Transcriptional Activation of Mina by Sp1/3 Factors

Mina is an epigenetic gene regulatory protein known to function in multiple physiological and pathological contexts, including pulmonary inflammation, cell proliferation, cancer and immunity. We showed previously that the level of Mina gene expression is subject to natural genetic variation linked to 21 SNPs occurring in the Mina 5′ region [1]. In order to explore the mechanisms regulating Mina gene expression, we set out to molecularly characterize the Mina promoter in the region encompassing these SNPs. We used three kinds of assays – reporter, gel shift and chromatin immunoprecipitation – to analyze a 2 kb genomic fragment spanning the upstream and intron 1 regions flanking exon 1. Here we discovered a pair of Mina promoters (P1 and P2) and a P1-specific enhancer element (E1). Pharmacologic inhibition and siRNA knockdown experiments suggested that Sp1/3 transcription factors trigger Mina expression through additive activity targeted to a cluster of four Sp1/3 binding sites forming the P1 promoter. These results set the stage for comprehensive analysis of Mina gene regulation from the context of tissue specificity, the impact of inherited genetic variation and the nature of upstream signaling pathways.

Given clear evidence of Mina's involvement in immunity, cell proliferation and cancer, it is important to understand how Mina expression is regulated. We know from analysis of protein turnover and pre-mRNA transcription rate that Mina protein abundance is controlled largely at the transcriptional level [1]. However, the mechanisms governing Mina transcription remain poorly understood. To begin addressing this gap, we report here the molecular characterization of the Mina promoter region and its trans-acting factors in murine T cells. Using a dual luciferase reporter assay to interrogate nested deletions of a region spanning the Mina transcriptional start site (TSS), we defined a 144 bp minimal Mina promoter encompassing four potential Sp1/3 binding sites. Gel shift assays validated all sites as functional for Sp1 and Sp3 binding.
Furthermore, mutagenesis analysis demonstrated that full reporter activity required WT sequence at all 4 Sp1/3 binding sites. Pharmacological inhibition and siRNA knockdown of Sp1/3 binding activity and level, respectively, substantially diminished Mina mRNA expression. Finally, chromatin immunoprecipitation (ChIP) assays in primary T helper cells revealed the Mina promoter region to be enriched in bound Sp1 and Sp3 as well as lysine-4 trimethylated histone H3 (H3K4me3), a marker of transcriptionally active chromatin. Together, these results indicate a physiological requirement of Sp1 and Sp3 for Mina transcription and provide a stimulus for analysis of potential distal regulatory elements and the upstream pathways responsible for the tight regulation of Mina expression in its diverse physiological contexts.

Ethics Statement
Mice used in this study were maintained in specific pathogenfree conditions in accordance with the guidelines of the Mice BALB/c and C57BL/6 mice were purchased from Jackson Lab.

Luciferase Assay
Lipofectamine LTX with Plus reagent (Invitrogen) was used to cotransfect EL4 cells (2610 5 ) with PGL3 reporter constructs expressing firefly luciferase (FL) (750 ng) and pRL-TK expressing control renilla luciferase (RL) (40 ng). Following 48 h culture in 24 well plates, cells were harvested and assayed for FL and RL activity using the Dual-Luciferase Reporter Assay System (Promega).

Electrophoretic Mobility Shift Assay
EL4 cell nuclear extracts were prepared with the NER-PER extraction kit (Pierce, 78833). Protein concentrations were determined using the Bradford Assay (Thermo Scientific, 1856209) and bovine serum albumin as a standard (Sigma A9418). Probes were generated by annealing 59 biotin-labeled oligonucleotides at 95uC in annealing buffer (100 mM Tris pH 7.5, 10 mM EDTA, 2 M NaCl, 50 mM MgCl2) followed by slow cooling to 22uC (,3 hrs). Probe sequences are given in Table  S1. Nuclear extracts (10 mg), Poly (dA:dT) (1 mg), and biotinylated probes (200 fmol) were incubated together at 22uC for 30 min in 20 ml binding buffer (10 mM Tris pH7.5, 60 mM KCl, 2 mM MgCl2, 0.15 mM dithiothreitol). For competition, 100-fold molar excess of unlabeled probe (20 pmol) was added to the reaction before adding biotinylated probe. For supershift, antibody (2 mg) was included in the reaction for 15 min before biotinylated probe was added. Binding reactions were resolved on 0.5X TBE 4% polyacrylamide gels at 150 V for 2-3 hr, transferred to nitrocellulose for 30 min at 4uC, and then UV crosslinked for 45-60

Chromatin Immunoprecipitation (ChIP) Assay
EL4 cells were fixed in 1% formaldehyde for 15 min at RT. 10610 6 nuclei (isolated as described for FAIRE in [15]) were resuspended in 350 ml Buffer 3 (21) and sonicated using a BioRuptor (Diagenode) until the average chromatin fragment size was ,500 bp. Sonicated samples were centrifuged at 16,000 g for 10 min at RT and supernatants diluted 1 to 8 in dilution buffer (1% Triton X-100, 2 mM EDTA, 150 mM NaCl, 20 mM TrisHCl pH8, 1X protease inhibitor). Immunoprecipitation was performed by combining 1 ml of diluted supernatant (,4610 6 nuclei) with 4 mg of antibody or corresponding IgG control and overnight rotation at 4uC. Then 30 ml of Protein G magnetic beads was added and rotation continued for another 3 hours at 4uC. Beads were separated by incubation on a magnet rack and washed sequentially with 1 ml of the following buffers: Par Wash 1# (1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM TrisHCl pH8, 150 mM NaCl, 1X protease inhibitor), Par Wash 2# (1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM TrisHCl pH8, 500 mM NaCl, 1X protease inhibitor), Par Wash 3# (1 mM EDTA, 10 mM TrisHCl pH8, 250 mM LiCl, 1% NP-40, 1% DOC, 1X protease inhibitor), and TE buffer. Each wash was performed by 5 min rotation at RT. Elution of DNA/protein complexes was performed by adding 100 ml freshly made elution buffer (0.1 M NaHCO3/1% SDS), vortexing, and 10 min rotation at RT. A second elution was combined with the first for a total volume of 200 ml. Crosslinks were reversed by adding 8 ml of 5 M NaCl to the combined eluates followed by overnight incubation at 65uC. DNA was purified by adding 20 ml of protein digestion buffer (0.4 M Tris-HCl, pH 6.8, 0.1 M EDTA, 0.8 mg/ ml Proteinase K), mixing, incubation for 1 h at 50uC and phenol/ chloroform extraction. The purified DNA pellet was dried and resuspended in 20 ml of water. Quantitative real time PCR was performed to measure the relative enrichment of Mina promoter and Mina intron 2 DNA sequences. Primer sequences are given in Table S1.
For ChIP performed with CD4 T cells, lymphocytes were isolated from spleen and lymph nodes of C57BL/6 or BALB/c mice, sorted for CD4+/CD62L+/CD442/CD252 T cells, and cultured for 6-8 h in 6-well plates coated with 0.1 mg/ml anti-TCRb and 10 mg/ml anti-CD28 (Biolegend). Cells were harvested and fixed in 1% formaldehyde at RT for 10 min and then processed as described above.

Two Promoters and an Enhancer in the Mina TSSproximal Region
To map the Mina promoter, a ,2 kb-fragment spanning the transcriptional start site (TSS) from position 21588 to +354 was cloned into PGL3 basic vector and tested for reporter activity by dual luciferase assay in the thymoma cell line EL4 (Fig. 1, gray filled arrow). Fragment (21588/+354) contained strong reporter activity, respectively ,6-and ,60-fold higher than that driven by the SV40 promoter and by vector alone (Fig. 1). To characterize this activity, we interrogated a panel of 59 nested deletions of fragment (21588/+354) by reporter assay. Whereas deletions extending as far as 264 had no impact, removal of an additional 83 base pairs to position +19 caused a dramatic drop but not complete abolition of activity (Fig. 1). These results suggested: (1) that region (264/+19) contains an element that mediates strong reporter activity suggestive of an enhancer; and (2) that fragment (+19/+354) contains a Mina promoter (termed P2; Fig. 1, rightmost dark gray rectangle) comparable in activity to the SV40 promoter. To further localize P2 within fragment (+19/ +354), two additional 59 deletions were generated and tested. One, extending as far as position +150, maintained P2 activity, whereas a second deletion extending to +280 completely abolished it  (Fig. 1). These results narrow the location of P2 to region (+150/ +280) (Fig. 1, rightmost dark gray rectangle).
In order to reconcile the enhancer-like magnitude of the activity mapping to region (264/+19) and its promoter-like position spanning the Mina TSS, we hypothesized that it might encompass a second Mina promoter that is selectively targeted by a distal enhancer contained within fragment (264/+354). To test this hypothesis, we generated the TSS-spanning fragment (264/+80) and found it to exhibit low-level reporter activity similar to the SV40 promoter and to P2 (Fig. 1). Thus, the high-level reporter activity located within fragment (264/+354) and mapping to region (264/+19) is likely due to a second promoter (termed P1; Fig. 1, leftmost dark gray rectangle) targeted by a P1-specific distal enhancer (termed E1) mapping to region (+80/+354). The lowrather than high-level activity of fragment (+19/+354) -lacking P1 and containing both P2 and E1 -provides further evidence for the P1-specificity of E1.

Four Sp1/3 Binding Sites Spanning the Mina P1 Promoter
Given its TSS-spanning location and its specific regulation by E1, we decided to focus on characterizing P1. Using Transfac software (Biobase), we identified a cluster of four predicted Sp1/3 binding sites spanning region (264/+19) ( Fig. 2A). To begin exploring whether any of these sites were functional, we performed gel shift assays with DNA probes (p1-p4) spanning each site and nuclear extract from the Mina-expressing EL4 murine thymoma cell line ( Fig. 2A). Each of the four probes formed a similar pair of nucleoprotein complexes (Fig. 2B, complex 1 and 2 as indicated by arrows). Formation of the paired complexes was inhibited in the presence of a 100-fold molar excess of unlabeled cognate probe, as well as two known Sp1-binding sequences from the OX40 promoter [16] (Fig. 2B, Sp1A and Sp1B). These results show that each of the four predicted Sp1/3 sites in Mina P1 can form complexes with EL4 nuclear factors.

Binding of Sp1/3 to the Mina P1 Promoter in EL4 Cells
Next, we investigated whether Sp1 and/or Sp3 occurred within nucleoprotein complexes 1 and 2. Using gel shift assays with Mina P1 probes and EL4 nuclear extract we found that antibodies against YY1, Runx3 and Mina failed to perturb either complex (Fig. 3A). By contrast, an Sp1-specific antibody caused a supershift of the upper band of the complex 1 doublet, while an Sp3specific antibody abolished the lower band of the complex 1 doublet and the upper of band of the complex 2 doublet. Combining antibodies against Sp1 and Sp3 led to a summation of their individual effects on complexes 1 and 2 (Fig. 3A). Together, these data indicate that probes p1-p4 each form 3 distinct nucleoprotein complexes, one containing Sp1 and two containing Sp3.
Next, to examine whether Sp1 and/or Sp3 bind the Mina promoter in EL4 cells, we performed chromatin immunoprecipitation experiments using control IgG and antibodies against Sp1 and Sp3. As shown in Figure 3B and C, a Mina intron 2 sequence ,6 kb downstream of the Mina P1 promoter was similarly enriched in chromatin immunoprecipitated with Sp1-and Sp3specific antibodies as with control IgG. By contrast, a Mina P1proximal sequence was enriched to a significantly greater extent in chromatin immunoprecipitated with Sp1 or Sp3 antibodies as compared to control IgG. These data indicate that both Sp1 and Sp3 bind the Mina P1 promoter in EL4 cells.

Regulation of Mina Expression by Sp1/3
To extend our cell-free and cellular evidence that Sp1 and Sp3 bind to the Sp1/3 sites in the Mina P1 promoter, we asked next whether Sp1/3 factors regulate Mina P1 promoter activity. First, to map the key nucleotides required for Sp1/3 binding to each of the four Sp1/3 sites (Sp1.1-1.4), we performed gel shift assays on sets of six probe-spanning mutants. For Sp1.1, we found that probe p1 mutants M3, M4 and M5 exhibited impaired formation of complexes 1 and 2 (Fig. 4A). Similarly, we identified mutants of p1, p2 and p3 that disrupted Sp1/3-binding to Sp1.2, Sp1.3 and Sp1.4, respectively (Fig. 4B-D). Together, these data provide further evidence that each Sp1/3 site is functional for Sp1/3 binding and set the stage for functional analysis.
To explore whether Sp1/3 regulates Mina P1 promoter activity via any of its four Sp1/3 sites, we performed site-directed mutagenesis to introduce the relevant mutations (as determined by gel shift assay above) in each or combinations of the four Sp1/3 sites (Fig. 4E). Individual Sp1/3 site mutations did not significantly decrease reporter activity. By contrast, combinations of two or three Sp1/3 site mutations led to impaired reporter activity, while simultaneous mutation of all four Sp1/3 sites completely abolished P1 promoter activity (Fig. 4E). Together, these data show that binding of Sp1/3 to each of the four Sp1/3 sites contributes additively to the activation of the Mina P1 promoter.
To explore whether Sp1/3 contributes to endogenous Mina transcription in EL4 cells, we treated cells for 24 h with Mithramycin A or vehicle alone. Mithramycin A is known to specifically inhibit Sp1 binding to its consensus GC-rich binding sites and impair its ability to promote transcription [17]}. In cells treated with 1 mM Mithramycin Mina transcript level decreased more than 50% in comparison to control-treated cells (Fig. 4F). Mithramycin did not inhibit Hprt expression (data not shown). To further extend these results, we performed siRNA knockdown of Sp1/3 in EL4 cells. Treatment of cells with non-specific siRNA had no detectable effects on Sp1, Sp3 or Mina expression (Fig. 5). An Sp1-targeted siRNA abolished Sp1 but not Sp3 expression whereas an Sp3-targeted siRNA was found to impair expression of both Sp3 and Sp1. Strikingly, both the Sp1-and Sp3-targeted siRNAs abolished Mina expression. Together, these data suggest that in T cells Mina gene transcription from its P1 promoter requires Sp1.

Binding of Sp1/3 to the H3K4me3-enriched Mina P1 Promoter in Primary T Cells
We chose to explore Sp1/3 binding to the endogenous Mina promoter in primary naïve (CD4 + CD25 2 CD44 low CD62L high ) CD4 T lymphocytes as they are known to express Mina at high level [1]. We assessed the enrichment of Mina P1 promoterproximal and distal intron 2 sequences in chromatin immunoprecipitated with control IgG and antibodies against Sp1 and Sp3. As was the case with EL4 cells (Fig. 3B and C) we found that in primary naïve CD4 T cells both Sp1 and Sp3 bound to the Mina P1 promoter but not to distal intron2 ( Fig. 6A and B).
Histone modifications are implicated in regulating gene expression by modulating chromatin structure, DNA accessibility and recruitment of transcription regulatory machinery [18,19]. Tri-methylation of histone 3 at lysine 4 (H3K4me3) marks the promoters of actively transcribed genes, whereas tri-methylation of histone 3 at lysine 27 (H3K27me3) is associated with gene repression and silenced chromatin [20]. To determine whether the epigenetic landscape of the Mina locus correlated with its transcriptional status and Sp1/3 binding to the P1 promoter, chromatin fragments from naive CD4 T cells of C57BL/6 mice were immunoprecipitated with H3K4me3 or H3K27me3 antibody and subjected to high-throughput sequencing (ChIPseq). In region (2111/+576), spanning the location of both the P1 and P2 Mina promoters, we observed a striking peak of H3K4me3 enrichment (Fig. 6C), consistent with the chromatin structure in this region being open and permissive for Sp1/3 binding. Interestingly, although H3K27me3 and H3K4me3 are usually enriched at inactive and active chromatin regions, respectively, in Mina intron 1 both marks were co-enriched (Fig. 6C). Such so-called bivalent domains have been suggested to poise genes for either activation or repression during lineage commitment in hematopoietic stem cells [21].

Discussion
In the current study we sought to characterize the Mina promoter region by exploring the transcriptional regulatory elements occurring within a 2 kb interval spanning the Mina TSS. We identified two promoters, termed P1 and P2, respectively mapping to region (264/+19) spanning the TSS and to region (+150/+280) spanning exon 1 and intron 1. Three main murine Mina mRNA isoforms are documented in AceView [22]. The first of these initiates 14 bp downstream of P2. The 2 nd and 3 rd initiate, respectively, 11 bp downstream of and within P1. The correlation between the locations of P1 and P2 mapped in our study and the TSSs documented in AceView support our finding that Mina contains 2 promoters. Interestingly, we detected an enhancer-like element (termed E1) occurring within region (+80/+150) that could strongly promote reporter activity from P1 but not P2. P1 was found to contain four functional Sp1/3 sites that acted synergistically to promote reporter activity and bound both Sp1 and Sp3 in EL4 cells and primary CD4 T cells. Pharmacological inhibition of Sp1/3 binding in EL4 cells diminished the level of Mina transcription. Finally, the epigenetic landscape of H3K4me3 and H3K27me3 modifications (marking transcriptionally active and silent chromatin regions, respectively) across the Mina locus in primary CD4 T cells revealed a striking peak of H3K4me3 at the promoter region and an unexpected bivalent domain (simultaneously enriched in both H3 modifications) spanning most of intron 1. Taken together, our study provides strong support for a central role of Sp1/3 factors in regulating Mina transcription through binding to a cluster of 4 Sp1/3 sites in the Mina P1 promoter.
The precise location of the E1 enhancer (up-or down-stream of P2), the mechanism underlying its specificity for P1 versus P2, the nature of the transcription factor(s) that bind to and mediate P2 promoter activity and the relative roles of Mina P1 and P2 are currently under investigation. Interestingly, we detected a potential myeloid zinc finger 1 (Mzf1) binding site in P2 (data not shown). Mzf1 is preferentially expressed in myeloid cells and controls hematopoiesis [23]. It is tempting to postulate that through P2 Mzf1 may control Mina transcription in dendritic cells, another cell type where Mina is highly expressed and may function.
Sp1 and Sp3 are well-known transcription factors that modulate transcription of TATA-less genes by interacting directly with and mediating recruitment to basal transcription machinery [24][25][26][27]. They are deregulated in various types of human cancer. Sp1 mRNA and DNA-binding activities are shown to increase in epithelial tumors, suggesting that increased Sp1 activity contributes to skin tumor progression [28]. Further, Sp1/3 is constitutively overexpressed in pancreatic and gastric cancers and many signal transduction pathways terminating on Sp1-regulated genes are linked to cancer progression [29][30][31]. Given that Mina is aberrantly expressed in many cancers, it is tempting to speculate that its transcriptional regulation by Sp1/3 may contribute to its roles in cell proliferation and oncogenesis.
Sp1/3 factors also play a prominent role in the immune system, being known to regulate transcription of cytokines and regulators of T cell homeostasis, activation and differentiation, including GM-CSF [32], IL-10 [33,34] and OX40 [16]. Interestingly, Sp1/ 3 contributes to the differential expression of Fas ligand (a critical component of peripheral T cell homeostasis and cytotoxic effector mechanisms), Eta-1 (a cytokine that regulates IL-10 expression and Th1 polarization), and FUT7 (the rate-limiting enzyme for synthesis of a lymphocyte homing ligand) in Th1 and Th2 cells either through differential binding to a promoter polymorphism [35] or through differential promoter recruitment of transcription factors [36,37]. Further, Sp1/3 regulates the promoter activity of the IL4 receptor alpha chain [38], through which IL4 exerts its critical biological effects, including promotion of Th2 differentiation. We postulate that orchestration of effector T cell differentiation by master transcriptional regulators involves Sp1/ 3-dependent transcription of Mina.
The Mina P1 promoter appears to be able to buffer genetic variation insofar as mutations abolishing Sp1/3 binding to as many as two binding sites still permitted a level of activity not significantly different from the WT promoter with 4 functional Sp1/3 binding sites. This might also explain why 1 mM Mithramycin A was required to impair endogenous Mina transcription, whereas 10-fold less was sufficient to inhibit transcription of an Sp1/3 target gene (TINF2) that contained only two Sp1/3 sites [39].
This work sets the stage for a comprehensive analysis of Mina regulatory elements (enhancers and silencers) and for the discovery of regulatory SNPs that modulate Mina expression level across strains with high and low Th2-bias.

Supporting Information
Table S1 Electromobility shift and ChIP assay primers and probes. (DOCX)