A bioinformatics screen reveals hox and chromatin remodeling factors at the Drosophila histone locus

Abstract

Background: Cells orchestrate histone biogenesis with strict temporal and quantitative control. To efficiently regulate histone biogenesis, the repetitive Drosophila melanogaster replication-dependent histone genes are arrayed and clustered at a single locus. Regulatory factors concentrate in a nuclear body known as the histone locus body (HLB), which forms around the locus. Historically, HLB factors are largely discovered by chance, and few are known to interact directly with DNA. It is therefore unclear how the histone genes are specifically targeted for unique and coordinated regulation.

Results: To expand the list of known HLB factors, we performed a candidate-based screen by mapping 30 publicly available ChIP datasets of 27 unique factors to the Drosophila histone gene array. We identified novel transcription factor candidates, including the Drosophila Hox proteins Ultrabithorax (Ubx), Abdominal-A (Abd-A), and Abdominal-B (Abd-B), suggesting a new pathway for these factors in influencing body plan morphogenesis. Additionally, we identified six other factors that target the histone gene array: JIL-1, hormone-like receptor 78 (Hr78), the long isoform of female sterile homeotic (1) (fs(1)h) as well as the general transcription factors TBP associated factor 1 (TAF-1), Transcription Factor IIB (TFIIB), and Transcription Factor IIF (TFIIF).

Conclusions: Our foundational screen provides several candidates for future studies into factors that may influence histone biogenesis. Further, our study emphasizes the powerful reservoir of publicly available datasets, which can be mined as a primary screening technique.

Keywords: ChIP-seq; Course-based Undergraduate Research Experience; Drosophila; Galaxy; Histone locus; Histone locus body; Hox factors.

Fig. 1

Known HLB factor CLAMP localizes to the GA-repeat cis elements in the H3/H4 promoter. (A) A diagram detailing the validated cis elements in the histone array including the TATA-box elements (maroon boxes), the TATA-less motif (teal box), and the CLAMP binding GA-repeat elements (green boxes). (B) We mapped ChIP-seq data for the known HLB factor CLAMP (green) from 2–4 h embryos [12]. The ChIP signal was normalized to its respective ChIP input signal (blue)

Fig. 2

Expected general transcription factors localize to the histone array. (A) We mapped ChIP-exo data for TRF2 (maroon, [30]) from S2 cells to the histone gene array, which recapitulates results from Isogai et al. 2007 showing localization specifically to the H1 promoter, validating our bioinformatics pipeline. We also mapped ChIP-exo data for M1BP (yellow, [30]) which did not localize to the histone gene array, further validating our pipeline. We compared ChIP-exo data to an IgG control (blue, [30]. (B) We aligned ChIP-exo data for TAF-1 (maroon, [30]) from S2 cells to the histone gene array and compared to a corresponding IgG control. We aligned ChIP-seq datasets for TFIIB (teal, two replicates overlayed, [31]) and TFIIF (pink, one replicate, [31]) from OregonR mixed population embryos to the histone gene array and normalized to the provided input (blue). TFIIB shows localization to the H3/H4 promoter and the H2A/H2B promoter, and TFIIF shows localization to both core promoters and the H1 promoter, confirming that our bioinformatics pipeline can be used to identify novel factors that localize to the histone gene array

Fig. 3

DNA-binding factors from different categories that did not pass the bioinformatics screen. We aligned ChIP-seq datasets for (A) Scm (pink, two replicates overlayed, [76]) from S2 cells, (B) MSL1 (yellow, one replicate, [39]) from S2 cells, (C) CP190 (maroon, two replicates overlayed, [40]) from Kc cells, and (D) Opa (teal, two replicates overlayed, [51]) from 3 h mixed population embryos to the histone array. We normalized each ChIP signal to its respective ChIP input signal (blue)

Fig. 4

JIL-1, Hr78, and Fs(1)hL localize to the histone gene array. We mapped ChIP datasets for (A) JIL-1 (pink, two replicates overlayed, [57]) from male third instar larvae, (B) Hr78 (maroon, two replicates overlayed, [73]) from 8–16 h mixed population embryos, and (C) the long (L, teal) and short (S, yellow) isoforms of fs(1)h from Kc cells [60] to the histone gene array. We normalized each ChIP-seq dataset to its respective input (blue)

Fig. 5

Hox factors Ubx, Abd-A, and Abd-B localize to the histone array. (A) Diagram of relative tissue expression patterns for Ubx (maroon), Abd-A (teal) and Abd-B (yellow). (B) We aligned ChIP-seq datasets from Kc cells expressing Ubx (maroon, two replicates overlayed, [63] ), Abd-A (teal, two replicates overlayed, [63] ), and Abd-B (yellow, two replicates overlayed, [63] ) to the histone gene array. We normalized each ChIP-seq dataset to the provided input (blue, two replicates overlayed, [63] ). (C) Enlarged signal from (B) of Ubx (maroon), Abd-A (teal), and Abd-B (yellow) over the H3/H4 promoter

Fig. 6

Ubx localizes to the H3/H4 promoter in embryos and 3rd instar larva. We mapped Ubx ChIP-seq datasets from (A) mixed population embryos (maroon, two replicates overlayed, [77] ) and (B) imaginal wing discs in third instar larva (maroon, two replicates overlayed, [78] ) to the histone gene array. We normalized ChIP-seq datasets to the provided inputs (blue, two replicates overlayed)

Fig. 7

ChIP-seq datasets from different tissues can show different alignment results. We mapped two different ChIP-seq datasets for Nejire (Nej) to the histone gene array. ChIP data from 2–4 h embryos (maroon, one replicate, [74]), showed localization to the H3/H4 promoter and the H2A/H2B promoter, while ChIP-seq data from S2 cells (pink, one replicate, [75] ) showed no localization to the histone gene array. We also aligned ChIP-seq data for Pnt from stage 11 embryos [55] to the histone gene array. We normalized the ChIP-seq signals to their respective input signals (blue)