dCas9-targeted locus-specific protein isolation method identifies histone gene regulators

Abstract

Eukaryotic gene regulation is a complex process, often coordinated by the action of tens to hundreds of proteins. Although previous biochemical studies have identified many components of the basal machinery and various ancillary factors involved in gene regulation, numerous gene-specific regulators remain undiscovered. To comprehensively survey the proteome directing gene expression at a specific genomic locus of interest, we developed an in vitro nuclease-deficient Cas9 (dCas9)-targeted chromatin-based purification strategy, called "CLASP" (Cas9 locus-associated proteome), to identify and functionally test associated gene-regulatory factors. Our CLASP method, coupled to mass spectrometry and functional screens, can be efficiently adapted for isolating associated regulatory factors in an unbiased manner targeting multiple genomic loci across different cell types. Here, we applied our method to isolate the Drosophila melanogaster histone cluster in S2 cells to identify several factors including Vig and Vig2, two proteins that bind and regulate core histone H2A and H3 mRNA via interaction with their 3' UTRs.

Keywords: CRISPR/Cas9; gene expression; histone regulation; reverse ChIP.

Fig. 1.

Layout of the CLASP. Graphic depiction of the CLASP method. (1) Cells of interest are crosslinked with a crosslinker of choice. (2) Small fragments of chromatin are generated by mechanical shearing of the fixed cells. (3) Recombinant dCas9-3×FLAG loaded with the chosen guide RNA is added to the chromatin mixture. (4) Anti-FLAG antibody conjugated to resin is added to RNP/chromatin and washed; enriched chromatin is eluted with 3×FLAG peptide. (5) Through the use of either Proteinase K or nuclease treatment, enriched DNA or protein samples can be isolated for downstream applications.

Fig. 2.

Purification and identification of proteins associated with the Drosophila HisC. (A) Graphic representation of the eight different genomic targets of the sgRNA used in the mixed HisC pool. (B) qPCR analysis of the HisC IP sample vs. the nontargeting IP sample. Each sample was prepared from two billion synchronized S2 cells fixed with 1% formaldehyde for 15 min. A dCas9-targeted ChIP sample loaded with a mix of HisC-targeting guide RNAs showed significant enrichment of the H2A/H2B promoter region and the 5′ end of the H2A gene promoter. (C) Venn diagram of the total proteins found only in HisC-specific IP samples, nontargeting IP samples, and within both samples. Results are from the first HisC pull-down sample. (D) Table of proteins that showed a dNSAF ratio ≥2 between the first HisC-specific sample and nontargeting sample and that have an association with nucleic acids within the DAVID database. (E) Table of proteins that were identified as having a LisH domain from the second HisC-specific pulldown and their relative abundance ratio in HisC-specific samples and nontargeting samples. A ratio of >999 indicates that the protein was not identified in the nontargeting sample in D and E.

Fig. 3.

dsRNA knockdown assays reveal potential histone gene expression regulators. (A) Proteins enriched by a dNSAF ratio of 2 and associated with nucleic acids are knocked down by dsRNA over 72 h. cDNA is synthesized with iScript reverse transcriptase with a mixture of poly-A and random hexamer primers. H2A mRNA levels for each knockdown were calculated relative to the nonspecific dsRNA control using Tub84b as a reference gene. Compared with the nonspecific knockdown, CG11844, vig, and brahma show a negative effect on H2A mRNA expression. (B) LisH domain-containing proteins that are enriched in the HisC sample are knocked down by dsRNA as described in A. Compared with the nonspecific knockdown, Smu1, mahj, and mxc all showed a negative effect on the mRNA expression of H2A. Data plotted are averages and SDs from three separate knockdowns.

Fig. 4.

Vig and Vig2 bind specifically to histone mRNA. (A) Enrichment of HisC mRNA from V5 IPs using cells stably overexpressing Vig-5 and Vig2-V5 fusion protein. Compared with empty vector controls, Vig and Vig2 protein IPs significantly enrich for histone mRNA as opposed to reference genes such as Tub84b, Actin, and Rpl32. (B) Graphic depiction of deletion mutants made across the Vig protein. The deletion mutants are used for a transient transfection of S2 cells, and after 72 h the fusion protein is immunoprecipitated via the V5 tag. Reverse transcription is performed as described in A. The enrichment of core histone mRNA through Vig IP remains undisrupted for the majority of deletion mutants, with the exception of MUT4, which deletes the C terminus of Vig containing the putative RGG box motif. (C) Recombinant Vig protein is overexpressed and purified. Recombinant Vig is then added to a mixture of a P³² end-labeled H2A 3′ UTR probe with and without 100-fold excess of nonspecific competitor yeast tRNA. The mixture is run on a 6% acrylamide gel in 1× Tris–glycine–EDTA buffer and visualized by exposing on a phosphoimager screen. Recombinant VIG protein binds and shifts the H2A 3′ UTR probe in the presence of a large excess of nonspecific competitor RNA. (D) ChIP is performed on S2 cells stably expressing Vig-V5 and Vig2-V5. qPCR analysis of Vig-V5 and Vig2-V5 ChIP samples shows that both Vig-V5 and Vig2-V5 is enriched at the 3′ end of the H2A gene compared with other regions of the Drosophila HisC. All data plotted are averages and SDs from three separate pull-down experiments.