Long-range massively parallel reporter assay reveals rules of distal enhancer-promoter interactions

Long-range activation is an essential property of enhancers, yet the features determining long-range enhancer activities have not been systematically investigated due to a lack of high-throughput methods to measure long-range enhancer activities efficiently. To address this gap, we present a long-range massively parallel reporter assay (long-range MPRA), a genome-integrated assay allowing the measurement of hundreds of enhancers at multiple distances away from a promoter in parallel. The long-range MPRA assay features two independent landing pads, which allow for modular control over the genome-integrated promoter and enhancer libraries. We showcased the capability of long-range MPRA by testing over 300 K562 enhancers, as well as a set of enhancer combinations, at distances up to 100kb. We found that enhancers' long-range activities are primarily determined by their intrinsic strength, with strong enhancers retaining more activity over long distances, while weak enhancers rapidly lose activity. Additionally, we found that GATA1-bound enhancers are more resistant to distance-dependent loss of activity, suggesting that TF binding also modulates long-range function. Finally, testing long-range enhancer activities with three different promoters (HBE, HBG and GAPDH) revealed that long-range E-P interactions rely on not only enhancer properties but also promoter responsiveness. Overall design: Genome-integrated, long-range massively parallel reporter assays of enhancers in K562 cells. To explore features of long-range enhancers, we designed two enhancer libraries: (1) GATA1/non-GATA1 enhancer library which contains 383 enhancers (249bp) selected from the K562 genome based on their GATA1 binding status; (2) LCR HS combination library which contains differet combinations of four DNase hypersensitive sites (HS1-4) in the beta-globin locus. For both libraries, we tagged each library member with a 10 bp unique enhancer barcode (eBC) and a 16bp random barcode (rBC). We also made a series of long-range landing pad cells, consisting of a promoter landing pad at a fixed location (chr18:58014732) and an enhancer landing pad at 0,3,10,20,50 or 100kb upstream. The promoter landing pad contains an mScarlet fluorescence reporter to report enhancer activity, and the enhancer landing pad contains counterselection markers for further selection of enhancer integrations. Each enhancer landing pad is tagged with a 12 bp distance barcode (dBC) to specificy the E-P distance. To measure enhancer activity at selected distances, we first mixed long-range landing pad cells with selected E-P distances in equal ratios and then integrated the enhancer library using Flpase-mediated site-directed recombination. We sorted cells into four bins based on their mScarlet levels, and sequenced the barcodes in each sorted population.