Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2009 Jun;5(6):e1000525.
doi: 10.1371/journal.pgen.1000525. Epub 2009 Jun 19.

Genome-wide Identification of Alternative Splice Forms Down-Regulated by Nonsense-Mediated mRNA Decay in Drosophila

Free PMC article

Genome-wide Identification of Alternative Splice Forms Down-Regulated by Nonsense-Mediated mRNA Decay in Drosophila

Kasper Daniel Hansen et al. PLoS Genet. .
Free PMC article


Alternative mRNA splicing adds a layer of regulation to the expression of thousands of genes in Drosophila melanogaster. Not all alternative splicing results in functional protein; it can also yield mRNA isoforms with premature stop codons that are degraded by the nonsense-mediated mRNA decay (NMD) pathway. This coupling of alternative splicing and NMD provides a mechanism for gene regulation that is highly conserved in mammals. NMD is also active in Drosophila, but its effect on the repertoire of alternative splice forms has been unknown, as has the mechanism by which it recognizes targets. Here, we have employed a custom splicing-sensitive microarray to globally measure the effect of alternative mRNA processing and NMD on Drosophila gene expression. We have developed a new algorithm to infer the expression change of each mRNA isoform of a gene based on the microarray measurements. This method is of general utility for interpreting splicing-sensitive microarrays and high-throughput sequence data. Using this approach, we have identified a high-confidence set of 45 genes where NMD has a differential effect on distinct alternative isoforms, including numerous RNA-binding and ribosomal proteins. Coupled alternative splicing and NMD decrease expression of these genes, which may in turn have a downstream effect on expression of other genes. The NMD-affected genes are enriched for roles in translation and mitosis, perhaps underlying the previously observed role of NMD factors in cell cycle progression. Our results have general implications for understanding the NMD mechanism in fly. Most notably, we found that the NMD-target mRNAs had significantly longer 3' untranslated regions (UTRs) than the nontarget isoforms of the same genes, supporting a role for 3' UTR length in the recognition of NMD targets in fly.

Conflict of interest statement

The authors have declared that no competing interests exist.


Figure 1
Figure 1. Isoform deconvolution.
(A) Probe placement and gene structure for 3 NMD–affected genes: glorund, RpL10Ab, and squid. Gene structures are shown with exons as boxes and introns represented by peaked lines. Dark blue regions indicate the coding region and grey boxes show the untranslated regions (UTRs). Each probe is represented by a vertical colored line, and its complementary site on an isoform is shown by a half circle (exon probe) or full circle (splice junction probe). The different colors indicate the combination of isoforms each probe targets. The NMD–affected isoform of each gene is indicated. The coding sequences (CDSs) of CG6946-RC, CG7283-RB, and CG16901-RD were identified as described in the text. (B) Normalized log2 fold-changes for the probes in the upf1 experiment, grouped by which isoforms they target with colors corresponding to panel A. Each colored circle is the measurement of one probe on one array. The black circle is the group-wise mean of the fold-changes. (C) Deconvolved fold-change for the individual isoforms; “possibly absent” isoforms are not plotted (see text). (D) Estimated relative abundance of each present isoform in both the control and the NMD inhibited samples. We estimate that the NMD–target isoform of squid was a negligible fraction of total squid mRNA.
Figure 2
Figure 2. Experimental validation of NMD status.
RNA samples isolated from control cells or cells depleted of UPF1 or UPF2 were analyzed by RT-PCR using primers flanking the alternative region of each gene. Bands corresponding to each isoform are labeled to the right and the exon/intron structure of each isoform is depicted along with the log2 fold-change estimated for that isoform from the array data when NMD is inhibited. “?” indicates a band of unknown origin and “*” shows the expected location of a band that is not observed. The isoform corresponding to the missing band is shown beneath the gel. “x” indicates an isoform called “possibly absent” on the array. Full gels are shown in Figure S10. (A) The action of NMD was confirmed on four genes called NMD–affected based on the array. For each gene, the ratio of NMD–target∶nontarget mRNA increased upon NMD inhibition. (B) Isoform classifications were confirmed for three genes called unaffected based on the array. (C) RT-PCR of CG8046 shows that NMD affects one isoform that was called “possibly absent” based on the array.
Figure 3
Figure 3. CDS reannotation.
For each gene, the top diagram shows the gene structure including the CDS (blue rectangles), stop codon (red octagon), and UTRs (gray rectangles) of the isoform unaffected by NMD. The bottom diagrams show the gene structure of the NMD–affected isoform; (1) depicts the annotated CDS from FlyBase and (2) depicts the re-annotated CDS. (A) Re-annotation of glorund shows the existence of a likely CDS that shares its start codon with the unaffected isoform and has an early stop codon. (B) FlyBase annotates the same ORF in the unaffected and NMD–affected isoforms of RpL8. Re-annotation shows that alternative splicing introduces a uORF. (C) Our method does not find an alternate CDS in the NMD–affected isoform of VhaSFD, which differs from the unaffected isoform by skipping an in-frame cassette exon.
Figure 4
Figure 4. Features correlated with NMD status.
(A) Boxplots of the 3′ UTR length comparing the strict set of upf1 NMD–target mRNAs to the set of upf1 NMD nontarget mRNAs from the same genes. The lower box indicates the second quartile of values and the upper box the third quartile, the belt shows the median, and the whiskers indicate the largest value within 1.5× the size of the box. (B) The 3′ UTR length, (C) CDS length, (D) number of introns in the 3′ UTR, and (E) distance between the stop codon and the last intron, compared per gene for each gene in the strict set of upf1-affected genes. These pairwise comparisons show more significant differences than the comparison in panel (A).
Figure 5
Figure 5. Isoform classification.
Flowchart for isoform classification. Each isoform is classified separately. formula image is the non-log fold-change associated with isoform formula image and formula image is the relative proportion of the isoform in the control sample.

Similar articles

See all similar articles

Cited by 49 articles

See all "Cited by" articles


    1. Culbertson MR, Leeds PF. Looking at mRNA decay pathways through the window of molecular evolution. Curr Opin Genet Dev. 2003;13:207–214. - PubMed
    1. Cali BM, Anderson P. mRNA surveillance mitigates genetic dominance in Caenorhabditis elegans. Mol Gen Genet. 1998;260:176–184. - PubMed
    1. Mendell JT, Sharifi NA, Meyers JL, Martinez-Murillo F, Dietz HC. Nonsense surveillance regulates expression of diverse classes of mammalian transcripts and mutes genomic noise. Nat Genet. 2004;36:1073–1078. - PubMed
    1. Pan Q, Saltzman AL, Kim YK, Misquitta C, Shai O, et al. Quantitative microarray profiling provides evidence against widespread coupling of alternative splicing with nonsense-mediated mRNA decay to control gene expression. Genes Dev. 2006;20:153–158. - PMC - PubMed
    1. Isken O, Maquat LE. Quality control of eukaryotic mRNA: safeguarding cells from abnormal mRNA function. Genes Dev. 2007;21:1833–1856. - PubMed

Publication types

LinkOut - more resources