Unproductive alternative splicing and nonsense mRNAs: a widespread phenomenon among plant circadian clock genes

Abstract

Background: Recent mapping of eukaryotic transcriptomes and spliceomes using massively parallel RNA sequencing (RNA-seq) has revealed that the extent of alternative splicing has been considerably underestimated. Evidence also suggests that many pre-mRNAs undergo unproductive alternative splicing resulting in incorporation of in-frame premature termination codons (PTCs). The destinies and potential functions of the PTC-harboring mRNAs remain poorly understood. Unproductive alternative splicing in circadian clock genes presents a special case study because the daily oscillations of protein expression levels require rapid and steep adjustments in mRNA levels.

Results: We conducted a systematic survey of alternative splicing of plant circadian clock genes using RNA-seq and found that many Arabidopsis thaliana circadian clock-associated genes are alternatively spliced. Results were confirmed using reverse transcription polymerase chain reaction (RT-PCR), quantitative RT-PCR (qRT-PCR), and/or Sanger sequencing. Intron retention events were frequently observed in mRNAs of the CCA1/LHY-like subfamily of MYB transcription factors. In contrast, the REVEILLE2 (RVE2) transcript was alternatively spliced via inclusion of a "poison cassette exon" (PCE). The PCE type events introducing in-frame PTCs are conserved in some mammalian and plant serine/arginine-rich splicing factors. For some circadian genes such as CCA1 the ratio of the productive isoform (i.e., a representative splice variant encoding the full-length protein) to its PTC counterpart shifted sharply under specific environmental stress conditions.

Conclusions: Our results demonstrate that unproductive alternative splicing is a widespread phenomenon among plant circadian clock genes that frequently generates mRNA isoforms harboring in-frame PTCs. Because LHY and CCA1 are core components of the plant central circadian oscillator, the conservation of alternatively spliced variants between CCA1 and LHY and for CCA1 across phyla [2] indicates a potential role of nonsense transcripts in regulation of circadian rhythms. Most of the alternatively spliced isoforms harbor in-frame PTCs that arise from full or partial intron retention events. However, a PTC in the RVE2 transcript is introduced through a PCE event. The conservation of AS events and modulation of the relative abundance of nonsense isoforms by environmental and diurnal conditions suggests possible regulatory roles for these alternatively spliced transcripts in circadian clock function. The temperature-dependent expression of the PTC transcripts among members of CCA1/LHY subfamily indicates that alternative splicing may be involved in regulation of the clock temperature compensation mechanism.

Figure 1

Examples of diurnal expression patterns of someArabidopsiscircadian genes surveyed for alternative splicing events. The time points when tissues were collected for RNA isolation are indicated by triangles. Gene expression patterns under different environmental conditions were obtained using the DIURNAL portal and microarray database. Detailed descriptions of diurnal conditions and experimental set up are available at http://diurnal.cgrb.oregonstate.edu/.

Figure 2

Examples of the discovery and characterization of alternative splicing events in circadian genes and identification and validation of the I4R events inLHYandCCA1. (A) Potential alternative splicing events were suggested by the distribution of RNA-seq reads density along gene features. The screen shots of the CCA1 and LHY loci are from the Arabidopsis RNA-seq GBrowse. Histograms represent the density of coverage of gene features by Illumina reads (GBrowse tracks of HTS reads) under different abiotic stress conditions (e.g., salt, drought, and cold treatments). The tracks "WT control" and "upf1-1" represent read density in the wild-type control and in the upf1-1 NMD mutant, respectively. The fourth introns in both the CCA1 and LHY transcripts display an increase in read coverage (bracketed) under specific treatments suggesting that these introns may be retained more often under certain conditions. The I9R event in LHY is also represented by an increase in read coverage on the 3' UTR distal portion (bracketed) in the upf1-1 NMD mutant. The numbers of LHY introns 4 and 9 correspond to the TAIR10 gene models AT1G01060.2 and AT1G01060.3, respectively. The CCA1 intron 4 corresponds to AT2G46830.1. All three IR events were confirmed by RT-PCR and qRT-PCR (Figures 2, 3 and 5, Table 1). The scale of the read depth histogram is limited to a maximum of 50 reads. (B) Primer design strategy developed for validation of the I4R events in CCA1 and LHY transcripts. Dark boxes depict protein-coding exons. The exons in the 5' and 3' UTRs are indicated by light boxes. The oligonucleotide primers used for the amplification of the constitutive and alternative splicing events are indicated by green and red arrows, respectively. The I4R events are depicted by dashed boxes. A small nested intron 4a in CCA1 is indicated by a dashed line. Normal and premature termination codons are shown by top and bottom star symbols, respectively. The expected sizes of RT-PCR products are given in base pairs (bp). The gene models are not drawn to scale. The inset in (B) shows the RT-PCR confirmation of the LHY and CCA1 transcripts with spliced and retained intron 4. Expected RT-PCR product sizes for the LHY transcript when intron 4 is spliced correctly is 264 bp; the size of the I4R isoform (if intron 4 is retained and introns 2 and 3 spliced out) is 224 bp. For CCA1 the predicted product sizes are 226 bp (introns 4/4a and 5 spliced out) and 225 bp for the I4R isoform (introns 4a and 5 are spliced out but intron 4 is not). RT-PCR products were separated on a 2 % agarose gels and stained with ethidium bromide. ‘M’ - molecular size markers (markers correspond from bottom to top to 100, 200, and 300 bp).

Figure 3

Oscillating profiles of IR events inLHYandLCL1are regulated by photoperiodic conditions with expression peaks at dark/light transitions under constant temperature (20 °C).LHY-I9S and LHY-I9R transcripts correspond to those with spliced and retained intron 9 events, respectively. LCL1-I7S and LCL1-I7R are transcripts with spliced and retained intron 7 events, respectively. The sampling and diurnal conditions are described in Methods. The normalized fold change in expression was calculated using a –ΔΔCt method and the CFX Manager software as described in Methods.

Figure 4

Alternative splicing ofRVE2pre-mRNA introduces an in-frame nonsense codon via a PCE event. (A) The schematic representation of the AS event in intron 1 introducing a PCE in the RVE2 transcript. The 5' UTR is shown by a light box, protein coding exons and the PCE are indicated by dark and hatched boxes, respectively. Binding sites for primers F1-R1 and F2-R2 used in RT-PCR are shown by arrows. Normal and premature termination codons are shown by the top and bottom stars. Dashed lines indicate primer portions spanning splice junctions. The gene model is not drawn to scale. (B) The alignment of the 5’ portion of RVE2 cDNA (top) and Sanger sequences of the PCR products (bottom). The PCE sequence is highlighted in grey. The initiation codon is boxed. (C) Quantification of alternatively spliced RVE2 mRNA harboring PCE event using qRT-PCR. The diurnal conditions and the sampling scheme are described in detail in Methods. Arrows indicate the sampling time points (ZT, hours). The relative transcript quantities were calculated using the –ΔΔCt method and CFX Manager software (BioRad). GOG mRNA was used as an internal reference. Vertical bars represent the standard error of the mean.

Figure 5

Regulation of the expression levels of alternatively splicedCCA1isoforms by environmental stress. (A) Semi-quantitative RT-PCR analysis of the I4R event in CCA1 transcripts under different abiotic stress treatments. The relative abundance of the PTC-harboring CCA1 I4R isoform changes compared to the full-length spliced variant (I4S) after cold stress treatment. Two-week-old seedlings were treated as previously described [2]. Ctrl - untreated seedlings, Hi light - high intensity light, Heat - heat stress (42 °C), Cold - cold stress (4 °C), Salt - high salinity (0.5 M sodium chloride), Desiccation - dehydration (polyethylene glycol treatment). eF1α mRNA was used to demonstrate an equal PCR amplification of cDNAs. PCR was carried out for 16 cycles for all the reactions. PCR products were separated in 2 % agarose gels and stained by ethidium bromide. (B) qRT-PCR analysis of cold stress-induced changes in relative abundance of the CCA1 splice variants. The normalized expression of the CCA1 transcripts with spliced intron 4 (CCA1 I4S) increased after 12 and 168 hours of the cold treatment (4 °C). In contrast, the normalized fold expression of the I4R transcripts sharply decreased following the treatment at 4 °C. GOG was used as a reference housekeeping transcript. The sampling and conditions of the time course are described in detail in the Methods. The normalized fold change of expression of the target transcripts was calculated using a –ΔΔCt method and CFX Manager software. Vertical bars denote the standard error of the mean. Both RT-PCR and qRT-PCR were performed using splicing event-specific oligonucleotide primers (see Additional file 2). ZT - Zeitgeber time (hours).