Given previous findings indicating that PADI4 plays a repressive role in gene transcription, we were initially surprised by our global analysis showing that PADI4 binding is primarily correlated with active gene transcription. However, the prediction that PADI4 also appears to play an important role in the activation of a broad range of target genes is supported by the following new lines of evidence. Firstly, we found that the PADI4 binding pattern near the TSSs is highly correlated with actively transcribed genes in MCF-7 cells and that most genes not expressed in the array are not bound by PADI4. Secondly, our computational ChIP-chip dataset analysis found that PADI4 is highly positively correlated with datasets from factors associated with gene activation and negatively correlated with datasets from factors associated with gene repression. Lastly, our computational analysis of transcription factor binding sites found that PADI4 is significantly enriched on promoters which contain DNA binding elements for a broad range of activating transcription factors. Interestingly, several of the elements identified in our computational analysis are bound by transcription factors such as STAT1, STAT3, STAT5, ETS1, PAX and E2F, which play important roles in mammary function and in breast cancer 
. Thus, the finding that PADI4 binding is enriched near these elements in MCF-7 cells supports the hypothesis that PADI4 plays a role in regulating target gene expression in a range of mammary gland signaling pathways.
Given that PADI4 binding was frequently coincident with the Elk-1 DNA binding element, and that PADI4 interacts with Elk-1, we next focused on investigating the potential relationship between PADI4 and Elk-1 on the c-Fos promoter, in order to better understand how PADI4 may activate gene transcription. Our findings that endogenous c-Fos levels were suppressed in PADI4 depleted and inhibitor-treated lines and elevated in PADI4 overexpressing cells, supported our genome-wide finding that PADI4 plays a role in transcriptional activation. Next, the observations that PADI4 interacts with Elk-1 at c-Fos and that PADI4 facilitates Elk-1-mediated activation of both a ubiquitous, and c-Fos-specific, reporter systems suggested that a functional relationship between Elk-1 and PADI4 confers activation of c-Fos. As to potential mechanisms behind this functional relationship, we found that PADI4 can citrullinate Elk-1 in vitro and that PADI4 enzymatic activity facilitates Elk-1 phosphorylation on c-Fos. Our ERK kinase assay then demonstrated that citrullination of Elk-1 by PADI4 enhances ERK-mediated Elk-1 phosphorylation. Finally, the observation that histone H4K5 acetylation was suppressed at c-Fos following PADI4 inhibition or depletion suggested that the observed PADI4 enzymatic activity on Elk-1 phosphorylation at c-Fos then facilitated histone acetylation and subsequent c-Fos activation.
The potential role for PADI4 in c-Fos
activation is partially clarified when the above findings are put in the context of what is currently known about the role of Elk-1 in this process. Recent reports have demonstrated that unphosphorylated Elk-1 and p300 form complexes and bind to gene promoters in the absence of stimuli. Following stimulation, Elk-1 is then phosphorylated by ERKs and this phosphorylation event strengthens the interaction between Elk-1 and p300, leading to enhanced histone acetyltransferase activity at target promoters and gene activation 
. In addition to observed Elk-1/p300 interactions, p300 has also been found to be targeted by PADI4 for citrullination 
. Given these observations, and our findings that PADI4 interacts with Elk-1 and that Elk-1 depletion reduces PADI4 levels at c-Fos
, it seems likely that Elk-1, p300, and PADI4 are constitutively bound to the c-Fos
promoter (). Our current working model then predicts that treatment of MCF-7 cells with EGF then activates PADI4, which, in turn, citrullinates Elk-1. Upon citrullination, Elk-1 is then phosphorylated by ERK2 leading to stronger associations between Elk-1 and p300, increased histone acetylation, and subsequent transcriptional activation 
. Given that citrullination neutralizes positively charged arginine residues, it seems possible that PADI4-mediated citrullination facilitates ERK-mediated phosphorylation of Elk-1 by allowing this kinase closer access to its target serine residues. As an alternate hypothesis to our working model, Lee et al. 
has found that PADI4 can also directly citrullinate p300 to enhance p300 activity. Thus, it is possible that a direct functional relationship exist between PADI4 and p300 on c-Fos
. However, we found that treatment of p300 with PADI4 did not enhance p300's histone acetyltransferase activity in vitro
(data not shown). Additionally, we do not believe that PADI4 is mediating c-Fos
activity via histone citrullination as we were not able to detect significant increases of citrullinated histones (H3 or H4) on the c-Fos
promoter after EGF stimulation (data not shown). Given these observations, and our findings presented here, the most likely mechanism by which PADI4 is regulating c-Fos
is by targeting Elk-1 for citrullination.
Working model depicting role of PADI4 in Elk-1–mediated activation of c-Fos.
In conclusion, our genomic analysis has identified a new role for PADI4 in mediating gene transactivation. We present evidence that a broad range of transcription factors may utilize PADI4 as a cofactor to mediate their transactivation functions, thus suggesting that the gene regulatory role for PADI4 is much broader than previously thought. We also provide new mechanistic insights into how PADI4 mediates gene transactivation by showing that PADI4 interacts with and citrullinates Elk-1 and that PADI4 enzymatic activity potentiates two key features of c-Fos activation, Elk-1 phosphorylation and histone H4 acetylation. Studies are currently underway to further dissect the mechanism by which PADI4 regulates gene expression and possibly mammary tumorigenesis.