Massively parallel cis-regulatory analysis in the mammalian central nervous system

Cis-regulatory elements (CREs, e.g., promoters and enhancers) are DNA regions that regulate gene expression, and variants within CREs can contribute to phenotypic diversity, including disease susceptibility (Wray 2007; Albert and Kruglyak 2015). In the past several years, vast amounts of genomic data have been generated that predict the locations of hundreds of thousands of CREs in cell lines and primary tissues (The ENCODE Project Consortium 2012; Shen et al. 2012; Romanoski et al. 2015). As an avenue for the experimental validation of these predictions, massively parallel reporter assays (MPRAs, e.g., CRE-seq) have been developed, in which barcoded plasmid reporters are introduced into cells. Next-generation sequencing of the resulting barcoded transcripts provides a quantitative measure of CRE activity (Kwasnieski et al. 2012; Melnikov et al. 2012; Patwardhan et al. 2012; Arnold et al. 2013; White et al. 2013; Levo and Segal 2014; Shlyueva et al. 2014). Thus far, MPRAs have been largely restricted to assaying short CRE fragments (<150 bp) synthesized as oligonucleotide libraries on microarrays (Patwardhan et al. 2009; Baker 2011; White et al. 2013) and delivered into select mammalian cells accessible by transfection or electroporation. However, CREs are often hundreds of base pairs in length, and CRE activity depends crucially on the assayed cell type and its particular complement of transcription factors (TFs) (Davidson 2001). Therefore, we sought to expand the biological relevance and applicability of MPRAs by increasing the length of assayed CREs and by widening the repertoire of assayable cell types. The retina and cerebral cortex are two parts of the central nervous system (CNS) with a shared forebrain origin, whose gene regulatory networks are topics of intense research interest (Swaroop et al. 2010; Wright et al. 2010; Bae et al. 2015; Nord et al. 2015). The genome-wide locations of putative CREs have been mapped in both tissues, using methods such as ChIP-seq and DNase-seq (Visel et al. 2009; Corbo et al. 2010; The ENCODE Project Consortium 2012; Wilken et al. 2015). Compared to the cortex, the retina is more experimentally amenable to cis-regulatory analysis, in part because its cellular composition is more completely understood (Livesey and Cepko 2001; London et al. 2013). Electroporation can be used to efficiently deliver plasmid DNA into rod photoreceptors, which constitute the majority (∼80%) of the cells in the retina (Jeon et al. 1998). We previously conducted CRE-seq by electroporating thousands of short CREs into the neonatal mouse retina ex vivo (Kwasnieski et al. 2012; White et al. 2013). Although hundreds of putative developmental forebrain enhancers have been assayed with one-at-a-time transgenic mouse reporter assays (Nord et al. 2013; Visel et al. 2013), never before has massively parallel cis-regulatory analysis been conducted in the mammalian CNS in vivo. Here, we sought to overcome current technological hurdles by developing a “capture-and-clone” approach for synthesizing CRE-seq libraries with a selectable range of fragment sizes for targeted cis-regulome analysis. As a built-in feature, our approach allows for truncation mutation analyses, which can identify regions within CREs that are critical for activity. We furthermore demonstrate the feasibility of conducting in vivo CRE-seq in the adult cerebral cortex by AAV-mediated delivery. Our approach provides a framework for the massively parallel functional analysis of CREs in a broad repertoire of organs and species in vivo.