift assay (lanes 1-5 in Fig. 4A). Similarly, monomeric p65⌬ proteins were mixed at very low concentrations (0.5 g in 50 l; 7 M) with Myo/V1 (0 -100 ng), the formation of p65⌬-p65⌬ homodimers Nuclei were stained with DAPI, and images of FITC and DAPI staining (60ϫ; A and B) were captured using a phase contrast Nikon Eclipse ⌭800 microscope. A nonfluorescent blue color was assigned for DAPI image during image capture and superimposed with FITC image using Metaview software. B, HeLa cells treated with TNF for 2 h. C, image of untreated HeLa cells was acquired using a laser-scanning microscope. White arrows denote the cytoplasmic and nuclear localization of Myo/V1.

FIG. 2. Myo/V1 is translocated to nucleus during superinduction of NFB.
A, GSA showing the B DNA binding activity of HeLa cell nuclear extracts. Cells were treated for 2 h with TNF (50 ng/ml) (lanes 2 and 3) or TNF plus CHX (10 g/ml) (lanes 4 and 5) to superinduce NFB. B DNA binding activity was performed with 20 g of nuclear extracts. B, Western blot analysis of native Myo/V1. HeLa cells were treated with TNF (50 ng/ml) plus CHX (10 g/ml) for 2 h and cytoplasmic (CYT; 40 g), mitochondrial (MC; 35 g), and nuclear (NE; 30 g) extracts were prepared, fractionated on a 10% Tris-Tricine SDS-PAGE, and processed for ECL immunoblotting. CON, control.
were studied, and the effect of Myo/V1 was studied (lanes 6 -10 in Fig. 4A). Although the addition of Myo/V1 (0 -100 ng) led to an increased formation of p50-p50 homodimers (lanes 1-5 in Fig. 4A), there was no enhanced formation of p65⌬-p65⌬ homodimers (lanes 6 -10 in Fig. 4A). We additionally tested the monomeric p65⌬ proteins with higher concentrations of Myo/V1 (0 -1500 ng), which also did not show any formation of additional p65⌬-p65⌬ homodimers (data not shown). These results indicate that Myo/V1 actively promotes the formation of p50-p50 homodimers by dynamically interacting with the p50 subunit of NFB. Although Myo/V1 was found to be physically associated with p65 and its homologues in NFB-activated HeLa cells (Fig. 3), we do not observe the active promotion of p65-p65 homodimers by Myo/V1 in vitro.
Later, to further confirm the nature of Myo/V1-generated dimers, the effect of NFB antibodies on these complexes was studied (Fig. 4C). The p50-p65⌬ heterodimers were pre-assembled at equimolar ratio as previously described, and Myo/V1 was added at two (100 and 400 ng) concentrations (lanes 2 and 3 in Fig. 4C). These reactions were performed in triplicate for p50 and p65 antibody supershifts (lanes 5, 6, 8, and 9 in Fig  4C). As expected, we observed both an upward shift in the mobility of Myo/V1-shifted dimers occupying the position of p50-p50 homodimers (lanes 2 and 3 in Fig. 4C) as well as an increased formation of p50-p50 homodimers with increasing amounts of Myo/V1 (lanes 2 and 3 in Fig. 4C) in these reactions. Lanes 4 -6 confirm that these dimers have p50 proteins as revealed by p50 antibody supershifted complexes. Lanes 7-9 exhibit the inhibitory effects of p65 antibody on Myo/V1-generated dimers. Because the p65 antibody exhibits inhibitory activity on p65 proteins (Rel DNA binding domain antibodies are not supershiftable), the formation of heterodimers as well as the conversion to homodimers was also affected (lanes 7-9). Hence, the subtle upward shift observed in lanes 2 and 3 was not observed in lanes 8 and 9. Lanes 8 and 9 contain a mixture of p50-p50 homodimers and p50-p65⌬ heterodimers. Close examination of the migratory patterns and the levels of these dimers (lanes 8 and 9 in Fig. 4C) reveals that, in the presence of p65 antibody, Myo/V1 promoted the formation of p50-p50 homodimers from the residual free monomeric p50 proteins, although at a reduced level. The splitting of the heterodimers as observed in lanes 2 and 3 did not occur in these reactions (lanes 8 and 9) because of the presence of p65 antibody. The NFB dimers that appear in lanes 7-9 are predominantly p50-p50 homodimers, as evidenced by its location.
To further confirm the chaperone functions of Myo/V1, we conducted another experiment (Fig. 4D), where the amount of p65⌬ was kept excess by 100-fold to that of p50 so that, in addition to the preformed p50-p65⌬ heterodimers, free p65⌬ proteins will be present in the reaction. The effect of Myo/V1 (lanes 1-3 in Fig. 4D), and the effect of NFB antibodies on these Myo/V1-shifted complexes were studied (lanes 6 -8 and 10 -12 in Fig. 4D). The rationale for this experiment is that, when the p50-p65 heterodimers are split by Myo/V1, there will be a competition among the released monomeric proteins to form either homodimers or heterodimers. With excess p65 protein in the reaction and because of its high affinity toward p50 protein to form heterodimers (43), there will be a continuous  7) and diluent (lanes 8 -10) were immunoprecipitated with IgG, p50, and p65 antibodies, and the bound proteins were eluted and immunoblotted for Myo/V1. C, cellular extracts from HeLa cells treated with phorbol ester (lanes [11][12][13] and AdMyo/V1 (lanes 14 and 15) were immunoprecipitated with p50, p65, and c-Rel antibodies, and the bound proteins were eluted and immunoblotted for Myo/V1. formation of new p50-p65 heterodimers as the old heterodimers are being converted by Myo/V1. Thus, in this experiment Myo/ V1-generated "intermediate products" can be identified. The p50-p65⌬ heterodimers were pre-assembled at 1:100 ratios, and Myo/V1 was added at two (100 and 400 ng) concentrations (lanes 2 and 3 in Fig. 4D). These reactions were performed in triplicate for p50 and p65 antibody supershifts (lanes 5-7 and 9 -11 in Fig. 4D). As expected (shown in Fig. 4D), both heterodimers and homodimers were formed when Myo/V1 was present at high concentrations with excess p65 (lane 3). The complete subtle upward shift observed in lane 3 of Fig. 4C was not similarly observed in lane 3 of Fig. 4D, strongly indicating the presence of both heterodimers and homodimers. The presence of both of these dimers is more clearly seen in lanes 7 and 11. Additionally, we also observed a simultaneous reduction in p65⌬-p65⌬ homodimers (lanes 1-3) with the addition of Myo/ V1. Because Myo/V1 does not have any direct effect on p65 per se (lanes 6 -10 in Fig. 4A), the reduction in p65⌬-p65⌬ ho-modimers in lanes 1-3 is a result of its incorporation into p50 protein in the reaction to form p50-p65⌬ heterodimers. Furthermore, the absence of p50-p65⌬ heterodimers in lanes 6 and 10 and the presence of these dimers in lanes 7 and 11 correlate with the reduction of p65⌬-p65⌬ homodimers in lanes 7 and 11. Lanes 5-7 confirm that the Myo/V1-generated dimers have p50 proteins, as revealed by supershifted complexes. Because of excess p65, p50 antibody supershifts did not occur completely. Because the p65 antibody used in lanes 9 -11 exhibits inhibitory activity on p65 proteins, the complete conversion of heterodimers to homodimers is affected. Thus, these results further confirm the chaperone functions of Myo/V1 in promoting NFB homodimers and the formation of p50-p65⌬ heterodimers is an indirect event promoted by the chaperone functions of Myo/V1 on p50.
Results show that overexpression of AdMyo/V1 enhanced the generation of p50-p50 (lane 2 in Fig. 5A) and p65-p65 (lane 4 in Fig. 5A) homodimers compared with the control Ad␤gal (lanes 1 and 3 in Fig. 5A). Interestingly, p50-p65 heterodimers are less abundant in AdMyo/V1 compared with the Ad␤gal-overexpressing cells (lanes 5 and 6 in Fig. 5A). Thus, the observation of a simultaneous decrease in the levels of p50-p65 heterodimers and an increase in p50-p50 and p65-p65 homodimers suggests a conversion from heterodimers to homodimers (Fig.  5B). Moreover, an unequal quantitative shift (Fig. 5B) occurred between p50-p50 homodimers and p50-p65 heterodimers in Myo/V1-overexpressing cells. Although the levels of p50-p50 homodimers were elevated by 70%, the p50-p65 heterodimers declined by 26% (Fig. 5B). This observation suggests that Myo/ V1, in addition to generating the p50-p50 homodimers from the p50-p65 heterodimeric substrates, might directly generate nascent p50-p50 homodimers from other in vivo substrates, further confirming our previous in vitro results (Fig. 4B). Furthermore, analysis of the relative levels of NFB dimers (Fig. 5C) Fig. 5A) was unexpected. We believe this is because of "antibody-induced effect" rather than residual p50-p65 heterodimers binding to these B oligos. The p50 antibody after binding to a p50-p65 heterodimer could change the affinity of the "supershifted heterodimeric ternary complex" now to recognize and bound to a otherwise p65-p65 homodimer specific oligo. Although antibodies used in gel-shift assays are generally described as "supershifting antibodies," these antibodies change the affinity of homo-and heterodimers and thereby affect their binding to B sites. Evidence of these artifacts is seen in lanes 12 and 15 where increased (lane 12) and decreased (lane 15) binding is observed after the antibody was added. Therefore, as described previously (44,45) without the NFB antibodies, B#seqB and B#u-iNOS oligos exhibit high affinity and exclusive specificity toward p50-p50 and p65-p65 homodimers. Additionally, we conducted heterologous B oligo chase experiments (B#u-iNOS-bound complexes chased with 100ϫ cold B#SeqB and B-Ig oligos), and the results revealed that Myo/V1-overexpressing HeLa cells still produced higher levels of p65-p65 homodimers compared with Ad␤gal (data not shown). Additional chase experiments (B#SeqB complexes chased with 100ϫ cold B#u-iNOS and B-Ig oligos; B-Ig complexes chased with 100ϫ cold B#50 -2 and B#u-iNOS oligos) were done, and the results were exactly the same as in Fig. 5A (data not shown).
Myo/V1 Represses NFB-mediated Transcription Process-To identify the functional relevance of Myo/V1 with respect to NFB dimers, we studied the role of Myo/V1 on NFBmediated transcription. HeLa cells were cotransfected with pB-tk-luc, p65 pRSV-RelA, and pcDNA-AM1.1-Myo/V1 expression vector. Forty-eight hours after transfection, HeLa cells were harvested and lysed and luciferase enzyme activity was measured. The results (Fig. 6) show that overexpression of Myo/V1 significantly reduced the NFB-mediated transcription by 42% (p Յ 0.05) on p65-mediated luciferase reporter activity. Because p50-p65 heterodimers are not available for Twelve hours after infection, nuclear extracts were prepared and GSAs were conducted with three high affinity NFB dimer-specific oligonucleotides (B#seqB for p50-p50 homodimers; B#u-iNOS for p65-p65 homodimers; IgGB oligonucleotide for p50-p65 heterodimers) (lanes 1-6). Supershift experiments with NFB p50 and p65 antibodies were conducted to confirm the nature of the NFB dimers (lanes 7-15). p50 supershift complexes are indicated by *, and p65 supershift complexes are indicated by arrows. B, quantitative comparison of NFB dimers between Ad␤gal and AdMyo/V1. The bar graph shows the -fold change in the levels of individual NFB dimers in relation to Ad␤gal-infected HeLa cells. C, relative levels of NFB dimers in Ad␤gal-and AdMyo/V1-infected cells. The bar graph shows the NFB dimer ratio in relation to p50-p65 heterodimers in Ad␤galand AdMyo/V1-infected HeLa cells. These results are representative of six different experiments, conducted four times at a multiplicity of infection of 10 and twice at a multiplicity of infection of 50.

Three-dimensional Structural Alignment of Myo/V1 with Rel-interacting Ankyrin Repeats of IB␣ Reveals the Potential Chaperone Functions of Myo/V1 on NFB-
To correlate the observed chaperone functions of Myo/V1 with its structural domains, we conducted the three-dimensional superimposition analysis (60) using the three-dimensional NMR structure of Myo/V1 (Fig. 7A) (11,12) and the three-dimensional structure of IB␣ (Fig. 7B) (13, 14), where NFB-interacting domains were well characterized. The superimposed C␣ traces of Myo/V1 with IB␣ (Fig. 7, C and D) show that entire Myo/V1 aligned with ankyrin repeats 3-6 of IB␣ protein, further confirming our earlier primary structural homologies with IB␣ (3). This finding is highly significant because these specific ankyrin repeats of IB␣ interact with the Rel domain of p50 and p65 subunits (Fig. 7, C and D) that are known for dimerization and B DNA binding. Moreover, the ankyrin repeats 1 and 2 of IB␣ (aa 67-142), which mask the NLS signal of p65, the amino-terminal SRD (aa 1-66 of IB␣), and the carboxyl-terminal acidic PEST domain (aa 281-317) that is responsible for degradation of IB␣, are all absent in the Myo/V1 protein. Thus, the fact that Myo/V1 resembles only the Rel-interacting ankyrin repeat domains of IB␣ (AR 3-6) strongly suggests that Myo/V1 might similarly interact with the Rel DNA binding domain of NFB but, unlike IB␣, Myo/V1 might mediate its chaperone functions in a zipper-like fashion using the terminal residues of its three finger loops (Fig. 7, C and D). One of the overlooked functions of IB␣ is its use of these ankyrin repeats (AR 3-6) to assemble the inactive NFB dimers in the cytoplasm. Similar to Myo/V1, IB␣ might also possess zipper chaperone function; however, it is only used to assemble the inactive NFB dimers in the cytoplasm. Because of its other NLS masking ankyrin repeats (repeats 1 and 2) and SRD domain, it behaves as an inhibitor or a carrier of NFB. Because of the observed similarity in Rel-interacting ankyrin repeats, one could envision that IB and Myo/V1 proteins could behave as lock and key of NFB dimers in the mammalian cell.

DISCUSSION
Myo/V1 was first identified as an up-regulated intracellular protein in pathophysiological conditions (1, 4 -6) such as human heart failure, as well as during early organ developmental processes such as in postnatal mammalian cerebellum (2,4,8,10). However, later it was identified to express in every mammalian organ and cell type. Moreover, Myo/V1 is an evolutionarily conserved protein from multicellular organisms to mammals (Fig. 8). Among mammals it shows 97% homology, and with invertebrates it exhibits 67% homology. Putative nuclear import and export signals as well as potential phosphorylation sites were mostly conserved between invertebrates and vertebrates (Fig. 8). Our three-dimensional structural alignment studies (Fig. 7C) further revealed that Myo/V1 possessed only Rel domain-interacting ankyrin repeats and resembles a truncated form of IB␣ protein without the SRD, NLS masking, and PEST degradation domains (Fig. 7D).
Results presented in this study indicate that, under basal conditions, cellular Myo/V1 predominantly present in the cytoplasm and probably associates with cytoskeletal structures (Fig. 1) similar to IB␣ (61, 62). Additionally, low levels of Myo/V1 are also present in the nucleus (Fig. 1C)  Rasmol plug-in software, and captured. D, primary sequence alignment of Myo/V1 with IB␣ using the three-dimensional superimposition analysis data. Purple and dark blue colored residues in IB␣ are known to interact with p50 and p65 proteins, respectively (13,14). Red AR1-6 denote ankyrin repeats of IB␣, and purple AR1-3 denote ankyrin repeats of Myo/V1. PEST is the degradation domain of IB␣. and rat neonatal cardiac myocytes. 2 Our studies also demonstrated that, upon stimulation with TNF, Myo/V1 relocates within the cytoplasm (Fig. 1B) and further translocates to the nucleus during sustained induction of NFB (Figs. 1B and 2B). Because Myo/V1 possesses both nuclear import and export signals (Fig. 8), it is expected that Myo/V1 might shuttle between cytoplasm and nucleus similar to IB family of proteins (32)(33)(34)(35)(36). Co-immunoprecipitation studies showed that, although Myo/V1 interacts with all NFB proteins in vitro, it remains physically associated only with p65 and c-Rel proteins in vivo (Fig. 3). The differences between in vitro and in vivo results might also reflect the potential function of Myo/V1 protein. For example, Myo/V1 may interact dynamically with p50 proteins in vivo in a transient fashion to mediate its zipper chaperone function; however, it remains physically associated with p65 or its homologue c-Rel protein to prevent them from reassociating with p50. In vitro NFB interaction studies indicated that Myo/V1 can generate p50-p50 homodimers from monomeric p50 proteins and can convert the preformed p50-p65 heterodimers into p50-p50 and p65-p65 homodimers (Figs. 4 and 9A). Compared with control Ad␤gal, overexpression of AdMyo/V1 resulted in an increase in both p50-p50 and p65-p65 homodimers exceeding the levels of p50-p65 heterodimers (Fig.  5). A simultaneous decrease in the levels of p50-p65 heterodimers was also observed, suggesting a conversion from transactivating heterodimers to transcriptionally repressive homodimers (17)(18)(19)(20)(21)(22)(23)(24)(25). However, the observation of disproportionate increase in the levels of p50-p50 homodimers compared with the decreased levels of heterodimers (70% increase versus 26% decrease) raises the possibility that Myo/V1 might also generate p50-p50 homodimers from other inactive NFB substrates. Because a majority of the Myo/V1 is localized in the cytoplasm, it could very well be involved in the activation of p50-p50 homodimers by acting on p50-p105 substrates (63) in vivo (Fig. 9B). Alternatively, Myo/V1 could assemble the p50-p50 homodimers from free monomeric p50 proteins in the cytoplasm similar to the conditions mentioned in Fig. 4A. Additionally, Myo/V1 could act on any p50 subunit containing complexes (p50-p65, p50-c-Rel, p50-RelB, and "p50-non-NFB" dimers) in the nucleus and further generate the transcriptionally repressive p50-p50 homodimers (Fig. 9B). In accordance with these observations, overexpression of Myo/V1 also re-sulted in significantly reduced NFB-mediated luciferase reporter gene expression (Fig. 6), which is probably a result of the generation of p50-p50 homodimers. Although we observed an increased level of p65-p65 homodimers during AdMyo/V1 overexpression (Fig. 5A), and Myo/V1 being physically associated with p65 and its homo-2 N. Sivasubramanian, unpublished data.  Model shows potential involvement of Myo/V1 in multiple NFB activation pathways in HeLa cells. B, 1, potential involvement of Myo/V1 in generating p50-p50 homodimers from p50-p105 complexes, and 2, conversion of heterodimers to homodimers by Myo/V1 during sustained activation of NFB. HDAC, histone deacetylase. logues (Fig. 3), in vitro interaction experiments (Fig. 4A) do not suggest an active role for Myo/V1 in the generation of p65-p65 homodimers. Moreover, in overexpression of Myo/V1 in HeLa cells through transient transfection methods (Fig. 6), where "active p50-p65 heterodimers" are not available for Myo/V1 to act upon, we only observed an inhibition of B driven luciferase reporter activity. If Myo/V1 would have actively promoted the generation of p65-p65 homodimers in these cells, we should have observed an increased B-luciferase reporter activity during Myo/V1 overexpression. Because Myo/V1 does not retain or sequester p65 in the cytoplasm 2 under basal conditions (Fig.  3B), the observed inhibition of B-luciferase reporter activity is the result of the active generation of p50-p50 homodimers by Myo/V1, as observed in our in vitro and in vivo experiments (Figs. 4 and 5). Therefore, we believe the increased generation of p65-p65 homodimers during AdMyo/V1 expression is an indirect effect of the conversion of p50-p65 heterodimers by Myo/V1 and the physical association of Myo/V1 with p65 is probably to prevent the p65 from reassociating with p50 (Fig. 9A).
Even though our studies indicate that Myo/V1 could favor the active generation of NFB homodimers in HeLa cells, it is possible that, under certain scenarios (similar to conditions mentioned in Fig. 4D), Myo/V1 could indirectly induce the formation of NFB heterodimers in other cell types. For example, Myo/V1 could potentially act on inactive p50-p65-IB␣ substrate complexes in the cytoplasm and generate active NFB heterodimers in immune cells. Because Myo/V1 has been identified as a "severe combined immunodeficiency complementing gene" (GenBank accession no. D78188), and because this protein is also found in the nuclear extracts of phorbol ester-treated Jurkat T cells, 2 the potential role of Myo/V1 in the activation of other NFB dimers is highly likely.
The data presented in this study were further strengthened when, from previously published independent studies (2, 64 -66), we identified a strong temporal relationship between the intracellular levels of Myo/V1 and NFB activation in vivo in mammalian organs. Myo/V1 was found to occur at elevated levels in failing human hearts (1, 4 -6) and found to be transiently elevated in postnatal mammalian cerebellum (2,7). Interestingly, NFB dimers were also found to be activated in failing human hearts 2 (67) as well as in developing postnatal cerebellum (64 -66) during this period. Specifically, in postnatal cerebellum, the intracellular level of Myo/V1 protein was found to rise at the 2nd day, peak at the 7th day, and rapidly decrease by the 12th day (2). Similarly, independent studies have reported the transient activation of NFB dimers begin to occur around postnatal day 2, peaking at the 7th day, and declining by 12th day in rat (64,65) and mouse cerebellum (66). This strong temporal observation that the intracellular levels of Myo/V1 being elevated exactly at the same time during NFB dimer activation and its decline during the disappearance of NFB dimers in postnatal cerebellum further highlights the important potential role of Myo/V1 even in initial NFB activation. The earlier observation of disproportionate increase in the levels of p50-p50 homodimers compared with the decreased levels of heterodimers in AdMyo/V1-infected cells (Fig. 5) also raises the possibility that Myo/V1 might be involved in the initial activation of p50-p50 homodimers in vivo (Fig. 9B). Therefore, the results presented in this study strongly suggest that not only Myo/V1 is involved in converting heterodimers into homodimers during sustained NFB activation, it could potentially be involved in the initial activation of NFB dimers. The recent identification that entire NFB machinery (NIK, IKKs, IB␣, ubiquitination, and proteasome) is shuttling (68, 69) between cytoplasm and the nucleus of resting mammalian cells (32)(33)(34)(35) and the observation of NFB activation without IB degradation (70,71) changes the central tenet of NFB signaling mechanism. In this changed scenario, several alternate IB-independent pathways (70 -72) have been proposed for NFB activation, and Myo/V1 could potentially be involved in such pathways.
The identification of Myo/V1 as a NFB zipper chaperone (Fig. 9A) to shift the NFB dimer ratio from transcriptionally active heterodimers to transcriptionally repressive homodimers is also highly significant. In fact, in many biological scenarios like B lymphocyte cell differentiation (40 -42), lipopolysaccharide tolerance (73), T-lymphocyte anergy (74), and chronic TNF exposure to mouse myocardium (75), the activated NFB dimers are known to change both quantitatively and qualitatively during these events. For example, during B lymphocyte cell differentiation (40 -42), dynamic exchange of NFB subunits are known to occur among the activated NFB dimers resulting in altered gene expression appropriate for the stage of differentiation. During endotoxin tolerance in monocytes (73), T-lymphocyte anergy (74), and chronic TNF exposure to mouse myocardium (75), the levels of p50-p50 homodimers are known to increase during late stages of these events, suggesting a dynamic qualitative and quantitative change in the activated NFB dimers. Moreover, In human sepsis, large amounts of p50-p50 homodimers exceeding the levels of p50-p65 heterodimers were observed in monocytes of non-survivors compared with the survivors, further suggesting an important role for NFB dimer ratio in the pathophysiology of sepsis (76). Furthermore, the NFB dimer ratio has been shown to determine the ultimate level of target gene expression (18,22). NFB-mediated gene transcription has been proposed in models of myocardial diseases (3,77,78). Taken together with the previous finding that elevated levels of Myo/V1 are found in failing mammalian hearts (5,6), this study collectively suggests that Myo/V1 might play a significant NFB regulatory role at the transcriptional level during chronic pathophysiological conditions such as human heart failure.