Control of pre-mRNA accumulation by the essential yeast protein Nrd1 requires high-affinity transcript binding and a domain implicated in RNA polymerase II association

Abstract

Nrd1 is an essential yeast protein of unknown function that has an RNA recognition motif (RRM) in its carboxyl half and a putative RNA polymerase II-binding domain, the CTD-binding motif, at its amino terminus. Nrd1 mediates a severe reduction in pre-mRNA production from a reporter gene bearing an exogenous sequence element in its intron. The effect of the inserted element is highly sequence-specific and is accompanied by the appearance of 3'-truncated transcripts. We have proposed that Nrd1 binds to the exogenous sequence element in the nascent pre-mRNA during transcription, aided by the CTD-binding motif, and directs 3'-end formation a short distance downstream. Here we show that highly purified Nrd1 carboxyl half binds tightly to the RNA element in vitro with sequence specificity that correlates with the efficiency of cis-element-directed down-regulation in vivo. A large deletion in the CTD-binding motif blocks down-regulation but does not affect the essential function of Nrd1. Furthermore, a nonsense mutant allele that produces truncated Nrd1 protein lacking the RRM has a dominant-negative effect on down-regulation but not on cell growth. Viability of this and several other nonsense alleles of Nrd1 appears to require translational readthrough, which in one case is extremely efficient. Thus the CTD-binding motif of Nrd1 is important for pre-mRNA down-regulation but is not required for the essential function of Nrd1. In contrast, the RNA-binding activity of Nrd1 appears to be required both for down-regulation and for its essential function.

Figure 1

RNAs and recombinant Nrd1 protein used for RNA-binding assays. (A) Schematic diagram of U6R* element that confers Nrd1-dependent RNA down-regulation (1). The arrow represents antisense U6, the shaded box represents the 5′-flanking element, and the black box a 14-nt region of antisense U6 defined by point-mutant suppressors of down-regulation. The sequence of the 43-nt wild-type transcript is shown with three guanosine residues contributed by the T7 promoter in lowercase, and the location and identity of the point mutants analyzed here are shown below. (B) Schematic diagram of wild-type Nrd1 and recombinant Nrd1_307–560 protein. Domains indicated are: CTD-bd, domain implicated in binding to RNA Pol II CTD; RE-RS, arginine-glutamate and arginine-serine dipeptide-rich region; RRM, RNA recognition motif; P+Q, proline- and glutamine-rich region. Sequences contributed to Nrd1_307–560 by the pET21b expression vector are indicated by the lines at the amino (N) and carboxyl (C) termini.

Figure 2

Gel-mobility-shift assays for RNA binding by Nrd1_307–560. (A) Representative gels showing retardation of wild-type, G92′U, A89′U, U95′G, and U83′A RNAs by binding to Nrd1_307–560. Protein concentrations assayed were 0, 5, 10, 20, 40, 80, and 160 nM, increasing from left to right for each RNA. Positions of the well, RNA–protein complex, and free RNA are identified by marks labeled w, c, and f, respectively, and the asterisk marks a nonspecific complex observed at the highest protein concentrations. (B) Competition assay for RNA binding to Nrd1, using unlabeled wild-type or A90′G mutant RNA as competitors for binding to labeled 43-nt wild-type RNA. Competitor RNA concentrations were 0, 3, 15, 75, 150, and 300 nM, increasing from left to right. The leftmost lane in each set (− Nrd1) contained no Nrd1 protein. (C) Formation of two discrete complexes on 68-nt RNAs containing the 5′-flanking portion of the U6R* element. Protein concentrations were 0, 10, 20, 40, and 80 nM. The mobilities of complexes interpreted to have 1 or 2 sites on the RNA occupied by Nrd1 are indicated by c1 and c2, respectively.

Figure 3

Correlation between the affinity of Nrd1_307–560 for 43-nt RNAs in vitro and down-regulation of ACT-CUP mRNA in vivo. (A) Steady-state levels of mRNA produced from ACT-CUP fusion genes with the U6R* insert or selected point mutant inserts. The mRNA level from the wild-type ACT-CUP fusion gene (no U6R* insert) is defined as 1.0. RNA levels are averages from two primer extension experiments, normalized to internal controls provided by extension of U5 and U6 RNAs. (B) Binding curves for 43-nt wild-type and selected mutant RNAs. Data are from representative single experiments.

Figure 4

An amino-terminal domain of Nrd1 is dispensable for viability but required for U6R*-directed RNA down-regulation. (A) Schematic of Nrd1 protein (see Fig. 1B), showing location of nonsense codons in the nrd1-1, nrd1-2, and nrd1-3 alleles. The location of the Δ39–169 deletion is also indicated. (B) Immunoblot showing expression of Nrd1Δ39–169 and wild-type Nrd1 proteins from centromere plasmids in the nrd1Δ39-575 strain, EJS101–9d. Numbers to the left indicate the sizes, in kDa, of molecular mass markers. (C) Primer extension analysis of ACT-CUP mRNA, showing suppression of U6R*-directed down-regulation by the nrd1Δ39-169 allele in EJS101–9d. Products from reverse transcription of ACT-CUP mRNA and U5 RNA are indicated. Lane 1, expression of mRNA from the U6R*-containing ACT-CUP fusion gene in the presence of wild-type NRD1 allele on a CEN plasmid; lane 2, expression of mRNA from the U6R*-containing fusion gene in the presence of nrd1Δ39-169; lane 3, expression of mRNA from the wild-type ACT-CUP fusion gene in the presence of wild-type NRD1. (D) Copper plate assay for ACT-CUP expression, showing dominant suppression of down-regulation by overexpressed nrd1-1 and nrd1-2 in the wild-type NRD1 strain, 46α. The plate shown contained 0.5 mM CuSO₄.

Figure 5

Expression and function of nrd1 nonsense alleles. (A) Viability of nrd1 alleles provided on CEN and 2μ plasmids in the nrd1Δ strain, EJS101–9d. The plate shown contained 5-fluoroorotic acid to select for loss of the URA3-marked plasmid carrying wild-type NRD1 (pRS316NRD1). (B) Immunoblots showing accumulation of Nrd1 protein in strains expressing nrd1 nonsense alleles. (Upper) Lane 1, wild-type NRD1 strain, 46α; lane 2, nrd1-1 strain; lane 3, nrd1-2 strain; and lane 4, nrd1-3 strain. The arrowhead indicates full-length Nrd1 protein, and the asterisk indicates the truncated product of nrd1-2. (Lower) Lane 5, nrd1Δ strain EJS101–9d expressing wild-type NRD1 from a CEN plasmid; lane 6, nrd1Δ strain expressing nrd1-3 from a CEN plasmid. (C) Primer extension analysis of nrd1-3 mRNA. The diagram shows the RNA sequence surrounding the nrd1-3 stop codon and indicates the expected reverse transcription products from wild-type and nrd1-3 mRNAs. Extension of the Nrd1-Bam169 primer in the presence of ddATP (lanes 1 and 2) shows the presence of a U residue in the mRNA at the position corresponding to the CAA to TAA mutation in the nrd1-3 gene. Extension in the presence of ddCTP (lanes 3 and 4) yields the same product from both the wild-type and nrd1-3 mRNAs, allowing quantitation of the mRNA levels.