MYB Elongation Is Regulated by the Nucleic Acid Binding of NFκB p50 to the Intronic Stem-Loop Region

MYB transcriptional elongation is regulated by an attenuator sequence within intron 1 that has been proposed to encode a RNA stem loop (SLR) followed by a polyU tract. We report that NFκBp50 can bind the SLR polyU RNA and promote MYB transcriptional elongation together with NFκBp65. We identified a conserved lysine-rich motif within the Rel homology domain (RHD) of NFκBp50, mutation of which abrogated the interaction of NFκBp50 with the SLR polyU and impaired NFκBp50 mediated MYB elongation. We observed that the TAR RNA-binding region of Tat is homologous to the NFκBp50 RHD lysine-rich motif, a finding consistent with HIV Tat acting as an effector of MYB transcriptional elongation in an SLR dependent manner. Furthermore, we identify the DNA binding activity of NFκBp50 as a key component required for the SLR polyU mediated regulation of MYB. Collectively these results suggest that the MYB SLR polyU provides a platform for proteins to regulate MYB and reveals novel nucleic acid binding properties of NFκBp50 required for MYB regulation.


Introduction
The MYB proto-oncogene encodes a transcription factor that plays an important role in cellular proliferation and differentiation [1]. While much of the pioneering research on MYB focused on hematopoiesis and malignancies [1], its expression has subsequently been shown to be important in the context of epithelial cancer biology, most notably breast and colon cancer [1]. MYB is essential for the proliferation of ERα positive breast cancer cells [2] and is required for mammary carcinogenesis in murine models [3]. In colorectal cancer (CRC), MYB is frequently over-expressed, a property that correlates with poor prognosis for patients with CRC [4]. MYB co-operates with Wnt to generate intestinal cancers in mice [5] and drives the expression of the gastrointestinal stem cell genes Lgr5, olfm4 and Bmi1 [6].
The regulation of transcription elongation has emerged as an important mechanism used to control eukaryotic gene expression [7][8][9][10][11][12][13][14]. In line with this view, accumulated data has revealed the important role of transcriptional elongation in the control of MYB expression [1]. In the murine hematopoietic system, a direct correlation exists between the expression levels of MYB RNA and transcription elongation arrest (attenuation) that operates in the first intron of the gene [15][16][17][18][19]. Similarly, in CRC lines the rapid down-regulation of MYB expression that occurs during induced differentiation by sodium butyrate, was found to result from transcription elongation arrest within intron 1,~1.7kb downstream of the transcription initiation site [20,21]. Furthermore, MYB expression in breast cancer cells is regulated by a similar mechanism, whereby attenuation is overcome by estrogen receptor binding [2,22]. We previously proposed that MYB elongation arrest was regulated by RNA secondary structure [1,20,21]. We modeled the RNA encoded by the attenuation region and showed that it potentially formed an energetically stable RNA stem loop region (SLR) followed by a 1920-ribonucleotide polyU tract [20,21]. In the case of CRC a high frequency of mutations occur in the MYB SLR that correlate with elevated levels of MYB mRNA observed in primary CRCs and derived cell lines [20,21]. These data suggest that the dysregulation of SLR function may be an important component in the over expression of MYB in colorectal cancer [1].
Despite the body of data indicating that the MYB SLR is an integral mediator of MYB transcriptional elongation, our understanding of how the SLR regulates the elongation process remains poor. The implication of NFκB proteins in the regulation of murine MYB transcriptional elongation via intron 1 sequences [23][24][25] led us to consider a role for NFκB proteins in regulating MYB via the MYB SLR. Here we demonstrate that NFκB proteins p50 and p65 (RelA) promote elongation through the MYB intron 1 attenuation region in an SLR and polyU dependent manner. Furthermore, NFκBp50 directly associates with the MYB SLR RNA through its amino terminal Rel homology domain (RHD). We identify a conserved lysine rich motif within the RHD, that is homologous to the TAR RNA-binding region of HIV-1 Tat and HEXIM 1 proteins [26,27]. Mutation of the lysine rich motif abolished p50 interaction with the SLR and impaired NFκBp50-mediated transcriptional elongation through the SLR. Furthermore, we identify the DNA binding activity of NFκBp50 as a key component NFκBp50 mediated stimulation of MYB SLR polyU elongation. Collectively these results suggest that the MYB SLR and polyU provide a platform for proteins to bind and regulate MYB elongation. The nucleic acid binding activity of NFκBp50 that includes its RNA and DNA binding functions may provide a novel mechanism for the regulation of MYB expression with important implications for NFκB function in other contexts.

Materials and Methods
Cell culture HEK293 (ATCC CRL-11268) cells were cultured in RPMI1640 10%FBS. The propagation and derivation of LIM1215 cells has been described [6]. Murine colon organoids were generated using previously established protocols [6]. For TNFα stimulation, 293 cells were plated overnight and TNFα (Abbiotec, USA) added at the concentrations specified. For 5,6-Dichloro-1-beta-D-ribofuranosylbenzimidazole (DRB; Sigma) treatment of 293 cells, DRB was dissolved in DMSO as recommended. DRB was diluted in RPMI to a final concentration of 30 μM and added to the cells for the times indicated in the figure legends. For NFκB inhibition, 293 cells were plated overnight and BAY 11-7082 (Millipore) added at the concentrations and times indicated in the figure legends.
Plasmids MYB CAT reporters have been described [21]. The MYB TAR CAT reporter was generated by removing the SLR and polyU region from the MYB SLR polyCAT reporter via a XhoI and XbaI digest and subsequently cloning in the HIV-1 TAR sequence. The pBluescriptIIKS-MYB SLR polyU, MYB ΔSLR polyU, MYB SLR ΔpolyU and MYB SLR Δ9 polyU were generated by cloning PCR fragments generated from pALTER1-MYB into pBluescriptIIKS via BamHI and EcoRV sites. pGEM-3Zf MYB SLR polyU, MYB ΔSLR polyU and MYB SLR ΔpolyU were constructed by cloning MYB SLR polyU sequences into pGEM via EcoRI and HindIII sites. pCMVTAT(72) was supplied by Malcolm Martin (NIAID, NIH). pCMV CDK9 and dominant negative pCMV DNCDK9 were gifts from Andy Rice (Baylor College, USA). pcDNA NFκBp50 and NFκBp65 were created by Gateway cloning (Invitrogen) cDNAs into pcDNA3.2/V5 and in the case of NFκBp50 also into pET15b. pCMVβ-galactosidase and pACTCAT-β-actin were a gift from Shunsuke Ishii (RIKEN, Japan). pHIVLTRCAT was obtained from the NIH AIDS Research and Reference Reagent Program. WT and bulge mutant TAR pGEM-3Zf vectors were a gift of Melanie Ott (Gladstone Institute of Virology and Immunology, USA) [28]. pNFκB p50-65 was a gift of Craig Rosen (Roche Institute of Molecular Biology) [29]. Mutagenesis of NFκBp50 and NFκBp50-65 was performed using the Gene Tailor site-directed mutagenesis system (Invitrogen). All oligonucleotide sequences are provided S1 Table. Ectopic protein expression from all mammalian expression plasmids was confirmed by Western blot analysis S1 Fig. CAT reporter assays 293 cells were transfected with 2-3 μg total DNA using Fugene (Roche). pCMVβ-gal was used as a transfection control. Cells were harvested and CAT activity measured as described [6]. Data shown represent an average of at least three experiments.

Chromatin Immunopreciptation (ChIP) assays
ChIP assays were performed as previously described [6]. Anti-RNA polymerase II (H224X; Santa Cruz) was used to detect RNA polymerase II at the MYB SLR polyU.

Nuclear run-on transcription
Nuclear run on assays were performed as described [20,21]. Transcriptional activity was normalized to GAPDH amplicon and the steady state rate of transcription for each transcript length. Densitometric analysis of the radioactivity bound to the filters was performed using Imagequant software and represents the mean values obtained from duplicate filters.

MYB SLR RNA probes and analysis
Radiolabeled RNA transcripts were prepared using the MEGAshortscript kit (Ambion) and 500 ng of EcoRV-linearized pBluescriptIIKS-MYB SLR polyU, MYB ΔSLR polyU, MYB SLR ΔpolyU, MYB SLR Δ9 polyU or 500 ng of Hind III-linearized pGEM-TAR, bulge mutant TAR, MYB SLR polyU, MYB ΔSLR polyU or MYB SLR ΔpolyU and 1 μl [α-32 P] UTP (3,000 Ci/mmol; Perkin Elmer). Transcription products were precipitated with ethanol and loaded onto a 0.5X TBE/4% PAGE gel. Denaturing electrophoresis used gels containing 6M urea and samples were heated to 95°C in formamide prior to loading. Gels were dried and the radiolabeled RNA detected by Phosphorimager analysis. Mfold version 3.0 was used to determine RNA secondary structures [31]. For RNA EMSAs probes were gel purified from 8M urea polyacrylamide gels.
analysis. Depleted extracts were subsequently incubated with radiolabeled MYB SLR polyU RNA probe and subjected to UV-cross linking followed by SDS-PAGE. Two phosphorimaging exposures of the same SDS-PAGE gel are shown to highlight the reduction in the 50kDa signal as indicated by the asterisk. Phosphorimaging quantitation confirmed reduction of the 50kDa signal. Error bars represent mean ± SEM, ** P <0.01.

Quantitation and Statistics
Statistical analysis was calculated using Graphpad Prism version 5.0 (Graphpad Software Inc, USA). Assays were repeated in triplicate to calculate the mean ± SEM unless otherwise stated. Student-T test was used to compare groups.

MYB SLR RNA has a secondary structure
We modeled the SLR polyU RNA transcribed from intron 1 of MYB and showed that this sequence can block elongation ( Fig 1A) [20,21]. To investigate whether the MYB SLR polyU encoded an RNA transcript with secondary structure wt (MYB SLR polyU), polyU tract-deleted (MYB SLR ΔpolyU), SLR deleted (MYB ΔSLR polyU) or MYB SLR polyU in which 9 of the 19 polyU residues were deleted (MYB SLR Δ9 polyU) transcripts were generated and their electrophoretic mobility examined under denaturing and non-denaturing conditions. Despite its larger mass, MYB SLR polyU RNA displayed faster mobility under native conditions, consistent with predictions that the MYB SLR has a high degree of secondary structure. Removal of the SLR (MYB ΔSLR polyU) or the polyU region (MYB SLR ΔpolyU), but not partial deletion of the polyU (MYB SLR Δ9 polyU) created mixed populations of RNAs under native conditions (Fig 1B), suggesting that the polyU tract may influence secondary structure. Similarly

Cellular factors engage the MYB SLR RNA
We sought to identify cellular factors that engage the MYB SLR RNA. Given the CRC cell line LIM1215 expresses relatively high levels of MYB [20], we reasoned that this cell line might express factors that engage the SLR and regulate MYB transcriptional elongation. UV cross-linking analysis was performed with LIM1215 extracts and intronic MYB RNA probes containing the SLR and polyU (MYB SLR polyU), a deletion of the SLR (MYB ΔSLR polyU) or a deletion of the polyU (MYB SLR ΔpolyU). Two bands of~50 and~38 kDa were resolved when the MYB SLR polyU probe was used ( Fig 1C). A similarly migrating band of~38 kDa was observed in the presence of MYB SLR ΔpolyU probe. The~50 kDa band was absent when MYB SLR ΔpolyU or MYB ΔSLR polyU probes were employed ( Fig 1C).
Previous studies have implicated NFκB in the regulation of MYB elongation via intron 1 [23][24][25]. Accordingly we hypothesized that the~50 kDa binding activity observed in the UV cross-linking experiments represented the interaction of NFκBp50 with the SLR (Fig 1C). In order to directly show that the 50kDa UV cross-linked band was indeed NFκBp50, UV crosslinking experiments were performed with MYB SLR polyU RNA probe and LIM1215 protein extracts in which NFκBp50 was depleted. Western blot analysed confirmed depletion of NFκBp50 from the cell extracts ( Fig 1D) and formation of the 50kDa UV cross-linked band was reduced two fold in reactions that contained NFκBp50 depleted cell extracts ( Fig 1D). Collectively these data suggest that the 50kDa band observed in UV cross-linking experiments represented the binding of NFκBp50 to MYB SLR polyU RNA probe.

NFκBp50 binds the MYB SLR RNA
The above experiments suggested that NFκBp50 could bind to the MYB SLR RNA. To examine this possibility RNA EMSAs were performed with recombinant NFκBp50 and intronic MYB RNA probes MYB SLR polyU, MYB ΔSLR polyU or MYB SLR ΔpolyU. NFκBp50 formed a complex with MYB SLR polyU RNA but was unable to shift probe in which the SLR was deleted (MYB ΔSLR polyU). Removal of the polyU tract (MYB SLR ΔpolyU) substantially reduced NFκBp50 binding (Fig 2A).
In order to establish that NFκBp50 was binding directly to the SLR, EMSAs were performed with shorter RNA probes that encompassed only the SLR and polyU (MYB SLR polyU) or only the SLR (MYB SLR ΔpolyU) ( Fig 2B). NFκBp50 formed a complex with the MYB SLR polyU probe ( Fig 2B). However, removal of the polyU (MYB SLR ΔpolyU) reduced NFκBp50 binding ( Fig 2B). Finally, in order to confirm that the complex formed with the MYB SLR polyU probe was indeed NFκBp50 we performed EMSAs with MYB SLR polyU RNA probe and included NFκBp50 antibody ( Fig 2C). NFκBp50 formed a complex with the MYB SLR polyU probe that was super-shifted in the presence of the NFκBp50 antibody ( Fig 2C). Furthermore, when reactions were subjected to UV cross-linking a single band of 50 kDa was resolved ( Fig 2C).
We next assessed whether the Rel homology domain (RHD), responsible for the DNA binding activity of NFκBp50 [32], could engage the MYB SLR RNA. Indeed the NFκBp50 RHD (aa 42-365) formed a complex with MYB SLR polyU RNA and when increasing amounts of RHD were added the MYB SLR polyU RNA probe was moved into multiple complexes that were super-shifted by NFκBp50 antibody (Fig 2D and 2E).
Within the RHD we pinpointed two motifs (aa 139-141; LGI) and (aa 146-148; KKK) that were highly conserved amongst the various human and mouse NFκB family members ( Fig  3A). Small clusters of highly basic amino acids rich in Lys residues are features of proteins that bind to RNA, with the positive charge of these residues complementing the negative charged surface of RNA [34]. We therefore sought to determine if the KKK residues contributed to the MYB SLR polyU RNA binding activity of NFκBp50. NFκBp50 RHD mutants were generated in which the Lys residues were substituted singly or collectively with Ala or Asp ( Fig 3B). Single substitution of the Lys residues 146, 147 or 148 subtly reduced NFκBp50 RHD binding to MYB SLR polyU RNA ( Fig 3C). In contrast, binding of p50 RHD was substantially reduced by substituting all three Lys for Ala or Asp ( Fig 3C). The LGI residues were also each substituted   for Ala (Fig 3B). Substituting the Leu or Ile residues reduced NFκBp50 binding to MYB SLR polyU RNA probe while substituting the Gly had little effect (Fig 3C).
To determine if the KKK and LGI motifs represented a modules that only bound RNA or whether they could also contribute to the DNA binding activity of NFκBp50, EMSAs were performed with the NFκBp50 RHD mutants and an IgGκB DNA probe [32]. Single or collective substitution of the Lys residues for Ala or Asp reduced binding of NFκBp50 RHD to IgGκB DNA S3 Fig Substituting the Leu, Ile or Gly residues for Ala reduced NFκBp50 RHD binding and the effect of these mutations mirrored that observed for the DNA-binding residues Arg 56 and 58, Tyr 59, Glu 62 and His 66 substituted for Ala [35] and S3 Fig We also observed that substituting Arg 56 for Ala had no effect on RNA binding (Fig 3C). These results indicate that NFκBp50 shares conserved sequences (aa 139-141; LGI) and (aa 146-148; KKK) that contribute to the DNA and RNA binding activity of NFκBp50.

NFκBp50 and p65 induce transcriptional elongation via the MYB SLR
The above experiments suggested that NFκBp50 might engage the MYB SLR polyU to regulate MYB elongation. To examine this possibility, the potential of NFκBp50 to regulate elongation in gene reporter assays was assessed using a series of CAT reporter constructs containing the promoter, exon 1 and intron 1 with or without the MYB SLR or polyU [21] (Fig 4A).
NFκBp50 did not stimulate transcription from the MYB SLR polyU elongation CAT construct ( Fig 4A). In contrast transfected NFκBp65 stimulated transcription in a MYB SLR and polyU tract dependent manner ( Fig 4A). Furthermore, consistent with previous published results [21], the activity of MYB ΔSLR polyU CAT and MYB SLR ΔpolyU CAT was higher than MYB SLR polyU CAT, confirming that the MYB SLR and polyU sequences negatively impacted on MYB transcriptional activity (Fig 4A). Because of these observations we tested whether NFκBp65 could associate with the MYB SLR polyU in RNA EMSA experiments. However, we did not detect a robust interaction between NFκBp65 and MYB SLR polyU RNA S4 Fig Fur-thermore, we confirmed the nucleic acid binding activity of the recombinant NFκBp65 preparation by performing DNA EMSAs with an HIV-1 LTR radiolabeled probe previously shown to support binding of NFκBp65 [33]. Our DNA EMSA results also showed that recombinant NFκBp65 could bind DNA and promote binding of NFκBp50, under conditions where the concentration of NFκBp50 was limiting S4 Fig The ability of NFκBp65 to activate the MYB SLR polyU CAT but not directly bind MYB SLR polyU RNA in vitro led us to conclude that the induction of MYB SLR polyU elongation by NFκBp65 may reflect a mechanism in which NFκBp65 is recruited to the MYB SLR polyU through its direct interaction with NFκBp50 that is bound to the MYB SLR thus forming an NFκBp50-65 heterodimer on the MYB SLR polyU. Indeed it is well established that NFκBp65 binds DNA inefficiently and that the NFκBp50-65 heterodimer is the most abundant form of the NFκB dimers that stimulate transcription [32,36] We next sought to further explore the scenario of a NFκBp50 and p65 complex on the MYB SLR polyU by exploiting a functional NFκBp50-65 heterodimeric fusion protein [29]. The NFκBp50-65 chimeric protein contains the RHD of NFκBp50 and the C-terminal activation domain (AD) of NFκBp65 [29]. The NFκBp50-65 fusion induced a dramatic stimulation of MYB SLR polyU CAT transcription in a SLR and polyU tract-dependent manner (Fig 4B). In contrast the NFκBp50-65 fusion failed to induce MYB promoter activity from the MYB construct containing only the MYB proximal promoter and exon 1 sequence (MYB promoter p50 Regulates MYB Transcriptional Elongation CAT) (Fig 4B). We also examined NFκBp50-65 regulation through the MYB SLR using MYB SLR polyU CAT constructs containing point mutations predicted to disrupt the MYB SLR structure, i.e. MYB SLR 3L, 5C or 23C mutation polyU CAT (Fig 4C). In each case NFκB p50-65 mediated stimulation of MYB SLR polyU CAT was not impacted on by the MYB SLR mutations (Fig 4C). Collectively these data show that the NFκBp50-65 chimera mediated its effect on MYB elongation through the MYB SLR polyU sequence and that the polyU was a major determinant regulating this activity.
We next examined whether the conserved lysine sequence within the NFκBp50 RHD (aa 146-148 (Fig 3) influenced the effect of the NFκBp50-65 fusion on MYB elongation. Lysine mutations that disrupted both the RNA and DNA binding activity of NFκBp50 (K146-148A) ( Fig 3C) and S3 Fig reduced the stimulation of MYB elongation by the p50-65 chimera (Fig  4D). The K148A substitution that modestly reduced the RNA binding activity, but not the DNA binding activity, of NFκBp50 ( Fig 3C) and S3 Fig also reduced the NFκBp50-65 mediated stimulation of MYB SLR polyU elongation (Fig 4D). Importantly, we also observed that mutation of the DNA binding reside Arg 56 also reduced the NFκBp50-65 mediated stimulation of MYB SLR polyU elongation S3 Fig Together these data suggest that the nucleic acid binding activity of NFκBp50-65 that includes its RNA and DNA binding functions, mediates stimulation of MYB SLR polyU CAT elongation.
We next examined whether the elongation factor P-TEFb influenced MYB expression. Previous studies have established that the P-TEFb inhibitory drug DRB inhibits elongation that leads to the synthesis of longer MYB transcripts [22]. We therefore treated 293 cells with DRB for 6 h and assessed the expression of MYB by QPCR. MYB expression was significantly down regulated in drug treated cells (Fig 5A). Furthermore, DRB treatment of 293 cells inhibited the NFκBp50-65 mediated induction of MYB elongation from the MYB SLR polyU CAT reporter (Fig 5B). Collectively these data suggest that P-TEFb could influence NFκBp50-65 mediated induction of MYB elongation.

NFκBp50 and p65 induce MYB expression
Having established a relationship between NFκBp50, p65 and MYB SLR polyU transcriptional elongation, we next sought to determine whether these factors could influence endogenous MYB expression. To this end the NFκBp50-65 chimera was transfected into 293 cells and MYB expression assessed. Q-RT-PCR analysis demonstrated that endogenous MYB levels were increased upon transfection of cells with NFκBp50-65 ( Fig 5C). Conversely, reduced induction of MYB was observed using NFκBp50-65 containing Lys mutations that disrupted both the RNA and/or DNA binding activity of NFκBp50-65 (K146-148A) ( Fig 5C) and NFκBp50-65 (R56A) S3 Fig However, in this context the K148A alone substitution did not effect NFκBp50-65 mediated stimulation of MYB SLR polyU elongation (Fig 5C). In addition, NFκBp50-65 enhanced the synthesis of intronic pre-mRNA MYB transcript past the MYB SLR polyU (post SLR) indicating that NFκB p50-p65 could relieve the transcriptional attenuation of MYB ( Fig  5D). Consistent with this finding we observed by chromatin immunopreciptation (ChIP) analysis an increase in the enrichment of RNA polymerase II at the MYB SLR poly U in the presence of NFκBp50-65 (Fig 5E).
We next determined whether endogenous NFκBp50 and p65 could induce MYB expression in vivo. Previous studies have established that NFκB pathway can be activated by the pro-inflammatory cytokine TNFα [37]. Q-RT-PCR analysis demonstrated that endogenous MYB levels were increased upon exposure of small intestinal organoid cultures and 293 cells to TNFα (Fig 5F). Similarly, MYB SLR polyU CAT activity was stimulated when 293 cells were treated with TNFα and when the SLR was removed (MYB ΔSLR polyU CAT) the addition of TNFα p50 Regulates MYB Transcriptional Elongation did not induce reporter activity (Fig 5G). Furthermore, inhibition of the NFκB pathway with the BAY inhibitor BAY 11-7082 reduced endogenous MYB levels in a time dependent manner (Fig 5H). Taken together, the above data indicate that the NFκB factors p50 and p65 can induce MYB expression via the intron 1 MYB SLR polyU.

Tat induces MYB transcriptional elongation
Comparative sequence alignment revealed that residues (aa118-157) of the NFκBp50 RHD were similar to a highly conserved basic region of Tat (aa 48-57), which functions as an RNA binding domain (Arg rich motif; ARM) [26,38] (Fig 6A). The NFκBp50 RHD shares two Lys residues (KK:aa 146-147) with the ARM of Tat that are important for TAR RNA binding [39]. A third Lys residue (K:aa 148) was a conserved change. Upstream of the ARM, Tat and NFκBp50 also shared Leu, Gly and Ile residues (LGI: aa 139-141) (Fig 6A). This region of Tat is required for its transactivation function and is part of a sequence required for interaction with CyclinT1 [40,41].
Because of the similarities between NFκBp50 RHD and the Tat RNA binding domain we reasoned that HIV-Tat might also influence endogenous MYB expression via elongation. To examine this possibility, nuclear run-on assays were performed on 293 cells expressing Tat. Probes 3´to the SLR region showed increased signal in the presence of exogenous expressed Tat suggesting that Tat increased elongation into regions downstream of the SLR (Fig 6B and  6C). Furthermore, in the presence of exogenous expressed Tat the level of post-SLR RNA was higher (Fig 6D).
When MYB SLR polyU CAT or MYB SLR ΔpolyU CAT was co-transfected with Tat an induction of CAT activity was observed (Fig 7A). In contrast Tat failed to induce MYB ΔSLR polyU CAT reporter activity, instead reducing activity below basal levels ( Fig 7A). Together these CAT reporter data suggested that Tat influenced MYB expression and elongation via the MYB SLR.
We next examined whether the MYB SLR and TAR where inter-changeable. A CAT reporter construct in which the MYB SLR was replaced with the HIV TAR (MYB TAR CAT) ( Fig  7A) was co-transfected with Tat and in this context an induction of CAT activity was observed ( Fig 7A). Together these data suggest that in the context of the MYB intron 1, the SLR and HIV TAR are inter-changeable with respect to Tat mediated induction of MYB.
Tat recruits CyclinT1 and CDK9 to promote elongation paused at the TAR RNA [42]. To determine whether the effects of Tat on MYB SLR elongation were influenced by CDK9, 293 cells were co-transfected with MYB SLR polyU CAT and Tat as well as CDK9 or a CDK9 dominant negative (DN) form. When MYB SLR polyU CAT was co-transfected with CDK9 a slight induction of CAT activity was observed (Fig 7B). Expression of CDK9 did not significantly enhance Tat mediated induction of MYB SLR polyU CAT. In contrast, expression of the DNCDK9 negated Tat-mediated activation of MYB SLR polyU CAT (Fig 7B). Together these CAT reporter data suggested that CDK9 can influence Tat mediated induction of MYB SLR polyU CAT.

Tat binds the MYB SLR
The above experiments suggested that Tat, like NFκBp50, influenced MYB expression and elongation via direct binding to the MYB SLR. To examine this possibility EMSAs were TNFα for 10 h and CAT reporter activity assessed. (H) 293 cells were exposed to BAY inhibitor for the times indicated and total RNA was isolated and analyzed by Q-PCR to measure MYB expression levels. Error bars represent mean ± SEM, * P <0.05, ** P <0.01, *** P <0.001, **** P <0.0001. doi:10.1371/journal.pone.0122919.g005 Nuclear run on assays were performed as described [20,21]. Transcriptional activity was normalized to GAPDH signal and the steady state rate of transcription for each transcript length. Relative transcription between untransfected cells and cell transfected with Tat is shown. Densitometric analysis of the radioactivity bound to the filters was performed using Imagequant software and represents the mean values obtained from duplicate filters. (D) Nuclear RNA was isolated from 293 cells transfected with 100 ng pCMVTat (72) and subjected to RT-PCR to detect intron 1 RNA pre-and post-MYB SLR attenuator region.  conducted with recombinant Tat and an RNA probe comprising the SLR and polyU tract (MYB SLR polyU). The RNA binding activity of Tat was confirmed by EMSAs in which Tat bound HIV-1 TAR RNA (wt TAR) but not TAR RNA containing a mutation in the Tat binding region [28] (Bulge mutant) (Fig 7C). Under identical binding conditions Tat formed a complex with the MYB SLR polyU RNA that was super-shifted by an anti-Tat antibody (Fig 7D). To establish that Tat was binding directly to the SLR sequence, gel shifts were also performed with RNA probes that contained a deletion of the polyU (MYB SLR ΔpolyU) or a deletion of the SLR (MYB ΔSLR polyU) (Fig 7D). Tat bound to MYB SLR ΔpolyU RNA but not to the MYB ΔSLR polyU RNA (Fig 7D). These observations were consistent with the capacity of Tat to induce MYB SLR ΔpolyU CAT activity but not MYB ΔSLR polyU CAT activity (Fig 7A).

Discussion
Many eukaryotic genes are regulated in part by transcriptional pausing [8,9,11,12,43,44]. MYB is one such gene and an attenuation region responsible for pausing has been mapped within the first intron in both human and mouse genes [18,20,21,23]. Two views are held regarding the pausing mechanism that operates via the intronic attenuation sequences; one dependent upon the DNA template itself and the other employing an RNA stem-loop region (SLR) [19-21, 23-25, 45]. In this study we provide evidence that the nucleic acid binding activity of NFκBp50 that includes its RNA and DNA binding functions can regulate MYB via the SLR polyU region.
Several transcription factors regulate transcription by binding both DNA and RNA sequences [46]. Our analysis adds to this list by confirming that NFκBp50 is an RNA binding factor and for the first time suggest that MYB transcript is an RNA target for NFκBp50. Collectively our observations also imply that MYB is a direct target of the NFκB signaling pathway. Consistent with this view, recent studies have demonstrated that in EBV-infected nasopharyngeal carcinoma cell lines, NFκB signaling regulates MYB expression [47]. While the focus here has been on the interaction of NFκBp50 with the MYB SLR polyU, of relevance to this study is the possibility that NFκB p50 may also bind HIV TAR RNA. In CAT reporter assays we observed that the MYB TAR CAT reporter was responsive to NFκB p50-65 and in band shift assays recombinant NFκB p50 shifted a TAR RNA probe S5 Fig In support of these data are recent structure based analyses, which have shown that NFκBp50 is capable of binding TAR RNA [48]. Collectively these observations provide a strong case that NFκBp50 can act as an RNA binding factor.
Previous analysis used MFold to predict the potential secondary structure of the MYB SLR polyU [21]. We have now extended this analysis by examining the electrophoretic mobility of MYB SLR polyU RNA versus a panel point mutants. Our results indicate that mutation of the SLR region alters its mobility under native conditions, consistent with the prediction that this region forms a secondary structure. Furthermore, our data suggest that the polyU region may play a previously unrecognized role in influencing the formation of the MYB SLR secondary structure.
We provide several observations that demonstrate NFκBp50-mediated regulation of MYB elongation by an RNA dependent mechanism. First, our UV cross-linking experiments isolated a 50kDa MYB SLR polyU RNA binding activity in colon carcinoma cell line LIM1215 extracts. Furthermore, depletion of NFκBp50 from LIM1215 extracts saw a decrease in the level of 50kDa signal in UV cross-linking experiments. Second, in gel shift experiments purified recombinant NFκBp50 bound to a MYB SLR polyU RNA probe. Third, reporter assays demonstrated a role for NFκBp50 in relieving the transcriptional elongation block through MYB intron 1 in an SLR-and CDK9-dependent manner. This activity was dependent on the nucleic acid binding activity of NFκBp50 that includes both its RNA and DNA binding activity. Finally, over-expression of TNFα and BAY inhibitor studies indicated that NFκBp50-p65 could induce endogenous MYB expression and enhance the synthesis of intronic pre-mRNA MYB transcript past the MYB SLR polyU tract (post SLR). An RNA binding activity for NFκBp50 was also underscored by comparative sequence alignments in which residues (aa118-157) of the NFκBp50 RHD were similar to a highly conserved basic region of Tat (aa [48][49][50][51][52][53][54][55][56][57], which functions as an RNA binding domain [26,38]. In this context we demonstrated that Tat could induce the expression and transcriptional elongation of MYB and interact directly with MYB SLR polyU RNA. Previous studies have implicated NFκB family members in the regulation of transcriptional elongation through intron 1 of the murine MYB gene [23][24][25]. Notably, an NFκB enhancer element resides within intron 1 that is capable of binding NFκBp50, RelA, RelB and c-Rel in vitro [23][24][25]. In these studies, binding of NFκB complexes to this region was linked to the activation of endogenous MYB or MYB-CAT expression, suggesting that in case of the murine MYB gene transcriptional elongation through intron 1 is mediated by a DNA-dependent mechanism. Although the NFκB enhancer is potentially important to the regulation of the endogenous gene, it resides approximately 1.2 kbp downstream of the region investigated here and was excluded from the reporters examined in the current study. Nevertheless, this suggests that NFκBp50 may also contribute to the regulation of MYB via a region(s) upstream of the MYB SLR polyU. Consistent with this observation other reports have shown a role for a NFκB site found in the MYB promoter [49]. Collectively these observations suggest that NFκBp50 may regulate MYB through multiple regions/sites within the gene. In this case NFκBp50 bound at upstream sequences may stimulate transcription from the multiple start sites upstream of exon 1. Transcripts that are then stalled at the downstream MYB SLR polyU are released by the action of NFκBp50 at the MYB SLR polyU. In this way NFκBp50 may regulate MYB at at least two points within the gene using its nucleic acid binding activity (Fig 8).
This model of NFκBp50 activity on the MYB gene raises the interesting possibility that upstream bound NFκBp50 may also interact with the intron 1 based MYB SLR polyU (Fig 8). Indeed recent data indicate that upstream enhancer elements loop towards murine Myb intron 1 to bring the transcription apparatus to the vicinity of the pausing region and regulate Myb attenuation/elongation [19]. These observations are consistent with data from chromosome conformation capture (3C) assays that we have performed using human Colo201 cells. We found that regions upstream of the MYB proximal promoter lie in proximity to the intron 1 MYB SLR polyU S5 Fig This effect was lost when Colo201 cells were induced to differentiate with sodium butyrate, concomitant with the downregulation of MYB S5 Fig Importantly, the downregulation of Myb is an early event in differentiation that we have previously associated with transcriptional attenuation in intron 1 [20]. The interaction of NFκBp50 bound to upstream DNA regions with downstream RNA elements is supported by recent structure based analyses of HIV-1 LTR interactions in which pre-formed NFκBp50-DNA complexes where found to interact with downstream HIV TAR RNA [48]. We also note recent data in which Tat preassembled at the HIV promoter in a P-TEFb:7SK snRBP complex is able to engage and transition to downstream TAR RNA to regulate P-TEFb mediated stimulation of transcription elongation [51].
We have mapped amino acids 118-157 within the RHD of NFκBp50 as constituting a MYB SLR polyU binding domain. Within this region we have pinpointed two motifs (aa 139-141; LGI) and (aa 146-148; KKK) required for the association of NFκBp50 with MYB SLR polyU RNA and are highly conserved amongst the various human and mouse NFκB family members. Remarkably, the 118-157aa segment bears a striking similarity with the RNA binding motif of Tat, particularly across the LGI and KKK motifs. Small clusters of highly basic amino acids rich in Arg or Lys residues are features of proteins that bind to RNA, with the positive charge of these residues complementing the negative charged surface of RNA [34]. These sequence features of NFκBp50, coupled with the observations described above reinforce the notion that NFκBp50 can function as an RNA binding factor.
The finding that NFκBp50 is an RNA binding protein is consistent with previous studies that have analyzed the potential of NFκBp50 to bind in vitro selected RNA ligands (aptamers) Fig 8. Model for the NFκBp50 and p65 regulation of MYB elongation through the intron 1 MYB SLR polyU region and upstream sequences. NFκBp50 binding at upstream sequences stimulates transcription from the multiple start sites. Transcripts are then paused/attenuated at the downstream MYB SLR polyU. In the absence of NFκBp65, NFκBp50 occupies the MYB SLR polyU region and MYB transcription is paused. In contrast the formation of a NFκBp50-p65 heterodimer on the MYB SLR polyU contributes to the stimulation of MYB transcription elongation by the NFκBp65-mediated recruitment of P-TEFb and the subsequent Ser2 phosphorylation of Pol II CTD (elongating form) by CDK9. This model is consistent with i) previous findings that NFκBp65 mediates transcriptional elongation through the direct recruitment of P-TEFb [50] and ii) our recent observations that ERα recruits P-TEFb to a region near the MYB SLR polyU and that this interaction is concordant with the accumulation of Ser2 CTD phosphorylated pol II bound to this region and the relief of MYB transcriptional attenuation [22]. This model also raises the possibility that the regulation of MYB may also involve the interaction of upstream bound NFκBp50/65 with the downstream intron 1 MYB SLR polyU. Recent data indicate that enhancer elements loop towards murine Myb intron 1 to bring the transcription apparatus to the vicinity of the pausing region and regulate Myb attenuation/elongation [19] and we have shown that the upstream regions of MYB lie in proximity to the MYB SLR polyU S5 Fig Furthermore, recent structure based analyses of HIV-1 LTR interactions suggest that pre-formed NFκBp50-DNA complexes can interact with downstream HIV TAR RNA [48].
doi:10.1371/journal.pone.0122919.g008 [52][53][54]. Characterization of an optimal NFκBp50 binding RNA aptamer revealed a highly pre-structured RNA, folded as stem-loop with an asymmetric internal loop [55]. However, the structure of the MYB SLR polyU described here differs with little or no sequence similarity with the RNA aptamer. The crystal structure of the NFκBp50-aptamer complex reveals a conserved protein-nucleic acid interface in which the RNA aptamer mimics the κB DNA consensus element by contacting amino acids located in loop 1 and 2 that are involved in DNA binding [56]. Similarly, our study indicates the importance of DNA binding residues within loop 2 for NFκBp50-MYB SLR polyU complex formation. Our mutational analysis suggests that at least in the case of Arg 56, loop 1 residues do not make a significant contribution to the NFκBp50-MYB SLR polyU complex.
It is important to note that while our data are consistent with the idea that the MYB SLR polyU acts as a scaffold to recruit RNA binding factors such as NFκBp50, it is likely that these proteins function with additional co-factors in order to promote elongation through intron 1. P-TEFb (CDK9/ CyclinT1) is an essential co-factor for the Tat-TAR RNA complex with the CDK9 component playing a key role in mediating the effects of Tat on transcriptional elongation [57][58][59]. Given our observations that NFκBp50-driven MYB elongation is CDK9 dependent, it is possible that a similar mechanism may exist for the MYB SLR polyU involving an NFκBp50-SLR-P-TEFb complex. Our data also show a role for NFκBp65 (RelA) in regulating transcriptional elongation through intron 1 of MYB in a SLR and polyU tract dependent manner. RNA shift analysis indicated that NFκBp65 did not bind efficiently to MYB SLR polyU RNA. These observations suggest that NFκBp65 might be recruited to MYB SLR polyU via its association with NFκBp50 that is bound to the SLR. Indeed this scenario is reminiscent of a large body of published data that have shown that NFκBp65 binds DNA poorly but elicits its transcriptional activity by forming a heterodimer with NFκBp50 on promoter DNA [32,36]. While we were unable to detect NFκBp50-65 heterodimers in our RNA experiments, our gene reporter assays suggest that NFκBp50-65 fusion protein in which the NFκBp65 AD is tethered to the NFκBp50 RHD, is a potent activator of MYB elongation via the MYB SLR and polyU tract. Collectively these data support the idea that NFκBp65 can activate the MYB SLR polyU without directly binding the MYB SLR polyU. Interestingly, NFκBp65 has been shown to mediate transcriptional elongation through the direct recruitment of P-TEFb [50]. Taken together, these observations invoke a model (Fig 8) where the NFκBp50-p65-SLR complex contributes to the stimulation of MYB transcription elongation by the NFκBp65-mediated recruitment of P-TEFb and the subsequent Ser2 phosphorylation of PolII CTD (elongating form) by CDK9. This model is consistent with our recent observations that ERα recruits P-TEFb to a region near the MYB SLR polyU and that this interaction is concordant with the accumulation of Ser2 CTD phosphorylated pol II bound to this region and the relief of MYB transcriptional attenuation [22]. In addition, the conservation of key amino acid residues required for MYB SLR polyU binding among NFκB family members suggests that alternative combinations of NFκB complexes may form with the MYB SLR polyU. This hypothesis raises the intriguing possibility that the formation of multiple types of NFκB-MYB SLR polyU complexes may elicit specific effects on MYB elongation and expression. Future studies identifying other NFκB componentsand indeed other factors-associated with MYB SLR polyU and confirming the role of P-TEFb may shed light on the function of these complexes in cellular processes where control of MYB expression is critical.  [33] was incubated with 100 ng of purified NFκBp65 and the NFκBp65-DNA complexes resolved on a 6% non-denaturing polyacrylamide/bisacrylamide (29:1) gel buffered with 0.5 X TBE. NFκBp65-DNA complexes were super-shifted by the addition of anti-NFκBp65 antibody. Anti-NFκBp50 and anti-Myc IgG was used as a control. The black arrows indicate the position of the NFκBp65-DNA complex; the grey arrow indicates free HIV-1 LTR DNA probe. The white arrows show the position of the complexes in the presence of the antibody. Right panel. EMSA HIV-1 LTR DNA probe [33] was incubated with 100 ng of purified NFκBp65 with or without the addition of 10 ng purified NFκBp50 and the NFκB-DNA complexes resolved on a 6% non-denaturing polyacrylamide/bisacrylamide (29:1) gel buffered with 0.5 X TBE. NFκB-DNA complexes were super-shifted by the addition of anti-NFκBp65 or anti-NFκBp50 antibody. Anti-Myc IgG was used as a control. Long range interaction between the MYB promoter and intron 1 MYB SLR polyU region. 3C analysis of the MYB promoter and intron 1 regions in vivo was performed using Colo201 cells in the presence or absence of sodium butyrate (20 μM; 6 h). A schematic of the MYB gene is shown with the BamH1 sites that were exploited in the 3C assay. Primers were located in the 5´regions flanking the BamH1 sites to allow identification of long range interaction between the MYB promoter and intron 1 MYB SLR polyU region. Error bars represent mean ± SEM, Ã P <0.05. (TIF) S1 Table. Oligonucleotide sequences used in this study. Oligonucleotide sequences were purchased from Geneworks (Australia) and used as primers in quantitative RT-PCRs or as probes in electrophoretic mobility shift assays. (DOCX)