Fig. 2: eRNAs trigger NELF release from the paused elongation complex in a length- and the guanosine-dependent manner in vitro.

a Nucleic acid scaffold used for radioactive electrophoretic mobility shift assays (EMSAs). Template DNA (49 nt) is shown in dark blue, fully-complementary non-template DNA (49 nt) is shown in light blue, and nascent RNA (25 nt) is shown in red. b EMSAs demonstrate the ability of Arc eRNA variants to release NELF from the PEC (paused elongation complex). Results for Arc eRNAs of three different lengths are shown. Top: Arc eRNA (1–200), middle: Arc eRNA (1–100), bottom: Arc eRNA (1–50). c EMSAs with Nr4a1-(a) variants, experimental setup as for (b). d EMSAs with Nr4a1-(b) variants, experimental setup as for (b). For all EMSAs the PEC was assembled on the nucleic acid scaffold shown in (a) using 0.8 pmol ³²P-labeled RNA. The Pol II-DSIF complex was then formed by using 1.2 pmol Pol II and 2.4 pmol DSIF. Subsequently, 1.2 pmol NELF were added to form the PEC (final concentration 0.1 µM). The addition of increasing amounts of eRNAs (final concentrations: 0.15, 0.3, 0.6, 0.9, 1.2 and 1.8 µM) triggered detachment of NELF from the PEC. e NELF release from the PEC is quantified by plotting the formation of the Pol II-DSIF complex against the eRNA concentration (mean of two experimental replicates). The resulting pseudo-binding curve was fitted using a single-site binding model, apparent K_d values are indicated. f EMSAs performed with the highly-structured Arc eRNA (96–200) fragment (104 nt), a mostly single-stranded Arc eRNA Δstem mutant (100 not), and a Nr4a1-(a) Δloop 12 mutant (102 nt). The Nr4a1-(a) Δloop 12 mutant lacks prominent single-stranded regions present in Nr4a1-(a) eRNA (see Supplementary Fig. 2d). g EMSAs performed with single-stranded, low complexity RNAs, such as poly(A), poly(C), poly(U), poly(CA), poly(UA), and poly(CU) RNAs. h EMSAs were performed with 80 nt long poly(GU) RNA and different lengths of poly(GA) RNA (40 and 96 nt). i EMSAs were performed with 96 nt long poly(G₂A), (G₂A₃), and (G₂A₆) RNA. EMSAs shown in (f–i) were carried out as described for (b–d), except for an additional RNA concentration step of 2.4 µM in (g–i). j Nucleotide frequency plot for all eRNAs, whose secondary structure was determined by SHAPE-MaP (Fig. 1d and Supplementary Data 2). To prevent biases only a single TSS was utilized for this analysis in the case eRNAs possessed two alternative TSSs within a distance of <40 nt. eRNA sequences were extended to 1 kb and divided into bins of 200 nt. Guanosines are only significantly overrepresented (as determined by a pairwise t-test with p values (A/G) = 0.018; (C/G) = 0.020; (U/G) = 0.036) in the 5′-terminal 200 nt. Source data for (b–j) are provided in a Source Data file.