A Comprehensive Atlas of Arabidopsis Regulatory DNA [DAP-seq]

We have developed a high throughput, next-generation DNA sequencing assay for rapid transcription factor binding site (TFBS) discovery in a genomic context. DNA affinity purification sequencing (DAP-seq), which uses affinity-purified transcription factors (TFs) to capture genomic DNA fragments, was applied to all 1,725 Arabidopsis thaliana TFs. High confidence TFBS motifs for 529 TFs and genome-wide enrichment maps for 349 factors were identified. In total,~ 2.7 million TFBS were identified which predict thousands of TF target genes enriched for known and novel functions.. Comparison of TF-binding using cytosine-methylated and -unmethylated genomic DNA revealed a 2-50 fold inhibition at methylated motifs for ~82% (264) of factors tested while 4.6% (15) showed stronger binding to methylated motifs. Finally, we describe how binding of Arabidopsis and maize Auxin Response Factors (ARFs) at phased motif repeats is highly enriched at ARF target gene promoters and how this architecture may allow for stabilization of dimers/multimers. Overall design: Identification of sequence motifs for 530 transcription factors (529 Arabidopsis thaliana, 1 Zea mays) and genome-wide binding sites for 350 transcription factors (349 Arabidopsis thaliana, 1 Zea mays) by direct sequencing of affinity purified genomic DNA fragments. Idenfication of binding sites for 343 transcription factors (Arabidopsis thaliana) by direct sequencing of affinity purified, PCR amplified DNA libraries. Comparison to ChIP-Seq for ABI5 (AT2G36270). The 530 transcription factors refer to the number of unique proteins assayed on non-amplified DNA libraries (i.e. "characteristics: DNA source" = "col" or "zm"). Additional 5 factors have only data on amplified libraries (i.e. "characteristics: DNA source = colamp"), which do not represent the natural DNA methylation state of the organism (therefore not counted; more details in the "description" field). A subset of 350 of the 530 have both motif and TFBS identification (i.e. "characteristics: subset = TFBS+motif"), and the rest 180 have only motif identification (i.e. "characteristics: subset = motif").

本研究开发了一种高通量下一代DNA测序检测方法，可在基因组背景下快速发现转录因子结合位点（transcription factor binding site, TFBS）。DNA亲和纯化测序（DNA affinity purification sequencing, DAP-seq）通过亲和纯化的转录因子（transcription factor, TF）捕获基因组DNA片段，该方法被应用于全部1725个拟南芥转录因子。最终鉴定得到529个转录因子的高可信度TFBS基序，以及349个转录因子的全基因组富集图谱。总计鉴定得到约270万个TFBS，这些位点可预测数千个转录因子靶基因，这些靶基因富集了已知及新发现的功能。

对比使用胞嘧啶甲基化与未甲基化基因组DNA的转录因子结合情况发现，约82%（264个）的受试转录因子在甲基化基序处的结合受到2至50倍的抑制，而4.6%（15个）的转录因子对甲基化基序的结合能力更强。最后，本研究阐明了拟南芥与玉米的生长素响应因子（Auxin Response Factors, ARFs）在阶段性基序重复序列处的结合，在ARF靶基因启动子处高度富集，并揭示了这种结构可稳定二聚体/多聚体的形成机制。

整体实验设计：通过直接对亲和纯化的基因组DNA片段进行测序，鉴定530个转录因子（529个拟南芥、1个玉米）的序列基序，以及350个转录因子（349个拟南芥、1个玉米）的全基因组结合位点。通过直接对亲和纯化并经PCR扩增的DNA文库进行测序，鉴定343个拟南芥转录因子的结合位点。并与ABI5（AT2G36270）的染色质免疫共沉淀测序（Chromatin Immunoprecipitation Sequencing, ChIP-Seq）结果进行对比。

该530个转录因子指的是在未扩增DNA文库中检测的独特蛋白数量（即"characteristics: DNA source" = "col"或"zm"）。另有5个转录因子仅在扩增文库中获得数据（即"characteristics: DNA source = colamp"），此类数据无法反映生物体的天然DNA甲基化状态，因此未被计入统计，详细说明参见"description"字段。在这530个转录因子中，有350个同时完成了基序与TFBS鉴定（即"characteristics: subset = TFBS+motif"），剩余180个仅完成了基序鉴定（即"characteristics: subset = motif"）。