Domains and residues of the Saccharomyces cerevisiae hnRNP protein Hrp1 important for transcriptional autoregulation and noncoding RNA termination

Abstract

Proteins that bind the nascent transcript exiting RNA polymerase II can regulate transcription elongation. The essential Saccharomyces cerevisiae hnRNP protein Hrp1 is one such protein and participates in both cleavage and polyadenylation-coupled and Nrd1-Nab3-Sen1-dependent RNA polymerase II termination. Prior evidence that Hrp1 is a positive RNA polymerase II elongation factor suggests that its release from the elongation complex promotes termination. Here we report the effects of deletions and substitutions in Hrp1 on its autoregulation via an Nrd1-Nab3-Sen1-dependent transcription attenuator in the 5'-UTR of its mRNA and on the function of an Hrp1-dependent Nrd1-Nab3-Sen1 terminator in the SNR82 snoRNA gene. Deletion of either of two central RNA recognition motifs or either of the flanking low-sequence complexity domains is lethal. Smaller, viable deletions in the amino-terminal low-sequence complexity domain cause readthrough of both the HRP1 attenuator and SNR82 terminator. Substitutions that cause readthrough localized mostly to the RNA recognition motifs, although not always to the RNA-binding face. We found that autoregulation of Hrp1 mRNA synthesis is surprisingly robust, overcoming the expected lethal effects of the start codon and frameshift mutations via overexpression of the mRNA up to 40-fold. Our results suggest a model in which binding of attenuator or terminator elements in the nascent transcript by RNA recognition motifs 1 and 2 disrupts interactions between RNA recognition motif 2 and the RNA polymerase II elongation complex, increasing its susceptibility to termination.

Keywords: NNS termination; RNA polymerase II; hnRNP proteins; transcription; transcription termination; yeast.

Fig. 1.

Strong attenuator readthrough by hrp1-7 requires more than one substitution. a) Domain structure of S. cerevisiae  Hrp1. The N-terminal region contains an acidic sequence (TLM) that is highly conserved across fungi, a potential NES, and Asn/Ser-rich (N/S) and Gln-rich (Q) LCDs. The RRMs are shown in green and blue. The C-terminal region contains Met/Gln- (M/Q) and Asp/Asn/Ser-rich (D/N/S) LCDs, two Arg-Gly-Gly repeats (RGG₂), and a PY-type NLS. The hrp1-7 substitutions are shown below. A horizontal black line marks the RT-qPCR amplicon. b) (top) Schematic diagram of the HRP1-CUP1 reporter. +1 indicates the most upstream transcription start site. (bottom) Serial dilutions of haploid strains containing the indicated alleles of HRP1 and the HRP1-CUP1 reporter were spotted on medium containing the indicated concentration of copper sulfate. A biological replicate is shown in Supplementary Fig. 5a. c) RT-qPCR of total RNA from the indicated strains to produce the HRP1 amplicon indicated in panel A. Values were normalized to a CDC19 amplicon. Error bars represent the standard error of the mean (SEM) for 3 biological replicates. Significance between each mutant and wild-type was calculated by a 2-tailed Student's t-test indicated with P-value < 0.05 (*), <0.01 (**), or <0.001 (***).

Fig. 2.

Confirmation of SNR82 terminator readthrough in the presence of hrp1-7. a) Transcriptome data of NNS mutants including hrp1-7 from (Chen et al. 2017). The y-axis is a log2 scale with wild-type RNA levels in black and fold change in transcript levels for the mutants in blue. Annotated genes and mapped poly(A) 3′ ends (Ozsolak et al. 2010) in the top (w) strand are shown in red. The SNR82 terminator region is indicated by a white box and RT-qPCR amplicons by black lines. b) SNR82 terminator readthrough in strains containing the indicated HRP1 alleles calculated as the ratio of 3′ to 5′ RT-qPCR amplicons, normalized to wild-type HRP1. Error bars represent the SEM for 3 biological replicates. Significance between each mutant and wild-type was calculated by a 2-tailed Student's t-test indicated with P-value < 0.05 (*), <0.01 (**), or <0.001 (***). c) Schematic of the ACT1-CUP1 reporter construct containing the SNR82 terminator region (white box in panel a). d) Serial dilutions of haploid strains containing the indicated alleles of HRP1 and the SNR82  ACT1-CUP1 reporter on separate plasmids were spotted on medium containing the indicated concentration of copper sulfate. Biological replicate shown in Supplementary Fig. 5b.

Fig. 3.

Effects of Hrp1 domain deletions on cell viability, protein level, and readthrough. a) Schematic of Hrp1 primary structure and alleles tested in this study. Abbreviations are as in Fig. 1a. Viability in a haploid strain after 5 days of growth at 30° on medium containing 5-FOA is indicated at right (see Supplementary Fig. 7). Amplicon for RT-qPCR shown with a black line. b) Anti-Hrp1 Western blot of cell extracts from merodiploid strains containing chromosomal wild-type HRP1 and the indicated HRP1 alleles on a low copy number plasmid. A duplicate Coomassie blue stained gel, part of which is shown below, was used to normalize for total protein loaded. The average protein levels normalized to endogenous Hrp1 (lane 1) from 2 biological replicates (see Supplementary Fig. 11a) are shown in the graph below the gel with the FL Hrp1 shown in black and the domain deletions shown in gray for each strain. Error bars represent the SEM and significance between each mutant and wild-type was calculated by a 2-tailed Student's t-test indicated with P-value < 0.05 (*), <0.01 (**), or <0.001 (***). c) Anti-Hrp1 Western blot of cell extracts as in Panel B except from haploid strains containing the indicated viable HRP1 alleles. d, e) RT-qPCR of total RNA from the indicated haploid strains to produce d) the HRP1 amplicon indicated in panel a normalized to a CDC19 amplicon or e) the SNR82 5′ and 3′ amplicons indicated in Fig. 2a. Error bars represent the SEM for 3 biological replicates for ΔTLM and ΔNES and 2 biological replicates for ΔN/S and ΔQ. Significance between each mutant and wild-type was determined as in panel b.

Fig. 4.

Viable terminator readthrough substitutions in RRMs 1 and 2 are mostly in the RNA-binding surface. a) Individual substitutions in Hrp1 RRMs 1 and 2 obtained in a selection for viable mutations causing readthrough of the SNR82 terminator. Secondary structure elements are labeled. b) Viable readthrough mutations shown in magenta on structure of Hrp1 RRM1 (green), linker helix (gray), and RRM2 (blue) bound to (UA)₄ RNA (gold) (PDB: 2cjk, Pérez-Cañadillas 2006). Panels i–v show interactions that may be disrupted by substitutions from the selection (labeled in bold). Other relevant residues are labeled in gray. Nucleotides are numbered from 5′ to 3′. (*) indicates substitutions from alleles with multiple mutations (see Table 1). i) Stacking interaction between A4 and W168 and a salt bridge between RRM1 and RRM2. ii) Apparent “three-bridge cluster” around M191 involving F162, F202, and F204 that interacts with U5, A6, and U7. iii) Potential hydrogen bonds between A6 and R232 backbone carbonyl, and R236 and RNA backbone. iv) “Three-bridge cluster” in RRM2 around M275 involving F246, F286, and F288 that appears to interact with A2 and U3. v) Substitutions in residues participating in or near the potential salt bridge between K231 (RRM1) and D271 (RRM2).

Fig. 5.

Potentially lethal mutations in Hrp1 are rescued by autoregulation. a) Start codon substitutions and the N418 frameshift mutation mapped onto Hrp1. Out of frame ATGs before the first in frame ATG downstream of the start codon are indicated with red dots. The predicted effect of the frameshift mutation is shown with red indicating the new reading frame and (*) indicating a premature stop codon. b) Anti-Hrp1 Western blot of cell extracts from haploid strains containing the indicated HRP1 alleles on low copy plasmids. N418 fs (1) and (2) indicate selection alleles hrp1-125 and 127, respectively (see Table 1). Coomassie blue stained gel shown below was used to normalize for total protein loaded. The average, relative protein levels normalized to HRP1 (lane 2) from 2 biological replicates (see Supplementary Fig. 11b) are shown in black for FL Hrp1 and gray for truncated Hrp1 in the graph below the gel. Error bars represent the SEM and significance between each mutant and wild-type was calculated by a 2-tailed Student's t-test indicated with P-value < 0.05 (*), <0.01 (**), or <0.001 (***). c) RT-qPCR of total RNA from the indicated haploid strains to produce the HRP1 amplicon indicated in panel A. Values were normalized to a CDC19 amplicon. Error bars represent the SEM for 3 biological replicates.

Fig. 6.

Dominant readthrough mutations are largely in RRM1. a) Substitutions in the Hrp1 RRMs from the SNR82 dominant readthrough selection. Recessive lethal substitutions are shown in red above and viable substitutions in black below. (*) indicate that allele contained multiple substitutions and lines connect substitutions in the same allele when present in the RRMs (see Table 2). b) Substitutions mapped onto Hrp1 RRM structure with RRM1 (green), linker helix (gray), and RRM2 (blue) bound to (UA)₄ RNA (gold) (PDB: 2cjk, Pérez-Cañadillas 2006). Panels i–iv show interactions that may be disrupted by substitutions from the dominant selection shown in black (viable) and red (lethal). Other relevant residues are labeled in gray. Nucleotides are numbered from 5′ to 3′. (*) indicates substitutions from alleles with multiple mutations. i) A4 binding pocket and a salt bridge between RRM1 and RRM2. ii) Apparent “three-bridge cluster” around M191 involving F162, F202, and F204 that interacts with multiple bases in the RNA. iii) Substitutions in residues participating in or near the potential salt bridge between K231 (RRM1) and D271 (RRM2). iv) Recognition of A6 by R232 and A233. c) Serial dilutions of haploid strains containing the indicated alleles of HRP1 and the SNR82  ACT1-CUP1 reporter were spotted on medium containing the indicated concentration of copper sulfate. Biological replicate shown in Supplementary Fig. 12b.

Fig. 7.

Model for Hrp1 function in RNAP II transcription. Hrp1 is recruited to the RNAP II EC (gray) near the transcription start site, potentially through interaction with RRM2 (blue). This interaction may promote RNAP II elongation indicated by (>>>) and RRM1 (green) scanning the nascent RNA. Loss of Hrp1 from the EC by recognition of termination signals in the RNA by both RRMs could promote NNS termination. The Hrp1 RRMs could have a similar function in CPA termination at protein-coding genes.