Aberrant mRNA transcripts and the nonsense-mediated decay proteins UPF2 and UPF3 are enriched in the Arabidopsis nucleolus

Abstract

The eukaryotic nucleolus is multifunctional and involved in the metabolism and assembly of many different RNAs and ribonucleoprotein particles as well as in cellular functions, such as cell division and transcriptional silencing in plants. We previously showed that Arabidopsis thaliana exon junction complex proteins associate with the nucleolus, suggesting a role for the nucleolus in mRNA production. Here, we report that the plant nucleolus contains mRNAs, including fully spliced, aberrantly spliced, and single exon gene transcripts. Aberrant mRNAs are much more abundant in nucleolar fractions, while fully spliced products are more abundant in nucleoplasmic fractions. The majority of the aberrant transcripts contain premature termination codons and have characteristics of nonsense-mediated decay (NMD) substrates. A direct link between NMD and the nucleolus is shown by increased levels of the same aberrant transcripts in both the nucleolus and in Up-frameshift (upf) mutants impaired in NMD. In addition, the NMD factors UPF3 and UPF2 localize to the nucleolus, suggesting that the Arabidopsis nucleolus is therefore involved in identifying aberrant mRNAs and NMD.

Figure 1.

Distribution and Nature of Transcript Types in cDNA Libraries. The distribution of fully spliced, aberrantly spliced, and single exon transcripts in whole-cell, nuclear, and nucleolar cDNA libraries expressed as percentage of total full-length cDNAs from the same library (226, 139, and 132 total cDNAs from the nucleolar, nuclear, and whole-cell libraries, respectively). Only full-length, nonredundant cDNAs were scored.

Figure 2.

Schematic Diagrams of Gene Structures of Representative Aberrantly Spliced Transcripts. The examples demonstrate different splicing phenotypes: unspliced introns (A), use of cryptic 5′ splice sites (B), use of cryptic 3′ splice sites (C), use of both cryptic 5′ and 3′ splice sites (D), aberrant splicing in a single intron gene (no downstream exon-exon junction/EJC) (E), aberrant splicing in the last intron/exon (no downstream EJC) (F), aberrant splicing in a 5′ UTR intron (this example produces a short open reading frame in the UTR) (G), aberrant splicing in a 3′ UTR intron (H), and alternative splicing leaving an intronic fragment in frame (no PTC) (I). The cloned transcript is shown below the exon-intron structure of the gene. The splicing events result in the removal of parts of exon sequences, inclusion of whole or partial intron sequences, or exon skipping. Open boxes, exons; thick black lines, UTRs; solid lines, introns; dashed vertical lines, site of cryptic splice site in exon. Aberrant splicing events are indicated by diagonal lines; black hexagons indicate that PTCs are generated in the transcript.

Figure 3.

Distribution of Splicing and NMD Transcript Phenotypes. (A) Distribution of different splicing phenotypes among 160 aberrantly spliced transcripts from the three cDNA libraries. (B) Distribution of transcripts by potential consequence of the aberrant splicing event(s). NMD phenotypes were classified as potential NMD substrates with or without a downstream EJC, or non-NMD substrates where no PTC is generated or where the PTC remains close to the 3′ UTR due to a frame change in the last exon. ss, splice site.

Figure 4.

RT-PCR of mRNA Transcripts in Nuclear, Nucleoplasmic, and Nucleolar RNA. Products of RT-PCR (RT+) with gene-specific primers from nuclear (N), nucleoplasmic (Np), and nucleolar (No) RNA for different Arabidopsis genes showing enrichment of higher molecular weight products and a reduction in fully spliced product in the nucleolar fraction. M, size markers; RT−, PCR controls without reverse transcriptase. The fractionation was assessed by RT-PCR of U1 snRNA (U1) and U2 snRNA (U2) for the nucleoplasmic and U3 snoRNA (U3) for nucleolar fractions.

Figure 5.

Relative Abundance of Fully Spliced and Aberrantly Spliced RT-PCR Products for Different Genes in Nuclear, Nucleoplasmic, and Nucleolar RNA Fractions. Histograms showing the mean and standard errors of the relative abundances (expressed as a percentage) of different higher molecular weight, aberrantly spliced products (AS1-5) and fully spliced (FS) products for the different fractions (white, nuclear; gray, nucleoplasmic; black, nucleolar). The number of different fractions used for RT-PCR is given in the inset for each histogram (N, nuclear; Np, nucleoplasmic; No, nucleolar). The combined means of the aberrantly spliced bands and mean of the fully spliced band were compared between the nuclear and nucleolar preparations by analysis of variance. Probability values for AS>FS in the nucleolus and FS>AS in the nucleoplasm are presented (sig, significant, P < 0.05; hs, highly significant, P < 0.01; ns, not significant).

Figure 6.

RT-PCR of mRNA Transcripts in Nuclear, Nucleoplasmic, and Nucleolar RNA and upf Mutants. (A) and (B) RT-PCR with gene-specific primers for At2g21660 (A) and At3g61860 (B) from RNA from nucleoplasmic (Np) and nucleolar (No) fractions of cell culture cells (lanes 1 and 2) and seedlings of wild-type plants (lane 3), upf1-5 (lane 4), and upf3-1 (lane 5). Transcript structures are shown below the exon/intron structure of the genes based on The Arabidopsis Information Resource (TAIR). (C) The fractionation was assessed by RT-PCR of U1 snRNA (lanes 1 and 2) and U2 snRNA (lanes 3 and 4) for the nucleoplasmic fraction and U3 snoRNA (lanes 5 and 6) for the nucleolar fraction. M, size markers; large arrowheads, RT-PCR products that are enriched in the nucleolus and in upf mutants; small arrowhead with question mark, unknown product; asterisk, fully spliced product; open boxes, exons; thick black lines, UTRs; solid lines, introns; diagonal lines, splicing events; small arrowheads, primer positions. RT-PCR controls without reverse transcriptase are not shown.

Figure 7.

Subcellular Localization of NMD Proteins, UPF1, UPF2, and UPF3. Confocal images of single cells coexpressing GFP-UPF (green signal) and fibrillarin-mRFP (red signal) protein fusions and merged images. GFP-UPF3 (A), GFP-UPF2 (B), and GFP-UPF1 (C). The right-hand panel shows transmission light images. Nucleoli are indicated by arrowheads. Bars = 10 μm.