Fig 1.
The RhoA enhancer activates gene expression in Drosophila abdominal C1-SOPs.
(A, B) Lateral view of Drosophila RhoBAD-LacZ (A) or RhoAAA-LacZ (B) embryos (stage 11) immunostained for β-gal (green) and AbdA (purple). Both reporters are active in a specific cell type (C1-SOP) with higher levels observed in abdominal segments (stained by AbdA, first abdominal segment marked by “A1”) than thoracic segments. (C) The RhoA sequence has binding sites for Pax2, Sens, Exd, Hth, and AbdA that are critical for proper RhoBAD-LacZ and RhoAAA-LacZ activity in Drosophila embryos [17, 22]. (D) Schematic model of competition between activator (Pax2/Exd/Hth/AbdA) and repressor (Sens) TFs. Sens binds and represses RhoA activity in the thorax; whereas AbdA and the activators outcompete Sens to promote gene activation in C1-SOP cells of the abdomen.
Fig 2.
RhoA contains low affinity Pax2 and Sens binding sites.
(A, D) Alignment of Pax2 (A) and Sens (D) logos derived from SELEX-seq [23] to RhoA and selected B1H sites [24]. Mismatches to the logos are highlighted in red. (B, E) Pax2 (B) and Sens (E) binding to RhoA and selected B1H hits using EMSAs. Each probe was incubated with 0, 106, or 212 ng of Sens or 0, 48, or 96 ng of Pax2. Full gels are shown in S2 Fig. (C, F) Correlation between proportion of probe bound in EMSAs versus proportion predicted by PWM energy models. The Spearman-rank correlation (ρ) and coefficient-of-determination (r2) are indicated on the plots. Linear regression of this relationship is shown in blue.
Fig 3.
A high affinity Pax2 binding site results in ectopic RhoA activity in additional PNS cells.
(A) The SELEX-seq [23] Pax2 logo aligned with RhoA-WT and RhoA-PS. Mis-matches are in red font, and nucleotides that improve the match are in green font. The Sens, Exd, Hth, and Hox TFBSs are highlighted. (B) EMSAs using purified Pax2 protein (0, 10.25, 20.5, 41, and 82 ng) on RhoA-WT and RhoA-PS probes reveal Pax2 has a higher affinity for RhoA-PS (Full gels are shown in S3 Fig). (C, D) Lateral view of RhoBAD-LacZ (C) and RhoBAD-PS-lacZ (D) embryos (stage 11) immunostained for β-gal. β-gal intensity is represented by heat-map at left. “A1” indicates the first abdominal segment. (C’, D’) Close-up of an abdominal C1-SOP with arrowheads highlighting non-C1-SOPs that activate RhoBAD-PS-lacZ. (C”, D”) Same close-up showing β-gal (green) and Pax2 (red). (E) Boxplot of β-gal immunostain intensities in thoracic and abdominal C1-SOPs. One-tailed Welch’s t-test was used to compare mean β-gal intensity per embryo (* p < 0.05, *** p < 0.001), n = 20 (WT) and 23 (PS). Each box represents measurements from a single embryo. (E’) Boxplot of β-gal immunostain intensities in non-C1 PNS (Sens+) cells. Dotted line represents 5th percentile of β-gal intensity in C1-SOPs–a threshold to define “C1-like reporter activity”. (F) Proportion of non-C1 PNS cells per embryo with C1-like β-gal intensities in RhoBAD-lacZ and RhoBAD-PS-lacZ embryos (** p < 0.01, One-tailed Wilcoxon Rank Sum Test), n = 20 (WT) and 23 (PS). (G-H) Lateral view of RhoBAD-rhocDNA; rho7M (G) and RhoBAD-PS-rhocDNA; rho7M (H) embryos (stage 15) immunostained for an oenocyte marker (HNF4). Note, in the absence of rho, embryos do not develop HNF4+ oenocytes [15, 17]. (I) Violin plots of the number of oenocytes (HNF4+) per embryonic segment for all RhoBAD-rhocDNA or RhoBAD-PS-rhocDNA embryos. Lines represent range of oenocytes per segment for each embryo, while dots represent individual segments (* p < 0.05, One-tailed Wilcoxon Rank Sum Test), n = 95 (WT) and 98 (PS). (J, K) Lateral views of RhoBAD-rhocDNA; rho7M/+ (J) and RhoBAD-PS-rhocDNA; rho7M/+ (K) embryos immunostained for HNF4. Arrowheads indicate ectopic HNF4+ cells. (J’, K’) Close-up of A1-A3 abdominal segments of panels J and K.
Fig 4.
A high affinity Sens site results in repression of RhoA in abdominal C1-SOPs.
(A) The SELEX-seq [23] Sens logo aligned with RhoA variants. Mis-matches are in red font, and sequence variants that improve the match are in green font. The Pax2, Exd, Hth, and Hox TFBSs are highlighted. (B, C) EMSAs using the indicated RhoA probes with either purified Sens (0, 23.5, 57, 114, and 228 ng) or Pax2 (0, 10.25, 20.5, 41, and 82 ng). Full gels are shown in S3 Fig. (D-G) Lateral view of stage 11 RhoBAD-lacZ (D), RhoBAD-SS-lacZ (E), RhoBAD-PM-lacZ (F), and RhoBAD-PSSS-lacZ (G) embryos immunostained for β-gal. β-gal intensity is represented by a heat-map at left. “A1” indicates the first abdominal segment. (H) Quantification of β-gal intensity in abdominal C1-SOPs in age-matched embryos. Each box represents measurements from a single embryo. RhoBAD-SS-lacZ, RhoBAD-PM-lacZ, and RhoBAD-PSSS-lacZ embryos were processed and imaged separately, each with RhoBAD-lacZ control embryos. Quantification for a representative set of RhoBAD-lacZ embryos are shown. β-gal intensities for each variant are reported as relative to the average β-gal intensity of control embryos. Two-tailed Welch’s T-test with Bonferroni correction was done to compare β-gal intensities to RhobAD-SS (* p < 0.05, ** p < 0.001, *** p < 0.0001), n = 12 (WT), 9 (SS), 13 (PM), and 19 (PSSS). (I-K) Lateral view of RhoBAD-rhocDNA (I), RhoBAD-SS-rhocDNA (J), and RhoBAD-PSSS-rhocDNA embryos in a rho7M background (stage 15) immunostained for an oenocyte marker (HNF4). Note, at least 10 embryos with transgenes containing high affinity Sens sites were analyzed and no oenocytes were observed.
Fig 5.
Overlapping activator and repressor binding sites are required for abdomen-specific RhoA activity.
(A) Sequences of tested RhoA variants. RhoA-RDM contains random nucleotides downstream of the Hox site. RhoA-SM contains mutations that decrease Sens binding, and RhoA-SM/SWT, RhoA-SM/SS and RhoA-SM/SM add either a low affinity (WT), high affinity (SS), or mutant (SM) Sens site downstream of the Hox site. (B-C) Quantification of β-gal immunostaining intensities in C1-SOPs in RhoBAD-LacZ versus RhoBAD-RDM-LacZ (B) or RhoBAD-SM-LacZ (C). Each box summarizes measurements from a single embryo. Two-tailed Welch’s T-test was used to compare RhoBAD-SM and RDM mutants to wildtype, n = 12 (WT) and 18 (RDM) in (B) and n = 12 (WT) and 15 (SM) in (C). (D) EMSAs comparing binding of purified Sens to RhoA probes (0, 57, 114, and 228 ng of Sens). (E) EMSAs assessing competition between purified Sens (114 or 228 ng) against purified AbdA (189 ng) and Exd/Hth (59.2 ng) on RhoA-SS and RhoA-SM-SS. (E’) Close-up view of Exd/Hth/Hox and Exd/Hth/Hox/Sens complexes on DNA probes. Schematics denote the formation of each transcription factor complex. (F-I) Lateral view of stage 11 RhoBAD-SM-lacZ (F), RhoBAD-SM/SWT-lacZ (G), RhoBAD-SM/SM-lacZ (H), and RhoBAD-SM/SS-lacZ (I) embryos immunostained for β-gal. Intensity of β-gal stain is represented by heat-map at left. “A1” indicates first abdominal segment. Note, no RhoBAD-SM/SS activity is detected in the PNS and the activity that is observed is in cells of the gut. (J) Quantification of β-gal intensities in thoracic and abdominal C1-SOPs of noted RhoBAD-lacZ embryos. Each box represents measurements from a single embryo. Statistical analysis was done using Kruskal-Wallis test followed by post-hoc pairwise Mann-Whitney U test with Bonferroni correction, n = 25 (SM), 23 (SM/SM), 22 (SM/SWT), and 24 (SM/SS). For all statistical comparisons, n.s. p ≥ 0.05; * p < 0.05, ** p < 0.001, *** p < 0.0001.
Fig 6.
Inverse correlation between PWM information content and the ability to identify low-affinity Pax2 and Sens TFBSs.
(A, B) Top Panel: Published Sens and Pax2 PWMs placed in order from lowest to highest information content (left-to-right). PWMs were derived from published B1H (low and high stringency) and SELEX-seq assays [23, 24]. Flanking low information positions were removed to make all PWMs the same length (shaded boxes). Total information content (bits) of the trimmed PWMs is indicated above each PWM. The Relative Log-Likelihood (RLL) score of each PWM on RhoA is indicated below the PWMs. Bottom Panel: AUROC of each published PWM for discriminating PBM probes (binned by fluorescence, as indicated on x-axis) from 10 sets of non-specific probes (matched number of control probes randomly selected from the 50% of probes with the lowest fluorescence). Sens and Pax2 PWMs were tested on PBMs for the vertebrate homologs H. sapiens Gfi1b and D. rerio Pax2b. Statistical comparisons were conducted using Kruskal-Wallis test. P-values were Bonferroni-adjusted to correct for multiple comparisons (* p < 0.05, ** p < 0.01, *** p < 0.001).
Fig 7.
High information content PWMs are less accurate at identifying TFBSs obtained from both in vitro and in vivo binding events.
(A) Schematic describing how PWMs were created by sub-sampling Sens and Pax2 B1H hits. Each B1H hit was placed into quartiles based on 8-mer sequence frequency within the pool of B1H hits. 100 PWMs were generated by iteratively sampling 50 B1H hits from each quartile. 100 PWMs were also generated by sampling 50 B1H hits from the entire pool (Control PWMs). The range of total information content (I.C.) for PWMs in each quartile are indicated below the motifs. (B) Relative log-likelihood (RLL) score of each PWM for the RhoA sequence. (C) AUROC of each PWM for discriminating low-stringency B1H hits from shuffled sequences. (D) AUROC of each PWM for discriminating bound PBM probes (binned by fluorescence, as indicated on x-axis) from non-specifically bound probes (matched number of control probes randomly selected from the 50% of probes with the lowest fluorescence). (E) AUROC of each Sens PWM for discriminating M. musculus Gfi1 and Gfi1b ChIP-seq peaks from random, non-repetitive genomic sequences. Gfi1b ChIP-seq was conducted using multipotent Hematopoietic Progenitor cells (HPC-7) and Gfi1 ChIP was conducted using innate Type-2 Lymphocytes (ILC2) [32, 33]. Analysis was limited to the 1000 peaks with greatest fold enrichment per ChIP dataset, and ChIP peaks were binned by fold enrichment as indicated on x-axis. For panels C-E, AUROCs represent the median using 10 different sets of negative sequences. All violin plots are scaled to have the same width. Statistical analysis was performed using Kurskal-Wallis test followed by a post-hoc pairwise Mann-Whitney U test. P-values were Bonferroni-adjusted due to multiple comparisons arising from groups of PWMs (all panels) and binning of sequences (panels D and E) (n.s. p ≥ 0.05; * p < 0.05; ** p < 0.01, *** p < 0.001).