CREB Is a Target for Transcriptional Repression by Foxp3
Although Foxp3 could down-regulate Tax-dependent transactivation of the HTLV-I LTR (Figure 4B) and inhibit Tax expression from an infectious molecular clone (Figure 5), Foxp3 failed to modulate Tax function in the absence of the viral promoter (Figure 4D). These results led us to hypothesize that Foxp3 acts on HTLV-I gene expression by interacting with proteins important for driving HTLV-I LTR activity in vivo. Previous studies have demonstrated that the Tax-responsive elements within the HTLV-I LTR play a crucial role in driving Tax-mediated transactivation of the HTLV-I LTR [18]. The Tax-responsive elements have been shown to resemble CREB binding sites, bind CREB in vitro and in vivo, and facilitate HTLV-I LTR activation both in the presence and in the absence of Tax [35,36]. Ching and colleagues [18] demonstrated that addition of a dominant-negative CREB expression vector resulted in nearly complete inhibition of Tax-mediated activation of the HTLV-I LTR, while blocking NF-kappaB activation by addition of a dominant-negative IKKbeta expression vector had no effect on Tax transactivation of the HTLV-I LTR. Therefore, we hypothesized that Foxp3 may inhibit Tax transactivation of the HTLV-I LTR via disruption of the CREB signaling pathway. To test this possibility, HEK 293T cells were transfected with an HTLV-I LTR or synthetic CREB reporter vector along with a control expression vector (EGFP) or expression vectors encoding Foxp3 or DeltaFKH. As shown in Figure 6A, Foxp3 down-regulated basal activation of the HTLV-I LTR and transcription of a synthetic CREB reporter vector, suggesting that Foxp3 down-regulates HTLV-I LTR activation by targeting the CREB pathway. Deletion of the FKH domain of Foxp3 dampened the suppressive effect of Foxp3, but did not completely abrogate suppression, as is seen with NF-kappaB-responsive promoters in HEK 293T cells. Like NF-kappaB activation, CREB transcriptional activation was also suppressed by expression of Foxp3, and to a similar extent DeltaFKH, in healthy donor CD4+ T cells (Figure 6B). Similarly, Foxp3 and DeltaFKH also repressed basal HTLV-I LTR activation in primary human CD4+ T cells (Figure 6C). To our knowledge, this is the first evidence implicating CREB as a molecular target of Foxp3. As observed with NF-kappaB activation, DeltaFKH was a more potent inhibitor of CREB activation in CD4+ T cells than in HEK 293T cells, further indicating that a cell type-specific mechanism of action may govern the function of this Foxp3 mutant. 
To determine whether Foxp3 functioned by directly signaling through CREB, we utilized expression vectors encoding CREB-1 or c-Jun (a member of the activator protein 1 family of transcription factors) fused in-frame to the Gal4-BD (Gal4-BD-CREB-1 and Gal4-BD-c-Jun). As shown in Figure 6D, activation of a Gal4-responsive reporter vector by Gal4-BD-CREB-1 was down-regulated by Foxp3 compared to control vector (EGFP), indicating that Foxp3 functions by directly or indirectly interacting with CREB-1. However, Foxp3 failed to markedly affect transcriptional activation of Gal4-BD-c-Jun (c-Jun has also been demonstrated to bind to the HTLV-I LTR) and Gal4-BD-Tax (see Figure 5). Importantly, the mechanism of Foxp3-mediated inhibition of CREB-dependent transcription was not due to a block in CREB-1 protein expression, as determined by Western blot analysis (Figure 6E). Although these results demonstrate that Foxp3 functions as a co-repressor of CREB activation (in addition to NF-kappaB and NF-AT), we were unable to detect a direct physical interaction between CREB-1 and Foxp3 by coimmunoprecipitation or mammalian two-hybrid analysis (unpublished data). Therefore, our data suggest that Foxp3 may interfere with CREB signaling at an upstream event, such as phosphorylation of CREB or recruitment/function of coactivator proteins CREB-binding protein (CBP)/p300. 
