Uterine Chromatin Immunoprecipitation-Sequencing Profile of Estrogen Receptor Alpha DNA Binding Mutant Reveals Novel Interactions Between Estrogen Receptor Alpha and Progesterone Receptor Signaling.

Abstract

Mutation of the first zinc finger of the mouse estrogen receptor a (ERa) disrupts its interaction with estrogen responsive element (ERE) sequences in target gene DNA and leads to altered uterine transcriptional responses. ERa binding sites were profiled using chromatin immunoprecipitation followed by comprehensive sequencing (ChIP-seq) with uteri from the zinc finger mutant ERa “KIKO” mice. Mice were ovariectomized and injected with vehicle (V) or estradiol (E2) for one hour. As was previously seen with WT ERa ChIP-seq, numerous binding sites were found with V treatment (3510 peaks) and E2 increased ERa binding (13773 peaks). Interestingly, only 30% of KIKO and WT ERa binding peaks overlapped. KIKO ERa ChiP-seq DNA sequences were divided into those with ERE motifs (6781 sequences) or without ERE motifs (6992 sequences) and evaluated for enrichment of other transcription factor motifs. Surprisingly, the most significant computed motifs were similar to androgen receptor (AR) or progesterone receptor (PR) binding motifs with 5685 of the KIKO ERa ChIP-seq sites showing one of the AR/PR like motifs. Possible mechanisms accounting for this finding include: 1) Interaction between ERa and PR signaling mediated by PRE/ARE sites. 2) Loss of ERE binding but acquisition of ARE/PRE affinity by the KIKO ERa mutant. The first mechanism was considered by scanning WT and KIKO ERa ChIP-seq peak sequences to predict PREs. 1364 of 17241 WT sites (7.9%) and 2922 of 13773 KIKO sites (21.2%) contained PREs, confirming the presence of PRE motifs in the WT ERa ChIP-seq binding sequences. However, only 449 of the predicted PREs were common between WT and KIKO, indicating that although some interactions are consistent with the first mechanism, not all KIKO ERa interaction with ARE/PRE are described by the scheme. To address the second mechanism, overlap between WT ERa, KIKO ERa and WT PR ChIP-seq peaks were evaluated. Interestingly, 20% of KIKO 1h E2 ERa peaks overlapped with PR 1h progesterone (P) peaks. Transcripts associated with the PRE-motif containing ERa binding peaks were examined in our microarray datasets. Many of these uterine transcripts exhibited ERa mediated E2 regulation in WT and KIKO, including some responses unique to KIKO. ER and/or PR regulated transcripts were selected for further study. Cdkn1a and Errfi1 transcripts are increased by E2 in both WT and KIKO. E2 dependent WT or KIKO ERa binding and P dependent PR recruitment was seen by ChIP-seq at common sites. Fkbp5 and Ihh transcripts are both increased by P and, interestingly, both have increased basal levels and enhanced E2 mediated increases in KIKO compared to WT. PR or ERa ChIP-seq indicates moderate E2 dependent WT ERa binding near both of these genes, but greater E2 dependent KIKO ERa binding and P dependent PR binding at the same sites. Hdc transcript is increased by P, but not by E2, in WT; however, either P or E2 increases Hdc in KIKO. WT ERa ChIP-seq peaks are not seen in the Hdc locus, but KIKO ERa and PR exhibit hormone-dependent recruitment to a common site. In conclusion, we show that the KIKO ERa exhibits novel responses, which suggest this receptor interacts with ARE/PRE like motifs. Together, our observations suggest that WT ERa and PR share sites of interaction in uterine chromatin, which are revealed by use of the ERa DNA binding mutant, consistent with a combination of both of the mechanisms suggested above.