H3K36 methylation and DNA-binding both promote Ioc4 recruitment and Isw1b remodeler function

Abstract The Isw1b chromatin-remodeling complex is specifically recruited to gene bodies to help retain pre-existing histones during transcription by RNA polymerase II. Recruitment is dependent on H3K36 methylation and the Isw1b subunit Ioc4, which contains an N-terminal PWWP domain. Here, we present the crystal structure of the Ioc4-PWWP domain, including a detailed functional characterization of the domain on its own as well as in the context of full-length Ioc4 and the Isw1b remodeler. The Ioc4-PWWP domain preferentially binds H3K36me3-containing nucleosomes. Its ability to bind DNA is required for nucleosome binding. It is also furthered by the unique insertion motif present in Ioc4-PWWP. The ability to bind H3K36me3 and DNA promotes the interaction of full-length Ioc4 with nucleosomes in vitro and they are necessary for its recruitment to gene bodies in vivo. Furthermore, a fully functional Ioc4-PWWP domain promotes efficient remodeling by Isw1b and the maintenance of ordered chromatin in vivo, thereby preventing the production of non-coding RNAs.


Figure S2
The PWWP aromatic cage is important for H3K36me3 recognition. EMSAs (A) and competitive EMSAs (B) were performed to assess the ability of the aromatic cage mutant W22A to specifically recognise H3K36 trimethylated nucleosomes. As expected, mutation of W22 led to a loss of discrimination in both settings. Furthermore, the mutant PWWP domain displayed lower affinities towards nucleosomes in general, when compared to the wildtype domain. The positions of the NCP and free DNA bands on the gels are indicated, as are the bands denoting the complexes formed by the PWWP proteins with NCPs (*) as well as with free DNA (**).

Yeast strains and media
All yeast strains used in this study are listed in Table S2. Wildtype IOC4 was tagged with a 3xFLAG epitope by targeted homologous integration of a PCR product derived from amplification of plasmid p3xFLAG-HIS3 with gene-specific primers. In order to generate yeast strains bearing different mutations of IOC4, mutant IOC4 constructs were first cloned into plasmid p3xFLAG-HIS3, followed by PCR amplification of these cassettes and transformation into wildtype yeast. To generate TAP-tagged yeast strains, the 3xFLAG tag was replaced by targeted homologous integration of a PCR product derived from amplification of plasmid pBS1539 with construct-specific primers. Single deletion of CHD1 was done by targeted homologous recombination of PCR fragments containing either the hygromycin (HphB) or kanamycin (KanMX) resistance marker. All strains generated in this study were confirmed by PCR and/or sequencing. Cells were grown at 30°C in YPD (1% yeast extract, 2% bacto-peptone, 2% dextrose) medium.

Yeast growth assay
Yeast strains were inocculated at an OD600 of 0.1 from overnight cultures and grown for ca. 5 hours at 30°C in YPD until they were growing exponentially. Equal numbers of cells were harvested by centrifugation, washed with ddH2O and resuspended at an OD600 of 0.5 in ddH2O. Five 6-fold dilutions were prepared for each strain and spotted onto YPD plates +/-1µM propiconazole. Plates were incubated at 30°C for 3-5 days. Full-length GST-Ioc4 was produced as described previously (1). All GST-tagged proteins were purified using glutathione-sepharose as described previously (3). If needed, GST tags were removed by cleavage with 3C protease while bound to glutathione-sepharose. Proteins were dialyzed exhaustively against 50 mM phosphate, pH 7.0, 500 mM NaCl, 10% Glycerol, flash-frozen in liquid nitrogen and stored at -80 °C. Full-length Ioc4 (aa1-475), Ioc42KE and Ioc4DPWWP (aa 179-475) were cloned into a modified pCoofy vector with an N-terminal 6xHis-MBP tag. Proteins were overexpressed in E. coli BL21 RIL by addition of 0.25 mM IPTG for 20 hours at 16 °C. Proteins were batch purified using nickel-NTA resin, followed by adsorption chromatography using a heparin column and gel filtration chromatography. Proteins were dialyzed exhaustively against 50 mM phosphate, pH 8.0, 500 mM NaCl, 10% Glycerol, 50 mM Arg, 50 mM Glu, flash-frozen in liquid nitrogen and stored at -80 °C.

Electromobility shift assays (EMSA)
Binding of Ioc4PWWP to DNA was assessed by EMSA. Double-stranded (ds) DNA with a length of 30 bp was prepared by heating and annealing complimentary, single-stranded (ss) DNA (Table S3). Gels were stained with ethidium bromide and visualized using a Tanon 1600 gel imaging system (Tanon Inc.). Alternatively, DNA was labeled with Cy5 and visualized by scanning on a Typhoon FLA9500 Imaging system and quantitated using ImageQuant TL software (GE Healthcare). DNA bands were quantitated for each lane. All lanes containing PWWP were normalized against input lanes containing DNA only. The percentage of DNA bound was expressed as 100 -% DNA for each lane. Mean values ± SEM were plotted for at least three independent experiments.
EMSA assays with reconstituted, K36me0-or K36me3-containing mononucleosomes were performed as described (3). For binding reactions 15 fmol of reconstituted nucleosomes per reaction were incubated with increasing concentrations (0-800 nM) of wildtype and mutant PWWP domains. For binding reactions with full-length Ioc4 15 fmol of reconstituted nucleosomes per reaction were incubated with increasing concentrations (0-35 nM) of wildtype or mutant Ioc4. For binding reactions with unmodified wildtype nucleosomes 30 fmol of reconstituted nucleosomes per reaction were incubated with increasing concentrations (0-4 µM) of wildtype or mutant PWWP. Complexes were separated by electrophoresis on 5.0 % native polyacrylamide gels (37.5:1), run in 0.4x TBE, 2 % glycerol. Gels were scanned using a Typhoon Imaging FLA9500 system and quantitated using ImageQuant TL software (GE Healthcare).
Mononucleosome bands were quantitated for each lane. All lanes containing PWWP were normalized against input lanes containing mononucleosomes only. The percentage of nucleosomes bound was expressed as 100 -% mononucleosomes for each lane. Mean values ± SEM were plotted for at least three independent experiments.

Antibodies
The following antibody was used in this study: αFlag M2 (Sigma #F1804).

Chromatin immunoprecipitation assays
For ChIPs of FLAG-tagged Ioc4 yeast strains were grown in 200 ml of YPD at 30 ºC, crosslinked and processed for ChIP as described earlier (7,8). FLAG-tagged Ioc4 was immunoprecipitated and processed as described before (7,8). Three biological repeats were done for all experiments and used for subsequent ChIP-qPCR and/or ChIP-chip analyses.

ChIP-qPCR analysis
Immunoprecipitated DNA was quantitated by qPCR using the PowerTrack SYBR Green Master Mix (Thermo Fisher Scientific) and a QuantStudio 5 Real-Time PCR System (Applied Biosystems). Primers are listed in Table S3. The mean signal was calculated for each experiment and normalized against input samples at each primer positions as internal controls. Input-normalized values were further corrected for variation by normalising against the mean signals for two control regions (STE3, subtelomeric region on chromosome V (ChrV)).

ChIP-chip microarray analysis
ChIP-chip assays for the genome-wide distribution of FLAG-tagged Ioc4 were performed as described previously (1), using 8x60K yeast genome DNA arrays (Agilent, Array #031697) with an average probe spacing of ca. 200 bp. 20-50 ng of input and IP samples were used for double T7 linear amplification and labeling. Inputs were labeled with Cy3 dye and IPs with Cy5 dye. 4 μg of input and IP were combined and used for hybridization. Arrays were scanned (Agilent DNA Microarray Scanner Model G2505B) and extracted using Feature Extraction software (Agilent) and normalized using median normalization in R software.

Data analysis
The normalized data were analyzed using a modified gene average analysis (1). ORFs were subdivided into 14 equal sised bins each. Intergenic regions (480 bp up-and downstream of genes) were allocated into three bins each. Microarray enrichment ratios [log2(IP/Input)] for each probe were assigned to the closest bin. For whole-genome average gene plots all probes within a bin were averaged and plotted as mean ± standard error (SEM). Genes that are not regulated by RNAPII, including tRNA and snRNA genes as well as the majority of dubious ORFs (ca. 450 genes) were removed from the analysis.

Isolation of total RNA
Yeast strains were grown in YPD at 30ºC until they reached an OD600 of 0.8. Total RNA was isolated using acid phenol extraction as described previously (9). RNA quality and quantity were assessed by UV spectroscopy using a NanoDrop 2000 (Thermo Scientific).

Northern blots
Northern blotting and hybridization were done as described previously (10). 20 μg of total RNA were used to assess the cryptic transcript phenotype. Blots were exposed onto phosphorimaging screens and scanned using a Typhoon FLA9500 Imaging system.

Strand-specific multiplex RT-qPCR
Total RNA was treated with TURBO DNA-free DNase I (ThermoScientific) to remove genomic DNA according to the manufacturer's instructions. 3.5 μg of DNase-treated RNA were used for subsequent reverse transcription (RT) reactions containing 2 pmol of each gene-specific primer, 60 U of SuperScript III (ThermoScientific) and 0.3 μg of Actinomycin D (AppliChem) to prevent antisense cDNA artefacts. Two RT reactions containing "forward" primers (Table S3) annealing to antisense transcripts derived from target genes VTH2 and YEN1, or FAA2 and ARO80 were set up. Each reaction also contained 2 pmol "reverse" primer annealing to the canonical reference ACT1 transcript. Annealing was done at 70°C for 10 minutes. First strand synthesis was performed at 55°C for one hour, followed by heat inactivation at 70°C for 15 min. All samples were quantitated by qPCR using the PowerTrack SYBR Green Master Mix (Thermo Fisher Scientific) and a QuantStudio 5 Real-Time PCR System (Applied Biosystems). Ctvalues were first normalized to those of ACT1. Subsequently, results for mutant yeast strains were normalized to the mean signal for wildtype yeast samples using the comparative CT (ΔΔCT) method. Five biological replicates were performed for each strain.