Fluorescent labeling of genomic loci in Drosophila imaginal discs with heterologous DNA-binding proteins

Abstract

Using the Drosophila melanogaster Hox gene Ultrabithorax (Ubx) as an example, we demonstrate the use of three heterologous DNA-binding protein systems-LacI/LacO, ParB1/ParS1, and ParB2/ParS2-to label genomic loci in imaginal discs with the insertion of a small DNA tag. We compare each system, considering the impact of labeling in genomic regions (1) inside versus outside of a transcribed gene body and (2) with varying chromatin accessibility. We demonstrate the value of this system by interrogating the relationship between gene expression level and enhancer-promoter distance, as well as inter-allelic distance at the Ubx locus. We find that the distance between an essential intronic cis-regulatory element, anterobithorax (abx), and the promoter does not vary with expression level. In contrast, inter-allelic distance correlates with Ubx expression level.

Keywords: 3D nuclear architecture; LacI/LacO; ParB/ParS; fluorescence microscopy; gene regulation; genomic foci.

Graphical abstract

Figure 1

A two-step targeting strategy to tag Ultrabithorax at two ROIs (A) Schematic of two-step GE process. First, CRISPR replaces an ROI with a fluorescent reporter flanked by recombinase-specific attP sites. gRNAs are targeted to regions of low conservation (lack of red bars). Second, RMCE replaces fluorescent reporter with the ROI sequence plus DNA tag. Two recombinases (PhiC31, Bxb1) and three fluorescent reporters (ubiDsRed, ubiGFP, P3RFP) were used. For additional information see Figure S1. (B) Genome browser image of Ubx gene. Tracks show chromatin accessibility in wing and haltere discs as determined by FAIRE (McKay and Lieb, 2013). UbxP and abx are boxed in green and magenta, respectively. A 6.8 kb fragment (abx6.8) that recapitulates Ubx expression in haltere discs is shown (Simon et al., 1990). (B′) Zoom of abx and UbxP. FAIRE peaks and conservation tracks (red bars) are shown. The region replaced with CRISPR and the location of DNA tag insertion are shown. CRISPR targeting and DNA tag insertion occur in regions of low conservation.

Figure 2

LacI/LacO labeling of genomic ROIs (A) Schematic of details of LacI-FP constructs. Wild-type LacI has an N-terminal DNA-binding domain and a C-terminal tetramerization domain. We removed the tetramerization domain and added a 3× SV40 NLS and fluorescent protein fused to LacI with a specified linker. LacI FP binds as a dimer to each LacO binding site. FPs include HaloTag and Neon. Sequences for linkers and NLS are shown. (B) Sequence of the LacO unit. (C) Protocol to express all FP fusions: a 15-min HS at 37°C followed by a 4-h rest period at 25°C prior to dissection. Theoretical protein levels following the HS and rest period are shown. (D) Examples of LacI-HaloTag (L1) foci in haltere and wing discs at each ROI. LacI-HaloTag (L1) is stained with HaloLigand TMR. Scale bars, 1 μm. (E) SNR of LacI-HaloTag (L1) stained with TMR at each ROI in wing and haltere discs. (F) Labeling efficiency of LacI-HaloTag (L1) stained with TMR at each ROI in wing and haltere discs. In (E) and (F), Tukey box plots are used. L2 is derived from Mir et al. (2018). Significance for (E) and (F) was tested using a one-way ANOVA followed by Tukey’s multiple comparisons test with α = 0.05 in GraphPad Prism. Reported values are adjusted p values. ns, not significant.

Figure 3

ParB2/ParS2 labeling of genomic ROIs (A) Schematic of details of ParB2 FP constructs. A 3× SV40 NLS and FP (HaloTag, Neon) is fused to the C terminus using two linkers (L1, L2). ParB2 binds ParS2 sequences and nucleates additional ParB2 with protein-protein interactions. The ParS2 binding sites from Dubarry et al. (2006) are shown, as well as sequences for the linkers and NLS. (B) Examples of ParB2-Neon(L2) spot formation at each ROI in haltere and wing discs. Scale bars, 1 μm. (C) SNR of ParB2-Neon(L2) at each ROI in haltere and wing discs. (D) Labeling efficiency of ParB2-Neon(L2) at each ROI in haltere and wing discs. In (C) and (D), Tukey box plots are used. Significance for (C) and (D) was tested using a one-way ANOVA followed by Tukey’s multiple comparisons test with α = 0.05 in GraphPad Prism. Reported values are adjusted p values. ns, not significant.

Figure 4

ParB1/ParS1 labeling of genomic ROIs (A) Schematic of details of ParB1 FP constructs. A 3× SV40 NLS and FP (HaloTag, Neon) is fused to the C terminus using two possible linkers (L1, L2). ParB1 binds ParS1 sequences and nucleates additional ParB1 with protein-protein interactions. The ParS1 binding sites from Dubarry et al. (2006) are shown, as well as sequences for linkers used and NLS. (B) Examples of ParB1-HaloTag (L1) spot formation at each ROI in haltere and wing discs. (C) SNR of ParB1-HaloTag (L1) at each ROI in haltere and wing discs. (D) Labeling efficiency of ParB1-HaloTag (L1) at each ROI in haltere and wing discs. (E) A comparison of SNR (left y axis) and labeling efficiency (right y axis) of each FP with L1 and L2. Tukey box plots are used in (C), (D), and (E), where one-way ANOVA followed by Tukey’s multiple comparisons test with α = 0.05 in GraphPad Prism was used to determine statistical significance. ∗∗∗∗p < 0.0001; ns, not significant (p > 0.05). Reported p values are adjusted p values. Scale bar is 1micron.

Figure 5

Clonal analysis of the effects of FP labeling on Ubx expression (A) Mitotic clonal analysis of FP labeling at abx. Top two panels: clones homozygous for 20× LacO abx with one copy of LacI-HaloTag (L1). Third panel: clones homozygous for ParS2-abx with one copy of ParB2-Neon. Fourth panel: clones homozygous for ParS1-abx with one copy of ParB1-HaloTag (L1). Quantification of clones with reduced Ubx: 4/8 LIHT2 clones, 0/13 B2N2 clones, 0/11 B1HT1 clones. (B) Mitotic clonal analysis of FP labeling at Ubx promoter. Top panel: clones homozygous for 20× LacO UbxP with one copy of LacI-HaloTag (L1). Middle panel: clones homozygous for ParS2-UbxP with one copy of ParB2-Neon (L2). Bottom panel: clones homozygous for ParS1-UbxP with one copy of ParB1-HaloTag (L1). Quantification of clones with reduced Ubx: 0/16 LIHT2 clones, 0/12 B2N2 clones, 0/2 B1HT1 clones. All clones were made 48 h after egg laying. Ubx immunostain, native GFP fluorescence, and a MERGE are shown, with zoom of a single clone. Scale bars, 50 μm.

Figure 6

Assaying enhancer-promoter distance within the Ubx locus (A) (Left) Ubx immunostain in imaginal discs. A wing (white dotted line), T3 leg, and haltere disc are shown. (Right) Schematic of haltere disc with high distal Ubx and low proximal Ubx. Dotted boxes show approximate position of proximal, distal ROIs. Scale bar, 50 μm. (B) Ubx immunostain in haltere discs with 4 kb deletion of abx encompassing major FAIRE peaks. GFP marks clones, outlined with dotted line. Quantification of clones shows 14/32 clones with reduced Ubx and 5/32 with no Ubx. All other clones show normal Ubx. Scale bars, 50 μm. (C) Distribution of distances between UbxP and abx. A schematic shows labeling of single Ubx allele with LacI-HaloTag (L1) (magenta) and ParB2-Neon (L2) (green). Median values and Ubx transcription level are stated. Foci images shown are representative pairs with a measured distance approximating the median. Statistical significance was tested using a one-way ANOVA followed by Tukey’s multiple comparisons test with α = 0.05 in GraphPad Prism. ns, not significant. (D) Normalized observed/expected Hi-C interaction frequency matrix of the bithorax complex (top) and Ubx locus (bottom) in wing discs at 10 kb resolution. The interaction frequency represents the number of contacts between genomic loci captured with Hi-C. The TAD encompassing the bithorax complex is denoted by black lines. The approximate position of abx (red box) is defined by ATAC-seq peaks (Loker et al., 2021). The Ubx promoter is shown by the green arrow next to the first exon.

Figure 7

Assaying inter-allelic distance at the Ubx locus (A) Distribution of inter-allelic distances between Ubx alleles, measured at UbxP. A schematic shows labeling of each allele with LacI-HaloTag (L1) (magenta) and ParB2-Neon (L2) (green). (B) Distribution of inter-allelic distances between Ubx alleles, measured at abx. A schematic shows labeling of each allele with LacI-HaloTag (L1) (magenta) and ParB2-Neon (L2) (green). For both (A) and (B), median values and Ubx transcription state are stated. Foci images shown are representative pairs with a measured distance approximating the median. Statistical significance was tested using a one-way ANOVA followed by Tukey’s multiple comparisons test with α = 0.05 in GraphPad Prism.