The expression of Rpb10, a small subunit common to RNA polymerases, is modulated by the R3H domain-containing Rbs1 protein and the Upf1 helicase

Abstract The biogenesis of eukaryotic RNA polymerases is poorly understood. The present study used a combination of genetic and molecular approaches to explore the assembly of RNA polymerase III (Pol III) in yeast. We identified a regulatory link between Rbs1, a Pol III assembly factor, and Rpb10, a small subunit that is common to three RNA polymerases. Overexpression of Rbs1 increased the abundance of both RPB10 mRNA and the Rpb10 protein, which correlated with suppression of Pol III assembly defects. Rbs1 is a poly(A)mRNA-binding protein and mutational analysis identified R3H domain to be required for mRNA interactions and genetic enhancement of Pol III biogenesis. Rbs1 also binds to Upf1 protein, a key component in nonsense-mediated mRNA decay (NMD) and levels of RPB10 mRNA were increased in a upf1Δ strain. Genome-wide RNA binding by Rbs1 was characterized by UV cross-linking based approach. We demonstrated that Rbs1 directly binds to the 3′ untranslated regions (3′UTRs) of many mRNAs including transcripts encoding Pol III subunits, Rpb10 and Rpc19. We propose that Rbs1 functions by opposing mRNA degradation, at least in part mediated by NMD pathway. Orthologues of Rbs1 protein are present in other eukaryotes, including humans, suggesting that this is a conserved regulatory mechanism.


RNA isolation and northern hybridization
Total RNA was isolated by heating and freezing the cells in the presence of SDS and phenol as described previously (6). Samples containing 20 µg of total RNA were denaturated in a glyoxal reaction mixture at 55C for 1 h (7) and were resolved by electrophoresis in 1.2 % agarose gel in 1BPTE buffer (10 mM PIPES, 30 mM Bis-Tris, 1 mM EDTA pH 8.0). RNA samples were transferred into a Hybond-N+ membrane (Amersham) with 10SSC using the TurboBlotter downward capillary transfer system (Schleicher & Schull) and crosslinked by UV radiation (0.14 J/cm 2 ). The blot was prehybridized for 3 h at 65C in buffer containing 7% SDS, 0.5 M Na2HPO4 pH 7.4, 1 mM EDTA, 1% BSA, and hybridized with RPB10, ACT1 and SCR1 probes. DNA probes were amplified by PCR using oligonucleotides listed in the Supplementary Table  1 and were labeled with [-32 P]-dATP by random priming using the HexaLabel DNA labeling kit (Fermentas). Hybridization was carried out overnight at 65C in the same buffer as prehybridization. Filters were washed three times (15 min each) with 2SSC, 0.1% SDS at 65C. Hybridization signals were exposed to a phosphorimager screen. RNA was quantified using the PhosphorImager STORM 820 (Molecular Dynamics). Band intensities were quantified using the MultiGauge v3.0 software (Fujifilm).

Protein extraction and western blot analysis
The protein extraction method was described earlier (8). Protein extracts were separated by 6% or 10% sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE). After electrophoresis proteins were transferred onto the nitrocellulose membrane (Milipore), which was then blocked in TBST (10mM Tris, 150 mM NaCl, 0.05% Tween 20) containing 5% fatfree dry milk for 30 min and subsequently incubated with the appropriate antibody: mouse monoclonal antibodies 9E10 anti-Myc (Roche) at a 1:2,000 dilution for 1 h, anti-HA (Covance) at a 1:5,000 dilution for 2h, anti-Pgk1 (Abcam) at a 1:20,000 dilution for 1 h, anti-Nab2 (gift from Torben Heick Jensen) at a 1:1,000,000 dilution for 1 h and rabbit polyclonal antibodies anti-Rbs1 (Gramsh) at a 1:1,000 dilution for 1h. Blots with TAP-tagged Upf1 protein were incubated with anti-PAP antibody at a 1:3,000 dilution. PAP is a peroxidase anti-peroxidase antibody which does not necessitate the use of a secondary antibody. Then membrane was incubated for 1 h with a secondary anti-mouse or anti-rabbit antibody coupled to horseradish

Quantification and statistical analysis of CRAC data
Pre-processing and data alignment Illumina sequencing data were demultiplexed using in-line barcodes and in this form were submitted to GEO. First quality control step was performed using FastQC software (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/) considering specificity of CRAC data. Raw reads were trimmed with flexbar v3.4.0 (Dodt et al., 2012) with parameters -q TAIL -qf i1.8 -qt 20 to remove bases with QC<20. Subsequently reads were collapsed to remove PCR duplicates using FASTX-collapser v0.0.14 (http://hannonlab.cshl.edu/fastx_toolkit/) then inline barcodes were removed using pyBarcodeFilter.py script from pyCRAC package v3.0 (9). The 3' adapter were removed using flexbar v3.4.0 (10) with parameters -as TGGAATTCTCGGGTGCCAAGGC -u 3 -m 17 -n 16 -bt RIGHT. All datasets were aligned to the yeast genome using Novoalign v2.07.00 (http://www.novocraft.com) with parameter -r random and saved in novo format to calculate classes of bound RNAs and BigWig files for visual inspection in IGV genome browser. Second quality control step was performed using pyReadCounters script (pyCRAC package) which calculates overlaps between aligned cDNAs and yeast genomic features. The 1 nt resolution BigWig files were generated using bamCoverage v3.1.3 script from deepTools package (11). Sam file operations were performed using SAMtools v1.9 (12). Additionally all datasets were aligned to yeast transcriptome with parameter using STAR aligner v.2.7.3a (13) and only unambiguously mapped reads were used to generate binding profiles. Downstream analyses were performed using python 3.

Identifying Rbs1 enriched genes
To compare CRAC data and RNA-seq all reads mapping to mRNAs were normalized to reads per million (RPM) and converted to log2 RPM. For each transcript p-value was calculated using a two-sided T-test. High confidence Rbs1 targets were selected using following criteria: p-value <0.01, ratio between Rbs1 binding and RNA-seq >1.5 and >128 uniquely mapped RPM.

Rbs1 metagene representation
For metagene analysis data were processed independently for each replicate. Reads mapped to each transcript were summed up to 1 and fraction of reads was used further. This excluded risk that the obtained profile is biased by the most abundant transcripts, as each transcript has a value of 1. To combine transcripts of different length, for metagene representation fraction of